Symposium on Intestinal Infections

The Pathophysiology of
Secretory Diarrheas
Secretory Diarrheas

Charles c.]. Carpenter, M.D.*

Secretory diarrheas are those diarrheas that are caused by net move-
ment of fluid into the gut lumen, without disruptive changes in the intes-
tinal mucosa. Secretory diarrhea can result from increased movement of
electrolytes and water into the lumen, from decreased movement of electro-
lytes and water out of the lumen, or from a combination of both abnormali-
ties. Net secretion generally occurs as a result of a biochemical alteration in
the small bowel mucosal cells. Diarrhea occurs whenever the fluid secreted
into the small bowel exceeds the absorptive capacity of the colon. ll In gen-
eral, neither the secretory nor the absorptive activity of the colon is signifi-
cantly affected during secretory diarrheas, with the notable exceptions of
diarrhea caused by villous adenomas of the colon and that caused by bile
salt and fatty acid malabsorption (in which increased colonic secretion ap-
pears to play a contributory role).
Secretory diarrheas are characterized by a stool that is virtually isotonic
with plasma and has many of the characteristics of an ultrafiltrate
ultra filtrate of plasma.
In adults, the osmolarity of the stool in true secretory diarrhea is almost en-
tirely accounted for by lueasurable
measurable electrolytes. R8 In small children there is
often a significant gap between the osmolarity and the measured electro-
lytes; this gap consists largely of organic acids.
Generally, any acute diarrheal illness in which intestinal fluid loss ex-
ceeds 10 ml per kg per 24 hours can be classified as a secretory diarrhea. 88
The converse is not true, in that each of the mechanisms of secretory diar-
rheas may result in a wide spectrum of clinical illnesses. 22
Table 1 presents a pathogenetic classification of the secretory diarrheas.
By far the great majority of secretory diarrheas are caused by bacterial en-
terotoxins. Such illnesses, of which Escherichia coli is the most common
etiology, may be the most common cause of death throughout the world to-
day.2 The noninfectious causes of secretory diarrheas are numerically far
less common, but are important in terms of infornlation
information that they provide
about abnormal intestinal function.

·Physician-in-Chief, Case Western Reserve University, University Hospitals of Cleveland, Cleve-

*Physician-in-Chief,
land, Ohio

Medical Clinics of North America-Vo!.

America-VD!. 66, No. 3, May 1982 597
598 CHARLES C. J. CARPENTER

Table 1. Pathogenesis of Secretory Diarrheas

INFECTIOUS:
r:':FECTIOUS: BACTERIAL ENTEHOTOXINS
ENTEROTOXINS

Vihrio
Vibrio cholerae
V. cholerae, non 0 group 1
Escherichia coli
Clostridium perfringens
Bacillus cereus
Staphylococcus aureus

NONINFECTIOUS
NO!\INFECnous

Hormones secreted by tunlors

tumors
Surreptitious ingestion of laxatives
Direct mucosal hypersecretion (villous adenomas)
Bile salt and/or fatty acid malabsorption

THE BACTERIAL SECRETORY DIARRHEAS

The bacterial secretory diarrheas depend on the delivery of specific en-

terotoxins to intestinal mucosal cells. In the case of cholera and enterotoxi-
genic E. coli, diarrhea depends on colonization of the small bowel by the
pathogenic organisms, which, in turn, produce the secretory enterotoxins. 2
The active peristalsis of the small bowel is an effective deterrent to the
successful colonization of most microorganisms entering from the stomach.
The organisms (Vibrio cholerae, E. coli) that are able to colonize this area
have developed special colonization factors
hlctorS such as fimbria (hair-line projec-
tions from the bacterial cell wall surface) or lectins (special proteins that at-
tach to specific carbohydrate binding sites), which allow them to associate
intimately with the mucosal cell surface, and not be washed downstream by
peristalsis. .
V. cholerae and E. coli that have colonized the small bowel do not in-
vade mucosal cells, but grow to large numbers (1066 to 1088 organisms per ml
of fluid) and produce enterotoxins, which cause the diarrhea. The enterotox-
ins bind to the mucosal cells and cause hypersecretion of isotonic fluid at a
rate that overwheln1s
overwhelms the reabsorptive capacity of the colon. Effortless diar-
rhea results. In this type of diarrhea, there is little or no inflammatory re-
sponse by the host, and the patient generally has no fever or other systemic
symptoms. The diarrhea that results is watery, often volun1inous,
voluminous, with an
electrolyte content that reflects its source of production, and without protein
or inflammatory cells. This loss of diarrheal fluid results in predictable sa-
line depletion, base-deficit acidosis, and potassium deficiency. The alTIOunt
amount
and rate of fluid loss determine the severity of the physiologic derangements
in the host. The electrolyte pattern in the stool is ren1arkably
remarkably constant when-
ever voluminous diarrhea is caused by any of the enterotoxigenic pathogens.
The mean stool electrolyte concentrations, as shown in Table 2, can there-
fore serve as a guide to fluid repletion needs in all the enterotoxigenic diar-
rheal illnesses.
THE PATHOPHYSIOLOGY OF SECRETORY DIARRHEAS 599
Table 2. Stool Electrolyte Patterns in 38 Adult Patients with Severe
Diarrheal Disease Caused by V. cholerae*
Sodiun1t
Sodiumt 128
Potassiun1
PotassiUln 16
Bicarhonate
Bicarbonate 48
Chloridet lJO
90

*Mean values are given in n1Eq

mEq per L.
tln very young children « 10 kg), mean sodium and chloride values are roughly 15
tIn 1.5 n1Eq
mEq per
L less than those in adults.

In the enterotoxigenic secretory diarrheas, caused characteristically by

V. cholerae and E. coli enterotoxins, the intestinal fluid losses may be mas-
sive, exceeding one liter per hour in adults.
In the case of the staphylococcal, clostridial, and Bacillus cereus diar-
rheas, pathogenesis largely depends on the ingestion of preformed toxin that
has been produced by bacteria in heavily contaminated foodstuffs. Once
these specific toxins have come in contact with the intestinal mucosa, diar-
rhea results by one of the several mechanisms discussed below. In the diar-
rheal illnesses caused by these bacterial enterotoxins, just as in diarrheal
diseases caused by viruses and protozoa, the fluid losses seldom exceed
750 ml per d_ay
day in adults.

Vibrio cholerae
Cholera is the prototype of the secretory bacterial diarrheas because the
bacterium neither invades nor damages the gut mucosa, and all pathophys-
iologic effects are mediated by a single enterotoxin that acts only on the
celp5
small bowel mucosal ce11. 1.5 In the pathogenesis of cholera, those ingested
microorganisms that survive passage through the gastric acid barrier multi-
ply rapidly in the small intestine and produce an enterotoxin. The molecular
mechanisms by which cholera enterotoxin causes gut fluid loss have been
clearly defined, and the importance of, these observations transcends their
relevance to cholera, in that one of the two E. coli enterotoxins has been
shown to act by the san1e
same molecular pathways.
The same protein enterotoxin, which has a molecular weight of 84,000
daltons, is produced by all toxigenic strains of V. cholerae. This enterotoxin
is composed of two moieties: a binding (B) moiety
n10iety that consists of 5 identi-
cal subunits of approxilnately
approximately 11,500 daltons; and an activating (A) moiety
that is composed of 2 unequal subunits, linked through a single disulfide
bond. The A2 subunit (about 5000 daltons) probably serves only to connect
the Al subunit (about 23,000 daltons) to the binding moiety.lO The amino
acid sequences of the binding subunits have been determined. The terminal
sequences share considerable homology with the binding units of certain
peptide hormones (for example, thyroid-stimulating hormone). Each of the
B subunits of the cholera toxin has the ability to bind to a GM-l
GM-I monosialo-
ganglioside molecule in the small intestinal cell wall. This binding occurs
rapidly and becomes irreversible within minutes after the initial contact of
the binding subunits with the GM-l monosialogangliosides.ll,18
monosialogangliosides.11,18 Following
600 CHAHLES C.
CHARLES J. CAHPEr\TEH
CARPENTER

the rapid binding of the B subunits to the cell wall, the Al subunit, which
contains adenosine diphosphate (ADP)-ribosyltransferase activity, migrates
through the epithelial cell; it then stimulates, by a catalytic process, the
transfer of ADP-ribose from nicotinamide adenine dinucleotide (NAD) to a
guanosine triphosphate (GTP)-binding protein that regulates adenylate cy-
clase activity.7 This ADP ribosylation of the GTP-binding protein inhibits
the GTP "turnoff' reaction and causes a sustained increase in adenylate cy-
clase activity. The resultant increase in intracellular levels of cyclic adeno-
sine 3', 5' -monophosphate (cyclic AMP) leads to rapid secretion of elec-
trolytes into the small bowel lumen.
lumen,l414 The putative mechanism of action of

the cholera enterotoxin is illustrated in Figure 1.

The GM-l
GM -1 monosialoganglioside receptors are present in the walls of
virtually all mammalian cells and occur in high concentration in the gut mu-
cosal cell walls. There are roughly 1 million potential binding sites per gut
mucosal cell, and the occupation of only 50,000 such binding sites by the
cholera enterotoxin appears to cause maximum stimulation of cellular aden-
ylate cyclase and maximum gut fluid secretion.
The rapid cellular binding of the toxin is followed by a lag period that
has been shown, in vitro, to be dependent on temperature,7 the' concentra-
tion of enterotoxin, and the density of GM-lGM -1 ganglioside receptors in the
cell wall. 18
I8 The lag period presumably reflects the time required for the A

subunit to penetrate the lipid bilayer and become reduced to the catalytic
Al peptide. The lag period is followed by a gradual increase in adenylate
cyclase activity and gut fluid secretion. In the mammalian gut at body tem-
perature, the lag period varies from 10 to 30 minutes, and adenylate cyclase
activity and fluid production generally reach maximal levels within three to
four hours after the binding event. Adenylate cyclase activity and fluid pro-
duction then persist at a maximal level for a period of 8 to 12 hours, follow-
ing which fluid production gradually subsides. The intestinal fluid transport
is generally back to normal levels within 36 hours after a single exposure to
toxin, which probably reflects the fact that the cells originally exposed to
toxin have been shed in the normal turnover of gut epithelial cells.
Current evidence indicates that the increased cyclic AMP causes in-
creased sodium-dependent chloride secretion; the increased intracellular
cyclic AMP also prevents sodium and chloride absorption across the brush
border via the sodium chloride cotransport mechanism. 66 Current data sup-
port the concept that the cyclic AMP exerts its direct chloride secretory ac-
tion primarily on crypt cells, and its inhibition of neutral sodium chloride
absorption primarily on villous cells (Fig. 2). The net effect of the increased
intracellular cyclic AMP is to cause secretion of an isotonic fluid throughout
the entire small bowel. The cholera toxin does not cause fluid secretion by
the stomach, and its effects on fluid movement in the human colon are min-
imal,144 The cholera stool, essentially the sum of the fluids produced by each
imal.l
of the small bowel segments, slightly modified by the colon, is an isotonic
fluid with a bicarbonate concentration roughly twice that of normal plasma
and a mean potassium concentration four to eight times that of plasma (see
Table 2).3 The rate of intestinal fluid loss by the adult patient with cholera
may reach 1 liter per hour. The duration of the diarrheal disease, if un-
treated by antimicrobial agents, may exceed five days.
 DIARHHEAS
THE PATHOPHYSIOLOGY OF SECRETORY DIARRHEAS 601
1) Approach 2) Binding

Cholera ~ clIr1
Toxin ~

/ '/ Plasma /
/ / Membrane

Latent Adenylate
Cyclase

3) Conformational Change 4) Dissociation and Entry

5) Penetration and "Activation" 6) Activation of Cyclase

of A Subunit

ATP
---.....cAMP
~

sitnation with enterotoxin

Figure 1. Mechanism of action of cholera enterotoxin. 1: Initial situation
free in intestinal lumen. 2: Multivalent binding of B subunits of enterotoxin to accessible oligo-
saccharide chains of Gml
Gm, monosialoganglioside. 3: Conformational change in enterotoxin struc-
ture that allows hydrophobic regions in A moiety to interact with mucosal cell membrane. 4: Dis-
sociation of enterotoxin components and entry of A moiety into fluid lipid bilayer of n1embrane.
membrane.
5: Penetration of A moiety into membrane and reduction of disulfide bond to generate active Al
peptide. 6: Cleavage of NAD to nicotinamide and ADP-ribose by Al peptide; transfer of ADP-
ribose to adenylate cyclase complex and activation thereof. (From Fishman,
Fishman. P.H.7)

All pathophysiologic effects of cholera can be readily reversed by the

administration of appropriate replacement fluids by either the intravenous
or oral route, as discussed by Black in this symposium.
In light of the increased understanding of the pathogenetic mechanisms
in cholera, a large number of potential antisecretory
anti secretory agents have been eval-
uated. Of these, only three have shown potential therapeutic usefulness,
602 CHARLES C. J. CARPENTER

VILLOUS CELL

C(
\
\
\
\
\
,

LUMEN CRYPT CELL

,,
I

/
~"
<;------
'CAMP +
No

Figure 2. Postulated separate and additive actions of cyclic AMP

A:\fP on ion transport in intestinal
crypt and villous cells (from Field 99 ).

since they can reverse the secretory process after the enterotoxin has been
bound to the intestinal mucosa. These substances are chlorpromazine, ni-
cotinic acid, and berberine. Of these, only chlorpromazine has been studied
in humans, under controlled conditions, as an anti secretory agent. It is
antisecretory
promising in that it rapidly decreases the rate of intestinal fluid loss, but it
has the significant disadvantage of causing drowsiness.I.'j
drowsiness. 15 Neither of the other
two agents has been evaluated in a controlled field trial.

v. cholerae, Non 0 Group 1

Previously called noncholera vibrios or nonagglutinating vibrios (NAGs),
these organisms are now being recognized increasingly as a cause of acute
diarrheal disease. Such vibrios may cause up to 5 per cent of acute diarrheal
illness in areas endemic for cholera, and have occasionally been incrimi-
nated in waterborne microepidemics of diarrheal disease. These organisms
are known to be associated with aquatic environments, but their ecologic
niche has not been studied extensively. A A large number (approximately 50)
of different serotypes have been identified. This group of vibrios is patho-
genetically heterogeneous in that some of these strains produce an entero-
toxin indistinguishable from cholera enterotoxin, while other strains pro-
duce an enterotoxin similar to the heat-stable E. coli enterotoxin.

Enterotoxigenic E. coli
In the past ten years enterotoxigenic E. coli (ETEC) have been recog-
nized as common etiologic agents of diarrheal disease in man. ETEC have
 SECHETOHY DIARRHEAS
THE PATHOPHYSIOLOGY OF SECRETORY DIAHRHEAS 603
603

a worldwide distribution and have been incriminated as the Inost most COlnmon
common
cause of acute diarrhea in small children in many of the developing nations
of the world 1616 They have also clearly been demonstrated to be the most
common etiologic agent in travelers' diarrhea occurring in visitors to the de-
veloping nations. In addition, they are responsible for occasional common
source (food or water) outbreaks of diarrheal disease in developed countries,
and they have occasionally been incriminated in severe cholera-like disease
in adults in the Indian subcontinent. Diarrheal disease caused by ETEC is
far more variable than that caused by cholera because of the multiple path-
ogenetic mechanisms that may be involved. At least two distinct toxins, a
heat-labile toxin (LT) and a heat-stable toxin (ST), may be produced by E.
coli that have colonized the small bowel.
bowel,16
16
Unlike cholera, in which the abil-
ity to produce enterotoxin is chromosomally mediated and a single toxin is
produced by all pathogenic strains, the ability of E. coli to produce entero-
toxins is plasmid mediated, and specific plasmids code for the production of
LT and ST. Genes for either or both of these toxins may be on a single
plasmid. 22
The E. coli LT has a molecular weight of approximately 83,000 daltons
and is structurally similar, but not identical, to the cholera toxin. 44 As in the
case with cholera enterotoxin, the primary binding site for the E. coli LT is
the GM-I monosialoganglioside.
monosialoganglioside,l1l1 Having bound to the gut mucosal cell,
the E. coli LT also stimulates adenylate cyclase through ADP ribosylation,
with kinetics similar to those of cholera toxin, causing delayed secretion of
isotonic fluid that persists for many hours after binding of the toxin. 14 14 Like
cholera enterotoxin, the E. coli LT also has binding and activating moieties.
The binding subunits are antigenically similar, but not identical, to those of
the cholera toxin. The activating, or AI, subunit of the E. coli LT is also sim-
ilar, but not identical, to the Al subunit of the cholera enterotoxin. Although
the LT may be produced by a large number of E. coli serotypes, current data
suggest that an identical LT molecule is produced by all such strains.
The E. coli ST is quite different from LT in that it exhibits a rapid onset
of action, does not bind to ganglioside&
ganglioside~ of the mucosal cell wall, and is of
low molecular weight (less than 2000 daltons). It is not antigenic in its nat-
ural state. The ST appears to act by stimulation of guanylate cyclase with
accumulation in mucosal cells. 66,12
resultant cyclic GMP accunlulation ,12 The kinetics of ST
action are strikingly different froIn
fi'om LT, in that ST causes an almost imme-
diate increase in gut fluid secretion after exposure to the gut mucosa; its se-
cretory effect rapidly disappears when ST is washed out of the gut lumen.
The increased levels of intracellular guanylate cyclase cause chloride secre-
tion by gut mucosal epithelial cells in a manner similar to that seen with the
cholera enterotoxin, but the ST does not not alter neutral sodium chloride ab-
sorption by the brush border transport route. 66 Although the binding proper-
ties have not yet been established, the binding appears to be specific for
animals, the E. coli ST
small bowel mucosal cells. In certain experimental anilnals,
mucosa, while showing its character-
has no detectable effect on duodenal nlucosa,
istic effect on jejunal and ileal mucosa. Unlike the cholera enterotoxin and
E. coli LT, the E. coli ST does not exhibit its characteristic effect on cyclase
activation in nonintestinal tissues. Despite the different kinetics, the major
physiologic effect of ST is also to increase net movement of isotonic fluid
604 CHARLES
CHAHLES C. J. CARPENTER
CAHPENTEH

from blood to gut, resulting in a watery diarrhea that may be qualitatively

indistinguishable from cholera.
An E. coli strain may produce LT, ST, or both enterotoxins, depending
on the plasmids that it carries. 99 The enterotoxin types may vary with the
geographic location and serotype. In certain geographic areas, diarrheal dis-
ease is more commonly produced by E. coli that produce both ST and LT;
in other regions, E. coli that produce ST alone predominate. The plasmids
carrying the genetic material controlling enterotoxin production have been
well characterized. These are for the most part transferable plasmids (such
as R factors) and can be transferred among strains through bacterial conju-
gation. The plasmid DNA sequences controlling enterotoxin production
have now been studied extensively by investigators using recombinant
DNA technology. The LT and ST genes have been cloned; the ST gene has
been shown to be a transposon (a DNA replicon that can insert itself into
the chromosome as well as on different areas of the plasmid). Methods are
being developed using DNA probes for the direct recognition of the gene
sequences in the bacteria as a diagnostic tool; such methodology could re-
place the current techniques for identifying the LT and ST themselves. 22
Furthermore, DNA responsible for the production of only the A or the B
subunits can be placed on separate plasmids, which may be useful in the
eventual production of a vaccine.
Enterotoxigenic E. coli also produce colonization factors that allow
them to adhere to the small bowel mucosa, rather than being washed
intestinal lumen. 55 Some of these factors have been identified as
through the intestinallumen.
fimbria or pili (hairlike projections of protein material extending from the
cell surface) and designated as colonization factor antigens (CFA-I and
CF A-I1). These occur, however, only on certain ETEC, and are related to
CFA-II).
the serotype of the organism; it is suspected that other colonization factors
will be found in relation to the remainder of the ETEC strains. Colonization
factors are also plasmid-mediated, and a single plasmid may carry genes for
the production of both enterotoxin and CFA. CF A.
Although the production of enterotoxin and CFA CF A is'
is mediated by trans-
ferable plasmids, and any E. coli strain could theoretically harbor them, in
nature this does not seem to occur. Certain serotypes of E. coli seem to be
especially adept at carrying these plasmids, although the reasons for this are
not known. Furthermore, it seems that there must be a combination of 0, K,
and H antigens (as well as a certain characteristic fermentation pattern, or
biotype) in order for enterotoxin plasmid to be harbored successfully. Some
serotypes more often produce both LT and ST, such as 06:K15:H16, and
others ST only, such as 0128:H (antigens variable). Likewise, some sero-
types carry CFA-l, such as 078:Hll, and others CFA-I1,
CFA-II, such as 06:K1.5:H16.
06:K15:H16.
Some enterotoxigenic serotypes are found worldwide, whereas others are
distributionY
more localized in distribution. 17 In fact, each geographic area studied ap-

pears to have a characteristic spectrum of typical ETEC serotypes. The

ETEC serotypes, with a few exceptions, are different from the enteropath-
ogenic serotypes of E. coli (EPEC) previously described as the cause of
nursery outbreaks of diarrhea; the EPEC do not produce LT or ST. How-
ever, recent data from adult volunteer challenge experiments and from ani-
mal models suggest that at least some of the epidemic-associated EPEe
THE PATHOPHYSIOLOGY OF SECRETORY DIARRHEAS 605
strains are enterotoxigenic, although the enterotoxins they produce are dif-
ferent from the LT and ST of ETEC strains.
Other Enterotoxigenic Bacteria
the enterotoxigenic diarrheas, from the
Although the most important of the,
standpoint of both severity and frequency, are those caused by V. cholerae
and E. coli enterotoxins, there are three well-recognized exceptions to the
above scheme of diarrheas caused by enterotoxins. 22 In each of these excep-
tions, the culpable enterotoxin exerts its effect by some means other than
stimulation of one of the known secretory mechanisms of the small bowel,
and the resultant fluid losses are seldom voluminous. (1) Enterotoxins may
be ingested directly in food, as with Staphylococcus aureus and Bacillus cer-
eus food poisoning. These organisms produce the enterotoxins while grow-
ing to high concentration in the food rather than the small intestine. The
which the staphylococcal enterotoxin causes diarrhea
precise mechanism by \Vhich
is not known. B. cereus appears to have the capacity to produce several en-
terotoxins, at least one of which stimulates adenylate cyclase in gut mucosal
cells in a manner similar to the cholera enterotoxin and the LT of ETEC. (2)
Clostridium perfringens enterotoxin, which is produced by ingested ba-
cilli in the small bowel, clearly damages gut mucosa and produces a short-
lived diarrhea, in which the volume is characteristically small and the pro-
tein content high. (3) Clostridium difficile heavily colonizes the large bowel,
rather than the small bowel, and elaborates its enterotoxin, which is cyto-
toxic. The enterotoxin in turn causes severe mucosal damage to the colon;
the lesions are similar to those of shigellosis. Clostridium difficile is not, per
se, an invasive microorganism; the resultant diarrhea is not voluminous and
is not truly a secretory diarrhea.

NONINFECTIOUS SECRETORY DIARRHEAS

The known causes of noninfectious 'secretory diarrheas are (1) hormone-

secreting tumors, (2) villous adenomas of the colon, (3) surreptitious inges-
tion of laxatives, and, possibly, (4) malabsorption of bile or fatty acids (Table
3). The mechanisms of fluid loss caused by certain of the hormones secreted
by tumors (vasoactive intestinal peptide, prostaglandins) are similar to
those caused by the cholera and E. coli enterotoxins in that the secretion of
fluid occurs as the result of stimulation of adenylate cyclase in the small in-
testinal mucosal cells, and when the volume of the fluid entering the colon
occurs.88 Other hormones of tumor
exceeds its reabsorptive capacity, diarrhea occurs.
origin (gastrin, histamine) act at an entirely different site, causing hyperse-
cretion of acid by the gastric mucosa. The mechanism is strikingly different
with villous adenomas of the colon, in which fluid is secreted directly into
the lumen of the large intestine without mediation by humoral agents that
have been secreted in another site. Surreptitious laxative ingestion may be
the most common cause of chronic diarrhea in the United States today. The
mechanisms of diarrhea caused by laxatives are multiple, ranging frof!l from the
simple osmotic effect of poorly absorbed cations (for example, milk of mag-
nesia) to actual stimulation of small bowel secretion by phenolphthalein, as
-nesia)
606 CHARLES C. J. CARPE~TER
CARPENTER

Table 3. Noninfectious Secretory Diarrheas

mediated: of neoplastic origin
H umorally Inediated:
Humoral Agent
HUlnoral Source Target Organ
VIP or VIP-like Non-islet cell neoplasn1
neoplasm or ganglioneuroma Small bowel
peptide
Prostaglandins Medullary carcinoma of thyroid Small bowel
Gastrin Non-islet cell neoplasm Stomach
Ston1ach
Histamine hasophils
Mast cells, basophils Stomach
Surreptitious ingestion of laxatives
Direct mucosal hypersecretion
Villous adenomas of the colon
Bile acid- and fatty acid-induced diarrheas*

*Active electrolyte secretion is only one of several mechanisn1s

*Active mechanisms responsible for these ill-
nesses, and they are not generally classified as true secretory cliarrheas.
diarrheas.

contained in several over-the-counter cathartics. Although active secretion

of electrolytes by the colonic mucosa has also been demonstrated in bile~ bile-
acid-induced and fatty-acid-induced diarrheas,l changes in gut mucosal
permeability appear to be more important etiologic factors than active elec-
trolyte secretion in these disease entities. These latter illnesses will, there-
fore, not be specifically addressed in this article.

Humorally Mediated Small Intestinal Secretory Diarrheas

As indicated in Table 3, two distinct pathogenetic mechanisms give rise
to tumor-associated chronic secretory diarrheas, depending on whether the
target organ is the stomach or the small bowel. The diarrhea associated with
certain non-beta islet cell pancreatic tumors, ganglioneuromas, and medul-
lary carcinomas of thyroid results from elaboration by the tumor of the agent
or agents that cause hypersecretion of fluid and electrolytes by the small in-
testine in a manner analogous to the mechanism of action of the cholera en-
terotoxin. The so-called pancreatic cholera syndrome is associated \vith with
non-beta islet cell pancreatic neoplasms; in the profuse watery diarrhea in
this syndrome, as well as in the secretory diarrhea associated with gangli-
oneuroma, the etiologic agent appears to be vasoactive intestinal peptide
(VIP) or a VIP-like substance. VIP, which has 28 amino acid residues, stim-
ulates adenylate cyclase in the intestinal mucosal cells and therefore causes
secretion by a mechanism similar to cholera enterotoxin and E. coli LT. In
the patients with secretory diarrhea secondary to medullary carcinoma of the
thyroid, the humoral agent that causes diarrhea has not been definitely de-
termined; of the active humoral agents secreted by these tumors, both cal-
citonin and prostaglandins cause net secretion of electrolytes from the small
intestine; current data suggest that prostaglandins, which also stimulate
adenylate cyclase in the small bowel mucosal cells, are probably the culpa-
ble agents in the secretory diarrhea associated with this disease.
The pathogenesis of secretory diarrhea by VIP and prostaglandins is
therefore similar to that caused by the cholera and the heat labile E. coli
enterotoxins. There is, however, one striking difference between the kinet-
THE PATHOPHYSIOLOGY OF SECRETORY DIARRHEAS 607
ics of secretion mediated by humoral agents of tumor origin and that me-
diated by cholera and E. coli LT enterotoxins. There is a consistent lag
phase between the binding of the bacterial enterotoxins to the intestinal
mucosal cells and the maximal intestinal secretory response to these enter-
otoxins. Net secretion of isotonic fluid generally occurs within 30 minutes
after cholera or E. coli enterotoxin are bound to intestinal mucosal cells, but
a maximal rate of intestinal secretion does not occur until three or four hours
after enterotoxin attachment to the cells. Secretion then continues at a max-
imal rate for a period of 8 to 12 hours, even if no further exposure of the
mucosal cells to the enterotoxin occurs. In a quite different manner, intes-
tinal secretion occurs almost simultaneously when either prostaglandin or
vasoactive intestinal peptide comes into contact with small intestinal mu-
cosal cells, and secretion ceases abruptly if the cells are washed free of the
culpable humoral agents. By the same token, adenylate cyclase stimulation
and intracellular cyclic AMP accumulation parallel the fluid secretion in
both instances, occurring almost instantaneously after the application of
either prostaglandin or VIP to the intestinal mucosal cell and ceasing ab-
ruptly when the intestinal mucosal cells are washed free of the two humoral
agents.
The fact that the secretory diarrheas, caused by humoral stimulation of
electrolyte secretion by the small bowel, are chronic illnesses indicates that
the very large volumes of intestinal secretion characteristic of the more se-
vere bacterial diarrheas are seldom associated with the noninfectious secre-
tory diarrheas. Although the humorally mediated noninfectious secretory
diarrheas may indeed cause saline depletion, the rate of isotonic fluid loss is
seldom great enough to cause the hypovolemic shock characteristic of se-
vere cholera. Therefore, the pathophysiologic features of the noninfectious
secretory diarrheas are more often related to chronic hypokalemia than to
clinically significant saline depletion.
Humorally Mediated Gastric Secretory Diarrheas.
A second pathogenetic mechanism of secretory diarrhea occurs in pa-
tients with Zollinger-Ellison syndrome (another form of non-beta islet cell
neoplasm), basophilic leukemia, or systemic mastocytosis. s8 In each of these
clinical disorders, the diarrhea results from secretion by the tumor of a hu-
moral substance that stimulates gastric acid secretion.
The gastric hypersecretion causes large volumes of acidic fluid to enter
the upper small intestine. Absorption of fluid by the small intestine appears
to be impaired because of a direct adverse effect of the increased hydrogen
ion concentration on the small intestinal mucosa. The high concentration of
hydrogen ion not only impairs absorption but also interferes with the normal
digestive process. Maldigestion of fat results from the low pH of the duo-
denal contents, which causes irreversible inactivation of intestinal lipase, as
well as precipitation of bile salts. Maldigestion of protein and starch also ap-
pears to occur because the enzymes that degrade these substances are inac-
tivated at the acid pH. The diarrhea associated with gastric hypersecretion
is therefore more complex than that caused by the bacterial enterotoxins,
VIP, and prostaglandins, in that malabsorption of proteins, carbohydrates,
and fats, as well as hypersecretion of electrolytes, contributes to the diar-
608 CHARLES C. J. CARPENTER

rheal losses. In the patients with diarrheas caused by gastric hypersecretion,

rheallosses.
the diarrhea can be abolished by continuous gastric aspiration or by total
gastrectomy, even when the tumor is still present. The specific role of in-
creased hydrogen ion concentration is emphasized by the fact that diarrhea
does not occur if the gastric juice is neutralized and reinstilled into the
stomach of such patients.
The Zollinger-Ellison syndrome, which is characterized by gastric hy-
persecretion and duodenal ulceration, is caused by secretion of gastrin by a
non-beta islet cell pancreatic tumor. This neoplasm is histologically indis-
tinguishable from the pancreatic tumor in patients with the pancreatic chol-
era syndrome, in which hypersecretion of VIP causes small intestinal hy-
persecretion. Patients with systemic mastocytosis or basophilic leukemia
may also have secretory diarrhea; the mechanism of diarrhea in these pa-
tients is indistinguishable from that in patients with Zollinger-Ellison syn-
drome. In patients with systemic mastocytosis or basophilic leukemia, how-
ever, the humoral agent responsible for the gastric hypersecretion is
histamine rather than gastrin. Histamine is present in basophils as well as
mast cells; with neoplasms involving these tissues, increased amounts of
histamine are released from the circulation with a resultant increase in gas-
tric acid secretion.
The cellular basis of gastric hypersecretion caused by histamine is par-
tially explained by activation of adenylate cyclase and increase in intracel-
lular cyclic AMP of gastric parietal cells in response to histamine stimula-
tion. It is not, however, clear that adenylate cyclase stimulation is the only,
or even the most important, factor in histamine-induced gastric hypersecre-
tion. The cellular basis by which gastrin interacts with parietal cells to in-
crease hydrogen ion secretion has not yet been delineated.
In patients in whom the secretory target organ is the stomach (those
with Zollinger-Ellison syndrome, basophilic leukemia, and systemic masto-
cytosis) the stool volume is generally less than in those with noninfectious
secretory diarrheas of small intestinal origin, and clinically significant saline
depletion is even less common. Hypokalemic alkalosis is the major resultant
metabolic disturbance.
The role of the colon in secretory diarrheas caused by humoral agents is
identical to that in diarrheas caused by cholera and E. coli enterotoxins.
Whenever the colon is presented with volumes of isotonic fluids that exceed
its reabsorption capacity, diarrhea occurs because of incomplete absorption
of fluid. As the volumes of diarrheal fluids increase, there is progressive in-
crease in the stool concentrations of sodium and chloride, accompanied by
a progressive decrease in the stool concentration of potassium, reflecting the
fact that the colonic absorptive mechanism for sodium is working at maxiInal
maximal
potasslUlll
capacity and is associated with the maximal movement of potassl UHl into the
colonic fluid. The colonic mechanism for chloride absorption is also working
at maximal capacity, with resultant maximal colonic secretion of bicarbon-
ate. This ultimately causes saline depletion (resulting from incomplete co-
lonic absorption of sodium, chloride and water) and hypokalemia (resulting
from colonic secretion of potassium).
There is a distinct difference between the electrolyte pattern seen in
the patient with a chronic noninfectious secretory diarrhea and the patient
THE PATHOPHYSIOLOGY OF SECRETORY DIARRHEAS 609
acutely ill with cholera. Because of the severity of the acidosis in the patient
with cholera, serum potassium is usually high at the time the patient is first
seen, despite total body potassium depletion. This contrasts with both the
hypokalemic acidosis characteristic of the chronic noninfectious secretory
diarrheas of small bowel origin and the hypokalemic alkalosis that occurs in
secretory diarrheas of gastric origin.

SUMMARY
The mechanisms by which bacterial enterotoxins cause secretory diar-
rheas have been well defined, and the definitions of such mechanisms have
been important in developing a consistently successful therapeutic ap-
proach. The less common secretory diarrheas, caused by the interaction of
hormones of tumor origin with the gut small intestinal mucosa have also
been clearly defined, and their pathogenetic mechanisms are similar to
those by which the cholera and E. coli enterotoxins
enter.otoxins cause secretory diarrhea.
The mechanisms by which histamine and gastrin of tumor origin cause gas-
tric hypersecretion are less clearly delineated; secretory diarrhea caused by
both of these agents can be stopped by total gastrectomy without removal of
the responsible tumor. The secretory diarrhea caused by villous adenomas
of the colon, which does not appear to be related to a distally produced hu-
moral agent, results in the same picture of hypokalemic acidosis that is char-
acteristic of the nonbacterial secretory diarrheas originating in the small in-
testine and is cured by resection of the responsible tumor.

REFERENCES
1. Binder, H. J.: j.: Pathophysiology of bile acid- and fatty hltty acid.,...-induced
acid-induced diarrhea. In Field, M., ~l.,
Fordtran, J. j. S., and Schultz,
Scbultz, S. G.
C. (eds.):
leek): Secretory Diarrhea, Bethesda, Am. Physiol. Soc.,
p. 159, 1980.
2. Carpenter, C. C C.. .T., and Sack, R. B.: Infectious diarrheal syndronles.
syndromes. In Isselhacher,
Isselbacher, K., Adams,
R., Braunwald, E. et al. (eds.): Update 1. 'Harrison'sHarrison's Principles of Internal Medicine, New
York, I\lcCraw-Hill,
McGraw-Hill, 1981, p. 209.
3. Carpenter, C. C. C . .T., Greenough,
Creenough, W. B., and Gordon,
Cordon, R. S.: Pathogenesis and pathophysiology
of cholera. In Barua,Barna, D., and Burrows, W. (eds.): Cholera. Philadelphia, W. B. Saunders Co.,
1974, p. 129.
Clen1ents, J.
4. Clements, j. D., and Finkelstein, R. A.: Isolation and characterization of hon10genous homogenous heat-
labite-
labile- enterotoxins with high specific activity fron1 from Escherichia coli cultures. Infect. Im-
mun., 24:760, 1979.
5. Evans, D. G., C., Evans, D. D . J.,
.T., and DuPont, H. L.: Virulence factors
f'lCtors of enterotoxigenic Escher-
ichia coli. J. j. Infect. Dis., 136:S118, 1977.HJ77.
6. Field, M.: Regulation of sn1all calcium. In
small intestinal ion transport by cyclic nucleotides and calciun1.
Field, Fordtran, and Schultz: op. cit., p. 21.
Fishman, P. H.: Mechanism of action of cholera toxin: events on the cell surface. In Field,
7. Fishn1an,
Fordtran, and Schultz: op. cit., p. 85. 8,5.
Gardner, J.
8. Cardner, .T. D.: Pathogenesis of secretory diarrhea. In Field, Fordtran, and Schultz: op. cit.,
p.153.
p. 153.
9. Gyles,
Cyles, C. S. M., and Falkow, S.: The enterotoxin plasn1ids plasmids of Escherichia coli. J. j. Infect. Dis.,
130:40, 1974.
Holmgren, J.,
10. Holn1gren, j., and Lonroth, 1.: Oligon1eric
Oligomeric structure of cholera toxin: characteristics of Hand
subllnits. J. Gen.
L subunits. Cen. Microbiol.,
~Hcrohiol., 86:49, 1975.
11. Holmgren, J.: Comparison of the tissue receptors of Vibrio cholerae and Escherichia coli en-
j.: Con1parison
terotoxins by hy means of gangliosides and natural cholera toxoid. Infect. In1nlun., Immun., 8:851, 1973.
12. Hughes, J. j. M., Murad,
~1urad, F., and Cuerrant,
Guerrant, R. L.:
L,: Role of cyclic GMP
C~lP in the action of heat-stable
enterotoxin of Escherichia coli. Nature (London), 271 :755, :7,55, 1978.
610 CHARLES C. J. CARPEKTER
CARPENTER

13. Moss, J., and Vaughan, M.:

~1oss. ]., ~1echanism of action of adenylate cyclase by choleragen and E.
~.: Mechanisn1
coli heat-labile enterotoxin. Field, Fordtran, and Schultz: op cit., p. 107.
14. Pierce, N. F., Greenough, W. ::md Carpenter, C. C. ].:
\V. B., and J.: Vibrio cholerae enterotoxin and its
Bacterio!. Rev., 35:
mode of action. Bacteriol. 3.5: 1, 1971.
J., Greenough, W. B., Holmgren, ].,
15. Rabbini, G. ]., J., et al.:
a!.: Chlorpromazine reduces fluid-loss in
cholera. Lancet, 1:410,
1 :410, 1979.
16. Sack, R. B.: Human diarrheal disease caused by enterotoxigenic Escherichia coli. Ann. Rev.
Microbio!., 29:333, 1975.
Microbiol.,
17. Sack, R. B.: Enterotoxigenic Escherichia coli: identification and characterization. ]. J. Infect.
Dis., 142:990, 1980.
18. van Heyningen, S.: Cholera toxin. BioI.Bio!. Rev., 52:509,
.52:509, 1977.

University Hospitals of Cleveland

2065
206,5 Adelbert Road
Cleveland, Ohio 44106