The impact of antihypertensive treatment of mild to moderate hypertension during pregnancy on maternal and neonatal outcomes: An updated meta‐analysis of randomized controlled trials

Abstract Currently, there is controversy regarding the treatment of pregnant patients with mild hypertension (blood pressure 140–159/90–109 mm Hg). While guidelines do not recommend this treatment, results from recent clinical trials are supportive of the treatment. This meta‐analysis aimed to clarify if active treatment of mild hypertension during pregnancy results in better maternal and fetal outcomes. All of the potentially eligible randomized controlled trials were retrieved through a systematic database search investigating the impact of pharmacological treatment in mild hypertensive patients on maternal, fetal, and neonatal outcomes. Relative risk (RR) and 95% confidence interval (CI) were calculated using a random‐effects model. Data from 12 trials comprising 4461 pregnant women diagnosed with mild to moderate hypertension (2395 in the intervention group and 2066 in the control group) were extracted for quantitative synthesis. Antihypertensive treatment was associated with better outcomes in seven out of the 19 analyzed outcomes: Severe hypertension (RR = 0.53; 95% CI = [0.38;0.75]), preeclampsia (RR = 0.71; 95% CI = [0.54; 0.93]), placental abruption (RR = 0.48; 95% CI = [0.26; 0.87]), changes in electrocardiogram (RR = 0.43; 95% CI = [0.25; 0.72]), renal impairment (RR = 0.42; 95% CI = [0.34; 0.51]), pulmonary edema (RR = 0.46; 95% CI = [0.25; 0.84]), and neonatal mortality (RR = 0.72; 95% CI = [0.57; 0.92]). The primary safety outcome of small for gestational age was not different between the treatment group and the control group (RR = 1.12; 95% CI = [0.80; 1.57]). The results of this meta‐analysis are in favor of the beneficial impact of pharmacological treatment of mild hypertension on both maternal and neonatal outcomes and without significant adverse events for the fetus.


| INTRODUCTION
Hypertensive disorders of pregnancy have been associated with a significantly higher risk of adverse pregnancy outcomes including fetal and neonatal death and small for gestational age (SGA). [1][2][3] Although guidelines have reached a consensus on treating cases with severe hypertension (blood pressure ≥ 160/110 mm Hg), 4 uncertainty still remains regarding the decision to treat patients with mild to moderate hypertension. Previously, the results of the CHIPS trial 5 showed that tight control of gestational and chronic hypertension during pregnancy (target diastolic blood pressure [DBP] < 85 mm Hg) could not lower the risk of adverse maternal and perinatal outcomes when compared to the less-tight control group (target DBP of 100 mm Hg). The recently published CHAP trial assigned 2408 patients with mild chronic hypertension during pregnancy to either antihypertensive treatment (first-line drugs) with a blood pressure goal of <140/90 mm Hg or no treatment for hypertension unless the blood pressure reached severe hypertension (control group). They demonstrated that the risk of the primary composite outcome of preeclampsia with severe features, preterm birth, placental abruption, or perinatal death was significantly lower in patients treated with antihypertensive medications. Also, the safety outcome which was SGA below the 10th percentile did not differ between the two groups proposing better outcomes of pregnancy without any observed harm in mild hypertensive patients treated with medications compared to no treatment. 6 The results of this trial suggest the beneficence of pharmacologic treatment of mild chronic hypertension during pregnancy to a blood pressure goal of below 140/90 mm Hg as supported by the latest statement made by the Society for Maternal-Fetal Medicine (SMFM). 7 Herein, we presented a meta-analysis of the impact of pharmacological treatment in mild to moderate hypertension during pregnancy on maternal, neonatal, and fetal outcomes with randomized controlled trials (RCTs) being stratified by their type of hypertension (chronic and gestational or pregnancy-induced).

| METHODS
This is a systematic review and meta-analysis aiming to investigate the potential impact of antihypertensive treatment in patients with mild chronic or gestational hypertension on pregnancy outcomes. We followed and reported this study based on the recommendations made by the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) 2020 Checklist. 8

| Information sources and search strategy
Relevant keywords related to hypertension and outcomes of pregnancy in patients receiving antihypertensive treatment in combination with the medical subject heading terms were used in a systematic search through PubMed, Embase, and Scopus from the database inception to August 2022 with no specific filters. The detailed search terms in each of the databases are provided in the Supporting Information: Material. Additionally, the bibliographies of the previous meta-analyses were screened for further potentially eligible studies.

| Selection process and eligibility criteria
All the records found in database searching were retrieved for further screening. After the removal of the duplicate records, two reviewers  5. Studies or groups that were designed to compare aspirin use during pregnancy. Since we wanted to investigate the impact of only antihypertensive medications on pregnancy outcomes, aspirin use may be a confounding factor given the fact that lowdose aspirin has been shown to be effective in the prevention of adverse maternal and perinatal outcomes such as preeclampsia, preterm birth, and perinatal mortality 10 and the mentioned variables were among our outcomes of interest.

| Outcomes of interest, data collection process, and risk of bias
The endpoints of this meta-analysis included all the maternal, fetal, and neonatal outcomes mentioned in the trials such as severe hypertension, superimposed preeclampsia, placental abruption, and low birth weight (LBW). The definition of each of the study variables is presented in Supporting Information: Table S1. For the safety outcome, our primary outcome was the incidence of SGA. The data were collected by two investigators (A. H. and M. M.) independently and rechecked after the process with discrepancies being resolved through discussion. We extracted data on the year of publication, country of the trial that took place, study design, type of hypertension (chronic, gestation, and mixed), the sample size of the experimental and control groups, maternal outcomes (including severe hypertension, preeclampsia, preterm labor, premature rupture of membrane, placental abruption, hospital admission, heart failure,

| Statistical analysis
We investigated the impact of pharmacological treatment on maternal, neonatal, and fetal outcomes in mild to moderate hypertensive patients during pregnancy. Thus, we compared the effect of treatment with a control group using the random-effects model and generated a risk ratio (RR) and its 95% confidence interval (CI). Antihypertensive treatment was associated with better outcomes if the RR < 1 and in the case of RR > 1 treatment was considered to increase the RR of that specific outcome. If the data on an outcome were repeated in two or more studies, it was included for analysis. Meta-analysis of proportions was conducted to estimate the overall proportions of all the outcomes included for analysis using the inverse variance method. The results of this meta-analysis were visualized by generating forest plots. A subgroup analysis was performed for all the analyses based on the type of hypertension (chronic, gestational, or both) included in each of the trials. All the analyses were performed using Mantel-Haenszel method. The level of heterogeneity between studies was quantified using I 2 statistics and the studies were categorized as considerable, substantial, and moderate heterogeneity if I 2 > 75%, 50%-75%, and 30%-50%, respectively. For the qualitative assessment of publication bias, we generated funnel plots and for the quantitative assessment, we calculated Peter's test p value if there were ≥10 studies included in the analysis. In each of the analyses, if the overall pooled estimate of RR did not cross the line of 1, it was considered to be statistically significant. Sensitivity analysis was conducted by deleting the outlier studies and reanalyzing the remaining ones. The data were analyzed using RStudio software version 1.3.959 with "meta" and "dmetar" packages being used.

| Search results and description of studies
The primary search through databases yielded a total of 5171 records across PubMed, Scopus, and Embase. Following the removal of 1462 duplicate studies, titles, and abstracts of 3709 articles were retrieved for checking the potential eligibility by applying the exclusion criteria.
After discarding the irrelevant papers (n = 3597), we carefully evaluated the full text of the remaining 112 studies and finally, 13 studies 6,12-23 from 12 RCTs including 4461 pregnant patients diagnosed with mild to moderate hypertension (2395 participants in the experimental group and 2066 in the control group) were chosen for inclusion in the analyses. The data were extracted from six trials 6,12,18,19,21,22 including patients with chronic hypertension, three with gestational hypertension, [13][14][15] and three with both types of hypertensive patients. 16,17,23 The included studies varied in the sample size with the smallest trial 22 including 58 pregnant patients and the largest one 6 comprising 2408 patients. PRISMA flowchart of the study is provided in Figure 1. Table 1 summarizes the details of the included trials.

| Quality assessment, publication bias, and sensitivity analysis
Almost all the studies were at low risk of bias due to using a clear method for random sequence generation including randomization in blocks of six with serial numbers, 23 labeled allocations cards numbered 1-150, 13 a list generated by a computer, 17 a list of random numbers, 15 stratified blocks of 10 in 2 centers, 14 computergenerated simple random tables, 18 random numbers generated by the computer, 21 web-based randomization program with variable block sizes of 2, 4, and 6, 6 block sizes of 6 generated with the computer, 19 and simple random table. 12 Allocation concealment was at unclear or high risk of bias in half of the included studies. 12,13,16,17,21,22 Three studies were open-label trials and hence were rated as high risk for performance bias. 6,16,19 Blinding of outcome assessors was not determined in the majority of the trials, thus the detection bias was rated as unclear. [12][13][14][15][16][17][19][20][21][22][23] The risk of bias summary and graph are displayed in Supporting Information:     Figure S6).  Figure S8).    28 The mentioned meta-analysis did not stratify patients

CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.