Bodnar, Lisa M.*†‡; Catov, Janet M.*†‡; Klebanoff, Mark A.§; Ness, Roberta B.*†‡; Roberts, James M.*†‡
Obesity is a substantial and escalating health problem in the United States. It is the second leading cause of preventable death,1 and dramatically increases morbidity associated with many chronic diseases.2 About 6–7% of national health expenditures in the United States are attributable to obesity.3 Obesity affects approximately 1 in 3 women in their prime childbearing years.4 Maternal obesity is linked to a host of perinatal problems that increase risk of pregnancy complications and adverse birth outcomes.5
Among the extensive list of consequences of maternal obesity are preeclampsia and transient hypertension of pregnancy, 2 pregnancy-specific hypertensive disorders that are associated with high perinatal morbidity and mortality for both mother and infant.6,7 Prevention of preeclampsia and transient hypertension of pregnancy has remained elusive, owing largely to their complex, heterogeneous nature.8 Currently, maternal prepregnancy overweight is one of the strongest modifiable risk factors.9–14 Recent studies have shown a dose-dependent relation between prepregnancy BMI and the risk of preeclampsia.10,11
Inferences drawn from previous studies are almost entirely limited to preeclampsia and to transient hypertension of pregnancy with mild maternal symptoms. However, it is the severe subclassifications of these disorders that are associated with the highest relative risks of preterm birth, fetal growth restriction, and other adverse outcomes.15–19 Few investigators have studied BMI in relation to severe disease.12,20–23 Moreover, a majority of these studies have suffered from substantial methodologic limitations, including poor case definition,22 measurement of BMI during gestation rather than prepregnancy,12,21 or lack of control for important confounders.22 None of these has examined racial/ethnic differences in this relation. Understanding the associations between prepregnancy BMI and the risk of severe preeclampsia and severe transient hypertension of pregnancy among black and white women will allow us to determine if the rising prevalence of obesity may predict an escalation in the incidence of severe hypertensive disorders and ensuing maternal and infant morbidity.
Our objective was to assess the independent relation between maternal prepregnancy BMI and the risk of both the mild and severe subclassifications of preeclampsia and transient hypertension of pregnancy among white and black women.
Data came from the Collaborative Perinatal Project (1958–1964).24 Women who attended prenatal care at 12 hospitals in the United States were enrolled into this large prospective study. At enrollment, demographic, socioeconomic, and behavioral information and medical history were collected by in-person interview. Detailed data were also collected at each prenatal visit, during labor and delivery and during the postpartum period.
Over 55,000 pregnancies were included in the study. We selected singleton pregnancies from women who identified their race/ethnicity as non-Hispanic white or non-Hispanic black, gave birth from 20 to 45 weeks’ gestation, and did not have chronic hypertension or elevated blood pressure before 24 weeks’ gestation (n = 43,890). From this sample, we excluded women with implausible height or prepregnancy weight values (n = 236) and women who were missing data on prepregnancy weight or height (n = 3866), preeclampsia or transient hypertension of pregnancy (n = 82), or covariates in the final model (n = 1518). The final analytic sample was 38,188. This sample was similar to the entire cohort with respect to the incidence of severe and mild preeclampsia and transient hypertension of pregnancy, prepregnancy BMI, age, race, socioeconomic status, marital status, smoking status, and parity (data not shown).
Blood pressures were measured at entry and each prenatal visit, during labor and delivery, and postpartum. Korotkoff phase 4 (muffling) or phase 5 (disappearance) was used for diastolic blood pressure.25 Urine samples were tested for albumin at each prenatal visit. A validation study in which information on blood pressure and urinary albumin was checked against that in the original medical records showed a high level of accuracy.25
We applied contemporary definitions6 of preeclampsia and transient hypertension of pregnancy to measurements of blood pressure and urinary protein taken at the time of the study. Preeclampsia was defined as gestational hypertension and proteinuria, and return of abnormalities to normal in the postpartum period.6 Gestational hypertension was defined as 2 or more measurements of systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg for the first time after 24 weeks of gestation. In the intrapartum period, the first 5 pressures obtained after hospital admission for delivery were averaged. Proteinuria was defined as 2 random urine dipsticks of 1+ protein or one dipstick of 2+ protein. Cases of preeclampsia were considered severe if they had at least one of the following symptoms: systolic blood pressure ≥160 mmHg, diastolic blood pressure ≥110 mmHg, proteinuria of 5 g/24 hours, proteinuria of 3+ or more, oliguria, pulmonary edema, or convulsions/eclampsia. All other cases were considered mild.
Transient hypertension of pregnancy was defined as gestational hypertension in the absence of proteinuria, followed by a return to normal blood pressure postpartum.6 Cases of transient hypertension of pregnancy were considered severe if they had at least one measurement of systolic blood pressure ≥160 mmHg or diastolic blood pressure ≥110 mm Hg. All other cases were considered mild. This definition has been used previously,15,16 though in past papers this entity was termed “severe gestational hypertension.”
Prepregnancy BMI (weight [kg]/height [m]2) was based on maternal self-report at enrollment of height and prepregnancy weight. BMI was categorized as underweight (BMI <18.5), normal weight (BMI = 18.5–24.9), overweight (BMI = 25.0–29.9), and obese (BMI = ≥30.0).26 Race/ethnicity was categorized as white or black. Data on marital status (unmarried, married), and smoking status at entry (smoker, nonsmoker) were available. Education, occupation, and family income data were combined into a composite socioeconomic status score (range 1 to 5).27
Multinomial logistic regression was used to model the outcome (eg, mild preeclampsia, severe preeclampsia, no preeclampsia) as a function of prepregnancy BMI and the following potential confounders, which were selected a priori from the literature: maternal age (<20, 20–29, ≥30 years), race (black, white), socioeconomic status (range = 1 to 5), marital status (married, unmarried), parity (primiparous, multiparous), height (continuous), and smoking status (smoker, nonsmoker at entry). Separate models were run for preeclampsia and transient hypertension of pregnancy. BMI was curvilinear in the logit of both outcomes. Therefore, it was specified as a series of spline terms in both models. After testing many splines, a restricted quadratic spline with knots at 20 and 30 was determined to have the best fit for both models. Although odds ratios can be calculated for any BMI value along the continuum, we chose 17, 20, 25, 30, and 35 as representative values because they span the BMI distribution and include the knots (the inflection points in the risk curve). Alternate specifications of covariates had no meaningful impact on the results.
Effect modification by race/ethnicity was assessed using a likelihood ratio test (α = 0.10) comparing models with and without BMI-race/ethnicity interaction terms. Because race/ethnicity met our criterion for effect modification in both models, interaction terms were included, with the referent being white women with BMI of 20. The homogeneity test of equivalence was used to determine if the BMI effect estimate differed across the mild and severe outcome subtypes in the multinomial logistic model.28 An alpha of 0.10 was used to reject the null hypothesis that the BMI effect was homogeneous across subtypes.
There were 6991 observations that represented repeated pregnancies to the same woman in the study. We compared results obtained when both models were fitted with and without accounting for clustering on patient identification number. There were no meaningful differences in results for mild preeclampsia or mild transient hypertension. For each of the severe subtypes, the point estimates remained the same and confidence intervals widened when ignoring the clustering. As has been done previously,19,29,30 we chose to ignore the clustering for 2 reasons. First, it had little impact on our conclusions. Second, use of clustering would have precluded us from conducting likelihood ratio tests needed for testing effect modification and the homogeneity test of equivalence.
Preeclampsia occurred in 3.6% of the total population, and was more common in blacks (4.5%) than whites (2.7%) (Table 1). Of all preeclampsia cases, 22% in blacks and 18% in whites were severe. The incidence of transient hypertension of pregnancy was 7.5% in the total population, 7.8% in blacks, and 7.2% in whites. The proportions of transient hypertension cases classified as severe were similar in white (4.5%) and black (4.4%) women. A majority of women in both racial/ethnic groups were normal weight before pregnancy. Black women were more likely than white women to be overweight or obese before pregnancy; they were also more likely to be 21 years of age or younger, unmarried, of low socioeconomic status, and nonsmokers. Black women entered care later in pregnancy and attended fewer prenatal visits than white women.
Among white and black women, there was a striking monotonic, dose-response relation between prepregnancy BMI and the risk of both mild and severe preeclampsia (Fig. 1). The risk of severe preeclampsia in both black and white women was about 2-fold for a BMI of 25 and 3-fold for a BMI of 30 compared with white women who had a BMI of 20 (Table 2). At a BMI of 35, the risk of severe preeclampsia was at least 5-fold greater in both racial/ethnic groups compared with the reference group, although point estimates were imprecise due to a small sample of morbidly obese women. The effect of BMI on the risk of severe preeclampsia was similar to its effect on mild disease (homogeneity test of equivalence, P = 0.28). We observed racial/ethnic differences at the extremes of the BMI distribution (Fig. 1; Table 2). A BMI of 17 was protective against both severe and mild preeclampsia in white but not black women. Furthermore the increase in risk of both severe and mild preeclampsia among women with a BMI of 35 was slightly less among black than among white women. The difference between black and white women in the BMI-preeclampsia association met our definition of effect modification (P = 0.08).
The risk of the mild and severe subtypes of transient hypertension of pregnancy also increased in a monotonic dose-response manner with prepregnancy BMI in both racial/ethnic groups (Fig. 2, Table 3). However, for both mild and severe transient hypertension, black women had a higher incidence and greater adjusted odds ratios than white women at BMI values less than 25, but beyond 25, their incidence and adjusted odds ratios were lower than whites (Fig. 2; Table 3). The race difference in the BMI-transient hypertension association was statistically significant (P < 0.01). The effect of BMI was also different for mild and severe transient hypertension of pregnancy (homogeneity test of equivalence P < 0.01). Inspection of the adjusted odds ratios suggested that high BMI was a stronger risk factor for severe than for mild transient hypertension.
We found no meaningful differences in our results for preeclampsia or transient hypertension of pregnancy after excluding women who entered care in the third trimester.
Our results confirm previous findings of a strong, monotonic dose–response relation between prepregnancy BMI and the risk of mild preeclampsia and mild transient hypertension of pregnancy, and extend previous findings to suggest that a similar dose-dependent association is evident for the severe subtypes of these disorders among both white and black women. Further, our results demonstrate that high BMI is as strong risk factor for severe and mild preeclampsia and severe transient hypertension of pregnancy among both black and white women.
Our findings are consistent with one previous investigation20 that showed a 3.5-fold (95% CI = 1.7–7.5) greater risk of severe preeclampsia among women with a prepregnancy BMI ≥32.3 kg/m2 compared with women who had a BMI <32.3 kg/m2. Other investigators have found that prepregnancy BMI is an important predictor of severe preeclampsia.23 Our results conflict with 2 studies12,21 that reported no association between maternal weight and severe preeclampsia. However, neither of these studies used a measure of prepregnancy adiposity. One study12 assessed maternal BMI at 13 to 21 weeks’ gestation, but this measure is confounded by gestational weight gain and by contribution of intravascular and extravascular fluid increases to maternal weight, which are near maximal at this time. The other investigation21 examined maternal weight at an unspecified time before 18 weeks’ gestation, and did not adjust for height. We are aware of only one study22 that assessed prepregnancy BMI in relation to severe gestational hypertension. Investigators reported that BMI was positively associated with “severe hypertension,” but they did not eliminate women with pre-existing hypertension from the case group, nor did they report any effect estimates to allow one to evaluate the strength or precision of the association.22
The mechanisms explaining why maternal overweight predisposes women to mild and severe hypertensive disorders of pregnancy have not been well studied. In a cohort of predominantly mild preeclampsia cases, we previously showed that the strong linear relation between prepregnancy BMI and preeclampsia risk was partially mediated by C-reactive protein, a marker of systemic inflammation, and triglycerides.10 Other important mediators may be oxidative stress, insulin resistance, endothelial dysfunction, reduced immune function, other markers of dyslipidemia, or lifestyle factors such as poor prenatal diet and prenatal physical inactivity. Research is needed to delineate the relevance of these or other pathways in the development of severe forms of transient hypertension of pregnancy and preeclampsia in overweight women, and whether the pathways vary by race/ethnicity.
One of our most intriguing findings was the interaction between BMI and race/ethnicity. Lean black women had a higher risk of preeclampsia and transient hypertension of pregnancy than lean white women, but at high BMI values, the trend was reversed: overweight black women had a lower risk of these disorders than overweight white women. While the effect modification by race met our a priori definition for both outcomes, the ethnic differences were more striking for transient hypertension of pregnancy. For preeclampsia, the differential effect of ethnicity was restricted to the tails of the BMI distribution.
To our knowledge, the interaction between overweight and race/ethnicity on the risk of transient hypertension of pregnancy has not been investigated previously, and only one study assessed this interaction in relation to preeclampsia. In a retrospective cohort study of 22,658 women, investigators reported that the effect of obesity (defined as a BMI >29 kg/m2) on preeclampsia risk was slightly lower among black women (adjusted OR = 1.6; 95% CI = 1.1–2.4) than white women (2.5; 1.9–3.5).31 However, the results are difficult to interpret because BMI was based on weight at the initial visit, which may have been in mid-to-late gestation in some women (no data on gestational age at entry are provided), and therefore is likely confounded by gestational weight gain. Other investigators have suggested that adiposity is a less important predictor of certain nonperinatal health outcomes in black than white women. For instance, a high BMI among black women has a substantially weaker association with all-cause mortality,2,32 coronary heart disease mortality,33,34 and incident cardiovascular disease35 than among white women.
The BMI–race/ethnicity interaction may be explained by the tendency for black women to have central and abdominal body fat distribution compared with white women.35,36 It is thought that atherogenic risk factors such as dyslipidemia and degree of peripheral insulin resistance are not as greatly influenced by central and abdominal obesity.36 Because hypertensive disorders of pregnancy share many similarities with cardiovascular disease,37 these mechanisms may likely be relevant to our findings. Nevertheless, little is currently known about how racial/ethnic differences in fat patterning relate to preeclampsia and transient hypertension of pregnancy. Given the racial/ethnic disparity in the incidence of these disorders,38,39 studies should assess racial/ethnic interactions with prepregnacy adiposity and strive to understand mechanisms underlying these differences.
As with other studies of maternal overweight and hypertensive disorders, our study was limited by the potential for right-censoring. Underweight women were more likely to deliver preterm than heavier women (data not shown), so they had less time to be at risk for severe preeclampsia. This may have biased our results upwards and away from the null. Nevertheless, we found a similar monotonic dose–response association between BMI and risk of mild and severe forms of both hypertensive disorders even after restricting the analysis to women who delivered at term (≥37 weeks). Because the Collaborative Perinatal Project was conducted in the 1960s, blood pressure cuffs may not have been available in extra large sizes to fit obese women. Cuffs that were too small would cause falsely elevated blood pressure in obese subjects, leading to differential misclassification and bias away from the null. Nevertheless, this may not be a major problem since we observed similar strong, dose-dependent relations when we eliminated obese women from our analyses (data not shown). We were limited by a relatively small proportion of obese women in this dataset, which caused unstable point estimates at high BMI values. We also lacked data on other surrogate markers of prepregnancy adiposity, such as prepregnancy waist circumference or skinfold measurements, which may be more potent measures of the adiposity phenotype associated with pregnancy hypertensive disorders.
This study had many notable strengths. The Collaborative Perinatal Project afforded us the unique opportunity to prospectively study the rare outcomes of severe preeclampsia and severe transient hypertension of pregnancy separately by race/ethnicity because of the large sample size and the proven validity of the hypertension and proteinuria data.25 While the standard of care for hypertensive disorders of pregnancy in the 1960s, when the study was conducted, differs from present day, these data offered the chance to study the natural progression of the disease. Our use of spline regression makes no unreasonably strong assumptions about the shape of the BMI dose-response curve, and allows for greater flexibility and more biologically plausible and relevant results than linear or categorical models.40,41 Maternal self-report of prepregnancy weight was unlikely to be an important source of bias in our analysis because self-reported weight has been shown to be highly correlated with measured body weight for most populations.42,43
Our results suggest that the escalating rates of obesity in U.S. childbearing-aged women may predict an increase in perinatal morbidity and mortality associated with severe hypertensive disorders of pregnancy. A preconceptional weight loss or lifestyle intervention should be explored to study simultaneously the prevention of hypertensive disorders of pregnancy and other adverse perinatal outcomes related to obesity.5 Moreover, future studies should explore the mechanisms underlying the effect of overweight on hypertensive disorders of pregnancy so as to aid in the design of appropriate prenatal interventions for women who enter pregnancy with a high BMI.
1. Allison DB, Fontaine KR, Manson JE, et al. Annual deaths attributable to obesity in the United States. JAMA
2. Calle EE, Thun TJ, Petrelli JM, et al. Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med
3. Thompson D, Wolf AM. The medical-care cost burden of obesity. Obes Rev
4. Flegal KM, Carroll MD, Ogden CL, et al. Prevalence and trends in obesity among US adults, 1999–2000. JAMA
5. Cogswell ME, Perry GS, Schieve LA, et al. Obesity in women of childbearing age: Risks, prevention, and treatment. Prim Care Update Ob/Gyns
6. National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy.Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy. Am J Obstet Gynecol
7. Roberts JM. Pregnancy related hypertension. In: Creasy RK, Resnik R, eds. Maternal Fetal Medicine
. 4th ed. Philadelphia: W.B. Saunders; 1998:883–872.
8. Ness RB, Roberts JM. Heterogeneous causes constituting the single syndrome of preeclampsia: a hypothesis and its implications. Am J Obstet Gynecol
9. Baeten JM, Bukusi EA, Lambe M. Pregnancy complications and outcomes among overweight and obese nulliparous women. Am J Public Health
10. Bodnar LM, Ness RB, Harger GF, et al. Inflammation and triglycerides partially mediate the effect of prepregnancy body mass index on the risk of preeclampsia. Am J Epidemiol
11. Bodnar LM, Ness RB, Markovic N, et al. The risk of preeclampsia rises with increasing prepregnancy body mass index. Ann Epidemiol
12. Sibai BM, Ewell M, Levine RJ, et al. Risk factors associated with preeclampsia in healthy nulliparous women. The Calcium for Preeclampsia Prevention (CPEP) Study Group. Am J Obstet Gynecol
13. Ros HS, Cnattingius S, Lipworth L. Comparison of risk factors for preeclampsia and gestational hypertension in a population-based cohort study. Am J Epidemiol
14. Dekker G, Sibai B. Primary, secondary, and tertiary prevention of pre-eclampsia. Lancet
15. Hauth JC, Ewell MG, Levine RJ, et al. Pregnancy outcomes in healthy nulliparas who developed hypertension. Calcium for Preeclampsia Prevention Study Group. Obstet Gynecol
16. Buchbinder A, Sibai BM, Caritis S, et al. Adverse perinatal outcomes are significantly higher in severe gestational hypertension than in mild preeclampsia. Am J Obstet Gynecol
17. Sibai BM, Caritis SN, Thom E, et al. Prevention of preeclampsia with low-dose aspirin in healthy, nulliparous pregnant women. The National Institute of Child Health and Human Development Network of Maternal-Fetal Medicine Units. N Engl J Med
18. Sibai BM, Hauth J, Caritis S, et al. Hypertensive disorders in twin versus singleton gestations. National Institute of Child Health and Human Development Network of Maternal-Fetal Medicine Units. Am J Obstet Gynecol
19. Zhang J, Klebanoff MA, Roberts JM. Prediction of adverse outcomes by common definitions of hypertension in pregnancy. Obstet Gynecol
20. Stone JL, Lockwood CJ, Berkowitz GS, et al. Risk factors for severe preeclampsia. Obstet Gynecol
21. Odegard RA, Vatten LJ, Nilsen ST, et al. Risk factors and clinical manifestations of pre-eclampsia. BJOG
22. Hrazdilova O, Unzeitig V, Znojil V, et al. Relationship of age and the body mass index to selected hypertensive complications in pregnancy. Int J Gynaecol Obstet
23. Lee LC, Sheu BC, Shau WY, et al. Mid-trimester beta-hCG levels incorporated in a multifactorial model for the prediction of severe pre-eclampsia. Prenat Diagn
24. Niswander K. The Collaborative Perinatal Study of the National Institute of Neurological Diseases and Stroke: The Women and Their Pregnancies
. Philadelphia, PA: WB Saunders; 1972.
25. Friedman EA, Neff RK. Pregnancy Hypertension: A Systematic Evaluation of Clinical Diagnostic Criteria
. Littleton, MA: PSG Publishing Co., 1977.
26. NHLBI Expert Panel. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: the evidence report. Obes Res
. 1998;6(Suppl 2):51S–209S.
27. Myrianthopoulos NC, French KS. An application of the U.S. Bureau of the Census socioeconomic index to a large, diversified patient population. Soc Sci Med
28. StataCorp. Stata Statistical Software: Release 8.0. College Station, TX: Stata Corporation; 2003.
29. Newman TB, Klebanoff MA. Neonatal hyperbilirubinemia and long-term outcome: another look at the Collaborative Perinatal Project. Pediatrics
30. Zhang J, Klebanoff MA. Low blood pressure during pregnancy and poor perinatal outcomes: an obstetric paradox. Am J Epidemiol
31. Ramos GA, Caughey AB. The interrelationship between ethnicity and obesity on obstetric outcomes. Am J Obstet Gynecol
32. Stevens J, Plankey MW, Williamson DF, et al. The body mass index-mortality relationship in white and African American women. Obes Res
33. Johnson JL, Heineman EF, Heiss G, et al. Cardiovascular disease risk factors and mortality among black women and white women aged 40–64 years in Evans County, Georgia. Am J Epidemiol
34. Stevens J, Juhaeri, Cai J. Changes in body mass index prior to baseline among participants who are ill or who die during the early years of follow-up. Am J Epidemiol
35. Freedman DS, Williamson DF, Croft JB, et al. Relation of body fat distribution to ischemic heart disease. The National Health and Nutrition Examination Survey I (NHANES I) Epidemiologic Follow-up Study. Am J Epidemiol
36. Stevens J, Gautman SP, Keil JE. Body mass index and fat patterning as correlates of lipids and hypertension in an elderly, biracial population. J Gerontol
37. Roberts JM, Lain KY. Recent Insights into the pathogenesis of pre-eclampsia. Placenta
38. Samadi AR, Mayberry RM, Zaidi AA, et al. Maternal hypertension and associated pregnancy complications among African-American and other women in the United States. Obstet Gynecol
39. Zhang J, Meikle S, Trumble A. Severe maternal morbidity associated with hypertensive disorders in pregnancy in the United States. Hypertens Pregnancy
40. Greenland S. Dose-response and trend analysis in epidemiology: alternatives to categorical analysis. Epidemiology
41. Witte JS, Greenland S. A nested approach to evaluating dose-response and trend. Ann Epidemiol
42. Palta M, Prineas RJ, Berman R, et al. Comparison of self-reported and measured height and weight. Am J Epidemiol
43. Stewart AL. The reliability and validity of self-reported weight and height. J Chronic Disease
© 2007 Lippincott Williams & Wilkins, Inc.