From Williams Obstetrics:

SECTION VII - COMMON COMPLICATIONS OF PREGNANCY

24. Hypertensive Disorders in Pregnancy

Hypertensive disorders complicating pregnancy are common and form one of the deadly triad, along with hemorrhage and infection, that results in much of the maternal morbidity and mortality related to pregnancy. According to the National Center for Health Statistics in 1998, hypertension associated with pregnancy was the most common medical risk factor (Ventura and colleagues, 2000). It was identified in 146,320 women, or 3.7 percent of all pregnancies that ended in live births. In 12,345 of these women eclampsia was diagnosed, and maternal deaths from this complication still remain a threat. Berg and colleagues (1996) reported that almost 18 percent of 1450 maternal deaths in the United States from 1987 to 1990 were from complications of pregnancy-related hypertension.

How pregnancy incites or aggravates hypertension remains unsolved despite decades of intensive research, and hypertensive disorders remain among the most significant unsolved problems in obstetrics. Important ongoing research currently is sponsored by the National Institutes of Child Health and Human Development (NICHD) and its Maternal-Fetal Medicine Units Network. Another important stimulus for research is the International Society for the Study of Hypertension in Pregnancy (ISSHP). Also, the National Heart, Lung, and Breast Institute (NHLBI) encourages ongoing research and coordination through the National High Blood Pressure Education Program (NHBPEP) and its Working Group Report on High Blood Pressure in Pregnancy.

TERMINOLOGY

In former editions of this textbook, the authors preferred to use the term pregnancy-induced hypertension to describe any new-onset pregnancy-related hypertension. This designation was chosen because it served to emphasize the cause-and-effect connection between pregnancy and a unique form of hypertension manifest in women only during reproduction. It was also intended that pregnancy-induced hypertension would include the development of hypertension without proteinuria, including in nulliparous women. In this latter circumstance, pregnancy-induced hypertension was also a potential precursor to preeclampsia or eclampsia, which require proteinuria for diagnosis. Our purpose was to communicate that the development of hypertension in a previously normotensive pregnant woman should and must be considered potentially dangerous to both her and her fetus. The designation pregnancy-induced hypertension also had the advantage of signifying that most hypertensive nulliparous women had only transient uncomplicated hypertension that subsided promptly after delivery. Recently, the working group of the National High Blood Pressure Education Program (2000) has proposed a classification system that accomplishes all the foregoing goals of the authors and thus will be adopted in this edition of Williams Obstetrics. The authors have taken this approach in an effort to deal with the nonuniform and confusing terminology that has long plagued the diagnosis of hypertension in pregnancy (American College of Obstetricians and Gynecologists, 1996).

DIAGNOSIS

The diagnosis of hypertensive disorders complicating pregnancy, as outlined by the Working Group (2000), is shown in Table 24-1. There are five types of hypertensive disease that include:

1. Gestational hypertension (formerly pregnancy-induced hypertension or transient hypertension).

2. Preeclampsia.

3. Eclampsia.

4. Preeclampsia superimposed on chronic hypertension.

5. Chronic hypertension.

An important consideration in this classification is differentiating hypertensive disorders that precede pregnancy from preeclampsia, which is a potentially more ominous disease.

Hypertension is diagnosed when blood pressure is 140/90 mm Hg or greater, using Korotkoff phase V to define diastolic pressure. Edema has been abandoned as a diagnostic criteria because it occurs in too many normal pregnant women to be discriminant. In the past, it had been recommended that an increment of 30 mm Hg systolic or 15 mm Hg diastolic blood pressure be used as a diagnostic criterion, even when absolute values were below 140/90 mm Hg. This criterion is no longer recommended because evidence shows that women in this group are not likely to suffer increased adverse pregnancy outcomes (Levine, 2000; North and colleagues, 1999). That said, women who have a rise of 30 mm Hg systolic or 15 mm Hg diastolic warrant close observation.

GESTATIONAL HYPERTENSION

As shown in Table 24-1, the diagnosis of gestational hypertension is made in women whose blood pressure reaches 140/90 mm Hg or greater for the first time during pregnancy, but in whom proteinuria has not developed. Gestational hypertension is termed transient hypertension if preeclampsia does not develop and the blood pressure has returned to normal by 12 weeks' postpartum. In this classification the final diagnosis that the woman does not have preeclampsia is made only postpartum. Thus, gestational hypertension is a diagnosis of exclusion. Importantly, however, women with gestational hypertension may develop other signs associated with preeclampsia, for example, headaches, epigastric pain, or thrombocytopenia, which influence management.

When blood pressure rises appreciably during the latter half of pregnancy, it is dangerous—especially to the fetus—not to take action simply because proteinuria has not yet developed. As Chesley (1985) emphasized, 10 percent of eclamptic seizures develop before overt proteinuria. Thus, it is clear that when blood pressure begins to rise, both mother and fetus are at increased risk. Proteinuria is a sign of worsening hypertensive disease, specifically preeclampsia; and when it is overt and persistent, maternal and fetal risks are increased even more.

PREECLAMPSIA

Preeclampsia is a pregnancy-specific syndrome of reduced organ perfusion secondary to vasospasm and endothelial activation. Proteinuria is an important sign of preeclampsia, and Chesley (1985) rightfully concluded that the diagnosis is questionable in its absence. Proteinuria is described as 300 mg or more of urinary protein per 24 hours or persistent 30 mg/dL (1+ dipstick) in random urine samples. The degree of proteinuria may fluctuate widely over any 24-hour period, even in severe cases. Therefore, a single random sample may fail to demonstrate significant proteinuria.

McCartney and co-workers (1971), in their extensive study of renal biopsy specimens obtained from hypertensive pregnant women, invariably found that proteinuria was present when the glomerular lesion considered to be characteristic of preeclampsia was evident. Importantly, both proteinuria and alterations of glomerular histology develop late in the course of hypertensive disorders due to pregnancy. In fact, preeclampsia becomes evident clinically only near the end of a covert pathophysiological process that may begin 3 to 4 months before hypertension develops (Gant and associates, 1973). As shown in Table 24-1, the minimum criteria for the diagnosis of preeclampsia are hypertension plus minimal proteinuria. The more severe the hypertension or proteinuria, the more certain is the diagnosis of preeclampsia (Table 24-2). Similarly, abnormal laboratory findings in tests of renal, hepatic, and hematological function increase the certainty of preeclampsia. Persistent premonitory symptoms of eclampsia such as headache and epigastric pain also increase the certainty of preeclampsia.

The combination of proteinuria and hypertension during pregnancy markedly increases the risk of perinatal mortality and morbidity (Ferrazzani and associates, 1990). Results from a 13-year prospective study reported by Friedman and Neff (1976) in over 38,000 pregnancies are shown in Table 24-3. Hypertension alone, defined by a diastolic blood pressure of 95 mm Hg or greater, was associated with a threefold increase in the fetal death rate. Worsening hypertension, especially if accompanied by proteinuria, was more ominous. Conversely, proteinuria without hypertension had little overall effect on the fetal death rate. Naeye and Friedman (1979) concluded that 70 percent of the excess fetal deaths in these same women were due to large placental infarcts, markedly small placental size, and abruptio placentae. They concluded that these causes usually develop late in the course of the disease. Certainly, persistent proteinuria of 2+ or more, or 24-hour urinary excretion of 2 g or more, is severe preeclampsia. With severe renal involvement, glomerular filtration may be impaired, and plasma creatinine may rise.

Epigastric or right upper quadrant pain likely results from hepatocellular necrosis, ischemia, and edema that stretches Glisson's capsule. This characteristic pain is frequently accompanied by elevated serum liver enzymes, and usually is a sign to terminate the pregnancy. The pain presages hepatic infarction and hemorrhage as well as catastrophic rupture of a subcapsular hematoma. Fortunately, hepatic rupture is rare and most often associated with hypertension in older and multiparous women.

Thrombocytopenia is characteristic of worsening preeclampsia, and probably is caused by platelet activation and aggregation and microangiopathic hemolysis induced by severe vasospasm. Evidence for gross hemolysis such as hemoglobinemia, hemoglobinuria, or hyperbilirubinemia is indicative of severe disease.

Other factors indicative of severe hypertension include cardiac dysfunction with pulmonary edema as well as obvious fetal growth restriction.

SEVERITY OF PREECLAMPSIA. The severity of preeclampsia is assessed by the frequency and intensity of the abnormalities listed in Table 24-2. The more profound these aberrations, the more likely is the need for pregnancy termination. Importantly, the differentiation between mild and severe preeclampsia can be misleading because apparently mild disease may progress rapidly to severe disease.

Although hypertension is a requisite to diagnosing preeclampsia, blood pressure alone is not always a dependable indicator of its severity. For example, a thin adolescent woman may have 3+ proteinuria and convulsions while her blood pressure is 140/85 mm Hg, whereas most women with blood pressures as high as 180/120 mm Hg do not have seizures. Convulsions are usually preceded by an unrelenting severe headache or visual disturbances; thus, these symptoms are considered ominous.

ECLAMPSIA

Eclampsia is the occurrence of seizures in a woman with preeclampsia that cannot be attributed to other causes. The seizures are grand mal and may appear before, during, or after labor. Seizures that develop more than 48 hours postpartum, however, especially in nulliparas, may be encountered up to 10 days postpartum (Brown and colleagues, 1987; Lubarsky and associates, 1994).

PREECLAMPSIA SUPERIMPOSED UPON CHRONIC HYPERTENSION

All chronic hypertensive disorders, regardless of their cause, predispose to development of superimposed preeclampsia or eclampsia. These disorders can create difficult problems with diagnosis and management in women who are not seen until after midpregnancy. The diagnosis of chronic underlying hypertension is suggested by:

1. Hypertension (140/90 mm Hg or greater) antecedent to pregnancy.

2. Hypertension (140/90 mm Hg or greater) detected before 20 weeks (unless there is gestational trophoblastic disease).

3. Persistent hypertension long after delivery (Table 24-1).

Additional historical factors that help support the diagnosis are multiparity and hypertension complicating a previous pregnancy other than the first. There is also usually a strong family history of essential hypertension.

The diagnosis of chronic hypertension may be difficult to make if the woman is not seen until the latter half of pregnancy. This is because blood pressure decreases during the second and early third trimesters in both normotensive and chronically hypertensive women (Chap. 45, p. 1211). Thus, a woman with chronic vascular disease, who is seen for the first time at 20 weeks, will frequently have a normal blood pressure. During the third trimester, however, blood pressure may return to its former hypertensive level, thus presenting a diagnostic problem as to whether the hypertension is chronic or induced by pregnancy.

Some of the many causes of underlying hypertension that are encountered during pregnancy are listed in Table 24-4. Essential hypertension is the cause of underlying vascular disease in more than 90 percent of pregnant women. McCartney (1964) studied renal biopsies from women with "clinical preeclampsia," and found chronic glomerulonephritis in 20 percent of nulliparas and in nearly 70 percent of multiparas. Fisher and co-workers (1969), however, did not confirm this high prevalence of chronic glomerulonephritis.

Chronic hypertension causes morbidity whether or not a woman is pregnant. Specifically, as discussed in Chapter 45, it may lead to ventricular hypertrophy and cardiac decompensation, cerebrovascular accidents, or intrinsic renal damage. In some young women, hypertension develops as a consequence of underlying renal parenchymal disease. Dangers specific to pregnancy complicated by chronic hypertension include the risk of superimposed preeclampsia, which may develop in up to 25 percent of these women (Sibai and colleagues, 1998). Additionally, the risk of abruptio placentae is increased substantively especially in those women who develop superimposed preeclampsia (Chap. 25, p. 623). Moreover, the fetus of the woman with chronic hypertension is at increased risk for growth restriction and death.

Preexisting chronic hypertension worsens in some women, typically after 24 weeks. If accompanied by proteinuria, superimposed preeclampsia is diagnosed. Often, superimposed preeclampsia develops earlier in pregnancy than "pure" preeclampsia, and it tends to be quite severe and accompanied in many cases by fetal growth restriction.

The diagnosis requires documentation of chronic underlying hypertension. Superimposed gestational hypertension is characterized by worsening hypertension, keeping in mind that both systolic and diastolic pressures normally rise after 26 to 28 weeks. Preeclampsia is accompanied by proteinuria. Indicators of severity shown in Table 24-2 are also used to further characterize these disorders.

INCIDENCE AND RISK FACTORS

Gestational hypertension more often affects nulliparous women. Older women, who accrue an increasing incidence of chronic hypertension with advancing age, are at greater risk for superimposed preeclampsia. Thus, women at either end of reproductive age are considered to be more susceptible (Chap. 9, p. 207).

The incidence of preeclampsia is commonly cited to be about 5 percent, although remarkable variations are reported. The incidence is markedly influenced by parity; it is related to race and ethnicity—and thus to genetic predisposition; and environmental factors may also have a role. For example, Palmer and colleagues (1999) reported that high altitude in Colorado increased the incidence of preeclampsia. Some investigators have concluded that socioeconomically advantaged women have a lesser incidence of preeclampsia, even after racial factors are controlled. Conversely, in carefully controlled epidemiological studies in Scottish women, Baird and colleagues (1969) found that the incidence of preeclampsia was not different among five social classes.

The incidence of hypertensive disorders due to pregnancy in healthy nulliparous women has been carefully studied in a recent randomized trial of daily maternal dietary calcium supplementation (Hauth and colleagues, 2000). Of 4302 nulliparous women delivered at or beyond 20 weeks' gestation, a fourth developed a pregnancy-related hypertensive disorder. Of all nulliparas, preeclampsia was diagnosed in 7.6 percent and severe disease as defined in Table 24-2 developed in 3.3 percent.

Other risk factors associated with preeclampsia include multiple pregnancy, history of chronic hypertension, maternal age over 35 years, obesity, and African-American ethnicity (Conde-Agudelao and Belizan, 2000; Sibai and colleagues, 1997; Walker, 2000). The relationship between maternal weight and risk of preeclampsia is progressive, increasing from 4.3 percent for women with a body mass index less than 19.8 kg/m² to 13.3 percent for those greater than or equal to 35 kg/m². In women with twin gestations compared with those with singletons, the incidence of gestational hypertension (13 versus 6 percent) and preeclampsia (13 versus 5 percent) are both significantly increased (Sibai and co-authors, 2000). Moreover, women with twins and hypertensive disorders due to pregnancy experience higher rates of adverse neonatal outcomes than do those with singletons. Although maternal smoking causes a variety of adverse pregnancy outcomes, ironically, smoking during pregnancy has consistently been associated with a reduced risk of hypertension during pregnancy (Zhang and colleagues, 1999). Placenta previa has also been claimed to reduce the risk of hypertensive disorders due to pregnancy (Ananth and colleagues, 1997).

ECLAMPSIA. In general, eclampsia is preventable, and it has become less common in the United States because most women now receive adequate prenatal care. For example, in the 15th edition of Williams Obstetrics (1976), the incidence of eclampsia at Parkland Hospital was cited to be 1 in 700 deliveries for the prior 25-year period. For the 4-year period 1983 to 1986, the incidence was 1 in 1150 deliveries, and for 1990 to 2000 the incidence was approximately 1 in 2300 deliveries. Using figures from the National Vital Statistics Report, Ventura and colleagues (2000) estimated an incidence of about 1 in 3250 for the United States in 1998. Douglas and Redman (1994) cite an incidence of 1 in 2000 for the United Kingdom in 1992.

Mattar and Sibai (2000) have chronicled the hazards in 399 consecutive eclamptic women delivered between 1977 and 1998 at their center in Memphis. Major complications included abruptio placentae (10 percent), neurological deficits (7 percent), aspiration pneumonia (7 percent), pulmonary edema (5 percent), cardiopulmonary arrest (4 percent), acute renal failure (4 percent), and maternal death (1 percent).

PATHOLOGY

Pathological deterioration of function in a number of organs and systems, presumably as a consequence of vasospasm and ischemia, has been identified in severe preeclampsia and eclampsia. For descriptive purposes, these effects are separated into maternal and fetal consequences; however, these aberrations often occur simultaneously. Although there are many possible maternal consequences of hypertensive disorders due to pregnancy, for simplicity these effects are considered by analysis of cardiovascular, hematological, endocrine and metabolic, and regional blood flow changes with subsequent end-organ derangements. The major cause of fetal compromise occurs as a consequence of reduced uteroplacental perfusion.

CARDIOVASCULAR CHANGES

Severe disturbance(s) of normal cardiovascular function is common with preeclampsia or eclampsia. These basically are related to increased cardiac afterload caused by hypertension, cardiac preload which is substantively affected by pathologically diminished hypervolemia of pregnancy or iatrogenically increased by intravenous crystalloid or oncotic solutions, and endothelial activation with extravasation into the extracellular space, especially the lung. These interactions are discussed further in Chapter 43 (p. 1164).

HEMODYNAMIC CHANGES. The cardiovascular changes due to preeclampsia have been studied using invasive hemodynamic monitoring. Once preeclampsia has become clinically evident, however, such invasive hemodynamic studies are unlikely to provide meaningful information about the nature of the disease in earlier pregnancy. Bosio and colleagues (1999) used noninvasive Doppler hemodynamic monitoring in a longitudinal study commencing early in pregnancy in 400 nulliparous women. Gestational hypertension developed in 24 women, and 20 developed preeclampsia. Compared with normotensive women, those who developed preeclampsia had significantly elevated cardiac outputs before clinical diagnosis, but total peripheral resistance was not significantly different during this preclinical phase. With clinical preeclampsia, there was a marked reduction in cardiac output and increased peripheral resistance. By contrast, women with gestational hypertension had significantly elevated cardiac outputs before and during the development of clinical hypertension.

Much of the data based upon invasive hemodynamic studies (Table 24-5) are confounded because (1) women with preeclampsia often have different severity and duration of disease, (2) underlying disease may modify the clinical presentation, or (3) therapeutic interventions may significantly alter these findings. Variables that define cardiovascular status range from high cardiac output with low vascular resistance to low cardiac output with high vascular resistance. Similarly, left ventricular filling pressures, estimated by pulmonary capillary wedge pressure determination, range from low to pathologically high. At least three factors may explain these differences:

1. Women with preeclampsia might present with a spectrum of cardiovascular findings dependent upon both severity and duration.

2. Chronic underlying disease may modify the clinical presentation.

3. Therapeutic interventions may significantly alter these findings.

It is likely that more than one of these is operative.

The studies listed in Table 24-5 are separated into three groups based on clinical management prior to initial hemodynamic observations:

1. No therapy for preeclampsia.

2. Magnesium sulfate and hydralazine without large volumes of intravenous fluid.

3. Magnesium sulfate and hydralazine plus intravenous volume loading.

Ventricular function from the studies listed in Table 24-5 is plotted in Figure 24-1. Cardiac function was hyperdynamic in all women, but filling pressures varied markedly.

Hemodynamic data obtained prior to active treatment of preeclampsia (Table 24-5) identified normal left ventricular filling pressures, high systemic vascular resistances, and hyperdynamic ventricular function. Benedetti (1980a), Hankins (1984), and their associates reported similar findings in women with severe preeclampsia or eclampsia who were being treated with magnesium sulfate, hydralazine, and intravenous crystalloid given at 75 to 100 mL/hour. Cardiac function in these women was appropriate, and the lower systemic vascular resistance was most likely due to hydralazine treatment.

Women similarly treated with magnesium sulfate and hydralazine plus aggressive intravenous therapy or volume expansion had the lowest systemic vascular resistances and highest cardiac outputs. A comparison of volume restriction with aggressive hydration shows hyperdynamic ventricular function in most women in both groups, and two responses with respect to left ventricular stroke work index and pulmonary capillary wedge pressure (Fig. 24-2). Fluid restriction resulted in wedge pressures of less than 10 mm Hg, and most were less than 5 mm Hg. Thus, hyperdynamic ventricular function was largely a result of low wedge pressures and not a result of augmented left ventricular stroke work index, which more directly measures myocardial contractility. By comparison, women given appreciably larger volumes of fluid commonly had pulmonary capillary wedge pressures that exceeded normal; however, ventricular function remained hyperdynamic because of increased cardiac output. Subsequently, Visser and Wallenburg (1995) reported findings from 87 women with severe preeclampsia or eclampsia and described high systemic vascular resistance and hyperdynamic ventricular function in most.

From these studies, it is reasonable to conclude that aggressive fluid administration given to women with severe preeclampsia causes normal left-sided filling pressures to become substantively elevated, while increasing an already normal cardiac output to supranormal levels.

BLOOD VOLUME. It has been known for over 75 years that hemoconcentration is a hallmark of eclampsia. Pritchard and co-workers (1984) reported that in eclamptic women the normally expected hypervolemia is usually absent (Table 24-6). Women of average size should have a blood volume of nearly 5000 mL during the last several weeks of a normal pregnancy, compared with about 3500 mL when nonpregnant. With eclampsia, however, much or all of the anticipated 1500 mL of blood normally present late in pregnancy is absent. The virtual absence of an expanded blood volume is likely the consequence of generalized vasoconstriction made worse by increased vascular permeability. In women with preeclampsia, these differences are not as marked, and women with gestational hypertension usually have a normal blood volume (Silver and Seebeck, 1996; Silver and colleagues, 1998). Silver and associates (2001) presented preliminary data that blood volume is decreased in women homozygous for the T235 angiotensinogen genotype associated with preeclampsia. An acute fall in hematocrit is more likely the consequence of blood loss at delivery in the absence of normal pregnancy hypervolemia; or occasionally it is the result of intense erythrocyte destruction, as described next.

In the absence of hemorrhage, the intravascular compartment in eclamptic women is usually not underfilled. Vasospasm has contracted the space to be filled and the reduction persists until after delivery when the vascular system typically dilates, blood volume increases, and hematocrit falls. The woman with eclampsia, therefore, is unduly sensitive to vigorous fluid therapy administered in an attempt to expand the contracted blood volume to normal pregnancy levels. She is sensitive as well to even normal blood loss at delivery. Management of blood loss in these circumstances is considered in Chapter 25.

HEMATOLOGICAL CHANGES

Hematological abnormalities develop in some, but certainly not all, women who develop hypertensive disorders due to pregnancy. Among these are thrombocytopenia, which at times may become so severe as to be life threatening; the level of some plasma clotting factors may be decreased; and erythrocytes may be so traumatized that they display bizarre shapes and undergo rapid hemolysis.

COAGULATION. Subtle changes consistent with intravascular coagulation, and less often erythrocyte destruction, commonly are found with preeclampsia and especially eclampsia (Baker and Cunningham, 1999). Since the early description by Pritchard and co-workers (1954) of an eclamptic coagulopathy, we have found little evidence that it is clinically significant. As presented in Table 24-7, thrombocytopenia, infrequently severe, was the most common finding. Unless some degree of placental abruption develops, plasma fibrinogen does not differ remarkably from levels found late in normal pregnancy and fibrin degradation products were elevated only occasionally. Barron and colleagues (1999) found routine laboratory evaluation of coagulation, including prothrombin time, activated partial thromboplastin time, and plasma fibrinogen level, to be unnecessary in the management of pregnancy-associated hypertensive disorders.

The thrombin time is somewhat prolonged in a third of the cases of eclampsia even when elevated levels of fibrin degradation products are not identified. The reason for this elevation is not known, but it has been attributed to hepatic derangements discussed subsequently (Leduc and associates, 1992). The coagulation changes just described are also identified in women with severe preeclampsia, but are certainly no more common. These observations in eclampsia are most consistent with the concept that coagulation changes are the consequence of preeclampsia-eclampsia, rather than the cause.

THROMBOCYTOPENIA. Maternal thrombocytopenia can be induced acutely by preeclampsia-eclampsia. After delivery, the platelet count begins to increase progressively to reach a normal level within 3 to 5 days. The frequency and intensity of maternal thrombocytopenia vary in different studies, apparently dependent upon the intensity of the disease process, the length of delay between the onset of preeclampsia and delivery, and the frequency with which platelet counts are performed. Overt thrombocytopenia, defined by a platelet count less than 100,000/uL, indicates severe disease. In most cases, delivery is indicated because the platelet count continues to decrease.

The cause of thrombocytopenia likely results from platelet activation and consumption at the same time that platelet production is increased. Thrombopoietin, a cytokine that promotes proliferation of platelets from existing megakaryocytes, is increased in preeclamptic women with thrombocytopenia (Frolich and associates, 1998). In most studies, platelet aggregation is decreased compared with the normal increase seen in pregnancy (Baker and Cunningham, 1999). This likely is due to platelet "exhaustion" following in vivo activation. While the cause(s) is unknown, immunological processes or simply platelet deposition at sites of endothelial damage may be implicated (Pritchard and colleagues, 1976). Samuels and colleagues (1987) performed direct and indirect antiglobulin tests and found that platelet-bound and circulating platelet-bindable immunoglobulin were increased in preeclamptic women and their neonates. They interpreted these findings to suggest platelet surface alterations.

The clinical significance of thrombocytopenia, in addition to the obvious impairment in coagulation, is that it reflects the severity of the pathological process. In general, the lower the platelet count, the greater are maternal and fetal morbidity and mortality (Leduc and co-workers, 1992). The addition of elevated liver enzymes to this clinical picture is even more ominous. Weinstein (1982) referred to this combination of events as the HELLP¹ syndrome—that is, hemolysis (H), elevated liver enzymes (EL), and low platelets (LP) (see "Fragmentation Hemolysis" below and "HELLP Syndrome" on p. 579).

NEONATAL THROMBOCYTOPENIA. Thiagarajah and co-workers (1984) and Weinstein (1985) reported thrombocytopenia in neonates whose mothers had preeclampsia. Conversely, Pritchard and colleagues (1987), in a large clinical study, did not observe severe thrombocytopenia in the fetus or infant at or very soon after delivery. In fact, no cases of fetal or neonatal thrombocytopenia were identified, despite severe maternal thrombocytopenia. Thrombocytopenia did develop later in some of these infants after hypoxia, acidosis, and sepsis developed. Hence, maternal thrombocytopenia in hypertensive women is not a fetal indication for cesarean delivery.

FRAGMENTATION HEMOLYSIS. Thrombocytopenia that accompanies severe preeclampsia and eclampsia may be accompanied by evidence of erythrocyte destruction characterized by hemolysis, schizocytosis, spherocytosis, reticulocytosis, hemoglobinuria, and occasionally hemoglobinemia (Pritchard and colleagues, 1954, 1976). These derangements result in part from microangiopathic hemolysis, and human and animal studies are suggestive that intense vasospasm causes endothelial disruption, with platelet adherence and fibrin deposition. Cunningham and associates (1985) described erythrocyte morphological characteristics using scanning electron microscopy. Women with eclampsia, and to a lesser degree those with severe preeclampsia, demonstrated schizocytosis and echinocytosis but not spherocytosis when compared with normally pregnant women. Sanchez-Ramos and colleagues (1994a) described increased erythrocyte membrane fluidity in women with HELLP¹ syndrome and postulated that these changes predispose to hemolysis. Grisaru and associates (1997) have shown that erythrocytic membrane changes may facilitate the hypercoagulable state.

OTHER CLOTTING FACTORS. A severe deficiency of any of the soluble coagulation factors is very uncommon in severe preeclampsia-eclampsia unless another event coexists that predisposes to consumptive coagulopathy, such as placental abruption or profound hemorrhage due to hepatic infarction.

Antithrombin III has been reported to be lower in women with preeclampsia compared with normally pregnant women and those with chronic hypertension (Chang and co-workers, 1992). Unfortunately, early hope that antithrombin III levels could be used to predict the future development of preeclampsia and separate chronic hypertensive women from those with preeclampsia has not proven to be true (Sen and colleagues, 1994). Fibronectin, a glycoprotein associated with vascular endothelial cell basement membrane, is elevated in women with preeclampsia (Brubaker and colleagues, 1992). This observation is consistent with the view that preeclampsia causes vascular endothelial injury with subsequent hematological aberrations.

A number of clotting factor deficiencies or mutations lead to hypercoagulability that may be associated with early-onset preeclampsia. These are termed thrombophilias and are discussed in Chapter 49 (p. 1330).

ENDOCRINE AND METABOLIC CHANGES

ENDOCRINE CHANGES. Plasma levels of renin, angiotensin II, and aldosterone are increased during normal pregnancy. Hypertensive disorders due to pregnancy result in a decrease of these values toward the normal nonpregnant range (Weir and colleagues, 1973). With sodium retention, hypertension, or both, renin secretion by the juxtaglomerular apparatus decreases. Because renin catalyzes the conversion of angiotensinogen to angiotensin I (which is then transformed into angiotensin II by converting enzyme), angiotensin II levels decline, resulting in a decrease in aldosterone secretion. Despite this, women with preeclampsia avidly retain infused sodium (Brown and colleagues, 1988b).

Another potent mineralocorticoid, deoxycorticosterone (DOC), is increased strikingly in third-trimester plasma (Chap. 8, p. 194). This does not result from increased maternal adrenal secretion, but from conversion from plasma progesterone. Thus, it is not reduced by sodium retention or hypertension, and it may serve to explain why women with preeclampsia retain sodium.

Vasopressin levels are normal despite decreased plasma osmolality (Durr and Lindheimer, 1999). As discussed in Chapter 8 (p. 177), atrial natriuretic peptide increases slightly during normal pregnancy. This peptide is released upon atrial wall stretching that results from blood volume expansion. It is vasoactive and promotes sodium and water excretion likely by inhibiting aldosterone, renin activity, angiotensin II, and vasopressin. Atrial natriuretic peptide is further increased in women with preeclampsia (Gallery and Lindheimer, 1999). Increases in atrial natriuretic peptide following volume expansion result in comparable increases in cardiac output and decreases in peripheral vascular resistance in both normotensive and preeclamptic women (Nisell and associates, 1992). This observation may in part explain observations of a fall in peripheral vascular resistance following volume expansion in preeclamptic women.

FLUID AND ELECTROLYTE CHANGES. Commonly, the volume of extracellular fluid, manifest as edema, in women with severe preeclampsia-eclampsia has expanded beyond the normally increased volume that characterizes pregnancy. The mechanism responsible for the pathological expansion is not clear. Women with endothelial injury—manifest by significant proteinuria—have reduced plasma oncotic pressure which creates a filtration imbalance, displacing intravascular fluid into the surrounding interstitium.

Electrolyte concentrations do not differ appreciably in women with preeclampsia compared with those of normal pregnancy unless there has been vigorous diuretic therapy, sodium restriction, or administration of water with sufficient oxytocin to produce antidiuresis. Edema does not ensure a poor prognosis, and absence of edema does not ensure a favorable outcome.

Following an eclamptic convulsion, the bicarbonate concentration is lowered due to lactic acid acidosis and compensatory respiratory loss of carbon dioxide. The intensity of acidosis relates to the amount of lactic acid produced and its metabolic rate, as well as the rate at which carbon dioxide is exhaled.

KIDNEY

During normal pregnancy, renal blood flow and glomerular filtration rate are increased appreciably (Chap. 8, p. 186). With development of preeclampsia, renal perfusion and glomerular filtration are reduced. Levels that are much below normal nonpregnant values are the consequence of severe disease. Plasma uric acid concentration is typically elevated, especially in women with more severe disease. The elevation exceeds the reduction in glomerular filtration rate and creatinine clearance that accompanies preeclampsia (Chesley and Williams, 1945).

In the majority of preeclamptic women, mild to moderately diminished glomerular filtration appears to result from a reduced plasma volume resulting in plasma creatinine values approximately twice those expected for normal pregnancy of about 0.5 mg/dL. In some cases of severe preeclampsia, however, renal involvement is profound, and plasma creatinine may be elevated several times over nonpregnant normal values or up to 2 to 3 mg/dL. This is likely due to intrinsic renal changes caused by severe vasospasm (Pritchard and colleagues, 1984). Lee and associates (1987) reported normal ventricular filling pressures in seven severely preeclamptic women with oliguria, and concluded that this was consistent with intrarenal vasospasm. In most, urine sodium concentration was elevated abnormally, also suggesting an intrinsic renal etiology. Urine osmolality, urine:plasma creatinine ratio, and fractional excretion of sodium were also indicative that a prerenal mechanism was involved. Importantly, intensive intravenous fluid therapy was not indicated for these women with oliguria. When dopamine was infused into oliguric preeclamptic women, this renal vasodilator caused increased urine output, fractional sodium excretion, and free water clearance (Kirshon and co-workers, 1988).

Taufield and associates (1987) reported that preeclampsia is associated with diminished urinary excretion of calcium because of increased tubular reabsorption. This mechanism would explain the decreased calcium excretion in hypertensive pregnant women.

After delivery, in the absence of underlying chronic renovascular disease, complete recovery of renal function usually can be anticipated. This would not be the case, of course, if renal cortical necrosis, an irreversible but rare lesion, develops (Sibai and associates, 1990; also see Chap. 47, p. 1266).

PROTEINURIA. There should be some degree of proteinuria to establish the diagnosis of preeclampsia-eclampsia. Because proteinuria develops late, however, some women may be delivered before it appears. Meyer and colleagues (1994) emphasized that 24-hour urine excretion should be measured. They found that a urinary dipstick of 1+ proteinuria or greater was predictive of at least 300 mg per 24 hours in 92 percent of cases. Conversely, trace or negative proteinuria had a negative predictive value of only 34 percent in hypertensive women. Urine dipstick values of 3+ to 4+ were positively predictive of severe preeclampsia in only 36 percent of cases.

Albuminuria is an incorrect term to describe proteinuria of preeclampsia. As with any other glomerulopathy, there is increased permeability to most large-molecular-weight proteins; thus, abnormal albumin excretion is accompanied by other proteins, such as hemoglobin, globulins, and transferrin. Normally, these large protein molecules are not filtered by the glomerulus, and their appearance in urine signifies a glomerulopathic process. Some of the smaller proteins that usually are filtered but reabsorbed are also detected in urine.

ANATOMICAL CHANGES. Changes identifiable by light and electron microscopy are commonly found in the kidney. Sheehan (1950) observed that the glomeruli were enlarged by about 20 percent. The capillary loops variably are dilated and contracted. The endothelial cells are swollen, and deposited within and beneath them are fibrils that have been mistaken for thickening of the basement membrane.

Most electron microscopy studies of renal biopsies are consistent with glomerular capillary endothelial swelling. These changes, accompanied by subendothelial deposits of protein material, were called glomerular capillary endotheliosis by Spargo and associates (1959). The endothelial cells are often so swollen that they block or partially block the capillary lumens. Homogeneous deposits of an electron-dense substance are found between basal lamina and endothelial cells and within the cells themselves. On the basis of immunofluorescent staining, Lichtig and co-workers (1975) identified deposited fibrinogen or its derivatives in 13 of 30 renal biopsy specimens from women with preeclampsia. Kincaid-Smith (1991) found that these deposits disappear progressively in the first week postpartum.

Renal tubular lesions are common in women with eclampsia, but what has been interpreted as degenerative changes may represent only an accumulation within cells of protein reabsorbed from the glomerular filtrate. The collecting tubules may appear obstructed by casts from derivatives of protein, including, at times, hemoglobin.

Acute renal failure from tubular necrosis may develop. Such kidney failure is characterized by oliguria or anuria and rapidly developing azotemia (approximately 1 mg/dL increase in serum creatinine per day). Although this is more common in neglected cases, it is invariably induced by hypovolemic shock, usually associated with hemorrhage at delivery, for which adequate blood replacement is not given (Chap. 47, p. 1266). Haddad and colleagues (2000) reported that 5 percent of 183 women with hemolysis, elevated liver enzymes, and thrombocytopenia—HELLP¹ syndrome—developed acute renal failure. Moreover, half of these also had a placental abruption, and most had postpartum hemorrhage. Rarely, renal cortical necrosis develops when the major portion of the cortex of both kidneys undergoes necrosis. Renal cortical necrosis is irreversible, and although it develops in nonpregnant women and in men, the lesion has most often been associated with pregnancy.

LIVER

With severe preeclampsia, at times there are alterations in tests of hepatic function and integrity, including delayed excretion of bromosulfophthalein and elevation of serum aspartate amniotransferase levels (Combes and Adams, 1972). Severe hyperbilirubinemia is uncommon even with severe preeclampsia (Pritchard and colleagues, 1976). Much of the increase in serum alkaline phosphatase is due to heat-stable alkaline phosphatase of placental origin. Oosterhof and co-workers (1994) described increased hepatic artery resistance using Doppler sonography in 37 women with preeclampsia.

Periportal hemorrhagic necrosis in the periphery of the liver lobule is the most likely reason for increased serum liver enzymes (Fig. 24-3). Such extensive lesions are seldom identified in nonfatal cases with liver biopsy (Barton and colleagues, 1992). Bleeding from these lesions may cause hepatic rupture, or they may extend beneath the hepatic capsule and form a subcapsular hematoma. Such hemorrhages without rupture may be more common than previously suspected. Using computed tomography, Manas and colleagues (1985) showed that five of seven women with preeclampsia and upper abdominal pain had hepatic hemorrhage (Fig. 24-4). Prompt surgical intervention may be life saving. Rinehart and co-workers (1999) reviewed 121 cases of spontaneous hepatic rupture associated with preeclampsia, and the mortality rate was 30 percent. One woman at Parkland Hospital survived hepatic rupture after receiving blood and blood products from more than 200 donors. Hunter and co-workers (1995) described a similar women in whom liver transplant was considered life saving.

HELLP¹ SYNDROME. Liver involvement in preeclampsia-eclampsia is serious and is frequently accompanied by evidence of other organ involvement, especially the kidney and brain, along with hemolysis and thrombocytopenia (De Boer and co-workers, 1991; Pritchard and associates, 1954; Weinstein, 1985). This is commonly referred to as HELLP¹ syndrome—Hemolysis, ELevated liver enzymes, and Low Platelets. The Memphis group identified this constellation in almost 20 percent of women with severe preeclampsia or eclampsia (Sibai and co-workers, 1993b). Five of the 437 women died. In 13 of 33 women (40 percent) with laboratory evidence of HELLP syndrome and severe right upper quadrant pain, they found evidence of a subcapsular hematoma using imaging studies (Barton and Sibai, 1996). From the same group, Audibert and associates (1996) cited other complications, including placental abruption (7 percent), acute renal failure (2 percent), pulmonary edema (6 percent), and subcapsular liver hematoma (1 percent). Isler and co-authors (1999) have identified factors contributing to the death of 54 women with the HELLP syndrome, and these are shown in Figure 24-5.

Adverse outcomes in subsequent pregnancies are increased in women with HELLP¹ syndrome. Sibai and colleagues (1995) observed a 3 percent incidence of recurrence of HELLP syndrome in 192 subsequent pregnancies, and Sullivan and associates (1994) found this to be 27 percent. Both groups confirmed a high incidence of recurrent preeclampsia, preterm delivery, fetal growth restriction, placental abruption, and cesarean delivery.

BRAIN

Central nervous system manifestations of preeclampsia, and especially the convulsions of eclampsia, have been long known. In particular, visual symptoms have received much attention. The earliest description of brain involvement came from gross and histological examination, but with modern noninvasive techniques, imaging and Doppler studies have added new insight into cerebrovascular involvement.

ANATOMICAL PATHOLOGY. Two distinct but related types of cerebral pathology include gross hemorrhages due to ruptured arteries caused by severe hypertension. These can be seen in any woman with gestational hypertension, and preeclampsia is not necessary for their development. These complications are more common with underlying chronic hypertension (Chap. 45).

The other lesions, variably demonstrated with preeclampsia, but likely universal with eclampsia, are more widespread and seldom fatal. The principal postmortem cerebral lesions are edema, hyperemia, focal anemia, thrombosis, and hemorrhage. Sheehan (1950) examined the brains of 48 eclamptic women very soon after death, and hemorrhages, ranging from petechiae to gross bleeding, were found in 56 percent. According to Sheehan, if the brain is examined within an hour after death, most often it is as firm as normal, and there is no obvious edema. This correlates with our findings that only 10 of 175 eclamptic women had evidence for cerebral edema (Cunningham and Twickler, 2000).

In another anatomical study, Govan (1961) concluded that cerebral hemorrhage was the cause of death in 39 of 110 fatal cases of eclampsia. In 40 of 47 women who died of cardiorespiratory failure, small cerebral hemorrhagic lesions were also found. A regular finding was fibrinoid changes in the walls of cerebral vessels. The lesions sometimes appeared to have been present for some time, as judged from the surrounding leukocytic response and hemosiderin-pigmented macrophages. These findings are consistent with the view that prodromal neurological symptoms and convulsions may be related to these lesions.

NEUROIMAGING STUDIES. The earliest abnormal imaging studies emerged with the use of computed tomographic scanning. In the earliest report from Parkland Hospital, Brown and colleagues (1988a) found that nearly half of eclamptic women studied had abnormal radiological findings. The most common were hypodense areas in the cerebral cortex, which corresponded to the petechial hemorrhages and infarction sites described at autopsy by Sheehan and Lynch (1973). While providing valuable insight to their number and locations, these studies did not answer the question concerning the cause of these localized areas of edema. It was still unknown if they were due to ischemic necrosis or hyperperfusion. The advent of magnetic resonance imaging allowed better resolution, but again the basic cause was not elucidated. For example, in another study from Parkland Hospital, Morriss and colleagues (1997) confirmed remarkable changes, especially in the area of the posterior cerebral artery.

These findings help to provide an explanation of why some women with preeclampsia convulse but others do not. The brain, like the liver and kidney, appears to be more involved in some women than in others. The extent and location of ischemic and petechial subcortical lesions likely influences the incidence of eclampsia. Their extent also explains more worrisome neurological complications such as blindness or coma. Recent findings indicate that these symptoms represent a continuum of involvement.

BLINDNESS. Although visual disturbances are common with severe preeclampsia, blindness, either alone or accompanying convulsions, is not. Most women with varying degrees of amaurosis are found to have radiographic evidence of extensive occipital lobe hypodensities. An example is shown in Figure 24-6 and is likely an exaggeration of the lesions described earlier. Over a 14-year period, we described 15 women with severe preeclampsia or eclampsia who also had blindness (Cunningham and associates, 1995). This persisted for 4 hours to 8 days, but in all it resolved completely.

Retinal artery vasospasm may also be associated with visual disturbances (Ohno and colleagues, 1999). Fortuitously, Belfort and associates (1992) showed that a 6-g bolus of magnesium sulfate caused retinal artery vasodilation. Retinal detachment may also cause altered vision, although it is usually one sided and seldom causes total visual loss as in some women with cortical blindness. Surgical treatment is seldom indicated; prognosis is good, and vision usually returns to normal within a week.

CEREBRAL EDEMA. Central nervous system manifestations from more widespread cerebral edema are worrisome. In some cases, obtundation and confusion are major features, and symptoms wax and wane. In a few cases, overt coma develops. Prognosis for these latter women is guarded and brainstem herniation is a serious complication.

During a 13-year period, we identified 10 of 175 eclamptic women at Parkland Hospital who had symptomatic cerebral edema (Cunningham and Twickler, 2000). Their symptoms ranged from lethargy, confusion, and blurred vision to obtundation and coma. Mental status changes correlated with the degree of involvement seen with computed tomographic and magnetic resonance imaging studies (Fig. 24-7). Three women with generalized cerebral edema were comatose and had impending transtentorial herniation on imaging studies and one of these died from herniation. It seems reasonable that this degree of involvement is related to both ischemic (cytotoxic) as well as hyperperfusion (vasogenic) edema. Conversely, Apollon and co-workers (2000) used single-photon-emission computed tomography and provided evidence for hyperperfusion and vasogenic (hydrostatic) edema.

CEREBRAL BLOOD FLOW. As discussed, it is not known precisely what effects preeclampsia or eclampsia have on cerebral blood flow. Williams and Wilson (1999) used transcranial Doppler ultrasonography to study cerebral blood flow in six women with severe preeclampsia and in three women with eclampsia. Preeclampsia was associated with increased cerebral perfusion pressure counterbalanced by increased cerebrovascular resistance with no net change in cerebral blood flow. In eclampsia, and presumably due to loss of autoregulation of cerebral blood flow manifest as decreased vascular resistance, there was cerebral hyperperfusion similar to that seen in hypertensive encephalopathy unrelated to pregnancy. Belfort and colleagues (1999) also used transcranial Doppler ultrasound to estimate cerebral perfusion pressure in the middle cerebral artery in 79 preeclamptic women with and without headache. Women with headaches were more likely to have abnormal cerebral perfusion (either increased or decreased) than those without headaches. Those women with severe headaches tended to have high cerebral perfusion. Another important finding was that cerebral perfusion pressure may be normal in one hemisphere and very disordered in the other.

This evidence suggests that women with preeclampsia have cerebral vasospasm characterized by high or low cerebral perfusion pressure which varies from one hemisphere to the other. Women who develop eclampsia, however, ostensibly have suffered a transient loss of cerebral vascular autoregulation. This conclusion is also supported by evidence of widespread low-density areas confirmed by computed tomographic and magnetic resonance imaging (Cunningham and Twickler, 2000). According to Apollon and co-workers (2000), hyperperfusion likely causes vasogenic edema. Brackley and colleagues (2000) attributed cerebral vasospasm in preeclamptic women to increased cerebral arterial wall stiffness and vasoconstriction.

ELECTROENCEPHALOGRAPHY. Nonspecific electroencephalographic abnormalities can usually be demonstrated for some time after eclamptic convulsions. Sibai and colleagues (1985a) observed that 75 percent of 65 eclamptic women had abnormal electroencephalograms within 48 hours of seizures. Half of these abnormalities persisted past 1 week, but most were normal by 3 months. An increased incidence of electroencephalographic abnormalities has been described in family members of eclamptic women, a finding suggestive that some eclamptic women have an inherited predisposition to convulse (Rosenbaum and Maltby, 1943).

UTEROPLACENTAL PERFUSION

Compromised placental perfusion from vasospasm is almost certainly a major culprit in the genesis of increased perinatal morbidity and mortality associated with preeclampsia. For example, Brosens and associates (1972) reported that the mean diameter of myometrial spiral arterioles of 50 normal pregnant women was 500 um. The same measurement in 36 women with preeclampsia was 200 um. Attempts to measure human maternal and placental blood flow have been hampered by several obstacles, including inaccessibility of the placenta, the complexity of its venous effluent, and the unsuitability of certain investigative techniques for humans.

INDIRECT METHODS. Everett and colleagues (1980) presented evidence that the clearance rate of dehydroisoandrosterone sulfate through placental conversion to estradiol-17b was an accurate reflection of maternal placental perfusion. Fritz and colleagues (1985) reported that the technique paralleled uteroplacental perfusion in primates. Normally, as pregnancy advances, this measurement increases greatly. The placental clearance rate decreases before the onset of overt hypertension (Worley and associates, 1975). Finally, placental clearance is decreased in women given diuretics or hydralazine (Gant and co-workers, 1976).

DOPPLER VELOCIMETRY. Doppler measurement of blood velocity through uterine arteries has been used to estimate uteroplacental blood flow (Chap. 41, p. 1133). Vascular resistance is estimated by comparing arterial systolic and diastolic velocity waveforms. Ducey and associates (1987) described systolic-diastolic velocity ratios from both uterine and umbilical arteries in 136 pregnancies complicated by hypertension. Among 51 women considered to have preeclampsia, 20 percent had normal umbilical artery velocity ratios; 15 percent had normal umbilical but abnormal uterine artery ratios; and in 40 percent both ratios were abnormal. Fleischer and colleagues (1986) and Trudinger and associates (1990) have also reported increased systolic-diastolic ratios in uterine arteries of women with preeclampsia. Others have not confirmed this (Hanretty and colleagues, 1988). These studies, in the aggregate, can be interpreted to imply that only a few women with preeclampsia have compromised uteroplacental circulations.

HISTOLOGICAL CHANGES IN THE PLACENTAL BED. Hertig (1945) identified in preeclamptic pregnancies a lesion of uteroplacental arteries characterized by prominent lipid-rich foam cells. Zeek and Assali (1950) termed this acute atherosis (Fig. 24-8). Most investigators are now in accord that there is a lesion, but they do not agree on its precise nature. Classically, in normal pregnancy, spiral arteries are invaded by endovascular trophoblast (Fig. 24-9). It seems that, in preeclampsia, decidual vessels, but not myometrial vessels, are invaded by endovascular trophoblasts. Using electron microscopy studies of arteries taken from the uteroplacental implantation site, De Wolf and co-workers (1980) reported that early preeclamptic changes included endothelial damage, insudation of plasma constituents into vessel walls, proliferation of myointimal cells, and medial necrosis. They also found that lipid accumulates first in myointimal cells and then in macrophages. Madazli and colleagues (2000) showed that the magnitude of defective trophoblastic invasion of the spiral arteries correlated with the severity of the hypertensive disorder.

PATHOPHYSIOLOGY

In 1903, and in some ways remarkably prescient as will be surmised in reading this section, J. Whitridge Williams reported on the pathophysiology of preeclampsia as follows:

The present status of the question may therefore be summarized as follows: The clinical history and anatomical findings afford presumptive evidence that the disease is due to the circulation of some poisonous substance in the blood which gives rise to thrombosis in many of the smaller vessels, with consequent degenerative necrosis in the various organs.

Any satisfactory theory on the pathophysiology of preeclampsia must account for the observation that hypertensive disorders due to pregnancy are very much more likely to develop in the woman who:

1. Is exposed to chorionic villi for the first time.

2. Is exposed to a superabundance of chorionic villi, as with twins or hydatidiform mole.

3. Has preexisting vascular disease.

4. Is genetically predisposed to hypertension developing during pregnancy.

Although chorionic villi are essential, they need not support a fetus or be located within the uterus. An overview of the pathophysiology to be considered in this section is shown in Figure 24-10.

Vasospasm is basic to the pathophysiology of preeclampsia-eclampsia. This concept, first advanced by Volhard (1918), is based upon direct observations of small blood vessels in the nail beds, ocular fundi, and bulbar conjunctivae, and it has been surmised from histological changes seen in various affected organs (Hinselmann, 1924; Landesman and co-workers, 1954). Vascular constriction causes resistance to blood flow and accounts for the development of arterial hypertension. It is likely that vasospasm itself also exerts a damaging effect on vessels. Moreover, angiotensin II causes endothelial cells to contract. These changes likely lead to endothelial cell damage and interendothelial cell leaks through which blood constituents, including platelets and fibrinogen, are deposited subendothelially (Brunner and Gavras, 1975). These vascular changes, together with local hypoxia of the surrounding tissues, presumably lead to hemorrhage, necrosis, and other end-organ disturbances that have been observed at times with severe preeclampsia. With this scheme, fibrin deposition is then likely to be prominent, as seen in fatal cases (McKay, 1965).

INCREASED PRESSOR RESPONSES. Normally pregnant women develop refractoriness to infused vasopressors (Abdul-Karim and Assali, 1961). Increased vascular reactivity to pressors in women with early preeclampsia has been identified by Raab and co-workers (1956) and Talledo and associates (1968) using either norepinephrine or angiotensin II, and by Dieckmann and Michel (1937) and Browne (1946) using vasopressin.

Gant and co-workers (1973) demonstrated that increased vascular sensitivity to angiotensin II clearly preceded the onset of pregnancy-induced hypertension. As shown in Figure 24-11, nulliparas who remained normotensive were refractory to the pressor effect of infused angiotensin II, while women who subsequently became hypertensive lost this refractoriness weeks before the onset of hypertension. Of women who required more than 8 ng/kg/min of angiotensin II to provoke a standardized pressor response between 28 and 32 weeks, 90 percent remained normotensive throughout pregnancy. Conversely, among normotensive nulliparas who required less than 8 ng/kg/min at 28 to 32 weeks, 90 percent subsequently developed overt hypertension. Women with underlying chronic hypertension have almost identical responses (Gant and colleagues, 1977).

PROSTAGLANDINS. Based on the findings of a number of studies, it has been concluded that the blunted pressor response described earlier is due principally to decreased vascular responsiveness mediated in part by vascular endothelial synthesis of prostaglandins or prostaglandin-like substances (Cunningham and associates, 1975; Gant and co-workers, 1974a). For example, refractoriness to angiotensin II in pregnant women is abolished by large doses of the prostaglandin synthase inhibitors (Everett and colleagues, 1978).

The exact mechanism by which prostaglandin(s) or related substances mediate vascular reactivity during pregnancy is unknown. From a number of observations, there is evidence that compared with normal pregnancy, prostacyclin production is decreased significantly and thromboxane A₂ significantly increased in preeclampsia (Walsh, 1985). Thus, in preeclamptic women, thromboxane is increased and prostacyclin and prostaglandin E₂ are decreased, resulting in vasoconstriction and sensitivity to infused angiotensin II.

Spitz and colleagues (1988) reported that 81 mg of aspirin given daily to future hypertensive women restored angiotensin II refractoriness by suppressing synthesis of thromboxane A₂ by about 75 percent; however, prostacyclin synthesis was decreased by only 20 percent and prostaglandin E₂ by 30 percent. These observations indicate that vessel reactivity may be mediated through a delicate balance of production and metabolism of these vasoactive prostaglandins. In this scheme, preeclampsia may follow inappropriately increased production or destruction of one prostaglandin, diminished synthesis or release of the other, or perhaps both. Unfortunately, these observations were not beneficial in clinical studies discussed subsequently.

NITRIC OXIDE. Previously termed endothelium-derived relaxing factor (EDRF), nitric oxide is synthesized by endothelial cells from L-arginine (Palmer and associates, 1988). It is a potent vasodilator whose absence or decreased concentration might play a role in the etiology of hypertensive disorders due to pregnancy. Its production appears to be increased in severe preeclampsia (Benedetto and associates, 2000). Withdrawal of nitric oxide from some pregnant animals results in the development of a clinical picture similar to preeclampsia (Conrad and Vernier, 1989; Weiner and associates, 1989). Inhibition of nitric oxide has been shown to increase mean arterial pressure, decrease heart rate, and reverse the pregnancy-induced refractoriness to vasopressors in some animals. Equally important, it appears to maintain the normal low-pressure vasodilated state characteristic of fetoplacental perfusion in the human (Chang and colleagues, 1992; Myatt and co-workers, 1992; Weiner and associates, 1992). Decreased nitric oxide release or production has not been shown to develop prior to the onset of hypertension (Anumba and colleagues, 1999). Thus, the changes in nitric oxide concentrations in women with hypertensive disorders due to pregnancy appear to be the consequence of hypertension and not the inciting event (Morris and colleagues, 1996).

ENDOTHELINS. These polypeptides are potent vasoconstrictors, and endothelin-1 is the only species produced by human endothelium (Mastrogiannis and co-workers, 1991). Plasma endothelin-1 is increased in normotensive laboring and nonlaboring women, and even higher levels have been reported in preeclamptic women (Clark, 1992; Nova, 1991; Schiff, 1992, and their associates). Otani and colleagues (1991), however, did not observe increased plasma endothelin levels, and Barton and associates (1993) did not find increased urinary endothelin-1 levels in preeclamptic women.

VASCULAR ENDOTHELIAL GROWTH FACTOR. This is a glycosalated glycoprotein that is selectively mitogenic for endothelial cells. Vascular endothelial growth factor (VEGF) is important in vasculogenesis and control of microvascular permeability and has been identified in the human placenta. Serum levels of VEGF² increase in the first half of pregnancy concurrent with trophoblast and uterine vascular events characteristic of pregnancy. VEGF has been reported to be increased in serum from women with preeclampsia (Baker and associates, 1995). Simmons and co-workers (2000) studied uteroplacental vascular resistance, measured with Doppler ultrasound, and placental VEGF in pregnant women with and without preeclampsia. They found an increase in VEGF parallel to increased uteroplacental vessel resistance in women with preeclampsia. They concluded that increased placental VEGF may represent a compensatory mechanism attempting to restore uteroplacental blood flow toward normal.

GENETIC PREDISPOSITION. The tendency for preeclampsia-eclampsia is inherited. Chesley and Cooper (1986) studied the sisters, daughters, granddaughters, and daughters-in-law of eclamptic women delivered at the Margaret Hague Maternity Hospital from 1935 to 1984. They concluded that preeclampsia-eclampsia is highly heritable, and that the single-gene model, with a frequency of 0.25, best explained their observations. A multifactorial inheritance was also considered possible. Kilpatrick and associates (1989), but not Hayward and co-workers (1992), reported an association between the histocompatibility antigen HLA-DR4 and proteinuric hypertension. Hoff and associates (1992) concluded that a maternal humoral response directed against fetal anti-HLA-DR immunoglobulin antibody might influence the development of gestational hypertension.

Cooper and Liston (1979) examined the possibility that susceptibility to preeclampsia is dependent upon a single recessive gene. They calculated the expected first-pregnancy frequencies of daughters of women with eclampsia; daughters-in-law served as controls. The frequencies calculated by them and those actually observed by Chesley and co-workers (1968) in daughters and daughters-in-law of women with eclampsia are remarkably close. Ward and associates (1993) reported that women carrying the angiotensinogen gene variant T235 had a higher incidence of hypertensive disorders due to pregnancy. Morgan and colleagues (1995), however, could not confirm these findings. Morgan and colleagues (1999) subsequently showed that spiral arteries obtained at 8 weeks' gestation in women homozygous for the angiotensinogen gene failed to undergo the remodeling characteristic of normal implantation. They hypothesized that the failure to undergo such normal remodeling of the spiral arteries may predispose these women to preeclampsia. Inherited thrombophilias, discussed in Chapter 49, predispose some women—either heterozygotes or homozygotes depending on the factor—to these syndromes.

IMMUNOLOGICAL FACTORS. The risk of hypertensive disorders due to pregnancy is appreciably enhanced in circumstances where formation of blocking antibodies to antigenic sites on the placenta might be impaired. This may arise where effective immunization by a previous pregnancy is lacking, as in first pregnancies; or where the number of antigenic sites provided by the placenta is unusually great compared with the amount of antibody, as with multiple fetuses (Beer, 1978). Strickland and associates (1986), however, provided data that do not support "immunization" by a previous pregnancy. They analyzed the outcomes of over 29,000 pregnancies at Parkland Hospital and reported that hypertensive disorders were decreased only slightly (22 versus 25 percent) in women who previously had miscarried (and thus were "immunized") and were now having their first baby. The immunization concept is supported by the observation that preeclampsia develops more frequently in multiparous women impregnated by a new consort (Trupin and colleagues, 1996).

Dekker and Sibai (1998) have reviewed the possible role of immune maladaptation in the pathophysiology of preeclampsia. Beginning in the early second trimester, women destined to develop preeclampsia have a significantly lower proportion of T-helper cells compared with women who remain normotensive (Bardeguez and associates, 1991). Antibodies against endothelial cells have been found in 50 percent of women with preeclampsia versus 15 percent of normotensive controls (Rappaport and colleagues, 1990). Immunological pathogenesis of acute atherosis shown in Figure 24-8 has also been suggested because of the morphological similarities to lesions seen in allograft rejection (Labarrere, 1988).

INFLAMMATORY FACTORS. Redman and colleagues (1999) have proposed that the endothelial cell dysfunction associated with preeclampsia can result from a "generalized perturbation of the normal, generalized maternal intravascular inflammatory adaptation to pregnancy" (Fig. 24-10). In this hypothesis, preeclampsia is considered a disease due to an extreme state of activated leukocytes in the maternal circulation (Faas and colleagues, 2000; Gervasi and co-workers, 2001).

The decidua contains an abundance of cells that, when activated, can release noxious agents (Staff and colleagues, 1999). These then serve as mediators to provoke endothelial cell injury (Fig. 24-10). Dekker and Sibai (1998) have reviewed this aspect of the pathophysiology of hypertensive disorders due to pregnancy.

Briefly, cytokines, to include tumor necrosis factor-alpha (TNF-a) and the interleukins, may contribute to the oxidative stress associated with preeclampsia. In this scheme, oxygen-free radicals lead to the formation of self-propagating lipid peroxides that in turn propagate highly toxic radicals which, in turn, injure endothelial cells. Such injury modifies endothelial cell production of nitric oxide, as well as interfering with prostaglandin balance. Other consequences of oxidative stress include production of the lipid-laden macrophage foam cells characteristics of atherosis (Fig. 24-8), activation of microvascular coagulation (thrombocytopenia), and increased capillary permeability (edema and proteinuria). These observations on the effects of oxidative stress in preeclampsia have given rise to increased interest in the potential benefit of antioxidant therapy given for the prevention of hypertensive disorders due to pregnancy. Antioxidants are a diverse family of components that function to prevent overproduction of and damage caused by noxious free radicals. Examples of antioxidants include vitamins E (a-tocopherol) and C (ascorbic acid) and b-carotene. Dietary supplementation with these antioxidants is discussed on page 590.

ENDOTHELIAL CELL ACTIVATION. Prevailing evidence is that endothelial cell activation is the centerpiece in the contemporary understanding of the pathogenesis of preeclampsia (Fig. 24-10). In this scheme, preeclampsia is an immunologically mediated deficiency in trophoblastic invasion of spiral arteries that leads to fetoplacental hypoperfusion. This results in the release of a factor(s) into the maternal circulation. These changes in turn provoke "activation" of the vascular endothelium, with the clinical syndrome of preeclampsia resulting from widespread changes in endothelial cell function (Hayman and associates, 2000; Ness and Roberts, 1996; Roberts, 2000; Walker, 2000). Intact endothelium has anticoagulant properties and blunts the response of vascular smooth muscle to agonists. Damaged endothelium, on the other hand, activates endothelial cells to promote coagulation, and increases sensitivity to vasopressor agents.

Further evidence of endothelial activation in preeclampsia includes the characteristic changes in glomerular capillary endothelial morphology, increased capillary permeability, and elevated blood levels of substances associated with such activation (see preceding section). Serum from preeclamptic women stimulates cultured endothelial cells to produce greater amounts of prostacyclin than serum from normotensive controls. Hyperhomocysteinemia is of interest in preeclampsia because elevated levels in men and nonpregnant women are an independent risk factor for atherosclerosis, which is very similar to implantation-site atherosis (Rogers and colleagues, 1999). Cotter and colleagues (2001) have presented preliminary data that elevated serum homocysteine levels in early pregnancy increase preeclampsia risk threefold. Conversely, Laivuori and associates (2000) found no increased risk in carriers of the T677 allele for methylenetetrahydrofolate reductase.

PREDICTION AND PREVENTION

PREDICTION

A variety of biochemical and biophysical markers, based primarily on rationales implicated in the pathology and pathophysiology of hypertensive disorders due to pregnancy, have been proposed for the purpose of predicting the development of preeclampsia later in pregnancy. Investigators have attempted to identify early markers of faulty placentation, reduced placental perfusion, endothelial cell dysfunction, and activation of coagulation. Virtually all these attempts have resulted in testing strategies with low sensitivity for the prediction of preeclampsia. Friedman and Lindheimer (1999) concluded in their review that, at the present time, there are no screening tests for preeclampsia that are reliable, valid, and economical. Stamilio and colleagues (2000) reached a similar conclusion. Selected tests for prediction of preeclampsia are discussed in the following section.

ANGIOTENSIN II INFUSION. In this test, angiotensin II is infused in a stepwise fashion until there is a 20 mm Hg rise in diastolic blood pressure. Women requiring less than 8 ng/kg/min of angiotensin II had a positive-predictive value (true positive) of developing preeclampsia of 20 to 40 percent (Friedman and Lindheimer, 1999). While quite good when compared with other tests for prediction, angiotensin II infusion is difficult to perform and is therefore not used clinically.

ROLL-OVER TEST. A hypertensive response induced by having the woman assume the supine position after lying laterally recumbent was demonstrated in some pregnant women by Gant and colleagues (1974b). The majority of nulliparous women at 28 to 32 weeks who had increased diastolic pressure of at least 20 mm Hg when the maneuver ("roll-over test") was performed, later developed hypertension due to pregnancy. Conversely, most women whose blood pressure did not become elevated when this was done remained normotensive. Women who demonstrated a positive roll-over test were also abnormally sensitive to infused angiotensin II. It is hypothesized that a positive test result is a manifestation of increased vascular responsivity or sympathetic overactivity in women who will develop hypertension later in pregnancy. Using preeclampsia, rather than gestational hypertension as the end point, the positive-predictive value (true positive) was 33 percent, which is similar to that of the angiotensin infusion test (Dekker and colleagues, 1990).

URIC ACID. Elevated uric acid levels in maternal blood, presumably due to decreased renal urate excretion, are frequently found in women with preeclampsia. Jacobson and colleagues (1990) studied 135 pregnant women to determine if plasma uric acid levels, determined at 24 weeks' gestation, might predict the subsequent development of preeclampsia. Values exceeding 5.9 mg/dL were considered predictive. The positive-predictive value (true positive) for this test was 33 percent. It is unlikely that uric acid levels will prove very useful in predicting development of preeclampsia later in pregnancy given the observation that such levels have not even proven to be useful in differentiating established gestational hypertension from preeclampsia (Lim and co-authors, 1998).

CALCIUM METABOLISM. Alterations in calcium metabolism as well as deficiencies in dietary intake of calcium have been implicated in the pathophysiology of preeclampsia. Hypocalciuria has been identified with preeclampsia (Taufield and co-authors, 1987). Several investigators have thus performed studies to determine if midpregnancy urinary calcium levels might predict the development of preeclampsia. Sanchez-Ramos and colleagues (1991) measured 24-hour urinary calcium excretion in 103 nulliparous women between 10 and 24 weeks. The sensitivity in predicting preeclampsia was 88 percent and the positive-predictive value was 32 percent.

URINARY KALLIKREIN EXCRETION. Kallikrein is an important regulator of blood pressure, and it has been hypothesized its diminished excretion might precede the development of preeclampsia. Although Millar and colleagues (1996) found the predictive value of this test to be very good (sensitivity 83 percent, 91 percent positive-predictive value), others have not confirmed these findings (Kyle and associates, 1996).

FIBRONECTIN. There are many reports describing plasma fibronectin levels in women with impending or established preeclampsia. Endothelial cell injury, a centerpiece in the hypothesized pathophysiology of preeclampsia discussed previously, is the presumed source of elevated cellular fibronectin levels found in women with established hypertensive disease (Brubaker and associates, 1992). For example, Halligan and colleagues (1994) reported that fibronectin levels were elevated in the first trimester in women destined to develop preeclampsia. Paarlberg and colleagues (1998) measured plasma fibronectin levels during the second trimester in 347 healthy nulliparous women to predict hypertensive disorders due to pregnancy. The sensitivity of this test was low (69 percent) as was the positive-predictive value (12 percent).

COAGULATION ACTIVATION. Thrombocytopenia and abnormalities of platelet function (aggregation) appear to be an integral feature of preeclampsia. Excessive platelet activation has been linked to maternal vasoconstriction, endothelial cell injury, placental infarction (atherosis and fetal growth restriction), and transient renal dysfunction. Platelet activation also leads to thromboxane A₂ release, which promotes vasospasm, further platelet aggregation, and endothelial cell injury. Thromboxane from platelets increases the thromboxane/prostacyclin ratio in women with preeclampsia (Fitzgerald and colleagues, 1987). This finding formed the basis for the concept that prophylactic administration of low-dose aspirin might prevent preeclampsia (see later discussion).

Clearly, the platelet count can be substantially decreased in women with severe preeclampsia. Platelet volume increases due to platelet consumption and resultant production of relatively younger (and therefore larger) platelets. Ahmed and associates (1993) found that high platelet volumes may be a marker of impending preeclampsia. Similarly, fibrinolytic activity is normally decreased in pregnancy due to increased plasminogen activator inhibitors (PAI) 1 and 2. In preeclampsia, PAI³-1 is increased relative to PAI-2, and it may be a marker of endothelial cell dysfunction (Caron and colleagues, 1991).

MARKERS OF OXIDATIVE STRESS. Increased levels of lipid peroxides coupled with decreased activity of antioxidants in women with preeclampsia has raised the possibility that markers of oxidative stress might prove useful in the prediction of preeclampsia (Walsh, 1994). Possible markers include malondialdehyde lipid peroxidation (Hubel and co-authors, 1989); a variety of pro-oxidants or potentiators of pro-oxidants such as iron (Herbert and colleagues, 1994); homocysteine (Cotter and associates, 2001; Powers and colleagues, 2000); blood lipids to include triglycerides, free fatty acids, and lipoproteins (Hubel and colleagues, 1996); and antioxidants to include ascorbic acid and vitamin E (Mikhail and colleagues, 1994).

The decrease in prostaglandin synthesis described earlier has been reported to be detectable as early as the first trimester in women destined to develop preeclampsia. However, measuring prostacyclin metabolites in urine presents significant technical problems. Other prostaglandin isomers, for example, 8-isoprostane, a potent vasoconstrictor and a result of lipid peroxidation, has been also studied as a potential marker of impending preeclampsia (Regan and Fitzgerald, 1997).

IMMUNOLOGICAL FACTORS. Cytokines are protein messengers released by immune cells that serve to regulate the function of other immune cells and are produced by macrophages and lymphocytes at the interface of trophoblast and decidua. There are at least 50 cytokines, to include interleukins, interferons, growth factors, and tumor necrosis factors. Several of these cytokines have been found to be elevated in women with preeclampsia and are of interest as possible markers for the development of preeclampsia (Benyo, 2000; Dudley, 1996; Greer, 1994; Kupferminc, 1994, and all their colleagues).

PLACENTAL PEPTIDES. A variety of peptides produced by the placenta are of interest as possible markers for the prediction of preeclampsia. Examples include corticotropin-releasing hormone (Petraglia and colleagues, 1996), chorionic gonadotropin (Ashour and colleagues, 1997), and activin A and inhibin A (Aquilina and co-authors, 1999; Cuckle and associates, 1998; Muttukrishna and colleagues, 1997). Inhibin A and activin A appear to be particularly promising in the search for early pregnancy markers for the development of preeclampsia (Lindheimer and Woodruff, 1997).

DOPPLER VELOCIMETRY OF THE UTERINE ARTERIES. Uteroplacental vascular resistance can be assessed by Doppler ultrasound (Chap. 41, p. 1133). This has prompted studies to use Doppler measurements of uterine artery impedance in the second trimester as an early screening test for preeclampsia (Bewley and colleagues, 1991; Chappell and Bewley, 1998). The rationale for this is based upon the presumption that the pathophysiology of preeclampsia includes impaired trophoblastic invasion of the spiral arteries leading to obstruction in uteroplacental blood flow. Bower and colleagues (1993) screened 2026 pregnant women at 18 to 22 weeks using continuous-wave Doppler. Those women with increased uterine artery resistance (13 percent) underwent repeat testing using color Doppler at 24 weeks. The sensitivity of this two-stage test for prediction of preeclampsia was 78 percent, but the positive-predictive value was only 28 percent (Friedman and Lindheimer, 1999). Irion and associates (1998) found uterine artery Doppler velocimetry to be an unreliable screening test for preeclampsia in low-risk pregnancies. Clearly, none of the tests for prediction of preeclampsia described in this section are ideal.

PREVENTION

A variety of strategies have been used in attempts to prevent preeclampsia. Usually these strategies involve manipulation of diet and pharmacological attempts to modify the pathophysiological mechanisms thought to play a role in the development of preeclampsia. The latter includes use of low-dose aspirin and antioxidants.

DIETARY MANIPULATION. One of the earliest efforts aimed at preventing preeclampsia was salt restriction during pregnancy (De Snoo, 1937). The first randomized trial, however, was not published until 1998 (Knuist and colleagues). In this study of 361 women, prescribing a sodium-restricted diet was proven to be ineffective in preventing hypertensive disorders due to pregnancy.

Based primarily upon studies outside the United States, women with low dietary calcium were found to be at significantly increased risk for developing hypertension due to pregnancy (Belizan and Villar, 1980; Lopez-Jaramillo and associates, 1989; Marya and colleagues, 1987). This has led to at least 14 randomized trials and resultant meta-analysis that showed calcium supplementation during pregnancy resulted in a significant reduction in blood pressure as well as prevention of preeclampsia (Bucher and colleagues, 1996). The apparently definitive study, however, was completed by Levine and co-workers (1997). This was a randomized trial sponsored by the National Institute of Child Health and Human Development. In this trial, using double-masking, 4589 healthy nulliparous women were randomly administered either 2 g per day of supplemental calcium or placebo. Supplemental calcium did not prevent any of the hypertensive disorders due to pregnancy, including gestational hypertension or preeclampsia.

Other dietary manipulations that have been tested for the purpose of preventing preeclampsia include administering four to nine capsules containing fish oil each day (Olsen and colleagues, 2000). This dietary supplement was chosen in an effort to modify the prostaglandin balance implicated in the pathophysiology of preeclampsia. Fish oil was ineffective in this study of 1474 pregnancies conducted at 19 hospitals in Europe.

LOW-DOSE ASPIRIN. In 1986, Wallenburg and co-workers reported their experiences with either 60 mg of aspirin or placebo given to angiotensin-sensitive primigravid women at 28 weeks' gestation. The reduced incidence of preeclampsia in the treated group was attributed to selective suppression of thromboxane synthesis by platelets and sparing of endothelial prostacyclin production. As a result of this report and others with similar results, multicenter randomized trials have been completed in both low-risk and high-risk women in the United States as well as in other countries (Caritis, 1998; CLASP Collaborative Group, 1994; Hauth, 1993; Sibai, 1993a, and their many colleagues). These trials have consistently shown that low-dose aspirin was ineffective in preventing preeclampsia. Hauth and colleagues (1998) in a secondary analysis of the high-risk intervention trial by Caritis and co-workers (1998), showed that administration of low-dose aspirin significantly reduced maternal thromboxane B₂ levels, but that this was without benefit because the incidence of preeclampsia was not decreased compared with placebo.

ANTIOXIDANTS. Sera of normal pregnant women contain antioxidant mechanisms that function to control lipid peroxidation which have been implicated in endothelial cell dysfunction associated with preeclampsia. Davidge and co-authors (1992) have shown that sera of women with preeclampsia have markedly reduced antioxidant activity. Schiff and colleagues (1996) tested the hypothesis that diminished antioxidant activity is involved in preeclampsia by studying dietary consumption as well as plasma concentration of vitamin E in 42 pregnancies complicated by preeclampsia compared with 90 controls. They found high plasma vitamin E levels in women with preeclampsia, but that the dietary vitamin E consumption was unrelated to preeclampsia. They speculated that the high vitamin E levels they observed were a response to the oxidative stress of preeclampsia.

Chappell and associates (1999) performed the first systematic study designed to test the hypothesis that treatment of pregnant women with antioxidants would alter endothelial cell injury linked to preeclampsia. A total of 283 women at risk for preeclampsia were randomized at 18 to 22 weeks' gestation to treatment with antioxidants or placebo. Antioxidant therapy significantly reduced endothelial cell activation, suggesting that such therapy might indeed be beneficial in the prevention of preeclampsia. There was also a significant reduction in the incidence of preeclampsia in those women given vitamins C and E compared with the control group (17 versus 11 percent, P < .02). A larger trial must be performed before concluding that such antioxidant therapy prevents preeclampsia.

MANAGEMENT

Basic management objectives for any pregnancy complicated by preeclampsia are:

1. Termination of pregnancy with the least possible trauma to mother and fetus.

2. Birth of an infant who subsequently thrives.

3. Complete restoration of health to the mother.

In certain cases of preeclampsia, especially in women at or near term, all three objectives are served equally well by induction of labor. Therefore, the most important information that the obstetrician has for successful management of pregnancy, and especially a pregnancy that becomes complicated by hypertension, is precise knowledge of the age of the fetus.

EARLY PRENATAL DETECTION

Traditionally, the timing of prenatal examinations has been scheduled at intervals of 4 weeks until 28 weeks, and then every 2 weeks until 36 weeks, and weekly thereafter. Increased prenatal visits during the third trimester facilitates early detection of preeclampsia. Women with overt hypertension (³ 140/90 mm Hg) are frequently admitted to the hospital for 2 to 3 days to evaluate the severity of new-onset pregnancy hypertension. Those with persistent severe disease are observed closely and many are delivered. Conversely, women with mild disease are often managed as outpatients.

Management of women without overt hypertension, but in whom early preeclampsia is suspected during routine prenatal visits, is primarily based upon increased surveillance. The protocol used successfully for many years at Parkland Hospital in women during the third trimester and with new-onset diastolic blood pressure readings between 81 and 89 mm Hg or sudden abnormal weight gain (more than 2 pounds per week) includes return visits at 3 to 4 day intervals. Such outpatient surveillance is continued unless overt hypertension, proteinuria, visual disturbances, or epigastric discomfort supervene.

HOSPITAL MANAGEMENT

Hospitalization is considered at least initially for women with new-onset hypertension if there is persistent or worsening hypertension or development of proteinuria. A systematic evaluation is instituted to include the following:

1. Detailed examination followed by daily scrutiny for clinical findings such as headache, visual disturbances, epigastric pain, and rapid weight gain.

2. Weight on admittance and every day thereafter.

3. Analysis for proteinuria on admittance and at least every 2 days thereafter.

4. Blood pressure readings in sitting position with an appropriate-size cuff every 4 hours, except between midnight and morning.

5. Measurements of plasma or serum creatinine, hematocrit, platelets, and serum liver enzymes, the frequency to be determined by the severity of hypertension.

6. Frequent evaluation of fetal size and amnionic fluid volume either clinically or with sonography.

If these observations lead to a diagnosis of severe preeclampsia (Table 24-2), further management is the same as described subsequently for eclampsia.

Reduced physical activity throughout much of the day is beneficial. Absolute bed rest is not necessary, and sedatives and tranquilizers are not prescribed. Ample, but not excessive, protein and calories should be included in the diet. Sodium and fluid intakes should not be limited or forced. Further management depends upon:

1. Severity of preeclampsia, determined by presence or absence of conditions cited.

2. Duration of gestation.

3. Condition of the cervix.

Fortunately, many cases prove to be sufficiently mild and near enough to term that they can be managed conservatively until labor commences spontaneously or until the cervix becomes favorable for labor induction. Complete abatement of all signs and symptoms, however, is uncommon until after delivery. Almost certainly, the underlying disease persists until after delivery!

TERMINATION OF PREGNANCY

Delivery is the cure for preeclampsia. Headache, visual disturbances, or epigastric pain are indicative that convulsions are imminent, and oliguria is another ominous sign. Severe preeclampsia demands anticonvulsant and usually antihypertensive therapy followed by delivery. Treatment is identical to that described subsequently for eclampsia. The prime objectives are to forestall convulsions, to prevent intracranial hemorrhage and serious damage to other vital organs, and to deliver a healthy infant.

When the fetus is known or suspected to be preterm, however, the tendency is to temporize in the hope that a few more weeks in utero will reduce the risk of neonatal death or serious morbidity. As discussed, such a policy certainly is justified in milder cases. Assessments of fetal well-being and placental function have been attempted, especially when there is hesitation to deliver the fetus because of prematurity. Most investigators recommend frequent performance of various tests currently used to assess fetal well-being as described by the American College of Obstetricians and Gynecologists (1999). These include the nonstress test or the biophysical profile, which are discussed in Chapter 40 (p. 1104). Measurement of the lecithin-sphingomyelin ratio in amnionic fluid may provide evidence of lung maturity. Even when this ratio is less than 2.0, however, respiratory distress may not develop; and if it does, it is usually not fatal (Chap. 39, p. 1006).

With moderate or severe preeclampsia that does not improve after hospitalization, delivery is usually advisable for the welfare of both mother and fetus. Labor should be induced by intravenous oxytocin. Many clinicians favor preinduction cervical ripening with a prostaglandin or osmotic dilator (see Chap. 20). Whenever it appears that labor induction almost certainly will not succeed, or attempts at induction of labor have failed, cesarean delivery is indicated for more severe cases.

For a woman near term, with a soft, partially effaced cervix, even milder degrees of preeclampsia probably carry more risk to the mother and her fetus-infant than does induction of labor by carefully monitored oxytocin induction. This is not likely to be the case, however, if the preeclampsia is mild but the cervix is firm and closed, indicating that abdominal delivery might be necessary if pregnancy is to be terminated. The hazard of cesarean delivery may be greater than that of allowing the pregnancy to continue under close observation until the cervix is more suitable for induction.

ELECTIVE CESAREAN DELIVERY. Once severe preeclampsia is diagnosed, the obstetrical propensity is for prompt delivery. Induced labor to effect vaginal delivery has traditionally been considered to be in the best interest of the mother. Several concerns, including an unfavorable cervix precluding successful induction of labor, a perceived sense of urgency because of the severity of preeclampsia, and the need to coordinate neonatal intensive care have led some practitioners to advocate cesarean delivery. Alexander and colleagues (1999) reviewed 278 singleton liveborn infants weighing 750 to 1500 g delivered of women with severe preeclampsia at Parkland Hospital. Half of the women had labor induced and the remainder were delivered by cesarean without labor. Induction of labor was not successful in 35 percent of the women in the induced group, but was not harmful to their very low-birthweight infants. Similar results were reported by Nassar and colleagues (1998).

ANTIHYPERTENSIVE DRUG THERAPY

The use of antihypertensive drugs in attempts to prolong pregnancy or modify perinatal outcomes in pregnancies complicated by various types and severities of hypertensive disorders has been of considerable interest.

Drug treatment for early mild preeclampsia has been disappointing (Table 24-8). Sibai and associates (1987a) performed a well-designed randomized study to evaluate the effectiveness of labetalol and hospitalization compared with hospitalization alone. They evaluated 200 nulliparous women with preeclampsia diagnosed between 26 and 35 weeks. Although women given labetalol had significantly lower mean blood pressure, there were no differences between the groups for mean pregnancy prolongation, gestational age at delivery, or birthweight. The cesarean delivery rates were similar, as were the number of infants admitted to special-care nurseries. Growth-restricted infants were twice as frequent in women given labetalol compared with those treated by hospitalization alone (19 versus 9 percent).

At least three other studies have been done to compare either the b-blocking agent, labetalol, or calcium-channel blockers, nifedipine and isradipine, with placebo, and the results are shown in Table 24-8. In none of these studies were any benefits of antihypertensive treatment shown. Von Dadelszen and associates (2000) performed a meta-analysis that included the aforementioned trials for the purpose of determining the relation between fetal growth and antihypertensive therapy. These investigators concluded that treatment-induced decreases in maternal blood pressure may adversely affect fetal growth.

In a somewhat unusual study, Easterling and colleagues (1999) identified 58 women at risk for preeclampsia because they had a high cardiac output at 24 weeks. Cardiac output was measured by Doppler technique and the women were randomized to either prophylactic atenolol or placebo. The incidence of preeclampsia was 18 percent in the control group compared with 4 percent in the atenolol group (P = .04).

The use of angiotensin-converting-enzyme (ACE) inhibitors during the second and third trimesters of pregnancy should be avoided (Chap. 38, p. 1014). Reported complications include oligohydramnios, fetal growth restriction, bony malformations, limb contractures, persistent patent ductus arteriosus, pulmonary hypoplasia, respiratory distress syndrome, prolonged neonatal hypotension, and neonatal death (Nightingale, 1992). Lip and colleagues (1997) reported that ACE⁴ inhibitors taken during early pregnancy do not carry an adverse outlook as long as these drugs are discontinued as soon as possible.

DELAYED DELIVERY WITH SEVERE PREECLAMPSIA

Women with severe preeclampsia are usually delivered without delay. In recent years, a different approach in the treatment of women with severe preeclampsia remote from term has been advocated by several investigators worldwide (Many and colleagues, 1999). This approach advocates conservative or "expectant" management in a selected group of women with the aim of improving infant outcome without compromising the safety of the mother. Aspects of such conservative management always include careful daily and more frequent, monitoring of the pregnancy in the hospital with or without use of drugs to control hypertension.

Theoretically, antihypertensive therapy has potential usefulness when preeclampsia severe enough to warrant termination of pregnancy develops before neonatal survival is likely. Such management is controversial, and it may be catastrophic. Sibai and colleagues (1985b) from the University of Tennessee attempted to prolong pregnancy because of fetal immaturity in 60 women with severe preeclampsia diagnosed between 18 and 27 weeks. The total perinatal mortality rate was 87 percent, and although no mothers died, 13 suffered placental abruption, 10 eclampsia, five consumptive coagulopathy, three renal failure, two hypertensive encephalopathy, one intracerebral hemorrhage, and one a ruptured hepatic hematoma.

Sibai and associates (1994) subsequently performed a randomized controlled trial of expectant versus aggressive management of severe preeclampsia in 95 women at more advanced gestations of 28 to 32 weeks. Women with HELLP¹ syndrome were specifically excluded from this trial. Aggressive management included glucocorticoid administration for fetal lung maturation followed by delivery in 48 hours. Expectantly managed women were treated with bed rest and either labetalol or nifedipine given orally. Pregnancy was prolonged for a mean of 15.4 days in the expectant management group with an improvement in neonatal outcome. Importantly, 4 percent in each group sustained placental abruption.

In a follow-up nonrandomized study from Memphis, Abramovici and colleagues (1999) compared infant outcomes for deliveries between 24 and 36 weeks in 133 women with HELLP¹ syndrome to 136 women with severe preeclampsia. Women with HELLP syndrome were subdivided into those with hemolysis plus elevated liver enzymes plus low platelets and those with partial HELLP syndrome defined as either one or two, but not three, of these laboratory findings. It was concluded that women with partial HELLP syndrome, as well as those with severe preeclampsia, could be managed expectantly. They also concluded that infant outcomes were related to gestational age rather than the hypertensive disorder per se. As shown in Table 24-9, these women were indeed severely hypertensive and had mean diastolic blood pressures of 110 mm Hg. The distinguishing feature between those with HELLP syndrome appears to be the platelet count: the mean value was 52,000/uL in women with complete HELLP syndrome compared with 113,000/uL in those with partial HELLP syndrome. Gestational age was about 2 weeks more advanced in women with severe preeclampsia compared with those with some degree of HELLP syndrome. Accordingly, neonatal outcomes, in terms of need for mechanical ventilation and neonatal death, were better in women with severe preeclampsia. Fetal growth restriction was not related to the severity of maternal disease and was prevalent in all three groups. Maternal morbidity was not described. However, Witlin and associates (2000) later reported that growth restriction adversely affected infant survival in such infants delivered in Memphis. Most importantly, the median elapsed time from admission to delivery was 0, 1 day, and 2 days for women with HELLP syndrome, partial HELLP syndrome, or severe preeclampsia, respectively.

Taken in toto, these three studies require comment about repeated claims that expectant management of severe preeclampsia and partial HELLP¹ syndrome is beneficial. First, with expectant management, the interval from admission to delivery is very short at their center, particularly in pregnancies managed subsequent to their report claiming efficacy and safety. Second, it seems possible, even likely, that the gestational age difference between severe preeclampsia and HELLP syndrome is related to the timing of onset of the disease itself. That is, HELLP syndrome may develop earlier in pregnancy than severe preeclampsia. Third, fetal growth restriction is prevalent in women with severe disease and it adversely affects infant survival, which is not improved by the severity of maternal disease. Lastly, and most importantly, the authors overlook that the overriding reason to terminate pregnancies with severe preeclampsia is maternal safety. Results in Table 24-9 clearly confirm that severe preeclampsia is deleterious for fetal outcome. Conversely, these data in Table 24-9 do not prove that expectant management is beneficial for the mother.

Visser and Wallenburg (1995) compared clinical outcomes of expectant management in 256 women with severe preeclampsia prior to 34 weeks. Half of these women had HELLP¹ syndrome and the other half had severe preeclampsia. They were able to prolong pregnancy in both groups for 10 to 14 days, but 5 percent had a placental abruption and three women developed eclampsia.

Hall and colleagues (2000b) managed 360 severely preeclamptic women with early-onset preeclampsia—before 34 weeks—by careful observation and blood pressure control. While these women gained a mean duration of 11 days, a fourth had major complications—20 percent had placental abruption, 2 percent pulmonary edema, and 1.2 percent had eclampsia. Reports such as these just cited serve to emphasize that it is prudent for clinicians to be concerned about maternal safety in women with severe hypertensive disorders due to pregnancy before term. We are reluctant to advise clinicians that it is safe to expectantly manage women with persistent severe hypertension or significant hematological, cerebral, or liver abnormalities due to preeclampsia. Such women are not managed expectantly at Parkland Hospital.

GLUCOCORTICOIDS

In attempts to enhance fetal lung maturation, glucocorticoids have been administered to severely hypertensive pregnant women remote from term. Treatment does not seem to worsen maternal hypertension, and a decrease in the incidence of respiratory distress and improved fetal survival has been cited (Chap. 39, p. 1006). Thiagarajah and co-authors (1984) were the first to suggest that glucocorticoids might also have a role in treatment of the laboratory abnormalities associated with the HELLP¹ syndrome. Tompkins and Thiagarajah (1999) more recently reported that glucocorticoids produced significant but transient improvement in the hematological abnormalities associated with the HELLP syndrome diagnosed upon admission in 52 women between 24 and 34 weeks. While platelet counts increased by an average of 23,000/uL, this salutary effect was short lived, and counts decreased by an average of 46,000/uL within 48 hours after completion of the glucocorticoid regimen. Importantly, few of the women studied had platelet counts less than 100,000/uL before glucocorticoid therapy, and hence, the efficacy of such treatment was not extensively tested in women with more severe hematological abnormalities.

One interpretation of these reports is that such use of glucocorticoids specifically administered as a therapy for the hematological abnormalities due to severe preeclampsia will not significantly delay the need for delivery. Most certainly, these reports should not be taken to imply that it is advisable to administer glucocorticoids to significantly delay delivery in women with severe laboratory abnormalities.

HIGH-RISK PREGNANCY UNIT

An inpatient High-Risk Pregnancy Unit was established at Parkland Hospital in 1973 in large part to provide care for women with hypertensive disorders. Initial results from this Unit were reported by Hauth (1976) and Gilstrap (1978) and their colleagues. The majority of women hospitalized have a salutary response characterized by disappearance or improvement of hypertension. Importantly, these women are not "cured," because nearly 90 percent have recurrent hypertension before or during labor. Through 2000, almost 8000 nulliparous women with mild to moderate early-onset hypertension during pregnancy have been successfully managed in the High-Risk Pregnancy Unit. The provider costs (as opposed to charges) for the relatively simple physical facility, modest nursing care, no drugs other than iron and folate supplements, and the very few laboratory tests that are essential, are slight compared with the cost of neonatal intensive care for a preterm infant.

HOME HEALTH CARE

Many clinicians feel that further hospitalization is not warranted if hypertension abates within a few days, and this has legitimized third-party payors in the United States to refuse hospital reimbursement. Thus, many women with mild to moderate hypertension, and without proteinuria, are consequently managed at home. Such management may continue as long as the disease does not worsen and if fetal jeopardy is not suspected. Sedentary activity throughout the greater part of the day is recommended. These women should be instructed in detail about reporting symptoms. Home blood pressure and urine protein monitoring or frequent evaluations by a visiting nurse may be necessary.

Barton and colleagues (1994) managed 592 predominately nulliparous women with mild hypertension in such a manner. They were 24 to 36 weeks at enrollment and a fourth had proteinuria. Gestation was prolonged for a mean of 4 weeks and 60 percent were delivered after 37 weeks. Importantly, half developed severe hypertension and 10 percent had fetal jeopardy. These investigators concluded that in highly motivated women such management produced results similar to inpatient care. Helewa and associates (1993) reached similar conclusions with a home-care program.

In a study from Parkland Hospital, Horsager and associates (1995) randomized 72 nulliparas with new-onset hypertension from 27 to 37 weeks to continued hospitalization or outpatient care. In all of these women, proteinuria had receded to less than 500 mg per day when randomized. Outpatient management included daily blood pressure monitoring by the patient or her family and weight and spot urine protein were determined three times weekly. A home health nurse visited twice weekly and the women were seen weekly in the clinic. Although perinatal outcomes were similar, recurrence of severe preeclampsia was significantly more common in the home group than in hospitalized women (40 versus 25 percent, respectively).

Another approach that has been evaluated on a limited basis is day-care. Tuffnell and colleagues (1992) randomized 54 women with gestational hypertension after 26 weeks to either day-care or routine management by their individual physicians. Hospitalizations, the development of preeclampsia, and labor inductions were significantly increased in the routine management group.

ECLAMPSIA

Preeclampsia that is complicated by generalized tonic-clonic convulsions is termed eclampsia. Fatal coma without convulsions has also been called eclampsia; however, it is better to limit the diagnosis to women with convulsions and to regard deaths in nonconvulsive cases as due to severe preeclampsia. Once eclampsia has ensued, the risk to both mother and fetus is appreciable.

Almost without exception, preeclampsia precedes the onset of eclamptic convulsions. Depending on whether convulsions appear before, during, or after labor, eclampsia is designated as antepartum, intrapartum, or postpartum. Eclampsia is most common in the last trimester and becomes increasingly more frequent as term approaches. In 254 eclamptic women cared for at the University of Mississippi Medical Center, about 3 percent first developed seizures more than 48 hours postpartum (Miles and colleagues, 1990). Other diagnoses should be considered in women with the onset of convulsions more than 48 hours postpartum.

The convulsive movements usually begin about the mouth in the form of facial twitchings. After a few seconds, the entire body becomes rigid in a generalized muscular contraction. This phase may persist for 15 to 20 seconds. Suddenly the jaws begin to open and close violently, and soon after, the eyelids as well. The other facial muscles and then all muscles alternately contract and relax in rapid succession. So forceful are the muscular movements that the woman may throw herself out of her bed, and, if not protected, her tongue is bitten by the violent action of the jaws (Fig. 24-12). This phase, in which the muscles alternately contract and relax, may last about a minute. Gradually, the muscular movements become smaller and less frequent, and finally the woman lies motionless. Throughout the seizure the diaphragm has been fixed, with respiration halted. For a few seconds the woman appears to be dying from respiratory arrest, but then she takes a long, deep, stertorous inhalation, and breathing is resumed. Coma then ensues. She will not remember the convulsion or, in all probability, events immediately before and afterward. Over time, these memories return.

The first convulsion is usually the forerunner of others, which may vary in number from one or two in mild cases to even 100 or more in untreated severe cases. In rare instances, convulsions follow one another so rapidly that the woman appears to be in a prolonged, almost continuous convulsion.

The duration of coma after a convulsion is variable. When the convulsions are infrequent, the woman usually recovers some degree of consciousness after each attack. As the woman arouses, a semiconscious combative state may ensue. In very severe cases, the coma persists from one convulsion to another, and death may result before she awakens. In rare instances, a single convulsion may be followed by coma from which the woman may never emerge, although, as a rule, death does not occur until after frequent convulsions.

Respirations after an eclamptic convulsion are usually increased in rate and may reach 50 or more per minute, in response presumably to hypercarbia from lactic acidemia, as well as to varying intensities of hypoxia. Cyanosis may be observed in severe cases. Fever of 39°C or more is a very grave sign, because it is probably the consequence of a central nervous system hemorrhage.

Proteinuria is almost always present and frequently pronounced. Urine output is likely diminished appreciably, and occasionally anuria develops. Hemoglobinuria is common, but hemoglobinemia is observed only rarely. Often, as shown in Figure 24-13, the edema is pronounced—at times, massive—but it may also be absent.

As with severe preeclampsia, after delivery an increase in urinary output is usually an early sign of improvement. Proteinuria and edema ordinarily disappear within a week (Fig. 24-13). In most cases, blood pressure returns to normal within a few days to 2 weeks after delivery. The longer hypertension persists postpartum, the more likely that it is the consequence of chronic vascular or renal disease (Table 24-1).

In antepartum eclampsia, labor may begin spontaneously shortly after convulsions ensue and progress rapidly, sometimes before the attendants are aware that the unconscious or stuporous woman is having effective uterine contractions. If the convulsion occurs during labor, contractions may increase in frequency and intensity, and the duration of labor may be shortened. Because of maternal hypoxemia and lactic acidemia caused by convulsions, it is not unusual for fetal bradycardia to follow a seizure (Fig. 24-14). This usually recovers within 3 to 5 minutes; if it persists more than about 10 minutes, another cause must be considered, such as placental abruption or imminent delivery.

Pulmonary edema may follow eclamptic convulsions. There are at least two sources:

1. Aspiration pneumonitis may follow inhalation of gastric contents if simultaneous vomiting accompanies convulsions.

2. Cardiac failure may be the result of a combination of severe hypertension and vigorous intravenous fluid administration.

In some women with eclampsia, sudden death occurs synchronously with a convulsion or follows shortly thereafter, as the result of a massive cerebral hemorrhage (Fig. 24-15). Hemiplegia may result from sublethal hemorrhage. Cerebral hemorrhages are more likely in older women with underlying chronic hypertension. Rarely they may be due to a ruptured berry aneurysm or arteriovenous malformation (Witlin and co-workers, 1997a).

In about 10 percent of women, some degree of blindness will follow a seizure. Blindness may also develop spontaneously with preeclampsia. There are at least two causes:

1. Varying degrees of retinal detachment.

2. Occipital lobe ischemia, infarction, or edema.

Whether due to cerebral or retinal pathology, the prognosis for return of normal vision is good and usually complete within a week (Cunningham and associates, 1995). About 5 percent of women will have substantively altered consciousness, including persistent coma, following a seizure. This is due to extensive cerebral edema, and transtentorial uncal herniation may cause death (Cunningham and Twickler, 2000).

Rarely, eclampsia is followed by psychosis, and the woman becomes violent. This usually lasts for several days to 2 weeks, but the prognosis for return to normal is good, provided there was no preexisting mental illness. Chlorpromazine in carefully titrated doses has proved effective in the few cases of posteclampsia psychosis treated at Parkland Hospital.

DIFFERENTIAL DIAGNOSIS. Generally, eclampsia is more likely to be diagnosed too frequently rather than overlooked, because epilepsy, encephalitis, meningitis, cerebral tumor, cysticercosis, and ruptured cerebral aneurysm during late pregnancy and the puerperium may simulate eclampsia. Until other such causes are excluded, however, all pregnant women with convulsions should be considered to have eclampsia.

PROGNOSIS. The prognosis for eclampsia is always serious; this is one of the most dangerous conditions that can afflict a pregnant woman and her fetus. Fortunately, maternal mortality due to eclampsia has decreased in the past three decades from 5 to 10 percent to less than 3 percent of cases (Eclampsia Trial Collaborative Group, 1995; Mattar and Sibai, 2000; Pritchard and associates, 1984). These experiences clearly underscore that eclampsia, as well as severe preeclampsia, are to be considered overt threats to maternal life. Indeed, 23 percent of maternal deaths recorded in the United States during 1997 were related to pregnancy hypertension and accounted for at least 64 deaths (Hoyert and colleagues, 1999).

TREATMENT. In 1955 Pritchard initiated a standardized treatment regimen at Parkland Hospital, and this was used through 1999 to manage over 400 women with eclampsia. The carefully analyzed results of treatment of 245 cases of eclampsia were reported by Pritchard and associates (1984). Most eclampsia regimens used in the United States adhere to a similar philosophy, the tenets of which include:

1. Control of convulsions using an intravenously administered loading dose of magnesium sulfate. This is followed by either a continuous infusion of magnesium sulfate or by an intramuscular loading dose and periodic intramuscular injections.

2. Intermittent intravenous or oral administration of an antihypertensive medication to lower blood pressure whenever the diastolic pressure is considered dangerously high. Some clinicians treat at 100 mm Hg, some at 105 mm Hg, and some at 110 mm Hg.

3. Avoidance of diuretics and limitation of intravenous fluid administration unless fluid loss is excessive. Hyperosmotic agents are avoided.

4. Delivery.

MAGNESIUM SULFATE TO CONTROL CONVULSIONS. In more severe cases of preeclampsia, as well eclampsia, magnesium sulfate administered parenterally is the effective anticonvulsant agent without producing central nervous system depression in either the mother or the infant. It may be given intravenously by continuous infusion or intramuscularly by intermittent injection (Table 24-10). The dosage schedule for severe preeclampsia is the same as for eclampsia. Because labor and delivery is a more likely time for convulsions to develop, women with preeclampsia-eclampsia usually are given magnesium sulfate during labor and for 24 hours postpartum. Magnesium sulfate is not given to treat hypertension.

Based on a number of studies cited subsequently, as well as extensive clinical observations, magnesium most likely exerts a specific anticonvulsant action on the cerebral cortex. Typically, the mother stops convulsing after the initial administration of magnesium sulfate, and within an hour or two regains consciousness sufficiently to be oriented as to place and time.

The magnesium sulfate dosage schedules presented in Table 24-10 usually result in plasma magnesium levels illustrated in Figures 24-16 and 24-17. When magnesium sulfate is given to arrest and prevent recurrent eclamptic seizures, about 10 to 15 percent of women will have a subsequent convulsion. An additional 2-g dose of magnesium sulfate in a 20 percent solution is administered slowly intravenously. In a small woman, an additional 2-g dose may be used once, and twice if needed in a larger woman. In only 5 of 245 women with eclampsia at Parkland Hospital was it necessary to use supplementary medication to control convulsions (Pritchard and associates, 1984). Sodium amobarbital is given slowly intravenously in doses up to 250 mg in women who are excessively agitated in the postconvulsion phase. Thiopental is suitable also. Maintenance magnesium sulfate therapy for eclampsia is continued for 24 hours after delivery. For eclampsia that develops postpartum, magnesium sulfate is administered for 24 hours after the onset of convulsions.

PHARMACOLOGY AND TOXICOLOGY OF MAGNESIUM SULFATE. Magnesium sulfate USP is MgSO₄ · 7H₂O and not MgSO₄. Parenterally administered magnesium is cleared almost totally by renal excretion, and magnesium intoxication is avoided by ensuring that urine output is adequate, the patellar or biceps reflex is present, and there is no respiratory depression. Eclamptic convulsions are almost always prevented by plasma magnesium levels maintained at 4 to 7 mEq/L (4.8 to 8.4 mg/dL or 2.0 to 3.5 mmol/L).

When administered as described in Table 24-10, the drug will practically always arrest eclamptic convulsions and prevent their recurrence. The initial intravenous infusion of 4 to 6 g is used to establish a prompt therapeutic level that is maintained by the nearly simultaneous intramuscular injection of 10 g of the compound, followed by 5 g intramuscularly every 4 hours, or by continuous infusion at 2 to 3 g per hour. With these dosage schedules, therapeutically effective plasma levels of 4 to 7 mEq/L are achieved compared with pretreatment plasma levels of less than 2.0 mEq/L.

Sibai and co-workers (1984) performed a prospective study in which they compared continuous intravenous magnesium sulfate and intramuscular magnesium sulfate. There was no significant difference between mean magnesium levels observed after intramuscular magnesium sulfate and those observed following a maintenance intravenous infusion of 2 g per hour (Fig. 24-17). In our experience many women will require 3-g-per-hour infusions to maintain effective plasma levels of magnesium.

Patellar reflexes disappear when the plasma magnesium level reaches 10 mEq per L (about 12 mg/dL), presumably because of a curariform action. This sign serves to warn of impending magnesium toxicity, because a further increase will lead to respiratory depression.

When plasma levels rise above 10 mEq/L, respiratory depression develops, and at 12 mEq/L or more, respiratory paralysis and arrest follow. Somjen and co-workers (1966) induced in themselves, by intravenous infusion, marked hypermagnesemia, achieving plasma levels up to 15 mEq/L. Predictably, at such high plasma levels, respiratory depression developed that necessitated mechanical ventilation, but depression of the sensorium was not dramatic as long as hypoxia was prevented. Treatment with calcium gluconate, 1 g intravenously, along with the withholding of magnesium sulfate usually reverses mild to moderate respiratory depression. Unfortunately, the effects of intravenously administered calcium may be short lived. For severe respiratory depression and arrest, prompt tracheal intubation and mechanical ventilation are life saving. Direct toxic effects on the myocardium from high levels of magnesium are uncommon. It appears that the cardiac dysfunction associated with magnesium is due to respiratory arrest and hypoxia. With appropriate ventilation, cardiac action is satisfactory even when plasma levels are exceedingly high (McCubbin and colleagues, 1981).

IMPAIRED RENAL FUNCTION. Because magnesium is cleared almost exclusively by renal excretion, plasma magnesium concentration, using the doses described previously, will be excessive if glomerular filtration is decreased substantively. The initial standard dose of magnesium sulfate can be safely administered without knowledge of renal function. Renal function is thereafter estimated by measuring plasma creatinine, and whenever it is 1.3 mg/dL or higher, we give only half of the maintenance intramuscular magnesium sulfate dose outlined in Table 24-10. With this renal impairment dosage, plasma magnesium levels are usually within the desired range of 4 to 7 mEq/L. Whether magnesium sulfate is being given intravenously by continuous infusion, serum magnesium levels are used to adjust the infusion rate. With either method, when there is renal insufficiency, plasma magnesium levels must be checked periodically.

Acute cardiovascular effects of parenteral magnesium ion in women with severe preeclampsia have been studied by Cotton and associates (1986b), who obtained data using pulmonary and radial artery catheterization. Following a 4-g intravenous dose given over 15 minutes, mean arterial blood pressure fell slightly, and this was accompanied by a 13 percent increase in cardiac index. Thus, magnesium decreased systemic vascular resistance and mean arterial pressure, and at the same time increased cardiac output, without evidence of myocardial depression. These findings were coincidental with transient nausea and flushing, and the cardiovascular effects persisted for only 15 minutes despite continued infusion of magnesium sulfate at 1.5 g per hour.

Thurnau and colleagues (1987) showed that there was a small but highly significant increase in cerebrospinal fluid magnesium concentration after magnesium therapy for preeclampsia. The magnitude of the increase was directly proportional to the corresponding serum concentration. This increase cannot be due to the disease itself, because cerebrospinal fluid magnesium levels are unchanged in untreated severely preeclamptic women when compared with normotensive controls (Fong and associates, 1995).

Lipton and Rosenberg (1994) attribute anticonvulsant effects to blocked neuronal calcium influx through the glutamate channel. Cotton and associates (1992) induced seizure activity in the hippocampus region of rats because it is a region with a low seizure threshold and a high density of N-methyl-D-aspartate receptors. These receptors are linked to various models of epilepsy. Because hippocampal seizures can be blocked by magnesium, it is believed that this implicated the N-methyl-D-aspartate receptor in eclamptic convulsions (Hallak and colleagues, 1998). Importantly, results such as these suggest that magnesium has a central nervous system effect in blocking seizures.

UTERINE EFFECTS. Magnesium ions in relatively high concentration will depress myometrial contractility both in vivo and in vitro. With the regimen described earlier and the plasma levels that have resulted, no evidence of myometrial depression has been observed beyond a transient decrease in activity during and immediately after the initial intravenous loading dose. Indeed, Leveno and colleagues (1998) compared labor and delivery outcomes in 480 nulliparous women given phenytoin for preeclampsia with outcomes in 425 similar women given magnesium sulfate. Magnesium sulfate did not significantly alter oxytocin stimulation of labor, admission-to-delivery intervals, or route of delivery. Similar results have been reported by others (Atkinson and associates, 1995; Szal and co-workers, 1999; Witlin and colleagues, 1997b).

The mechanisms by which magnesium might inhibit uterine contractility is not established, but is generally assumed to depend on its effect on intracellular calcium (Watt-Morse and associates, 1995). The regulatory pathway leading to uterine contraction begins with an increase in the intracellular free Ca²⁺ concentration, which activates myosin light chain kinase (Mizuki and associates, 1993). High concentrations of extracellular magnesium have been reported not only to inhibit calcium entry into myometrial cells but to also lead to high intracellular magnesium levels.

This latter effect has been reported to inhibit calcium entry into the cell—presumably by blocking calcium channels (Mizuki and associates, 1993). These mechanisms for inhibition of uterine contractility appear to be dose dependent because serum magnesium levels of at least 8 to 10 mEq/L are necessary to inhibit uterine contractions (Watt-Morse and associates, 1995). This likely explains why there is no uterine effect clinically when magnesium sulfate is given for eclampsia treatment or prophylaxis. Specifically, magnesium sulfate when given either intravenously or intramuscularly for preeclampsia or eclampsia, consistently results in serum magnesium levels less than the 8 to 10 mEq/L necessary to inhibit uterine contractility (Fig. 24-16).

FETAL EFFECTS. Magnesium administered parenterally to the mother promptly crosses the placenta to achieve equilibrium in fetal serum and less so in amnionic fluid (Hallak and colleagues, 1993). The neonate may be depressed only if there is severe hypermagnesemia at delivery. We have not observed neonatal compromise after therapy with magnesium sulfate (Cunningham and Pritchard, 1984), nor have Green and associates (1983). Whether magnesium sulfate affects the fetal heat rate pattern, specifically beat-to-beat variability (Chap. 14, p. 336), is controversial. Hallak and colleagues (1999b), in a randomized investigation comparing an infusion of magnesium sulfate to saline, found that magnesium sulfate was associated with a small but clinically insignificant decrease in variability of the fetal heart rate.

There is a suggestion of a possible protective effect of magnesium against cerebral palsy in very-low-birthweight infants (Nelson and Grether, 1995; Schendel and colleagues, 1996). Murphy and colleagues (1995) found that preeclampsia, rather than magnesium sulfate, was protective against cerebral palsy in these infants. Kimberlin and colleagues (1996), however, found no advantage of maternal magnesium sulfate tocolysis in infants born weighing less than 1000 g.

CLINICAL EFFICACY OF MAGNESIUM SULFATE THERAPY. In 1995, results were reported from the multinational clinical trial of eclampsia therapy. The Eclampsia Trial Collaborative Group (1995) study was funded in part by the World Health Organization and coordination was provided by the National Perinatal Epidemiology Unit in Oxford, England. This study included 1687 women with eclampsia who were randomly allocated to different anticonvulsant regimens. The primary outcome measures were recurrence of convulsions and maternal deaths. In one study, 453 women were randomized to be given magnesium sulfate and compared with 452 given diazepam. Another 388 eclamptic women were randomized to be given magnesium sulfate and compared with 387 women given phenytoin.

As shown in Figure 24-18, women allocated to magnesium sulfate therapy had a 50 percent reduction in incidence of recurrent seizures compared with those given diazepam. Importantly, as shown in Table 24-11, maternal deaths were reduced in women given magnesium sulfate, and although these differences are clinically impressive, they are not statistically significant. Specifically, there were 3.8 percent deaths in 453 women randomized to magnesium sulfate compared with 5.1 percent of 452 given diazepam. Maternal and perinatal morbidity were not different between these two groups, and there was no difference in the number of labor inductions or cesarean deliveries.

In a second comparison, also shown in Figure 24-18, women randomized to receive magnesium sulfate compared with phenytoin had a 67 percent reduction in recurrent convulsions. As shown in Table 24-11, maternal mortality was lower in the magnesium compared with the phenytoin group (2.6 versus 5.2 percent). This clinically impressive decreased maternal mortality of 50 percent again was not significant statistically.

In other comparisons, women allocated to magnesium sulfate therapy were less likely to be artificially ventilated, to develop pneumonia, and to be admitted to intensive care units than those given phenytoin. Neonates of women given magnesium sulfate were significantly less likely to require intubation at delivery and to be admitted to the neonatal intensive care unit compared with infants whose mothers received phenytoin.

The Collaborative Group concluded: "There is now compelling evidence in favour of magnesium sulfate, rather than diazepam or phenytoin, for the treatment of eclampsia." These results are even more impressive when it is emphasized that women in this study who received intravenous magnesium sulfate received only 1 g per hour!

ANTIHYPERTENSIVE TREATMENT ALONE TO CONTROL SEIZURES. There continues to be a question of whether antihypertensive medications alone can be used to prevent recurrent seizures (Duley and Johanson, 1994; Ramsay and colleagues, 1994). It is difficult and unwise to recommend withholding magnesium sulfate because even with the excellent results reported using both anticonvulsants and antihypertensive agents in the Collaborative Eclampsia Trial (1995), maternal mortality was still 4.1 percent of 1690 women.

PREVENTION OF ECLAMPSIA. Magnesium sulfate therapy also is superior to phenytoin in preventing eclamptic seizures. Lucas and colleagues (1995) reported results of a prospective study from Parkland Hospital in which women with gestational hypertension were randomized to receive magnesium sulfate or phenytoin during labor. The magnesium sulfate therapy consisted of the intramuscular regimen presented in Table 24-10. The phenytoin regimen consisted of a 1000-mg loading dose infused over a 1-hour period, followed by a 500-mg oral dose 10 hours later. Anticonvulsant therapy in both groups was continued for 24 hours postpartum. Ten of the 1089 women randomly assigned to the phenytoin regimen had eclamptic convulsions. There were no convulsions in the 1049 women given magnesium sulfate (P = .004). There were no significant differences in any risk factors for eclampsia between the two groups of women studied. Maternal and neonatal outcomes were similar in the two study groups. Women given phenytoin, and who developed eclampsia, did so despite "therapeutic" serum levels (10 to 25 ug/mL).

Debate still ensues—mostly outside the United States—over whether magnesium sulfate prophylaxis should be given routinely to all women in labor who have hypertension (Nelson and Grether, 1995; Robson, 1996). Burrows and Burrows (1995) described 467 women with preeclampsia in whom seizure prophylaxis was not given. A total of 3.9 percent developed eclamptic seizures. Hall and colleagues (2000a) reported this to be 1.5 percent in 318 preeclamptic women. This is in contrast to a failure rate of prophylaxis of 1 in 750 women treated with the regimen shown in Table 24-10 and reported by Cunningham and Leveno (1988). The debate hinges primarily on the position of some clinicians that a convulsion due to eclampsia does no immediate great harm to most mothers and fetuses. Presumably, those who embrace this view would reserve magnesium sulfate therapy for only those women who develop eclampsia. There may be, however, other previously unmeasured fetal effects of maternal eclampsia that could justify attempts to prevent as well as treat eclampsia. For example, Hallak and associates (1999a) have reported in experimental animals that maternal seizures were associated with fetal brain injury due to maternal hypoxia during the convulsion. Magnesium sulfate prevented this fetal damage.

The debate currently centers around which preeclamptic women should be given prophylaxis. To assess this, Coetzee and colleagues (1998) randomized 699 South African women with severe preeclampsia to intravenous magnesium sulfate or to saline placebo. Eclampsia developed in 1 of 30 of the women given saline, and although the maternal and fetal outcomes were good, the study was stopped. Benefits of prophylactic magnesium sulfate for women with mild preeclampsia is disputable because the estimated risk of eclampsia is 1 in 100 or less (Lucas and colleagues, 1995). Witlin and Sibai (1998) recently reviewed the evidence for the efficacy of magnesium sulfate for treatment and prevention of convulsions resulting from hypertensive disorders due to pregnancy. They concluded that although there is little question about the benefits of magnesium sulfate in women with severe preeclampsia and eclampsia, the need for its prophylactic administration to women with mild disease is unclear. We certainly agree that it is prudent to prevent eclampsia in women at high risk for convulsions.

HYDRALAZINE TO CONTROL SEVERE HYPERTENSION. At Parkland Hospital, hydralazine is given intravenously whenever the diastolic blood pressure is 110 mm Hg or higher. Some recommend treatment of diastolic pressures higher than 100 mm Hg and some use 105 mm Hg as a limit (Cunningham and Lindheimer, 1992; Sibai, 1996). Because precise data are lacking, the Working Group (National High Blood Pressure Education Program, 2000) recommended the compromise value of persistent systolic pressure greater than or equal to 160 mm Hg and/or diastolic pressure greater than 105 mm Hg.

A number of regimens have been used. Hydralazine is administered in 5- to 10-mg doses at 15- to 20-minute intervals until a satisfactory response is achieved. A satisfactory response antepartum or intrapartum is defined as a decrease in diastolic blood pressure to 90 to 100 mm Hg, but not lower lest placental perfusion be compromised. Hydralazine so administered has proven remarkably effective in the prevention of cerebral hemorrhage. At Parkland Hospital, approximately 8 percent of all women with hypertensive disorders are given hydralazine as described, and we estimate that more than 4000 women have been treated. Seldom was another antihypertensive agent needed because of poor response to hydralazine. In many European centers, hydralazine is also favored (Redman and Roberts, 1993).

The tendency to give a larger initial dose of hydralazine when the blood pressure is higher must be avoided. The response to even 5- to 10-mg doses cannot be predicted by the level of hypertension; thus we always give 5 mg as the initial dose. An example of very severe hypertension in a woman with chronic hypertension complicated by superimposed eclampsia that responded to repeated intravenous injections of hydralazine is shown in Figure 24-19. Hydralazine was injected more frequently than recommended in the protocol, and blood pressure decreased in less than 1 hour from 240-270/130-150 mm Hg to 110/80 mm Hg. Uteroplacental insufficiency fetal heart rate decelerations were evident when the pressure fell to 110/80 mm Hg, and persisted until maternal blood pressure increased.

LABETALOL. Intravenous labetalol is also used to treat acute hypertension of pregnancy. In 1994, hydralazine was not always available in the United States and more experience was gained with this a₁- and nonselective b-blocker. Mabie and associates (1987) compared intravenous hydralazine with labetalol for blood pressure control in 60 peripartum women. Labetalol lowered blood pressure more rapidly, and associated tachycardia was minimal, but hydralazine lowered mean arterial pressure to safe levels more effectively. We have evaluated labetalol given intravenously for women with severe preeclampsia and our results are very similar. Our protocol calls for 10 mg intravenously initially. If the blood pressure has not decreased to the desirable level in 10 minutes, then 20 mg is given. The next 10-minute incremental dose is 40 mg followed by another 40 mg, and then 80 mg if a salutary response is not yet achieved.

The Working Group (2000) recommends starting with a 20-mg intravenous bolus. If not effective within 10 minutes, this is followed by 40 mg, then 80 mg every 10 minutes but not to exceed a 220-mg total dose.

OTHER ANTIHYPERTENSIVE AGENTS. The Working Group recommends nifedipine in a 10-mg oral dose to be repeated in 30 minutes if necessary. Scardo and colleagues (1996) gave 10-mg nifedipine orally to 10 women with preeclamptic hypertensive emergencies and reported no hypotension or fetal compromise. Mabie and colleagues (1988) administered nifedipine sublingually to 34 women with peripartum hypertension. Its antihypertensive effects were potent and rapid, and two women developed worrisome hypotension. Similar effects in nonpregnant patients have caused cerebrovascular ischemia, myocardial infarction, conduction disturbances, and death, leading Grossman and colleagues (1996) to call for a moratorium on its use in hypertensive emergencies. Vermillion and colleagues (1999) and Scardo and co-workers (1999) compared nifedipine with labetalol in randomized trials and found neither definitively superior to the other. Belfort and associates (1990) administered the calcium antagonist, verapamil, by intravenous infusion at 5 to 10 mg per hour. Mean arterial pressure was lowered by 20 percent. Belfort and co-workers (1996) reported that nimodipine given by continuous infusion as well as orally (1998) was effective to lower blood pressure in women with severe preeclampsia. Bolte and colleagues (1998) reported good results in 169 preeclamptic women treated with intravenous ketanserin, a selective serotonin₂ receptor blocker.

Nitroprusside is not recommended by the Working Group of the National High Blood Pressure Education Program (2000) unless there is no response to hydralazine, labetalol, or nifedipine. A continuous infusion is begun with a dose of 0.25 ug/kg/min increased as necessary to 5 ug/kg/min. Fetal cyanide toxicity may occur after 4 hours.

PERSISTENT IMMEDIATE SEVERE POSTPARTUM HYPERTENSION. The potential problem of antihypertensive agents causing serious compromise of placental perfusion and fetal well-being is obviated by delivery. If there is a problem after delivery in controlling severe hypertension and intravenous hydralazine or another agent is being used repeatedly early in the puerperium to control persistent severe hypertension, then other regimens can be used. We have had success with intramuscular hydralazine, usually in 10- to 25-mg doses at 4- to 6-hour intervals. Once repeated blood pressure readings remain near normal, hydralazine is stopped.

If hypertension of appreciable intensity persists or recurs in these postpartum women, oral labetalol or a thiazide diuretic are given for as long as necessary. A variety of other antihypertensive agents have been utilized for this purpose, including other b-blockers and calcium-channel antagonists. The persistence or refractoriness of hypertension is likely due to at least two mechanisms:

1. Underlying chronic hypertension.

2. Mobilization of edema fluid with redistribution into the intravenous compartment.

Labetalol and a diuretic are effective treatment for both mechanisms.

PLASMA EXCHANGE. Over the years, the group at the University of Mississippi Medical Center has described an atypical syndrome in which severe preeclampsia-eclampsia persists despite delivery. Martin and associates (1995) described 18 such women over a 10-year period during which time they delivered nearly 43,000 patients. They advocate single or multiple plasma exchange for these women and in some cases, 3 L of plasma (representing 12 to 15 donors) were exchanged three times before a response was forthcoming. In our experiences of over 50,000 hypertensive women in nearly 350,000 pregnancies, we have not encountered this syndrome. In a very few women, persistent hypertension, thrombocytopenia, and renal dysfunction was found to be due to thrombotic microangiopathy (Dashe and colleagues, 1998).

DIURETICS AND HYPEROSMOTIC AGENTS. Potent diuretics further compromise placental perfusion, because their immediate effects include intravascular volume depletion, which most often is already reduced compared with normal pregnancy. Therefore, diuretics are not used to lower blood pressure lest they enhance the intensity of the maternal hemoconcentration and its adverse effects on the mother and the fetus (Zondervan and co-authors, 1988).

Once delivery is accomplished, in almost all cases of severe preeclampsia and eclampsia there is a spontaneous diuresis that usually begins within 24 hours and results in the disappearance of excessive extravascular fluid over the next 3 to 4 days, as demonstrated in Figure 24-13.

With infusion of hyperosmotic agents, the potential exists for an appreciable intravascular influx of fluid and, in turn, subsequent escape of intravascular fluid in the form of edema into vital organs, especially the lungs and brain. Moreover, an osmotically active agent that leaks through capillaries into lungs and brain promotes accumulation of edema at these sites. Most importantly, a sustained beneficial effect from their use has not been demonstrated. For all of these reasons, hyperosmotic agents have not been administered, and use of furosemide or similar drugs has been limited to the rare instances in which pulmonary edema was identified or strongly suspected.

FLUID THERAPY. Lactated Ringer solution is administered routinely at the rate of 60 mL to no more than 125 mL per hour unless there was unusual fluid loss from vomiting, diarrhea, or diaphoresis, or more likely, excessive blood loss at delivery. Oliguria, common in cases of severe preeclampsia and eclampsia, coupled with the knowledge that maternal blood volume is very likely constricted compared with normal pregnancy, makes it tempting to administer intravenous fluids more vigorously. The rationale for controlled, conservative fluid administration is that the typical eclamptic woman already has excessive extracellular fluid that is inappropriately distributed between the intravascular and extravascular spaces. Infusion of large fluid volumes could and does enhance the maldistribution of extravascular fluid and thereby appreciably increases the risk of pulmonary and cerebral edema (Sibai and colleagues, 1987b).

PULMONARY EDEMA. Women with severe preeclampsia-eclampsia who develop pulmonary edema most often do so postpartum (Cunningham and colleagues, 1986). Aspiration of gastric contents, the result of convulsions or perhaps from anesthesia, or oversedation, should be excluded; however, the majority of these women have cardiac failure. Some normal pregnancy changes, magnified by preeclampsia-eclampsia, predispose to pulmonary edema. Importantly, plasma oncotic pressure decreases appreciably in normal term pregnancy because of decreases in serum albumin, and oncotic pressure falls even more with preeclampsia (Zinaman and associates, 1985). Moreover, Oian and colleagues (1986) described increased extravascular fluid oncotic pressure in preeclamptic women, and this favors capillary fluid extravasation. Brown and associates (1989) verified increased capillary permeability in preeclamptic women. Bhatia and associates (1987) found a correlation between plasma colloid osmotic pressure and fibronectin concentration; this suggested to them that vascular protein loss was the result of increased vascular permeability caused by vessel injury.

The frequent findings of hemoconcentration, as well as the identification of reduced central venous and pulmonary capillary wedge pressures in women with severe preeclampsia, have tempted some investigators to infuse various fluids, starch polymers, or albumin concentrates, or all three, in attempts to expand blood volume and thereby somehow to relieve vasospasm and reverse organ deterioration. Thus far, clear-cut evidence of benefits from this approach is lacking; however, serious complications, especially pulmonary edema, have been reported. Lopez-Llera (1982) reported that vigorous volume expansion was associated with a high incidence of pulmonary edema in more than 700 eclamptic women. Benedetti and colleagues (1985) described pulmonary edema in 7 of 10 severely preeclamptic women who were given colloid therapy. Sibai and colleagues (1987b) cited excessive colloid and crystalloid infusions as causing most of their 37 cases of pulmonary edema associated with severe preeclampsia-eclampsia. Finally, Lehmann and co-workers (1987) reported that pulmonary edema caused nearly a third of maternal deaths due to hypertensive disorders at Charity Hospital in New Orleans.

For these reasons, until it is understood how to contain more fluid within the intravascular compartment and, at the same time, less fluid outside the intravascular compartment, we remain convinced that, in the absence of marked fluid loss, fluids can be administered safely only in moderation. To date, no serious adverse effects have been observed from such a policy. Importantly, dialysis for renal failure was not required for any of the more than 400 cases of eclampsia so managed.

INVASIVE HEMODYNAMIC MONITORING. Much of what has been learned within the past decade about cardiovascular and hemodynamic pathophysiological alterations associated with severe preeclampsia-eclampsia has been made possible by invasive hemodynamic monitoring using a flow-directed pulmonary artery catheter. The need for clinical implementation of such technology for the woman with preeclampsia-eclampsia, however, has not been established. Gilbert and colleagues (2000) recently described a retrospective review of pulmonary artery catheterization in 17 women with eclampsia. Although they found this procedure subjectively "helpful" in clinical management, all of these women had undergone "multiple interventions," including volume expansion, prior to catheterization.

Use of pulmonary-artery catheterization has been reviewed by Nolan and colleagues (1992), Hankins and Cunningham (1991), and Clark and Cotton (1988). Two conditions frequently cited as indications for such monitoring are preeclampsia associated with oliguria and preeclampsia associated with pulmonary edema. Perhaps somewhat paradoxically, it is usually vigorous treatment of the former that results in most cases of the latter.

Because vigorous intravenous hydration and osmotically active agents are avoided at Parkland Hospital in women with severe preeclampsia and eclampsia, hemodynamic monitoring has not been used for the vast majority of these women. Such measures are usually reserved for women with accompanying severe cardiac disease and/or renal disease or in cases of refractory hypertension, oliguria, and pulmonary edema. Similar indications are used by Clark and associates (1997), Clark and Cotton (1988), Cowles (1994), and Easterling and co-workers (1989), as well as recommended by the American College of Obstetricians and Gynecologists (1988, 1996). The routine use of such monitoring even if pulmonary edema develops is questionable. Most of these women respond quickly to furosemide given intravenously. Afterload reduction with intermittent doses of intravenous hydralazine to lower blood pressure, as described earlier, may also be necessary, because women with chronic hypertension and severe superimposed preeclampsia are more likely to develop heart failure (Cunningham and colleagues, 1986). Obese women in these circumstances are even more likely to develop heart failure (Mabie and associates, 1988).

Invasive monitoring should be considered for those women with multiple clinical factors such as intrinsic heart disease and/or advanced renal disease that might cause pulmonary edema by more than one mechanism. This is particularly relevant if pulmonary edema is inexplicable or refractory to treatment. Still, in most of these cases it is not necessary to perform pulmonary artery catheterization for clinical management.

DELIVERY. To avoid maternal risks from cesarean delivery, steps to effect vaginal delivery are employed initially in women with eclampsia. After an eclamptic seizure, labor often ensues spontaneously or can be induced successfully even in women remote from term. An immediate cure does not immediately follow delivery by any route, but serious morbidity is less common during the puerperium in women delivered vaginally.

BLOOD LOSS AT DELIVERY. Hemoconcentration, or lack of normal pregnancy-induced hypervolemia, is an almost predictable feature of severe preeclampsia-eclampsia. These women, who consequently lack normal pregnancy hypervolemia, are much less tolerant of blood loss than are normotensive pregnant woman. It is of great importance to recognize that an appreciable fall in blood pressure very soon after delivery most often means excessive blood loss and not sudden dissolution of vasospasm. When oliguria follows delivery, the hematocrit should be evaluated frequently to help detect excessive blood loss that, if identified, should be treated appropriately by careful blood transfusion.

ANALGESIA AND ANESTHESIA

In the past, both spinal and epidural analgesia were avoided in women with severe preeclampsia and eclampsia. Physiological changes leading to these concerns centered on the hypotension induced by sympathetic blockade and, in turn, on dangers from pressor agents or large volumes of intravenous fluid used to correct iatrogenically induced hypotension. For example, rapid infusion of large volumes of crystalloid or colloid, given to counteract maternal hypovolemia caused by a variety of causes, including epidural analgesia, has been implicated as a cause of pulmonary edema (Sibai and colleagues, 1987b). There have also been concerns about fetal safety because sympathetic blockade-induced hypotension can dangerously lower uteroplacental perfusion (Montan and Ingemorsson, 1989). Another concern is that attempts to restore blood pressure pharmacologically with vasopressors may be hazardous because women with preeclampsia are extremely sensitive to such agents.

As regional analgesia techniques were improved during the past decade, epidural analgesia was promoted by some proponents for women with severe preeclampsia to ameliorate vasospasm and lower blood pressure (Gutsche and Creek, 1993). Moreover, many who favored epidural blockade believed that general anesthesia was inadvisable because stimulation caused by tracheal intubation may result in sudden hypertension, which may cause pulmonary edema, cerebral edema, or intracranial hemorrhage (Lavies and colleagues, 1989). Others have also cited that tracheal intubation may be particularly hazardous in women with airway edema due to preeclampsia (Chadwick and Easterling, 1991).

These differing perspectives on the advantages, disadvantages, and safety of the anesthetic method used in the cesarean delivery of women with severe preeclampsia have evolved so that most authorities believe that epidural analgesia is the preferred method. Wallace and colleagues (1995) evaluated these important issues by conducting a randomized investigation in women with severe preeclampsia cared for at Parkland Hospital. There were 80 women with severe preeclampsia who were to be delivered by cesarean and who were randomized to general anesthesia or epidural or combined spinal-epidural analgesia. Their mean preoperative blood pressure was approximately 170/110 mm Hg, and all had proteinuria. Anesthetic and obstetrical management included antihypertensive drug therapy and limited intravenous fluids and other drug therapy. The infants, whose mean gestational age at delivery was 34.8 weeks, all were born in good condition as assessed by Apgar scores and umbilical arterial blood gas determinations. Maternal hypotension resulting from regional analgesia was managed without excessive intravenous fluid administration. Similarly, maternal blood pressure was managed without severe hypertensive effects in women undergoing general anesthesia (Fig. 24-20). There were no serious maternal or fetal complications attributable to any of the three anesthetic methods. It was concluded that general as well as regional anesthetic methods are equally acceptable for cesarean delivery in pregnancies complicated by severe preeclampsia if steps are taken to ensure a careful approach to either method.

The immense popularity and increasing availability of epidural analgesia for labor has led many anesthesiologists as well as obstetricians to develop the viewpoint that epidural analgesia is an important factor in the intrapartum treatment of women with preeclampsia (Chadwick and Easterling, 1991; Ramanatham, 1991). Although epidural analgesia is widely used during labor in women with preeclampsia, the effects of such analgesia on the mother and fetus have not been extensively investigated. In a study by Lucas and associates (2001) from Parkland Hospital, 738 laboring women at 36 weeks or more who had gestational hypertension of varying severity were randomized to epidural analgesia or patient-controlled intravenous meperidine analgesia. Maternal and infant outcomes were similar in the two study groups. As shown in Table 24-12, although epidural analgesia significantly lowered the maternal blood pressure compared with the meperidine-treated group, this provided no significant benefit in terms of preventing severe hypertension later in labor. It was concluded that epidural analgesia during labor was safe for women with pregnancy-associated hypertensive disorders; it should not be misconstrued to be a therapy for hypertension.

These results were similar to those published by the Maternal-Fetal Medicine Units Network of the National Institute of Child Health and Human Development. Hogg and colleagues (1999) performed a retrospective comparison of 327 women with severe hypertension, of whom 209 (65 percent) were given epidural analgesia during labor. There were no significant differences in the cesarean delivery rate or in neonatal outcomes in those given epidural analgesia compared with those not receiving such analgesia. Importantly, 4.3 percent of the women given epidural analgesia developed pulmonary edema, presumably due in part to the intravenous crystalloid required to alleviate epidural-induced hypotension.

Newsome and colleagues (1986) studied the lowered mean arterial pressure that follows epidural blockade in a group of women with severe preeclampsia. Despite decreased blood pressure, the cardiac index did not fall and intravenous fluid loading caused elevation of pulmonary capillary wedge pressures as compared with women in whom fluids were restricted. It is clear that aggressive volume replacement in these women increases their risk for pulmonary edema, especially in the first 72 hours postpartum (Clark and colleagues, 1985; Cotton and associates, 1986a). When pulmonary edema develops there is also concern for development of cerebral edema. Finally, Heller and co-workers (1983) demonstrated that the majority of cases of pharyngolaryngeal edema seen in women with severe preeclampsia were related to aggressive volume therapy.

LONG-TERM CONSEQUENCES

Women who develop hypertension during pregnancy should be evaluated during the immediate postpartum months and counseled about future pregnancies and also their cardiovascular risk later in life (Working Group of the National High Blood Pressure Education Program, 2000). As mentioned earlier, the longer hypertension diagnosed during pregnancy persists postpartum, the greater the likelihood that the cause is underlying chronic hypertension. Indeed, the Working Group has taken the position that hypertension attributable to pregnancy must resolve within 12 weeks of delivery. Persistence of hypertension beyond this time is considered evidence of chronic hypertension (Chap. 45).

Although hypertension during pregnancy is common in occurrence, there are few reports concerning the long-term maternal consequences of hypertensive disorders during pregnancy. Most reports are focused on the risk of recurrence of hypertension during a subsequent pregnancy in women typically ascertained during the first months of such a pregnancy.

COUNSELING FOR FUTURE PREGNANCIES

Women who had preeclampsia are more prone to hypertensive complications in future pregnancies (Working Group of the National High Blood Pressure Education Program, 2000). Generally the earlier preeclampsia is diagnosed during the index pregnancy, the greater the likelihood of recurrence. For example, Sibai and colleagues (1986, 1991) found that nulliparous women diagnosed to have preeclampsia before 30 weeks have a recurrence risk as high as 40 percent during a subsequent pregnancy. The recurrence rate for women with one episode of HELLP¹ syndrome is approximately 5 percent (Sibai and colleagues, 1995). Multiparous women who develop preeclampsia are at increased risk for recurrence of preeclampsia in subsequent pregnancy compared with nulliparas who develop preeclampsia (Trupin and associates, 1996).

Women with early-onset severe preeclampsia may be at risk for underlying thrombophilias such as factor V Leiden, protein S and C deficiency, and antiphospholipid antibodies, (Dekker and colleagues, 1995; Kupferminc and associates, 1999; Sibai, 1999; van Pampus and co-workers, 1999). These disorders, discussed in Chapter 49 (p. 1330), not only complicate subsequent pregnancies but also have an impact overall long-term health.

LONG-TERM PROGNOSIS

There is one truly remarkable and unparalleled effort to measure long-term maternal consequences of hypertension due to pregnancy. Chesley and co-workers (1976) identified 270 women with eclampsia delivered at the Margaret Hague Maternity Hospital between 1931 and 1951 and meticulously followed these women through 1974—entailing a follow-up period of up to 43 years in some of the women. As shown in Table 24-13, the long-term cardiovascular prognosis depends on whether eclampsia occurred in nulliparous, compared with multiparous women. Chesley and co-workers (1976) also analyzed the long-term outcome in 54 nulliparous women with eclampsia in their index pregnancy and hypertension again during a subsequent pregnancy compared with 100 eclamptic nulliparous who were normotensive during all subsequent pregnancies. Those with recurrent pregnancy hypertension were at increased risk for chronic hypertension whereas those who remained normotensive during subsequent pregnancies were at decreased risk. Thus, normal blood pressure during subsequent pregnancies serves to define women not at risk for future chronic hypertension. According to the Working Group of the National High Blood Pressure Education Program (2000), women experiencing normotensive births have a reduced risk for remote hypertension. Thus, in some respects, repeated pregnancy serves as a screening test for future hypertension. Finally, and importantly, preeclampsia does not cause chronic hypertension (Fisher and colleagues, 1981).