Association of antidepressant use during pregnancy and pregnancy-induced hypertension: a systematic review and meta-analysis

Abstract

Background

The association between antenatal antidepressant use and pregnancy-induced hypertension (GH/PE) remains controversial, with conflicting evidence on whether depression itself confounds this risk.

Methods

We searched PubMed, Web of Science, Medline, Embase, Cochrane Library, and the China National Knowledge Infrastructure electronic databases from inception to Dec. 31, 2024. Eligibility criteria for selecting studies are randomized trials, cohort studies, and case-control studies exploring the association between antidepressant use during pregnancy and pregnancy-induced hypertension. Two reviewers extracted data and assessed quality. Outcomes were expressed as odds ratios with 95% confidence intervals.

Results

Seventeen observational studies were analyzed, indicating a potential link between antidepressant use in pregnancy and an elevated risk of gestational hypertension/preeclampsia (GH/PE). Pooled odds ratios (ORs) stratified by drug class demonstrated significant associations for: Selective serotonin reuptake inhibitors (OR = 1.34; 95% CI: 1.14–1.58); Serotonin-norepinephrine reuptake inhibitors (OR = 1.94; 95% CI: 1.40–2.70); Tricyclic antidepressants (OR = 2.00; 95% CI: 1.32–3.04). Following adjustment for depression as a confounding variable, sensitivity analysis indicated: SNRIs maintained significant association with GH/PE (adjusted OR = 1.53; 95% CI: 1.25–1.88); SSRIs showed no significant risk (adjusted OR = 1.03; 95% CI: 0.97–1.10); TCAs lost statistical significance (adjusted OR = 1.80; 95% CI: 0.86–3.74). Subgroup analyses suggesting geographical variations in effect magnitude.

Conclusion

This study demonstrates differential associations between antenatal exposure to three antidepressant classes (SSRIs, SNRIs, TCAs) and gestational hypertension/preeclampsia risk. Enhanced blood pressure monitoring is advised during SNRI therapy, with clinical decisions requiring individualized risk-benefit assessment. Residual confounding constrains causal inference, necessitating prospective studies with standardized depression scales and systematic exposure documentation to resolve methodological limitations.

PROSPERO registration

PROSPERO: CRD42025640646.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-025-07974-6.

Introduction

Among pregnant women, the incidence of hypertension is about 13% [1]. However, the rise in blood pressure during pregnancy is primarily due to physiological causes, and 70% of pregnant women will gradually return to normal blood pressure after delivery [2]. Factors affecting hypertension during pregnancy include pregestational obesity, diabetes, maternal age, and polycystic pregnancy [3]. Gestational hypertension (GH) and preeclampsia (PE) are the most common hypertensive disorders during pregnancy. They are not only the leading causes of maternal/neonatal morbidity and mortality [4–6] but also risk factors for developing cardiovascular diseases later in life [7]. GH is defined as hypertension that occurs for the first time after 20 weeks of gestation. PE is defined as GH accompanied by one or more of the following conditions: proteinuria, thrombocytopenia, renal insufficiency, impaired liver function, pulmonary edema, or cerebral or visual symptoms [8]. Severe pre-eclampsia can also result in hemolysis, elevated liver enzymes, and a low platelet count syndrome [9]. GH is a known risk factor for PE, as patients with GH often progress to PE [10]. Reports indicate that approximately 12–25% of pregnant women exhibit some signs of depressive symptoms during pregnancy [11], and mental pathology often heightens the risk of hypertension [12]. On the other hand, pregnant women with chronic hypertension may have a higher risk of depression [13]. About 2–8% of pregnant women are treated with antidepressants, primarily selective serotonin reuptake inhibitors (SSRIs), with other subtypes including selective serotonin-norepinephrine reuptake inhibitors (SNRIs) and tricyclic antidepressants (TCAs) [14–18].

Prior studies have reported that exposure to antidepressants during the perinatal period might be associated with pregnancy complications and adverse neonatal outcomes, such as preterm birth, being small for gestational age, persistent pulmonary hypertension of the newborn, and delayed motor development in children [19–22]. However, the association between antidepressant use during pregnancy and GH or PE remains unclear. Some studies have reported that women who use antidepressants during pregnancy have a higher risk for GH or PE [23–25], while others have not found such an association [26, 27]. Although Uguz et al. conducted a systematic review to explore the relationship between gestational use of antidepressants and PE/GH [10], the study was limited by qualitative descriptions of previous studies without synthesized data analysis, preventing a definitive conclusion on this topic. A subsequent meta-analysis of seven cohort studies identified increased GH/PE risks with SSRI monotherapy [28], yet limitations included a narrow focus on SSRI agents and insufficient exploration of exposure timing/duration. Gumusoglu and colleagues performed a comprehensive systematic review on the risks associated with SSRIs. Despite identifying an increased risk of preeclampsia, they advised against discontinuing depression treatment due to the potential risk of gestational hypertension [29].

Most current studies focus on the association of SSRIs with pregnancy-induced hypertension. Other antidepressants, such as SNRIs or TCAs, have been less studied. Therefore, this systematic review and meta-analysis aim to explore the association between antidepressant administration during pregnancy and the risk of GH/PE. Furthermore, we differentiated between various reference and control groups in the literature we reviewed to mitigate the influence of confounding factors related to depression. This study may potentially provide evidence for a more definitive conclusion to guide recommendations for managing depression and pregnancy complications.

Materials and methods

This systematic review and meta-analysis of scientific peer-reviewed literature were performed by following the recommendations from the Preferred Reporting Items for Systematic Reviews and Meta-analysis statements. The PRISMA Checklist is presented in supplementary file 1. The protocol was registered in advance with PROSPERO (registration number CRD42025640646).

Literature search strategy

In this systematic review and meta-analysis, two researchers systematically searched Pubmed, Web of Science, Medline, Embase, Cochrane Library, and the China National Knowledge Infrastructure electronic databases for studies published until December 31, 2024. Grey literature was not included in the present study. The following search strategy was applied: (‘antidepressant’ OR ‘serotonin and noradrenaline reuptake inhibitors’ OR ‘tricyclic antidepressants’ OR ‘5-hydroxytryptamine uptake inhibitors’) AND (‘pregnancy induced hypertension’ OR ‘preeclampsia’ OR ‘gestational hypertension’). The reference lists of the included study were afterward hand-searched to identify additional relevant papers. All the search results were imported into Endnote (X9) and duplicates, if any, were removed.

Eligibility criteria and study selection

The exposure of interest was the use of antidepressants during pregnancy. The primary outcome was pregnancy-induced hypertension. The inclusion criteria were listed as follows: (a) the study was an observational study including cohort study or case-control study; (b) The non-exposed group was defined as pregnant women who did not use any type of antidepressants. This group included both unmedicated individuals with depression and those without any psychiatric diagnoses. (c) the article reported the association between the use of antidepressants and pregnancy-induced hypertension; (d) the report was not a review, comment, or letter.

Data extraction

Titles, abstracts, and full texts were reviewed by two researchers independently, and data were extracted from included articles. In case of discordance between the two researchers, the consensus was made with the help of the third researcher. The following data were extracted from individual studies, including first author, publication year, study location, study design, sample size, types of antidepressants, outcome, and assessment of outcome. Characteristic of the included studies was summarized in Table 1.

Table 1.

Baseline characteristics of included studies

First author and year	Country	Study design	Sample Size	Age(years)	Assessment of exposure	Outcome	Assessment of outcome
Toh 2009 [24]	USA and Canada	retrospective cohort study	5731	< 25(18.9%)	Telephone interview	GH/PE	Self-reported
Katon 2012 [13]	USA	prospective study	1656	≥15 years	Self-report	GH	ICD-9 diagnoses
Palmsten 2012 [33]	Canada	retrospective cohort study	69,448	Median age was 31	pharmacy dispensing record	PE	ICD-9
De Vera 2012 [17]	Canada	case–control study	13,376	Mean age was 27.3	Prescription Drug File	GH	ICD-9
Palmsten 2013 [39]	Canada	prospective cohort study	85,767	Median age was 23–27	medical records	PE	ICD-9CM diagnosis code
Avalos 2015 [23]	USA	retrospective cohort study	18,134	18 years and older	electronic medical records	PE	ICD-9CM diagnosis code
Malm 2015 [18]	Finland	prospective study study	25,381	≤40 years	Drug Reimbursement Register	GH	Medical Birth Register
De Ocampo 2016 [62]	USA and Canada	retrospective cohort study	3471	Mean age was 32.0 for unexposed	interview	GH/PE	Self-reported
Lupattelli 2017 [26]	Norway	prospective cohort study	5745	Mean age was 30 for exposed	record	PE	ICD-9CM diagnosis code
Yonkers 2017 [22]	USA	retrospective cohort study	2654	≥35(26.6%)	Telephone interview	GH	Medial records
Zakiyah 2018 [11]	Netherlands	retrospective cohort study	28,020	Mean age was 31.1 for exposed	The prescription database IADB	GH	Dutch Society of Obstetrics and Gynaecology
Bernard 2019 [27]	Canada	prospective cohort study	6761	18 years and older	record	PE	ICD-9CM diagnosis code
Huybrechts 2020 [14]	USA	cohort study	1.3 to 4.1 million	18–55	outpatient prescription	PE	ICD-9 diagnostic
Tran 2021 [6]	Netherlands	retrospective cohort study	94,788	≥35(11.61%)	Netherlands Perinatal Registry and the PHARMO Database	PE	medical records
Yang 2021 [61]	Taiwan, China	retrospective cohort study	5664	Mean age was 31.3	prescription drug utilization of Health and Welfare Database	GH/PE	ICD-9CM diagnosis code
Galbally 2023 [9]	Australia	retrospective cohort study	229	Mean age was 32	Mercy Pregnancy Emotional Wellbeing Study	GH	Australian Guidelines
Benevent 2023 [25]	France	retrospective cohort study	143,391	Mean age was 30.5 for unexposed	prescription drug utilization of EFEMERIS database	GH/PE	Medical records

SSRIs selective serotonin reuptake inhibitors, SNRIs selective serotonin-norepinephrine reuptake inhibitors, TCAs tricyclic antidepressants, GH/PE gestational hypertension or preeclampsia

Quality assessment

The quality of each included study was assessed by two researchers independently using a nine-point Newcastle-Ottawa Scale (NOS). The following three perspectives were assessed in each study: the selection of patients, the comparability of groups, and the ascertainment of either the exposure or outcome of interest. The higher the score, the better the methodological quality of the study. A score of ≥ 6 indicated good quality. The quality assessment of the included studies was summarized in Supplemental Table 1.

Statistical analysis

All statistical analyses were performed using the Stata software version 18.0. Homogeneity was estimated by using Cochran’s Q test and Higgins’s I-squared (I²) statistic. The I² statistic estimates the proportion of total variance across studies attributed to true between-study heterogeneity rather than sampling error. Heterogeneity magnitude was characterized using widely adopted thresholds (Low: I² ≤ 25%; Moderate: 25% < I² ≤ 50%; High: I² >50%) proposed by Higgins et al. [30]. To obtain conservative and reliable results, all pooled risk estimates were calculated by the random-effect model. Furthermore, the literature highlights various indications for antidepressants [31, 32], such as depression, anxiety, neuropathic pain, obsessive-compulsive disorder, and insomnia. To mitigate the influence of confounding factors related to depression, a clear distinction was made between the reference and exposure groups based on the presence or absence of a depression diagnosis, followed by sensitivity analyses. Publication bias was examined by the use of funnel plots and Egger’s test.

Results

Baseline characteristics of included studies

A total of 994 potentially eligible records were identified from the electronic databases and reference lists. Finally, 17 eligible studies (consisting of 11 retrospective cohort studies, 5 prospective cohort studies, and 1 case-control studies) were included. The flow chart of study selection is presented in Fig. 1. Briefly, the studies were published during the period from 2009 to 2023. Among the 17 included studies, 4 studies were conducted in the United States of America, 4 in Canada, 2 in the Netherlands, 1 in Norway, 1 in Finland,1 in France, 1 in Australia,1 in Chinese Taiwan, 1 in Australia and Canada, and 1 in the USA and Canada. The assessment of outcomes in 15 studies was obtained from medical records and 2 from self-reports. The characteristics of the included studies are shown in Table 1, and their adjusted effect sizes and their respective adjusted covariates are shown in Table 2.

Fig. 1.

Literature screening flowchart

Table 2.

Effect sizes of studies of antidepressants and associated gestational hypertension or preeclampsia outcomes

Study	Antidepressants	Adjusted effect size (95%CI)	Covariates	Limitation	Exclusions	depressive disorder (exposed vs. control group)
Toh 2009 [24]	SSRIs	aRR 4.86 (2.70–8.76)	age, diabetes, smoking, BMI	Sev, Dose, med	chronic hypertension, abortion	unidentified vs. unidentified
Katon 2012 [13]	SSRIs	aOR 1.06 (0.65–1.73)	age, education, chronic disease	n, med	chronic hypertension	depression vs. no depression
Palmsten 2012 [33]	SSRIs SNRIs TCAs	aRR 1.22 (0.97–1.54) aRR 1.95 (1.25–3.03) aRR 3.23 (1.87–5.59)	age, diabetes, multiple pregnancy, BMI, anticonvulsants, benzodiazepines	Dose	chronic hypertension	depression vs. depression
De Vera 2012 [17]	SSRIs SNRIs	aOR 1.60 (1.00-2.55) aOR 0.75 (0.17–3.25)	age, diabetes, asthma, cardiovascular diseases	Sev	chronic hypertension	unidentified vs. unidentified
Palmsten 2013 [39]	SSRIs SNRIs TCAs	aRR 1.00 (0.93–1.07) aRR 1.52 (1.26–1.83) aRR 1.62 (1.23–2.12)	age, diabetes, multiple pregnancy, other antidepressants	Sev	preeclampsia before medication administration	depression vs. depression
Avalos 2015 [23]	SSRIs SNRIs	aRR 1.34 (1.00-1.81) aRR 1.49 (0.57–3.90)	age, BMI, smoking, diabetes, parity	Dose, n, med	multiple pregnancy, chronic hypertension	depression vs. no depression
Malm 2015 [18]	SSRIs	aOR 1.10 (0.97–1.26)	smoking, parity, diabetes, tranquilliser	Dose, Sev	chronic hypertension	depression vs. depression
De Ocampo 2016 [62]	SSRIs SNRIs	aOR 1.27 (0.71–2.26) aOR 4.21 (1.44–12.31)	age, BMI, smoking, diabetes, parity, asthma	n, Sev	chronic hypertension, abortion	unidentified vs. unidentified
Lupattelli 2017 [26]	SSRIs	aRR 0.96 (0.64–1.45)	age, BMI, smoking, parity, education, analgesics	Dose	multiple pregnancy, abortion	depression vs. depression
Yonkers 2017 [22]	SSRIs	aOR 2.82 (1.58–5.04)	age, education, smoking	Dose	diabetes, chronic hypertension	unidentified vs. unidentified
Zakiyah 2018 [11]	SSRIs TCAs	aOR 2.07 (1.25–3.44) aOR 1.60 (0.50–5.09)	age, antibiotic prescription	Sev	diabetes, chronic hypertension	unidentified vs. unidentified
Bernard 2019 [27]	SSRIs SNRIs	aOR 3.09 (1.22–7.85) aOR 6.46 (2.49–16.78)	age, BMI, smoking, parity	Dose, Sev, med	multiple pregnancy	unidentified vs. unidentified
Huybrechts 2020 [14]	SNRIs	aRR 1.12 (0.96–1.31)	diabetes, multiple pregnancy, bipolar affective disorder, schizophrenia, opioid analgesics	Sev, Dose, med	chronic hypertension	unidentified vs. unidentified
Tran 2021 [6]	SSRIs TCAs	aRR 1.08 (0.81–1.44) aRR 2.46 (1.51–4.02)	diabetes, smoking, multiple pregnancy, vasoconstrictors	Sev, Dose	chronic hypertension	unidentified vs. unidentified
Yang 2021 [61]	SSRIs SNRIs TCAs	aRR 0.94 (0.67–1.31) aRR 1.12 (0.64–1.97) aRR 0.31 (0.04–2.22)	age, BMI, diabetes, asthma, epilepsy, renal diseases, liver disease	Sev	chronic hypertension	depression vs. depression
Galbally 2023 [9]	SSRIs SNRIs	aRR 1.54 (0.53–4.41) aRR 9.10 (3.82–21.67)	age, BMI, parity, smoking, education	n	chronic hypertension	depression vs. no depression
Benevent 2023 [25]	SNRIs	aOR 1.89 (1.13–3.18)	age, parity, diabetes, multiple pregnancy, medication used to treat neuropathic pain and anxiety	n, Sev, Dose	chronic hypertension	unidentified vs. unidentified

Sev lack of data on the severity of depression, Dose lack of data on drug dosage, n small exposure group, med no control over concurrent medication use

Synthesis of the included studies

Of the 17 articles assessed (Table 2), 15 analyzed the risk of SSRIs, 10 analyzed the risk of SNRIs, and 5 analyzed the risk of TCAs. Of these, six studies have reported that the use of SSRIs during pregnancy has a significant statistical impact on the increased risk of gestational hypertension or preeclampsia; six studies reported a statistically significant effect of gestational SNRIs use on an increased risk of gestational hypertension or preeclampsia; and three studies reported an increased risk of preeclampsia or gestational hypertension associated with the use of TCAs during pregnancy. Although nearly all primary studies excluded patients with chronic hypertension or multifetal pregnancies to minimize confounding, several critical limitations persist: (a) insufficient sample sizes in antidepressant-exposure subgroups; (b) inadequate control for concomitant medications; (c) lack of standardized assessment for depression severity, preventing dose-response and indication bias analyses; (d) absence of quantitative dose-effect evaluations. Collectively, these constraints may attenuate the robustness of pooled effect estimates.

SSRIs and the risk of GH/PE

The overall pooled OR for the association between SSRIs use and GH/PE was 1.34 (OR = 1.34; 95% CI: 1.14–1.58) (Fig. 2). However, substantial heterogeneity was observed (I² = 76.29%, P < 0.001). Regarding GH or PE, the current study revealed that taking SSRIs during pregnancy was significantly associated with an increased risk of GH (OR = 1.53; 95% CI: 1.10–2.13) (Supplemental Fig. 1). However, SSRI use did not significantly influence the risk of PE (OR: 1.14; 95% CI: 0.97–1.33) (Supplemental Fig. 2).

Fig. 2.

Forest plot of the association between SSRIs and risk of GH/PE

SNRIs and the risk of GH/PE

The pooled estimates for the association between SNRIs use and the total risk of GH/PE was 1.94 (OR = 1.94; 95%CI: 1.40–2.70) (Fig. 3). In terms of PE, this meta-analysis indicated that pregnant women taking SNRIs had a lower risk for PE (OR = 1.67; 95% CI: 1.18–2.35) (Supplemental Fig. 3). Since data on GH and SNRI use during pregnancy were limited in the included studies, the analysis of their association was not performed.

Fig. 3.

Forest plot of the association between SNRIs and risk of GH/PE

TCAs and the risk of GH/PE

For the association between the use of TCAs and the overall risk of GH/PE, the composite estimate was 2.00 (OR = 2.00; 95%CI: 1.32–3.04) (Fig. 4). For PE, this meta-analysis showed that pregnant women taking TCAs had a higher risk of preeclampsia (OR = 2.22; 95%CI: 1.45–3.40) (Supplemental Fig. 4). Due to limited data on GH and TCAs use during pregnancy in the included studies, an analysis of their association was not performed.

Fig. 4.

Forest plot of the association between TCAs and risk of GH/PE

Subgroup analysis

We conducted a subgroup analysis based on the characteristics of the included studies. The studies were categorized by study quality, study location, and sample size. As only one study utilized a case-control design, subgroup analyses based on study design were not conducted. Furthermore, due to the limited number of studies available, our subgroup analysis focused solely on SSRIs. Although the interval of pooled OR values for subgroup in Europe included null values, differences between regions did not show significant significance. This might be because there are only four studies conducted in the European region. The results of the subgroup analysis indicate that the combined effect size is stable and unaffected by the factors above (Table 3). The forest plots for the subgroup analysis are presented in Supplemental Figs. 5 through 7.

Table 3.

Subgroup analysis of the association between antidepressants and risk of GH/PE

	No. of study	Pooled OR	95% CI	I2 (%)	Group differences(P)
Quality score
≤ 6	3	2.36	1.05–5.33	77.92	0.11
>7	12	1.20	1.05–1.37	64.04
Study area
Europe	4	1.16	0.94–1.44	52.66	0.09
North America	9	1.59	1.20–2.11	84.39
Sample size
≤ 10,000	8	1.65	1.07–2.57	80.33	0.15
>10,000	7	1.18	1.03–1.35	63.60

Sensitivity analysis

Antidepressants are prescribed for various conditions such as anxiety, neuropathic pain, fibromyalgia, and depression. In our meta-analysis, we selected five studies that focused on depression diagnosis in exposed and control groups to prioritize the use of these medications for depression. Our analysis indicates that the use of SNRIs during pregnancy is associated with an increased risk of GH/PE (OR = 1.53; 95%CI: 1.25–1.88). However, the confidence intervals for SSRIs and TCAs encompass neutral values (OR = 1.03; 95%CI: 0.97–1.10) (OR = 1.80; 95%CI: 0.86–3.74). More details are shown in Fig. 5.

Fig. 5.

Forest plot of sensitivity analysis (A SSRIs; B SNRIs; C TCAs)

Publication bias

The Egger test was employed to assess publication bias and small sample effects within the included studies. Analysis of both SNRIs and TCAs studies through visual examination of funnel plots (Supplemental Fig. 8; Supplemental Fig. 9) and Egger test outcomes (P = 0.0671; P = 0.2792) revealed no definitive indications of publication bias. Conversely, for SSRIs studies, the Egger test indicated potential publication bias (P = 0.003), accompanied by a slightly skewed distribution of studies in the funnel plot (Supplemental Fig. 10). This asymmetry may be attributed to the limited number of participants included in the study.

Discussion

This meta-analysis found that the use of antidepressants during pregnancy may increase the risk of gestational hypertension/preeclampsia (GH/PE). Our results are consistent with previous studies [33]; however, the strength of these associations based on observational studies remains debatable. The heterogeneity of our findings was high, and although we used adjusted OR values, the interference of confounding factors cannot be entirely ruled out. Additionally, although the odds ratios (ORs) in subgroup analyses exhibited some fluctuations, these variations may be related to the number of included studies.

In the primary analysis, we examined the potential associations between different classes of antidepressants (SSRIs, SNRIs, and TCAs) and GH/PE, and calculated the pooled odds ratios (ORs) for these three therapeutic agents. TCAs, SNRIs, and SSRIs inhibit serotonin transporters or both serotonin and norepinephrine transporters, increasing the extracellular concentration of these monoamines [34]. These increased substances induce constriction of the uterine and umbilical blood vessels leading to ischemia, which may contribute to preeclampsia [35, 36]. Alternatively, SSRIs cause an increase in serotonin levels, which in turn enhances vasoconstriction of the umbilical-placental vessels. They also inhibit the synthesis of an important vasodilator, nitric oxide [37, 38].

In our sensitivity analysis, the studies were grouped based on the depression diagnosis status of the exposed and control groups. We identified five studies where all participants were confirmed to have depression, aiming to minimize potential confounding effects related to depression. Within this cohort of women with depression, the utilization of antidepressants showed associations with the risk of GH/PE (SSRIs: OR, 1.03, 95% CI, 0.97–1.10; SNRIs: OR, 1.53, 95% CI, 1.25–1.88; TCAs: OR, 1.80, 95% CI, 0.86–3.74). These results indicate that in pregnant women with diagnosed depression, the use of SSRIs and TCAs may not be linked to GH/PE. Nevertheless, further investigation is warranted regarding the potential adverse effects of SNRIs. Several studies have established that depression itself poses a risk for pregnancy-induced hypertension [39–43]. A systematic review establishes an independent association between prenatal depression and hypertensive disorders of pregnancy, indicating that depression with onset before 20 weeks gestation is a significant risk factor for the development of hypertensive disorders of pregnancy [44]. The underlying mechanism may involve alterations in the excretion of vasoactive hormones or other neuroendocrine transmitters, which could contribute to the increased risk of preeclampsia [40, 45]. Frayne and Galbally et al. also identified an association between depression and hypertensive disorders during pregnancy in their retrospective study [1, 9].

The sensitivity analysis findings aligned with Newport et al.‘s case-cohort study [46], which indicated that antenatal SNRI exposure significantly increased hypertensive disorder risk (OR = 2.57; 95%CI:1.34–4.93), whereas gestational SSRI use showed no significant association. The distinct cardiovascular profiles of these agents arise from divergent pharmacodynamic mechanisms [47]. SSRIs exclusively inhibit serotonin (5-HT) reuptake, yielding divergent hemodynamic outcomes: pressor responses mediated by peripheral 5-HT_2a receptor activation (sympathetic excitation) versus depressor effects resulting from central 5-HT_1a stimulation (cardiac vagal tone enhancement and sympathetic suppression) [48–50]. Conversely, SNRIs exhibit dual reuptake inhibition of both 5-HT and norepinephrine (NE). Critically, NE activates vascular α₁-adrenergic receptors, directly inducing arterial vasoconstriction. This noradrenergic potentiation, mediated through NE transporter blockade, triggers sympathetic activation and subsequent blood pressure elevation [51, 52]. Consequently, the absence of NE modulation in SSRIs explains their reduced hypertensive liability compared to SNRIs. Concerning the association between SNRIs and preeclampsia, Staff suggested that the vasoconstrictive properties of these drugs and their effects on embryonic development could be potential underlying mechanisms [53]. In the case of TCAs, these medications can interact with various receptors, such as muscarinic receptors, cholinergic receptors, histamine H₁ receptors, and α₁-adrenergic receptors, which may contribute to their etiology [54]. In the sensitivity analysis, although the confidence interval (CI) for TCAs was wide and included the null value (OR = 1.0), a potential association between TCA exposure and both GH and PE cannot be ruled out. This statistical uncertainty likely stems from the limited sample size of TCA-exposed cases in the analyzed studies. Larger-sample studies are needed in the future to validate the association of TCA with hypertension in pregnancy.

Robust evidence regarding antidepressant use in pregnant women with depression remains limited, necessitating further high-quality studies to clarify risk-benefit profiles. Importantly, SNRI may require more rigorous blood pressure monitoring, whereas SSRI may have a lower risk (based on differences in existing OR values). According to an Italian expert consensus, depression is not a contraindication to pharmacotherapy during pregnancy, and patients may continue antidepressant treatment if clinically indicated. And to mitigate the risk of potential adverse effects, priority should be given to agents with established safety profiles and extensive epidemiological data, such as SSRIs, prescribed at the lowest effective dose [55]. This approach aligns with Canadian clinical practice guidelines indicating that after adjustment for confounding by indication, associations between most antidepressants and adverse pregnancy outcomes are significantly attenuated [56].

In the study comparing the use of antidepressants to untreated depression during pregnancy, it was found that untreated depression was associated with significantly increased risks of premature birth and low birth weight [57, 58]. On the other hand, not receiving treatment might also lead to suicidal tendencies and recurrence of depression in pregnant women [59, 60]. Considering the impact of untreated depression during pregnancy on maternal and infant health, whether to continue taking antidepressants must be determined by clarifying the benefits of the treatment plan.

Notably, the positive correlation between longer duration/higher cumulative doses of antidepressants and preeclampsia risk [23, 61, 62] may be particularly relevant for SNRIs, given their noradrenergic mechanism. Consequently, regulating the timing and dosage of SNRIs during pregnancy is essential to mitigate pregnancy-induced hypertension, whereas SSRIs may require less stringent management. Given the evidence gap in depression research highlighted in Table 2, dosage, course of treatment, and drug blood concentration data will need to be included in future studies to refine clinical guidance. This is critical to disentangle confounding by indication. To operationalize individualized treatment, clinicians should: select antidepressants by risk profile (prefer SSRIs for mild-moderate depression; limit SNRIs to refractory cases with cardiovascular screening), and implement structured monitoring. Additionally, antidepressant use is not the sole factor influencing hypertensive disorders during pregnancy. Research indicates a clear genetic predisposition to pre-eclampsia due to alterations near the fumes-like tyrosine kinase 1 locus in the human fetal genome. Other contributing factors include the number of pregnancies, maternal age, smoking history, diabetes, and chronic kidney disease [63].

There are differences in the baseline characteristics of the populations involved in the included studies. Certain key confounding factors—such as body mass index (BMI), smoking status, and diabetes status—were inconsistently adjusted for: BMI was unadjusted in 8 studies, smoking status in 8 studies, and diabetes status in 6 studies. These differences in covariate adjustment may introduce residual confounding bias, limiting the internal validity of the pooled estimates. Although random-effects models partially addressed between-study heterogeneity, differential adjustment practices across studies could still bias the effect estimates and compromise precision. Additionally, although we controlled for depression as a confounding variable, we cannot rule out the influence of other unidentified confounders that may affect the association between antidepressants and GH/PE. We were unable to determine whether pregnant women who initiated antidepressant treatment had more severe depression prior to therapy. Compared to controls, the exposed groups receiving medication may have exhibited more severe depressive symptoms. Notably, the occurrence of hypertensive disorders of pregnancy is also influenced by factors such as chronic hypertension, diabetes, and multifetal gestation. Although the adjusted covariates for OR values included many confounders, the treatment of antenatal depression is complex, and the findings across these studies are often contradictory. Overall, the causal relationship between antidepressant uses and hypertensive disorders of pregnancy requires validation through methodologically rigorous clinical trials.

Strengths and limitations

This review offers several contributions. First, all procedures adhered to the Cochrane Handbook for Systematic Reviews criteria. The exhaustive search spanned multiple databases and incorporated numerous high-quality original studies, lending credibility to our findings. Our analysis focused on the relationship between conventional antidepressants and pregnancy-induced hypertension. Second, our review encompasses a vast sample size and the association trends remained robust across subgroup analyses, ensuring stable and reliable results. Finally, our sensitivity analysis eliminated the confounding interference of depression, providing new insights into whether to continue antidepressant treatment during pregnancy.

However, the study also has some limitations. First, antidepressant exposure data was retrospectively gathered through registries and healthcare databases. Moreover, some studies relied on self-reported antidepressant exposure, raising concerns about the authenticity of drug exposure. Second, significant heterogeneity was observed for certain outcomes. Even though we conducted sensitivity and subgroup analyses, the high heterogeneity might still influence the results. Furthermore, we were unable to conduct subgroup analyses based on other confounding factors (e.g., the duration of medication use, the dosage of the drug, or the severity of depression), thus we were unable to assess the impact of these factors on our research results.

Conclusion

This study indicates that the use of antidepressants during pregnancy may elevate the likelihood of developing gestational hypertension or preeclampsia. However, there was notable variability among the studies included, and limitations existed due to several confounding variables. Upon adjusting for depression as a confounder, SNRIs were consistently linked to gestational hypertension/preeclampsia, while SSRIs and TCAs did not exhibit such an association. It is noteworthy that this study could not definitively establish the impact of various confounding factors, such as the severity of depression or medication dosage, on the emergence of gestational hypertension. When considering the prescription of antidepressants to pregnant individuals, healthcare providers should carefully assess the potential risks in comparison to the benefits. Subsequent research endeavors should adhere to standardized methodologies and uphold high-quality standards to provide further insights into these variables.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

GH

Gestational hypertension

PE

Preeclampsia

SSRIs

Selective serotonin reuptake inhibitors

SNRIs

Selective serotonin-norepinephrine reuptake inhibitors

TCAs

Tricyclic antidepressants

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

PROSPERO

International Prospective Register of Systematic Reviews

NOS

Newcastle-Ottawa Scale

Author contributions

SH: Writing–review & editing, Data curation. TL: Writing–review & editing, Data curation. LL: Writing–original draft, Formal analysis. HX: Writing–original draft, Supervision, Conceptualization. FZ: Writing–original draft, Supervision, Conceptualization. JL: Writing–original draft, Supervision, Data curation, Conceptualization. All authors reviewed the final version of the manuscript and provided their consent for publication.

Funding

This study was funded by the Changsha Municipal Health Committee (No. KJ-B2023032); and the Natural Science Foundation of Hunan Province (No. 2020JJ8044).

Human Ethics and Consent to Participate declarations: not applicable.

Data availability

The raw data of this study can be obtained from the corresponding author under reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.