Preeclampsia is an important cause of maternal multiorgan failure and death, preterm birth, intrauterine growth restriction, and increased perinatal mortality; it is estimated to affect 3% to 5% of pregnant women in the developed world.1 Worldwide, hypertensive disorders of pregnancy are responsible for 12% of maternal deaths during pregnancy and the puerperium.2 Preeclampsia is a pregnancy-specific disorder characterized by hypertension (blood pressure ≥140/90 mmHg) and proteinuria in the second half of pregnancy.3 The pathophysiology of preeclampsia involves endothelial dysfunction, vasoconstriction, activated hemostasis, and multiorgan injury affecting the liver, kidney, and brain.4 The clinical picture of preeclampsia exhibits great variability from prolonged mild forms to rapidly developing multiorgan failure.
The causes of preeclampsia are not well understood, and its prevention remains a challenge. The prevalence of preeclampsia varies by ethnicity, with higher rates among African American and Latina women,5 and a genetic component is likely to influence risk.6 Several cardiovascular risk factors (including dyslipidemia,7 obesity,8 and diabetes9) are also risk factors for preeclampsia. A systematic review8 reported a doubling in the risk of preeclampsia for a 5- to 7-unit increase in body mass index (BMI) (kg/m2). Women with preeclampsia have an increased risk of cardiovascular disease, hypertension, and overall mortality later in life.10 The risk seems to be particularly strong for women with preterm preeclampsia, with one study reporting a eight-fold increase in the risk of cardiovascular mortality.11
Physical activity is a protective factor for hypertension and cardiovascular disease.12,13 Given that preeclampsia and cardiovascular disease share several risk factors, it has been hypothesized that physical activity may also protect against preeclampsia.14 Several cohort15–26 and case–control studies27–30 have investigated the association between physical activity and preeclampsia. Some studies have found an inverse association between physical activity in early pregnancy and risk of preeclampsia,18,24,25,27,28 whereas other studies have not.15–17,19,22,29,30 Studies on physical activity before pregnancy have reported both inverse associations17,28 and no association.20,21 However, some of these studies may have had a sample size that was too small to detect an association.15,16,20,22,24,29,30 A Cochrane review of two randomized trials found a summary relative risk (RR) of 0.31 (95% confidence interval [CI] = 0.01–7.09),31 but the review was underpowered to detect an association. A more recent meta-analysis of case–control and cohort studies did find an inverse association in case–control studies, but not among cohort studies.32 However, the analysis compared any physical activity with none or low activity, pooling all levels of activity above the reference category, without assessing the dose–response effect of various levels of physical activity. We conducted a systematic review and meta-analysis of case–control studies and cohort studies of physical activity and preeclampsia, with a specific aim of clarifying the possible dose–response relationship between physical activity and preeclampsia.
Search Strategy and Inclusion Criteria
We developed a systematic review protocol for the project. We searched the PubMed, Embase, and Ovid databases up to 2 November 2012 for eligible studies. We used the following search terms in the PubMed search: (physical activity OR vigorous physical activity OR moderate physical activity OR exercise OR physical exercise OR sports OR leisure-time physical activity OR occupational physical activity OR walking OR biking OR running OR physical fitness OR fitness OR exercise test OR inactivity OR sedentary OR risk factor) AND (pregnancy induced hypertension OR pre-eclampsia OR preeclampsia). The search strategy was modified and adapted for searches in Embase and Ovid (eAppendix, http://links.lww.com/EDE/A745). We followed standard criteria for conducting and reporting meta-analyses.33 In addition, we searched the reference lists of the identified publications and reviews32 for further studies. The studies were screened by one of the authors (D.A.).
We included published retrospective case–control studies and cohort, case-cohort, and nested case–control studies that investigated the association between physical activity and the risk of preeclampsia. Estimates of RR had to be provided (with the 95% CIs). For the dose–response analysis, a quantitative measure of the physical activity level was required. We identified 11 cohort studies15–25 and four case–control studies27–30 that were eligible for inclusion. Of the publications that we considered to be possibly relevant for the analysis, the following were excluded: one duplicate publication,34 two randomized trials of stretching exercises (one of which was reported in three publications) or yoga that did not provide risk estimates,35–38 one combined lifestyle intervention,39 one study that reported on a nonspecific outcome of hypertensive complications that also included gestational hypertension (not only preeclampsia),40 and one publication that reported only on severe preeclampsia.41 In addition, one study (two publications)26,42 was excluded because the CIs were considered unreliable due to small sample size (n = 238) with only seven cases. Another publication was excluded because it presented unreliable risk estimates (no cases in the lowest category and no controls in the highest category, with three categories of activity), with no reference category in the analysis43 (Figure 1). We contacted the authors of three publications25,30,40 for additional information, and two of these responded.30,40 We received adjusted overall risk estimates for preeclampsia from one of the studies.30 It was not possible to retrieve results for preeclampsia alone from a study that reported on hypertensive complications during pregnancy.40
The following data were extracted from each study: the first author’s last name, publication year, country where the study was conducted, study period, sample size, number of cases/controls/participants, physical activity type, level of activity, odds ratios or RRs (with 95% CIs) for the highest versus the lowest level of physical activity, and the variables adjusted for in the analysis.
We calculated summary RRs for the highest versus the lowest level of physical activity using the random effects model by DerSimonian and Laird,44 which takes into account both within- and between-study variation (heterogeneity). The average of the natural logarithm of the RRs was estimated, and the RR from each study was weighted by the inverse of its variance and then unweighted by a random effects variance component that corresponds to the amount of heterogeneity in the analysis. For one study that reported results stratified by parity, we pooled the stratified risk estimates using a fixed effects model before inclusion in the overall analysis.15 For one study that used the intermediate category as the reference category,29 we converted the risk estimates so that the lowest category became the reference category, using the method by Hamling et al.45 We conducted separate analyses for prepregnancy and early pregnancy physical activity because two studies that presented results for both time periods17,28 found large differences in the number of subjects in the lowest category between the two periods, with a reduced physical activity level in early pregnancy compared with prepregnancy.
To investigate whether a specific level of physical activity was associated with preeclampsia, we used the method described by Greenland and Longnecker46 to conduct dose–response analysis by computing study-specific slopes (linear trends) and 95% CIs from the natural log of the RRs and CIs across categories of physical activity level. The median or mean level of physical activity in each category of physical activity was assigned to the corresponding RR for each study. For studies that reported physical activity by ranges of activity, we estimated the mean physical activity level in each category by calculating the average of the upper and lower boundaries. Because the studies did not report the results in a consistent manner (eg, some used hours per week, frequency per month or week, metabolic equivalent task [MET]-hours, or other measures), we arbitrarily converted results reported as frequency per month or week to hours per week by assigning a dose of 45 minutes per session for two studies.18,29
We conducted separate analyses for studies that reported results in MET-hours per week. The MET is an index of the intensity of physical activity and is defined as the caloric expenditure per kilogram of body weight per hour of activity, divided by the equivalent per hour at rest.47 One MET is considered to be equal to the energy cost of a person during quiet sitting; walking at a slow pace has a MET value of 2; and jogging and bicycling have MET values of 7–8. MET-hours are the number of hours engaged in each activity multiplied by the MET value of that activity; these are then summed across the various types of activities. Studies that did not quantify the physical activity level were excluded from the dose–response analysis. We assessed a potential nonlinear dose–response relationship between physical activity and preeclampsia using fractional polynomial models.48 We determined the best-fitting second-order fractional polynomial regression model, defined as the one with the lowest deviance. A likelihood ratio test was used to assess the difference between the nonlinear and linear models to test for nonlinearity.48
Heterogeneity among studies was evaluated using Q and I2 statistics.49I2 is a measure of how much heterogeneity is due to between-study variation rather than chance. I2 values of 25%, 50%, and 75% indicates low, moderate, and high heterogeneity, respectively. We conducted subgroup analyses by study characteristics such as study design, sample size, number of cases, and geographic location and by adjustment for confounding factors to investigate potential sources of heterogeneity.
Publication bias was assessed when there were at least five studies in the analysis, using Egger’s test50 and Begg-Mazumdar’s test51 and by inspection of funnel plots. Statistical analyses were conducted using Stata, version 10 software (StataCorp, College Station, TX).
We identified 15 studies that could be included in the analysis of physical activity and preeclampsia, including 11 cohort studies15–25 and four case–control studies27–30 (Tables 1 and 2). Figure 1 shows a flowchart of the study selection. Seven of the studies were from North America, seven were from Europe, and one was from Asia.
Prepregnancy Physical Activity
High Versus Low Analysis
Four cohort studies17,20,21,23 and one case–control study28 investigated the association between prepregnancy physical activity and preeclampsia, including 621 cases among 10,317 participants (Tables 1 and 2). The summary RR for high versus low physical activity was 0.65 (95% CIs = 0.47–0.89) with no evidence of heterogeneity (I2 = 0%; test for heterogeneity, P = 0.91; it was similar among cohort studies; Figure 2A). There was no evidence of heterogeneity by study design (P = 0.93) nor evidence of publication bias, either with Egger’s test (P = 0.12) or with Begg’s test (P = 0.62), although there were few studies (eFigure 1, http://links.lww.com/EDE/A745).
One cohort study17 and one case–control study28 were included in the dose–response analysis of MET-hours per week. The summary RR was 0.78 (95% CI = 0.63–0.96; I2 = 0%; test for heterogeneity, P = 0.80) per 20 MET-hours per week (Figure 2B). Two cohort studies17,21 and one case–control study28 could be included in the dose–response analysis of hours per week. The summary RR was 0.72 (0.53–0.99; I2 = 0%; test for heterogeneity, P = 0.70) per 1 hour of physical activity per day and was similar in the two cohort studies (Figure 2C). There was no evidence of heterogeneity by study design (P = 0.83). There was an indication that the association between prepregnancy physical activity and preeclampsia was nonlinear (test for nonlinearity, P = 0.03), with a 40% reduction in risk up to 5–6 hours per week but no further reductions at higher activity levels (Figure 2D, eTable 1, http://links.lww.com/EDE/A745).
Physical Activity in Early Pregnancy
High Versus Low Analyses
Seven cohort studies16–19,22,23,25 and four case–control studies27–30 investigated the association between early pregnancy physical activity (defined in most studies as physical activity up to gestational weeks 16–24, physical activity up to first antenatal visit, or the first trimester of pregnancy) and preeclampsia (Tables 1 and 2). These studies included 5702 cases among 168,602 participants. The summary RR was 0.79 (95% CI = 0.70–0.91). There was no indication of heterogeneity (I2 = 0%; test for heterogeneity, P = 0.55) (Figure 3A). Results were similar to this overall summary estimate in cohort studies but were slightly stronger in case–control studies (Figure 3A); however, there was no heterogeneity by study design (P = 0.38). There was no evidence of publication bias either with Egger’s test (P = 0.90) or with Begg’s test (P = 1.00) (eFigure 2, http://links.lww.com/EDE/A745).
Two cohort studies17,19 and one case–control study28 reported activity in MET-hours per week of activity. The summary RR per 20 MET-hours per week was 0.85 (95% CI = 0.68–1.07; I2 = 69%; test for heterogeneity, P =0.04; Figure 3B).
Four cohort studies17–19,22 and three case–control studies27–29 were included in the dose–response analysis of hours per week. The summary RR per 1 hour per day of leisure-time physical activity was 0.83 (0.72–0.95; I2 = 21%; test for heterogeneity, P = 0.27) for all studies combined (Figure 3C), 0.87 (0.78–0.98; I2 = 0%; test for heterogeneity, P = 0.48) for cohort studies and 0.61 (0.43–0.86; I2 = 0%; test for heterogeneity, P = 0.50) for case–control studies. There was a marginal heterogeneity between study designs (P = 0.054). There was no evidence for a nonlinear association between physical activity in early pregnancy and risk of preeclampsia (test for nonlinearity, P = 0.37) (Figure 3D, eTable 2, http://links.lww.com/EDE/A745).
Intensity of Physical Activity
One case–control study28 and one cohort study21 reported on intensity of physical activity in the prepregnancy period in relation to preeclampsia; two case–control studies27,28 and one cohort study22 reported on intensity of physical activity in early pregnancy and risk of preeclampsia. The summary RR for high- versus low-intensity activity was 0.55 (95% CI = 0.25–1.21; I2 = 56%; test for heterogeneity, P = 0.13) in prepregnancy and 0.51 (0.37–0.71; I2 = 0%; test for heterogeneity, P = 0.84) in early pregnancy.
Combined Physical Activity Before and During Early Pregnancy
One cohort study17 and two case–control studies16,28 reported results for combined physical activity before and during early pregnancy. Compared with women who did not participate in physical activity in either period (before and during early pregnancy), the summary RR was 0.89 (95% CI = 0.59–1.35; I2 = 0%; test for heterogeneity, P = 0.79) for women who participated in physical activity before pregnancy only, 1.32 (0.62–2.78; I2 = 0%; test for heterogeneity, P = 0.40) for women who participated in physical activity in early pregnancy only, and 0.64 (0.44–0.93; I2 = 0%; test for heterogeneity, P = 0.91) for women who participated in physical activity both before and during early pregnancy.
Three case–control studies27–29 and one cohort study24 reported on walking in early pregnancy and preeclampsia risk. The summary RR was 0.68 (95% CI = 0.51–0.89; I2 = 0%; test for heterogeneity, P = 0.75) for high versus low walking (Figure 4A). Restricting the analysis to the three studies24,27,29 that reported on occupational walking gave a summary RR of 0.68 (0.50–0.93; I2 = 0%; test for heterogeneity, P = 0.54).
Occupational Physical Activity
Four cohort studies15,16,23,24 and two case–control studies27,29 assessed occupational physical activity in early pregnancy and risk of preeclampsia. The summary RR for high versus low occupational activity was 0.82 (95% CI = 0.66–1.03; I2 = 0%; test for heterogeneity, P = 0.68), with an inverse association among cohort studies, summary RR = 0.75 (0.56–1.00; I2 = 0%; test for heterogeneity, P = 1.00) (Figure 4B). There was no evidence of publication bias with Egger’s test (P = 0.78) or with Begg’s test (P = 1.00) (eFigure 3, http://links.lww.com/EDE/A745).
Subgroup and Sensitivity Analyses
Physical activity in both prepregnancy and early pregnancy was protective in most strata of subgroup analyses defined by study design, geographic location, number of cases, and adjustment for confounding factors (Table 3). With meta-regression analysis, there was no evidence that the results differed among these subgroups (Table 3). Using alternative risk estimates based on frequency instead of METs from one study25 did not materially alter the conclusions (results not shown). Alternatively, excluding one abstract-only publication25 did not change the results: summary RR = 0.81 (95% CI = 0.71–0.93). Inclusion of one study that reported hypertensive complications (not specifically preeclampsia)40 and one study that was originally excluded because of unreliable CIs26 did also not alter the conclusions: summary RR = 0.77 (0.68–0.88; I2 = 0%; test for heterogeneity, P = 0.52). Restricting the analyses to the three studies that reported on both prepregnancy and early pregnancy physical activity17,23,28 gave summary estimates of 0.70 (95% CI = 0.46–1.05) for prepregnancy physical activity and 0.69 (0.45–1.07) for early pregnancy physical activity. Only three studies could be included in subgroup analyses of severe onset or early-onset preeclampsia18,19,30 and two in analyses of mild onset or late-onset preeclampsia18,30; summary RRs were 1.06 (0.54–2.11; I2 = 71%; test for heterogeneity, P = 0.03) and 0.85 (0.68–1.08; I2 = 0%; test for heterogeneity, P = 0.67), respectively. In sensitivity analyses removing one study at a time, the size of the summary estimates was similar for most analyses (eFigures 4–6, http://links.lww.com/EDE/A745).
To investigate how much of the association between physical activity and preeclampsia is potentially mediated through BMI, we conducted additional analyses among studies that provided results both unadjusted and adjusted for BMI. The summary RR was 0.53 (95% CI = 0.33–0.86) for two studies17,20 of prepregnancy physical activity without BMI adjustment and 0.60 (0.36–1.00) with adjustment. For three studies17,18,23 of early pregnancy physical activity, the summary was 0.64 (0.53–0.77) without BMI adjustment and 0.77 (0.64–0.92) with BMI adjustment.
This dose–response meta-analysis of observational studies of physical activity and risk of preeclampsia confirms an inverse association of higher levels of physical activity (both before and during early pregnancy) and lower risk of preeclampsia. There was a ~20–35% RR reduction for the women with the highest prepregnancy physical activity level and a ~20% RR reduction with high physical activity in early pregnancy. A stronger inverse association was observed among women who were physically active both before and during early pregnancy (36% RR reduction) and for high-intensity physical activity in early pregnancy (50% RR reduction), although there were few studies in these analyses. In the dose–response analyses, there was evidence of a nonlinear association between prepregnancy physical activity and preeclampsia, with the most benefit observed at levels of 5–6 hours of exercise per week (40% RR reduction); in contrast, the association between physical activity in early pregnancy and preeclampsia appeared to be linear. There was evidence that walking in early pregnancy reduced the risk as well. We found little evidence of an association between occupational physical activity and preeclampsia overall—although a suggestion of reduced risk was observed among cohort studies. A previous systematic review of two randomized trials of physical activity during pregnancy and preeclampsia31 did not find an association (summary RR = 0.31 [95% CI = 0.01–7.09]), but there was low statistical power to detect an association.
Several biological mechanisms could explain a protective effect of physical activity on the risk of preeclampsia. In a transgenic mouse model, exercise training both before and during gestation reduced features of preeclampsia (including normalization of mean arterial pressure, proteinuria, placental disease and cardiac hypertrophy, vascular reactivity, and placental vascular endothelial growth factor expression).52 Physical activity reduces blood pressure and triglyceride levels in nonpregnant women53 and has been associated with lower levels of total cholesterol, triglycerides, leptin, and improved glycemic control and reduced insulin resistance in pregnant women.54–57 Several studies have found that physical activity reduces the risk of gestational hypertension as well,22,27,40 although other studies found no association.16,23,29
The features of preeclampsia often include dyslipidemia, inflammation, and oxidative stress; it has been shown that regular weight-bearing exercise influences the level of plasma tumor necrosis factor-α58 and reduces oxidative stress by reducing lipid peroxidation59 and increasing iron-binding capacity, antioxidant enzyme levels, and prostacyclin while reducing thromboxane levels.60,61 In addition, physical activity protects against obesity,62 diabetes,63 gestational diabetes,64 and cardiovascular disease,65 which are established risk factors for preeclampsia. Recent cohort studies found reduced risk of excessive gestational weight gain among physically active women.66,67 However, many of the studies included in our meta-analysis had adjusted for BMI, and yet, the results persisted in the subgroup analysis with adjustment for BMI, suggesting an association independent of overweight or obesity. If BMI is an intermediate variable, it is possible that we may have somewhat underestimated the association between physical activity and preeclampsia in the overall analyses. When we conducted analyses among studies that reported results both with and without BMI adjustment, the results were slightly (12–36%) stronger without adjustment for BMI. We found a slightly stronger association between prepregnancy physical activity and preeclampsia compared with activity in early pregnancy. It is possible that this may be partly explained by different studies being included in the two analyses; when we restricted the analyses to the three studies that reported on both types of physical activity, the summary estimates were similar. However, it is also possible that the differences in the strength of the association might be real in terms of what level or intensity of physical activity that is practically achievable in early pregnancy compared with the nonpregnant state. There were fewer participants in the highest category and more participants in the lowest category of physical activity in analyses of early pregnancy physical activity compared with prepregnancy physical activity.17,28
The present systematic review and meta-analysis have some limitations. The number of available studies was moderate, and even fewer studies could be included in the dose–response analysis because physical activity level was not quantified or quantified in a manner that could not be combined with the remaining studies. Further studies should aim to clarify the dose–response relationship between physical activity and preeclampsia and to report sufficient details for inclusion in future dose–response analyses.68 In addition, more consistent assessment and reporting on various types and intensities of physical activity would be useful. Many of the studies included had adjusted for important confounding factors, and we did not find evidence that the results differed when stratified by adjustment for confounding factors in subgroup analyses. However, we cannot exclude the possibility that residual confounding by unidentified or unmeasured risk factors could have influenced the results. Physical activity was assessed using self-reported questionnaires, and therefore, measurement error in the exposure variable is possible. Such errors are most likely to be random and to bias results toward the null in prospective studies, although differential bias is possible in case–control studies. However, the results persisted in the subgroup analyses of prospective studies. Because of the small number of studies that reported on subtypes of preeclampsia, we were not able to clarify with sufficient statistical power whether the association between physical activity and preeclampsia varies according to subtype of preeclampsia.
In most of the analyses, there was little evidence of heterogeneity; this was true in several subgroup analyses as well. Overall, results were similar when excluding studies one at a time. Together with the biological plausibility of the findings, the weight of the current evidence points toward an inverse association. Our results are partly consistent with a recent meta-analysis that found an inverse association between physical activity and preeclampsia in case–control studies,32 although no association was found in cohort studies (unlike our findings). However, that meta-analysis had used different methods than in our analysis and had pooled risk estimates for all categories above the reference category; thus, the comparison was for any versus low or no physical activity, while our analysis compared high versus low level of physical activity. We examined whether specific levels of physical activity were related to risk, as well, by conducting dose–response analyses. The differences for cohort studies between the previous meta-analysis and ours show the importance of dose–response analyses. Publication bias is a possibility. We did not find evidence of such bias with the statistical tests used or by inspection of the funnel plots, although the number of studies was low. Most of the included studies were from Europe and North America, and most participants were white. Studies would be needed from other geographic locations and among other ethnicities to clarify whether these findings can be generalized to other populations. Strengths of the present meta-analysis include the comprehensive search strategy, detailed dose–response, subgroup and sensitivity analyses, and the large sample size providing a more robust estimate of the association between physical activity and risk of preeclampsia.
In conclusion, women with higher levels of prepregnancy or early pregnancy physical activity have a 20% to 35% reduction in the RR of developing preeclampsia. Considering the few modifiable risk factors that have been established for preeclampsia, as well as the other health benefits of physical activity, promotion of physical activity for pregnant women may be a promising approach for reducing the risk of preeclampsia. Further studies are needed to clarify whether increasing physical activity in early pregnancy reduces the risk among previously inactive women and to further define the dose–response association for various types and intensities of physical activity in relation to preeclampsia. Furthermore, possible interactions with other risk factors and additive effects of weight control, dietary factors, and physical activity need to be clarified.
We thank Vitool Lohsoonthorn and Larissa R. Brunner Huber for replying to our inquiries.
1. Roberts JM, Cooper DW. Pathogenesis and genetics of pre-eclampsia. Lancet. 2001;357:53–56
2. World. Health Organization Collaboration. The World Health Report: Make Every Mother and Child Count. 2005. 2011 Geneva: Department of Reproductive Health and Research, WHO
3. . Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy. Am J Obstet Gynecol. 2000;183:S1–S22
4. Roberts JM, Pearson GD, Cutler JA, Lindheimer MDNational Heart Lung and Blood Institute. . Summary of the NHLBI Working Group on research on hypertension during pregnancy. Hypertens Pregnancy. 2003;22:109–127
5. Shahabi A, Wilson ML, Lewinger JP, Goodwin TM, Stern MC, Ingles SA. Genetic admixture and risk of hypertensive disorders of pregnancy among Latinas in Los Angeles County. Epidemiology. 2013;24:285–294
6. Mogren I, Högberg U, Winkvist A, Stenlund H. Familial occurrence of preeclampsia. Epidemiology. 1999;10:518–522
7. Ray JG, Diamond P, Singh G, Bell CM. Brief overview of maternal triglycerides as a risk factor for pre-eclampsia. BJOG. 2006;113:379–386
8. O’Brien TE, Ray JG, Chan WS. Maternal body mass index and the risk of preeclampsia: a systematic overview. Epidemiology. 2003;14:368–374
9. Duckitt K, Harrington D. Risk factors for pre-eclampsia at antenatal booking: systematic review of controlled studies. BMJ. 2005;330:565
10. Bellamy L, Casas JP, Hingorani AD, Williams DJ. Pre-eclampsia and risk of cardiovascular disease and cancer in later life: systematic review and meta-analysis. BMJ. 2007;335:974
11. Irgens HU, Reisaeter L, Irgens LM, Lie RT. Long term mortality of mothers and fathers after pre-eclampsia: population based cohort study. BMJ. 2001;323:1213–1217
12. Sofi F, Capalbo A, Cesari F, Abbate R, Gensini GF. Physical activity during leisure time and primary prevention of coronary heart disease: an updated meta-analysis of cohort studies. Eur J Cardiovasc Prev Rehabil. 2008;15:247–257
13. Whelton SP, Chin A, Xin X, He J. Effect of aerobic exercise on blood pressure: a meta-analysis of randomized, controlled trials. Ann Intern Med. 2002;136:493–503
14. Belo L, Santos-Silva A, Quintanilha A, Rebelo I. Similarities between pre-eclampsia and atherosclerosis: a protective effect of physical exercise? Curr Med Chem. 2008;15:2223–2229
15. Irwin DE, Savitz DA, St André KA, Hertz-Picciotto I. Study of occupational risk factors for pregnancy-induced hypertension among active duty enlisted Navy personnel. Am J Ind Med. 1994;25:349–359
16. Saftlas AF, Logsden-Sackett N, Wang W, Woolson R, Bracken MB. Work, leisure-time physical activity, and risk of preeclampsia and gestational hypertension. Am J Epidemiol. 2004;160:758–765
17. Rudra CB, Sorensen TK, Luthy DA, Williams MA. A prospective analysis of recreational physical activity and preeclampsia risk. Med Sci Sports Exerc. 2008;40:1581–1588
18. Magnus P, Trogstad L, Owe KM, Olsen SF, Nystad W. Recreational physical activity and the risk of preeclampsia: a prospective cohort of Norwegian women. Am J Epidemiol. 2008;168:952–957
19. Østerdal ML, Strøm M, Klemmensen AK, et al. Does leisure time physical activity in early pregnancy protect against pre-eclampsia? Prospective cohort in Danish women. BJOG. 2009;116:98–107
20. Hegaard HK, Ottesen B, Hedegaard M, et al. The association between leisure time physical activity in the year before pregnancy and pre-eclampsia. J Obstet Gynaecol. 2010;30:21–24
21. Tyldum EV, Romundstad PR, Slørdahl SA. Pre-pregnancy physical activity and preeclampsia risk: a prospective population-based cohort study. Acta Obstet Gynecol Scand. 2010;89:315–320
22. Vollebregt KC, Wolf H, Boer K, van der Wal MF, Vrijkotte TG, Bonsel GJ. Does physical activity in leisure time early in pregnancy reduce the incidence of preeclampsia or gestational hypertension? Acta Obstet Gynecol Scand. 2010;89:261–267
23. Fortner RT, Pekow PS, Whitcomb BW, Sievert LL, Markenson G, Chasan-Taber L. Physical activity and hypertensive disorders of pregnancy among Hispanic women. Med Sci Sports Exerc. 2011;43:639–646
24. Nugteren JJ, Snijder CA, Hofman A, Jaddoe VW, Steegers EA, Burdorf A. Work-related maternal risk factors and the risk of pregnancy induced hypertension and preeclampsia during pregnancy. The Generation R Study. PLoS One. 2012;7:e39263
25. Liu J, Trivedi T, Blair SN, Ness A, Macdonald-Wallis C, Lawlor DA. Physical activity and hypertensive disorders of pregnancy among British women. Am J Epidemiol. 2012;175(11 suppl):S22 (Abstract 086)
26. Longo-Mbenza B, Tshimanga KB, Buassa-bu-Tsumbu B, Kabangu MJ. Diets rich in vegetables and physical activity are associated with a decreased risk of pregnancy induced hypertension among rural women from Kimpese, DR Congo. Niger J Med. 2008;17:265–269
27. Marcoux S, Brisson J, Fabia J. The effect of leisure time physical activity on the risk of pre-eclampsia and gestational hypertension. J Epidemiol Community Health. 1989;43:147–152
28. Sorensen TK, Williams MA, Lee IM, Dashow EE, Thompson ML, Luthy DA. Recreational physical activity during pregnancy and risk of preeclampsia. Hypertension. 2003;41:1273–1280
29. Haelterman E, Marcoux S, Croteau A, Dramaix M. Population-based study on occupational risk factors for preeclampsia and gestational hypertension. Scand J Work Environ Health. 2007;33:304–317
30. Fang R, Dawson A, Lohsoonthorn V, Williams MA. Risk factors of early and late onset preeclampsia among Thai women. Asian Biomed (Res Rev News). 2009;3:477–486
31. Meher S, Duley L. Exercise or other physical activity for preventing pre-eclampsia and its complications. Cochrane Database Syst Rev. 2006:CD005942
32. Kasawara KT, do Nascimento SL, Costa ML, Surita FG, e Silva JL. Exercise and physical activity in the prevention of pre-eclampsia: systematic review. Acta Obstet Gynecol Scand. 2012;91:1147–1157
33. Moher D, Liberati A, Tetzlaff J, Altman DGPRISMA Group. . Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ. 2009;339:b2535
34. Rudra CB, Williams MA, Lee IM, Miller RS, Sorensen TK. Perceived exertion during prepregnancy physical activity and preeclampsia risk. Med Sci Sports Exerc. 2005;37:1836–1841
35. Yeo S, Davidge S, Ronis DL, Antonakos CL, Hayashi R, O’Leary S. A comparison of walking versus stretching exercises to reduce the incidence of preeclampsia: a randomized clinical trial. Hypertens Pregnancy. 2008;27:113–130
36. Yeo S. Adherence to walking or stretching, and risk of preeclampsia in sedentary pregnant women. Res Nurs Health. 2009;32:379–390
37. Yeo S. Prenatal stretching exercise and autonomic responses: preliminary data and a model for reducing preeclampsia. J Nurs Scholarsh. 2010;42:113–121
38. Rakhshani A, Nagarathna R, Mhaskar R, Mhaskar A, Thomas A, Gunasheela S. The effects of yoga in prevention of pregnancy complications in high-risk pregnancies: a randomized controlled trial. Prev Med. 2012;55:333–340
39. Asbee SM, Jenkins TR, Butler JR, White J, Elliot M, Rutledge A. Preventing excessive weight gain during pregnancy through dietary and lifestyle counseling: a randomized controlled trial. Obstet Gynecol. 2009;113(2 pt 1):305–312
40. Martin CL, Brunner Huber LR. Physical activity and hypertensive complications during pregnancy: findings from 2004 to 2006 North Carolina Pregnancy Risk Assessment Monitoring System. Birth. 2010;37:202–210
41. Spinillo A, Capuzzo E, Colonna L, Piazzi G, Nicola S, Baltaro F. The effect of work activity in pregnancy on the risk of severe preeclampsia. Aust N Z J Obstet Gynaecol. 1995;35:380–385
42. Longo-Mbenza B, Kadima-Tshimanga B, Buassa-bu-Tsumbu B, M’buyamba K Jr. Diets rich in vegetables and physical activity are associated with a decreased risk of pregnancy induced hypertension among rural women from Kimpese, DR Congo. Niger J Med. 2008;17:45–49
43. Nanjundan P, Bagga R, Kalra JK, Thakur JS, Raveendran A. Risk factors for early onset severe pre-eclampsia and eclampsia among north Indian women. J Obstet Gynaecol. 2011;31:384–389
44. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188
45. Hamling J, Lee P, Weitkunat R, Ambühl M. Facilitating meta-analyses by deriving relative effect and precision estimates for alternative comparisons from a set of estimates presented by exposure level or disease category. Stat Med. 2008;27:954–970
46. Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol. 1992;135:1301–1309
47. Ainsworth BE, Haskell WL, Leon AS, et al. Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc. 1993;25:71–80
48. Bagnardi V, Zambon A, Quatto P, Corrao G. Flexible meta-regression functions for modeling aggregate dose-response data, with an application to alcohol and mortality. Am J Epidemiol. 2004;159:1077–1086
49. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21:1539–1558
50. Egger M, Davey SG, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315:629–634
51. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994;50:1088–1101
52. Falcao S, Bisotto S, Michel C, et al. Exercise training can attenuate preeclampsia-like features in an animal model. J Hypertens. 2010;28:2446–2453
53. Durstine JL, Grandjean PW, Davis PG, Ferguson MA, Alderson NL, DuBose KD. Blood lipid and lipoprotein adaptations to exercise: a quantitative analysis. Sports Med. 2001;31:1033–1062
54. Butler CL, Williams MA, Sorensen TK, Frederick IO, Leisenring WM. Relation between maternal recreational physical activity and plasma lipids in early pregnancy. Am J Epidemiol. 2004;160:350–359
55. Ning Y, Williams MA, Butler CL, Muy-Rivera M, Frederick IO, Sorensen TK. Maternal recreational physical activity is associated with plasma leptin concentrations in early pregnancy. Hum Reprod. 2005;20:382–389
56. Oken E, Ning Y, Rifas-Shiman SL, Radesky JS, Rich-Edwards JW, Gillman MW. Associations of physical activity and inactivity before and during pregnancy with glucose tolerance. Obstet Gynecol. 2006;108:1200–1207
57. Clapp JF. Effects of diet and exercise on insulin resistance during pregnancy. Metab Syndr Relat Disord. 2006;4:84–90
58. Clapp JF 3rd, Kiess W. Effects of pregnancy and exercise on concentrations of the metabolic markers tumor necrosis factor alpha and leptin. Am J Obstet Gynecol. 2000;182:300–306
59. Kharb S. Lipid peroxidation in pregnancy with preeclampsia and diabetes. Gynecol Obstet Invest. 2000;50:113–116
60. Yeo S, Davidge ST. Possible beneficial effect of exercise, by reducing oxidative stress, on the incidence of preeclampsia. J Womens Health Gend Based Med. 2001;10:983–989
61. Wennmalm A, Fitzgerald GA. Excretion of prostacyclin and thromboxane A2 metabolites during leg exercise in humans. Am J Physiol. 1988;255(1 pt 2):H15–H18
62. Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB. Changes in diet and lifestyle and long-term weight gain in women and men. N Engl J Med. 2011;364:2392–2404
63. Jeon CY, Lokken RP, Hu FB, van Dam RM. Physical activity of moderate intensity and risk of type 2 diabetes: a systematic review. Diabetes Care. 2007;30:744–752
64. Tobias DK, Zhang C, van Dam RM, Bowers K, Hu FB. Physical activity before and during pregnancy and risk of gestational diabetes mellitus: a meta-analysis. Diabetes Care. 2011;34:223–229
65. Nocon M, Hiemann T, Müller-Riemenschneider F, Thalau F, Roll S, Willich SN. Association of physical activity with all-cause and cardiovascular mortality: a systematic review and meta-analysis. Eur J Cardiovasc Prev Rehabil. 2008;15:239–246
66. Jiang H, Qian X, Li M, et al. Can physical activity reduce excessive gestational weight gain? Findings from a Chinese urban pregnant women cohort study. Int J Behav Nutr Phys Act. 2012;9:12
67. Haakstad LA, Voldner N, Henriksen T, Bø K. Physical activity level and weight gain in a cohort of pregnant Norwegian women. Acta Obstet Gynecol Scand. 2007;86:559–564
68. Bekkering GE, Harris RJ, Thomas S, et al. How much of the data published in observational studies of the association between diet and prostate or bladder cancer is usable for meta-analysis? Am J Epidemiol. 2008;167:1017–1026