= History and Epidemiology of Preeclampsia-Eclampsia Leon G. Chesley, PhD State University of New York Downstate Medical Center Brooklyn, New York the hypertensive disorders, chiefly eclamp- sia, are the leading cause of maternal deaths in the United States,’ England and Wales,? and many other parts of the world. Moreover, they predispose to potentially lethal complications such as abruptio pla- centae, acute renal failure, cerebral hemor- rhage, disseminated intravascular coagula- tion, and circulatory collapse. They also weigh heavily in fetal and neonatal morbid- ity and mortality. Much of the neonatal mortality is attributable to premature de- livery, either spontaneous or induced in the mother’s interest. The combined incidence and prevalence of the various hypertensive disorders in pregnancy is commonly said to be about 6-8%, with wide variations, depending chiefly upon the closeness and recording of observations, who makes the diagnosis, and the proportion of primigravidas in the population. History Eclampsia was not differentiated from epi- lepsy until 1789, when de Sauvages wrote that epilepsy was chronic, with recurrences ‘Correspondence: Leon C. Chesley, PhD, 450 Clark- son Ave., Box 24, Brooklyn, NY 11203. of convulsions through the years; all con- vulsions of acute causation he called “eclampsia.” Twenty years later he defined several species of the genus Eclampsia, in relation to various acute causes that Hip- pocrates had described, such as marked hemorrhage, severe pain, vermicular infes- tation, and the like. De Sauvages added Eclampsia parturientium, which he said had been described by Mauriceau. Eclampsia is dramatic, and its convul- sions terrify the beholder. Yet it is seldom mentioned in the classical medical writings of old, probably because midwives had hada monopoly on obstetrics for countless centu- ries and tradition excluded men from the scene. Ancient Egyptian, Chinese, and Indian writings have been alleged to have men- tioned eclampsia, but perusal of the sources cited is unconvincing. They are discussed in a much more extensive review published elsewhere.’ The pre-Hippocratic Coan Prognosis alluded to eclampsia: “In pregnancy, drows- iness and headache accompanied by heavi- ness and convulsions, is generally bad” (XXXI, 507). The ancient Greeks also recognized preeclampsia: “In pregnancy, the onset of drowsy headaches with heavi- Clinical Obstetrics and Gynecology, Vol. 27, No. 4, December 1984 801Sv CHESLEY ness is bad; such cases are perhaps liable to some sort of fits at the same time” (XXXI, 523). There seems to be no mention of eclamp- sia in any of the authenticated writings of Hippocrates, and he did not mention pregnancy in his book On the Sacred Disease (epilepsy). In section V of his aphorisms, number 80, he did write: “It proves fatal toa woman in the state of pregnancy if she be seized with any of the acute diseases.” Galen, in the second century A.D., commented on the aphorism and wrote that epilepsy, apoplexy, and tetanus are especially lethal. His specification of convulsive disorders suggests that he may have had in mind what, was to be called eclampsia 1600 years later. The standard textbook of midwifery in Europe and England for nearly two centu- ries was Résslin’s Der Swangern Frawen und Hebammen Rosengarten, first pub- lished in 1513. The Byrth of Mankinde was an English translation of what was probably a Latin version of the second edition of the original. It first appeared in 1540 and went through many editions, with little change after the second, Résslin wrote that loss of consciousness and convulsions were omi- nous signs in women with difficult labor and fetal death. He had paraphrased or perhaps merely translated what Paul of Aegina had written in the seventh century AD. Gabelchoverus, in 1596, specified four kinds of epilepsy: they arose in the head, the stomach, in chilled extremities, or in the Pregnant uterus, respectively. In_ uterine epilepsy, the mother feels as if a rat were gnawing at her heart; that sounds like epigastric pain, the description of which is usually attributed to Chaussier, 228 years later. The literature of eclampsia really begins with the advent of the male midwife, when physicians in France invaded the field of obstetrics in the seventeenth century. Mauri- ceau was the most eminent of the new breed. His only mention of convulsions in the first edition of his book, published in 1668, was of those associated with severe obstetric hemorrhage, of which his sister had died. In the succeeding editions he wrote more and more about what we now call eclampsia, and in 1694 he set out several aphorisms. Among them were the following: number 228, The mortal danger to mother and fetus is greater when the mother does not recover consciousness between convulsions; num- ber 229, Primigravidas are at far greater risk of convulsions than are multiparas; num- ber 230, Convulsions during pregnancy are more dangerous than those beginning after delivery; number 231, Convulsions are more dangerous when the fetus is dead than when it is alive. Mauriceau did not not unequivocally differentiate eclampsia from epilepsy, but he may have recognized it as an entity, for he attributed it to an excess of hot blood flowing from the uterus to stimulate the nervous system, with aggravation by cervi- cal irritation. He also wrote that when the fetus dies, malignant vapors arising from its decomposition might cause convulsions. Edema Demanet, in 1797, wrote that all six of his eclamptic patients had anasarca and sug- gested that edema be added to the three recognized causes of convulsions: depletion, repletion, and the pains of labor. Proteinuria In 1840, Rayer observed proteinuria in three edematous pregnant women. From his descriptions, it is likely that at least the first one had preeclampsia. The discovery of proteinuria in eclampsia was made inde- pendently by Lever and Simpson in 1843. Lever was struck by the resemblance be- tween his eclamptic patients and those with nephritis on the service of his colleague, Richard Bright, which led him to look for proteinuria, He found it in 9 of 10 convul- sive women, and the diagnosis of eclampsia was questionable in the one without pro-i HISTORY OF PREECLAMPSIA teinuria. Lever found no proteinuria in “upwards of 50” women in labor unless “symptoms have presented themselves, which are readily recognized as precursors of puerperal fits.”” Lever observed that the proteinuria of eclampsia and preeclampsia abated and disappeared after delivery, and he therefore concluded that eclampsia was not nephritis. Others were not so astute, and they seized upon the proteinuria as further evidence for the popular view that eclampsia was a manifestation of chronic nephritis. Unfor- tunately, one of Simpson's proteinuric eclamptic patients came to autopsy and was found to have granular kidneys, which misled him. In 1851, Frerichs published an influential book on nephritis in which he wrote that eclampsia is a form of uremia, an opinion that held sway for half a century thereafter. Frerichs had suggested that an enzyme converted urea to toxic ammonium carbon- ate; others identified the toxin asa precursor of urea, carbamic acid. Among other sub- stances postulated were xanthine, creatine, creatinine, acetone, bilirubin, lactic acid, globulins, leucomaines, and a host of others, including noxious material normally shed during menstruation, but retained because of amenorrhea, Even water was thought by some to be the toxin. Pathologists often failed to find renal abnormalities in eclamptic women coming to autopsy. Spiegelberg countered that by postulating intense renal vascular spasm, elicited by a reflex arising from the distended uterus, and he wrote, in italics: “True eclampsia depends upon uremic poisoning in consequence of deficient renal excretion.” Hypertension The hard bounding pulse of eclamptic women had suggested arterial hypertension to the old-time clinicians. Sphygmographic tracings made in the late 19th century were interpreted as showing high blood pressure in eclampsia, but Mahomed, for instance, interpreted the tracings as showing hyper- tension in nearly all pregnant women. Ballantyne inferred vascular collapse from sphygmographic tracings made in a dying eclamptic woman. Vinay, in 1894, used a primitive sphygmomanometer and found blood pressures ranging from 180 to 200 mmHg in pregnant proteinuric women; pressures of up to 160 mmHg were normal as. estimated by his instrument. The discovery of eclamptic hypertension is generally cred- ited to Vaquez and Nobécourt in 1897, but they remarked that they had confirmed Vinay's observation. Vinay thought that his patients had nephritis, but the equation of preeclampsia-eclampsia with nephritis was common in those days; actually, the term “nephritic toxemia” persisted until about 1940. Primary, or essential, hypertension was not recognized until 1896, when Allbutt observed that middle-aged and older per- sons, especially women, often develop hy- pertension with no other evidence of renal disease. He called the disorder “senile plethora,” an appellation that had an unfortunate and lingering effect because obstetricians thought that their patients were not old enough to have it. W. W. Herrick was the medical consultant in the Sloane Hospital for Women, in New York City, and between 1926 and 1936 he and his co-workers established that essential hypertensioniscommon in pregnant women and that it is the usual form of hypertension in women found with high blood pressure following hypertensive disorders in preg- nancy. Most obstetricians, led by Henricus J. Stander, opposed his views and clung to the belief that hypertension in a young woman must be either preeclampsia-eclampsia or nephritis. In one of his‘last papers, Herrick, with Tillman, wrote: “When these are fully delineated it is our opinion that we shall find nephritis concerned in but a small fraction of the toxemias; that the larger number, including the eclampsias, the pre- eclampsias, and the variously designatedCHESLEY milder types of late toxemia . . . will be found to have unit characteristics based upon cardiovascular diseases with hyperten- sion.” Herrick had been goaded into going too far by the inclusion of eclampsia and “true” preeclampsia, but he was correct for most of the other forms. Fishberg, in the fourth edition of his monumental Hypertension and Nephritis, denied the specificity of preeclampsia- eclampsia and regarded itas a manifestation of essential hypertension, brought to light and peculiarly colored by pregnancy. That view was held by several of the leading internists of the 1930s. Ifeclampsia were a manifestation of latent essential hypertension, most or all survivors of eclampsia eventually should develop chronic hypertension if they live long enough. That is not the case. There are only two large, long-term follow-up studies of eclamptic women, and they agree that the remote prevalence of hypertension follow- ing eclampsia in the first pregnancy is‘ not increased over that in unselected women matched for age and race. The one follow-up study of women with preeclampsia, proved by renal biopsies, came to the same conclu- sion.’ Preeclampsia-eclampsia and essen- tial hypertension are unrelated. Women with preeclampsia or eclampsia are neither more nor less likely than other women to develop chronic hypertension. Conversely, women who eventually develop essential hypertension are neither more nor less likely than other women to have preeclampsia- eclampsia. Women having eclampsia as multiparas are not included in the generali- zations, for many of them have antecedent chronic hypertension that predisposes them to eclampsia in the first place. Mortality in Relation to Management Although phlebotomy and purgation were the sheet anchors of physicians’ treatment of, eclampsia in the seventeenth, eighteenth, and nineteenth centuries, their beginnings are obscure. It seems unlikely that they were used by midwives, and they probably were introduced when physicians entered the field of obstetrics. Mauriceau, in 1694, recommended phlebotomy and indicated that his colleagues used it. Hippocrates wrote: “Convulsions take place from either repletion or depletion” (section VI, Number 29). He referred to convulsions generally, and physicians were divided as to which accounted for convul- sions in childbirth. Phlebotomy and purga- tion probably were of late origin, even for physicians, if they were ever called for an obstetric case, because Hippocrates had written that they were contraindicated in pregnant women. Merriman, in 1820, wrote: “Dr. Hunter, Dr. Lowden, and other teachers of mid- wifery, used to state in their lectures, that more than half of the women died, who were attacked with convulsions in their labours. Dr. Parr, in his Medical Dictionary (1809— L.C.C.) states an even larger amount of fatal cases; he says that ‘six or seven in ten elude the most active and best concerted measures’ and Jacobs, in his Ecole Pratique des Accouchments (1785—L.C.C.) says that it is always fatal, scarce an instance of recovery is known.* In modern practice the proportion of deaths is by no means so great.” Merriman had lost 8 of 36 mothers and 25 of their 38 infants, mortality rates of 23% and 66%, respectively. Review of the literature published be- tween 1837 and 1867, which I tabulated in 1978,? indicated an average maternal mortal- ity of about 30% in eclampsia, with a range of from 27% to 37%. In that period the predominant treatment might be character- ized as “purge, puke, plaster, and phle- botomy”—all were aimed at depletion. In the latter half of the nineteenth century heavy sedation with narcotic and anesthetic drugs was added to the earlier measures. Tincture of Veratrum viride was used in some clinics, and some practiced forced ‘French obstetrics seems to have retrogressed; Mauri- ceau, a century earlier, lost 21 of 45 mothers, for a mor- tality rate of 478.HISTORY OF PREECLAMPSIA delivery, although most did not. Theadvent of antisepsis and asepsis reduced the num- bers of eclamptic women dying of infection in the later years. Data from large clinics and surveys of the literature show an average maternal mortality of about 24% The highest mortality was in a clinic where two- thirds of the patients were delivered opera- tively. Biittner conducted an epidemiologic study in the Grand Duchy of Mecklenburg- Schwerin and reported that from 1881 through 1891 the maternal mortality in eclampsia was 28.3% in the cities and 45.8% in the rural areas. From 1892 through 1898, the rates were 16.8% and 27.4%, respectively. Two widely divergent approaches to the management of eclampsia were developed near the end of the nineteenth century. Halbertsma, in the Netherlands, advocated prompt abdominal cesarean section; and Diihrssen, in Germany, practiced vaginal cesarean section by using the cervical incisions that bear his name. Despite the eponym, Van Swieten had used similar incisions for expediting delivery of eclamp- tic women a century and a half earlier; and exactly a century before Duhrssen’s paper, Lauverjat had published a book describing what he called “vaginal cesarean section” for the same purpose. Tweedy, in Dublin, and Stroganoff, in Russia, adopted con- servative, expectant courses and treated their patients medically. In Germany, especially, the radical ap- proach predominated, and the eclamptic woman was on the operating table within 10 minutes of her admittance to the hospital. In those days many of the women did not go to a hospital for many hours, or even days after the onset of convulsions, and their acidosis and dehydration made them poor operative risks. German obstetricians came to differenti- ate Schnellentbindung (quick delivery) from Frihentbindung (early delivery). Quick delivery meant delivery as soon as the patient was seen, regardless of the duration of eclamptic convulsions and coma, and early delivery meant delivery soon after the first convulsion. Needless to say, quick delivery carried a high rate of maternal mortality, but early delivery gave astonish- ingly good results. Freund, for instance, reported no deaths in 47 women delivered within an hour of the first convulsion, but 11 deaths in 88 (12.5%) who were delivered at from 1 to 4 hours. The rate of maternal mortality increased with prolongation of the interval from convulsion to delivery. ‘The disadvantage of early delivery was that it could not be practiced in the majority of eclamptic women because of their late hospitalization. Even Freund was able to effect early delivery in only 34% of his patients, Nevertheless, most of the German clinics continued with quick delivery, pre- sumably with gratitude for early cases. The keystone of the conservative manage- ment was sedation, which had been used less systematically for nearly half a century. Tweedy injected 82 mg of morphine sulfate and 0.65 mg of atropine subcutaneously, lavaged the stomach to remove toxins, and gave a strong purgative agent. He then irrigated the colon for up toan hour, to wash out toxins, and applied a hot linseed poul- tice to the loins, renewed every 2 hours. If the patient was unconscious, he injected a pint of a 0.4% or 0.8% solution of sodium bicar- bonate under the breasts, which might be repeated along the high colonic lavage after 8hours. Morphine sulfate, 16 mg, wasgiven when convulsions recurred; and up to 140 mg might be given within 24 hours if the respiratory rate did not fall below 6 per minute. Tweedy originally performed two phle- botomies about 2 hours apart, the first to reduce the circulating toxins and the second to remove more that had been mobilized from the tissues during the interim. The patient was kept on her right side and was starved. When labor occurred, forceps were applied as soon as they could be pushed through the cervix, or a breech extraction was performed. Stroganoff made all examinations, injec- tions, catheterizations, rectal instillations,CHESLEY and the like with the patient under light anesthesia with chloroform. The patient was kept under constant observation in a quiet, dark room; and all sensory stimuli were minimized. She lay on her right side, and her position was changed every 4 hours. Treatment began with 16 mg of morphine hydrochloride injected subcutaneously with the patient under light anesthesia. An hour later, 2 g of chloral hydrate in warm milk was given by mouth or by rectal instillation. At 8 hours, the dose of morphine was repeated, and chloral hydrate was given at 7, 18, and 21 hours, with a reduction in dose to 1.5 g for the last two times. The treatment was continued through labor, delivery, and for 24 hours thereafter. Larger doses were given to big women with severe eclampsia; and, if they were not anemic, 400 ml of blood might be withdrawn. Oxygen was admin- istered after each convulsion, and digitalis was given if the pulse was weak and rapid. When the cervix was dilated about 6 cm, the membranes were ruptured artificially, as they were in severe intrapartum eclampsia without regard for the degree of dilatation; if necessary in such cases, Hegar’s dilators were used to gain access to the membranes. Labor was often shortened by a conservative use of forceps. Tweedy and especially Stroganoff re- ported marked reductions in maternal mor- tality, as compared with the overall results of radical management, and they won fol- lowers slowly. It was not until the 1920s that, two publications convinced most of the obstetric world that the conservative man- agement was better. Eden, in 1922, analyzed the hundreds of cases of eclampsia collected by the British Congress of Obstetrics and Gynaecology; his findings are shown in Table 1. The overall experience of many clinics had been that the lowest rate of mortality was associated with simple vaginal delivery in both mild and severe eclampsia. Cesarean section or forced delivery increased the mortality greatly. Plass, in 1927, collected from the literature more than 10,000 cases of eclampsia reported TABLE 1. Maternal Mortality in Eclampsia, Related to Severity and Method of Delivery Maternal Mortality (%) Method of ‘Mild Severe Delivery Eclampsia — Eclampsia Natural or assisted delivery or induction of labor 5 34 Cesarean Section n 46 Forced Delivery 18 63 From Eden JW. Eclampsia: A commentary on the reports presented to the British Congress of Obstetrics and Gynaecology, June 29, 1922. J Obstet Gynaecol Br Emp 1922;29:586. between 1911 and 1926. The maternal mortality in 5976 cases managed conserva- tively had been 11.1%, but it was nearly doubled (21.7%) in the 4607 delivered operatively. The pendulum swung to ultraconserva- tism as obstetricians treated the eclampsia and ignored the pregnancy. For a time, they awaited the spontaneous onset of labor, and 50 years ago many pregnancies were carried for weeks after the onset of convulsions, with a soaring rate of stillbirths. I remember one patient who went on for 12 weeks with severe hypertension, gross proteinuria, and ana- sarca, She had a second set of convulsions, abruptio placentae, and a stillbirth. Typi- cally, the patient’s condition appears to improve once convulsions cease, but the improvement is transient, and in the 1980s significant numbers had recurrence of con- vulsions, sometimes with a fatal outcome. As a result, the usual practice came to be termination of the pregnancy at some arbitrary time after the control of convul- sions. That interval has been shortened progressively from several days to a few hours in many modern clinics. The introduction of parenteral magne- sium sulfate was epochal in the manage- ment of preeclampsia-eclampsia. Intrathe- cal injections had been reported to control the convulsions of tetanus, which prompted Horn to uy it in eclampsia, in 1906. Sporadically, others reported the use of intrathecal, intramuscular, subcutaneous,OOOO —————————————————— HISTORY OF PREECLAMPSIA and intravenous magnesium sulfate, but its parenteral use was popularized by Lazard, in 1925, and Dorsett, in 1926. Lazard injected 20 ml of 10% solution intravenously at once and repeated the dose every hour until convulsions ceased; but there is a question as to what limit, if any, he put on the number of injections. In his several publications he wrote that the attending physician should be called if three doses did not stop the convulsions, but he did not indicate what the physician said about further doses. McNeile, in the same hospital, wrote that “an ignorant interne” had given a patient six doses (12 g) in 10 hours, without any untoward effect. Mc- Neile reported that the maternal mortality in 259 cases of eclampsia treated by Lazard’s method was 12.0%, as compared with 36.2% in the 91 immediately preceding cases. Ware and Noblin, whose report was updated by Rucker, reduced their maternal mortality in eclampsia from 25.5% to 5.6% when they changed from a modified Stroganoff regi- men to the use of Lazard’s method. Dorsett used intramuscular magnesium sulfate, injecting an initial dose of 15 ml of 25% solution (3.75 g) and total doses of up to 25 g in 20 hours. The only other treatment was purgation by oral magnesium sulfate. Two of 38 eclamptic women died, one of a cerebral hemorrhage 10 days postpartum. Pritchard and Chesley and Tepper ascer- tained that magnesium injected intrave- nously as a bolus leaves the bloodstream within 15 minutes and that the absorption from intramuscular depots is so slow that 90-120 minutes are required for maximal levels of magnesium to be attained in plasma, They, therefore, combined an intra- venous dose (Pritchard used 4 g; Chesley and Tepper, 3 g) with an initial intramuscular dose of 10 g. The plasma level rose immediately to the desired level and was maintained for about 4 hours; hence, they gave 5 g intramuscularly every 4 hours. The renal excretion of magnesium is balanced by the absorption from the instramuscular depot, and fairly constant plasma levels can be maintained for as long as the treatments continued. Pritchard’ reported 154 consecutive cases of eclampsia so treated without a maternal death; the series was extended to 179 before an accident befell the 180th. Moreover, the perinatal mortality was 15.4%, corrected to 9.9% by the conventional exclusion of those weighing less than 1000 g at birth. The maternal and perinatal salvages are un- matched in the history of eclampsia. Mag- nesium sulfate and delivery as soon as the patient was oriented was the sole treatment except in women whose diastolic pressures exceeded 110 mmHg; they were given hydralazine, Rational Treatment of Eclampsia Once physicians decided that eclamptic convulsions were caused by repletion, the depleting measures of phlebotomy, purga- tion, and the induction of diaphoresis and diuresis seemed rational. When eclampsia came to be regarded as a toxemia, those means seemed to be logical ways in which to remove toxins, uremic or other. Meat, especially red meat, had had a bad name for centuries; and when proteins were identified and associated with meat, they shared the opprobrium. Toxic “split prod- ucts” of protein were postulated as the cause of eclampsia; other variations were that amino acids were decarboxylated but not deaminated and that the amines were the toxic agents, or that autointoxication was the cause. Accordingly, low protein diets were prescribed in both prophylaxis and treatment of preeclampsia and were advo- cated as late as 1946, in the ninth edition of Williams Obstetrics. Measures designed to remove the toxic products of protein in- cluded the induction of vomiting, gastric lavage, high colonic irrigation, purgation, sweating, and diuresis. ‘The bulky uterus was thought to com- press the kidneys, renal veins, or ureters; and logical treatments included keeping the eclamptic woman prone with her belly in a hole in the mattress, ventral suspension ofCHESLEY the uterus, ureteral catheterization, trans- plantation of the ureters to the gut, and renal decapsulation A Scandinavian physician found large lutein cysts in a woman with hydatidiform mole and severe preeclampsia, which prompted him to treat eclampsia by oopho- rectomy. Bovine parturient paresis was mis- taken for eclampsia in cows; and some obstetricians adopted the veterinarians’ treatment in their patients with eclampsia, such asinjections of potassium chloride into the breasts. An efficacious treatment of parturient paresis was inflation of the udder with air. Selheim tried it, found the breasts resistant to inflation, and lopped them off. His prestige was such that bilateral mastec- tomy was tried in several clinics, but the results were unsatisfactory, both medically and cosmetically. Some thought that increased intracranial pressure caused eclamptic convulsions; and they practiced drainage of spinal fluid, cisternal puncture, or even craniotomy. Early in the present century, postpartum psychosis wasattributed to the absorption of toxic products from residual trophoblastic and decidual tissues. It occurred to Latzko that here might be the source of the eclamptic toxin, and he performed postpar- tum curettage. There were reports from several countries between 1909 and 1913, but it fell out of use until revived briefly in 1961. The drugs used in thetreatment of eclamp- sia would include nearly all in the older pharmacopoeias, including strychnine. Even today, many physicians are impelled to try any new drug that controls or modifies any sign or symptom of preeclampsia- eclampsia. In time, many of our present practices will seem as bizarre as those just rehearsed, with the difference being that the former were rational in the light of hypotheses as to the cause and nature of eclampsia; whereas ours are empiric, too often symptomatic, and in some respects, based upon imitative magic. Prophylaxis It does not seem likely that preeclampsia can be prevented on the basis of present know!- edge. A major purpose of prenatal care is to detect incipient preeclampsia and to prevent its progression, which usually, but not always, can be done. Early detection of preeclampsia depends upon frequent obser- vation, looking for rapid gains in weight, increases in blood pressure, and proteinuria. The first prophylactic measure directed specifically against eclampsia that I have found was advocated by Mauriceau, in 1694. He recommended two or three phlebotomies during the course of pregnancy. The ancient Chinese imposed dietary restrictions upon pregnant women, prob- ably as a hygienic measure rather than as a preventive of eclampsia, which was not mentioned. Hou translated Fu Ssu Hui's admonition as: “Meats from animals such as rabbit, goat, sparrow, donkey, turtle and so on are not allowed.” Interestingly, Hip- pocrates, in On the Sacred Disease, wrote that charlatans treating epilepsy forbade meat, especially that of goats, fowl, and turtle. As previously mentioned, low protein diets were advocated as late as 1946, with emphasis on the restriction of meat. Miquel, in 1824, recommended a diet of cereals in the form of a slop or, better, one of milk and cheese; spices should be avoided. A of milk and its products was popular for nearly acentury thereafter, although Tweedy regarded milk as harmful (protein!) and starved his eclamptic and severely pre- eclamptic patients. Johns, in 1843, advocated prenatal exam- inations at intervals during late pregnancy. He described some of the signs and symp- toms of what we now call preeclampsia, such as edema of the hands and face, giddiness, visual disturbances, headaches, ringing in the ears, and pain in the “stomach.” He recommended as preventive measures a diet of fruit, vegetables, andi HISTORY OF PREECLAMPSIA milk; laxative or purgative agents; fresh air and moderate exercises; phlebotomy; and, if the signs and symptoms were marked, an emetic agent. Lever, who discovered the proteinuria of eclampsia in 1843, recommended periodic testing of the urine for protein during late pregnancy. Sinclair and Johnson, in 1858, wrote that women were admitted to the Dublin Lying-in Hospital by prearrange- ment and that they were examined when the arrangement was made. If the patient had edema, headaches, dizziness, or proteinuria, she was either admitted at once or seen frequently as an outpatient. She was told to stay in bed, wasallowed only the mildest and lightest foods, and purged repeatedly. The authors wrote that they often had prevented eclampsia and had decreased its severity when it did occur. Cook and Briggs, in 1903, observed that the detection of hypertension in a pregnant woman should “excite the apprehension of eclampsia,” and that the blood pressure was more reliable than proteinuria asa guide to the likelihood of convulsions. Zangemeister, 1916, initiated the periodic weighing of pregnant women and equated rapid gains in weight with the abnormal retention of fluids that sometimes precedes other clinical signs of preeclampsia. Rapid gains in weight are not specific, however, and they do not always occur. If they do occur, they may precede, coincide with, or follow increases in blood pressure. “Preeclampsia” Often Is Not Preeclampsia Before discussing the epidemiology of pre- eclampsia-cclampsia, we must consider its diagnosis, which is so often erroneous that we might confuse fact with fiction. Clumsy diction in the revised classification of the hypertensive disorders in pregnancy set out by the American Committee on Maternal Welfare in 1952 permitted the diagnosis of preeclampsia on the basis of the appearance of any one sign after the 24th week of gestation: hypertension, proteinuria, edema, or even a gain in weight of 5 pounds or more within a week. Nelson's classification, which is widely used in the British Com- monwealth and elsewhere, defines mild preeclampsia as a rise in diastolic blood pressure to 90 mmHg or higher after the 26th week, observed on two or more occasions a day apart, or a progressive rise in pressure during labor. Edema is ignored, and pro- teinuria is either absent or does not exceed 0.25 g/l. Thus, a large proportion and perhaps most of the diagnoses of mild preeclampsia have been made on the sole basis of rises in blood pressure, without proteinuria or edema. The diagnosis often is erroneous and usually is so in multiparas. Gestational Hypertension The Committee on Terminology of the American College of Obstetricians and Gynecologists sought to improve the accu- racy of the diagnosis of preeclampsia by relegating mere rises in blood pressure to a new category called “gestational hyperten- sion” (an inappropriate term; surely, pre- eclamptic hypertension literally is gesta- tional hypertension), “Gestational hyper- tension” is defined as the apparently acute onset of hypertension in pregnancy, labor, or the early puerperium, without protein- uria or abnormal edema; the blood pressure returns to normal within 10 days after delivery. It corresponds almost exactly with what once was called “low reserve kidney.” There is abundant evidence that such hypertension has a high rate of recurrence in later pregnancies and is the usual basis for the diagnosis of preeclampsia in multiparas. Also, gestational hypertension foretells a high likelihood of eventual essential hyper- tension. Berman reexamined 225 women who had had pregnancies following hypertensiveCHESLEY disorders in earlier ones. Among those who had had no proteinuria during pregnancy the prevalence of hypertension at follow-up was far greater than that in those who had had proteinuria, and 80% had had recur- rence of hypertension in the later pregnan- cies. Herrick and Tillman followed 188 women with “low reserve kidney” for an average of 7 years. Half had hypertension, a third had retinal angiosclerosis, and 88% of those pregnant again had had recurrent hypertension. Significantly, 55% were not yet 40 at follow-up, which means that many more would become hypertensive as they aged. Gestational hypertension sometimes is early, mild preeclampsia, but more often it seems to be latent essential hypertension brought to light by pregnancy. Sometimes it is nothing more than the return of the blood pressure to hypertensive levels that ante- dated conception. Most, if not all, women have decreases in the diastolic pressure during the first and second trimesters, with rises in the third; the changes are often more marked in women with chronic hyperten- sion. There are four papers that among them describe the course of more than 1000 pregnancies in women with known chronic hypertension. Blood pressures decreased significantly in between 30% and 40% of them, often falling to normotensive levels. The pattern is variable, but commonly the pressure falls in the first trimester and rises again early in the third, In the absence of a reliable history, the rise can be mistaken for an acute onset of hypertension, and an erroneous diagnosis of preeclampsia or gestational hypertension is made. Preeclampsia For the diagnosis of preeclampsia, The Commitiee on Terminology requires acute hypertension arising after the 20th week of gestation, together with abnormal edema or proteinuria, or both. Given the high preva- lence of facial and digital edema in normal pregnancies, it seems likely that it is common in nonpreeclamptic women with acute rises in blood pressure, but the presence of such edema permits the diag- nosis of preeclampsia in them. “Abnormal edema’ is defined as edema of the hands, face, or generalized, although such edema is common and physiologic in normal pregnancies. Dexter and Weiss found edema of the hands, face, or both in 64 of 100 consecutive women examined in late normal pregnancies, Robertson, in a pro- spective and objective study of 83 women from early pregnancy, observed occult edema in 17% and frank edema in the remaining 83%, Moreover, he could find no correlation between the development of hypertension in late pregnancy and the presence or absence of edema, or the degree, or the distribution of the edema. The reason for ignoring edema in Nelson’s classifica tion is that it is so common in normal pregnancy. It is more frequent and often more pronounced in preeclampsia than in normal pregnancy, but differentiation of the two kinds of edema, if there are two, defies us. The characteristic renal lesions of pre- eclampsia are almost never found in the absence of proteinuria, and the diagnosis of preeclampsia must always be questionable if the patient does not have proteinuria. The reason for the Committee's not requiring it for the diagnosis of preeclampsia is that proteinuria usually is a late sign in the course of preeclampsia, so late that from 5% to 10% of eclamptic women have the onset of convulsions before its appearance. It fol- lows, therefore, that not every woman with “true” preeclampsia has the typical renal lesions, and that the lesions are secondary. Preeclampsia and eclampsia may occur before the 20th week of gestation in associa- tion with a large, rapidly growing hydatidi- form mole or fetal hydrops, and rarely without those trophoblastic abnormalities. During the 1960s and 1970s, renal biopsies were taken from many women with hyper- tensive disorders in pregnancy at the Chi- cago Lying-in Hospital. Fisher, Luger, Spargo, and Lindheimer’ reviewed theHISTORY OF PREECLAMPSIA obstetric records and found that the clinical diagnoses of preeclampsia, mild and severe combined, were sustained by the anatomic findings in only 55% of cases. Some had no renal lesions, but most not having those of preeclampsia had the lesions of nephro- sclerosis, chronic _glomerulonephritis, chronic interstitial nephritis, chronic pyelo- nephritis, and a scattering of other renal diseases. The Chicago Lying-in Hospital has been a leader in research on the hypertensive disorders in pregnancy for half a century, and if their diagnostic error is 45% for preeclampsia, what must it be in routine diagnosis? Fisher et al. observed that the degree of proteinuria reflected the severity of the preeclamptic lesions; but because of the wide range of proteinuria in most of the diseases, its degree was of little value for differential diagnosis. Similarly, the aver- ages of blood pressure, serum creatinine, and urea nitrogen were different from group to group, but the wide overlap precludes differential diagnostic value. Plasma urate levels are often increased in preeclampsia, relative to urea nitrogen, but that change is, not specific although it would be better than creatinine or urea nitrogen for diagnosis. In summary, what is called preeclampsia, especially mild preeclampsia, may be as follows: 1, Preeclampsia; the diagnosis may be correct, in up to half of cases, or to about 70% if multiparas are excluded 2. Chronic renal disease 3. Chronic hypertension that had abated in midpregnancy 4, Latent essential hypertension, manifested by gestational hypertension, perhaps ac- companied by physiologic edema that had been interpreted as abnormal In the day-to-day managementof patients, acute hypertension should be regarded as preeclampsia with the threat of convulsions. Dieckmann observed that about 22% of eclamptic women had their first convulsions with systolic blood pressures that had not exceeded 140 mmHg, and that about 12% of that group died. As previously mentioned, convulsions may precede proteinuria as well. Barely diagnosable preeclampsia is potentially lethal. It must be emphasized that the working diagnosis for the purpose of management is quite different from a reliable diagnosis for the selection of patients in whom one can study preeclampsia. Now that eclampsia has become rare in centers equipped to study it, investigators have turned their attention to preeclampsia; in their impatience to get on with it, they too often include women with the diagnosis of “mild preeclampsia” and even multiparas with that diagnosis, which usually is erroneous. The inevitable result is confusion and conflicting conclusions. The minimal requirements in selecting patients for study should be as follows: Nulliparity Abundant proteinuria Age less than 25 years Reliable history of cardiovascular-renal normality or follow-up proving it 5. Perhaps hyperuricemia ere Ideally, the diagnoses should be con- firmed by renal biopsy, but that is hardly practicable. The clinical criteria specified would greatly reduce erroneous diagnoses, but not entirely so. Epidemiology The older German literature, generated when eclampsia was common, provides much information about the epidemiology of the disorder, some of which was reviewed by Hinselmann in 1924. Factors Predisposing to Preeclampsia-Eclampsia Nulliparity Mauriceau, in 1694, observed that primi- gravidas are far more likely to develop convulsions than are multiparas. Hinsel- mann collected 6498 cases of eclampsia for which the parity of the patients and the totala senna CHESLEY numbers of deliveries by parity had been published. On the average, 74% of the women had had eclampsia in the first preg- nancy carried to viability. From the pro- portions of deliveries in primiparas and multiparas, he calculated that the previ- ously nulliparous woman is eight times more likely to develop eclampsia than is the previously parous patient. MacGillivray compared the incidence of preeclampsia, as, defined by Nelson, in viable first preg- nancies with that in 516 women completing second pregnancies after abortion in the first, He reported that the incidence of severe preeclampsia was reduced by two-thirds and that of mild preeclampsia (gestational hy- pertension) by one-third in the pregnancies following abortion. That is, a little bit of pregnancy seemed to confer some immunity to preeclampsia in the next pregnancy. Campbell et al.,* in the same clinic, have repeated the study with 754 women having had spontaneous or induced abortions in the first pregnancy. They failed to find the alleged protective effect of an aborted pregnancy against preeclampsia in the next, gestation. When eclampsia does occur in a multi- para, she often has some predisposing factor that may not have been present in earlier gestations; chronic hypertension, diabetes, and plural gestation are examples. Familial History of Preeclampsia-Eclampsia ‘The older literature contains many case reports of eclampsia in close relatives; but Hinselmann seconded Scanzoni, in writing that eclampsia was so common that mere coincidence could explain such cases. Chesley, Cosgrove, and Annitto presented an analysis of pregnancies in 147 sisters and the first 110 daughters of eclamptic women in 1960, at the Seventh Conferences of the International Society for Geographic Pa- thology, in what I believe to be the first systematic study of the familial factor. The data for sisters and daughters, shown in Table 2, were published in 1968. All data in TABLE 2. Preeclampsia or Eclampsia in Sisters, Daughters, and Daughters-in-law of Women Having Had Eclampsia Relationship Cases Expected Observed Sisters 147 58.0 55 Daughters 187 489 48 I daughter R 184 18 2 daughters 28 12.5 13 Sto Sdaughters 12 83 8 Daughters-in-law (controls) 6 49 6 From Chesley LC. Hypertension in pregnancy: definitions, familial factor, and remote prognosis. Kidney Int 1980;18:284. By permission. “Numbers (not percentages) expected if a single recessive gene determines the development of pre- eclampsia. the table are absolute numbers, not per- centages. The first and last columns were republished by Cooper and Liston,’ who put in the middle column showing the numbers of cases of preeclampsia that we should have seen if a single recessive gene determined the development of preeclampsia. There is a close correspondence between the num- bers of observed and expected cases in all six categories: sisters; daughters, singly and by families; and daughters-in-law (con trols). We now have many more cases. Six sisters of eclamptic women have had eclampsia, and 87% of the sisters have had preeclamp- sia-eclampsia in their first pregnancies. Of 248 daughters, 26.2% have had preeclampsia or eclampsia in first pregnancies; 7 had eclampsia, or 1 in 35, as compared with the expected incidence of about 1 in 800. The data are more impressive when presented in another way. In families in which only one daughter had been tested by pregnancy, 24.7% have had preeclampsia. In families with two daughters (sisters), one or both daughters have had preeclampsia in 35.3%; with three daughters, at least one had preeclampsia in 46.7%; and with four or five, the familial incidence has been 66.7%. In daughters-in-law, used as controls, the incidence has been 6.1%.HISTORY OF PREECLAMPSIA The close agreement between the numbers of observed and expected cases is disturbing, for it seems to indicate that every woman homozygous for the gene will develop preeclampsia, That leaves no room for the operation of predisposing factors, several of which are beyond doubt. Thus, we are faced with the necessity of reconciling the genetic hypothesis with the operation of predispos- ing factors. Further evidence for a genetic factor has been adduced by Sutherland et al.,’ who compared the incidences of proteinuric preeclampsia in the first pregnancies of the mothers and mothers-in-law of women with the same disorder (probands) and of a control group with normal pregnancies matched with the probands for parity, marital status, sex of child, social class, and smoking habits. The incidences of protein- uric preeclampsia in the mothers of the women with the same disorder had been 15.9%, but was only 4% in the mothers-in- law of each group. The incidence in the mothers of the controls had been 8%. The data are compatible with the hypothesis that a single recessive gene determines the development of preeclampsia, but they do not exclude the possibility of polygenic inheritance. If a single recessive gene is responsible, certain predictions can be made about the pregnancies of the granddaughters of eclamptic women, of whom thereare several subgroups with differing prognoses. Thus far, we have the obstetric records of only 74, not nearly enough for analysis, but 12 have had preeclampsia. Plural Gestation Hamilton, in 1775, noted that win gesta- tions predispose to eclampsia. Among 7748 cases of eclampsia collected by Hinselmann, 6.4% were associated with plural gestation. That is, the incidence of plural gestation in eclamptic women was about 5.8 times that in all pregnancies. He could find few data for the incidence of eclampsia in plural gestations, but they did show a great increase. Thus, he concluded that plural gestation increases the risk of eclampsia about six times, When Mrs. Dionne bore her famous quintuplets, she is said to have had severe, nearly fatal preeclampsia; she was para 7 at the time (she needed five more), but her relative immunity as a multipara was overridden by the plural gestation. Diabetes Mellitus Although case reports of the occasional pregnancies of diabetic women carried to viability mentioned eclampsia or “albumin- uria”, the first clear indication that diabetes predisposes to eclampsia came in 1985 when Priscilla White described 257 pregnancies in 180 women with diabetes. Two of 15 juvenile diabetics had eclampsia, the overall incidence of which was 5%. Many later publications have confirmed the high inci- dence of hypertensive disorders in pregnant diabetic women, with combined incidences and prevalences as high as 50%. Chronic renal and vascular diseases and essential hypertension often are associated with dia- betes, and problems of differential diagnosis prohibit the specification of a multiple by which diabetes increases the incidence of preeclampsia, but certainly diabetes is a predisposing factor. Chronic Hypertension Seitz, in 1916, recognized that Schwanger- schaftsnephropathie (preeclampsia) is some- times superimposed upon chronic hyperten- sion (then regarded as chronic nephritis). The American Committee on Maternal Welfare did not recognize superimposed preeclampsia-eclampsia until 1952, al- though several studies of pregnancies in women with chronic hypertension had de- scribed the aggravation of hypertension and the appearance of proteinuria and abnormal edema, and even of convulsions. Most of the authors regarded the syndrome as super- imposed preeclampsia or eclampsia, butCHESLEY Dieckmann thought it to be merely an “intensification” of the underlying disorder with a clinical course different from that of preeclampsia. Although superimposed preeclampsia is grossly over-diagnosed, as indicated by McCartney's findings in renal biopsies, the incidence of preeclampsia is considerably higher in chronically hypertensive multi- paras than in normotensive ones; the same is true of primigravidas. Hydatidiform Mole Malichecq described a case of eclampsia associated with a large hydatidiform mole in 1866, and many more have been reported since, A noteworthy feature is the early onset of preeclampsia-eclampsia, even as early as the 12th week of gestation. In 1928 Hitsch- mann, in his monograph on hydatidiform mole and chorioepithelioma, specified the onset of preeclampsia before mid pregnancy as one of the diagnostic signs of mole. Page did not see preeclampsia in the presence of small, slowly growing moles, but wrote that the incidence is about 70% in association with large, rapidly growing ones. Such a mole overrides the relative immunity of multiparas to preeclampsia. Some have doubted that the triad of hypertension, proteinuria, and edema as- sociated with mole is really preeclampsia. It does seem to be, for more than 60 cases of apparent eclampsia have been reported and Sanchez-Torres and Santamaria found the lesions characteristic of preeclampsia in renal biopsies from all 14 of their cases with mole and the syndrome. Fetal Hydrops In 1892, Ballantyne described “triple edema,” a syndrome in which lethal fetal hydrops is associated with gross edema of the fetus and placenta and edema and proteinuria in the mother (blood pressures were not obtainable in those days). Jann reviewed the literature and found triple edema together with maternal proteinuria, and hypertension in up to 70% of cases. Most of the instances occurred in multiparas with Rh isoimmunization, although preeclamp- sia has a high incidence in cases of fetal hydrops of other causes. Rh isoimmuniza- tion is not associated with an increased incidence of preeclampsia in the absence of fetal hydrops. Typically, the preeclampsia hasan early onset and is severe; several cases of eclampsia have been reported, one as early as the 16th week of gestation. Extremes of Age Lehmann’ analyzed all cases of eclampsia occurring in Denmark during the 10-year period, 1918 through 1927, in more than 781,000 deliveries. Lehmann’s monograph is in Danish and contains a wealth of information, but I have not seen a direct reference to it in the American or English literature, presumably because of the lan- guage barrier. Figure 1, based upon Lehmann’s Table 52, shows the age-specific incidences for all cases of eclampsia. It may be seen ata glance that the incidences are highest at the extremes of the childbearing ages. Interpre- tation of the Figure is complicated by the fact that nearly all of the youngest women must have been primigravidas, and that alone might account for the high incidence of eclampsia among them. Lehmann’s Table 59 provides the age-specific inci- dences for primigravidas and shows the same pattern as Figure 1. He did not tabulate the date for multiparas, but they can be ECLAMPSIA PER 1000 's 20 25 30 35 40 45 50 AGE FIG. 1. Age-specific incidences of eclampsia, (From Lehmann.?)HISTORY OF PREECLAMPSIA derived by subtracting the numbers in his Table 59 (primigravidas) from those in Table 52 (all cases). Again, the age-specific incidences follow the same pattern. In both primigravidas and multiparas the highest incidences of eclampsia occur at the ex- tremes of age. The increase after age 35 may represent, in part, the superimposition of eclampsia upon chronic hypertension. Other Factors Alleged to Predispose to Preeclampsia-Eclampsia Hydramnios Most writers have accepted Miquel’s state- ment, made in 1824, that hydramnios predisposes to preeclampsia-eclampsia. The issue is confused by the fact that hydramnios often is associated with plural gestation, fetal hydrops, and diabetes, which of themselves predispose to preeclampsia, as pointed out by Scott. Among Scott's 169 cases of hydramnios, 18 were in twin pregnancies, 4 were associated with fetal hydrops, and 4 were associated with dia- betes; 11 of the 21 women developed preeclampsia, Of the remaining 148 women, 16 developed preeclampsia, but the expected number was 13, based upon the proportions of primigravidas and multiparas. Thirty (of 169? of 1487) developed gross edema, but no other signs of preeclampsia. Habitus. ‘The literature of eclampsia in the nine- teenth century abounds in references to an alleged relation between bodily build and susceptibility to eclampsia, but opposite views were held. Denman, for instance, wrote that frail, delicate, highly intelligent, educated city women who cultivated music are especially susceptible. Others believed that short, squat, obese, plethoric dump- lings are at the highest risk. Hinselmann concluded from his review of the literature that there probably is no relation, although a hypoplastic cardiovascular system might be a factor. There are few systematic studies, Bublit- schenko compared 187 of his 226 Russian eclamptic women with unmatched controls and without separation of primigravid from multiparous women. He concluded that eclampsia typically occurs in short, heavy women with wide pelvises. Sserdjukoff and Melnikoff conducted an extensive anthro- pometric study of 1000 Russian pregnant women among whom there were 16 cases of eclampsia, mostly in pyknic subjects. Rhodes compared 20 English eclamptic women with those having normal preg- nancies and with others having mild and severe preeclampsia as defined by Nelson. ‘The eclamptic women weighed less before pregnancy and their ponderal indices (height//weight) were higher than those in the other groups; that is, they were thinner. Rhodes suggested that the more slender the woman, the greater the risk of eclampsia. I analyzed the habitus of 242 eclamptic women whose heights and weights were known; data were not available for the remaining 57 women in the series, Prim gravid and multiparous women were signif- icantly different, in that the multiparas were heavier, had higher weight/height ratios, and had lower indices of height/*/weight. On the basis of stated prepregnancy weights, only 8% of the primigravidas were over- weight, whercas from 28% to 36% of the multiparas were (the range depends upon the index used). At follow-up, from 28% to 38% of the primiparas were overweight, as were from 52% to 60% of the multiparas. Sixty-three of the primigravidas were compared with 68 controls with normal pregnancies, matched for age; race; parity; clinic, nonclinic, or private status; and as nearly as possible with the time of delivery. The eclamptic women were more slender, with only 8% being overweight, as con- trasted to from 21% to 30% of the controls. In sum, eclampsia spares no habitus, but slender, underweight women are at slightly greater risk. Several studies of women thought to havehv CHESLEY preeclampsia were published in the 1980s, and there was another flurry of such papers in the 1960s. Most studies included large proportions of multiparas, and most cases were diagnosed as mild preeclampsia. Not astonishingly, nearly all writers concluded that short, heavyset, obese, thick- and short- necked women are the most susceptible. That is the habitus often associated with essential hypertension, and the selection of patients for study leaves the conclusion highly suspect. Lowe recognized that apparently acute hypertension in late pregnancy is not necessarily preeclampsia. The diagnoses among his 1096 primigravidas were as follows: proteinuric preeclampsia, 5.5%; proteinuria without mention of hyperten- sion, 8.6%; hypertension alone, 6.4%. In combining the three groups Lowe found the usual correlation with height/?/weight, with the highest incidence of abnormality among the overweight women. When he analyzed each group separately, however, only the women with hypertension alone conformed to the pattern, The incidence of proteinuric hypertension increased as the index increased; that is, the more slender the women, the higher the incidence of pre- eclampsia. Race It seems to be almost generally believed that black women are more susceptible than white women to the development of pre- eclampsia-eclampsia in the United States, but it is doubtful. From the opening of the Margaret Hague Maternity Hospital, in Jersey City, New Jersey, on October 16, 1981, through 1951, black women constituted 9.0% of the eclamptic patients and 8.5% of all women delivered; they were not disproportionately represented in the eclamptic group. Since 1958, I have made all of the final diagnoses of the hypertensive disorders in pregnancy in the Kings County and State University Hospital; the diagnoses may be questioned: but, however erroneous, my criteria have remained essentially constant. From 1960 to 1970 the incidences of preeclampsia, eclampsia, and preeclampsia superimposed upon chronic hypertension were virtually identical in black and white women. The incidences during that period were extraor- dinarily low, as is discussed in the section on temporal variations. The prevalence of chronic hypertension in black women was more than twice that in white patients, as is well known. It seems probable that the alleged racial predisposition of blacks to preeclampsia stems from problems of differ- ential diagnosis. Mengert, who had vast experience in three major medical centers, also concluded that blacks are not more susceptible than whites to preeclampsia- eclampsia. The incidence of preeclampsia-eclampsia has been said to be low in native Fijians but high in Indians, who make up about half the population of the Fiji Islands, When I was in Suva in 1980 I was told that there is no racial difference, and that one of the sources of the allegation had never been in Fiji. Magee reporteda striking racial difference in the Colonial Hospital in Trinidad. The incidences of preeclampsia and eclampsia in East Indian women (immigrants) were 14.6% and 2.7%, respectively, in 1701 deliver- ies. In contrast, the incidence in 7152 deliveries by native black women were 6.3% and 0.6%, respectively. About half of the admittances of each group were emergen- cies, and the incidences do not reflect those in the whole populations, but the difference is so large as to suggest a real racial factor. Davies et al. reported a unique survey in which a team of investigators made all of the final diagnoses in every pregnant woman in Jerusalem from 1964 through 1966. The lowest incidence of preeclampsia, 1.6%, was in Jews born in Africa and Israel; the incidence was doubled in Moslem women (8.4%) and in Jews born in Iraq (3.3%). If preeclampsia is determined genetically, one might expect racial differences, but the study by Davies et al. is the only one that makes a convincing case.HISTORY OF PREECLAMPSIA Socioeconomic Status Eclampsia was thought to be a disorder of upper-class women until well into the present century. In part, it may have been because affluent women were attended by physicians, who published, whereas poor women were delivered by midwives, who seldom wrote papers. Yet Fitzgibbon, who attended women of all classes in Dublin's Rotunda Hospital, wrote, in 1922; “Tox- aemia is unquestionably a disease of the well-to-do classes of society.” In the same year, Ruiz Contreras wrote that the inci- dences of eclampsia and albuminuria of pregnancy were much greater in his private patients than in the poor women whose cases he managed in the charity wards in Barcelona. Tt was not until about 1930 that impover- ished women were recognized as equally susceptible to eclampsia, and the current cry is that they are much more susceptible. Nelson, who had access to the details of nearly all pregnancies and deliveries in Aberdeen, Scotland, found that the inci- dences of preeclampsia were essentially the same in each of the five social classes, ranging from the professional and affluent to unskilled laborers. Physicians often are reluctant to diagnose preeclampsia in their private patients. In one extreme case the diagnosis was “Normal pregnancy, normal delivery and normal puerperium” in a woman with severe eclampsia (nine convulsions), who had been on the Danger List for 24 hours. Illegitimate Pregnancy Ryan wrote in 1831 that the incidence of eclampsia was higher in unmarried than in married mothers. The observation has been confirmed repeatedly, by Lehmann? among others, and it has been used as an argument for a psychosomatic cause of eclampsia. Lehmann went a step further, however, in showing that the incidences were virtually identical in married and unmarried primi- gravidas; the incidence was much lower in multiparas, but the same in married and unmarried mothers. Preeclampsia- eclampsia is predominantly a disease of primigravidas; and until fairly recently most unmarried mothers were primigravi- das. Lehmann’s findings have been con- firmed by later writers. Clty Versus Country Women Eclampsia has been thought for more than two centuries to afflict city women more frequently than their country cousins. Leh- mann? agreed; but, again, when he related the incidence of eclampsia to parity, the difference almost disappeared. The trans- port of rural patients to hospitals in cities could account for the residual difference. Until fairly recently, farmers begat their future farmhands; and their families were larger than those in the city. Consequently, there was a larger proportion of primigravi- das in the cities. Secondary Sex Ratio Buttner observed that there was an excess of male fetuses in eclamptic pregnancies. A tabulation of 4799 cases of eclampsia from the literature shows the average ratio of males to females to be 122.5:1.00, with an excess of 354 boys over the expected num- ber.’ The observation has excited several speculations, such as HLA antigens on the Y chromosome being responsible for pre- eclampsia-eclampsia, but the fact is that nearly half of all cases of eclampsia are associated with female fetuses. Fetal Malformations Gabelchoverus, in 1596, wrote that fetal malformations predispose to what he called epilepsy of uterine origin. Miquel and later writers agreed. De Watteville wrote that severe fetal malformations occurred in about 1% of 16,000 normal pregnancies, but in 4.2% of women with hypertension. Conversely, the combined incidences and prevalences of hypertensive disorders was 3.0% in all pregnancies, but 6.2% in women bearing malformed fetuses.CHESLEY Nelson reported that the incidence of severe fetal anomalies was lower in all cases of preeclampsia occurring in Aberdeen, Scotland, between 1988 and 1958 than in the whole obstetric population of the city. Stevenson analyzed 539 cases of hydramnios from Belfast and Oxford, and found a significantly lower incidence of fetal mal- formations in hypertensive than in normo- tensive mothers. It appears that any association there may be between fetal anomalies and preeclamp- sia is weak at best. Weather Hippocrates wrote in his Aphorism 1, Section II: "The changes of seasons mostly engender diseases, and in the same season great changes either of heat or cold, and the rest agreeably to the same rule.” Many writers since Smellie in 1756 have observed clustering of cases of eclampsia and attrib- uted it to changes in weather. In the early years of the present century there were hundreds of cases of eclampsia each year in New York City. Davis and Harrar found that the incidence reached a peak in March and April, which the nutritionists seized upon as evidence for malnutrition as the cause of eclampsia; the meteorologists cited it as evidence for un- settled weather as the cause. It may be that the brides of the previous June had provided a surge of primigravidas. Miscellaneous Associations Sserdjukoff and Melnikoff wrote thatall but 1 of their 16 eclamptic women were blondes. Dieckmann, recognizing that the color of a woman's hair is not necessarily natural, recorded the color of the eyes; 58% of his eclamptic women had blue eyes, versus the 37% of his patients with chronic hyperten- sion, The difference was offset, however, by the 87% of women with severe preeclampsia having blue eyes. He provided no informa- tion as to the color of the eyes in the whole clinic, I have been unable to find any documentation for the statement that women with red hair are more susceptible to eclampsia than are either blondes or bru- nettes, or for the allegation that separation of the upper incisors marks the women at high risk. Persianinov observed that the incidence of “toxemia” was 8% in women who slept more than 8 hours day, 38% in those sleeping 7-8 hours, and 70% in those sleeping less than 7 hours. Temporal Trend Accurate estimates of the incidence of eclampsia in populations are difficult to ascertain, and they do not exist for pre- eclampsia. The older data are unreliable because the incidences in hospitals were loaded with women admitted as emergen- cies, and who would not have been seen except for their complications. Statewide eclamptic deaths have depended more upon the availability, quality, and utilization of prenatal care than upon the incidence of preeclampsia. The early experience of several major German clinics had been that the incidence of eclampsia followed a temporally undulat- ing curve, with the incidence decreasing for several years, then rising progressively for several more. In Denmark, Lehmann’ noted a fairly steady increase in the incidence of eclampsia, year by year, from 1.38 per 1000 deliveries in 1918 to 2.09 in 1927. In Saxony, where eclampsia became a reportable disease in 1920, Kistner noted a similar wend between 1920 and 1928. Figure 2 depicts the temporal trend in the incidence of preeclampsia in the Kings County and State University Hospitals in Brooklyn, New York, where I have made all the diagnoses with constant criteria. The incidence fell progressively from 6% in 1954 to 1.8% in 1966; it has risen progressively since 1966 and reached 5.8% in 1974. The increase since 1966 appears to be largely accountable to the increasing proportion of primigravid deliveries, indicated by the broken line in the graph. The decrease between 1954 and 1966 cannot be explainedHISTORY OF PREECLAMPSIA Nuliparas, percent i388 BTS Years FIG. 2. Yearly incidences of preeclampsia in the Kings County Hospital, 1953-1974, in relation to the proportionate representation of nulliparas coming to delivery. (From Chesley.°) on the basis of the patients’ parity, or on any other basis that occurs to me. The temporal trend is hard to reconcile with the genetic hypothesis as to the cause of preeclampsia. Eclampsia In Times of War The incidence of eclampsia in hospitals decreased in some parts of Germany and the other Central Powers during World War I, but not in others such as Bavaria, Mecklen- burg, and Altmark. Lichtenstein questioned the reality of the decrease on two grounds. He wrote that between 1895 and 1914 the incidence of eclampsia in the Leipzig clinic had varied from 1.5% to 3.5% with the undulating curve described in the preceding section. The annual incidences had been decreasing just before the war, and during the war the incidence did not fall so low as had been observed in earlier years. He also noted that the admittance of patients from the outlying areas of Leipzig had decreased by half during the war, and suggested that it might be because of the severe shortage of physicians in the countryside and the extraordinary difficulties of transport. Hinselmann collected from the literature 161,443 births during the 4 years preceding the war, 131,704 during the war, and 113,077 in the following 4 years, and found the incidences of eclampsia to be 1.44%, 0.85%, and 1.27%, respectively. Presumably the data were hospital statistics, and Lichtenstein's criticism may be valid. Nonetheless, the apparent reduction has been accepted as real, and several speculations have been advanced in explanation. Some believed that the women’s working in field and factory had a beneficial effect; others postu- lated a decreased exposure to antibodies or toxins in semen, The nutritionists prevailed for a time. They argued that Germans overeat in good times and that during the war they had, of necessity, reduced their intake of proteins, An editorial in the Journal of the American Medical Associa- tion pontificated: “The conclusion seems inevitable that the restriction of fatand meat tends to ward off eclampsia.” Proteins had been condemned again, and the conclusion reinforced the centuries-old belief that di- etary proteins should be restricted during pregnancy. Also, the limitation of weight gained in pregnancy as prophylaxis against preeclampsia stemmed from the same source. Actually, the decrease in eclampsia in Germany began before there were shortages of food, and the incidence rose markedly after the war, when the shortages were most severe because of the blockade that the Allies unconscionably maintained for a year after the armistice. During the counterrevolution in Spain, about 20 years later, the incidence of eclampsia soared in Madrid while it was, under siege and severely short of food. Lehmann obtained statistics from the Landsamt in Karlsruhe from which he found that the incidence of eclampsia in all of Baden had paralleled the proportion of primigravid deliveries. Diaz del Castillo made the same correlation in the statistics from Madrid. In Germany the young men were at the front, and the nulligravidas were not conceiving at the usual rate. In MadridCHESLEY the young men were cooped up with the young women, and in due course there were parallel surges of primigravid deliveries and eclampsia. Many of the contradictory papers on the incidence of eclampsia during World War II are reviewed elsewhere’. Geographic Distribution Davies" collected more than 170 reports from 50 countries and concluded: “. . . the vast majority of the reports in the literature are not representative of the populations studied and because of variations in classifi- cation, the absence of data on age and parity structure and so forth cannot be used for any international comparison, even of the gross- est sort.” True. References Note: References for undocumented statements in the text can be found in Chesley.? 1, Quilligan EJ, Little AB, Oh W, Caz, Alex- ander D, Cornblath M. Pregnancy, birth and infant. Washington, DC: US Department of Health and Human Services, Public Health Service, 1981. 10. . Editorial. Treatment of moderate hyperten- sion in pregnancy. Br Med J 1980;280:1483. . Chesley LC, Hypertensive disorders in preg- nancy. New York: Appleton-Century-Crofts, 1978. Fisher KA, Luger A, Spargo BH, Lindheimer MD. Hypertension in pregnancy: clinico- pathological correlations and late prognosis. Medicine 1981;60:267. . Pritchard JA. Standardized treatment of 154 consecutive cases of eclampsia. Am J Obstet Gynecol 1975;128:543. 3. Campbell DM, Carr-Hill R, Orisaseyi AE. Pre-eclampsia in a second pregnancy. Clin Exp Hypertension 1988;B2:308. . Cooper DW, Liston WA. Genetic control of severe pre-eclampsia. J Med Genet 1979;16: 409. Sutherland A, Cooper DW, Howie PW, Liston WA, MacGillivray I, The incidence of severe pre-eclampsia amongst mothers and mothers-in-law of pre-eclamptics and controls. Br J Obstet Gynaecol 1981;88:785. ). Lehmann K. Eklampsien i Danmark i arene 1918-1927. Copenhagen: Busck, 1933. Davies AM. Geographic epidemiology of the toxemias of pregnancy. Springfield, Ill: Charles C Thomas, 1971.