Keywords
preeclampsia, epidemiology, incidence, recurrence, risk factors, cardiovascular disease, placenta
Editors’ comment: The epidemiology literature relating to hypertension in pregnancy grew substantially since the third edition, much of it focusing on preeclampsia’s relationship to subsequent cardiovascular and renal disease. There is now little doubt that formerly preeclamptic women have significantly more remote disease, especially if their preeclampsia occurred preterm, when compared to women with normotensive deliveries, but whether preeclampsia is a “marker” of already at risk populations, or that preeclampsia itself is causal, remains to be determined. Either way there is a need for balanced programs that focus on prevention, forestalling, and early intervention for this at-risk population. Finally, we welcome a new first author, Janet Rich-Edwards an epidemiologist and co-director of the Reproductive, Perinatal, and Pediatric Track at the Harvard University School of Public Health. Her research focus includes effects of the perinatal environment on future cardiovascular disease in mothers and their offspring.
Introduction
Hypertensive disorders specific to pregnancy include gestational hypertension (formerly known as transient hypertension of pregnancy), preeclampsia, and eclampsia. These syndromes, marked by elevations in blood pressure that return to normal after delivery, are distinguished clinically by the presence (preeclampsia and eclampsia) or absence (gestational hypertension) of proteinuria or systemic findings. Eclampsia is a more life-threatening syndrome, characterized by seizures and thought to represent the progression of preeclampsia. Preeclampsia arising in the context of chronic hypertension preceding pregnancy is known as superimposed preeclampsia; in this case, blood pressures do not return to normal after pregnancy. As many women’s blood pressures are unknown before pregnancy, when hypertension fails to resolve after pregnancy, superimposed preeclampsia is assumed post hoc .
Here, we will review the epidemiologic data that characterize the frequency of occurrence, risk factors and predictors for, and the natural history and later life implications of hypertensive disorders in pregnancy. It is the purpose of this chapter to review these epidemiologic data, critically assess the methods used in conducting previous studies, and suggest new research areas, focusing on preeclampsia. We will review evidence suggesting that preeclampsia, in itself, is a syndrome composed of at least two underlying pathophysiologies, usually interactive. One pathophysiological pathway may be the result of reduced placental perfusion and is the component of preeclampsia unique to pregnancy. The other may be related to preexisting maternal pathology, in particular, an often subclinical maternal predisposition to cardiovascular risk. We have proposed in the past that these causes converge to predispose to preeclampsia, with its hallmark placental lesions and deficit of angiogenesis, by the common mechanism of endothelial dysfunction. The combination of both maternal and placental factors is particularly damaging. As such, we will discuss the natural heterogeneity among the risk factors and clinical predictors as well as among the outcomes for mother and infant within the spectrum of preeclampsia.
Definitions of the Hypertensive Disorders of Pregnancy
There are differences in reports of the frequency with which preeclampsia and gestational hypertension occur, due in part to disparities in the many definitions used to classify the hypertensive disorders of pregnancy. In recent years, there has been increasing convergence in definitions of preeclampsia, including those of the American College of Obstetrics and Gynecology (ACOG) Task Force on Hypertension in Pregnancy and the International Society for the Study of Hypertension in Pregnancy. Both bodies require the new onset of systolic blood pressure of≥140 mm Hg or a diastolic blood pressure of≥90 mm Hg as the foundation for both gestational hypertension and preeclampsia. Both definitions recognize the combination of hypertension and proteinuria as preeclampsia, but both bodies now recognize new definitions of preeclampsia that do not require proteinuria. In recognition of the syndromic nature of preeclampsia other organ involvement may substitute for proteinuria and can be used to complete the diagnosis of preeclampsia. They recommended documentation of the elevated blood pressure and proteinuria on at least two occasions. For research purposes, both bodies require documentation of normotension before 20 weeks gestation and after 12 weeks postpartum; in 2000, the International Society for the Study of Hypertension in Pregnancy (ISSHP) recommended researchers increase the specificity of preeclampsia diagnosis by requiring proteinuria. Other, less well-accepted definitions, often including clinical symptoms as diagnostic criteria, have also been used. Thus, the frequency of hypertensive disorders has been variously estimated based on different definitions that vary across geography and time. It is likely that the changes in diagnostic criteria by the ACOG and ISSHP will increase the frequency of preeclampsia.
Most, but not all, studies of hypertension in pregnancy exclude women with preexisting hypertension, so that few estimates of the prevalence of preeclampsia include superimposed preeclampsia. Recently, studies have begun to consider superimposed preeclampsia in its own right, due to its apparent increase with rising obesity and chronic hypertension in young women.
Recently, the ISSHP issued consensus definitions for severe and early-onset preeclampsia. ‘Severe’ preeclampsia required systolic blood pressure>160 mm Hg or diastolic blood pressure>110 mm Hg, with evidence of proteinuria (spot urine>30 mg/mmol creatinine). Degree of proteinuria was not deemed a criterion for severity of preeclampsia. The ACOG definition of severe preeclampsia has similar blood pressure criteria, but departs from the ISSHP definition by not requiring proteinuria, and instead requiring any of the following: thrombocytopenia, impaired liver function, renal insufficiency, pulmonary edema, or new onset cerebral of visual disturbances. The ISSHP has defined early onset preeclampsia as that arising before 34 weeks gestation. “Preterm preeclampsia” was defined as that occurring between 34+1 to<37+0 gestation, and “term” preeclampsia as that occurring after 37 weeks. It should be noted that most large studies use the gestational age at delivery, rather than the gestational age at diagnosis of preeclampsia, to distinguish “preterm” from “term” preeclampsia, as the latter is not usually available and is subject to the frequency of prenatal visits.
Finally, changes in the international classification of hypertensive pregnancy have made it difficult to compare preeclampsia prevalence across time and place; for example, the ICD-10 system combines mild preeclampsia and gestational hypertension, while the ICD-9 system did not, resulting in a seeming drop in the prevalence of preeclampsia when statistics are compared across old and new versions. With these caveats in mind, we review below the literature on the prevalence of the hypertensive disorders of pregnancy. Illustrative studies, organized by case ascertainment method, are presented in Table 3.1 .
| Study | Population | Size (ICD) | P(HDP) | P (GH) | P (PE) | P (ECL) |
|---|---|---|---|---|---|---|
| Hypertensive disorders ascertained by general population registry or administrative database | ||||||
| Wallis, 2008 | US National Hospital Discharge Survey, 1987–2004 | ~200,000 births per year (ICD-9) | 2.1% 3.0% in 2004 | 2.7% 3.2% in 2004 | 0.09% | |
| Goldenberg, 2011 | Iceland, USA, UK, Singapore since 1980 | 0.02–0.1% | ||||
| Roberts, 2011 | Alberta, Canada | 256,137 (ICD-10) | 6.0% | 1.4% | ||
| New South Wales, Australia | 732,288 (ICD-10) | 8.8% | 3.3% | |||
| Western Australia | 149,624 (ICD-10) | 9.1% | 2.9% | |||
| Denmark | 645,993 (ICD-10) | 3.6% | 2.7% | |||
| Norway | 456,353 (ICD-10) | 5.8% | 4.0% | |||
| Scotland | 531,622 (ICD-10) | 5.9% | 2.2% | |||
| Sweden | 913,779 (ICD-10) | 3.9% | 2.9% | |||
| Massachusetts, USA | 762,723 (ICD-9) | 7.0% | 3.3% | |||
| Gaio, 2001 | Brazilian Multicentre Cohort: 5 state capitals | 4,892 hospital discharge records | 7.5% | 0.7% | 2.3% | 0.16% |
| Roberts CL, 2005 | New South Wales, 2000–2002 | 250,173 | 9.8% | 4.3% | 4.2% | 0.06% |
| Hypertensive disorders ascertained by medical record review or study protocol (in placebo or standard care arms, where relevant) | ||||||
| Douglas, 1994 | UK, 1992. Case review. | All UK births | 0.05% | |||
| Knight, 2007 | UK, Feb 2005 – Feb 2006. Case review. | All UK births | 0.03% | |||
| Souza, 2013 | WHO hospital survey in 29 countries in Asia, Africa, and Latin America | 314,623 | 2.5% | 0.3% | ||
| Conde-Agudelo, 2000 | Perinatal information system, Latin America & Caribbean (ICD-10) | 878,680 | 4.8% | 0.2% | ||
| Sibai, 1993 | Aspirin trial: USA, healthy nulliparas | 1,565 | 5.9% | 6.3% | ||
| Villar, 2001 | WHO Antenatal Care trial: Argentina, Cuba, Saudi Arabia, Thailand | 11,121 | 5.0% | 1.3% | 0.08% | |
| Lumbiganon, 2007 | 22 hospitals, Mexico City | 18,288 | 5.5% | 0.6% | ||
| 18 hospitals, Thailand | 17,525 | 1.9% | 0.3% | |||
| Chalumeau, 2002 | 6 West African countries, 1994–1996 | 20,326 | 8.0% | 0.02% | ||
| Rotchell, 1998 | Barbados Aspirin Study: Maternity Hospital | 1,822 | 7.3% | 4.6% | ||
| Roberts, 2010 | Antioxidant trial: USA, healthy nullipara | 4,993 | 19.9% | 6.7% | 0.1% | |
| Hypertensive disorders defined by mixed methods | ||||||
| Abalos, 2013 | Africa (AFRO) | 93,613 (57% Nigerian) | 4.0% | 2.7% | ||
| Americas (AMRO) | 36,693,594 (99% U.S.) | 2.3% | 1.1% | |||
| Eastern Mediterranean (EMRO) | 148,909 | 1.2% | 0.5% | |||
| European (EURO) | 1,093,782 | 3.8% | 0.1% | |||
| Southeast Asian (SEARO) | 203,159 | 2.7% | 1.3% | |||
| Western Pacific (WPRO) | 361,402 (69% Australian) | 4.2% | 0.1% | |||
Prevalence of Hypertensive Disorders of Pregnancy
Eclampsia
The most reliable estimate of the prevalence of eclampsia is probably that reported from a national survey in the United Kingdom by Douglas and Redman. All obstetricians and all hospitals with an obstetrics unit were asked to participate in an active surveillance program in 1992. Each presumptive case was reviewed by a single obstetrician and was defined as eclamptic if there were seizures in the setting of hypertension, proteinuria, and either thrombocytopenia or an increased plasma aspartate transaminase concentration. The prevalence of eclampsia was estimated at 0.049% pregnancies. Most seizures occurred despite prenatal care (70%) and even after admission to the hospital (77%).
Table 3.1 shows a broad range of estimates of eclampsia incidence, ranging from 0.02% to 0.1% when ascertained in birth and other statistical registries and from 0.02% to 0.6% in studies with medical record review. The prevalence of eclampsia has declined rapidly during the 20th century in high-income countries, from prevalences of 0.3% or more before 1930 to prevalences of 0.03% or less in the past decade. Chesley showed a marked reduction at Margaret Hague Maternity Hospital in Jersey City from 1931 to 1951 ( Table 3.2 ). In recent decades, the risk of eclampsia dropped in the USA from an average annual rate of 0.10% in 1987–1995 to 0.08% of deliveries from 1996–2004, even as rates of preeclampsia have risen. As pointed out several times by Leon Chesley, the reduction in eclampsia is largely related to improved medical care rather than a changing natural history of preeclampsia. The prevalence of eclampsia and eclampsia-associated case-fatality in many lower-income countries remains high. A recent systematic review estimated rates as high as 2.7% in Africa and as low as 0.1% in Europe.
| 1931–1934 | 1935–1939 | 1940–1945 | 1946–1951 | Totals | |
|---|---|---|---|---|---|
| Registrations | 12,604 | 17,407 | 12,022 | 12,208 | 54,241 |
| Cases of eclampsia | 51 | 41 | 11 | 4 | 107 |
| % incidence | 0.40 | 0.23 | 0.09 | 0.03 | 0.20 |
Preeclampsia
Because of the decline in eclampsia in the developed world, much of the recent epidemiologic research has focused on preeclampsia. Table 3.1 shows estimates derived from registries, ranging from 1% to 4%, and from medical record review, ranging somewhat higher, at 1% to 7%. In the USA, where reporting of preeclampsia is not mandatory, Wallis, Saftlas and colleagues estimated the prevalence of hypertensive pregnancies from 1979 through 2004 using a nationally representative sample of hospital discharge records. Women with discharge diagnoses of gestational hypertension (ICD-9 642.3), preeclampsia (ICD-9 642.4 or 643.5) or eclampsia (ICD-9 643.6) were included. The authors estimated that preeclampsia complicated 2.9% of pregnancies in 2003–2004, a 25% increase since 1987. Similar prevalence of preeclampsia was estimated from hospital discharge data in the Danish National Birth Cohort of over 100,000 women recruited early in pregnancy, where the overall prevalence of preeclampsia was 3.0%, including 4.2% of nulliparous women and 1.3% of parous women.
The recent survey of multiple datasets published by Abalos (summarized in Table 3.1 ) reveals variability in regional preeclampsia rates, but does not suggest a pattern of especially high preeclampsia rates in lower-income regions. Observational studies and randomized clinical trials that follow women prospectively with the express intent of documenting hypertensive disorders of pregnancy with standardized protocols often report modestly higher prevalences of these disorders than are captured retrospectively by hospital discharge or birth registry databases. In a somewhat more selected population in the United States, the incidence of preeclampsia was estimated among control women enrolled in the NICHD Maternal Fetal Medicine Network for Clinical Trials (MFMU Network) trial of low-dose aspirin to prevent preeclampsia terminating in 1993. Included in the study were nulliparous women presenting to a series of academic medical centers for prenatal care. Women with a history of chronic hypertension, diabetes mellitus, renal disease, and other medical illnesses as well as women with a baseline blood pressure above 135/85 mm Hg were excluded. Among the 1500 women in the placebo group who were followed throughout pregnancy, 94 (6.3%) developed preeclampsia as defined by hypertension (systolic blood pressure of≥140 mm Hg or a diastolic blood pressure of≥90 mm Hg) plus proteinuria (either≥300 mg/24 h or 2+or more by dipstick on two or more occasions 4 hours apart). In a more recent 2010 study of more than 4500 control pregnancies using the same patient inclusion and exclusion and diagnostic criteria, the preeclampsia rate was remarkably similar, 6.7%. Other randomized clinical trials in the United States and Europe have demonstrated similar incidence rates for preeclampsia.
Superimposed preeclampsia , the development or worsening features of preeclampsia among women with chronic hypertension, has often been excluded from the study of hypertensive pregnancies. Estimates of the prevalence of superimposed preeclampsia are hard to come by, although many commentators have observed that it may be increasing as the prevalence of overweight and chronic hypertension rises among women of reproductive age. Among women with preexisting hypertension, preeclampsia will develop in 10–25%, as compared to a general population rate of 3–7%. The more severe and longstanding the hypertension prior to pregnancy, the greater is the risk of preeclampsia. In data from the MFMU Network, women with hypertension for at least 4 years duration had a remarkably high rate of preeclampsia: 31%.
Gestational hypertension has been especially difficult to characterize, as it is notoriously under-reported in registry databases. Estimated prevalences range from 1% to 4% from registries, but are considerably higher in studies that review medical records, on the order of 5% to 7% ( Table 3.1 ). Wallis et al.’s analysis of the U.S. National Hospital Discharge survey estimated that gestational hypertension complicated 3.0% of pregnancies in 2003–2004, a prevalence that had increased by 184% since 1987. Probably the best estimates come from trials. The MFMU Network aspirin trial in 1993 estimated that 5.9% of control women developed gestational hypertension (hypertension as defined above without proteinuria) but the percentage was much higher, 19.9%, in a 2010 MFMU Network study and in another 1993 NICHD trial of healthy US nulliparous women (17.3%). Rates in most other studies are in agreement with the aspirin trial.
Discussion of Differential Frequency Estimates
Hospital-based incidence estimates for preeclampsia systematically differ from national estimates. There are probably five main reasons for this discrepancy. First, many studies, such as that by Saftlas et al., rely on discharge diagnoses. Ales and Charlson performed a validation study of medical records at the New York Hospital and found that 25% of ICD-9 codes incorrectly diagnosed preeclampsia and that 53% of true cases were missed by ICD-9 coding. Similarly, Eskenazi et al. found that 47% of 263 women who received a discharge diagnosis of severe preeclampsia or eclampsia did not meet a rigorous set of criteria. Klemmensen et al. reported 2.9% prevalence of PE recorded in the Danish National Patient Registry, compared with 2.7% by medical record review and 3.4% by maternal recall. However, 26% of cases in the registry proved not to have preeclampsia by record review, and 31% of preeclampsia cases identified by record review were not recorded as such by the registry. The registry severely underestimated gestational hypertension compared with medical record review. Such under-reporting of milder forms of pregnancy hypertension is widely reported. Second, the MFMU Network studies included only nulliparous women, a group known to be at five-fold or greater risk of developing preeclampsia as compared to parous women. Third, women electing to enroll in a randomized clinical trial to prevent preeclampsia may also be a group with characteristics that would suggest a tendency to developing hypertension in pregnancy. However, the strict definition of preeclampsia used in the MFMU Network trials should have resulted in a decreased incidence estimate. Fourth, women seeking prenatal care at academic medical centers are a selected group who are probably at higher risk of developing pregnancy complications than would be reflected in a national sample. Fifth, data from the British Commonwealth in years past considered de novo hypertension without proteinuria as mild preeclampsia, and in these studies the diagnosis is substantially contaminated with women with transient hypertension. Nonetheless, a reasonable estimated range for the rate of preeclampsia in developed countries is 3–7%.
Risk Factors for Preeclampsia
Risk factors consistently shown to be associated with an increased rate of preeclampsia ( Table 3.3 ) include elevated prepregnancy or early pregnancy blood pressure, prepregnancy adiposity, age ( Fig. 3.1 ), family history of preeclampsia or of cardiovascular disease, African-American ethnicity ( Fig 3.2 ), preexisting medical conditions such as hypertension or diabetes, obstetric characteristics such as multiple gestation and hydrops fetalis, nulliparity, and history of a previous preeclamptic pregnancy ( Table 3.4 ).
| Risk factor ref. | Rating of Evidence |
|---|---|
| Older age | ++ |
| High blood pressure in 2nd/early 3rd trimester | ++ |
| Prepregnancy hypertension | +++ |
| Prepregnancy diabetes | +++ |
| Caloric excess | + |
| Elevated body mass index | +++ |
| Weight gain during pregnancy | + |
| Family history | +++ |
| Nulliparity | +++ |
| New paternity | ++ |
| Lack of previous abortion | ++ |
| Barrier contraception | ++ |
| Excessive placental size | +++ |
| Lack of smoking | +++ |
| Specific dietary factors | + |
| African American | + |
| Study | Population | Size (ICD) | P (GH) | P (PE) | P (ECL) |
|---|---|---|---|---|---|
| By parity | |||||
| Wu, 2009 | Denmark, 1978–2004 | 1.6 million | 6.3% nulliparous 1.9% parous | 0.10% nulliparous 0.03% parous | |
| Hernandez-Diaz, 2009 | Sweden, 1987–2004 | 763,795 (ICD-9 & 10) | 4.1% nulliparous 1.7% parous | ||
| By singleton versus multiple pregnancies | |||||
| Ros, 1998 | Sweden, 1987–1993 | 10,666 nulliparas (ICD-9) | 4.4% single 6.4% multi | 6.4% singleton 18.0% multiple pregnancy | |
| Douglas, 1994 | UK, 1992 | All UK births | 0.05% single 0.28% multi | ||
| Prevalence of preeclampsia in second pregnancies following first pregnancies that were preeclamptic | |||||
| P (PE) | |||||
| Mostello, 2008 | Missouri, USA, 1989–1997 | 103,860 | 14.7% Risk falling with increased length of 1st pregnancy: 12.9–38.6% | ||
| Lykke, 2009 | Denmark | 22,198 | 14.1–37.9% Risk falling with increased length of 1st pregnancy | ||
| Hernandez-Diaz, 2009 | Sweden, 1987–2004 | 763,795 (ICD-9 & 10) | 14.7% | ||
| Klungsoyr, 2012 | Norway, 1967–2006 | 2,416,501 | 15.2% | ||
| Hnat, 2002 | USA, 1991–1995 | 598 | 17.9% | ||
Most of these factors can be understood in relation to a maternal predisposition to cardiovascular disease. The last factors, obstetric characteristics and nulliparity, may represent the placental or uniquely pregnancy-related component of preeclampsia. For each of these factors, we will relate the data suggesting its association with preeclampsia. We will also evaluate each factor’s ability to accurately predict the development of preeclampsia.
Cardiovascular Risk Factors
Cardiovascular risk factors related to preeclampsia include: preexisting hypertension, elevated prepregnancy or early pregnancy blood pressure, family history of heart disease, diabetes mellitus, and in this category we also place African-American ethnicity, adiposity, lack of physical activity, and older age. Biochemical markers associated with cardiovascular risk, including elevated glucose/insulin, lipids, coagulation factors, inflammatory mediators, homocysteine, uric acid, and angiogenic and endothelial cell activation markers, also differentiate women developing preeclampsia from women with unaffected pregnancies. Many of these cardiovascular risk factors predate preeclamptic pregnancies, albeit at subclinical levels. Even in healthy nulliparous women, higher blood pressures prior to 27 weeks gestation are associated with an elevated risk of developing preeclampsia; the higher the blood pressure, the higher the risk of preeclampsia. Prepregnancy diabetes also increases the risk for developing preeclampsia. The more severe the diabetes prior to pregnancy, the greater is the risk of developing preeclampsia. Among women without preexisting diabetes, hyperglycemia and insulin resistance elevate the risk of developing preeclampsia. Women with polycystic ovarian syndrome (PCOS) have an excess of circulating androgens, are insulin resistant and have an increased frequency of preeclampsia. It is posited that increased insulin resistance explains the increased risk; however, in women without PCOS, preeclampsia is also associated with increased circulating androgens.
Body Mass Index
Elevations in body mass index, a marker of adiposity, have repeatedly been associated with risk of preeclampsia. Measured either at the first prenatal visit or prior to pregnancy, the magnitude of this effect ranges from 2.5 to 5-fold. This increase is in mild, severe, and early-onset preeclampsia, and is present in both black and white women. Even a modest increase in adiposity can increase preeclampsia risk. Women who fail to return to their prepregnant weight between pregnancies are at increased risk of preeclampsia in their second pregnancy compared to women who return to prepregnancy weight. Among morbidly obese women, those who obtained gastric bypass surgery before pregnancy had a 2.5% risk of preeclampsia or eclampsia, compared with a 14.5% risk among those who had their surgery after pregnancy; similar reductions were observed for gestational hypertension and superimposed preeclampsia/eclampsia. As with cardiovascular disease, it is likely that visceral obesity is most relevant. This is supported by the observation that waist circumference in excess of 80 cm, a more specific marker of biologically active abdominal fat, was associated with a 2.7-fold elevation in the development of preeclampsia in one recent study.
Physical Activity
Paralleling the known benefits of physical activity in reducing the risk of cardiovascular disease and type 2 diabetes mellitus, limited evidence suggests that physical activity may protect against preeclampsia. A recent review found that the role of exercise in reducing preeclampsia was supported in six observational studies, but not in prospective studies. Nonetheless, the weight of opinion favors a beneficial effect. The mechanisms are not established, but it has been shown that exercise can reduce weight gain during pregnancy. In a study of acute and chronic exercise in relationship to angiogenic activity, acute exercise in pregnant women was associated with an increase in the pro-angiogenic placental growth factor, and a reduction in the antiangiogenic factors, soluble endoglin and soluble fms-like tyrosine kinase.
Diet
Diet may also affect preeclampsia risk in a similar fashion to its effect on coronary artery disease risk. Women with low concentrations of erythrocyte membrane omega-3 (marine oil) fatty acids are reportedly more likely to have had preeclamptic pregnancies than women with high erythrocyte omega-3 fatty acids. Dietary intake studies largely from Scandinavia support an observation from older and smaller studies that increased sugar intake, largely from soft drinks, is associated with increased risk of preeclampsia. Interestingly, in a very large study the uncorrected effect of added sugar but not sugars in food was associated with an increased risk of preeclampsia. However, after adjustment for maternal age at delivery, education, prepregnant BMI, height, smoking, leisure exercise in the first pregnancy trimester, total energy intake and dietary fiber, only the effects of sweetened drinks persisted. A unique feature of such drinks is sweetening with high-fructose corn syrup, which has been suggested to have unique metabolic features increasing the risk of obesity and perhaps cardiovascular disease. In a study of dietary patterns from Norway, women with high intake of fresh and cooked vegetables, olive oil, fruits and berries and poultry had a reduced incidence of preeclampsia while women with an intake of predominantly processed food (including sweetened drinks) had an increased risk. Studies of micronutrients also suggest a relationship of increased risk with deficiencies of certain micronutrients. These deficits unfortunately have not yet resulted in effective therapy with supplements. Women who ingested less than 85 mg of vitamin C daily (below the recommended daily allowance) were at two-fold increased risk of developing preeclampsia, while women at high risk of preeclampsia with the highest quartile of vitamin C concentration at 18 weeks gestation had a 40% lower risk of preeclampsia. Nonetheless, there is no evidence that in unselected high- or low-risk populations the administration of large doses of vitamins C and E reduces the frequency of preeclampsia. One of the older relationships is that of low calcium intake and preeclampsia; however, replacement studies with up to 2 grams of calcium have had only a small effect. Other micronutrients may also be relevant. A self-reported history of ingestion of multivitamins prior to gestation was associated with a 71% reduction in preeclampsia. Interestingly, this reduction was not present in obese women.
Vitamin D
In recent years there has been extensive attention paid to the relationship of circulating concentrations of vitamin D to adverse pregnancy outcomes, including preeclampsia. Based on data from Pittsburgh, 25-hydroxyvitamin D (25(OH)D) was deficient (<37.5 nmol/L) in 5.0 % of white women and insufficient (37.5–80 nmol/L) in 42.1%. The frequency of inadequate circulating concentrations of 25(OH)D was even greater in black women, with 29.2% at deficient concentrations and 54.1% insufficient concentrations, perhaps reflecting the lower conversion of vitamin D from 7-dehydrocholesterol in dark skin or lower intake of vitamin D-rich or enriched foods or supplements. Lower plasma vitamin D concentrations have been posited as a potential explanation for the increase in adverse pregnancy outcomes in African-Americans. In keeping with the importance of ultraviolet light exposure, in northern countries with less ultraviolet exposure the rates of deficiency are even higher. In white individuals the rate of deficiency was 47%. A recent meta-analysis found that in 31 studies of the relationship of vitamin D with preeclampsia the rate of preeclampsia was 80% higher (1.79, 1.25 to 2.58) in individuals with vitamin D insufficiency. There are no large published randomized controlled trials of vitamin D to prevent preeclampsia. However, in a large Norwegian study it was possible to estimate vitamin D intake from diet and supplements and as vitamin D intake increased from 3 µg to 15–20 µg/day there was a 23% reduction in the frequency of preeclampsia. Nonetheless, increasing the dose of vitamin D by supplementation remains controversial, with two major scientific organizations presenting conflicting recommendations.
Smoking
The one traditional coronary artery disease risk factor associated with a reduction, rather than an elevation, in risk of preeclampsia is cigarette smoking. The reason for this inverse association with preeclampsia is unclear but may reflect an idiosyncratic effect of smoking on angiogenesis, a survival bias wherein the most affected fetuses are aborted and thus do not develop preeclampsia (M. Williams, personal communication), or an antiinflammatory effect of pregnancy. There is also evidence that the effect of smoking to reduce preeclampsia is not present in women beyond the age of 30. This could suggest that there is an early pharmacological effect to prevent preeclampsia, which is overcome by “damage” with more prolonged smoking. Although the identity of this pharmacological agent is not established, it does not appear to be nicotine. Unlike smoking, the intake of nicotine in snuff does not reduce preeclampsia.
Family Patterns
The best estimates of heritability and non-shared environmental effect on preeclampsia are an 0.54 index for heritability and a 0.46 index for non-shared environment, suggesting equivalent contributions of genes and environment in determining preeclampsia risk; corresponding estimates for gestational hypertension are 0.24 for heritability and 0.76 for non-shared environment. In the Swedish Twin Registry that examined female twin pairs in which at least one twin had a pregnancy complicated by preeclampsia, there were 16 of 61 monozygotic pairs in which both twins were affected (concordance rate 0.25); and 4 of 63 dizygotic pairs in which both sisters were affected (concordance rate 0.06), These results suggest an important role of maternal genes in determining preeclampsia, and a lesser role in determining gestational hypertension. These estimates rely on linkage of mothers and their sisters, and do not comment upon the contribution of paternal (and therefore fetal) genes.
That fathers and fetuses help drive the risk of preeclampsia is suggested by the observation that men who father preeclamptic pregnancies with one woman are at 80% higher likelihood of fathering a preeclamptic pregnancy with another woman. Furthermore, there appears to be a “mother-in-law effect” in which a son born of a preeclamptic pregnancy has an elevated chance of fathering a preeclamptic pregnancy. In fact, women born of a preeclamptic pregnancy have a 3-fold higher and men born of preeclamptic pregnancy have a 2-fold higher risk of triggering severe preeclampsia in their own (or their partner’s) pregnancy.
These patterns imply an interacting role for both maternal and fetal genes in preeclampsia. Further complexity is added by the possibility that the impact of parentally imprinted genes controlling trophoblast growth and fetal development depends on whether the genes are of maternal or paternal origin.
First Birth and Other Placental Factors
Nulliparity is a particularly strong risk factor for the development of preeclampsia ( Fig. 3.1 , Table 3.3 , Table 3.4 ). Large studies from Denmark and Sweden indicate that nulliparous women have preeclampsia risks on the order of 4% to 6%, in contrast to parous women, whose risk falls to roughly 2% ( Table 3.4 ). MacGillivray noted in 1958 in a well-characterized population in Scotland that preeclampsia occurred in 5.6% of nulliparas and in only 0.3% of secundi-paras. Other authors confirmed the observation that nulliparas were anywhere from five to ten times more likely to experience preeclampsia than parous women. Unadjusted data on the relationship between age and preeclampsia have frequently shown younger women to be at an increased risk. However, nulliparity may be the underlying factor driving this apparent association as younger women are more likely to be nulliparous. In fact, after accounting for parity, preeclampsia risk increases with maternal age ( Fig. 3.1 ). There is evidence that the risk associated with advanced maternal age may be narrowing over time.
The relationship between nulliparity and preeclampsia suggests something about the uniqueness of a first placentation vs. a later baby’s placental implantation. Redman suggested that in later pregnancies there is the development of protective immunologic mechanisms against paternal antigens. Other epidemiologic data support this suggestion. For example, studies have shown that the relative protection afforded by further pregnancies was reduced or eliminated with new paternity, that previous induced abortion was protective, that women using barrier methods of contraception were at increased risk, and that risk was reduced with increased duration of sexual activity antedating pregnancy. All of these risk factors would reduce maternal recognition of paternal antigens prior to pregnancy and all of the protective factors would enhance maternal recognition of paternal antigens. Other studies have reported that change in sexual partner elevated the risk of preeclampsia in a second pregnancy to a level almost as high as for first pregnancies. However, subsequent cohort studies demonstrated that a longer interval between pregnancies might account for the apparent partner change effect. Further supporting the antigenic theory are observations that teenage pregnancies, out-of-wedlock pregnancies, and donor sperm insemination each represent reproductive events wherein the female is likely to be relatively naïve to sperm antigens and each has been shown to increase the risk of preeclampsia. These observations must be viewed with caution since age, educational level, carriage of sexually transmitted infections, and access to healthcare may all confound the relationship between shorter-duration partnerships and preeclampsia.
Other observations also support the idea that novel exposure to paternal antigens may elevate preeclampsia risk. Women undergoing intrauterine insemination with washed donor sperm have a higher risk of preeclampsia than women undergoing intrauterine insemination with washed partner (autologous) sperm, implicating exposure to unfamiliar sperm. A particularly interesting comparison of in vitro fertilization or intracytoplasmic sperm injections (ICSI) with ICSI involving surgically obtained sperm showed a two-fold elevation in preeclampsia risk with the latter, suggesting that exposure to sperm, in the absence of semen, might increase preeclampsia risk.
Even if implantation is normal, relative placental perfusion may be reduced because of excessive placental size. Obstetric complications associated with increased placental size such as twin pregnancies, hydatidiform moles, and hydrops fetalis all markedly increase the risk of developing preeclampsia. Zhang et al. pooled the results from six studies that compared both twin and singleton gestations and found that pregnant women with twin pregnancies were three times more likely to develop hypertension in pregnancy than women with singletons ( Table 3.4 ). Women with these hypertensive twin pregnancies do not appear to have heightened risk of recurrence in subsequent singleton pregnancies.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
