Section V. Drugs Affecting Renal and Cardiovascular Function Chapter 29.

Diuretics
Overview Diuretics increase the rate of urine flow and sodium excretion and are used to adjust the volume and/or composition of body fluids in a variety of clinical situations, including hypertension, heart failure, renal failure, nephrotic syndrome, and cirrhosis. The objective of this chapter is to provide the reader with unifying concepts as to how the kidney operates and how diuretics modify renal function. The chapter begins with a description of renal anatomy and physiology, as this information is prerequisite to a discussion of diuretic pharmacology. Categories of diuretics are introduced and then described with regard to chemistry, mechanism of action, site of action, effects on urinary composition, and effects on renal hemodynamics. Near the end of the chapter, diuretic pharmacology is integrated with a discussion of mechanisms of edema formation and the role of diuretics in clinical medicine. Therapeutic applications of diuretics are expanded upon in Chapters 33: Antihypertensive Agents and the Drug Therapy of Hypertension (hypertension) and 34: Pharmacological Treatment of Heart Failure (heart failure).

Renal Anatomy and Physiology Renal Anatomy The main renal artery branches close to the renal hilum into segmental arteries, which, in turn, subdivide to form interlobar arteries that pierce the renal parenchyma. The interlobar arteries curve at the border of the renal medulla and cortex to form arc-like vessels known as arcuate arteries. Arcuate arteries give rise to perpendicular branches, called interlobular arteries, which enter the renal cortex and supply blood to the afferent arterioles. A single afferent arteriole penetrates the glomerulus of each nephron and branches extensively to form the glomerular capillary nexus. These branches coalesce to form the efferent arteriole. Efferent arterioles of superficial glomeruli ascend toward the kidney surface before splitting into peritubular capillaries that service the tubular elements of the renal cortex. Efferent arterioles of juxtamedullary glomeruli descend into the medulla and divide to form the descending vasa recta, which supply blood to the capillaries of the medulla. Blood returning from the medulla via the ascending vasa recta drains directly into the arcuate veins, and blood from the peritubular capillaries of the cortex enters the interlobular veins, which, in turn, connect with the arcuate veins. Arcuate veins drain into interlobar veins, which in turn drain into segmental veins, and blood leaves the kidney via the main renal vein. The basic urine-forming unit of the kidney is the nephron, which consists of a filtering apparatus, the glomerulus, connected to a long tubular portion that reabsorbs and conditions the glomerular ultrafiltrate. Each human kidney is composed of approximately 1 million nephrons. The nomenclature for segments of the tubular portion of the nephron has become increasingly complex as renal physiologists have subdivided the nephron into shorter and shorter named segments. These subdivisions initially were based on the axial location of the segments but increasingly have been based on the morphology of the epithelial cells lining the various nephron segments. Figure 291 illustrates the currently accepted subdivision of the nephron into 14 subsegments. Commonly encountered names that refer to these subsegments and to combinations of subsegments are

included. Figure 291. Anatomy and Nomenclature of the Nephron.

Glomerular Filtration In the glomerular capillaries, a portion of the plasma water is forced through a filter that has three basic components: the fenestrated capillary endothelial cells, a basement membrane lying just beneath the endothelial cells, and the filtration slit diaphragms formed by the epithelial cells that cover the basement membrane on its urinary space side. Solutes of small size flow with filtered water (solvent drag) into the urinary (Bowman's) space, whereas formed elements and macromolecules are retained by the filtration barrier. For each nephron unit, the rate of filtration (single-nephron glomerular filtration rate, SNGFR) is a function of the hydrostatic pressure in the glomerular capillaries (PGC), the hydrostatic pressure in Bowman's space (which can be equated with pressure in the proximal tubule, PT), the mean colloid osmotic pressure in the glomerular capillaries ( GC ), the colloid osmotic pressure in the proximal tubule ( T ), and the ultrafiltration coefficient (Kf ), according to the equation: SNGFR = Kf [(PGC (
GC T )]

(291)
T

If PGCPT is defined as the transcapillary hydraulic pressure difference ( P), and if (as it usually is since little protein is filtered), then: SNGFR = Kf ( P
GC )

is negligible

(292)

This latter equation succinctly expresses the three major determinants of SNGFR. However, each of these three determinants can be influenced by a number of other variables. K f is determined by the physicochemical properties of the filtering membrane and by the surface area available for filtration. P is determined primarily by the arterial blood pressure and by the proportion of the arterial pressure that is transmitted to the glomerular capillaries. This is governed by the relative resistances of preglomerular and postglomerular vessels. GC is determined by two variables, i.e., the concentration of protein in the arterial blood entering the glomerulus and the single-nephron blood flow (QA). QA influences GC because, as blood transverses the glomerular capillary bed, filtration concentrates proteins in the capillaries, causing GC to rise with distance along the glomerular bed. When QA is high, this effect is reduced; however, when QA is low, GC may increase to the point that GC = P and filtration ceases (a condition known as filtration equilibrium; seeDeen et al., 1972). Overview of Nephron Function Approximately 120 ml of ultrafiltrate is formed each minute, yet only 1 ml/min of urine is produced. Therefore, greater than 99% of the glomerular ultrafiltrate is reabsorbed at a staggering energy cost. The kidneys consume 7% of total-body oxygen intake despite the fact that the kidneys make up only 0.5% of body weight. The kidney is designed to filter large quantities of plasma, reabsorb those substances that the body must conserve, and leave behind and/or secrete substances that must be eliminated. The proximal tubule is contiguous with Bowman's capsule and takes a tortuous path until finally forming a straight portion that dives into the renal medulla. The proximal tubule has been subdivided into S1, S2, and S3 segments based on the morphology of the epithelial cells lining the tubule. Normally, approximately 65% of filtered Na+ is reabsorbed in the proximal tubule, and since this part of the tubule is highly permeable to water, reabsorption is essentially isotonic. Between the outer and inner strips of the outer medulla, the tubule abruptly changes morphology to

become the descending thin limb (DTL), which penetrates the inner medulla, makes a hairpin turn, and then forms the ascending thin limb (ATL). At the juncture between the inner and outer medulla, the tubule once again changes morphology and becomes the thick ascending limb, which is made up of three segments: a medullary portion (MTAL), a cortical portion (CTAL), and a postmacular segment. Together, the proximal straight tubule, DTL, ATL, MTAL, CTAL, and postmacular segment are known as the loop of Henle. The DTL is highly permeable to water, yet its permeability to NaCl and urea is low. In contrast, the ATL is permeable to NaCl and urea but is impermeable to water. The thick ascending limb actively reabsorbs NaCl but is impermeable to water and urea. Approximately 25% of filtered Na+ is reabsorbed in the loop of Henle, mostly in the thick ascending limb, which has a large reabsorptive capacity. The thick ascending limb passes between the afferent and efferent arterioles and makes contact with the afferent arteriole via a cluster of specialized columnar epithelial cells known as the macula densa. The macula densa is strategically located to sense concentrations of NaCl leaving the loop of Henle. If the concentration of NaCl is too high, the macula densa sends a chemical signal (perhaps adenosine) to the afferent arteriole of the same nephron, causing it to constrict. This in turn causes a reduction in PGC and QA and decreases SNGFR. This homeostatic mechanism, known as tubuloglomerular feedback (TGF), serves to protect the organism from salt and volume wasting. Besides causing a TGF response, the macula densa also regulates renin release from the adjacent juxtaglomerular cells in the wall of the afferent arteriole. Approximately 0.2 mm past the macula densa, the tubule changes morphology once again to become the distal convoluted tubule (DCT). The postmacular segment of the thick ascending limb and the distal convoluted tubule often are referred to as the early distal tubule. Like the thick ascending limb, the DCT actively transports NaCl and is impermeable to water. Since these characteristics impart the ability to produce a dilute urine, the thick ascending limb and the DCT are collectively called the diluting segment of the nephron, and the tubular fluid in the DCT is hypotonic regardless of hydration status. However, unlike the thick ascending limb, the DCT does not contribute to the countercurrent-induced hypertonicity of the medullary interstitium (see below). The collecting duct system (connecting tubule + initial collecting tubule + cortical collecting duct + outer and inner medullary collecting duct) is an area of fine control of ultrafiltrate composition and volume. It is here that final adjustments in electrolyte composition are made, a process modulated by the adrenal steroid, aldosterone. In addition, permeability of this part of the nephron to water is modulated by antidiuretic hormone (ADH; seeChapter 30: Vasopressin and Other Agents Affecting the Renal Conservation of Water). The more distal portions of the collecting duct pass through the renal medulla, where the interstitial fluid is markedly hypertonic. In the absence of ADH, the collecting duct system is impermeable to water, and a dilute urine is excreted. However, in the presence of ADH, the collecting duct system is permeable to water, so that water is reabsorbed. The movement of water out of the tubule is driven by the steep concentration gradient that exists between the tubular fluid and the medullary interstitium. The hypertonicity of the medullary interstitium plays a vital role in the ability of mammals and birds to concentrate urine and is therefore a key adaptation necessary for living in a terrestrial environment. This is accomplished via a combination of the unique topography of the loop of Henle and the specialized permeability features of the loop's subsegments. Although the precise mechanism giving rise to the medullary hypertonicity has remained elusive, the passive countercurrent multiplier hypothesis of Kokko and Rector (1972) is an intuitively attractive model that is qualitatively accurate (seeSands and Kokko, 1996). According to this hypothesis, the process

begins with active transport in the thick ascending limb, which concentrates NaCl in the interstitium of the outer medulla. Since this segment of the nephron is impermeable to water, active transport in the ascending limb dilutes the tubular fluid. As the dilute fluid passes into the collecting duct system, water is extracted if and only if ADH is present. Since the cortical and outer medullary collecting ducts have a low permeability to urea, urea is concentrated in the tubular fluid. The inner medullary collecting duct, however, is permeable to urea, so that urea diffuses into the inner medulla where it is trapped by countercurrent exchange in the vasa recta. Since the DTL is impermeable to salt and urea, the high urea concentration in the inner medulla extracts water from the DTL and concentrates NaCl in the tubular fluid of the DTL. As the tubular fluid enters the ATL, NaCl diffuses out of the salt-permeable ATL, thus contributing to the hypertonicity of the medullary interstitium. General Mechanism of Renal Epithelial Transport Figure 292 illustrates seven mechanisms by which solute crosses renal epithelial cell membranes. If bulk water flow occurs across a membrane, solute molecules will be transferred by convection across the membrane, a process known as solvent drag. Solutes with sufficient lipid solubility may also dissolve in the membrane and diffuse across the membrane down their electrochemical gradients (simple diffusion). Many solutes, however, have limited lipid solubility, and transport must rely on integral proteins embedded in the cell membrane. In some cases, the integral protein merely provides a conductive pathway (pore) through which the solute may diffuse passively (channel-mediated diffusion). In other cases, the solute may bind to the integral protein and, due to a conformational change in the protein, be transferred across the cell membrane down an electrochemical gradient (carrier-mediated or facilitated diffusion, also called uniport). However, this process will not result in net movement of solute against an electrochemical gradient. If solute must be moved "uphill" against an electrochemical gradient, then either primary active transport or secondary active transport is required. With primary active transport, ATP hydrolysis is coupled directly to conformational changes in the integral protein, thus providing the necessary free energy (ATP-mediated transport). Often, ATP-mediated transport is used to create an electrochemical gradient for a given solute, and the free energy of that solute gradient is then released to drive the "uphill" transport of other solutes. This process requires symport (cotransport of solute species in the same direction) or antiport (countertransport of solute species in opposite directions) and is known as secondary active transport. Figure 292. Seven Basic Mechanisms for Transmembrane Transport of Solutes. 1, convective flow in which dissolved solutes are "dragged" by bulk water flow; 2, simple diffusion of lipophilic solute across membrane; 3, diffusion of solute through pore; 4, transport of solute by carrier protein down electrochemical gradient; 5, transport of solute by carrier protein against electrochemical gradient with ATP hydrolysis providing driving force; 6 and 7, cotransport and countertransport, respectively, of solutes with one solute traveling "uphill" against an electrochemical gradient and the other solute traveling down an electrochemical gradient.

The kinds of transport achieved in a particular nephron segment depend mainly on which transporters are present and whether they are embedded in the luminal or basolateral membrane. A general model of renal tubular transport is shown in Figure 293 and can be summarized as follows: Figure 293. Generic Mechanism of Renal Epithelial Cell Transport (See Text for Details). S, symporter; A, antiporter; CH, ion channel; WP, water pore; U, uniporter; ATPase, Na+ ,K+ ATPase (sodium pump); X and Y, transported solutes; P, membrane-permeable (reabsorbable) solutes; I, membrane-impermeable (nonreabsorbable) solutes; PD, potential difference across indicated membrane or cell.

1. Na+,K+ATPase (sodium pump) in the basolateral membrane hydrolyzes ATP, which results in the transport of Na+ into the intercellular and interstitial spaces and the movement of K+ into the cell. Although other ATPases exist in selected renal epithelial cells and participate in the transport of specific solutes (e.g., Ca2+ATPase and H+ ATPase), the bulk of all transport in the kidney is due to the abundant supply of Na+,K+ ATPase in the basolateral membranes of the renal epithelial cells. 2. Na+ may diffuse across the luminal membrane via Na+ channels into the epithelial cell down the

3. 4.

5.

6.

7.

electrochemical gradient for Na+ that is established by the basolateral Na +,K+ATPases. In addition, the free energy available in the electrochemical gradient for Na+ is tapped by integral proteins in the luminal membrane, resulting in cotransport of various solutes against their electrochemical gradients by symporters (e.g., Na+glucose, Na+Pi, Na+ amino acid). This process results in movement of Na+ and cotransported solutes out of the tubular lumen into the cell. Also, antiporters (e.g., Na+ H+ ) countertransport Na+ out of and some solutes into the tubular lumen. Na+ exits the basolateral membrane into the intercellular and interstitial spaces via the Na+ pump or via symporters or antiporters in the basolateral membrane. The action of Na+-linked symporters in the luminal membrane causes the concentration of substrates for these symporters to rise in the epithelial cell. These electrochemical gradients then permit simple diffusion or mediated transport (symporters, antiporters, uniporters, and channels) of solutes into the intercellular and interstitial spaces. + Accumulation of Na and other solutes in the intercellular space creates a small osmotic pressure differential across the epithelial cell. In water-permeable epithelium, water moves into the intercellular spaces driven by the osmotic pressure differential. Water moves through aqueous pores in both the luminal and the basolateral cell membranes as well as through the tight junctions (paracellular pathway). Bulk water flow carries some solutes into the intercellular space by solvent drag. Movement of water into the intercellular space concentrates other solutes in the tubular fluid, resulting in an electrochemical gradient for these substances across the epithelium. Membranepermeable solutes then move down their electrochemical gradients into the intercellular space via both the transcellular (simple diffusion, symporters, antiporters, uniporters, and channels) and paracellular pathways. Membrane-impermeable solutes remain in the tubular lumen and are excreted in the urine with an obligatory amount of water. As water and solutes accumulate in the intercellular space, the hydrostatic pressure increases, thus providing a driving force for bulk water flow. Bulk water flow carries solute (solute convection) out of the intercellular space into the interstitial space and, finally, into the peritubular capillaries. The movement of fluid into the peritubular capillaries is governed by the same Starling forces that determine transcapillary fluid movement for any capillary bed.

Mechanism of Organic Acid and Organic Base Secretion The kidney is a major organ involved in the elimination of organic chemicals from the body. Organic molecules may enter the renal tubules by glomerular filtration of molecules not bound to plasma proteins or may be actively secreted directly into the tubules. The proximal tubule has a highly efficient transport system for organic acids and an equally efficient but separate transport system for organic bases. Current models for these secretory systems are illustrated in Figure 294. Both systems are powered by the sodium pump in the basolateral membrane, involve secondary and tertiary active transport, and utilize a poorly characterized facilitated-diffusion step. The antiporter that exchanges -ketoglutarate for organic acids has been cloned from several species, including human beings (Lu et al., 1999). The optimal substrate for transport by the organic acid secretory mechanism is a molecule with a negative or partial negative charge, separated by 6 to 7 from a second negative charge, and a hydrophobic region 8 to 10 long. However, much flexibility exists around this optimal motif, and the system transports a large variety of organic acids. The organic base secretory mechanism is even less discriminating and may involve a family of related transport mechanisms. This system(s) transports many drugs containing an amine nitrogen positively charged at physiological pH.

Figure 294. Mechanisms of Organic Acid (A) and Organic Base (B) Secretion in the Proximal Tubule. The numbers 1, 2, and 3 refer to primary, secondary, and tertiary active transport. A, organic acid (anion); C+ , organic base (cation); 2 KG , -ketoglutarate, but also other dicarboxylates.

Renal Handling of Specific Anions and Cations Reabsorption of Cl generally follows reabsorption of Na . In segments of the tubule with lowresistance tight junctions (i.e., "leaky" epithelium), such as the proximal tubule and thick ascending limb, Cl movement can occur paracellularly. With regard to transcellular Cl flux, Cl crosses the
+

luminal membrane via antiport with formate and oxalate (proximal tubule), symport with Na+/K+ (thick ascending limb), symport with Na+ (DCT), and antiport with HCO3 (collecting duct system). Cl crosses the basolateral membrane via symport with K+ (proximal tubule and thick ascending limb), antiport with Na+/HCO3 (proximal tubule), and Cl channels (thick ascending limb, DCT, collecting duct system). Eighty to ninety percent of filtered K+ is reabsorbed in the proximal tubule (diffusion and solvent drag) and thick ascending limb (diffusion), largely via the paracellular pathway. In contrast, the DCT and collecting duct system secrete variable amounts of K +via a conductive (channel-mediated) pathway. Modulation of the rate of K+ secretion in the collecting duct system, particularly by aldosterone, allows urinary excretion of K+ to be matched with dietary intake. The transepithelial potential difference (VT ), lumen-positive in the thick ascending limb and lumen-negative in the collecting duct system, provides an important driving force for K+ reabsorption and secretion, respectively. Most of the filtered Ca2+ (approximately 70%) is reabsorbed by the proximal tubule by passive diffusion, probably via a paracellular route. Another 25% of filtered Ca2+ is reabsorbed by the thick ascending limb, mostly via a paracellular route driven by the lumen-positive VT, although a component of active Ca2+ reabsorption also may exist. The remaining Ca2+ is reabsorbed in the distal convoluted tubule and the connecting tubule via a transcellular pathway that is modulated by parathyroid hormone (PTH; seeChapter 62: Agents Affecting Calcification and Bone Turnover: Calcium, Phosphate, Parathyroid Hormone, Vitamin D, Calcitonin, and Other Compounds). PTH appears to increase Ca2+ channels in the luminal membrane, thereby facilitating the passive movement of Ca2+ into the epithelial cell. Ca2+ is extruded across the basolateral membrane by a Ca2+ATPase and via Na+ Ca 2+ antiport. Inorganic phosphate (Pi ) is largely reabsorbed (80% of filtered load) by the proximal tubule. A Na+Pi symporter uses the free energy of the Na+ electrochemical gradient to effect secondary active transport of Pi into the cell. The Na+Pi symporter is inhibited by PTH. Pi exits the basolateral membrane down its electrochemical gradient by a poorly understood transport system. Only 20% to 25% of Mg2+ is reabsorbed in the proximal tubule, and only 5% is reabsorbed by the DCT and collecting duct system. The bulk of Mg2+ is reabsorbed in the thick ascending limb via a paracellular pathway driven by the lumen-positive VT. However, transcellular movement of Mg2+ also may occur with basolateral exit via Na+Mg2+ antiport or via a Mg2+ATPase. The renal tubules play an extremely important role in the reabsorption of HCO3 and secretion of protons (tubular acidification) and thus participate critically in the maintenance of acidbase balance. A description of these processes is presented in the section on carbonic anhydrase inhibitors. Principles of Diuretic Action By definition, diuretics are drugs that increase the rate of urine flow; however, clinically useful diuretics also increase the rate of excretion of Na+ (natriuresis) and of an accompanying anion, usually Cl. NaCl in the body is the major determinant of extracellular fluid volume, and most clinical applications of diuretics are directed toward reducing extracellular fluid volume by decreasing total-body NaCl content. A sustained imbalance between dietary Na+ intake and Na+ loss is incompatible with life. A sustained positive Na+ balance would result in volume overload with pulmonary edema, and a sustained negative Na+ balance would result in volume depletion and cardiovascular collapse. Although continued administration of a diuretic causes a sustained net

deficit in total-body Na+ , the time course of natriuresis is finite as renal compensatory mechanisms bring Na+ excretion in line with Na+ intake, a phenomenon known as "diuretic braking." These compensatory, or braking, mechanisms include activation of the sympathetic nervous system, activation of the reninangiotensinaldosterone axis, decreased arterial blood pressure (which reduces pressure-natriuresis), hypertrophy of renal epithelial cells, increased expression of renal epithelial transporters, and perhaps alterations in natriuretic hormones such as atrial natriuretic peptide. Historically, the classification of diuretics was based on a mosaic of ideas such as site of action (loop diuretics), efficacy (high-ceiling diuretics), chemical structure (thiazide diuretics), similarity of action with other diuretics (thiazide-like diuretics), effects on potassium excretion (potassiumsparing diuretics), etc. However, since the mechanism of action of each of the major classes of diuretics is now reasonably well understood, a classification scheme based on mechanism of action is now possible and is used in this chapter. Diuretics not only alter the excretion of Na+, but also may modify renal handling of other cations (e.g., K+, H+, Ca2+, and Mg2+), anions (e.g., Cl , HCO 3, and H2PO4), and uric acid. In addition, diuretics may indirectly alter renal hemodynamics. Table 291 gives a comparison of the general effects of the major classes of diuretics. Inhibitors of Carbonic Anhydrase Acetazolamide (DIAMOX) is the prototype of a class of agents that have limited usefulness as diuretics but have played a major role in the development of fundamental concepts of renal physiology and pharmacology. Chemistry When sulfanilamide was introduced as a chemotherapeutic agent, metabolic acidosis was recognized as a side effect. This observation led to in vitro and in vivo studies demonstrating that sulfanilamide is an inhibitor of carbonic anhydrase. Subsequently, an enormous number of sulfonamides were synthesized and tested for the ability to inhibit carbonic anhydrase; of these compounds, acetazolamide has been most extensively studied. Table 292 lists the chemical structures of the three carbonic anhydrase inhibitors currently available in the United States acetazolamide, dichlorphenamide (DARANIDE ), and methazolamide (NEPTAZANE). The common molecular motif of available carbonic anhydrase inhibitors is an unsubstituted sulfonamide moiety. Mechanism and Site of Action Proximal tubular epithelial cells are richly endowed with the zinc metalloenzyme carbonic anhydrase, which is found in the luminal and basolateral membranes (type IV carbonic anhydrase, an enzyme tethered to the membrane by a glycosylphosphatidylinositol linkage) as well as in the cytoplasm (type II carbonic anhydrase). Davenport and Wilhelmi (1941) were the first to discover this enzyme in the mammalian kidney, and subsequent studies revealed the key role played by carbonic anhydrase in NaHCO3 reabsorption and acid secretion (seeMaren, 1967 and 1980). In the proximal tubule, the free energy in the Na+ gradient established by the basolateral Na+ pump is used by a Na+H+ antiporter (also referred to as a Na+ H+ exchanger or NHE) in the luminal membrane to transport H+ into the tubular lumen in exchange for Na+ (Figure 295). In the lumen, H+ reacts with filtered HCO3 to form H2CO3, which rapidly decomposes to CO2 and water in the presence of carbonic anhydrase in the brush border. Normally the reaction between CO2 and water

occurs slowly, but carbonic anhydrase reversibly accelerates this reaction several thousandfold. CO2 is lipophilic and rapidly diffuses across the luminal membrane into the epithelial cell where it reacts with water to form H2CO3, a reaction catalyzed by cytoplasmic carbonic anhydrase. (The actual reaction catalyzed by carbonic anhydrase is OH+ CO2 HCO3; however, H2O OH+ H + and HCO3+ H+ H2CO3, so that the net reaction is H2O + CO2 H2CO3.) Continued operation of the Na+H+ antiporter maintains a low proton concentration in the cell, so that H2CO3 spontaneously ionizes to form H+ and HCO3, creating an electrochemical gradient for HCO3 across the basolateral membrane. The electrochemical gradient for HCO3 is used by a Na+HCO3 symporter (also called Na+ HCO3 cotransporter or NBC) in the basolateral membrane to transport NaHCO3 into the interstitial space. The net effect of this process is transport of NaHCO3 from the tubular lumen to the interstitial space followed by movement of water (isotonic reabsorption). Removal of water concentrates Cl in the tubular lumen, and consequently Cl diffuses down its concentration gradient into the interstitium via the paracellular pathway. Figure 295. NaHCO3 Reabsorption in Proximal Tubule and Mechanism of Diuretic Action of Carbonic Anhydrase (CA) Inhibitors. A, antiporter; S, symporter; CH, ion channel. (The actual reaction catalyzed by carbonic + + anhydrase is OH + CO2 HCO3 ; however, H2O OH + H , and HCO3 + H H2CO3, so that the net reaction is H2O + CO2 H2CO3.) Numbers in parentheses indicate stoichiometry.

Carbonic anhydrase inhibitors potently inhibit (IC50 for acetazolamide is 10 nM) both the membrane-bound and cytoplasmic forms of carbonic anhydrase, resulting in nearly complete abolition of NaHCO3 reabsorption in the proximal tubule (Cogan et al., 1979). Studies with a highmolecular-weight carbonic anhydrase inhibitor that only inhibits luminal enzyme because of limited cellular permeability indicate that inhibition of both the membrane-bound and cytoplasmic pools of carbonic anhydrase contributes to the diuretic activity of carbonic anhydrase inhibitors (Maren et al., 1997). Because of the large excess of carbonic anhydrase in proximal tubules, a high percentage of enzyme activity must be inhibited before an effect on electrolyte excretion is observed. Although

the proximal tubule is the major site of action of carbonic anhydrase inhibitors, carbonic anhydrase also is involved in secretion of titratable acid in the collecting duct system (a process that involves a proton pump); therefore the collecting duct system is a secondary site of action for this class of drugs. Effects on Urinary Excretion Inhibition of carbonic anhydrase is associated with a rapid rise in urinary HCO3 excretion to approximately 35% of filtered load. This, along with inhibition of titratable acid and ammonia secretion in the collecting duct system, results in an increase in urinary pH to approximately 8 and development of a metabolic acidosis. However, even with a high degree of inhibition of carbonic anhydrase, 65% of HCO3 is rescued from excretion by poorly understood mechanisms that may involve carbonic anhydraseindependent HCO3 reabsorption at downstream sites. Inhibition of the transport mechanism described in the preceding section results in increased delivery of Na+ and Cl to the loop of Henle, which has a large reabsorptive capacity and captures most of the Cl and a portion of the Na+. Thus, only a small increase in Cl excretion occurs, HCO3 being the major anion excreted along with the cations Na+ and K+. The fractional excretion of Na + may be as much as 5%, and the fractional excretion of K+ can be as much as 70%. The increased excretion of K+ is secondary to increased delivery of Na+ to the distal nephron. The mechanism by which increased distal delivery of Na+ enhances K+ excretion is described in the section on inhibitors of sodium channels. Carbonic anhydrase inhibitors also increase phosphate excretion (mechanism unknown), but have little or no effect on the excretion of Ca2+ or Mg2+. The effects of carbonic anhydrase inhibitors on renal excretion are self-limiting, probably because, as metabolic acidosis develops, the filtered load of HCO3 decreases to the point that the uncatalyzed reaction between CO2 and water is sufficient to achieve HCO3 reabsorption. Effects on Renal Hemodynamics By inhibiting proximal reabsorption, carbonic anhydrase inhibitors increase delivery of solutes to the macula densa. This triggers tubuloglomerular feedback (TGF), which increases afferent arteriolar resistance and reduces renal blood flow (RBF) and glomerular filtration rate (GFR) (Persson and Wright, 1982). Other Actions Carbonic anhydrase is present in a number of extrarenal tissues including the eye, gastric mucosa, pancreas, central nervous system (CNS), and red blood cells (RBCs). Carbonic anhydrase in the ciliary processes of the eye mediates the formation of large amounts of HCO3 in aqueous humor. For this reason, inhibition of carbonic anhydrase decreases the rate of formation of aqueous humor and consequently reduces intraocular pressure. Acetazolamide frequently causes paresthesias and somnolence, suggesting an action of carbonic anhydrase inhibitors in the CNS. The efficacy of acetazolamide in epilepsy is in part due to the production of metabolic acidosis; however, direct actions of acetazolamide in the CNS also contribute to its anticonvulsant action. Due to interference with carbonic anhydrase activity in RBCs, carbonic anhydrase inhibitors increase CO2 levels in peripheral tissues and decrease CO2 levels in expired gas. Large doses of carbonic anhydrase inhibitors reduce gastric acid secretion, but this has no therapeutic applications. Absorption and Elimination The oral bioavailability, plasma half-life, and route of elimination of the three currently available carbonic anhydrase inhibitors are listed in Table 292. Carbonic anhydrase inhibitors are avidly

bound by carbonic anhydrase and, accordingly, tissues rich in this enzyme will have higher concentrations of carbonic anhydrase inhibitors following systemic administration. Toxicity, Adverse Effects, Contraindications, Drug Interactions Serious toxic reactions to carbonic anhydrase inhibitors are infrequent; however, these drugs are sulfonamide derivatives and, like other sulfonamides, may cause bone-marrow depression, skin toxicity, and sulfonamide-like renal lesions and may cause allergic reactions in patients hypersensitive to sulfonamides. With large doses, many patients exhibit drowsiness and paresthesias. Most adverse effects, contraindications, and drug interactions are secondary to urinary alkalinization or metabolic acidosis, including: (1) diversion of ammonia of renal origin from urine into the systemic circulation, a process that may induce hepatic encephalopathy (the drugs are contraindicated in patients with hepatic cirrhosis); (2) calculus formation and ureteral colic due to precipitation of calcium phosphate salts in an alkaline urine; (3) worsening of metabolic or respiratory acidosis (the drugs are contraindicated in patients with hyperchloremic acidosis or severe chronic obstructive pulmonary disease); (4) interference with the urinary tract antiseptic methenamine; and (5) reduction of the urinary excretion rate of weak organic bases. Therapeutic Uses Although acetazolamide is used for treatment of edema, the efficacy of carbonic anhydrase inhibitors as single agents is low, and carbonic anhydrase inhibitors are not widely employed in this regard. However, studies by Knauf and Mutschler (1997) indicate that the combination of acetazolamide with diuretics that block Na+ reabsorption at more distal sites in the nephron causes a marked natriuretic response in patients with low basal fractional excretion of Na+ (<0.2%) who are resistant to diuretic monotherapy. Even so, the long-term usefulness of carbonic anhydrase inhibitors often is compromised by development of metabolic acidosis. The major indication for carbonic anhydrase inhibitors is open-angle glaucoma. Carbonic anhydrase inhibitors also may be employed for secondary glaucoma and preoperatively in acute angle-closure glaucoma to lower ocular pressure before surgery (seeChapter 66: Ocular Pharmacology). Acetazolamide also is used for the treatment of epilepsy (seeChapter 21: Drugs Effective in the Therapy of the Epilepsies). The rapid development of tolerance, however, may limit the usefulness of carbonic anhydrase inhibitors for epilepsy. Acetazolamide may provide symptomatic relief in patients with acute mountain sickness; however, it is more appropriate to give acetazolamide as a prophylactic measure (Coote, 1991). Acetazolamide also is useful in patients with familial periodic paralysis (Links et al., 1988). The mechanism for the beneficial effects of acetazolamide in mountain sickness and familial periodic paralysis is not clear, but it may be related to the induction of a metabolic acidosis. Finally, carbonic anhydrase inhibitors can be useful for correcting a metabolic alkalosis, especially an alkalosis caused by diuretic-induced increases in H+ excretion. Osmotic Diuretics Osmotic diuretics are agents that are freely filtered at the glomerulus, undergo limited reabsorption by the renal tubule, and are relatively inert pharmacologically. Osmotic diuretics are administered in large enough doses to increase significantly the osmolality of plasma and tubular fluid. Table 29 3 gives the molecular structures of the four currently available osmotic diureticsglycerin, isosorbide, mannitol, and urea. Mechanism and Site of Action

For many years it was thought that osmotic diuretics act primarily in the proximal tubule (Wesson and Anslow, 1948). By acting as nonreabsorbable solutes, it was reasoned that osmotic diuretics limit the osmosis of water into the interstitial space and thereby reduce luminal Na+ concentration to the point that net Na+ reabsorption ceases. Indeed, early micropuncture studies supported this concept (Windhager et al., 1959). However, subsequent studies suggest that this mechanism, while operative, may be of only secondary importance. For instance, mannitol only slightly increases the delivery of Na+ and moderately increases the delivery of water out of the proximal tubule (Seely and Dirks, 1969), and urea does not alter proximal tubular reabsorption in rats at the time of a large osmotic diuresis (Kauker et al., 1970). Mannitol, on the other hand, markedly increases the delivery of Na+ and water out of the loop of Henle (Seely and Dirks, 1969), suggesting that the major site of action is the loop of Henle. By extracting water from intracellular compartments, osmotic diuretics expand the extracellular fluid volume, decrease blood viscosity, and inhibit renin release. These effects increase RBF, and the increase in renal medullary blood flow removes NaCl and urea from the renal medulla, thus reducing medullary tonicity. Also, under some circumstances, prostaglandins may contribute to the renal vasodilation and medullary washout induced by osmotic diuretics (Johnston et al., 1981). A reduction in medullary tonicity causes a decrease in the extraction of water from the DTL, which in turn limits the concentration of NaCl in the tubular fluid entering the ATL. This latter effect diminishes the passive reabsorption of NaCl in the ATL. In addition, the marked ability of osmotic diuretics to inhibit reabsorption of Mg2+, a cation that is mainly reabsorbed in the thick ascending limb, suggests that osmotic diuretics also interfere with transport processes in the thick ascending limb. The mechanism of this effect is unknown. In summary, osmotic diuretics act both in the proximal tubule and the loop of Henle, with the latter being the primary site of action. Also, osmotic diuretics probably act by an osmotic effect in the tubules and by reducing medullary tonicity. Effects on Urinary Excretion Osmotic diuretics increase the urinary excretion of nearly all electrolytes, including Na+, K+ , Ca2+, Mg2+, Cl , HCO 3, and phosphate. Effects on Renal Hemodynamics As indicated in the preceding section, osmotic diuretics increase RBF by a variety of mechanisms. Osmotic diuretics dilate the afferent arteriole, which increases PGC , and dilute the plasma, which decreases GC . These effects would increase GFR were it not for the fact that osmotic diuretics also increase PT. In general, superficial SNGFR is increased but total GFR is little changed. Absorption and Elimination The oral bioavailability, plasma half-life, and route of elimination of the four currently available osmotic diuretics are listed in Table 293. Glycerin and isosorbide can be given orally, whereas mannitol and urea must be administered intravenously. Toxicity, Adverse Effects, Contraindications, Drug Interactions Osmotic diuretics are distributed in the extracellular fluid and contribute to the extracellular osmolality. Thus, water is extracted from intracellular compartments, and the extracellular fluid volume becomes expanded. In patients with heart failure or pulmonary congestion, this may cause

frank pulmonary edema. Extraction of water also causes hyponatremia, which may explain common adverse effects, including headache, nausea, and vomiting. On the other hand, loss of water in excess of electrolytes can cause hypernatremia and dehydration. In general, osmotic diuretics are contraindicated in patients who are anuric due to severe renal disease or who are unresponsive to test doses of the drugs. Urea may cause thrombosis or pain if extravasation occurs, and it should not be administered to patients with impaired liver function because of the risk of elevation of blood ammonia levels. Both mannitol and urea are contraindicated in patients with active cranial bleeding. Glycerin is metabolized and can cause hyperglycemia. Therapeutic Uses A rapid decrease in GFR, i.e., acute renal failure (ARF), is a serious medical condition that occurs in 5% of hospitalized patients and is associated with a significant mortality rate. ARF can be caused by diverse conditions both extrinsic (prerenal and postrenal failure) and intrinsic to the kidney. Acute tubular necrosis (ATN), i.e., damage to tubular epithelial cells, accounts for the majority of cases of intrinsic ARF. In animal models, mannitol is effective in attenuating the reduction in GFR associated with ATN when administered before the ischemic insult or offending nephrotoxin. The renal protection afforded by mannitol may be due to removal of obstructing tubular casts, dilution of nephrotoxic substances in the tubular fluid, and/or reduction of swelling of tubular elements via osmotic extraction of water. Although prophylactic mannitol is effective in animal models of ATN, the clinical efficacy of mannitol is less well established. Most published clinical studies have been uncontrolled, and controlled studies have not shown a benefit over hydration per se (seeKellum, 1998). In patients with mild-to-moderate renal insufficiency, hydration with 0.45% sodium chloride is as good as or better than either mannitol or furosemide in protection against decreases in GFR induced by radiocontrast agents (Soloman et al., 1994). Studies of prophylactic mannitol indicate effectiveness in jaundiced patients undergoing surgery (Dawson, 1965). However, in vascular and open heart surgery, prophylactic mannitol maintains urine flow but not GFR. In established ATN, mannitol will increase urine volume in some patients, and those patients converted from oliguric to nonoliguric ATN appear to recover more rapidly and require less dialysis compared with patients who do not respond to mannitol (Levinsky and Bernard, 1988). However, it is not clear whether these benefits are due to the diuretic or whether "responders" have lesser degrees of renal damage from the outset compared with "nonresponders." Repeated administration of mannitol to nonresponders is not recommended, and nowadays loop diuretics are more frequently used to convert oliguric to nonoliguric ATN. Another use for mannitol and urea is in the treatment of dialysis disequilibrium syndrome. Too rapid a removal of solutes from the extracellular fluid by hemodialysis or peritoneal dialysis results in a reduction in the osmolality of the extracellular fluid. Consequently, water moves from the extracellular compartment into the intracellular compartment, causing hypotension and CNS symptoms (headache, nausea, muscle cramps, restlessness, CNS depression, and convulsions). Osmotic diuretics increase the osmolality of the extracellular fluid compartment and thereby shift water back into the extracellular compartment. By increasing the osmotic pressure of the plasma, osmotic diuretics extract water from the eye and brain. All four osmotic diuretics are used to control intraocular pressure during acute attacks of glaucoma and for short-term reductions in intraocular pressure, both preoperatively and postoperatively, in patients who require ocular surgery. Also, mannitol and urea are used to reduce cerebral edema and brain mass before and after neurosurgery. Inhibitors of Na K 2 Cl Symport (Loop Diuretics; High-Ceiling Diuretics)
+ +

Inhibitors of Na+K+2Cl symport are a group of diuretics that have in common an ability to block the Na+ K+ 2Cl symporter in the thick ascending limb of the loop of Henle; hence these diuretics also are referred to as loop diuretics. Although the proximal tubule reabsorbs approximately 65% of the filtered Na+, diuretics acting only in the proximal tubule have limited efficacy because the thick ascending limb has a great reabsorptive capacity and reabsorbs most of the rejectate from the proximal tubule. Diuretics acting predominantly at sites past the thick ascending limb also have limited efficacy, because only a small percentage of the filtered Na+ load reaches these more distal sites. In contrast, inhibitors of Na+ K+ 2Cl symport are highly efficacious, and for this reason they often are called high-ceiling diuretics. The efficacy of inhibitors of Na+K+2Cl symport in the thick ascending limb of the loop of Henle is due to a combination of two factors: (1) Approximately 25% of the filtered Na+ load normally is reabsorbed by the thick ascending limb; and (2) nephron segments past the thick ascending limb do not possess the reabsorptive capacity to rescue the flood of rejectate exiting the thick ascending limb. Chemistry Inhibitors of Na+K+2Cl symport are a chemically diverse group of drugs (seeTable 294). Furosemide, bumetanide, azosemide, piretanide, and tripamide all contain a sulfonamide moiety, whereas ethacrynic acid is a phenoxyacetic acid derivative. Muzolimine has neither of these structural features, and torsemide is a sulfonylurea. Only furosemide ( LASIX ), bumetanide ( BUMEX ), ethacrynic acid (EDECRIN), and torsemide ( DEMADEX ) are available currently in the United States. Mechanism and Site of Action Inhibitors of Na+K+2Cl symport act primarily in the thick ascending limb. Micropuncture of the DCT demonstrates that loop diuretics increase the delivery of solutes out of the loop of Henle (Dirks and Seely, 1970). Also, in situ microperfusion of the loop of Henle (Morgan et al., 1970) and in vitro microperfusion of the CTAL (Burg et al., 1973) indicate inhibition of transport by low concentrations of furosemide in the perfusate. Some inhibitors of Na+ K+2Cl symport may have additional effects in the proximal tubule; however, the significance of these effects is unclear. It was initially thought that Cl was transported by a primary active electrogenic transporter in the luminal membrane independent of Na+. Discovery of furosemide-sensitive Na+K+2Cl symport in other tissues caused Greger (1981) to investigate more carefully the Na+ dependence of Cl transport in the isolated perfused rabbit CTAL. By scrupulously removing Na+ from the luminal perfusate, Greger demonstrated the dependence of Cl transport on Na+. It is now well accepted that, in the thick ascending limb, flux of Na+, K+ , and Cl from the lumen into the epithelial cell is mediated by a Na+K+2Cl symporter (seeFigure 296). This symporter captures the free energy in the Na+ electrochemical gradient established by the basolateral Na+ pump and provides for "uphill" transport of K+ and Cl into the cell. K+ channels in the luminal membrane (called ROMK) provide a conductive pathway for the apical recycling of this cation (Ho et al., 1993; Kohda et al., 1998), and basolateral Cl channels (called CLCN) provide a basolateral exit mechanism for Cl. The luminal membranes of epithelial cells in the thick ascending limb have conductive pathways (channels) only for K+ ; therefore the apical membrane voltage is determined by the equilibrium potential for K+ (EK). In contrast, the basolateral membrane has channels for both K+ and Cl, so that the basolateral membrane voltage is less than EK ; i.e., conductance for Cl depolarizes the basolateral membrane. Depolarization of the basolateral membrane results in a transepithelial potential difference of approximately 10 mV, with the lumen positive with respect to the interstitial + 2+ 2+ space. This lumen-positive potential difference repels cations (Na , Ca , and Mg ) and thereby provides an important driving force for the paracellular flux of these cations into the interstitial

space. Figure 296. NaCl Reabsorption in Thick Ascending Limb and Mechanism of Diuretic Action of Na+ K+2Cl Symport Inhibitors. S, symporter; CH, ion channel. Numbers in parentheses indicate stoichiometry. Designated voltages are the potential differences across the indicated membrane or cell.

As the name implies, inhibitors of Na+K+2Cl symport bind to the Na+K+2Cl symporter in the thick ascending limb (Koenig et al., 1983) and block its function, bringing salt transport in this segment of the nephron to a virtual standstill (Burg et al., 1973). The molecular mechanism by which this class of drugs blocks the Na+K+2Cl symporter is unknown, but evidence suggests that these drugs attach to the Clbinding site (Hannafin et al., 1983) located in the symporter's transmembrane domain (Isenring and Forbush, 1997). Inhibitors of Na+K+2Cl symport also inhibit Ca2+ and Mg2+ reabsorption in the thick ascending limb by abolishing the transepithelial potential difference that is the dominant driving force for reabsorption of these cations. Na+K+2Cl symporters are an important family of transport molecules found in many secretory and absorbing epithelia. The rectal gland of the dogfish shark is a particularly rich source of the protein, and a cDNA encoding a Na+K+2Cl symporter was isolated from a cDNA library obtained from the dogfish shark rectal gland by screening with antibodies to the shark symporter (Xu et al., 1994). Molecular cloning revealed a deduced amino acid sequence of 1191 residues containing 12 putative membrane-spanning domains flanked by long N and C termini in the + + cytoplasm. Expression of this protein resulted in Na K 2Cl symport that was sensitive to + + bumetanide. The shark rectal gland Na K 2Cl symporter cDNA subsequently was used to screen + + a human colonic cDNA library, and this provided Na K 2Cl symporter cDNA probes from this tissue. These latter probes were used to screen rabbit renal cortical and renal medullary libraries, + + which allowed cloning of the rabbit renal Na K 2Cl symporter (Payne and Forbush, 1994). This + + symporter is 1099 amino acids in length, is 61% identical to the dogfish shark secretory Na K 2Cl symporter, has 12 predicted transmembrane helices, and contains large N- and C-terminal

cytoplasmic regions. Subsequent studies demonstrated that Na+ K+ 2Cl symporters are of two varieties (seeKaplan et al., 1996). The "absorptive" symporter (called ENCC2, NKCC2, or BSC1) is expressed only in the kidney, is localized to the apical membrane of the thick ascending limb, and is regulated by cyclic AMP (Obermller et al., 1996; Kaplan et al., 1996; Nielsen et al., 1998; Plata et al., 1999). At least six different isoforms of the absorptive symporter are generated by alternative mRNA splicing (Mount et al., 1999). The "secretory" symporter (called ENCC3, NKCC1, or BSC2) is a "housekeeping" protein that is widely expressed and, in epithelial cells, is localized to the basolateral membrane. A model of Na+K+2Cl symport has been proposed based on ordered binding of ions to the symporter (Lytle et al., 1998). Mutations in the genes coding for the absorptive Na+K+2Cl symporter, the apical K+ channel, or the basolateral Cl channel give rise to Bartter's syndrome (inherited hypokalemic alkalosis with salt wasting and hypotension) (seeSimon and Lifton, 1998). Effects on Urinary Excretion Due to blockade of the Na +K+ 2Cl symporter, loop diuretics cause a profound increase in the urinary excretion of Na+ and Cl (i.e., up to 25% of the filtered load of Na +). Abolition of the transepithelial potential difference also results in marked increases in the excretion of Ca2+ and Mg2+. Some (e.g., furosemide), but not all (e.g., bumetanide and piretanide), sulfonamide-based loop diuretics have weak carbonic anhydraseinhibiting activity. Those drugs with carbonic anhydraseinhibiting activity increase the urinary excretion of HCO3 and phosphate. The mechanism by which inhibition of carbonic anhydrase increases phosphate excretion is not known. All inhibitors of Na+ K+ 2Cl symport increase the urinary excretion of K+ and titratable acid. This effect is due in part to increased delivery of Na+ to the distal tubule. The mechanism by which increased distal delivery of Na+ enhances excretion of K+ and H+ is discussed in the section on inhibitors of Na+ channels. Acutely, loop diuretics increase the excretion of uric acid, whereas chronic administration of these drugs results in reduced excretion of uric acid. The chronic effects of loop diuretics on uric acid excretion may be due to enhanced transport in the proximal tubule secondary to volume depletion, leading to increased uric acid reabsorption, or to competition between the diuretic and uric acid for the organic acid secretory mechanism in the proximal tubule, leading to reduced uric acid secretion. By blocking active NaCl reabsorption in the thick ascending limb, inhibitors of Na+K+2Cl symport interfere with a critical step in the mechanism that produces a hypertonic medullary interstitium. Therefore, loop diuretics block the kidney's ability to concentrate urine during + hydropenia. Also, since the thick ascending limb is part of the diluting segment, inhibitors of Na + K 2Cl symport markedly impair the kidney's ability to excrete a dilute urine during water diuresis. Effects on Renal Hemodynamics If volume depletion is prevented by replacing fluid losses, inhibitors of Na+K+2Cl symport generally increase total RBF and redistribute RBF to the midcortex (Stein et al., 1972). However, the effects on RBF are variable. The mechanism of the increase in RBF is not known, but prostaglandins have been implicated (Williamson et al., 1974). In fact, nonsteroidal antiinflammatory drugs (NSAIDs) attenuate the diuretic response to loop diuretics, most likely by preventing prostaglandin-mediated increases in RBF (Brater, 1985). Loop diuretics block TGF by inhibiting salt transport into the macula densa, so that the macula densa no longer can "sense" NaCl concentrations in the tubular fluid. Therefore, unlike carbonic anhydrase inhibitors, loop diuretics do not decrease GFR by activating TGF. Loop diuretics are powerful stimulators of renin release. This effect is due to interference with NaCl transport by the macula densa and, if volume depletion

occurs, to reflex activation of the sympathetic nervous system and to stimulation of the intrarenal baroreceptor mechanism. Prostaglandins, particularly prostacyclin, may play an important role in mediating the renin release response to loop diuretics (Oates et al., 1979). Other Actions Loop diuretics may cause direct vascular effects (seeDormans et al., 1996). Loop diuretics, particularly furosemide, acutely increase systemic venous capacitance and thereby decrease left ventricular filling pressure. This effect, which may be mediated by prostaglandins and requires intact kidneys (Johnston et al., 1983), benefits patients with pulmonary edema even before diuresis ensues. Furosemide and ethacrynic acid can inhibit Na+, K+ATPase, glycolysis, mitochondrial respiration, the microsomal Ca 2+ pump, adenylyl cyclase, phosphodiesterase, and prostaglandin dehydrogenase; however, these effects do not have therapeutic implications. In vitro, high doses of inhibitors of Na+K+2Cl symport can inhibit electrolyte transport in many tissues. Only in the inner ear, where alterations in the electrolyte composition of endolymph may contribute to druginduced ototoxicity, is this effect clinically important. Absorption and Elimination The oral bioavailability, plasma half-life, and route of elimination of the four inhibitors of Na+ K+ 2Cl symport available in the United States are listed in Table 294. Because furosemide, bumetanide, ethacrynic acid, and torsemide are extensively bound to plasma proteins, delivery of these drugs to the tubules by filtration is limited. However, they are efficiently secreted by the organic acid transport system in the proximal tubule and thereby gain access to their binding sites on the Na+K +2Cl symport in the luminal membrane of the thick ascending limb. Probenecid shifts the plasma concentrationresponse curve to furosemide to the right by competitively inhibiting furosemide secretion by the organic acid transport system (Brater, 1983). The most recent loop diuretic to receive FDA approval is torsemide, which has a longer half-life than the other loop diuretics available in the United States (Brater, 1991: seeKnauf and Mutschler, 1997). Toxicity, Adverse Effects, Contraindications, Drug Interactions Adverse effects unrelated to the diuretic efficacy are rare, and most adverse effects are due to abnormalities of fluid and electrolyte balance. Overzealous use of loop diuretics can cause serious depletion of total body Na+. This may be manifest as hyponatremia and/or extracellular fluid volume depletion associated with hypotension, reduced GFR, circulatory collapse, thromboembolic episodes, and, in patients with liver disease, hepatic encephalopathy. Increased delivery of Na+ to the distal tubule, particularly when combined with activation of the reninangiotensin system, leads to increased urinary excretion of K+ and H +, causing a hypochloremic alkalosis. If dietary K+ intake is not sufficient, hypokalemia may develop, and this may induce cardiac arrhythmias, particularly in patients taking cardiac glycosides. Increased Mg2+ and Ca2+ excretion may result in hypomagnesemia (a risk factor for cardiac arrhythmias) and hypocalcemia (rarely leading to tetany). Loop diuretics can cause ototoxicity that manifests itself as tinnitus, hearing impairment, deafness, vertigo, and a sense of fullness in the ears. Hearing impairment and deafness are usually, but not always, reversible. Ototoxicity occurs most frequently with rapid intravenous administration and least frequently with oral administration. Ethacrynic acid appears to induce ototoxicity more often than do other loop diuretics. Loop diuretics also can cause hyperuricemia (rarely leading to gout) and hyperglycemia (rarely precipitating diabetes mellitus) and can increase plasma levels of lowdensity lipoprotein (LDL) cholesterol and triglycerides, while decreasing plasma levels of high-

density lipoprotein (HDL) cholesterol. Other adverse effects include skin rashes, photosensitivity, paresthesias, bone marrow depression, and gastrointestinal disturbances. Contraindications to the use of loop diuretics include severe Na+ and volume depletion, hypersensitivity to sulfonamides (for sulfonamide-based loop diuretics), and anuria unresponsive to a trial dose of loop diuretic. Drug interactions may occur when loop diuretics are coadministered with: (1) aminoglycosides (synergism of ototoxicity caused by both drugs); (2) anticoagulants (increased anticoagulant activity); (3) digitalis glycosides (increased digitalis-induced arrhythmias); (4) lithium (increased plasma levels of lithium); (5) propranolol (increased plasma levels of propranolol); (6) sulfonylureas (hyperglycemia); (7) cisplatin (increased risk of diuretic-induced ototoxicity); (8) NSAIDs (blunted diuretic response; salicylate toxity when given with high doses of salicylates); (9) probenecid (blunted diuretic response); (10) thiazide diuretics (synergism of diuretic activity of both drugs leading to profound diuresis); and (11) amphotericin B (increased potential for nephrotoxicity and toxicity and intensification of electrolyte imbalance). Therapeutic Uses A major use of loop diuretics is in the treatment of acute pulmonary edema. A rapid increase in venous capacitance in conjunction with a brisk natriuresis reduces left ventricular filling pressures and thereby rapidly relieves pulmonary edema. Loop diuretics also are widely used for the treatment of chronic congestive heart failure when diminution of extracellular fluid volume is desirable to minimize venous and pulmonary congestion (seeChapter 34: Pharmacological Treatment of Heart Failure). Diuretics are widely used for the treatment of hypertension (seeChapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension), and controlled clinical trials demonstrating reduced morbidity and mortality have been conducted with Na+Cl symport (thiazides and thiazide-like diuretics), but not Na+ K+ 2Cl symport, inhibitors. Nonetheless, Na+K+2Cl symport inhibitors appear to lower blood pressure as effectively as Na+ Cl symport inhibitors while causing smaller perturbations in the lipid profile (van der Heijden et al., 1998). The edema of nephrotic syndrome often is refractory to other classes of diuretics, and loop diuretics often are the only drugs capable of reducing the massive edema associated with this renal disease. Loop diuretics also are employed in the treatment of edema and ascites of liver cirrhosis; however, care must be taken not to induce encephalopathy or hepatorenal syndrome. In patients with a drug overdose, loop diuretics can be used to induce a forced diuresis to facilitate more rapid renal elimination of the offending drug. Loop diureticscombined with isotonic saline administration to prevent volume depletionare used to treat hypercalcemia. Loop diuretics interfere with the kidney's ability to produce a concentrated urine. Consequently, loop diuretics combined with hypertonic saline are useful for the treatment of life-threatening hyponatremia. Loop diuretics also are used to treat edema associated with chronic renal insufficiency. However, animal studies have demonstrated that loop diuretics increase PGC by activating the reninangiotensin system, an effect that could accelerate renal injury (Lane et al., 1998). Most patients with ARF receive a trial dose of a loop diuretic in an attempt to convert oliguric ARF to nonoliguric ARF. However, there is no evidence that loop diuretics prevent ATN or improve outcome in patients with ARF (seeKellum, 1998). Inhibitors of Na+ Cl Symport (Thiazide and Thiazide-Like Diuretics) The benzothiadiazides were synthesized in an effort to enhance the potency of inhibitors of carbonic anhydrase. However, unlike carbonic anhydrase inhibitors, which primarily increase NaHCO3 excretion, benzothiadiazides were found predominantly to increase NaCl excretion (Beyer, 1958),

an effect shown to be independent of carbonic anhydrase inhibition. Chlorothiazide was the first serious challenge to the mercurial diuretics, a now-obsolete class of organometallic compounds that dominated diuretic therapy for more than thirty years. Chemistry Inhibitors of Na+Cl symport are sulfonamides (seeTable 295) and many are analogs of 1,2,4benzothiadiazine-1,1-dioxide. Because the original inhibitors of Na +Cl symport were benzothiadiazine derivatives, this class of diuretics became known as thiazide diuretics. Subsequently, drugs that are pharmacologically similar to thiazide diuretics but are not thiazides were developed and are called thiazide-like diuretics. The term thiazide diuretics often is used to refer to all members of the class of inhibitors of Na+Cl symport, and this usage is employed in the present chapter. Mechanism and Site of Action Some studies using split-droplet and stationary-microperfusion techniques have described reductions in proximal tubule reabsorption by thiazide diuretics; however, free-flow micropuncture studies have not consistently demonstrated increased solute delivery out of the proximal tubule following administration of thiazides. In contrast, micropuncture (Kunau et al., 1975) and in situ microperfusion studies (Costanzo and Windhager, 1978) clearly indicate that thiazide diuretics inhibit NaCl transport in the DCT. Furthermore, the renal cortex has a high-affinity receptor for thiazide diuretics (Beaumont et al., 1988), and binding of thiazides localizes to the DCT (Beaumont et al., 1989). It is now well accepted that the primary site of action of thiazide diuretics is the DCT, whereas the proximal tubule may represent a secondary site of action. Figure 297 illustrates the current model of electrolyte transport in the DCT. As with other nephron segments, transport is powered by a Na+ pump in the basolateral membrane. The free energy in the electrochemical gradient for Na+ is harnessed by a Na +Cl symporter in the luminal membrane, which moves Cl into the epithelial cell against its electrochemical gradient. Cl then passively exits the basolateral membrane via a Cl channel. Thiazide diuretics inhibit the Na+Cl symporter, perhaps by competing for the Cl binding site (Beaumont et al., 1988). Figure 297. NaCl Reabsorption in Distal Convoluted Tubule and Mechanism of Diuretic Action of Na+ Cl Symport Inhibitors. S, symporter; CH, ion channel. Numbers in parentheses indicate stoichiometry.

Using a functional expression strategy (Cl -dependent Na uptake in Xenopus oocytes), Gamba et al. (1993) isolated a cDNA clone from the urinary bladder of the winter flounder that codes for a + + Na Cl symporter. This Na Cl symporter is inhibited by a number of thiazide diuretics (but not by furosemide, acetazolamide, or an amiloride derivative), has 12 putative membrane-spanning domains, and its sequence is 47% identical to the cloned dogfish shark rectal gland Na+K+2Cl symporter. Subsequently, Gamba et al. (1994) cloned the rat and Mastroianni et al. (1996) cloned the human Na+Cl symporter. The Na+Cl symporter (called ENCC1 or TSC) is expressed predominantly in the kidney (Chang et al., 1996) and is localized to the apical membrane of DCT epithelial cells (Bachmann et al., 1995; Obermller et al., 1995; Plotkin et al., 1996). Expression of the Na+ Cl symporter is regulated by aldosterone (Velzquez et al., 1996; Kim et al., 1998; Bostonjoglo et al., 1998). Mutations in the Na+Cl symporter cause a form of inherited hypokalemic alkalosis called Gitelman's syndrome (seeSimon and Lifton, 1998). Effects on Urinary Excretion As would be expected from their mechanism of action, inhibitors of Na+Cl symport increase Na + and Cl excretion. However, thiazides are only moderately efficacious (i.e., maximum excretion of filtered load of Na+ is only 5%), since approximately 90% of the filtered Na+ load is reabsorbed before reaching the DCT. Some thiazide diuretics also are weak inhibitors of carbonic anhydrase, an effect that increases HCO3 and phosphate excretion and probably accounts for the weak proximal tubular effects of some thiazide diuretics. Like inhibitors of Na+K+2Cl symport, inhibitors of Na+Cl symport increase the excretion of K+ and titratable acid due to increased delivery of Na+ to the distal tubule. Acute administration of thiazides increases the excretion of uric acid. However, uric acid excretion is reduced following chronic administration by the same mechanisms discussed for loop diuretics. The acute effects of inhibitors of Na+Cl symport on Ca2+ excretion are variable; when administered chronically, thiazide diuretics decrease Ca2+ excretion. The mechanism is unknown but may involve increased proximal reabsorption due to volume depletion as well as direct effects of thiazides to increase Ca2+ reabsorption in the DCT. Thiazide diuretics may cause a mild magnesuria by a poorly understood mechanism, and there is increasing awareness that longterm use of thiazide diuretics may cause magnesium deficiency, particularly in the elderly (Martin and Milligan, 1987). Since inhibitors of Na+Cl symport inhibit transport in the cortical diluting segment, thiazide diuretics attenuate the ability of the kidney to excrete a dilute urine during water diuresis. However, since the DCT is not involved in the mechanism that generates a hypertonic

medullary interstitium, thiazide diuretics do not alter the kidney's ability to concentrate urine during hydropenia. Effects on Renal Hemodynamics In general, inhibitors of Na+Cl symport do not affect RBF and only variably reduce GFR due to increases in intratubular pressure. Since thiazides act at a point past the macula densa, they have little or no influence on TGF. Other Actions Thiazide diuretics may inhibit phosphodiesterase, mitochondrial oxygen consumption, and renal uptake of fatty acids; however, these effects are not of clinical significance. Absorption and Elimination The relative potency, oral bioavailability, plasma half-life, and route of elimination of inhibitors of Na+Cl symport currently used in the United States are listed in Table 295. Of special note is the wide range of half-lives for this class of drugs. Sulfonamides are organic acids and therefore are secreted into the proximal tubule by the organic acid secretory pathway. Since thiazides must gain + access to the tubular lumen to inhibit the Na Cl symporter, drugs such as probenecid can attenuate the diuretic response to thiazides by competing for transport into the proximal tubule. However, plasma protein binding varies considerably among thiazide diuretics, and this parameter determines the contribution that filtration makes to tubular delivery of a specific thiazide. Toxicity, Adverse Effects, Contraindications, Drug Interactions Thiazide diuretics rarely cause CNS (vertigo, headache, paresthesias, xanthopsia, weakness), gastrointestinal (anorexia, nausea, vomiting, cramping, diarrhea, constipation, cholecystitis, pancreatitis), hematological (blood dyscrasias), and dermatological (photosensitivity, skin rashes) disorders. The incidence of sexual dysfunction (i.e., erection problems) is greater with Na+ Cl symport inhibitors than with several other antihypertensive agents ( -adrenergic receptor antagonists, calcium channel blockers, angiotensin converting enzyme inhibitors, 1-receptor antagonists) (Grimm et al., 1997), but usually is tolerable. However, like loop diuretics, most serious adverse effects of thiazides are related to abnormalities of fluid and electrolyte balance. These adverse effects include extracellular volume depletion, hypotension, hypokalemia, hyponatremia, hypochloremia, metabolic alkalosis, hypomagnesemia, hypercalcemia, and hyperuricemia. Thiazide diuretics have caused fatal or near fatal hyponatremia, and some patients are at recurrent risk of hyponatremia when rechallenged with thiazides. Thiazide diuretics also decrease glucose tolerance, and latent diabetes mellitus may be unmasked during therapy. The mechanism of the reduced glucose tolerance is not completely understood but appears to involve reduced insulin secretion and alterations in glucose metabolism. Hyperglycemia + + may be related in some way to K depletion, in that hyperglycemia is reduced when K is given along with the diuretic (Tannen, 1985). Thiazide diuretics also may increase plasma levels of LDL cholesterol, total cholesterol, and total triglycerides. Thiazide diuretics are contraindicated in individuals who are hypersensitive to sulfonamides. With regard to drug interactions, thiazide diuretics may diminish the effects of anticoagulants, uricosuric agents used to treat gout, sulfonylureas, and insulin and may increase the effects of anesthetics, diazoxide, digitalis glycosides, lithium, loop diuretics, and vitamin D. The effectiveness

of thiazide diuretics may be reduced by NSAIDs, bile acid sequestrants (reduced absorption of thiazides), and methenamines (alkalinization of urine may decrease effectiveness of thiazides). Amphotericin B and corticosteroids increase the risk of hypokalemia induced by thiazide diuretics. A potentially lethal drug interaction warranting special emphasis is that involving thiazide diuretics with quinidine (Roden, 1993). Prolongation of the QT-interval by quinidine can lead to the development of polymorphic ventricular tachycardia (torsades de pointes) due to triggered activity originating from early afterdepolarizations (seeChapter 35: Antiarrhythmic Drugs). Although usually self-limiting, torsades de pointes may deteriorate into fatal ventricular fibrillation. Hypokalemia increases the risk of quinidine-induced torsades de pointes, and thiazide diuretics cause hypokalemia. It is likely, therefore, that thiazide diureticinduced K+ depletion accounts for many cases of quinidine-induced torsades de pointes. Therapeutic Uses Thiazide diuretics are used for treatment of the edema associated with heart (congestive heart failure), liver (hepatic cirrhosis), and renal (nephrotic syndrome, chronic renal failure, acute glomerulonephritis) disease. With the exceptions of metolazone and indapamide, most thiazide diuretics are ineffective when GFR is <30 to 40 ml/min. Thiazide diuretics decrease blood pressure in hypertensive patients by increasing the slope of the renal pressurenatriuresis relationship (Saito and Kimura, 1996), and thiazide diuretics are widely used for the treatment of hypertension, either alone or in combination with other antihypertensive drugs (seeChapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension). In this regard, thiazide diuretics are inexpensive, as efficacious as other classes of antihypertensive agents, and well tolerated. Thiazides can be administered once daily, do not require dose titration, and have few contraindications. Moreover, thiazides have additive or synergistic effects when combined with other classes of antihypertensive agents. Although thiazides may marginally increase the risk of sudden death (seeHoes and Grobbee, 1996) and renal cell carcinoma (Grossman et al., 1999), in general these agents are safe and reduce cardiovascular morbidity and mortality in hypertensive patients. Because the adverse effects of thiazides increase progressively in severity at doses higher than maximally effective antihypertensive doses, only low doses should be prescribed for hypertension (see Ramsey, 1999). Thiazide diuretics, which reduce urinary excretion of Ca2+, sometimes are employed to treat calcium nephrolithiasis and may be useful for the treatment of osteoporosis (seeChapter 62: Agents Affecting Calcification and Bone Turnover: Calcium, Phosphate, Parathyroid Hormone, Vitamin D, Calcitonin, and Other Compounds). Thiazide diuretics are also the mainstay for treatment of nephrogenic diabetes insipidus, reducing urine volume by up to 50%. The mechanism of this paradoxical effect remains unknown (Grnbeck et al., 1998). Since other halides are excreted by renal processes similar to those for Cl, thiazide diuretics may be useful for the management of Br intoxication. Inhibitors of Renal Epithelial Na+ Channels (K+ -Sparing Diuretics) Triamterene (DYRENIUM , MAXZIDE ) and amiloride (MIDAMOR ) are the only two drugs of this class in clinical use. Both drugs cause small increases in NaCl excretion and usually are employed for their antikaluretic actions to offset the effects of other diuretics that increase K+ excretion. Consequently, triamterene and amiloride, along with spironolactone (see next section), often are classified as potassium (K+ )-sparing diuretics. Chemistry

Amiloride is a pyrazinoylguanidine derivative, and triamterene is a pteridine (Table 296). Both drugs are organic bases and are transported by the organic base secretory mechanism in the proximal tubule. Mechanism and Site of Action Available data suggest that triamterene and amiloride have similar mechanisms of action. Of the two, amiloride has been studied much more extensively, so its mechanism of action is known with a higher degree of certainty. As illustrated in Figure 298, principal cells in the late distal tubule and collecting duct have in their luminal membranes a Na+ channel that provides a conductive pathway for the entry of Na+ into the cell down the electrochemical gradient created by the basolateral Na+ pump. The higher permeability of the luminal membrane for Na+ depolarizes the luminal membrane, but not the basolateral membrane, creating a lumen-negative transepithelial potential difference. This transepithelial voltage provides an important driving force for the secretion of K+ into the lumen via K+ channels (ROMK) in the luminal membrane. Carbonic anhydrase inhibitors, loop diuretics, and thiazide diuretics increase the delivery of Na+ to the late distal tubule and collecting duct, a situation that is often associated with increased K+ and H + excretion. It is likely that the elevation in luminal Na+ concentration in the distal nephron induced by such diuretics augments depolarization of the luminal membrane and thereby enhances the lumen-negative VT , which facilitates K+ excretion. In addition to principal cells, the collecting duct also contains type A intercalated cells that mediate the secretion of H+ into the tubular lumen. Tubular acidification is driven by a luminal H+ATPase (proton pump), and this pump is aided by the lumen-negative transepithelial voltage. However, increased distal delivery of Na+ is not the only mechanism by which diuretics increase K+ and H+ excretion. Activation of the reninangiotensinaldosterone axis by diuretics also contributes to diuretic-induced K+ and H+ excretion by a mechanism explained in the section on mineralocorticoid antagonists. Figure 298. Na Reabsorption in Late Distal Tubule and Collecting Duct and Mechanism of Diuretic Action of Epithelial Na+-Channel Inhibitors. Cl reabsorption (not shown) occurs both paracellularly and transcellularly, and precise mechanism of Cl transport appears to be species specific. A, antiporter; CH, ion channel; CA, carbonic anhydrase. Numbers in parentheses indicate stoichiometry. Designated voltages are the potential differences across the indicated membrane or cell.
+

Considerable evidence indicates that amiloride blocks Na+ channels in the luminal membrane of principal cells in the late distal tubule and collecting duct. This evidence includes data from epithelia of nonrenal origin (amphibian skin and toad bladder) (Garty and Palmer, 1997) as well as a number of electrophysiological studies in isolated mammalian collecting ducts (O'Neil and Boulpaep, 1979). Amiloride produces half-maximal inhibition at concentrations <1 mM and, depending on the study, amiloride may interact with Na+ in the channel either competitively or noncompetitively. It is important, however, to bear in mind that renal epithelial Na+ channels inhibited by this class of diuretics are not the same as voltage-gated Na + channels found in many cell types (for instance, neurons and myocytes). Molecular cloning studies have revealed that the amiloride-sensitive Na + channel (called ENaC) consists of three subunits ( , , ) (Canessa et al., 1994). Although the subunit is sufficient for channel activity, maximal Na+ permeability is induced when all three subunits are coexpressed in the same cell, suggesting a minimal oligomeric structure in which one copy of each subunit is associated in a heterotrimeric protein. Studies in Xenopus oocytes expressing ENaC suggest that triamterene and amiloride bind to ENaC by similar mechanisms (Busch et al., 1996). Liddle's syndrome (pseudohyperaldosteronism) is an autosomal dominant form of low-renin, volumeexpanded hypertension that is due to mutations in the or subunits, leading to increased basal activity of ENaC (Ismailov et al., 1999). Effects on Urinary Excretion Since the late distal tubule and collecting duct have a limited capacity to reabsorb solutes, blockade of Na+ channels in this part of the nephron results in only a mild increase in the excretion rates of

Na+ and Cl (approximately 2% of filtered load). Blockade of Na+ channels hyperpolarizes the luminal membrane, reducing the lumen-negative transepithelial voltage. Since the lumen-negative potential difference normally opposes cation reabsorption and facilitates cation secretion, attenuation of the lumen-negative voltage decreases the excretion rates of K+ , H+, Ca2+, and Mg2+. Volume contraction may increase reabsorption of uric acid in the proximal tubule; hence, chronic administration of amiloride and triamterene may decrease uric acid excretion. Effects on Renal Hemodynamics Amiloride and triamterene have little or no effect on renal hemodynamics and do not alter TGF. Other Actions Amiloride, at concentrations higher than needed to elicit therapeutic effects, also blocks the Na +H+ and Na+Ca2+ antiporters and inhibits the Na + pump. Absorption and Elimination The relative potency, oral bioavailability, plasma half-life, and route of elimination for amiloride and triamterene are listed in Table 296. Amiloride is eliminated predominantly by urinary excretion of intact drug. Triamterene is extensively metabolized to an active metabolite, 4hydroxytriamterene sulfate, and this metabolite is excreted in the urine. The pharmacological activity of 4-hydroxytriamterene sulfate is comparable to that of the parent drug. Therefore, the toxicity of triamterene may be enhanced in both hepatic disease (decreased metabolism of triamterene) and renal failure (decreased urinary excretion of active metabolite). Toxicity, Adverse Effects, Contraindications, Drug Interactions The most dangerous adverse effect of Na+-channel inhibitors is hyperkalemia, which can be lifethreatening. Consequently, amiloride and triamterene are contraindicated in patients with hyperkalemia as well as in patients at increased risk of developing hyperkalemia (e.g., patients with + renal failure, patients receiving other K -sparing diuretics, patients taking angiotensin converting + enzyme inhibitors, or patients taking K supplements). Even NSAIDs can increase the likelihood of + hyperkalemia in patients receiving Na -channel inhibitors. Cirrhotic patients are prone to megaloblastosis because of folic acid deficiency, and triamterene, a weak folic acid antagonist, may increase the likelihood of this adverse event. Triamterene also can reduce glucose tolerance and induce photosensitization and has been associated with interstitial nephritis and renal stones. Both drugs can cause CNS, gastrointestinal, musculoskeletal, dermatological, and hematological adverse effects. The most common adverse effects of amiloride are nausea, vomiting, diarrhea, and headache; those of triamterene are nausea, vomiting, leg cramps, and dizziness. Therapeutic Uses Because of the mild natriuresis induced by Na+-channel inhibitors, these drugs seldom are used as sole agents in the treatment of edema or hypertension. Rather, their major utility is in combination with other diuretics. Coadministration of a Na+ -channel inhibitor augments the diuretic and antihypertensive response to thiazide or loop diuretics. More importantly, the ability of Na+-channel inhibitors to reduce K+ excretion tends to offset the kaliuretic effects of thiazide and loop diuretics; + consequently, the combination of a Na -channel inhibitor with a thiazide or loop diuretic tends to + result in normal values of plasma K (Hollenberg and Mickiewicz, 1989). Liddle's syndrome can be + treated effectively with Na -channel inhibitors. Aerosolized amiloride has been shown to improve

mucociliary clearance in patients with cystic fibrosis (Zahaykevich, 1991). By inhibiting Na + absorption from the surface of airway epithelial cells, amiloride augments hydration of respiratory secretions and thereby improves mucociliary clearance. Amiloride also is useful for lithium-induced nephrogenic diabetes insipidus because it blocks Li+ transport into the cells of the collecting tubules. Antagonists of Mineralocorticoid Receptors (Aldosterone Antagonists; K+ -Sparing Diuretics) Mineralocorticoids cause retention of salt and water and increase the excretion of K+ and H + by binding to specific mineralocorticoid receptors. Kagawa et al. (1957) observed that some spirolactones block the effects of mineralocorticoids; this finding led to the synthesis of specific antagonists for the mineralocorticoid receptor (MR). Spironolactone (ALDACTONE), a 17spirolactone, is the only member of this class available in the United States (Table 297). Mechanism of Action Epithelial cells in the late distal tubule and collecting duct contain cytoplasmic MRs that have a high affinity for aldosterone. This receptor is a member of the superfamily of receptors for steroid hormones, thyroid hormones, vitamin D, and retinoids (seeChapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect). Aldosterone enters the epithelial cell from the basolateral membrane and binds to MRs; the MR aldosterone complex translocates to the nucleus, where it binds to specific sequences of DNA (hormone-responsive elements) and thereby regulates the expression of multiple gene products called aldosterone-induced proteins (AIPs). Figure 299 illustrates some of the proposed effects of AIPs, including: activation of "silent" Na + channels and "silent" Na+ pumps that preexist in the cell membrane; alterations in the cycling of Na+ channels and Na+ pumps between the cytosol and cell membrane so that more channels and pumps are located in the membrane; increased expression of Na+ channels and Na + pumps; changes in permeability of the tight junctions; and increased activity of enzymes in the mitochondria that are involved in ATP production. The precise mechanisms by which AIPs alter transport are incompletely understood. However, the net effect of AIPs is to increase Na+ conductance of the luminal membrane and sodium pump activity of the basolateral membrane. Consequently, transepithelial NaCl transport is enhanced and the lumen-negative transepithelial voltage is increased. The latter effect increases the driving force for secretion of K+ and H+ into the tubular lumen. Figure 299. Effects of Aldosterone on Late Distal Tubule and Collecting Duct and Diuretic Mechanism of Aldosterone Antagonists. AIP, aldosterone-induced proteins; ALDO, aldosterone; MR, mineralocorticoid receptor; CH, ion channel; + + 1, activation of membrane-bound Na channels; 2, redistribution of Na channels + from cytosol to membrane; 3, de novo synthesis of Na channels; 4, activation of membrane-bound Na+,K+ATPase; 5, redistribution of Na+,K+ ATPase from + + cytosol to membrane; 6, de novo synthesis of Na ,K ATPase; 7, changes in permeability of tight junctions; 8, increased mitochondrial production of ATP.

Drugs such as spironolactone competitively inhibit the binding of aldosterone to the MR (Marver et al., 1974). Unlike the MRaldosterone complex, the MRspironolactone complex is not able to induce the synthesis of AIPs. Since spironolactone and other drugs in this class block the biological effects of aldosterone, these agents also are referred to as aldosterone antagonists. Effects on Urinary Excretion The effects of spironolactone on urinary excretion are very similar to those induced by renal epithelial Na+-channel inhibitors. However, unlike that of the Na+-channel inhibitors, the clinical efficacy of spironolactone is a function of endogenous levels of aldosterone. The higher the levels of endogenous aldosterone, the greater the effects of spironolactone on urinary excretion. Effects on Renal Hemodynamics Spironolactone has little or no effect on renal hemodynamics and does not alter TGF. Other Actions High concentrations of spironolactone have been reported to interfere with steroid biosynthesis by inhibiting 11 - and 18-, 21-, and 17 -hydroxylase. These effects have limited clinical relevance (seeChapter 60: Adrenocorticotropic Hormone; Adrenocortical Steroids and Their Synthetic Analogs; Inhibitors of the Synthesis and Actions of Adrenocortical Hormones).

Absorption and Elimination Spironolactone is partially absorbed (approximately 65%), is extensively metabolized (even during its first passage through the liver), undergoes enterohepatic recirculation, is highly protein-bound, and has a short half-life (approximately 1.6 hours). However, an active metabolite of spironolactone, canrenone, has a half-life of approximately 16.5 hours, which prolongs the biological effects of spironolactone. Although not available in the United States, canrenone and the K+ salt of canrenoate also are in clinical use. Canrenoate is not active per se but is converted to canrenone in the body. MR antagonists are the only diuretics that do not require access to the tubular lumen to induce a diuresis. Toxicity, Adverse Effects, Contraindications, Drug Interactions As with other K+-sparing diuretics, spironolactone may cause life-threatening hyperkalemia. Therefore, spironolactone is contraindicated in patients with hyperkalemia and in patients at increased risk of developing hyperkalemia, either because of disease or because of administration of other medications. Spironolactone also can induce metabolic acidosis in cirrhotic patients. Salicylates may reduce the tubular secretion of canrenone and decrease the diuretic efficacy of spironolactone, and spironolactone may alter the clearance of digitalis glycosides. Due to its steroid structure, spironolactone may cause gynecomastia, impotence, decreased libido, hirsutism, deepening of the voice, and menstrual irregularities. Spironolactone also may induce diarrhea, gastritis, gastric bleeding, and peptic ulcers (the drug is contraindicated in patients with peptic ulcers). CNS adverse effects include drowsiness, lethargy, ataxia, confusion, and headache. Spironolactone may cause skin rashes and, rarely, blood dyscrasias. Breast cancer has occurred in patients taking spironolactone chronically (cause and effect not established), and high doses of spironolactone have been associated with malignant tumors in rats. Whether or not therapeutic doses of spironolactone can induce malignancies remains an open question. Therapeutic Uses As with other K+-sparing diuretics, spironolactone often is coadministered with thiazide or loop diuretics in the treatment of edema and hypertension. Such combinations result in increased mobilization of edema fluid while causing lesser perturbations of K+ homeostasis. Spironolactone is particularly useful in the treatment of primary hyperaldosteronism (adrenal adenomas or bilateral adrenal hyperplasia) and of refractory edema associated with secondary aldosteronism (cardiac failure, hepatic cirrhosis, nephrotic syndrome, severe ascites). Spironolactone is considered the diuretic of choice in patients with hepatic cirrhosis. Pitt et al. (1999) have reported that spironolactone, when added to standard therapy, substantially reduces morbidity and mortality in patients with New York Heart Association (NYHA) class III and class IV heart failure (seeChapter 34: Pharmacological Treatment of Heart Failure). Mechanisms of Edema Formation and the Role of Diuretics in Clinical Medicine Mechanism of Edema Formation A complex set of interrelationships (Figure 2910) exists among the cardiovascular system, the kidneys, the CNS (Na+ appetite, thirst regulation), and the tissue capillary beds [distribution of extracellular fluid volume (ECFV)], so that perturbations at one of these sites can affect all of the remaining sites. A primary law of the kidney is that Na+ excretion is a steep function of mean arterial blood pressure (MABP) such that small increases in MABP cause marked increases in Na+ excretion (Guyton, 1991). Over any given time interval, the net change in total body Na+ (either

positive or negative) is simply the dietary Na+ intake minus the urinary excretion rate minus other losses (e.g., sweating, fecal losses, vomiting). When a net positive Na+ balance occurs, the concentration of Na+ in the ECF will increase, stimulating water intake (thirst) and reducing urinary water output (via ADH release). Opposite changes occur during a net negative Na+ balance. Changes in water intake and output adjust ECFV concentration toward normal, thereby expanding or contracting total ECFV. Total ECFV is distributed among many body compartments; however, since the volume of extracellular fluid on the arterial side of the circulation pressurizes the arterial tree, it is this fraction of ECFV that determines MABP, and it is this fraction of ECFV that is "sensed" by the cardiovascular system and kidneys. Since MABP is a major determinant of Na+ output, a closed loop is established (Figure 2910). This loop cycles until net Na+ accumulation is zero; i.e., in the long run, Na+ intake must equal Na+ loss. Figure 2910. Interrelationships among Renal Function, Na Intake, Water Homeostasis, Distribution of Extracellular Fluid Volume (ECFV), and Mean Arterial Blood Pressure (MABP). Pathophysiological mechanisms of edema formation: 1, rightward shift of renalpressure natriuresis curve; 2, excessive dietary Na+ intake; 3, increased distribution of ECFV to peritoneal cavity (e.g., liver cirrhosis with increased hepatic sinusoidal hydrostatic pressure) leading to ascites formation; 4, increased distribution of ECFV to lungs (e.g., left heart failure with increased pulmonary capillary hydrostatic pressure) leading to pulmonary edema; 5, increased distribution of ECFV to venous circulation (e.g., right heart failure) leading to venous congestion; 6, peripheral edema caused by altered Starling forces causing increased distribution of ECFV to interstitial space (e.g., diminished plasma proteins in nephrotic syndrome, severe burns, liver disease).
+

The above discussion implies that three fundamental types of perturbations contribute to venous congestion and/or edema formation: (1) A shift to the right in the renal pressurenatriuresis relationship (e.g., chronic renal failure) causes reduced Na+ excretion for any level of MABP. If all other factors remain constant, this would increase total body Na +, ECFV, and MABP. The additional ECFV would be distributed throughout various body compartments, according to the state of cardiac function and prevailing Starling forces, and would predispose toward venous congestion and/or edema. Even so, in the absence of any other predisposing factors for venous congestion and/or edema, a rightward shift in the renal pressurenatriuresis curve generally causes hypertension with only a slight (usually immeasurable) increase in ECFV. As elucidated by Guyton and coworkers (Guyton, 1991), ECFV expansion triggers the following series of events: expanded ECFV augmented cardiac output enhanced vascular tone (i.e., total body autoregulation) increased total peripheral resistance elevated MABP pressure natriuresis reduction of ECFV and cardiac output toward normal. Most likely, a sustained rightward shift in the renal pressure natriuresis curve is a necessary and sufficient condition for long-term hypertension but is only a predisposing factor for venous congestion and/or edema. (2) An increase in dietary Na+ intake would have the same effects as a rightward shift in the renal pressurenatriuresis relationship (i.e., increased MABP and predisposition to venous congestion/edema). However, changes in salt intake may have minimal or large effects depending on the shape of the patient's renal pressurenatriuresis curve. (3) Any pathophysiological alterations in the forces that govern the distribution of ECFV among the various body compartments would cause abnormal amounts of ECFV to be trapped at the site of altered forces. This would deplete the "sensed" ECFV, which would be restored back to normal by the mechanisms described in the preceding paragraph. ECFV may be trapped at several sites by different mechanisms. For instance, cirrhosis of the liver increases lymph in the space of Disse, leading to spillover via the glissonian wall into the peritoneal cavity (ascites). Left heart failure, both acute and chronic, increases hydrostatic pressure in the lung capillaries, leading to pulmonary edema. Chronic right heart failure redistributes ECFV from the arterial to the venous circulation, resulting in venous, hepatic, and splenic congestion and peripheral tissue edema. Decreased levels of plasma protein, particularly albumin (e.g., in nephrotic syndrome, severe burns, hepatic disease), increase the distribution of ECFV into the interstitial spaces, causing generalized peripheral edema. Peripheral edema also may be "idiopathic" due to unknown alterations in the Starling forces at the capillary bed. The Role of Diuretics in Clinical Medicine Another implication of the mechanisms illustrated in Figure 2910 is that three fundamental + strategies exist for mobilizing edema fluidcorrect the underlying disease, restrict Na intake, or administer diuretics. The most desirable course of action would be to correct the primary disease; however, this is often impossible. For instance, the increased hepatic sinusoidal pressure in cirrhosis of the liver and the urinary loss of protein in nephrotic syndrome are due to structural alterations in + the portal circulation and glomeruli, respectively, which may not be remediable. Restriction of Na intake is the favored nonpharmacological approach for the treatment of edema and hypertension and should usually be attempted; however, compliance is a major obstacle. Diuretics, therefore, remain the cornerstone for the treatment of edema or volume overload, particularly that due to congestive heart failure, ascites, chronic renal failure, and nephrotic syndrome. With regard to heart failure, diuretics reduce pulmonary edema and venous congestion, and it may be possible to manage mild cardiac failure with diuretics alone. However, most patients ultimately will require additional therapy with digitalis and/or angiotensin converting enzyme inhibitors (Chapter 34: Pharmacological Treatment of Heart Failure). Periodic administration of diuretics to cirrhotic patients with ascites may eliminate the necessity for or reduce the interval

between paracenteses, adding to patient comfort and sparing protein reserves that are lost by paracenteses. Although diuretics can reduce the edema associated with chronic renal failure, increased doses of the more powerful loop diuretics may be required. In the nephrotic syndrome, the response to diuretics often is disappointing. Whether a patient should receive diuretics and, if so, what therapeutic regimen should be used (i.e., type of diuretic, route of administration, and speed of mobilization of edema fluid) depends on the clinical situation. Massive pulmonary edema in patients with acute left heart failure is a medical emergency requiring rapid, aggressive therapy including intravenous administration of a loop diuretic. In this setting, use of oral diuretics or diuretics with lesser efficacy is inappropriate. On the other hand, mild pulmonary and venous congestion associated with chronic heart failure is best treated with an oral loop diuretic, the dosage of which should be titrated carefully to maximize the benefit-to-risk ratio. In many situations, edema will not pose an immediate health risk. Even so, uncomfortable, oppressive, and/or disfiguring edema can greatly reduce quality of life, and the decision to treat will be based in part on quality-of-life issues. In such cases, only partial removal of edema fluid should be attempted, and the fluid should be mobilized slowly using a diuretic regimen that accomplishes the task with minimal perturbation of normal physiology. Brater (1998) has provided an algorithm for diuretic therapy (specific recommendations for drug, dose, route, and drug combinations) in patients with edema caused by renal, hepatic, or cardiac disorders. In many clinical situations, edema is not caused by an abnormal intake of Na+ or by an altered renal handling of Na +. Rather, edema is the result of altered Starling forces at the capillary beds, i.e., a "Starling trap." Use of diuretics in these clinical settings represents a judicious compromise between the edematous state and the hypovolemic state. In such conditions, reducing total ECFV with diuretics will decrease edema but also will cause depletion of "sensed" ECFV, possibly leading to hypotension, malaise, and asthenia. Diuretic resistance refers to edema that is or has become refractory to a given diuretic. If diuretic resistance develops against a less efficacious diuretic, a more efficacious diuretic should be substituted, e.g., a loop diuretic for a thiazide. However, resistance to loop diuretics is not uncommon and can be due to several causes (Brater, 1985). NSAIDs block prostaglandin-mediated increases in RBF, resulting in resistance to loop diuretics. In chronic renal failure, a reduction in RBF decreases the delivery of diuretics to the kidney, and accumulation of endogenous organic acids competes with loop diuretics for transport at the proximal tubule. Consequently, the concentration of diuretic at the active site in the tubular lumen is diminished. In nephrotic syndrome, urinary protein binds diuretics and thereby limits the response. In hepatic cirrhosis or heart failure, the kidney may have a diminished responsiveness to diuretics because of increased + + proximal tubular Na reabsorption, leading to diminished delivery of Na to the distal nephron segments (Knauf and Mutschler, 1997). Faced with resistance to loop diuretics, the clinician has several options: 1. Bed rest may restore drug responsiveness by improving the renal circulation. 2. An increase in the dose of loop diuretic may restore responsiveness. 3. Administration of smaller doses more frequently or a continuous intravenous infusion of a loop diuretic (Rudy et al., 1991; Dormans et al., 1996; Ferguson et al., 1997) will increase the length of time that an effective concentration of the diuretic is at the active site. 4. Use of combination therapy to sequentially block more than one site in the nephron may result in a synergistic interaction between two diuretics. For instance, a combination of a loop diuretic + with a K -sparing or a thiazide diuretic may improve therapeutic response; however, nothing is gained by the administration of two drugs of the same type. Thiazide diuretics with significant

proximal tubular effects, e.g., metolazone, are particularly well suited for sequential blockade when coadministered with a loop diuretic. 5. Scheduling of diuretic administration shortly before food intake will provide effective concentrations of diuretic in the tubular lumen when the salt load is highest. Prospectus All currently available diuretics perturb K+ homeostasis. However, studies in animals have established that blockade of adenosine A1 receptors induces a brisk natriuresis without significantly increasing urinary K+ excretion (Kuan et al., 1993). Two clinical studies with FK453, a highly selective A1-receptor antagonist, confirm that blockade of A 1 receptors induces natriuresis in human beings with minimal effects on K+ excretion (Balakrishnan et al., 1993; van Buren et al., 1993). The natriuretic mechanism of this novel class of diuretics has been partially elucidated (Takeda et al., 1993). Endogenous adenosine acts on A1 receptors in the proximal tubule to inhibit adenylyl cyclase. As cyclic AMP inhibits basolateral Na+ HCO3 symport activity, a reduced level of cyclic AMP increases Na+ HCO3 symport activity in the basolateral membrane. Blockade of A 1 receptors prevents the inhibition of adenylyl cyclase by endogenous adenosine, increases epithelial cyclic AMP levels, and consequently decreases Na+HCO3 symport activity in the basolateral membrane of the proximal tubule. Because A1 receptors are involved in TGF, A1-receptor antagonists uncouple increased distal delivery of Na+ from activation of TGF (Wilcox et al., 1999). Other mechanisms, including an effect in the collecting tubules, contribute to the natriuretic response to A1-receptor antagonists; however, it is not known why this class of diuretics has little effect on K + excretion. An A1-receptor antagonist is under clinical development for the treatment of edema due to heart failure. Recently, the water channels of the proximal tubule (aquaporin 1) and of the collecting duct (aquaporins 2, 3, and 4) were cloned and their functional characteristics examined (Agre et al., 1993; Fushimi et al., 1993; seeYamamoto and Sasaki, 1998). The cloning of these proteins represents an important step in our understanding of water homeostasis. Currently, there are no specific inhibitors of aquaporins; however, these proteins are important targets for the development of novel diuretics. Inhibition of water channels in the proximal tubule would greatly diminish the flux of water across proximal tubular epithelial cells, which would reduce luminal Na+ concentration to the point where Na+ reabsorption would cease. Therefore, inhibitors of proximal tubular aquaporin 1 may be useful natriuretic diuretics. On the other hand, inhibitors of collecting duct aquaporins 2, 3, and 4 would prevent water reabsorption in the collecting duct and therefore would be highly efficacious "aquaretic" diuretics, i.e., diuretics with a predominant effect on water, + rather than on Na , excretion. Since aquaporin 2 is regulated by the vasopressin V2 receptor, another potential class of aquaretic diuretics would be nonpeptide, orally active V2-receptor antagonists. Exciting progress has been made, and this subject is developed in Chapter 30: Vasopressin and + Other Agents Affecting the Renal Conservation of Water. Finally, apical K channels (ROMK) are + another potential molecular target for the development of K -sparing diuretics with high efficacy.

Chapter 30. Vasopressin and Other Agents Affecting the Renal Conservation of Water
Overview Precise regulation of body fluid osmolality is essential. It is controlled by a finely tuned, intricate

homeostatic mechanism that operates by adjusting both the rate of water intake and the rate of solute-free water excretion by the kidneysi.e., water balance. Abnormalities in this homeostatic system can be caused by genetic diseases, acquired diseases, or drugs and may result in serious and potentially life-threatening deviations in plasma osmolality. The goals of this chapter are to describe the physiological mechanisms that regulate plasma osmolality, to discuss the diseases that perturb those mechanisms, and to examine pharmacological approaches for treating disorders of water balance. Argininevasopressin (the antidiuretic hormone in human beings) is the main hormone involved in regulation of body fluid osmolality. Many diseases of water homeostasis and many pharmacological strategies for correcting such disorders pertain to vasopressin. Accordingly, vasopressin is the major focus of this chapter and is discussed with regard to: (1) chemistry (including the chemistry of vasopressin agonists and antagonists); (2) physiology (including anatomical considerations; the synthesis, transport, and storage of vasopressin and the regulation of vasopressin secretion); (3) basic pharmacology (including vasopressin receptors and their signal transduction pathways, renal actions of vasopressin, pharmacological modification of the antidiuretic response to vasopressin, and nonrenal actions of vasopressin); (4) diseases affecting the vasopressin system (diabetes insipidus, syndrome of inappropriate secretion of antidiuretic hormone, and other water-retaining states); and (5) clinical pharmacology of vasopressin peptides (therapeutic uses, pharmacokinetics, toxicities, adverse effects, contraindications, and drug interactions). A small number of other drugs can be used to treat abnormalities of water balance; a discussion of these agents is integrated into the section on diseases affecting the vasopressin system. Introduction to Vasopressin Immunoreactive vasopressin has been observed in neurons from organisms belonging to the first animal phylum with a nervous system (e.g., Hydra attenuata), and vasopressin-like peptides have been isolated and characterized from both mammalian and nonmammalian vertebrates as well as from invertebrates (Table 301). Genes encoding vasopressin-like peptides probably evolved more than 700 million years ago. With the emergence of life on land, vasopressin became the mediator of a remarkable regulatory system for the conservation of water. The hormone is released by the posterior pituitary whenever water deprivation causes an increased plasma osmolality or whenever the cardiovascular system is challenged by hypovolemia and/or hypotension. In amphibians, the target organs for vasopressin are skin and the urinary bladder, whereas in other vertebrates, including human beings, the site of action is primarily the renal collecting duct. In each of these target tissues, vasopressin acts by increasing the permeability of the cell membrane to water, thus permitting water to move passively down an osmotic gradient across skin, bladder, or collecting duct into the extracellular compartment. In view of the long evolutionary history of vasopressin, it is not surprising that vasopressin acts at sites in the nephron other than the collecting duct and on tissues other than the kidney. Vasopressin is a potent vasopressor, and its name was originally chosen in recognition of its vasoconstrictor action. Vasopressin is a neurotransmitter; among its actions in the central nervous system (CNS) are apparent roles in the secretion of adrenocorticotropic hormone (ACTH) and in the regulation of the cardiovascular system, temperature, and other visceral functions. Vasopressin also promotes the release of coagulation factors by the vascular endothelium and increases platelet aggregability; therefore, it may play a role in hemostasis. Chemistry of Vasopressin Receptor Agonists and Antagonists

Chemistry of Vasopressin Receptor Agonists du Vigneaud and coworkers (1954) determined the structures of vasopressin and oxytocin and accomplished the complete synthesis of each. A number of vasopressin-like peptides occur naturally (Table 301). All are nonapeptides; contain cysteine residues in positions 1 and 6; have an intramolecular disulfide bridge between the two cysteine residues (essential for agonist activity); have additional conserved amino acids in positions 5, 7, and 9 (asparagine, proline, and glycine, respectively); contain a basic amino acid in position 8; and are amidated on the carboxyl terminus. In all mammals except swine, the neurohypophyseal peptide is 8-arginine vasopressin, and the terms vasopressin, arginine vasopressin (AVP), and antidiuretic hormone (ADH) are used interchangeably. The chemical structure of oxytocin is closely related to that of vasopressin, i.e., oxytocin is [Ile3, Leu8]AVP. Oxytocin binds to specific oxytocin receptors on myoepithelial cells in the mammary gland and on smooth muscle cells in the uterus, causing milk ejection and uterine contraction, respectively. Inasmuch as vasopressin and oxytocin are structurally similar, it is not surprising that vasopressin and oxytocin agonists and antagonists can bind to each other's receptors. Therefore, most of the available peptide vasopressin agonists and antagonists have some affinity for oxytocin receptors; at high doses, they may block or mimic the effects of oxytocin (Manning and Sawyer, 1989). With the advent of solid-phase peptide synthesis, many vasopressin analogs were synthesized with the goal of increasing duration of action and selectivity for vasopressin receptor subtypes (V 1versus V2 vasopressin receptors, which mediate pressor responses and antidiuretic responses, respectively). In 1967, Zaoral and coworkers announced the synthesis of desmopressin: 1-deamino-8-D-arginine vasopressin (DDAVP) (Table 301). Deamination at position 1 increases duration of action and increases antidiuretic activity without increasing vasopressor activity. Substitution of D-arginine for L-arginine greatly reduces vasopressor activity without reducing antidiuretic activity. Thus, the antidiuretic-to-vasopressor ratio for desmopressin is approximately 3000-fold greater than that for vasopressin, and desmopressin now is the preferred drug for the treatment of central diabetes insipidus (Robinson, 1976). Substitution of valine for glutamine in position 4 further increases the 4 8 antidiuretic selectivity, and the antidiuretic-to-vasopressor ratio for deamino[Val , D-Arg ] AVP (Table 301) is approximately 11,000-fold greater than that for vasopressin. Increasing V1 selectivity has proved more difficult than increasing V2 selectivity (Thibonnier, 1990). However, a limited number of agonists have been developed with modest selectivity for V1 receptors (seeTable 301). Vasopressin receptors in the adenohypophysis that mediate vasopressin-induced ACTH release are neither classical V1 nor V2 receptors. Since the vasopressin receptors in the adenohypophysis appear to share a common signal transduction mechanism with classical V1 receptors, and since many vasopressin analogs with vasoconstrictor activity release ACTH, V1 receptors have been subclassified into V1a (vascular/hepatic) and V1b (pituitary) receptors (Jard et al., 1986). Vasopressin agonists selective for both V1a (Thibonnier, 1990) and V1b receptors (Schwartz et al., 1991) have been described (seeTable 301). Chemistry of Vasopressin Receptor Antagonists The impetus for the development of specific vasopressin receptor antagonists is the belief that such drugs may be useful in a number of clinical settings. Selective V1a antagonists may be beneficial when total peripheral resistance is increased (e.g., congestive heart failure and hypertension); selective V2 antagonists could be useful whenever reabsorption of solute-free water is excessive (e.g., the syndrome of inappropriate secretion of antidiuretic hormone and hyponatremia associated with a reduced effective blood volume). Combined V1a/V2 receptor antagonists might be beneficial

in diseases associated with a combination of increased peripheral resistance and dilutional hyponatremia (e.g., congestive heart failure). Shortly after the synthesis of vasopressin, du Vigneaud and coworkers began designing antagonists of vasopressin's pharmacological effects. Since that time, numerous vasopressin receptor antagonists have been synthesized (Manning et al., 1993; Lszlet al., 1991). Highly selective V1 and V2 peptide antagonists that are structural analogs of vasopressin have been synthesized (seeTable 302 for examples), including both cyclic and linear peptides. [1-( -mercapto- , cyclopentamethyleneproprionic acid),2-O-methyltyrosine]. Arginine vasopressin, also known as d(CH2)5[Tyr(Me) 2]AVP, has a greater affinity for V1a receptors than for either V1b or V2 receptors; this antagonist has been widely employed in physiological and pharmacological studies (Manning and Sawyer, 1989). Although [1-deaminopenicillamine, 2-O-methyltyrosine] arginine vasopressin, also called dP[Tyr(Me)2]AVP, is a potent V1b receptor antagonist with little affinity for the V2 receptor, it also blocks V1a receptors. No truly selective V1b receptor antagonist currently is available. Peptide antagonists currently available have limited oral activity, and the potency of peptide V2 antagonists is species-dependent. Also, with prolonged infusion, peptide V2 antagonists appear to express significant agonist activity (Kinter et al., 1993). However, in the early 1990s a nonpeptide, orally active V1-selective antagonist (OPC-21268) (Yamamura et al., 1991) and a nonpeptide, orally active V1a -/V 2-selective antagonist (OPC-31260) (Yamamura et al., 1992) were synthesized. Neither compound had partial agonist activity. Other highly potent nonpeptide vasopressin receptor antagonists, such as SR 49059, SR 121463A, VPA-985, and YM 087 (Mayinger and Hensen, 1999) have been synthesized (seeTable 302). Anatomy The antidiuretic mechanism in mammals involves two anatomical components: a CNS component for the synthesis, transport, storage, and release of vasopressin, and a renal collecting duct system composed of epithelial cells that respond to vasopressin by increasing their permeability to water. The relevant anatomy of the renal collecting duct system is described in Chapter 29: Diuretics. The CNS component of the antidiuretic mechanism is called the hypothalamiconeurohypophyseal system and consists of neurosecretory neurons with perikarya located predominantly in two specific hypothalamic nuclei, the supraoptic nucleus (SON) and the paraventricular nucleus (PVN). The long axons of neurons in the SON and PVN traverse the supraopticohypophyseal tract to terminate in the median eminence and pars nervosa of the posterior pituitary. Synthesis Vasopressin and oxytocin are synthesized in the perikarya of magnocellular neurons in the SON and PVN; the two hormones are synthesized predominantly in separate neurons. Vasopressin synthesis appears to be regulated solely at the transcriptional level (Robinson and Fitzsimmons, 1993), and the molecular mechanism of vasopressin synthesis has been elucidated in considerable detail (Archer, 1993). In human beings, a 168-amino-acid preprohormone (Figure 301) is synthesized, and a signal peptide (residues -23 to -1) assures incorporation of the nascent polypeptide into ribosomes. During synthesis, the signal peptide is removed to form the vasopressin prohormone, and vesicle-mediated translocations maneuver the prohormone through the rough endoplasmic reticulum and cis-, medial-, and trans-Golgi compartments, so that the prohormone emerges incorporated into large (0.1- to 0.3-micron) membrane-enclosed granules. The prohormone consists of three domains: vasopressin (residues 1 to 9), vasopressin (VP)neurophysin (residues 13 to 105), and VPglycopeptide (residues 107 to 145). The vasopressin domain is linked to the VP neurophysin domain through a glycinelysinearginine processing signal, and the VPneurophysin domain is linked to the VPglycopeptide domain by an arginine-processing signal. In the secretory

granules, an endopeptidase, exopeptidase, monooxygenase, and lyase act sequentially on the prohormone to produce vasopressin, VPneurophysin (sometimes referred to as neurophysin II or MSELneurophysin), and VPglycopeptide (sometimes called copeptin). The synthesis and transport of vasopressin are dependent on the conformation of the preprohormone. In particular, VPneurophysin binds vasopressin and is critical to the correct processing, transport, and storage of vasopressin (Breslow, 1993). Genetic mutations in either the signal peptide or VPneurophysin give rise to central diabetes insipidus (Raymond, 1994). Figure 301. Processing of the 168Amino Acid Human 8-Arginine Vasopressin (AVP) Preprohormone to AVP, Vasopressin (VP)Neurophysin, and VP Glycopeptide.

Transport and Storage The process of axonal transport of vasopressin-containing granules is rapid, and newly synthesized neurohypophyseal hormones arrive at the posterior lobe within 30 minutes of a stimulus. The axons involved in transport of granules have two destinations, carrying vasopressin not only to classical storage sites in the neurohypophysis but also to the external zone of the median eminence, where vasopressin enters the adenohypophyseal portal circulation and plays a role as a corticotropinreleasing factor. Maximal release of vasopressin occurs when impulse frequency is approximately 12 spikes per second for 20 seconds. Higher frequencies or longer periods of stimulation lead to diminished hormone release (fatigue). Appropriately, vasopressin-releasing cells demonstrate an atypical pattern of spike activity characterized by rapid phasic bursts (5 to 12 spikes per second for 15 to 60 seconds) separated by quiescent periods (15 to 60 seconds in duration). This pattern is orchestrated

by activation and inactivation of ion channels in the magnocellular neurons and provides for optimal release of vasopressin (Leng et al., 1992). Vasopressin Synthesis Outside of the CNS Vasopressin also is synthesized by the heart (Hupf et al., 1999) and adrenal gland (Guillon et al., 1998). In the heart, elevated wall stress increases vasopressin synthesis several-fold. Cardiac synthesis of vasopressin is predominantly vascular and perivascular and may contribute to impaired ventricular relaxation and coronary vasoconstriction. Vasopressin synthesis in the adrenal medulla stimulates catecholamine secretion from chromaffin cells and may promote adrenal cortical growth and stimulate aldosterone synthesis. Regulation of Vasopressin Secretion An increase in plasma osmolality is the principal physiological stimulus for vasopressin secretion. Severe hypovolemia/hypotension also is a powerful stimulus for vasopressin release. In addition, pain, nausea, and hypoxia can stimulate vasopressin secretion, and several endogenous hormones and pharmacological agents can modify vasopressin release. Hyperosmolality The relationship between plasma osmolality and plasma vasopressin concentration is shown in Figure 302A, and the relationship between plasma vasopressin levels and urine osmolality is illustrated in Figure 302B. The osmolality threshold for secretion is approximately 280 mOsm/kg. Below the threshold, vasopressin is barely detectable in plasma, and above the threshold, vasopressin levels are a steep and linear function of plasma osmolality. In fact, a 2% elevation in plasma osmolality causes a two- to threefold increase in plasma vasopressin levels. Therefore, a small increase in plasma osmolality leads to enhanced vasopressin secretion, which, in turn, causes increased solute-free water reabsorption (as evidenced by the increased urine osmolality). Increases in plasma osmolality (due to insensible water losses) above 290 mOsm/kg lead to an intense desire for water (thirst). Thus, the vasopressin system affords the organism longer thirst-free periods and, in the event that water is unavailable, allows the organism to survive longer periods of water deprivation. It is important to point out, however, that above a plasma osmolality of approximately 290 mOsm/kg, plasma levels of vasopressin exceed 5 pM. Since urinary concentration is maximal (about 1200 mOsm/kg) when vasopressin levels exceed 5 pM, further defense against hypertonicity is entirely dependent on water intake rather than on decreases in water loss. Figure 302. A. The relationship between plasma osmolality and plasma vasopressin levels. Plasma osmolality associated with thirst is indicated by arrow. B. The relationship between plasma vasopressin levels and urine osmolality. (From Robertson et al., 1977, and Kovacs and Robertson, 1992, with permission.)

Several CNS structures are involved in osmotic stimulation of vasopressin release; these structures are collectively referred to as the osmoreceptive complex. Although magnocellular neurons in the SON and PVN are osmosensitive, afferent inputs from other components of the osmoreceptive complex are required for a normal vasopressin response. The SON and PVN receive projections from the subfornical organ (SFO) and the organum vasculosum of the lamina terminalis (OVLT) either directly or indirectly [via the median preoptic nucleus (MnPO)]. Subgroups of neurons in the SFO, OVLT, and MnPO are either osmoreceptors or osmoresponders (i.e., are stimulated by osmoreceptive neurons located at other sites). Thus, a web of interconnecting neurons contributes to osmotically induced vasopressin secretion. Aquaporin 4, a water-selective channel, is associated with CNS structures involved in osmoregulation and may confer osmosensitivity. In the CNS, aquaporin 4 resides on glial and ependymal cells rather than neurons, suggesting that osmotic status may be communicated to neuronal cell by a glialneuron interaction (Wells, 1998). Hypovolemia and Hypotension Vasopressin secretion also is regulated hemodynamically by changes in effective blood volume and/or arterial blood pressure (Robertson, 1992). Reductions in effective blood volume and/or arterial blood pressure, regardless of the cause (e.g., hemorrhage, sodium depletion, diuretics, heart failure, hepatic cirrhosis with ascites, adrenal insufficiency, hypotensive drugs), may be associated with high circulating concentrations of vasopressin. However, unlike osmoregulation, hemodynamic regulation of vasopressin secretion is exponential, i.e., small decreases (5% to 10%) in blood volume and/or pressure have little effect on vasopressin secretion, whereas larger decreases (20% to 30%) can increase vasopressin levels to 20 to 30 times normal levels (exceeding the concentration of vasopressin required to induce maximal antidiuresis). Vasopressin is one of the most potent vasoconstrictors known, and the vasopressin response to hypovolemia or hypotension serves as a mechanism to stave off cardiovascular collapse during periods of severe blood loss and/or hypotension. Importantly, hemodynamic regulation of vasopressin secretion does not disrupt osmotic regulation; rather, hypovolemia/hypotension alters the setpoint and slope of the plasma osmolalityplasma vasopressin relationship (Figure 303).

Figure 303. Interactions between Osmolality and Hypovolemia/Hypotension. Numbers in circles refer to percentage increase (+) or decrease () in blood volume or arterial blood pressure. N indicates normal blood volume/blood pressure. (From Robertson, 1992, with permission.)

The neuronal pathways that mediate hemodynamic regulation of vasopressin release are completely different from those involved in osmoregulation. Baroreceptors in the left atrium, left ventricle, and pulmonary veins sense blood volume (filling pressures), and baroreceptors in the carotid sinus and aorta monitor arterial blood pressure. Nerve impulses reach brainstem nuclei predominantly through the vagus and glossopharyngeal nerves; these signals are relayed to the nucleus of the solitary tract, then to the A1-noradrenergic cell group in the caudal ventrolateral medulla, and finally to the SON and PVN (Cunningham and Sawchenko, 1991). Hormones and Neurotransmitters There is a large, sometimes contradictory, body of literature on modulation of vasopressin secretion by hormones and neurotransmitters (Renaud and Bourque, 1991). Vasopressin-synthesizing magnocellular neurons have a large array of receptors on both perikarya and nerve terminals; therefore, vasopressin release can be accentuated or attenuated by chemical agents acting at both ends of the magnocellular neuron. Also, hormones and neurotransmitters can modulate vasopressin secretion by stimulating or inhibiting neurons in nuclei that project, either directly or indirectly, to the SON and PVN. Because of these complexities, the results of any given investigation may depend critically on the route of administration of the agent and on the experimental paradigm. In many cases, the precise mechanism by which a given agent modulates vasopressin secretion is either unknown or controversial, and the physiological relevance of modulation of vasopressin secretion by most hormones and neurotransmitters is unclear. Nonetheless, several agents are known to stimulate vasopressin secretion, including acetylcholine (via nicotinic receptors), histamine (via H1 receptors), dopamine (via both D1 and D2 receptors), glutamine, aspartate, cholecystokinin, neuropeptide Y, substance P, vasoactive intestinal polypeptide, prostaglandins, and angiotensin II. Inhibitors of vasopressin secretion include atrial natriuretic peptide, gamma-aminobutyric acid, and opioids (particularly dynorphin via receptors). Of all the aforementioned hormones/neurotransmitters, angiotensin II has received the most attention (Phillips, 1987). Angiotensin II, when applied directly to magnocellular neurons in the SON and PVN, increases neuronal excitability; when applied to the MnPO, angiotensin II indirectly

stimulates magnocellular neurons in the SON and PVN. In addition, angiotensin II stimulates angiotensin-sensitive neurons in the OVLT and SFO (circumventricular nuclei lacking a blood brain barrier) that project to the SON/PVN. Thus, both angiotensin II synthesized in the brain and that formed in the circulation may stimulate vasopressin release. Inhibition of the conversion of angiotensin II to angiotensin III blocks angiotensin IIinduced vasopressin release, suggesting that angiotensin III is the main effector peptide of the brain reninangiotensin system controlling vasopressin release (Zini et al., 1996). Pharmacological Agents A number of drugs alter the osmolality of urine. It has been hypothesized that the action of many agents involves stimulation or inhibition of the secretion of vasopressin (Robertson, 1992). In some cases, the mechanism by which a drug alters vasopressin secretion involves direct effects on one or more CNS structures involved in the regulation of vasopressin secretion. In other cases, vasopressin secretion is indirectly altered by the effects of a drug on blood volume, arterial blood pressure, pain, or nausea. In most cases, the mechanism is not known. Stimulators of vasopressin secretion include vincristine, cyclophosphamide, tricyclic antidepressants, nicotine, epinephrine, and high doses of morphine. Lithium, which inhibits the renal effects of vasopressin, also enhances vasopressin secretion. Inhibitors of vasopressin secretion include ethanol, phenytoin, low doses of morphine, glucocorticoids, fluphenazine, haloperidol, promethazine, oxilorphan, and butorphanol. Carbamazepine has a renal action to produce antidiuresis in patients with central diabetes insipidus but actually inhibits vasopressin secretion via a central action.

Basic Pharmacology of Vasopressin Vasopressin Receptors The cellular effects of vasopressin are mediated by interactions of the hormone with the two principal types of receptors, V1 and V2. V1 receptors have been subclassified further as V1a and V1b . The V1a receptor is the most widespread subtype of vasopressin receptor; it is found in vascular smooth muscle, the adrenal gland, myometrium, the bladder, adipocytes, hepatocytes, platelets, renal medullary interstitial cells, vasa recta in the renal microcirculation, epithelial cells in the renal cortical collecting duct, spleen, testis, and in many CNS structures. The adenohypophysis and the adrenal medulla are known to contain V1b receptors, whereas V2 receptors are located predominantly in principal cells of the renal collecting duct system. Although originally defined by pharmacological criteria, V1a (Morel et al., 1992), V 1b (Sugimoto et al., 1994), and V2 receptors (Birnbaumer et al., 1992; Lolait et al., 1992) have been cloned, and vasopressin receptors now are defined by their primary amino acid sequences. The cloned vasopressin receptors are typical G proteincoupled receptors containing seven transmembrane-spanning domains. Manning and coworkers (1999) have synthesized novel, hypotensive vasopressin peptide agonists that do not interact with V1a, V1b , or V2 receptors and may stimulate a putative vasopressin vasodilatory receptor. V1 ReceptorEffector Coupling Considerable progress has been made in defining the mechanisms by which vasopressin receptors are coupled to biological responses (Thibonnier et al., 1993; Holtzman and Ausiello, 1994). Figure 304 summarizes the current model of V 1 receptoreffector coupling. When vasopressin binds to V1 receptors, a G proteinmediated activation of several membrane-bound phospholipases ensues.

Activation of phospholipase C- , probably via Gq, is responsible for hydrolysis of phosphatidylinositol-1,4,5-bisphosphate with the resultant generation of inositol-1,4,5-trisphosphate (IP3) and diacylglycerol (DAG). IP3 binds to a receptor located on Ca2+-release channels in intracellular IP3-sensitive Ca2+ stores, an event that triggers intracellular release of Ca2+. Although the mechanism is unclear, V1 receptors also cause Ca2+ influx from the extracellular compartment via Ca 2+ channels located on the cell membrane. Ca2+ binds to and activates a number of intracellular proteins that contribute to the ultimate cellular response. Stimulation of phospholipase D by V1 receptors mediates the hydrolysis of other phospholipids to produce phosphatidic acid, which is further metabolized to DAG. G proteincoupled receptors, such as the V1 receptor, may activate phospholipase D via Ca2+, protein kinase C, or small GTP-binding proteins such as ARF and Rho. Activation of protein kinase C by DAG leads to the phosphorylation of key proteins, and these phosphoproteins also contribute to the biological response. Finally, stimulation of phospholipase A2, either in the membrane or in cytosol, mobilizes arachidonic acid from phospholipids, which is metabolized to various prostaglandins and epoxyeicosatrienoic acids via cyclooxygenase and epoxygenase, respectively (Chapter 26: Lipid-Derived Autacoids: Eicosanoids and Platelet-Activating Factor). Metabolites of arachidonic acid modulate biological responses to vasopressin at least in part by stimulating their own eicosanoid receptors. The biological effects mediated by the V1 receptor include vasoconstriction, glycogenolysis, platelet aggregation, ACTH release, and growth of vascular smooth muscle cells. The effects of vasopressin on cell growth appear to involve increased expression of the proto-oncogenes c-fos and c-jun. The products of these proto-oncogenes, Fos and Jun, activate the transcription of other genes involved in the regulation of cellular growth. Figure 304. Mechanism of V1 ReceptorEffector Coupling. See text for details. "?" indicates that mechanism of coupling is unclear. V1, V1vasopressin receptor; AVP, 8-arginine vasopressin; q, , , subunits of G protein; PLD, phospholipase D; PLC- , phospholipase C- ; PLA2, phospholipase A2; DAG, 1,2diacylglycerol; PKC, protein kinase C; PIP2, phosphatidyl-inositol-4-5bisphosphate; IP3, 1,4,5-inositol trisphosphate; PA, phosphatidic acid; PPH, phosphatidate phosphohydrolase; PC, phosphatidylcholine; AA, arachidonic acid; PGs, prostaglandins; EPs, epoxyeicosatrienoic acids; CO, cyclooxygenase; EPO, epoxygenase; AP-1, transcription factor consisting of heterodimer of FOS and JUN; c-fos and c-jun are proto-oncogenes; FOS and JUN are products of cfos and c-jun gene expression, respectively.

V2 ReceptorEffector Coupling Principal cells in the renal collecting duct have V2 receptors on their basolateral membranes that are coupled to adenylyl cyclase via the timulatory G protein Gs (seeFigure 305). Consequently, when vasopressin binds to V2 receptors, adenylyl cyclase is stimulated, and intracellular levels of cyclic AMP are increased. Activation of cyclic AMPdependent protein kinase (protein kinase A) mediates the hydroosmotic effects of vasopressin (Snyder et al., 1992) via protein phosphorylation. In this regard, protein kinase Amediated protein phosphorylation triggers an increased rate of exocytosis of water channelcontaining vesicles (WCVs) into the apical membrane and a decreased rate of endocytosis of WCVs from the apical membrane. The distribution of WCVs between the cytosolic compartment and the apical membrane compartment is thus shifted in favor of the apical membrane compartment (Knepper and Nielsen, 1993; Nielsen et al., 1999). Because WCVs contain preformed, functional water channels (aquaporin 2) their increased rate of insertion into and

decreased rate of removal from the apical membrane greatly increases the water permeability of the apical membrane. Figure 305. Mechanism of V2 ReceptorEffector Coupling. See text for details. V2, V2vasopressin receptor; AVP, 8-arginine vasopressin; s, , and , subunits of G protein; ATP, adenosine triphosphate; ADP, adenosine diphosphate; cAMP, adenosine 3',5'-monophosphate (cyclic AMP).

Aquaporins are a family of water channel proteins that allow water molecules to cross biological membranes (Nielsen et al., 1999; Marples et al., 1999). Aquaporins have six membrane domains connected by five loops (A to E). Loops B and E dip into the cell membrane, and the asparagineproline-alanine sequences in each B and E loop interact to form a water pore. Aquaporins generally form tetrameric complexes in cell membranes. Of the nine cloned mammalian aquaporins, six are found in the kidney. Aquaporin 1 is present in the apical and basolateral membrane of the proximal tubule and in the thin descending limb. Aquaporin 2 resides in the apical membrane and WCVs of the collecting duct principal cells, whereas aquaporins 3 and 4 are present in the basolateral membrane of principal cells. Aquaporin 7 is in the apical brush border of the straight proximal tubule. Although aquaporin 6 is found in the kidney, its distribution is unknown. Aquaporin 2, the water channel in WCVs, is phosphorylated on serine 256 by protein kinase A, and this is the first step leading to vasopressin-induced insertion of WCVs into apical membranes (Nishimoto et al., 1999). In addition to increasing the insertion of aquaporin 2 into apical membranes in collecting duct principal cells, vasopressin also increases the expression of aquaporin 2 mRNA and protein (Marples et al., 1999). This effect is mediated by protein kinase A phosphorylation of cyclic AMP response element binding protein (CREB). Phosphorylated CREB is a transcription factor that binds to the cyclic AMPresponse element (CRE) in the 5'-untranslated region of the gene encoding aquaporin 2 and increases its transcription. Thus, chronic dehydration leads to long-term upregulation of aquaporin 2 and water transport in the collecting duct. For maximum concentration of urine, large amounts of urea must be deposited in the interstitium of the inner medullary collecting duct. It is not surprising, therefore, that V2 receptor activation also increases urea permeability by 400% in the terminal portions of the inner medullary collecting duct. The mechanism by which V2 receptors increase urea permeability involves activation of a vasopressin-regulated urea transporter (termed VRUT, UT1, or UT-A1), most likely by protein kinase Ainduced phosphorylation (Star et al., 1988; Sands, 1999). The kinetics of vasopressininduced water and urea permeability are different (Nielsen and Knepper, 1993), and vasopressininduced regulation of VRUT does not entail vesicular trafficking to the plasma membrane (Inoue et al., 1999). In addition to increasing the permeability of the collecting duct to water and the permeability of the inner medullary collecting duct to urea, V2 receptor activation also increases both short- and longterm NaCl transport in the thick ascending limb. Increased transport in the thick ascending limb augments the countercurrent multiplication system and thereby increases the osmolality of the medullary interstitium and the reabsorption of water (Knepper et al., 1999). This effect, which is most likely mediated by cyclic AMP and protein kinase Ainduced phosphorylation, involves the + + immediate stimulation of Na K 2Cl symporter expression (Kim et al., 1999). The multiple mechanisms by which vasopressin increases water reabsorption are summarized in Figure 306. Figure 306. Mechanisms by Which Vasopressin Increases the Renal Conservation of Water. IMCD, inner medullary collecting duct; CD, collecting duct; TAL, thick ascending limb; VRUT, vasopressin-regulated urea transporter.

Renal Actions of Vasopressin There are several sites of vasopressin action in the kidney involving both V1 and V2 receptors. V 1 receptors mediate contraction of mesangial cells in the glomerulus and contraction of vascular smooth muscle cells in the vasa recta and efferent arteriole (Edwards et al., 1989). Indeed, V1 receptormediated reduction in inner medullary blood flow contributes to the maximum concentrating capacity of the kidney (Franchini and Cowley, 1996) (seeFigure 306). V1 receptors also stimulate prostaglandin synthesis by medullary interstitial cells. Since prostaglandin E2 inhibits adenylyl cyclase in the collecting duct, stimulation of prostaglandin synthesis by V1 receptors may function to restrain V2 receptormediated antidiuresis (Sonnenberg and Smith, 1988). V1 receptors on principal cells in the cortical collecting duct (Burnatowska-Hledin and Spielman, 1989) may directly inhibit V2 receptormediated water flux via activation of protein kinase C (Schlondorff and Levine, 1985). Without question, V2 receptors mediate the most prominent response to vasopressin, i.e., increased water permeability of the collecting duct. Indeed, vasopressin can increase water permeability in the collecting duct at concentrations as low as 50 fM. Thus, V2 receptor-mediated effects of vasopressin occur at concentrations far lower than are required to engage the V1 receptor-mediated actions; however, this differential sensitivity may not be due to differences in receptor affinities, since cloned rat V1a and V 2 receptors have similar affinities for vasopressin (Kd= 0.7 and 0.4 nM, respectively). The differential sensitivity may instead be due to differential amplification of signal transduction pathways evoked by these receptors. The collecting duct system is critical for the conservation of water. By the time tubular fluid arrives at the cortical collecting duct, it has been rendered hypotonic by the upstream diluting segments of the nephron that reabsorb NaCl without reabsorbing water. In the well-hydrated subject, plasma osmolality is in the normal range, concentrations of vasopressin are low, the entire collecting duct is

relatively impermeable to water, and the urine is dilute. Under conditions of dehydration, plasma osmolality is increased, concentrations of vasopressin are elevated, and the collecting duct becomes permeable to water. The osmotic gradient between the dilute tubular urine and the hypertonic renal interstitial fluid (which becomes progressively more hypertonic in deeper regions of the renal medulla) provides for the osmotic flux of water out of the collecting duct. The final osmolality of urine may be as high as 1200 mOsm/kg in human beings, and a significant saving of solute-free water is thus possible. Other renal actions mediated by V2 receptors include increased urea transport in the inner medullary collecting duct and increased NaCl transport in the thick ascending limb; both effects contribute to the urine-concentrating ability of the kidney. V2 receptors also increase Na+ transport in the cortical collecting duct (Schafer and Troutman, 1990). Pharmacological Modification of the Antidiuretic Response to Vasopressin Nonsteroidal antiinflammatory drugs (NSAIDs) (seeChapter 27: Analgesic-Antipyretic and Antiinflammatory Agents and Drugs Employed in the Treatment of Gout), particularly indomethacin, enhance the antidiuretic response to vasopressin. Since prostaglandins attenuate antidiuretic responses to vasopressin and NSAIDs inhibit prostaglandin synthesis, reduced prostaglandin production probably accounts for the potentiation of vasopressin's antidiuretic response. Other drugs that enhance the antidiuretic effects of vasopressin include carbamazepine and chlorpropamide; however, the mechanisms by which these agents potentiate the antidiuretic response to vasopressin are not known. In rare instances, chlorpropamide can induce water intoxication. A number of drugs inhibit the antidiuretic actions of vasopressin. Lithium is of particular importance because of its wide use in the treatment of manic-depressive disorders. Lithium-induced polyuria is usually, but not always, reversible (Ramsey and Cox, 1982). Acutely, lithium appears to reduce V2 receptormediated stimulation of adenylyl cyclase. The mechanism of this effect may involve attenuation of Gs-mediated activation of adenylyl cyclase (Cogan and Abramow, 1986; Goldberg et al., 1988) and/or enhancement of Gi-mediated inhibition of adenylyl cyclase (Yamaki et al., 1991). Also, lithium increases plasma levels of parathyroid hormone, and parathyroid hormone is a partial antagonist to vasopressin (Carney et al., 1996). In most patients, the antibiotic demeclocycline attenuates the antidiuretic effects of vasopressin, and this action of demeclocycline is probably due to decreased accumulation and action of cyclic AMP (Singer and Rotenberg, 1973). Nonrenal Actions of Vasopressin Vasopressin and related peptides are ancient hormones in evolutionary terms, and they are found in species that do not concentrate urine. It is thus not surprising that vasopressin has nonrenal actions in mammals. Cardiovascular System The cardiovascular effects of vasopressin are complex, and vasopressin's role in physiological situations is ill-defined. Vasopressin is a potent vasoconstrictor (V 1 receptormediated), and resistance vessels throughout the circulation may be affected (Lszlet al., 1991). Vascular smooth muscle in the skin, skeletal muscle, fat, pancreas, and thyroid gland appear most sensitive, with significant vasoconstriction also occurring in the gastrointestinal tract, coronary vessels, and brain (Liard et al., 1982). However, despite the potency of vasopressin as a direct vasoconstrictor, vasopressin-induced pressor responses in vivo are minimal and occur only with concentrations of

vasopressin significantly higher than those required for maximal antidiuresis. To a large extent, this is due to circulating vasopressin that acts on V1 receptors to inhibit sympathetic efferents and potentiate baroreflexes (Abboud et al., 1990). In addition, in some blood vessels, V 2 receptors cause vasodilation, perhaps via release of nitric oxide (a potent vasodilator) from the vascular endothelium (Aki et al., 1994), and studies by Manning and coworkers (1999) suggest the existence of a unique vasopressin vasodilator receptor. A large body of data from experimental animals supports the conclusion that vasopressin helps maintain arterial blood pressure during episodes of severe hypovolemia/hypotension (Lszlet al., 1991). However, antagonism of V 1 receptors does not alter the hypotensive response to lower-body negative pressure in normal subjects despite a fivefold elevation in circulating vasopressin levels (Hirsch et al., 1993). Vasopressin also may increase total peripheral resistance in heart failure, and administration of a peptide V1 receptor antagonist improves hemodynamic function in such patients (Thibonnier, 1988). At present, there is no convincing evidence for a role of vasopressin in essential hypertension in human beings (Kawano et al., 1997). The effects of vasopressin on the heart (reduced cardiac output and heart rate) are largely indirect and are the result of coronary vasoconstriction, decreased coronary blood flow, and alterations in vagal and sympathetic tone (Lszlet al., 1991). In human beings, the effects of vasopressin on coronary blood flow can be easily demonstrated, especially if large doses are employed. The cardiac actions of the hormone are of more than academic interest. Some patients with coronary insufficiency experience anginal pain even in response to the relatively small amounts of vasopressin required to control diabetes insipidus, and vasopressin-induced myocardial ischemia has led to severe reactions and even death. Central Nervous System It is likely that vasopressin plays a role as a neurotransmitter and/or neuromodulator (Gash et al., 1987; Jolles, 1987). Vasopressin may participate in the acquisition of certain learned behaviors (Dantzer and Bluth, 1993), in the development of some complex social processes (Insel et al., 1993; Young et al., 1998), and in the pathogenesis of specific psychiatric diseases (Legros et al., 1993). However, the physiological/pathophysiological relevance of these findings is controversial, and some of the actions of vasopressin on memory and learned behavior may be due to visceral autonomic effects. In 1931, Cushing reported on the antipyretic effects of pituitary extracts injected into the lateral ventricle of febrile patients. Since then, many studies have supported a physiological role for vasopressin as a naturally occurring antipyretic factor (Kasting, 1989; Cridland and Kasting, 1992). Although vasopressin can modulate CNS autonomic systems controlling heart rate, arterial blood pressure, respiration rate, and sleep patterns, the physiological significance of these actions is unclear. Finally, secretion of ACTH is enhanced by vasopressin that is delivered to the anterior pituitary via a neuronal pathway leading to secretion of peptide into the hypophyseal portal blood. However, vasopressin is not the principal corticotropin-releasing factor. The CNS effects of vasopressin appear to be mediated predominantly by V1 receptors. Blood Coagulation Activation of V2 receptors by desmopressin or vasopressin increases circulating levels of procoagulant factor VIII and of von Willebrand factor (David, 1993). This effect is mediated by extrarenal V 2 receptors (Bernat et al., 1997). Presumably, vasopressin stimulates the secretion of von Willebrand factor and of factor VIII from storage sites in vascular endothelium. However, since release of von Willebrand factor does not occur when desmopressin is applied directly to cultured endothelial cells or to isolated blood vessels, intermediate factors are likely to be involved. In this

regard, it has been hypothesized that desmopressin releases interleukin 1 (IL-1) from monocytes, and IL-1 can then release von Willebrand factor (Breit and Green, 1988). Other Nonrenal Effects of Vasopressin At high concentrations, vasopressin stimulates uterine (viaoxytocin receptors) and gastrointestinal (via V1 receptors) smooth muscle. Vasopressin is stored in platelets, and V1 receptors stimulate platelet aggregation (Inaba et al., 1988). Also, V 1 receptors located on hepatocytes stimulate glycogenolysis (Keppens and de Wulf, 1975). The physiological significance of these effects of vasopressin is not known. Diseases Affecting the Vasopressin System Diabetes Insipidus (DI) DI is a disease of impaired renal conservation of water due either to an inadequate secretion of vasopressin from the neurohypophysis (central or cranial DI) or to an insufficient renal response to vasopressin (nephrogenic DI). Very rarely, DI can be caused by an abnormally high rate of degradation of vasopressin by circulating vasopressinases (Durr et al., 1987). Pregnancy may accentuate or reveal central and/or nephrogenic DI by increasing plasma levels of vasopressinase and by reducing the renal sensitivity to vasopressin. Patients with DI excrete large volumes (more than 30 ml/kg per day) of dilute (less than 200 mOsm/kg) urine and, if their thirst mechanism is functioning normally, are polydipsic. In contrast to the sweet urine excreted by patients with diabetes mellitus, urine from patients with DI is tasteless, hence the name, insipidus. Fortunately, the urinary taste test for DI devised by Willis in the seventeenth century has been supplanted by the more palatable approach of simply observing whether the patient is able to reduce urine volume and increase urine osmolality after a period of carefully observed fluid deprivation. Central DI can be distinguished from nephrogenic DI by administration of desmopressin, which will increase urine osmolality in patients with central DI but have little or no effect in patients with nephrogenic DI. DI can be differentiated from primary polydipsia by measuring plasma osmolality, which will be low to low-normal in patients with primary polydipsia and high to high-normal in patients with DI. For a more complete discussion of diagnostic procedures, see Vokes and Robertson (1988). Central DI Head injury, either surgical or traumatic, in the region of the pituitary and/or hypothalamus may cause central DI. Postoperative central DI may be transient, permanent, or triphasic (recovery followed by permanent relapse) (Seckl and Dunger, 1992). Other causes include hypothalamic or pituitary tumors, cerebral aneurysms, CNS ischemia, and brain infiltrations and infections. Finally, central DI may be idiopathic or familial. Familial central DI is usually autosomal dominant (chromosome 20) and has been associated with point mutations in the signal peptide and VPneurophysin (seeFigure 301), leading to defects in the synthesis, processing, and transport of the preprohormone complex (Raymond, 1994). Since familial central DI is autosomal dominant, the defective preprohormone coded by the mutant allele must interfere in some way with the synthesis of hormone coded by the normal allele. Antidiuretic peptides are the primary treatment for central DI, with desmopressin being the peptide of choice. (See"Clinical Pharmacology of Vasopressin Peptides," below, for discussion of antidiuretic peptides in the treatment of central DI.) For patients with central DI who cannot tolerate antidiuretic peptides because of side effects or allergic reactions, other treatment options are available. Chlorpropamide, an oral sulfonylurea that potentiates the action of small or residual

amounts of circulating vasopressin, will cause a reduction in urine volume in more than half of all patients with central DI at a dose of 125 to 500 mg daily and is particularly effective in patients with partial central DI, where it potentiates the action of low levels of circulating vasopressin. If polyuria is not satisfactorily controlled with chlorpropamide alone, addition of a thiazide diuretic (Chapter 29: Diuretics) to the regimen usually results in an adequate reduction in the volume of urine. Carbamazepine (800 to 1000 mg daily in divided doses) and clofibrate (1 to 2 g daily in divided doses) also reduce urine volume in patients with central DI. Long-term use of these agents may induce serious adverse effects; therefore carbamazepine and clofibrate are rarely used to treat central DI. The antidiuretic mechanism(s) of chlorpropamide, carbamazepine, and clofibrate are not clear. These agents are not effective in nephrogenic DI, which indicates that functional V2 receptors are required for the antidiuretic mechanism to be expressed. Since carbamazepine inhibits and chlorpropamide has little effect on vasopressin secretion, it is likely that carbamazepine and chlorpropamide act directly on the kidney to enhance V2 receptormediated antidiuresis. Nephrogenic DI Nephrogenic DI may be acquired or genetic. Hypercalcemia, hypokalemia, postobstructive renal failure, lithium, and demeclocycline can induce nephrogenic DI. As many as one in three patients treated with lithium may develop nephrogenic DI. Familial, X-linked, recessive nephrogenic DI is caused by mutations in the gene coding for the V2 receptor (a gene located in the q28 region of the X chromosome). A number of missense, nonsense, and frame-shift mutations in the gene encoding the V2 receptor have been identified in patients with familial, X-linked nephrogenic DI (Oksche and Rosenthal, 1998; Morello et al., 2000). Of all mutant alleles reported to date, mutations encoded result predominantly in abnormal synthesis, processing, or intracellular transport of V2 receptors. Autosomal recessive nephrogenic DI is caused by inactivating mutations in aquaporin 2 (13 different mutations have been identified to date). These findings indicate that aquaporin 2 is essential for the antidiuretic effect of vasopressin in human beings (Deen et al., 1994). Although the mainstay of treatment of nephrogenic DI is assurance of an adequate intake of water, drugs also can be used to reduce polyuria. Amiloride (Chapter 29: Diuretics) blocks the uptake of lithium by the sodium channel in the collecting duct system and is therefore the drug of choice for lithium-induced nephrogenic DI. Paradoxically, thiazide diuretics cause a reduction in the polyuria of patients with DI and often are used to treat nephrogenic DI. The use of thiazide diuretics in infants with nephrogenic DI may be crucially important, since uncontrolled polyuria may exceed the child's capacity to imBIBe and absorb fluids. The antidiuretic mechanism of thiazides in DI is incompletely understood. It is possible that the natriuretic action of thiazides and resultant depletion of extracellular fluid volume play an important role in the thiazide-induced antidiuresis in DI. In this regard, whenever extracellular fluid volume is reduced, compensatory mechanisms increase reabsorption of NaCl in the proximal tubule, with a resultant reduction of volume delivered to the distal tubule. Consequently, less free water can be formed, and this should diminish polyuria. However, recent studies in rats with vasopressin-deficient DI seriously challenge this hypothesis (Grnbeck et al., 1998). Nonetheless, the antidiuretic effects appear to parallel the ability of thiazides to cause natriuresis, and the drugs are given in doses similar to those used for mobilization of edema fluid. In patients with DI, a 50% reduction of urine volume is a good response to thiazides. Moderate restriction of sodium intake can enhance the antidiuretic effectiveness of thiazides. A number of case reports describe the effectiveness of indomethacin in the treatment of nephrogenic DI (Libber et al., 1986); however, other prostaglandin synthase inhibitors (e.g., ibuprofen) appear to be less effective. The mechanism of the effects of indomethacin is unclear but may involve a decrease in glomerular filtration rate, an increase in medullary solute concentration,

and/or enhanced proximal reabsorption of fluid (Seckl and Dunger, 1992). Also, since prostaglandins attenuate vasopressin-induced antidiuresis in patients with at least a partially intact V2 receptor system, a portion of the antidiuretic response to indomethacin may be due to enhancement of the effects of vasopressin on the principal cells of the collecting duct. Syndrome of Inappropriate Secretion of Antidiuretic Hormone (SIADH) SIADH is a disease of impaired water excretion with accompanying hyponatremia and hypoosmolality caused by the inappropriate secretion of vasopressin. The clinical manifestations of plasma hypotonicity resulting from SIADH may include lethargy, anorexia, nausea/vomiting, muscle cramps, coma, convulsions, and death. A multitude of disorders can induce SIADH (Zerbe et al., 1980) including malignancies, pulmonary diseases, CNS injuries/diseases (head trauma, infections, tumors), general surgery, and drugs (e.g., cisplatin, vinca alkaloids, cyclophosphamide, chloropropamide, thiazide diuretics, phenothiazines, carbamazepine, clofibrate, nicotine, narcotics, and tricyclic antidepressants). In a normal individual, an elevation in plasma vasopressin levels per se does not induce plasma hypotonicity, because the person simply stops drinking due to an osmotically induced aversion to fluids. Therefore, plasma hypotonicity only occurs when excessive fluid intake (oral or intravenous) accompanies inappropriate secretion of vasopressin. Treatment of hypotonicity in the setting of SIADH includes water restriction, intravenous administration of hypertonic saline, loop diuretics (which interfere with the concentrating ability of the kidneys), and drugs that inhibit the ability of vasopressin to increase water permeability in the collecting ducts. To inhibit vasopressin's action in the collecting ducts, demeclocycline is the preferred drug (however, see"Prospectus," below). Although lithium can inhibit the renal actions of vasopressin, it is effective in only a minority of patients, may induce irreversible renal damage when used chronically, and has a low therapeutic index. Therefore, lithium should be used only in patients with symptomatic SIADH who cannot be controlled by other means or in whom tetracyclines are contraindicated, e.g., patients with liver disease. It is important to stress that the majority of patients with SIADH do not require therapy because plasma Na+ stabilizes in the range of 125 to 132 mM; such patients usually are asymptomatic. Only when symptomatic hypotonicity ensues, generally when plasma Na+ levels drop below 120 mM, should therapy with demeclocycline be initiated. Since hypotonicity, which causes an influx of water into cells with resulting cerebral swelling, is the cause of symptoms, the goal of therapy is simply to increase plasma osmolality toward normal. For a more complete description of the diagnosis and treatment of SIADH, see Kovacs and Robertson (1992). Other Water-Retaining States In patients with congestive heart failure, liver cirrhosis, and nephrotic syndrome, effective blood volume often is reduced, and hypovolemia is frequently exacerbated by the liberal use of diuretics in such patients. Since hypovolemia stimulates vasopressin release, patients may become hyponatremic due to vasopressin-mediated retention of water. The development of potent, orally active V2 receptor antagonists and specific inhibitors of water channels in the collecting duct would provide an effective therapeutic strategy, not only in patients with SIADH but also in the much more common setting of hyponatremia in patients with heart failure, liver cirrhosis, and nephrotic syndrome.

Clinical Pharmacology of Vasopressin Peptides

Therapeutic Uses Only three antidiuretic peptides are available for clinical use in the United States. (1) Vasopressin (synthetic 8-L-arginine vasopressin; PITRESSIN SYNTHETIC) is available as a sterile aqueous solution; it may be administered subcutaneously, intramuscularly, or intranasally. (2) Lypressin (synthetic 8-lysine vasopressin; DIAPID) is supplied as an aqueous nasal spray. (3) Desmopressin acetate ( DDAVP ) is available as a sterile aqueous solution packaged for intravenous or subcutaneous injection, in a nasal solution for intranasal administration with either a nasal spray pump or rhinal tube delivery system, and in tablets for oral administration. The therapeutic uses of vasopressin and its congeners can be divided into two main categories according to the type of vasopressin receptor involved. V1 receptormediated therapeutic applications are based on the rationale that V1 receptors cause contraction of gastrointestinal and vascular smooth muscle. V1 receptormediated contraction of gastrointestinal smooth muscle is useful to treat postoperative ileus and abdominal distension and to dispel intestinal gas before abdominal roentgenography to avoid interfering gas shadows. V1 receptormediated vasoconstriction of the splanchnic arterial vessels reduces blood flow to the portal system and thereby attenuates pressure and bleeding in esophageal varices (Burroughs, 1998). Although endoscopic sclero-therapy is the treatment of choice for bleeding esophageal varices, V1 receptor agonists can be used in an emergency setting until endoscopy can be performed. Simultaneous administration of nitroglycerin with vasopressin has been reported to reverse the cardiotoxic effects of vasopressin while enhancing the beneficial splanchnic effects of the drug (Gimson et al., 1986). Also, V1 receptor agonists can be used during abdominal surgery in patients with portal hypertension to diminish the risk of hemorrhage during the procedure. Finally, V1 receptormediated vasoconstriction of the gastric vascular bed reduces bleeding in acute hemorrhagic gastritis (Peterson, 1989). 8-Arginine vasopressin should be used for all V1 receptor mediated therapeutic applications. Although not yet available in the United States, terlipressin (GLYPRESSIN) appears to be effective for bleeding esophageal varices, with reduced side effects compared with vasopressin (Soederlund, 1993). V2 receptormediated therapeutic applications are based on the rationale that V2 receptors cause water conservation and release of blood coagulation factors. Central, but not nephrogenic, DI can be treated with V2 receptor agonists, and polyuria and polydipsia are usually well controlled. Some patients experience transient DI (e.g., in head injury or surgery in the area of the pituitary); however, for most patients with DI, therapy is lifelong. Desmopressin is the drug of choice for the vast majority of patients, and numerous clinical trials (Robinson, 1976; Cobb et al., 1978) have demonstrated that desmopressin is an effective agent in both adults and children and has few side effects. The duration of effect from a single intranasal dose is from 6 to 20 hours, and twice-daily administration has proven to be effective in most patients. There is considerable variability in the intranasal dose of desmopressin required to maintain normal urine volume, and the dosage must be individually tailored. The usual intranasal dosage range in adults is 10 to 40 g daily, either as a single dose or divided into two or three doses. In view of the high cost of the drug and the importance of avoiding water intoxication, the schedule of administration should be adjusted to determine the minimal amount required. An initial dose of 2.5 g can be used, and therapy should first be directed toward the control of nocturia. An equivalent or higher morning dose controls daytime polyuria in most patients, although a third dose occasionally may be needed in the afternoon. In some patients, chronic allergic rhinitis or other nasal pathology may preclude reliable absorption of the peptide following nasal administration. Oral administration of desmopressin in doses 10 to 20 times the intranasal dose provides adequate blood levels of desmopressin and controls polyuria (Fjellestad-Paulsen et al., 1993). Subcutaneous administration of 1 to 2 g daily of

desmopressin also is effective in central DI. Lypressin nasal spray also can be used to treat central DI; however, lypressin's duration of action is short (4 to 6 hours), making it less convenient than desmopressin. Also, lypressin, like vasopressin, can induce V1 receptormediated adverse effects. Nonetheless, for patients refractory to desmopressin or who experience side effects with desmopressin, lypressin provides an alternative. 8-Arginine vasopressin has little if any place in the long-term therapy of DI because of its short duration of action and V1 receptormediated side effects. Vasopressin can be used as an alternative to desmopressin in the initial diagnostic evaluation of patients with suspected DI and to control polyuria in patients with DI who have recently undergone surgery or experienced head trauma. Under these circumstances, polyuria may be transient, and long-acting agents may produce water intoxication. An additional V2 receptormediated therapeutic application is the use of desmopressin in bleeding disorders (Mannucci, 1997; Sutor, 1998). In most patients with type I von Willebrand's disease (vWD) and in some with type IIN vWD, desmopressin will elevate von Willebrand factor and shorten bleeding time. However, desmopressin is generally ineffective in patients with types IIa, IIb, and III vWD. Desmopressin may cause a marked, transient thrombocytopenia in individuals with type IIb vWD and may be dangerous in such patients. Desmopressin also increases factor VIII levels in patients with moderately severe hemophilia A. Desmopressin is not indicated in patients with severe hemophilia A, those with hemophilia B, or those with factor VIII antibodies. The response of any given patient with type I vWD or hemophilia A to desmopressin should be determined at the time of diagnosis or 1 to 2 weeks before elective surgery to assess the extent of increase in factor VIII or von Willebrand factor. Desmopressin is employed widely to treat the hemostatic abnormalities induced by uremia (Mannucci et al., 1983). In patients with renal insufficiency, desmopressin shortens bleeding time and increases circulating levels of factor VIII coagulant activity, factor VIIIrelated antigen, and ristocetin cofactor. It also induces the appearance of larger von Willebrand factor multimers. Desmopressin is effective in some patients with liver cirrhosisinduced or druginduced (heparin, hirudin, antiplatelet agents) bleeding disorders. Desmopressin, given intravenously at a dose of 0.3 g/kg, increases factor VIII and von Willebrand factor for more than 6 hours. Desmopressin can be given at intervals of 12 to 24 hours, depending on the clinical response and the severity of bleeding. Tachyphylaxis to desmopressin usually occurs after several days (due to depletion of factor VIII and von Willebrand factor storage sites) and limits its usefulness to preoperative preparation, postoperative bleeding, excessive menstrual bleeding, and emergency situations. Another V2 receptormediated therapeutic application is the use of desmopressin for primary nocturnal enuresis (Sukhai, 1993). In this regard, desmopressin should be used only intranasally and may be used alone or in combination with behavioral conditioning. Finally, desmopressin has been found to relieve postlumbar puncture headache, probably by causing water retention and thereby facilitating rapid fluid equilibration in the CNS. Pharmacokinetics When vasopressin, lypressin, and desmopressin are given orally, they are quickly inactivated by trypsin, which cleaves the peptide bond between amino acids 8 and 9. Inactivation by peptidases in various tissues (particularly the liver and kidneys) results in a plasma half-life of vasopressin of 17 to 35 minutes. The plasma half-life of desmopressin has two components, a fast component of 6.5 to 9 minutes and a slow component of 30 to 117 minutes.

Toxicity, Adverse Effects, Contraindications, Drug Interactions Most adverse effects are mediated through the V1 receptor acting on vascular and gastrointestinal smooth muscle; consequently such adverse effects are much less common and less severe with desmopressin than with vasopressin or lypressin. After the injection of large doses of vasopressin, marked facial pallor as a result of cutaneous vasoconstriction commonly is observed. Increased intestinal activity is likely to cause nausea, belching, cramps, and an urge to defecate. Most serious, however, is the effect on the coronary circulation. Vasopressin and lypressin should be administered only at low doses and with extreme caution in individuals suffering from vascular disease, especially disease of the coronary arteries. Other cardiac complications include arrhythmia and decreased cardiac output. Peripheral vasoconstriction and gangrene have been encountered in patients receiving large doses of vasopressin. The major V2 receptormediated adverse effect is water intoxication, which can occur with desmopressin, lypressin, or vasopressin. In this regard, carbamazepine, chlorpropamide, and NSAIDs can potentiate the antidiuretic effects of these peptides. Desmopressin, lypressin, and vasopressin should be used cautiously in disease states in which a rapid increase in extracellular water may impose risks (e.g., in angina, hypertension, heart failure) and should not be used in patients with acute renal failure. Also, it is imperative that these peptides not be administered to patients with primary or psychogenic polydipsia, because severe hypotonic hyponatremia will ensue. Allergic reactions, ranging from urticaria to anaphylaxis, may occur with desmopressin, lypressin, or vasopressin. Intranasal administration may cause local adverse effects in the nasal passages, such as edema, scarring, rhinorrhea, congestion, irritation, pruritus, and ulceration. Prospectus It is likely that nonpeptide, orally active V2-selective receptor antagonists will become available in the near future for the treatment of dilutional hyponatremia induced by heart failure, liver cirrhosis, nephrotic syndrome, and SIADH (Schrier et al., 1998; Mayinger and Hensen, 1999). In addition, orally active combined V1a-/V2-selective antagonists may become available for diseases such as congestive heart failure, in which both increased peripheral vascular resistance and dilutional hyponatremia are present (Yatsu et al., 1999). Whether or not sporadic administration of membrane-permeant, nonpeptide V2 receptor antagonists will achieve successful in vivo cell surface-receptor delivery of certain alleles of X-linked nephrogenic DI, as demonstrated in vitro, remains to be ascertained in clinical trials (Morello et al., 2000). Several important new clinical applications of the ancient hormone vasopressin are under active investigation. Vasopressin may prove superior to epinephrine as a pressor agent in cardiopulmonary resuscitation (Chugh et al., 1997), and it may be particularly effective in refractory hypotension after cardiopulmonary bypass (Overland and Teply, 1998). Vasopressin levels in patients with septic shock are inappropriately low, and septic shock patients are extraordinarily sensitive to the pressor actions of vasopressin (Landry et al., 1997). Vasopressin also reverses intractable hypotension in the late phase of hemorrhagic shock (Morales et al., 1999). Thus, vasopressin may become the preferred pressor agent in emergency and critical care medicine.

Chapter 31. Renin and Angiotensin

Overview In the early 1970s, pharmacological interruption of the reninangiotensin system was considered sensible only for patients with high-renin hypertension. In the ensuing years, however, an extraordinary expansion of clinical indications for pharmacological interruption of the renin angiotensin system occurred, beginning in the late 1970s with the advent of angiotensin converting enzyme (ACE) inhibitors. Surprisingly, ACE inhibitors proved to be effective not only in patients with high-renin hypertension, but also in many patients with essential hypertension who have normal levels of plasma renin activity. More recently, ACE inhibitors have gained widespread use in patients with congestive heart failure, myocardial infarction, and diabetic nephropathy. The realization that the reninangiotensin system participates significantly in the pathophysiology of several highly prevalent diseases has led to an enormous effort to explore all aspects of the renin angiotensin system and to devise new approaches for inhibiting its actions. These efforts have paid off handsomely, particularly in recent years, and knowledge concerning the reninangiotensin system and pharmacological methods for manipulating it have grown impressively. The objective of this chapter is to provide an up-to-date account of: (1) the biochemistry, molecular and cellular biology, and physiology of the reninangiotensin system; (2) the pharmacology of drugs that interrupt the reninangiotensin system; and (3) the clinical utility of inhibitors of the renin angiotensin system. Therapeutic applications of drugs covered in this chapter also are discussed in Chapters 32: Drugs Used for the Treatment of Myocardial Ischemia, 33: Antihypertensive Agents and the Drug Therapy of Hypertension, and 34: Pharmacological Treatment of Heart Failure.

The ReninAngiotensin System History In 1898, Tiegerstedt and Bergman found that crude saline extracts of the kidney contained a pressor substance, which they named renin. Their discovery had an obvious bearing on the problem of arterial hypertension and its relation to kidney disease that had been posed by Richard Bright's work some 60 years earlier. However, relatively little interest was generated until 1934, when Goldblatt and his colleagues showed convincingly that it was possible to produce persistent hypertension in dogs by constricting the renal arteries. In 1940, Braun-Menndez and his colleagues in Argentina and Page and Helmer in the United States reported that renin was an enzyme that acted on a plasma protein substrate to catalyze the formation of the actual pressor material, a peptide, which was named hypertensin by the former group and angiotonin by the latter. These two terms persisted for nearly 20 years, until it was agreed to rename the pressor substance angiotensin and to call the plasma substrate angiotensinogen. In the mid-1950s, two forms of angiotensin were recognized, the first a decapeptide (angiotensin I) and the second an octapeptide (angiotensin II) formed by enzymatic cleavage of angiotensin I by another enzyme, termed angiotensin converting enzyme (ACE). The octapeptide was shown to be the more active form, and its synthesis in 1957 by Schwyzer and by Bumpus made the material available for intensive study. Further progress came in 1958, when Gross suggested that the reninangiotensin system was involved in the regulation of aldosterone secretion. It was soon shown that the kidneys are important for such regulation and that synthetic angiotensin, in minute amounts, stimulates the production of aldosterone in human beings. Moreover, elevated rates of renin secretion were noted upon experimental depletion of Na+ (Gross, 1968). Thus, the reninangiotensin system came to be

recognized as a mechanism to stimulate aldosterone synthesis and secretion and an important physiological mechanism in the homeostatic regulation of blood pressure and electrolyte composition of body fluids. In the early 1970s, polypeptides (not orally active) were discovered that either inhibited the formation of angiotensin II or blocked angiotensin II receptors, and experimental studies with these inhibitors revealed important physiological and pathophysiological roles for the reninangiotensin system. These findings inspired the development of a new and broadly efficacious class of antihypertensive drugs, the orally active ACE inhibitors. Subsequent experimental and clinical studies with ACE inhibitors uncovered additional roles for the reninangiotensin system in the pathophysiology of hypertension, heart failure, vascular disease, and renal failure, findings that provided further impetus for the development of additional classes of drugs that inhibit the renin angiotensin system. In patents granted in 1982 (Furakawa et al., 1982), it was reported that derivatives of imidazole-5-acetic acid attenuated vasoconstriction induced by angiotensin II. Two of these compounds, S-8307 and S-8308, subsequently were shown to be selective and competitive antagonists of angiotensin II receptors. These lead compounds rapidly were improved upon and gave rise to losartan, an orally active, highly selective, and potent nonpeptide angiotensin II receptor antagonist (Carini and Duncia, 1988). Subsequently, many other angiotensin II receptor antagonists have been developed. Overview The reninangiotensin system is an important participant in both the short- and long-term regulation of arterial blood pressure. Factors that decrease arterial blood pressure, such as decreases in effective blood volume (caused by, for example, a low-sodium diet, diuretics, blood loss, congestive heart failure, liver cirrhosis, or nephrotic syndrome) or reductions in total peripheral resistance (caused by, for example, vasodilators), activate renin release from the kidneys. Renin is an enzyme that acts on angiotensinogen (renin substrate) to catalyze the formation of the decapeptide angiotensin I. This decapeptide is then cleaved by ACE to yield the octapeptide angiotensin II. A representation of the biochemical pathways of the reninangiotensin system is shown in Figure 311. Figure 311. Formation of Angiotensin Peptides. The solid arrows show the classical pathways, and the dashed arrows indicate minor alternative pathways. The structures of the angiotensins shown are those found in human beings, horse, rat, and pig; the bovine form has valine in the 5 position. The N-terminal sequence of human angiotensinogen is depicted.

Angiotensin II acts via diverse, yet coordinated, mechanisms to raise arterial blood pressure toward normal. The peptide acts in several ways to increase total peripheral resistance and thereby contributes to the short-term regulation of arterial blood pressure. Perhaps more important is the ability of angiotensin II to inhibit excretion of Na+ and water by the kidneys. Angiotensin II induced changes in renal function play an important role in long-term stabilization of arterial blood pressure in the face of large swings in dietary Na+ intake (Hall et al., 1980). As with its effects on peripheral resistance, the renal actions of angiotensin II also involve multiple interacting mechanisms (see below). Components of the ReninAngiotensin System Renin The major determinant of the rate of angiotensin II production is the amount of renin released by the kidney. Renin is synthesized, stored, and secreted into the renal arterial circulation by the granular juxtaglomerular cells that lie in the walls of the afferent arterioles as they enter the glomeruli. There is morphological and functional evidence that renin is secreted by exocytosis of storage granules, and studies by Friis et al. (1999) provide direct evidence of exocytosis in isolated juxtaglomerular cells. Renin is an aspartyl protease that attacks a restricted number of substrates. Its principal natural substrate is a circulating 2-globulin, angiotensinogen. Renin cleaves the bond between residues 10 and 11 at the amino terminus of this protein to generate angiotensin I. The active form of renin is a

glycoprotein that contains 340 amino acids. It is synthesized as a preproenzyme of 406 amino acid residues that is processed to prorenin, a mature but inactive form of the protein (Imai et al., 1983). Prorenin is finally activated by an as yet uncharacterized enzyme that removes 43 amino acids from the amino terminus of prorenin. Similar to other aspartyl proteases, renin has a bilobal structure with a cleft that forms the active site (Inagami, 1989; Sielecki et al., 1989). A truncated, nonsecreted form of renin is expressed in the brain as a result of an alternative promoter within intron I of the renin gene (Lee-Kirsch et al., 1999). Renin and prorenin both are stored in the juxtaglomerular cells and, when released, circulate in the blood. The concentration of prorenin in the circulation is approximately tenfold greater than that of the active enzyme. The half-life of circulating renin is approximately 15 minutes. The physiological status of circulating prorenin is unclear. Control of Renin Secretion (See Figure 312.) The secretion of renin from juxtaglomerular cells is controlled predominantly by three pathways: two acting locally within the kidney, and the third acting through the central nervous system (CNS) and mediated by norepinephrine release from renal noradrenergic nerves. Figure 312. A. A Schematic Portrayal of the Three Major Physiological Pathways Regulating Renin Release. See text for details. MD, macula densa; PGI2, prostaglandin I2; PGE2, prostaglandin E2; NSAIDs, nonsteroidal antiinflammatory drugs; Ang II, angiotensin II; ACE, angiotensin converting enzyme, AT1-R, angiotensin subtype 1 receptor; NE/Epi, norepinephrine/epinephrine; -blockers, -adrenergic receptor antagonists; AT1 blockers, AT1-R antagonists; JGCs, juxtaglomerular cells. B. Possible mechanisms by which the macula densa regulates renin release. Chronic sodium depletion upregulates neuronal nitric oxide synthase (nNOS) and inducible cyclooxygenase (COX-2) in the macula densa. nNOS increases nitric oxide (NO) production, and NO reacts with superoxide anion (O2) to form peroxynitrite, an activator of COX-2. In addition, COX-2 may be rapidly although indirectly inhibited and stimulated by increases and decreases in NaCl transport, respectively, across the macula densa. Arachidonic acid (AA) is converted to prostaglandins (PGs), which diffuse to nearby juxtaglomerular cells to stimulate adenylyl cyclase (AC) via a stimulatory G-protein (Gs). Cyclic AMP (cAMP) augments renin release. Increased NaCl transport depletes ATP and increases adenosine (ADO) levels in the macula densa. ADO diffuses to the juxtaglomerular cells and inhibits adenylyl cyclase via adenosine A1 receptor mediated activation of an inhibitory G protein (Gi). Thus, both acute changes in tubular delivery of NaCl to the macula densa and chronic changes in dietary sodium intake cause appropriate signals to be conveyed from macula densa to the juxtaglomerular cells.

One intrarenal mechanism controlling renin release is called the macula densa pathway (top of Figure 312 A). The macula densa lies adjacent to the juxtaglomerular cells and is composed of specialized columnar epithelial cells located in the wall of that portion of the cortical thick ascending limb that passes between the afferent and efferent arterioles of the glomerulus (see Figure 291). A change in NaCl reabsorption by the macula densa results in the transmission to nearby juxtaglomerular cells of chemical signals that modify renin release. Increases in NaCl flux across the macula densa inhibit renin release, and decreases in NaCl flux stimulate renin release. The chemical signals mediating the macula densa pathway involve both adenosine (Itoh et al., 1985; Weihprecht et al., 1990) and prostaglandins (Gerber et al., 1981; Greenberg et al., 1993), with the former being released when NaCl transport increases and the latter being released when NaCl transport decreases. In this regard, adenosine, acting via an A1 adenosine receptor, inhibits renin release (Jackson, 1991), and prostaglandins stimulate renin release (Jackson et al., 1982). Several lines of evidence suggest a critical role for inducible cyclooxygenase (COX-2) and neuronal nitric oxide synthase (nNOS) in the mechanism of macula densastimulated renin release. Although constitutive cyclooxygenase (COX-1) is the most abundant cyclooxygenase isoform in the mammalian kidney, inducible COX-2 is the only cyclooxygenase form expressed in the macula densa, and the amount of COX-2 in the macula densa is upregulated by chronic dietary sodium restriction (Harris et al., 1994). Moreover, selective inhibition of COX-2 blocks macula densa

mediated renin release (Traynor et al., 1999). Like COX-2, nNOS is expressed in the macula densa; nNOS expression in the macula densa is upregulated by dietary sodium restriction (Singh et al., 1996), and selective inhibition of nNOS reduces renin release in response to chronic dietary sodium restriction (Beierwaltes, 1997). Together, these findings suggest a biochemical interplay between COX-2 and nNOS in the mechanism of macula densamediated renin release. Since nitric oxide reacts with superoxide anion to generate peroxynitrite, and peroxynitrite markedly activates cyclooxygenase activity (Landino et al., 1996), it is conceivable that activation of macula densa mediated renin release by sodium depletion involves the following events: upregulation of nNOS and COX-2 in the macula densa by sodium depletion; increased nitric oxide, and hence peroxynitrite, biosynthesis in the macula densa; peroxynitrite-induced activation of COX-2 in the macula densa; increased prostaglandin production in the macula densa; activation of prostaglandin receptors in the juxtaglomerular cells by prostaglandins released from the macula densa. Possible mechanisms by which the macula densa regulates renin release are summarized in Figure 312 B. Although a change in NaCl transport by the macula densa is the key event that modulates the macula densa pathway, regulation of this pathway is more dependent on the luminal concentration of Cl than Na +. NaCl transport into the macula densa is mediated by the Na+ K+2 Cl symporter, and the half-maximal concentrations of Na+ and Cl required for transport via this symporter are 2 to 3 mEq/liter and 40 mEq/liter, respectively. Since the luminal concentration of Na+ at the macula densa is usually much greater than the level required for half-maximal transport, physiological variations in luminal Na+ concentrations at the macula densa have little effect on renin release (i.e., the symporter remains saturated with respect to Na+ ). On the other hand, physiological changes in Cl concentration at the macula densa have profound effects on macula densamediated renin release (Lorenz et al., 1991). The second intrarenal mechanism controlling renin release is called the intrarenal baroreceptor pathway (middle of Figure 312 A). Increases and decreases in blood pressure in the preglomerular vessels inhibit and stimulate renin release, respectively. The immediate stimulus to secretion is believed to be a reduction in the tension within the wall of the afferent arteriole. Increases and decreases in renal perfusion pressure may inhibit and stimulate, respectively, the release of renal prostaglandins, which perhaps mediate in part the intrarenal baroreceptor pathway (Data et al., 1978; Linas, 1984). In fact, in renin-dependent renovascular hypertension, renin secretion and blood pressure are reduced by selective inhibition of COX-2 (Wang et al., 1999). However, biomechanical coupling via stretch-activated ion channels may play an important role in the intrarenal baroreceptor pathway (Carey et al., 1997). The third mechanism, called the -adrenergic receptor pathway (bottom of Figure 312 A), is mediated by the release of norepinephrine from postganglionic sympathetic nerve terminals; activation of 1-receptors on juxtaglomerular cells enhances renin secretion. The three pathways regulating renin release are embedded in a physiological network. Increases in renin secretion enhance the formation of angiotensin II, and angiotensin II stimulates angiotensin subtype 1 (AT1) receptors on juxtaglomerular cells to inhibit renin release. This feedback system has been termed the short-loop negative feedback mechanism. Angiotensin II also increases arterial blood pressure via stimulation of AT1 receptors (see below). Increases in blood pressure inhibit renin release by: (1) activating high-pressure baroreceptors, thereby reducing renal sympathetic tone; (2) increasing pressure in the preglomerular vessels; and (3) reducing NaCl reabsorption in the proximal tubule (pressure natriuresis), which increases tubular delivery of NaCl to the macula densa. The inhibition of renin release due to angiotensin IIinduced increases in blood pressure has been termed the long-loop negative feedback mechanism.

The physiological pathways regulating renin release can be influenced by a number of pharmacological agents. Loop diuretics (Chapter 29: Diuretics) stimulate renin release in part by blocking the reabsorption of NaCl at the macula densa. Nonsteroidal antiinflammatory drugs (NSAIDs; Chapter 27: Analgesic-Antipyretic and Antiinflammatory Agents and Drugs Employed in the Treatment of Gout) inhibit the formation of prostaglandins and thereby decrease renin release (Frlich et al., 1979). ACE inhibitors, AT1 receptor blockers, and renin inhibitors interrupt both the short- and long-loop negative feedback mechanisms and therefore increase renin release. In general, diuretics and vasodilators increase renin release by decreasing arterial blood pressure. Centrally acting sympatholytic drugs as well as -adrenergic receptor antagonists decrease renin secretion by inhibiting the -adrenergic receptor pathway. Angiotensinogen The substrate for renin is angiotensinogen, an abundant globular glycoprotein (MW = 55,000 to 60,000 Da), containing 13% to 14% carbohydrate. High-molecular-weight (350,000 to 500,000 Da) angiotensinogen also circulates in plasma and represents a complex of angiotensinogen with other proteins. The relevant portion of angiotensinogen is the amino terminus, from which angiotensin I is cleaved. Molecular cloning studies (Kageyama et al., 1984) demonstrate that human angiotensinogen contains 452 amino acids and is synthesized as preangiotensinogen, which has a 24 or 33amino acid signal peptide. Angiotensinogen is synthesized primarily in the liver, although mRNA that encodes the protein also is abundant in fat, certain regions of the CNS, and kidney (Campbell and Habener, 1986; Cassis et al., 1988). Angiotensinogen is continuously synthesized and secreted by the liver, and its synthesis is stimulated by inflammation, insulin, estrogens, glucocorticoids, thyroid hormone, and angiotensin II (Ben-Ari and Garrison, 1988). During pregnancy, plasma levels of angiotensinogen increase severalfold under the influence of estrogen. Circulating levels of angiotensinogen are approximately equal to the K m of renin for its substrate (about 1 M). Consequently, the rate of angiotensin II synthesis, and therefore blood pressure, can be influenced by changes in angiotensinogen levels. For instance, mice with no angiotensinogen gene are hypotensive (Tanimoto et al., 1994), and there is a progressive relationship in genetically engineered mice among the number of copies of the angiotensinogen gene, plasma levels of angiotensinogen, and arterial blood pressure (Kim et al., 1995). Moreover, oral contraceptives increase circulating levels of angiotensinogen and can induce hypertension. Increases in angiotensinogen levels are associated with essential hypertension (Jeunemaitre et al., 1992), and there is genetic linkage between essential hypertension, angiotensinogen levels, and the angiotensinogen gene (Jeunemaitre et al., 1992; Caulfield et al., 1994, 1995). Furthermore, a specific mutation in the angiotensinogen gene (a methionine to threonine point mutation at position 235 of angiotensinogen) is associated with increased plasma levels of angiotensinogen and essential (Jeunemaitre et al., 1992; Kunz et al., 1997; Staessen et al., 1999) and pregnancy-induced (Ward et al., 1993) hypertension. Angiotensin Converting Enzyme (ACE; Kininase II; Dipeptidyl Carboxypeptidase) ACE is an ectoenzyme and a glycoprotein with an apparent MW of 170,000 Da. Human ACE contains 1277 amino acid residues and has two homologous domains, each with a catalytic site and 2+ a region for binding Zn (Soubrier et al., 1988; Bernstein et al., 1989). ACE has a large aminoterminal extracellular domain, a short carboxyl-terminal intracellular domain, and a 17amino acid hydrophobic stretch that anchors the ectoenzyme to the cell membrane. Circulating ACE represents membrane ACE that has undergone proteolysis at the cell surface by a secretase (Beldent et al., 1995). ACE is rather nonspecific and cleaves dipeptide units from substrates with diverse amino

acid sequences. Preferred substrates have only one free carboxyl group in the carboxyl-terminal amino acid, and proline must not be the penultimate amino acid; thus the enzyme does not degrade angiotensin II. Bradykinin is one of the many natural substrates for ACE, and ACE is identical to kininase II, which inactivates bradykinin and other potent vasodilator peptides. Although slow conversion of angiotensin I to angiotensin II occurs in plasma, the very rapid metabolism that occurs in vivo is due largely to the activity of membrane-bound ACE present on the luminal aspect of endothelial cells throughout the vascular system. The ACE gene codes for both a somatic and a testis-specific isozyme. The testis ACE is found exclusively in developing spermatids and mature sperm and is encoded by the second half of the ACE gene, driven by a testis-specific promoter. Studies by Hagaman et al. (1998) demonstrate that testis ACE is involved in the transport of sperm in the oviduct and in binding of the sperm to the zonae pellucidae. These effects of testis-specific ACE are not mediated by angiotensin II. The ACE gene contains, in intron 16, an insertion/deletion polymorphism that explains 47% of the phenotypic variance in serum ACE levels (Rigat et al., 1990). The deletion allele is associated with higher levels of serum ACE and may confer an increased risk of ischemic heart disease (Cambien et al., 1992; Gardemann et al., 1995; Mattu et al., 1995), coronary artery spasm (Oike et al., 1995), restenosis after coronary stenting (Amant et al., 1997; Ribichini et al., 1998), vascular endothelial dysfunction (Butler et al., 1999), left ventricular hypertrophy (Iwai et al., 1994; Schunkert et al., 1994), exercise-induced left ventricular growth (Montgomery et al., 1997), carotid artery hypertrophy (Hosoi et al., 1996), ischemic stroke (Kario et al., 1996), hypertension in males (O'Donnell et al., 1998; Fornage et al., 1998), diabetic nephropathy (Marre et al., 1997), deterioration of renal function in IgA nephropathy (Yoshida et al., 1995), renal artery stenosis (Olivieri et al., 1999), and thrombosis in patients undergoing hip arthroplasy (Philipp et al., 1998). Surprisingly, however, the deletion allele is more frequent in centenarians (Schachter et al., 1994), a finding that may be explained by the strong association of the deletion allele with protection against Alzheimer's disease (Kehoe et al., 1999). Angiotensin Peptides When given intravenously, angiotensin I is so rapidly converted to angiotensin II that the pharmacological responses of these two peptides are indistinguishable. However, angiotensin I per se is less than 1% as potent as angiotensin II on smooth muscle, heart, and the adrenal cortex. As shown in Figure 311, angiotensin III, also called [des-Asp1] angiotensin II or angiotensin (28), can be formed either by the action of aminopeptidase on angiotensin II or by the action of ACE on [des-Asp1] angiotensin I. Angiotensin III and angiotensin II cause qualitatively similar effects. Angiotensin III is approximately as potent as angiotensin II in stimulating the secretion of aldosterone; however, angiotensin III is only 25% and 10% as potent as angiotensin II in elevating blood pressure and stimulating the adrenal medulla, respectively (Peach, 1977; Bell et al., 1984). Angiotensin I can be metabolized to angiotensin (17) by the enzymes metalloendopeptidase 24.15, endopeptidase 24.11, and prolylendopeptidase 24.26, and angiotensin II can be converted to angiotensin (17) by prolylcarboxypeptidase (Ferrario et al., 1997). ACE inhibitors increase, rather than decrease, tissue and plasma levels of angiotensin (17), because angiotensin I levels are increased and diverted away from angiotensin II formation (Figure 311) and because ACE contributes importantly to the plasma clearance of angiotensin (17) (Yamada et al., 1998). The pharmacological profile of angiotensin (17) is distinct from that of angiotensin II. Unlike angiotensin II, angiotensin (17) does not cause vasoconstriction, aldosterone release, or facilitation of noradrenergic neurotransmission. Angiotensin (17) releases vasopressin, stimulates prostaglandin biosynthesis, elicits depressor responses when microinjected into certain brainstem

nuclei, dilates some blood vessels, and exerts a natriuretic action on the kidneys. Angiotensin (17) also inhibits proliferation of vascular smooth muscle cells (Tallant et al., 1999). The effects of angiotensin (17) may be mediated by a specific angiotensin (17) receptor (Tallant et al., 1997). Ferrario et al. (1997) propose that angiotensin (17) serves to counterbalance the actions of angiotensin II. Putative receptors for angiotensin (38), also called angiotensin IV, are detectable in a number of tissues (Swanson et al., 1992), and angiotensin (38) stimulates the expression of plasminogen activator inhibitor 1 in endothelial (Kerins et al., 1995) and proximal tubular (Gesualdo et al., 1999) cells. The physiological significance of both angiotensin (17) and angiotensin (38) remains uncertain. There is considerable information on the structureactivity relationships of angiotensin-related peptides (Samanen and Regoli, 1994). In general, phenylalanine in position 8 is critical for most agonist activity, and the aromatic residues in positions 4 and 6, the guanido group in position 2, and the C-terminal carboxyl are thought to be involved in binding to the receptor site. Position 1 is not critical, but replacement of aspartic acid in position 1 with sarcosine enhances binding to angiotensin receptors and slows hydrolysis by rendering the peptide refractory to a subgroup of aminopeptidases (angiotensinase A). Such a substitution, combined with that of alanine or isoleucine in place of phenylalanine in position 8, yields potent angiotensin II receptor antagonists. Angiotensinases This term is applied to various peptidases that are involved in the degradation and inactivation of angiotensin peptides; none is specific. Among them are aminopeptidases, endopeptidases, and carboxypeptidases. Local (Tissue) ReninAngiotensin Systems The traditional view of the reninangiotensin system is that of a classical endocrine system. Circulating renin of renal origin acts on circulating angiotensinogen of hepatic origin to produce angiotensin I in the plasma; circulating angiotensin I is converted by plasma ACE and by pulmonary endothelial ACE to angiotensin II; angiotensin II is then delivered to its target organs via the bloodstream, where it induces a physiological response. Recent evidence suggests that this traditional view is an oversimplification and should be expanded to include local (tissue) renin angiotensin systems. In this regard, it is important to distinguish between extrinsic and intrinsic, local reninangiotensin systems. Extrinsic, Local ReninAngiotensin Systems Since ACE is present on the luminal face of vascular endothelial cells throughout the circulation, and since circulating renin of renal origin can be taken up (sequestered) by the arterial wall as well as by other tissues, the conversion of hepatic angiotensinogen to angiotensin I and the conversion of angiotensin I (both circulating and locally produced) to angiotensin II may occur primarily within or at the surface of the blood vessel wall, not in the circulation per se. Indeed, studies by Danser et al. (1991; 1994) demonstrate that many vascular beds produce angiotensins I and II locally and that a substantial fraction of local production does not occur in the plasma as it transverses the vascular bed. In this regard, local sequestration of renal renin in both vascular and cardiac tissues participates in the local production of angiotensins (Kato et al., 1993; Taddei et al., 1993; Danser et al., 1994). However, studies by Hu et al. (1998) do not support an important role for tissue binding of circulating renin.

Intrinsic, Local ReninAngiotensin Systems Many tissuesincluding the brain, pituitary, blood vessels, heart, kidney, and adrenal gland express mRNAs for renin, angiotensinogen, and/ or ACE, and various cultured cell types from these tissues produce renin, angiotensinogen, ACE, and/or angiotensins I, II, and III (Phillips et al., 1993; Saavedra, 1992; Dzau, 1993; Baker et al., 1992). Therefore, it appears that local reninangiotensin systems exist independently of the renal/hepatic-based system. Although these local systems do not contribute significantly to circulating levels of active renin or angiotensins (Campbell et al., 1991), there is evidence that the local production of angiotensin II by intrinsic, local reninangiotensin systems influences vascular, cardiac, and renal function and structure. This hypothesis, although controversial (von Lutterotti et al., 1994), is an active area of investigation. Alternative Pathways for Angiotensin Biosynthesis Some tissues contain nonrenin angiotensinogen-processing enzymes that convert angiotensinogen to angiotensin I (nonrenin proteases) or directly to angiotensin II (e.g., cathepsin G, tonin) and nonACE angiotensin Iprocessing enzymes that convert angiotensin I to angiotensin II (e.g., cathepsin G, chymostatin-sensitive angiotensin IIgenerating enzyme, heart chymase) (Dzau et al., 1993). There is mounting evidence that chymase, possibly mast cellderived, importantly contributes to the local tissue conversion of angiotensin I to angiotensin II, particularly in the heart (Wolny et al., 1997; Wei et al., 1999) and kidneys (Hollenberg et al., 1998); however, the role of chymase as a non-ACE angiotensin Iprocessing enzyme is species- and organ-dependent (Akasu et al., 1998). Angiotensin Receptors The effects of angiotensins are exerted through specific cell surface receptors. Studies by Whitebread et al. (1989) and Chiu et al. (1989) have identified two subtypes of angiotensin receptors that are now designated AT1 and AT2 (Bumpus et al., 1991). The AT1 receptor has a high affinity for losartan (and related biphenyl tetrazole derivatives), a low affinity for PD 123177 (and related 1-benzyl spinacine derivatives), and a low affinity for CGP 42112A (a peptide analog). In contrast, the AT2 receptor has a high affinity for PD 123177 and CGP 42112A but a low affinity for losartan. Both the AT1 (Sasaki et al., 1991; Murphy et al., 1991) and AT2 (Mukoyama et al., 1993) receptors are members of the G proteincoupled receptor family (see Chapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect) with seven putative transmembrane regions. The AT1 receptor is 359 amino acids long, and the AT2 receptor is 363 amino acids long. The AT1 and AT2 receptors have little sequence homology. Most of the biological effects of angiotensin II are mediated by the AT1 receptor; functional roles for the AT2 receptor are poorly defined. Nonetheless, evidence suggests that AT 2 receptors may exert antiproliferative, proapoptotic, and vasodilatory effects (Inagami et al., 1999; Ardaillou, 1999; Horiuchi et al., 1999). The AT2 receptor is widely distributed in fetal tissues, but its distribution is more restricted in adults. In adults, some tissues contain primarily either AT1 receptors or AT2 receptors, whereas other tissues contain the receptor subtypes in similar amounts. In this regard, tissue and species differences are the rule, not the exception (Timmermans et al., 1993). The AT1 receptor gene contains a polymorphism (A to C transversion at position 1166) that may be associated with hypertension (Kainulainen et al., 1999), hypertrophic cardiomyopathy (Osterop et al., 1998), coronary artery vasoconstriction (Amant et al., 1997), and aortic stiffness (Benetos et al., 1996). Moreover, the C allele synergizes with the ACE deletion allele with regard to increased risk of

coronary artery disease (Tiret et al., 1994; lvarez et al., 1998). Preeclampsia is associated with development of agonistic autoantibodies against the AT1 receptor (Wallukat et al., 1999). Angiotensin ReceptorEffector Coupling AT1 receptors activate a large array of signal transduction systems, including calcium release and influx pathways, phospholipases, mitogen-activated protein kinase pathways, Janus kinase pathways, serine/ threonine protein kinases, nonreceptor tyrosine kinases, small GTP-binding proteins, inducible transcription factors, factors affecting translational efficiency, and pathways leading to the production of reactive oxygen species (Griendling et al., 1997; Berk, 1999; Inagami, 1999; Blume et al., 1999) (Figure 313). The most proximate coupling of AT 1 receptors to signal transduction systems is mediated by heterotrimeric GTP-binding proteins (G proteins; see Chapter 2: Pharmacodynamics: Mechanisms of Drug Action and the Relationship Between Drug Concentration and Effect) such as G q, Gi, G12, and G13 . A few proximate interactions engage a network of signal transduction systems as the initial signals fan out to include a wide array of signaling processes. Stimulation of AT1 receptors leads to activation, via the G protein Gq, of phospholipase C- . Phospholipase C- is a membrane-bound enzyme that hydrolyzes phosphatidylinositol-4,5-bisphosphate to generate inositol-1,4,5-trisphosphate (IP3) and diacylglycerol. IP3 binds to receptors on Ca2+-release channels in the IP 3-sensitive Ca2+ stores, an event that triggers the intracellular release of Ca2+. Additional Ca2+ enters the cell from outside due to opening of voltage-sensitive Ca2+ channels located in the cell membrane. Ca2+ binds to calmodulin, and the Ca2+/calmodulin complex activates a number of intracellular enzymes, such as Ca2+-calmodulindependent protein kinases, that contribute to the ultimate cellular response. Although the initial surge of intracellular diacylglycerol is derived from phosphatidylinositol-4,5bisphosphate, increases in diacylglycerol are sustained via activation of phospholipase D, which hydrolyzes phosphatidylcholine to generate phosphatidic acid. Phosphatidic acid, which has important cellular effects per se, is transformed into diacylglycerol by the enzyme phosphatidate phosphohydrolase. Ca2+ and/or diacylglycerol activate a family of serine/threonine kinases called 2+ protein kinase C. Activation of protein kinase C and increases in intracellular Ca lead to the phosphorylation and activation of a network of signaling pathways, including nonreceptor tyrosine kinases. Nonreceptor tyrosine kinases, in turn, activate phospholipase C- , leading to additional IP3 and diacylglycerol generation, and engage the mitogen-activated protein kinase pathways that participate importantly in the regulation of gene expression and mRNA translational efficiency. Some mitogen-activated protein kinases activate cytoplasmic phospholipase A2, an enzyme that releases arachidonic acid from phospholipid stores and thereby increases the biosynthesis of biologically active substances such as prostaglandins, hydroxyeicosatetraenoic acids, leukotrienes, and epoxides and diols derived from arachidonic acid. Gene transcription also is regulated by the AT1 receptor via activation of the Janus kinase/signal transducers and activators of transcription pathways. Although the precise mechanism by which the AT1 receptor couples to this pathway is unknown, Janus kinases and signal transducers and activators of transcription physically associate with the cytoplasmic tail of the AT1 receptor. Gene transcription results in the expression of a number of inducible transcription factors that, in turn, orchestrate changes in the expression of a host of gene products that regulate cell growth and extracellular matrix protein biosynthesis. In some cell types, AT1 receptor activation decreases intracellular cyclic AMP by inhibiting adenylyl cyclase. This effect, which is mediated by the inhibitory G protein Gi, reduces the activity of protein kinase A and thereby decreases the phosphorylation state of substrates for protein kinase A. AT1 receptors also stimulate the activity of a membrane-bound NADH/NADPH oxidase that generates superoxide anion. Superoxide anion is transformed to hydrogen peroxide via catalase, and hydrogen peroxide generated by this pathway may induce important effects, such as activation of mitogenactivated protein kinase pathways and expression of monocyte chemoattractant protein-1. The relative importance of each of the aforementioned signal transduction pathways in mediating a

given biological response to angiotensin II depends on the specific tissue and response under investigation. Although AT1 receptoreffector coupling mechanisms have been studied in considerable detail, much less is known regarding AT2 receptoreffector coupling. Signal transduction mechanisms for the AT2 receptor include activation of phosphatases, potassium channels, and nitric oxide production and inhibition of calcium channels (Horiuchi et al., 1999). The effects of AT2 receptors are mediated mostly by Gi . Figure 313. Mechanisms of AT1 ReceptorEffector Coupling. The general categories of signal transduction molecules activated by the AT1 receptor are summarized. The most proximate coupling of AT 1 receptor to signal transduction systems is mediated by heterotrimeric G proteins (such as Gq, Gi, G12, and G13 ). However, only some of the indicated systems directly couple with the AT1 receptor via heterotrimeric G proteins. Rather, a few such proximate interactions engage a network of signal transduction systems as the initial signals fan out to include a wide array of signaling processes. Abbreviations: AT1-R, angiotensin type I receptor; Gq, Gi, G12, and G13 , different types of heterotrimeric GTPbinding proteins; RhoA, ARF, Rac, and p21ras, different types of smallmolecular-weight GTP-binding proteins; PKA (protein kinase A), PKC (protein 2+ kinase C), and Ca -CaM-dependent kinase, different types of serine/threonine protein kinases; Ca2+-CaM-, calcium bound to calmodulin; PLC- , PLC- , PLD, and PLA2, different types of phospholipases; PAK, Pky2, pp60c-src, and pp125FAK, different types of nonreceptor tyrosine kinases; ERK1, ERK2, JNK, and p38, different types of mitogen-activated protein (MAP) kinases; c-Fos, FosB, c-Jun, JunB, JunD, Krox-20, Krox-24, c-Myc, family of inducible transcription factors; pp90rsk and p70s6K, S6 kinases (S6 is a protein that binds to ribosomes and regulates translation); Jak2 and Tyk2, members of the Janus kinase family of tyrosine kinases; Stat1 and Stat2, members of the family of signal transducers and activators of transcription; O2, superoxide anion.

Functions of the ReninAngiotensin System The reninangiotensin system plays a major role in regulating arterial blood pressure over both the short and long term. Modest changes in plasma concentrations of angiotensin II acutely increase blood pressure; on a molar basis, angiotensin II is approximately 40 times more potent than norepinephrine in this regard. When a single moderate dose of angiotensin II is injected intravenously, systemic blood pressure begins to rise within seconds, rapidly reaches maximum, and returns to normal within minutes. This rapid pressor response to angiotensin II is due to a swift increase in total peripheral resistancea response that helps maintain arterial blood pressure in the face of an acute hypotensive challenge (e.g., blood loss, vasodilation). Although angiotensin II directly increases cardiac contractility (via opening voltage-gated Ca2+ channels in cardiac myocytes) and indirectly increases heart rate (via facilitation of sympathetic tone, enhanced noradrenergic neurotransmission, and adrenal catecholamine release), the rapid increase in arterial blood pressure activates a baroreceptor reflex that decreases sympathetic tone and increases vagal tone. Thus, angiotensin II may increase, decrease, or not change cardiac contractility, heart rate, and cardiac output, depending on the physiological state. Therefore, changes in cardiac output contribute little if at all to the rapid pressor response induced by angiotensin II. Angiotensin II also causes a slow pressor response that helps stabilize arterial blood pressure over the long term. A continuous infusion of initially subpressor doses of angiotensin II gradually increases arterial blood pressure, with the maximum effect requiring days to achieve (Brown et al., 1981). Most likely, the slow pressor response to angiotensin II is mediated by a decrement in renal excretory function that shifts the renal pressurenatriuresis curve to the right (see below). Angiotensin IIinduced stimulation of endothelin-1 (Laursen et al., 1997) and superoxide anion (Rajagopalan et al., 1997) production mediates, in part, the slow pressor response.

In addition to buffering short- and long-term changes in arterial blood pressure, angiotensin II significantly alters the morphology of the cardiovascular system; i.e., it causes hypertrophy of vascular and cardiac cells. The pathophysiological implications of this effect of the renin angiotensin system are under intensive investigation. The effects of angiotensin II on total peripheral resistance, renal function, and cardiovascular structure are mediated by a number of direct and indirect mechanisms. Figure 314 summarizes the three major effects of angiotensin II and how they are mediated. Figure 314. Summary of the Three Major Effects of Angiotensin II and the Mechanisms That Mediate Them. Abbreviation: NE, norepinephrine.

Mechanisms by Which Angiotensin II Increases Total Peripheral Resistance Angiotensin II increases total peripheral resistance (TPR) via direct and indirect effects on blood vessels. Direct Vasoconstriction Angiotensin II constricts precapillary arterioles and, to a lesser extent, postcapillary venules by

activating AT1 receptors located on vascular smooth muscle cells. Angiotensin II has differential effects on the tone of vascular beds throughout the circulation. Direct vasoconstriction is strongest in the kidneys and somewhat less in the splanchnic vascular bed; blood flow in these regions falls sharply when angiotensin II is infused. Angiotensin IIinduced vasoconstriction is much less in vessels of the brain and still weaker in those of the lung and skeletal muscle. In these regions, blood flow actually may increase, especially following small changes in the concentration of the peptide, because the relatively weak vasoconstrictor response is opposed by the elevated systemic blood pressure. Nevertheless, with high circulating concentrations of angiotensin II, cerebral and coronary blood flow may decrease. Enhancement of Peripheral Noradrenergic Neurotransmission Angiotensin II facilitates peripheral noradrenergic neurotransmission by augmenting norepinephrine release from sympathetic nerve terminals, by inhibiting the reuptake of norepinephrine into nerve terminals, and by enhancing the vascular response to norepinephrine (see Jackson et al., 1985). High concentrations of the peptide stimulate ganglion cells directly. Facilitation of noradrenergic neurotransmission by endogenous angiotensin II occurs in animals with renin-dependent renovascular hypertension (Zimmerman et al., 1987), and in human beings, intracoronary angiotensin II potentiates sympathetic nervous systeminduced coronary vasoconstriction (Saino et al., 1997). Effects on the Central Nervous System Small amounts of angiotensin II infused into the vertebral arteries cause an increase in arterial blood pressure. This effect is mediated by increased sympathetic outflow due to an effect of the hormone on circumventricular nuclei that are not protected by a bloodbrain barrier (area postrema, subfornical organ, and organum vasculosum of the lamina terminalis). Blood-borne angiotensin II also attenuates baroreceptor-mediated reductions in sympathetic discharge, thereby increasing arterial pressure. The CNS is affected both by blood-borne angiotensin II and by angiotensin II formed within the brain (see Saavedra, 1992; Bunnemann et al., 1993). The brain contains all the components of a reninangiotensin system. Moreover, there is angiotensin-like immunoreactivity at many sites within the CNS, suggesting that angiotensin II serves as a neurotransmitter or modulator. In addition to increasing sympathetic tone, angiotensin II also causes a centrally mediated dipsogenic effect (see Fitzsimons, 1980) and enhances the release of vasopressin from the neurohypophysis (see Ganong, 1984). Increased drinking and vasopressin secretion are produced more consistently following intraventricular than intravenous injections. Release of Catecholamines from the Adrenal Medulla Angiotensin II stimulates the release of catecholamines from the adrenal medulla by depolarizing chromaffin cells. Although this response is of minimal physiological importance, intense and dangerous reactions have followed the administration of angiotensin II to individuals with pheochromocytoma. Mechanisms by Which Angiotensin II Alters Renal Function Angiotensin II has pronounced effects on renal function to reduce the urinary excretion of Na+ and water while increasing the excretion of K+ . The overall effect of angiotensin II on the kidneys is to shift the renal pressurenatriuresis curve to the right (see below). Like the effects of angiotensin II on TPR, the effects of this peptide on renal function are multifaceted.

Direct Effects of Angiotensin II on Sodium Reabsorption in the Proximal Tubule Very low concentrations of angiotensin II stimulate Na+/H + exchange in the proximal tubulean effect that increases Na+, Cl , and bicarbonate reabsorption. Approximately 20% to 30% of the bicarbonate handled by the nephron may be affected by this mechanism (Liu and Cogan, 1987). Paradoxically, at high concentrations, angiotensin II may inhibit Na+ transport in the proximal tubule. Release of Aldosterone from the Adrenal Cortex Angiotensin II stimulates the zona glomerulosa of the adrenal cortex to increase the synthesis and secretion of aldosterone, and angiotensin II exerts trophic and permissive effects that augment other stimuli (e.g., ACTH and K+). Increased output of aldosterone is elicited by very low concentrations of angiotensin II that have little or no acute effect on blood pressure. As described in Chapter 29: Diuretics, aldosterone acts on the distal and collecting tubules to cause retention of Na+ and excretion of K+ and H+. The stimulant effect of angiotensin II on the synthesis and release of aldosterone is enhanced under conditions of hyponatremia or hyperkalemia and is reduced when concentrations of Na+ and K+ in plasma are altered in the opposite direction. Such changes in sensitivity are due in part to alterations in the number of receptors for angiotensin II on zona glomerulosa cells as well as to adrenocortical hyperplasia in the Na+-depleted state. Altered Renal Hemodynamics Reductions in renal blood flow markedly attenuate renal excretory function, and angiotensin II reduces renal blood flow by directly constricting the renal vascular smooth muscle, by enhancing renal sympathetic tone (a CNS effect), and by facilitating renal noradrenergic neurotransmission (an intrarenal effect). Autoradiographic and in situ hybridization studies indicate a high concentration of AT1 receptors in the vasa recta of the renal medulla, and angiotensin II may reduce Na+ excretion in part by diminishing medullary blood flow. Angiotensin II also influences glomerular filtration rate (GFR); however, this effect of angiotensin II is variable. Angiotensin II exerts several effects that may alter GFR: (1) constriction of the afferent arterioles, which reduces intraglomerular pressure and tends to reduce GFR; (2) contraction of mesangial cells, which decreases the capillary surface area within the glomerulus available for filtration and also tends to decrease GFR; and (3) constriction of efferent arterioles, which increases intraglomerular pressure and tends to increase GFR. The outcome of these opposing effects on GFR depends on the physiological state. Normally, GFR is slightly reduced by angiotensin II; however, during renal artery hypotension, the effects of angiotensin II on the efferent arteriole predominate so that, in this setting, angiotensin II increases GFR (Hall et al., 1981; Kastner et al., 1984). Thus, blockade of the reninangiotensin system may cause acute renal failure in patients with bilateral renal artery stenosis or in patients with unilateral stenosis who have only a single kidney (Hricik et al., 1983). Mechanisms by Which Angiotensin II Alters Cardiovascular Structure Several cardiovascular diseases are accompanied by changes in the morphology of the heart and/or blood vessels, and these changes pose an increased risk of morbidity and mortality. Pathological alterations in cardiovascular structures may involve hypertrophy (an increase in tissue mass) and/or remodeling (redistribution of mass within a structure). Examples include (1) increased wall-tolumen ratio in blood vessels (associated with hypertension); (2) cardiac concentric hypertrophy (also associated with hypertension); (3) cardiac eccentric hypertrophy and cardiac fibrosis (associated with congestive heart failure and myocardial infarction); and (4) thickening of the intimal surface of the blood vessel wall (associated with atherosclerosis and angioplasty). These

morbid changes in cardiovascular structure are due to increased migration, proliferation (hyperplasia), and hypertrophy of cells as well as to increased extracellular matrix. The cells involved include vascular smooth muscle cells, cardiac myocytes, and fibroblasts. The renin angiotensin system may contribute importantly to the aforementioned morbid changes in cardiovascular structure. In this regard, angiotensin II: (1) stimulates the migration (Bell and Madri, 1990; Dubey et al., 1995), proliferation (Daemen et al., 1991), and hypertrophy of vascular smooth muscle cells (Itoh et al., 1993); (2) increases extracellular matrix production by vascular smooth muscle cells (Scott-Burden et al., 1990); (3) causes hypertrophy of cardiac myocytes (Baker et al., 1992); and (4) increases extracellular matrix production by cardiac fibroblasts (Villarreal et al., 1993; Crawford et al., 1994). Nonhemodynamically Mediated Effects of Angiotensin II on Cardiovascular Structure Angiotensin II stimulates migration, proliferation, hypertrophy, and/or synthetic capacity of vascular smooth muscle cells, cardiac myocytes, and/or fibroblasts in part by acting directly on cells to induce the expression of specific proto-oncogenes. In cell culture, angiotensin II rapidly (within minutes) increases steady-state levels of mRNA for the protooncogenes c-fos, c-jun, c-myc, and egr1. FOS and JUN, the proteins coded by c-fos and c-jun, combine to form AP-1, and AP-1 alters the expression of several genes involved in stimulating cell growth (hypertrophy and hyperplasia), including basic fibroblast growth factor, platelet-derived growth factor, and transforming growth factor . In addition, the expression of genes coding for extracellular matrix proteins, such as collagen, fibronectin, and tenascin, is increased. Hemodynamically Mediated Effects of Angiotensin II on Cardiovascular Structure In addition to the direct cellular effects of angiotensin II on cardiovascular structure, changes in cardiac preload (volume expansion due to Na+ retention) and afterload (increased arterial blood pressure) probably contribute to cardiac hypertrophy and remodelling. Arterial hypertension also contributes to hypertrophy and remodeling of blood vessels. Role of the ReninAngiotensin System in Long-Term Maintenance of Arterial Blood Pressure Despite Extremes in Dietary Na + Intake Arterial blood pressure is a major determinant of Na+ excretion (see Guyton, 1990). This can be illustrated graphically by plotting urinary Na+ excretion versus mean arterial blood pressure (Figure 315), a plot known as the renal pressure natriuresis curve. Over the long term, Na+ excretion must equal Na+ intake; therefore, the set point for long-term levels of arterial blood pressure can be obtained as the intersection of a horizontal line representing Na+ intake with the renal pressure natriuresis curve (Guyton, 1991) (Figure 315). If the renal pressurenatriuresis curve were fixed, then long-term levels of arterial blood pressure would be greatly affected by dietary Na+ intake. However, as illustrated in Figure 315, the reninangiotensin system plays a major role in maintaining a constant set point for long-term levels of arterial blood pressure despite extreme changes in dietary Na+ intake. When dietary Na+ intake is low, renin release is stimulated, and angiotensin II acts on the kidneys to shift the renal pressurenatriuresis curve to the right. Conversely, when dietary Na+ is high, renin release is inhibited, and the withdrawal of angiotensin II causes the renal pressurenatriuresis curve to shift to the left. Consequently, the intersection of salt intake with the renal pressurenatriuresis curve remains near the same set point despite large swings in dietary Na+ intake. When modulation of the reninangiotensin system is pharmacologically prevented, changes in salt intake markedly affect long-term levels of arterial blood pressure (Hall et al., 1980).

Figure 315. Interactions among Salt Intake, the Renal PressureNatriuresis Mechanism, and the ReninAngiotensin System to Stabilize Long-Term Levels of Arterial Blood Pressure Despite Large Variations in Dietary Sodium Intake. (Modified from Jackson et al., 1985, with permission.)

Other Roles of ReninAngiotensin System Expression of the reninangiotensin system is required for the development of normal kidney morphology, particularly the maturational growth of the renal papilla (Niimura et al., 1995). Angiotensin II causes a marked anorexigenic effect and weight loss, and high circulating levels of angiotensin II may contribute to the anorexia, wasting, and cachexia of heart failure (Brink et al., 1996) Inhibitors of the ReninAngiotensin System Angiotensin II itself has limited therapeutic utility and is not available for therapeutic use in the United States. Instead, clinical interest focuses on inhibitors of the reninangiotensin system. Angiotensin Converting Enzyme (ACE) Inhibitors History In the 1960s, Ferreira and colleagues found that the venoms of pit vipers contain factors that intensify responses to bradykinin. These bradykinin-potentiating factors proved to be a family of peptides that inhibit kininase II, an enzyme that inactivates bradykinin. Erds and coworkers established that ACE and kininase II are in fact the same enzyme, which catalyzes both the synthesis of angiotensin II, a potent pressor substance, and the destruction of bradykinin, a potent vasodilator. Following the discovery of bradykinin-potentiating factors, the nonapeptide teprotide was synthesized and tested in human subjects. It was found to lower blood pressure in many patients with essential hypertension more consistently than did peptide angiotensin II receptor antagonists, such as saralasin, which have partial agonist activity. Teprotide also exerted beneficial effects in patients with heart failure. These key observations encouraged the search for ACE inhibitors that

would be active orally. The orally effective ACE inhibitor captopril (Cushman et al., 1977) was developed by a rational approach that involved analysis of the inhibitory action of teprotide, inference about the action of ACE on its substrates, and analogy with carboxypeptidase A, which was known to be inhibited by D-benzylsuccinic acid. Ondetti, Cushman, and colleagues argued that inhibition of ACE might be produced by succinyl amino acids that corresponded in length to the dipeptide cleaved by ACE. This hypothesis proved to be true and led ultimately to the synthesis of a series of carboxy alkanoyl and mercapto alkanoyl derivatives that acted as potent competitive inhibitors of ACE (see Petrillo and Ondetti, 1982). Most active was captopril (see Vane, 1999, for an insider's perspective on the history of the discovery of ACE inhibitors). Pharmacological Effects in Normal Laboratory Animals and Human Beings The essential effect of these agents on the reninangiotensin system is to inhibit the conversion of the relatively inactive angiotensin I to the active angiotensin II (or the conversion of [desAsp1]angiotensin I to angiotensin III). Thus, ACE inhibitors attenuate or abolish responses to angiotensin I but not to angiotensin II (see Figure 311). In this regard, ACE inhibitors are highly selective drugs. They do not interact directly with other components of the reninangiotensin system, and the principal pharmacological and clinical effects of ACE inhibitors seem to arise from suppression of synthesis of angiotensin II. Nevertheless, ACE is an enzyme with many substrates, and inhibition of ACE therefore may induce effects unrelated to reducing the levels of angiotensin II. Since ACE inhibitors increase bradykinin levels, and since bradykinin stimulates prostaglandin biosynthesis, bradykinin and/or prostaglandins may contribute to the pharmacological effects of ACE inhibitors. Indeed, Gainer et al. (1998) demonstrated that blockade of bradykinin receptors in humans attenuates the acute blood pressure reduction induced by ACE inhibition. Recent studies, however, fail to demonstrate a role for bradykinin in the vascular or cardiac effects of ACE inhibitors (Davie et al., 1999; Campbell et al., 1999; Rhaleb et al., 1999). ACE inhibitors increase by fivefold the circulating levels of the natural stem-cell regulator N-acetyl-seryl-aspartyl-lysylproline (Ac-SDKP; Azizi et al., 1997); the long-term consequences of this effect are unknown. In addition, ACE inhibitors interfere with both short- and long-loop negative feedbacks on renin release (Figure 312 A). Consequently, ACE inhibitors increase renin release and the rate of formation of angiotensin I. Since the metabolism of angiotensin I to angiotensin II is blocked by ACE inhibitors, angiotensin I is directed down alternative metabolic routes, resulting in the increased production of peptides such as angiotensin (17). Whether or not biologically active peptides such as angiotensin (17) contribute to the pharmacological effects of ACE inhibitors is unknown. In healthy, Na+-replete animals and human beings, a single oral dose of an ACE inhibitor has little effect on systemic blood pressure (Atlas et al., 1983); but repeated doses over several days cause a small reduction in blood pressure. By contrast, even a single dose of these inhibitors lowers blood pressure substantially in normal subjects when they have been depleted of Na+. Clinical Pharmacology Many ACE inhibitors have been synthesized. These drugs can be classified into three broad groups based on chemical structure: (1) sulfhydrylcontaining ACE inhibitors structurally related to captopril (e.g., fentiapril, pivalopril, zofenopril, alacepril); (2) dicarboxyl-containing ACE inhibitors structurally related to enalapril (e.g., lisinopril, benazepril, quinapril, moexipril, ramipril, spirapril, perindopril, pentopril, cilazapril); and (3) phosphorus-containing ACE inhibitors structurally related to fosinopril. Many ACE inhibitors are estercontaining prodrugs that are 100 to 1000 times less

potent ACE inhibitors than the active metabolites but which have a much better oral bioavailability than the active molecules. Currently 12 ACE inhibitors are approved for use (11 marketed) in the United States (see Figure 316 for chemical structures). In general, ACE inhibitors differ with regard to three properties: (1) potency; (2) whether ACE inhibition is due primarily to the drug itself or to conversion of a prodrug to an active metabolite; and (3) pharmacokinetics (i.e., extent of absorption, effect of food on absorption, plasma half-life, tissue distribution, and mechanisms of elimination). Figure 316. Chemical Structures of Selected Angiotensin Converting Enzyme Inhibitors. Captopril, lisinopril, and enalaprilat are active molecules. Benazepril, enalapril, fosinopril, moexipril, perindopril, quinapril, ramipril, and trandolapril are relatively inactive until converted to their corresponding di-acids. The structures enclosed within blue boxes are removed by esterases and replaced with a hydrogen atom to form the active molecule in vivo (e.g., enalapril to enalaprilat or ramipril to ramiprilat).

There is no compelling reason to favor one ACE inhibitor over another, since all ACE inhibitors effectively block the conversion of angiotensin I to angiotensin II and all have similar therapeutic

indications, adverse-effect profiles, and contraindications. However, the Quality-of-Life Hypertension Study Group reported that, although captopril and enalapril are indistinguishable with regard to antihypertensive efficacy and safety, captopril may have a more favorable effect on quality of life (Testa et al., 1993). Since hypertension usually requires lifelong treatment, qualityof-life issues are an important consideration in comparing antihypertensive drugs. ACE inhibitors differ markedly in tissue distribution, and it is possible that this difference could be exploited to inhibit some local reninangiotensin systems while leaving others relatively intact. Whether or not site-specific inhibition actually would confer therapeutic advantages remains to be established. With the notable exceptions of fosinopril and spirapril (which display balanced elimination by the liver and kidneys), ACE inhibitors are cleared predominantly by the kidneys. Therefore, impaired renal function significantly diminishes the plasma clearance of most ACE inhibitors, and dosages of these ACE inhibitors should be reduced in patients with renal impairment. Elevated plasma renin activity (PRA) renders patients hyperresponsive to ACE inhibitorinduced hypotension, and initial dosages of all ACE inhibitors should be reduced in patients with high plasma levels of renin (e.g., patients with heart failure, salt-depleted patients). Captopril (CAPOTEN) Captopril, the first ACE inhibitor to be marketed, is a potent ACE inhibitor with a K i of 1.7 nM. It is the only ACE inhibitor approved for use in the United States that contains a sulfhydryl moiety. Given orally, captopril is rapidly absorbed and has a bioavailability of about 75%. Peak concentrations in plasma occur within an hour, and the drug is cleared rapidly (it has a half-life of approximately 2 hours). Most of the drug is eliminated in urine, 40% to 50% as captopril and the rest as captopril disulfide dimers and captoprilcysteine disulfide. The oral dose of captopril ranges from 6.25 to 150 mg two to three times daily, with 6.25 mg three times daily and 25 mg twice daily being appropriate for the initiation of therapy for heart failure and hypertension, respectively. Most patients should not receive daily doses in excess of 150 mg. Since food reduces the oral bioavailability of captopril by 25% to 30%, the drug should be given 1 hour before meals. Enalapril (VASOTEC) Enalapril maleate, the second ACE inhibitor approved in the United States, is a prodrug that is not highly active and must be hydrolyzed by esterases in the liver to produce the active parent dicarboxylic acid, enalaprilat. Enalaprilat is a highly potent inhibitor of ACE with a K i of 0.2 nM. Although it also contains a "proline surrogate," enalaprilat differs from captopril in that it is an analog of a tripeptide rather than of a dipeptide. Enalapril is rapidly absorbed when given orally and has an oral bioavailability of about 60% (not reduced by food). Although peak concentrations in plasma occur within an hour, enalaprilat concentrations peak only after 3 to 4 hours. Enalapril has a half-life of only 1.3 hours, but enalaprilat, because of tight binding to ACE, has a plasma half-life of about 11 hours. Nearly all of the drug is eliminated by the kidneys either as intact enalapril or enalaprilat. The oral dosage of enalapril ranges from 2.5 to 40 mg daily (single or divided dosage), with 2.5 mg and 5 mg daily being appropriate for the initiation of therapy for heart failure and hypertension, respectively. The initial dose for hypertensive patients who are taking diuretics, are water- or Na +-depleted, or have heart failure is 2.5 mg daily. Enalaprilat (VASOTEC INJECTION) Enalaprilat is not absorbed orally but is available for intravenous administration when oral therapy is not appropriate. For hypertensive patients, the dosage is 0.625 to 1.25 mg given intravenously

over 5 minutes. This dosage may be repeated every 6 hours. Lisinopril (PRINIVIL, ZESTRIL) Lisinopril, the third ACE inhibitor approved for use in the United States, is the lysine analog of enalaprilat; unlike enalapril, lisinopril itself is active. In vitro, lisinopril is a slightly more potent ACE inhibitor than is enalaprilat. Lisinopril is slowly, variably, and incompletely (about 30%) absorbed after oral administration (not reduced by food); peak concentrations in plasma are achieved in about 7 hours. It is cleared as the intact compound by the kidney, and its half-life in plasma is about 12 hours. Lisinopril does not accumulate in tissues. The oral dosage of lisinopril ranges from 5 to 40 mg daily (single or divided dosage), with 5 and 10 mg daily being appropriate for the initiation of therapy for heart failure and hypertension, respectively. A daily dose of 2.5 mg is recommended for patients with heart failure who are hyponatremic or have renal impairment. Benazepril (LOTENSIN) Cleavage of the ester moiety by hepatic esterases transforms benazepril hydrochloride, a prodrug, into benazeprilat, an ACE inhibitor that in vitro is more potent than captopril, enalaprilat, or lisinopril. Benazepril is rapidly, yet incompletely (37%), absorbed after oral administration (only slightly reduced by food). Benazepril is nearly completely metabolized to benazeprilat and to the glucuronide conjugates of benazepril and benazeprilat, which are excreted into both the urine and bile; peak concentrations of benazepril and benazeprilat in plasma are achieved in about 0.5 to 1 hour and 1 to 2 hours, respectively. Benazeprilat has an effective half-life in plasma of about 10 to 11 hours. With the exception of the lungs, benazeprilat does not accumulate in tissues. The oral dosage of benazepril ranges from 5 to 80 mg daily (single or divided dosage). Fosinopril (MONOPRIL) Fosinopril sodium is the only ACE inhibitor approved for use in the United States that contains a phosphinate group that binds to the active site of ACE. Cleavage of the ester moiety by hepatic esterases transforms fosinopril, a prodrug, into fosinoprilat, an ACE inhibitor that in vitro is more potent than captopril yet less potent than enalaprilat. Fosinopril is slowly and incompletely (36%) absorbed after oral administration (rate but not extent reduced by food). Fosinopril is nearly completely metabolized to fosinoprilat (75%) and to the glucuronide conjugate of fosinoprilat. These are excreted in both the urine and bile; peak concentrations of fosinoprilat in plasma are achieved in about 3 hours. Fosinoprilat has an effective half-life in plasma of about 11.5 hours, and its clearance is not significantly altered by renal impairment. The oral dosage of fosinopril ranges from 10 to 80 mg daily (single or divided dosage). The dose is reduced to 5 mg daily in patients with Na+ or water depletion or renal failure. Trandolapril (MAVIK) Approximately 10% and 70% of an oral dose of trandolapril is bioavailable (absorption rate but not extent is reduced by food) as trandolapril and trandolaprilat, respectively. Trandolaprilat is about 8 times more potent than trandolapril as an ACE inhibitor. Trandolapril is metabolized to trandolaprilat and to inactive metabolites (mostly glucuronides of trandolapril and deesterification products), and these are recovered in the urine (33%, mostly trandolaprilat) and feces (66%). Peak concentrations of trandolaprilat in plasma are achieved in 4 to 10 hours. Trandolaprilat displays biphasic elimination kinetics with an initial half-life of about 10 hours (the major component of elimination), followed by a more prolonged half-life due to slow dissociation of trandolaprilat from tissue ACE. Plasma clearance of trandolaprilat is reduced by both renal and hepatic insufficiency.

The oral dosage ranges from 1 to 8 mg daily (single or divided dosage). The initial dose is 0.5 mg in patients who are taking a diuretic or who have renal impairment. Quinapril (ACCUPRIL) Cleavage of the ester moiety by hepatic esterases transforms quinapril hydrochloride, a prodrug, into quinaprilat, an ACE inhibitor that in vitro is about as potent as benazeprilat. Quinapril is rapidly absorbed (peak concentrations are achieved in 1 hour, but the peak may be delayed after food), and its rate but not extent of oral absorption (60%) may be reduced by food. Quinapril is metabolized to quinaprilat and to other minor metabolites, and quinaprilat is excreted in the urine (61%) and feces (37%). Peak concentrations of quinaprilat in plasma are achieved in about 2 hours. Conversion of quinapril to quinaprilat is reduced in patients with diminished liver function. The initial half-life of quinaprilat is about 2 hours; a prolonged terminal half-life of about 25 hours may be due to high-affinity binding of the drug to tissue ACE. The oral dosage of quinapril ranges from 5 to 80 mg daily (single or divided dosage). Ramipril (ALTACE) Cleavage of the ester moiety by hepatic esterases transforms ramipril into ramiprilat, an ACE inhibitor that in vitro is about as potent as benazeprilat and quinaprilat. Ramipril is rapidly absorbed (peak concentrations of ramipril achieved in 1 hour), and the rate but not extent of its oral absorption (50% to 60%) is reduced by food. Ramipril is metabolized to ramiprilat and to inactive metabolites (glucuronides of ramipril and ramiprilat and the diketopiperazine ester and acid), and these are excreted predominantly by the kidneys. Peak concentrations of ramiprilat in plasma are achieved in about 3 hours. Ramiprilat displays triphasic elimination kinetics with half-lives of 2 to 4 hours, 9 to 18 hours, and greater than 50 hours. This triphasic elimination is due to extensive distribution to all tissues (initial half-life), clearance of free ramiprilat from plasma (intermediate half-life), and dissociation of ramiprilat from tissue ACE (terminal half-life). The oral dosage of ramipril ranges from 1.25 to 20 mg daily (single or divided dosage). Moexipril (UNIVASC ) Moexipril is another prodrug whose antihypertensive activity is almost entirely due to its deesterified metabolite, moexiprilat. Moexipril is incompletely absorbed, with bioavailability as moexiprilat of about 13%. Bioavailability is markedly decreased by food; therefore, the drug should be taken one hour before meals. The time to peak plasma concentration of moexiprilat is almost 1.5 hours, and the elimination half-life varies between 2 and 12 hours. The recommended dosage range is 7.5 to 30 mg daily in one or two divided doses. The dosage range is halved in patients who are taking diuretics or who have renal impairment. Perindopril (ACEON) Perindopril erbumine is a prodrug, and 30% to 50% of systemically available perindopril is transformed to perindoprilat by hepatic esterases. Although the oral bioavailability of perindopril (75%) is not affected by food, the bioavailability of perindoprilat is reduced by approximately 35%. Perindopril is metabolized to perindoprilat and to inactive metabolites (glucuronides of perindopril and perindoprilat, dehydrated perindopril, and diastereomers of dehydrated perindoprilat), and these are excreted predominantly by the kidneys. Peak concentrations of perindoprilat in plasma are achieved in 3 to 7 hours. Perindoprilat displays biphasic elimination kinetics with half-lives of 3 to 10 hours (the major component of elimination) and 30 to 120 hours (due to slow dissociation of perindoprilat from tissue ACE). The oral dosage ranges from 2 to 16 mg daily (single or divided

dosage). Therapeutic Uses of ACE Inhibitors Drugs that interfere with the reninangiotensin system play a prominent role in the treatment of the major cause of mortality in modern societies, i.e., cardiovascular disease. ACE Inhibitors in Hypertension (See Chapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension.) Inhibition of ACE lowers systemic vascular resistance and mean, diastolic, and systolic blood pressures in various hypertensive states. The effects are readily observed in animal models of renal and genetic hypertension. In human subjects with hypertension, ACE inhibitors commonly lower blood pressure (except when high blood pressure is due to primary aldosteronism). The initial change in blood pressure tends to be positively correlated with PRA and angiotensin II plasma levels prior to treatment. However, several weeks into treatment, a greater percentage of patients show a sizable reduction in blood pressure, and the antihypertensive effect then correlates poorly or not at all with pretreatment values of PRA. It is possible that increased local (tissue) production of angiotensin II and/or increased responsiveness of tissues to normal levels of angiotensin II in some hypertensive patients make them sensitive to ACE inhibitors despite normal PRA. Regardless of the mechanisms, ACE inhibitors have a breadth of clinical utility as antihypertensive agents. The long-term fall in systemic blood pressure observed in hypertensive individuals treated with ACE inhibitors is accompanied by a leftward shift in the renal pressurenatriuresis curve (Figure 315) and a reduction in total peripheral resistance in which there is variable participation by different vascular beds. The kidney is a notable exception to this variability in that there is a prominent vasodilator effect, and increased renal blood flow is a relatively constant finding. This is not surprising, since the renal vessels are exceptionally sensitive to the vasoconstrictor actions of angiotensin II. Increased renal blood flow occurs without an increase in glomerular filtration rate; in fact, the filtration fraction is reduced. Both the afferent and efferent arterioles are dilated. Blood flows in the cerebral and coronary beds, where autoregulatory mechanisms are powerful, are generally well maintained. Besides causing systemic arteriolar dilatation, ACE inhibitors increase the compliance of large arteries, which contributes to a reduction of systolic pressure. Cardiac function in patients with uncomplicated hypertension is generally little changed, although stroke volume and cardiac output may increase slightly with sustained treatment. Baroreceptor function and cardiovascular reflexes are not compromised, and responses to postural changes and exercise are little impaired. Yet, surprisingly, even when a substantial lowering of blood pressure is achieved, heart rate and concentrations of catecholamines in plasma generally increase only slightly if at all. This perhaps reflects an alteration of baroreceptor function with increased arterial compliance and the loss of the normal tonic influence of angiotensin II on the sympathetic nervous system. Secretion of aldosterone in the general population of hypertensive individuals is reduced but not seriously impaired by ACE inhibitors. Aldosterone secretion is maintained at adequate levels by other steroidogenic stimuli, such as adrenocorticotropic hormone and K+ . The activity of these secretagogues on the zona glomerulosa of the adrenal cortex requires, at most, only very small trophic or permissive amounts of angiotensin II, which are always present because inhibition of ACE is never complete. Excessive retention of K+ is encountered only in patients taking supplemental K+, in patients with renal impairment, or in patients taking other medications that

reduce K+ excretion. ACE inhibitors alone normalize blood pressure in approximately 50% of patients with mild-tomoderate hypertension, and many consider ACE inhibitors first-line drugs for the treatment of high blood pressure except for elderly African-American patients. Ninety percent of patients with mildto-moderate hypertension will be controlled by the combination of an ACE inhibitor with either a Ca2+ channel blocker, -adrenergic receptor blocker, or diuretic (see Zusman, 1993). Diuretics, in particular, augment the antihypertensive response to ACE inhibitors by rendering the patient's blood pressure renin-dependent. The goal of antihypertensive therapy is not just to lower blood pressure but, more importantly, to diminish the patient's overall risk of cardiovascular disease. Long-term clinical trials with diuretics and -adrenergic receptor antagonists demonstrate that, while reducing blood pressure does decrease cardiovascular morbidity and mortality, the beneficial effects of lowering blood pressure are due primarily to reductions in the incidence of stroke, with only modest reductions in the incidence of myocardial infarction. This result may be due to the adverse metabolic effects of diuretics and -adrenergic receptor antagonists and/or to the inability of these drugs to reverse the structural changes in the heart and/or blood vessels that may be mediated by the circulating (endocrine) and/or local (paracrine/autocrine/intracrine) reninangiotensin systems. It is possible that ACE inhibitors reduce the incidence of heart disease in hypertensive patients more than do other antihypertensive agents, and clinical trials are underway to test this hypothesis. This hypothesis is based on the lack of adverse metabolic effects of ACE inhibitors and the ability of ACE inhibitors to cause regression of left ventricular hypertrophy in hypertensive patients (De Castro et al., 1996), to prevent ventricular remodeling following myocardial infarction, and to double the life span of hypertensive rats (Linz et al., 1997). There is increasing evidence that ACE inhibitors are superior to other antihypertensive drugs in hypertensive patients with diabetes, in whom they improve endothelial function (O'Driscoll et al., 1997) and reduce cardiovascular events more so than do Ca2+ channel blockers (Estacio et al., 1998; Tatti et al., 1998) or diuretics and -adrenergic receptor antagonists (Hansson et al., 1999). ACE Inhibitors in Left Ventricular Systolic Dysfunction (See Chapter 34: Pharmacological Treatment of Heart Failure.) Left ventricular systolic dysfunction ranges from a modest, asymptomatic reduction in systolic performance to a severe impairment of left ventricular systolic function with New York Heart Association grade IV congestive heart failure. It is now clear that, unless contraindicated, ACE inhibitors should be given to all patients with impaired left ventricular systolic function whether or not they are experiencing symptoms of overt heart failure. Several large, prospective, randomized, placebo-controlled clinical studies support the usefulness of ACE inhibitors in patients with varying degrees of left ventricular systolic dysfunction. These studies are summarized in Table 311 and are described in more detail in Chapter 34: Pharmacological Treatment of Heart Failure. The combined results of these studies strongly indicate that inhibition of ACE in patients with systolic dysfunction prevents or delays the progression of heart failure, decreases the incidence of sudden death and myocardial infarction, decreases hospitalization, and improves quality of life. The more severe the ventricular dysfunction, the greater the benefit from ACE inhibition. Although the mechanisms by which ACE inhibitors improve outcome in patients with systolic

dysfunction are not completely understood, the induction of a more favorable hemodynamic state most likely plays an important role. Inhibition of ACE commonly reduces afterload and systolic wall stress, and both cardiac output and cardiac index increase, as do indices of stroke work and stroke volume. Heart rate generally is reduced. Systemic blood pressure falls, sometimes steeply at the outset, but tends to return toward initial levels. Renovascular resistance falls sharply and renal blood flow increases. Natriuresis occurs as a result of the improved renal hemodynamics, the reduced stimulus to secretion of aldosterone by angiotensin II, and the diminished direct effects of angiotensin II on the kidney. The excess volume of body fluids contracts, which reduces venous return to the right heart. A further reduction results from venodilation and an increased capacity of the venous bed. Venodilation is a somewhat unexpected effect of ACE inhibition, as angiotensin II has little acute venoconstrictor activity. Nevertheless, long-term infusion of angiotensin II increases venous tone, perhaps by central or peripheral interactions with the sympathetic nervous system (Schwartz and Chatterjee, 1983; Johns and Ayers, 1984). The response to ACE inhibitors also involves reductions of pulmonary arterial pressure, pulmonary capillary wedge pressure, and left atrial and left ventricular filling volumes and pressures. Consequently, preload and diastolic wall stress are diminished. The better hemodynamic performance results in increased exercise tolerance and suppression of the sympathetic nervous system (Grassi et al., 1997). Cerebral and coronary blood flows usually are well maintained, even when systemic blood pressure is reduced (Romankiewicz et al., 1983; Schwartz and Chatterjee, 1983). The beneficial effects of ACE inhibitors in systolic dysfunction also involve improvements in ventricular geometry. In heart failure, ACE inhibitors reduce ventricular dilation and tend to restore the heart to its normal elliptical shape. ACE inhibitors may reverse ventricular remodeling via changes in preload/afterload, by preventing the growth effects of angiotensin II on myocytes, and/or by attenuating aldosterone-induced cardiac fibrosis. Although the role of ACE inhibitors in left ventricular systolic dysfunction is firmly established, whether or not these drugs improve diastolic dysfunction is an important yet open question. Infusions of enalaprilat into the left coronary arteries of patients with left ventricular hypertrophy significantly improve diastolic function (Friedrich et al., 1994; Kyriakidis et al., 1998). ACE Inhibitors in Acute Myocardial Infarction Several large, prospective, randomized clinical studies involving thousands of patients (see Table 311) provide convincing evidence that ACE inhibitors reduce overall mortality when treatment is begun during the periinfarction period. The beneficial effects of ACE inhibitors in acute myocardial infarction are particularly large in hypertensive (Borghi et al., 1999) and diabetic (Zuanetti et al., 1997; Gustasson et al., 1999) patients. Unless contraindicated (e.g., cardiogenic shock or severe hypotension), ACE inhibitors should be started immediately during the acute phase of myocardial infarction and can be administered along with thrombolytics, aspirin, and -adrenergic receptor antagonists (ACE Inhibitor Myocardial Infarction Collaborative Group, 1998). After several weeks, ACE-inhibitor therapy should be reevaluated. In high-risk patients (e.g., large infarct, systolic ventricular dysfunction), ACE inhibition should be continued long-term. ACE Inhibitors in Patients Who Are at High Risk of Cardiovascular Events ACE inhibitors tilt the fibrinolytic balance toward a profibrinolytic state by reducing plasma levels of plasminogen activator inhibitor-1 (Vaughan et al., 1997; Brown et al., 1999) and improve endothelial vasomotor dysfunction in patients with coronary artery disease (Mancini et al., 1996). The landmark HOPE study demonstrated that patients who were at high risk of cardiovascular events benefited considerably from treatment with ACE inhibitors (Yusuf et al., 2000). In this

regard, ACE inhibition significantly decreased the rate of myocardial infarction, stroke, and death in a broad range of patients who did not have left ventricular dysfunction but had evidence of vascular disease or diabetes and one other risk factor for cardiovascular disease. Thus, the HOPE study suggests that the use of ACE inhibitors should be expanded to the large population of patients at risk for ischemic cardiovascular events. ACE Inhibitors in Chronic Renal Failure Diabetes mellitus accounts for about one-third of all end-stage renal disease. The landmark study by Lewis et al. (1993) demonstrates that, in patients with type I diabetes mellitus and diabetic nephropathy, captopril prevents or delays the progression of renal disease. These findings are generalizable to other ACE inhibitors and to patients with both type I or type II diabetes regardless of baseline renal function or arterial blood pressure (Ravid et al., 1993, 1996, 1998; EUCLID Study Group, 1997). The nephroprotective effects of ACE inhibitors and Ca2+ channel blockers in diabetes may be additive (see Bretzel, 1997). In addition to preventing diabetic nephropathy, ACE inhibitors also may decrease retinopathy progression in type I diabetics (Chaturvedi et al., 1998). ACE inhibitors also attenuate the progression of renal insufficiency in patients with a variety of nondiabetic nephropathies (Maschio et al., 1996; Gisen Group, 1997; Ruggenenti et al., 1998, 1999b) and may arrest the decline in glomerular filtration rate even in patients with severe renal disease (Ruggenenti et al., 1999a). Several mechanisms participate in the renal protection afforded by ACE inhibitors. Increased glomerular capillary pressure induces glomerular injury, and ACE inhibitors reduce this parameter both by decreasing arterial blood pressure and by dilating renal efferent arterioles. ACE inhibitors increase the permeability selectivity of the filtering membrane, thereby diminishing exposure of the mesangium to proteinaceous factors that may stimulate mesangial cell proliferation and matrix production, two processes that contribute to expansion of the mesangium in diabetic nephropathy. Since angiotensin II is a growth factor, reductions in the intrarenal levels of angiotensin II may further attenuate mesangial cell growth and matrix production. ACE Inhibitors in Scleroderma Renal Crisis Before the use of ACE inhibitors, patients with scleroderma renal crisis generally died within several weeks. A few small, observational studies have suggested that captopril markedly improved this otherwise grim prognosis. Adverse Effects of ACE Inhibitors Metabolic side effects are not encountered during long-term therapy with ACE inhibitors. The drugs do not alter plasma concentrations of uric acid or Ca2+ (Frohlich, 1989) and may actually improve insulin sensitivity in patients with insulin resistance and decrease cholesterol levels and lipoprotein(a) levels in proteinuric renal disease. Serious untoward reactions to ACE inhibitors are rare (see Materson, 1992), and in general ACE inhibitors are well tolerated. Hypotension A steep fall in blood pressure may occur following the first dose of an ACE inhibitor in patients with elevated PRA. In this regard, care should be exercised in patients who are salt-depleted, in patients being treated with multiple antihypertensive drugs, and in patients who have congestive heart failure. In such situations, treatment should be initiated with very small doses of ACE

inhibitors, or salt intake should be increased and diuretics withdrawn before beginning therapy. Cough In 5% to 20% of patients, ACE inhibitors induce a bothersome, dry cough; it is usually not doserelated, occurs more frequently in women than in men, usually develops between 1 week and 6 months after initiation of therapy, and sometimes requires cessation of therapy. This adverse effect may be mediated by the accumulation in the lungs of bradykinin, substance P, and/or prostaglandins. Thromboxane antagonism reduces ACE inhibitorinduced cough (Malini et al., 1997). Once ACE inhibitors are stopped, the cough disappears, usually within 4 days (Israili and Hall, 1992). Hyperkalemia Despite some reduction in the concentration of aldosterone, significant retention of K+ is rarely encountered in patients with normal renal function who are not taking other drugs that cause K+ retention. However, ACE inhibitors may cause hyperkalemia in patients with renal insufficiency or in patients taking K+ -sparing diuretics, K+ supplements, -adrenergic receptor blockers, or NSAIDs. Acute Renal Failure Angiotensin II, by constricting the efferent arteriole, helps maintain adequate glomerular filtration when renal perfusion pressure is low. Consequently, inhibition of ACE can induce acute renal insufficiency in patients with bilateral renal artery stenosis, stenosis of the artery to a single remaining kidney, heart failure, or dehydration due to diarrhea or diuretics. Older patients with congestive heart failure are particularly susceptible to ACE inhibitorinduced acute renal failure. However, in nearly all patients who receive appropriate treatment, recovery of renal function occurs without sequelae (Wynckel et al., 1998). Fetopathic Potential Although ACE inhibitors are not teratogenic during the early period of organogenesis (first trimester), continued administration of ACE inhibitors during the second and third trimesters can cause oligohydramnios, fetal calvarial hypoplasia, fetal pulmonary hypoplasia, fetal growth retardation, fetal death, neonatal anuria, and neonatal death. These fetopathic effects may be due in part to fetal hypotension. While ACE inhibitors are not contraindicated in women of reproductive age, once pregnancy is diagnosed, it is imperative that ACE inhibitors be discontinued as soon as possible. If necessary, an alternative antihypertensive regimen should be instituted. The fetus is not at risk of ACE inhibitorinduced pathology if ACE inhibitors are discontinued during the first trimester of pregnancy (Brent and Beckman, 1991). Skin Rash ACE inhibitors occasionally cause a maculopapular rash that may or may not itch. The rash may resolve spontaneously and may respond to a reduction in dosage or a brief course of antihistamines. This side effect was initially attributed to the presence of the sulfhydryl group in captopril; however, it also occurs with other ACE inhibitors, albeit less frequently. Proteinuria ACE inhibitors have been associated with proteinuria (more than 1 g/day); however, a causal

relationship has been difficult to establish. In general, proteinuria is not a contraindication for ACE inhibitors, as ACE inhibitors are renoprotective in certain renal diseases associated with proteinuria, e.g., diabetic nephropathy. Angioneurotic Edema In 0.1% to 0.2% of patients, ACE inhibitors induce a rapid swelling in the nose, throat, mouth, glottis, larynx, lips, and/or tongue. This untoward effect, called angioneurotic edema, apparently is not dose-related and nearly always develops within the first week of therapy, usually within the first few hours after the initial dose. Airway obstruction and respiratory distress may lead to death. Although the mechanism of angioneurotic edema is unknown, it may involve accumulation of bradykinin, induction of tissue-specific autoantibodies, or inhibition of complement 1esterase inactivator. Once ACE inhibitors are stopped, angioneurotic edema disappears within hours; meanwhile the patient's airway should be protected, and, if necessary, epinephrine, an antihistamine, and/or a corticosteroid should be administered (Israili and Hall, 1992). African Americans have a 4.5 times greater risk of ACE inhibitorinduced angioneurotic edema than do Caucasians (Brown et al., 1996). Dysgeusia An alteration in or loss of taste can occur in patients receiving ACE inhibitors. This adverse effect, which may occur more frequently with captopril, is reversible. Neutropenia Neutropenia is a rare, but serious, side effect of ACE inhibitors. Although the frequency of neutropenia is low, it occurs predominantly in hypertensive patients with collagen-vascular or renal parenchymal disease. If serum creatinine is 2 mg/dl or greater, the dose of ACE inhibitor should be kept low, and the patient should be counseled to watch for symptoms of neutropenia (e.g., sore throat, fever). Glycosuria An exceedingly rare and reversible side effect of ACE inhibitors is spillage of glucose into the urine in the absence of hyperglycemia (Cressman et al., 1982). The mechanism is unknown. Hepatotoxicity Also exceedingly rare and reversible is hepatotoxicity, usually of the cholestatic variety (Hagley et al., 1993). The mechanism is unknown. Drug Interactions Antacids may reduce the bioavailability of ACE inhibitors; capsaicin may worsen ACE inhibitor induced cough; NSAIDs, including aspirin (Guazzi et al., 1998), may reduce the antihypertensive response to ACE inhibitors; and K+ -sparing diuretics and K+ supplements may exacerbate ACE inhibitorinduced hyperkalemia. ACE inhibitors may increase plasma levels of digoxin and lithium and may increase hypersensitivity reactions to allopurinol. Nonpeptide Angiotensin II Receptor Antagonists

History Attempts to develop therapeutically useful angiotensin II receptor antagonists date to the early 1970s, and these initial endeavors concentrated on angiotensin peptide analogs. Saralasin, 1sarcosine, 8-isoleucine angiotensin II, and other 8-substituted angiotensins were potent angiotensin II receptor antagonists but were of no clinical value because of lack of oral bioavailability and because all peptide angiotensin II receptor antagonists expressed unacceptable partial agonist activity. Although initial efforts to develop nonpeptide angiotensin receptor antagonists were unsuccessful, a breakthrough came in the early 1980s with the issuance of patents (Furakawa et al., 1982) on a series of imidazole-5-acetic acid derivatives that attenuated pressor responses to angiotensin II in rats. Two of the compounds described in the patents, S-8307 and S-8308, later were found to be highly specific, albeit very weak, nonpeptide angiotensin II receptor antagonists that were devoid of partial agonist activity (Wong et al., 1988; Chiu et al., 1988). In an instructive example of drug design, molecular modeling of these lead compounds gave rise to the hypothesis that their structures would have to be extended to mimic more closely the pharmacophore of angiotensin II (Figure 31 7 A). Through an insightful series of stepwise modifications (Figure 317 B), the orally active, potent, and selective nonpeptide AT1 receptor antagonist losartan was developed (Timmermans et al., 1993). Losartan was approved for clinical use by the United States Food and Drug Administration in 1995. Since then, the Food and Drug Administration has approved five additional AT1 receptor antagonists (Figure 318). Hundreds of AT 1 receptor antagonists have been synthesized, representing a diverse array of chemical structures (Weinstock and Keenan, 1994). However, AT1 receptor antagonists approved in the United States are either biphenylmethyl derivatives or thienylmethylacrylic acid derivatives (Figure 318). Although approved AT 1 receptor antagonists are devoid of partial agonist activity, nonpeptide AT1 receptor agonists have been synthesized, and structural modifications as minor as a methyl group can transform a potent antagonist into an agonist (Perlman et al., 1997). Figure 317. A. Hypothesized Relationship between S-8308 (Takeda Lead Compound) and Angiotensin II, and Design Strategies to Enhance Binding Affinity of Nonpeptide Antagonists to the Angiotensin II Receptor. Letters Indicate Corresponding Regions of S-8308 and Angiotensin II. B. Pathway Leading to the Discovery of Losartan. (Modified from Timmermans et al., 1993, with permission.)

Figure 318. FDA-Approved Angiotensin II Receptor Antagonists.

Pharmacological Effects The angiotensin II receptor blockers (ARBs) available for clinical use bind to the AT1 receptor with high affinity and are generally >10,000-fold selective for AT 1versus the AT2 receptor. The rank order affinity of the AT1 receptor for ARBs is: candesartan > irbesartan > telmisartan = valsartan = EXP 3174 (the active metabolite of losartan) > losartan (Mimran et al., 1999). Although binding of ARBs to the AT1 receptor is competitive, the inhibition by ARBs of biological responses to angiotensin II is often insurmountable, i.e., the maximal response to angiotensin II cannot be restored in the presence of the ARB regardless of the concentration of angiotensin II added to the experimental preparation. Of the currently available ARBs, candesartan suppresses the maximal response to angiotensin II the most, whereas insurmountable blockade by irbesartan, eprosartan, telmisartan, and valsartan is less. Although losartan per se demonstrates surmountable antagonism,

EXP 3174, an active metabolite of losartan, causes some degree of insurmountable blockade. The mechanism of insurmountable antagonism by ARBs may be due to slow dissociation kinetics of the compounds from the AT1 receptor; however, a number of other factors may contribute, such as ARB-induced receptor internalization and alternative binding sites for ARBs on the AT1 receptor (McConnaughey et al., 1999). Regardless of the mechanism, insurmountable antagonism has the theoretical advantage of sustained receptor blockade even with increased levels of endogenous ligand and with missed doses of drug. Whether this theoretical advantage translates into an enhanced clinical performance remains to be determined. The pharmacology of ARBs is well described (Timmermans et al., 1993; Csajka et al., 1997). ARBs potently and selectively inhibit, both in vitro and in vivo, most of the biological effects of angiotensin II, including angiotensin II-induced: (1) contraction of vascular smooth muscle; (2) rapid pressor responses; (3) slow pressor responses; (4) thirst; (5) vasopressin release; (6) aldosterone secretion; (7) release of adrenal catecholamines; (8) enhancement of noradrenergic neurotransmission; (9) increases in sympathetic tone; (10) changes in renal function; and (11) cellular hypertrophy and hyperplasia. ARBs reduce arterial blood pressure in animals with renovascular and genetic hypertension as well as in transgenic animals overexpressing the renin gene. ARBs, however, have little effect on arterial blood pressure in animals with low-renin hypertension (e.g., rats with hypertension induced by NaCl and deoxycorticosterone). A critical issue is whether or not ARBs are equivalent to ACE inhibitors with regard to therapeutic efficacy. Although both classes of drugs block the reninangiotensin system, ARBs differ from ACE inhibitors in several important aspects: (1) ARBs reduce activation of AT1 receptors more effectively than do ACE inhibitors. ACE inhibitors reduce the biosynthesis of angiotensin II produced by the action of ACE on angiotensin I, but do not inhibit alternative non-ACE angiotensin IIgenerating pathways. Because ARBs block the AT 1 receptor, the actions of angiotensin II via the AT1 receptor are inhibited regardless of the biochemical pathway leading to angiotensin II formation. (2) In contrast to ACE inhibitors, ARBs indirectly activate AT2 receptors. ACE inhibitors increase renin release; however, because ACE inhibitors block the conversion of angiotensin I to angiotensin II, ACE inhibition is not associated with increased levels of angiotensin II. ARBs also stimulate renin release; however, with ARBs, this translates into a several-fold increase in circulating levels of angiotensin II. Because AT2 receptors are not blocked by clinically available ARBs, ARBs indirectly stimulate AT2 receptors by increasing angiotensin II levels. (3) ACE inhibitors may increase angiotensin (1-7) levels more than do ARBs. ACE is involved in the clearance of angiotensin (1-7), so inhibition of ACE may increase angiotensin (1-7) levels more so than do ARBs. (4) ACE inhibitors increase the levels of a number of ACE substrates, including bradykinin and Ac-SDKP. ACE is a nondiscriminating enzyme that processes a wide array of substrates; inhibiting ACE therefore increases the levels of ACE substrates and decreases the levels of their corresponding products. Whether or not the pharmacological differences between ARBs and ACE inhibitors result in significant differences in therapeutic outcomes is an open question. Clinical Pharmacology Oral bioavailability of ARBs is generally low <50% except for irbesartan (70%), and protein binding is high (>90%). Candesartan Cilexetil (ATACAND) Candesartan cilexetil is an inactive ester prodrug that is completely hydrolyzed to the active form, candesartan, during absorption from the gastrointestinal tract. Peak plasma levels are obtained 3 to 4 hours after oral administration, and the plasma half-life is about 9 hours. Plasma clearance of

candesartan is due to renal elimination (33%) and biliary excretion (67%). The plasma clearance of candesartan is affected by renal insufficiency but not by mild-to-moderate hepatic insufficiency. Candesartan cilexetil should be administered orally once or twice daily for a total daily dosage of 4 to 32 mg. Eprosartan (TEVETEN) Peak plasma levels are obtained approximately 1 to 2 hours after oral administration, and the plasma half-life ranges from 5 to 9 hours. Eprosartan is in part metabolized to the glucuronide conjugate, and the parent compound and its glucuronide conjugate are cleared by renal elimination and biliary excretion. The plasma clearance of eprosartan is affected by both renal insufficiency and hepatic insufficiency. The recommended dosage of eprosartan is 400 to 800 mg/day in one or two doses. Irbesartan (AVAPRO ) Peak plasma levels are obtained approximately 1.5 to 2 hours after oral administration, and the plasma half-life ranges from 11 to 15 hours. Irbesartan is in part metabolized to the glucuronide conjugate, and the parent compound and its glucuronide conjugate are cleared by renal elimination (20%) and biliary excretion (80%). The plasma clearance of irbesartan is unaffected by either renal or mild-to-moderate hepatic insufficiency. The oral dosage of irbesartan is 150 to 300 mg once daily. Losartan (COZAAR) Approximately 14% of an oral dose of losartan is converted to the 5-carboxylic acid metabolite, designated EXP 3174, which is more potent than losartan as an AT1 receptor antagonist. The metabolism of losartan to EXP 3174 and to inactive metabolites is mediated by CYP2C9 and 3A4. Peak plasma levels of losartan and EXP 3174 are obtained approximately 1 to 3 hours after oral administration, respectively, and the plasma half-lives are 2.5 and 6 to 9 hours, respectively. The plasma clearances of losartan and EXP 3174 (600 and 50 ml/min, respectively) are due to renal clearance (75 and 25 ml/min, respectively) and hepatic clearance (metabolism and biliary excretion). The plasma clearance of losartan and EXP 3174 is affected by hepatic but not renal insufficiency. Losartan should be administered orally once or twice daily for a total daily dose of 25 to 100 mg. Telmisartan (MICARDIS) Peak plasma levels are obtained approximately 0.5 to 1 hour after oral administration, and the plasma half-life is about 24 hours. Telmisartan is cleared from the circulation mainly by biliary secretion of intact drug. The plasma clearance of telmisartan is affected by hepatic, but not renal, insufficiency. The recommended oral dosage of telmisartan is 40 to 80 mg once daily. Valsartan (DIOVAN) Peak plasma levels are obtained approximately 2 to 4 hours after oral administration, and the plasma half-life is about 9 hours. Food markedly decreases absorption. Valsartan is cleared from the circulation by the liver (about 70% of total clearance). The plasma clearance of valsartan is affected by hepatic but not renal insufficiency. The oral dosage of valsartan is 80 to 320 mg once daily.

Therapeutic Uses of Angiotensin II Receptor Antagonists ARBs are approved only for the treatment of hypertension. The efficacy of ARBs with regard to lowering blood pressure is comparable to that of other established antihypertensive drugs, with an adverse effect profile similar to that of placebo (Mimran et al., 1999). Several ARBs also are available as fixed-dose combinations with hydrochlorothiazide. Whether ARBs reduce cardiovascular morbidity and mortality as well as or better than other classes of antihypertensive drugs is unknown but under active investigation. The ongoing Losartan Intervention For Endpoint (LIFE) Reduction in Hypertension Study is a comparison of losartan with atenolol on cardiovascular morbidity and mortality in hypertensive patients with left ventricular hypertrophy (Dahlf et al., 1997). Losartan is well tolerated in patients with heart failure and is comparable to enalapril with regard to improving exercise tolerance (Lang et al., 1997). Moreover, losartan improves peak exercise capacity and attenuates symptoms in heart failure patients who have severe symptoms despite treatment with large doses of ACE inhibitors (Hamroff et al., 1999). The Evaluation of Losartan in the Elderly (ELITE) study reported that, in elderly patients with heart failure, losartan was as effective as captopril in improving symptoms and reduced mortality more than did captopril (Pitt et al., 1997). However, the greater reduction in mortality by losartan was not confirmed in the larger Losartan Heart Failure Survival Study (ELITE II) trial (Pitt et al., 1999); in fact, captopril tended to have a more favorable effect on several outcome measures. A similar trial comparing candesartan and an ACE inhibitor in heart failure patients (CHARM study) is in progress. Current recommendations are to use ACE inhibitors as first-line agents for the treatment of heart failure and to reserve ARBs for treatment of heart failure in patients who cannot tolerate or have an unsatisfactory response to ACE inhibitors. The ongoing Optimal Therapy in Myocardial Infarction with the Angiotensin II Antagonist Losartan (OPTIMAAL) study is comparing losartan with captopril on all-cause mortality in high-risk patients after acute myocardial infarction (Dickstein and Kjekshus, 1999). Losartan is reported to be safe and highly effective in the treatment of portal hypertension in patients with cirrhosis and portal hypertension (Schneider et al., 1999) without compromising renal function. The efficacy of ARBs in diabetic or nondiabetic patients with chronic renal insufficiency is unknown. Adverse Effects The incidence of discontinuation of ARBs due to adverse reactions is comparable to that of placebo. Unlike ACE inhibitors, ARBs do not cause cough, and the incidence of angioneurotic edema with ARBs is much less than with ACE inhibitors. As with ACE inhibitors, ARBs have fetopathic potential and should be discontinued before the second trimester of pregnancy. ARBs should be used cautiously in patients whose arterial blood pressure or renal function is highly dependent on the reninangiotensin system. In such patients, ARBs can cause hypotension, oliguria, progressive azotemia, or acute renal failure. ARBs may cause hyperkalemia in patients with renal disease or in patients taking K+ supplements or K+ -sparing drugs. ARBs enhance the blood pressure-lowering effect of other antihypertensive drugs, a desirable effect but one that may necessitate dosage adjustment. Prospectus Vasopeptidase Inhibitors

Vasopeptidase inhibitors are drugs that inhibit enzymes that metabolize vasoactive peptides. One class of vasopeptidase inhibitors are compounds that block both ACE and neutral endopeptidase (NEP). Although ACE inhibitors are effective antihypertensive agents in high-renin and normalrenin hypertension, they are generally much less effective in hypertensive states associated with volume expansion and suppressed renin levels. NEP metabolizes various natriuretic peptides, including ANP, BNP, and CNP, as well as bradykinin. Because the secretion of natriuretic peptides is elevated by volume expansion, inhibition of NEP increases levels of natriuretic peptides during volume expansion. Thus, inhibition tends to cause natriuresis and lowers blood pressure in volumeexpanded hypertension. Merging ACE and NEP inhibitory actions into a single molecular entity might afford antihypertensive drugs that, on average, lower blood pressure more and in a higher percentage of hypertensive patients than do currently available agents. These predictions recently were confirmed in hypertensive patients using the dual ACE/NEP inhibitor omapatrilat. Many dual ACE/NEP inhibitors are under preclinical and clinical investigation, and it is likely that such agents will become available in the near future. No doubt the efficacy of dual ACE/NEP inhibitors in patients with heart failure, myocardial infarction, and renal insufficiency, as well as in patients at high risk of cardiovascular disease, will be the focus of many outcome trials. In addition to vasopeptidase inhibitors that are dual ACE/NEP inhibitors, it is likely that other classes of vasopeptidase inhibitors will be developed and tested, such as agents that block some combination of ACE, NEP, endothelin-converting enzyme, or TNF -converting enzyme. Renin Inhibitors Renin inhibitors have been the focus of drug development efforts for over two decades. Inefficient absorption and high first-pass metabolism and biliary excretion have stymied clinical development of this group of drugs. Nonetheless, development efforts continue, and some transition-state analogs of angiotensinogen hold significant promise (Lin and Frishman, 1996). Gene Therapy for Hypertension Many drugs now are available for the treatment of hypertension, and control of high blood pressure is no longer limited by the lack of effective drugs. However, since hypertension is usually a painless, lifelong disorder, compliance is the major obstacle to controlling hypertension in a significant percentage of patients. Consequently, the next breakthrough in hypertension research could be the development of treatments for hypertension that are semipermanent, requiring only occasional attention. Gene therapy (see Chapter 5: Gene Therapy) aimed at nullifying the renin angiotensin system is a speculative, though not inconceivable, approach to the semipermanent or permanent treatment of hypertension. In this regard, two approaches have been suggested, a mutant angiotensinogen gene strategy (Jackson, 1992) and an antisense strategy; viral delivery of AT1 receptor (Iyer et al., 1996) and ACE antisense (Wang et al., 1999) causes a sustained reduction of blood pressure in hypertensive animals. For further discussion of hypertension, see Chapter 230 in Harrison's Principles of Internal Medicine, 16th ed., McGraw-Hill, New York, 2005.

Chapter 32. Drugs Used for the Treatment of Myocardial Ischemia

Overview This chapter briefly reviews the pathophysiology underlying angina pectoris, the most common symptom of chronic ischemic heart disease. The causes of the myocardial ischemia that produces

angina are defined in terms of the myocardial oxygen supplydemand relationship. Stable and unstable angina, silent ischemia, variant angina, and myocardial infarction are considered; the contributions of fixed atherosclerotic coronary narrowings, of active coronary vasospasm, and of intracoronary thrombosis in these syndromes are discussed to clarify the roles of antianginal agents. Also discussed is the angina of autonomic dysfunction, for which therapy must be specifically tailored to ameliorate the underlying problem of great fluctuations in coronary perfusion pressure, and in which conventional antianginal therapy is not efficacious. For each class of antianginal therapy (organic nitrates, Ca2+ channel antagonists, -adrenergic receptor antagonists, and antiplatelet/antithrombotic agents), the effects of the class and its most important agents on the factors determining myocardial oxygen demand and myocardial oxygen supply are reviewed. The use of organic nitrates (also covered in Chapter 34: Pharmacological Treatment of Heart Failure) in sublingual, oral, buccal, and intravenous forms is discussed, with attention to the issue of nitrate tolerance and the relationship of nitrovasodilators to endogenous endothelium-derived vasodilator(s). Interactions of organic nitrates with sildenafil also are discussed. The multiple classes of Ca2+ channel antagonists (also discussed in Chapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension) have distinct effects on vascular smooth muscle and cardiac tissue, and these effects are placed in the context of the ischemic cardiac syndromes. -Adrenergic receptor antagonists, which also are addressed in Chapters 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists, 33: Antihypertensive Agents and the Drug Therapy of Hypertension, 34: Pharmacological Treatment of Heart Failure, and 35: Antiarrhythmic Drugs, are dealt with here in terms of their ability to improve survival in ischemic heart disease as well as their efficacy in improving exercise tolerance in stable angina. Considering unstable angina as a thrombotic disease leads to a description of the role of antiplatelet therapy in ischemic heart disease (see also Chapter 55: Anticoagulant, Thrombolytic, and Antiplatelet Drugs). Agents currently under investigation and the potential role of gene-based therapy in the treatment of myocardial ischemia also are described. Drugs Used for the Treatment of Myocardial Ischemia: Introduction The primary symptom of ischemic heart disease is angina pectoris, caused by transient episodes of myocardial ischemia. These episodes of ischemia are due to an imbalance in the myocardial oxygen supplydemand relationship and may be caused by an increase in myocardial oxygen demand (determined by heart rate, ventricular wall tension, and ventricular contractility), by a decrease in myocardial oxygen supply (primarily determined by coronary blood flow, but occasionally modified by the oxygen-carrying capacity of the blood), or sometimes by both (Friesinger and Robertson, 1985, 1986; Kaplinsky, 1992; see Figure 321). Regardless of the precipitating factors, the sensation of angina is similar in most patients. Both typical and variant (Prinzmetal's) angina are commonly experienced as a heavy, pressing, substernal discomfort (rarely called "pain"), often radiating to the left shoulder, flexor aspect of the left arm, jaw, or epigastrium, with a significant minority of patients noting discomfort in a different location or of a different character. Women, the elderly, and diabetics are more likely to have ischemia with atypical symptoms. Figure 321. Ischemic Episodes: An Imbalance in the Myocardial Oxygen SupplyDemand Relationship. The figure illustrates the principal determinants of myocardial oxygen consumption and mechanisms for increasing oxygen delivery. The arteriovenous oxygen difference always is near maximum in the coronary circulation; thus, widening of the arteriovenous oxygen difference cannot significantly enhance oxygen delivery. Redistribution of regional myocardial flow probably is of major importance. (Adapted with permission from Ross, 1971.)

Angina pectoris is a common symptom, affecting 6,400,000 Americans (American Heart Association, 2001). It may occur in a stable pattern over many years or may become unstable, increasing in frequency or severity and even occurring at rest. In typical stable angina, the pathological substrate is usually fixed atherosclerotic narrowing of an epicardial coronary artery, upon which exertion, emotional stress, etc., superimpose an increase in myocardial oxygen consumption. In variant angina, focal or diffuse coronary vasospasm episodically reduces coronary flow. Patients also may display a mixed pattern of angina with the addition of altered vessel tone on a background of atherosclerotic narrowing. In the majority of patients with unstable angina, rupture of an atherosclerotic plaque, with consequent platelet adhesion and aggregation, decreases coronary blood flow. Plaques with thinner fibrous caps are recognized to be more "vulnerable" to rupture (Fuster et al. 1996). Myocardial ischemia also may be "silent," with electrocardiographic, echocardiographic, or radionuclide evidence of ischemia appearing in the absence of symptoms. While some patients have only silent ischemia, the majority of patients who have silent ischemia have symptomatic episodes as well. The precipitants of silent ischemia appear to be the same as those of symptomaticischemia. We now know that the "ischemic burden,"i.e., the total time a patient is ischemic each day, is greater in many patients than was recognized previously. The agents that are efficacious in conventional angina appear, in most trials, to be efficacious in reducing silent ischemia. Adrenergic receptor antagonists appear to be more effective than the Ca2+ channel blockers in the prevention of episodes. Therapy directed at abolishing all silent ischemia has not been shown to be of additional benefit over conventional therapy. An unusual form of angina is seen in patients with autonomic dysfunction and faulty control of the circulation in the upright posture (Hines et al. 1981). The marked orthostatic hypotension seen in these patients can reduce coronary perfusion pressure sufficiently to cause myocardial ischemia even in patients with normal coronary arteries. As this form of angina is precipitated by upright posture and is relieved when the blood pressure and coronary perfusion pressure rise with sitting or lying, it may appear to be typical exertional angina if the history is not taken carefully and the blood pressure determined in the upright posture. The specific therapy needed for this form of angina is discussed below. This chapter describes the pharmacological agents used in the treatment of angina. The major drugs are nitrovasodilators (see also Chapter 34: Pharmacological Treatment of Heart Failure), adrenergic receptor antagonists (see also Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists), Ca2+ channel antagonists (see also Chapter 33:

Antihypertensive Agents and the Drug Therapy of Hypertension), and, in both stable and unstable angina, antiplatelet agents (see Chapters 27: Analgesic-Antipyretic and Antiinflammatory Agents and Drugs Employed in the Treatment of Gout and 55: Anticoagulant, Thrombolytic, and Antiplatelet Drugs) as well as 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors, which may have a role in stabilizing the vulnerable plaque (see Chapter 36: Drug Therapy for Hypercholesterolemia and Dyslipidemia). All approved antianginal agents function by improving the balance of myocardial oxygen supply and demand: increasing supply by dilating the coronary vasculature or decreasing demand by reducing cardiac work (see Figure 321). Increasing the cardiac extraction of oxygen from the blood is not a practical therapeutic goal. Agents used in typical angina either increase blood flow to the heart; decrease left ventricular wall tension, heart rate, and/or contractility; or do both. In variant or vasotonic angina, prevention of coronary vasospasm is the therapeutic aim. In unstable angina, correcting the tendency to intracoronary thrombosis is the most important therapeutic maneuver. Antianginal agents may provide prophylactic or symptomatic treatment, but -adrenergic receptor antagonists appear, in addition, to reduce mortality by decreasing the incidence of sudden cardiac death associated with myocardial ischemia and infarction. The treatment of cardiac risk factors can reduce the progression or even lead to the regression of atherosclerosis. Aspirin is used routinely in patients with myocardial ischemia, and lipid-lowering agents also have been demonstrated to reduce the incidence of clinical events (Gibbons et al., 1999). Coronary artery bypass surgery and percutaneous coronary interventions such as angioplasty, atherectomy, and stent deployment are alternatives to pharmacological treatment; in some subsets of patients, bypass surgery and other forms of revascularization have been demonstrated to have a survival advantage over medical treatment alone. Novel therapies modifying the expression of vascular or myocardial cell genes are expected to be an important part of the therapy of ischemic heart disease in the future. Organic Nitrates History Nitroglycerin was first synthesized in 1846 by Sobrero, who observed that a small quantity of the oily substance placed on the tongue elicited a severe headache. Constantin Hering, in 1847, developed the sublingual dosage form. In 1857, the eminent physician T. Lauder Brunton of Edinburgh administered amyl nitrite, a known vasodepressor, by inhalation and noted that anginal pain was relieved within 30 to 60 seconds. The action of amyl nitrite was transitory, however, and the dosage was difficult to adjust. Subsequently, William Murrell decided that the action of nitroglycerin mimicked that of amyl nitrite, and he established the use of sublingual nitroglycerin for relief of the acute anginal attack and as a prophylactic agent to be taken prior to exertion (Murrell, 1879). The empirical observation that organic nitrates could be used safely for the rapid, dramatic alleviation of the symptoms of angina pectoris led to their widespread acceptance by the medical profession. Basic investigations led to an understanding of the role of nitric oxide (Moncada et al., 1988) in both the vasodilation produced by nitrates and endogenous vasodilation. The importance of nitric oxide as a signaling molecule in the cardiovascular system and elsewhere was recognized by the awarding of the Nobel Prize in Medicine or Physiology to Furchgott, Ignarro, and Murad in 1998. Chemistry Organic nitrates are polyol esters of nitric acid, whereas organic nitrites are esters of nitrous acid (Table 321). Nitrate esters (CONO2) and nitrite esters (CONO) are characterized by a sequence of carbonoxygennitrogen, whereas nitro compounds possess carbonnitrogen bonds

(CNO2). Thus, glyceryl trinitrate is not a nitro compound, and it is erroneously called nitroglycerin; however, this nomenclature is both widespread and official. Amyl nitrite is a highly volatile liquid that is administered by inhalation. Organic nitrates of low molecular mass (such as nitroglycerin) are moderately volatile, oily liquids, whereas the high-molecular-mass nitrate esters (e.g., erythrityl tetranitrate, pentaerythritol tetranitrate, isosorbide dinitrate) are solids. The fully nitrated polyols are lipid soluble, whereas their incompletely nitrated metabolites are more soluble in water. In the pure form (without an inert carrier such as lactose), nitroglycerin is explosive. The organic nitrates and nitrites and several other compounds that are capable of denitration to release nitric oxide (NO) have been collectively termed nitrovasodilators. Nitric oxide activates guanylyl cyclase, increasing intracellular levels of cyclic guanosine 3',5'-monophosphate (cyclic GMP), and thereby produces vasodilation (Murad, 1986; Molina et al., 1987; Thadani, 1992). Endogenous NO is formed when L-arginine is converted to citrulline by nitric oxide synthases. Both constitutive and inducible forms of these synthases are found in vascular endothelial and smooth muscle cells as well as in other cell types throughout the body, including the central nervous system (Lowenstein et al., 1994). In the setting of human atherosclerosis, the expression of endothelial nitric oxide synthase and the production of NO are reduced (Oemar et al., 1998). Pharmacological Properties Cardiovascular Effects Hemodynamic Effects The nitrovasodilators relax most smooth muscle, including that in arteries and veins. Low concentrations of nitroglycerin produce dilation of the veins that predominates over that of arterioles. Venodilation results in decreased left and right ventricular chamber size and end-diastolic pressures but little change in systemic vascular resistance. Systemic arterial pressure may fall slightly, and heart rate is unchanged or slightly increased reflexly. Pulmonary vascular resistance and cardiac output both are slightly reduced. Even doses of nitroglycerin that do not alter systemic arterial pressure often produce arteriolar dilation in the face and neck, resulting in a flush, or dilation of meningeal arterial vessels, causing headache. There is an enrichment of the enzyme that converts nitroglycerin to NO in venous compared to arterial smooth muscle cells, likely the basis of the partially venoselective properties of the nitrates (Bauer and Fung, 1996). Higher doses of organic nitrates cause further venous pooling and may decrease arteriolar resistance as well, decreasing systolic and diastolic blood pressure and cardiac output and resulting in pallor, weakness, dizziness, and activation of compensatory sympathetic reflexes. The resultant tachycardia and peripheral arteriolar vasoconstriction tend to restore systemic vascular resistance; this is superimposed on sustained venous pooling. Coronary blood flow may increase transiently as a result of coronary vasodilation, but with a subsequent decrease if cardiac output and blood pressure decrease sufficiently. In patients with autonomic dysfunction and an inability to increase sympathetic outflow (multiplesystem atrophy and pure autonomic failure are the most common forms), the fall in blood pressure consequent to the venodilation produced by nitrates cannot be compensated. Nitrates may reduce arterial pressure and coronary perfusion pressure significantly and actually aggravate angina in addition to producing potentially life-threatening hypotension. The correct therapy in patients with orthostatic angina and normal coronary arteries is to correct the orthostatic hypotension by increasing volume retention (fludrocortisone and a high-sodium diet) by preventing venous pooling with fitted support garments and by the carefully titrated use of oral vasopressors. As patients with autonomic dysfunction occasionally may have coexistent coronary artery disease, the coronary

anatomy should be defined before therapy is undertaken. Effects on Total and Regional Coronary Blood Flow Ischemia is a powerful stimulus to coronary vasodilation, and regional blood flow is adjusted by autoregulatory mechanisms. In the presence of atherosclerotic coronary narrowing, ischemia distal to the lesion is a stimulus for vasodilation, and, if the degree of narrowing is severe, much of the capacity to dilate is utilized to maintain resting blood flow. When situations arise that increase demand, further dilation may not be possible. After demonstration of direct coronary vasodilation in experimental animals, it became generally accepted that nitrates relieved anginal pain by dilating coronary arteries and thereby increasing coronary blood flow. This hypothesis was questioned by Gorlin and associates (1959), who were unable to demonstrate increases in coronary blood flow in patients with angina pectoris following the administration of nitroglycerin. However, organic nitrates do appear to cause redistribution of blood flow in the heart when the coronary circulation is partially occluded. Under these circumstances, there is a disproportionate reduction in blood flow to the subendocardial regions of the heart, which are subjected to the greatest extravascular compression during systole; organic nitrates tend to restore blood flow in these regions toward normal (Horwitz et al., 1971). The hemodynamic mechanisms responsible for these effects are not entirely clear. Most hypotheses have focused on the ability of organic nitrates to cause dilation and prevent vasoconstriction of large epicardial vessels without impairing autoregulation in the small vessels, which are responsible for about 90% of the overall coronary vascular resistance. The vessel diameter is an important determinant of the response to nitroglycerin; vessels larger than 200 m in diameter are highly responsive, whereas those less than 100 m are minimally responsive (Sellke et al., 1990). Experimental evidence in patients undergoing coronary bypass surgery indicates that nitrates do have a relaxant effect on large coronary vessels. Collateral flow to ischemic regions also is increased (Goldstein et al., 1974). Moreover, analyses of coronary angiograms in human beings have shown that sublingual nitroglycerin can dilate epicardial stenoses and reduce the resistance to flow through such areas (Brown et al., 1981; Feldman et al., 1981). The resultant increase in blood flow would be distributed preferentially to ischemic myocardial regions as a consequence of vasodilation induced by autoregulation. An important indirect mechanism for a preferential increase in subendocardial blood flow is the nitroglycerin-induced reduction in intracavitary systolic and diastolic pressures that oppose blood flow to the subendocardium (see below). To the extent that organic nitrates decrease myocardial requirements for oxygen (see below), the increased blood flow in ischemic regions could be balanced by decreased flow in nonischemic areas, and an overall increase in coronary blood flow need not occur. Recent studies suggest that dilation of cardiac veins may result in an improvement in the perfusion of the coronary microcirculation (Darius, 1999). Redistribution of blood flow to subendocardial tissue is not typical of all vasodilators. Dipyridamole, for example, dilates resistance vessels nonselectively by distorting autoregulation; it is ineffective in patients with typical angina. In patients with angina due to coronary spasm, the ability of organic nitrates to dilate epicardial coronary arteries, and particularly regions affected by spasm, may be the primary mechanism by which they are of benefit. Effects on Myocardial Oxygen Requirements By their effects on the systemic circulation, the organic nitrates also can reduce myocardial oxygen demand. The major determinants of myocardial oxygen consumption include left ventricular wall tension, heart rate, and the contractility of the myocardium. Ventricular wall tension is affected by a

number of factors that may be considered under the categories of "preload" and "afterload."Preload is determined by the diastolic pressure that distends the ventricle (ventricular end-diastolic pressure). Increasing end-diastolic volume augments the ventricular wall tension (by the law of Laplace, tension is proportional to pressure x radius). Increasing venous capacitance with nitrates decreases venous return to the heart, decreases ventricular end-diastolic volume, and thereby decreases oxygen consumption. An additional benefit of reducing preload is that it increases the pressure gradient for perfusion across the ventricular wall; this favors subendocardial perfusion (Parratt, 1979). Afterload is the impedance against which the ventricle must eject. In the absence of aortic valvular disease, afterload is related to peripheral resistance. Decreasing peripheral arteriolar resistance reduces afterload and thus myocardial work and oxygen consumption. Organic nitrates do not directly alter the inotropic or chronotropic state of the heart. They do decrease both preload and afterload as a result of respective dilation of venous capacitance and arteriolar resistance vessels. Since the primary determinants of oxygen demand are reduced by the nitrates, their net effect usually is to decrease myocardial consumption of oxygen. In addition, an improvement in the lusitropic state of the heart may be seen, with more rapid early diastolic filling (Breisblatt et al., 1988). This may be secondary to the relief of ischemia, rather than primary, or may be due to a reflex increase in sympathetic activity. Nitrovasodilators also increase cyclic GMP in platelets with consequent inhibition of platelet function (De Caterina et al., 1988; Lacoste et al., 1994) and decreased deposition of platelets in animal models of arterial wall injury (Lam et al., 1988). While this may contribute to their antianginal efficacy, the effect appears to be modest and may in some settings be confounded by the potential of nitrates to alter the pharmacokinetics of heparin, reducing its antithrombotic effect. When nitroglycerin is injected or infused directly into the coronary circulation of patients with coronary artery disease, anginal attacks (induced by electrical pacing) are not aborted, even when coronary blood flow is increased. However, sublingual administration of nitroglycerin does relieve anginal pain in the same patients (Ganz and Marcus, 1972). Furthermore, venous phlebotomy that is sufficient to reduce left ventricular end-diastolic pressure can mimic the beneficial effect of nitroglycerin. Patients are able to exercise for considerably longer periods after the administration of nitroglycerin. Nevertheless, angina occurs, with or without nitroglycerin, at the same value of the "triple product" (aortic pressure x heart rate x ejection time is proportional to the myocardial consumption of oxygen). The observation that angina occurs at the same level of myocardial oxygen consumption suggests that the beneficial effects of nitroglycerin are the result of a reduced cardiac oxygen demand, rather than an increase in the delivery of oxygen to ischemic regions of myocardium. However, these results do not preclude the possibility that a favorable redistribution of blood flow to ischemic subendocardial myocardium may contribute to relief of pain in a typical anginal attack, nor do they preclude the possibility that direct coronary vasodilation may be the major effect of nitroglycerin in situations where vasospasm compromises myocardial blood flow. Mechanism of Relief of Symptoms of Angina Pectoris Brunton ascribed the nitrate-induced relief of anginal pain to a decrease in cardiac work secondary to the fall in systemic arterial pressure. As described above, the ability of nitrates to dilate epicardial coronary arteries, even in areas of atherosclerotic stenosis, is modest, and the bulk of evidence continues to favor a reduction in myocardial work and thus in myocardial oxygen demand as their primary effect in chronic stable angina. Paradoxically, high doses of organic nitrates may reduce blood pressure to such an extent that

coronary flow is compromised; reflex tachycardia and adrenergic enhancement of contractility also occur. These effects may override the salutary action of the drugs on myocardial oxygen demand and can aggravate ischemia. Additionally, sublingual nitroglycerin administration may produce bradycardia and hypotension, probably due to activation of the Bezold-Jarisch reflex (Gibbons et al., 1999). Other Effects The nitrovasodilators act on almost all smooth muscle. Bronchial smooth muscle is relaxed irrespective of the cause of the preexisting tone. The muscles of the biliary tract, including those of the gallbladder, biliary ducts, and sphincter of Oddi, are effectively relaxed. Smooth muscle of the gastrointestinal tract, including that of the esophagus, can be relaxed and its spontaneous motility decreased by nitrates both in vivo and in vitro. The effect may be transient and incomplete in vivo, but abnormal "spasm" frequently is reduced. Indeed, many incidences of atypical chest pain and "angina" are due to biliary or esophageal spasm, and these too can be relieved by nitrates. Similarly, nitrates can relax ureteral and uterine smooth muscle, but these effects are somewhat unpredictable. Mechanism of Action Nitrites, organic nitrates, nitroso compounds, and a variety of other nitrogen oxidecontaining substances (including nitroprusside; see Chapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension) lead to the formation of the reactive free radical nitric oxide (NO), which can activate guanylyl cyclase and increase the synthesis of cyclic GMP in smooth muscle and other tissues (see Murad, 1986; Molina et al., 1987). The exact mechanism(s) of denitration of the organic nitrates, with the subsequent liberation of NO, is controversial (Harrison and Bates, 1993). A cyclic GMPdependent protein kinase catalyzes the phosphorylation of various proteins in smooth muscle. Eventually, the light chain of myosin is dephosphorylated (Waldman and Murad, 1987). Phosphorylation of the myosin light chain regulates the maintenance of the contractile state in smooth muscle. The pharmacological and biochemical effects of the nitrovasodilators appear to be identical to those of an endothelium-derived relaxing factor, which has been shown to be NO (Moncada et al., 1988; Ignarro et al., 1987; Murad, 1996; Furchgott, 1996). The relationship of endogenous nitric oxide to its precursor, L-arginine, has been reviewed (Moncada and Higgs, 1993). Nitric oxide appears to function as a biological signal in many cell types (Lowenstein et al., 1994; Vane, 1994). Absorption, Fate, and Excretion The biotransformation of organic nitrates is the result of reductive hydrolysis catalyzed by the hepatic enzyme glutathioneorganic nitrate reductase. The enzyme converts the lipid-soluble organic nitrate esters into more water-soluble denitrated metabolites and inorganic nitrite. The partially denitrated metabolites are considerably less potent vasodilators than are the parent compounds. However, under certain conditions their activity may become important. Since the liver has an enormous capacity to catalyze the reduction of organic nitrates, their biotransformation is a major factor in determining oral bioavailability and duration of action. The pharmacokinetic properties of nitroglycerin and isosorbide dinitrate have been studied in the greatest detail. Nitroglycerin One molecule of nitroglycerin reacts with two molecules of reduced glutathione to produce 1,3- or 1,2-glyceryl dinitrate and oxidized glutathione (Needleman, 1975). A comparison of the maximal velocities of metabolism of the clinically used nitrates by this reductase indicates that erythrityl

tetranitrate is degraded three times faster than is nitroglycerin, while isosorbide dinitrate and pentaerythritol tetranitrate are denitrated at one-sixth and one-tenth the rate of nitroglycerin. In human beings, peak concentrations of nitroglycerin are found in plasma within 4 minutes of sublingual administration; the drug has a half-life of 1 to 3 minutes. The onset of action of nitroglycerin may be more rapid if it is delivered as a sublingual spray, rather than as a sublingual tablet (Ducharme et al., 1999). Dinitrate metabolites, which are about ten times less potent as vasodilators, appear to have a half-life of about 40 minutes. Isosorbide Dinitrate The major route of metabolism of isosorbide dinitrate in human beings is by enzymatic denitration followed by formation of glucuronide conjugates. Sublingual administration produces maximal concentrations of the drug in plasma by 6 minutes, and the fall in concentration is rapid (half-life approximately 45 minutes). The primary initial metabolites, isosorbide-2-mononitrate and isosorbide-5-mononitrate, have longer half-lives (3 to 6 hours) and are presumed to be responsible, at least in part, for the therapeutic efficacy of the drug. Isosorbide-5-Mononitrate This agent is available in tablet form. It has excellent bioavailability after oral administration as it does not undergo significant first-pass metabolism. It has a significantly longer half-life than does isosorbide dinitrate and has been formulated as a plain tablet and as a sustained-release preparation; both of which have longer durations of action than the corresponding dosage forms of isosorbide dinitrate. Correlation of Plasma Concentrations of Drug and Biological Activity Intravenous administration of nitroglycerin or the long-acting nitrates (isosorbide dinitrate, pentaerythritol tetranitrate, and erythrityl tetranitrate) in anesthetized animals produces the same transient (1 to 4 minutes) decrease in blood pressure. Relative to nitroglycerin, the potency of erythrityl tetranitrate as a vasodepressor in dogs is about 12% and that of isosorbide dinitrate 3.5%. Since denitration markedly reduces the activity of the organic nitrates, their rapid clearance from blood indicates that the transient duration of action under these conditions correlates with the concentrations of the parent compounds. The rate of hepatic denitration is characteristic of each nitrate and is influenced by hepatic blood flow or the presence of hepatic disease. In experimental animals, injection of moderate amounts of organic nitrates into the portal vein results in little or no vasodepressor activity, indicating that a substantial fraction of drug can be metabolized during its first circulation through the liver (isosorbide mononitrate is an exception). Tolerance Sublingual organic nitrates should be taken at the time of an anginal attack or in anticipation of exercise or stress. Such intermittent treatment results in reproducible cardiovascular effects. However, frequently repeated or continuous exposure to high doses of organic nitrates leads to a marked attenuation in the magnitude of most of their pharmacological effects (Anonymous, 1992; Thadani, 1992). The magnitude of tolerance is a function of dosage and the frequency of administration of the preparation. Tolerance may result from an inability of the vascular smooth muscle to convert nitroglycerin to NO, "true vascular tolerance," or to the activation of mechanisms extraneous to the vessel wall,

"pseudotolerance" (Munzel et al., 1996). Multiple mechanisms have been proposed to account for nitrate tolerance, including volume expansion, neurohumoral activation, cellular depletion of sulfhydryl groups, and the generation of free radicals (Thadani, 1992; Rutherford, 1995; Parker and Parker 1998). The administration of organic nitrates to healthy volunteers is associated within 24 hours with the activation of the renin-angiotensin-aldosterone system and with increases in plasma norepinephrine (Parker et al., 1991). Clinical data relating to the ability of agents that modify the renin-angiotensin-aldosterone system to prevent nitrate tolerance are contradictory (Dakak et al., 1990; Muisan et al., 1993; Parker and Parker 1993; Pizzulli et al., 1996; Heitzer et al., 1998). Important factors that may influence the ability of angiotensin converting enzyme (ACE) inhibitors to prevent nitrate tolerance and that can influence the interpretation of clinical trials include the dose, treatment with the ACE inhibitors prior to the initiation of nitrates, and the tissue specificity of the agent. Despite the rationale for the depletion of sulfhydryl groups leading to impaired biotransformation of nitrates to NO and thereby resulting in nitrate tolerance, experimental results to date with sulfhydryl donors have been disappointing. A more recent proposal has linked nitroglycerin tolerance to endothelium-derived superoxide generation (Munzel et al., 1995b). In the setting of patients with congestive heart failure, administration of carvedilol, but not metoprolol, doxazosin, or placebo, has resulted in the prevention of tolerance to the effects of nitroglycerin on forearm blood flow (Watanabe et al., 1998). In addition to its properties as an - and -adrenergic receptor antagonist, carvedilol possesses antioxidant properties. A similar effect on vascular superoxide generation in the reduction of nitrate tolerance has been proposed for hydralazine (Elkayam et al., 1998). Other changes that are observed in the setting of nitroglycerin tolerance include an enhanced response to vasocontrictors such as angiotensin II, serotonin, and phenylephrine. This increased sensitivity to vasoconstrictors may result from a priming effect of endothelin-1 derived from the vasculature (Mnzel et al., 1995a). Administration of nitroglycerin is associated with plasma volume expansion, which may be reflected by a decrease in hematocrit. Although diuretic therapy with hydrochlorothiazide can improve a patient's exercise duration, appropriately designed crossover trials have failed to demonstrate an effect of diuretics on nitrate tolerance (Parker et al., 1996). In contrast to these adjunctive approaches, a more effective approach is to interrupt therapy for 8 to 12 hours each day, which allows the return of efficacy. It is usually most convenient to omit dosing at night in patients with exertional angina, either by adjusting dosing intervals of oral or buccal preparations or by removing cutaneous nitroglycerin. However, patients who have angina in a pattern suggesting its precipitation by increased left ventricular filling pressures (i.e., occurring with or near episodes of orthopnea or paroxysmal nocturnal dyspnea) may benefit from continuing nitrates at night and omitting them during a quiet period during the day. Tolerance to isosorbide dinitrate may be minimized by giving this drug two or three times daily, at 7 A.M. and noon or at 7 A.M. , noon, and 5 P.M., respectively (Parker et al., 1987). Although it was thought that tolerance might be less of a problem with isosorbide-5-mononitrate, tolerance has been seen with this agent as well; an eccentric (7 or 8 A.M. and 2 or 3 P.M.) twice-daily dosing schedule appears to maintain efficacy (Parker, 1993; Thadani et al., 1994). While these approaches appear to be effective, some patients develop an increased frequency of nocturnal angina when a nitrate-free interval is employed using nitroglycerin patches; such patients may need to receive another class of antianginal agent during this period. In addition, a phenomenon referred to as the "zero-hour effect" has added to the complexity of designing an appropriate dosing regimen. In a 29-day study, the early morning exercise tolerance (before patch application) of patients using intermittent patch therapy was less than that of patients using placebo patches, though the antianginal efficacy of the patch itself was maintained (DeMots and Glasser, 1989). The clinical significance of this finding and its applicability to other dosage forms of nitroglycerin are not known. Tolerance does not appear to be a uniform phenomenon, as in some patients only partial tolerance seems to develop.

The problem of anginal rebound during nitrate-free intervals is especially troublesome in the treatment of unstable angina with intravenous nitroglycerin. If coverage of the interval with other agents is ineffective, an alternative approach that often is used is to increase gradually the dose of intravenous nitroglycerin in an attempt to overcome tolerance. This approach has not been studied carefully. A special aspect of tolerance has been observed among individuals exposed to nitroglycerin in the manufacture of explosives. If protection is inadequate, workers may experience severe headaches, dizziness, and postural weakness during the first several days of employment. Tolerance then develops, but headache and other symptoms may reappear after a few days away from the jobthe "Monday disease." The most serious effect of chronic exposure is a form of organic nitrate dependence. Workers without demonstrable organic vascular disease have been reported to have an increase in the incidence of acute coronary syndromes during the 24- to 72-hour periods away from the work environment (Morton, 1977, Parker et al., 1995). Coronary and digital arteriospasm during withdrawal and its relaxation by nitroglycerin also have been demonstrated radiographically. Because of the potential problem of nitrate dependence, it seems prudent not to withdraw nitrates abruptly from a patient who has received such therapy chronically. Toxicity and Untoward Responses Untoward responses to the therapeutic use of organic nitrates are almost all secondary to actions on the cardiovascular system. Headache is common and can be severe. It usually decreases over a few days if treatment is continued and often can be controlled by decreasing the dose. Transient episodes of dizziness, weakness, and other manifestations associated with postural hypotension may develop, particularly if the patient is standing immobile, and may occasionally progress to loss of consciousness. This reaction appears to be accentuated by alcohol. It may be seen with very low doses of nitrates in patients with autonomic dysfunction. Even in the most severe nitrate syncope, positioning and other procedures to facilitate venous return are the only therapeutic measures required. It was widely believed that nitrates can increase intraocular pressure and precipitate glaucoma, but this fear appears to be unfounded (Robertson and Stevens, 1977). All of the organic nitrates occasionally can produce drug rash. Interaction of Nitrates with Sildenafil Erectile dysfunction is a frequently encountered problem, risk factors for which parallel those of coronary artery disease [diabetes mellitus, hypertension, known heart disease (Johannes et al., 2000) and a low level of high-density lipoprotein (Feldman et al., 1994)]. Thus, it is likely that many men requiring therapy for erectile dysfunction already may be receiving (or may require, especially if they increase physical activity) antianginal therapy. The past decade has seen remarkable advances in our understanding of the physiology of penile erection (Andersson and Wagner, 1995). Cells in the corpus cavernosum produce nitric oxide during sexual arousal in response to nonadrenergic, noncholinergic neurotransmission (Kim et al., 1991; Rajfer et al., 1992; and Burnett et al., 1992). Nitric oxide stimulates the formation of cyclic GMP, which leads to relaxation of smooth muscle of the corpus cavernosum and penile arteries. The accumulation of cyclic GMP can be enhanced by inhibition of the cyclic GMP-specific phopshophodiesterase-5 (PDE5) family (Beavo et al., 1994). Sildenafil (VIAGRA) was developed as an inhibitor of PDE5 (Boolell et al., 1996) and has been demonstrated to improve erectile function in patients with various causes of erectile dysfunction (Goldstein et al., 1998). The side effects of sildenafil are largely predictable on the basis of its effects on PDE5. Headache, flushing, and rhinitis may be observed, as may dyspepsia due to relaxation of the lower esophageal

sphincter, all thought to be consequences of the inhibition of PDE5. Sildenafil also is a weak inhibitor of PDE6, the isoenzyme involved in photoreceptor signal transduction (Beavo et al., 1994), and sildenafil has been associated with visual disturbances, most notably changes in the perception of color hue or brightness (Wallis et al., 1999; Goldstein et al., 1998). Sildenafil's most important toxicity is hemodynamic. When given alone to men with severe coronary artery disease, sildenafil has modest effects on blood pressure, producing less than a 10% fall in systolic, diastolic, and mean systemic pressures and in pulmonary artery systolic and mean pressures (Herrmann et al., 2000). However, it has a significant and potentially dangerous interaction with nitrates. As discussed above, the therapeutic actions of organic nitrates are mediated via their conversion to NO with resultant increases in cyclic GMP. In the presence of a PDE5 inhibitor, nitrates cause profound increases in cyclic GMP and can produce dramatic reductions in blood pressure. Healthy male subjects pretreated with sildenafil exhibited a much greater decrease in systolic blood pressure when treated with sublingual glyceryl trinitrate, and in many a fall of more than 25 mm Hg was detected (Webb et al., 1999). This interaction between sildenafil and nitrates is the basis for the warning that sildenafil should not be prescribed to patients receiving any form of nitrate (Cheitlen et al., 1999) and dictates that patients should be questioned about the use of sildenafil within 24 hours before nitrates are administered. A period of longer than 24 hours may be needed after sildenafil for safe use of nitrates. In the event that patients develop significant hypotension following combined administration of sildenafil and a nitrate, fluids and -adrenergic receptor agonists, if needed, should be used for support (Cheitlin et al., 1999). Sildenafil is metabolized via CYP3A4, and its toxicity may be enhanced in patients who receive other substrates of this enzyme, including macrolide and imidazole antibiotics, some HMG-CoA reductase inhibitors, and highly active antiretroviral therapy (HAART; see Chapter 51: Antiretroviral Agents: Antiretroviral Agents) (Hall and Ahmad, 1999). Sildenafil also has been demonstrated to prolong cardiac repolarization by blocking the IKr (Geelen et al., 2000). Although these interactions and effects are clinically important, the overall incidence and profile of adverse events observed with sildenafil, when used without nitrates, are consistent with the expected background frequency of the same events in the treated population (Zusman et al., 1999). In patients with coronary artery disease who are not currently taking nitrates and whose exercise capacity indicates that usual sexual activity is unlikely to precipitate angina, the use of sildenafil can be considered. Such therapy needs to be individualized and appropriate warnings given about the risk of toxicity if nitrates are taken during the next 24 hours for angina. Alternative nonnitrate antianginal therapy, such as -adrenergic receptor antagonists, should be used during this time period (Cheitlin et al., 1999). Therapeutic Uses Angina Diseases that predispose to angina should be treated as part of a comprehensive therapeutic program with the primary goal being to prolong life. Such conditions as hypertension, anemia, thyrotoxicosis, obesity, heart failure, cardiac arrhythmias, and acute anxiety can precipitate anginal symptoms in many patients. The patient should be asked to stop smoking and overeating, hypertension and hyperlipidemia should be corrected (see Chapters 33: Antihypertensive Agents and the Drug Therapy of Hypertension and 36: Drug Therapy for Hypercholesterolemia and Dyslipidemia), and daily aspirin (or a thienopyridine such as clopidogrel or ticlopidine, if aspirin is not tolerated; see Chapter 55: Anticoagulant, Thrombolytic, and Antiplatelet Drugs) should be prescribed. Exposure to sympathomimetic agents (e.g., those in nasal decongestants) should be avoided. The use of drugs that modify the perception of pain is a poor approach to the treatment of angina, since the underlying myocardial ischemia is not relieved. See Table 321 for preparations

and dosages of the nitrites and organic nitrates. The rapidity of onset, the duration of action, and the likelihood of developing tolerance are related to the method of administration. Sublingual Administration Because of its rapid action, long-established efficacy, and low cost, nitroglycerin is the most useful drug among the organic nitrates that can be given sublingually. The onset of action is within 1 to 2 minutes, but the effects are undetectable by 1 hour after administration. An initial dose of 0.3 mg of nitroglycerin often will relieve pain within 3 minutes. Absorption may be limited in patients with dentures or with dry mouths. Tablets of nitroglycerin are stable but should be dispensed in glass containers and protected from moisture, light, and extremes of temperature. Active tablets usually produce a burning sensation under the tongue, but the absence of a burning sensation does not reliably predict loss of activity. Patients, especially elderly ones, differ in the ability to detect the burning sensation. Anginal pain may be prevented when the drug is used prophylactically immediately prior to exercise or stress. The smallest effective dose should be prescribed. Patients should be taught to seek medical attention immediately when three tablets taken over a 15-minute period do not relieve a sustained attack, since this situation may be indicative of myocardial infarction or another cause of the pain. The patient also should be advised that there is no virtue in trying to avoid taking sublingual nitroglycerin for anginal pain. Other nitrates that can be taken sublingually do not appear to be longer acting than nitroglycerin, as their half-lives depend only on the rate at which they are delivered to the liver. They are not more effective than nitroglycerin and often are more expensive. Oral Administration Oral nitrates often are used to provide prophylaxis against anginal episodes in patients who have more than occasional angina. They must be given in sufficient dosage to provide effective plasma levels after first-pass hepatic degradation. At low doses (e.g., 5 to 10 mg of isosorbide dinitrate) they are no more effective than placebo in decreasing the frequency of angina or increasing the patient's exercise tolerance. Clinical studies that have used higher doses of either isosorbide dinitrate (e.g., 20 mg or more orally every 4 hours) or sustained-release preparations of nitroglycerin indicate that such regimens decrease the frequency of attacks of angina and improve exercise tolerance. Effects peak at 60 to 90 minutes and last for 3 to 6 hours. Under these circumstances, the activities of less potent metabolites also may contribute to the therapeutic effect. Chronic oral administration of isosorbide dinitrate (120 to 720 mg daily) results in persistence of the parent compound and higher concentrations of metabolites in plasma. However, these doses are more likely to cause troublesome side effects and tolerance. Significant, prolonged (up to 4 hours) improvement of exercise tolerance also can be demonstrated with a sustained-release oral form of nitroglycerin, but high doses (e.g., 6.5 mg) of nitroglycerin are required. Cutaneous Administration Application of nitroglycerin ointment can relieve angina, prolong exercise capacity, and reduce ischemic ST depression with exercise for 4 hours or more. Nitroglycerin ointment (2%) is applied to the skin [2.5 to 5 cm (1 to 2 in.) as it is squeezed from the tube; it is then spread in a uniform layer]; the dosage must be adjusted for each patient. Effects are apparent within 30 to 60 minutes (although absorption is variable) and last for 4 to 6 hours. The ointment is particularly useful for controlling nocturnal angina, which commonly develops within 3 hours after the patient goes to sleep. Transdermal nitroglycerin discs utilize a nitroglycerin-impregnated polymer (bonded to an adhesive bandage) that permits gradual absorption and a continuous plasma nitrate concentration over 24 hours. The onset of action is slow, with peak effects occurring at 1 to 2 hours. To avoid tolerance

and loss of the therapeutic effect, therapy should be interrupted for at least 8 hours each day. With this regimen, long-term prophylaxis of ischemic episodes often can be attained. Transmucosal or Buccal Nitroglycerin This formulation is inserted under the upper lip above the incisors, where it adheres to the gingiva and gradually dissolves in a uniform manner. Hemodynamic effects are seen within 2 to 5 minutes, and it is therefore useful for short-term prophylaxis of angina. Nitroglycerin continues to be released into the circulation for a prolonged period, and exercise tolerance may be enhanced for up to 5 hours. Congestive Heart Failure The utility of nitrovasodilators to relieve pulmonary congestion and to increase cardiac output in congestive heart failure is well established and is addressed in Chapter 34: Pharmacological Treatment of Heart Failure. Unstable Angina Unstable angina has been considered a single entity in most therapeutic trials; it has included patients with new-onset exertional angina, with an increase in their usual pattern of angina, and with rest angina, with or without a preceding history of exertional angina. The electrocardiogram (ECG) may show either elevation or depression of the ST segment, with variable T-wave abnormalities. In patients with left main or three-vessel disease, revascularization leads to improved survival (Multicenter Study, 1978). In the remainder of patients, and in all patients prior to determination of the coronary anatomy, appropriate medical therapy provides important benefits. The pathophysiology in most patients studied involves thrombosis overlying a ruptured atherosclerotic plaque. However, there is some variability in the anatomic substrate of unstable angina, with gradually progressive atherosclerosis accounting for some cases of new-onset exertional angina, and vasospasm occurring in minimally atherosclerotic coronary vessels accounting for some cases where rest angina has never been preceded by or associated with exertional angina. It is likely that this variability accounts for the differences in therapeutic response seen in studies with differing inclusion criteria. Multiple agents are employed in the acute phase of treatment, although few have been demonstrated conclusively to reduce mortality. Aspirin (see below), by inhibiting platelet aggregation, has been shown clearly to improve survival (Kerins and FitzGerald, 1991). Heparin (either unfractionated or low-molecular-weight heparin) also appears to reduce angina and prevent infarction. These and related agents are discussed in detail in Chapters 27: Analgesic-Antipyretic and Antiinflammatory Agents and Drugs Employed in the Treatment of Gout and 55: Anticoagulant, Thrombolytic, and Antiplatelet Drugs. Nitrates are useful in reducing vasospasm and controlling angina; their administration should be initiated intravenously. Intravenous administration of nitroglycerin allows high concentrations of drug to be attained rapidly. As nitroglycerin is promptly degraded, the plasma concentration can be titrated quickly and safely using this route. If coronary vasospasm is present, intravenous nitroglycerin is likely to be effective, although the addition of a Ca 2+ channel blocker is required to achieve complete control in some patients. Because of the potential risks of profound hypotension, nitrates should be withheld and alternate antianginal therapy administered if patients have consumed sildenafil within the prior 24 hours (see above). Myocardial Infarction

Therapeutic maneuvers in myocardial infarction are directed at reducing the ultimate size of the infarct and preserving or retrieving viable tissue by reducing the oxygen demand of the myocardium. Since the proximate cause of myocardial infarction is intracoronary thrombosis, reperfusion therapies are critically important and include thrombolytic agents and direct percutaneous coronary angioplasty (Ryan et al., 1999). Thrombolytic and antiplatelet therapy are discussed in Chapter 55: Anticoagulant, Thrombolytic, and Antiplatelet Drugs. A drug that favorably alters the oxygen balance could decrease the area of myocardial damage if it were given soon after infarction. In the past, nitroglycerin was considered to be contraindicated for use in patients with acute myocardial infarction because of its ability to induce hypotension and reflex tachycardia, although it may be highly efficacious if the infarction is due to prolonged coronary spasm. Nevertheless, evidence that nitrates improve mortality in myocardial infarction is sparse. Nitrates may be most helpful in patients in whom reperfusion does not occur, despite thrombolytic agents, and may prevent adverse remodeling. The effects of nitrates in patients with acute myocardial infarction were assessed in two large trials (GISSI-3, 1994, and ISIS-4 Collaborative Group, 1995). In the GISSI-3 study, 19,394 patients with acute myocardial infarction were randomized to receive either transdermal nitroglycerin (10 mg daily with a 10-hour nitrate-free interval overnight) or placebo for 6 weeks. There were no significant effects of nitroglycerin on mortality, reinfarction, revascularization procedures, or renal dysfunction. There was a slight reduction in postinfarction angina in the nitroglycerin-treated group (20.7% vs. 19.7%). A similar lack of benefit on mortality was observed in the ISIS-4 study of 58,050 patients, some of whom were randomized to receive an oral controlled-release form of isosorbide mononitrate, 60 mg each morning for 28 days, or placebo. In the ISIS-4 study there was no effect of the nitrate on postinfarction angina. Thus transdermal or oral nitrates are safe and well tolerated in the setting of myocardial infarction and may have a beneficial effect on pain, but they do not provide a survival advantage. Variant (Prinzmetal's) Angina The large coronary arteries normally contribute little to coronary resistance. However, in variant angina, coronary constriction results in reduced blood flow and ischemic pain. Multiple mechanisms have been hypothesized to be involved in the initiation of vasospasm, including endothelial cell injury (Freisinger and Robertson, 1986). It does not seem likely that abnormalities of sympathetic input are etiologic (Robertson et al., 1979), and -adrenergic receptor antagonists can be deleterious (Robertson et al., 1982). Despite the presence of abnormal coronary anatomy in all cases that have come to autopsy, it does not appear that active intravascular platelet aggregation is a precipitating factor, and aspirin does not appear to provide benefit (Robertson et al., 1981). Whereas long-acting nitrates alone are occasionally efficacious in abolishing episodes of variant 2+ 2+ angina, more often additional therapy with Ca channel blockers is required. Because Ca channel blockers, but not nitrates, have been shown to favorably influence mortality and the incidence of myocardial infarction in variant angina, they should be included in therapy. Ca2+ Channel Antagonists History The work in the 1960s of Fleckenstein, Godfraind, and their colleagues led to the concept that drugs can alter cardiac and smooth muscle contraction by blocking the entry of Ca2+ into myocytes. Godfraind and associates showed that the ability of the diphenylpiperazine analogs cinnarizine and lidoflazine to prevent vascular smooth muscle contraction induced by some agonists could be overcome by raising the concentration of Ca2+ in the extracellular medium; they used the term

"calcium antagonist" to describe these agents (see Godfraind and Kaba, 1972; Godfraind et al., 1986). Hass and Hartfelder reported in 1962 that verapamil, a putative coronary vasodilator, possessed negative inotropic and chronotropic effects that were not seen with other vasodilatory agents, such as nitroglycerin. In 1967, Fleckenstein suggested that the negative inotropic effect resulted from inhibition of excitationcontraction coupling and that the mechanism involved reduction of the movement of Ca2+ into cardiac myocytes (see Fleckenstein, 1983). A derivative of verapamil, gallopamil, and other compounds, such as nifedipine (Kohlhardt and Fleckenstein, 1977), also were shown to block the movement of Ca2+ through the cardiac myocyte Ca2+ channel, or the slow channel (see Chapter 35: Antiarrhythmic Drugs), and thereby alter the plateau phase of the cardiac action potential. Subsequently, many drugs in several chemical classes have been shown to alter cardiac and smooth muscle contraction by blocking or "antagonizing" the entry of Ca2+ through channels in the myocyte membrane. Chemistry The ten Ca2+ channel antagonists that have been approved for clinical use in the United States have diverse chemical structures. Five classes of compounds have been examined: phenylalkylamines, dihydropyridines, benzothiazepines, diphenylpiperazines, and a diarylaminopropylamine. At present, verapamil (a phenylalkylamine); diltiazem (a benzothiazepine); nicardipine, nifedipine, isradipine, amlodipine, felodipine, nisoldipine, and nimodipine (dihydropyridines); and bepridil (a diarylaminopropylamine ether) are approved for clinical use in the United States. Their structures are shown in Table 322. Pharmacological Properties Cardiovascular Effects Actions in Vascular Tissue Although there is some involvement of Na currents, depolarization of vascular smooth muscle cells 2+ is primarily dependent on the influx of Ca (Bolton, 1979). At least three distinct mechanisms may 2+ be responsible for contraction of vascular smooth muscle cells. First, voltage-sensitive Ca 2+ channels open in response to depolarization of the membrane, and extracellular Ca moves down its electrochemical gradient into the cell. After closure of Ca2+ channels, a finite period of time is required before the channels can open again in response to a stimulus. Second, agonist-induced contractions that occur without depolarization of the membrane result from the hydrolysis of membrane phosphatidylinositol with the formation of inositol trisphosphate, which acts as a second messenger to release intracellular Ca2+ from the sarcoplasmic reticulum (see Berridge, 1993). This receptor-mediated release of intracellular Ca 2+ may trigger further influx of extracellular Ca2+. Third, receptor-operated Ca2+ channels allow the entry of extracellular Ca2+ in response to receptor occupancy. An increase in cytosolic Ca2+ results in enhanced binding of Ca2+ to the protein calmodulin. The Ca2+calmodulin complex in turn activates myosin light-chain kinase, with resultant phosphorylation of the light chain of myosin. Such phosphorylation promotes interaction between actin and myosin and contraction of smooth muscle. Ca2+ channel antagonists or blockers inhibit the voltage-dependent Ca2+ channels in vascular smooth muscle at significantly lower concentrations than are required to interfere with the release of intracellular Ca2+ or to block receptor-operated Ca2+ channels. All Ca2+ channel blockers relax arterial smooth muscle, but they have little effect on most
+

venous beds and hence do not affect cardiac preload significantly. Actions in Cardiac Cells The mechanisms involved in excitationcontraction coupling in the heart differ from those in vascular smooth muscle in that a portion of the two inward currents is carried by Na+ through the fast channel in addition to that carried by Ca2+ through the slow channel. In the sinoatrial (SA) and atrioventricular (AV) nodes, depolarization is largely dependent on the movement of Ca2+ through the slow channel. Within the cardiac myocyte, Ca2+ binds to troponin, the inhibitory effect of troponin on the contractile apparatus is relieved, and actin and myosin interact to cause contraction. Thus, Ca2+ channel blockers can produce a negative inotropic effect. Although this is true of all classes of Ca2+ channel blockers, the greater degree of peripheral vasodilation seen with the dihydropyridines is accompanied by sufficient baroreflex-mediated increase in sympathetic tone to overcome the negative inotropic effect. Diltiazem also may inhibit mitochondrial Na +Ca 2+ exchange (Schwartz, 1992). The effect of a Ca2+ channel blocker on atrioventricular conduction and on the rate of the sinus node pacemaker is dependent on whether or not the agent delays the recovery of the slow channel (Henry, 1983). Although nifedipine reduces the slow inward current in a dose-dependent manner, it does not affect the rate of recovery of the slow Ca2+ channel (Kohlhardt and Fleckenstein, 1977). The channel blockade caused by nifedipine and related dihydropyridines also shows little dependence on the frequency of stimulation. At doses used clinically, nifedipine does not affect conduction through the node. In contrast, verapamil not only reduces the magnitude of the Ca2+ current through the slow channel but also decreases the rate of recovery of the channel. In addition, channel blockade caused by verapamil (and to a lesser extent by diltiazem) is enhanced as the frequency of stimulation increases, a phenomenon known as "frequency dependence" or "use dependence." Verapamil and diltiazem depress the rate of the sinus node pacemaker and slow AV conduction; the latter effect is the basis for their use in the treatment of supraventricular tachyarrhythmias (see Chapter 35: Antiarrhythmic Drugs). Bepridil, like verapamil, inhibits both slow inward Ca2+ current and fast inward Na + current. It has a direct negative inotropic effect. Its electrophysiologic properties lead to slowing of the heart rate, prolongation of the AV nodal effective refractory period, and, importantly, prolongation of the QTc interval. Particularly in the setting of hypokalemia, the last effect can be associated with torsades de pointes, a potentially lethal ventricular arrhythmia (see Chapter 35: Antiarrhythmic Drugs). Hemodynamic Effects All of the Ca2+ channel blockers that have been approved for clinical use decrease coronary vascular resistance and increase coronary blood flow. The dihydropyridines are more potent vasodilators in vivo and in vitro than is verapamil, which is more potent than diltiazem. The hemodynamic effects of each of these agents vary, depending on the route of administration and the extent of left ventricular dysfunction. Nifedipine given intravenously increases forearm blood flow with little effect on venous pooling; this indicates a selective dilation of arterial resistance vessels. The decrease in arterial blood pressure elicits sympathetic reflexes, with resultant tachycardia and positive inotropy. Nifedipine also has direct negative inotropic effects in vitro. However, nifedipine relaxes vascular smooth muscle at significantly lower concentrations than those required for prominent direct effects on the heart. Thus, arteriolar resistance and blood pressure are lowered, contractility and segmental ventricular function are improved, and heart rate and cardiac output are increased modestly (Serruys et al., 1981; Theroux et al., 1980). After oral administration of nifedipine, arterial dilation increases

peripheral blood flow; venous tone does not change. The other dihydropyridinesnicardipine, amlodipine, isradipine, felodipine, nisoldipine, and nimodipineshare many of the cardiovascular effects of nifedipine. There may be some selectivity of nicardipine for coronary vessels compared with peripheral vessels (Pepine and Lambert, 1988); in comparative studies, nicardipine appears to produce fewer side effects, such as dizziness, than does nifedipine but has equivalent antianginal efficacy (DeWood and Wohlbach, 1990). Intravenous or oral administration of nicardipine results in decreases in systolic and diastolic blood pressure that are accompanied by an increase in cardiac output because of the reduction in afterload and compensatory increases in heart rate and ejection fraction. It also appears to reduce left ventricular diastolic dysfunction (Hanet et al., 1990). Nicardipine decreases the frequency of anginal attacks and improves exercise tolerance in patients with effort-induced angina (Pepine and Lambert, 1988). Amlodipine is a dihydropyridine that has slow absorption and a prolonged effect. With a plasma half-life of 35 to 50 hours, plasma levels and effect increase over 7 to 10 days of therapy. Amlodipine produces both peripheral arterial vasodilation and coronary dilation, with a hemodynamic profile similar to that of nifedipine. However, there is less reflex tachycardia with amlodipine, possibly because the long half-life produces minimal peaks and troughs in plasma concentrations (van Zwieten and Pfaffendorf, 1993; Taylor, 1994; Lehmann et al., 1993). Felodipine appears to have even greater vascular specificity than does nifedipine or amlodipine. At concentrations producing vasodilation, there is no negative inotropic effect. Like nifedipine, felodipine produces activation of the sympathetic nervous system, leading to an increase in heart rate (Todd and Faulds, 1992). Isradipine also produces the typical peripheral vasodilation seen with other dihydropyridines, but because of its inhibitory effect on the SA node, little or no rise in heart rate is seen. This inhibitory effect does not extend to the myocardium, however, as no cardiodepressant effect is seen. Despite the negative chronotropic effect, isradipine appears to have little effect on the AV node, so it may be used in patients with AV block or combined with a adrenergic receptor antagonist. In general, because of their lack of myocardial depression and, to a greater or lesser extent, lack of negative chronotropic effect, dihydropyridines are less effective as monotherapy in stable angina than are verapamil, diltiazem, or a -adrenergic receptor antagonist. Nisoldipine is more than 1000-times as potent in preventing contraction of human vascular smooth muscle than in preventing contraction of human cardiac muscle in vitro, suggesting a very high degree of vascular selectivity (Godfraind et al., 1992). This selectivity has been confirmed by in vitro studies. Although nisoldipine has a short elimination half-life, a sustained-release preparation has been developed, nisoldipine coat-core, that has been demonstrated to be as effective an antianginal agent as amlodipine or diltiazem (Langtry and Spencer, 1997). Nimodipine, because of its high lipid solubility, was developed as an agent to relax the cerebral vasculature. It is effective in inhibiting cerebral vasospasm and is used primarily to treat patients with neurological defects thought to be caused by vasospasm after subarachnoid hemorrhage. Bepridil has been demonstrated to reduce blood pressure and heart rate in patients with stable exertional angina. It also produces an increase in left ventricular performance in patients with angina, but its side-effect profile (see below) limits its use to truly refractory patients (Zusman et al., 1993; Hollingshead et al., 1992). Verapamil is a less potent vasodilator in vivo than are the dihydropyridines. Like the latter agents, verapamil causes little effect on venous resistance vessels at concentrations that produce arteriolar dilation. With doses of verapamil sufficient to produce peripheral arterial vasodilation, there are more direct negative chronotropic, dromotropic, and inotropic effects than with the dihydropyridines. Intravenous verapamil causes a decrease in arterial blood pressure due to a decrease in vascular resistance, but the reflex tachycardia is blunted or abolished by the direct negative chronotropic effect of the drug. The intrinsic negative inotropic effect of verapamil is

partially offset by both a decrease in afterload and the reflex increase in adrenergic tone. Thus, in patients without congestive heart failure, ventricular performance is not impaired and may actually improve, especially if ischemia is limiting performance. In contrast, in patients with congestive heart failure, intravenous verapamil can cause a marked decrease in contractility and left ventricular function. Oral administration of verapamil results in reduction of peripheral vascular resistance and blood pressure with no change in heart rate (Theroux et al., 1980). The relief of pacing-induced angina seen with verapamil is due primarily to a reduction in myocardial oxygen demand (Rouleau et al., 1983). Intravenous administration of diltiazem can result initially in a marked decrease in peripheral vascular resistance and arterial blood pressure, which elicits a reflex increase in heart rate and cardiac output. Heart rate then falls below initial levels because of the direct negative chronotropic effect of the agent. Oral administration of diltiazem results in a sustained fall in both heart rate and mean arterial blood pressure (Theroux et al., 1980). Despite the fact that diltiazem and verapamil produce similar effects on the SA and AV nodes, the negative inotropic effect of diltiazem is more modest. The effect of Ca2+ channel blockers on diastolic ventricular relaxation (the lusitropic state of the ventricle) is complex. The direct effect of several of these agents, assessed when they are given by the intracoronary route, is to impair relaxation (Rousseau et al., 1980; Amende et al., 1983; Serruys et al., 1983; Walsh and O'Rourke, 1985). Although several clinical studies have suggested an improvement in peak left ventricular filling rates when verapamil, nifedipine, nisoldipine, or nicardipine was given systemically (Bonow et al., 1982; Paulus et al., 1983; Rodrigues et al., 1987; DEFIANT-II Research Group, 1997), one must be cautious in extrapolating this change in filling rates to enhancement of relaxation. Indeed, in studies by Nishimura et al. (1993), verapamil increased peak filling rate but also increased left ventricular end-diastolic pressure. Because ventricular relaxation is modulated at several levels (Brutsaert et al., 1993), the effect of even a single agent may be complex. If reflex stimulation of sympathetic tone increases myocardial cyclic AMP levels, increased lusitropy will result and may outweigh a direct negative lusitropic effect. Likewise, a reduction in afterload will improve the lusitropic state. In addition, if ischemia is improved, the negative lusitropic effect of asymmetrical left ventricular contraction will be reduced. However, in any given patient, the sum total of these effects cannot be determined a priori. Thus, caution should be exercised in the use of Ca2+ channel blockers for this purpose; it is ideal if the end result can be determined objectively before committing the patient to therapy. Mechanisms of Action Increased concentrations of cytosolic Ca2+ cause increased contraction of cardiac and vascular smooth muscle cells. The entry of extracellular Ca2+ is more important in initiating the contraction of myocardial cells, while the release of Ca2+ from intracellular storage sites also participates in contraction of vascular smooth muscle, particularly in some vascular beds. In addition, the entry of extracellular Ca2+ can trigger the release of additional Ca2+ from intracellular stores. Cytosolic Ca2+ concentrations may be increased by various contractile stimuli. Thus, many hormones and neurohormones increase Ca2+ influx through so-called receptor-operated channels, while high external concentrations of K+ and depolarizing electrical stimuli increase Ca2+ influx through voltage-sensitive, or "potential-operated," channels (Bevan et al., 1982). Voltage-sensitive channels contain domains of homologous sequence that are arranged in tandem within a single large subunit. In addition to the major channel-forming subunit (termed 1), Ca2+

channels contain several other associated subunits (termed

2,

, , and ; see Schwartz, 1992).

Voltage-sensitive Ca2+ channels have been divided into at least three subtypes based on their conductances and sensitivities to voltage (Schwartz, 1992; Tsien et al., 1988). The channels best characterized to date are the L, N, and T subtypes, although P/Q and R channels have been identified. Only the L-type channel is sensitive to the dihydropyridine Ca2+ channel blockers. Large divalent cations such as Cd2+ and Mn2+ block a wider range of Ca2+ channels. All approved Ca2+ channel blockers bind to the 1 subunit of the L-type calcium channel, which is the main poreforming unit of the channel. This 200,000- to 250,000-dalton subunit is associated with a disulfidelinked 2 subunit of approximately 140,000 daltons and an intracellular subunit of 55,000 to 72,000 daltons. The 1 subunits share a common topology of four homologous domains (I, II, III and IV), each of which is composed of six putative transmembrane segments (S1S6). The 2 and subunits modulate the 1 subunit. The phenylalkylamine Ca2+ channel blockers bind to transmembrane segment 6 of domain IV (IVS6), the benzothiazepine Ca2+ channel blockers bind to the cytoplasmic bridge between domain III (IIIS) and domain IV (IVS), and the dihydropyridine Ca2+ channel blockers bind to transmembrane segment of both domain III (IIIS6) and domain IV (IVS6). These three separate receptor sites are allosterically linked (Hockerman et al., 1997, Abernethy and Schwartz, 1999). The vascular and cardiac effects of some of the Ca2+ channel blockers are summarized below and in Table 322. Absorption, Fate, and Excretion Although the absorption of these agents is nearly complete after oral administration, their bioavailability is reduced, in some cases markedly, because of first-pass hepatic metabolism. The effects of these drugs are evident within 30 to 60 minutes of an oral dose, with the exception of the more slowly absorbed and longer-acting agents amlodipine, isradipine, and felodipine. For comparison, peak effects of verapamil occur within 15 minutes of its intravenous administration. These agents all are bound to plasma proteins to a significant extent (70% to 98%); their elimination half-lives are widely variable and may range from 1.3 to 64 hours. During repeated oral administration, bioavailability and half-life may increase because of saturation of hepatic metabolism. A major metabolite of diltiazem is desacetyldiltiazem, which has about one-half of diltiazem's potency as a vasodilator. N-Demethylation of verapamil results in production of norverapamil, which is biologically active but much less potent than the parent compound. The half-life of norverapamil is about 10 hours. The metabolites of the dihydropyridines are inactive or weakly active. In patients with hepatic cirrhosis, the bio availabilities and half-lives of the Ca 2+ channel blockers may be increased, and dosage should be decreased accordingly. The half-lives of these agents also may be longer in older patients. Except for diltiazem and nifedipine, all of the Ca2+ channel blockers are administered as racemic mixtures (Abernethy and Schwartz, 1999). Toxicity and Untoward Responses The most common side effects caused by the Ca2+ channel antagonists, particularly the dihydropyridines, are due to excessive vasodilation. These effects may be expressed as dizziness, hypotension, headache, flushing, digital dysesthesia, and nausea. Patients also may experience constipation, peripheral edema, coughing, wheezing, and pulmonary edema. Nimodipine may produce muscular cramps when given in the large doses required for a beneficial effect in patients with subarachnoid hemorrhage. Less common side effects include rashes, somnolence, and occasional minor elevations of liver function tests. These side effects usually are benign and may abate with time or with adjustment of the dose. Aggravation of myocardial ischemia has been

observed in two studies with the dihydropyridine nifedipine (Schulz et al., 1985; Egstrup and Anderson, 1993). In both of these studies, worsening of angina was observed in patients with an angiographically demonstrable coronary collateral circulation. The worsening of angina may have resulted from excessive hypotension and decreased coronary perfusion, selective coronary vasodilation in nonischemic regions of the myocardium (i.e., coronary steal, since vessels perfusing ischemic regions may already be maximally dilated), or an increase in oxygen demand owing to increased sympathetic tone and excessive tachycardia. In a study of monotherapy with an immediate-release formulation of nisoldipine, the dihydropyridine was not superior to placebo therapy and was associated with a trend toward an increased incidence of serious adverse events (Thadani et al., 1991), a process described by Waters (1991) as proischemia. Although bradycardia, transient asystole, and exacerbation of heart failure have been reported with verapamil, these responses usually have occurred after intravenous administration of verapamil, in patients with disease of the SA node or AV nodal conduction disturbances, or in the presence of adrenergic receptor blockade. The use of intravenous verapamil with a -adrenergic receptor antagonist is contraindicated because of the increased propensity for atrioventricular block and/or severe depression of ventricular function. Patients with ventricular dysfunction, SA or AV nodal conduction disturbances, and systolic blood pressures below 90 mm Hg should not be treated with verapamil or diltiazem, particularly intravenously. Some Ca2+ channel antagonists can cause an increase in the concentration of digoxin in plasma, although toxicity from the cardiac glycoside rarely develops. The use of verapamil to treat digitalis toxicity is thus contraindicated; AV nodal conduction disturbances may be exacerbated. Bepridil, because of its antiarrhythmic properties and its ability to prolong the QTc interval, can produce serious arrhythmic side effects. Especially in the setting of hypokalemia and/or bradycardia, polymorphic ventricular tachycardia (torsades de pointes), a potentially lethal arrhythmia, can be seen. Agranulocytosis also has been reported. Because of these serious side effects, this agent should be reserved for patients refractory to all other appropriate medical and surgical therapy (Hollingshead et al., 1992). A novel Ca channel blocker, mibefradil, is an example of an agent that inhibits both the T- and L2+ 2+ type Ca channels. The T-type Ca channel contributes to the spontaneous contractile function of smooth muscle cells (Mishra and Hermsmeyer, 1994). Mibefradil was demonstrated to be effective in reducing the frequency and duration of asymptomatic ischemic episodes in patients with stable exertional angina pectoris and asymptomatic ischemia (Braun et al., 1996) and received approval from the United States Food and Drug Administration (FDA) for use as an antianginal agent. However, it subsequently was withdrawn from the market due to adverse drug interactions, possibly due to its dual inhibition of both the P-glycoprotein and CYP3A systems (Wandel et al., 2000). Recent observational studies and a metaanalysis have raised concerns about the long-term safety of 2+ the Ca channel blockers, and in particular, short-acting nifedipine preparations (Psaty et al., 1995; Pahor et al., 1995; Furberg et al., 1995). Authors of a recent analysis based on a total of 100 clinical 2+ studies of Ca channel blockers concluded that observational studies and randomized clinical trials give concordant evidence linking adverse safety effects to short-acting Ca2+ channel blockers, specifically to short-acting nifedipine (Opie et al., 2000). The proposed hypothesis for this adverse effect lies in abrupt vasodilation with reflex sympathetic activation. A similar conclusion was reached by Stason et al. (1999), who performed a systematic review of the literature on nifedipine and determined that adverse effects were observed in patients on monotherapy with an immediaterelease formulation of nifedipine. Therapeutic Uses
2+

Variant Angina Variant angina is a direct result of a reduction in flow, not the result of an increase in oxygen demand. Controlled clinical trials have demonstrated efficacy of the Ca2+ channel blocking agents for the treatment of variant angina (Antman et al., 1980; Severi et al., 1980). These drugs can attenuate ergonovine-induced vasospasm in patients with variant angina, which suggests that protection in variant angina is due to coronary dilation rather than to alterations in peripheral hemodynamics (Waters et al., 1981). Exertional Angina Ca2+ channel antagonists also are effective in the treatment of exertional, or exercise-induced, angina. The utility of these agents may result from an increase in blood flow due to coronary arterial dilation, from a decrease in myocardial oxygen demand (secondary to a decrease in arterial blood pressure, heart rate, or contractility), or from both. Numerous double-blind placebo-controlled studies have shown that these drugs decrease the number of anginal attacks and attenuate exerciseinduced depression of the ST segment. The "double product," which is calculated as heart rate x systolic blood pressure, is an indirect measure of myocardial oxygen demand. Since these agents reduce the level of the double product (or oxygen demand) at a given external work load, and the value of the double product at peak exercise is not altered, the beneficial effect of Ca2+ channel blockers likely is due primarily to a decrease in oxygen demand rather than to an increase in coronary flow. As described above, Ca2+ channel antagonists, particularly the dihydropyridines, may aggravate anginal symptoms in some patients when used without a -adrenergic receptor antagonist. This adverse effect is not prominent with verapamil or diltiazem because of their limited ability to induce marked peripheral vasodilation and reflex tachycardia. Concurrent therapy with nifedipine and the -adrenergic receptor antagonist propranolol, or with amlodipine and any of several -adrenergic receptor antagonists, has proven more effective than either agent given alone in exertional angina, presumably because the -adrenergic receptor antagonist suppresses reflex tachycardia (Bassan et al., 1982; Lehmann et al., 1993). This concurrent drug therapy is particularly attractive, since the dihydropyridines, unlike verapamil and diltiazem, do not delay atrioventricular conduction and will not enhance the negative dromotropic effects associated with -adrenergic receptor blockade. Although concurrent administration of verapamil or diltiazem with a -adrenergic receptor antagonist also may reduce angina, the potential for atrioventricular block, severe bradycardia, and decreased left ventricular function requires that these combinations be used judiciously (Packer, 1989), especially if left ventricular function is compromised prior to therapy. Amlodipine produces less reflex tachycardia than does nifedipine, probably because of a flatter plasma concentration profile. Isradipine, approximately equivalent to nifedipine in enhancing exercise tolerance, also produces less rise in heart rate, possibly because of its slow onset of action. Unstable Angina Medical therapy for unstable angina involves the administration of aspirin, which reduces mortality, and nitrates, -adrenergic receptor blocking agents, and heparin, which are effective in controlling pain and ischemic episodes. Since vasospasm occurs in some patients with unstable angina (Hugenholtz et al., 1981), Ca2+ channel blockers offer an additional approach to the treatment of unstable angina. However, there is insufficient evidence to assess whether or not such treatment actually decreases mortality except in patients in whom the principal mechanism is vasospasm. In a randomized, double-blind, clinical trial, the short-acting dihydropyridine nifedipine was found to be

less effective than was metoprolol (Muller et al., 1984), and there are no studies supporting the administration of a dihydropyridine to patients with unstable angina. One small study of 121 patients reported a benefit of intravenous diltiazem, compared to nitroglycerin, on the end-points of refractory angina or event-free survival (Gbel et al., 1995). In contrast, therapy directed toward reduction of platelet function and thrombotic episodes clearly decreases morbidity and mortality in patients with unstable angina (see Chapters 27: Analgesic-Antipyretic and Antiinflammatory Agents and Drugs Employed in the Treatment of Gout and 55: Anticoagulant, Thrombolytic, and Antiplatelet Drugs). Myocardial Infarction There is no evidence that Ca2+ channel antagonists are of benefit in the early treatment or secondary prevention of acute myocardial infarction, and in several trials, the short-acting formulation of the dihydropyridine nifedipine appears to have had a detrimental effect on mortality at higher doses (Kloner, 1995; Opie and Messerli, 1995; Yusuf, 1995; Furberg et al., 1995). Diltiazem and verapamil may reduce the incidence of reinfarction in patients with a first non-Q-wave infarction who are not candidates for a -adrenergic receptor antagonist (Ryan et al., 1999). Other Uses The use of Ca2+ channel antagonists as antiarrhythmic agents is discussed in Chapter 35: Antiarrhythmic Drugs, and their use for the treatment of hypertension is discussed in Chapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension. Clinical trials are under way to evaluate the capacity of Ca2+ channel blockers to slow the progression of renal failure and to protect the transplanted kidney. Verapamil has been demonstrated to improve left ventricular outflow obstruction and symptoms in patients with hypertrophic cardiomyopathy. Verapamil also has been used in the prophylaxis of migraine headaches. While several studies suggest that dihydropyridines may suppress the progression of mild atherosclerosis, there is no evidence that this alters mortality or reduces the incidence of ischemic events. Nimodipine has been approved for use in patients with neurological deficits secondary to cerebral vasospasm after the rupture of a congenital intracranial aneurysm. Nifedipine, diltiazem, and felodipine have been shown to provide symptomatic relief in Raynaud's disease. Ca2+ channel antagonists cause relaxation of the myometrium in vitro and markedly inhibit the amplitude of spontaneous and oxytocin-induced contractions. Clinical studies have shown Ca2+ channel blockers to be effective in stopping preterm uterine contractions (Murray et al., 1992; Childress and Katz, 1994; Evidence Report/Technology Assessment, 2000). In studies comparing nifedipine with the 2-adrenergic receptor agonist ritodrine, nifedipine has been found to be at least as effective as ritodrine in stopping contractions and to be associated with fewer maternal side effects and a lower incidence of neonatal morbidity (Koks et al., 1998; Garcia-Velasco and Gonzalez Gonzalez, 1998; Oei et al., 1999; Papatsonis et al., 2000). -Adrenergic Receptor Antagonists The -adrenergic receptor antagonists are effective in reducing the severity and frequency of attacks of exertional angina and improve survival in patients who have had a myocardial infarction. In contrast, these agents are not useful for vasospastic angina and, if used in isolation, may worsen the condition (Robertson et al., 1982). Most -adrenergic receptor antagonists appear to be equally effective in the treatment of exertional angina (Thadani et al., 1980). Timolol, metoprolol, atenolol, and propranolol have been shown to exert cardioprotective effects. The effectiveness of adrenergic receptor antagonists in the treatment of exertional angina is attributable primarily to a

fall in myocardial oxygen consumption at rest and during exertion, although there also is some tendency for increased flow toward ischemic regions. The decrease in myocardial oxygen consumption is due to a negative chronotropic effect (particularly during exercise), a negative inotropic effect, and a reduction in arterial blood pressure (particularly systolic pressure) during exercise. Not all the actions of -adrenergic receptor antagonists are beneficial in all patients. The decrease in heart rate and contractility causes an increase in the systolic ejection period and an increase in left ventricular end-diastolic volume; this tends to increase oxygen consumption. However, the net effect of -adrenergic blockade is usually to decrease myocardial oxygen consumption, particularly during exercise. Nevertheless, in patients with limited cardiac reserve who are critically dependent on adrenergic stimulation, -adrenergic receptor blockade can result in profound decreases in left ventricular function. Despite this, several -adrenergic receptor antagonists have been shown to reduce mortality in patients with congestive heart failure (see Chapter 34: Pharmacological Treatment of Heart Failure). Numerous -adrenergic receptor antagonists are approved for clinical use in the United States. They are considered in detail in Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists. Therapeutic Uses Unstable Angina -Adrenergic receptor antagonists are effective in reducing recurrent episodes of ischemia and reduce the risk of progression to acute myocardial infarction (Braunwald et al., 2000). Currently available results of clinical trials have not had sufficient statistical power to define effects of adrenergic receptor antagonists on mortality. On the other hand, if the underlying pathophysiology is coronary vasospasm, nitrates and Ca2+ channel blockers will be effective, and -adrenergic receptor antagonists should not be used alone. In some patients, there is a combination of severe fixed disease and superimposed vasospasm; if adequate antiplatelet therapy and vasodilation have been provided by other agents and angina continues, the addition of a -adrenergic receptor antagonist may be helpful. Myocardial Infarction -Adrenergic receptor antagonists that do not have intrinsic sympathomimetic activity have been demonstrated clearly to improve mortality in myocardial infarction. They should be given early and continued for 2 to 3 years in all patients who can tolerate them (Ryan et al., 1999). Comparison of Antianginal Therapeutic Strategies In evaluating trials in which different forms of antianginal therapy are compared, careful attention must be paid to the patient population studied, including the pathophysiology and stage of the disease. It also is important to realize that an important placebo effect may be seen in these trials. The efficacy of antianginal treatment will depend on the severity of angina, on the presence of coronary vasospasm, and on the factors underlying myocardial oxygen demand. It also is most helpful if the dose of each agent is titrated to maximum benefit. Considering the precipitants of angina in a given patient often is helpful. In patients with normal left ventricular function who have predictable angina with exertion despite nitrate therapy, -adrenergic receptor antagonists often will be beneficial due to their effects on heart rate and blood pressure. However, in patients with impaired ventricular performance and severe coronary disease, adrenergic receptor blockade may lead to further elevation of end-diastolic pressure and an increase in oxygen demand. Task forces from the European Society of Cardiology (Anonymous, 1997) and

both the American College of Cardiology (ACC) and the American Heart Association (AHA) (Gibbons et al., 1999) have developed current guidelines that are useful in the selection of an appropriate initial form of therapy for patients with chronic stable angina pectoris. Table 323 summarizes the issues that the ACC/AHA task force considered to be relevant in choosing between -adrenergic receptor antagonists and Ca2+ channel blockers in patients with angina and other medical conditions. A recent metaanalysis of publications that compared two or more antianginal therapies has been conducted by Heidenreich et al., (1999). From the comparison of -adrenergic receptor antagonists with Ca2+ channel blockers (72 studies), the authors concluded that adrenergic receptor antagonists were associated with fewer episodes of angina per week and a lower rate of withdrawal because of adverse events. However, there were no overall differences in effects on time to ischemia during exercise, and in the rate of adverse events when Ca2+ channel blockers other than nifedipine were compared with -adrenergic receptor antagonists. There were no significant differences in outcome between the studies comparing long-acting nitrates and Ca2+ channel blockers and the studies comparing long-acting nitrates with -adrenergic receptor antagonists. Combination Therapy Since the different categories of antianginal agents utilize different mechanisms of action, it has been suggested that combinations of these agents would allow the use of lower doses, increasing effectiveness and reducing the incidence of side effects. However, despite the potential advantages, combination therapy in practice rarely fully achieves this potential and may be accompanied by serious side effects. Nitrates and -Adrenergic Receptor Antagonists The concurrent use of organic nitrates and -adrenergic receptor antagonists can be very effective in the treatment of typical exertional angina. The additive efficacy is primarily a result of one drug blocking the adverse effects of the other agent on net myocardial oxygen consumption. Adrenergic receptor antagonists can block the reflex tachycardia and positive inotropic effects that are sometimes associated with nitrates. Nitrates can attenuate the increase in left ventricular enddiastolic volume associated with -adrenergic blockade by increasing venous capacitance. Concurrent administration of nitrates also can alleviate the increase in coronary vascular resistance associated with blockade of -adrenergic receptors. Ca2+ Channel Blockers and -Adrenergic Receptor Antagonists When angina is not adequately controlled by nitrates and a -adrenergic receptor antagonist, additional improvement sometimes can be achieved by the addition of a Ca2+ channel blocker, especially if there is a component of coronary vasospasm. If the patient already is being treated with maximal doses of verapamil or diltiazem, it is difficult to demonstrate any additional benefit of adrenergic blockade, and excessive bradycardia, heart block, or heart failure may ensue. However, in patients treated with a dihydropyridine, such as nifedipine, or with nitrates, there often is substantial reflex tachycardia that limits the effectiveness of these agents. A -adrenergic receptor antagonist may be a helpful addition in this situation, resulting in a lower heart rate and blood pressure with exercise. The efficacy of amlodipine also has been shown to be improved by combination with a -adrenergic receptor antagonist. However, in the Total Ischaemic Burden European Trial (TIBET), which compared the effects of atenolol, a sustained-release form of nifedipine, and their combination on exercise parameters and ambulatory ischemia in 608 patients with mild angina, there were no differences between the agents, either singly or in combination, on any of the measured ischemic parameters (Fox et al., 1996). On the other hand, in two studies of

patients with more severe but still stable angina, atenolol and propranolol were shown to be superior to nifedipine, and the combination of propranolol and nifedipine was more effective than nifedipine alone (Fox et al., 1993). In the IMAGE trial, the combination of metoprolol and sustainedrelease nifedipine was compared to each agent alone (Savonitto et al., 1996). Although an additional effect was observed in the group of patients receiving combination therapy, this was the result of an increase in exercise tolerance in individual patients who had not exhibited an increase during monotherapy; this recruitment effect was more marked in the patients in whom metoprolol was added to nifedipine. Fluctuations in coronary tone long have been recognized as primary in variant angina. It also is likely that increased tone superimposed on fixed disease plays a role in the variable anginal threshold seen in many patients with otherwise chronic stable angina and possibly in ischemic episodes precipitated by cold and by emotion (Zeiher et al., 1991). Increased coronary tone also may be important in the anginal episodes occurring early after myocardial infarction (Bertrand et al., 1982) and after coronary angioplasty, and it probably accounts for those patients with unstable angina who respond to dihydropyridines (Hugenholtz et al., 1981). Atherosclerotic arteries have abnormal vasomotor responses to a number of stimuli (Kaplinsky, 1992; Oemar et al., 1998), including exercise, other forms of sympathetic activation, and cholinergic agonists; in such vessels, stenotic segments actually may become more severely stenosed during exertion. This implies that the normal exercise-induced increase in coronary flow is lost in atherosclerosis. Similar exaggerated vascular contractile responses are seen in hyperlipidemia, even before anatomic evidence of atherosclerosis develops. Because of this, coronary vasodilators (nitrates and/or Ca2+ channel blockers) are an important part of the therapeutic program in the majority of patients with ischemic heart disease. Ca Channel Blockers and Nitrates In severe exertional or vasospastic angina, the combination of a nitrate and a Ca2+ channel blocker may provide additional relief over that obtained with either type of agent alone. Since nitrates primarily reduce preload, whereas Ca2+ channel blockers reduce afterload, the net effect on reduction of oxygen demand should be additive. However, excessive vasodilation and hypotension can occur. The concurrent administration of a nitrate and nifedipine has been advocated in particular for patients with exertional angina with heart failure, the sick-sinus syndrome, or AV nodal conduction disturbances, but excessive tachycardia may be seen. Ca2+ Channel Blockers, -Adrenergic Receptor Antagonists, and Nitrates In patients with exertional angina that is not controlled by the administration of two types of antianginal agents, the use of all three may provide improvement, although the incidence of side effects increases significantly (Tolins et al., 1984; Asirvatham et al., 1998). The dihydropyridines and nitrates dilate epicardial coronary arteries; the dihydropyridines decrease afterload; the nitrates decrease preload; and the -adrenergic receptor antagonists decrease heart rate and myocardial contractility. Therefore, there is theoretical, and sometimes real, benefit with their combination. Combining verapamil or diltiazem with a -adrenergic receptor antagonist greatly increases the risk of conduction system and left ventricular dysfunction-related side effects, and should be undertaken only with extreme caution and only if no other alternatives exist. Antiplatelet and Antithrombotic Agents Unlike other antianginal agents, aspirin clearly has been demonstrated to reduce mortality in patients with unstable angina, reducing the incidence of myocardial infarction and death. In
2+

addition, low doses of aspirin appear to reduce the incidence of myocardial infarction in patients with chronic stable angina, and aspirin, given in doses of 160 to 325 mg at the onset of treatment of myocardial infarction, clearly reduces mortality, presumably by inhibiting the increased platelet aggregation that accompanies thrombolytic therapy (Kerins and FitzGerald, 1991). Heparin, in its unfractionated form and as low-molecular-weight heparin, also has been shown to reduce angina and prevent infarction in unstable angina. More direct thrombin inhibitors, such as hirudin, which directly inhibit even clot-bound thrombin, are not affected by circulating inhibitors, and function independently of antithrombin III, are being investigated. A metaanalysis of the GUSTO-IIB, TIMI9B, OASIS-1, and OASIS-2 trials showed that in patients with acute myocardial ischemia without ST segment elevation there was a modest reduction in the risk of death or myocardial infarction at 35 days when hirudin was compared with unfractionated heparin, but an increase in bleeding, with the majority of the benefit occurring in patients not receiving thrombolytic agents (OASIS-2 Investigators, 1999). Further studies seem warranted. Thrombolytic agents, on the other hand, have been shown to be of no benefit in unstable angina (Anonymous, 1992). Intravenous inhibitors of the platelet glycoprotein IIb/IIIa receptor (abciximab, tirofiban, and eptifibatide) have proven to be effective in preventing the complications of percutaneous coronary interventions and in the treatment of patients presenting with acute coronary syndromes (Bhatt and Topol, 2000). In contrast, orally active platelet glycoprotein IIb/IIIa antagonists have not proven to be effective in the chronic treatment of patients with ischemic heart disease and may result in worse outcomes. It is possible that there are subsets of patients with unstable angina who have different response profiles. Prospectus It is anticipated that new therapeutic agents in ischemic heart disease will fall into two categories. The first will include agents that modify cellular actions via cell surface or intracellular receptors but have no effect on gene expression. The second will include agents that either permanently or transiently alter gene expression, either by enhancing or inhibiting the production of a normal cell product, or by rendering the cell capable of producing an entirely new product. This gene-based therapy (see Chapter 5: Gene Therapy) will assume increasing importance in the treatment of ischemic heart disease. Emerging agents in the nitrate category include molsidomine, a nitrate-like agent that appears to produce vascular smooth muscle relaxation utilizing mechanisms similar to those of the nitrates themselves. A number of Ca2+ channel antagonists are under development in the classes of dihydropyridines (nitrendipine), phenylalkylamines (gallopamil, a verapamil derivative), and piperazines (flunarizine, which is marketed in some countries outside the United States, trimetazidine, and ranolazine). Whereas the first two agents share the general pharmacological characteristics of their classes, the piperazines exhibit some cytoprotective effects on myocardial energy metabolism and, in early studies, appear to exert an antianginal effect in the absence of significant hemodynamic effects. Although the first agent to block the T-type Ca2+ channel was withdrawn due to adverse drug interactions, the antianginal efficacy of this agent supports the future development of other members of this class. K+ channel activators, such as cromakalim, pinacidil, and nicorandil, have been proposed for use as direct coronary vasodilators in the treatment of both vasospastic and chronic stable angina (Hamilton and Weston, 1989; Why and Richardson, 1993; Lablanche et al., 1993). Studies of nicorandil (Kukovetz et al., 1992) demonstrate that its relaxant effect on coronary arterioles is + + inhibited by the K channel blocker, glyburide, and thus is likely due to K channel activation, with attendant cellular hyperpolarization of vascular smooth muscle. However, this agent also exerts a

nitrate-like effect, stimulating guanylyl cyclase to increase cyclic GMP, primarily in epicardial coronary arteries, including stenotic segments. The relative importance of these separate effects in human beings is not known. In studies of relatively small sample size, nicorandil had antianginal efficacy similar to that of nitrates; the -adrenergic receptor antagonists atenolol, metoprolol, and propranolol; and the Ca 2+ channel blockers amlodipine, diltiazem, and nifedipine (Markham et al., 2000). Drugs that have their primary site of vasodilation at the arteriolar level generally are not of benefit in the treatment of angina (Anonymous, 1992). The efficacy of nicorandil may be primarily due to its nitrate-like effect. Angiotensin converting enzyme (ACE) inhibitors (see Chapters 31, 33, and 34) are useful not only in the treatment of hypertension but also in reducing morbidity and mortality in symptomatic and asymptomatic congestive heart failure (Cohn et al., 1986; CONSENSUS Trial Study Group, 1987; Pitt, 1994). It also has been suggested that these agents may have a beneficial effect in angina pectoris over and above their effects on blood pressure, but it has been difficult to demonstrate this effect in normotensive patients. In addition to the ability of these agents to reduce blood pressure, which would be expected to have a favorable effect on ventricular wall stress, they also may reduce the coronary vascular response to angiotensin II and may prevent deleterious ventricular remodeling. Lisinopril has been suggested to reduce mortality when used in the treatment of acute infarction in the GISSI-III trial. However, if coronary perfusion pressure is lowered, ACE inhibitors have the potential to have a deleterious effect on angina; this may account for the patients who have become worse with these drugs in some trials (Vogt et al., 1993). In the HOPE trial, the ACE inhibitor rampiril resulted in a reduction in myocardial infarction and death from cardiovascular causes in patients with vascular disease or with diabetes plus one other cardiovascular risk factor, but without conventional indications for the use of an ACE inhibitor (Yusuf et al., 2000). Our appreciation of the role of inflammation in the development and progression of atherosclerosis and of unstable coronary syndromes is evolving (Libby, 2000; Ridker et al., 2000). Clinical studies are in progress that will address the ability of antibiotics, antiinflammatory agents, and HMG-CoA reductase inhibitors to retard the atherosclerotic process and reduce cardiac events. There has been great interest in evaluating the therapeutic utility of angiogenic factors, stem cells, and endothelial progenitor cells to facilitate growth of new vessels in areas of ischemia (Asahara et al., 2000; Isner, 2000; Kalka et al., 2000). Promising results have been achieved using vascular endothelial growth factor (VEGF) and related agents in human patients (Schumacher et al., 1998; Hendel et al., 2000). Several problems in the treatment of myocardial ischemia lend themselves to the use of gene-based therapies. Because coronary angioplasty is so commonly followed by restenosis (30% to 40% of cases) and because conventional pharmacological therapy has not abolished this problem, the concept of altering the biology of the cellular response to angioplasty has emerged. Techniques for delivering vectors to localized segments of peripheral and coronary vessels and vein grafts have been developed, and functional genes have been transferred (Chapman et al., 1992; Lemarchand et al., 1993; Nabel, 1995; Duckers and Nabel, 2000; O'Blenes et al., 2000). In animal models, the tendency for restenosis has been significantly inhibited (Ohno et al., 1994), and clinical trials are planned. Building on the success with angiogenic factors mentioned above, studies with animals have demonstrated that transferring the gene coding for VEGF also is efficacious (Isner, 2000), and trials in human beings are under way. In an attack on one mechanism underlying the development of atherosclerosis, it recently has been shown that retroviral therapy restores apo-E levels in apo-E deficient mice and reduces the extent of atherosclerosis (Hasty et al., 1999). Alternatively, disruption of the 12/15-lipoxygenase gene also was protective in this mouse model, consistent with a role for inflammation in the atherosclerotic process (Cyrus et al., 1999). These areas of rapid

progress suggest that therapy with genotypic strategies will play a significant role in the future. For further discussion of acute myocardial infarction and ischemic heart disease see Chapters 228 and 226 in Harrison's Principles of Internal Medicine, 16th ed., McGraw-Hill, New York, 2005.

Chapter 33. Antihypertensive Agents and the Drug Therapy of Hypertension

Overview Arterial pressure is the product of cardiac output and peripheral vascular resistance. Drugs lower pressure by actions on either the peripheral resistance or the cardiac output or both. The cardiac output may be reduced by drugs that either inhibit myocardial contractility or decrease ventricular filling pressure. Many of the antihypertensive drugs that affect adrenergic receptors, the reninangiotensin system, Ca2+channels, and Na+ and water balance are also discussed in Chapters 9: Agents Acting at the Neuromuscular Junction and Autonomic Ganglia, 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists, 29: Diuretics, 31: Renin and Angiotensin, 32: Drugs Used for the Treatment of Myocardial Ischemia, and 34: Pharmacological Treatment of Heart Failure. The pharmacology of antihypertensive agents that are not discussed elsewhere is presented here; in addition, the properties of all of the major drugs that are particularly relevant to their use in hypertension are reviewed, and an overview of the therapy of hypertension is provided. Antihypertensive Agents and the Drug Therapy of Hypertension: Introduction Hypertension is the most common cardiovascular disease. As many as 43 million adults in the United States have systolic and/or diastolic blood pressure above 140/90. Elevated arterial pressure causes pathological changes in the vasculature and hypertrophy of the left ventricle. As a consequence, hypertension is the principal cause of stroke, leads to disease of the coronary arteries with myocardial infarction and sudden cardiac death, and is a major contributor to cardiac failure, renal insufficiency, and dissecting aneurysm of the aorta. Hypertension is defined conventionally as blood pressure 140/90; this serves to characterize a group of patients who carry a risk of hypertension-related cardiovascular disease that is high enough to merit medical attention. However, from the standpoint of health promotion, it should be noted that the risk of both fatal and nonfatal cardiovascular disease in adults is lowest with systolic blood pressures of less than 120 mm Hg and diastolic of less than 80 mm Hg; these risks increase progressively with higher levels of both systolic and diastolic blood pressure. Although many of the clinical trials classify the severity of hypertension by diastolic pressure, progressive elevations of systolic pressure are similarly predictive of adverse cardiovascular events; at every level of diastolic pressure, risks are greater with higher levels of systolic blood pressure. Indeed, in elderly patients, systolic blood pressure predicts outcome better than diastolic blood pressure. At very severe levels of hypertension (systolic 210 and/or diastolic 120), a subset of patients develops fulminant arteriolopathy characterized by endothelial injury and a marked proliferation of cells in the intima, leading to intimal thickening and ultimately to arteriolar occlusion. This is the pathological basis of the syndrome of malignant hypertension, which is associated with rapidly progressive microvascular occlusive disease in the kidney (with renal failure), brain (hypertensive encephalopathy), retina (hemorrhages, exudates, and discedema), and other organs. The severe

endothelial disruption can lead to microangiopathic hemolytic anemia. Untreated malignant hypertension is rapidly fatal and requires in-hospital management on an emergency basis. The presence of certain target organ changes confers on a patient a worse prognosis than that for a patient with the same level of blood pressure lacking these findings. Thus, retinal hemorrhages, exudates, and discedema indicate a far worse short-term prognosis for a given level of blood pressure. Left ventricular hypertrophy defined by electrocardiogram, or more accurately by echocardiography, is associated with a substantially worse long-term outcome that includes a higher risk of sudden cardiac death. The risk of cardiovascular disease, disability, and death in hypertensive patients also is increased markedly by concomitant cigarette smoking and by elevated low-density lipoprotein; the coexistence of hypertension with these risk factors increases cardiovascular morbidity and mortality to an extent that is supraadditive. Robust evidence from multiple controlled trials indicates that pharmacological treatment of patients with diastolic pressures of 95 mm Hg or greater will reduce morbidity, disability, and mortality from cardiovascular disease. Effective antihypertensive therapy will almost completely prevent the hemorrhagic strokes, cardiac failure, and renal insufficiency due to hypertension. There is a marked reduction in total strokes. Moreover, several recent clinical trials suggest that reduction of diastolic blood pressure to 85 mm Hg confers a greater therapeutic benefit than reduction to 90 mm Hg, particularly in patients with diabetes (Hansson et al., 1998). The usual approach to patients with diastolic blood pressure in the range of 85 to 94 mm Hg is to use nonpharmacological therapy as an initial strategy. Because blood pressures in this range predict a clear increase in cardiovascular risk, the recommendations for nonpharmacological therapy should be accompanied by careful observation; in addition to the value of regular follow-up visits to maintain surveillance of blood pressure, they also afford an opportunity to assist and support patients in their efforts to achieve the changes in lifestyle required for effective nonpharmacological reduction in blood pressure. Antihypertensive drugs can be classified according to their sites or mechanisms of action (see Table 331). As arterial pressure is the product of cardiac output and peripheral vascular resistance, it can be lowered by actions of drugs on either the peripheral resistance or the cardiac output, or both. Drugs may reduce the cardiac output by either inhibiting myocardial contractility or decreasing ventricular filling pressure. Reduction in ventricular filling pressure may be achieved by actions on the venous tone or on blood volume via renal effects. Drugs can reduce peripheral resistance by acting on smooth muscle to cause relaxation of resistance vessels or by interfering with the activity of systems that produce constriction of resistance vessels (e.g., the sympathetic nervous system). The hemodynamic consequences of long-term treatment with antihypertensive agents are presented in Table 332, which also provides a framework for potential complementary effects of concurrent therapy with two or more drugs. The simultaneous use of drugs with similar mechanisms of action and hemodynamic effects often produces little additional benefit. However, concurrent use of drugs from different classes is a strategy for achieving effective control of blood pressure while minimizing dose-related adverse effects. Diuretics One of the earliest strategies for the management of hypertension was to alter Na+ balance by restriction of salt in the diet. Pharmacological alteration of Na+ balance became practical in the 1950s with the development of the orally active thiazide diuretics (see Chapter 29: Diuretics). These and related diuretic agents have antihypertensive effects when used alone, and they enhance the

efficacy of virtually all other antihypertensive drugs. The exact mechanism for reduction of arterial blood pressure by diuretics is not certain. Initially, the drugs decrease extracellular volume and cardiac output. However, the hypotensive effect is maintained during long-term therapy because of reduced vascular resistance; cardiac output returns to pretreatment values and extracellular volume remains somewhat reduced. Because of the persistent reduction in vascular resistance, some investigators have postulated that the diuretics have a direct effect on vascular smooth muscle that is independent of their saluretic effect. However, substantial data indicate that this is not the case. Thus, anephric patients and nephrectomized animals do not show a reduction in blood pressure when given diuretics (Bennett et al., 1977); a high salt intake or an infusion of saline (but not dextran) to counteract the net negative Na+ balance produced by diuretics reverses the antihypertensive effect; during effective therapy, plasma volume remains about 5% below pretreatment values and the plasma renin activity remains elevated, confirming a persistent reduction in body Na+ (Shah et al., 1978); diuretics do not relax vascular smooth muscle in vitro; and the hemodynamic effects of the diuretics to reduce vascular resistance are reproduced by restriction of salt (Freis, 1983). Potential mechanisms for reduction of vascular resistance by a persistent, albeit small, reduction in body Na+ include a decrease in interstitial fluid volume; a fall in smooth muscle Na+ concentration that may secondarily reduce the intracellular Ca2+ concentration, such that the cells are more resistant to contractile stimuli; and a change in the affinity and response of cell surface receptors to vasoconstrictor hormones (Insel and Motulsky, 1984). Benzothiadiazines and Related Compounds Benzothiadiazines ("thiazides") and related diuretics make up the most frequently used class of antihypertensive agents in the United States. Following the discovery of chlorothiazide, the first benzothiadiazine, a number of oral diuretics were developed that have an aryl-sulfonamide structure and block the Na+-Cl symporter. Some of these are not benzothiadiazines, but because they have structural features and molecular functions that are similar to the original benzothiadiazine compounds, they have been designated as members of the "thiazide class" of diuretics. For example, chlorthalidone, one of the nonbenzothiadiazines in the thiazide class of diuretics, is widely used in the treatment of hypertension. Because the thiazide class of drugs has the same pharmacological effects, they are generally interchangeable with appropriate adjustment of dosage (see Chapter 29: Diuretics). Regimen for Administration of the Thiazide-Class Diuretics in Hypertension When a thiazide-class diuretic is utilized as the sole antihypertensive drug (monotherapy), it should be administered in a low dose. Further, there is mounting evidence that the administration of these diuretics in the long-term treatment of hypertension should be in conjunction with a K +-sparing agent. Antihypertensive effects can be achieved in many patients with as little as 12.5 mg of chlorthalidone (HYGROTON) or hydrochlorothiazide (HYDRODIURIL) daily. This should be the initial dose in most elderly patients who do not require urgent reduction of pressure. Furthermore, when used as monotherapy, the maximal daily dose of thiazide-class diuretics usually should not exceed 25 mg of hydrochlorothiazide or chlorthalidone (or equivalent). Even though more diuresis can be achieved with higher doses of these diuretics, abundant evidence indicates that doses higher than this are not required for monotherapy of hypertension and probably are not as safe.

A large study comparing 25 and 50 mg hydrochlorothiazide daily in an elderly population did not show a greater decrease in blood pressure with the larger dose (MRC Working Party, 1987). In the randomized, controlled trials of antihypertensive therapy in the elderly (SHEP Cooperative Research Group, 1991; Dahlf et al., 1991; MRC Working Party, 1992) that demonstrate the best outcomes in cardiovascular morbidity and mortality, 25 mg of hydrochlorothiazide or chlorthalidone was the maximum dose given; if greater reduction of blood pressure than achieved with this dose was required, treatment with a second drug was initiated. With respect to safety, a case-control study (Siscovick et al., 1994) found a dose-dependent increase in the occurrence of sudden death at doses of hydrochlorothiazide greater than 25 mg daily. This finding supports the hypothesis engendered by a retrospective analysis of the Multiple Risk Factor Intervention Trial (Multiple Risk Factor Intervention Trial Research Group, 1982), suggesting that increased cardiovascular mortality is associated with higher diuretic doses. Further, the drug-specific metabolic effects of the thiazide-class diuretics, as well as the side effects perceived by patients, are to some degree dose-dependent, providing additional reasons not to administer more than the 25-mg dose of hydrochlorothiazide/chlorthalidone required to achieve nearly maximum blood pressure reduction. Taken together, clinical studies to date indicate that, if adequate blood pressure reduction is not achieved with the 25-mg daily dose of hydrochlorothiazide or chlorthalidone, a second drug should be added rather than increasing the dose of diuretic. When used in the treatment of hypertension, thiazide-class diuretics usually should be administered in conjunction with a K+-sparing agent. Attenuation of the kaliuretic effect of the thiazide-class diuretics can be achieved by drugs that block the Na+ channels in the late distal tubule and collecting duct (amiloride and triamterene), or by inhibition of aldosterone action (spironolactone). + + Oral K supplementation in the usual doses is not as effective as these K -sparing agents. At this time, the data from clinical trials most strongly support the use of amiloride in combination with a thiazide-class diuretic. In the two large clinical trials that demonstrate the best results in terms of cardiovascular morbidity and mortality (Dahlf et al., 1991; MRC Working Party, 1992), amiloride was the drug employed together with hydrochlorothiazide, used in a ratio of 1 mg amiloride/10 mg hydrochlorothiazide. Angiotensin converting enzyme inhibitors and angiotensin receptor antagonists will attenuate diuretic-induced loss of potassium to some degree, and this is a consideration if a second drug is required to achieve further blood pressure reduction beyond that attained with the diuretic alone. Because the diuretic and hypotensive effects of these drugs are greatly enhanced when they are given in combination, care should be taken to initiate combination therapy with low doses of each of these drugs. Administration of angiotensin converting enzyme inhibitors or angiotensin receptor antagonists together with other K+-sparing agents or with K+ supplements requires caution; combination of K+-sparing agents with each other or with K + supplementation can cause serious hyperkalemia in occasional patients. In contrast with the limitation on dose of thiazide-class diuretics used as monotherapy, the treatment of severe hypertension that is unresponsive to three or more drugs may require larger doses of the thiazide-class diuretics. Indeed, hypertensive patients may become refractory to drugs that block the sympathetic nervous system or to vasodilator drugs, because these drugs engender a state in which the blood pressure is very volume-dependent. Therefore, it is appropriate to consider the use of thiazide-class diuretics in doses of 50 mg of daily hydrochlorothiazide equivalent when treatment with appropriate combinations and doses of three or more drugs fails to yield adequate control of the blood pressure. Dietary Na+ restriction is a valuable adjunct to the management of such refractory patients and will minimize the dose of diuretic that is required. This can be achieved by a modest restriction of Na+ intake to 2 g daily. A more strict Na + restriction is not feasible for most patients. Since the degree of K+ loss relates to the amount of Na+ delivered to the distal tubule, such

restriction of Na+ can minimize the development of hypokalemia and alkalosis. The effectiveness of the thiazide class of drugs as diuretic or antihypertensive agents is progressively diminished when the glomerular filtration rate falls below 30 ml/min. One exception is metolazone, which retains efficacy in patients with this degree of renal insufficiency. Most patients will respond to the thiazide class of diuretics with a reduction in blood pressure within 2 to 4 weeks, although a minority will not achieve maximum reduction in arterial pressure for up to 12 weeks on a given dose. Therefore, doses should not be increased more often than every 2 to 4 weeks. Although the blood pressure of patients who have suppressed plasma renin activity is almost uniformly sensitive to diuretics in the thiazide class, most other patients also respond. There is no way to predict the antihypertensive response from the duration or severity of the hypertension in a given patient, although diuretics are unlikely to be effective as a sole therapy in patients with severe hypertension. Since the effect of the thiazide class of diuretics is additive with that of other antihypertensive drugs, combination regimens that include these diuretics are common and rational. Diuretics also have the advantage of minimizing the retention of salt and water that is commonly caused by vasodilators and some sympatholytic drugs. Adverse Effects and Precautions The adverse effects of diuretics are discussed in Chapter 29: Diuretics. Some of these are effects that determine whether patients can tolerate and comply with diuretic treatment. Sexual impotence is the most common troublesome side effect of the thiazide-class diuretics, and physicians should inquire specifically regarding its occurrence in conjunction with treatment with these drugs. Gout may be a consequence of the hyperuricemia induced by these diuretics. The occurrence of either of these adverse effects is a reason for considering alternative approaches to therapy. Muscle cramps also are related to diuretic therapy in a dose-dependent manner. Other effects of the thiazide-class diuretics are laboratory observations that are of concern primarily because they are putative surrogate markers for adverse drug effects on morbidity and mortality. Surrogate markers are factors that, in epidemiological studies, have been found to correlate with disease outcomes and therefore are used in investigations of drug therapy as a surrogate of the actual effect of a drug on disease outcome. Thus, reduction in systolic blood pressure is an example of a surrogate marker for the reduction of stroke by antihypertensive therapy that has been extensively validated (SHEP Cooperative Research Group, 1991). In contrast, the epidemiological evidence linking frequent ventricular ectopic depolarizations to sudden cardiac death had suggested that these ventricular arrhythmias might be surrogate markers for a beneficial effect of antiarrhythmic drugs on sudden cardiac death. However, reduction of ventricular ectopic depolarizations and nonsustained ventricular tachycardia by the antiarrhythmic drugs encainide and flecainide was associated with an increase in sudden cardiac death, demonstrating the fallibility of using surrogate markers to predict the outcome of a specific pharmacological intervention (CAST Investigators, 1989). Accordingly, the effect of drugs on surrogate markers cannot be considered as convincing evidence; rather, surrogate markers form the basis for hypotheses that require testing with controlled clinical trials. Until such trials are carried out, however, the effect of drugs on surrogate markers that predict adverse outcomes do cause concern and influence decisions regarding dose. The effects of diuretic drugs on several surrogate markers for adverse outcomes merit consideration. + The K depletion produced by thiazide-class diuretics is dose-dependent over a wide range of doses + and is variable among individuals, such that a subset of patients may become substantially K + depleted on diuretic drugs. Even small doses given chronically, however, lead to some K depletion.

There are two types of ventricular arrhythmia that are thought to be enhanced by K+ depletion. One of these is polymorphic ventricular tachycardia (torsades de pointes), which is induced by a number of drugs, including quinidine. Such drug-induced polymorphic ventricular tachycardia is an arrhythmia initiated by abnormal ventricular repolarization, and it is markedly enhanced by drugs that produce K+ depletion, which elicits abnormal repolarization (see Chapter 35: Antiarrhythmic Drugs). Accordingly, thiazide diuretics should not be given together with drugs that can cause polymorphic ventricular tachycardia. The most important concern regarding K+ depletion is its influence on ischemic ventricular fibrillation, the leading cause of sudden cardiac death and a major contributor to cardiovascular mortality in treated hypertensive patients. Studies in experimental animals have demonstrated that K+ depletion lowers the threshold for electrically induced ventricular fibrillation in the ischemic myocardium and also increases spontaneous ischemic ventricular fibrillation (Curtis and Hearse, 1989; Yano et al., 1989). A case-control study of hypertensive patients has found a positive correlation between diuretic dose and sudden cardiac death and an inverse correlation between the use of adjunctive K+-sparing agents and sudden cardiac death (Siscovick et al., 1994). One controlled clinical trial has demonstrated a significantly greater occurrence of sudden cardiac death in patients treated with 50 mg of hydrochlorothiazide daily in comparison with the -adrenergic antagonist metoprolol (MRC Working Party, 1992). Controlled clinical trials comparing the effect of diuretics with other antihypertensive drugs on cardiovascular mortality are now under way and should further illuminate this question. In the meantime, the available data on sudden cardiac death support the limitation of diuretic dose as monotherapy to 25 mg of hydrochlorothiazide daily (or + equivalent) and the use of an adjunctive K -sparing agent. The thiazide class of diuretics elevates the levels of low-density lipoprotein (LDL) and increases the ratio of LDL/high-density lipoprotein (HDL). A linkage between increased LDL and increased coronary heart disease has been demonstrated in epidemiological studies and in investigations of lipid-lowering drugs, justifying, the consideration that LDL is a surrogate marker for coronary heart disease morbidity and mortality. Whether the increase in the LDL/HDL ratio caused by diuretics is in fact predictive of increased coronary heart disease in patients on diuretics is not presently known, and until adequate morbidity and mortality trials comparing antihypertensive drugs are completed, the effect of the thiazide-class diuretics on plasma lipids can be a basis for concern but not for definitive recommendations regarding the choice of these drugs relative to other antihypertensive agents. In epidemiological studies, left ventricular hypertrophy is a powerful predictor of an increase in cardiac death in hypertensive patients. The thiazide-class diuretics are less effective in reducing left ventricular hypertrophy than are other antihypertensive drugs such as the angiotensin converting enzyme inhibitors (Dahlf et al., 1992). The thiazide-class diuretics increase glycosylated hemoglobin in patients with diabetes mellitus (Gall et al., 1992). This finding, taken together with evidence that angiotensin converting enzyme inhibitors delay the deterioration of renal function in diabetic patients, suggests that thiazide-class diuretics are not the first drugs of choice in the monotherapy of hypertensive patients with diabetes mellitus. All of the thiazide-like drugs cross the placenta, but they have not been found to have direct adverse effects on the fetus. However, if administration of a thiazide is begun during pregnancy, there is a risk of transient volume depletion that may result in placental hypoperfusion. Since the thiazides appear in breast milk, they should be avoided by nursing mothers.

The Choice of a Thiazide-Type Diuretic as the Initial Drug in the Treatment of Hypertension Few issues in hypertension are more controversial than whether or not patients should be placed on a diuretic as the initial or only drug for the treatment of hypertension (Joint National Committee, 1997; Tobian et al., 1994). A definitive answer to this question awaits the completion of a large clinical trial (the ALLHAT trial) conducted by the National Institutes of Health comparing a thiazide-class diuretic with other antihypertensive drugs as the initial or only therapeutic agents. In the interim, decisions will be made based on interpretations of the available data. There are several general types of data, from which interpretations have been inferred, but none provides convincing evidence. One set of data is the group of metabolic effects of the thiazide-class of diuretics discussed above: the increase in LDL, depletion of K+ , impairment of diabetes control, and a reduction of left ventricular hypertrophy that is less than achieved with other drugs. The case-control study demonstrating a dose-dependent linkage of these diuretics to sudden cardiac death strengthens the inference that K+ depletion is not benign (Siscovick et al., 1994). Another type of data is that obtained from the clinical trials that have established the benefit of antihypertensive therapy. In controlled clinical trials conducted before 1991 in predominantly middle-aged patients, the beneficial effect of antihypertensive therapy on coronary heart disease and cardiac death was found to be less than the effect on stroke. Long-term prospective observational studies predict that a decrease of 5 to 6 mm Hg in diastolic pressure should lead to a 35% to 40% reduction in stroke; indeed, an overview of 14 randomized trials of pharmacological therapy (Collins et al., 1990) reveals that the prevalence of fatal plus nonfatal stroke is lowered by 42% (p<0.0002) as a consequence of that change in diastolic pressure. Whereas the observational studies predict a 20% to 25% reduction in fatal plus nonfatal coronary heart disease with a difference of 5 to 6 mm Hg in diastolic pressure, reduction of pressure to this degree with antihypertensive treatment lowered total coronary heart disease by only 14% (p<0.01). Fatal stroke was lowered by 45% (p<0.0001), whereas fatal coronary heart disease was reduced by only 11% (not significant) during antihypertensive treatment. The failure of antihypertensive therapy to yield the expected reduction in coronary heart disease, and particularly coronary heart disease mortality, has led to considerable speculation, including the possibility that the diuretic drugs that were the primary agents used in these trials may impose adverse effects on coronary heart disease, including sudden cardiac death. These disappointing results in terms of coronary heart disease were obtained in trials on mostly middle-aged patients that predominantly employed hydrochlorothiazide or chlorthalidone + in doses up to 50 mg without a K -sparing agent. Subsequent to this meta-analysis of the 14 controlled trials, 2 trials in elderly patients that used a low dose of hydrochlorothiazide (25 mg daily) together with the K+ -sparing agent amiloride demonstrated more favorable trends in reducing coronary mortality, lowering it by 50% (Dahlf et al., 1991) and 40% (MRC Working Party, 1992). In the latter study, the reduction in stroke and coronary events was significant on diuretic therapy, whereas no significant reductions in these endpoints were seen in a comparison group randomized to the -adrenergic blocker atenolol. The following approach to selection of a diuretic as the initial drug is derived from the data above. In elderly patients (age 65 or greater), selection of a thiazide-class diuretic as the initial drug is rational if it is given in doses in the range of 12.5 to 25 mg of hydrochlorothiazide (or equivalent) + daily and together with a K -sparing agent. This is based on the high level of efficacy in the clinical trials with such regimens, the relatively low profile of side effects, and the fact that elderly patients are likely to have a good response to diuretic therapy.

In patients under age 65, a greater individualization of choice of the initial antihypertensive drug seems warranted. The earlier clinical trials (conducted in this age group with high doses of the thiazide-class drug given without K+-sparing agents) provide neither reassurance regarding the thiazide-class diuretics nor rejection of them. Individualization of initial drug choice can consider several factors. In diabetic patients, angiotensin converting enzyme inhibitors attenuate the decline in renal function, making these drugs preferable to diuretics, which may impair glucose tolerance in diabetes. Other drugs, particularly the angiotensin converting enzyme inhibitors, are more effective than diuretics in reducing left ventricular mass in patients with left ventricular hypertrophy. Caucasian race, age less than 65, and a high or normal renin status at baseline predict an antihypertensive response to an angiotensin converting enzyme inhibitor or a -adrenergic receptor antagonist that is superior to that with a diuretic. Other Diuretic Antihypertensive Agents The thiazide-type diuretics are more effective antihypertensive agents than are the loop diuretics, such as furosemide and bumetanide, in patients who have normal renal function (Ram et al., 1981). This differential effect is most likely related to the short duration of action of loop diuretics, such that a single daily dose does not cause a significant net loss of Na+ for an entire 24-hour period. The spectacular efficacy of the loop diuretics in producing a rapid and profound natriuresis is a potential detriment for the treatment of hypertension. When a loop diuretic is given twice daily, the acute diuresis can be excessive and lead to more side effects than occur with a slower-acting, milder thiazide diuretic. The loop diuretics produce hypercalciuria, rather than the hypocalciuria associated with the thiazides. However, other metabolic consequences of the thiazides are shared with the loop diuretics, including hypokalemia, hyperuricemia, glucose intolerance, and potentially adverse effects on plasma concentrations of lipids. Loop diuretics may be particularly useful in patients with azotemia and in patients with severe edema associated with a vasodilator such as minoxidil. Although spironolactone in doses up to 100 mg per day is equivalent to hydrochlorothiazide in its hypotensive effect (Jeunemaitre et al., 1988), higher doses produce an unacceptable incidence of side effects (Schrijver and Weinberger, 1979). Spironolactone may be particularly useful for individuals with clinically significant hyperuricemia, hypokalemia, or glucose intolerance, and it is the agent of choice for management of primary aldosteronism. In contrast to thiazide diuretics, spironolactone does not affect plasma concentrations of Ca2+ or glucose. The effects of spironolactone on plasma lipids have not been studied extensively, but data indicate that the changes in triglycerides, LDL cholesterol, and total cholesterol are less than those seen with the thiazides. However, spironolactone may decrease the concentration of HDL cholesterol (Falch and + Schreiner, 1983). The other K -sparing diuretics, triamterene and amiloride, are used primarily to reduce the kaliuresis and potentiate the hypotensive effect of a thiazide (DeCarvalho et al., 1980; Multicenter Diuretic Cooperative Study Group, 1981). These agents should be used cautiously with + frequent measurements of K concentrations in plasma in patients predisposed to hyperkalemia. Patients taking spironolactone, amiloride, or triamterene should be cautioned regarding the possibility that concurrent use of K+ -containing "salt substitutes" could produce hyperkalemia. Renal insufficiency is a relative contraindication to the use of K+-sparing diuretics. Diuretic-Associated Drug Interactions Since the antihypertensive effects of diuretics are frequently additive with those of other antihypertensive agents, a diuretic commonly is used in combination with other drugs. As discussed above, the concurrent administration of diuretics with quinidine and other drugs that cause polymorphic ventricular tachycardia greatly increases the risk of this drug-induced arrhythmia. The + 2+ K - and Mg -depleting effects of the thiazide-like and loop diuretics also can potentiate

arrhythmias that arise from digitalis toxicity. Corticosteroids can amplify the hypokalemia produced by the diuretics. All diuretics can decrease the clearance of Li+, resulting in increased plasma concentrations of Li+ and potential toxicity (Amdisen, 1982). Nonsteroidal antiinflammatory drugs (see Chapter 27: Analgesic-Antipyretic and Antiinflammatory Agents and Drugs Employed in the Treatment of Gout) that inhibit the synthesis of prostaglandins reduce the antihypertensive effects of diuretics. It is not known if this interaction is due to Na+ retention resulting from blockade of the natriuretic effect of the diuretic by the antiinflammatory agent or is related to inhibition of vascular synthesis of prostaglandins (Webster, 1985). Based on recent studies, it would appear that the effects of selective COX-2 inhibitors on renal prostaglandin synthesis and function are similar to those of the nonselective nonsteroidal antiinflammatory drugs. Nonsteroidal antiinflammatory drugs, -adrenergic receptor antagonists, and angiotensin converting enzyme inhibitors reduce plasma concentrations of aldosterone and can potentiate the hyperkalemic effects of a K+-sparing diuretic. Sympatholytic Agents Since the demonstration in 1940 that bilateral excision of the thoracic sympathetic chain could lower blood pressure, the search for effective chemical sympatholytic agents has been intensive. Many compounds were tolerated poorly because they produced symptomatic orthostatic hypotension, sexual dysfunction, diarrhea, and fluid retention, with subsequent reduction of the antihypertensive effect. However, newer agents and rational combinations of these drugs with diuretics and vasodilators have overcome many of these difficulties. The subgroups of sympatholytic agents are shown in Table 331. Methyldopa Methyldopa (ALDOMET) is a centrally acting antihypertensive agent. It is a prodrug that exerts its antihypertensive action via an active metabolite. Methyldopa ( -methyl-3,4-dihydroxy-L-phenylala-nine), an analog of 3,4-dihydroxyphenylalanine (DOPA), is metabolized by the L-aromatic amino acid decarboxylase in adrenergic neurons to methyldopamine, which then is converted to -methylnorepinephrine (Figure 331). Methylnorepinephrine is stored in the neurosecretory vesicles of adrenergic neurons, substituting for norepinephrine itself. Thus, when the adrenergic neuron discharges its neurotransmitter, methylnorepinephrine is released instead of norepinephrine. Figure 331. The Metabolism of Methyldopa in Adrenergic Neurons. Methylnorepinephrine Replaces Norepinephrine in Neurosecretory Vesicles.

Because -methylnorepinephrine is as potent as norepinephrine as a vasoconstrictor, its substitution for norepinephrine in peripheral adrenergic neurosecretory vesicles does not alter the

vasoconstrictor response to peripheral adrenergic neurotransmission. Rather, methylnorepinephrine acts in the brain to inhibit adrenergic neuronal outflow from the brainstem, and this central effect is principally responsible for its antihypertensive action. It is probable that methylnorepinephrine acts as an 2-adrenergic receptor agonist in the brainstem to attenuate the output of vasoconstrictor adrenergic signals to the peripheral sympathetic nervous system. A body of evidence supports the conclusion that methyldopa acts in the brain via an active metabolite to lower blood pressure (Bobik et al., 1988; Granata et al., 1986; Reid, 1986). In experimental animals, the hypotensive effect of methyldopa is blocked by DOPA decarboxylase inhibitors that have access to the brain, but not by inhibitors that are excluded from the central nervous system (CNS). The hypotensive effect also is abolished by inhibitors of dopamine hydroxylase and by centrally acting -adrenergic receptor antagonists. Small doses of methyldopa that do not lower blood pressure when injected systemically elicit a hypotensive effect when injected into the vertebral artery. Selective microinjection of -methylnorepinephrine into the C-1 area of the rostral ventrolateral medulla of the rat elicits a hypotensive response that is prevented by -adrenergic receptor blockade. It is presumed that methylnorepinephrine inhibits the neurons in this area that are responsible for maintaining tonic discharge of peripheral sympathetic nerves and also for transmission of baroreflex-initiated tone. The excess -adrenergic inhibition of sympathetic output may be a consequence of the accumulation of methylnorepinephrine in quantities larger than the norepinephrine that it displaces; this could result from the fact that the methylnorepinephrine is not a substrate for monoamine oxidase, the enzyme principally responsible for norepinephrine disposition in the brain. In addition to the -adrenergic receptormediated inhibition of sympathetic output by methylnorepinephrine in the C-1 area of the rostral ventrolateral medulla, it also may exert inhibitory effects at other sites such as the nucleus tractus solitarius. Pharmacological Effects Methyldopa reduces vascular resistance without causing much change in cardiac output or heart rate in younger patients with uncomplicated essential hypertension. In older patients, however, cardiac output may be decreased as a result of a reduction in heart rate and stroke volume; this is secondary to relaxation of veins and a reduction in preload. The fall in arterial pressure is maximal 6 to 8 hours after an oral or intravenous dose. Although the decrease in supine blood pressure is less than that in the upright position, symptomatic orthostatic hypotension is less common with methyldopa than with drugs that act exclusively on peripheral adrenergic neurons or autonomic ganglia; this is because methyldopa attenuates but does not completely block baroreceptor-mediated vasoconstriction. For this reason, it is well tolerated during surgical anesthesia. Any severe hypotension is reversible with volume expansion. Renal blood flow is maintained and renal function is unchanged during treatment with methyldopa. Plasma concentrations of norepinephrine fall in association with the reduction in arterial pressure, and this reflects the decrease in sympathetic tone. Renin secretion also is reduced by methyldopa, but this is not a major effect of the drug and is not necessary for its hypotensive effects. Salt and water often are gradually retained with prolonged use of methyldopa, and this tends to blunt the antihypertensive effect. This has been termed "pseudotolerance," and it can be overcome with concurrent use of a diuretic. Of interest, treatment with methyldopa may reverse left ventricular hypertrophy within 12 weeks without any apparent relationship to the degree of change of arterial pressure (Fouad et al., 1982). Absorption, Metabolism, and Excretion Since methyldopa is a prodrug that is metabolized in the brain to the active form, its concentration

in plasma has less relevance for its effects than is true for many other drugs. When administered orally, methyldopa is absorbed by an active amino acid transporter. Peak concentrations in plasma occur after 2 to 3 hours. The drug is distributed in a relatively small apparent volume (0.4 liter/kg) and is eliminated with a half-life of about 2 hours. The transport of methyldopa into the CNS is apparently also an active process (Bobik et al., 1986). Methyldopa is excreted in the urine primarily as the sulfate conjugate (50% to 70%) and as the parent drug (25%). The remaining fraction is excreted as other metabolites, including methyldopamine, methylnorepinephrine, and O-methylated products of these catecholamines (Campbell et al., 1985). The half-life of methyldopa is prolonged to 4 to 6 hours in patients with renal failure. In spite of its rapid absorption and short half-life, the peak effect of methyldopa is delayed for 6 to 8 hours, even after intravenous administration, and the duration of action of a single dose is usually about 24 hours; this permits once- or twice-daily dosing (Wright et al., 1982). The discrepancy between the effects of methyldopa and the measured concentrations of the drug in plasma is most likely related to the time required for transport into the CNS, conversion to the active metabolites, and accumulation of these metabolites in central adrenergic neurons. Patients with renal failure are more sensitive to the antihypertensive effect of methyldopa, but it is not known if this is due to alteration in excretion of the drug or to an increase in transport into the CNS. Adverse Effects and Precautions In addition to lowering blood pressure, the active metabolites of methyldopa act on 2-adrenergic receptors in the brainstem to inhibit the centers that are responsible for wakefulness and alertness. Thus, methyldopa produces sedation that is largely transient. A diminution in psychic energy may be a persistent effect in some patients, and depression occurs occasionally. Medullary centers that control salivation also are inhibited by -adrenergic receptors, and methyldopa may produce dryness of the mouth. Other side effects that are related to the pharmacological effects in the CNS include a reduction in libido, parkinsonian signs, and hyperprolactinemia that may become sufficiently pronounced to cause gynecomastia and galactorrhea. In individuals who have sinoatrial node dysfunction, methyldopa may precipitate severe bradycardia and sinus arrest, including that which occurs with carotid sinus hypersensitivity. Methyldopa also produces some adverse effects that are not related to its pharmacological action. Hepatotoxicity, sometimes associated with fever, is an uncommon but potentially serious toxic effect of methyldopa. Prompt diagnosis of hepatotoxicity requires a low threshold for considering the drug as a cause for hepatitis-like symptoms (e.g., nausea, anorexia) and screening for hepatotoxicity (e.g., with determination of gamma-glutamyl transpeptidase or alanine aminotransferase) at about 3 weeks and again at about 3 months following initiation of treatment with this drug. The incidence of methyldopa-induced hepatitis is unknown, but about 5% of patients will have transient increases in alanine aminotransferase activity in plasma. Hepatic dysfunction usually is reversible with prompt discontinuation of the drug, but it will recur if methyldopa is given again, and a few cases of fatal hepatic necrosis have been reported. Hepatitis may occur only after long-term therapy with methyldopa, but it usually appears within 3 months of starting the drug. It is advisable to avoid the use of methyldopa in patients with hepatic disease. Methyldopa can cause hemolytic anemia. At least 20% of patients who receive methyldopa for a year develop a positive Coombs test (antiglobulin test) that is due to autoantibodies directed against the Rh locus on the patients' erythrocytes. The development of a positive Coombs test per se, however, is not an indication to stop treatment with methyldopa; 1% to 5% of these patients will develop a hemolytic anemia which requires prompt discontinuation of the drug. The Coombs test may remain positive for as long as a year after discontinuation of methyldopa, but the hemolytic

anemia usually resolves within a matter of weeks. Severe hemolysis may be attenuated by treatment with glucocorticoids. Adverse effects that are even more rare include leukopenia, thrombocytopenia, red cell aplasia, lupus erythematosuslike syndrome, lichenoid and granulomatous skin eruptions, myocarditis, retroperitoneal fibrosis, pancreatitis, diarrhea, and malabsorption. Therapeutic Uses Methyldopa is an effective antihypertensive agent when given in conjunction with a diuretic. It is generally well tolerated by patients with ischemic heart disease and by those with diastolic dysfunction, in whom it reduces left ventricular mass. However, frequent side effects and the potential for immunological abnormalities and organ toxicity are such that it is not used as the initial drug in monotherapy but is reserved for patients in whom it may have special value. Methyldopa is the preferred drug for treatment of hypertension during pregnancy based on its effectiveness and safety for both mother and fetus. The usual initial dose of methyldopa is 250 mg twice daily, and there is little additional effect with doses above 2 g per day. Administration of a single daily dose of methyldopa at bedtime minimizes sedative effects, but administration twice daily may be required for some patients. A parenteral preparation of the ethyl ester of methyldopa, methyldopate hydrochloride (ALDOMET), also is available. It is usually given by intermittent intravenous infusion of 250 to 500 mg every 6 hours. The rate of deesterification of the methyldopate is variable among patients, and the doses given intravenously may deliver less methyldopa to the circulation than the same dose given orally. Clonidine, Guanabenz, and Guanfacine The detailed pharmacology of the 2-adrenergic agonists clonidine (CATAPRES), guanabenz (WYTENSIN ), and guanfacine (TENEX ), is discussed in Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists. These drugs stimulate the 2A subtype of 2-adrenergic receptors in the brainstem, resulting in a reduction in sympathetic outflow from the CNS (Sattler and van Zwieten, 1967; Langer et al., 1980; MacMillan et al., 1996). The decrease in plasma concentrations of norepinephrine is correlated directly with the hypotensive effect (Goldstein et al., 1985; Sorkin and Heel, 1986). Patients who have had a spinal cord transection above the level of the sympathetic outflow tracts do not display a hypotensive response to clonidine (Reid et al., 1977). At doses higher than those required to stimulate central 2Aadrenergic receptors, these drugs can activate the 2B subtype of 2-adrenergic receptors on vascular smooth muscle cells (Link et al., 1996; MacMillan et al., 1996). This effect accounts for the initial vasoconstriction that is seen when overdoses of these drugs are taken, and it has been postulated to be responsible for the loss of therapeutic effect that is observed with high doses (Frisk-Holmberg et al., 1984; Frisk-Holmberg and Wibell, 1986). Pharmacological Effects The 2-adrenergic agonists lower arterial pressure by an effect on both cardiac output and peripheral resistance. In the supine position, when the sympathetic tone to the vasculature is low, the major effect is to reduce both heart rate and stroke volume; however, in the upright position, when sympathetic outflow to the vasculature is normally increased, these drugs reduce vascular resistance. Some degree of orthostatic hypotension always occurs because of a reduction in venous return (secondary to systemic venodilatation), but symptomatic postural hypotension is uncommon in the absence of volume depletion. Sympathetic reflexes are damped but not entirely inhibited, and the sympathetic responses that are associated with the use of arteriolar vasodilators such as

hydralazine and minoxidil are blunted. However, the 2-adrenergic agonists do not interfere with the hemodynamic response to exercise, and exercise-induced hypotension is unusual. The decrease in cardiac sympathetic tone leads to a reduction in myocardial contractility and heart rate. Renal blood flow and glomerular filtration rate are maintained. Secretion of renin often is reduced, although it will respond to volume depletion or maintenance of an upright posture; there is no correlation between the hypotensive response and the effect on plasma renin activity. Retention of salt and water may occur with the 2-adrenergic agonists, and it may be necessary to use a diuretic concurrently. Centrally acting 2-adrenergic agonists have either no effect on plasma lipids or produce a slight reduction of total cholesterol, LDL cholesterol, and triglycerides (Lardinois and Neuman, 1988). When guanabenz was first introduced, there was considerable interest in observations that the drug could be natriuretic in experimental animals. However, studies in human subjects have given variable results. With long-term therapy, there is usually a small loss of weight with no clinically significant changes in salt and water balance, suggesting that the "pseudotolerance" (Na+ retention) seen with methyldopa and guanadrel may not occur with guanabenz. Nonetheless, the antihypertensive effects of diuretics and guanabenz are additive. If individuals are given guanabenz after a salt load, the drug has a natriuretic effect, and a new steady-state of Na+ balance is attained by 1 week. This short-term effect is thought to be related to a reduction in renal sympathetic stimulation, with a consequent reduction in Na + reabsorption in the proximal nephron (Gehr et al., 1986). Guanabenz also has been shown to cause a water diuresis in some situations, which may be due to inhibition of the release and the renal actions of vasopressin (Strandhoy, 1985). Stimulation of renal 2-adrenergic receptors by guanabenz may inhibit vasopressin-induced accumulation of cyclic AMP (Gellai and Edwards, 1988). Adverse Effects and Precautions Although the 2-adrenergic agonists rarely cause life-threatening adverse reactions, many patients experience annoying and sometimes intolerable side effects. Sedation and xerostomia occur in at least 50% of patients upon initiation of therapy with clonidine and guanabenz and in 25% of patients who receive guanfacine (Wilson et al., 1986). Although these symptoms may diminish after several weeks of therapy, at least 10% of patients discontinue the drug because of persistence of these effects or because of impotence, nausea, or dizziness. The xerostomia may be accompanied by dry nasal mucosa, dry eyes, and parotid gland swelling and pain. Clonidine may produce a lower incidence of dry mouth and sedation when given transdermally, perhaps because high peak concentrations are avoided. Less common CNS side effects include sleep disturbances with vivid dreams or nightmares, restlessness, and depression. Cardiac effects related to the sympatholytic action of these drugs include symptomatic bradycardia and sinus arrest in patients with dysfunction of the sinoatrial node and atrioventricular (AV) block in patients with AV nodal disease or in patients taking other drugs that depress the AV node. Some 15% to 20% of patients who receive transdermal clonidine may develop contact dermatitis. Sudden discontinuation of clonidine and related 2-adrenergic agonists may cause a withdrawal syndrome consisting of headache, apprehension, tremors, abdominal pain, sweating, and tachycardia. The arterial blood pressure may rise to levels above those that were present prior to treatment, but the syndrome may occur in the absence of an overshoot in pressure. Symptoms typically occur 18 to 36 hours after the drug is stopped, and they are associated with increased sympathetic discharge, as evidenced by elevated plasma and urine concentrations of catecholamines. The exact incidence of the withdrawal syndrome is not known, but it is doserelated, occurring rarely in patients taking 0.3 mg or less daily of clonidine and more frequently and severely upon discontinuation of higher doses. It has been reported with all of the drugs of this

class, but it may be milder with guanfacine, perhaps because of this drug's longer half-life. Rebound hypertension also has been seen after discontinuation of transdermal administration of clonidine (Metz et al., 1987). Treatment of the withdrawal syndrome depends on the urgency of reducing the arterial blood pressure. In the absence of hypertensive encephalopathy, patients can be treated with their usual dose of antihypertensive drug, which should reduce the pressure within 2 hours. If a more rapid effect is required, sodium nitroprusside or a combination of an - and -adrenergic blocker is appropriate. -Adrenergic blocking agents should not be used alone in this setting, since they will accentuate the hypertension by allowing unopposed -adrenergic vasoconstriction caused by the elevated circulating concentrations of epinephrine. Because perioperative hypertension has been described in patients when clonidine was withdrawn the night before surgery, surgical patients who are being treated with an 2-adrenergic agonist either should be switched to another drug prior to elective surgery or should receive their morning dose and/or transdermal clonidine prior to the procedure. All patients who receive one of these drugs should be apprised of the potential danger of discontinuing the drug abruptly, and patients suspected of being noncompliant with medications should not be given 2-adrenergic agonists for hypertension. Adverse drug interactions with 2-adrenergic agonists are rare. Diuretics potentiate the hypotensive effect of these drugs in a predictable manner. Tricyclic antidepressants may inhibit the antihypertensive effect of clonidine, but the mechanism of this interaction is not known. Overdosage with an 2-adrenergic agonist causes depression of the sensorium and transient hypertension followed by hypotension, bradycardia, and respiratory depression. The depressed respiration (with miosis) resembles the effects of an opioid. Treatment consists of ventilatory support, atropine or a sympathomimetic for bradycardia, and circulatory support with expansion of the blood volume and dopamine or dobutamine if needed. Therapeutic Uses The 2-adrenergic agonists are usually used in conjunction with diuretics for the treatment of hypertension, but they may be effective when given alone; all of the drugs in this class are equally efficacious (Holmes et al., 1983). The CNS effects are such that this class of drugs is not a leading option for monotherapy of hypertension, nor is it the first choice for use together with a diuretic. These drugs also are effective in blunting the reflex increase in sympathetic activity produced by vasodilators, and they may be used instead of a -adrenergic antagonist for this purpose. Because the clonidine withdrawal syndrome occurs predominantly in patients taking higher doses of the drug, clonidine at doses of more than 0.3 mg daily is not optimal therapy for patients with severe hypertension. Clonidine also has been used in hypertensive patients for the diagnosis of pheochromocytoma. The lack of suppression of the plasma concentration of norepinephrine to less than 500 pg/ml 3 hours after an oral dose of 0.3 mg of clonidine suggests the presence of such a tumor. A modification of this test, wherein overnight urinary excretion of norepinephrine and epinephrine is measured after administration of a 0.3-mg dose of clonidine at bedtime, may be useful when results based on plasma norepinephrine concentrations are equivocal (MacDougall et al., 1988). Other uses for 2adrenergic agonists are discussed in Chapters 10, 14, and 24.

Guanadrel Guanadrel (HYLOREL) specifically inhibits the function of peripheral postganglionic adrenergic neurons. The structure of guanadrel, which contains the strongly basic guanidine group, is as follows:

Locus and Mechanism of Action Guanadrel is targeted uniquely to the peripheral adrenergic neuron, where it inhibits sympathetic function. The drug reaches its site of action by active transport into the neuron by the same transporter that is responsible for the reuptake of norepinephrine (see Chapter 6: Neurotransmission: The Autonomic and Somatic Motor Nervous Systems). In the neuron, guanadrel is concentrated within the neurosecretory vesicles, where it replaces norepinephrine. During chronic administration, guanadrel acts as a "substitute neurotransmitter," in that it is present in storage vesicles, it depletes the normal transmitter, and it can be released by stimuli that normally release norepinephrine. This replacement of norepinephrine with an inactive transmitter is probably the principal mechanism of its neuron-blocking action. When given intravenously, guanadrel initially can release norepinephrine in an amount sufficient to increase arterial blood pressure. This does not occur with oral administration, since norepinephrine is released only slowly from the vesicles under this circumstance and is degraded within the neuron by monoamine oxidase. Nonetheless, because of the potential for norepinephrine release, guanadrel is contraindicated in patients with pheochromocytoma. During adrenergic neuron blockade with guanadrel, effector cells become supersensitive to norepinephrine. The supersensitivity is similar to that produced by postganglionic sympathetic denervation. Pharmacological Effects Essentially all of the therapeutic and adverse effects of guanadrel result from sympathetic blockade. The antihypertensive effect is achieved by a reduction in peripheral vascular resistance that results from inhibition of sympathetically mediated vasoconstriction. Thus, the arterial pressure is reduced modestly in the supine position when sympathetic activity is normally low, but the pressure can fall to a greater extent during situations where reflex sympathetic activation is a mechanism for maintaining arterial pressure, such as assumption of the upright posture, exercise, and depletion of plasma volume. Renal blood flow and glomerular filtration rate are modestly decreased during therapy with guanadrel, but this is without clinical consequence; renin secretion is not reduced. Plasma volume often becomes expanded, which may diminish the antihypertensive efficacy of guanadrel and require administration of diuretic to restore the antihypertensive effect. Absorption, Distribution, Metabolism, and Excretion Guanadrel is rapidly absorbed, leading to maximal levels in plasma at 1 to 2 hours. Because guanadrel must be transported into and accumulate in adrenergic neurons, the maximum effect on

blood pressure is not seen until 4 to 5 hours. Although the -phase of its elimination has an estimated half-life of 5 to 10 hours, this almost certainly does not reflect the longer half-life of drug stored at its site of action in the neurosecretory vesicles of adrenergic neurons. The half-life of the pharmacological effect of guanadrel is determined by the drug's persistence in this neuronal pool, and that is probably at least 10 hours. Guanadrel is administered in a regimen of twice-daily doses. Guanadrel is cleared from the body by both renal and nonrenal disposition. Its elimination is impaired in patients with renal insufficiency; total-body clearance was reduced by 4- to 5-fold in a group of patients with a clearance of creatinine averaging 13 ml per minute. Adverse Effects Guanadrel produces undesirable effects that are related entirely to sympathetic blockade. Symptomatic hypotension during standing, exercise, ingestion of alcohol, or hot weather is the result of the lack of sympathetic compensation for these stresses. A general feeling of weakness and lassitude is partially, but not entirely, related to postural hypotension. Rarely, guanadrel can precipitate congestive heart failure in patients with limited cardiac reserve as a result of druginduced fluid retention. Sexual dysfunction usually presents as delayed or retrograde ejaculation. Diarrhea also may occur. Because guanadrel is actively transported to its site of action, drugs that block neuronal uptake of norepinephrine will inhibit the effect of guanadrel. Such drugs include the tricyclic antidepressants, cocaine, chlorpromazine, ephedrine, phenylpropanolamine, and amphetamine. Therapeutic Uses Because of the availability of a number of drugs that lower blood pressure without producing orthostatic hypotension, guanadrel is not employed in the monotherapy of hypertension, and is used chiefly as an additional agent in patients who have not achieved a satisfactory antihypertensive effect on two or more other agents. The usual starting dose is 10 mg daily, and side effects can be minimized by not exceeding 20 mg daily. Reserpine Reserpine is an alkaloid extracted from the root of Rauwolfia serpentina (Benth.), a climbing shrub indigenous to India. Descriptions of the medicinal use of the root of this plant are present in ancient Hindu Ayurvedic writings. "Modern" use of the whole root for the treatment of hypertension and psychoses was described in the Indian literature in 1931 (Sen and Bose, 1931). However, rauwolfia alkaloids were not used in western medicine until the mid-1950s. Reserpine was the first drug that was found to interfere with the function of the sympathetic nervous system in human beings, and its use began the modern era of effective pharmacotherapy of hypertension. The structure of reserpine is as follows:

Locus and Mechanism of Action Reserpine binds tightly to storage vesicles in central and peripheral adrenergic neurons, and the drug remains at such sites for prolonged periods of time (Giachetti and Shore, 1978). The storage vesicles are rendered dysfunctional as a result of their interaction with reserpine, and nerve endings lose their ability to concentrate and store norepinephrine and dopamine. Catecholamines leak into the cytoplasm, where they are destroyed by intraneuronal monoamine oxidase, and little or no active transmitter is discharged from nerve endings when they are depolarized. A similar process occurs at storage sites for 5-hydroxytryptamine. Reserpine-induced depletion of biogenic amines correlates with evidence of sympathetic dysfunction and antihypertensive effects. Recovery of sympathetic function requires synthesis of new storage vesicles, which takes days to weeks after discontinuation of the drug. Since reserpine depletes amines in the CNS as well as in the peripheral adrenergic neuron, it is probable that its antihypertensive effects are related to both a central and a peripheral action; it is certain that many of the side effects of reserpine are related to its effects in the CNS. Pharmacological Effects Both cardiac output and peripheral vascular resistance are reduced during long-term therapy with reserpine. Orthostatic hypotension may occur but does not usually cause symptoms. Heart rate and renin secretion fall. Salt and water are retained, which commonly results in "pseudotolerance." Absorption, Metabolism, and Excretion Few data are available on the pharmacokinetic properties of reserpine because of the lack of an assay capable of detecting low concentrations of the drug or its metabolites. Reserpine that is bound to isolated storage vesicles cannot be removed by dialysis, indicating that the binding is not in equilibrium with the surrounding medium. Because of the irreversible nature of reserpine binding, the amount of drug in plasma is unlikely to bear any consistent relationship to drug concentration at the site of action. Reserpine is entirely metabolized, and none of the parent drug is excreted unchanged. Toxicity and Precautions Most of the adverse effects of reserpine are due to its effect on the CNS. Sedation and inability to concentrate or perform complex tasks are the most common adverse effects. More serious is the occasional psychotic depression that can lead to suicide. Depression usually appears insidiously over many weeks or months and may not be attributed to the drug because of the delayed and gradual onset of symptoms. Reserpine must be discontinued at the first sign of depression, and the

drug should never be given to patients with a history of depression. Reserpine-induced depression may last several months after the drug is discontinued. Depression appears to be uncommon, but not unknown, with doses of 0.25 mg per day or less. Other side effects include nasal stuffiness and exacerbation of peptic ulcer disease, which is uncommon with small oral doses. Therapeutic Uses Reserpine was the sympatholytic drug used in the landmark Veterans Administration Cooperative Study that demonstrated the beneficial effects of treatment of hypertension (Veterans Administration Cooperative Study Group on Antihypertensive Agents, 1967, 1970), but with the availability of newer drugs that are both effective and well tolerated, the use of reserpine has diminished because of its CNS side effects. However, in comparative studies, low doses of reserpine given concurrently with a diuretic were as well tolerated as combinations of a diuretic with propranolol or methyldopa. The major advantage of reserpine is that it is much less expensive than other antihypertensive drugs. Reserpine is used once daily with a diuretic, and several weeks are necessary to achieve a maximum effect. The daily dose should be limited to 0.25 mg or less, and as little as 0.05 mg per day may be efficacious when a diuretic is also used. Metyrosine Metyrosine (DEMSER) is (-)- -methyl-L-tyrosine. It has the structure shown below. Metyrosine is an inhibitor of tyrosine hydroxylase, the enzyme that catalyzes the conversion of tyrosine to DOPA; this is the rate-limiting step in catecholamine biosynthesis (see Chapter 6: Neurotransmission: The Autonomic and Somatic Motor Nervous Systems). At a dose of 1 to 4 g per day, metyrosine decreases catecholamine biosynthesis by 35% to 80% in patients with pheochromocytoma. The maximal decrease in synthesis occurs only after several days, and the effect may be assessed by measurements of urinary catecholamines and their metabolites.

Metyrosine is used as an adjuvant to phenoxybenzamine and other -adrenergic blocking agents for the management of malignant pheochromocytoma and in the preoperative preparation of patients for resection of pheochromocytoma (Brogden et al., 1981). Metyrosine carries a risk of crystalluria, which can be minimized by maintaining a daily urine volume of more than 2 liters. Other adverse effects include orthostatic hypotension, sedation, extrapyramidal signs, diarrhea, anxiety, and psychic disturbances. Doses must be titrated carefully to achieve significant inhibition of catecholamine biosynthesis and yet minimize these substantive side effects. -Adrenergic Receptor Antagonists -Adrenergic receptor blocking drugs were not expected to have antihypertensive effects when they were first investigated in patients. However, pronethalol, a drug that was never marketed, was found to reduce arterial blood pressure in hypertensive patients with angina pectoris. This antihypertensive effect was subsequently demonstrated for propranolol and all other -adrenergic receptor antagonists. The pharmacology of these drugs is discussed in Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists; characteristics relevant to their use in hypertension will be described here.

Locus and Mechanism of Action Antagonism of -adrenergic receptors affects the regulation of the circulation through a number of mechanisms, including a reduction in myocardial contractility and cardiac output. An important consequence of blocking -adrenergic receptors is reduction in the secretion of renin with a resulting fall in the levels of angiotensin II. The weight of the evidence supports the concept that the reduction in angiotensin II, with its multiple effects on circulatory control and on aldosterone, contributes importantly to the antihypertensive action of this class of drugs, acting in concert with the cardiac effects. There clearly are effects of -adrenergic blockers, particularly in higher doses, that do not seem to be dependent on renin. A number of mechanisms have been postulated to account for a non-renin-dependent reduction in blood pressure, including alteration of the control of the sympathetic nervous system at the level of the CNS, a change in baroreceptor sensitivity, an alteration in peripheral adrenergic neuron function, and an increase in prostacyclin biosynthesis. Because all -adrenergic antagonists are effective antihypertensive agents and (+)-propranolol, which has little -adrenergic receptor blocking activity, has no effect on blood pressure, the antihypertensive therapeutic effect of these agents is undoubtedly related to blockade of receptors. Pharmacological Effects The -adrenergic blockers vary in their lipid solubility, selectivity for the 1-adrenergic receptor subtype, presence of partial agonist or intrinsic sympathomimetic activity, and membranestabilizing properties. Regardless of these differences, all of the -adrenergic receptor antagonists are equally effective as antihypertensive agents. Drugs without intrinsic sympathomimetic activity produce an initial reduction in cardiac output and a reflex-induced rise in peripheral resistance with no net change in arterial pressure. In patients who respond with a reduction in blood pressure, peripheral resistance returns to pretreatment values in a few hours to a few days. It is this delayed normalization of vascular resistance in the face of a persistently reduced cardiac output that accounts for the reduction in arterial pressure (van den Meiracker et al., 1988). Drugs with intrinsic sympathomimetic activity produce less of an effect on resting heart rate and cardiac output, and the fall in arterial pressure is correlated with a fall in vascular resistance below pretreatment levels, possibly because of stimulation of vascular 2-adrenergic receptors that mediate vasodilation. Renal blood flow is reduced in the short term by most -adrenergic antagonists, but reports of deterioration of renal function associated with long-term administration of these drugs are rare. Nevertheless, small reductions in renal plasma flow and glomerular filtration rate may persist, particularly with the nonselective drugs that block both 1- and 2-adrenergic receptors. Adverse Effects and Precautions The adverse effects of -adrenergic blocking agents are discussed in Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists. These drugs should be avoided in patients with reactive airway disease (asthma) or with sinoatrial or atrioventricular nodal dysfunction. -Adrenergic receptor antagonists should not be the initial drugs employed in hypertensive patients with cardiac failure because of the deleterious combination of a drop in myocardial contractility in conjuction with a rise in peripheral vascular resistance. After the cardiac failure has been addressed diagnostically and therapeutically, including the reduction of peripheral vascular resistance with another drug, -blockers may then be considered as rational components of long-term antihypertensive therapy. Patients with insulin-dependent diabetes also are better treated with other drugs. -Adrenergic receptor antagonists without intrinsic sympathomimetic activity increase

concentrations of triglycerides in plasma and lower those of HDL choles terol without changing total cholesterol concentrations. -Adrenergic blocking agents with intrinsic sympathomimetic activity have little or no effect on blood lipids or increase HDL cholesterol. The long-term consequences of these effects are unknown. Sudden discontinuation of some -adrenergic blockers can produce a withdrawal syndrome that is reminiscent of sympathetic hyperactivity; this can exacerbate the symptoms of coronary artery disease. Rebound hypertension to levels higher than those that existed before treatment has been noted with discontinuation of -adrenergic receptor antagonists in hypertensive patients (Houston and Hodge, 1988). Thus, -adrenergic blockers should not be discontinued abruptly except under close observation; dosage should be tapered over 10 to 14 days prior to discontinuation. Nonsteroidal antiinflammatory drugs such as indo-methacin can blunt the antihypertensive effect of propranolol and probably other -adrenergic receptor antagonists. This effect may be related to inhibition of vascular synthesis of prostacyclin, as well as to retention of Na+ (Beckmann et al., 1988). Epinephrine can produce severe hypertension and bradycardia when a nonselective -adrenergic receptor antagonist is present. This is due to the unopposed stimulation of -adrenergic receptors when vascular 2-receptors are blocked, and the bradycardia is the result of reflex vagal stimulation. Such "paradoxical" hypertensive responses to -adrenergic receptor antagonists have been observed in patients with hypoglycemia or pheochromocytoma or during withdrawal from clonidine or administration of epinephrine as a therapeutic agent. Therapeutic Uses The -adrenergic receptor antagonists provide effective therapy for all grades of hypertension. Despite marked differences in their pharmaco-kinetic properties, the antihypertensive effect of all the blockers is of sufficient duration to permit twice daily administration. Populations that have a lesser antihypertensive response to -blocking agents include the elderly and African Americans, but some individuals in these groups may have an excellent response. The -adrenergic receptor antagonists do not usually cause retention of salt and water, and administration of a diuretic is not necessary to avoid edema or the development of tolerance. However, diuretics do have additive antihypertensive effects when combined with blockers. The combination of a -adrenergic receptor antagonist, a diuretic, and a vasodilator is effective for patients who require a third drug. When minoxidil is the vasodilator, this combination can control the arterial pressure of most patients, even if they are resistant to other regimens.
1-Adrenergic

Receptor Antagonists

The development of drugs that selectively block 1-adrenergic receptors without affecting 2adrenergic receptors has added another group of antihypertensive agents. The pharmacology of these drugs is discussed in detail in Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists. Prazosin (MINIPRESS), terazosin (HYTRIN), and doxazosin (CARDURA) are the agents that are available for the treatment of hypertension. Additionally, investigational drugs such as ketanserin, indoramin, and urapidil may owe a major portion of their antihypertensive effects to blockade of 1-adrenergic receptors (Cubeddu, 1988). Pharmacological Effects Initially, the
1-adrenergic

receptor antagonists reduce arteriolar resistance and venous capacitance;

this causes a sympathetically mediated reflex increase in heart rate and plasma renin activity. During long-term therapy, vasodilation persists, but cardiac output, heart rate, and plasma renin activity return to normal. Renal blood flow is unchanged during therapy with an 1-adrenergic receptor antagonist. The 1-adrenergic blockers cause a variable amount of postural hypotension, depending on the plasma volume. Retention of salt and water occurs in many patients during continued administration, and this attenuates the postural hypotension. 1-Adrenergic receptor antagonists reduce plasma concentrations of triglycerides and total and LDL cholesterol and increase HDL cholesterol. These potentially favorable effects on lipids persist when a thiazide-type diuretic is given concurrently. The long-term consequences of these small, drug-induced changes in lipids are unknown. Adverse Effects The use of doxazosin as monotherapy for hypertension increases the risk for developing congestive heart failure. There is every reason to assume that this is a "class effect" and represents an adverse effect of all of the 1-adrenergic receptor antagonists. A major precaution regarding the use of the 1- adrenergic receptor antagonists for hypertension is the so-called first-dose phenomenonsymptomatic orthostatic hypotension that occurs within 90 minutes of the initial dose of the drug or when the dosage is increased rapidly. This effect may be seen in up to 50% of patients, and it is particularly likely to occur in patients who are already receiving a diuretic or a -adrenergic receptor antagonist. After the first few doses, patients develop a tolerance to this marked hypotensive response. Therapeutic Uses Because 1-adrenergic receptor antagonists increase the risk of cardiac failure, they are not recommended as monotherapy for hypertensive patients. Thus, they are used primarily in conjunction with diuretics, -adrenergic receptor blockers, and other antihypertensive agents. Adrenergic receptor antagonists enhance the efficacy of the 1 blockers. 1-Adrenergic receptor antagonists are not the drugs of choice in patients with pheochromocytoma, because a vasoconstrictor response to epinephrine can still result from activation of unblocked vascular 2adrenergic receptors. Combined
1-and

-Adrenergic Receptor Antagonists

Labetalol (NORMODYNE, TRANDATE) (see Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists) is an equimolar mixture of four stereoisomers. One isomer is an 1-adrenergic receptor antagonist (like prazosin), another is a nonselective -adrenergic receptor antagonist with partial agonist activity (like pindolol), and the other two isomers are inactive. The isomer that is the -adrenergic receptor antagonist has been under development as a separate drug (dilevalol) (Lund-Johansen, 1988). Labetalol lowers arterial pressure by reducing vascular resistance as a consequence of blockade of 1-adrenergic receptors and stimulation of 2-adrenergic receptors. Cardiac output at rest is not reduced. Because of its capacity to block 1-adrenergic receptors, labetalol given intravenously can reduce pressure sufficiently rapidly to be useful for the treatment of hypertensive emergencies. Given over the long term, labetalol has efficacy and side effects that would be expected with any combination of - and 1-adrenergic receptor antagonists; it also has the disadvantages that are inherent in fixed-dose combination products. Carvedilol (COREG) (see Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists) is a -adrenergic receptor antagonist with 1-adrenergic receptor antagonist

activity that has been approved for the treatment of essential hypertension and for the treatment of symptomatic heart failure. The ratio of 1- to -adrenergic receptor antagonist potency for carvedilol is 1:10. Carvedilol undergoes oxidative metabolism and glucuronidation in the liver; the oxidative metabolism occurs via cytochrome CYP2D6. Carvedilol has been shown to reduce mortality in patients with systolic dysfunction and New York Heart Association class I, II, or III symptoms when used as an adjunct to therapy with diuretics and angiotensin converting enzyme inhibitors. It should not be given to those patients with decompensated heart failure who are dependent on sympathetic stimulation. As with labetalol, the long-term efficacy and side effects of carvedilol in hypertension are predictable based on its properties as a - and 1-adrenergic receptor antagonist. In addition, mild reversible hepatocellular injury has been reported with carvedilol. Vasodilators Hydralazine Hydralazine (APRESOLINE) was one of the first orally active antihypertensive drugs to be marketed in the United States; however, the drug initially was used infrequently because of tachycardia and tachyphylaxis. With a better understanding of the compensatory cardiovascular responses that accompany use of arteriolar vasodilators, hydralazine was combined with sympatholytic agents and diuretics with greater therapeutic success. Numerous phthalazines have been synthesized in the hope of producing vasoactive agents, but only those with hydrazine moieties in the 1 or 4 position of the ring have vasodilatory activity (Reece, 1981). None of the analogs has any advantage over hydralazine. Hydralazine (1-hydrazinophthalazine) has the following structural formula:

Locus and Mechanism of Action Hydralazine causes direct relaxation of arteriolar smooth muscle. The molecular mechanism of this effect is not known. It is not a dilator of capacitance vessels (e.g., the epicardial coronary arteries) and does not relax venous smooth muscle. Hydralazine-induced vasodilation is associated with powerful stimulation of the sympathetic nervous system, which results in increased heart rate and contractility, increased plasma renin activity, and fluid retention; all of these effects counteract the antihypertensive effect of hydralazine. Although most of the sympathetic activity is due to a baroreceptor-mediated reflex, hydralazine may stimulate the release of norepinephrine from sympathetic nerve terminals and augment myocardial contractility directly (Azuma et al., 1987). Pharmacological Effects Most of the effects of hydralazine are confined to the cardiovascular system. The decrease in blood pressure after administration of hydralazine is associated with a selective decrease in vascular resistance in the coronary, cerebral, and renal circulations, with a smaller effect in skin and muscle. Because of preferential dilation of arterioles over veins, postural hypotension is not a common problem; hydralazine lowers blood pressure equally in the supine and upright positions. Although hydralazine lowers pulmonary vascular resistance, the greater increase in cardiac output can cause mild pulmonary hypertension. It is difficult to predict which patients will respond in this manner, but the increase in cardiac output can be attenuated by the use of -adrenergic receptor blocking

agents. Absorption, Metabolism, and Excretion Hydralazine is N-acetylated in the bowel and/or the liver. The rate of acetylation is genetically determined; about half of the people in the United States acetylate rapidly and half do so slowly. Since the acetylated compound is inactive, the dose necessary to produce a systemic effect is larger in fast acetylators. Hydralazine is well absorbed through the gastrointestinal tract, but the systemic bioavailability is low (16% in fast acetylators and 35% in slow acetylators). The half-life of hydralazine is 1 hour, and the systemic clearance of the drug is about 50 ml/kg per minute. Since the systemic clearance exceeds hepatic blood flow, extrahepatic metabolism must occur. Indeed, hydralazine rapidly combines with circulating -keto acids to form hydrazones, and the major metabolite recovered from the plasma is hydralazine pyruvic acid hydrazone. This metabolite has a longer half-life than hydralazine, but it does not appear to be very active (Reece et al., 1985). Although the rate of acetylation is an important determinant of the bioavailability of hydralazine, it does not play a role in the systemic elimination of the drug, probably because the hepatic clearance is so high that systemic elimination is principally a function of hepatic blood flow. The peak concentration of hydralazine in plasma and the peak hypotensive effect of the drug occur within 30 to 120 minutes of ingestion. Although its half-life in plasma is about an hour, the duration of the hypotensive effect of hydralazine can last as long as 12 hours. There is no clear explanation for this discrepancy. Toxicity and Precautions Two types of side effects occur after the use of hydralazine. The first, which are extensions of the pharmacological effects of the drug, include headache, nausea, flushing, hypotension, palpitation, tachycardia, dizziness, and angina pectoris. Myocardial ischemia occurs because of the increased oxygen demand imposed by the baroreflex-induced stimulation of the sympathetic nervous system and also because hydralazine does not dilate the epicardial coronary arteries; thus, the arteriolar dilation it produces may cause a "steal" of blood flow away from the ischemic region. Following parenteral administration to patients with coronary artery disease, the myocardial ischemia may be sufficiently severe and protracted to cause frank myocardial infarction. For this reason, parenteral administration of hydralazine is contraindicated in hypertensive patients with coronary artery disease and inadvisable for most hypertensive patients over 40 years old. In addition, if the drug is used alone, there may be salt retention with development of high-output congestive heart failure. These symptoms were common during the early clinical use of hydralazine; because tachyphylaxis developed, the daily dose of the drug was frequently increased to 400 to 1000 mg. When combined with a -adrenergic receptor blocker and a diuretic, hydralazine is better tolerated, although side effects such as headache are still commonly described and may necessitate discontinuation of the drug. The second type of side effect is caused by immunological reactions, of which the drug-induced lupus syndrome is the most common. Administration of hydralazine also can result in an illness that resembles serum sickness, hemolytic anemia, vasculitis, and rapidly progressive glomerulonephritis. The mechanism of these autoimmune reactions is unknown, but hydralazine has been shown to inhibit methylation of DNA and induce self-reactivity in T cells (Cornacchia et al., 1988). The drug-induced lupus syndrome usually occurs after at least 6 months of continuous treatment with hydralazine, and its incidence is related to dose, sex, acetylator phenotype, and race (Perry,

1973). In one study, after three years of treatment with hydralazine, drug-induced lupus occurred in 10.4% of patients who received 200 mg daily, 5.4% who received 100 mg daily, and none who received 50 mg daily (Cameron and Ramsay, 1984). The incidence is four times higher in women than in men, and the syndrome is seen more commonly in Caucasians than in African Americans. The rate of conversion to a positive antinuclear antibody test is faster in slow acetylators than in rapid acetylators, suggesting that the native drug or a nonacetylated metabolite is responsible. However, since the majority of patients with positive antinuclear antibody tests do not develop the drug-induced lupus syndrome, hydralazine need not be discontinued unless clinical features of the syndrome appear. These features are similar to those of other drug-induced lupus syndromes and consist mainly of arthralgia, arthritis, and fever. Pleuritis and pericarditis may be present, and pericardial effusion can occasionally cause cardiac tamponade. Discontinuation of the drug is all that is necessary for most patients with the hydralazine-induced lupus syndrome, but symptoms may persist in a few patients and administration of corticosteroids may be necessary. Hydralazine also can produce a pyridoxine-responsive polyneuropathy. The mechanism appears to be related to the ability of hydralazine to combine with pyridoxine to form a hydrazone. This side effect is very unusual with doses up to 200 mg per day. Therapeutic Uses Hydralazine generally is not used as the sole drug for the long-term treatment of hypertension because of the development of tachyphylaxis secondary to an increase in cardiac output and fluid retention. In addition, the drug should be used with the greatest of caution in elderly patients and in hypertensive patients with coronary artery disease because of the possibility of precipitation of myocardial ischemia. The usual oral dosage of hydralazine is 25 to 100 mg twice daily. Twice-daily administration is as effective as administration four times a day for control of blood pressure, regardless of acetylator phenotype. The maximum recommended dose of hydralazine is 200 mg per day to minimize drug-induced lupus syndrome. Slow acetylators show a better response to this dosage than do fast acetylators because of the greater bioavailability of the drug. Hydralazine has been used widely to treat hypertension that occurs during pregnancy. However, the drug should be used cautiously during early pregnancy, since hydralazine can combine with DNA and cause a positive Ames test (Williams et al., 1980). Parenteral administration of hydralazine has been used for the treatment of hypertensive emergencies in pregnancy, but is not recommended for the treatment of hypertensive emergencies in patients in the age range for coronary artery disease. The drug is contraindicated for the short-term production of hypotension in patients with dissecting aortic aneurysm or in those with symptomatic ischemic heart disease. Minoxidil The discovery in 1965 of the hypotensive action of minoxidil (LONITEN) was a significant advance in the treatment of hypertension, since the drug has proven to be efficacious in patients with the most severe and drug-resistant forms of hypertension. The chemical structure of minoxidil is as follows:

Locus and Mechanism of Action Minoxidil is not active in vitro but must be metabolized by hepatic sulfotransferase to the active molecule, minoxidil N-O sulfate (McCall et al., 1983); the formation of this active metabolite is a minor pathway in the metabolic disposition of minoxidil. Minoxidil sulfate relaxes vascular smooth muscle in isolated systems where the parent drug is inactive. Minoxidil sulfate activates the ATPmodulated potassium channel. By opening potassium channels in smooth muscle and thereby permitting potassium efflux, it causes hyperpolarization and relaxation of smooth muscle (Leblanc et al., 1989). Pharmacological Effects Minoxidil produces arteriolar vasodilation with essentially no effect on the capacitance vessels; the drug resembles hydralazine and diazoxide in this regard. Minoxidil increases blood flow to skin, skeletal muscle, the gastrointestinal tract, and the heart more than to the CNS. The disproportionate increase in blood flow to the heart may have a metabolic basis, in that administration of minoxidil is associated with a reflex increase in myocardial contractility and in cardiac output. The cardiac output can increase markedly, as much as three- to fourfold. The principal determinant of the elevation in cardiac output is the action of minoxidil on peripheral vascular resistance to enhance venous return to the heart; by inference from studies with other drugs, the increased venous return probably results from enhancement of flow in the regional vascular beds with a fast time constant for venous return to the heart (Ogilvie, 1985). The adrenergically mediated increase in myocardial contractility contributes to the increased cardiac output, but is not the predominant causal factor. The effects of minoxidil on the kidney are complex. Minoxidil is a renal vasodilator, but systemic hypotension produced by the drug occasionally can decrease renal blood flow. However, in the majority of patients who take minoxidil for the treatment of hypertension, renal function improves, especially if renal dysfunction is secondary to hypertension (Mitchell et al., 1980). Minoxidil is a very potent stimulator of renin secretion; this effect is mediated by a combination of renal sympathetic stimulation and activation of the intrinsic renal mechanisms for regulation of renin release. Absorption, Metabolism, and Excretion Minoxidil is well absorbed from the gastrointestinal tract. Although peak concentrations of minoxidil in blood occur 1 hour after oral administration, the maximal hypotensive effect of the drug occurs later, possibly because formation of the active metabolite is delayed. Only about 20% of the absorbed drug is excreted unchanged in the urine, and the main route of elimination is by hepatic metabolism. The major metabolite of minoxidil is the glucuronide conjugate at the N-oxide position in the pyrimidine ring. This metabolite is less active than minoxidil, but it persists longer in the body. The extent of biotransformation of minoxidil to its active metabolite, minoxidil N-O sulfate, has not been evaluated in human beings. Minoxidil has a half-life in plasma of 3 to 4 hours, but its duration of action is 24 hours or occasionally even longer. It has been proposed that

persistence of minoxidil in vascular smooth muscle is responsible for this discrepancy. However, without knowledge of the pharmacokinetic properties of the active metabolite, an explanation for the prolonged duration of action cannot be given. Adverse Effects and Precautions The adverse effects of minoxidil are predictable and can be divided into three major categories: fluid and salt retention, cardiovascular effects, and hypertrichosis. Retention of salt and water results from increased proximal renal tubular reabsorption, which is in turn secondary to reduced renal perfusion pressure and to reflex stimulation of renal tubular adrenergic receptors. Similar antinatriuretic effects can be observed with the other arteriolar dilators (e.g., diazoxide and hydralazine). Although administration of minoxidil causes increased secretion of renin and aldosterone, this is not an important mechanism for retention of salt and water in this case. Fluid retention usually can be controlled by the administration of a diuretic. However, thiazides may not be sufficiently efficacious, and it may be necessary to use a loop diuretic. This is especially true if the patient has any degree of renal dysfunction. The cardiac consequences of the baroreceptor-mediated activation of the sympathetic nervous system during minoxidil therapy are similar to those seen with hydralazine; there is an increase in heart rate, myocardial contractility, and myocardial oxygen consumption. Thus, myocardial ischemia can be induced by minoxidil in patients with coronary artery disease. The cardiac sympathetic responses are attenuated by concurrent administration of a -adrenergic blocker. The adrenergically induced increase in renin secretion also can be ameliorated by a -adrenergic receptor antagonist or an angiotensin converting enzyme inhibitor, with enhancement of the blood pressure control. The increased cardiac output evoked by minoxidil has particularly adverse consequences in those hypertensive patients who have left ventricular hypertrophy and diastolic dysfunction. Such poorly compliant ventricles respond suboptimally to increased volume loads, with a resulting increase in left ventricular filling pressure. This probably is a major contributor to the increased pulmonary artery pressure seen with minoxidil (and hydralazine) therapy in hypertensive patients, and is compounded by the retention of salt and water caused by minoxidil. Cardiac failure can result from minoxidil therapy in such patients; the potential for this complication can be reduced but not prevented by effective diuretic therapy. Pericardial effusion is an uncommon but serious complication of minoxidil. Although more commonly described in patients with cardiac failure and renal failure, pericardial effusion can occur in patients with normal cardiovascular and renal function. Mild and asymptomatic pericardial effusion is not an indication for discontinuing minoxidil, but the situation should be monitored closely to avoid progression to tamponade. Effusion usually clears when the drug is discontinued, but it will recur if treatment with minoxidil is resumed (Reichgott, 1981). Flattened and inverted T waves frequently are observed in the electrocardiogram following the initiation of minoxidil treatment. These are not ischemic in origin and are seen with other drugs that activate potassium channels. These drugs accelerate myocardial repolarization, shorten the refractory period, and one of them, pinacidil, lowers the ventricular fibrillation threshold and increases spontaneous ventricular fibrillation in the setting of myocardial ischemia (Chi et al., 1990). The effect of minoxidil on the refractory period and ischemic ventricular fibrillation has not been investigated; whether or not such findings enhance the risk of ventricular fibrillation in human myocardial ischemia is unknown.

Hypertrichosis occurs in all patients who receive minoxidil for an extended period and is probably a consequence of potassium channel activation. Growth of hair occurs on the face, back, arms, and legs and is particularly offensive to women. Frequent shaving or depilatory agents can be used to manage this problem. Topical minoxidil (ROGAINE) is now marketed for the treatment of malepattern baldness. The topical use of minoxidil can cause measurable cardiovascular effects in some individuals (Leenen et al., 1988). Other side effects of the drug are rare and include rashes, StevensJohnson syndrome, glucose intolerance, serosanguinous bullae, formation of antinuclear antibodies, and thrombocytopenia. Therapeutic Uses Minoxidil is best reserved for the treatment of severe hypertension that responds poorly to other antihypertensive medications (Campese, 1981). It has been used successfully in the treatment of hypertension in both adults and children. Minoxidil should never be used alone; it must be given concurrently with a diuretic to avoid fluid retention and with a sympatholytic drug (usually a receptor antagonist) to control reflex cardiovascular effects. The drug usually is administered either once or twice a day, but some patients may require more frequent dosage for adequate control of blood pressure. The initial daily dose of minoxidil may be as little as 1.25 mg, which can be increased gradually to 40 mg in one or two daily doses. Sodium Nitroprusside Although sodium nitroprusside has been known since 1850 and its hypotensive effect in human beings was described in 1929, its safety and usefulness for the short-term control of severe hypertension were not demonstrated until the mid-1950s. Several investigators subsequently demonstrated that sodium nitroprusside also was effective in improving cardiac function in patients with left ventricular failure (see Chapter 34: Pharmacological Treatment of Heart Failure). The structural formula of sodium nitroprusside is as follows:

Locus and Mechanism of Action Nitroprusside is a nitrovasodilator. It is metabolized by blood vessels to its active metabolite, nitric oxide. Nitric oxide activates guanylyl cyclase, leading to the formation of cyclic GMP and vasodilation (Murad, 1986). The metabolic activation of nitroprusside is catalyzed by a different nitric oxidegenerating system than that for nitroglycerin, probably accounting for the difference in the potency of these drugs at different vascular sites and the fact that tolerance develops to nitroglycerin but not to nitroprusside (Kowaluk et al., 1992). Pharmacological Effects Nitroprusside dilates both arterioles and venules, and the hemodynamic response to its administration results from a combination of venous pooling and reduced arterial impedance.

Because of its effect on venules, the hypotensive effect of sodium nitroprusside is greater when the patient is upright. In subjects with normal left ventricular function, venous pooling affects cardiac output more than does the reduction of afterload; cardiac output thus tends to fall. In contrast, in patients with severely impaired left ventricular function and diastolic ventricular distention, the reduction of arterial impedance is the predominant effect, leading to a rise in cardiac output (see Chapter 34: Pharmacological Treatment of Heart Failure). Sodium nitroprusside is a nonselective vasodilator, and regional distribution of blood flow is little affected by the drug. In general, renal blood flow and glomerular filtration are maintained, and plasma renin activity increases. Unlike minoxidil, hydralazine, diazoxide, and other arteriolar vasodilators, sodium nitroprusside usually causes only a modest increase in heart rate and an overall reduction in myocardial demand for oxygen. Absorption, Metabolism, and Excretion Sodium nitroprusside is an unstable molecule that decomposes under strongly alkaline conditions and when exposed to light. The drug must be given by continuous intravenous infusion to be effective. Its onset of action is within 30 seconds; the peak hypotensive effect occurs within 2 minutes, and when the infusion of the drug is stopped, the effect disappears within 3 minutes. The metabolism of nitroprusside by smooth muscle is initiated by its reduction, which is followed by the release of cyanide and then nitric oxide (Bates et al., 1991; Ivankovich et al., 1978). Cyanide is further metabolized by liver rhodanase to thiocyanate, which is eliminated almost entirely in the urine. The mean elimination half-time for thiocyanate is 3 days in patients with normal renal function, and it can be much longer in patients with renal insufficiency. Toxicity and Precautions The short-term side effects of nitroprusside are due to excessive vasodilation, with hypotension and the consequences thereof. Close monitoring of blood pressure and the use of a continuous variablerate infusion pump will prevent an excessive hemodynamic response to the drug in the majority of cases. Less commonly, toxicity may result from conversion of nitroprusside to cyanide and thiocyanate. Toxic accumulation of cyanide leading to severe lactic acidosis can occur usually if sodium nitroprusside is infused at a rate greater than 5 g/kg per minute, but such toxicity can occur in some patients receiving doses about 2 g/kg per minute. The limiting factor in the metabolism of cyanide appears to be the availability of sulfur-containing substrates in the body (mainly thiosulfate). The concomitant administration of sodium thiosulfate can prevent accumulation of cyanide in patients who are receiving higher than usual doses of sodium nitroprusside; the efficacy of the drug is unchanged (Schulz, 1984). The risk of thiocyanate toxicity increases when sodium nitroprusside is infused for more than 24 to 48 hours, especially if renal function is impaired. Signs and symptoms of thiocyanate toxicity include anorexia, nausea, fatigue, disorientation, and toxic psychosis. The plasma concentration of thiocyanate should be monitored during prolonged infusions of nitroprusside and should not be allowed to exceed 0.1 mg/ml. Rarely, excessive concentrations of thiocyanate may cause hypothyroidism by inhibiting iodine uptake by the thyroid gland. In patients with renal failure, thiocyanate can be removed readily by hemodialysis. Nitroprusside can worsen arterial hypoxemia in patients with chronic obstructive pulmonary disease because the drug interferes with hypoxic pulmonary vasoconstriction and therefore promotes mismatching of ventilation with perfusion. Rebound hypertension may occur after abrupt cessation of short-term nitroprusside infusions (Packer et al., 1979); this may be caused by persistently

elevated concentrations of renin in the plasma. Therapeutic Uses Sodium nitroprusside is used primarily to treat hypertensive emergencies, but the drug can be used in many situations when short-term reduction of cardiac preload and/or afterload is desired. Thus, nitroprusside has been used to lower blood pressure during acute aortic dissection, to increase cardiac output in congestive heart failure (see Chapter 34: Pharmacological Treatment of Heart Failure), and to decrease myocardial oxygen demand after acute myocardial infarction. In addition, nitroprusside is the drug most often used to induce controlled hypotension during anesthesia in order to reduce bleeding in surgical procedures. In the treatment of acute aortic dissection, it is important to administer a -adrenergic receptor antagonist with nitroprusside, since reduction of blood pressure with nitroprusside alone can increase the rate of rise in pressure in the aorta as a result of increased myocardial contractility, thereby enhancing propagation of the dissection. Sodium nitroprusside is available in vials that contain 50 mg. The contents of the vial should be dissolved in 2 to 3 ml of 5% dextrose in water. Addition of this solution to 250 to 1000 ml of 5% dextrose in water produces a concentration of 50 to 200 g/ml. Because the compound decomposes in light, only fresh solutions should be used, and the bottle should be covered with an opaque wrapping. The drug must be administered as a controlled, continuous infusion, and the patient must be closely observed. The majority of hypertensive patients respond to an infusion of 0.25 to 1.5 g/kg per minute. Higher rates of infusion are necessary to produce controlled hypotension in normotensive patients under surgical anesthesia. Infusion of nitroprusside at rates exceeding 5 g/kg per minute over a prolonged period can cause cyanide and/or thiocyanate poisoning. Patients who are receiving other antihypertensive medications usually require less nitroprusside to lower blood pressure. If infusion rates of 10 g/kg per minute do not produce adequate reduction of blood pressure within 10 minutes, the rate of administration of nitroprusside should be reduced to minimize potential toxicity. Diazoxide Diazoxide (HYPERSTAT IV ) is used in the treatment of hypertensive emergencies. Sodium nitroprusside is the drug of choice for this indication, but diazoxide maintains a place in the treatment of hypertensive emergencies in situations in which accurate infusion pumps are not available and/or close monitoring of blood pressure is not feasible. The drug is a benzothiadiazine derivative, like the thiazide diuretics, but it does not cause diuresis, apparently because it lacks a sulfonamido group. Its structural formula is as follows:

Mechanism of Action and Pharmacological Effects Diazoxide hyperpolarizes arterial smooth muscle cells by activating ATP-sensitive K+ channels; this causes relaxation of the vascular smooth muscle (Standen et al., 1989). The effect of the drug in vivo is exclusively arteriolar, with negligible effect on capacitance vessels. This produces substantial reflex activation of the sympathetic nervous system. Cardiac output may double from

stimulation of heart rate and myocardial contractility. The avid retention of salt and water is probably a result of stimulation of renal sympathetic nerves and changes in intrarenal hemodynamics, as with other arteriolar vasodilators. Diazoxide increases coronary blood flow, and cerebral and renal blood flows are maintained by autoregulation. Renin secretion is enhanced, and the combination of an increased cardiac output, salt and water retention, and elevated concentrations of angiotensin II counteract the antihypertensive effects of diazoxide. Absorption, Metabolism, and Excretion Although well absorbed orally, diazoxide is administered only intravenously for the treatment of severe hypertension. Approximately 20% to 50% of the drug is eliminated as such by the kidney, and the rest is metabolized in the liver to the 3-hydroxymethyl and 3-carboxy derivatives (Pruitt et al., 1974). Although the plasma half-life of diazoxide is 20 to 60 hours, the duration of the hypotensive response to the drug is variable and can be as short as 4 hours or as long as 20 hours; the development of a brisk rise in renin secretion may antagonize the early hypotensive effect of diazoxide. The main indication for the use of diazoxide is for the treatment of hypertensive emergencies. Injection of an intravenous bolus lowers blood pressure within 30 seconds, and a maximum effect is achieved within 3 to 5 minutes. Although initial recommendations were to administer a 300-mg bolus of diazoxide, excessive hypotension with resultant cerebral and cardiovascular damage has resulted from this practice. Hypotension can be minimized by the administration of a "minibolus" of 50 to 150 mg at intervals of 5 to 15 minutes until the desired blood pressure is achieved (Wilson and Vidt, 1978). Diazoxide also can be given by slow intravenous infusion at a rate of 15 to 30 mg per minute (Garrett and Kaplan, 1982). Prior administration of a -adrenergic receptor antagonist will enhance the hypotensive effect of the drug. Diazoxide should not be used to treat hypertension associated with aortic coarctation, arteriovenous shunts, or aortic dissection. Similarly, risks outweigh benefits in its use for acute pulmonary edema and ischemic heart disease. Adverse Effects and Precautions The most common side effects caused by diazoxide are myocardial ischemia, salt and water retention, and hyperglycemia. Myocardial ischemia may be precipitated or aggravated by diazoxide, and it results from the reflex adrenergic stimulation of the heart and from increased flow to nonischemic regions that "steal" blood flow from the regions supplied by stenotic vessels. Retention of fluid can be avoided by restriction of salt and water. The routine use of diuretic agents with diazoxide is not recommended because patients with malignant hypertension are frequently volumedepleted. Hyperglycemia results from diazoxide's capacity to inhibit the secretion of insulin from pancreatic cells. This effect also appears to result from stimulation of ATP-sensitive K+ channels (Znkler et al., 1988). The drug does not alter the response to administration of insulin. Thus, hyperglycemia is mainly a problem in non-insulin-dependent diabetic patients who are being treated with oral hypoglycemic agents. Severe hyperglycemia with hyperosmolar, nonketotic coma has been described. Cerebral ischemia may be caused by excessive hypotension. Diazoxide relaxes uterine smooth muscle and may arrest labor when used to treat the hypertensive crisis of eclampsia. Rare side effects include gastrointestinal disturbances, flushing, local pain and inflammation after extravasation, altered ability to taste and smell, excessive salivation, and dyspnea. Ca2+-Channel Antagonists Ca2+-channel blocking agents are an important group of drugs for the treatment of hypertension. The general pharmacology of these drugs is presented in Chapter 32: Drugs Used for the Treatment

of Myocardial Ischemia; their use in heart failure is discussed in Chapter 34: Pharmacological Treatment of Heart Failure; and their use in cardiac arrhythmia is covered in Chapter 35: Antiarrhythmic Drugs. The logic behind their use in hypertension comes from the understanding that fixed hypertension is the result of increased peripheral vascular resistance. Since contraction of vascular smooth muscle is dependent on the free intracellular concentration of Ca2+, inhibition of transmembrane movement of Ca2+ should decrease the total amount of Ca 2+ that reaches intracellular sites. Indeed, all of the Ca 2+-channel blockers lower blood pressure by relaxing arteriolar smooth muscle and decreasing peripheral vascular resistance (Lehmann et al., 1983). As a consequence of a decrease in peripheral vascular resistance, the Ca2+-channel blockers evoke a baroreceptor-mediated sympathetic discharge. In the case of the dihydropyridines, mild to moderate tachycardia ensues from the adrenergic stimulation of the sinoatrial node, whereas tachycardia is minimal to absent with verapamil and diltiazem because of the direct negative chronotropic effect of these two drugs. The increased adrenergic stimulation of the heart serves to counter the negative 2+ inotropic effect of Ca -channel blockers such as verapamil, diltiazem, and nifedipine; the importance of this compensatory support of myocardial contractility should be considered in decisions regarding possible concurrent use of -adrenergic receptor antagonists, particularly in 2+ patients who may be prone to hypertensive cardiac failure. The adrenergic reflex response to Ca channel blockers also acts to attenuate the hypotensive effect of these drugs; thus, when the reflex vasoconstriction is diminished, as in the elderly or during treatment with -adrenergic receptor 2+ antagonists, the hypotensive effect of the Ca -channel blockers is increased, sometimes excessively. In considering the cardiovascular effects of the Ca -channel blockers, it is essential to evaluate both the hemodynamic effects in the normal heart and the interaction of these drugs with cardiac disease, given that both cardiac failure and coronary artery diseases are important consequences of hypertension and that left ventricular hypertrophy is a harbinger for sudden cardiac death in 2+ hypertensive patients. As a consequence of the peripheral vasodilation, Ca -channel blockers may increase venous return, which will result in an increased cardiac output except in the case of those that exert substantial negative inotropic effects (e.g., verapamil and diltiazem). The increased venous return is not as great as with minoxidil or hydralazine, but is a consideration in the management of patients with diastolic dysfunction due to hypertensive cardiomyopathy who are at risk of left ventricular failure. The Ca2+-channel blockers do not improve the diastolic function of the ventricle. Although earlier noninvasive studies had demonstrated that peak filling rates of the left ventricle of hypertensive patients were shortened by Ca 2+-channel blockers, direct hemodynamic evaluation of ventricular function has demonstrated that verapamil causes an increase in left ventricular end-diastolic pressure, an undesirable hemodynamic consequence that occurs in conjunction with, and probably as a contributor to, the acceleration of peak filling rate (Nishimura et al., 1993). In addition to these findings that Ca2+-channel blockers do not improve and have the potential for worsening the hemodynamics in diastolic dysfunction, the long-term effects of the Ca2+-channel blockers on left ventricular hypertrophy, a major contributor to diastolic dysfunction, should be considered. An overview of all trials evaluating the effects of antihypertensive agents on left ventricular mass concludes that, although Ca2+-channel blockers do reduce left ventricular mass and do so more effectively than diuretics, they are less effective than angiotensin converting enzyme inhibitors and methyldopa (Dalhf et al., 1993). Based on the sum of this evidence, Ca2+-channel blockers probably are not the first choice as the initial drug in the treatment of patients whose hypertension is accompanied by left ventricular hypertrophy nor as the predominant drug in a combination for their treatment. All Ca2+ -channel blockers are equally effective when used alone for the treatment of mild to
2+

moderate hypertension; and in comparative trials, Ca2+-channel blockers are as effective in lowering blood pressure as -adrenergic receptor antagonists or diuretics (Doyle, 1983; Inouye et al., 1984). The presence of ischemic heart disease in conjunction with hypertension raises a special set of concerns regarding some of the Ca 2+-channel blockers. Dihydropyridine Ca 2+-channel blockers do not improve survival in patients following myocardial infarction. Neither verapamil nor diltiazem improves mortality in the entire group of postinfarction patients. Data suggesting that diltiazem may exert a favorable effect on mortality in the subset of patients who exhibit no abnormality in systolic ventricular function does not form a strong basis for its use in such patients because of the post hoc selection of this group for analysis (Multicenter Diltiazem Postinfarction Trial Research Group, 1988). This is in contrast to the clear benefit to the survival of patients after myocardial infarction that is conferred by treatment with -adrenergic receptor antagonists and angiotensin converting enzyme inhibitors. Accordingly, the Ca2+-channel blockers are not the initial or even the second drugs to be used in the treatment of the hypertension in patients who have had a myocardial infarction. In diabetic patients with hypertension, the available evidence favors an angiotensin converting enzyme inhibitor as the initial antihypertensive drug (Estacio et al., 1998), following which it is appropriate to consider adding a Ca2+-channel blocker if a second drug is required. The profile of adverse reactions to the Ca2+-channel blockers varies among the drugs in this class, but only a small fraction of patients discontinue these drugs because of perceived adverse reactions. The dihydropyridines cause the highest incidence of vascular side effects. Approximately 10% of patients receiving the standard formulation of nifedipine (immediate-release capsules) develop headache, flushing, dizziness, and peripheral edema. Dizziness and flushing are much less of a problem with the sustained-release formulations and with the dihydropyridines having a long halflife and relatively constant concentrations of drug in plasma. The edema usually is not the result of fluid retention; it most likely results from increased hydrostatic pressure in the lower extremities owing to precapillary dilation and reflex postcapillary constriction. Contraction of the lower esophageal sphincter is inhibited by the Ca2+-channel blockers. Accordingly, all Ca2+ -channel blockers can cause gastroesophageal reflux. Constipation is a common side effect of verapamil, but it occurs less frequently with other Ca2+-channel blockers. Inhibition of sinoatrial node function by diltiazem and verapamil can lead to bradycardia and even sinoatrial node arrest, particularly in patients with sinoatrial node dysfunction; this effect is exaggerated by concurrent use of adrenergic receptor antagonists. Oral administration of nifedipine as an approach to urgent reduction of blood pressure has been abandoned. Sublingual administration does not achieve the maximum plasma concentration any more quickly than does oral administration. Moreover, in the absence of deleterious consequences of high arterial pressure, data do not support the rapid lowering of blood pressure. Short-acting, parenterally administered agents should be used in the setting of hypertensive emergency. Nifedipine is a suboptimal choice for treatment of hypertensive pulmonary edema; nitroprusside produces a greater reduction in left ventricular end-diastolic pressure than do equihypotensive doses of nifedipine (Aroney et al., 1991), and the magnitude of nitroprusside's pharmacological effect can be regulated more effectively. There is no place in the treatment of hypertension for the use of 2+ nifedipine or other dihydropyridine Ca -channel blockers with short half-lives when administered in a standard (immediate release) formulation, because of the oscillation in blood pressure and concurrent surges in sympathetic reflex activity within each dosage interval. Ca -channel blockers are versatile drugs with proven efficacy in all types of patients (Kiowski et al., 1985). They seem to be especially efficacious in low-renin hypertension. Compared with other
2+

classes of antihypertensive agents, there is a greater frequency of achieving blood pressure control with Ca2+-channel blockers as monotherapy in elderly subjects and in African Americans, population groups in which the low renin status is more prevalent. Long-acting dihydropyridine Ca2+-channel blockers have been found to reduce cardiovascular mortality in older patients (Staessen et al., 1997). The efficacy of Ca 2+-channel blockers is enhanced by the concomitant use of an angiotensin converting enzyme inhibitor, methyldopa, or -adrenergic receptor antagonists. When -adrenergic receptor antagonists are administered concurrently, the preferred Ca2+-channel blocker would be one from the group that is relatively vasoselective (e.g., amlodipine, isradipine, nicardipine). Diuretics also may enhance the efficacy of Ca2+-channel blockers, but the data have not been consistent. Significant drugdrug interactions may be encountered when Ca 2+-channel blockers are used to treat hypertension. Verapamil blocks the drug transporter, P-glycoprotein. Both the renal and hepatic disposition of digoxin occur via this transporter. Accordingly, verapamil inhibits the elimination of digoxin and other drugs that are cleared from the body by the P-glycoprotein (see Chapter 1: Pharmacokinetics: The Dynamics of Drug Absorption, Distribution, and Elimination) (Pedersen et al., 1981). When used with quinidine, Ca2+-channel blockers may cause excessive hypotension, particularly in patients with idiopathic hypertrophic subaortic stenosis. Ca2+-channel blockers should not be used in patients with SA or AV nodal abnormalities or in patients with overt congestive heart failure. These drugs usually are safe, however, in hypertensive patients with asthma, hyperlipidemia, diabetes mellitus, and renal dysfunction. Unlike -adrenergic receptor antagonists, Ca2+-channel blockers do not alter exercise tolerance; nor do they alter plasma concentrations of lipids, uric acid, or electrolytes. Angiotensin Converting Enzyme Inhibitors Angiotensin II is an important regulator of cardiovascular function (see Chapter 31: Renin and Angiotensin). The ability to reduce levels of angiotensin II with orally effective inhibitors of angiotensin converting enzyme (ACE) represents an important advance in the treatment of hypertension. Captopril (CAPOTEN) was the first such agent to be developed for the treatment of hypertension. Since then, enalapril (VASOTEC), lisinopril (PRINIVIL), quinapril (ACCUPRIL ), ramipril (ALTACE), benazepril (LOTENSIN), moexipril (UNIVASC), fosinopril (MONOPRIL), trandolapril (MAVIK), and perindopril (ACEON) also have become available. These drugs have proven to be very useful for the treatment of hypertension because of their efficacy and their very favorable profile of side effects, which enhances compliance. The angiotensin converting enzyme inhibitors appear to confer a special advantage in the treatment of patients with diabetes, slowing the development of diabetic glomerulopathy. They also have been shown to be effective in slowing the progression of other forms of chronic renal disease, such as glomerulosclerosis, and many of these patients also have hypertension. An angiotensin converting enzyme inhibitor is probably the preferred initial agent in the treatment of hypertensive patients with left ventricular hypertrophy. Patients with hypertension and ischemic heart disease are candidates for treatment with angiotensin converting enzyme inhibitors; this includes treatment in the immediate postmyocardial infarction period which has been shown to lead to improved ventricular function and reduced morbidity and mortality (see also Chapter 34: Pharmacological Treatment of Heart Failure). The endocrine consequences of inhibiting the biosynthesis of angiotensin II are of importance in a number of facets of hypertension treatment. Because angiotensin converting enzyme inhibitors blunt the normal aldosterone response to Na+ loss, the normal role of aldosterone to oppose diuretic-

induced natriuresis is diminished. Thus, angiotensin converting enzyme inhibitors enhance the efficacy of diuretic drugs. This means that even very small doses of diuretics may substantially improve the antihypertensive efficacy of angiotensin converting enzyme inhibitors; and on the other end of the spectrum, the use of high doses of diuretics together with angiotensin converting enzyme inhibitors may lead to excessive reduction in blood pressure and to Na+ loss in some patients. The attenuation of aldosterone production by the angiotensin converting enzyme inhibitors also influences K+ homeostasis. There is only a very small and clinically unimportant rise in serum K+ when angiotensin converting inhibitors are used alone in patients with normal renal function. However, substantial retention of K+ can occur in some patients with renal insufficiency. Furthermore, the potential for developing hyperkalemia should be considered when angiotensin converting enzyme inhibitors are used with other drugs that can cause K+ retention; these include the K+ -sparing diuretics (amiloride, triamterene, spironolactone), nonsteroidal antiinflammatory drugs, K+ supplements, and -adrenergic receptor antagonists. There are several cautions in the use of angiotensin converting enzyme inhibitors in patients with hypertension. Angioedema is an infrequent but serious and potentially fatal adverse effect of all of the angiotensin converting enzyme inhibitors. Thus, patients starting treatment with these drugs should be explicitly warned to discontinue their use with the advent of any signs of angioedema. The angiotensin converting enzyme inhibitors should not be used during pregnancy, a fact that should be communicated to patients of childbearing age. In most patients there is no appreciable change in glomerular filtration rate following the administration of a converting enzyme inhibitor. However, in renovascular hypertension, the glomerular filtration rate is maintained as the result of increased resistance in the postglomerular arteriole caused by angiotensin II. Accordingly, in patients with bilateral renal artery stenosis or stenosis in a sole kidney, the administration of a converting enzyme inhibitor will reduce the filtration fraction and cause a substantial reduction in glomerular filtration rate. Converting enzyme inhibitors lower the blood pressure to some extent in most patients with hypertension. Following the initial dose of a converting enzyme inhibitor, there may be a considerable fall in blood pressure in some patients; this response to the initial dose is a function of pretreatment plasma renin activity. The potential for a large initial drop in blood pressure is the reason for using a low dose for initiating therapy. On continuing treatment, there usually is a progressive fall in blood pressure that in most patients does not reach a maximum for about one week. The level of blood pressure seen during chronic treatment is not strongly correlated with the level of pretreatment plasma renin activity. Young and middle-aged Caucasian patients have a higher probability of responding to the angiotensin converting enzyme inhibitors. Elderly AfricanAmerican patients as a group are more resistant to the hypotensive effect of these drugs, but concurrent use of a diuretic in low doses overcomes this relative resistance. These drugs are discussed in detail in Chapter 31: Renin and Angiotensin. Angiotensin IIReceptor Antagonists The importance of angiotensin II in regulating cardiovascular function has led to the development of nonpeptide antagonists of the angiotensin II receptor for clinical use. Losartan (COZAAR ), candesartan (ATACAND), irbesartan (AVAPRO ), valsartan (DIOVAN), telmisartan (MICARDIS), and eprosartan (TEVETEN) have been approved for the treatment of hypertension. By preventing effects of angiotensin II, these agents relax smooth muscle and thereby promote vasodilation, increase renal salt and water excretion, reduce plasma volume, and decrease cellular hypertrophy. Angiotensin IIreceptor antagonists also theoretically overcome some of the disadvantages of ACE

inhibitors, which not only prevent conversion of angiotensin I to angiotensin II but also prevent ACE-mediated degradation of bradykinin and substance P. Cough, an adverse effect of ACE inhibitors, has not been associated with angiotensin IIreceptor antagonists. Angioedema occurs rarely. Two distinct subtypes of angiotensin II receptors have been cloned, designated as type 1 (AT1) and type 2 (AT2). The AT1 angiotensin II receptor subtype is located predominantly in vascular and myocardial tissue and also in brain, kidney, and adrenal glomerulosa cells, which secrete aldosterone (see Chapter 31: Renin and Angiotensin). The AT2 subtype of angiotensin II receptor is found in the adrenal medulla, kidney, and in the CNS, and may play a role in vascular development (Horiuchi et al., 1999). Because the AT 1 receptor mediates feedback inhibition of renin release, renin and angiotensin II concentrations are increased during AT1-receptor antagonism. The clinical consequences of increased angiotensin II effects on an uninhibited AT2 receptor are unknown; however, emerging data suggest that the AT2 receptor may elicit antigrowth and antiproliferative responses. Adverse Effects and Precautions The adverse effects of AT1-receptor antagonists may be considered in the context of those known to be associated with the ACE inhibitors. ACE inhibitors cause problems of two major types, those related to a diminished level of angiotensin II and those due to molecular actions independent of abrogating the function of angiotensin II. Adverse effects of ACE inhibitors that result from inhibiting angiotensin II-related functions (see above) occur also with AT1-receptor antagonists. These include hypotension, hyperkalemia, and reduced renal function, including that associated with bilateral renal artery stenosis and stenosis in the artery of a solitary kidney. Hypotension is most likely to occur in patients in whom the blood pressure is highly dependent on angiotensin II, including those with volume depletion (e.g., with diuretics), renovascular hypertension, cardiac failure, and cirrhosis; in such patients initiation of treatment with low doses and attention to blood volume is essential. Hyperkalemia will occur only in conjunction with other factors that alter K+ homeostasis, such as renal insufficiency, ingestion of excess K+, and the use of drugs that promote K + retention. In contrast with ACE inhibitors, the AT1-receptor antagonists do not cause cough. Angioedema has been reported, but it is not clear whether or not the rate of angioedema in patients taking the AT1receptor antagonists is any higher than that in the general population. Hepatic dysfunction has been reported with the AT1-receptor antagonists. AT1-receptor antagonists should not be administered during the second or third trimester of pregnancy and should be discontinued as soon as pregnancy is detected. Although it is not yet known whether or not AT1-receptor antagonists are secreted in human breast milk, significant amounts are detected in the milk of animals; consequently, AT1-receptor antagonists should not be administered to patients who are breast-feeding. Therapeutic Uses When given in adequate doses, the AT1-receptor antagonists appear to be as effective as ACE inhibitors in the treatment of hypertension. As with ACE inhibitors, these drugs may be less effective in African-American and low-renin patients. The full effect of AT1-receptor antagonists on blood pressure typically is not observed until 3 to 6

weeks after the initiation of therapy. If blood pressure is not controlled by an AT 1-receptor antagonist alone, a low dose of a hydrochlorothiazide or other diuretic may be added. In several randomized, double-blind studies of patients with mild to severe hypertension, the addition of hydrochlorothiazide to an AT1-receptor antagonist produced significant additional reductions in blood pressure in patients who demonstrated an insufficient response to hydrochlorothiazide alone. A smaller initial dosage is preferred for patients who have already received diuretics and therefore have an intravascular volume depletion, and for other patients whose blood pressure is highly dependent on angiotensin II. Ongoing clinical trials should shed light on the relative efficacy of ACE inhibitors and AT1-receptor antagonists in patients with diabetic nephropathy, coronary artery disease, and left ventricular dysfunction (Pitt et al., 1999a). Given the different mechanisms by which they act, there is no assurance that the effects of ACE inhibitors and antagonists of the AT1 receptor will be equivalent. Nonpharmacological Therapy of Hypertension Nonpharmacological approaches to the reduction of blood pressure generally are advisable as the initial approach to treatment of patients with diastolic blood pressures in the range of 90 to 95 mm Hg. Further, these approaches will augment the effectiveness of pharmacological therapy in patients with higher levels of blood pressure. Also, for patients with diastolic blood pressures in the range of 85 to 90 mm Hg, the epidemiological data on cardiovascular risks support the institution of nonpharmacological therapy. To maintain compliance with a therapeutic regimen, the intervention should not lessen the quality of life. All drugs have side effects. If minor alterations of normal activity or diet can reduce blood pressure to a satisfactory level, the complications of drug therapy can be avoided. In addition, nonpharmacological methods to lower blood pressure allow the patient to participate actively in the management of his or her disease. Reduction of weight, restriction of salt, and moderation in the use of alcohol may reduce blood pressure and improve the efficacy of drug treatment. In addition, regular isotonic exercise also lowers blood pressure in hypertensive patients. Smoking per se does not cause hypertension. However, smokers do have a higher incidence of malignant hypertension (Isles et al., 1979), and smoking is a major risk factor for coronary heart disease. Hypertensive patients have an exceptionally great incentive to stop smoking. Consumption of caffeine can raise blood pressure and elevate plasma concentrations of norepinephrine, but longterm consumption of caffeine causes tolerance to these effects and has not been associated with the development of hypertension. An increased intake of Ca2+ has been reported by some investigators to lower blood pressure. The mechanism of this effect is not understood, but suppression of the secretion of parathyroid hormone apparently is involved. However, supplemental Ca2+ does not lower blood pressure when populations of hypertensive subjects are studied. Although it is possible that there are some hypertensive patients who have a hypotensive response to Ca2+, there is no easy way to identify such individuals. Supplemental use of Ca 2+ for this purpose cannot be recommended at the present time (Kaplan, 1988). Reduction of Body Weight Obesity and hypertension are closely associated, and the degree of obesity is positively correlated with the incidence of hypertension. Obese hypertensives may lower their blood pressure by losing weight regardless of a change in salt consumption (Maxwell et al., 1984). The mechanism by which obesity causes hypertension is unclear, but increased secretion of insulin in obesity could result in insulin-mediated enhancement of renal tubular reabsorption of Na + and an expansion of extracellular volume. Obesity also is associated with increased activity of the sympathetic nervous

system; this is reversed by weight loss. Maintenance of weight loss is difficult for many. A combination of aerobic physical exercise and dietary counseling may enhance compliance. Sodium Restriction Severe restriction of salt will lower the blood pressure in most hospitalized hypertensive patients; this treatment method was advocated prior to the development of effective antihypertensive drugs (Kempner, 1948). However, severe salt restriction is not practical from a standpoint of compliance. Several studies have shown that moderate restriction of salt intake to approximately 5 g per day (2 g Na+) will, on average, lower blood pressure by 12 mm Hg systolic and 6 mm Hg diastolic. The higher the initial blood pressure, the greater the response. In addition, subjects over 40 years of age are more responsive to the hypotensive effect of moderate restriction of salt (Grobbee and Hofman, 1986). Even though not all hypertensive patients respond to restriction of salt, this intervention is benign and can easily be advised as an initial approach in all patients with mild hypertension. An additional benefit of salt restriction is improved responsiveness to some antihypertensive drugs. Alcohol Restriction Consumption of alcohol can raise blood pressure, but it is unclear how much alcohol must be consumed to observe this effect (MacMahon et al., 1984). Heavy consumption of alcohol increases the risk of cerebrovascular accidents but not coronary heart disease (Kagan et al., 1985). In fact, small amounts of ethanol have been found to protect against the development of coronary artery disease. The mechanism by which alcohol raises blood pressure is unknown, but it may involve increased transport of Ca2+ into vascular smooth muscle cells. Excessive intake of alcohol also may result in poor compliance with antihypertensive regimens. All hypertensive patients should be advised to restrict consumption of ethanol to no more than 30 ml per day. Physical Exercise Increased physical activity lowers rates of cardiovascular disease in men (Paffenbarger et al., 1986). It is not known if this beneficial effect is secondary to an antihypertensive response to exercise. Lack of physical activity is associated with a higher incidence of hypertension (Blair et al., 1984). Although consistent changes in blood pressure are not always observed, meticulously controlled studies have demonstrated that regular isotonic exercise reduces both systolic and diastolic blood pressures by approximately 10 mm Hg (Nelson et al., 1986). The mechanism by which exercise can lower blood pressure is not clear, but several hemodynamic and humoral changes have been documented. Regular isotonic exercise reduces blood volume and plasma catecholamines and elevates plasma concentrations of atrial natriuretic factor. The beneficial effect of exercise can occur in subjects who demonstrate no change in body weight or salt intake during the training period. Relaxation and Biofeedback Therapy The fact that long-term stressful stimuli can cause sustained hypertension in animals has given credence to the possibility that relaxation therapy will lower blood pressure in some hypertensive patients. A few studies have generated positive results but, in general, relaxation therapy has inconsistent and modest effects on blood pressure (Jacob et al., 1986). In addition, the long-term efficacy of such treatment has been difficult to demonstrate, presumably in part because patients must be highly motivated to respond to relaxation and biofeedback therapy. Only those few patients with mild hypertension who wish to use this method should be encouraged to try, and these patients

should be closely followed and receive pharmacological treatment if necessary. Potassium Therapy There is a positive correlation between total body Na+ and blood pressure and a negative correlation between total body K+ and blood pressure in hypertensive patients (Lever et al., 1981). In addition, dietary intake, plasma concentrations, and urinary excretion of K+ are reduced in various populations of hypertensive subjects. Increased intake of K+ might reduce blood pressure by increasing excretion of Na+, suppressing renin secretion, causing arteriolar dilation (possibly by stimulating Na+ , K+ -ATPase activity and decreasing intracellular concentrations of Ca 2+), and impairing responsiveness to endogenous vasoconstrictors. In hypertensive rats, supplementation with K+ decreases blood pressure and reduces the incidence of stroke, irrespective of blood pressure (Tobian, 1986). In mildly hypertensive patients, oral K+ supplements of 48 mmol per day reduce both systolic and diastolic blood pressure (Siani et al., 1987). Supplementation with K+ also may protect against ventricular ectopy and stroke (Khaw and Barrett-Connor, 1987). Based on all of these data, it seems prudent to use a high-K+ diet in conjunction with moderate restriction of Na+ in the nonpharmacological treatment of hypertension. However, a high-K+ diet should not be recommended for patients on angiotensin converting enzyme inhibitors. Prospectus The most anticipated development in antihypertensive therapy is the new knowledge expected from clinical trials comparing the effectiveness of drugs on the important endpoints of morbidity and mortality. One clinical trial addressing these endpoints has been launched by the National Heart, Lung, and Blood Institute (Davis et al., 1996). It is the Antihypertensive and Lipid Lowering treatment to prevent Heart Attack Trial (ALLHAT) and is comparing the outcomes of treatment with a thiazide-class diuretic (chlorthalidone), an angiotensin converting enzyme inhibitor (lisinopril), a Ca2+-channel blocker (amlodipine), and an 1-adrenergic receptor antagonist (doxazosin). This trial is evaluating the effects of these drugs in patients over the age of 55 who are at high risk for vascular occlusive events. Concurrently, the benefit of a cholesterol-lowering drug, pravastatin, is being assessed in the same population. The Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) study is comparing the AT1-receptor antagonist losartan and adrenergic receptor blocker atenolol on cardiovascular mortality and morbidity in patients, aged 55 to 88 years, with hypertension and left ventricular hypertrophy (Dahlf et al., 1998). The AfricanAmerican Study of Kidney Disease (AASK) will determine the efficacy of two different levels of blood pressure control and three different antihypertensive regimens on the progression of renal disease in African Americans with hypertensive nephropathy (Wright et al., 1996). A growing recognition of the contribution of the renin-angiotensin-aldosterone system (RAAS) to the development and progression of hypertensive end-organ damage promises to bring a number of studies examining new strategies for interrupting the RAAS. For example, studies comparing the efficacy of ACE inhibition, AT1-receptor antagonism, and the combination of ACE inhibition and AT1-receptor antagonism are ongoing. The finding that the addition of an aldosterone antagonist decreases mortality in patients with congestive heart failure treated with an ACE inhibitor (Pitt et al., 1999b) has sparked a renewed interest in aldosterone antagonists in the treatment of hypertension. These studies, if they are successful, could profoundly influence the approach to treating hypertension. For further discussion of hypertension, see Chapter 230 in Harrison's Principles of Internal Medicine, 16th ed., McGraw-Hill, New York, 2005.

Chapter 34. Pharmacological Treatment of Heart Failure

Overview Heart failure is one of the most common causes of death and disability in industrialized nations and is among the syndromes most commonly encountered in clinical practice. Over 4.6 million patients in the United States alone carry this diagnosis, and it is the cause of death in several hundred thousand patients each year (American Heart Association, 1999). The diagnosis of heart failure carries a risk of mortality comparable to that of the major malignancies. Patients with newly diagnosed heart failure have an average five-year survival of only 35%. In the past 20 years, advances in understanding the pathophysiology of heart failure and new developments in pharmacotherapy have added substantially to the physician's ability to alleviate the symptoms of this disease and slow the natural progression of the underlying myocardial process. This chapter deals with drug therapy of heart failure due to systolic and/or diastolic ventricular dysfunction. Systolic dysfunction due to idiopathic dilated or ischemic cardiomyopathies usually is characterized by large, dilated ventricular chambers. Diastolic dysfunction due to long-standing hypertension, stenotic valvular disease, or a primary hypertrophic cardiomyopathy generally leads to thickened, poorly compliant ventricular walls with small ventricular volumes. In reality, many patients exhibit abnormal hemodynamics comprising significant degrees of both systolic and diastolic dysfunction. Treatment must therefore be tailored to the underlying pathophysiological process in the individual patient. The basic pharmacology of many of the classes of drugs described in this chapter is discussed in detail in other chapters and cross-referenced in the text. Therefore, the pharmacology of these agents is discussed here only in the context of the treatment of heart failure. The main elements in the pharmacotherapy of heart failure are angiotensin converting enzyme inhibitors (seeChapter 31: Renin and Angiotensin) and -adrenergic receptor antagonists (seeChapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists). Accordingly, their use is discussed in detail in this chapter. Goals of Therapy Relief of Symptoms A primary goal in the treatment of heart failure is the alleviation of symptoms, which, in turn, are a direct result of the underlying hemodynamic disorder. Intravascular volume expansion and elevated ventricular filling pressures result in systemic and pulmonary venous hypertension, which causes dyspnea on exertion and orthopnea. Reduced cardiac output results in fatigue and decreased exercise capacity. In the short term, symptomatic treatment is directed at improving hemodynamic function through the use of drugs that increase cardiac output and reduce ventricular filling pressures. In the patient hospitalized because of severe symptoms of heart failure, rapid treatment may include the use of intravenous diuretics, positive inotropic agents (e.g., -adrenergic receptor agonists or phosphodiesterase inhibitors) and vasodilators (e.g., nitroprusside or nitroglycerin). In the ambulatory patient in whom there is less urgency, similar goals are approached through the use of oral diuretics, digitalis, and vasodilators [e.g., angiotensin converting enzyme (ACE) inhibitors]. Myocardial Remodeling

Even in the absence of recurrent damage to the heart, the severity of the underlying myocardial dysfunction often is progressive. This is due to ventricular "remodeling," a process that results in progressive maladaptive changes in the structure and function of the ventricle (Cohn, 1995). Therefore, a second major goal of therapy is to slow or prevent the progression of myocardial remodeling. Initially, myocardial dysfunction causes intravascular volume expansion and the activation of neurohormonal systems, particularly the sympathetic nervous and reninangiotensin systems. These primitive, compensatory responses defend the perfusion of vital organs by increasing left ventricular preload, stimulating myocardial contractility, and increasing arterial tone. However, with time each plays a role in the pathophysiology of the disease by promoting the progression of the underlying myocardial dysfunction. Expansion of the intravascular volume results in elevated ventricular filling pressures which increase ventricular wall stresses. Neurohormonal activation causes arterial and venous constriction which also leads to increased ventricular wall stresses. In addition, some neurohormones (e.g., norepinephrine and angiotensin) may act directly on the myocardium to promote remodeling by causing myocyte apoptosis, abnormal gene expression, and/or alterations in the extracellular matrix (Colucci and Braunwald, 2000). Drugs that reduce ventricular wall stresses (e.g., vasodilators) and/or inhibit the reninangiotensin system (e.g., angiotensin converting enzyme inhibitors) or the sympathetic nervous system (e.g., adrenergic receptor antagonists) have been found to decrease pathological ventricular remodeling, and are therefore a mainstay in the long-term treatment for heart failure. Some agents that slow progression also exert an immediate beneficial effect on hemodynamic function and symptoms (e.g., vasodilators and angiotensin converting enzyme inhibitors). Other drugs that slow the progression of myocardial remodeling (e.g., -adrenergic receptor antagonists), may actually exert an adverse effect on hemodynamic function and can worsen symptoms in the short term. Figure 34 1 provides an overview of the pathophysiological mechanisms of heart failure and the sites of action of the major drug classes used in treatment. Figure 341. Pathophysiological Mechanisms of Heart Failure and Major Sites of Drug Action. Heart failure is accompanied by compensatory neurohormonal responses including activation of the sympathetic nervous and reninangiotensin systems. Although these responses initially help to maintain cardiovascular function by increasing ventricular preload and systemic vascular tone, with time they contribute to the progression of myocardial failure. Increased ventricular afterload, due to systemic vasoconstriction and chamber dilation, causes a depression in systolic function. In addition, increased afterload and the direct effects of angiotensin and norepinephrine on the ventricular myocardium cause pathological remodeling characterized by progressive chamber dilation and loss of contractile function. The figure illustrates several mechanisms that appear to play an important role in the pathophysiology of heart failure, and the sites of action of pharmacological therapies that have been shown to be of clinical value. ACE, angiotensin converting enzyme; ARB, angiotensin-receptor blocker.

The pharmacotherapeutic management of patients with heart failure is described in two sections. The first describes the use of oral drugs in ambulatory patients. The second describes intravenously administered agents primarily used for the treatment of hospitalized patients Oral Drugs for the Management of Ambulatory Heart Failure Diuretics For several decades, and especially following the introduction of "loop" diuretics, this group of drugs has played a central role in the pharmacological management of the "congestive" symptoms in heart failure. Diuretics are discussed in detail in Chapter 29: Diuretics. Therefore, only those aspects of their pharmacology that are relevant to the treatment of heart failure will be dealt with in this chapter. The importance of these drugs is due to the central role of the kidney as the target organ for many of the hemodynamic, hormonal, and autonomic nervous system changes that occur in response to a failing myocardium. The net effect of these changes is the retention of salt and water and expansion of the extracellular fluid volume, which serves in the short run to sustain cardiac output and tissue perfusion by allowing the heart to operate higher on its ventricular function (i.e., FrankStarling) curve (Figure 342). However, this response incurs the cost of higher end-diastolic filling pressures and increasing ventricular chamber dimensions and wall stress, which eventually limit any further increase in cardiac output and also result in pulmonary venous congestion and peripheral edema.

Figure 342. Hemodynamic Responses to Pharmacological Interventions in Heart Failure. The relationships between diastolic filling pressure (or preload) and stroke volume (or ventricular performance) are illustrated for a normal heart (black line) and for a patient with heart failure due to predominant systolic dysfunction (blue line). Note that positive inotropic agents, such as cardiac glycosides or dobutamine, move patients to a higher ventricular function curve (dashed line), resulting in greater cardiac work for a given level of ventricular filling pressure. Vasodilators, such as angiotensin converting enzyme inhibitors or nitroprusside, also move patients to improved ventricular function curves while reducing cardiac filling pressures. Diuretics improve symptoms of congestive heart failure by moving patients to lower cardiac filling pressures along the same ventricular function curve. Combinations of drugs often will yield additive effects on hemodynamics.

Diuretics act to reduce extracellular fluid volume and ventricular filling pressures (or "preload") but usually do not cause a clinically important reduction in cardiac output, particularly in patients with advanced heart failure who have an increased left ventricular filling pressure, unless there has been a profound and sustained natriuresis that results in a rapid decline in intravascular volume. Despite the clear efficacy of diuretics in controlling congestive symptoms and improving exercise capacity, it is worth noting that, with the exception of low-dose spironolactone, the use of diuretics has not been demonstrated to improve survival in heart failure. Indeed, monotherapy with a diuretic may cause increased neurohormonal activation due to volume depletion, with potentially deleterious effects on the progression of heart failure. For this reason, it may be preferable to avoid the use of diuretics in the subset of patients who have mild heart failure without evidence of fluid retention. Dietary Sodium Restriction All patients with clinically significant ventricular dysfunction, regardless of whether or not they are symptomatic, should be advised to limit dietary intake of NaCl. Most patients will tolerate moderate reductions in salt intake (2 to 3 g/day total intake). More stringent salt restriction is seldom necessary and may be counterproductive in many heart-failure patients as it may lead to hyponatremia, hypokalemia, and a metabolic alkalosis due to chloride depletion when combined with loop diuretics, as well as loss of lean body mass due to reduced appetite.

Loop Diuretics Of the loop diuretics currently available, only furosemide (LASIX ), bumetanide (BUMEX), and torsemide (DEMADEX) are indicated in the treatment of most patients with heart failure. Due to the increased risk of ototoxicity, ethacrynic acid should be reserved for patients who are allergic to sulfonamides or who have developed interstitial nephritis on alternative drugs. Loop diuretics act to inhibit a specific ion transport protein, the Na+K+2Cl symporter (seeChapter 29: Diuretics; Gamba et al., 1994) on the apical membrane of renal epithelial cells in the ascending limb of Henle's loop. They rely for their efficacy on adequate renal plasma flow and proximal tubular secretion to deliver the diuretics to their site of action. These drugs also reduce the tonicity of the medullary interstitium by preventing the resorption of solute in excess of water in the thick ascending limb of the loop of Henle, and this may contribute to the development of hyponatremia in heart-failure patients. The increased delivery of Na+ and fluid to distal nephron segments also markedly enhances K+ secretion, particularly in the presence of elevated aldosterone levels, as is typically the case in heart failure. The bioavailability of orally administered furosemide ranges from 40% to 70%, and adjustments in dosage may be required before it is deemed ineffective. In contrast, bumetanide and torsemide have oral bioavailabilities of more than 80% and provide more consistent absorption, albeit at a considerably greater cost. Furosemide and bumetanide are short-acting drugs. Avid renal Na+ retention by all nephron segments following a decline in renal tubular diuretic levels can limit or prevent a negative Na+ balance. In many patients with heart failure, this necessitates the use of two or more daily doses of these diuretics to induce and sustain a negative salt balance. This is an acceptable strategy for outpatient management of heart failure, provided that there is adequate monitoring of daily weights and blood electrolyte levels. Thiazide Diuretics The principal site of action of the thiazide diuretics is now known to be the Na+Cl cotransporter (seeChapter 29: Diuretics) present in renal tubular epithelial cells in the distal convoluted tubule. The thiazide diuretics generally are useful as single drugs for the therapy of volume retention only in patients with relatively mild heart failure, largely because their site of action in the distal nephron permits rapid adjustment of water and solute absorption by other, more proximal nephron segments. Thiazide diuretics also are ineffective at glomerular filtration rates below 30 ml/minute. However, these drugs exhibit true synergism with loop diuretics (i.e., a natriuresis that is greater than the sum of either class of drugs given individually). This is useful when patients become refractory to loop diuretics (see"Diuretic Resistance in Heart Failure," below). K+-Sparing Diuretics K+-sparing diuretics (seeChapter 29: Diuretics) are divided into those agents that inhibit apical membrane Na+ conductance channels in epithelial cells of the collecting duct (e.g., amiloride, triamterene) and aldosterone antagonists that also have their principal pharmacological effect in the collecting duct (e.g., spironolactone, canrenone). Although these agents generally are not effective as diuretic agents when used alone, they may be useful in limiting renal K+ and Mg2+ wasting and/or in augmenting the response to other classes of diuretics. As discussed below, there is evidence that a low dose of spironolactone may improve survival in patients with advanced symptoms of heart failure, apparently via a mechanism independent of diuresis (Pitt et al., 1999). Use of Diuretics in Clinical Practice

The majority of patients with heart failure will require the chronic administration of a loop diuretic to maintain euvolemia. In patients with fluid retention, furosemide typically is started at a dose of 40 mg once or twice per day, and the dosage is increased until an adequate diuresis (increased urine output and weight loss of 0.5 to 1.5 kg daily) is achieved. A larger initial dose may be required if there is significant renal impairment, or in severe heart failure. Serum electrolytes and renal function should be monitored, especially if there is preexisting renal insufficiency or a brisk diuresis is desirable in a severely symptomatic patient. Once fluid retention has resolved, the dose of diuretic should be reduced to the minimum necessary to maintain euvolemia. Electrolyte abnormalities or worsening azotemia may supervene before euvolemia is achieved. Hypokalemia may occur and may be corrected by potassium supplementation or addition of a potassium-sparing diuretic. In general, diuresis should be slowed only if azotemia or renal impairment become progressive or the patient is symptomatic. Diuretic Resistance in Heart Failure The response to diuretics frequently is impaired in heart failure. Following prolonged administration of a loop diuretic, a process of adaptation occurs in which there is a compensatory increase in sodium reabsorption in the distal nephron and blunting of net sodium and water loss. Furthermore, while there may be a brisk response following a single dose of diuretic, a compensatory increase in sodium reabsorption for the remaining part of the 24-hour period may prevent effective diuresis. Patients who have impaired renal function typically require higher doses of diuretic to ensure adequate delivery of the diuretic to its site of action. A poor response to diuretics also may be due to edema and decreased motility of the bowel wall as well as to reduced splanchnic blood flow. This may cause slowed absorption and a delay in the peak effect of orally administered diuretics, although the total amount absorbed usually is unchanged. The more common causes of diuretic resistance are listed in Table 341. It often is difficult to determine clinically whether an increasing diuretic requirement is due to intravascular volume depletion following aggressive diuretic and vasodilator therapy or to a decrease in cardiac output and blood pressure due to the underlying cardiac failure. Invasive monitoring to determine the left ventricular filling pressure may be required to make this distinction. However, a more marked decline in urea clearance than in creatinine clearance (resulting in an increase in the BUN to creatinine ratio) suggests intravascular volume depletion. Vasodilators commonly employed as "unloading" agents in heart failure may reduce renal blood flow despite an increase in cardiac output, thereby reducing diuretic effectiveness. Also, since some patients with heart failure also may have renal arterial atherosclerotic disease, vasodilator therapy may lower renal perfusion pressure below that necessary to maintain normal autoregulation and glomerular filtration. The caveats about vasodilators also apply to ACE inhibitors and to angiotensin II type 1 receptor (i.e., AT1 receptor) antagonists (seeChapter 31: Renin and Angiotensin). However, because of the unique role of angiotensin II as an intrarenal signaling autocoid, these drugs can either augment the effectiveness of diuretics by mechanisms independent of their ability to reduce systemic vascular resistance or diminish their effectiveness by reducing the transglomerular perfusion pressure to the point that the glomerular filtration rate declines abruptly. The latter response is observed most commonly in patients with decreased renal arterial perfusion pressure, due either to renal artery stenosis and/or a limited cardiac output, for whom a high angiotensin IImediated glomerular efferent arteriolar tone is necessary to maintain glomerular filtration. This cause of diuretic resistance generally is accompanied by a decline in creatinine clearance and should be distinguished from the more modest, limited rise in serum creatinine levels and improved responsiveness to diuretics that commonly accompany ACE-inhibitor administration.

Diuretic resistance due primarily to poor cardiac function generally improves when cardiac output is increased by the use of positive inotropic agents (e.g., dobutamine) or vasodilators. Decreased responsiveness to loop diuretics in patients otherwise receiving optimal medical management should be managed initially by increasing the frequency of doses and by more stringent dietary salt restriction. If this is ineffective, a thiazide diuretic (e.g., hydrochlorothiazide or metolazone) administered with the loop diuretic often is effective (Ellison, 1991). However, this diuretic combination can result in an unpredictable and at times excessive diuresis, which can cause intravascular volume depletion and renal K+ wasting; the combination therefore should be used cautiously. Spironolactone also may be effective in these patients when combined with a loop diuretic. For a detailed discussion on the subject of diuretic resistance, the reader is referred to a recent comprehensive review (Ellison, 1999). Metabolic Consequences of Diuretic Therapy The side effects of diuretic agents are discussed in Chapter 29: Diuretics and in recent overviews of diuretic use (e.g., Brater, 1998). With regard to diuretic use in heart failure, the most important adverse sequelae of diuretics are electrolyte abnormalities, including hyponatremia, hypokalemia, and hypochloremic metabolic alkalosis. The clinical importance, or even the existence, of significant Mg2+ deficiency with chronic diuretic use remains controversial (Bigger, 1994; Davies and Fraser, 1993). Both hypokalemia and renal Mg2+ wasting can be limited by administration of oral KCl supplements or a K+-sparing diuretic. Aldosterone Antagonists One of the principal features of heart failure is marked activation of the renin-angiotensinaldosterone system with elevation of the plasma aldosterone concentration to as much as 20 times the normal level. As mentioned above, when used alone spironolactone exerts only a very weak diuretic effect in patients with heart failure. However, aldosterone has a range of biological effects in addition to salt retention (seeTable 342), and it has been suggested that antagonism of aldosterone, per se, may be beneficial in patients with heart failure. For a review of the subject of aldosterone and spironolactone in heart failure, see the recent review by Struthers (1999). Clinical Use of Spironolactone in Heart Failure The Randomized Aldactone Evaluation Study (RALES) randomized patients with moderate to severe heart failure [New York Heart Association (NYHA) class III to IV] to receive 25 mg daily of spironolactone or placebo in addition to conventional therapy, which included, in the large majority of patients, an ACE inhibitor (Pitt, 1999). Patients with serum creatinine concentrations greater than 2.5 mg/dl (221 M) were excluded from the study, and only a very small number of patients received 50 mg of spironolactone daily. Patients randomized to spironolactone had a significant 30% reduction in mortality and hospitalization for heart failure. The decrease in the risk of death was due to reductions in both progressive heart failure and sudden cardiac death and, notably, was achieved in the apparent absence of a measurable diuretic effect. Treatment was generally well tolerated, and although 10% of men in the spironolactone group developed gynecomastia, withdrawal of treatment was necessary in less than 2%. Severe hyperkalemia occurred in only 2% of patients on spironolactone, and there were no clinically important effects on renal function. The RALES trial suggests that the beneficial effects of spironolactone are additive to those of ACE inhibitors, and its use should be considered in patients with NYHA class III to IV heart failure. However, caution should be exercised in its use when significant renal impairment is present. Treatment should be initiated at a dose of 12.5 or 25 mg daily. Higher doses should be avoided, as

they may lead to hyperkalemia, particularly in patients receiving an ACE inhibitor. Serum potassium levels and electrolytes should be checked after initiation of treatment, and vigilance should be exercised for potential drug interactions and medical disorders that may cause elevations in serum potassium concentration (e.g., potassium supplements, ACE inhibitors, and worsening renal function). Vasodilators Although several classes of drugs exhibit vasodilator activity and may improve symptoms in heart failure (Table 343), only ACE inhibitors and the hydralazineisosorbide dinitrate combination have been shown to improve survival in prospective randomized trials. Some classes of drugs, such as 1-adrenergic antagonists, have no demonstrated effect on mortality, while other agents, such as prostacyclin and flosequinan, appear to decrease long-term survival. The basic and clinical pharmacology of most of the vasodilators discussed in this chapter are considered in more detail in Chapters 32: Drugs Used for the Treatment of Myocardial Ischemia and 33: Antihypertensive Agents and the Drug Therapy of Hypertension. The rationale for the use of drugs with vasodilatory activity in heart failure grew out of experience with parenteral phentolamine and nitroprusside in patients with severe heart failure. Cohn and Franciosa, in an influential article in 1977, reviewed the evidence supporting this approach. Studies of ACE inhibitors in the following decade showed that these drugs generally were well tolerated and effective in improving symptoms in heart failure, while two randomized, prospective trials verified the effectiveness of an isosorbide dinitratehydralazine combination (Cohn et al., 1986) and enalapril (CONSENSUS, 1987) in reducing mortality in patients with heart failure. Subsequent clinical trials have reinforced the results of these two trials and have provided evidence supporting expanded indications for the use of ACE inhibitors to patients with ventricular dysfunction but without overt symptoms of heart failure (see below). Principles of Vasodilator Therapy The principles of vasodilator therapy in heart failure are reviewed in detail in texts of cardiovascular medicine and physiology (e.g., Smith et al., 1997). Briefly, the hemodynamic responses to heart failure are similar in some respects to those that accompany a fall in blood pressure due to hypovolemia; they include tachycardia and an increase in venous and arterial vasoconstriction, with shunting of blood toward the thorax and brain and away from the periphery and splanchnic and renal vascular beds. While this provides a clear evolutionary advantage to the organism to survive dehydration or hemorrhage, it is maladaptive and deleterious in chronic heart failure. The concepts of preload and afterload reduction provide a convenient framework in which to address treatment options in heart failure, which, in many respects, attempt to overcome these inappropriate compensatory hemodynamic responses. While this discussion focuses on heart failure due to left ventricular dysfunction, the general principles of preload and afterload reduction are applicable to failure of either ventricle despite differences in the specific drugs or other forms of therapy that can be employed. Although drugs may be classified as either "arterial" or "venous" vasodilators, most vasodilators exhibit activity on both vascular beds. Also, classes of vasodilators differ in the specific arterial beds that are affected. This has important implications in preserving, for example, renal blood flow and diuretic effectiveness and may explain, in part, the superior efficacy of certain classes of vasodilator agents in heart failure. Preload Reduction

The principle of preload reduction can be expressed in the form of the FrankStarling relationship illustrated in Figure 342. In early heart failure, increases in intraventricular volume and pressure, as well as heart rate, compensate for a decline in ventricular systolic performance due to underlying cardiac disease. In more advanced heart failure, there may be little or no further augmentation of stroke volume with increasing filling pressures (i.e., a "flat FrankStarling curve"), while the transmission of increased pressure into the pulmonary and systemic venous beds produces congestive symptoms. This usually is accompanied by worsening myocardial energetics due to an increase in ventricular wall stress and a decrease in diastolic coronary blood flow. Agents that reduce ventricular filling pressures by decreasing intravascular volume (e.g., diuretics) or by increasing venous capacitance (e.g., venodilators) decrease pulmonary venous congestion and may improve myocardial metabolism with minimal effects on stroke volume and cardiac output. While these measures clearly improve symptoms due to systolic ventricular dysfunction, they also benefit patients with congestive symptoms due to impaired diastolic compliance (i.e., "diastolic dysfunction"), whether due to ischemia or structural changes in the myocardium. However, patients with poorly compliant hypertrophied ventricles due, for example, to aortic stenosis, often require elevated end-diastolic filling pressures to support an adequate forward stroke volume. An excessive decrease in preload may markedly reduce cardiac output in these patients. Afterload Reduction The importance of systemic arterial vasodilation in improving cardiac hemodynamics in heart failure was succinctly described by Cohn and Franciosa (1977). Afterload, the sum of forces opposing ventricular emptying during systole, is dependent on aortic and aortic outflow tract (including valvular) impedance, systemic vascular resistance, ventricularvascular coupling (i.e., the harmonics of reflected arterial pressure waves during systole), and the volume of blood in the ventricle at the initiation of systole. Hypertrophy of ventricular muscle is a compensatory mechanism that reduces wall stress and preserves ventricular systolic function despite a primary abnormality in one or more of the determinants of afterload (e.g., a stenotic aortic valve). In the failing heart, as illustrated in Figure 343, any reduction in ventricular wall stress during systole, whether achieved by corrective surgery, intraaortic balloon counterpulsation, or vasodilator drugs, results in improved systolic contractile function. In addition, any reduction in aortic or arterial determinants of afterload will improve forward stroke volume and improve signs and symptoms of mitral regurgitation, which is often present in patients with severe heart failure due to systolic dysfunction, even in the absence of primary disease of the mitral valve. Figure 343. Relationship between Ventricular Outflow Resistance and Stroke Volume in Patients with Systolic Ventricular Dysfunction. An increase in ventricular outflow resistance, a principal determinant of "afterload," has little effect on stroke volume in normal hearts, as illustrated by the relatively flat curve. In contrast, in patients with systolic ventricular dysfunction, an increase in outflow resistance often is accompanied by a sharp decline in stroke volume. With more severe ventricular dysfunction, this curve becomes steeper. Because of this relationship, a reduction in systemic vascular resistance, one component of outflow resistance, following administration of an arterial vasodilator may markedly increase stroke volume in patients with severe myocardial dysfunction. The resultant increase in stroke volume may be sufficient to offset the decrease in systemic vascular resistance, thereby preventing a fall in systemic arterial pressure. (Adapted from Cohn and Franciosa, 1977, with permission.)

Inhibition of the ReninAngiotensin System The reninangiotensin system has a central role in the pathophysiology of heart failure (for detailed description of the reninangiotensin system, seeChapter 31: Renin and Angiotensin). Angiotensinogen is cleaved by kidney-derived renin to form the decapeptide angiotensin I; ACE converts angiotensin I to the octapeptide angiotensin II (seeFigure 344). Angiotensin II is a potent arterial vasoconstrictor which promotes sodium and water retention through its role in the regulation of renal hemodynamics and release from the adrenal cortex of aldosterone. Additionally, angiotensin II potentiates catecholamine release, is arrhythmogenic, promotes vascular hyperplasia and pathologic myocardial hypertrophy, and stimulates myocyte death. These actions of angiotensin II contribute to the pathophysiology of heart failure and, in many cases, to pathological remodeling of the myocardium leading to disease progression. Figure 344. The Renin-Angiotensin-Aldosterone System. Angiotensin II is formed through the cleavage of angiotensin I by angiotensin converting enzyme (ACE). Most of the known biological effects of angiotensin II are mediated by the type 1 angiotensin receptor (AT1). In general, the type 2 angiotensin receptor (AT2) appears to counteract the actions of the AT 1 receptor. Angiotensin II also may be formed through non-ACE-dependent pathways. These pathways, and possibly incomplete inhibition of tissue ACE, may account for persistence of angiotensin in patients treated with ACE inhibitors. AT1 receptor antagonists have been postulated to provide more complete blockade of the reninangiotensin-aldosterone system than ACE inhibition alone. ACE also is a kininase, and therefore ACE inhibition reduces bradykinin degradation, thus enhancing its levels and biological effects including the release of nitric oxide and prostacyclin. It has been suggested that bradykinin may mediate many of the important biological effects of ACE inhibitors.

ACE inhibitors have been shown to suppress angiotensin II and aldosterone production, decrease sympathetic nervous system activity, and potentiate the effects of diuretics in heart failure. However, angiotensin II levels frequently return to baseline values following chronic treatment with ACE inhibitors (Juillerat et al., 1990). This is believed to be due to production of angiotensin II through non-ACE-dependent pathwaysfor example, through the action of chymase, a tissue protease. The sustained clinical effectiveness of ACE inhibitors despite failure to maintain angiotensin II suppression has raised the possibility that there may be additional or alternate mechanisms by which ACE inhibitors work in heart failure. ACE is identical to kininase II, which degrades bradykinin and other kinins; bradykinin stimulates production of nitric oxide, cyclic GMP and vasoactive prostaglandins which result in vasodilation and oppose the effects of angiotensin II on vascular- and myocardial-cell growth. Thus, it has been suggested that increased levels of bradykinin as a result of ACE inhibition may play an important role in the hemodynamic and antiremodeling effects of ACE inhibitors. ACE inhibitors cause both arterial and venous dilation. There are reductions in systemic and pulmonary arterial resistances. Mean arterial pressure may be unchanged or decrease, but heart rate usually is unchanged, even when there is a decrease in systemic pressure, perhaps due to a decrease in sympathetic nervous system activity. The decrease in left ventricular afterload results in an increase in cardiac output due to increases in stroke volume and ejection fraction. Venodilation results in decreases in right and left heart filling pressures and end-diastolic volumes. An alternative means of inhibiting the reninangiotensin system is through inhibition of angiotensin receptors. Most of the known clinical actions of angiotensin II, including its deleterious effects in heart failure, are mediated through the type 1 angiotensin receptor (AT1). Type 2 angiotensin receptors (AT2) also are present throughout the cardiovascular system, where they are believed to counterbalance the biological effects of AT 1-receptor stimulation. Due to their more distal site of action, AT1 angiotensin-receptor blockers (ARBs) theoretically may allow more complete interruption of angiotensin's actions than can be obtained with ACE inhibitors. Furthermore, since

AT1-receptor blockade has been associated with a compensatory increase in angiotensin II levels, AT2-receptor stimulation is increased, potentially enhancing the beneficial effects of ARBs. On the other hand, a theoretical drawback of the ARBs is their failure to increase bradykinin production. The relative merits of these possible mechanisms of action remain to be seen, as does the value of combined therapy with an ACE inhibitor and an ARB, which has the potential to provide more complete reninangiotensin system blockade while increasing bradykinin production. Angiotensin Converting Enzyme Inhibitors The first orally active ACE inhibitor, captopril (CAPOTEN), was introduced in 1977, and currently five other ACE inhibitorsenalapril (VASOTEC), ramipril (ALTACE), lisinopril (PRINIVIL, ZESTRIL), quinapril (ACCUPRIL ), and fosinopril (MONOPRIL), are approved by the U.S. Food and Drug Administration (FDA) for the treatment of heart failure. ACE inhibitors are now indicated for the treatment of heart failure of any severity, including asymptomatic left ventricular dysfunction. ACE-inhibition therapy should be initiated at a low dose (e.g., 6.25 mg of captopril or 5 mg of lisinopril), as some patients may experience an abrupt drop in blood pressure, particularly if they are volume depleted. Unacceptable hypotension usually can be reversed by intravascular volume expansion, although this may be counterproductive in patients with symptomatic heart failure. ACE inhibitor doses are customarily titrated upwards over several days in hospitalized patients or a few weeks in ambulatory patients, with careful observation of blood pressure, serum electrolytes, and serum creatinine levels. There is no precisely defined relationship between dose and long-term clinical effectiveness of these drugs. The target doses of these drugs in several large prospective trials in which a positive effect was demonstrated on mortality and other endpoints were 50 mg of captopril three times a day (Pfeffer et al., 1992); 10 mg of enalapril twice daily (SOLVD Investigators, 1991; Cohn et al., 1991); 10 mg of lisinopril once daily (GISSI-3 Investigators, 1994); or 5 mg twice daily of ramipril (AIRE Study Investigators, 1993). The question of optimal dosage of ACE inhibitor was addressed in the Assessment of Treatment with Lisinopril and Survival (ATLAS) study (Packer, 1999). Highdose lisinopril (32.5 or 35 mg) reduced the combined endpoint of mortality and hospitalization when compared to a low dose (2.5 or 5 mg). Based on the available evidence, the initial dosage of an ACE inhibitor should be titrated to the dosage that was shown to be of benefit in the major heartfailure trials. In patients who have not achieved an adequate clinical response, further uptitration to higher doses, as tolerated, may be of value. In patients with heart failure and reduced renal blood flow, ACE inhibitors, unlike other vasodilators, limit the kidney's ability to autoregulate glomerular perfusion pressure due to their selective effects on efferent arteriolar tone. If this occurs, the dose of ACE inhibitor should be reduced or another class of vasodilator added or substituted. Rarely, worsening of renal function following initiation of therapy with an ACE inhibitor will be due to the presence of bilateral renal artery stenosis. Another class of vasodilator should be substituted if this occurs. Likewise, angioneurotic edema secondary to ACE inhibitors should prompt immediate cessation of therapy. A small rise in serum potassium levels occurs frequently with ACE inhibitors; this may infrequently be substantial, especially in patients with renal impairment. Mild hyperkalemia is best managed by institution of a low potassium diet, but may require adjustment of dosage. A troublesome cough may occur, believed related to the effects of bradykinin. Substitution of an ARB often alleviates this problem. Effect of ACE Inhibitors on Survival in Heart Failure

A number of placebo-controlled trials have demonstrated that ACE inhibitors improve survival in patients with overt heart failure due to systolic ventricular dysfunction regardless of the etiology or severity of symptoms. The Cooperative Northern Scandinavian Enalapril Survival Study (CONSENSUS, 1987) demonstrated a 40% reduction in mortality after 6 months in patients with severe heart failure randomized to enalapril rather than placebo. These results were confirmed by a 16% reduction in mortality in the treatment arm of the Studies On Left Ventricular Dysfunction (SOLVD Investigators, 1991) trial, in which patients with symptomatic mild to moderate heart failure and left ventricular ejection fractions less than 35% were randomized to receive either enalapril or placebo. The second Veterans Administration Cooperative VasodilatorHeart Failure Trial (V-HeFT II; Cohn et al., 1991) showed a small but clear survival benefit in patients with mild to moderate heart failure who were randomized to enalapril rather than the combination of hydralazine and isosorbide dinitrate. A smaller randomized trial comparing captopril to hydralazine and isosorbide dinitrate in patients with moderate to severe heart failure also demonstrated a significant survival advantage to patients receiving the ACE inhibitor (Fonarow et al., 1992). These data convinced many clinicians that ACE inhibitors clearly improved survival of patients with symptomatic heart failure. The prevention arm of the SOLVD trial subsequently examined whether or not asymptomatic patients with left ventricular systolic dysfunction also derive a survival benefit (SOLVD Investigators, 1992). Although this study failed to demonstrate a statistically significant reduction in mortality among enalapril-treated patients, there was a significant (29%) reduction in the combined end point of the development of symptomatic heart failure and death due to any cause. ACE inhibitors also may prevent the development of clinically significant ventricular dysfunction and mortality following myocardial infarction. The Survival And Ventricular Enlargement trial (SAVE; Pfeffer et al., 1992), which examined patients with recent, acute anterior myocardial infarction and ejection fractions of 40% or less, showed a 20% reduction in mortality and a 36% reduction in the rate of progression to severe heart failure in the captopril-treated group after 12 months of follow-up. Both the SOLVD trials (Konstam et al., 1992) and the SAVE trial (St. John Sutton et al., 1994) also demonstrated that enalapril and captopril, respectively, markedly reduced or prevented the increases in left ventricular end-diastolic and end-systolic volumes and decline in ejection fraction observed in patients randomized to receive placebo. The Acute Infarction Ramipril Efficacy trial (AIRE Investigators, 1993), which had a study design similar to that of the SAVE trial, also demonstrated a significant (27%) reduction in mortality in the ACE inhibitortreated group. These studies attest to the safety and efficacy of ACE-inhibitor therapy initiated early in the postinfarct period regardless of whether clinically significant left ventricular dysfunction is present at the time of randomization. Angiotensin II Receptor Antagonists Six orally active ARBs now have been approved for the treatment of hypertension, but none has been approved for use in heart failure. ARBs are discussed in detail in Chapter 31: Renin and Angiotensin and a recent review (Burnier and Brunner, 2000). In patients with heart failure, ARBs exert hemodynamic effects similar to those of the ACE inhibitors; however, there is much less information about their effects on long-term outcomes such as hospitalization and survival. The Evaluation of Losartan in the Elderly (ELITE) trial was primarily designed to examine the effects of the ARB losartan (COZAAR) on renal function in elderly patients with heart failure (Pitt et al., 1997). Patients were randomized to receive either the ACE inhibitor captopril, 50 mg three times daily, or losartan, 50 mg once daily; there were no significant differences in renal function between the two groups, but an unexpected reduction in

mortality and hospitalizations from heart failure was noted in the losartan group. This was followed by the much larger ELITE II study which, like ELITE I, compared losartan and captopril in elderly patients, but had mortality as the primary end point (Pitt et al., 2000). In contrast to the findings in ELITE I, no significant difference in outcome was noted between the treatment groups. However, treatment with the ARB was better tolerated and associated with fewer adverse effects than treatment with the ACE inhibitor. Thus, ELITE II failed to confirm the superiority of an ARB over an ACE inhibitor. In addition, this study was not designed to test the equivalency of these two drug classes, and its interpretation has been confounded by concern that the dosage of losartan studied was not sufficient to achieve adequate blockade of AT1 receptors. Thus, until further data are available, the much larger body of data showing the benefits of ACE inhibitors in heart failure supports their routine use as first line agents. Conversely, although the present data do not allow the conclusion that ARBs are equivalent to ACE inhibitors, it appears reasonable to use ARBs as an alternative in patients intolerant to ACE inhibitors. Large trials are in progress that should provide more definitive data regarding the relative roles of ACE inhibitors and ARBs in the treatment of heart failure. The use of ACE inhibitors and ARBs in combination offers the interesting possibility of additive therapeutic effects due to their different modes of action on the reninangiotensin system. Preliminary results suggest that combined therapy with the ARB candesartan and the ACE inhibitor enalapril had favorable effects on hemodynamics, ventricular remodeling, and neurohormonal profile compared to therapy with either agent alone (McKelvie et al., 1999). Again, more definitive data are required before combination therapy is routinely applied in clinical practice. Nitrovasodilators Nitrovasodilators are among the oldest and most widely used vasodilators in clinical practice. The mechanism underlying the ability of these drugs to activate soluble guanylyl cyclase and relax vascular smooth muscle has become apparent only in the past 15 years. These drugs mimic the activity of nitric oxide (NO), an intracellular and paracrine signaling autocoid that is formed by the conversion of arginine to citrulline mediated by a family of enzymes termed nitric oxide synthases. This family of enzymes is found in endothelial and smooth muscle cells throughout the vasculature as well as in many other cell types. The basis for the differential sensitivity of selected regions of the vasculature to specific nitrovasodilators (e.g., the sensitivity of the epicardial coronary arteries to nitroglycerin, for example) remains controversial. Unlike nitroprusside, which is spontaneously converted to nitric oxide by reducing agents such as glutathione, nitroglycerin and other organic nitrates undergo a more complex enzymatic biotransformation to nitric oxide or bioactive Snitrosothiols. The activities of specific enzyme(s) and cofactors required for this biotransformation, while not yet clearly identified, appear to differ within the vascular beds among organs and at different levels of the vasculature within an organ (Kelly and Smith, 1996). The basic pharmacology of the organic nitrates is discussed in Chapter 32: Drugs Used for the Treatment of Myocardial Ischemia. Organic Nitrates Organic nitrate preparations, most commonly isosorbide dinitrate (ISORDIL, DILATRATE, SORBITRATE), intravenous nitroglycerin, and nitroglycerin ointment, sublingual tablets, transdermal patch, and lingual spray, are relatively safe and effective agents in reducing ventricular filling pressures in acute as well as chronic congestive heart failure. The predominant effect at conventional doses is preload reduction due to an increase in peripheral venous capacitance. Nitrates also will cause a decline in pulmonary and systemic vascular resistance, particularly at

higher doses, although this response is less marked and less predictable than with nitroprusside. Due to their relatively selective vasodilating effects on the epicardial coronary vasculature, these drugs may enhance systolic and diastolic ventricular function by increasing coronary flow in patients with underlying ischemia. Isosorbide dinitrate has been shown to be more effective than placebo in improving exercise capacity and in reducing symptoms when administered chronically to heart-failure patients. However, limited effects on systemic vascular resistance and the problem of pharmacological tolerance have restricted the use of organic nitrates as single agents in the pharmacotherapy of heart failure. In a number of smaller trials, isosorbide dinitrate has been shown to increase the clinical effectiveness of other vasodilators such as hydralazine, resulting in a sustained improvement in hemodynamics that exceeded those of either drug given alone. Importantly, the combination of isosorbide dinitrate (20 mg four times daily) and hydralazine in the V-HeFT I trial reduced overall mortality compared to either placebo or the 1-adrenergic receptor antagonist prazosin in patients with mild to moderate heart failure concurrently treated with digoxin and diuretics (Cohn et al., 1986). Nitrate tolerance limits the long-term effectiveness of these drugs when administered throughout the day for heart-failure symptoms. Blood levels of these drugs should be permitted to fall to negligible levels for at least 6 to 8 hours each day. The timing of nitrate withdrawal (e.g., removal of a transdermal nitroglycerin patch or skipping a dose of isosorbide dinitrate) can be adjusted to the patient's symptoms. Patients with recurrent orthopnea or paroxysmal nocturnal dyspnea, for example, would probably benefit most by using nitrates at night. Orally bioavailable compounds containing sulfhydryl groups, such as N-acetylcysteine, may diminish tolerance to the hemodynamic effects of nitrates in heart failure (Mehra et al., 1994). Likewise, hydralazine may decrease nitrate tolerance by a strong antioxidant effect reducing superoxide formation (which reacts rapidly with nitric oxide) and hence increasing bioavailability of nitric oxide (Gogia et al., 1995). Hydralazine The vasodilator activity of hydralazine (APRESOLINE) is not mediated by any known neural or hormonal agent, and its cellular mechanism of action in vascular smooth muscle remains poorly understood. Hydralazine is an effective antihypertensive drug (Chapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension), particularly when combined with other agents that blunt compensatory increases in sympathetic tone and salt and water retention. In heart failure, hydralazine reduces right and left ventricular afterload by reducing systemic and pulmonary vascular resistance. This results in an augmentation of forward stroke volume and a reduction in ventricular systolic wall stress and regurgitant fraction in mitral insufficiency. Hydralazine also appears to have moderate "direct" positive inotropic activity in cardiac muscle unrelated to afterload reduction. Hydralazine has minimal effects on venous capacitance and therefore is most effective when combined with agents with venodilating activity (e.g., organic nitrates). Importantly, hydralazine is effective in reducing renal vascular resistance and in increasing renal blood flow to a greater degree than are most other vasodilators, with the exception of ACE inhibitors. Therefore, hydralazine may be useful in heart-failure patients with renal dysfunction who cannot tolerate an ACE inhibitor. The combination of hydralazine (300 mg/day) and isosorbide dinitrate was less effective than enalapril in reducing mortality in heart-failure patients in the V-HeFT II trial (Cohn et al., 1991), although this combination of agents did increase survival compared to placebo or the 1-adrenergic antagonist prazosin in V-HeFT I (Cohn et al., 1986). This is an important point, because a number

of promising vasodilator drugs, some of which also have direct effects on cardiac contractility, have been shown to increase mortality (e.g., milrinone, flosequinan, prostacyclin) in heart failure. Hydralazine, with or without nitrates, may provide additional hemodynamic improvement for patients with advanced heart failure who already are being treated with conventional doses of an ACE inhibitor, digoxin, and diuretics (Cohn, 1994). Side effects that may necessitate dose adjustment or withdrawal of hydralazine are common. Twenty percent of patients in the V-HeFT I trial (Cohn et al., 1986) complained of symptoms that could have been related to hydralazine, although the most common complaints, headache and dizziness, also could have been due to the concomitantly administered nitrates. Usually, the symptoms diminish with time or respond to a reduction in dose. The oral bioavailability and pharmacokinetics of elimination of hydralazine do not appear to be importantly affected by heart failure unless there is severe hepatic congestion or hypoperfusion. Intravenous hydralazine is available but provides little practical advantage over oral formulations except for urgent use during pregnancy, in which relative contraindications exist for most other vasodilators. Hydralazine is typically started at a dose of 10 to 25 mg three times a day and the dosage titrated to 100 mg three times a day over several days as clinical needs dictate and side effects allow. Ca2+ Channel Antagonists The Ca2+ channel antagonists are effective arterial vasodilators that have been widely used to treat hypertension (seeChapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension). Although these agents offer theoretical advantages in the management of heart failure, the clinical experience has been disappointing. The use of the first-generation Ca2+ channel antagonists (verapamil, diltiazem, nifedipine) has not been shown to produce sustained improvement in symptoms in patients with predominant systolic ventricular dysfunction. Indeed, these drugs may worsen symptoms and increase mortality in patients with systolic dysfunction, including patients with heart failure due to ischemic disease (Elkayam et al., 1993). The reason for this adverse effect of Ca2+ channel blockers in heart failure is unclear, but may be related to their known negative inotropic effects or to reflex neurohumoral activation. Amlodipine (NORVASC) and felodipine (PLENDIL ) are second-generation dihydropyridines with greater vascular selectivity, and hence, fewer negative inotropic effects than the first-generation agents. In the Prospective Randomized Amlodipine Survival Evaluation Study (PRAISE), 1153 patients with severe heart failure were randomized to receive amlodipine (up to 10 mg daily) or placebo (Packer et al., 1996a). A trend toward a decrease in mortality was noted in the amlodipine group, with a more pronounced survival benefit in the subgroup of patients with nonischemic cardiomyopathy. Therefore, PRAISE II further evaluated amlodipine in patients with nonischemic cardiomyopathy; preliminary results of this study, however, do not suggest a survival benefit with amlodipine, but confirmed its safety in this patient population. Similarly, felodipine was shown to have a neutral effect on survival and exercise tolerance in the third Vasodilator-Heart Failure Trial (V-HeFT III) (Cohn et al., 1997). The 2+ aforementioned Ca channel antagonists act on the voltage-sensitive L-type channel. In contrast, mibefradil is a Ca2+ channel antagonist that is selective for the non-voltage-regulated, T-type channel. Although this agent does not appear to have clinically significant negative inotropic activity, its use was associated with an increased risk of death (believed secondary to major adverse drug interactions) that prompted its withdrawal from the market. The available clinical evidence does not support the routine use of Ca channel antagonists as firstline therapy in patients with heart failure. However, the use of amlodipine or felodipine may be considered in certain circumstances in which additional control of blood pressure or afterload reduction is required, or if other vasodilators (e.g., ACE inhibitors, angiotensin receptor antagonists or hydralazine) are contraindicated or poorly tolerated.
2+

In contrast to results in patients with predominant systolic dysfunction, Ca2+ channel antagonists appear to be useful agents for the treatment of heart failure due predominantly to diastolic dysfunction, such as hypertensive or idiopathic hypertrophic cardiomyopathies. By slowing heart rate, which is an important determinant of diastolic filling time, verapamil and diltiazem may facilitate diastolic relaxation and lower diastolic filling pressures. These agents also can be useful in the acute management of patients in heart failure due to most supraventricular tachyarrhythmias in the absence of severe right or left ventricular systolic dysfunction, or a known or suspected extranodal atrioventricular accessory pathway (Wolff-Parkinson-White syndrome). Other Vasodilators As noted in Table 343, other vasodilator drugs are effective in reducing ventricular preload and afterload and improving symptoms in heart failure. None of these agents, however, has been shown to improve survival in heart-failure patients. Their use should be restricted to the treatment of patients who are intolerant to, or not adequately treated by, the agents discussed above. For the treatment of acute or chronic decompensated heart failure refractory to standard drug regimens and not complicated by significant aortic insufficiency, a mechanical aortic counterpulsation device (i.e., intraaortic balloon pump) often provides the most effective short-term means to reduce left ventricular afterload and directly increase cardiac output. -Adrenergic Receptor Antagonists The pharmacology of -adrenergic receptor antagonists is discussed in detail in Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists. The initial use of these agents for the treatment of heart failure was based largely on empirical evidence from small clinical trials, despite clinical and experimental animal data that these drugs can exert a negative inotropic effect and thereby worsen ventricular function. In the 1970s, Waagstein and associates in Sweden reported that -adrenergic antagonists (most commonly, the relatively 1-selective adrenergic receptor antagonist metoprolol) improved symptoms, exercise tolerance, and several measures of ventricular function over a period of several months in patients with heart failure due to idiopathic dilated cardiomyopathy (reviewed by Swedberg, 1993). With few exceptions, a number of small clinical trials over the next decade reinforced these initial observations (Bristow, 2000). Although none of these studies was sufficiently large to be definitive regarding symptoms, effort tolerance, or mortality, they did demonstrate a consistent increase in left ventricular ejection fraction (Figure 345). Serial measurements have shown that, immediately after starting a receptor antagonist in such patients, there is a decrease in systolic function as reflected by a decrease in ejection fraction. However, with continued treatment over the ensuing 2 to 4 months, systolic function gradually recovers and then improves beyond the baseline level (Hall et al., 1995). Since this time-dependent effect of -receptor antagonist therapy cannot be attributed to a direct hemodynamic action, it has been suggested that improved ventricular function with chronic receptor antagonist therapy is due to prevention of adverse effects of norepinephrine on the myocardium that are mediated by -adrenergic receptors (Eichhorn and Bristow, 1996).

Figure 345. Time-Dependent Effects of Metoprolol on Left Ventricular Ejection Fraction in Patients with Heart Failure. In patients with severe left ventricular dysfunction, the initial administration of a low dose of metoprolol caused an immediate depression in ejection fraction (day 1). However, over time and despite uptitration of metoprolol to full therapeutic levels, ejection fraction returned to baseline (1 month) and by 3 months was significantly higher than at baseline. In the group given standard therapy, ejection fraction did not change significantly. An increase in left ventricular systolic function between 2 and 4 months after initiation of therapy is seen consistently with a variety of adrenergic receptor antagonists used in patients with heart failure. This observation confirms that the direct hemodynamic effect of a -receptor antagonist in patients with heart failure is to depress contractile function. Thus, the improvement in function with chronic therapy cannot be attributed to a direct hemodynamic effect, and likely reflects a beneficial effect of treatment on the biology of the myocardium. (Adapted from Hall et al., 1995, with permission.)

Early Outcome Trials The Metoprolol in Dilated Cardiomyopathy (MDC) trial (Waagstein et al., 1993) was a multicenter, prospective, randomized trial that examined metoprolol versus placebo in 383 patients with mild to moderate idiopathic dilated cardiomyopathy (i.e., patients with clinically evident coronary artery disease or active myocarditis were excluded) who already were receiving optimal medical management including ACE inhibitors. Although there was no difference in mortality at 12 months of follow-up between the placebo- and metoprolol-treated groups, 19 patients receiving placebo deteriorated to the point of being listed for cardiac transplantation, a primary end point of the trial, compared to 2 patients receiving metoprolol. There was significant improvement in left ventricular ejection fraction, exercise tolerance, NYHA classification status, and the patients' own assessment of their quality of life (Waagstein et al., 1993; Andersson et al., 1994). The mean dose of metoprolol achieved was approximately 100 mg/day following a 6-week period of gradual upward titration beginning at 10 mg/day. Subsequently, the Cardiac Insufficiency Bisoprolol (CIBIS) trial examined the effect of another 1-selective antagonist, bisoprolol, in 641 patients with both ischemic and nonischemic dilated cardiomyopathy. This trial also failed to find an effect on overall mortality (Cibis Investigators, 1994). However, as in the MDC trial, there was evidence of

symptomatic and functional improvement. Carvedilol Carvedilol (COREG) is a nonselective -adrenergic receptor antagonist and an 1-selective adrenergic receptor antagonist. The U.S. Carvedilol Trial randomized 1094 patients with ischemic and nonischemic dilated cardiomyopathy and an ejection fraction <35% to carvedilol (target dose of 25 mg twice per day) or placebo (Packer et al., 1996b). All of the patients were ambulatory, clinically stable (essentially all were in NYHA class II or III), and receiving an angiotensin converting enzyme inhibitor. In patients randomized to carvedilol there was a 65% reduction in allcause mortality which was independent of age, gender, etiology of heart failure, or ejection fraction. The mortality benefit was due primarily to a decrease in deaths due to refractory pump failure, and to a lesser extent, sudden deaths. In a portion of the patients who were randomized to three dose levels of carvedilol (6.25, 12.5, or 25 mg twice per day), there was a strong relationship between the dose of carvedilol and the improvements in mortality and ejection fraction (Figure 346) (Bristow et al, 1996). Patient- and physician-reported symptoms were improved, and there was a 27% decrease in hospitalizations, although exercise function, as assessed by a 6-minute walk test, was not improved. In this trial, the effect of carvedilol on disease progression was assessed in a subgroup of 366 patients with good exercise function and mild symptoms at baseline (Colucci et al., 1996). Approximately 85% of these patients were in NYHA class II and the rest were in class III. The primary end point of clinical progression was defined as the occurrence of death due to heart failure, hospitalization for heart failure, or the need for a sustained increase in heart-failure medications. Carvedilol caused a 48% decrease in clinical progression, so defined, due to parallel reductions in death, hospitalization, and medication increase. The Australia/New Zealand Carvedilol Trial, which enrolled 415 patients with mild heart failure due to ischemic cardiomyopathy, likewise found a 26% reduction in the combined end point of all-cause mortality and hospitalization over a follow-up period of 18 to 24 months (Australia/New Zealand Heart Failure Research Collaborative Group, 1997). Based largely on these trials, in 1997 the FDA approved carvedilol for the treatment of patients with NYHA class II or III symptoms of heart failure and an ejection fraction <35%. Figure 346. Dose-Dependent Effect of Carvedilol on Left Ventricular Ejection Fraction. In the U.S. Carvedilol Trials Program a subgroup of patients were randomized to placebo or carvedilol in the standard dose (25 mg twice per day [bid]) or in a reduced dose of 12.5 or 6.25 mg twice per day. After 6 months of treatment, left ventricular ejection fraction ( LVEF) increased with all three doses of carvedilol, but not with placebo. Of note, the increase in ejection fraction was strongly related to the dose of carvedilol. These data emphasize the importance of titrating doses of -adrenergic receptor antagonists to the target or the highest tolerated dose. (Adapted from Bristow, 1996, with permission.)

Bisoprolol (ZEBETA) CIBIS-II, a follow-up trial to CIBIS I (discussed above), enrolled 2647 patients with moderate to severe heart-failure symptoms due to ischemic or nonischemic dilated cardiomyopathy (CIBIS-II Investigators, 1999). This trial found a 34% reduction in all-cause mortality in bisoprolol-treated patients that was due primarily to a decrease in sudden deaths (44% reduction) and, to a lesser extent, a decrease in pump failure (26% reduction). Because of the way deaths were classified, many patients with pump failure may have been placed in the "unknown" category, which accounted for a large fraction of all deaths. The mortality benefit of bisoprolol was independent of the etiology of heart failure and was associated with an approximately 36% decrease in hospitalizations for heart failure. Metoprolol (LOPRESSOR, TOPROL XL) The Metoprolol Randomized Intervention Trial in Congestive Heart Failure (MERIT-HF) randomized 3991 patients with NYHA functional class II to IV symptoms and an ejection fraction <40% to metoprolol (target dose, 200 mg/day) or placebo (MERIT-HF Study Group, 1999). Metoprolol caused a 34% decrease in all-cause mortality, which was attributable to similar reductions in sudden death (41% decrease) and death from worsening heart failure (49% decrease). As in the U.S. Carvedilol and CIBIS II trials, the beneficial effects of treatment on mortality were independent of age, gender, etiology of heart failure, or ejection fraction. Mechanism of Action It is not completely clear how -adrenergic receptor antagonists exert their benefits in heart failure. Although initial theories focused on resensitization of the -adrenergic pathway, this appears not to be critical, since some -receptor antagonists that have proven to be of clinical utility (e.g., carvedilol) do not cause this effect. A consistent finding in the large mortality trials is a reduction in the incidence of sudden death, presumably reflecting a decrease in malignant ventricular arrhythmias. Thus, an antiarrhythmic benefit seems likely. However, another consistent finding is an improvement in left ventricular structure and function with a decrease in chamber size and an increase in ejection fraction. This apparent antiremodeling action may reflect favorable effects on the molecular and cellular processes that underlie pathological remodeling. In support of this thesis, it has been shown that -adrenergic stimulation can cause apoptosis of cardiac myocytes

(Communal et al., 1998), and that mice overexpressing the 1-adrenergic receptor in the myocardium develop a dilated cardiomyopathy associated with myocyte apoptosis (Bisognano et al., 2000). Likewise, -adrenergic receptors may affect remodeling through their effects on gene expression and the turnover of extracellular matrix in the myocardium. It also has been suggested that -receptor antagonists could improve myocardial energetics (Eichhorn et al., 1994) and/or reduce oxidative stress in the myocardium (Sawyer and Colucci, 2000). Clinical Use of -Adrenergic Receptor Antagonists in Heart Failure The extensive body of data regarding the use of -receptor antagonists in chronic heart failure, reflecting more than 15,000 patients enrolled in controlled trials, has provided compelling evidence that -adrenergic receptor antagonists improve symptoms, reduce hospitalization, and decrease mortality in patients with mild and moderate heart failure. Accordingly, -receptor antagonists are now recommended for routine use in patients with an ejection fraction <35% and NYHA class II or III symptoms despite standard therapy with diuretics and an ACE inhibitor. This seemingly broad recommendation should be tempered by certain limitations in the experimental database. First, the large majority of the data that underlie this recommendation were obtained in relatively stable patients with mild to moderate symptoms. Therefore, the role of -receptor antagonists in patients with more severe symptoms, or with recent decompensation, is not yet clear. Likewise, the utility of -receptor blockade in patients with asymptomatic left ventricular dysfunction has not been studied. Finally, although it appears likely that the beneficial effects of these drugs is related to -receptor blockade, it cannot be assumed that all -receptor antagonists will exert similar effects. Thus, as discussed in Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists, within this general class of drugs there is marked heterogeneity in pharmacological characteristics such as -adrenergic receptor selectivity, direct or receptor-mediated vasodilation, and other nonreceptor-mediated actions (e.g., antioxidant effects). These properties may play a role determining the overall efficacy and utility of a given -receptor antagonist. Since -receptor antagonists have the potential to worsen both ventricular function and symptoms in patients with heart failure, several caveats should be remembered. First, -receptor antagonists should be initiated at very low doses, generally less than a tenth of the final target dose. Second, these drugs should be titrated upward slowly, over the course of weeks, and under careful supervision. The rapid institution of the usual -receptor-blocking doses used for hypertension or coronary artery disease may cause decompensation in many patients who otherwise would be able to tolerate a slower uptitration. Even when therapy is initiated with low doses of a -receptor antagonist, there may be an increased tendency to retain fluid that will require adjustments in the diuretic regimen. Third, although limited experience with NYHA class IIIB and IV patients suggests that they may tolerate -receptor blockers and benefit from their use, this group of patients should be approached with a high level of caution. Fourth, there is almost no experience in patients with new-onset, recently decompensated heart failure. There are theoretical reasons for caution in such patients, and at present they should not be treated with -receptor blockers until after they have stabilized for several days to weeks. Cardiac Glycosides The cardiac glycosides possess a common molecular motif, a steroid nucleus containing an unsaturated lactone at the C 17 position and one or more glycosidic residues at C 3 (seeFigure 34 7). Digoxin (LANOXIN, LANOXICAPS) and digitoxin (CRYSTODIGIN ) are both orally active, but only digoxin is in widespread clinical use today. Digitoxin differs from digoxin only by the absence of a hydroxyl group at C 12, resulting in a less hydrophilic compound with altered pharmacokinetics compared to digoxin. The cardiac glycosides have been used for centuries as therapeutic agents.

The beneficial effects in heart failure were believed to derive from a positive inotropic effect on failing myocardium and efficacy in controlling the ventricular rate response to atrial fibrillation. However, it is now recognized that the cardiac glycosides also modulate sympathetic nervous system activity, an additional mechanism that may contribute importantly to their efficacy in heart failure. Figure 347. Structure of Digoxin.

Mechanisms of Action Inhibition of Na+ , K+ ATPase All cardiac glycosides are potent and highly selective inhibitors of the active transport of Na+ and K+ across cell membranes, by binding to a specific site on the extracytoplasmic face of the subunit of Na+, K+ATPase, the enzymatic equivalent of the cellular "Na+ pump." The binding of cardiac glycosides to Na+, K+ATPase and inhibition of the cellular ion pump is reversible and entropically driven. These drugs bind preferentially to the enzyme following phosphorylation at a -aspartate on the cytoplasmic face of the subunit and stabilize this conformation (known as E2P). Extracellular K+ promotes dephosphorylation of the enzyme as an initial step in this cation's active translocation into the cytosol, thereby decreasing the affinity of the enzyme for binding cardiac glycosides. This provides one explanation for why increased extracellular K+ reverses some of the toxic effects of these drugs. The regulation of Na+, K+ATPase by digitalis has been reviewed in detail (Eisner and Smith, 1992). Positive Inotropic Effect Both Na+ and Ca2+ ions enter cardiac muscle cells during each cycle of depolarization, contraction, and repolarization (Figure 348). Ca2+ that enters the cell via the L-type Ca2+ channel during depolarization triggers the release of additional Ca2+ into the cytosol from an intracellular compartment, the sarcoplasmic reticulum (SR). The greater the amount of activating Ca2+, the greater the force of contraction. During myocyte repolarization and relaxation, Ca2+ is pumped back into the SR by a Ca2+ATPase and also is removed from the cell by the Na+Ca2+ exchanger and by a sarcolemmal Ca2+ATPase.

Figure 348. Sarcolemmal Exchange of Na+ and Ca 2+ during Cell Depolarization and Repolarization. Na+ and Ca 2+ ions enter mammalian cardiac muscle cells during each cycle of membrane depolarization, triggering the release, through 2+ 2+ Ca release channels (G), of larger amounts of Ca from internal stores in the 2+ sarcoplasmic reticulum (SR). The resulting increase in intracellular Ca interacts with troponin C and hence is responsible for activating the cross-bridge interactions between actin filaments and myosin cross-bridges that result in + sarcomere shortening. The electrochemical gradient for Na across the + sarcolemma is maintained by active (i.e., ATP-consuming) transport of Na out + + + of the cell by the sarcolemmal Na , K ATPase (D). Na is actively extruded by Na+, K+ATPase, while the bulk of cytosolic Ca2+ is pumped back into the SR by a Ca2+ATPase (F1), where it is bound by the protein calsequestrin, and the remainder is removed from the cell by either a plasma membrane Ca2+ATPase (F2) or a high capacity Na+ Ca 2+cation exchange protein (B, E). This sarcolemmal membrane protein exchanges 3 Na+ ions for every Ca2+ ion, using the electrochemical potential of Na+ to drive Ca2+ extrusion. Note that the direction of cation transport may reverse briefly during depolarization (B), when the electrical gradient across the sarcolemma is transiently reversed. Adrenergic receptor agonists and phosphodiesterase inhibitors, by increasing intracellular cyclic AMP levels, activate protein kinase A, which enhances the contractile state by phosphorylating target proteins, including phospholamban and the subunit of the L-type Ca2+ channel. (Adapted from Smith et al., 1992, with permission.)

Importantly, the capacity of the exchanger to extrude Ca2+ from the cell depends on the intracellular Na+ concentration. Binding of cardiac glycosides to the sarcolemmal Na+, K+ATPase and inhibition of cellular Na+ pump activity results in a reduction in the rate of active Na+ extrusion and

driving the extrusion of intracellular Ca2+ during myocyte repolarization. Hence, some incremental Ca2+ is taken up into the SR to be made available to the contractile elements during the subsequent cell depolarization cycle, and contractility of the myocardium is augmented. Electrophysiological Actions (see alsoChapter 35: Antiarrhythmic Drugs) Atrial and ventricular muscle and specialized cardiac pacemaker and conduction fibers exhibit differing responses and sensitivities to cardiac glycosides that are a summation of the direct effects of these drugs on cardiac cells and their indirect, neurally mediated effects. At therapeutic, nontoxic serum or plasma concentrations (i.e., 1.0 to 2.0 ng/ml), digoxin decreases automaticity and increases maximal diastolic resting membrane potential predominantly in atrial and atrioventricular (AV) nodal tissues, due to an increase in vagal tone and a decrease in sympathetic nervous system activity. There also is a prolongation of the effective refractory period and a decrease in conduction velocity in AV nodal tissue. At higher concentrations, this may cause sinus bradycardia or arrest and/or prolongation of AV conduction or heart block. In addition, cardiac glycosides at higher concentrations can increase sympathetic nervous system activity and directly affect automaticity in cardiac tissue, actions that contribute to the generation of atrial and ventricular arrhythmias. Increased intracellular Ca2+ loading and increased sympathetic tone result in an increase in the spontaneous (phase 4) rate of diastolic depolarization as well as delayed afterdepolarizations that may reach the threshold for generation of a propagated action potential. This simultaneous nonuniform increase in automaticity and depression of conduction in HisPurkinje and ventricular muscle fibers predisposes to arrhythmias that may lead to ventricular tachycardia or fibrillation. Regulation of Sympathetic Nervous System Activity An increase in sympathetic nervous system activity is one of the physiological responses to a decline in heart function below that required for maintenance of a cardiac output adequate to meet the metabolic demands of body tissues (i.e., heart failure). This is due, in part, to a reduction in the sensitivity of the arterial baroreflex response to blood pressure, resulting in a decline in tonic baroreflex suppression of CNS-directed sympathetic activity (Ferguson et al., 1989). This desensitization of the normal baroreflex arc also is thought to be responsible in part for the sustained elevation in plasma norepinephrine, renin, and vasopressin levels in heart failure, as well as other indices of systemic neurohumoral activation that are characteristically observed in patients with heart failure. Increased sympathetic nervous system activity initially helps to maintain blood pressure and cardiac output by increasing heart rate, contractility, and systemic vascular resistance, and by decreasing the excretion of salt and water by the kidneys. However, when sustained chronically, these effects of sympathetic overactivity contribute to the pathophysiology of heart failure and progression of the underlying myocardial disease. A direct effect of cardiac glycosides on carotid baroreflex responsiveness to changes in carotid sinus pressure has been demonstrated in isolated baroreceptor preparations from animals with experimental heart failure (Wang et al., 1990). In addition, Ferguson et al. (1989) demonstrated in patients with moderate to advanced heart failure that infusion of the cardiac glycoside deslanoside increased forearm blood flow and cardiac index and decreased heart rate, while markedly decreasing skeletal muscle sympathetic nerve activity, an indicator of the centrally mediated sympathetic nervous system tone. This was unlikely to have been due predominantly to a direct inotropic effect of the drug, since dobutamine, a sympathomimetic drug that increases cardiac output to a comparable extent, did not affect muscle sympathetic nerve activity in these patients. A reduction in neurohumoral activation could represent an important additional mechanism contributing to the efficacy of cardiac glycosides in the treatment of heart failure.

Pharmacokinetics The elimination half-life for digoxin is 36 to 48 hours in patients with normal or near-normal renal function. This permits once-a-day dosing for patients with normal or mildly impaired renal function, and near steady-state blood levels are achieved 1 week after initiation of maintenance therapy. Digoxin is excreted for the most part unchanged with a clearance rate that is proportional to the glomerular filtration rate. In patients with congestive heart failure and marginal cardiac reserve, an increase in cardiac output and renal blood flow with vasodilator therapy or sympathomimetic agents may increase renal digoxin clearance, necessitating adjustment of daily maintenance doses. Nevertheless, digoxin is not removed effectively by peritoneal or hemodialysis due to the drug's large (4 to 7 liters/kg) volume of distribution. The principal tissue reservoir is skeletal muscle and not adipose tissue and, thus, dosing should be based on estimated lean body mass. Neonates and infants tolerate and appear to require higher doses of digoxin for an equivalent therapeutic effect than do older children or adults, although absorption and renal clearance rates are similar. Digoxin does cross the placenta, and drug levels in maternal and umbilical vein blood are similar. Most digoxin tablets average 70% to 80% oral bio- availability; however, approximately 10% of the general population harbors the enteric bacterium Eubacterium lentum, which can convert digoxin into inactive metabolites, and this may account for some cases of apparent resistance to standard doses of oral digoxin. Liquid-filled capsules of digoxin (LANOXICAPS) have a higher bioavailability than do tablets and require dosage adjustment if a patient is switched from one dosage form to the other. Parenteral digoxin is available for intravenous administration, and maintenance doses can be given by intravenous injection when oral dosing is impractical. Intramuscular digoxin administration is erratically absorbed, causes local discomfort, and is not recommended. A number of drug interactions (seeTable 344) and clinical conditions can alter digoxin's pharmacokinetics or alter a patient's susceptibility to toxic manifestations of these drugs. Chronic renal failure, for example, decreases digoxin's volume of distribution, necessitating a decrease in maintenance dosage of the drug. Electrolyte disturbances, especially hypokalemia, acid base imbalances, and type of underlying heart disease also may alter a patient's susceptibility to toxic manifestations of digoxin. Clinical Use of Digoxin in Heart Failure Since at least the turn of the century, there has been controversy surrounding the efficacy of cardiac glycosides in the treatment of patients with heart failure who are in sinus rhythm. Despite widespread use of digoxin, objective data from randomized, controlled trials on the safety and efficacy of digoxin had been lacking until the 1990s. The PROVED (Prospective Randomized study Of Ventricular failure and Efficacy of Digoxin; Uretsky et al., 1993) and RADIANCE (Randomized Assessment of Digoxin on Inhibition of Angiotensin Converting Enzyme; Packer et al., 1993) trials examined the effects of withdrawal of digoxin in patients with stable mild to moderate heart failure (i.e., NYHA class II and III) and systolic ventricular dysfunction (left ventricular ejection fraction <0.35). All patients studied were in normal sinus rhythm. Withdrawal of digoxin resulted in a significant worsening of heart-failure symptoms in patients who received placebo compared with patients who continued to receive active drug. Maximal treadmill exercise tolerance also declined significantly in patients withdrawn from digoxin in both trials despite continuation of other medical therapies for heart failure. The much larger Digoxin Investigators' Group (DIG) trial was designed to detect an effect of digoxin therapy on the survival of patients with heart failure (The Digitalis Investigation Group,

1997). In this randomized, double-blind trial, 6,800 patients with predominantly mild to moderate (NYHA class II to III) heart failure and a left ventricular ejection fraction <0.45 were assigned to receive either digoxin or placebo in addition to standard therapy including ACE inhibitors. A trend was seen toward a decrease in the risk of death attributed to worsening heart failure in the digoxintreated group. However, this was balanced by a small increase in the risk of death due to other cardiac causes (presumed to result from arrhythmia), and overall, no difference in mortality was seen between the treatment groups (seeFigure 349). However, fewer patients in the digoxin group were hospitalized due to worsening heart failure. This benefit was seen at all levels of ejection fraction but was greatest in patients with more severe degrees of heart failure. Interestingly, in a predefined substudy of patients with normal ejection fraction (i.e., presumed to have diastolic dysfunction), a similar pattern of benefit was seen with digoxin. Based on these data, it is recommended that digoxin be reserved for patients with heart failure who are in atrial fibrillation, or for patients in sinus rhythm who remain symptomatic despite therapy with adequate dosages of ACE inhibitors and -adrenergic receptor antagonists. Figure 349. Effect of Digoxin on Survival and Hospitalization for Heart Failure in the Digoxin Investigators Group (DIG) Trial. In the DIG trial, 6800 patients with New York Heart Association class II to III symptoms of heart failure and a left ventricular ejection fraction <0.45 were randomized to digoxin or placebo in addition to standard therapy including ACE inhibitors. There was no difference in mortality between the treatment groups (Panel A). However, fewer patients in the digoxin group were hospitalized due to worsening heart failure (Panel B). (Adapted from the Digitalis Investigation Group, 1997, with permission.)

Doses of Digoxin in Clinical Practice and Monitoring of Serum Levels Using indices of ventricular function, most studies suggest that the greatest increase in contractility is apparent at serum levels of digoxin around 1.4 ng/ml (1.8 nM) (Kelly and Smith, 1992a). The neurohormonal effects of digoxin may occur at lower serum levels, between 0.5 and 1.0 ng/ml; higher serum concentrations than this are not associated with further decreases in neurohormonal activation or with increased clinical benefit. Furthermore, a subgroup analysis of the DIG trial (The Digitalis Investigation Group, 1997) showed an apparent increased risk of death with increasing serum concentrations, even for values within the traditional therapeutic range. Therefore, many authorities advocate maintaining digoxin levels below 1.0 ng/ml. A common approach for initiating digoxin therapy is to begin at 0.125 to 0.25 mg/day, depending on lean body mass and creatinine clearance, and to measure serum digoxin levels a week later when a steady-state has been achieved. The blood sample should be obtained at least 6 hours following the last digoxin dose. Routine surveillance monitoring of digoxin levels need not be carried out, unless a significant deterioration in renal function occurs, or a new drug (e.g., amiodarone) which substantially alters digoxin pharmacokinetics, is started. Oral or intravenous loading with digoxin, while generally safe, is rarely necessary as other safer and more effective drugs exist for short-term

inotropic support. Digoxin Toxicity The incidence and severity of digoxin toxicity have declined substantially in the past two decades, due in part to the development of alternative drugs for the treatment of supraventricular arrhythmias and heart failure, to the increased understanding of digoxin pharmacokinetics, to the monitoring of serum digoxin levels, and to the identification of important interactions between digoxin and many commonly used drugs. Nevertheless, the recognition of digoxin toxicity remains an important consideration in the differential diagnosis of arrhythmias and/or neurological and gastrointestinal symptoms in patients receiving cardiac glycosides (Table 345). Vigilance for and early recognition of disturbances of impulse formation, conduction, or both are critically important. Among the more common electrophysiological manifestations are ectopic beats of AV junctional or ventricular origin, first-degree AV block, an excessively slow ventricular rate response to atrial fibrillation, or an accelerated AV junctional pacemaker. These often require only a dosage adjustment and appropriate monitoring. Sinus bradycardia, sinoatrial arrest or exit block, and second- or third-degree AV conduction delay usually respond to atropine, although temporary ventricular pacing may be necessary. Potassium administration should be considered for patients with evidence of increased AV junctional or ventricular automaticity, even when the serum K+ is in the normal range, unless high-grade AV block also is present. Lidocaine or phenytoin, which have minimal effects on AV conduction, may be used for the treatment of worsening ventricular arrhythmias that threaten hemodynamic compromise. Electrical cardioversion carries increased risk of inducing severe rhythm disturbances in patients with overt digitalis toxicity, and it should be used with particular caution. Antidigoxin Immunotherapy An effective antidote for digoxin or digitoxin toxicity is now available in the form of antidigoxin immunotherapy with purified Fab fragments from ovine antidigoxin antisera (DIGIBIND). A full neutralizing dose of Fab based on either the estimated total dose of drug ingested or the total body digoxin burden (Table 346) can be administered intravenously in saline solution over 30 to 60 minutes. For a more comprehensive review of the treatment of digitalis toxicity, seeKelly and Smith (1992b). Chronic Positive Inotropic Therapy Several oral inotropic agents have been developed, some with vasodilating properties. Although many of these agents cause a marked improvement in hemodynamic function and may alleviate symptoms and improve exercise capacity, their effect on mortality with long-term treatment has been disappointing. The dopaminergic agonist ibopamine, the cyclic AMP phosphodiesterase (PDE) inhibitors milrinone, inamrinone (formerly amrinone) and vesnarinone, and the benzimidazoline PDE inhibitor with calcium-sensitizing properties, pimobendan, have been associated with increased mortality (Hampton et al., 1997; Packer et al., 1991; Cohn et al., 1998). This increased mortality has been attributed to an increased risk of arrhythmia with these agents, and it underscores the observation that effects of a drug on hemodynamic function and survival need not be directly related. Therefore, digoxin remains the only oral inotropic agent that should be used in patients with heart failure. Continuous or intermittent outpatient therapy with dobutamine or milrinone, administered by a portable or home-based infusion pump through a central venous catheter, has been evaluated in

patients with end-stage heart failure and symptoms refractory to other classes of drugs. There is as yet no convincing evidence that chronic parenteral inotropic therapy improves the quality or length of life. Furthermore, there are concerns that this form of therapy may actually hasten death (reviewed by Gheorghiade, 2000). Outpatient inotropic therapy may be useful in patients with intractable heart failure who are awaiting heart transplantation or who are not candidates for further management in hospital. The use of parenteral inotropic agents in hospitalized patients with heart failure is discussed later in this chapter. Anticoagulation and Antiplatelet Drugs in Heart Failure Patients with heart failure have a significantly higher incidence of stroke and thromboembolism. Nonetheless, the rate of embolic events remains relatively low, and retrospective analyses have shown conflicting evidence of benefit from anticoagulation in patients who remain in sinus rhythm. Anticoagulation with warfarin (COUMADIN ; seeChapter 55: Anticoagulant, Thrombolytic, and Antiplatelet Drugs) is recommended for patients who have atrial fibrillation or a history of a previous embolic event, if there is evidence of a left ventricular thrombus, or possibly, if left ventricular dysfunction is severe and the ventricles are markedly dilated. Similarly, while aspirin therapy is effective in reducing the incidence of ischemic events in patients with coronary artery disease, concerns have been raised that aspirin may attenuate the benefits from ACE inhibitors in patients with heart failure (Nguyen et al., 1997; Al-Khadra et al., 1998). A potential mechanism for this postulated adverse interaction is antagonism of bradykinin-mediated prostaglandin generation by aspirin. Synthesis of vasodilatory prostaglandins and other potentially beneficial molecules is reduced, thereby attenuating the favorable effects of ACE inhibition. In support of this thesis is the finding that the vasodilatory effects of enalapril may be blunted by concomitant administration of aspirin (Hall et al., 1992). Furthermore, a recent study has shown a trend toward greater morbidity and mortality in patients with heart failure receiving aspirin compared to those receiving placebo or warfarin (Cleland, 1999). Accordingly, until more definitive data are available on the use of aspirin in heart failure, it should be reserved for patients who have a clear indication for its use (e.g., the presence of known or suspected coronary artery disease). Alternative antiplatelet agents that do not appear to interfere with prostaglandin metabolism are now available in the form of the adenosine diphosphate antagonists ticlopidine (TICLID ) and clopidogrel (PLAVIX ). While both agents have been demonstrated to be effective in the prevention of ischemic events, there are as yet few data on their use in patients who have heart failure. Antiarrhythmic Drugs in Heart Failure Sudden cardiac death accounts for a significant proportion of the mortality due to heart failure. In the majority of cases, these deaths have been attributed to a ventricular tachyarrhythmia. However, ventricular arrhythmias are common in patients in heart failure, and are usually asymptomatic. Antiarrhythmic drugs, while effective in suppressing ventricular arrhythmias, may be associated with an increased risk of mortality, believed related to the proarrhythmic and negative inotropic properties common to many of these agents (Echt et al., 1991) (seeChapter 35: Antiarrhythmic Drugs). Therefore, indiscriminate use of antiarrhythmic drugs is not recommended in patients with heart failure who have ventricular arrhythmias. A possible exception is amiodarone (CORDARONE; seeChapter 35: Antiarrhythmic Drugs), an antiarrhythmic agent that has additional -adrenergic suppressing properties which may contribute to its beneficial effect in heart failure. Initial studies using amiodarone were promising, suggesting its use improved survival in patients with heart failure (Doval et al., 1994); however, this has not

been confirmed in later studies (Singh et al., 1995; Buxton et al., 1999). In the Multicenter Unsustained Tachycardia Trial (MUSTT), patients with heart failure and asymptomatic ventricular tachycardia who had inducible ventricular arrhythmias documented during an electrophysiological study were randomized to receive either amiodarone, an implanted cardiac defibrillator (ICD), or routine management. Improved survival was demonstrated in patients who received an ICD compared to those receiving routine management; in contrast, no benefit was seen with amiodarone therapy (Buxton et al., 1999). Thus, the available evidence does not support the routine use of amiodarone to suppress ventricular arrhythmias in patients with heart failure. Patients who have experienced a life-threatening ventricular arrhythmia should receive an ICD as treatment of first choice, with amiodarone being reserved as alternative therapy for patients who are unable to receive an ICD. However, amiodarone is effective in preventing the recurrence of atrial fibrillation or other supraventricular arrhythmias in patients who have ventricular dysfunction. Guidelines for the Treatment of Ambulatory Heart Failure As the number of proven therapeutic interventions in heart failure grows, the complexity of managing these patients has correspondingly increased. A number of guidelines for the management of patients with heart failure have been published by learned societies. The main points of the most recent set of guidelines (Heart Failure Society of America, 1999) are summarized in Figure 3410. Figure 3410. Guidelines for Pharmacological Management of Ambulatory Patients with Heart Failure. As the number of options for the drug therapy of heart failure increases, it has become more important to determine, based in most cases on evidence from clinical trials, the optimal usage of these drugs. Shown are recommendations for the pharmacological management of left ventricular systolic dysfunction as formulated by the Heart Failure Society of America. Blue boxes represent groups in which drugs should be routinely administered. Gray boxes represent groups in whom drug use should be considered. (ACE inhibitor = angiotensin-converting enzyme inhibitor; ARB = angiotensin receptor blocker; HF = heart failure) (Adapted from Heart Failure Society of America, 1999, with permission.)

Parenteral Drugs for the Treatment of Hospitalized Patients with Heart Failure General Considerations Patients with heart failure are commonly hospitalized because of increased dyspnea and peripheral edema due to pulmonary and systemic congestion. Fatigue also is a frequent complaint and is related to a reduction in cardiac output and perfusion of skeletal muscle. Accordingly, relief of congestion through the use of diuretics and possibly venodilators is often a priority in patients hospitalized for heart failure. Treatment also is directed at improving ventricular function and increasing cardiac output by lowering ventricular afterload and enhancing myocardial contractility. Wherever possible, precipitating factors (e.g., fever, infection) should be identified and removed, and the underlying cause of heart failure (e.g., ischemia, valvular disease) corrected. Diuretics As discussed in the treatment of ambulatory heart failure, diuretics are important for the alleviation of intravascular and extravascular fluid overload. In patients with decompensated heart failure of sufficient severity to warrant admission to a hospital, it is generally desirable to initiate an effective diuresis by using intravenous doses of a loop diuretic. Intravenous administration provides a more rapid and predictable diuresis than does the oral route, which is susceptible to delayed or impaired gut absorption, particularly in patients with marked fluid accumulation. The loop diuretic may be administered as repetitive boluses which are titrated against the desired response, or by constant infusion. An advantage of the latter approach is that the same total daily dose of diuretic, when given as a continuous infusion, can result in a more sustained and continuous natriuresis due to

maintenance of high diuretic drug levels within the lumen of renal tubules. In addition, the constant infusion helps to decrease the risk of ototoxicity that occurs with transient, high blood levels of the drug following repetitive intermittent loop diuretic dosing (Lahav et al., 1992). A typical continuous furosemide infusion is initiated with a 40-mg bolus injection followed by a constant infusion of 10 mg/hour, with upward titration of the infusion as necessary. When there is a poor response to diuretics due to reduced renal perfusion, the short-term administration of sympathomimetic drugs or phosphodiesterase inhibitors to increase cardiac output may be necessary to achieve a response. Another useful approach (see section on Diuretic Resistance in Heart Failure) is the intravenous administration of dopamine at so-called "low" doses (i.e., less than 2 g/kg per minute, based on estimated lean body weight) that can cause selective dopaminergic receptor stimulation, thereby causing a selective increase in renal blood flow without causing systemic and venous constriction via -adrenergic receptor stimulation, as may occur at higher infusion rates. Parenteral Vasodilators Sodium Nitroprusside Sodium nitroprusside (NITROPRESS ) is a potent vasodilator that is effective in reducing both ventricular filling pressures and systemic and arterial resistance. It has a rapid onset of action, within 2 to 5 minutes, is quickly metabolized to cyanide and nitric oxide, and its dose usually can be titrated expeditiously to achieve an optimal and predictable hemodynamic effect. For these reasons, nitroprusside is commonly used in intensive-care settings for rapid control of hypertension (seeChapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension) and for the management of acutely decompensated heart failure. Nitroprusside reduces ventricular filling pressures by directly increasing venous compliance, resulting in a redistribution of blood volume from central to peripheral veins. Nitroprusside is among the most effective afterload-reducing drugs by virtue of the spectrum of pharmacodynamic actions the drug has on different vascular beds (seeFigure 3411). It causes a fall in peripheral vascular resistance as well as an increase in aortic wall compliance and, at optimal doses, improves ventricularvascular coupling. These effects decrease left-ventricular afterload, resulting in an increase in cardiac output. Nitroprusside also dilates pulmonary arterioles and reduces right ventricular afterload. This combination of preload- and afterload-reducing effects improves myocardial energetics by reducing wall stress, provided that blood pressure does not fall to the point of compromising diastolic coronary artery flow or of activating a marked reflex increase in sympathetic nervous system tone. Following the rapid withdrawal of nitroprusside infusion, there may be a transient deterioration in ventricular function associated with a "rebound" increase in systemic vascular resistance that is thought to reflect activation of neurohormonal systems. Nitroprusside is particularly effective in patients with congestive heart failure due to mitral regurgitation or left-to-right shunts through a ventricular septal defect.

Figure 3411. Comparative Effects of Dobutamine, Milrinone, and Nitroprusside on Left Ventricular Contractility and Systemic Vascular Resistance. Dobutamine, milrinone, and nitroprusside are parenteral agents used in the management of patients with severe heart failure. Shown are the effects of these drugs on left ventricular contractility, as reflected by left ventricular peak contractility (+dP/dt), and systemic vascular resistance (SVR) in patients with heart failure. Dobutamine and milrinone both increase left ventricular contractility due to their positive inotropic actions on the myocardium. Milrinone, and to a lesser extent dobutamine, also decreased SVR indicating that they also exert a vasodilatory action. However, for a comparable increase in contractility, milrinone causes a greater decrease in SVR. Nitroprusside, a pure vasodilator, decreases SVR but had no effect on contractility. (Adapted from Colucci et al., 1986, with permission.)

The most common adverse effect of nitroprusside, as with most vasodilators, is hypotension. The increase in renal blood flow that accompanies an increase in cardiac output following initiation of nitroprusside in patients with severe heart failure may improve glomerular filtration and diuretic effectiveness. However, the redistribution of blood flow from central organs to peripheral vascular beds may limit or prevent an increase in renal blood flow in some patients. Cyanide produced during the biotransformation of nitroprusside is rapidly metabolized by the liver to thiocyanate, which is cleared by the kidney. Thiocyanate and/or cyanide toxicity is uncommon but may occur in the face of hepatic or renal failure, or following prolonged high-dose infusions of nitroprusside. Typical symptoms include unexplained abdominal pain, mental status changes, convulsions or lactic acidosis. Methemoglobinemia is another unusual complication of prolonged, high-dose nitroprusside infusion. Nitroglycerin Nitroglycerin, like nitroprusside, is a potent vasodilator (see previous section on the treatment of ambulatory heart failure and Chapter 32: Drugs Used for the Treatment of Myocardial Ischemia). In contrast to nitroprusside, nitroglycerin is relatively selective for veins, particularly at low infusion

rates. Thus, intravenous nitroglycerin is most often used in the treatment of acute heart failure when a decrease in ventricular filling pressures is desired. At higher infusion rates, nitroglycerin also causes decreases in the systemic and pulmonary arterial resistance, thereby decreasing ventricular afterload. The use of nitroglycerin may be limited by headache and the development of nitrate tolerance, although the latter is generally overcome by uptitration of the infusion rate to the desired response. Since nitroglycerin is administered in ethanol, high infusion rates can be associated with significant blood-alcohol levels. -Adrenergic and Dopamine Receptor Agonists Dopamine and dobutamine are the positive inotropic agents most often used for the short-term support of the circulation in advanced heart failure. Isoproterenol, epinephrine, and norepinephrine, although useful in specific circumstances, have little role in the treatment of most cases of severe heart failure. The basic pharmacology of these and other adrenergic agonists is discussed in Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists. Dobutamine Dobutamine (DOBUTREX), in the formulation available for clinical use, is a racemic mixture that stimulates both 1- and 2-adrenergic receptor subtypes. In addition, the () enantiomer is an agonist for -adrenergic receptors, whereas the (+) enantiomer is a very weak partial agonist. At clinical infusion rates that result in a positive inotropic effect in humans, the 1-adrenergic effect in the myocardium predominates. In the vasculature, the -adrenergic agonist effect of the () enantiomer appears to be negated by the partial agonism of the (+) enantiomer and the vasodilatory action due to 2-receptor stimulation. Thus, the net pharmacological effect of dobutamine is to increase stroke volume due to a positive inotropic action. At doses that increase cardiac output, there is relatively little increase in heart rate. Dobutamine also generally causes a modest decrease in systemic resistance and venous filling pressures (seeFigure 3412). Of note, dobutamine does not activate dopaminergic receptors at any dose. Therefore, the increase in renal blood flow with dobutamine is due to an increase in renal blood flow that is proportional to the increase in cardiac output. Figure 3412. Comparative Hemodynamic Effects of Dopamine and Dobutamine in Patients with Heart Failure. Dopamine and dobutamine were titrated over the usual clinical doses in patients with severe heart failure. The numbers shown on the figures are infusion rates in g/kg per minute. Dobutamine increased cardiac output due to an increase in stroke volume (not shown). This effect was associated with modest decreases in pulmonary capillary wedge pressure and systemic vascular resistance, reflecting both direct vasodilation due to stimulation of 2-adrenergic receptors and reflex withdrawal of sympathetic tone in response to improved cardiovascular function. At infusion rates that exceeded 2 to 4 g/kg per minute, dopamine exerted a potent vasoconstrictor effect as evidenced by the increase in systemic vascular resistance. Pulmonary capillary wedge pressure also increased with dopamine due to venoconstriction and a decrease in left ventricular function caused by the increase in afterload. (Adapted from Stevenson and Colucci, 1996, with permission.)

Continuous infusions of dobutamine for up to several days generally are well tolerated, although pharmacological tolerance may limit the efficacy during long-term use. Infusion rates usually are initiated at 2 to 3 g/kg per minute, without a loading dose, and titrated upward according to the patient's symptoms, the hemodynamic goals, and the responsiveness to diuretics. Blood pressure may increase, not change, or fall, depending on the relative effects on vascular tone and cardiac output achieved. Heart rate often declines due to reflex withdrawal of sympathetic tone in response to improved cardiovascular function. Measurement of pulmonary capillary wedge pressure and cardiac output using a pulmonary artery catheter often allows more effective use of dobutamine

alone or in conjunction with other vasodilators and diuretics. The major side effects of dobutamine are excessive tachycardia and arrhythmias, which may require a reduction in dosage. Tolerance may occur after prolonged usage, necessitating switching to an intravenous class III cyclic AMP phosphodiesterase inhibitor (e.g., milrinone; see below). Likewise, in patients who have been receiving a -adrenergic receptor antagonist, the initial response to dobutamine may be attenuated until the -receptor blocker has been metabolized. Dopamine Dopamine, an endogenous catecholamine, has limited utility in the treatment of most patients with systolic ventricular dysfunction who are not in shock due to primary cardiac failure, hemorrhage, dehydration, or toxicity from vasodilatory drugs. The pharmacological and hemodynamic effects of dopamine are strongly dose- related. At low doses of less than or equal to 2 g/kg per minute (based on estimated lean body weight), dopamine causes vasodilation by direct stimulation of dopaminergic postsynaptic type 1 and presynaptic type 2 (i.e., D1 and D2) receptors in the peripheral vasculature and relatively selective vasodilation of splanchnic and renal arterial beds. This effect may prove useful in promoting renal blood flow and maintaining glomerular filtration rate in patients who are refractory to diuretics or whose glomerular filtration rate has begun to decline due to marginal renal perfusion. Dopamine also has direct effects on renal tubular epithelial cells that promote natriuresis. At intermediate infusion rates (2 to 5 g/kg per minute), dopamine directly stimulates -adrenergic receptors in the heart and induces norepinephrine release from vascular sympathetic neurons. At higher infusion rates (5 to 15 g/kg per minute), peripheral arterial and venous constriction caused by -adrenergic receptor stimulation occurs, which may be desirable for support of a critically reduced arterial pressure, but which otherwise may further suppress ventricular systolic function due to the increase in afterload. Even with intermediate infusion rates, some patients may have an increase in systemic vascular resistance. Tachycardia, which is more pronounced with dopamine than with dobutamine, may provoke ischemia in patients with coronary artery disease. It should be emphasized that the dosing ranges noted above are based on estimated lean body weight, not on the patient's actual weight. Unexplained tachycardia or newly apparent arrhythmias in a patient receiving, among other drugs, "renal range" dopamine should make the clinician suspect an inappropriately high dopamine infusion rate. Phosphodiesterase Inhibitors The cyclic AMP PDE inhibitors mimic the effects of adenylyl cyclase activation by inhibiting the breakdown of cyclic AMP. Increased levels of cyclic AMP result in increased myocardial contractility and vasodilation in the venous and arterial circulation. Although agents such as theophylline and caffeine had been identified as nonspecific cyclic GMP and cyclic AMP PDE inhibitors more than 30 years ago, the use of these drugs at hemodynamically effective doses is plagued by side effects. During the 1980s, amrinone (now called inamrinone), milrinone, and other PDE inhibitors with isoenzyme selectivity became available, largely alleviating these problems. Inamrinone and Milrinone Parenteral formulations of inamrinone (INOCOR) and milrinone (PRIMACOR) have been approved for short-term support of the circulation in advanced heart failure. Both drugs are bipyridine derivatives and relatively selective inhibitors of the cyclic GMP-inhibited, cyclic AMP PDE (type III) family. These drugs cause direct stimulation of myocardial contractility and acceleration of myocardial relaxation. In addition, they cause balanced arterial and venous dilation with a

consequent fall in systemic and pulmonary vascular resistances, and left and right heart filling pressures. Cardiac output increases due to the stimulation of myocardial contractility and the decrease in left ventricular afterload. As a result of this dual mechanism of action, the increase in cardiac output with milrinone, when compared at comparable decreases in systemic pressure, is greater than with the pure vasodilator nitroprusside; and conversely, the arterial and venous dilator effects of milrinone are greater, when compared for comparable increases in cardiac output, than with dobutamine (Colucci et al., 1986) (seeFigure 3411). Both drugs are effective when given either as single agents or, more commonly, in combination with other oral and/or intravenous drugs for short-term treatment of patients with severe heart failure due to systolic right or left ventricular dysfunction. Intravenous infusions of either drug should be initiated by a loading dose followed by a continuous infusion. For inamrinone, typically a 0.75-mg/kg bolus injection over 2 to 3 minutes is followed by a 2- to 20- g/kg per minute infusion. Milrinone is approximately 10-fold more potent than inamrinone. A loading dose of milrinone is usually 50 g/kg, and the continuous infusion rate ranges from 0.25 to 1.0 g/kg per minute. The elimination half-lives of inamrinone and milrinone in healthy subjects are 2 to 3 hours and 30 to 60 minutes, respectively, and are approximately doubled in patients with severe heart failure. Clinically significant thrombocytopenia occurs in 10% of patients receiving inamrinone but is rare with milrinone. Because of its greater selectivity for PDE III isoenzymes, shorter half-life, and fewer side effects, milrinone is the agent of choice among currently available PDE inhibitors for short-term, parenteral inotropic support in severe heart failure. General Guidelines for the Use of Parenteral Drugs In general, initial therapy in the patient with symptomatic pulmonary congestion should include diuresis to alleviate pulmonary and systemic vascular congestion (seeFigure 3413). The administration of an oral or intravenous nitrate preparation may provide additional rapid symptomatic relief of pulmonary congestion by causing venodilation. Cardiac output may be optimized by a variety of agents, the selection of which may be guided in part by the arterial pressure, or ideally, by knowledge of the systemic vascular resistance (SVR) as provided by a pulmonary artery catheter. Figure 3413. General Approach to the Treatment of Patients Hospitalized with Decompensated Heart Failure. Fluid overload is a common feature in patients hospitalized for decompensated heart failure. Therefore, the initial management of such patients generally involves the use of loop diuretics administered intravenously as boluses or continuous infusions. In patients who do not respond to loop diuretics alone, the addition of a thiazide diuretic (e.g., metolazone) or a low-dose infusion of dopamine may facilitate diuresis. Venodilators (e.g., nitroglycerin) may provide rapid alleviation of pulmonary congestion. Poor diuretic responsiveness and other sequelae of reduced organ hypoperfusion may respond to an increase in cardiac output caused by a positive inotropic agent (e.g., dobutamine or milrinone) and/or a vasodilator that reduces left ventricular afterload (e.g., nitroprusside or milrinone). Although it frequently is possible to use parenteral agents without invasive monitoring, in many patients therapy is best guided by knowledge of the left and right heart filling pressures and systemic vascular resistance (SVR). ACE = angiotensin converting enzyme.

In patients with an elevated arterial pressure, afterload reduction with nitroprusside may be very effective. Nitroprusside also may be effective when there is an elevated SVR, despite the presence of a low arterial pressure. On the other hand, in patients in whom the SVR is not elevated, afterload reduction may result in an excessive fall in arterial pressure. In such patients, increasing myocardial contractility with a positive inotropic agent such as dobutamine often is preferable. Milrinone, which exerts both inotropic and vasodilatory actions, may be tolerated in some patients who do not tolerate a pure vasodilator, and conversely, may provide better relief of pulmonary congestion than dobutamine. Likewise, in patients in whom the left ventricular filling pressure is known (or suspected) not to be markedly elevated, treatment with dobutamine may avoid an excessive decrease in preload that can occur with vasodilators. When heart failure is severe, in patients who fail to respond to standard therapy, or if there is progressive renal insufficiency in the face of persistent signs of elevated filling pressures, placement of a pulmonary artery catheter and hemodynamic monitoring may be important for guidance of drug selection and dosing. Once hemodynamic monitoring is established, therapy is directed at optimization of hemodynamics and improvement in clinical status of the patient; afterload and cardiac filling pressures should be reduced to target levels, and cardiac output augmented where necessary to ensure adequate perfusion of vital organs.

Prospectus Since the publication of the previous edition of this textbook, there has been continuing evolution in the drug treatment of heart failure. Several large randomized trials have shown consistent evidence of benefit and have firmly established the central role of -adrenergic receptor antagonism alongside ACE inhibition in the therapy of ambulatory patients with heart failure. Although neutral with regard to mortality, the efficacy of cardiac glycosides in improving symptoms and reducing

hospitalization for heart failure has been reaffirmed. Finally, an old drug has found new life with the revival of spironolactone as a useful adjunctive therapy in patients with severe heart failure. Despite these and other recent advances in the pharmacotherapy of heart failure, the mortality rate remains high. Thus, preventive strategies including aggressive control of risk factors for cardiovascular disease continue to be of great importance. Further insights into the cellular and molecular abnormalities that underlie the contractile, hemodynamic, and neurohormonal disturbances of heart failure have cleared the way for the development of innovative therapeutic agents that may impact positively on outcomes. New therapeutic strategies include agents that block or stimulate endogenous signaling pathways. Endothelin receptor blockers have been shown to exert beneficial hemodynamic effects in patients with heart failure during short-term administration and are being evaluated with regard to long-term effects on morbidity and mortality. Antagonists of the inflammatory cytokine, tumor necrosis factor- , which have been shown to be both safe and effective in the treatment of inflammatory conditions such as rheumatoid arthritis and enteritis, are being evaluated for the treatment of heart failure. The natriuretic peptides are endogenous molecules that cause vasodilation and oppose the reninangiotensin system. Nesiritide, a recombinant, human-brain natriuretic peptide, exerts favorable vasodilatory effects with short-term infusion and may be of value in the treatment of patients hospitalized with heart failure. Omapatrilat, an orally active neutral endopeptidase inhibitor which blocks both ACE and the degradation of natriuretic peptides, has been shown to exert beneficial clinical effects in patients with chronic heart failure. Ultimately, the role of these and other new pharmacologic approaches will need to be established through controlled trials that determine their efficacy and safety. For further discussion of normal and abnormal cardiac function and of heart failure seeChapters 215 and 216 in Harrison's Principles of Internal Medicine, 16th ed., McGraw-Hill, New York, 2005. Acknowledgment The authors wish to acknowledge Drs. Ralph A. Kelley and Thomas W. Smith, the authors of this chapter in the ninth edition of Goodman and Gilman's The Pharmacological Basis of Therapeutics, some of whose text has been retained in this edition.

Chapter 35. Antiarrhythmic Drugs

Overview Individual cardiac cells undergo depolarization and repolarization to form cardiac action potentials about sixty times per minute. The shape and duration of each action potential are determined by the activity of ion channel protein complexes on the surface of individual cells, and the genes encoding many of these proteins have now been identified. Thus, each heartbeat is the result of the highly integrated electrophysiological behavior of multiple gene products on multiple cardiac cells. Ion channel function can be perturbed by factors such as acute ischemia, sympathetic stimulation, or myocardial scarring to create abnormalities of cardiac rhythm, or arrhythmias. Available antiarrhythmic drugs suppress arrhythmias by blocking flow through specific ion channels or by altering autonomic function. Arrhythmias can range from incidental, asymptomatic clinical findings to life-threatening abnormalities. Mechanisms underlying cardiac arrhythmias have been identified in cellular and animal experiments. In some human cases, precise mechanisms are known, and treatment targeted

against those mechanisms can be used. In other cases, mechanisms can be only inferred, and the choice of drugs is based largely on results of prior experience. Antiarrhythmic drug therapy can have two goals: termination of an ongoing arrhythmia or prevention of an arrhythmia. It is now well recognized that antiarrhythmic drugs not only help to control arrhythmias but also can cause them, especially during long-term therapy. Thus, prescribing antiarrhythmic drugs requires that precipitating factors be excluded or minimized, that a precise diagnosis of the type of arrhythmia (and its possible mechanisms) be made, that the prescriber has reason to believe that drug therapy will be beneficial, and that the risks of drug therapy be minimized. In this chapter, the principles underlying normal and abnormal cardiac electrophysiology are outlined. Then, the mechanisms by which drugs modulate cardiac electrophysiology are presented, followed by a description of the important properties of individual agents. Principles of Cardiac Electrophysiology The flow of charged ions across cell membranes results in the ionic currents that make up cardiac action potentials. The factors that determine the magnitude of individual currents and how these are modified by drugs now can be elucidated at the cellular and molecular levels (Fozzard and Arnsdorf, 1991; Snyders et al., 1991; Priori et al., 1999). However, the action potential is a highly integrated entity: changes in one current almost inevitably produce secondary changes in other currents. Most antiarrhythmic drugs affect more than one ion current, and many exert ancillary effects such as modification of cardiac contractility or autonomic nervous system function. Thus, antiarrhythmic drugs usually exert multiple actions and can be beneficial or harmful in individual patients (Roden, 1994; Priori et al., 1999). The Cardiac Cell at Rest: A K+-Permeable Membrane Ions move across cell membranes in response to electrical and concentration gradients, not through the lipid bilayer but through specific ion channels or transporters. The normal cardiac cell at rest maintains a transmembrane potential approximately 80 to 90 mV negative to the exterior; this gradient is established by pumps, especially Na+,K+ATPase, and fixed anionic charges within cells. There is both an electrical and a concentration gradient that would move Na+ ions into resting cells (Figure 351). However, Na+ channels, which allow Na + to move along this gradient, are closed at negative transmembrane potentials, so Na+ does not enter normal resting cardiac cells. In contrast, a specific type of K+ channel protein (the inward rectifier channel) is in an open conformation at negative potentials: i.e., K+ can move across the cell membrane at negative potentials in response to either electrical or concentration gradients (Figure 351). For each individual ion, there is an equilibrium potential Ex at which there is no net driving force for the ion to move across the membrane. Ex can be calculated using the Nernst equation: Figure 351. Electrical and Chemical Gradients for K+ and for Na+ in a Resting Cardiac Cell. Inward rectifier K+ channels are open (left), allowing K+ ions to move across the membrane and the transmembrane potential to approach EK. In + + contrast, Na does not enter the cell despite a large net driving force because Na channel proteins are in the closed conformation (right) in resting cells.

Ex=61 log ([x]i/[x]o) (351) where [x]o is the extracellular concentration of the ion and [x]i is the intracellular concentration. For usual values for K+, [K]o= 4 mM and [K]i= 140 mM, the calculated K+ equilibrium potential EK is 94 mV. There is thus no net force driving K+ ions into or out of a cell when the transmembrane potential is 94 mV, which is close to the resting potential. If [K]o is elevated to 10 mM, as might occur in diseases such as renal failure or myocardial ischemia, the calculated EK rises to 70 mV. In this situation, K+ will tend to move down its concentration gradient. In fact, there is excellent agreement between changes in theoretical EK due to changes in [K]o and the actual measured transmembrane potential (Figure 352), indicating that the normal cardiac cell at rest is permeable to K+ (because inward rectifier channels are open) and that the concentration of K+ in the extracellular space is the major determinant of resting potential. Figure 352. The Influence of Extracellular K on Theoretical EK (Dotted Line) and on Measured Transmembrane Potential (Solid Line). At values of + extracellular K >4 mM, the two lines are identical, indicating that extracellular + K is the major factor influencing resting potential.
+

Na Channel Opening Initiates the Action Potential: Ion Currents If a cardiac cell at rest is depolarized above a threshold potential, Na+ channel proteins change conformation from the "closed" (or rest) to the conducting ("open") state, allowing up to 107 Na+ ions per second to enter each cell and moving the transmembrane potential towards ENa (+65 mV). This surge of Na+ ion movement lasts only about a millisecond, after which the Na+ channel protein rapidly changes conformation from the "open" state to an "inactivated," nonconducting state. Measuring Na+ current directly is technically demanding; therefore many studies report the maximum upstroke slope of phase 0 (dV/dtmax, or Vmax ) of the action potential (Figure 353), which + + is proportional to Na current. The traditional view is that Na channels, once inactivated, cannot reopen until they reassume the rested, or closed, conformation. Electrophysiological techniques capable of measuring the behavior of individual ion channel proteins are now revealing some of the detailed mechanisms of these state transitions, and the findings obtained are changing some traditional views. For example, a small population of Na+ channels may continue to open during the action potential plateau in some cells (Figure 353). In fact, a defect in the structural region of the + Na channel protein that has been implicated in control of channel inactivation is responsible for one form of the congenital long QT syndrome, a disease associated with abnormal repolarization and serious arrhythmias (Roden and Spooner, 1999). In general, however, as the cell membrane repolarizes, the changes in membrane potential to which Na+ channel proteins are subject moves them from inactivated to "closed" conformations. The relationship between Na+ channel availability and transmembrane potential is an important determinant of conduction and of refractoriness in many cells, as discussed below. Figure 353. The Relationship between a Hypothetical Action Potential from the Conducting System and the Time Course of the Currents That Generate It. The current magnitudes are not to scale; the Na+ current is ordinarily 50-fold larger than any other current, although the portion that persists into the plateau (phase 2) is small. Multiple types of Ca2+ current, transient outward current (ITO), and delayed rectifier (IK) have been identified; it is likely that each represents a different channel protein. 4-AP (4-aminopyridine) is a widely used in vitro blocker of K+ channels. ITO2 may be a Cl current in some species. Components of IK have been separated on the basis of how rapidly they activate: slowly (IKs), rapidly (IKr), or ultrarapidly (IKur). The voltage-activated, time-independent current may be carried by Cl (ICl) or K+(IKp, "p" for plateau). For all currents shown here (with the possible exception of ITO2), the genes encoding the major

poreforming proteins have been cloned. (Adapted from Task Force of the Working Group on Arrhythmias of the European Society of Cardiology, 1991, with permission.)

The changes in transmembrane potential generated by the inward Na+ current produce, in turn, a series of openings (and in some cases subsequent inactivation) of other channels (Figure 353). For example, when a cell from the epicardium or the HisPurkinje conducting system is depolarized by the Na+ current, "transient outward" K+ channels change conformation to enter an open, or conducting, state; since the transmembrane potential at the end of phase 0 is positive to EK, the opening of transient outward channels results in an outward, or repolarizing, K+ current (termed ITO ), which contributes to the phase 1 "notch" seen in some action potentials. Transient outward K+ channels, like Na+ channels, rapidly inactivate. During the phase 2 plateau of a normal cardiac action potential, an inward, depolarizing current primarily through Ca2+ channels is balanced by an outward, repolarizing current primarily through K+ ("delayed rectifier") channels. Delayed rectifier currents (termed IK) increase with time, while Ca2+ currents inactivate (and so decrease with time); the result is repolarization of the cardiac cell (phase 3) several hundred milliseconds after the initial Na+ channel opening. Mutations in the genes encoding repolarizing K+ channels are responsible for other forms of the congenital long QT syndrome (Roden and Spooner, 1999). Identification of these specific channels has allowed more precise characterization of the pharmacological effects of antiarrhythmic drugs. A common mechanism whereby drugs prolong cardiac action potentials is inhibition of specific delayed rectifier current, IKr.

Differing Action Potential Behaviors Among Cardiac Cells This general description of the action potential and the currents that underlie it must be modified for certain cell types (Figure 354), presumably because of variability in the number or products of ion channel genes expressed in individual cells. Endocardial ventricular cells lack a prominent transient outward current, while cells from the subendocardial HisPurkinje conducting system (and in some species from the midmyocardium) have very long action potentials (Antzelevitch et al., 1991). Atrial cells have very short action potentials, probably because ITO is larger, and an additional repolarizing K+ current, activated by the neurotransmitter acetylcholine, is present. As a result, vagal stimulation further shortens atrial action potentials. Cells of the sinus and atrioventricular (AV) nodes lack substantial Na+ currents. In addition, these cells, as well as cells from the conducting system, normally display the phenomenon of spontaneous diastolic, or phase 4, depolarization and thus spontaneously reach threshold for regeneration of action potentials. The rate of spontaneous firing usually is fastest in sinus node cells, which therefore serve as the natural pacemaker of the heart. Specialized K+ channels underlie the pacemaker current in heart. Figure 354. Normal Impulse Propagation. Action potentials from different regions of the heart are shown. In each panel, tissue that is depolarized is shown in light blue and the portion of the electrocardiogram to which it contributes is shown in black.

Modern molecular biological and electrophysiological techniques, by which the behavior of single ion channel proteins in an isolated patch of membrane can be studied, have refined the description of ion channels important for the normal functioning of cardiac cells and have identified channels that may be particularly important under pathological conditions. For example, it is now established that transient outward and delayed rectifier currents actually result from multiple ion channel subtypes (Figure 353; Tseng and Hoffman, 1989; Sanguinetti and Jurkiewicz, 1990), and that acetylcholine-evoked hyperpolarization results from activation of a K+ channel formed by heterooligomerization of multiple, distinct channel proteins (Krapivinsky et al., 1995). The understanding that molecularly diverse entities subserve regulation of the cardiac action potential is important because drugs may target one channel subtype selectively. Furthermore, ancillary function-modifying proteins (the products of diverse genes) have been identified for most

ion channels. In addition to the usual ("L-type") Ca2+ channels, a second type of Ca2+ channel, which is most prominent at relatively negative potentials, has been identified in some cardiac cells (Bean, 1985). This "T-type" Ca 2+ channel may be important in diseases such as hypertension and may play a role in pacemaker activity in some cells. A T-typeselective antihypertensive agent, mibefradil, was available briefly in the late 1990s but was withdrawn because it was involved in many serious, undesirable drugdrug interactions. Specific channels that transport Cl ions and result in repolarizing currents (ICl) have been identified in many species (Hume and Harvey, 1991); some of these are observed only under pathophysiological conditions, such as adrenergic stimulation. Some K+ channels are quiescent when intracellular ATP stores are normal and become active when these stores are depleted. Such ATP-inhibited K+ channels may become particularly important in repolarizing cells during states of metabolic stress such as myocardial ischemia (Weiss et al., 1991; Wilde and Janse, 1994). Maintenance of Intracellular Homeostasis With each action potential, the cell interior gains Na+ ions and loses K+ ions. An ATP-requiring Na+K+ exchange mechanism, or pump, is activated in most cells to maintain intracellular homeostasis. This Na+,K+ATPase extrudes three Na+ ions for every two K+ ions shuttled from the exterior of the cell to the interior; as a result, the act of pumping itself generates a net outward (repolarizing) current. Normally, intracellular Ca2+ is maintained at very low levels (<100 nM). In heart cells, the entry of Ca2+ during each action potential is a signal to the sarcoplasmic reticulum to release its Ca2+ stores. The resultant increase in intracellular Ca2+ then allows Ca2+-dependent contractile processes to occur. Removal of intracellular Ca2+ occurs by both an ATP-dependent Ca2+ pump (which moves Ca2+ ions back to storage sites in the sarcoplasmic reticulum) and an electrogenic Na+Ca2+ exchange mechanism on the cell surface, which exchanges three Na+ ions from the exterior for each Ca2+ ion extruded. The initial rise in Ca2+, which serves as the trigger for Ca2+ release from intracellular stores, is a result of the opening of Ca 2+ channels in the cell membrane or of Ca2+ entry through Na+Ca2+ exchange; i.e., in response to phase 0 entry of Na +, the Na +Ca 2+ exchange protein may transiently extrude Na+ ions in exchange for Ca2+ ions (Figure 353). Impulse Propagation and the Electrocardiogram Normal cardiac impulses originate in the sinus node. Impulse propagation in the heart depends on two factors: the magnitude of the depolarizing current (usually Na+ current) and the geometry of cellcell electrical connections. Cardiac cells are long and thin and well coupled through specialized gap junction proteins at their ends, whereas lateral ("transverse") gap junctions are sparser. As a result, impulses spread along cells two to three times faster than across cells. This "anisotropic" (direction-dependent) conduction may be a factor in the genesis of certain arrhythmias described below (Priori et al., 1999). Once impulses leave the sinus node, they propagate rapidly throughout the atria, resulting in atrial systole and the P wave of the surface electrocardiogram (ECG; Figure 354). Propagation slows markedly through the AV node, where the inward current (through Ca2+ channels) is much smaller than the Na+ current in atria, ventricles, or the subendocardial conducting system. This conduction delay allows the atrial contraction to propel blood into the ventricle, thereby optimizing cardiac output. Once impulses exit from the AV node, they enter the conducting + system, where Na currents are larger than in any other tissue. Hence, propagation is correspondingly faster, up to 0.75 meter/second longitudinally, allowing coordinated ventricular contraction, the QRS complex on the ECG, as impulses spread from the endocardium to the epicardium. Ventricular repolarization results in the T wave of the ECG. The ECG can be used as a rough guide to some cellular properties of cardiac tissue (Figure 354): (1) heart rate reflects sinus

node automaticity, (2) PR interval duration reflects AV nodal conduction time, (3) QRS duration reflects conduction time in the ventricle, and (4) the QT interval is a measure of ventricular action potential duration. Refractoriness: Fast-Response Versus Slow-Response Tissue If a single action potential, such as that shown in Figure 353, is restimulated very early during the plateau, no Na+ channels are available to open, so no inward current results and no action potential is generated: the cell is refractory. If, on the other hand, an extrastimulus occurs after the cell has repolarized completely, Na+ channels have recovered from inactivation, and a normal Na+ channel dependent upstroke results (Figure 355A). When an extrastimulus occurs during phase 3 of the action potential, the magnitude of the resultant Na+ current is dependent on the number of Na + channels that have recovered from inactivation (Figure 355A), which, in turn, is dependent on the voltage at which the extrastimulus was applied. Thus, in atrial, ventricular, and HisPurkinje cells ("fast-response cells"), refractoriness is determined by the voltage-dependent recovery of Na+ channels from inactivation. Refractoriness also frequently is measured by assessing whether premature stimuli applied to tissue preparations (or the whole heart) result in propagated impulses. While the magnitude of the Na+ current is one major determinant of such propagation, cellular geometry (see above) also is important in multicellular preparations. Ordinarily, each cell is connected to many neighbors, so that impulses spread rapidly, and the heart acts like a single large cell, a "syncytium." However, if the geometric arrangement is such that a single cell must supply depolarizing current to many neighbors, conduction can fail. The effective refractory period (ERP) is the shortest interval at which a premature stimulus results in a propagated response and is often used to describe drug effects in intact tissue. Figure 355. Qualitative Differences in Responses of Fast- and Slow-Response Tissues to Premature Stimuli. A. With a very early premature stimulus (black + arrow) in fast-response tissue, all Na channels are still in the inactivated state, and no upstroke results. As the action potential repolarizes, Na+ channels recover from the inactivated to the rest state, from which opening can occur. The phase 0 upstroke slope of the premature action potentials (blue) are greater with later stimuli because recovery from inactivation is voltage-dependent. B. The relationship between transmembrane potential and degree of recovery of Na+ channels from inactivation. The dotted line indicates 25% recovery. Most Na+ channelblocking drugs shift this relationship to the left. C. In slow-response tissues, premature stimuli delivered even after full repolarization of the action potential are depressed; recovery from inactivation is time-dependent.

The situation is different in Ca2+ channeldependent ("slow-response") tissue such as the AV node. The major factor controlling recovery from inactivation of Ca2+ channels is time (Figure 355C). Thus, even after a Ca2+ channeldependent action potential has repolarized back to its initial resting potential, Ca2+ channels are not all available for reexcitation. Therefore, an extrastimulus applied shortly after repolarization is complete results in a reduced Ca2+ current, which may propagate slowly to adjacent cells prior to extinction. An extrastimulus applied later will result in a larger Ca 2+ current and faster propagation. Thus, in Ca2+ channeldependent tissues, which include not only the AV node but also tissues whose underlying characteristics have been altered by factors such as myocardial ischemia, refractoriness is time-dependent, and propagation occurs slowly. Conduction that exhibits such dependence on the timing of premature stimuli is termed "decremental." By contrast, conduction velocity is independent of prematurity in fast-response tissues until a stimulus shorter than the effective refractory period is applied, when it fails completely ("all-or-none" response). Slow conduction in the heart, a critical factor in the genesis of reentrant arrhythmias + (below), also can occur when Na currents are depressed by disease or membrane depolarization + (e.g., elevated [K]o), resulting in decreased steady-state Na channel availability (Figure 355B).

Mechanisms of Cardiac Arrhythmias When the normal sequence of impulse initiation and propagation is perturbed, an arrhythmia occurs. Failure of impulse initiation may result in slow heart rates (bradyarrhythmias), and failure of impulses to propagate normally from atrium to ventricle results in dropped beats or "heart block," which usually reflects an abnormality in either the AV node or the HisPurkinje system. These abnormalities may be caused by drugs (Table 351) or by structural heart disease; in the latter case, permanent cardiac pacing may be required. Abnormally rapid heart rhythms (tachyarrhythmias) are common clinical problems that may be treated with antiarrhythmic drugs. Three major underlying mechanisms have been identified: enhanced automaticity, triggered automaticity, and reentry.

Enhanced Automaticity Enhanced automaticity may occur in cells that normally display spontaneous diastolic depolarizationthe sinus and AV nodes and the HisPurkinje system. -Adrenergic stimulation, hypokalemia, and mechanical stretch of cardiac muscle cells increase phase 4 slope and so accelerate pacemaker rate, whereas acetylcholine reduces pacemaker rate both by decreasing phase 4 slope and by hyperpolarization (making the maximum diastolic potential more negative). In addition, automatic behavior may occur in sites that ordinarily lack spontaneous pacemaker activity; for example, depolarization of ventricular cells (e.g., by ischemia) may produce such "abnormal" automaticity. When impulses propagate from a region of enhanced normal or abnormal automaticity to excite the rest of the heart, arrhythmias result. Afterdepolarizations and Triggered Automaticity Under some pathophysiological conditions, a normal cardiac action potential may be interrupted or followed by an abnormal depolarization (Figure 356). If this abnormal depolarization reaches threshold, it may, in turn, give rise to secondary upstrokes which then can propagate and create abnormal rhythms. These abnormal secondary upstrokes occur only after an initial normal, or "triggering," upstroke and so are termed triggered rhythms. Two major forms of triggered rhythms are recognized: (1) Under conditions of intracellular Ca2+ overload (myocardial ischemia, adrenergic stress, digitalis intoxication), a normal action potential may be followed by a "delayed afterdepolarization" (DAD; Figure 356A). If this afterdepolarization reaches threshold, a secondary triggered beat or beats may occur. DAD amplitude is increased in vitro by rapid pacing, and clinical arrhythmias thought to correspond to DAD-mediated triggered beats are more frequent when the underlying cardiac rate is rapid (Rosen and Reder, 1981). (2) The key abnormality in the second type of triggered activity is marked prolongation of the cardiac action potential. When this occurs, phase 3 repolarization may be interrupted by an "early afterdepolarization" (EAD; Figure 356B). EAD-mediated triggering in vitro and clinical arrhythmias are most common when the underlying heart rate is slow, extracellular K+ is low, and certain drugs (antiarrhythmics and others) that prolong action potential duration are present. EADs represent, by definition, an increase in net inward current during repolarization. However, it is not certain through which channel(s) current flows to generate EADs. When an EAD is present, sympathetic stimulation ( - or -adrenergic) can increase the likelihood of triggered beats. EAD-related triggered upstrokes probably reflect inward current through Na+ or Ca2+ channels. When cardiac repolarization is markedly prolonged, polymorphic ventricular tachycardia with a long QT interval, known as the torsades de pointes syndrome, may occur and is thought to be caused by early afterdepolarizations and resultant triggering (Roden and Hoffman, 1985; Jackman et al., 1988). As mentioned above, the congenital long QT syndrome, a disease in which torsades de pointes are common, is now known to be caused + by mutations in the genes encoding the Na channels or the channels underlying the repolarizing currents IKr and IKs (Roden and Spooner, 1999). Figure 356. Afterdepolarizations and Triggered Activity. A. Delayed afterdepolarization (DAD) arising after full repolarization. A DAD that reaches threshold results in a triggered upstroke (black arrow, right). B. Early afterdepolarization (EAD) interrupting phase 3 repolarization. Under some conditions, triggered beat(s) can arise from an EAD (black arrow, right).

Reentry Anatomically Defined Reentry The principles of cardiac reentry were first described early in the 1900s. Reentry can occur when impulses propagate by more than one pathway between two points in the heart and those pathways have heterogeneous electrophysiological properties. An "experiment of nature" is the Wolff ParkinsonWhite (WPW) syndrome, described in the 1930s. Patients with WPW have accessory connections between the atrium and ventricle (Figure 357). With each sinus node depolarization, impulses can excite the ventricle via the normal structures (AV node) or the accessory pathway. However, the electrophysiological properties of the AV node and accessory pathways are different: accessory pathways consist of fast-response tissue, whereas the AV node is composed of slowresponse tissue. Thus, with a premature atrial beat, conduction may fail in the accessory pathway and continue to conduct, albeit slowly, in the AV node and then through the HisPurkinje system, where the propagating impulse may encounter the ventricular end of the accessory pathway when it is no longer refractory. Note that the likelihood that the accessory pathway is no longer refractory increases as AV nodal conduction slows. When the impulse reenters the atrium, it can then reenter the ventricle via the AV node, reenter the atrium via the accessory pathway, and so on (Figure 35 7). Reentry of this type is therefore determined by (1) the presence of an anatomically defined circuit, (2) heterogeneity in refractoriness among regions in the circuit, and (3) slow conduction in one part of the circuit. Similar "anatomically defined" reentry commonly occurs in the region of the AV node (AV nodal reentrant tachycardia) and in the atrium (atrial flutter). The term paroxysmal supraventricular tachycardia (PSVT) includes both AV reentry and AV nodal reentry, which share many clinical features. In some of these cases, it is now possible to identify and nonpharmacologically ablate critical portions of reentrant pathways (or automatic foci), thus curing the patient and obviating the need for long-term drug therapy. The procedure is carried out through a catheter advanced to the interior of the heart and requires minimal convalescence. Figure 357. Atrioventricular Reentrant Tachycardia in the WolffParkinson White Syndrome. In these patients, an accessory atrioventricular connection is present (light blue). A premature atrial impulse blocks in the accessory pathway (1), and propagates slowly through the AV node and conducting system. Upon reaching the (by now no longer refractory) accessory pathway, it reenters the atrium (2), where it can then reenter the ventricle via the AV node and become self-sustaining (Figure 359C). AV nodal blocking drugs readily terminate this tachycardia. Recurrences can be prevented by drugs that prevent atrial premature beats, by drugs that alter the electrophysiological characteristics of tissue in the circuit (e.g., they prolong AV nodal refractoriness), or by nonpharmacological techniques that section the accessory pathway.

Functionally Defined Reentry Reentry also may occur in the absence of a distinct, anatomically defined pathway (Figure 358). For example, alterations in cellcell coupling following acute myocardial infarction in dogs result in reentrant ventricular tachycardia (VT) whose circuit is dependent not only on postinfarction scarring but also on the rapid longitudinal and slow transverse conduction properties of cardiac tissue (Wit et al., 1990). If ischemia or other electrophysiological perturbations result in an area of sufficiently slow conduction in the ventricle, impulses exiting from that area may find the rest of the myocardium reexcitable, in which case fibrillation may ensue. Atrial or ventricular fibrillation is an extreme example of "functionally defined" (or "leading circle") reentry: cells are reexcited as soon as they are repolarized sufficiently to allow enough Na+ channels to recover from inactivation. In this setting, neither organized activation patterns nor coordinated contractile activity is present. Figure 358. Two Types of Reentry. The border of a propagating wavefront is denoted by a heavy black arrowhead. In anatomically defined reentry (top), a fixed pathway is present (e.g., Figure 357). The black area denotes tissue in the reentrant circuit that is completely refractory because of the recent passage of the propagating wavefront; the gray area denotes tissue in which depressed upstrokes can be elicited (see Figure 355A), and the dark blue area represents tissue in which restimulation would result in action potentials with normal upstrokes. The dark blue area is termed an excitable gap. In functionally defined, or "leading circle," reentry (bottom), there is no anatomic pathway and no excitable gap. Rather, the circulating wavefront creates an area of inexcitable tissue at its core. In this type of reentry, the circuit does not necessarily remain in the same anatomic position during consecutive beats, and multiple such "rotors" may be present.

Common Arrhythmias and Their Mechanisms The primary tool for diagnosis of arrhythmias is the electrocardiogram, although more sophisticated approaches, such as recording from specific regions of the heart during artificial induction of arrhythmias by specialized pacing techniques, sometimes are used. Table 352 lists common arrhythmias, their likely mechanisms, and approaches that should be considered for their acute termination and for long-term therapy to prevent recurrence. Examples of some arrhythmias discussed here are shown in Figure 359. Some arrhythmias, notably ventricular fibrillation (VF), are best treated not with drugs but with DC cardioversionthe application of a large electric current across the chest. This technique also can be used to immediately restore normal rhythm in less serious cases; if the patient is conscious, a brief period of general anesthesia is required. Implantable cardioverter/defibrillators (ICDs), devices that are capable of detecting VF and automatically delivering a defibrillating shock, increasingly are being used in patients who have been resuscitated from one episode of VF. Often drugs are used with these devices to reduce the

need

for defibrillating shocks.

Figure 359. ECGs Showing Normal and Abnormal Cardiac Rhythms. The P, QRS, and T waves in normal sinus rhythm are shown in panel A. Panel B shows a premature beat arising in the ventricle (arrow). Paroxysmal supraventricular tachycardia (PSVT) is shown in panel C; this is most likely reentry utilizing an accessory pathway (Figure 357) or reentry within or near the AV node. In atrial fibrillation (panel D), there are no P waves and the QRS complexes occur irregularly (and at a slow rate in this example); electrical activity between QRS complexes shows small undulations (arrow), corresponding to fibrillatory

activity in the atria. In atrial flutter (panel E), the atria beat rapidly, approximately 250 beats per minute (arrows) in this example, and the ventricular rate is irregular. If a drug that slows the rate of atrial flutter is administered, 1:1 atrioventricular conduction (panel F) can occur. In monomorphic ventricular tachycardia (VT, panel G), identical, wide QRS complexes occur at a regular rate, 180 per minute. The electrocardiographic features of the torsades de pointes syndrome (panel H) include a very long QT interval (>600 ms in this example, arrow) and ventricular tachycardia in which each successive beat has a different morphology (polymorphic VT). Panel I shows the disorganized electrical activity characteristic of ventricular fibrillation. Mechanisms of Antiarrhythmic Drug Action Drug effects that may be antiarrhythmic can be demonstrated in vitro or in animal models, but the relationship between the multiple effects these drugs produce in patients and their effects on arrhythmias (whose mechanisms are only sometimes known) can be complex. A single arrhythmia may result from multiple mechanisms; for example, an automatic or triggered beat may result in a sustained reentrant arrhythmia in a patient with a potential reentrant circuit. Drugs may be antiarrhythmic by suppressing the initiating mechanism or by altering the reentrant circuit. In some cases, however, drugs also may suppress the initiator but nonetheless promote reentry (see below). Drugs may slow automatic rhythms by altering one of the four determinants of spontaneous pacemaker discharge (Figure 3510): maximum diastolic potential, phase 4 slope, threshold potential, or action potential duration. Block of Na + or Ca2+ channels usually results in altered threshold, block of cardiac K+ channels prolongs action potential, adenosine and acetylcholine may increase maximum diastolic potential, and -adrenergic receptor antagonists ( -blockers; see Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists) may decrease phase 4 slope. Figure 3510. Four Ways to Reduce the Rate of Spontaneous Discharge in Automatic Tissues.

Antiarrhythmic drugs may block arrhythmias due to DADs or EADs through two major mechanisms: (1) inhibition of the development of afterdepolarizations or (2) interference with the inward current (usually through Na+ or Ca2+ channels), which is responsible for the upstroke. Thus, for example, arrhythmias due to digitalis-induced DADs may be inhibited by verapamil (which blocks the development of DAD) or by quinidine (which blocks Na+ channels, thus elevating the threshold required to produce the abnormal upstroke). Similarly, two approaches are used in arrhythmias thought to be related to EAD-induced triggered beats (Tables 351 and 352). EADs can be inhibited by shortening action potential duration; in practice, heart rate is accelerated by isoproterenol infusion or by pacing. Triggered beats arising from EADs can be inhibited by Mg 2+, without normalizing repolarization in vitro or QT interval in patients, through mechanisms that are not well understood. In patients with a congenitally prolonged QT interval, torsades de pointes often occur with adrenergic stress; preventive treatment includes -adrenergic blockade (which does not shorten the QT interval) as well as pacing. In anatomically determined reentry, drugs may terminate the arrhythmia by blocking propagation of the action potential. Conduction usually fails in a "weak leak" in the circuit. In the example of the WPW-related arrhythmia described above, the weak link is the AV node, and drugs that prolong AV nodal refractoriness and slow AV nodal conduction, such as Ca2+ channel blockers, -

adrenergic receptor antagonists, or digitalis glycosides, are likely to be effective. On the other hand, conduction slowing in functionally determined reentrant circuits may produce only a change in the pathway without extinguishing the circuit. In fact, slow conduction generally promotes the development of reentrant arrhythmias, and the approach that is most likely to terminate functionally determined reentry is prolongation of refractoriness (Task Force, 1991). In fast-response tissues, refractoriness is prolonged by delaying the recovery of Na+ channels from inactivation. Drugs that act by blocking Na+ channels generally shift the voltage dependence of recovery from block (Figure 355B) and so prolong refractoriness (Figure 3511). Also, drugs that increase action potential duration (achieved without direct action on Na+ channels, e.g., by blocking delayed rectifier currents) also will prolong refractoriness (Figure 3511; Singh, 1993). In slow-response tissues, Ca2+ channel block prolongs refractoriness. Drugs that interfere with cellcell coupling also theoretically should increase refractoriness in multicellular preparations; amiodarone may exert this effect in diseased tissue (Levine et al., 1988). Acceleration of conduction in an area of slow conduction also could be antiarrhythmic in reentry; lidocaine may exert such an effect under some experimental conditions (Arnsdorf and Bigger, 1972). Figure 3511. Two Ways to Increase Refractoriness in Fast-Response Cells. In + this figure, the black dot indicates the point at which a sufficient number of Na channels (an arbitrary 25%, Figure 355B) have recovered from inactivation to allow a premature stimulus to produce a propagated response in the absence of a drug. Block of Na+ channels (A) shifts voltage-dependence of recovery (Figure 355B) and so delays the point at which 25% of channels have recovered (blue diamond), prolonging refractoriness. Note that, if the drug also dissociates slowly from the channel (see Figure 3512), refractoriness in fast-response tissues actually can extend beyond full repolarization ("postrepolarization refractoriness"). Drugs that prolong the action potential (B) also will extend the point at which an arbitrary percentage of Na+ channels have recovered from inactivation, even without directly interacting with Na+ channels.

Figure 3512. Recovery from Block of Na+ Channels during Diastole. This recovery is the critical factor determining extent of steady-state Na+ channel block. Na+ channel blockers bind to (and block) Na + channels in the open and/or inactivated states, resulting in phasic changes in the extent of block during the action potential. As shown in the middle panel, a decrease in the rate of recovery from block increases the extent of block. Different drugs have different rates of recovery, and depolarization reduces the rate of recovery. The right panel shows that increasing heart rate, which results in relatively less time spent in the rest state, also increases the extent of block. (Modified from Roden et al., 1993, with permission.)

State-Dependent Ion Channel Block A key concept in understanding differences in the clinical actions among antiarrhythmic drugs is state-dependent block. Experimental evidence strongly supports the idea that ion channelblocking drugs bind to specific receptor-like sites on the ion channel proteins to modify function (e.g., decrease current) and that, as an ion channel protein shuttles among functional conformations (or ion channel "states"), the affinity of the ion channel protein for the drug on its target site will vary (Hille, 1977; Hondeghem and Katzung, 1984; Snyders et al., 1991). Drugs are thought to gain access to target sites by at least two routes: through the pore (the "hydrophilic" pathway) and through the lipid bilayer (the "hydrophobic" pathway). Physicochemical characteristics, such as molecular weight or lipid solubility, are important determinants of state-dependent binding. Statedependent binding has been studied most extensively in the case of Na+ channelblocking drugs. Most useful agents of this type block open and/or inactivated Na+ channels and have very little affinity for channels in the resting state. Thus, with each action potential, drugs bind to Na+ channels (and block them), and with each diastolic interval, drugs dissociate and "unblock." Block may be due to a drug binding within the conduction pore or to binding at a remote site that then induces changes in the ability of the channel protein to form a pore (an "allosteric" effect). As + illustrated in Figure 3512, the unblocking rate is a key determinant of steady-state block of Na channels. When heart rate increases, the time available for unblocking decreases, and steady-state + Na channel block increases. The rate of recovery from block also slows as cells are depolarized, as in ischemia (Chen et al., 1975). This provides the basis for the finding that Na+ channel blockers + depress Na current, and hence conduction, to a greater extent in ischemic tissues than in normal tissues. Open-versus inactivated-state block also may be important in determining the effects of some drugs. For example, increased action potential duration, which results in a relative increase in time spent in the inactivated state, may increase block by drugs such as lidocaine or amiodarone, which bind to inactivated channels (Hondeghem and Katzung, 1984). The rate of recovery from block often is expressed as a time constant ( recovery, the time required for 63% of an exponentially determined process to be complete; Courtney, 1987). In the case of some

drugs such as lidocaine, recovery is so short (<<1 second) that recovery from block is very rapid, and substantial Na+ channel block occurs only in rapidly driven tissues, particularly in ischemia. At the other end of the spectrum are drugs such as flecainide, with such long recovery (>10 seconds) that roughly the same number of Na+ channels are blocked during systole and diastole. As a result, marked slowing of conduction occurs even in normal tissues, at normal rates. State-dependent block also can be demonstrated for Ca2+ or K+ channelblocking drugs, but the clinical relevance of these findings still is being evaluated. Classifying Antiarrhythmic Drugs Classifying drugs by common electrophysiological properties emphasizes the connection between basic electrophysiological actions and antiarrhythmic effects (Vaughan Williams, 1992). To the extent that the clinical actions of drugs can be predicted from their basic electrophysiological properties, such classification schemes have some merit. However, as each compound is better characterized in a range of in vitro and in vivo test systems, it becomes apparent that, even among drugs that share the same classification, differences in pharmacological effects occur, some of which may be responsible for the observed clinical differences in responses to drugs of the same broad "class" (Table 353). An alternative way of approaching antiarrhythmic therapy is to attempt to classify arrhythmia mechanisms and then to target drug therapy to the electrophysiological mechanism most likely to terminate or prevent the arrhythmia (Table 352; Task Force, 1991). Na+ Channel Block The extent of Na+ channel block is critically dependent on heart rate and membrane potential as well as on drug-specific physicochemical characteristics that determine recovery (Figure 3512). The following description applies when Na+ channels are blocked, i.e., at rapid rates in diseased tissue with a rapid-recovery drug such as lidocaine or even at normal rates in normal tissues with a slowrecovery drug such as flecainide. When Na+ channels are blocked, threshold for excitability is decreased, i.e., greater membrane depolarization is required to bring Na+ channels from the rest to open states. This change in threshold probably contributes to the clinical findings that Na + channel blockers tend to increase both pacing threshold and the energy required to defibrillate the fibrillating heart (Echt et al., 1989). These deleterious effects may be important if antiarrhythmic drugs are used in patients with pacemakers or implanted defibrillators. Na+ channel block decreases conduction velocity in fast-response tissue and increases QRS duration. Usual doses of flecainide prolong QRS intervals by 25% or more during normal rhythm, whereas lidocaine increases QRS intervals only if they are measured at very fast heart rates. Drugs with recovery values >10 seconds (e.g., flecainide) also tend to prolong the PR interval; it is not known whether this represents additional Ca2+ channel block (see below) or block of fast-response tissue in the region of the AV node. Drug effects on the PR interval also are highly modified by autonomic effects. For example, quinidine actually tends to shorten the PR interval, largely as a result of its vagolytic properties. Action potential duration is either unaffected or shortened by Na+ channel block; some Na+ channelblocking drugs do prolong cardiac action potentials, but by other mechanisms, usually K+ channel block (Table 353). By increasing threshold, Na+ channel block decreases automaticity (Figure 3510B) and can inhibit triggered activity arising from delayed afterdepolarizations (DADs) or early afterdepolarizations + (EADs). Many Na channel blockers also decrease phase 4 slope (Figure 3510A). In anatomically + defined reentry, Na channel blockers may decrease conduction sufficiently to extinguish the + propagating reentrant wavefront. However, as described above, conduction slowing due to Na + channel block may exacerbate reentry. Block of Na channels also shifts the voltage dependence of recovery from inactivation (Figure 355B) to more negative potentials, thereby tending to increase

refractoriness. Thus, whether a given drug exacerbates or suppresses a reentrant arrhythmia depends on the balance between its effects on refractoriness and on conduction in a particular reentrant circuit. In most studies, Na+ channel blockers prevent recurrence of reentrant ventricular tachycardia in 20% to 40% of patients. Combining drugs with fast and slow rates of recovery from Na+ channel block (e.g., mexiletine plus quinidine) can be effective, with fewer adverse effects, when neither drug alone is effective. Lidocaine and similar agents (mexiletine, tocainide, phenytoin) with short recovery values are not useful in atrial fibrillation or flutter, whereas quinidine, flecainide, propafenone, and similar agents are effective in some patients. Many of these agents owe part of their antiarrhythmic activity to blockade of K+ channels. Na+ Channel Blocker Toxicity Conduction slowing in potential reentrant circuits can account for toxicity due to Na+ channel block (Table 351). For example, Na+ channel block decreases conduction velocity and hence slows atrial flutter rate. Normal AV nodal function permits a greater number of impulses to penetrate the ventricle, and heart rate actually may increase (Figure 359). Thus, atrial flutter rate may drop from 300 per minute, with 2:1 or 4:1 atrioventricular conduction (i.e., a heart rate of 75 or 150 per minute), to 220 per minute, but with 1:1 transmission to the ventricle (i.e., a heart rate of 220 per minute). This form of drug-induced arrhythmia is especially common during treatment with quinidine because the drug also increases AV nodal conduction through its vagolytic properties; flecainide and propafenone also have been implicated. Therapy with Na+ channel blockers in patients with reentrant ventricular tachycardia after a myocardial infarction can increase the frequency and severity of arrhythmic episodes, presumably because conduction slowing allows persistence of the reentrant wavefront within the tachycardia circuit. Such drug-exacerbated arrhythmia can be very difficult to manage, and deaths due to intractable drug-induced ventricular tachycardia have been reported. Some studies suggest Na+ infusion may be beneficial. Several Na+ channel blockers (procainamide, quinidine) have been reported to exacerbate neuromuscular paralysis by d-tubocurarine (see Chapter 9: Agents Acting at the Neuromuscular Junction and Autonomic Ganglia). Action Potential Prolongation Most drugs producing this effect do so by blocking K+ channels, although enhanced inward Na+ current also can prolong action potentials. Enhanced inward current may underlie QT prolongation (and arrhythmia suppression) by ibutilide. Block of cardiac K + channels increases action potential duration and reduces normal automaticity (Figure 3510 D). Increased action potential duration, seen as an increase in QT interval, increases refractoriness (Figure 3511), which should be an effective way of treating reentry (Task Force, 1991; Singh, 1993). Some studies have shown that K+ channel block reduces heterogeneity of refractoriness, an effect that also should prevent reentrant arrhythmias. Experimentally, K+ channel block produces a series of desirable effects: reduced defibrillation energy requirement, inhibition of ventricular fibrillation due to acute ischemia, and increased contractility (Echt et al., 1989; Roden, 1993). As shown in Table 353, most K+ channel blocking drugs also interact with -adrenergic receptors (sotalol) or other channels (e.g., amiodarone, quinidine). Amiodarone and sotalol appear to be at least as effective as drugs with + predominant Na channelblocking properties in both atrial and ventricular arrhythmias. "Pure" action potentialprolonging drugs dofetilide, ibutilide) recently have become available (Murray, 1998; Torp-Pedersen et al., 1999). Toxicity of Drugs That Prolong QT Interval Most of these agents prolong cardiac action potentials to a disproportionate extent when underlying

heart rate is slow; this effect, in turn, results in EADs and related triggered activity in vitro, and can cause torsades de pointes (Table 351; Figure 359). In patients being treated for atrial fibrillation, torsades de pointes occur after conversion to sinus rhythm. For unknown reasons, this form of antiarrhythmic drug toxicity is significantly more common in women (Makkar et al., 1993). Ca2+ Channel Block The major electrophysiological effects resulting from block of cardiac Ca2+ channels are in slowresponse tissues, the sinus and AV nodes. Dihydropyridines, such as nifedipine, commonly used in angina and hypertension (Chapters 32: Drugs Used for the Treatment of Myocardial Ischemia and 33: Antihypertensive Agents and the Drug Therapy of Hypertension), preferentially block Ca 2+ channels in vascular smooth muscle; their cardiac electrophysiological effects, such as heart rate acceleration, are indirect and attributable to sympathetic activation. Only verapamil, diltiazem, and bepridil block Ca 2+ channels in cardiac cells at clinically used doses. With these drugs, heart rate generally is slowed (Figure 3510A), although hypotension, if marked, can cause reflex sympathetic activation and tachycardia. AV nodal conduction velocity decreases, so the PR interval increases. AV nodal block occurs as a result of decremental conduction as well as increased AV nodal refractoriness. These latter effects form the basis of the antiarrhythmic actions of Ca2+ channel blockers in reentrant arrhythmias whose circuit involves the AV node, such as AV reentrant tachycardia (Figure 357). Another important antiarrhythmic action is reduction of ventricular rate in atrial flutter or fibrillation. Rare forms of ventricular tachycardia appear to be DAD-mediated and respond to verapamil (Sung et al., 1983). Unlike -adrenergic receptor antagonists, Ca2+ channel blockers have not been shown to reduce mortality in patients convalescing from myocardial infarction (Singh, 1990). In contrast to other Ca 2+ channel blockers, bepridil increases action potential duration in many tissues and can exert an antiarrhythmic effect in atria and ventricles. However, because of the incidence of torsades de pointes following the administration of bepridil, the drug is not widely prescribed. Verapamil and Diltiazem The major adverse effect of intravenous verapamil or diltiazem is hypotension, particularly with bolus doses. This is a particular problem if the drugs are mistakenly used in patients with ventricular tachycardia (in which Ca2+ channel blockers are not usually effective) misdiagnosed as AV nodal reentrant tachycardia (Stewart et al., 1986). Hypotension also is frequent in patients receiving other vasodilators, including quinidine, and in patients with underlying left ventricular dysfunction, which the drugs can exacerbate. Severe sinus bradycardia or heart block also occurs, especially in susceptible patients, such as those also receiving -blockers. With oral therapy, these adverse effects tend to be less severe. Constipation can occur with oral verapamil. Verapamil (CALAN, ISOPTIN , VERELAN, COVERA -HS) is prescribed as a racemate. l-Verapamil is a more potent calcium channel blocker than is d-verapamil. However, with oral therapy, the lenantiomer undergoes more extensive first-pass hepatic metabolism. For this reason, a given concentration of verapamil prolongs the PR interval to a greater extent when the drug is administered intravenously (where concentrations of the l- and d-enantiomers are equivalent) than when it is administered orally (Echizen et al., 1985). Diltiazem (CARDIZEM, TIAZAC, DILACORXR , and others) also undergoes extensive first-pass hepatic metabolism, and both drugs have metabolites that exert Ca2+ channelblocking actions. In clinical practice, adverse effects during therapy with verapamil or diltiazem are determined largely by underlying heart disease and concomitant therapy; plasma concentrations of these agents are not routinely measured during therapy. Both drugs can

increase serum digoxin concentration, although the magnitude of this effect is variable; excess slowing of ventricular response in patients with atrial fibrillation can occur. Block of -Adrenergic Receptors -Adrenergic stimulation increases the magnitude of the Ca2+ current and slows its inactivation, increases the magnitude of repolarizing K+ and Cl currents (Sanguinetti et al., 1991; Hume and Harvey, 1991), increases pacemaker current (thereby increasing sinus rate; DiFrancesco, 1993), andunder pathophysiological conditionscan increase both DAD- and EAD-mediated arrhythmias. Also, increases in plasma epinephrine associated with severe stress, such as acute myocardial infarction or resuscitation from cardiac arrest, lower serum K+, especially in patients receiving chronic diuretic therapy (Brown et al., 1983). Thus, -adrenergic receptor antagonists (often referred to as -blockers), which inhibit these effects, can be antiarrhythmic by reducing heart rate, decreasing intracellular Ca2+ overload, and inhibiting afterdepolarization- mediated automaticity. Epinephrine-induced hypokalemia appears to be mediated by 2-adrenergic receptors and is blocked by "noncardioselective" antagonists such as propranolol (see Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists). In acutely ischemic tissue, -blockers increase the energy required to fibrillate the heart, an antiarrhythmic action (Anderson et al., 1983). These effects may contribute to the reduction in mortality observed in trials of chronic therapy with -blockersincluding propranolol, timolol, and metoprololafter myocardial infarction (Singh, 1990), although the precise mechanism underlying this effect has not been established. Atenolol and metoprolol have been shown to decrease mortality in the first week following myocardial infarction. As with Ca2+ channel blockers and digitalis, a major effect of -blocker therapy is increased AV nodal conduction time (increased PR interval) and prolonged AV nodal refractoriness. Hence, blockers are useful in terminating reentrant arrhythmias that involve the AV node and in controlling ventricular response in atrial fibrillation or flutter. In many (but not all) patients with the congenital long QT syndrome, as well as in many other patients, arrhythmias are triggered by physical or emotional stress; -blockers may be useful in these cases (Schwartz et al., 2000; Roden and Spooner, 1999). -Adrenergic receptor antagonists also have been reported to be effective in controlling arrhythmias due to Na+ channel blockers; this effect may be due in part to slowing of the heart rate, which then decreases the extent of rate-dependent conduction slowing by Na+ channel block (Myerburg et al., 1989). As described further in Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists, adverse effects due to -blocker therapy include fatigue, bronchospasm, impotence, depression, aggravation of heart failure, worsening of symptoms due to peripheral vascular disease, and inhibition of the symptoms of hypoglycemia in diabetic patients. In patients with arrhythmias due to excess sympathetic stimulation (e.g., pheochromocytoma, clonidine withdrawal), -blockers may in theory result in unopposed -adrenergic stimulation, with resultant severe hypertension and/or -adrenergic mediated arrhythmias. In such patients, arrhythmias should be treated with both - and -adrenergic antagonists. Abrupt discontinuation of chronic -blocker therapy can lead to "rebound" symptoms including hypertension, increased angina, and arrhythmias (see Chapter 33: Antihypertensive Agents and the Drug Therapy of Hypertension). Selected -Adrenergic Receptor Blockers It is likely that most -adrenergic antagonists share antiarrhythmic properties. Some, such as + propranolol, can be shown to exert Na channel-blocking ("membrane stabilizing") effects at high concentrations in vitro (Davis and Temte, 1968), but whether or not this is important in patients is uncertain. Similarly, drugs with intrinsic sympathomimetic activity may be less useful as

antiarrhythmics, at least in theory (Singh, 1990). Acebutolol is as effective as quinidine in suppressing ventricular ectopic beats, an arrhythmia that many clinicians now do not treat. Sotalol (see below) is more effective for many arrhythmias than are other -blockers, probably because of its additional K+ channelblocking actions. Esmolol (Frishman et al., 1988) is a cardioselective agent that is metabolized by red cell esterases and so has a very short elimination half-life (9 minutes). Although methanol is a metabolite, methanol intoxication has not been a clinical problem. Intravenous esmolol is useful in clinical situations in which immediate -adrenergic blockade is desired (e.g., for rate control of rapidly conducted atrial fibrillation). Because of esmolol's very rapid elimination, adverse effects due to -adrenergic blockadeshould they occurdissipate rapidly when the drug is stopped. With most -blockers, intravenous therapy can produce transient hypertension due to unopposed -adrenergic stimulation; with intravenous esmolol, hypotension is more common. Principles in the Clinical Use of Antiarrhythmic Drugs Drugs that modify cardiac electrophysiology often have a very narrow margin between the doses required to produce a desired effect and those associated with adverse effects. Moreover, adverse effects from antiarrhythmic drug therapy can include induction of new arrhythmias, with possibly fatal consequences. Nonpharmacological treatments, such as cardiac pacing, electrical defibrillation, or ablation of targeted regions (Morady, 1999), are indicated for some arrhythmias; in other cases no therapy is required even though an arrhythmia is detected. Therefore, the fundamental principles of therapeutics described here must be applied to optimize antiarrhythmic therapy. Identify and Remove Precipitating Factors Factors that commonly precipitate cardiac arrhythmias include hypoxia, electrolyte disturbances (especially hypokalemia), myocardial ischemia, and certain drugs. Antiarrhythmics, including digitalis glycosides, are not the only drugs that can precipitate arrhythmias (Table 351). For example, theophylline is a common cause of multifocal atrial tachycardia, which sometimes can be managed simply by reducing the dose of theophylline. Torsades de pointes can arise not only during therapy with action potential-prolonging antiarrhythmics, but also with other drugs not ordinarily classified as having effects on ion channels. These include the antihistamines terfenadine and astemizole (Chapter 25: Histamine, Bradykinin, and Their Antagonists); the antibiotic, erythromycin (Chapter 47: Antimicrobial Agents: Protein Synthesis Inhibitors and Miscellaneous Antibacterial Agents); the antiprotozoal, pentamidine (Chapter 41: Drugs Used in the Chemotherapy of Protozoal Infections: Amebiasis, Giardiasis, Trichomoniasis, Trypanosomiasis, Leishmaniasis, and Other Protozoal Infections); some antipsychotics, notably thioridazine (Chapter 20: Drugs and the Treatment of Psychiatric Disorders: Psychosis and Mania); and certain tricyclic antidepressants (Chapter 19: Drugs and the Treatment of Psychiatric Disorders: Depression and Anxiety Disorders). Establish the Goals of Treatment Some Arrhythmias Should Not Be Treated: The CAST Example Abnormalities of cardiac rhythm are readily detectable by a variety of recording methods. However, the mere detection of an abnormality should not be equated with the need for therapy. This was best illustrated in the Cardiac Arrhythmias Suppression Trial (CAST). The presence of asymptomatic ventricular ectopic beats is known to be a marker for an increased risk of sudden death due to ventricular fibrillation in patients convalescing from a myocardial infarction. In the CAST, patients

in whom ventricular ectopic beats were suppressed by the potent Na+ channel blockers encainide (no longer marketed) or flecainide were randomly assigned to receive those drugs or a matching placebo. Unexpectedly, the mortality rate was two-to threefold higher among patients treated with the drugs than those treated with placebo (CAST Investigators, 1989). While the explanation of this effect is not known, a number of lines of evidence suggest that, in the presence of these drugs, transient episodes of myocardial ischemia and/or sinus tachycardia can cause marked conduction slowing (because these drugs have a very long recovery), resulting in fatal reentrant ventricular tachyarrhythmias (Ruskin, 1989; Ranger et al., 1989; Akiyama et al., 1991). One consequence of this very important clinical trial was to reemphasize the concept that therapy should be initiated only when a clear benefit to the patient can be identified. When symptoms are obviously attributable to an ongoing arrhythmia, there usually is not much doubt that termination of the arrhythmia will be beneficial; when chronic therapy is used to prevent recurrence of an arrhythmia, risks may be greater (Roden, 1994). Among the antiarrhythmic drugs discussed here, only adrenergic blockers, and to a lesser extent amiodarone (Connolly, 1999), have been shown to reduce mortality during long-term therapy. Symptoms Due to Arrhythmias Some patients with an arrhythmia may be asymptomatic; in this case, establishing any benefit for treatment will be very difficult. Some patients may present with presyncope, syncope, or even cardiac arrest, which may be due to brady- or tachyarrhythmias. Other patients may present with a sensation of irregular heartbeats, which can be minimally symptomatic in some individuals and incapacitating in others. The irregular heartbeats may be due to intermittent premature contractions or to sustained arrhythmias such as atrial fibrillation (which results in an irregular ventricular rate; Figure 359). Finally, patients may present with symptoms due to decreased cardiac output attributable to arrhythmias. The most common symptom is breathlessness at rest or on exertion. Rarely, patients with sustained tachycardias will present with congestive heart failure, which can be controlled by treating the arrhythmia. Choosing Among Therapeutic Approaches Establishing the goals of therapy is especially important when different therapeutic options are available. For example, in patients with atrial fibrillation, three options are available: (1) Reduce the ventricular response, using AV nodal blocking agents such as digitalis, verapamil, diltiazem, or adrenergic antagonists (Table 351); (2) restore and maintain normal rhythm, using drugs such as quinidine, flecainide, or amiodarone; or (3) decide not to implement antiarrhythmic therapy, which may be the appropriate approach if the patient truly is asymptomatic. Most patients with atrial fibrillation also benefit from anticoagulation to reduce stroke incidence, regardless of symptoms (Singer, 1996; see Chapter 55: Anticoagulant, Thrombolytic, and Antiplatelet Drugs). Factors that contribute to choice of therapy include not only symptoms but also the type and extent of structural heart disease, the QT interval prior to drug therapy, the coexistence of conduction system disease, and the presence of noncardiac diseases (Table 354). In the rare patient with the WPW syndrome and atrial fibrillation, the ventricular response can be extremely rapid and can be paradoxically accelerated by AV nodal blocking drugs such as digitalis or Ca2+ channel blockers; deaths due to drug therapy have been reported under these circumstances. Frequency and reproducibility of arrhythmia should be established prior to initiating therapy, since inherent variability in the occurrence of arrhythmias can be confused with a beneficial or adverse drug effect. Techniques for this assessment include recording cardiac rhythm for prolonged periods or evaluating the response of the heart to artificially induced premature beats. It also is important to

recognize that drug therapy may be only partially effective: a marked decrease in the duration of paroxysms of atrial fibrillation may be sufficient to render a patient asymptomatic, even if an occasional episode still can be detected. Minimize Risks Antiarrhythmic Drugs Can Cause Arrhythmias One increasingly well-recognized risk of antiarrhythmic therapy is the possibility of provoking new arrhythmias, with potentially life-threatening consequences. Mechanistically distinct syndromes of arrhythmia provocation by antiarrhythmic drugs have been described (Table 351). These drugprovoked arrhythmias must be recognized, since further treatment with antiarrhythmic drugs often exacerbates the problem, whereas withdrawal of the causative agent often is curative. In addition, specific therapies targeted toward underlying mechanisms of these arrhythmias may be indicated. It also is critical to establish a precise diagnosis. For example, treating a ventricular tachycardia with verapamil not only may be ineffective but also can cause catastrophic cardiovascular collapse (Stewart et al., 1986). Monitoring of Plasma Concentration Some adverse effects of antiarrhythmic drugs are related to excessively elevated plasma drug concentration. Thus, measuring plasma concentration and adjusting the dose to maintain the concentration within a prescribed therapeutic range may be a useful way of minimizing some adverse effects. In many patients, the occurrence of serious adverse reactions appears to be related to interactions involving drug (often at usual plasma concentrations), transient factors such as electrolyte disturbances or myocardial ischemia, and the type and extent of the underlying heart disease (Ruskin, 1989; Morganroth et al., 1986; Roden, 1994). Factors such as generation of unmeasured active metabolites, variability in elimination of enantiomers (which may exert differing pharmacological effects), and disease-or enantiomer-specific abnormalities in drug binding to plasma proteins can complicate the interpretation of routine monitoring of plasma drug concentrations (see Chapter 1: Pharmacokinetics: The Dynamics of Drug Absorption, Distribution, and Elimination). Patient-Specific Contraindications Another way of minimizing the adverse effects of antiarrhythmic drugs is to avoid certain drugs in certain patient subsets altogether. For example, patients with a history of congestive heart failure are particularly prone to develop heart failure during disopyramide therapy. Often, adverse effects of drugs may be difficult to distinguish from exacerbations of underlying disease. Amiodarone may cause interstitial lung disease; its use is therefore undesirable in a patient with advanced pulmonary disease in whom the development of this potentially fatal adverse effect would be difficult to detect. Specific diseases that constitute relative or absolute contraindications to specific drugs are listed in Table 354. The Electrophysiology of the Heart as a "Moving Target" Cardiac electrophysiology varies in a highly dynamic fashion in response to external influences such as changing autonomic tone, myocardial ischemia, or myocardial stretch. For example, myocardial ischemia results in changes in extracellular K+ that, in turn, make the resting potential less negative, inactivate Na+ channels, decrease Na+ current, and slow conduction (Weiss et al., 1991). In addition, myocardial ischemia can result in the release of "metabolites of ischemia," such

as lysophosphatidylcholine, which can alter ion channel function (DaTorre et al., 1991); ischemia also may activate channels that otherwise are quiescent, such as the ATP-inhibited K+ channels (Wilde and Janse, 1994). Thus, a normal heart may display, in response to myocardial ischemia, changes in resting potential, conduction velocity, intracellular Ca2+ concentrations, and repolarization, any one of which may then create arrhythmias or alter response to antiarrhythmic therapy. Antiarrhythmic Drugs Summaries of important electrophysiological and pharmacokinetic features of the drugs considered here are presented in Tables 353 and 355. Ca2+ channel blockers and -adrenergic antagonists have been considered above and in Chapters 32: Drugs Used for the Treatment of Myocardial Ischemia and 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists, respectively. The drugs are presented in alphabetical order. Adenosine Adenosine (ADENOCARD) is a naturally occurring nucleoside which is administered as a rapid intravenous bolus for the acute termination of reentrant supraventricular arrhythmias (Lerman and Belardinelli, 1991). Rare cases of ventricular tachycardia in patients with otherwise normal hearts are thought to be DAD-mediated and can be terminated by adenosine. Adenosine also has been used to produce controlled hypotension during some surgical procedures and in the diagnosis of coronary artery disease. Intravenous ATP appears to produce effects similar to those of adenosine.

Pharmacological Effects The effects of adenosine are mediated by its interaction with specific G proteincoupled adenosine receptors. Adenosine activates acetylcholine-sensitive K+ current in the atrium and sinus and AV nodes, resulting in shortening of action potential duration, hyperpolarization, and slowing of normal automaticity (Figure 3510C). Adenosine also inhibits the electrophysiological effects of increased intracellular cyclic AMP, which occur with sympathetic stimulation. Because adenosine thereby reduces Ca2+ currents, it can be antiarrhythmic by increasing AV nodal refractoriness and by inhibiting DADs elicited by sympathetic stimulation. Administration of an intravenous bolus of adenosine to human beings transiently slows sinus rate and AV nodal conduction velocity and increases AV nodal refractoriness. It has been shown that a bolus dose of adenosine can produce transient sympathetic activation by interacting with carotid baroreceptors (Biaggioni et al., 1991); when a continuous infusion of the drug is administered,

hypotension results. Adverse Effects A major advantage of adenosine therapy is that adverse effects are short-lived, since the drug is eliminated so rapidly. Transient asystole (lack of any cardiac rhythm whatsoever) is common but usually lasts <5 seconds and is in fact the therapeutic goal. Most patients feel a sense of chest fullness and dyspnea when therapeutic doses (6 to 12 mg) of adenosine are administered. Rarely, an adenosine bolus can precipitate atrial fibrillation, presumably by heterogeneously shortening atrial action potentials, or bronchospasm. Clinical Pharmacokinetics Adenosine is eliminated with a half-life of seconds by carrier-mediated uptake, which occurs in most cell types including the endothelium, and subsequent metabolism by adenosine deaminase. Adenosine probably is the only drug whose efficacy requires a rapid bolus dose, preferably through a large central intravenous line; slow administration results in elimination of the drug prior to its arrival at the heart. The effects of adenosine are potentiated in patients receiving dipyridamole, an adenosine-uptake inhibitor, and in patients with cardiac transplants, due to denervation hypersensitivity. Methylxanthines (see Chapter 28: Drugs Used in the Treatment of Asthma) such as theophylline and caffeine block adenosine receptors; therefore larger-than-usual doses are required to produce an antiarrhythmic effect in patients who have consumed these agents in beverages or as therapy. Amiodarone Amiodarone (CORDARONE, PACERONE ) exerts a multiplicity of pharmacological effects, none of which is clearly linked to its arrhythmia-suppressing properties (Mason, 1987). Amiodarone is a structural analog of thyroid hormone, and some of its antiarrhythmic actions and its toxicity may be attributable to interaction with nuclear thyroid hormone receptors. Amiodarone is highly lipophilic, is concentrated in many tissues, and is eliminated extremely slowly; consequently, adverse effects may be very slow to resolve. In the United States, the drug currently is indicated for oral therapy in patients with recurrent ventricular tachycardia or fibrillation resistant to other drugs. Controlled clinical trials indicate that oral amiodarone also is effective in maintaining sinus rhythm in patients with atrial fibrillation (Connolly, 1999). An intravenous form has been approved in the United States for the acute termination of ventricular tachycardia or fibrillation (Kowey et al., 1995). Trials of oral amiodarone in patients convalescing from acute myocardial infarction have shown a modest beneficial effect on mortality (Amiodarone Trials Meta-Analysis Investigators, 1997). The drug is nevertheless not widely prescribed in this population, perhaps because of concerns about its potential for long-term toxicity.

Pharmacological Effects

Studies of the acute effects of amiodarone in in vitro systems are complicated by its insolubility in water, necessitating the use of solvents such as dimethyl sulfoxide. It has been suggested that amiodarone's effects are mediated by perturbation of the lipid milieu in which ion channels are placed (Herbette et al., 1988). Amiodarone blocks inactivated Na + channels and has a relatively rapid (time constant 1.6 seconds) rate of recovery from block. It also decreases Ca 2+ current and transient outward, delayed rectifier and inward rectifier K+ currents and exerts a noncompetitive (see Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists) adrenergic blocking effect. Amiodarone is a potent inhibitor of abnormal automaticity, and, in most tissues, it prolongs action potential duration. Amiodarone decreases conduction velocity by Na+ channel block as well as by a poorly understood effect on cellcell coupling that may be especially important in diseased tissue (Levine et al., 1988). PR, QRS, and QT prolongations and sinus bradycardia are frequent during chronic therapy. Amiodarone prolongs refractoriness in all cardiac tissues; Na+ channel block, delayed repolarization due to K+ channel block, and inhibition of cell cell coupling all may contribute to this effect of amiodarone. Adverse Effects Hypotension due to vasodilation and depression of myocardial performance is frequent with the intravenous form of amiodarone, and may be due in part to the solvent. While depression of contractility can occur during long-term therapy, it is unusual. During oral drug-loading regimens, which usually take place over several weeks, adverse effects are unusual, despite administration of high dosages that would cause serious toxicity if continued long-term. Occasional patients develop nausea, which responds to a decrease in daily dose during the loading phase. Adverse effects during long-term therapy have been related to both the size of daily maintenance doses as well as to cumulative dose (i.e., to duration of therapy), suggesting tissue accumulation may be responsible. The most serious adverse effect during chronic amiodarone therapy is pulmonary fibrosis, which can be rapidly progressive and fatal. Underlying lung disease, doses 400 mg/day, and recent pulmonary insults such as pneumonia appear to be risk factors (Dusman et al., 1990). Serial chest X-rays or pulmonary function studies may detect early amiodarone toxicity, but monitoring plasma concentrations has not been useful. With low doses, such as 200 mg/day used in atrial fibrillation, pulmonary toxicity is extremely unusual. Other adverse effects during long-term therapy include corneal microdeposits (which often are asymptomatic), hepatic dysfunction, hypo- or hyperthyroidism, neuromuscular symptoms (most commonly peripheral neuropathy or proximal muscle weakness), and photosensitivity. Treatment consists of withdrawal of the drug and supportive measures, including corticosteroids, for life-threatening toxicity; reduction of dosage may be sufficient if the drug is deemed necessary and the adverse effect is not life-threatening. Despite the marked QT prolongation and bradycardia typical of chronic amiodarone therapy, torsades de pointes and other drug-induced tachyarrhythmias are unusual. Clinical Pharmacokinetics Amiodarone is incompletely ( 30%) bioavailable, presumably because of poor absorption. This reduced bioavailability is important in calculating equivalent dosing regimens when converting from intravenous to oral therapy. The drug is distributed in lipid; for example, heart tissue to plasma concentration ratios >20:1 and lipid to plasma ratios >300:1 have been reported. After the initiation of amiodarone therapy, increases in refractoriness, a marker of pharmacological effect, require several weeks to develop. Amiodarone undergoes hepatic metabolism by cytochrome P450 3A4 (CYP3A4) to desethyl-amiodarone, a metabolite with pharmacological effects similar to those of the parent drug. When amiodarone therapy is withdrawn from a patient who has been receiving therapy for several years, plasma concentrations decline with a half-life of weeks to months. The

mechanism whereby amiodarone and desethyl-amiodarone are eliminated is not well established. A therapeutic plasma amiodarone concentration range of 0.5 to 2.0 g/ml has been proposed. However, efficacy appears to depend as much on duration of therapy as on plasma concentration, and elevated plasma concentrations are not useful in predicting toxicity (Dusman et al., 1990). Because of amiodarone's slow accumulation in tissue, a high-dose oral loading regimen (e.g., 800 to 1600 mg/day) is usually administered for several weeks and then maintenance therapy started. Maintenance dose is adjusted on the basis of adverse effects and the arrhythmias being treated. If the presenting arrhythmia is life-threatening, dosages >300 mg/day normally are used unless clear toxicity occurs. On the other hand, maintenance doses of 200 mg/day are used if recurrence of an arrhythmia would be tolerated, as in patients with atrial fibrillation. Because of the drug's very slow elimination, amiodarone is administered once daily, and omission of one or two doses during chronic therapy rarely results in recurrence of arrhythmia. Dose adjustments are not required in conditions such as hepatic, renal, or cardiac dysfunction. Amiodarone is a potent inhibitor of the hepatic metabolism or renal elimination of many compounds. Mechanisms identified to date include inhibition of CYP3A4 and CYP2C9 and of Pglycoprotein (see Chapter 1: Pharmacokinetics: The Dynamics of Drug Absorption, Distribution, and Elimination). Dosages of warfarin, other antiarrhythmics (flecainide, procainamide, quinidine), or digoxin usually require reduction during amiodarone therapy. Bretylium Bretylium (bretylium tosylate; BRETYLOL) is a quaternary ammonium compound that prolongs cardiac action potentials and interferes with reuptake of norepinephrine by sympathetic neurons; both actions may be antiarrhythmic (Heissenbuttel and Bigger, 1979). Bretylium loading and maintenance infusions are used to treat ventricular fibrillation and prevent its recurrence (see Table 355 for loading and maintenance doses).

Pharmacological Effects Bretylium prolongs action potentials in normal Purkinje cells to a greater extent than in cells that have survived a recent ischemic insult (in which action potentials already are prolonged abnormally). Thus, bretylium reduces heterogeneity of repolarization times, an effect that may suppress reentry (Cardinal and Sasyniuk, 1978). The mechanism whereby bretylium prolongs cardiac action potentials has not been established, although block of K+ channels seems likely. Bretylium has no effect on Na+ channels, except at high concentrations, and no direct effect on automaticity. In animals and human beings, administration of bretylium initially results in increased norepinephrine release from sympathetic neurons and inhibition of subsequent reuptake. Adverse Effects As a result of norepinephrine release, bretylium can produce transient hypertension and increased arrhythmias; this effect rarely is observed, since bretylium is used in critically ill patients who often

are hemodynamically unstable. In theory, bretylium should be avoided in patients who are especially prone to increased arrhythmias with norepinephrine release, such as those with digitalis intoxication. In contrast, hypotension due to inhibition of norepinephrine reuptake is a common problem during bretylium therapy. Bretylium-induced hypotension should be managed with judicious fluid replacement if possible. Since bretylium effectively results in sympathetic denervation, the administration of normal doses of catecholamines such as dopamine may cause marked hypertension. Bretylium should be used only with great caution when the drug's vasodilating effects may be particularly hazardous, as in patients with aortic stenosis, carotid occlusive disease, or hypertrophic cardiomyopathy. The occurrence of torsades de pointes is unusual during bretylium therapy. Clinical Pharmacokinetics Bretylium is excreted unchanged by the kidneys without undergoing significant hepatic metabolism. Reduction of a maintenance infusion rate has been recommended in patients with renal failure, although adverse effects due to accumulation of bretylium in plasma have not been seen. A lag time of 2 hours has been reported between peak plasma bretylium concentrations and peak prolongation of ventricular refractoriness after an intravenous dose in dogs (Anderson et al., 1980). This lag time suggests that bretylium is distributed to sites in peripheral tissues prior to exerting its pharmacological effect. An oral formulation has been investigated, but hypotension and reduced bioavailability ( 35%), presumably because of poor absorption, are problems. The hypotensive effects of bretylium can be inhibited by coadministration of tricyclic antidepressants such as protriptyline, which blocks bretylium's effects on sympathetic norepinephrine release and reuptake (Woosley et al., 1982). Results of a limited number of studies suggest that the antiarrhythmic effects of bretylium may be preserved during coadministration of tricyclic antidepressants. Digitalis Glycosides Pharmacological Effects Digitalis glycosides exert positive inotropic effects and are widely used in heart failure (Chapter 34: Pharmacological Treatment of Heart Failure). Their inotropic action is the result of increased intracellular Ca2+ (Smith, 1988), which also forms the basis for arrhythmias related to digitalis intoxication. Digitalis glycosides increase phase 4 slope (i.e., increase the rate of automaticity), especially if [K]o is low. Digitalis glycosides also exert prominent vagotonic actions, resulting in inhibition of Ca2+ currents in the AV node and activation of acetylcholine-mediated K+ currents in the atrium. Thus, the major "indirect" electrophysiological effects of digitalis glycosides are hyperpolarization, shortening of atrial action potentials, and increases in AV nodal refractoriness. The latter action accounts for the utility of digitalis in termination of reentrant arrhythmias involving the AV node, and in controlling ventricular response in patients with atrial fibrillation. Digitalis preparations may be especially useful in the latter situation, since many such patients have heart failure, which can be exacerbated by other AV nodal blocking drugs such as Ca2+ channel blockers or -adrenergic receptor antagonists. However, in many patients with advanced heart failure, sympathetic drive is markedly increased, so digitalis is not very effective in decreasing the rate; on the other hand, even a modest decrease in rate can sometimes help ameliorate heart failure. Similarly, in other conditions in which high sympathetic tone drives rapid atrioventricular conduction (e.g., chronic lung disease, thyrotoxicosis), digitalis therapy may be only marginally effective in slowing the rate. Increased sympathetic activity and hypoxia can potentiate digitalisinduced changes in automaticity and DADs, thus increasing the risk of digitalis toxicity. A further

complicating feature in thyrotoxicosis is increased digoxin clearance. The major ECG effects of cardiac glycosides are PR prolongation and a nonspecific alteration in ventricular repolarization (the ST segment), whose underlying mechanism is not well understood. Adverse Effects Digitalis intoxication is a common clinical problem (see also Chapter 34: Pharmacological Treatment of Heart Failure). Arrhythmias, nausea, disturbances of cognitive function, and blurred or yellow vision are the usual manifestations. Often, these are not immediately recognized, since they occur frequently in patients with advanced illnesses receiving multiple drugs. Elevated serum concentrations of digitalis, hypoxia (e.g., due to chronic lung disease), and hypokalemia, hypomagnesemia, and hypercalcemia predispose patients to digitalis-induced arrhythmias. While digitalis intoxication can cause virtually any arrhythmia, certain types of arrhythmias are characteristic. Arrhythmias that should raise a strong suspicion of digitalis intoxication are those in which DAD-related tachycardias occur along with impairment of sinus node or AV nodal function. Atrial tachycardia with AV block is "classic," but ventricular bigeminy (sinus beats alternating with beats of ventricular origin), "bidirectional" ventricular tachycardia (a very rare entity), AV junctional tachycardias, and various degrees of AV block also can occur. With advanced intoxication (e.g., with suicidal ingestion), severe hyperkalemia due to poisoning of Na+,K+ ATPase and profound bradyarrhythmias, which may be unresponsive to pacing therapy, are seen. In patients with elevated serum digitalis blood levels, the risk of precipitating ventricular fibrillation by DC cardioversion probably is increased; in those with therapeutic blood levels, DC cardioversion can be used safely. Minor forms of digitalis intoxication may require no specific therapy beyond monitoring cardiac rhythm until symptoms and signs of toxicity resolve. Sinus bradycardia and AV block often respond to intravenous atropine, but the effect is transient. Mg2+ has been used successfully in some cases of digitalis-induced tachycardia (Seller, 1971). Any serious arrhythmia should be treated with antidigoxin Fab fragments, which are highly effective in binding digoxin and digitoxin, greatly enhancing the renal excretion of these drugs (see Chapter 34: Pharmacological Treatment of Heart Failure). Serum glycoside concentrations rise markedly with antidigitalis antibodies, but these represent bound (nonpharmacologically active) drug. Temporary cardiac pacing may be required for advanced sinus node or AV node dysfunction. Other forms of antidotal therapy that have been used in the past, but have been largely supplanted by the use of Fab fragments, include lidocaine, + + phenytoin, or the cautious administration of K for ventricular arrhythmias; K can exacerbate AV block. Digitalis exerts direct arterial vasoconstrictor effects, which can be especially deleterious in patients with advanced atherosclerosis who receive intravenous drug; ischemia in the intestinal or coronary beds has been reported. Clinical Pharmacokinetics The most commonly used digitalis glycoside in the United States is digoxin (LANOXIN), although digitoxin (CRYSTODIGIN) is also used for chronic oral therapy. Digoxin tablets are incompletely (75%) bioavailable, but capsules are >90% bioavailable. In some patients, intestinal microflora may metabolize digoxin, causing a marked reduction in drug bioavailability. In these patients, higherthan-usual doses are required for clinical efficacy to be achieved; toxicity is a serious risk if antibiotics such as tetracycline or erythromycin, which destroy intestinal microflora, are administered. Inhibition of P-glycoprotein (see below) also may play a role. Digoxin is 20% to 30% protein-bound. The antiarrhythmic effects of digoxin can be achieved with intravenous or oral therapy. However, digoxin undergoes relatively slow distribution to effector site(s); therefore, even with intravenous therapy, there is a lag of several hours between drug administration and the

development of measurable pharmacological effects such as PR interval prolongation or slowing of the ventricular rate in atrial fibrillation. To avoid digitalis intoxication, a loading regimen of 1 to 1.5 mg given over 24 hours is administered. Measurement of postdistribution serum digoxin concentration and adjusting the daily dose (0.125 to 0.375 mg) to maintain concentrations of 0.5 to 2 ng/ml are useful during chronic digoxin therapy (see Table 355). Some patients may require and tolerate higher concentrations, but with an increased risk of adverse effects.

The elimination half-life of digoxin ordinarily is 36 hours, so maintenance doses are administered once daily. Renal elimination of unchanged drug accounts for <80% of digoxin elimination. Digoxin doses should be reduced (or dosing interval increased) and serum concentrations closely monitored in patients with impaired excretion due to renal failure or in patients who are hypothyroid. Digitoxin undergoes primarily hepatic metabolism and may be useful in patients with fluctuating or advanced renal dysfunction. Digitoxin metabolism is accelerated by drugs such as phenytoin or rifampin that induce hepatic metabolism (Chapter 1: Pharmacokinetics: The Dynamics of Drug Absorption, Distribution, and Elimination). Digitoxin's elimination half-life is even longer than that of digoxin (about 7 days); it is highly protein-bound, and its therapeutic range is 10 to 30

ng/ml. Quinidine elevates serum digoxin concentrations by decreasing clearance and volume of distribution; new steady-state digoxin concentrations are approached in 4 to 5 digoxin elimination half-lives, i.e., in about a week (Leahey et al., 1978). A similar effect has been reported with digitoxin when serum concentrations were monitored for sufficiently long periods. Digitalis toxicity results so often when quinidine is being administered that it is routine to decrease the dose of digoxin if quinidine is started. Other drugs that increase serum digoxin concentration include verapamil, diltiazem, amiodarone, cyclosporine, itraconazole, propafenone, flecainide, and spironolactone. In these cases, the effect is less predictable, and digoxin concentrations regularly are measured and the dose adjusted only if necessary. One common mechanism underlying digoxin clearance appears to involve P-glycoproteinmediated transport. Inhibition of this transport has been implicated as the mechanism whereby quinidine and other interacting drugs decrease digoxin clearance (Fromm et al., 1999). Hypokalemia, which can be caused by many drugs (e.g., diuretics, amphotericin B, corticosteroids), will potentiate digitalis-induced arrhythmias. Disopyramide Disopyramide (NORPACE, others; Morady et al., 1982) exerts electrophysiological effects very similar to those of quinidine, but the drugs have different adverse-effect profiles. Disopyramide is used for the maintenance of sinus rhythm in patients with atrial flutter or atrial fibrillation and for the prevention of recurrence of ventricular tachycardia or ventricular fibrillation. Disopyramide is prescribed as a racemate. Its structure is given below.

Pharmacological Actions and Adverse Effects The in vitro electrophysiological actions of S-(+)-disopyramide are similar to those of quinidine (Mirro et al., 1981). The R-()-enantiomer produces similar Na+ channel block but does not prolong cardiac action potentials. Unlike quinidine, racemic disopyramide is not an -adrenergic receptor antagonist, but it does exert prominent anticholinergic actions, which account for many of its adverse effects. These include precipitation of glaucoma, constipation, dry mouth, and urinary retention; the latter is most common in males with prostatism but can occur in females. Disopyramide commonly depresses contractility and can precipitate heart failure (Podrid et al., 1980), and it can cause torsades de pointes. Clinical Pharmacokinetics Disopyramide is well absorbed. Binding to plasma proteins is concentration-dependent, so a small increase in total concentration may represent a proportionately larger increase in free drug concentration (Lima et al., 1981). Disopyramide is eliminated by both hepatic metabolism (to a weakly active metabolite) and renal excretion of unchanged drug. The dose should be reduced in

patients with renal dysfunction. Higher-than-usual dosages may be required in patients receiving drugs that induce hepatic metabolism, such as phenytoin. Dofetilide Dofetilide(TIKOSYN) is a potent and "pure"IKr blocker. As a result of this specificity, it has virtually no extracardiac pharmacological effects. Dofetilide is effective in maintaining sinus rhythm in patients with atrial fibrillation. In the DIAMOND studies (Torp-Pedersen et al., 1999), dofetilide did not affect mortality in patients with advanced heart failure or in those convalescing from acute myocardial infarction. Dofetilide currently is available through a restricted distribution system that includes only those physicians, hospitals, and other institutions that have received special educational programs covering proper dosing and treatment initiation.

Adverse Effects Torsades de pointes occurred in 1% to 3% of patients in clinical trials, in whom strict exclusion criteria (e.g., hypokalemia) were applied and continuous ECG monitoring was used to detect marked QT prolongation in the hospital. The incidence of this adverse effect during more widespread use of the drug, marketed in 2000, is unknown. Other adverse effects are no more common than with placebo. Clinical Pharmacokinetics Most of a dose of dofetilide is excreted unchanged by the kidneys. In patients with mild to moderate renal failure, decreases in dosage based on creatinine clearance are required to minimize the risk of torsades de pointes. The drug should not be used in patients with advanced renal failure and should not be used with inhibitors of renal cation transport. Dofetilide also undergoes minor hepatic metabolism. Flecainide The effects of flecainide (TAMBOCOR) therapy are thought to be attributable to the drug's very long + recovery from Na channel block (Roden and Woosley, 1986a). In the CAST study, flecainide increased mortality in patients convalescing from myocardial infarction (CAST Investigators, 1989). However, it continues to be approved for the maintenance of sinus rhythm in patients with supraventricular arrhythmias, including atrial fibrillation, in whom structural heart disease is absent (Anderson et al., 1989; Henthorn et al., 1991). Encainide, a drug with very similar electrophysiological actions, is no longer available.

Pharmacological Effects Flecainide blocks Na+ current and delayed rectifier K+ current (IKr) at similar concentrations in vitro, 1 to 2 M (Ikeda et al., 1985; Follmer and Colatsky, 1990). It also blocks Ca2+ currents in vitro. Action potential duration is shortened in Purkinje cells (probably due to block of late-opening Na+ channels) but prolonged in ventricular cells, probably due to block of delayed rectifier current (Ikeda et al., 1985). Flecainide does not cause EADs in vitro or torsades de pointes. In atrial tissue, flecainide prolongs action potentials disproportionately at fast rates, an especially desirable antiarrhythmic drug effect; this effect contrasts with that of quinidine, which prolongs atrial action potentials to a greater extent at slower rates (Wang et al., 1990). Flecainide prolongs the duration of PR, QRS, and QT, even at normal heart rates. Adverse Effects Flecainide produces few subjective complaints in most patients; dose-related blurred vision is the most common noncardiac adverse effect. It can exacerbate congestive heart failure in patients with depressed left ventricular performance. The most serious adverse effects are provocation or exacerbation of potentially lethal arrhythmias. These include acceleration of ventricular rate in patients with atrial flutter, increased frequency of episodes of reentrant ventricular tachycardia, and increased mortality in patients convalescing from myocardial infarction (Morganroth et al., 1986; Crijns et al., 1988; CAST Investigators, 1989; Ranger et al., 1989). As discussed above, it is likely that all these effects can be attributed to Na+ channel block. Flecainide also can cause heart block in patients with conduction system disease. Clinical Pharmacokinetics Flecainide is well absorbed. The elimination half-life is shorter with urinary acidification (10 hours) than with urinary alkalinization (17 hours), but it is nevertheless sufficiently long to allow dosing twice daily (see Table 355). Elimination occurs by both renal excretion of unchanged drug and hepatic metabolism to inactive metabolites. The latter is mediated by the polymorphically distributed enzyme CYP2D6 (Chapter 1: Pharmacokinetics: The Dynamics of Drug Absorption, Distribution, and Elimination) (Gross et al., 1989). However, even in patients in whom this pathway is absent because of genetic polymorphism or inhibition by other drugs (i.e., quinidine, fluoxetine), renal excretion ordinarily is sufficient to prevent drug accumulation. In the rare patient with renal dysfunction and lack of active CYP2D6, flecainide may accumulate to toxic plasma concentrations. Flecainide is a racemate, but there are no differences in the electrophysiological effects or disposition kinetics of its enantiomers (Kroemer et al., 1989). Some reports have suggested that plasma flecainide concentrations >1000 ng/ml should be avoided to minimize the risk of flecainide toxicity; however, in susceptible patients, the adverse electrophysiological effects of flecainide therapy can occur at therapeutic plasma concentrations. Ibutilide

Ibutilide (CORVERT; Murray, 1998) is an IKr blocker that in some systems also activates an inward Na+ current. The action potential-prolonging effect of the drug may arise from either mechanism. Ibutilide is administered as a rapid infusion (1 mg over 10 minutes) for the immediate conversion of atrial fibrillation or flutter to sinus rhythm. The drug's efficacy rate is higher in patients with atrial flutter (50% to 70%) than in those with atrial fibrillation (30% to 50%). In atrial fibrillation, the conversion rate is lowest in those in whom the arrhythmia has been present for weeks or months compared with those in whom it has been present for days. The major toxicity with ibutilide is torsades de pointes, which occur in up to 6% of patients and requires immediate cardioversion in up to one-third of these. The drug undergoes extensive first-pass metabolism and so is not used orally. It is eliminated by hepatic metabolism and has a half-life of 2 to 12 hours (average of 6 hours).

Lidocaine Lidocaine (XYLOCAINE ) is a local anesthetic that also is useful in the acute intravenous therapy of ventricular arrhythmias. When lidocaine was administered to all patients with suspected myocardial infarction, the incidence of ventricular fibrillation was reduced (Lie et al., 1974). However, survival to hospital discharge tended to be decreased (Hine et al., 1989), perhaps because of lidocaineexacerbated heart block or congestive heart failure. Lidocaine is therefore no longer routinely administered to all patients in coronary care units.

Pharmacological Effects Lidocaine blocks both open and inactivated cardiac Na+ channels. Findings from in vitro studies suggest that lidocaine-induced block reflects an increased likelihood that the Na + channel protein assumes a nonconducting conformation in the presence of drug (Balser et al., 1996). Recovery from block is very rapid, so lidocaine exerts greater effects in depolarized (e.g., ischemic) and/or rapidly driven tissues. Lidocaine is not useful in atrial arrhythmias, possibly because atrial action potentials are so short that the Na+ channel is in the inactivated state only briefly compared with diastolic (recovery) times, which are relatively long (Hondeghem and Katzung, 1984). In some studies, lidocaine increased current through inward rectifier channels, but the clinical significance of this effect is not known. Lidocaine can hyperpolarize Purkinje fibers depolarized by low [K]o or stretch (Arnsdorf and Bigger, 1972); the resultant increased conduction velocity may be antiarrhythmic in reentry. Lidocaine decreases automaticity by reducing the slope of phase 4 and altering the threshold for excitability. Action potential duration usually is unaffected or is shortened; such shortening may be

due to block of the few Na+ channels that inactivate late during the cardiac action potential. Lidocaine usually exerts no significant effect on PR or QRS duration; QT is unaltered or slightly shortened. The drug exerts little effect on hemodynamic function, although rare cases of lidocaineassociated exacerbations of heart failure have been reported, especially in patients with very poor left ventricular function. Adverse Effects When a large intravenous dose of lidocaine is administered rapidly, seizures can occur. When plasma concentrations of the drug rise slowly above the therapeutic range, as may occur during maintenance therapy, tremor, dysarthria, and altered levels of consciousness are more common. Nystagmus is an early sign of lidocaine toxicity. Clinical Pharmacokinetics Lidocaine is well absorbed, but undergoes extensive, though variable, first-pass hepatic metabolism (Thompson et al., 1973); thus, oral use of the drug is inappropriate. In theory, therapeutic plasma concentrations of lidocaine may be maintained by intermittent intramuscular administration, but the intravenous route is preferred (see Table 355). Lidocaine's metabolites, glycine xylidide (GX) and mono-ethyl GX (MEGX), are less potent as Na+ channel blockers than the parent drug. GX and lidocaine appear to compete for access to the Na+ channel, suggesting that, with infusions during which GX accumulates, lidocaine's efficacy may be diminished (Bennett et al., 1988). With infusions lasting longer than 24 hours, the clearance of lidocaine fallsan effect that has been attributed to competition between parent drug and metabolites for access to hepatic drugmetabolizing enzymes (LeLorier et al., 1977; Suzuki et al., 1984). Plasma concentrations of lidocaine decline biexponentially after a single intravenous dose, indicating that a multicompartment model (Chapter 1: Pharmacokinetics: The Dynamics of Drug Absorption, Distribution, and Elimination) is necessary to analyze lidocaine disposition. The initial drop in plasma lidocaine following intravenous administration occurs rapidly, with a half-life of 8 minutes, and represents distribution from the central compartment to peripheral tissues. The terminal elimination half-life, usually 100 to 120 minutes, represents drug elimination by hepatic metabolism. Lidocaine's efficacy depends on maintenance of therapeutic plasma concentrations in the central compartment. Therefore, the administration of a single bolus dose of lidocaine can result in transient arrhythmia suppression, but this effect dissipates rapidly as the drug is distributed and concentrations in the central compartment fall. To avoid this distribution-related loss of efficacy, a loading regimen of 3 to 4 mg/kg over 20 to 30 minutes is usede.g., an initial 100 mg followed by 50 mg every 8 minutes for three doses. Subsequently, stable concentrations can be maintained in plasma with an infusion of 1 to 4 mg/min, which replaces drug removed by hepatic metabolism. The time to steady-state lidocaine concentrations is 8 to 10 hours. If the chosen maintenance infusion rate is too low, arrhythmias may recur hours after the institution of apparently successful therapy. On the other hand, if the rate is too high, toxicity may result. In either case, routine measurement of plasma lidocaine concentration at the time of expected steady state is useful in adjusting maintenance infusion rate. In heart failure, the central volume of distribution is decreased, so the total loading dose should be decreased (Thompson et al., 1973). Since lidocaine clearance also is decreased, the rate of the maintenance infusion should be decreased. Lidocaine clearance is also reduced in hepatic disease (Thompson et al., 1973), during treatment with cimetidine or -blockers, and during prolonged infusions (Nies et al., 1976; LeLorier et al., 1977; Feely et al., 1982). Frequent measurement of plasma lidocaine concentration and dose adjustment to ensure that plasma concentrations remain

within the therapeutic range (1.5 to 5 g/ml) are necessary to minimize toxicity in these settings. Lidocaine is bound to the acute phase reactant, 1-acid glycoprotein. Diseases such as acute myocardial infarction are associated with increases in 1-acid glycoprotein and protein binding, and hence a decreased proportion of free drug. These findings may explain why some patients require and tolerate higher-than-usual total plasma lidocaine concentrations to maintain antiarrhythmic efficacy (Alderman et al., 1974; Kessler et al., 1984). Magnesium The intravenous administration of 1 to 2 g of MgSO4 has been reported to be effective in preventing recurrent episodes of torsades de pointes, even if serum Mg2+ is normal (Tzivoni et al., 1988). However, controlled studies of this effect have not been performed. The mechanism of action is unknown; since QT interval is not shortened, an effect on the inward current, possibly a Ca2+ current, responsible for the triggered upstroke arising from EADs (black arrow, Figure 356B) is possible (Jackman et al., 1988; Roden, 1991b). Intravenous Mg 2+ also has been used successfully in arrhythmias related to digitalis intoxication (Seller, 1971). Large, placebo-controlled trials of intravenous magnesium to improve outcome in acute myocardial infarction have yielded conflicting results (Woods and Fletcher, 1994; ISIS-4 Collaborative Group, 1995). While oral Mg2+ supplements may be useful in preventing hypomagnesemia, there is no evidence that chronic Mg2+ ingestion exerts a direct antiarrhythmic action. Mexiletine and Tocainide Mexiletine (MEXITIL) and tocainide (TONOCARD) are analogs of lidocaine with structures that have been modified to reduce first-pass hepatic metabolism to make chronic oral therapy effective (Roden and Woosley, 1986b; Campbell, 1987). Their electrophysiological actions are similar to those of lidocaine. Tremor and nausea are the major dose-related adverse effects; these can be minimized by taking the drugs with food. Because tocainide can cause potentially fatal bone marrow aplasia and pulmonary fibrosis, it is rarely used.

Mexiletine undergoes hepatic metabolism, which is inducible by drugs such as phenytoin. Tocainide, on the other hand, is eliminated by renal excretion. Thus, in patients with renal disease, the dose of tocainide should be decreased. Both agents have been used for ventricular arrhythmias; combinations of mexiletine or tocainide with quinidine or sotalol may increase efficacy while reducing adverse effects. In vitro studies and clinical anecdotes have suggested a role for mexiletine (or flecainide) in correcting the molecular defect in the form of congenital long QT syndrome

caused by abnormal Na+ channel inactivation (Shimizu and Antzelevitch, 1997). Moricizine Moricizine (ETHMOZINE ) is a phenothiazine analog with Na+ channelblocking properties used in the chronic treatment of ventricular arrhythmias (Clyne et al., 1992). In a randomized, double-blind trial (CAST II), moricizine increased mortality in patients shortly after a myocardial infarction and did not improve survival during long-term therapy (Cardiac Arrhythmia Suppression Trial II Investigators, 1992). Moricizine undergoes extensive first-pass hepatic metabolism; despite its short elimination half-life, its antiarrhythmic effect can persist for many hours after a single dose, suggesting that some of its metabolites may be active.

Phenytoin The anticonvulsant phenytoin (DILANTIN; Chapter 21: Drugs Effective in the Therapy of the Epilepsies) also is a blocker of inactivated cardiac Na + channels. It has been used in the acute and chronic suppression of ventricular arrhythmias and in digitalis intoxication (Atkinson and Davison, 1974). Phenytoin has a short recovery, and little QRS prolongation is observed during chronic therapy. Phenytoin undergoes extensive, saturable, first-pass hepatic metabolism; therefore, small increases in dose can produce large increases in plasma concentration and toxicity (Richens, 1979). Phenytoin is highly bound to plasma proteins, but the extent of its binding may vary. For example, in patients with renal disease, phenytoin binding falls from 90% to 80%, effectively doubling the free fraction of drug. Phenytoin toxicity may ensue if doses are adjusted on the basis of total rather than free drug concentrations. Symptoms of phenytoin toxicity include CNS complaints, such as ataxia, nystagmus, or mental confusion, and gingival hyperplasia; serious dermatological and bone marrow reactions can occur. With intravenous use, hypotension and ventricular fibrillation have been reported. Phenytoin is an inducer of the hepatic metabolism of many other drugs, including quinidine, mexiletine, digitoxin, estrogens, theophylline, and vitamin D (Richens, 1979).

Procainamide Procainamide (PROCAN SR; others) is an analog of the local anesthetic, procaine. It exerts electrophysiological effects similar to those of quinidine, but lacks quinidine's vagolytic and adrenergic blocking activity. Procainamide is better tolerated than quinidine when given

intravenously. Loading and maintenance intravenous infusions are used in the acute therapy of many supraventricular and ventricular arrhythmias. However, long-term oral treatment is often stopped because of adverse effects.

Pharmacological Effects Procainamide is a blocker of open Na+ channels, with an intermediate time constant of recovery from block. It also prolongs cardiac action potentials in most tissues, probably by blocking outward K+ current(s). Procainamide decreases automaticity, increases refractory periods, and slows conduction. The major metabolite, N-acetyl procainamide, lacks the Na+ channelblocking activity of the parent drug but is equipotent in prolonging action potentials (Dangman and Hoffman, 1981). As the plasma concentrations of N-acetyl procainamide often exceed those of procainamide, increased refractoriness and QT prolongation during chronic procainamide therapy can be attributed at least in part to the metabolite. However, it is the parent drug that slows conduction and produces QRS interval prolongation. Although hypotension may occur at high plasma concentrations, this effect usually is attributable to ganglionic blockade rather than to any negative inotropic effect, which is minimal. Adverse Effects Hypotension and marked slowing of conduction are major adverse effects of high concentrations (>10 g/ml) of procainamide, especially during intravenous use. Dose-related nausea is frequent during oral therapy and may be attributable in part to high plasma concentrations of N-acetyl procainamide. Torsades de pointes can occur, particularly when plasma concentrations of N-acetyl procainamide rise to >30 g/ml. Procainamide produces potentially fatal bone marrow aplasia in 0.2% of patients; the mechanism is not known, but high plasma drug concentrations are not suspected. During long-term therapy, most patients will develop biochemical evidence of the drug-induced lupus syndrome, such as circulating antinuclear antibodies (Woosley et al., 1978). Therapy need not be interrupted merely because of the presence of antinuclear antibodies. However, many patients, perhaps as many as 25% to 50%, eventually will develop symptoms of the lupus syndrome; common early symptoms are rash and small-joint arthralgias. Other symptoms of lupus, including pericarditis with tamponade, can occur, although renal involvement is unusual. Clinical Pharmacokinetics Procainamide is rapidly eliminated (t1/2 = 3 to 4 hours) by both renal excretion of unchanged drug as well as by hepatic metabolism. The major pathway for hepatic metabolism is conjugation by Nacetyl transferase (see Chapter 1: Pharmacokinetics: The Dynamics of Drug Absorption, Distribution, and Elimination) to form N-acetyl procainamide. N-Acetyl procainamide is eliminated by renal excretion (t1/2 = 6 to 10 hours) and is not significantly converted back to procainamide. Because of the relatively rapid elimination rates of both the parent drug and its major metabolite, procainamide usually is administered as a slow-release formulation. In patients with renal failure, procainamide and/or N-acetyl procainamide can accumulate to potentially toxic plasma

concentrations (Drayer et al., 1977). Reduction of procainamide dose and dosing frequency and monitoring of plasma concentrations of both compounds are required in this situation. Because the parent drug and metabolite exert different pharmacological effects, the practice of using the sum of their concentrations to guide therapy is inappropriate. In individuals who are slow acetylators, the procainamide-induced lupus syndrome develops more often and earlier during treatment than among rapid acetylators (Woosley et al., 1978). In addition, the symptoms of procainamide-induced lupus resolve during treatment with N-acetyl procainamide. Both of these findings support results of in vitro studies suggesting that it is chronic exposure to the parent drug (or an oxidative metabolite) that results in the lupus syndrome; these findings also provided one rationale for the further development of N-acetyl procainamide and its analogs as antiarrhythmic agents (Roden, 1993). Propafenone Propafenone (RYTHMOL) is a Na+ channel blocker with a relatively slow time constant for recovery from block (Funck-Brentano et al., 1990). Some data suggest that, like flecainide, propafenone also blocks K+ channels. Its major electrophysiological effect is to slow conduction in fast-response tissues. The drug is prescribed as a racemate; while the enantiomers do not differ in their Na+ channelblocking properties, S-(+)-propafenone is a -adrenergic receptor antagonist in vitro and in some patients. Propafenone prolongs PR and QRS durations. Chronic therapy with oral propafenone is used to maintain sinus rhythm in patients with supraventricular tachycardias, including atrial fibrillation; it also can be used in ventricular arrhythmias but, like other Na+ channel blockers, it is only modestly effective.

Adverse effects during propafenone therapy include acceleration of ventricular response in patients with atrial flutter, increased frequency or severity of episodes of reentrant ventricular tachycardia, exacerbation of heart failure, and the adverse effects of -adrenergic blockade such as sinus bradycardia and bronchospasm (see above and Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists). Clinical Pharmacokinetics Propafenone is well absorbed and is eliminated by both hepatic and renal routes. The activity of cytochrome P450 2D6 (CYP2D6), an enzyme that functionally is absent in 7% of Caucasians and African-Americans (Chapter 1: Pharmacokinetics: The Dynamics of Drug Absorption, Distribution, and Elimination), is a major determinant of plasma propafenone concentration, and thus the clinical action of the drug. In most subjects ("extensive metabolizers"), propafenone undergoes extensive first-pass hepatic metabolism to 5-hydroxy propafenone, a metabolite equipotent to propafenone as a Na+ channel blocker but much less potent as a -adrenergic receptor antagonist. A second metabolite, N-desalkyl propafenone, is formed by non-CYP2D6-mediated metabolism and is a less potent blocker of Na+ channels and -adrenergic receptors. CYP2D6-mediated metabolism of propafenone is saturable, so small increases in dose can lead to disproportionate increases in plasma propafenone concentration. In "poor metabolizer" subjects, in whom functional CYP2D6 is absent,

the extent of first-pass hepatic metabolism is much less than in extensive metabolizers, and plasma propafenone concentrations are much higher. The incidence of adverse effects during propafenone therapy is significantly higher in poor than in extensive metabolizers. CYP2D6 activity can be markedly inhibited by a number of drugs, including quinidine and fluoxetine. In extensive metabolizer subjects receiving such inhibitors or in poor metabolizer subjects, plasma propafenone concentrations >1 g/ml are associated with clinical effects of -adrenergic receptor blockade, such as reduction of exercise heart rate (Lee et al., 1990; Mrike and Roden, 1994). It is recommended that dosage in patients with moderate to severe liver disease should be reduced to approximately 20% to 30% of the usual dose, with careful monitoring. It is not known if propafenone doses must be decreased in patients with renal disease. Quinidine As early as the eighteenth century, the bark of the cinchona plant was used to treat "rebellious palpitations" (Levy and Azoulay, 1994). Studies in the early twentieth century identified quinidine, a diastereomer of the antimalarial quinine, as the most potent of the antiarrhythmic substances extracted from the cinchona plant, and by the 1920s quinidine was used as an antiarrhythmic agent (Wenckebach, 1923). Quinidine is used for the maintenance of sinus rhythm in patients with atrial flutter or atrial fibrillation and in the prevention of recurrence of ventricular tachycardia or ventricular fibrillation (Grace and Camm, 1998).

Pharmacological Effects Quinidine (QUINAGLUTE , QUINIDEX , others) blocks Na+ current and multiple cardiac K+ currents. It is an open-state blocker of Na + channels, with a time constant of recovery in the intermediate ( 3 seconds) range; as a consequence, QRS duration increases modestly, usually from 10% to 20%, at therapeutic dosages. At therapeutic concentrations, quinidine routinely prolongs QT interval up to 25%, but the effect is highly variable. At concentrations as low as 1 M, quinidine blocks Na+ current and the rapid component of delayed rectifier (IKr); higher concentrations block the slow component of delayed rectifier, inward rectifier, transient outward current, and L-type Ca2+ current. Quinidine's Na+ channel-blocking properties result in an increased threshold for excitability and decreased automaticity. As a consequence of its K+ channelblocking actions, quinidine prolongs action potentials in most cardiac cells. This effect is most prominent at slow rates. In some cells, such as midmyocardial cells and Purkinje cells, quinidine consistently elicits EADs at slow heart rates, particularly when [K]o is low (Roden and Hoffman, 1985). Quinidine prolongs refractoriness in most tissues, probably as a result of both prolongation of action potential duration and its Na+ channel blockade.

In the intact animal and in human beings, quinidine also produces -adrenergic receptor blockade and vagal inhibition. Thus, the intravenous use of quinidine is associated with marked hypotension and sinus tachycardia. Quinidine's vagolytic effects tend to inhibit its direct depressant effect on AV nodal conduction, so that the effect of drug on the PR interval is variable. Moreover, quinidine's vagolytic effect can result in increased AV nodal transmission of atrial tachycardias such as atrial flutter (Table 351). Adverse Effects Noncardiac Diarrhea is the most common adverse effect during quinidine therapy, occurring in 30% to 50% of patients. The mechanism is not known. Diarrhea usually occurs within the first several days of quinidine therapy, but can occur later. In mild cases of diarrhea in which quinidine therapy is deemed vital, antidiarrheal drugs can be used. Diarrhea-induced hypokalemia may potentiate torsades de pointes due to quinidine. A number of immunological reactions can occur during quinidine therapy. The most common is thrombocytopenia, which can be severe but which resolves rapidly with discontinuation of the drug. Hepatitis, bone marrow depression, and a lupus syndrome occur rarely. None of these effects is related to elevated plasma quinidine concentrations. Quinidine also can produce cinchonism, a symptom complex that includes headache and tinnitus. In contrast to other adverse events during quinidine therapy, cinchonism usually is related to elevated plasma quinidine concentrations, and can be managed by dose reduction. Cardiac It is estimated that 2% to 8% of patients who receive quinidine therapy will develop marked QTinterval prolongation and torsades de pointes. In contrast to effects of sotalol, N-acetyl procainamide, and many other drugs, quinidine-associated torsades de pointes generally occur at therapeutic, or even subtherapeutic, plasma concentrations (Jackman et al., 1988; Roden, 1991). The reasons for individual susceptibility to this adverse effect are not known. At high plasma concentrations of quinidine, marked Na channel block can occur, with resultant ventricular tachycardia. This adverse effect formerly occurred when very high doses of quinidine were used to try to convert atrial fibrillation to normal rhythm, but this aggressive approach to quinidine dosing has now been abandoned, and quinidine-induced ventricular tachycardia due to excess Na+ channel block is unusual. Quinidine can exacerbate heart failure or conduction system disease. However, in most patients with congestive heart failure, quinidine is well tolerated, perhaps because of its vasodilating actions. Clinical Pharmacokinetics Quinidine is well absorbed and is 80% bound to plasma proteins including albumin and, like lidocaine, the acute phase reactant, 1-acid glycoprotein. As with lidocaine, greater-than-usual doses (and total plasma quinidine concentrations) may be required to maintain therapeutic concentrations of free quinidine in high-stress states such as acute myocardial infarction (Kessler et al., 1984). Quinidine undergoes extensive hepatic oxidative metabolism, and 20% is excreted unchanged by the kidneys. One metabolite, 3-hydroxy-quinidine, is nearly as potent as quinidine in
+

blocking cardiac Na+ channels or prolonging cardiac action potentials. Concentrations of unbound 3-hydroxyquinidine equal to or exceeding those of quinidine are tolerated by some patients. Other metabolites are less potent than quinidine, and their plasma concentrations are lower; thus, it is unlikely that they contribute significantly to the clinical effects of quinidine. There is substantial individual variability in the range of dosages required to achieve therapeutic plasma concentrations of 2 to 5 g/ml. Some of this variability may be assay-dependent, since not all assays have excluded quinidine metabolites. Among patients with advanced renal disease or advanced congestive heart failure, quinidine clearance is only modestly decreased. Thus, dosage requirements in these patients are similar to those in other patients. Drug Interactions Quinidine is a potent inhibitor of CYP2D6. As a result, the administration of quinidine to patients receiving drugs that undergo extensive CYP2D6-mediated metabolism may result in altered drug effects due to accumulation of parent drug and failure of metabolite formation. For example, inhibition of CYP2D6-mediated metabolism of codeine to its active metabolite, morphine, results in decreased analgesia. On the other hand, inhibition of CYP2D6-mediated metabolism of propafenone results in elevated plasma propafenone concentrations and increased -adrenergic receptor blockade. Quinidine reduces the clearance of digoxin and digitoxin; inhibition of Pglycoproteinmediated digoxin transport has been implicated (Fromm et al., 1999). This interaction is particularly important during the initiation of quinidine therapy in patients with atrial fibrillation, many of whom also are receiving digoxin (Leahey et al., 1978). Quinidine metabolism is inducible by drugs such as phenobarbital or phenytoin (Data et al., 1976). In patients receiving these agents, very high doses of quinidine may be required to achieve therapeutic concentrations. It is important to note that, if therapy with the inducing agent is then stopped, quinidine concentrations may rise to very high levels, and its dosage must be adjusted downward. Cimetidine and verapamil elevate plasma quinidine concentrations, but these effects usually are modest. Sotalol Sotalol (BETAPACE ) is a nonselective -adrenergic receptor antagonist that also prolongs cardiac action potentials by inhibiting delayed rectifier and possibly other K+ currents (Hohnloser and Woosley, 1994). Sotalol is prescribed as a racemate; the l-enantiomer is a much more potent adrenergic receptor antagonist than the d-enantiomer, but the two are equipotent as K + channel blockers. Its structure is shown below:

In the United States, racemic sotalol is approved for use in patients with both ventricular tachyarrhythmias and atrial fibrillation or flutter. Clinical trials suggest that it is at least as effective as most Na+ channel blockers in ventricular arrhythmias (Mason, 1993). Sotalol prolongs action potential duration throughout the heart and QT interval on the ECG. It decreases automaticity, slows AV nodal conduction, and prolongs AV refractoriness both by K+ channel block and block of -adrenergic receptors, but it exerts no effect on conduction velocity in

fast-response tissue. Sotalol causes EADs and triggered activity in vitro (Strauss et al., 1970) and can cause torsades de pointes, especially when serum K+ is low. Unlike the situation with quinidine, the incidence of torsades de pointes seems to depend on the dose of sotalol, and in fact torsades de pointes are the major toxicity with sotalol overdose. Occasional cases occur at low dosages, often in patients with renal dysfunction, since sotalol is eliminated by renal excretion of unchanged drug. The other adverse effects of sotalol therapy are those associated with -adrenergic receptor blockade (see above and Chapter 10: Catecholamines, Sympathomimetic Drugs, and Adrenergic Receptor Antagonists).

Prospectus Recent and continuing elucidation of the mechanisms underlying normal and abnormal cardiac electrical behavior, along with clinical investigations clearly delineating the dangers of indiscriminate use of potent ion channelblocking drugs in large groups of patients, have

significantly altered therapeutic strategies for the treatment of arrhythmias. In some instances, nonpharmacological therapies are increasingly used to avoid the dangers of chronic drug therapy. Techniques for ablation of areas critical for the genesis of arrhythmias are now sufficiently well developed that chronic drug therapy can be avoided in many patients. When arrhythmia recurrence would be lethal and drugs may be incompletely effective, as in patients resuscitated from an episode of ventricular fibrillation, implanted defibrillators often are an appropriate choice. Among drugs discussed here, only -blockers (and, to a lesser extent, amiodarone) reduce the incidence of sudden death. By contrast, while drugs with predominant ion channel-blocking properties can acutely terminate arrhythmias, long-term therapy has either increased or not altered the incidence of sudden death. Basic genetic, molecular, and cellular studies continue to improve our understanding of arrhythmia mechanisms; it is likely that this new knowledge will identify novel drug targets to intervene in specific ways to alleviate particular arrhythmic events and thus supplant currently available drugs. For further discussion of cardiac arrhythmias, see Chapters 213 and 214 in Harrison's Principles of Internal Medicine, 16th ed., McGraw-Hill, New York, 2005.

Chapter 36. Drug Therapy for Hypercholesterolemia and Dyslipidemia

Overview Hyperlipidemia is a major cause of atherosclerosis and atherosclerosis-associated conditions, such as coronary heart disease, ischemic cerebrovascular disease, and peripheral vascular disease. This chapter covers the normal metabolism of lipoproteins, the pathophysiology of dyslipidemia and atherosclerosis, and drugs used to treat dyslipidemia. Drugs covered include HMG-CoA reductase inhibitorsthe statinswhich are the most effective and best-tolerated drugs currently in use for treating dyslipidemia; bile acidbinding resins; nicotinic acid (niacin); and fibric acid derivatives. The chapter concludes with a brief discussion of potential new classes of antidyslipidemic drugs that are undergoing clinical or preclinical evaluation. Drug Therapy for Hypercholesterolemia and Dyslipidemia: Introduction Although the incidence of atherosclerosis-related vascular disease events is declining in the United States, coronary heart disease (CHD), ischemic cerebrovascular disease, and peripheral vascular disease still account for the majority of morbidity and mortality among middle-aged and older adults. Hyperlipidemia (hypercholesterolemia) is a major cause of increased atherogenic risk, and both genetic disorders and diets enriched in saturated fat and cholesterol contribute to the elevated lipid levels of our population and many other developed countries around the world. Recognition of hypercholesterolemia as a risk factor has led to the development of drugs that reduce cholesterol levels. These drugs have been used in well-controlled studies of patients with high cholesterol levels caused primarily by elevated levels of low-density lipoproteins (LDL). The results of these trials indicate that CHD mortality is reduced by as much as 30% to 40% and that nonfatal events are similarly reduced when hypercholesterolemic patients are treated with moderate doses of hypolipidemic drugs [Scandinavian Simvastatin Survival Study Group, 1994; Shepherd et al., 1995; The Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group, 1998]. Clinical trial data support extending the benefit of lipid-lowering therapy to high-risk patients

whose major lipid risk factor is a reduced plasma level of high-density-lipoprotein cholesterol (HDL-C) even if the LDL cholesterol (LDL-C) levels of these patients do not meet the existing threshold values for initiating hypolipidemic drug therapy (The Expert Panel, 1993). In patients with low HDL-C and average LDL-C levels, appropriate drug therapy reduced CHD endpoint events by 20% to 35% (Downs et al., 1998; Rubins et al., 1999). Since 40% of patients with CHD in the United States have low HDL-C levels, it is of obvious importance to include low-HDL patients in management guidelines for dyslipidemia, even if their LDL-C levels are in the "normal" range (Rubins et al., 1995). Hypertriglyceridemia (elevated levels of triglycerides), if severe (>1000 mg/dl), requires therapy to prevent pancreatitis. Moderately elevated triglyceride levels (150 to 400 mg/dl) also are of concern because they often occur as part of a syndrome distinguished by insulin resistance, obesity, hypertension, and substantially increased CHD risk. The atherogenic dyslipidemia in patients with this insulin resistance or metabolic syndrome is characterized by moderately elevated triglycerides, low HDL-C levels, and lipid-depleted LDL (sometimes referred to as "small, dense LDL") (Reaven, 1995; Grundy, 1998a). The metabolic syndrome is common in CHD patients; hence, identification of moderate hypertriglyceridemia in a patient, even if the total cholesterol level is normal, should trigger an evaluation to identify this disorder (National Cholesterol Education Program Expert Panel, 2001). Hyperlipidemia (elevated levels of triglycerides or cholesterol) and reduced HDL-C levels occur as a consequence of several factors that affect the concentrations of the various plasma lipoproteins. These factors may be lifestyle or behavioral (e.g., diet or exercise), genetic (e.g., mutations in a gene regulating lipoprotein levels), or metabolic conditions (e.g., diabetes mellitus) that influence plasma lipoprotein metabolism. An understanding of these factors requires a brief description of lipoprotein metabolism. More detailed descriptions can be found elsewhere (Breslow, 1994; Ginsberg and Goldberg, 1998; Mahley et al., 1998). Plasma Lipoprotein Metabolism Lipoproteins are macromolecules that contain lipids and proteins known as apolipoproteins or apoproteins. The lipid constituents include free and esterified cholesterol, triglycerides, and phospholipids. The apoproteins are very important since they provide structural stability to the lipoproteins, and a number of apoproteins function as ligands in lipoproteinreceptor interactions or are cofactors in enzymatic processes that regulate lipoprotein metabolism. In all spherical lipoproteins, the most water-insoluble lipids (cholesteryl esters and triglycerides) are core components, and the more polar, water-soluble components (apoproteins, phospholipids, and unesterified cholesterol) are located on the surface. Table 361 lists the major classes of lipoproteins and describes a number of their properties. Table 362 provides information about apoproteins that have well-defined roles in plasma lipoprotein metabolism. These apolipoproteins include apolipoprotein (apo) A-I, apoA-II, apoA-IV, apoB-100, apoB-48, apoC-I, apoC-II, apoC-III, apoE, and apo(a). Except for apo(a), the lipidbinding regions of all apoproteins contain structural features called amphipathic helices that interact with the polar, hydrophilic lipids (such as surface phospholipids) and with the aqueous plasma environment in which the lipoproteins circulate. Differences in the non-lipid-binding regions are responsible for the functional specificities of the apolipoproteins. Chylomicrons Chylomicrons are synthesized from the fatty acids of dietary triglycerides and cholesterol absorbed

from the small intestine by epithelial cells. Triglyceride synthesis is regulated by diacylglycerol transferase, an enzyme that regulates triglyceride synthesis in many tissues (Farese et al., 2000). After they are synthesized in the endoplasmic reticulum, triglycerides are transferred by microsomal triglyceride transfer protein (MTP) to the site where newly synthesized apoB-48 is available to form chylomicrons. Dietary cholesterol is esterified by one of two forms of the enzyme acyl coenzyme A:cholesterol acyltransferase (ACAT). This enzyme, ACAT-2, is found in the intestine and in the liver, where cellular free cholesterol is esterified before triglyceride-rich lipoproteins [chylomicrons and very-low-density lipoproteins (VLDL)] are assembled. In the intestine, ACAT-2 regulates the absorption of dietary cholesterol, and it may be a potential pharmacological target for reducing blood cholesterol levels (Cases et al., 1998). [A second ACAT enzyme, ACAT-1, is expressed in macrophages, including foam cells, adrenocortical cells, and skin sebaceous glands. Although ACAT-1 esterifies cholesterol and promotes foam-cell development, ACAT-1 knockout mice do not have a reduced susceptibility for developing atherosclerosis (Accad et al., 2000).] Chylomicrons are the largest of the plasma lipoproteins and are the only lipoproteins that float to the top of a tube of plasma that has been allowed to stand undisturbed for 12 hours. The buoyancy of chylomicrons reflects their 98% to 99% fat content, of which 85% is dietary triglyceride. In chylomicrons, the ratio of triglycerides to cholesterol is 10 or greater. In normolipidemic individuals, chylomicrons are present in plasma for 3 to 6 hours after a fat-containing meal has been ingested. After a fast of 10 to 12 hours, no chylomicrons remain. The apolipoproteins of chylomicrons include some (apoB-48, apoA-I, and apoA-IV) that are synthesized by intestinal epithelial cells and others (apoE and apoC-I, C-II, and C-III) acquired from HDL after chylomicrons have been secreted into the lymph and enter the plasma (Table 362). The apoB-48 of chylomicrons is one of two forms of apoB present in lipoproteins. ApoB-48, synthesized only by intestinal epithelial cells, is unique to chylomicrons, whereas apoB-100 is synthesized by the liver and incorporated into VLDL and intermediate-density lipoproteins (IDL) and LDL, which are products of VLDL catabolism. The apparent molecular weight of apoB-48 is 48% that of apoB-100, which accounts for the name "apoB-48." This is because the amino acid sequence of apoB-48 is identical to the first 2152 of the 4536 residues of apoB-100. An RNA editing mechanism unique to the intestine accounts for the premature termination of the translation of the apoB-100 mRNA (Innerarity et al., 1996). ApoB-48 does not contain the portion of the sequence of apoB-100 that allows apoB-100 to bind to the LDL receptor, so apoB-48 appears to function primarily as a structural component of chylomicrons. After gaining entry to the circulation via the thoracic duct, chylomicrons are metabolized initially at the capillary luminal surface of tissues that synthesize lipoprotein lipase (LPL), a triglyceride hydrolase (Figure 361). These tissues include adipose tissue, skeletal and cardiac muscle, and breast tissue of lactating women. As the triglycerides are hydrolyzed by LPL, the resulting free fatty acids are taken up and utilized by the adjacent tissues. The interaction of chylomicrons and LPL requires apoC-II as an absolute cofactor that mediates the interaction of LPL and chylomicron triglycerides. The absence of functional LPL or functional apoC-II prevents the hydrolysis of triglycerides in chylomicrons and results in severe hypertriglyceridemia and pancreatitis during childhood or even infancy (chylomicronemia syndrome). Recently, a variety of new, potentially atherogenic roles for LPL have been identified that affect the metabolism and uptake of atherogenic lipoproteins by the liver, the arterial wall, and the dyslipidemia of insulin resistance (Mead et al., 1999).

Figure 361. Summary of the Major Pathways Involved in the Metabolism of Chylomicrons Synthesized by the Intestine and VLDL Synthesized by the Liver. Chylomicrons are converted to chylomicron remnants by the hydrolysis of their triglycerides by LPL. Chylomicron remnants are rapidly cleared from the plasma by the liver. FFA released by LPL are used by muscle tissue as an energy source or taken up and stored by adipose tissue. FFA, free fatty acid; HL, hepatic lipase, IDL, intermediate-density lipoproteins; LDL, low-density lipoproteins; LPL, lipoprotein lipase; VLDL, very-low-density lipoproteins.

The concentration of chylomicrons can be controlled only by reducing dietary fat consumption. There is no current therapeutic approach that will enhance chylomicron catabolism except for insulin replacement in patients with type I diabetes mellitus (insulin has a "permissive effect" on LPL-mediated triglyceride hydrolysis). Chylomicron Remnants After LPL-mediated removal of much of the dietary triglycerides, the chylomicron remnants, which still contain all of the dietary cholesterol, detach from the capillary surface and within minutes are removed from the circulation by the liver in a multistep process mediated by apoE (Figure 361) (Mahley and Ji, 1999). First, the remnants are sequestered by the interaction of apoE with heparan sulfate proteoglycans on the surface of hepatocytes and are processed by hepatic lipase (HL), which further reduces the remnant triglyceride content. Next, apoE mediates remnant uptake by interacting with the hepatic LDL receptor or the LDL receptorrelated protein (LRP) (Krieger and Herz, 1994). The LRP is a receptor with multiple functions that recognizes a variety of ligandsincluding apoE, HL, and LPLand several ligands unrelated to lipid metabolism. In plasma lipid metabolism, the LRP is important because it is the backup receptor responsible for the uptake of apoE-enriched remnants of chylomicrons and VLDL. Cell-surface heparan sulfate proteoglycans facilitate the interaction of apoE-containing remnant lipoproteins with the LRP, which mediates uptake by hepatocytes (Mahley and Huang, 1999). Inherited absence of either functional HL (very rare) or functional apoE impedes remnant clearance by the LDL receptor and the LRP, resulting in a hyperlipidemia characterized by an increase of triglyceride- and cholesterol-rich remnant

lipoproteins in the plasma (type III hyperlipoproteinemia) (Mahley and Rall, 2001). Chylomicron remnants are not precursors of LDL. However, during the initial hydrolysis of chylomicron triglycerides by LPL, apoA-I and phospholipids are shed from the surface of chylomicrons and remain in the plasma. This is one mechanism by which nascent (precursor) HDL are generated. Very-Low-Density Lipoproteins VLDL are produced in the liver and are synthesized when triglyceride production is stimulated by an increased flux of free fatty acids or by increased de novo synthesis of fatty acids by the liver. The VLDL are 400 to 1000 in diameter and are large enough to cause plasma turbidity, but VLDL particles, unlike chylomicrons, do not float spontaneously to the top of a tube of plasma that is allowed to stand undisturbed for 12 hours. ApoB-100, apoE, and apoC-I, C-II, and C-III are synthesized constitutively by the liver and incorporated into VLDL (Table 362). If triglycerides are not available to form VLDL, the newly synthesized apoB-100 is degraded by hepatocytes. Triglycerides are synthesized in the endoplasmic reticulum and, along with other lipid constituents, are transferred by MTP to the site in the endoplasmic reticulum where newly synthesized apoB-100 is available to form nascent (precursor) VLDL. The nascent VLDL incorporate small amounts of apoE and the C apoproteins within the liver before secretion, but most of these apoproteins are acquired from plasma HDL after the VLDL are secreted by the liver. Without MTP, hepatic triglycerides cannot be transferred to apoB-100. As a consequence, patients with dysfunctional MTP fail to make any of the apoB-containing lipoproteins (VLDL, IDL, or LDL). MTP also plays a key role in the synthesis of chylomicrons in the intestine, and mutations of MTP that result in the inability of triglycerides to be transferred to either apoB-100 in the liver or apoB-48 in the intestine prevent VLDL and chylomicron production and cause the genetic disorder abetalipoproteinemia (Gregg and Wetterau, 1994). Plasma VLDL are then catabolized by LPL in the capillary beds in a process similar to the lipolytic processing described for chylomicrons (Figure 361). When triglyceride hydrolysis is nearly complete, the VLDL remnants, usually termed IDL, are released from the capillary endothelium and reenter the circulation. ApoB-100containing small VLDL and IDL (VLDL remnants), which have a half-life of less than 30 minutes, have two potential fates. About 40% to 60% are bound by LDL receptors or the LRP, which recognizes ligands (apoB-100 and apoE) on the remnants, and are cleared from the plasma primarily by the liver. The remainder of the IDL are further acted upon by LPL and HLwhich remove additional triglycerides, C apoproteins, and apoEand are converted to plasma LDL. Virtually all LDL particles in the plasma are derived from VLDL. ApoE plays a major role in the metabolism of triglyceride-rich lipoproteins (chylomicrons, chylomicron remnants, VLDL, and IDL) and has a number of major functions related to the binding and uptake of plasma lipoproteins and to the redistribution of lipids locally among cells (Mahley and Rall, 2000; Mahley, 1988; Mahley and Huang, 1999). About half of the apoE in the plasma of fasting subjects is associated with triglyceride-rich lipoproteins, and the other half is a constituent of HDL. ApoE controls the catabolism of the apoE-containing lipoproteins by mediating their binding to cell-surface heparan sulfate proteoglycans (especially in the liver) and to LDL receptors and the LRP (Mahley and Ji, 1999). About three-fourths of the apoE in plasma is synthesized by the liver and the remainder by a variety

of tissues. The brain is the second most abundant site of apoE mRNA synthesis, which occurs primarily in astrocytes. ApoE also is synthesized by macrophages, where it appears to play a role in modulating cholesterol accumulation. In transgenic mice, overexpression of apoE by macrophages inhibits hypercholesterolemia-induced atherogenesis (Bellosta et al., 1995; Hasty et al., 1999). There are three commonly occurring alleles of the apoE gene (designated 2, 3, and 4) that occur with a frequency of 8%, 77%, and 15%, respectively. These alleles code for the three major forms of apoE: E2, E3, and E4. Consequently, there are three homozygous apoE phenotypes (E2/2, E3/3, and E4/4) and three heterozygous phenotypes (E2/3, E2/4, and E3/4). Approximately 60% of the population is homozygous for apoE3. Single amino acid substitutions result from the genetic polymorphisms in the apoE gene (Mahley and Rall, 2000; Mahley, 1988). ApoE2, with a cysteine at residue 158, differs from apoE3, which has arginine at this site. ApoE3, with a cysteine at residue 112, differs from apoE4, which has arginine at this site. These single amino differences affect both receptor binding and lipid binding of the three apoE isoforms. Both apoE3 and apoE4 can bind to the LDL receptor, but apoE2 binds much less effectively and, as a consequence, causes the remnant lipoprotein dyslipidemia of type III hyperlipoproteinemia. ApoE2 and apoE3 bind preferentially to the phospholipids of HDL, whereas apoE4 binds preferentially to VLDL triglycerides. Low-Density Lipoproteins The LDL particles arising from the catabolism of IDL have a half-life of 1.5 to 2 days, which accounts for the higher plasma concentration of LDL than of VLDL and IDL. In subjects without hypertriglyceridemia, two-thirds of plasma cholesterol is found in the LDL. Plasma clearance of LDL particles is mediated primarily by LDL receptors; a small component is mediated by nonreceptor clearance mechanisms (Brown and Goldstein, 1986). Defective or absent LDL receptors cause high levels of plasma LDL and familial hypercholesterolemia (Brown and Goldstein, 1986; Hobbs et al., 1992). ApoB-100, the only apoprotein of LDL, is the ligand that binds LDL to its receptor. Residues 3000 to 3700 in the carboxyl-terminal sequence are critical for binding. Mutations in this region disrupt binding and are a cause of hypercholesterolemia (familial defective apoB-100) (Innerarity et al., 1990; Pullinger et al., 1995). The liver expresses a large complement of LDL receptors and removes 75% of all LDL from the plasma (Dietschy et al., 1993). Consequently, manipulation of hepatic LDL receptor expression is a most effective way to modulate plasma LDL and cholesterol levels. Thyroxine and estrogen enhance LDL receptor gene expression, which explains the LDL-C-lowering effects of these hormones (Windler et al., 1980; Wiseman et al., 1993). The most effective dietary (decreased consumption of saturated fat and cholesterol) and pharmacological (treatment with statins) treatments of hypercholesterolemia act by enhancing hepatic LDL receptor expression (Bilheimer et al., 1983; Woollett and Dietschy, 1994). Regulation of LDL receptor expression is part of a complex process by which cells regulate their free cholesterol content. This regulatory process is mediated by transcription factors called sterol regulatory binding element proteins (SREBPs) (Brown and Goldstein, 1998). SREBPs enhance LDL receptor expression when cellular cholesterol content is reduced. LDL become atherogenic when they are modified by oxidation (Steinberg, 1997), a required step for LDL uptake by the scavenger receptors of macrophages. This process leads to foam-cell formation in arterial lesions. At least two scavenger receptors (SRs) are involved (SR-AI/II and CD36). Knocking out either receptor in transgenic mice retards the uptake of oxidized LDL by

macrophages. Expression of the two receptors is differentially regulated; SR-AI/II appears to be more important in early atherogenesis, and CD36 more important as foam cells form during lesion progression (Nakata et al., 1999; Dhaliwal and Steinbrecher, 1999). High-Density Lipoproteins The metabolism of HDL is complex because of the multiple mechanisms by which HDL particles are modified in the plasma compartment and by which HDL particles are synthesized (Breslow, 1994; Segrest et al., 2000; Tall et al., 2000). ApoA-I is the major HDL apoprotein, and its plasma concentration is a more powerful inverse predictor of CHD risk than is the HDL-C level (Maciejko et al., 1983). ApoA-I synthesis is required for normal production of HDL. Mutations in the apoA-I gene that cause HDL deficiency are variable in their clinical expression and often are associated with accelerated atherogenesis (Assmann et al., 2001). Conversely, overexpression of apoA-I in transgenic mice protects against experimentally induced atherogenesis (Plump et al., 1994). Mature HDL can be separated by ultracentrifugation into HDL2 (d = 1.063 to 1.125 g/ml), which are larger, more cholesterol-rich lipoproteins (70 to 100 in diameter), and HDL3 (d = 1.125 to 1.21 g/ml), which are smaller particles (50 to 70 in diameter). In addition, two major subclasses of mature HDL particles in the plasma can be differentiated by their content of the major HDL apoproteins, apoA-I and apoA-II (Duriez and Fruchart, 1999). Mature HDL particles have electrophoretic mobility. Some -migrating HDL particles contain only apoA-I and no apoA-II and are called LpA-I HDL particles. Others contain both apoA-I and apoA-II and are called LpA-I/A-II HDL particles. These two particles usually are separated by electroimmunoassay and quantitated by assessment of their apoA-I content (Duriez and Fruchart, 1999). LpA-I particles are larger than LpA-I/A-II and are primarily associated with HDL2. LpA-I/A-II particles are smaller and are primarily associated with HDL3. Patients with reduced HDL-C levels and CHD have lower levels of LpA-I, but not of LpA-I/A-II, than subjects with normal HDL-C levels (Duriez and Fruchart, 1999). This finding suggests that HDL particles containing apoA-I and apoA-II may not be atheroprotective. In fact, overexpression of apoA-II in transgenic mice enhances susceptibility to atherosclerosis (Schultz et al., 1993). ApoA-II deficiency is not associated with any apparent deleterious effects in human beings (Deeb et al., 1990). The precursor of most of the plasma HDL is a discoidal particle containing apoA-I and phospholipid, called pre- 1 HDL because of its pre- 1 electrophoretic mobility. Pre- 1 HDL are synthesized by the liver and the intestine, and they also arise when surface phospholipids and apoAI of chylomicrons and VLDL are lost as the triglycerides of these lipoproteins are hydrolyzed. Phospholipid transfer protein plays an important role in the transfer of phospholipids to HDL (Tall et al., 2000). Discoidal pre- 1 HDL can then acquire free (unesterified) cholesterol from the cell membranes of tissues, such as arterial wall macrophages, by an interaction with the class B, type I scavenger receptor (SR-BI), to which the apoA-I of HDL docks, so that free cholesterol can be transferred to or from the HDL particle (Williams et al., 1999). SR-BI facilitates the movement of excess free cholesterol from cells with excess cholesterol (e.g., arterial wall foam cells) (Williams et al., 1999). In the liver, SR-BI facilitates the uptake of cholesteryl esters from the HDL without internalizing and degrading the lipoproteins. In mice, overexpression of SR-BI reduces susceptibility to atherosclerosis, and elimination of SR-BI significantly increases atherosclerosis (Krieger and Kozarsky, 1999). A homologue of SR-BI, CLA-1, has been identified in human beings (Dhaliwal and Steinbrecher, 1999). Modulation of CLA-1 expression may offer new avenues for the

management of atherogenesis (Krieger, 1999). The membrane transporter ATP-binding cassette protein 1 (ABC-1) facilitates the transfer of free cholesterol from cells to HDL (Young and Fielding, 1999; Oram and Vaughan, 2000). When ABC1 is defective, the acquisition of cholesterol by HDL is greatly diminished, and HDL levels are markedly reduced because poorly lipidated nascent HDL are metabolized rapidly. Dysfunctional mutants of ABC-1 cause the defect observed in Tangier disease, a genetic disorder characterized by extremely low levels of HDL and cholesterol accumulation in the liver, spleen, tonsils, and neurons of peripheral nerves. After free cholesterol is acquired by the pre- 1 HDL, it is esterified by lecithin:cholesterol acyltransferase. The newly esterified and nonpolar cholesterol moves into the core of the discoidal HDL. As the cholesteryl ester content increases, the HDL particle becomes spherical and less dense. These newly formed spherical HDL particles (HDL3) further enlarge by accepting more free cholesterol, which is in turn esterified by lecithin:cholesterol acyltransferase. In this way, HDL3 are converted to HDL2, which are larger and less dense than HDL3. As the cholesteryl ester content of the HDL2 increases, the cholesteryl esters of these particles begin to be exchanged for triglycerides derived from any of the triglyceride-containing lipoproteins (chylomicrons, VLDL, remnant lipoproteins, and LDL). This exchange is mediated by the cholesteryl ester transfer protein and, in human beings, accounts for the removal of about two-thirds of the cholesterol associated with HDL. The transferred cholesterol subsequently is metabolized as part of the lipoprotein into which it was transferred. The triglyceride that is transferred into HDL2 is hydrolyzed in the liver by HL, a process that regenerates smaller, spherical HDL3 particles that recirculate and acquire additional free cholesterol from tissues containing excess free cholesterol. HL activity is regulated and modulates HDL-C levels. Both androgens and estrogens affect HL gene expression, but with opposite effects (Haffner et al., 1983; Brinton, 1996). Androgens increase HL activity, which accounts for the lower HDL-C values observed in men than in women. Estrogens reduce HL activity, but their impact on HDL-C levels in women is substantially less than that of androgens on HDL-C levels in men. HL appears to have a pivotal role in regulating HDL-C levels, as HL activity is increased in many patients with low HDL-C levels. HDL are protective lipoproteins that decrease the risk of CHD; thus, high levels of HDL are desirable. This protective effect may result from the participation of HDL in reverse cholesterol transport, the process by which excess cholesterol is acquired from cells and transferred to the liver for excretion. HDL also may inhibit oxidative modification of LDL through the action of paraoxonase, an HDL-associated antioxidant protein. Lipoprotein(a) Lipoprotein(a) [Lp(a)] is composed of an LDL particle that has a second apoprotein in addition to apoB-100 (Berg, 1994). The second apoprotein, apo(a), is attached to apoB-100 by at least one disulfide bond and does not function as a lipid-binding apoprotein. Apo(a) of Lp(a) is structurally related to plasminogen and appears to be atherogenic by interfering with fibrinolysis of thrombi on the surface of plaques. Hyperlipidemia and Atherosclerosis Despite a continuing decline in the incidence of atherosclerosis-related deaths in the past 35 years, deaths from CHD, cerebrovascular disease, and peripheral vascular disease accounted for 30% of

the 2.3 million deaths in the United States during 1997. Two-thirds of atherosclerosis deaths were due to CHD. About 85% of CHD deaths occurred in individuals over 65 years of age. Among the 15% dying prematurely (below age 65), 80% died during their first CHD event. Among those dying of sudden cardiac death in 1997, 50% of the men and 63% of the women had been previously asymptomatic (American Heart Association, 1999). These statistics illustrate the importance of identifying and managing risk factors for CHD. The major known risk factors are elevated LDL-C, reduced HDL-C, cigarette smoking, hypertension, type 2 diabetes mellitus, advancing age, and a family history of premature (men < 55 years; women < 65 years) CHD events in a first-degree relative. Control of the modifiable risk factors is especially important in preventing premature CHD (events in men below 55 years or in women below 65 years). Observational studies suggest that modifiable risk factors account for 85% of excess risk (risk over and above that of individuals with optimal risk-factor profiles) for premature CHD (Stamler et al., 1986; Wilson et al., 1998). Furthermore, these studies indicate that, when total cholesterol levels are below 160 mg/dl, CHD risk is markedly attenuated, even in the presence of additional risk factors (Grundy et al., 1998). This pivotal role of hypercholesterolemia in atherogenesis gave rise to the almost universally accepted cholesterol-diet-CHD hypothesis (Thompson and Barter, 1999). The cholesterol-diet-CHD hypothesis states that elevated plasma cholesterol levels cause CHD, that diets rich in saturated fat (animal fat) and cholesterol raise cholesterol levels, and that the lowering of cholesterol levels reduces CHD risk. Although the relationship between cholesterol, diet, and CHD was recognized nearly 50 years ago, proof that cholesterol lowering was safe and prevented CHD death required extensive epidemiological studies and clinical trials. Epidemiological Studies Epidemiological studies have demonstrated the importance of the relationship between excess saturated fat consumption and elevated cholesterol levels. Reducing the consumption of dietary saturated fat and cholesterol is the cornerstone of population-based approaches to the management of hypercholesterolemia (National Cholesterol Education Program, 1990). In addition, it is clearly established that the higher the cholesterol level, the higher the CHD risk (Stamler et al., 1986). Clinical Trials Studies of the efficacy of cholesterol lowering began in the 1960s. However, it was not until the advent of powerful cholesterol-reducing drugs known as statins that clear-cut evidence of the benefit of cholesterol lowering became available (Illingworth and Durrington, 1999). Several important trials in the 1970s and 1980s showed that average cholesterol reductions of about 10% resulted in 20% reductions in nonfatal CHD events, but these trials were not large enough to detect an effect on mortality (Lipid Research Clinics Program, 1984a; Committee of Principal Investigators, 1984; Frick et al., 1987; Durrington and Illingworth, 1998). In fact, increases in noncardiac mortality in these trials raised concerns about the safety of cholesterol-lowering therapy (Wysowski and Gross, 1990). In 1994, the Scandinavian Simvastatin Survival Study (4S), a secondary prevention trial, proved for the first time that lowering cholesterol levels with simvastatin reduced total mortality among CHD patients with normal HDL levels and high mean baseline LDL-C levels (188 mg/dl). Simvastatin therapy reduced LDL-C by an average of 35%, CHD mortality by 42%, nonfatal CHD events by 40%, and total mortality by 30%. Simvastatin therapy did not increase noncardiac mortality from

any cause (Scandinavian Simvastatin Survival Study Group, 1994). Subsequently, the efficacy and safety of statin therapy in patients with established CHD at baseline was evaluated in the Cholesterol and Recurrent Events (CARE) trial and the Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) study [Sacks et al., 1996; The LongTerm Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group, 1998]. In CARE and LIPID, the average baseline LDL-C and HDL-C levels (139 and 39 mg/dl in CARE and 150 and 36 mg/dl in LIPID) were lower than in 4S. Treatment with pravastatin reduced LDL-C levels by 25% (CARE) and 28% (LIPID) and was associated with a 24% reduction in CHD death in LIPID and 29% and 28% reductions in nonfatal myocardial infarctions in LIPID and CARE, respectively. The results of the 4S, CARE, and LIPID trials indicated that CHD patients with baseline LDL-C values above 130 mg/dl benefit from lipid-lowering therapy. Subgroup analyses of patients in these trials with baseline LDL-C levels between 100 and 130 mg/dl did not consistently show a benefit from cholesterol-lowering drug therapy (Grundy, 1998b). More recently, however, the Veterans Affairs High Density Lipoprotein Intervention Trial (VA HIT) showed that CHD patients with LDL-C 104 mg/dl benefited from gemfibrozil therapy (Rubins et al., 1999). Taken together, however, the results of these four trials suggest that hypolipidemic drug therapy is beneficial in CHD patients with baseline LDL-C levels >100 mg/dl and that the goal of treatment of patients with CHD should be to reduce LDL-C to less than 100 mg/dl. There also have been clinical trials of lipid lowering in patients who had no evidence of vascular disease at baseline (primary prevention trials). The West of Scotland Coronary Prevention Study (WOSCOPS) (Shepherd et al., 1995) demonstrated a benefit of pravastatin therapy in male patients with baseline LDL-C >155 mg/dl. The average LDL-C in WOSCOPS was high (192 mg/dl), and the mean HDL-C level was 44 mg/dl. The average on-treatment LDL-C was 142 mg/dl, a 26% decrease from baseline, and this resulted in a 31% reduction in CHD death and nonfatal myocardial infarction. The second major statin trial in patients without vascular disease was the Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TexCAPS) (Downs et al., 1998). This trial included men and women who at baseline had only moderately elevated levels of LDL-C (average, 156 mg/dl) and who were primarily at risk because of age (men >45 years of age; women >55 years of age) and because of low HDL-C levels (average, 37 mg/dl). In this trial, lovastatin reduced LDL-C by 26% and primary endpoint events (fatal and nonfatal myocardial infarction and unstable angina pectoris) by 37%. The trials described above provided evidence that supported the 2001 revision of the National Cholesterol Education Program (NCEP) guidelines for the management of dyslipidemic patients. These trials demonstrated efficacy in the prevention of vascular disease events and provided convincing evidence for the short-term ( 5 years) safety of hypolipidemic therapy. National Cholesterol Education Program Guidelines for Treatment: Managing Patients with Dyslipidemia The current NCEP guidelines for management of patients with lipid disorders are of two types. One is a population-based approach, which is intended to lower blood cholesterol by dietary recommendations: Reduce total calories from fat to less than 30% and from saturated fat to less than 10%; consume less than 300 mg of cholesterol per day; and maintain desirable body weight (National Cholesterol Education Program, 1990). The second is the patient-based approach

described in the 2001 report of the NCEP Adult Treatment Panel III, which continues to focus on lowering LDL-C levels as the primary goal of therapy (National Cholesterol Education Program Expert Panel, 2001). The 2001 Adult Treatment Panel III guidelines for the management of adults 20 years and older recommend a complete lipoprotein profile (total cholesterol, LDL-C, HDL-C, and triglycerides) rather than screening for total cholesterol and HDL-C alone. Fasting for 12 hours is required to accurately measure the triglyceride and LDL-C levels [LDL-C = total cholesterol (triglyceride 5) HDL-C]. The classification of lipid levels is shown in Table 363. If the values for total cholesterol, LDL-C, and triglycerides are in the lowest category and the HDL-C level is not low, lifestyle recommendations (diet and exercise) should be made to ensure maintenance of a normal lipid profile. Other vascular disease risk factors (Table 364), if present, should be assessed and treated individually. For patients with elevated levels of total cholesterol, LDL-C, or triglycerides, or reduced HDL-C values, further treatment is based on the patient's risk-factor status (Table 364) and LDL-C levels (Table 365). All patients should receive instruction about dietary restriction of saturated fat and cholesterol. Patients with CHD or a CHD equivalent (symptomatic peripheral or carotid vascular disease, abdominal aortic aneurysm, >20% ten-year CHD risk or diabetes mellitus) should immediately start appropriate lipid-lowering therapy. Patients without CHD or CHD equivalent should be managed with lifestyle advice (diet, exercise, weight management) for 3 to 6 months before drug therapy is implemented. Before drug therapy is initiated, however, secondary causes of hyperlipidemia should be excluded. Most secondary causes (Table 366) can be excluded by ascertaining the patient's medication history and by measuring serum creatinine, liver function tests, fasting glucose, and thyroidstimulating hormone levels. Treatment of the disorder causing secondary dyslipidemia may preclude the necessity of treatment with hypolipidemic drugs. Risk Assessment Using Framingham Risk Scores The 2001 NCEP guidelines and those of the European Atherosclerosis Society (Wood et al., 1998) employ risk assessment tables devised from the Framingham Heart Study in an attempt to match the intensity of treatment to the severity of CHD risk in patients without a prior history of symptomatic atherosclerotic vascular disease. High risk or "CHD equivalent" status is defined as >20% chance of sustaining a CHD event in the next ten years. The tables used to determine a patient's absolute risk do not take into account risk associated with a family history of premature CHD or obesity. As a consequence, the risk may be seriously underestimated, resulting in insufficiently aggressive management. After calculation of the risk score, more aggressive therapy should be considered for obese patients or for patients with a family history of premature CHD. Arterial Wall Biology and Plaque Stability More effective lipid-lowering agents and a better understanding of atherogenesis have helped to prove that aggressive lipid-lowering therapy has many beneficial effects over and above those obtained by simply decreasing lipid deposition in the arterial wall. Arteriographic trials have shown that, although aggressive lipid lowering results only in very small increases in lumen diameter, it promptly decreases acute coronary events (Brown et al., 1993). Lesions causing less than 60% occlusion are responsible for more than two-thirds of the acute events. Aggressive lipid-lowering therapy may prevent acute events through its positive effects on the arterial wall; it corrects endothelial dysfunction, corrects abnormal vascular reactivity (spasms), and improves plaque stability.

Atherosclerotic lesions containing a large lipid core, large numbers of macrophages, and a poorly formed fibrous cap (Brown et al., 1993; Gutstein and Fuster, 1999) are prone to plaque rupture and acute thrombosis. Aggressive lipid lowering appears to alter plaque architecture, resulting in less lipid, fewer macrophages, and a larger collagen and smooth muscle cellrich fibrous cap. Stabilization of plaque susceptibility to thrombosis appears to be a direct result of LDL-C lowering or an indirect result of changes in cholesterol and lipoprotein metabolism or arterial wall biology (see below, "Potential Cardioprotective Effects Other Than LDL Lowering?"). Who and When to Treat? Large-scale trials with statins have provided new insights into which patients with dyslipidemia should be treated and when treatment should be initiated. Gender Both men and women benefit from lipid-lowering therapy. In fact, CARE and AFCAPS/TexCAPS showed greater benefit in women. Statins, rather than hormone-replacement therapy, are now recommended by the American Heart Association and the American College of Cardiology as the first-line drug therapy for lowering lipids in postmenopausal women. This recommendation reflects the increased CHD morbidity in older women with established CHD who were treated with hormone-replacement therapy (Hulley et al., 1998; Mosca et al., 1999). Age Age >45 years in men and >55 years in women is considered to be a CHD risk factor. The statin trials have shown that patients >65 years of age benefit from therapy as much as do younger patients. Old age per se is not a reason to withhold drug therapy in an otherwise healthy person. Cerebrovascular Disease Patients In most studies of patients with cerebrovascular disease, plasma cholesterol levels correlate positively with risk of ischemic stroke. In clinical trials, statins reduced stroke and/or transient ischemic attacks in patients with and without CHD (Hebert et al., 1997). Peripheral Vascular Disease Patients Statins have proved beneficial in patients with peripheral vascular disease. Hypertensive Patients and Smokers The risk reduction for coronary events in hypertensive patients and in smokers is similar to that in subjects without these risk factors. Type 2 Diabetes Mellitus Patients with type 2 diabetes benefit very significantly from aggressive lipid lowering (see"Treatment of Type 2 Diabetes," below). PostMyocardial Infarction or Revascularization Patients As soon as CHD is diagnosed, it is essential to begin lipid-lowering therapy (NCEP guidelines:

LDL-C < 100 mg/dl). Compliance with drug therapy is greatly enhanced if treatment is initiated in the hospital (Fonarow and Gawlinski, 2000). It remains to be determined if statin therapy alters restenosis after angioplasty; however, the NHLBI Post Coronary Artery Bypass Graft trial showed that statin therapy improved the long-term outcome after bypass surgery and that the lower the LDL-C, the better (The Post Coronary Artery Bypass Graft Trial Investigators, 1997). Can Cholesterol Levels Be Lowered Too Much? Are there total and LDL cholesterol levels below which adverse health consequences begin to increase? Observational studies initially were confusing. In the United States and western Europe, low cholesterol levels appeared to be associated with an increase in noncardiac mortality from chronic pulmonary disease, chronic liver disease, cancer (many primary sites), and hemorrhagic stroke. However, more recent data indicate that it is the noncardiac diseases that cause the low plasma cholesterol levels and not the low cholesterol levels that cause the noncardiac diseases (Law et al., 1994). One exception may be hemorrhagic stroke. In the Multiple Risk Factor Intervention Trial (MRFIT), hemorrhagic stroke occurred more frequently in hypertensive patients with total cholesterol levels below 160 mg/dl; however, the increased incidence of hemorrhagic stroke was more than offset by reduced CHD risk due to the low cholesterol levels (Neaton et al., 1992). In addition, in a study of the Chinese population, in which cholesterol levels rarely exceeded 160 mg/dl, lower levels of total cholesterol were not associated with increases in hemorrhagic stroke or any other cause of noncardiac mortality (Chen et al., 1991). Abetalipoproteinemia and hypobetalipoproteinemia, two rare disorders in human beings that are associated with extremely low total cholesterol levels, are instructive because affected individuals have reduced CHD risk and no increase in noncardiac mortality (Welty et al., 1998). Patients who are homozygous for the mutations that cause these disorders have total cholesterol levels below 50 mg/dl and triglyceride levels below 25 mg/dl. Individuals consuming very low levels of total fat (less than 5% of total calories) and vegetarians, who consume no animal fat, usually have total cholesterol levels below 150 mg/dl and have no increase in noncardiac mortality (Appleby et al., 1999). Based on the lack of harm associated with low total cholesterol levels in these various groups, reducing cholesterol levels to similarly low levels with drugs does not appear to be contraindicated. With the advent of more efficacious cholesterol-lowering agents, it soon may be possible to test the benefits and risks of lowering total cholesterol levels below 150 mg/dl. Will Greater Lipid Lowering Further Reduce CHD? Despite the remarkable results of the statin trials (a 25% to 40% reduction in events), there are still 60% to 75% as many events in the treatment groups of the statin trials as in the placebo groups. Do these results suggest that even more aggressive lipid lowering is required (Grundy, 1998b)? The investigators who conducted the AFCAPS/TexCAPS trial suggested that the 1993 NCEP guidelines are still too conservative for high-risk subjects without CHD, like those enrolled in the AFCAPS/TexCAPS study (Downs et al., 1998). They state that drug therapyalong with a prudent diet, regular exercise, and risk factor modificationshould be used to lower the risk of the first acute major coronary event in primary prevention candidates who are older (men 45 years, women 55 years), have HDL-C 50 mg/dl, and have LDL-C 130 mg/dl. However, subgroup analysis of baseline HDL-C levels in AFCAPS/TexCAPS indicated that only patients with HDL-C <40 mg/dl benefited, suggesting that it may not be cost-effective to treat patients with risk factor profiles like those in AFCAPS/TexCAPS if HDL-C exceeds 40 mg/dl. The 2001 NCEP guidelines would not

initiate treatment of many patients like those in AFCAPS/TexCAPS unless the LDL-C was >160 mg/dl in patients without CHD, and the 2001 NCEP guidelines have a higher target LDL-C level than AFCAPS/TexCAPS (130 versus <110 mg/dl). As more effective lipid-lowering drugs and better combinations of therapies are developed, we will be able to lower lipid levels more effectively. But will lower cholesterol levels translate into a further reduction of clinical events (Figure 362)? Many researchers believe that the answer is yes. In addition, as statins become generic drugs, more aggressive treatment of wider segments of the population will become more cost-effective. Figure 362. Reduction in Coronary Heart Disease Events in Clinical Trials Is Associated with the Extent of Cholesterol Lowering. As More Potent Cholesterol-Reducing Agents Become Available, Will It Be Possible to Reduce Events by 50% or More in a Typical 5-Year Trial? (Adapted from Thompson and Barter, 1999, and Used by Permission of Lippincott Williams & Wilkins.) AFCAPS, Air Force/Texas Coronary Atherosclerosis Prevention Study; CARE, Cholesterol and Recurrent Events trial; LIPID, Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) study; LRC, Lipid Research Clinics Coronary Primary Prevention Trial; POSCH, Program on the Surgical Control of the Hyperlipidemias; 4S, Scandinavian Simvastatin Survival Study; WOS, West of Scotland Coronary Prevention Study.

Treatment of Type 2 Diabetes Diabetes mellitus is an independent predictor of high risk for CHD. CHD morbidity is two to four times higher in patients with diabetes than in nondiabetics, and the mortality from CHD is up to 100% higher in diabetic patients than in nondiabetics over a 6-year period (Grundy et al., 1999). Glucose control is essential, but this provides only minimal benefit with respect to CHD prevention. Aggressive treatment of diabetic dyslipidemia through diet, weight control, and drugs (in most cases) is critical in reducing risk. Diabetic dyslipidemia is usually characterized by high triglycerides, low HDL-C, and moderate elevations of total cholesterol and LDL-C. Recent recommendations from the American Heart Association and the American Diabetes Association indicate that the treatment guidelines for diabetic patients should be the same as for patients with CHD (Grundy et al., 1999). The revised 2001 NCEP guidelines also will reflect this recommendation. Haffner et al. (1998) reported that diabetics without diagnosed CHD are at the same level of risk as nondiabetics with established

CHD. This is consistent with the recommendation to reduce plasma LDL-C levels of all diabetics to <100 mg/dl, irrespective of whether they have had a prior ischemic vascular disease event. The American Diabetes Association further recommends that the first line of treatment for a diabetic dyslipidemia usually should be a statin (Grundy et al., 1999). Clinical trials with simvastatin, pravastatin, and lovastatin have clearly established in post hoc analyses that diabetics profit from cholesterol lowering as much as other subgroups or even more. For example, diabetics in the 4S, CARE, AFCAPS/TexCAPS, and LIPID trials had 55%, 25%, 43%, and 19% reductions in events, respectively. The Diabetes Atherosclerosis Intervention Study recently showed a significant benefit of treating type 2 diabetics with fenofibrate. This 3-year arteriographic study demonstrated a 40% decrease of focal coronary stenoses (p= 0.029) (Diabetes Atherosclerosis Intervention Study Investigators, 2001). Metabolic Syndrome The 2001 NCEP guidelines recognize the increased CHD risk associated with the insulin-resistant, prediabetic state described under the rubric of "metabolic syndrome." This syndrome consists of a constellation of five CHD risk factors (Table 367). The 2001 NCEP guidelines arbitrarily define the presence of three or more of these risk factors as indicating that a patient is affected. Treatment should focus on weight loss and increased physical activity since overweight and obesity usually preclude optimal risk factor reduction. Specific treatment of increased LDL-C and triglyceride levels and low HDL-C levels should be undertaken as well. Treatment of Hypertriglyceridemia The 2001 NCEP guidelines reflect the increased CHD risk associated with the presence of triglyceride levels above 150 mg/dl. Three categories of hypertriglyceridemia are recognized (Table 363), and treatment is recommended based on the degree of elevation. Weight loss, increased exercise, and alcohol restriction are important for all hypertriglyceridemic patients. The LDL-C goal should be ascertained based on each patient's risk factor or CHD status (Table 365). If triglycerides remain above 200 mg/dl after the LDL-C goal is reached, further reduction in triglycerides may be achieved by increasing the dose of a statin or of niacin. Combination therapy (statin plus niacin or statin plus fibrate) may be required, but caution is necessary with these combinations to avoid myopathy (see Statins in Combination with Other Lipid-lowering Drugs, below). Treatment of Low HDL-C The most frequent risk factor for premature CHD is low HDL-C. In a study of 321 men with angiographically documented CHD, 60% had HDL-C levels of <35 mg/dl and only 25% had LDLC >160 mg/dl (Genest et al., 1991). In a separate study of more than 8500 older men with CHD, 38% had HDL-C levels <35 mg/dl (Rubins et al., 1995). Data from the Framingham Heart Study demonstrate that subjects with "normal" cholesterol levels of <200 mg/dl but with low HDL-C (<40 mg/dl) have as much CHD risk as subjects with higher total cholesterol levels (230 to 260 mg/dl) and more normal HDL-C (40 to 49 mg/dl) (Castelli et al., 1986). In low-HDL-C patients, the total cholesterol/HDL-C ratio is a particularly useful predictor of CHD risk. The Framingham data indicate that the ideal ratio is 3.5 and a ratio of >4.5 is associated with increased risk (Castelli, 1994). American men, who are a high-risk group, have a typical ratio of 4.5. Patients with low HDL-C may have what are considered to be "normal" total and LDL cholesterol levels and would not qualify for therapy, butbecause of the low HDL-Cmay be at

high risk based on the total cholesterol/HDL-C ratio (e.g., a total cholesterol of 180 mg/dl and an HDL-C of 30 mg/dl yields a ratio of 6.0). A desirable total cholesterol level in low-HDL-C patients may be considerably lower than 200 mg/dl. This is especially true because many low-HDL-C patients also have moderately elevated triglycerides, which may reflect increased levels of atherogenic remnant lipoproteins (Grundy, 1998a). Results from AFCAPS/TexCAPS (a primary prevention trial) and VA HIT (a secondary prevention trial) are particularly relevant. Patients in these trials had average or low LDL-C, low HDL-C, and high total cholesterol/ HDL-C ratios and treatment greatly reduced clinical events in both trials. The 2001 NCEP guidelines extend treatment to include some, but not all, patients typical of those who benefited in AFCAPS/TexCAPS and VA HIT. See Table 368 for a summary of trial results (Downs et al., 1998; Rubins et al., 1999). The treatment of low HDL-C patients according to the 2001 NCEP guidelines is focused on lowering LDL-C to the target level based on the patient's risk factor or CHD status (Table 365) and a reduction of VLDL cholesterol (estimated by dividing the plasma triglyceride level by 5) below 30 mg/dl. If this strategy results in a ratio of total cholesterol/HDL-C that is 4.0 or less, it will be optimal. Patients with higher total cholesterol/HDL-C ratios (>4.5) will still be at risk even if their "non-HDL-C" levels (LDL-C and VLDL cholesterol) are at the goal values recommended by the 2001 NCEP guidelines. Drug Therapy of Dyslipidemia In addition to the present discussion, the topic of drug therapy for dyslipidemia has been extensively reviewed by Durrington and Illingworth (1998). Statins The statins are the most effective and best-tolerated agents for treating dyslipidemia. These drugs are competitive inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, which catalyzes an early, rate-limiting step in cholesterol biosynthesis. Higher doses of the more potent statins (e.g., atorvastatin and simvastatin) also can reduce triglyceride levels caused by elevated VLDL levels. Some statins also are indicated for raising HDL-C levels, although the clinical significance of these effects on HDL-C remains to be proven. Five large, well-controlled clinical trials have documented the efficacy and safety of simvastatin, pravastatin, and lovastatin in reducing fatal and nonfatal CHD events, strokes, and total mortality [Scandinavian Simvastatin Survival Study Group, 1994; Shepherd et al., 1995; Sacks et al., 1996; Downs et al., 1998; The Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group, 1998]. Rates of adverse events in all five trials were the same in the placebo groups and in the groups receiving active drug. This was true with regard to noncardiac illness and the two laboratory tests, hepatic transaminases and creatine kinase (CK), that have been most frequently used to monitor patients taking statins. History Statins were isolated from a mold, Penicillium citrinium, and identified as inhibitors of cholesterol biosynthesis in 1976 by Endo and colleagues (Endo et al., 1976). Subsequently, Brown et al., (1978) established that statins act by inhibiting HMG-CoA reductase. The first statin studied in human beings was compactin, renamed mevastatin, which demonstrated the therapeutic potential of this class of drugs (Yamamoto et al., 1984). However, Alberts and colleagues at Merck developed

the first statin (lovastatin; formerly known as mevinolin) that was approved for use in human beings, which was isolated from Aspergillus terreus (Alberts et al., 1980; Bilheimer et al., 1983). Since the approval of lovastatin by the United States Food and Drug Administration (FDA) in 1987, five other statins have been approved. Two of these, pravastatin and simvastatin, are chemically modified derivatives of lovastatin (seeFigure 363). The more recently approved statins atorvastatin, fluvastatin, and cerivastatinare synthetic compounds. More statins are under development. Figure 363. Chemical Structures of the Statins and the Reaction Catalyzed by 3Hydroxy-3-Methylglutaryl Coenzyme a (HMG-CoA) Reductase.

Chemistry The structural formulas of the original statin (mevastatin) and the six statins currently available in the United States are shown in Figure 363 along with the reaction (conversion of HMG-CoA to mevalonate) catalyzed by HMG-CoA reductase, the enzyme they competitively inhibit. The statins possess a side group that is structurally similar to HMG-CoA. Mevastatin, lovastatin, simvastatin, and pravastatin are fungal metabolites, and each contains a hexahydronapthalene ring. Lovastatin differs from mevastatin in having a methyl group at carbon 3. There are two major side chains. One is a methylbutyrate ester (lovastatin and pravastatin) or a dimethylbutyrate ester (simvastatin). The other contains a hydroxy acid that forms a six-membered analog of the intermediate compound in the HMG-CoA reductase reaction (Figure 363). Fluvastatin, atorvastatin, and cerivastatin are entirely synthetic compounds containing a heptanoic acid side chain that forms a structural analog of the HMG-CoA intermediate. As a result of their structural similarity to HMG-CoA, statins are reversible competitive inhibitors of the enzyme's natural substrate, HMG-CoA. The inhibition constant (Ki) of cerivastatin is 0.01

nM (Bischoff et al., 1997); all other statins have a K i in the 1-nM range. The dissociation constant of HMG-CoA is three orders of magnitude higher than this value. Lovastatin and simvastatin are lactone prodrugs that are modified in the liver to active hydroxy acid forms. Since they are lactones, they are less soluble in water than are the other statins, a difference that appears to have little if any clinical significance. Pravastatin (an acid in the active form), fluvastatin and cerivastatin (sodium salts), and atorvastatin (a calcium salt), are all administered in the active, open-ring form. Mechanism of Action Statins exert their major effectreduction of LDL levelsthrough a mevalonic acidlike moiety that competitively inhibits HMG-CoA reductase by product inhibition (Alberts et al., 1980). Statins affect blood cholesterol levels by inhibiting cholesterogenesis in the liver, which results in increased expression of the LDL receptor gene. In response to the reduced free cholesterol content within hepatocytes, membrane-bound SREBPs are cleaved by a protease and translocated to the nucleus. The transcription factors are then bound by the sterol-responsive element of the LDL receptor gene, enhancing transcription and ultimately increasing the synthesis of LDL receptors (Brown and Goldstein, 1998). Degradation of LDL receptors also is reduced (Brown et al., 1978). The greater number of LDL receptors on the surface of hepatocytes results in increased removal of LDL from the blood (Bilheimer et al., 1983), thereby lowering LDL-C levels. Some studies suggest that statins also can reduce LDL levels by enhancing the removal of LDL precursors (VLDL and IDL) and by decreasing hepatic VLDL production (Grundy and Vega, 1985; Arad et al., 1990; Aguilar-Salinas et al., 1998). Since VLDL remnants and IDL are enriched in apoE, a statin-induced increase in the number of LDL receptors, which recognize both apoB-100 and apoE, enhances the clearance of these LDL precursors (Gaw et al., 1993). The reduction in hepatic VLDL production induced by statins is thought to be mediated by reduced synthesis of cholesterol, a required component of VLDL (Thompson et al., 1996). This mechanism also likely accounts for the triglyceride-lowering effect of statins (Ginsberg, 1998) and may account for the approximately 25% reduction of LDL-C levels in patients with homozygous familial hypercholesterolemia treated with 80 mg of either atorvastatin or simvastatin (Raal et al., 1997, 2000). Triglyceride Reduction by Statins Triglyceride levels greater than 250 mg/dl are reduced substantially by statins, and the percent reduction achieved is similar to the percent reduction in LDL-C (Stein et al., 1998). Accordingly, hypertriglyceridemic patients taking the highest doses (80 mg/day) of two of the most potent statins (simvastatin, atorvastatin) experience a 35% to 45% reduction in LDL-C and a similar reduction in fasting triglyceride levels (Bakker-Arkema et al., 1996; Ose et al., 2000). If baseline triglyceride levels are below 250 mg/dl, reductions in triglycerides do not exceed 25% irrespective of the dose or statin used (Stein et al., 1998). Similar reductions (35% to 45%) in triglycerides can be accomplished with usual doses of fibrates or niacin (see below), although these drugs do not reduce LDL-C to the same extent as 80-mg doses of atorvastatin or simvastatin. Effect of Statins on HDL-C Levels Most studies of patients treated with statins have systematically excluded patients with low HDL-C levels. In studies of patients with elevated LDL-C levels and gender-appropriate HDL-C levels (40 to 50 mg/dl for men; 50 to 60 mg/dl for women), an increase in HDL-C of 5% to 10% was observed, irrespective of the dose or statin employed. However, in patients with reduced HDL-C

levels (<35 mg/dl), preliminary studies suggest that statins differ in their effects on HDL-C levels. Simvastatin, at its highest dose of 80 mg, increases HDL-C and apoA-I levels more than a comparable dose of atorvastatin (Crouse et al., 1999; Crouse et al., 2000). However, more studies are needed to ascertain the effects of statins on HDL-C in patients with low HDL-C levels and to determine if statin effects on HDL-C are clinically meaningful. Effects of Statins on LDL-C Levels Statins lower LDL-C by 20% to 55%, depending on the dose and statin used. In large trials comparing the effects of the various statins, equivalent doses appear to be 5 mg of simvastatin = 15 mg of lovastatin = 15 mg of pravastatin = 40 mg of fluvastatin (Pedersen and Tobert, 1996); 20 mg of simvastatin = 10 mg of atorvastatin (Jones et al., 1998; Crouse et al., 1999), and 20 mg of simvastatin = 0.4 mg of cerivastatin. Analysis of dose-response relationships for all statins demonstrate that the efficacy of LDL-C lowering is log linear; LDL-C is reduced by 6% (from baseline) with each doubling of the dose (Pedersen and Tobert, 1996; Jones et al., 1998). Maximal effects on plasma cholesterol levels are achieved within 7 to 10 days. Table 369 provides information regarding the doses of the various statins that are required to reduce LDL-C by 20% to 55%. The percent reductions achieved with the various doses are the same regardless of the absolute value of the baseline LDL-C level. The statins are effective in virtually all patients with high LDL-C levels. The exception is patients with homozygous familial hypercholesterolemia, who have very attenuated responses to the usual doses of statins, because both alleles of the LDL receptor gene code for dysfunctional LDL receptors; the partial response in these patients is due to a reduction in hepatic VLDL synthesis associated with the inhibition of HMG-CoA reductasemediated cholesterol synthesis (Raal et al., 1997, 2000). Statin therapy does not reduce Lp(a) levels (Kostner et al., 1989). Potential Cardioprotective Effects Other Than LDL Lowering? Although the statins clearly exert their major effects on CHD by lowering LDL-C and improving the lipid profile as reflected in plasma cholesterol levels (Thompson and Barter, 1999) (Figure 36 2), a multitude of potentially cardioprotective effects are being ascribed to these drugs (Davignon and Laaksonen, 1999), largely on the basis of in vitro and ex vivo data. However, the mechanisms of action for nonlipid roles of statins have not been established, and it is not known whether these potential pleiotropic effects represent a class-action effect, differ among statins, or are biologically relevant. Until these questions are resolved, selection of a specific statin should not be based on any one of these effects. Nevertheless, the potential importance of the nonlipid roles of statins requires some discussion. Statins and Endothelial Function A variety of studies have established that the vascular endothelium plays a dynamic role in vasoconstriction/relaxation and that hypercholesterolemia modulates these processes directly. Acetylcholine-induced vasodilation of coronary arteries is depressed in patients with hypercholesterolemia and in patients with vascular disease (Treasure et al., 1995). Statin therapy improves coronary vasodilation in response to acetylcholine. The vasomotor response is controlled by nitric oxide, which is synthesized by endothelial cell nitric oxide synthase. Statins stabilize endothelial cell nitric oxide synthase mRNA, thereby enhancing synthesis of endothelial cell nitric oxide (Laufs et al., 1998). Statin therapy reverses endothelial dysfunction as monitored by vasoactivity within as short a period as one month (O'Driscoll et al., 1997), but similar results have been observed after a single acute reduction of LDL levels by apheresis (Tamai et al., 1997). In

nonhuman primates fed a high-cholesterol diet, statin therapy improved endothelial function independently of significant changes in plasma cholesterol levels (Williams et al., 1998). Statins and Plaque Stability As discussed earlier, the vulnerability of plaques to rupture and thrombosis is of greater clinical relevance than the degree of stenosis they cause (Gutstein and Fuster, 1999). Statins may affect plaque stability in a variety of ways. There are reports that statins inhibit monocyte infiltration into the artery wall in a rabbit model (Bustos et al., 1998) and inhibit macrophage secretion of matrix metalloproteinases in vitro (Bellosta et al., 1998). The metalloproteinases degrade all extracellular matrix components and thus weaken the fibrous cap of atherosclerotic plaques. Statins also appear to modulate the cellularity of the artery wall by inhibiting proliferation of smooth muscle cells and enhancing apoptotic cell death (Corsini et al., 1998). It is debatable whether these effects would be beneficial or harmful if they occurred in vivo. Reduced proliferation of smooth muscle cells and enhanced apoptosis could retard initial hyperplasia and restenosis but also could weaken the fibrous cap and destabilize the lesion. Interestingly, statin-induced suppression of cell proliferation and the induction of apoptosis have been extended to tumor biology. The effects of statins on isoprenoid biosynthesis and protein prenylation associated with reduced availability of mevalonate may alter the development of malignancies (Davignon and Laaksonen, 1999). Statins and Inflammation Appreciation of the importance of inflammatory processes in atherogenesis is growing (Ross, 1999), and statins have been suggested to have an antiinflammatory role (Rossen, 1997). In a retrospective analysis of blood samples from the CARE trial, Ridker et al. (1998) demonstrated that the C-reactive protein concentration was a marker for high CHD risk and that statin therapy decreased baseline C-reactive protein levels and risk of CHD independently of cholesterol lowering. It remains to be determined whether the C-reactive protein is simply a marker of inflammation or it contributes to the pathogenesis of atherosclerosis. Statins and Lipoprotein Oxidation Oxidative modification of LDL appears to play a key role in mediating the uptake of lipoprotein cholesterol by macrophages and in other processes, including cytotoxicity within lesions (Steinberg, 1997). Statins reduce the susceptibility of lipoproteins to oxidation both in vitro and ex vivo (Kleinveld et al., 1993; Hussein et al., 1997b). Furthermore, atorvastatin has been reported to stabilize or increase the plasma level of paraoxonase, the antioxidation enzyme associated with plasma HDL (Aviram et al., 1998). Statins and Coagulation Statins reduce platelet aggregation (Hussein et al., 1997a), and in vitro model systems indicate that statins reduce the deposition of platelet thrombi on porcine aorta (Lacoste et al., 1995). In addition, the different statins have variable effects on fibrinogen levels, the significance of which remains to be determined (Rosenson and Tangney, 1998). Elevated plasma fibrinogen levels are associated with an increase in the incidence of CHD (Ernst and Resch, 1993). However, it remains to be determined whether fibrinogen is involved in the pathogenesis or is a marker of disease.

Statins in Combination with Other Lipid-Lowering Drugs Statins plus the bile acidbinding resins cholestyramine and colestipol (see below) produce 20% to 30% greater reductions in LDL-C than can be achieved with statins alone (Tikkanen, 1996). Preliminary data indicate that colesevelam hydrochloride plus a statin lowers LDL-C by 8% to 16% more than do statins alone. Niacin also can enhance the effect of statins, but the occurrence of myopathy increases when statin doses greater than 25% of maximum (e.g., 20 mg of simvastatin or atorvastatin) are used with niacin (Guyton and Capuzzi, 1998). A fibrate (clofibrate, gemfibrozil, or fenofibrate; see below) plus a statin is particularly useful in patients with hypertriglyceridemia and a high LDL-C level. This combination does increase the risk of myopathy, but it usually is safe to use a fibrate at its usual maximal dose plus no more than 25% of each statin's maximal dose (Tikkanen, 1996; Athyros et al., 1997). Triple therapy with resins, niacin, and statins can reduce LDL-C by up to 70% (Malloy et al., 1987). Absorption, Fate, and Excretion All the statins are administered as the active -hydroxy acid form except for lovastatin and simvastatin, which are administered in the lactone form and must be hydrolyzed in vivo to the corresponding -hydroxy acid. All the statins are subject to extensive first-pass metabolism by the liver. Less than 5% to 20% of a dose reaches the general circulation. Greater than 95% of most of these drugs and their active metabolites, the -hydroxy acids, are bound to plasma proteins. After an oral dose, plasma concentrations peak in 1 to 4 hours. The liver biotransforms all statins, resulting in low systemic availability of the parent compounds. About 70% of statin metabolites are excreted by the liver (Corsini et al., 1999). Lovastatin and simvastatin are administered as prodrugs and are converted to their active hydroxy acid forms in the liver. About 30% of lovastatin and up to 85% of simvastatin are absorbed. About 50% of the active metabolites are protein bound. The liver is the primary route of excretion. Pravastatin is administered in its active -hydroxy acid form as a sodium salt, and 34% of oral doses are absorbed. There is extensive first-pass metabolism, and about 50% of the circulating drug is protein bound. Its metabolites possess little HMG-CoA reductase inhibitory activity. Pravastatin is excreted without extensive metabolic modification; the liver is the major route of excretion, although up to 20% can be excreted in the urine (Quion and Jones, 1994; Corsini et al., 1999). Fluvastatin also is administered in its active form as a sodium salt and is almost completely absorbed. The drug is metabolized (50% to 80%) to inactive metabolites, and >90% of its excretion is hepatic (Corsini et al., 1999). Fluvastatin is the only statin with saturable first-pass hepatic metabolism; consequently, it is the only statin to achieve peak plasma concentrations in the micromolar range. However, the clinical relevance of this is unclear, as fluvastatin and its metabolites have the shortest elimination half-life of available statins (Dain et al., 1993; Corsini et al., 1999). Atorvastatin is administered as a calcium salt. It is extensively transformed in the liver to ortho- and parahydroxylated derivatives, which account for about 70% of the circulating inhibitory activity of HMG-CoA reductase. Atorvastatin and its active metabolites are metabolized principally in the liver, where they are excreted into the bile. Atorvastatin has a half-life of about 20 hours, but the half-life of plasma HMG-CoA reductase inhibitory activity is up to 30 hours. All of the other statins have half-lives of only 1 to 4 hours. The clinical significance, if any, of the prolonged half-life of atorvastatin is unclear, but it is thought to play a role in the greater efficacy of atorvastatin

compared with the other statins (Christians et al., 1998; Corsini et al., 1999). Cerivastatin is given in its active form as a sodium salt. Absorption is greater than 98%. In the liver, cerivastatin is transformed into two major and two minor active metabolites. About 70% of these metabolites are excreted in the feces, and the remainder is excreted in the urine (Corsini et al., 1999). Adverse Effects and Drug Interactions Hepatotoxicity The initial postmarketing surveillance studies of the various statins revealed up to 1% incidence of elevations in hepatic transaminase to values greater than three times the upper limit of normal. The incidence appears to be dose related. However, in the placebo-controlled outcome trials, in which 20- to 40-mg doses of simvastatin, lovastatin, or pravastatin were used, the incidence of threefold elevations in hepatic transaminases was not significantly increased in the active drug treatment groups [Scandinavian Simvastatin Survival Study Group, 1994; Shepherd et al., 1995; Sacks et al., 1996; Downs et al., 1998; The Long-Term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group, 1998]. Serious hepatopathy must be rare, although it appears to occur (Nakad et al., 1999). Primarily because of the safety data from the clinical trials, it is reasonable to measure alanine aminotransferase (ALT) at baseline and 3 to 6 months after therapy is initiated or after increasing the dose. If the ALT values are normal, it is not necessary to repeat the ALT test more than every 6 to 12 months. Myopathy The only major adverse effect of clinical significance associated with statin use is myopathy. All statins have been associated with myopathy and rhabdomyolysis (Pogson et al., 1999; Physicians' Desk Reference, 2001). The incidence of myopathy is low (<0.1%) in patients taking statins without concomitant administration of drugs that enhance the risk of myopathy. Two classes of drugs, fibrates (gemfibrozil, clofibrate, and fenofibrate) and niacin, also are lipid-lowering drugs and can cause myopathy by themselves (see below). When statins are administered with fibrates or niacin, the myopathy is probably caused by an enhanced inhibition of skeletal muscle sterol synthesis (a pharmacodynamic interaction) (Christians et al., 1998). The other drugs are those that, like most statins, are metabolized by the 3A4 isoform of cytochrome P450 (CYP3A4) and include certain macrolide antibiotics (e.g., erythromycin), azole antifungals (e.g., itraconazole), cyclosporine, a phenylpiperazine antidepressant, nefazodone, and protease inhibitors (Christians et al., 1998; Fichtenbaum et al., 2000; Dresser et al., 2000). This interaction is pharmacokinetic and is associated with increased plasma concentrations of statins and their active metabolites (Christians et al., 1998). Atorvastatin, cerivastatin, lovastatin, and simvastatin are primarily metabolized by CYP3A4, but cerivastatin also can be metabolized by CYP2C8. Recently, cerivastatin plus gemfibrozil therapy has been contraindicated because of a number of case reports of myopathy (Physicians' Desk Reference, 2001). Fluvastatin is mostly (50% to 80%) metabolized by CYP2C9 to inactive metabolites, but CYP3A4 and CYP2C8 also contribute to the metabolism of fluvastatin. Pravastatin, however, is not metabolized to any appreciable extent by the CYP system (Corsini et al., 1999) but is excreted unchanged in the urine. Pravastatin and fluvastatin, which are not extensively metabolized by the CYP3A4 system, may be less likely to cause myopathy when used with one of the predisposing drugs. However, because cases of myopathy have been reported with both drugs, the benefits of combined therapy with any statin should be carefully weighed against the risk of myopathy.

The myopathy syndrome is characterized by intense myalgia, first in the arms and thighs and then in the entire body (similar to flu-related myalgia) along with fatigue. The symptoms progress as long as patients continue to take the drug or drugs that induce the myopathy. Myoglobinuria and renal failure have been reported (Pogson et al., 1999). Serum creatine kinase (CK) levels in affected patients typically are 10-fold higher than the upper limit of normal. As soon as myopathy is suspected, a blood sample should be drawn to document the presence of a 10-fold elevation of CK, as many patients complain of muscle pain that is not due to true statin-induced myopathy. The statin, and any other drug suspected of contributing to myopathy, should be discontinued if true myopathy is suspected, even if it is not possible to measure CK activity to document the presence of myopathy. Rhabdomyolysis should be excluded and renal function monitored. Since myopathy rarely occurs in the absence of combination therapy, routine CK monitoring is not recommended unless the statins are used with one of the predisposing drugs. Such monitoring is not sufficient to protect patients, as myopathy can occur months to years after combined therapy is initiated. As a rule, statins may be used in combination with one of these predisposing drugs with reduced risk of myopathy if the statin is administered at no more than 25% of its maximal dose (Christians et al., 1998). Although cataracts were a concern initially, careful monitoring of patients in the early days of statin use failed to document any statin-induced eye pathology (Bradford et al., 1991). Differences in the solubility of the statins (hydrophilic versus hydrophobic) prompted speculation that the more lipidsoluble drugs might be more likely to penetrate the central nervous system. However, cerebrospinal fluid concentrations of the two lipid-soluble statins, lovastatin and simvastatin, are very low, which is probably due to extensive plasma protein binding of these drugs. Differences in lipid solubility among statins do not appear to be clinically relevant. Pregnancy The safety of statins during pregnancy has not been established. Women wishing to conceive should not take statins. During their childbearing years, women taking statins should use highly effective contraceptive procedures. Nursing mothers also are advised to avoid taking statins. Therapeutic Uses Each statin has a low recommended starting dose that reduces LDL-C by 20% to 30%. Since studies have documented that a majority of dyslipidemic patients remain on their initial dose and are not titrated to achieve their target LDL-C level, this approach leads to undertreatment. For this reason, it is advisable to start patients on a dose that will achieve the patient's target goal for LDL-C lowering. For example, a patient with a baseline LDL-C of 150 mg/dl and a goal of 100 mg/dl requires a 33% reduction in LDL-C and should be started on a dose that will provide it (seeTable 369). The manufacturer's initial recommended dose of lovastatin ( MEVACOR ) is 20 mg and is slightly more effective if taken with the evening meal than if it is taken at bedtime. If it is more convenient, bedtime dosing is preferable to missing doses. The statins that have half-lives of 4 hours or less should be given in the evening because hepatic cholesterol synthesis is maximal between midnight and 2:00 a.m. The dose may be increased every 3 to 6 weeks up to a maximum of 80 mg per day. The 80-mg dose is slightly (2% to 3%) more effective if given as 40 mg twice daily. The FDA-approved starting dose of simvastatin ( ZOCOR ) for most patients is 20 mg at bedtime unless the required LDL-C reduction exceeds 45%, in which case a 40-mg starting dose is indicated. The maximal dose is 80 mg, and the drug should be taken at bedtime. In patients taking

cyclosporine, fibrates, or niacin, the daily dose should not exceed 20 mg. Pravastatin ( PRAVACHOL ) therapy is initiated with a 10- or 20-mg dose that may be increased to 40 mg. This drug should be taken at bedtime. Since pravastatin is a hydroxy acid, it is bound by bileacid sequestrants, which reduces its absorption. Practically, this is rarely a problem since the resins should be taken before meals and pravastatin should be taken at bedtime. The starting dose of fluvastatin ( LESCOL ) is 20 or 40 mg, and the maximum is 80 mg per day. Like pravastatin, it is administered as a hydroxy acid and should be taken at bedtime, several hours after ingesting a bile-acid sequestrant. Atorvastatin ( LIPITOR ) has a long half-life, which allows administration of this statin at any time of the day. The starting dose is 10 mg, and the maximum is 80 mg per day. Cerivastatin (BAYCOL) is available in doses ranging between 0.2 and 0.4 mg. It should be taken at bedtime and several hours after a dose of a bile-acid sequestrant. The maximal FDA-approved dose is 0.8 mg. The choice of statins should be based on efficacy and cost. Can the dose of a particular statin reduce the patient's LDL-C to the target level? Cost should be the next discriminating factor. However, three drugs (lovastatin, simvastatin, and pravastatin) have been used safely in clinical trials involving thousands of subjects for 5 or more years. The documented safety records of these statins should be considered, especially when initiating therapy in younger patients. Once drug treatment is initiated, it is almost always lifelong. Baseline determinations of ALT and repeat testing at 3 to 6 months are recommended. If ALT is normal after the initial 3 to 6 months, then it need not be repeated more than once every 6 to 12 months. CK measurements are not routinely necessary unless the patient also is taking a drug that enhances the risk of myopathy. However, even if CK levels are monitored every 3 to 4 months, patients receiving combined therapy may develop myopathy months to years after starting the therapy. Bile-Acid Sequestrants The two established bile-acid sequestrants or resins (cholestyramine and colestipol) are among the oldest of the hypolipidemic drugs, and they are probably the safest, since they are not absorbed from the intestine (West et al., 1980; Groot et al., 1983). These resins are the only hypocholesterolemic drugs currently recommended for children 11 to 20 years of age, although data now are emerging that document the safety of statin therapy of children in this age range (National Cholesterol Education Program, 1991; Stein et al., 1999). Because statins are so effective as monotherapy, the resins are most often used as second agents if statin therapy does not lower LDLC levels sufficiently. When used with a statin, cholestyramine and colestipol usually are prescribed at submaximal doses. Maximal doses can reduce LDL-C by up to 25% but are associated with unacceptable gastrointestinal side effects (bloating and constipation) that limit compliance. Colesevelam is a new bile-acid sequestrant that is prepared as an anhydrous gel and taken as a tablet. It lowers LDL-C by 18% at its maximum dose. The safety and efficacy of colesevelam have not been studied in pediatric patients or pregnant women. Cholestyramine was used in the Coronary Primary Prevention Trial, one of the first studies to document that lowering LDL-C prevents heart disease events (Lipid Research Clinics Program, 1984a; Lipid Research Clinics Program, 1984b). Cholestyramine therapy reduced total cholesterol and LDL-C by 13% and 20%, respectively, compared with diet-induced reductions of 5% in total cholesterol and 8% in LDL-C. CHD events (fatal and nonfatal) were reduced by 19%, suggesting

that a 1% reduction in total cholesterol is associated with at least a 2% reduction in CHD events. Chemistry Cholestyramine and colestipol are anion-exchange resins. Cholestyramine, a polymer of styrene and divinylbenzene with active sites formed from trimethylbenzylammonium groups, is a quaternary amine (Figure 364). Colestipol, a co-polymer of diethylenetriamine and 1-chloro-2,3epoxypropane, is a mixture of tertiary and quaternary diamines (Figure 364). Cholestyramine and colestipol are hygroscopic powders administered as chloride salts and are insoluble in water. Colesevelam is a polymer, poly(allylamine hydrochloride), cross-linked with epichlorohydrin and alkylated with 1-bromodecane and (6-bromohexyl)-trimethylammonium bromide (Figure 364). It is a hydrophilic gel and insoluble in water. Figure 364. Structures of Cholestyramine, Colestipol, and Colesevelam.

Mechanism of Action The bile-acid sequestrants are highly positively charged and bind negatively charged bile acids. Because of their large size, resins are not absorbed, and the bound bile acids are excreted in the stool. Since over 95% of bile acids are normally reabsorbed, interruption of this process depletes the liver's pool of bile acids, and hepatic bile-acid synthesis increases. As a result, hepatic cholesterol content declines, stimulating the production of LDL receptors, an effect similar to that of statins

(Bilheimer et al., 1983). The increase in hepatic LDL receptors increases LDL clearance and lowers LDL-C levels, but this effect is partially offset by the enhanced cholesterol synthesis caused by upregulation of HMG-CoA reductase (Shepherd et al., 1980; Reihnr et al., 1990). Inhibition of reductase activity by a statin substantially increases the effectiveness of the resins. The resininduced increase in bile-acid production is accompanied by an increase in hepatic triglyceride synthesis, which is of consequence in patients with significant hypertriglyceridemia (baseline triglyceride level >250 mg/dl). In such patients, bile-acid sequestrant therapy may cause striking increases in triglyceride levels. An initial report suggests that colesevelam may not raise triglyceride levels significantly; however, baseline triglyceride levels in the patients studied were not elevated (Davidson et al., 1999). Consequently, use of colesevelam to lower LDL-C levels in hypertriglyceridemic patients should be accompanied by frequent (every 1 to 2 weeks) monitoring of fasting triglyceride levels until the triglyceride level is stable, or its use should be avoided until data are published supporting the use of colesevelam in these patients. Effects on Lipoprotein Levels The reduction in LDL-C is dose-dependent. Doses of 8 to 12 g of cholestyramine or 10 to 15 g of colestipol are associated with 12% to 18% reductions in LDL-C (Casdorph, 1975; Hunninghake et al., 1981). Maximal doses (24 g of cholestyramine or 30 g of colestipol) may reduce LDL-C by as much as 25%, but cause gastrointestinal side effects and are poorly tolerated by most patients. One to two weeks is sufficient to attain maximal LDL-C reduction by a given resin dose. In patients with normal triglyceride levels, triglycerides may increase transiently and then return to baseline. HDL-C levels increase 4% to 5% (Witztum et al., 1979). Statins plus resins or niacin plus resins can reduce LDL-C by as much as 40% to 60% (Kane et al., 1981; Illingworth et al., 1981; Davignon and Pearson, 1998). Colesevelam in doses of 3.0 to 3.75 g reduces LDL-C levels significantly in a dosedependent manner by 9% to 19%. Adverse Effects and Drug Interactions The resins are quite safe, as they are not systemically absorbed (West et al., 1980). Since they are administered as a chloride salt, rare instances of hyperchloremic acidosis have been reported. Severe hypertriglyceridemia is a contraindication to the use of cholestyramine and colestipol since these resins increase triglyceride levels (Crouse, 1987). At present, there are insufficient data on the effect of colesevelam on triglyceride levels. Cholestyramine and colestipol both are available as a powder that must be mixed with water and drunk as a slurry. The gritty sensation is objectionable to patients initially but can be tolerated easily. Colestipol is available in a tablet form that reduces the complaint of grittiness but not the gastrointestinal symptoms. Colesevelam is available as a hard capsule that absorbs water and creates a soft, gelatinous material that allegedly minimizes the potential for gastrointestinal irritation. The main objections of patients taking cholestyramine and colestipol are bloating and dyspepsia. However, these symptoms can be substantially reduced if the drug is completely suspended in liquid several hours before ingestion. For example, evening doses can be mixed in the morning and refrigerated; morning doses can be mixed the previous evening and refrigerated. Constipation still may occur but sometimes can be prevented by adequate daily water intake and psyllium, if necessary. Fecal impaction has been reported. Limited data from 106 patients treated with colesevelam for six weeks suggest that it may not cause the dyspepsia, bloating, and constipation observed in patients treated with cholestyramine or colestipol (Davidson et al., 1999).

Cholestyramine and colestipol bind many other drugs and interfere with their absorption. Drugs affected include, but are not limited to, some thiazides, furosemide, propranolol, l-thyroxine, some cardiac glycosides, coumarin anticoagulants, and some of the statins (Farmer and Gotto, 1994). The effect of cholestyramine and colestipol on the absorption of most drugs has not been studied. For this reason, it is wise to administer all drugs either 1 hour before or 3 to 4 hours after a dose of cholestyramine or colestipol. Colesevelam does not appear to interfere with the absorption of fatsoluble vitamins or of drugs such as digoxin, lovastatin, warfarin, metoprolol, quinidine, and valproic acid. However, the maximum concentration and the area under the curve of sustainedrelease verapamil were reduced by 31% and 11%, respectively, when coadministered with colesevelam. Since the effect of colesevelam on the absorption of other drugs has not been tested, it seems prudent to recommend that patients take other medications 1 hour before or 3 or 4 hours after a dose of colesevelam. Therapeutic Uses Cholestyramine resin ( QUESTRAN, QUESTRAN LIGHT, LOCHOLEST LIGHT, PREVALITE ) is available in bulk (with scoops that measure a 4-g dose) or in individual packets of 4 g. Additional flavorings are added to increase palatability. The "light" preparations contain artificial sweeteners rather than sucrose. Colestipol hydrochloride ( COLESTID, FLAVORED COLESTID) is available in bulk, in individual packets containing 5 g of colestipol, or as 1-g tablets. The powdered forms of cholestyramine (4 g per dose) and colestipol (5 g per dose) are either mixed with a fluid (water or juice) and drunk as a slurry or mixed with crushed ice in a blender. Each patient should be instructed to determine the most palatable slurry. Resins should never be taken in the dry form. Ideally, patients should take the resins before breakfast and before supper, starting with one scoop or packet twice daily, and then increase the number of scoops or packets after several weeks or longer as needed and as tolerated. Most studies that have carefully monitored patients using resins have found that patients generally will not take more than two doses (scoops or packets) twice a day. Colesevelam hydrochloride (WELCHOL) is available as a solid tablet containing 0.625 g of colesevelam. The starting dose is either 3 tablets taken twice daily with meals or all 6 tablets taken with a meal. The tablets should be taken with a liquid. The maximum daily dose is 7 tablets (4.375 g). Nicotinic Acid (Niacin) Nicotinic acid (niacin, pyridine-3-carboxylic acid) is one of the oldest drugs used to treat dyslipidemia and is the most versatile in that it favorably affects virtually all lipid parameters (Altschul et al., 1955; Knopp, 1998). Niacin is a water-soluble B-complex vitamin that functions as a vitamin only after its conversion to nicotinamide adenine dinucleotide (seeChapter 63: WaterSoluble Vitamins: The Vitamin B Complex and Ascorbic Acid). Oral nicotinamide may be used as a source of niacin for its vitamin functions, but nicotinamide does not affect lipid levels. The hypolipidemic effects of niacin require larger doses than are required for its vitamin effects. Niacin is the best agent available for increasing HDL-C (increments of 30% to 40%); it also lowers triglycerides by 35% to 45% (as effectively as fibrates and the more potent statins) and reduces LDL-C levels by 20% to 30% (Knopp et al., 1985; Vega and Grundy, 1994; Martin-Jadraque et al., 1996). Niacin also is the only lipid-lowering drug that reduces Lp(a) levels significantly, by about 40% (Carlson et al., 1989). However, adequate control of other lipid abnormalities renders an elevation of Lp(a) harmless (Maher et al., 1995). The only other drugs that significantly lower Lp(a) levels are estrogen and neomycin (Gurakar et al., 1985; Espeland et al., 1998; Shlipak et al., 2000).

Despite its salutary effect on lipids, niacin has side effects that limit its use (see"Adverse Effects," below). Mechanism of Action Niacin has multiple effects on lipoprotein metabolism. In adipose tissue, niacin inhibits the lipolysis of triglycerides by hormone-sensitive lipase, which reduces transport of free fatty acids to the liver and decreases hepatic triglyceride synthesis (Grundy et al., 1981). In the liver, niacin reduces triglyceride synthesis by inhibiting both the synthesis and esterification of fatty acids, effects that increase apoB degradation (Jin et al., 1999). Reduction of triglyceride synthesis reduces hepatic VLDL production, which accounts for the reduced LDL levels. Niacin also enhances LPL activity, which promotes the clearance of chylomicrons and VLDL triglycerides. Niacin raises HDL-C levels by decreasing the fractional clearance of apoA-I in HDL rather than by enhancing HDL synthesis (Blum et al., 1977). This effect is due to a reduction in the hepatic clearance of HDL-apoA-I, but not of cholesteryl esters, thereby increasing the apoA-I content of plasma and augmenting reverse cholesterol transport (Jin et al., 1997). Effects on Plasma Lipoprotein Levels Niacin in doses of 2 to 6 g per day reduces triglycerides by 35% to 50%, and the maximal effect occurs within 4 to 7 days (Figge et al., 1988). Reductions of 25% in LDL-C levels are possible with doses of 4.5 to 6 g per day, but 3 to 6 weeks are required for maximal LDL reductions to be observed. HDL-C increases less in patients with low HDL-C levels (<35 mg/dl) than in those with higher levels. Average increases of 15% to 30% occur in patients with low HDL-C levels; greater increases may occur in patients with normal HDL-C levels at baseline (Vega and Grundy, 1994). Combination therapy with resins can reduce LDL-C levels by as much as 40% to 60% (Malloy et al., 1987). Absorption, Fate, and Excretion The pharmacological doses of regular (crystalline) niacin (>1 g per day) used to treat dyslipidemia are almost completely absorbed, and peak plasma concentrations (up to 0.24 mM) are achieved within 30 to 60 minutes. The half-life is about 60 minutes, which accounts for the necessity of twice or thrice daily dosing. At lower doses, most niacin is taken up by the liver; only the major metabolite, nicotinuric acid, is found in the urine. At higher doses, a greater proportion of the drug is excreted in the urine as unchanged nicotinic acid (Iwaki et al., 1996). Adverse Effects Two of niacin's side effects, flushing and dyspepsia, limit patient compliance. The cutaneous effects include flushing and pruritus of the face and upper trunk, skin rashes, and acanthosis nigricans. Flushing and associated pruritus are prostaglandin mediated (Stern et al., 1991). Flushing is worse when therapy is initiated or the dosage is increased, but after 1 to 2 weeks of a stable dose, most patients no longer flush. Taking an aspirin each day alleviates the flushing in many patients. Flushing recurs if only one or two doses are missed, and the flushing is more likely to occur when niacin is consumed with hot beverages (coffee, tea) or with ethanol-containing beverages. Flushing is minimized if therapy is initiated with low doses (100 to 250 mg twice daily) and if the drug is taken after breakfast or supper. Dry skin, a frequent complaint, can be dealt with by using skin moisturizers, and acanthosis nigricans can be dealt with by using lotions or creams containing salicylic acid. The dyspepsia and rarer episodes of nausea, vomiting, and diarrhea are less likely to occur if the drug is taken after a meal. Patients with a history (even a remote history) of peptic ulcer

disease should not take niacin because it reactivates ulcer disease. The most common, medically serious side effects are hepatotoxicity, which causes elevated serum transaminases, and hyperglycemia. Both regular (crystalline) niacin and sustained-release niacin, which was developed to reduce flushing and itching, have been reported to cause severe liver toxicity, and sustained-released niacin can cause fulminant hepatic failure (Christensen et al., 1961; Mullin et al., 1989; McKenney et al., 1994; Tatet al., 1998). A newer preparation of sustainedrelease niacin ( NIASPAN ) appears to be less likely to cause severe hepatotoxicity (Capuzzi et al., 1998), perhaps simply because it is administered once daily instead of more frequently (Guyton et al., 1998). The incidence of flushing and pruritus with this preparation is not substantially different from that with regular niacin. Patients using sustained-release niacin may develop hepatic toxicity at any time, but severe toxicity appears to occur when patients take more than 2 g of sustained-release, over-the-counter preparations. Affected patients experience flu-like fatigue and weakness. Usually, aspartate transaminase and ALT are elevated, serum albumin levels decline, and all serum lipid levels decline substantially. In fact, concurrent reductions in LDL-C and HDL-C in a patient taking niacin should prompt concern about niacin toxicity. In patients with diabetes mellitus, niacin-induced insulin resistance can cause severe hyperglycemia, and the drug usually is not recommended for use in these patients (Knopp et al., 1985; Henkin et al., 1991; Schwartz, 1993). In fact, niacin use in patients with diabetes mellitus often mandates a change to insulin therapy. If niacin is prescribed for patients with known or suspected diabetes, blood glucose levels should be monitored at least once a week until proven to be stable. Niacin also elevates uric acid levels and may reactivate gout. A history of gout is a relative contraindication for niacin use. More rare side effects include toxic amblyopia and toxic maculopathy, which are reversible. Atrial tachyarrhythmias and atrial fibrillation have been reported, more commonly in elderly patients. Niacin, at doses used in human beings, has been associated with birth defects in experimental animals and should not be taken by pregnant women. Therapeutic Uses Niacin is indicated for hypertriglyceridemia and elevated LDL-C; it is especially useful in patients with both hypertriglyceridemia and low HDL-C levels. There are two commonly available forms of niacin. Crystalline niacin (immediate-release or regular) refers to niacin tablets that dissolve quickly after ingestion. Sustained-release niacin refers to preparations that continuously release niacin for 6 to 8 hours after ingestion. Crystalline niacin tablets do not require a prescription and are available in a variety of strengths from 50-mg to 500-mg tablets. To minimize the flushing and pruritus, it is best to start with a low dose (e.g., 100 mg twice daily taken after breakfast and supper). The dose may be increased stepwise every 7 days by 100 to 200 mg to a total daily dose of 1.5 to 2.0 g. After 2 to 4 weeks at this dose, transaminases, serum albumin, fasting glucose, and uric acid levels should be measured. Lipid levels should be checked and the dose increased further until the desired effect on plasma lipids is achieved. After a stable dose is attained, blood should be drawn every 3 to 6 months to monitor for the various toxicities. Since concurrent use of niacin and statins can cause myopathy, the dose of the statin should be maintained at no more than 25% of each particular statin's maximal dose. Patients also should be instructed to expect flu-like muscle aches throughout the body if myopathy occurs. Routine measurement of CK in patients taking niacin and statins does not assure that severe myopathy will be prevented or detected, as patients have developed myopathy after several years of concomitant

use of niacin with a statin. Over-the-counter, sustained-release niacin preparations are effective up to a total daily dose of 2.0 g per day. All doses of sustained-release niacin, but particularly doses above 2 g per day, have been reported to cause hepatotoxicity, which may occur soon after beginning therapy or after several years of use (Knopp et al., 1985). Although the exact incidence of hepatotoxicity from use of sustained-release niacin is unknown because it can be purchased without a prescription, the potential for severe liver damage should preclude its use in most patients, including those who have taken an equivalent dose of crystalline niacin safely for many years and are considering switching to a sustained-release preparation (Mullin et al., 1989). Fibric Acid Derivatives History Thorp and Waring (1962) reported that ethyl chlorophenoxyisobutyrate lowered lipid levels in rats. In 1967, the ester form (clofibrate) was approved for use in the United States and was, for a number of years, the most widely prescribed hypolipidemic drug. Its use declined dramatically, however, after the results of the World Health Organization (WHO) trial were published in 1978. This trial found that, despite a 9% reduction in cholesterol levels, clofibrate treatment did not reduce fatal cardiovascular events, although nonfatal infarcts were reduced (Committee of Principal Investigators, 1978). Total mortality was significantly greater in the clofibrate group. The increased mortality was due to multiple causes, including cholelithiasis. Interpretation of these negative results was clouded by failure to analyze the data according to the intention-to-treat principle. A later analysis demonstrated that the apparent increase in noncardiac mortality did not persist in the clofibrate-treated patients after discontinuation of the drug (Heady et al., 1992). Clofibrate use was virtually abandoned after the 1978 WHO trial publication, although it, as well as two other fibrates, gemfibrozil and fenofibrate, remain available in the United States. Two subsequent trials, the Helsinki Heart Study and the Veterans Affairs HDL Intervention Trial, have reported favorable effects of gemfibrozil therapy on fatal and nonfatal cardiac events without an increase in morbidity or mortality (Frick et al., 1987; Rubins et al., 1999). Chemistry Clofibrate, the prototype of the fibric acid derivatives, is the ethyl ester of pchlorophenoxyisobutyrate. Gemfibrozil is a nonhalogenated phenoxypentanoic acid and thus is distinct from the halogenated fibrates. A number of fibric acid analogs (e.g., fenofibrate, bezafibrate, and ciprofibrate) have been developed and are used in Europe and elsewhere (seeFigure 365 for structural formulas). Figure 365. Structures of the Fibric Acids.

Mechanism of Action Despite extensive studies in human beings, the mechanisms by which fibrates lower lipoprotein levels, or raise HDL levels, remain unclear (Grundy and Vega, 1987; Illingworth, 1991). Recent studies suggest that many of the effects of these compounds on blood lipids are mediated by their interaction with peroxisome proliferatoractivated receptors (PPARs) (Kersten et al., 2000), which regulate gene transcription. Three PPAR isotypes ( , , and ) have been identified. Fibrates bind to PPAR , which is expressed primarily in the liver and brown adipose tissue and to a lesser extent in the kidney, heart, and skeletal muscle. Fibrates reduce triglycerides through PPAR -mediated stimulation of fatty acid oxidation, increased LPL synthesis, and reduced expression of apoC-III. An increase in LPL would enhance the clearance of triglyceride-rich lipoproteins. A reduction in hepatic production of apoC-III, which serves as an inhibitor of lipolytic processing and receptormediated clearance, would enhance the clearance of VLDL. Fibrate-mediated increases in HDL-C are due to PPAR stimulation of apoA-I and apoA-II expression (Staels and Auwerx, 1998), which increases HDL levels. LDL levels rise in many patients, especially hypertriglyceridemic patients, treated with gemfibrozil. However, LDL levels are unchanged or fall in others, especially those whose triglyceride levels are not elevated or who are taking a second-generation agent, such as fenofibrate, bezafibrate, or ciprofibrate. The fall of LDL levels may be due in part to changes in the cholesterol and triglyceride contents of LDL that are mediated by cholesteryl ester transfer protein activity; such changes can alter the affinity of LDL for the LDL receptor (Eisenberg et al., 1984). There also is evidence that a PPAR -mediated increase in hepatic SREBP-1 production enhances hepatic expression of LDL

receptors (Kersten et al., 2000). Lastly, gemfibrozil reduces the plasma concentration of small, dense, more easily oxidized LDL particles (Yuan et al., 1994). Most of the fibric acid agents have potential antiatherothrombotic effects, including inhibition of coagulation and enhancement of fibrinolysis. These salutary effects also could alter cardiovascular outcomes by mechanisms unrelated to any hypolipidemic activity (Watts and Dimmitt, 1999). Effects on Lipoprotein Levels The effects of the fibric acid agents on lipoprotein levels differ widely depending on the starting lipoprotein profile, the presence or absence of a genetic hyperlipoproteinemia, the associated environmental influences, and the drug used. Patients with type III hyperlipoproteinemia (dysbetalipoproteinemia) are among the most sensitive responders to fibrates (Mahley and Rall, 2001). Elevated triglyceride and cholesterol levels are dramatically lowered, and tuberoeruptive and palmar xanthomas may regress completely. Angina and intermittent claudication also improve (Kuo et al., 1988). In patients with mild hypertriglyceridemia (e.g., triglycerides <400 mg/dl), fibrate treatment decreases triglyceride levels by up to 50% and increases HDL-C concentrations about 15%; LDL-C levels may be unchanged or increase. The second-generation agents, such as fenofibrate, bezafibrate, and ciprofibrate, lower VLDL levels to a degree similar to that produced by gemfibrozil, but they also are more likely to decrease LDL levels by 15% to 20%. In patients with more marked hypertriglyceridemia (e.g., 400 to 1000 mg/dl), a similar fall in triglycerides occurs, but LDL increases of 10% to 30% are seen frequently. Normotriglyceridemic patients with heterozygous familial hypercholesterolemia usually experience little change in LDL levels with gemfibrozil; with the other fibric acid agents, reductions as great as 20% may occur in some patients. Fibrates usually are the drugs of choice for treating severe hypertriglyceridemia and the chylomicronemia syndrome. While the primary therapy is to remove alcohol and as much fat from the diet as possible, fibrates help both by increasing triglyceride clearance and by decreasing hepatic triglyceride synthesis. In patients with chylomicronemia syndrome, fibrate maintenance therapy and a low-fat diet keep triglyceride levels well below 1000 mg/dl and thus prevent episodes of pancreatitis. Gemfibrozil was used in the Helsinki Heart Study, a primary prevention trial of 4081 hyperlipidemic men who received either placebo or gemfibrozil for 5 years (Frick et al., 1987). Gemfibrozil reduced total cholesterol by 10% and LDL-C by 11%, raised HDL-C levels by 11%, and decreased triglycerides by 35%. Overall, there was a 34% decrease in fatal and nonfatal cardiovascular events without any effect on total mortality. No increased incidence of gallstones or cancers was observed. Subgroup analysis suggested that the greatest benefit occurred in the subjects with the highest levels of VLDL or combined VLDL and LDL and in those with the lowest HDL-C levels (<35 mg/dl). It also is possible that gemfibrozil affected the outcome by influencing platelet function, coagulation factor synthesis, or LDL size. In a recent secondary prevention trial, gemfibrozil reduced fatal and nonfatal CHD events by 22% despite a lack of effect on LDL-C levels. HDL-C levels increased by 6%, which may have contributed to the favorable outcome (Rubins et al., 1999). Absorption, Fate, and Excretion

All of the fibrate drugs are absorbed rapidly and efficiently (>90%) when given with a meal but less efficiently when taken on an empty stomach. The ester bond is hydrolyzed rapidly, and peak plasma concentrations are attained within 1 to 4 hours. More than 95% of these drugs in plasma are bound to protein, nearly exclusively to albumin. The half-lives of fibrates differ significantly (Miller and Spence, 1998), ranging from 1.1 hours (gemfibrozil) to 20 hours (fenofibrate). The drugs are widely distributed throughout the body, and concentrations in liver, kidney, and intestine exceed the plasma level. Gemfibrozil is transferred across the placenta. The fibrate drugs are excreted predominantly as glucuronide conjugates; 60% to 90% of an oral dose is excreted in the urine, with smaller amounts appearing in the feces. Excretion of these drugs is impaired in renal failure, though excretion of gemfibrozil was reported to be less severely compromised in renal insufficiency than was excretion of other fibrates (Evans et al., 1987). Nevertheless, the use of fibrates is contraindicated in patients with renal failure. Adverse Effects and Drug Interactions Fibric acid compounds usually are well tolerated (Miller and Spence, 1998). Side effects may occur in 5% to 10% of patients but most often are not sufficient to cause discontinuation of the drug. Gastrointestinal side effects occur in up to 5% of patients. Other side effects are reported infrequently and include rash, urticaria, hair loss, myalgias, fatigue, headache, impotence, and anemia. Minor increases in liver transaminases and decreases in alkaline phosphatase have been reported. Clofibrate, bezafibrate, and fenofibrate have been reported to potentiate the action of oral anticoagulants, in part by displacing them from their binding sites on albumin. Careful monitoring of the prothrombin time and reduction in dosage of the anticoagulant may be appropriate when treatment with a fibrate is begun. A myositis flu-like syndrome occasionally occurs in subjects taking clofibrate, gemfibrozil, and fenofibrate, and may occur in up to 5% of patients treated with a combination of an HMG-CoA reductase inhibitor and gemfibrozil, if higher doses of the reductase inhibitor are used. Patients on this combination should be instructed to be aware of the potential symptoms and should be followed at 3-month intervals with careful history and determination of CK values until a stable pattern is established. Use of fibrates with cerivastatin should be avoided because a number of cases of rhabdomyolysis have resulted from combined gemfibrozil-cerivastatin therapy (Guyton et al., 1999; Alexandridis et al., 2000). All of the fibrates increase the lithogenicity of bile. In the Coronary Drug Project and the WHO trial, clofibrate use was associated with increased risk of gallstone formation. However, no significant increase was seen in the Helsinki Heart Study with the use of gemfibrozil or in the VA HIT. Placebo-controlled clinical trial data for fenofibrate are not available. Renal failure is a relative contraindication to the use of fibric acid agents, as is hepatic dysfunction. Combined statin-fibrate therapy should be avoided in patients with compromised renal function. Gemfibrozil should be used with caution and at a reduced dosage to treat the hyperlipidemia of renal failure. Fibrates should not be used by children or pregnant women. Therapeutic Uses Clofibrate (ATROMID-S) is available for oral administration. The usual dose is 2 g per day in divided doses. This compound is little used, but it still may be useful in patients who do not tolerate gemfibrozil or fenofibrate. Gemfibrozil ( LOPID ) is usually administered as a 600-mg dose taken twice a day 30 minutes before the morning and evening meals. Fenofibrate (TRICOR) is available as

single-dose capsules of 67, 134, and 200 mg. Fibrates are the drugs of choice for treating hyperlipidemic subjects with type III hyperlipoproteinemia as well as subjects with severe hypertriglyceridemia (triglycerides >1000 mg/dl), who are at risk for pancreatitis. Based on the VA HIT results, fibrates appear to have an important role in subjects with familial combined hyperlipidemia, who predominantly have elevated VLDL levels and low HDL-C levels. When fibrates are used in such patients, the LDL levels need to be monitored; if LDL levels rise, the addition of a low dose of a statin may be needed. Alternatively, many experts now treat such patients first with a statin, and only subsequently add a 600-mg dose of gemfibrozil once or twice a day to further lower VLDL levels. If this combination is used, there should be careful monitoring for myositis. Prospectus New Lipid-Lowering Agents Statins More potent statins capable of lowering LDL-C by >65% are being developed. Some of these agents also may have greater efficacy in reducing triglycerides and raising HDL-C. A new statin, ZD4522, in developmental clinical trials has been reported to reduce LDL-C by 65% (Olsson et al., 2000). MTP Inhibitor MTP transfers triglycerides and other nonpolar lipids to the apolipoproteins of nascent lipoproteins as they form in the intestine and liver and is required for the synthesis and secretion of chylomicrons and VLDL. For example, an MTP inhibitor targeted to the liver would decrease VLDL production, thereby decreasing plasma triglyceride levels and ultimately reducing LDL production from VLDL. One such MTP inhibitor, BMS-201038, is in clinical trials (Wetterau et al., 1998). Dietary and Biliary Cholesterol Absorption Inhibitor Ezetimibe is an azetidione-based cholesterol absorption inhibitor that blocks the intestinal absorption of cholesterol, resulting in lowered plasma total cholesterol and LDL-C levels (van Heek et al., 2000). The drug undergoes glucuronidation in the intestine, and the absorbed glucuronide, an active metabolite, is excreted into the bile by the liver. Due to its enterohepatic circulation, the halflife of the active metabolite is 22 hours in human beings, which indicates that once daily dosing is sufficient (Zhu et al., 2000). The maximum effective dose is 10 mg daily, which provides a 19% reduction in LDL-C (Lipka et al., 2000). A 19% reduction of LDL-C by ezetimibe is equivalent to three doublings of a statin from its baseline dose, since each doubling of a statin dose after the starting dose provides an additional 6% decrease in LDL-C (Pedersen and Tobert, 1996). In human beings, the combination of simvastatin (20 mg; the starting dose of simvastatin) plus ezetimibe (10 mg) provides a 52% reduction in LDL-C (Kosoglou et al., 2000). This 50% reduction achieved with low-dose simvastatin plus ezetimibe (10 mg) is equivalent to that of 80 mg of simvastatin alone. Since there is virtually no myopathy when statins are used at their starting doses, the combination of ezetimibe plus low-dose statin promises an additional margin of safety. Early studies suggest that there are no dangerous interactions between ezetimibe and statins, but further investigations are required to adequately document this initial impression.

ACAT Inhibitors To date, ACAT inhibitors have failed to reach the marketplace. However, the recognition that there are at least two forms of this enzyme with specific tissue sites of expression suggests that it may be possible to develop an inhibitor that specifically inhibits the assimilation of dietary cholesterol. ACAT-1 is expressed in several tissues, including macrophages (Chang et al., 1993). Avasimibe, an inhibitor of this enzyme, appears to reduce macrophage and cholesteryl ester contents of lesions in cholesterol-fed rabbits and could affect atherosclerotic lesion development, an effect that could stabilize lesions (Bocan et al., 2000). Avasimibe reduced plasma triglycerides and LDL-C levels by up to 50% in miniature swine; however, hepatictriglyceride content increased two- to sevenfold in a dose-dependent manner (Burnett et al., 1999). Interestingly, ACAT-1 knockout mice did not have reduced susceptibility to developing atherosclerosis (Accad et al., 2000). ACAT-2 is expressed in the liver and intestine and appears to play a role in cholesteryl ester formation for VLDL and chylomicron production (Cases et al., 1998). An inhibitor of this form of ACAT could reduce plasma lipids. For further discussion of hyperlipoproteinemias and other disorders of lipid metabolism, seeChapter 335 in Harrison's Principles of Internal Medicine, 16th ed., McGraw-Hill, New York, 2005.