Hypertension associated with serotonin reuptake inhibitors: A new analysis in the WHO pharmacovigilance database and examination of dose-dependency

Hypertension is a major risk factor for cardiovascular disease and can be secondary to medical conditions or pharmacologic causes. Emerging evidence suggests antidepressants may influence blood pressure; MAOIs can precipitate hypertensive crises in the presence of tyramine or amphetamines, and tricyclic antidepressants have been linked to higher odds of stage 1–2 hypertension. Serotonin reuptake inhibitors (SRIs) are first-line antidepressants due to a more favorable adverse effect profile, yet whether SRIs are associated with hypertension remains debated and is inconsistently reflected in drug labels. Building on earlier pharmacovigilance signals, this study aimed to assess whether a class effect exists between SRIs and hypertension reports and to evaluate dose-dependency using the WHO VigiBase® database with improved confounder control.

Prior studies indicate depression/anxiety is associated with a higher risk of uncontrolled hypertension (adjusted OR 1.82). MAOIs are known to trigger hypertensive crises via monoamine potentiation, and tricyclic antidepressants have shown increased odds for hypertension in cohort data. Previous pharmacovigilance work (Humbert et al., 2019) suggested an SRI–hypertension signal but lacked multivariable adjustment and dose-effect assessment. Product labeling inconsistently lists hypertension risk (notably for sertraline and paroxetine). FAERS analyses have found cardiovascular event signals, including hypertension, with SRIs, particularly in women and older adults. Conversely, a meta-analysis suggested SRIs may reduce peak blood pressure in hypertensive patients with depression. Mechanistic hypotheses include increased extracellular serotonin leading to vasoconstriction, platelet serotonin uptake changes, and nitric oxide synthase inhibition (e.g., paroxetine). Experimental and clinical observations have reported fluoxetine-associated pressor effects under certain conditions. Overall, evidence is mixed, warranting robust real-world signal detection with confounder control.

Data source: VigiBase®, the WHO global Individual Case Safety Report (ICSR) database (>35 million reports from 120 countries since 1968). Adult ICSRs include administrative data, demographics, MedDRA-coded reactions (v27.1), causality assessments, drug details (indication, dosing, dates, dechallenge/rechallenge), and seriousness per WHO criteria. Design: Case/non-case disproportionality analysis. Cases were adult ICSRs reporting hypertension per the MedDRA SMQ "Hypertension (broad)" (SOC v27.1) from 1968 to July 29, 2024. Non-cases were adult reports without hypertension. Exposure was defined by at least one SRI (ATC classification) recorded in the ICSR, regardless of role (suspect, concomitant, or interaction). Statistical analysis: Reporting Odds Ratios (ROR) with 95% CIs were computed for each SRI–hypertension pair. Signals were defined as ROR>1 with lower 95% CI >1. Multivariable analyses estimated adjusted RORs (aRORs) controlling for age category, sex, co-reported antihypertensive drugs (ATC C02), and other drugs known to induce hypertension (per Foy et al.). Dose-response assessment: RORs were estimated for doses ≤ median vs > median for each SRI; linear regression tested dose-effect when ≥5 reports per stratum. Time to onset (TTO) and positive dechallenge/rechallenge were summarized. Sensitivity analyses: (1) restricted to suspected SRIs only; (2) restricted to patients treated with antidepressants (ATC N06A). Analyses followed READUS-PV guidelines and were conducted in R 4.4.1.

- Data set: 625,205 SRI-related ICSRs; 13,688 hypertensive cases (2.2%). - Demographics: 70.0% female; most frequent age 45–64 years (44.8%). Serious outcomes in 75.5% of evaluable reports (9,223/12,208). Concomitant antihypertensive drugs in 8.0%; concurrent CVD events in 1,752 reports (1,002 strokes; 750 myocardial infarctions). - Counts by SRI (hypertension cases): sertraline 3,879; fluoxetine 2,862; citalopram 2,516; escitalopram 2,586; paroxetine 2,441; fluvoxamine 201; zimeldine 8. - Disproportionality: Significant signals for the SRI class (aROR 1.33; 95% CI 1.31–1.35) and for individual SRIs in multivariable models—citalopram (aROR 1.16; 1.12–1.21), escitalopram (1.25; 1.20–1.30), fluoxetine (1.29; 1.25–1.34), paroxetine (1.40; 1.34–1.46), sertraline (1.30; 1.26–1.35). Fluvoxamine (0.99; 0.86–1.14) and zimeldine (0.59; 0.29–1.18) were not significant. Univariate RORs generally ranged 1.43–1.58 by drug. - Dose-response: No dose–effect relationship detected in linear regression or median-split analyses across SRIs (all p-values non-significant where estimable). - Time to onset (median [IQR], days): escitalopram 1.0 [0.0–12.5]; citalopram 2.0 [0.0–16.0]; sertraline 4.5 [0.0–56.0]; paroxetine 5.0 [1.0–119.0]; fluvoxamine 9.0 [1.0–38.0]; fluoxetine 13.0 [1.0–66.0]; zimeldine 1.0 [1.0–4.0]. - Dechallenge/rechallenge (number positive): dechallenge—sertraline 275, fluoxetine 253, paroxetine 212, escitalopram 168, citalopram 138, fluvoxamine 41, zimeldine 1; rechallenge—fluoxetine 15, sertraline 12, paroxetine 5, citalopram 3, escitalopram 2, fluvoxamine 2. - Sensitivity analyses: In suspected-SRI-only reports, no signal except sertraline (aROR 1.07; 1.00–1.15). In the antidepressant-treated population (ATC N06A), signals persisted for citalopram, escitalopram, fluoxetine, paroxetine, and sertraline (multivariable).

Findings support a pharmacovigilance signal linking SRI exposure with reported hypertension after adjusting for age, sex, antihypertensive co-medication, and other hypertensive drugs, suggesting a possible class effect without dose-dependency. Results extend prior VigiBase analyses by incorporating multivariable adjustments and explicit dose-effect testing. Literature is mixed: FAERS data show cardiovascular adverse event signals for SRIs, while some clinical and meta-analytic data in hypertensive depressed patients suggest potential blood pressure lowering. Biological plausibility includes serotonin-mediated vasoconstriction, altered platelet serotonin handling, and nitric oxide synthase inhibition (e.g., paroxetine), as well as sympathetic activation pathways observed experimentally with fluoxetine. Despite possible notoriety bias and confounding by depression (linked to incident hypertension), the short time to onset and positive dechallenge/rechallenge observations lend support to a drug-related effect in some cases. Clinically, awareness and monitoring of blood pressure in SRI-treated patients is warranted, particularly among middle-aged and older adults and women who were overrepresented in reports.

This real-world pharmacovigilance analysis reveals a significant safety signal associating SRI use with reported hypertension, consistent with a potential class effect and no evident dose–response relationship. Clinicians should monitor blood pressure in patients receiving SRIs and tailor dosing based on clinical indications rather than dose–effect concerns for hypertension. Further longitudinal and comparative studies with robust confounder control are needed to clarify causality, quantify risk, and identify susceptible subgroups.

Pharmacovigilance data are subject to under-reporting, reporting biases (including notoriety bias), and missing data. Case/non-case disproportionality quantifies reporting association rather than incidence or risk and cannot establish causality. Residual confounding is possible, notably from depression itself being associated with hypertension. The sensitivity analysis restricted to suspected SRIs showed minimal signal (only sertraline marginally), raising the possibility of confounding or co-medication effects. Dose information completeness may limit dose–response analyses. Clinical trial data cited often involve normotensive populations, limiting generalizability to real-world patients.