Proton Pump Inhibitors Use and Risk of Preeclampsia: A Meta-Analysis

Evidence from preclinical studies suggests a preventive effect of proton pump inhibitors (PPIs) in preeclampsia. Recently, several epidemiological studies have described a conflicting association between the use of PPIs during pregnancy and preeclampsia risk. This study aimed to evaluate the association between PPI use and the risk of preeclampsia. We searched databases, including MEDLINE, Embase, Scopus, Web of Science Core Collection, Emcare, CINAHL, and the relevant grey literature from inception until 13 September 2021. Studies reporting the preeclampsia risk with the use of PPIs were eligible for inclusion. Literature screening, data extraction, and the risk of bias assessment were performed independently by two investigators. Random-effect meta-analysis was performed to generate relative risks (RR) and 95% confidence intervals (CI). The risk of preeclampsia and preterm preeclampsia among women receiving PPIs during pregnancy were the primary outcomes of interest. This meta-analysis comprised three studies involving 4,877,565 pregnant women, of whom 119,017 were PPI users. The included studies were judged to have a low risk of bias. The risk of preeclampsia among pregnant women who received PPIs anytime during pregnancy was significantly increased (RR 1.27 (95% CI: 1.23–1.31)), although the increase was trivial in absolute terms (2 per 1000). The subgroup analysis revealed that the risk was increased in each of the three trimesters. The risk of preterm preeclampsia among pregnant women receiving PPIs anytime during pregnancy was not significantly increased (RR 1.04 (95% CI: 0.70–1.55)). The certainty evaluated by GRADE in these estimates was low. PPI use may be associated with a trivial increase in the risk of preeclampsia in pregnant women. There is no evidence supporting that PPI use decreases the risk of preeclampsia or preterm preeclampsia.


Introduction
Preeclampsia is one of the most severe complications of pregnancy characterized by high blood pressure. It is one of the leading causes of maternal morbidity and mortality worldwide. The global burden of preeclampsia is continuously rising; epidemiological trends showed a 10.9% increase in the incidence of preeclampsia from 1990 to 2019 [1]. Preeclampsia leads to adverse maternal and perinatal outcomes, including preterm birth, prolonged hospital stays, low birthweight babies, and a higher risk of neonatal intensive care unit admission [2]. Preterm birth imposes a significant mortality risk on the mother and the baby [3]. Several database studies have reported a positive association between preterm birth and mortality and morbidity in both the mother and the baby [4,5]. There are no definitive treatment options available for preeclampsia management, except for the timely delivery of the fetus and placenta.
Evidence from preclinical studies suggests a plausible preventive effect of proton pump inhibitors (PPIs) in preeclampsia [6][7][8]. The potential mechanism of this protective effect of PPIs in managing preeclampsia could be due to the reduction in the mRNA expression and secretion of antiangiogenic factors (sFlt1) and soluble endoglin (sEng) in placental endothelial cells, as these are the key component involved in the pathophysiology of preeclampsia [7,9,10]. This mechanistic association was supported by the findings of a recently published prospective cohort study where lower levels of sFLT-1 and sEng were noticed among pregnant PPI users with suspected preeclampsia [11]. PPIs are commonly used to treat gastroesophageal reflux disorder (GERD); however, in the last decade, the safety of PPIs has been a matter of scrutiny [12]. Our previous systematic reviews and meta-analyses found PPI use to be associated with several other non-pregnancy-related adverse health outcomes [13][14][15][16]. Nevertheless, PPIs are widely used by pregnant women due to their acceptable safety profiles and their availability as over-the-counter (OTC) drugs in many countries [17]. The evidence from cohort studies and a meta-analysis supported the PPI safety profiles among women who used PPIs during pregnancy and found no increased risk of congenital defects or preterm delivery [18,19].
Recently, several epidemiological studies examined the association of PPIs with preeclampsia risk [20][21][22]. A large cohort study from the US using the Truven Health MarketScan database found no association of PPIs with a decreased risk of preeclampsia or severe preterm preeclampsia [22]. Similar findings were reported by Choi et al. using the Korean Healthcare database [21]. However, a Swedish population register-based cohort study found reduced preterm and early preeclampsia risk in women who used PPIs in the third trimester [20]. These published studies presented conflicting evidence, and to date, no meta-analysis has been performed to explore this association, as confirmed through a preliminary search in multiple databases. Therefore, this systematic review and metaanalysis aimed to synthesize the evidence and assess the overall risk of preeclampsia in women using PPIs during pregnancy.

Protocol
The protocol for this systematic review was prospectively published as a preprint at medRxiv [23]. The principles laid down in the Cochrane Handbook of Systematic Review of Interventions and the JBI reviewers manual were utilized [24,25]. Preferred reporting items for systematic review and meta-analysis (PRISMA 2020) and meta-analysis of observational studies in epidemiology (MOOSE) reporting guidelines were followed [26,27]. Refer to Supplementary Table S1 for the detailed checklist.

Search Strategy
The literature search was conducted in each database from the inception date to 13 September 2021 to identify published and unpublished studies assessing preeclampsia risk in women receiving PPIs during pregnancy. The three-step search strategy was used; an initial limited search was conducted in MEDLINE (Ovid) and Embase (Ovid), using keywords and index terms related to PPIs and preeclampsia without restriction to any date or language; then, a detailed search was performed across all major databases by analyzing the text words and index terms used to describe the articles. We searched MEDLINE (Ovid), Embase (Ovid), Scopus, Web of Science Core Collection, Emcare (Ovand id), and CINAHL (EBSCO). The sources of grey literature were ProQuest Dissertations & Theses Global and clinical trials registers, ClinicalTrials.gov (accessed on 13 September 2021), and the WHO International Clinical Trials Registry Platform (ICTRP). Search strings were developed by a medical information specialist (AL). Lastly, bibliographies of the relevant articles were scanned manually for additional articles. The search strategies used for different databases were provided in Supplementary Table S2 with their respective hits.

Study Selection/Inclusion Criteria
Two reviewers independently reviewed the retrieved articles based on title and abstract screening, which was followed by second-level screening based on full-text articles. Covidence systematic review software was used for completing the article screening process [28].
Studies were eligible for inclusion if they met the inclusion criteria mentioned in Table 1:

Data Extraction and Risk of Bias
Two reviewers independently extracted all the relevant data based on the study characteristics (author, publication year, data source, and study period); details on patient characteristics; exposure; comparator; ascertainment of PPI use; confirmation of outcome; effect estimates (unadjusted and adjusted risk); and conclusions. Any discrepancy during the data extraction process was resolved by discussion. There was no missing data, so none of the primary authors was contacted for any additional data.
The risk of bias in the included studies was evaluated using the Newcastle-Ottawa Scale (NOS) independently by two reviewers [29]. According to the NOS, a study can achieve a maximum of 4 points in the selection, 2 points in the comparability, and 3 points in the exposure (case-control studies) or outcome (cohort studies) domain of the scale. Studies were classified as having a high, moderate, or low risk of bias, depending on the adjustment for appropriate confounders and the adequacy of the exposure and outcome ascertainment and not based on a numerical score [30].

Certainty of Evidence
We used the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology to assess the certainty of evidence [31]. The certainty assessment was judged as either high, moderate, low, or very low, based on the risk of bias, inconsistency, indirectness, imprecision, and publication bias.

Statistical Analysis
The primary outcome of interest was to assess the pooled relative risk of preeclampsia among women receiving PPIs during pregnancy. The risk ratio (RR) and odds ratio were used interchangeably, as PPI use and preeclampsia events were very rare [32]. Heterogeneity was determined based on Cochrane chi-square and I2 statistics [33]. The Cochrane chi-square value (p < 0.10) and I2 statistics ≥ 50% represent important heterogeneity [33]. Since we anticipated heterogeneity in terms of the population characteristics and settings of the studies, we used the random effect model. A subgroup analysis was performed based on the trimester of pregnancy and preterm and term preeclampsia. The leave-one-out method was used to understand the impact of each study on the pooled effect size. Metaregression and statistical evaluation of the publication bias using funnel plot approaches were not feasible due to the limited number of studies. Review Manager (RevMan) version 5.4.1 was used to perform the meta-analysis using the generic inverse variance method. Summary of the findings table was created using the GRADEpro GDT tool [34].

Studies Characteristics
The database search yielded 600 articles; three studies [20][21][22] involving 4,877,565 pregnant women, of whom 119,017 were PPI users, were included in the meta-analysis. A list of articles excluded with exclusion reasons during the full-text screening phase are presented in Supplementary Table S3. Refer to the PRISMA chart ( Figure 1) for the study inclusion process.
The design of the three studies was a retrospective cohort design, and they were published within the time frame of 2019-2021. All the eligible studies were published as a full text, except the study by Choi et al., published as a research letter. Included studies were conducted in the US, Sweden, and Korea. PPI exposure was defined as exposure to PPIs at any time during pregnancy or individually during the first, second, and third trimesters. Studies ascertained the PPI exposure through prescription records or claims data, and the outcome of preeclampsia was confirmed based on the International Classification of Diseases (ICD) code-9th or 10th edition. Refer to Table 2 for a detailed description of the included study characteristics.

Quality Assessment and Certainty of Evidence
The risk of bias in the three included studies was low. All the included studies selected the patients and control from the same databases and adjusted for several possible confounding factors. The details of this assessment are presented in Table 3. The certainty of the evidence on the risk of preeclampsia and preterm preeclampsia among women receiving PPIs during pregnancy was low to very low, as per the GRADE rating system (Table 4a,b). Of note, the absolute effects were trivial or very small.

Meta-Analysis (Preeclampsia Risk)
The risk of preeclampsia among pregnant women who received PPIs anytime during pregnancy was statistically significantly increased ( Figure 2a) with a pooled RR of 1.27 (95% CI: 1.23-1.31), p < 0.00001 in an adjusted analysis (adjusted for several possible confounding factors). The subgroup analysis revealed a statistically significantly higher risk of preeclampsia in pregnant women receiving PPIs in each of the three trimesters ( Figure 2b). All the analyses were adjusted for maternal age, chronic kidney disease, autoimmune disease, multiple gestation, pregestational diabetes, chronic hypertension, nulliparity, and multiple pregnancies. (a) Preeclampsia risk in women exposed to PPIs anytime during pregnancy. (b) Preeclampsia risk in women exposed to PPIs in different trimesters of pregnancy.

Meta-Analysis (Preterm Preeclampsia Risk)
Two studies reported data for the preterm preeclampsia risk among pregnant women receiving PPIs anytime during pregnancy, and the pooled estimate ( Figure 3a) revealed a nonsignificant association (p = 0.83).
The subgroup analysis based on the use of PPIs in various trimesters and the risk of preterm preeclampsia (Figure 3b) revealed a significantly higher risk in the second trimester, with a pooled relative risk of 1.32 (95% CI: 1.19-1.46), p < 0.00001. However, the association was nonsignificant in the first or third trimesters. (a) Preterm preeclampsia risk in women exposed to PPIs anytime during pregnancy. (b) Preterm preeclampsia risk in women exposed to PPIs in different trimesters of pregnancy.

Sensitivity Analysis
A sensitivity analysis was performed by excluding each study one by one (leave-oneout) from the pooled analysis. None of the studies had any significant influence on the pooled effect size. The RR for preeclampsia was identical to the main results.   Table 3. Quality assessment of the included studies.

Study author
Representation of the exposed cohort Selection of the non-exposed cohort

Discussion
This is the first systematic review and meta-analysis to investigate the risk of preeclampsia and preterm preeclampsia among women receiving PPIs during pregnancy. We found a significantly higher risk of preeclampsia in pregnant women exposed to PPIs anytime during pregnancy or during each specific trimester, although this risk was trivial or very small in absolute terms. PPI use was also associated with a significant increase in the risk of preterm preeclampsia only in the second trimester.
The certainty in the estimates was low, which suggests the need for randomized trials that evaluate patient important outcomes, as very few trials exist or are ongoing. One trial has demonstrated no prolongation in the gestation period with esomeprazole in women with preterm preeclampsia [35]. In addition, that trial showed no significant change in the relevant biomarkers levels (sFlt1, sEng, and placental growth factor) and maternal, fetal, or neonatal outcomes [35]. Similarly, a recently published randomized, double-blinded placebo-controlled trial by Abbas et al. [36] also confirmed no significant change in the antiangiogenic markers in women with early onset preeclampsia who received 40 mg of esomeprazole daily. This trial also found a nonsignificant effect of esomeprazole on the length of pregnancy and maternal and fetal complications [36]. Several trials are currently underway to assess the efficacy of PPIs, either alone or in combination with agents such as metformin or sildenafil for the management of preeclampsia. Table 5 provides a detailed list of the relevant ongoing trials.
The strength of the current systematic review was the exhaustive search for all the eligible published and unpublished studies from multiple literature sources, including gray literature, enabling a large sample size to estimate the risk of both preeclampsia and preterm preeclampsia based on the use of PPIs in different gestation ages.
However, a few important limitations need to be considered. First, the included studies did not specifically ascertain medication adherence, and the assessment of PPI exposure was based on the claims data and prescription records. Further, the availability of PPIs as OTC drugs in two of the included studies may lead to an inaccurate assessment of their exposure and, thus, raise the possibility of confounding [37]. Second, only one study adjusted the findings for the body mass index (BMI), although the BMI is known to be associated with preeclampsia [38]. Lastly, a small number of studies and the observational nature of the included studies and confounding by indication should be considered while interpreting the findings.

Conclusions
PPIs use may be associated with a trivial increase in the risk of preeclampsia in pregnant women. There is no evidence supporting that PPI use decreases the risk of preeclampsia or preterm preeclampsia. We recommend that future epidemiological studies consider all possible confounding factors, including the BMI. Furthermore, future population-based studies should ascertain the risk of preeclampsia and preterm preeclampsia separately by including a sufficiently large number of (preeclampsia and preterm preeclampsia) cases. The ongoing clinical trials of PPIs are expected to shed light on this important clinical question.  Institutional Review Board Statement: Ethical review and approval were waived for this study due to its observational nature and its use of publicly accessible data.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest.