Original Article

Original Article Preeclampsia and Future Cardiovascular Health A Systematic Review and Meta-Analysis Pensée Wu, MBChB, MD(Res); Randula Haththotuwa, MBChB; Chun Shing Kwok, MBBS; Aswin Babu, BM, BS; Rafail A. Kotronias, MBChB; Claire Rushton, PhD; Azfar Zaman, MBChB, MD; Anthony A. Fryer, PhD; Umesh Kadam, MBChB, PhD; Carolyn A. Chew-Graham, MBChB, MD; Mamas A. Mamas, BM BCh, DPhil

Downloaded from http://circoutcomes.ahajournals.org/ by guest on May 1, 2018

Background—Preeclampsia is a pregnancy-specific disorder resulting in hypertension and multiorgan dysfunction. There is growing evidence that these effects persist after pregnancy. We aimed to systematically evaluate and quantify the evidence on the relationship between preeclampsia and the future risk of cardiovascular diseases. Methods and Results—We studied the future risk of heart failure, coronary heart disease, composite cardiovascular disease, death because of coronary heart or cardiovascular disease, stroke, and stroke death after preeclampsia. A systematic search of MEDLINE and EMBASE was performed to identify relevant studies. We used random-effects meta-analysis to determine the risk. Twenty-two studies were identified with >6.4 million women including >258 000 women with preeclampsia. Meta-analysis of studies that adjusted for potential confounders demonstrated that preeclampsia was independently associated with an increased risk of future heart failure (risk ratio [RR], 4.19; 95% confidence interval [CI], 2.09–8.38), coronary heart disease (RR, 2.50; 95% CI, 1.43–4.37), cardiovascular disease death (RR, 2.21; 95% CI, 1.83–2.66), and stroke (RR, 1.81; 95% CI, 1.29–2.55). Sensitivity analyses showed that preeclampsia continued to be associated with an increased risk of future coronary heart disease, heart failure, and stroke after adjusting for age (RR, 3.89; 95% CI, 1.83–8.26), body mass index (RR, 3.16; 95% CI, 1.41–7.07), and diabetes mellitus (RR, 4.19; 95% CI, 2.09–8.38). Conclusions—Preeclampsia is associated with a 4-fold increase in future incident heart failure and a 2-fold increased risk in coronary heart disease, stroke, and death because of coronary heart or cardiovascular disease. Our study highlights the importance of lifelong monitoring of cardiovascular risk factors in women with a history of preeclampsia. (Circ Cardiovasc Qual Outcomes. 2017;10:e003497. DOI: 10.1161/CIRCOUTCOMES.116.003497.) Key Words: arteries ◼ atherosclerosis ◼ heart failure ◼ hypertension ◼ preeclampsia ◼ pregnancy

P

reeclampsia is a major cause of maternal mortality worldwide1 and affects 2% to 8% of all pregnancies.2,3 It is confined to pregnancy and defined as onset of hypertension after 20-week gestation with proteinuria, organ dysfunction, or uteroplacental dysfunction.4 The pathogenesis of preeclampsia remains poorly understood and is thought to be because of the failure of spiral artery remodeling in the placenta causing placental hypoperfusion and hypoxia. The resultant oxidative stress triggers an excessive systemic inflammatory response, which causes endothelial dysfunction and vasoconstriction leading to systemic hypertension and end-organ hypoperfusion.2,5 There is growing evidence that these effects on end organs persist after pregnancy. Cardiovascular disease is a leading cause of mortality globally and also of maternal death in the United Kingdom

and United States.6,7 Several studies have examined the relationship between preeclampsia and future incident cardiovascular disease, although the literature has been inconsistent. Some studies reported significantly higher risks of composite cardiovascular events or heart failure,8,9 whereas others have not demonstrated such relationships.10,11 It is unclear whether preeclampsia is an independent risk factor for future cardiovascular disease or an early marker of women with high-risk profiles for future cardiovascular disease. Factors that predispose women to preeclampsia are also found in the risk profile for cardiovascular diseases. These include obesity,12 metabolic abnormalities, dyslipidemia, insulin resistance,13 heightened inflammatory responses, hypercoagulable states, and endothelial dysfunction.14 Alternatively, the body may not fully recover from the damage to the

Received October 16, 2016; accepted January 24, 2017. From the Institute for Science and Technology in Medicine (P.W.), Keele Cardiovascular Research Group (P.W., C.S.K., A.B., R.A.K., C.R., U.K., M.A.M.), Institute for Primary Care and Health Sciences (R.H., C.A.C.-G.), Institute for Applied Clinical Sciences (A.A.F., C.S.K., C.R., U.K., M.A.M.), and NIHR Collaboration for Leadership in Applied Health Research and Care (CLAHRC) West Midlands (C.A.C.-G.), Keele University, Stoke-on-Trent, United Kingdom; Academic Obstetrics and Gynaecology (P.W.) and The Heart Centre (C.S.K., M.A.M.), Royal Stoke University Hospital, Stoke-on-Trent, United Kingdom; and Freeman Hospital and Institute of Cellular Medicine, Newcastle University, Newcastle-Upon-Tyne, United Kingdom (A.Z.). The Data Supplement is available at http://circoutcomes.ahajournals.org/lookup/suppl/doi:10.1161/CIRCOUTCOMES.116.003497/-/DC1. Correspondence to Randula Haththotuwa, MBChB, Institute for Primary Care and Health Sciences, Keele University, Stoke-on-Trent, ST5 5BG, United Kingdom. E-mail [email protected] © 2017 American Heart Association, Inc. Circ Cardiovasc Qual Outcomes is available at http://circoutcomes.ahajournals.org

1

DOI: 10.1161/CIRCOUTCOMES.116.003497

2 Wu et al Preeclampsia and Future Cardiovascular Health

Data Sources and Searches

WHAT IS KNOWN • Although preeclampsia is confined to pregnancy, it is associated with increased risks of hypertension in the future.

WHAT THE STUDY ADDS


• After adjusting for potential confounders (age, body mass index, and diabetes mellitus), we showed that preeclampsia is associated with an increased risk of heart failure, stroke, coronary heart disease, cardiovascular disease, and death because of coronary heart or cardiovascular disease. • The increase in risk for heart failure, stroke, and cardiovascular disease death is higher during the first 10 years after a pregnancy affected by preeclampsia compared with that beyond 10 years. • Future cost–benefit analysis regarding the optimal timing for cardiovascular risk screening after preeclampsia is needed in this high-risk population.

MEDLINE and EMBASE were searched using OVID SP for studies from January 2005 to August 2015. The detailed search strategy and search terms are outlined in Methods in the Data Supplement. The relevant primary studies for inclusion on this study were extracted from a comprehensive program of evidence synthesis, which explored the association between preeclampsia and adverse cardiovascular or metabolic outcomes. A search for additional articles was also conducted through manual searching of the bibliography of relevant review articles and meta-analyses.

Study Selection and Data Extraction Four reviewers (P.W., R.H., R.A.K., and A.B.) screened all titles that met the inclusion criteria. This was followed by a screen of the remaining abstracts. The full articles were screened by the same 4 reviewers, and the final decision to include studies was made by P.W., R.H., and C.S.K. Independent double data extraction was done by 4 reviewers (P.W., R.H., R.A.K., and A.B.). Data were collected on study design, year, country, number of participants, mean age, parity, cohort characteristics, definition and ascertainment of preeclampsia, ascertainment of outcomes, timing of assessment, adequacy of follow-up, and results. The information was obtained from published data.

Study Quality Assessment vascular, endothelial, and metabolic systems associated with preeclampsia and may manifest in later life with future cardiovascular events.2 Lipid deposition in the spiral artery walls is more commonly seen in preeclamptic than healthy pregnancies, which mimic the early stages of atherosclerosis.15 Although the national guidance in the United States16,17 and the United Kingdom18 recommend that after a diagnosis of preeclampsia, women should be counseled and followed up for cardiovascular risk modification, no conclusive evidence exists for an effective risk reduction strategy, and such followup is probably not done in practice. In addition, more recent studies have shown nonsignificant and conflicting data for the level of cardiovascular risk after preeclampsia.19,20 There is a need to reevaluate and quantify the risk of cardiovascular events after preeclampsia to guide future management and risk modification, to contribute to guidelines for clinicians. To this end, we conducted a systematic review and meta-analysis of contemporary studies (published 2005–2015) to quantify the future risk of cardiovascular events in women after preeclampsia.

Methods Eligibility Criteria We selected studies investigating the long-term cardiovascular outcomes of women with and without preeclampsia published in the English language between 2005 and August 2015. There was no restriction on the definition of preeclampsia. Primary cardiovascular outcomes were heart failure; coronary heart disease; death because of coronary heart disease; composite cardiovascular disease defined as a combination of cardiac, cerebrovascular, and peripheral vascular disease; death because of composite cardiovascular disease; stroke; and stroke death. The included studies had at least 2 groups (1 with preeclampsia and 1 without preeclampsia) and reported sufficient data to allow for accurate risk estimates to be calculated. Studies assessing outcomes during antepartum or before 6 weeks postpartum were excluded. There was no restriction based on cohort type, study design, or duration of follow-up.

Study quality was assessed against the Newcastle-Ottawa Quality Assessment Scale for cohort studies.21 Our gold standard for each of the criteria were selection of exposed cohort from the general population of pregnant women; selection of nonexposed cohort from the same population; reliable ascertainment of exposure such that the likelihood of controls being misclassified as having preeclampsia when they did not or cases being wrongly classified as not having preeclampsia was minimized; exclusion of women who had cardiovascular outcome of interest before or during pregnancy; comparable cohort where confounders, in particular, age, smoking, and other cardiovascular risk factors, were accounted for; assessment of outcomes prospectively or through linkage of records and independent blind assessment; follow-up duration for at least 1 year postpartum; and 10 studies and little evidence of heterogeneity, we performed funnel plots to assess for publication bias.

Results Description of Studies Included in Analysis The initial MEDLINE and EMBASE search produced 9964 titles and abstracts, after which 22 studies were included in the analysis (Figure 1). The studies examined 6 456 379 women in total (ranges from 137–2 066 230 women in each study). Studies recruiting patients from the same population were paired to avoid duplication of participant numbers.8,23–25


Figure 1. Flow diagram of study inclusion. CVD indicates cardiovascular disease. Downloaded from http://circoutcomes.ahajournals.org/ by guest on May 1, 2018

Details of study design and demographics are shown in Table I in the Data Supplement. From the 17 studies that reported the number of women in each group, there were 258 275 women with preeclampsia and 4 006 431 controls. Four studies recruited primiparous women only,8,19,24,26 whereas 17 studies included women of any parity.9–11,23,25,27–38 Studies reporting a mean or median age at enrollment ranged from 23.4 to 32.3 years, whereas follow-up ranged from 6 weeks postpartum to 39.4 years.

Quality Assessment of Included Studies The study quality was evaluated using the Newcastle-Ottawa Quality Assessment Scale for cohort studies21 as shown in Tables II and III in the Data Supplement. Nineteen studies had reliable methods of ascertaining preeclampsia from databases,8,9,19,23–26,30,32,35,36 medical records,10,11,20,28,34 prospective measurements,29,33 or through completion of questionnaire with trained staff.30,31 Nineteen studies used reliable methods of obtaining cardiovascular outcomes from databases,8–10,19,23–26,28,32,34–36 medical records,11 prospective measurements with echocardiography,33,38 or through completion of questionnaire with trained staff.29–31 There was adequate follow-up (>90%) in 17 studies.8,9,11,23–28,30–36,38 Eighteen studies used adjusted analyses.8,9,19,20,23–26,28–33,35–38

Determining Preeclampsia and Results of Studies Various methods were used to ascertain preeclampsia, with the most common being the International Society of the Studies

of Hypertension in Pregnancy (2014) definition.4 Table IV in the Data Supplement shows the results of the studies.

Pooled Analysis of Preeclampsia and Cardiovascular Outcomes The risk of heart failure with preeclampsia is shown in Figure 2. The pooled results of 7 studies8,10,23,31,33,35,38 suggest a 3.6-fold increased risk of heart failure with preeclampsia (RR, 3.62; 95% confidence interval [CI], 2.25–5.85; I2=83%; 2 764 824 participants). The risk increases to >4-fold for the adjusted studies10,23,35,38 (adjusted risk ratio [aRR], 4.19; 95% CI, 2.09–8.38; I2=71%; 1 986 285 participants). The factors that have been adjusted for in the studies are shown in Table III in the Data Supplement. We performed leave-out analyses to explore the sources of heterogeneity (Table VI in the Data Supplement). The heterogeneity was mainly driven by the study by Männistö et al10; if this study was excluded, heterogeneity was reduced to 46% in the adjusted analysis (aRR, 5.57; 95% CI, 3.14–9.88). The relationship between preeclampsia and future risk of coronary heart disease and coronary heart disease death is shown in Figures 3 and 4. For coronary heart disease, there was a 2-fold increased risk of events with preeclampsia (RR, 2.11; 95% CI, 1.60–2.77; I2=87%; 3 239 797 participants). The risk was even greater in the adjusted studies10,19,23,25,31,35,37 (aRR, 2.50; 95% CI, 1.43–4.37; I2=89%; 2 068 673 participants). The heterogeneity was mainly driven by the study by Lin et al23; if this study was excluded, heterogeneity was reduced to 66% in


Figure 2. Risk of heart failure with preeclampsia (PE).8,10,23–25,31,33,35,38 CI indicates confidence interval.


the adjusted analysis (aRR, 1.67; 95% CI, 1.19–2.33; Table VI in the Data Supplement). The 4 adjusted studies10,19,34,36 reporting coronary heart disease death also show a 2-fold increased risk with preeclampsia (aRR, 2.10; 95% CI, 1.25–3.51; I2=64%; 677 378 participants). The heterogeneity was mainly driven by the study by Bhattacharya et al19; if this study was excluded, heterogeneity was reduced to 2% in the adjusted analysis (aRR, 2.63; 95% CI, 1.74–3.98; Table VI in the Data Supplement). The risk of composite cardiovascular disease and cardiovascular disease death with preeclampsia is shown in Figure I in the Data Supplement and Figure 5. The pooled results from 6 studies10,11,20,23,27,32 with 1 398 119 participants suggest a significantly increased risk of cardiovascular disease (RR, 1.65; 95% CI, 1.36–2.01; I2=42%). However, the results were not statistically significant for the 3 studies10,20,23 that adjusted for baseline confounders before pregnancy (aRR, 1.85; 95% CI, 0.80–4.29; I2=72%). For cardiovascular disease death, the pooled results of 4 studies23–25,28,36 with 2 614 180 participants suggest a 2-fold increase in cardiovascular disease death with preeclampsia (aRR, 2.21; 95% CI, 1.83–2.66; I2=54%). All 4 studies adjusted for potential confounders. Figure 6 and Supplemental Figure II in the Data Supplement show the results for pooled analysis for studies on preeclampsia and stroke and stroke death. For stroke,

there was a 2-fold increased risk of events with preeclampsia (RR, 1.71; 95% CI, 1.38–2.11; I2=69%; 4 906 182 participants). This increase in risk persisted in studies that adjusted for confounders (aRR, 1.81; 95% CI, 1.29–2.55; I2=74%; 4 131 344 participants).9,10,19,25,35,37 The heterogeneity was mainly driven by the study by Bhattacharya et al19; if this study was excluded, heterogeneity was reduced to 24% in the adjusted analysis (aRR, 2.04; 95% CI, 1.60–2.60; Table VI in the Data Supplement). Two studies that reported stroke death showed a 2-fold increased risk with preeclampsia; however, the result was not statistically significant (aRR, 1.97; 95% CI, 0.80–4.88; I2=86%).

Sensitivity Analysis for Follow-Up Time We conducted sensitivity analyses to consider the effect of follow-up time for cardiovascular outcomes that were significant in the adjusted studies (Table 1). The risk of heart failure was the highest 1 to 10 years (aRR, 8.42; 95% CI, 4.39–16.17) after the preeclamptic pregnancy compared with 10 years (aRR, 1.60; 95% CI, 0.73–3.50) postpartum. For coronary heart disease (aRR, 3.10; 95% CI, 1.56–6.16) and stroke (aRR, 2.22; 95% CI, 1.73–2.85), the increase in risk was significant within the first year after delivery compared with other time points. In cardiovascular disease death, the increase in risk were similar

Figure 3. Risk of coronary heart disease with preeclampsia (PE).8,10,19,23–26,29–31,35,37 CI indicates confidence interval.


Figure 4. Risk of coronary heart disease death with preeclampsia (PE).10,19,34,36 CI indicates confidence interval.

at 1 to 10 years (aRR, 2.30; 95% CI, 1.65–3.20) and >10 years (aRR, 2.21; 95% CI, 1.73–2.81) post-delivery. All studies on coronary heart disease death had a follow-up of >10 years; therefore, we could not conduct further sensitivity analysis on duration of follow-up.


Sensitivity Analysis Considering Studies That Adjusted for or Had Exclusions Based on Baseline Age, Body Mass Index or Weight, Diabetes Mellitus or Gestational Diabetes Mellitus, Smoking, and Hypertension Between Preeclamptic and Control Groups Sensitivity analyses were performed to consider age as a confounding factor in the 5 cardiovascular outcomes that were significant in adjusted studies (Table 2). These showed that the risk remained significant in all outcomes: heart failure (aRR, 3.89; 95% CI, 1.83–8.26), coronary heart disease (aRR, 3.13; 95% CI, 1.45–6.75), coronary heart disease death (aRR, 2.63; 95% CI, 1.74–3.98), cardiovascular disease death (aRR, 2.21; 95% CI, 1.83–2.66), and stroke (aRR, 2.04; 95% CI, 1.60–2.60). The effect of pregestational body mass index (BMI) or weight and pregestational diabetes mellitus or gestational diabetes mellitus was examined for the heart failure, coronary heart disease, and stroke outcomes. The risk for all 3 outcomes remained significantly increased despite adjustment for BMI or weight (heart failure: aRR, 2.74; 95% CI, 1.10–6.83; coronary heart disease: aRR, 1.84; 95% CI, 1.23–2.74; stroke: aRR, 1.94; 95% CI, 1.42–2.65) and diabetes mellitus or gestational diabetes mellitus (heart failure: aRR 3.89; 95% CI, 1.83–8.26; coronary heart disease: aRR, 2.16; 95% CI, 1.03– 4.52; stroke: aRR, 2.46; 95% CI, 1.11–5.43). We considered the effect of pregestational smoking for coronary heart disease and stroke outcomes and found that the increased risk remained significant (coronary heart disease: aRR, 1.56; 95% CI, 1.11–2.20; stroke: aRR, 1.64; 95% CI, 1.12–2.40). However, when we examined the effect of pregestational hypertension for the coronary heart disease outcome, the increase in risk was nonsignificant. We could not examine other important confounding factors, such as family history

of cardiovascular disease or hypercholesterolemia, because of the lack of studies presenting these data. The full cardiovascular risk factor profile of the preeclampsia and the control population is shown in Table V in the Data Supplement. There were significant differences in age, BMI, diabetes mellitus, smoking, and blood pressure between the preeclamptic and control groups at baseline in 1,32 3,10,19,32 1,32 2,10,19 and 1 studies,11 respectively, which only contributed to 2% of total participant women. However, the cardiovascular risk factor profiles were not available in the majority of the studies included in this systematic review and meta-analysis. Only 3 studies had adjusted their results to account for the risk factor profile differences in age,32 BMI,10 diabetes mellitus,32 and smoking.10,19

Discussion In this systematic review and meta-analysis, 22 studies with >6.4 million women were included. We showed an association of preeclampsia with future incident coronary heart disease, composite cardiovascular disease, heart failure, stroke, and deaths because of coronary heart disease. The adjusted risk ranged between 1.8-and 2.5-fold compared with those without a history of preeclampsia in all cardiac outcomes, except in heart failure, where a 4-fold increase in risk was found. For coronary heart disease, heart failure, and stroke, the increase in risk remained significant after adjusting for age, BMI, and diabetes mellitus. Preeclampsia is a well-recognized risk factor for future hypertension. Our study confirms preeclampsia to be a risk factor for future cardiac disease, although we are unable to determine whether this relationship is confounded by an adverse cardiovascular risk factor profile in patients with preeclampsia at baseline or whether preeclampsia is an independent risk factor. Our meta-analysis supports previous literature in terms of a 2-fold increased risk of cardiovascular disease death39,40 and stroke,41 but unique to this study, we conducted sensitivity analyses to consider the effects of potential confounding factors, such as age, BMI, diabetes mellitus, hypertension, and smoking, as well as examining the cardiovascular risk factors profiles at baseline. Compared with previous meta-analyses,

Figure 5. Risk of cardiovascular disease death with preeclampsia (PE).8,23–25,28,36 CI indicates confidence interval.


Figure 6. Risk of stroke with preeclampsia (PE).8–10,19,23–25,35,37 CI indicates confidence interval.


we considered heart failure as a separate entity and showed a much higher risk than previously reported in composite cardiac outcomes. Because of the gaps in the current literature, it is difficult to ascertain whether confounding factors have contributed toward the associations we identified. After considering the effects of hypertension in the future coronary heart disease outcome, we found that the link between preeclampsia and future coronary heart disease was no longer statistically significant. This suggests that the association may be confounded by hypertension. We were unable to fully evaluate the effects of all confounding factors and undertake further sensitivity analysis because of the absence of such data in the included studies. The increased risk in future cardiovascular diseases may be driven by unmeasured confounders as none of the included studies have adjusted for all of the established cardiovascular risk factors, such as age, BMI, diabetes mellitus, family history of cardiovascular disease, hypercholesterolemia, hypertension, and smoking. For instance, only 2 of the 4 studies that examined heart failure had adjusted for pre-existing cardiovascular disease before delivery.23,35 Preeclampsia and cardiovascular diseases are known to share risk factors. According to the American College of Obstetricians and Gynecologists, the recognized risk factors for preeclampsia are obesity, chronic hypertension, diabetes mellitus, chronic renal disease, previous preeclampsia, systemic lupus erythematosus, age >40

years, primiparity, multiple pregnancy, in vitro fertilization, and family history of preeclampsia.42 Apart from the pregnancy-specific factors and age, all other risk factors overlap with those recognized by American Heart Association17 and American Stroke Association.16 These factors may be in the same causal pathways for cardiovascular diseases, and it is difficult to establish whether preeclampsia is a predictor of cardiovascular events through distinct pathways or through inherent adverse cardiovascular profiles, such as obesity and hypertension, in women who develop preeclampsia. There is also a potential of confounding by pregnancyinduced hypertension in 2 of the studies (Kessous et al32 and Hovsepian et al9) as women with pregnancy-induced hypertension could have been included in the preeclamptic group. Historically, it was thought that pregnancy-induced hypertension is a milder form of preeclampsia; however, it is increasing speculated that preeclampsia and pregnancy-induced hypertension have separate underlying pathophysiological mechanisms with distinct cardiovascular consequences for the women and their offspring. Therefore, we have conducted additional sensitivity analyses in the stroke and cardiovascular disease outcomes where these 2 studies were excluded and showed that the increased risk persisted (stroke: aRR, 1.75; 95% CI, 1.15–2.65 and cardiovascular disease: RR, 1.79; 95% CI, 1.40–2.31). A mechanistic link to explain the association between preeclampsia and future heart failure remains elusive.

Table 1. Sensitivity Analysis With Regard to Duration of Follow-Up Outcomes

10 y

Cardiovascular disease death

Adjusted

…

2.30 (1.65–3.20), n=1

2.21 (1.73–2.81), n=3

Coronary heart disease

Adjusted

3.10 (1.56–6.15), n=1

3.78 (0.43–77.30), n=2

1.46 (0.95–2.25), n=3

Unadjusted

…

…

2.09 (1.64–2.66), n=3

Adjusted

…

…

2.10 (1.25–3.51), n=4

Coronary heart disease death Heart failure

Stroke

Adjusted

4.10 (2.90–5.80), n=1

8.42 (4.39–16.17), n=2

1.60 (0.73–3.50), n=1

Unadjusted

…

4.27 (2.09–8.71), n=1

2.73 (1.30–5.74), n=2

Adjusted

2.22 (1.73–2.85), n=2

3.56 (0.52–24.28), n=2

1.18 (0.95–1.46), n=2

Unadjusted

…

…

1.60 (1.47–1.74), n=1

Values are represented as risk ratio (95% CI). CI indicates confidence interval.

7 Wu et al Preeclampsia and Future Cardiovascular Health Table 2. Sensitivity Analysis With Regard to Age, Pregestational Body Mass Index or Weight, Pregestational Smoking, Pregestational Diabetes Mellitus or Gestational Diabetes Mellitus, and Pregestational Hypertension Outcomes

Age

BMI/Weight

Diabetes Mellitus/GDM

Smoking

Hypertension

Cardiovascular mortality

2.21 (1.83–2.66), n=4

…

…

…

…

Coronary heart disease

3.13 (1.45–6.75), n=5

1.84 (1.23–2.74), n=3

2.16 (1.03–4.52), n=2

1.56 (1.11–2.20), n=4

3.84 (0.81–18.16), n=3

Coronary heart disease death

2.63 (1.74–3.98), n=3

…

…

…

…

Heart failure

3.89 (1.83–8.26), n=3

2.74 (1.10–6.83), n=2

3.89 (1.83–8.26), n=3

…

…

Stroke

2.04 (1.60–2.60), n=5

1.94 (1.42–2.65), n=3

2.46 (1.11–5.43), n=3

1.64 (1.12–2.40), n=4

…

Values are represented as risk ratio (95% CI). BMI indicates body mass index; CI, confidence interval; and GDM, gestational diabetes mellitus.


Preeclampsia has been found to be associated with a 3-fold risk of future hypertension40 and a doubling in risk of future diabetes mellitus.43 Therefore, heart failure in this group of women may be because of causes other than ischemic cardiomyopathy, for example, hypertensive heart disease or diabetic cardiomyopathy. It has been suggested that the left ventricular remodeling and hypertrophy seen during preeclamptic pregnancies continues after the pregnancy.33 On the contrary, women with pre-existing left ventricular dysfunction and placental insufficiency are more likely to develop preeclampsia than those with normal left ventricular function, suggesting a correlation rather than causation.44 Furthermore, in peripartum cardiomyopathy where a significant proportion of women continue to have left ventricular dysfunction years after delivery, the prevalence of preeclampsia is 4 times more common than that in healthy pregnancies.45 Risk stratification enables early identification of women at high risk of cardiovascular disease and allows optimization of targeted care toward prevention. There are several cardiovascular risk scores in use, although most were developed with an under-representation of women. The Reynolds risk score is sex specific46 and has been shown to perform better than the Framingham risk score in predicting cardiovascular events in women.47 However, currently, there are no risk calculators incorporating pregnancy complications. The American Heart Association has included preeclamptic and gestational hypertension as a risk factor for coronary artery disease and advocates active follow-up of risk factors.17 The association suggest that preeclampsia manifesting in pregnancy is akin to a failed stress test of future vascular or metabolic disease. Using the Framingham prediction score, preeclampsia has been found to be independently associated with an increased 10-year cardiovascular risk score both at 3- to 8-year48 and at 18-year49 follow-up intervals. The strength of our study is the large sample size from contemporary studies with a total of 45 938 256 patient-years of follow-up. As a part of a larger program of evidence synthesis, we used a comprehensive search strategy to examine the long-term cardiovascular outcomes in preeclampsia. The inclusion of more recent studies means that there is a greater likelihood of their findings being relevant and more generalizable to current practice. In our study, we used independent reviewers for performing double data extractions and for data analysis. The majority of the studies were designed to examine future cardiovascular diseases as their main outcome (n=21) and contribute to 99% of the women in our meta-analysis.

The main limitation of our study is that significant unmeasured confounders may have contributed to our reported association between preeclampsia and future cardiovascular risk. Although most of the studies (n=20) have attempted to adjust for some potential confounding factors, none of the studies included here have adequately adjusted for all relevant risk factors. In the few studies (2.6% of total participants) that reported baseline cardiovascular risk factor profiles, a majority of the cohort (2% of total participants) had significant baseline risk factor profile differences between the preeclamptic and the nonpreeclamptic populations at index pregnancy. Other limitations include an inherent limitation from publication bias, where studies with positive findings are more likely to be published than those with negative or neutral results. As the majority were from retrospective studies, we had limited control over the quality of data collected. There could have been incomplete, inaccurate, or inconsistent historical data on exposure and recall bias, which could have affected whether the case and control groups were ascribed correctly. In particular, 4 studies used questionnaire data to assess the cardiovascular outcomes.27,30,31,37 Furthermore, we limited the studies to those in English and may have missed important research data from non-English publications. Potential reasons for heterogeneity may be because of differences in the study population, research methodology, and inherent differences between studies. As shown in Table I in the Data Supplement, 3 studies were conducted in ethnically diverse populations to the other studies, as they were conducted on the continent of Asia (Funai et al,28 Kessous et al,32 Lin et al,23 and Tang et al25), whereas the remainder were in Europe or North America. Two studies examined primiparous women only (Bhattacharya et al19 and Wikström et al26), whereas the others studied women of any parity. Specific populations were analyzed in 2 studies, which were by Gordin et al29 (women with type 1 diabetes mellitus) and Stuart et al37 (nurses). In terms of methodology, Table III in the Data Supplement shows that there were differences between methods of data collection and the actual type of data collected. Most of the data were collected through medical records or databases, where codes were used to identify the outcomes. However, 5 studies used questionnaires or interviews (Andersgaard et al,27 Gordin et al,29 Haukkamaa et al,30 Kaaja et al,31 and Stuart et al37), and 2 studies performed echocardiography (GhosseinDoha et al38 and Melchiorre et al33). Although codes are more objective, the research team is relying on historical accounts



in the records from the clinician. However, though questionnaires, interviews, or echocardiograms ensure direct patient contact with reliable outcome data collection, these methods are prone to selection bias from the research team. The leave-out analyses for the main outcomes (Table VI in the Data Supplement) demonstrates that there are underlying differences in the study cohorts that may also have contributed to the heterogeneity we observed. In the heart failure outcome, the heterogeneity was mainly driven by the study by Männistö et al10 within the adjusted studies. Compared with the other 3 studies, the study by Männistö et al was the only study with a prospective design. For the coronary heart disease outcome, the heterogeneity was mainly from the study by Lin et al,23 which is the only research conducted in Asia within this group of studies. For both the coronary heart disease death and stroke outcomes, the heterogeneity was mainly driven by the study by Bhattacharya et al.19 In this study, the participants were younger in their index pregnancy with preeclampsia (out of the studies that included data on age during index pregnancy) compared with participants in the other studies. In the sensitivity analysis regarding duration of follow-up, the increased risk for heart failure was greater at 1 to 10 years compared with 10 years postnatally, the risk was lower than that at 1 to 10 years and no longer significant. This may be because of a higher absolute risk in the control group during longer follow-up periods (ie, >10 years). Therefore, the reduction in the relative risk may be a product of the higher baseline risk. Furthermore, the effect is exacerbated by the small sample size and number of events in the control group reported.10 Further research is required to determine whether women with preeclampsia have an adverse cardiovascular risk factor profile at baseline, which contributes to their increased risk of cardiovascular diseases in later life. The development of effective strategies toward reducing these unfavorable risk profiles are required to plan the logistics for follow-up of these women at high risk, such as where, when, and how they should be followed-up and the type of intervention needed to rescue the progression of adverse events. The perinatal period is an opportune time for health screening, education, intervention, and monitoring in at-risk women. In a focus group study, women with previous preeclampsia were found to be generally unaware of their increased risk of future cardiovascular diseases but were motivated to undertake lifestyle modifications to reduce risk.50 In view of the burden and impact of cardiovascular disease on women in our society, we recommend a detailed cost–benefit analysis to determine the postnatal timing for a screening program in this high-risk population.

Conclusions Over 258 000 women with preeclampsia were examined in this meta-analysis of 22 studies. We found that preeclampsia is linked with a 4-fold increase in future heart failure. Significant unmeasured confounders may have contributed to the association we identified. In keeping with current recommendations, the findings from our study highlight the importance of patient education about risk and lifestyle modifications to reduce risk, as well as regular monitoring of cardiovascular risk factors in women with a history of preeclampsia.

Sources of Funding This work was supported by a grant from the North Staffordshire Heart Committee. R. Haththotuwa and C.S. Kwok are funded by the National Institute for Health Research Academic Clinical Fellowships.

Disclosures None.

References 1. Say L, Chou D, Gemmill A, Tunçalp Ö, Moller AB, Daniels J, Gülmezoglu AM, Temmerman M, Alkema L. Global causes of maternal death: a WHO systematic analysis. Lancet Glob Health. 2014;2:e323–e333. doi: 10.1016/S2214-109X(14)70227-X. 2. Ahmed R, Dunford J, Mehran R, Robson S, Kunadian V. Pre-eclampsia and future cardiovascular risk among women: a review. J Am Coll Cardiol. 2014;63:1815–1822. doi: 10.1016/j.jacc.2014.02.529. 3. Duley L. The global impact of pre-eclampsia and eclampsia. Semin Perinatol. 2009;33:130–137. doi: 10.1053/j.semperi.2009.02.010. 4. Tranquilli AL, Dekker G, Magee L, Roberts J, Sibai BM, Steyn W, Zeeman GG, Brown MA. The classification, diagnosis and management of the hypertensive disorders of pregnancy: a revised statement from the ISSHP. Pregnancy Hypertens. 2014;4:97–104. doi: 10.1016/j.preghy.2014.02.001. 5. Roberts JM, Hubel CA. The two stage model of preeclampsia: variations on the theme. Placenta. 2009;30(suppl A):S32–S37. doi: 10.1016/j. placenta.2008.11.009. 6. Townsend N, Williams J, Bhatnagar P, Wickramasinghe K, Rayner M. Cardiovascular Disease Statistics. London: British Heart Foundation; 2014. 7. Knight M, Tuffnell D, Kenyon S, Shakespeare J, Gray R, Kurinczuk JJ (Eds.) on behalf of MBRRACE-UK. Saving Lives, Improving Mothers’ Care - Surveillance of maternal deaths in the UK 2011-13 and lessons learned to inform maternity care from the UK and Ireland Confidential Enquiries into Maternal Deaths and Morbidity 2009-13. Oxford: National Perinatal Epidemiology Unit, University of Oxford 2015. 8. Lykke JA, Langhoff-Roos J, Sibai BM, Funai EF, Triche EW, Paidas MJ. Hypertensive pregnancy disorders and subsequent cardiovascular morbidity and type 2 diabetes mellitus in the mother. Hypertension. 2009;53:944– 951. doi: 10.1161/HYPERTENSIONAHA.109.130765. 9. Hovsepian DA, Sriram N, Kamel H, Fink ME, Navi BB. Acute cerebrovascular disease occurring after hospital discharge for labor and delivery. Stroke. 2014;45:1947–1950. doi: 10.1161/STROKEAHA.114.005129. 10. Männistö T, Mendola P, Vääräsmäki M, Järvelin MR, Hartikainen AL, Pouta A, Suvanto E. Elevated blood pressure in pregnancy and subsequent chronic disease risk. Circulation. 2013;127:681–690. doi: 10.1161/ CIRCULATIONAHA.112.128751. 11. Nijdam ME, Timmerman MR, Franx A, Bruinse HW, Numans ME, Grobbee DE, Bots ML. Cardiovascular risk factor assessment after pre-eclampsia in primary care. BMC Fam Pract. 2009;10:77. doi: 10.1186/1471-2296-10-77. 12. O’Brien TE, Ray JG, Chan WS. Maternal body mass index and the risk of preeclampsia: a systematic overview. Epidemiology. 2003;14:368–374. 13. Innes KE, Wimsatt JH, McDuffie R. Relative glucose tolerance and subsequent development of hypertension in pregnancy. Obstet Gynecol. 2001;97:905–910. 14. Roberts JM. Endothelial dysfunction in preeclampsia. Semin Reprod Endocrinol. 1998;16:5–15. doi: 10.1055/s-2007-1016248. 15. Staff AC, Dechend R, Pijnenborg R. Learning from the placenta: acute atherosis and vascular remodeling in preeclampsia-novel aspects for atherosclerosis and future cardiovascular health. Hypertension. 2010;56:1026–1034. doi: 10.1161/HYPERTENSIONAHA.110.157743. 16. Bushnell C, McCullough LD, Awad IA, Chireau MV, Fedder WN, Furie KL, Howard VJ, Lichtman JH, Lisabeth LD, Piña IL, Reeves MJ, Rexrode KM, Saposnik G, Singh V, Towfighi A, Vaccarino V, Walters MR; American Heart Association Stroke Council; Council on Cardiovascular and Stroke Nursing; Council on Clinical Cardiology; Council on Epidemiology and Prevention; Council for High Blood Pressure Research. Guidelines for the prevention of stroke in women: a statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2014;45:1545–1588. doi: 10.1161/01.str.0000442009.06663.48. 17. Mosca L, Benjamin EJ, Berra K, Bezanson JL, Dolor RJ, Lloyd-Jones DM, Newby LK, Piña IL, Roger VL, Shaw LJ, Zhao D, Beckie TM, Bushnell C, D’Armiento J, Kris-Etherton PM, Fang J, Ganiats TG, Gomes



AS, Gracia CR, Haan CK, Jackson EA, Judelson DR, Kelepouris E, Lavie CJ, Moore A, Nussmeier NA, Ofili E, Oparil S, Ouyang P, Pinn VW, Sherif K, Smith SC Jr, Sopko G, Chandra-Strobos N, Urbina EM, Vaccarino V, Wenger NK. Effectiveness-based guidelines for the prevention of cardiovascular disease in women–2011 update: a guideline from the american heart association. Circulation. 2011;123:1243–1262. doi: 10.1161/ CIR.0b013e31820faaf8. 18. National Institute for Health and Care Excellence. Hypertension in Pregnancy: The Management of Hypertensive Disorders During Pregnancy. (NICE Clinical Guidelines, No. 107). UK: RCOG Press; 2010. 19. Bhattacharya S, Prescott GJ, Iversen L, Campbell DM, Smith WC, Hannaford PC. Hypertensive disorders of pregnancy and future health and mortality: a record linkage study. Pregnancy Hypertens. 2012;2:1–7. doi: 10.1016/j.preghy.2011.08.116. 20. Grandi SM, Vallée-Pouliot K, Eberg M, Platt RW, Arel R, Filion KB. Abstract MP20: hypertensive disorders in pregnancy and the risk of incident cardiovascular disease. Circulation. 2015;131:AMP20–AMP20. 21. Wells G, Shea B, O’connell D, Peterson J, Welch V, Losos M, Tugwell P. The Newcastle-Ottawa Scale (NOS) for Assessing the Quality of Nonrandomised Studies in Meta-Analyses. 2000. Available from: www. ohri.ca/programs/clinical_epidemiology/oxford.asp. Accessed October 16, 2016. 22. Higgins JP, Green S. Cochrane Handbook for Systematic Reviews of Interventions. Wiley Online Library; 2008. 23. Lin YS, Tang CH, Yang CY, Wu LS, Hung ST, Hwa HL, Chu PH. Effect of pre-eclampsia-eclampsia on major cardiovascular events among peripartum women in Taiwan. Am J Cardiol. 2011;107:325–330. doi: 10.1016/j. amjcard.2010.08.073. 24. Lykke JA, Langhoff-Roos J, Lockwood CJ, Triche EW, Paidas MJ. Mortality of mothers from cardiovascular and non-cardiovascular causes following pregnancy complications in first delivery. Paediatr Perinat Epidemiol. 2010;24:323–330. doi: 10.1111/j.1365-3016.2010.01120.x. 25. Tang CH, Wu CS, Lee TH, Hung ST, Yang CY, Lee CH, Chu PH. Preeclampsia-eclampsia and the risk of stroke among peripartum in Taiwan. Stroke. 2009;40:1162–1168. doi: 10.1161/STROKEAHA.108.540880. 26. Wikström AK, Haglund B, Olovsson M, Lindeberg SN. The risk of maternal ischaemic heart disease after gestational hypertensive disease. BJOG. 2005;112:1486–1491. doi: 10.1111/j.1471-0528.2005.00733.x. 27. Andersgaard AB, Acharya G, Mathiesen EB, Johnsen SH, Straume B, Øian P. Recurrence and long-term maternal health risks of hypertensive disorders of pregnancy: a population-based study. Am J Obstet Gynecol. 2012;206:143.e1–143.e8. doi: 10.1016/j.ajog.2011.09.032. 28. Funai EF, Friedlander Y, Paltiel O, Tiram E, Xue X, Deutsch L, Harlap S. Long-term mortality after preeclampsia. Epidemiology. 2005;16:206–215. 29. Gordin D, Hiilesmaa V, Fagerudd J, Rönnback M, Forsblom C, Kaaja R, Teramo K, Groop PH; FinnDiane Study Group. Pre-eclampsia but not pregnancy-induced hypertension is a risk factor for diabetic nephropathy in type 1 diabetic women. Diabetologia. 2007;50:516–522. doi: 10.1007/ s00125-006-0544-5. 30. Haukkamaa L, Moilanen L, Kattainen A, Luoto R, Kahonen M, Leinonen M, Jula A, Kesäniemi YA, Kaaja R. Pre-eclampsia is a risk factor of carotid artery atherosclerosis. Cerebrovasc Dis. 2009;27:599–607. doi: 10.1159/000216834. 31. Kaaja R, Kinnunen T, Luoto R. Regional differences in the prevalence of pre-eclampsia in relation to the risk factors for coronary artery disease in women in Finland. Eur Heart J. 2005;26:44–50. doi: 10.1093/eurheartj/ ehi015. 32. Kessous R, Shoham-Vardi I, Pariente G, Sergienko R, Sheiner E. Longterm maternal atherosclerotic morbidity in women with pre-eclampsia. Heart. 2015;101:442–446. doi: 10.1136/heartjnl-2014-306571. 33. Melchiorre K, Sutherland GR, Liberati M, Thilaganathan B. Preeclampsia is associated with persistent postpartum cardiovascular impairment. Hypertension. 2011;58:709–715. doi: 10.1161/ HYPERTENSIONAHA.111.176537.

34. Mongraw-Chaffin ML, Cirillo PM, Cohn BA. Preeclampsia and cardiovascular disease death: prospective evidence from the child health and development studies cohort. Hypertension. 2010;56:166–171. doi: 10.1161/ HYPERTENSIONAHA.110.150078. 35. Savitz DA, Danilack VA, Elston B, Lipkind HS. Pregnancy-induced hypertension and diabetes and the risk of cardiovascular disease, stroke, and diabetes hospitalization in the year following delivery. Am J Epidemiol. 2014;180:41–44. doi: 10.1093/aje/kwu118. 36. Skjaerven R, Wilcox AJ, Klungsøyr K, Irgens LM, Vikse BE, Vatten LJ, Lie RT. Cardiovascular mortality after pre-eclampsia in one child mothers: prospective, population based cohort study. BMJ. 2012;345:e7677. 37. Stuart JJ, Rimm EB, Missmer SA, Spiegelman D, Hibert EN, Rexrode KM, Mukamal KJ, Rich-Edwards JW. Hypertensive disorders in pregnancy and risk of myocardial infarction and stroke. Am J Epidemiol. 2013;177:S41–S41. 38. Ghossein-Doha C, Wissink B and Spaanderman ME. Preeclampsia relates to heart failure within 7 years after pregnancy. Reprod Sci. 2014;21:122A–123A. 39. McDonald SD, Malinowski A, Zhou Q, Yusuf S, Devereaux PJ. Cardiovascular sequelae of preeclampsia/eclampsia: a systematic review and meta-analyses. Am Heart J. 2008;156:918–930. doi: 10.1016/j. ahj.2008.06.042. 40. Brown MC, Best KE, Pearce MS, Waugh J, Robson SC, Bell R. Cardiovascular disease risk in women with pre-eclampsia: systematic review and meta-analysis. Eur J Epidemiol. 2013;28:1–19. doi: 10.1007/ s10654-013-9762-6. 41. 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. doi: 10.1136/bmj.39335.385301.BE. 42. American College of Obstetricians and Gynecologists; Task Force on Hypertension in Pregnancy. Hypertension in pregnancy. Report of the American College of Obstetricians and Gynecologists’ Task Force on Hypertension in Pregnancy. Obstet Gynecol. 2013;122:1122–1131. 43. Wu P, Kwok CS, Haththotuwa R, Kotronias RA, Babu A, Fryer AA, Myint PK, Chew-Graham CA, Mamas MA. Pre-eclampsia is associated with a two-fold increase in diabetes: a systematic review and meta-analysis. Diabetologia. 2016;59:2518–2526. 44. Melchiorre K, Sharma R, Thilaganathan B. Cardiovascular implications in preeclampsia: an overview. Circulation. 2014;130:703–714. doi: 10.1161/ CIRCULATIONAHA.113.003664. 45. Bello N, Rendon IS, Arany Z. The relationship between pre-eclampsia and peripartum cardiomyopathy: a systematic review and meta-analysis. J Am Coll Cardiol. 2013;62:1715–1723. doi: 10.1016/j.jacc.2013.08.717. 46. Ridker PM, Buring JE, Rifai N, Cook NR. Development and validation of improved algorithms for the assessment of global cardiovascular risk in women: the Reynolds Risk Score. JAMA. 2007;297:611–619. doi: 10.1001/jama.297.6.611. 47. Cook NR, Paynter NP, Eaton CB, Manson JE, Martin LW, Robinson JG, Rossouw JE, Wassertheil-Smoller S, Ridker PM. Comparison of the Framingham and Reynolds Risk scores for global cardiovascular risk prediction in the multiethnic Women’s Health Initiative. Circulation. 2012;125:1748–1756. doi: 10.1161/CIRCULATIONAHA.111.075929. 48. Bhasin P, Kapoor S. Pregnancy complications and calculated cardiovascular risk in urban women: do we envisage an association? J Urban Health. 2014;91:162–175. doi: 10.1007/s11524-013-9818-7. 49. Fraser A, Nelson SM, Macdonald-Wallis C, Cherry L, Butler E, Sattar N, Lawlor DA. Associations of pregnancy complications with calculated cardiovascular disease risk and cardiovascular risk factors in middle age: the Avon Longitudinal Study of Parents and Children. Circulation. 2012;125:1367–1380. doi: 10.1161/CIRCULATIONAHA.111.044784. 50. Seely EW, Rich-Edwards J, Lui J, Nicklas JM, Saxena A, Tsigas E, Levkoff SE. Risk of future cardiovascular disease in women with prior preeclampsia: a focus group study. BMC Pregnancy Childbirth. 2013;13:240. doi: 10.1186/1471-2393-13-240.

Preeclampsia and Future Cardiovascular Health: A Systematic Review and Meta-Analysis Pensée Wu, Randula Haththotuwa, Chun Shing Kwok, Aswin Babu, Rafail A. Kotronias, Claire Rushton, Azfar Zaman, Anthony A. Fryer, Umesh Kadam, Carolyn A. Chew-Graham and Mamas A. Mamas Downloaded from http://circoutcomes.ahajournals.org/ by guest on May 1, 2018

Circ Cardiovasc Qual Outcomes. 2017;10: doi: 10.1161/CIRCOUTCOMES.116.003497 Circulation: Cardiovascular Quality and Outcomes is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231 Copyright © 2017 American Heart Association, Inc. All rights reserved. Print ISSN: 1941-7705. Online ISSN: 1941-7713

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://circoutcomes.ahajournals.org/content/10/2/e003497

Data Supplement (unedited) at: http://circoutcomes.ahajournals.org/content/suppl/2017/02/21/CIRCOUTCOMES.116.003497.DC1

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation: Cardiovascular Quality and Outcomes can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: http://www.lww.com/reprints Subscriptions: Information about subscribing to Circulation: Cardiovascular Quality and Outcomes is online at: http://circoutcomes.ahajournals.org//subscriptions/

SUPPLEMENTAL MATERIAL

Supplemental Methods Assessment of exposure Exposure was classified as pre-eclampsia if prospectively ascertained or retrospectively through medical records or database or questionnaire not limited to a particular definition of pre-eclampsia. Where data were given as sub-cohorts of severe/mild pre-eclampsia or term/preterm pre-eclampsia or multiparous/ primiparous pre-eclampsia, these were combined as a single exposure (pre-eclampsia) group.

Assessment of outcomes Coronary heart disease outcomes included studies that defined outcomes such as ischaemic heart disease, myocardial infarction, angina, coronary heart disease or coronary artery disease. Outcomes were assessed as cardiovascular disease if outcomes were defined as cardiovascular disease or composite of cardiac, cerebrovascular and peripheral vascular disease. Outcomes of heart failure were defined as cardiac insufficiency or heart failure. Stroke outcomes were defined as composite stroke, any acute cerebrovascular event, stroke unspecified or stroke/transient ischaemic attack. Deaths due to coronary heart disease or cardiovascular disease outcomes were also examined. Where studies reported a composite of fatal and non-fatal cardiac outcomes, the data was included in the non-fatal cardiac outcome to avoid over estimating mortality rates. Outcome of stroke death was defined as cerebrovascular disease death or mortality.

1

Search terms Synonyms of pre-eclampsia (‘preeclampsia’ or ‘pre-eclampsia’ or ‘EPH’ or ‘pregnancy toxemia’ or ‘edema-proteinuria-hypertension gestos’) AND ‘hypertension’ or ‘diabetes’ or ‘ischaemic heart disease’ or ‘ischemic heart disease’ or ‘coronary artery disease’ or ‘coronary heart disease’ or ‘myocardial infarction’ or ‘acute coronary syndrome’ or ‘heart failure’ or ‘cardiac failure’ or ‘left ventricular systolic dysfunction’ or ‘stroke’ or ‘cerebrovascular disease’ or ‘cerebrovascular accident’ or ‘pulmonary embolus’ or ‘venous thromboembolism’ or ‘deep vein thrombosis’ or ‘cardiomyopathy’ or ‘renal impairment’ or ‘kidney disease’ or ‘peripheral vascular disease’.

2

Supplemental Table 1. Study design and participant characteristics. Study ID

Study Design;

Total No. of Participants

Mean age at

Country; Year

(PE/no PE)

Index

Parity

Participant Selection Criteria

A

Women in the Tromso study which focuses on cardiovascular risk

Pregnancy Andersgaard

Cross-sectional study;

8,088 (PE 901; controls

20121

Norway; 1994-1995.

7,187).

Bhattacharya

Register-based cohort

25,963 (PE 2,026; controls

20122

study; Scotland;

23,937).

23.4

factors. 24.4

P

(AMND) to include women born before 1968 and living in

1967-2008. Funai 20053

Selected from the Aberdeen Maternity and Neonatal Databank

Aberdeen during their first pregnancy.

Prospective cohort;

37,913 (PE 1,055; controls

Jerusalem; 1964-

36,858).

26.2

A

Women in the Jerusalem Perinatal Study who delivered between 1964-1976.

1976. Ghossein-Doha

Cohort study;

20144

Netherlands; Unclear.

Gordin 20075

Prospective cohort;

137 (PE 95; controls 42).

Unclear

A

Women with history of PE and healthy parous controls.

148 (PE 43; controls 105).

29.7

A

Women with Type 1 DM who have been followed throughout

Finland; 1988-1996.

their pregnancy and delivery at Helsinki University Central Hospital.

1

Grandi 20156

Retrospective cohort

156,967.

29

Unclear

study; Canada;

Women with first recorded pregnancy between the ages of 15-45 years.

Unclear. Haukkamaa

Cross-sectional

524 (PE 35; controls 489).

20097

survey; Finland;

hospitals who participated in a Finnish cross-sectional health

2000-2001.

survey.

Hovsepian


2,066,230 (PE 163,974;

2014 8

study; USA; 2005-

controls 1,902,256).

57.2*

A

28.3

A

Women aged 45-74 years within 150 km of 5 Finnish university

Women without prior cardiovascular disease who were admitted to hospitals in California for delivery from 2005 to 2011.

2011. Kaaja 20059

Cross-sectional


survey; Finland;

3,162).

26.7

A

Women in FINRISK-cross sectional survey which monitors cardiovascular risk factors in Finland every 5 years.

2002. Kessous 201510


96,370 (PE 7,824; controls

study; Israel;

88,546).

28.6

A

Women who delivered at Soroka University Medical Center from 1988 to 2012.

1988- 2012. Lin 201111 &


1,132,064 (after exclusions

Tang 200912

study; Taiwan; 1999-

in Lin 2011) or 1,132,019

2003.

(after exclusions in Tang

Unclear

A

2009).

2

Women giving birth in Taiwan between 1999 to 2003.

Lykke 200913 &


677,761 (PE 33,826;

Lykke 201014

study; Denmark;

controls 643,935).

26.8

P

Women age 15-50 who had first delivery from 1978-2007 in the National Patient Registry in Demark.

1978-2007. Mannisto 201315 Prospective cohort

Women without CVS RF:

study; Finland; 1966-


2006.

4,283).

Melchiorre

Prospective cohort

142 (PE 64; controls 78).

201116

study; UK; 2008-

26.7

A

Women from the prospective Northern Finland Birth Cohort (NFBC) 1966 which composed of all expected births for 1966.

31.8

A

Women with singleton pregnancy and PE from 2008 to 2011 and their matched controls.

2011. Mongraw-

Prospective cohort


Chaffin 201017

study; USA; 1959-

13,922).

Median 26

A

Women without a prior diagnosis of cardiovascular disease who were members of the Kaiser Permanente Health Plan in the East

1967.

Bay of California pregnant between 1959-1967 and took part in the Child Health and Development Studies (CHDS) cohort.

Nijdam 200918


185 (PE 35; controls 150).

32.3

A

study; Netherlands;

Women in four primary care centers with PE from January 2000 to July 2007 and random controls.

2000-2007. Savitz 201419


849,639.

Unclear

A

study; USA; 1995-

All women who gave birth in hospitals in New York City from 1995–2004.

2004.

3

Skjaerven 201220

Prospective cohort study; Norway; 1967-

836,147 (PE 34,824;

Unclear

A

controls 801,323).

2002 and second births that occurred within seven years identified

2002. Stuart 201321

Prospective cohort

Norwegian women with a first singleton birth between 1967 and

through the Medical Birth Registry. 53,003.

Unclear

A

study; USA; 2001-

Women of singleton live births who provided pregnancy history in 2001 for the Nurses' Health Study II.

2009. Wikstrom 200522 Cross-sectional study; Sweden; 1973-1982.

395,614 (PE 12,533;

Median 48*

P

controls 383,081).

A - Any parity; DM- Diabetes; PE - pre-eclampsia; P – primiparous; RF – risk factors; * age at follow-up.

4

Women in the Swedish Medical Birth Register from 1973 to 1982.

Supplemental Table 2. Study quality assessment overview.

Study ID

Representative

Selection of

Ascertainment

Demonstration

Comparability

Assessment

Follow-up

Adequacy of

Total

of the exposed

the non-

of exposure

that outcome of

of cohort

of outcome

duration to

follow-up

score

cohort

exposed

interest was

cohort

not present at

*

5

capture outcomes

start of study Andersgaard

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

7

*

*

*

*

4

*

*

*

20121 Bhattacharya

6

20122 Funai 20053 Ghossein-Doha 20144 Gordin 20075

*

*

*

*

*

Haukkamaa 20097 *

*

*

Hovsepian 20148

*

*

Grandi 20156

*

*

5

*

*

5

5

**

*

*

7

*

*

*

7

Kaaja 20059

*

*

*

Kessous 201510

*

*

*

Lin 201111 &

*

*

*

*

**

*

*

*

8

*

*

*

*

*

8

*

*

**

*

*

*

9

*

*

*

*

*

*

*

8

*

*

*

*

*

Tang 200912 Lykke 200913 & Lykke 201014 Mannisto 201315

Melchiorre 201116 * Mongraw-

6

*

*

*

*

*

*

7

*

**

*

*

*

9

*

*

*

6

*

7

*

7

*

*

*

Nijdam 200918

*

*

*

Savitz 201419

*

*

*

Skjaerven 201220

*

*

*

Chaffin 201017

Stuart 201321 Wikstrom 200522

* *

*

*

*

*

*

*

*

*

*

*

6

* *

*

*

4 *

7

Supplemental Table 3. Study quality assessment in detail.

Study ID

Representative

Selection of the

Ascertainmen

Demonstration

Comparabilit

Assessment of

Follow-up

Adequacy of

of the exposed

non-exposed

t of exposure

that outcome of

y of cohort

outcome

duration to

follow-up

cohort

cohort

interest was not

capture outcomes

present at start of study Andersgaard 20121

General cohort

Controls from

Completion of

of women.

same cohort.

questionnaire.

No.

Unadjusted.

Completion of

Mean 24.7 years.

questionnaire.

434/9,974 (4%) loss to followup.

Bhattacharya 20122

General cohort

Controls from

Recorded on

Excluded baseline

Adjusted for

ICD-9 and 10

of women.

same cohort.

Aberdeen

HTN.

women’s year

codes from the

follow-up data

Maternity and

of birth,

Scottish

for 25.2% of

Neonatal

smoking,

morbidity records

women.

Databank.

socioeconomi

and death

c status at time

registrations.

of first pregnancy.

7

Mean 34.5 years.

Unable to link

Funai 20053

General cohort

Controls from

Labour ward

Not applicable as

Adjusted for

ICD-9 codes from

Median 30 years.

of women.

same cohort.

logs, checked

mortality

age, insulin-

the population

(1.6%)

by nurses

outcomes.

dependent

registry for

emigrated/

DM,

deaths.

changed

gestational

653/39,802

identity.

DM, birth weight, education, socioeconomi c status, and year of baseline birth. Ghossein-Doha

Unclear.

Unclear.

Unclear.

No.

20144

Adjusted for

By

BMI, fasting

echocardiography

insulin, HDLc

.

systolic BP and diastolic BP.

8

5 -7 years.

All women followed-up.

Gordin 20075

Women with

Controls from

Measurement

type 1 DM.

same cohort.

Excluded HTN.

Adjusted for

Standardised

Mean 10.6 years.

203/396 (51%)

by

age, duration

questionnaire

invited had

nurses/midwiv

of DM,

completed by a

participated in

es.

smoking,

physician who

study.

follow-up time

verified with

and BMI.

patients’ medical files.

Grandi 20156

General cohort

Controls from

Medical and

Excluded prior

Adjusted for

Unclear.

Unclear.

Unclear.

of women.

same cohort.

records.

history of HTN or

unspecified

CVD.

time-varying

Adjusted for

Home health

Unclear.

All women

confounders. Haukkamaa 20097

General cohort

Controls from

Self-reported,

No.

of women.

same cohort.

then verified

age, CRP,

interview,

with ICD

fasting

completed by a

codes from the

glucose, age at

doctor.

National

menarche,

Hospital

systolic BP,

Discharge

BMI, parity,

Register.

insulin,

9

followed-up.

smoking, HDLc. Hovsepian 20148

General cohort

Controls from

ICD-9 codes

Excluded women

Adjusted for

ICD-9 codes

6 weeks post-

of women.

same cohort.

from

with prior CVD.

age, ethnicity,

used.

delivery.

California

insurance

administrative

status, PE ,

claims

eclampsia,

database

peripartum haemorrhage, peripartum infection, pregnancyrelated haematologic disorders, HTN, DM, congestive heart failure, chronic kidney

10

Database study.

disease, CHD, peripheral vascular disease, atrial fibrillation, tobacco use and alcohol abuse . Kaaja 20059

General cohort

Controls from

Completed

of women.

same cohort.

No.

Adjusted for

Completed

questionnaire

age at first

questionnaire

CVS outcomes

with trained

birth, age,

with trained staff.

available for

staff.

parity, BMI,

3,559/3,650

increased

(97.5%).

blood cholesterol (ever), HTN (ever), DM or impaired glucose

11

17 years.

Data on PE and

tolerance, CAD (angina pectoris), and mother’s myocardial infarction/apo plexy. Kessous 201510

General cohort

Controls from

From the

Excluded known

Adjusted for

ICD-9 codes from

of women.

same cohort.

perinatal

CVD, renal

maternal age,

hospital database.

database,

disease, and

parity, DM

where

congenital heart

and obesity

information is

or renal

entered by

malformations.

Mean 11.2 years.

Database study.

At least 3 years.

Database study.

obstetrician after the delivery. Lin 201111 &

General cohort

Controls from

ICD-9 codes

Excluded women

Adjusted for

ICD-9 codes from

Tang 200912

of women.

same cohort.

from National

with history of

age, years of

National Health

Health

education,

12

Insurance

major

marital status,

Insurance

database.

cardiovascular

multiple

database.

events 90 days

gestations,

before delivery

infant sex,

(Lin 2011).

birthweight, parity, longterm HTN, pregnancyrelated HTN, anaemia, DM, antepartum haemorrhage, postpartum haemorrhage, and systemic lupus erythematosus .

13

Excluded women

Adjusted for

with stroke 90

age, years of

days before

education,

delivery (Tang

marital status,

2009).

multiple gestation, birth weight, parity, anaemia, DM, caesarean delivery, chronic HTN, pregnancyrelated HTN, antepartum haemorrhage, and postpartum haemorrhage.

14

1 year.

Lykke 200913 &

General cohort

Controls from

Data from

Excluded women

Adjusted for

From National

Median 14.6 years

15,902/791,163

Lykke 201014

of women.

same cohort.

national

with

age, year of

Patient registry.

(Lykke 2009) or

(2.0%)

database.

cardiovascular

delivery,

14.8 years (Lykke

emigrated and

diagnoses prior to

placental

2010).

8,876/791,163

delivery.

abruption and

(1.1%)

stillbirth.

died.

Mannisto 201315

General cohort

Controls from

Medical

of women.

same cohort.

No.

Excluded

ICD codes from

39.4 years.

1,554/12,055

records

multiparous

National patient

(13%) missing

reviewed by 2

women, age

registries.

blood pressure

obstetricians.

>35, BMI

measurements

>25, smokers

or died.

and DM. Melchiorre 201116

General cohort

Women without

Prospectively

Echocardiograms

Unadjusted,

Standardised

of women.

HTN with

recruited after

at baseline to

but matched

echocardiography

singleton

diagnosis

ascertain left

for age,

.

pregnancy

using ISSHP

ventricular

gestational age

matched for

criteria.

function.

and ethnicity.

gestational age,

15

1 year.

8/150 (5%) loss to follow-up.

age and ethnicity. Mongraw-Chaffin

General cohort

Controls from

Review of

Excluded women

Adjusted for

ICD codes.

Median 37 years.

14,403/16,002

201017

of women.

same cohort.

medical notes.

with pre-existing

pre-existing

(90%) women

cardiac disease.

HTN, DM,

analysed after

delivering an

exclusions and

IUGR infant,

loss to follow-

maternal

up.

completion of high school, socioeconomi c status, age, BMI, and parity. Nijdam 200918

General cohort

Controls from

International

of women.

same cohort.

No.

Unadjusted.

International

Mean 2.9 years in

No loss to

classification

classification of

PE group and 2.5

follow-up.

of primary

primary care

years in control.

care codes

codes from

from

electronic medical

16

electronic

records in

medical

primary care.

records in primary care. Savitz 2014 19

General cohort

Controls from

Use of

Excluded women

Adjusted for

Identified using

of women.

same cohort.

hospital

with

year, age,

ICD-9 codes.

discharge

cardiovascular

ethnicity,

information.

disease prior to

health

delivery.

insurance, gestational DM, parity, socioeconomi c status, smoking, prenatal care, and prepregnancy weight.

17

Within 1 year.

Database study.

Skjaerven 201220

General cohort

Controls from

ICD codes

Not applicable as

Adjusted for

ICD codes from

Median 25 years.

< 1% loss to

of women.

same cohort.

from Medical

mortality

maternal

National Cause of

follow-up due to

Birth registry.

outcomes.

education,

Death Registry.

emigration.

maternal age at first birth, and year of first birth

Stuart 201321

Female nurses

Controls from

Self-reported

Excluded women

Adjusted for

Self-reported on

from USA.

same cohort.

hypertensive

with history of MI

age, race,

questionnaire.

disorders in

or stroke at

parental

pregnancy.

recruitment.

history of MI

8 years.

Unclear.

15 years.

3.15% died or

aged 90 mmHg for 2

Mean 34.5

PE/eclampsia vs controls.

20122

readings 4 hours apart or

years.

Cerebrovascular disease death: 29/2,026 vs 266/23,937. aRR 1.27 (0.87-1.87).

>110 mmHg single reading,

IHD death: 52/2,026 vs 467/23,937. aRR 1.38 (1.03-1.84).

and proteinuria 0.3g/24 hours.

Cerebrovascular disease admission: 94/2,026 vs 1,004/23,937. aRR 1.16 (0.93-1.45). IHD admission: 172/2,026 vs 1,882/23,937. aRR 1.18 (0.99-1.41).

Funai 20053

Ghossein-

ISSHP (2014) definition

Unclear.

Median 30

PE vs controls.

years.

CVD death: 41/1,055 vs 269/36,858. RR 3.07 (2.18-4.34).

5-7 years.

PE vs controls.

Doha 20144 Gordin 20075

Stage B asymptomatic heart failure: 28/95 (29%) vs 1/42 (3%). aOR 9.9 (1.0-93.2). ISSHP (2014) definition.

Mean 10.6

PE vs controls.

years.

CHD: 5/43 (12.2%) vs 2/105 (2.2%). MI: 3/43 (7.3%) vs 0/105 (0%).

20

Grandi 20156

PE not defined.

Unclear.

PE/eclampsia vs controls (total n=156,967). CVD: aHR 1.1 (0.5-2.2).

Haukkamaa

ISSHP (2014) definition.

Unclear.

20097 Hovsepian

PE vs controls. CAD: 2/35 (6%) vs 29/489 (6%).

ICD-9 codes.

20148

6 -12 weeks

PE vs controls (total n=2,066,230).

post-

Acute cerebrovascular disease: 57/163,974 vs 249/1,902,256. aOR 2.1 (1.6-2.8).

delivery.

Ischaemic stroke (n=75): aOR 3.7 (2.2-6.1). Haemorrhagic stroke (n=117): aOR 1.9 (1.2-3.0).

Kaaja 20059


Mean 17.4

PE vs controls.

years.

Cardiac insufficiency: 12/397 (2.9%) vs 22/3,162 (0.7%). CAD: 10/397 (2.5%) vs 25/3,162 (0.8%).

Kessous 201510 ICD-9 codes.

Mean 11.2

PE vs controls.

years.

Cardiovascular hospitalisation: 360/7,824 (4.6%) vs 2,391/88,546 (2.7%). aHR 1.7 (1.6-1.9).

Lin 201111

PE/eclampsia defined by

At least 3

PE/eclampsia vs controls (total n=1,132,064).

& Tang 200912

ICD-9 codes.

years (Lin

MI: HR 22.6 (8.7-58.4).

2011).

Heart failure: aHR 8.3 (4.2-16.4). Stroke: aHR 14.5 (1.3-165.1). Major CVD: aHR 12.6 (2.4-66.3). Major CVD related death: aHR 2.3 (1.6–3.1).

21

Major CVD excluding stroke: aHR 7.3 (5.5–9.7).

1 year (Tang

Ischaemic stroke: 16/8,814 vs 91/1,122,637. aRR 4.35 (0.58-32.92).

2009).

Haemorrhagic stroke: 17/8,815 vs 122/1,122,668. aRR 19.90 (7.75-51.11).

Lykke 200913 & ISSHP (2014) definition.

Median 14.6

PE vs controls.

Lykke 201014

years (Lykke

IHD: 651/33,826 vs 7,727/741,012.

2009).

Stroke: 600/33,826 vs 8,240/741,012. Congestive heart failure: 187/33,826 vs 2,050/741,012.

Median 14.8

Cardiovascular mortality: 116/25,184 vs 824/643,935. aHR 2.08 (1.63, 2.64).

years (Lykke 2010). Mannisto 201315 ≥145/95 mmHg with proteinuria ≥ 0.3g/l after 20 weeks’ gestation.

Mean 39.4

PE/eclampsia vs controls excluding those with CVS RF:

years.

CVD: 39/162 vs 893/4,283. aHR 1.39 (1.01–1.92). IHD: 21/162 vs 535/4,283. aHR 1.27 (0.82-1.95). MI: 7/162 vs 165/4,283. aHR 1.46 (0.69-3.09). MI death: 1/162 vs 17/4,283. aHR 2.06 (0.29-14.9). Heart failure: 7/162 vs 148/4,283. aHR 1.60 (0.73-3.49). Ischaemic cerebrovascular disease: 7/162 vs 144/4,283. aHR 1.40 (0.64-3.09).

22

Melchiorre


1 year.

201116 Mongraw-

Stage B heart failure: 28/64 vs 8/78. ISSHP (2014) definition.

Median 37

PE vs controls.

years.

Cardiac disease death: 24/481 vs 242/13,922. aHR 2.14 (1.29-3.57).

Records were screened for

Mean 2.9

PE vs controls.

history of PE or HTN and

years in PE

Cardio/cerebrovascular disease: 1/35 vs 3/150.

substantial proteinuria.

group and

Chaffin 201017 Nijdam 200918

PE vs controls.

2.5 years in controls. Savitz 201419

PE defined by ICD9 codes.

Within 1

PE vs control (total n=849,639).

year of

Heart failure (n=259): aOR 4.1 (2.9-5.8).

delivery.

Intracranial haemorrhage (n=68): aOR 2.8 (1.3-6.2). Stroke/TIA (n=126): aOR 2.8 (1.6-5.0). CHD (n=81): aOR 3.1 (1.6-6.3).

Skjaerven 201220


Median 25

PE vs controls.

years.

Cardiovascular death: 176/34,824 vs 2,380/801,323. aHR 1.9 (1.6-2.2).

IHD mortality: Controls: multiparous no PE: 675/599,973.

23

Primiparous Term PE: 34/4,758. aHR 4.7 (3.3-6.7). Primiparous Preterm PE: 15/1,426. aHR 9.3 (5.5-15.6). Multiparous Term PE: 37/21,950. aHR 1.7 (1.2-2.4). Multiparous Preterm PE: 12/4,460. aHR 3.7(2.1-6.6).

Cerebrovascular disease mortality: Controls: multiparous no PE: 675/599,973. Primiparous Term PE: 14/4,758. aHR 2.1 (1.2-3.6). Primiparous Preterm PE: 16/1,426. aHR 10.4 (6.3-17.2). Multiparous Term PE: 32/21,950. aHR 1.4 (0.95 - 1.91). Multiparous Preterm PE: 4/4,460. aHR 1.12 (0.42 to 3.0). Stuart 201321

Self-reported PE.

8 years.

PE vs controls (total n=53,003). MI: aHR 1.6 (1.2-2.2). Stroke: aHR 1.8 (1.3-2.4).

Wikstrom 200522


15 years.

PE vs controls. IHD hospitalisation or death: 176/12,533 vs 2,306/383,081.

BP – Blood Pressure; CAD – Coronary Artery Disease; CHD – Coronary Heart Disease; CVD – Cardiovascular Disease; HTN – Hypertension; IHD – Ischaemic Heart Disease; ISSHP - International Society for the study of Hypertension in Pregnancy; MI – Myocardial Infarction; PE – Pre eclampsia; TIA – Transient Ischaemic Attack.

24

Supplemental Table 5. Metabolic risk factor profile of PE and control groups in the included studies. * Total PE vs. control.

Study ID

Risk factor profile

During pregnancy

At follow-up

PE

Control

p value

PE

Control

p value

Andersgaard

Age (year)

-

-

-

48.8

47.4