Hypertensive disorders of pregnancy and subsequent maternal cardiovascular health

Women with a history of hypertensive pregnancy disorders exhibit a more unfavourable cardiovascular health after pregnancy compared to women with a previous normotensive pregnancy. Already 6 years after index pregnancy the prevalence of chronic hypertension shows an increased risk in both women with a history of GH or PE. The results further demonstrate that especially early pregnancy BP is strongly associated with the diagnosis chronic hypertension 6 years after pregnancy. Even though these results are strongly influenced by BMI at follow-up, effect estimates remained significant.

Strengths are the prospective data collection and large sample size. Complete information on pregnancies and pregnancy disorders that occurred in the years after the index pregnancy was not available. Instead we used gravidity at follow-up to account for pregnancies in between the index pregnancy and the follow-up measurement after 6 years. BP at study intake during pregnancy was higher in women with a history of GH than in women remaining normotensive throughout pregnancy. Pregnancy is associated with a physiologic decrease of BP which might suggest that non-random misclassification of the diagnosis chronic hypertension occurred in our study. Nevertheless, we think this is unlikely because information on chronic hypertension was cross-checked between multiple sources (maternal questionnaire in pregnancy and information from the original medical records and the Dutch obstetric database). Finally, we are aware of the fact that our approach to define chronic hypertension at follow-up might be suboptimal as 1) we defined chronic hypertension based upon BP measurements at one point in time at the 6 years follow-up visit (i.e. transient hypertension) and 2) international guidelines recommend that ambulatory BP monitoring should be used to define or confirm clinical diagnosis of hypertension because of prevalence of white coat hypertension. Transient hypertension is an established risk factor (i.e. early form) of chronic hypertension [16, 17]. Unfortunately, ambulatory BP monitoring was not available. Besides, BP was measured in a research setting by trained research assistants wearing normal clothing (i.e. no white coats). We also performed a sensitivity analyses in which we labelled women as chronic hypertensive, only if they used BP medication at follow-up. These analyses showed similar results (Supplementary Material S6).

Studies show that hypertensive pregnancy disorders are associated with a woman’s risk of CVD [2]. This provides opportunities to identify women at risk early in their lives when it may be possible to alter their risk trajectory. It has been shown that measures of arterial stiffness and left ventricular function are increased during pregnancy among women with hypertensive pregnancy disorders [18, 19]. Arterial stiffness and left ventricular function are thought to be independent predictors of chronic hypertension and CVD. Also aortic root dilatation may be secondary to hypertension [18, 20, 21]. Franz et al. [22] showed differences in PWV after index pregnancy among former early, but not late, onset PE women. Likewise, Ghossein-Doha et al. [23] showed increased left ventricular mass indices and decreased cardiac diastolic function among a population of mainly severe (early onset) PE postpartum as proxy for hypertension. We observed increased PWV 6 years after index pregnancy among women with a history of GH. Similarly, women with a history of GH had an increased LV mass, a larger AOD, a higher BP and a higher risk of chronic hypertension. Interestingly, no differences were seen when comparing these cardiovascular measurements between women with PE and normotensive women, with the exception of an increased risk of chronic hypertension. Regrettably, GH has only rarely been included in postpartum research and despite evidence that differences in CVD risk exist between women with severe early-onset PE, mild late-onset PE and GH, most studies do not differentiate between the subgroups of hypertensive pregnancy disorders. One large longitudinal study which differentiated between (mild) PE and GH found that both GH and PE were associated with greater CVD risk factors. However after controlling for various confounders, results for PE were not significant anymore [24]. In a study by Wikström et al. [25] the risk of developing ischaemic heart disease was higher in women with severe PE compared with GH and mild PE [25, 26]. This was similar to results reported by Lykke et al. [27]. Women with mild PE had a four-fold higher risk of chronic hypertension. However, this risk increased up to sixfold in women with a history of GH and severe PE, respectively [25‐27]. Similar patterns in mean BP, usage of anti-hypertensive medication and chronic hypertension were reported by Verbeek et al. [26] with the highest incidence of chronic hypertension in severe PE followed by GH and finally mild PE. In our study the majority of preeclamptic women had mild PE (92%) with only eight women suffering from severe PE. These results strengthen our findings on GH showing that not only severe PE women exhibit an increased risk of CVD but also that women with GH are at increased risk of an adverse cardiovascular health profile after pregnancy. However, the results could also indicate that, for women with GH, this phenotype already existed prior to the index pregnancy. Women with GH should in either case not be excluded from (secondary) preventive interventions. A meta-analysis found that lifestyle interventions may alter cardiovascular risk after a history of PE up to 13% with relatively simple intervention measures [28]. These may include exercise, dietary counseling and support for smoking cessation assistance.

Causal pathways relating hypertensive pregnancy disorders to chronic hypertension and CVD are unclear. One hypothesis focuses on common risk factors including among others obesity, chronic hypertension and genetic constitution [29]. Both PE and atherosclerosis arise from vascular inflammation with endothelial dysfunction. It has been also been hypothesized that hypertensive pregnancy disorders worsen pre-existing subclinical CVD risk factors already present before index pregnancy or even induce de novo risk [29]. A large population-based study showed that most CVD risk factors remain higher after PE following adjustment for pre-pregnancy values [30]. It is possible that products of the dysfunctional placenta in PE could permanently compromise maternal cardiovasculature [28]. A study demonstrated that increased sensitivity to infused Angiotensin II exists in the postpartum state in women with a history of new-onset hypertension in pregnancy and that this increased sensitivity to Angiotensin II is present in the vasculature and in the adrenal glands, with a suggestion of sFlt-1 responsiveness.[31] This study also showed that women with a history of new-onset hypertension in pregnancy were unable to modulate a response to infused Angiotensin II on the basis of salt intake. They suggested a dysregulation of the renin-angiotensin system.[31] Another study suggested persistence of left ventricular geometrical changes that herald the development of chronic hypertension [23]. In our study we observed significant associations regarding early pregnancy BP and all cardiac and vascular outcomes and the risk of chronic hypertension. Also the combined associations of maternal BP during early and late pregnancy were consistently associated with vascular, cardiac and hypertensive outcomes at follow-up. These findings may corroborate those of prior research supporting the theory that hypertensive pregnancy disorders share pathophysiology already programmed before the challenge of pregnancy that ultimately leads to CVD. However, they also cannot reject the hypothesis that hypertensive pregnancy disorders may cause permanent vascular damage thereby contributing to CVD risk.