Proton pump inhibitors in pregnancy and risk of preeclampsia: A nationwide cohort study of 1.9 million pregnancies

초록

Background: Preclinical studies have suggested that proton pumpinhibitors (PPIs) may have preventive effects on preeclampsia. How-ever, standing clinical evidence on the association between PPI useduring pregnancy and the risk of preeclampsia is limited.Objectives: To evaluate the association between PPI use in pregnancyand the risk of preeclampsia.Methods: We conducted a nationwide cohort study using Korea'shealthcare database (2011-2017) and included all pregnancies inwomen aged 19-44 years that resulted in live births between 2013and 2017. The date of conception was estimated by subtracting245 and 273 days from the date of delivery for preterm and all otherbirths, respectively. Exposure to PPIs was defined as ≥1 PPI prescrip-tion in 4 windows: any time during pregnancy, first, second, and thirdtrimester. We considered 2 comparator groups: 1) women unexposedto PPIs from 90 days before conception to delivery (non-user), and 2)women exposed to an active comparator, histamine 2 receptor antag-onists (H2RA user). Study outcomes were all preeclampsia and pre-eclampsia accompanied with preterm birth, which were identifiedfrom 20 weeks of gestation. As heartburn is one of the symptoms forpreeclampsia, PPI or H2RA exposure after the incident preeclampsiadiagnose were excluded from our analysis. Propensity score (PS) finestratification, based on 30 variables, was used to adjust for potentialconfounding, and relative risk (RR) with 95% confidence intervals(CI) was estimated using weighted generalized linear models.Results: Of 1,963,690 pregnancies, 30,018 (1.50%), 8,785 (0.45%)and 7,479 (0.38%) were exposed to PPI in the first, second, and thirdtrimester, respectively. The prevalence of all preeclampsia was 14.6per 1,000 pregnancies among those exposed to PPIs during preg-nancy, compared to 9.4 per 1,000 among unexposed pregnancies. InPS-adjusted analyses, the RR for women exposed to PPI at any time during pregnancy was 1.26 (95% CI 1.16-1.36) for all preeclampsiaand 1.29 (0.97-1.71) for preterm preeclampsia, compared to non-users. However, as compared with H2RA users, the RR attenuated to1.07 (0.96-1.20) and 1.03 (0.68-1.55) for all and preterm preeclamp-sia, respectively. A trend of reduced risk of preterm preeclampsia wasobserved when PPI use neared outcome onset (first: 1.14 [0.71-1.85],second: 0.83 [0.30-2.26], third: 0.39 [0.10-1.58]), although the esti-mates had wide CIs.Conclusions: In this large nationwide study, we found that PPI useduring pregnancy was not associated with reduced risk of preeclamp-sia, despite there being potential evidence of a lower risk of pretermpreeclampsia when exposed to PPIs late in pregnancy.