For each study, we calculated an unadjusted risk ratio (RR) for preeclampsia, comparing each higher BMI category with the lowest category, which served as the referent. An unadjusted RR and its 95% confidence interval (CI) were plotted for each. Because of differences between studies in the defined BMI categories and in participant inclusion and exclusion criteria, we chose, a priori, not to pool the risk estimates. Because some studies adjusted for various sets of covariates using multiple logistic regression analysis, those results are presented separately (Table 3).
For the second study objective, we performed a separate analysis to estimate the degree of change in the risk of preeclampsia according to increasing BMI. We plotted the data from all studies on a single scattergraph and generated a straight line of best fit. The x-axis represented the mid-point of each BMI category, and the y-axis the corresponding risk of preeclampsia. For example, for a BMI category of 24 to 26 kg/m2, a mid-point value of 25 kg/m2 was chosen. When a BMI category was defined as being greater (or less) than a specific value, then a BMI value one decimal place higher (or lower) than that value was used. For example, if the highest BMI category was “≥30.0 kg/m2”, then a value of “30.1 kg/m2” was used, whereas a BMI category of “<20 kg/m2” was designated as “19.9 kg/m2.” The latter approach, though potentially imprecise, would enable an estimation of the relation between BMI and the risk of preeclampsia.
For the second analysis, the mean and CI change in the risk of preeclampsia for each kg/m2-unit change in BMI were approximated using a weighted mixed-effects linear model. This method, analogous to a combination of linear regression and analysis of variance (ANOVA), models a data set by both the mean (ie, fixed-effects) and variance-covariance (ie, random-effects) parameters (see Appendix). Including covariance estimates in the model acknowledges that the experimental units on which the data are measured units (ie, each individual BMI-preeclampsia category) are grouped into clusters (ie, individual studies), and that the data from a common cluster are correlated. An inverse-variance–weighted mixed-effects linear model was created, accounting for the different numbers of women within each BMI-preeclampsia category. To obtain appropriate interpretations of intercepts and slopes, the BMI was centered to the study’s mean levels. Statistical significance testing was set at a 2-sided P-value less than 0.05. Mixed linear modeling was done using the MIXED Procedure in SAS Version 8 (SAS Institute Inc., Cary, North Carolina), whereas Meta-Analyst 0.988 (Lau J. Meta-Analyst 0.988. Boston: Meta-Analyst Statistical Software, 1995) and Microsoft Excel version 5.0c (Microsoft Corporation, 1985–1994) were used to generate the Figures.
The initial searches of Medline and Embase yielded 491 and 125 citations, respectively. Of the 13 studies that met the inclusion criteria, eight originated from the United States 8–15 and the remainder were from Sweden, 16 the Netherlands, 17 Latin and Caribbean Amer-ica, 18 Taiwan 19 and the United Kingdom (Table 1). 20 Five of the studies excluded women with chronic hypertension, 9,12,13,17,19 four excluded women with diabetes mellitus, 9,12,13,17 and nine excluded women with multiple gestations (Table 1). 8–12,14,15,17–20 All but one 15 of the study reports specified that preeclampsia was defined using established criteria defined as a blood pressure ≥140/90 mmHg or a rise in systolic blood pressure >30 mmHg or diastolic blood pressure >15 mmHg after 20 weeks’ gestation, in addition to either ≥1+ dipstick proteinuria on two separate occasions, ≥2+ dipstick proteinuria on one occasion or ≥300 mg proteinuria over 24 hours. 21,22 None of the studies had masked and independent assessment of exposure and outcome.
The variation in prevalence of preeclampsia among studies is presumably related in part to the exclusion of women at high risk for preeclampsia, 9,12,17 or to geographic and ethnic differences between participants (Table 1). For example, one study included Asian women, 19 whereas others solely recruited women from North America. 8–15
A total of 1,390,226 women were included in the 13 studies. The BMI categories and corresponding percentages of preeclampsia are listed in Table 2. The risk of preeclampsia characteristically rose with increasing BMI, as did the unadjusted RR of preeclampsia (Figure 1). In most studies, there was a step-up increase in the unadjusted RR of preeclampsia with each increasing BMI category, which was most evident in the largest studies. 9,15,18,19 The risk of preeclampsia typically doubled for each 5 to 7 kg/m2 increase in BMI (Figure 1).
In those five studies that excluded women with chronic hypertension there appeared a nonsubstantial 12,17 or substantial 9,13,19 trend of increasing preeclampsia risk with rising prepregnancy BMI (Figure 1). Nonsubstantial 12,17 and substantial 9,19 trends were also seen in the four studies that excluded women with diabetes mellitus (Figure 1).
Eight studies used multivariate analysis to adjust for potential confounders, as listed in Table 3. Using those comparison and referent BMI categories set by the authors, all eight studies observed at least a doubling of the risk of preeclampsia with excess maternal BMI (Table 3).
Considering all 43 BMI categories from the 13 studies, and using the weighted mixed-effects linear model, the risk of preeclampsia typically rose by 0.54% (CI = 0.27–0.80) for each 1-kg/m2 increase in BMI (Figure 2).
In a systematic overview of 13 cohort studies we observed a consistent and linear rise in the risk of preeclampsia with increasing prepregnancy BMI.
Weaknesses and Strengths
We avoided deriving a pooled estimate of the RR of preeclampsia, principally because each study defined the BMI categories differently. Although we chose to plot the relation among the 43 BMI categories and the corresponding preeclampsia rates from all 13 studies (Figure 2), we did so knowing that the result should be interpreted with caution. Even so it might be useful as a research tool (see below). Our use of the mid-point value for each BMI category may have approximated the true mean BMI imprecisely within that category. For the lowest and highest BMI categories, we used values that were closest to the respective category cutoff values, likely leading to an overestimate of the true relation between BMI and preeclampsia. We further acknowledge that study publication bias, with the omission of negative studies, also might have led to an overestimate of the relation between prepregnancy BMI and preeclampsia.
The observed association between prepregnancy BMI and preeclampsia may be confounded by the presence of chronic hypertension, 23 diabetes mellitus 24 and other elements of the dysmetabolic syndrome, 5 each of which are known risk factors for preeclampsia. When we evaluated studies that either excluded women with chronic hypertension or diabetes mellitus, or that adjusted for the presence of these and other confounders, the relation between maternal BMI and preeclampsia usually remained unchanged. 9,13,19,20 Thus, it is likely that elevated BMI is an independent predictor of preeclampsia risk, just as for other adverse pregnancy events. 7
We considered only those studies that assessed prepregnancy BMI as a primary exposure. BMI is a better overall measure of obesity than weight alone. 23 Although seven 9,10,12,13,15–17 of the 13 cohort studies used self-reported prepregnancy height and weight for the calculation of BMI, this method reliably approximates measured BMI. 25
Mechanisms of Disease
There is likely more than one mechanism underlying the relation of prepregnancy obesity to preeclampsia. Placental vasculopathy and endothelial dysfunction appear central to the pathogenesis of preeclampsia. 1,3,26,27 Individuals with the dysmetabolic syndrome, of which obesity is a major feature, also exhibit chronic hypertriglyceridemia, a risk factor for endothelial dysfunction. 28 Hyperlipidemia may also alter prostaglandin regulation, leading to arteriolar constriction, 27 whereas biomarkers of insulin resistance, including plasminogen activator inhibitor, leptin and tumor necrosis factor, appear higher in women with preeclampsia. 29–31 Finally, obesity, a major risk factor for obstructive sleep apnea and associated hypertension, 32 is commonly found in pregnant women with disordered breathing during sleep. 33 Preeclamptic women with obstructive sleep apnea have higher blood pressure during obstructive periods. 34
We observed a continuous relation between prepregnancy BMI and the risk of preeclampsia. Further work could explore the interplay between BMI and other common risk factors for preeclampsia, including maternal age, gravidity and ethnicity. Taken together, these risk factors might enable better prediction of a woman’s risk of preeclampsia.
It is not known whether a program of lifestyle modification, including prepregnancy weight loss, can reduce the risk of preeclampsia. A randomized clinical trial might usefully evaluate this question among obese women of reproductive age. Both the exposure (ie, obesity) 35 and the outcome (ie, preeclampsia) 1 are common, and screening for obesity requires little time or special equipment. 36 Second, there are other potential benefits to prepregnancy weight loss, including lower risk of gestational diabetes mellitus, Cesarean delivery, and perinatal morbidity and mortality. 37 Moreover, obesity before pregnancy is a major predictor of obesity later in life, which is commonly associated with the development of chronic hypertension, dyslipidemia and type 2 diabetes mellitus. 38
Two randomized clinical trials among middle-aged obese adults with glucose intolerance clearly demonstrated that dietary modification, increased physical activity, and weight loss can be achieved, with a 50% reduction in subsequent type 2 diabetes mellitus. 39,40–42 Based on results of our linear models, prepregnancy weight reduction might produce an approximately 0.54% decrease in the rate of preeclampsia per 1-kg/m2 decline in BMI.
We thank Marian J. Vermeulen for her helpful comments about the data analyses.
The weighted mixed effects linear model may be written as:EQUATION
fixed effects random effects
where, BMI ij denotes i th body mass index measure within the j th study, Y ij is the rate of preeclampsia for the i th BMI measure within the j th study, γ00 is the overall intercept, γ10 is the overall slope, u 0 j is the variation between study intercepts, u 1 j is the variation between study slopes and rij is the random error associated.
1. Roberts JM, Cooper DW. Pathogenesis and genetics of pre-eclampsia. Lancet 2001; 357: 53–56.
2. Dekker G, Sibai B. Primary, secondary, and tertiary prevention of pre-eclampsia. Lancet 2001; 357: 209–215.
3. Hayman R, Brockelsby J, Kenny L, Baker P. Preeclampsia
: the endothelium, circulating factor(s) and vascular endothelial growth factor. J Soc Gynecol Investig 1999; 6: 3–10.
4. Solomon CG, Seely EW. Brief review: hypertension
: a manifestation of the insulin resistance syndrome? Hypertension
2001; 37: 232–239.
5. Barden AE, Beilin LJ, Ritchie J, Walters BN, Micheal C. Does a predisposition to the metabolic syndrome sensitize women to develop pre-eclampsia? J Hypertens 1999; 17: 1307–1315.
6. Ford ES, Giles WH, Dietz WH. Prevalence of the metabolic syndrome among U.S. adults. Findings from the Third National Health and Nutrition Examination Survey. JAMA 2002; 287: 356–359.
7. Cnattingius S, Bergstrom R, Lipworth L, Kramer MS. Prepregnancy weight and the risk of adverse pregnancy
outcomes. N Engl J Med 1998; 338: 147–152.
8. Edwards LE, Hellerstedt WL, Alton IR, Story M, Himes JH. Pregnancy
complications and birth outcomes in obese and normal-weight women: effects of gestational weight change. Obstet Gynecol 1996; 87: 389–394.
9. Sibai BM, Ewell M, Levine RJ, et al
. Risk factors associated with preeclampsia
in healthy nulliparous women. Am J Obstet Gynecol 1997; 177: 1003–1010.
10. Ogunyemi D, Hullett S, Leeper J, Risk A. Prepregnancy body mass index
, weight gain during pregnancy
, and perinatal outcome in a rural Black population. J Matern Fetal Med 1998; 7: 190–193.
11. Bianco AT, Smilen SW, Davis Y, Lopez S, Lapinski R, Lockwood C. Pregnancy
outcome and weight gain recommendations for the morbidly obese woman. Obstet Gynecol 1998; 91: 97–102.
12. Bowers D, Cohen W. Obesity
and related pregnancy
complications in an inner-city clinic. J Perinatol 1999; 19: 216–219.
13. Thadhani R, Stampfer MJ, Hunter DJ, Manson JE, Solomon CG, Curhan GC. High body mass index
and hypercholesterolemia: risk of hypertensive disorders of pregnancy
. Obstet Gynecol 1999; 94: 543–550.
14. Steinfeld JD, Valentine S, Lerer T, Ingardia CJ, Wax JR, Curry JL. Obesity
-related complications of pregnancy
vary by race. J Matern Fetal Med 2000; 9: 238–241.
15. Baeten JM, Bukusi EA, Lambe M. Pregnancy
complications and outcomes among overweight and obese nulliparous women. Am J Public Health 2001; 91: 436–440.
16. Ros HS, Cnattingius S, Lipworth L. Comparison of risk factors for preeclampsia
and gestational hypertension
in a population-based cohort study
. Am J Epiemiol 1998; 147: 1062–1070.
17. Knuist M, Bonsel GJ, Zondervan HA, Treffers PE. Risk factors for preeclampsia
in nulliparous women in distinct ethnic groups: a prospective study. Obstet Gynecol 1998; 92: 174–178.
18. Conde-Agudelo A, Belizan J. Risk factors for pre-eclampsia in a large cohort of Latin American and Caribbean women. Br J Obstet Gynecol 2000; 107: 75–83.
19. Lee CJ, Hsieh TT, Chiu TH, Chen KC, Lo LM, Hung TH. Risk factors for pre-eclampsia in an Asian population. Int J Gynecol and Obstet 2000; 70: 327–333.
20. Sebire N, Jolly M, Harris JP, et al
. Maternal obesity
outcome: a study of 287,213 pregnancies in London. Int J Obes 2001; 25: 1175–1182.
21. National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy
. Report on high blood pressure in pregnancy
. Am J Obstet Gynecol 1990; 163: 1689–1712.
22. 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 of High Blood Pressure in Pregnancy
. Am J Obstet Gynecol 2000; 183: S1–S22.
23. Ray JG, Burrows R, Burrows EA, Vermeulen M. MOS HIP: McMaster outcome study of hypertension
1 (MOS HIP 1). Early Hum Dev 2001; 64: 129–143.
24. Ray JG, Vermeulen MJ, Shapiro JL, Kenshole AB. Maternal and neonatal outcomes in pregestational and gestational diabetes mellitus, and the influence of maternal obesity
and weight gain: the DEPOSIT study. QJM 2001; 94: 347–356.
25. Willett W, Stampfer MJ, Lipnick BC, et al
. Cigarette smoking, relative weight and menopause. Am J Epidemiol 1983; 117: 651–658.
26. DeWolf F, Brosens I, Renaer M. Fetal growth retardation and the maternal arterial supply in the human placenta in the absence of sustained hypertension
. Br J Obstet Gynaecol 1980; 87: 678–685.
27. Roberts JM, Taylor RN, Goldfein A. Clinical and biochemical evidence of endothelial cell dysfunction in the pregnancy
syndrome pre-eclampsia. Am J Hypertens 1991; 4: 700–708.
28. de Man FH, Weverling-Rijnsburger AW, van der Laarse A, Smelt AH, Jukema JW, Blauw GJ. Not acute but chronic hypertriglyceridemia is associated with impaired endothelium-dependent vasodilation: reversal after lipid-lowering therapy by atorvastatin. Arterioscler Thromb Vasc Biol 2000; 20: 744–750.
29. McCarthy JF, Misra DN, Roberts JM. Maternal plasma leptin is increased in preeclampsia
and positively correlates with fetal cord concentration. Am J Obstet Gynecol 1999; 180: 731–36.
30. Estelles A, Gilabert J, Grancha S et al. Abnormal expression of type 1 plasminogen activator and tissue factor in severe preeclampsia
. Thromb Haemost 1998; 79: 500–8.
31. Vince GS, Startkey PM, Austgulen R, Kwiatkowski D, Redman CW. Interleukin-6, tumour necrosis factor and soluble tumour factor receptors in women with pre-eclampsia. Br J Obstet Gynaecol 1995; 102: 20–5.
32. Berger HA, Somers VK, Phillips BG. Sleep disordered breathing and hypertension
. Curr Opin Pulm Med 2001; 7: 386–90.
33. Maasilta P, Bachour A, Teramo K, Polo O, Laitinen LA. Sleep-related disordered breathing during pregnancy
in obese women. Chest 2001; 120: 1448–1454.
34. Edwards N, Blyton DM, Kirjavainen TT, Sullivan CE. Hemodynamic responses to obstructive respiratory events during sleep are augmented in women with preeclampsia
. Am J Hypertens 2001; 14: 1090–1095.
35. Solomon CG, Willett WC, Carey VJ, et al
. A prospective study of pregravid determinants of gestational diabetes mellitus. JAMA 1997; 278: 1078–1083.
36. Kiernan M, Winkleby MA. Identifying patients for weight-loss treatment: an empirical evaluation of the NHLBI obesity
education initiative expert panel treatment recommendations. Arch Intern Med 2000; 160: 2169–2176.
37. Lu GC, Rouse DJ, DuBard M, Cliver S, Kimberlin D, Harith JC. The effect of the increasing prevalence of maternal obesity
on perinatal morbidity. Am J Obstet Gynecol 2001; 185: 845–849.
38. Brown CD, Higgins M, Donato KA, et al
. Body mass index
and the prevalence of hypertension
and dyslipidemia. Obes Res 2000; 8: 605–619.
39. Tuomilehto J, Lindstrom J, Eriksson JG, et al. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med 2001; 344: 1343–1350.
40. Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002; 346: 393–403.
Keywords:© 2003 Lippincott Williams & Wilkins, Inc.
body mass index; obesity; pregnancy; preeclampsia; toxemia; hypertension; cohort study; systematic overview