No clear link between early-pregnancy GLP-1 use and hypertensive disorders

GLP-1 receptor agonists are not approved for use during pregnancy, and current guidance generally recommends stopping them before conception. But because the drugs are increasingly prescribed to women of reproductive age, unplanned pregnancies during periconceptional or early GLP-1 exposure are a real and growing clinical scenario — one a meta-analysis published this month in Endocrine specifically investigates for its relationship to hypertensive pregnancy disorders [s1].

The design

Researchers searched the Cochrane Central Register of Controlled Trials, ClinicalTrials.gov, PubMed, Scopus, and Embase from inception through 15 December 2025, for studies examining GLP-1 receptor agonist exposure during the preconception period or first trimester and its association with hypertensive disorders of pregnancy (HDP), a category including conditions like preeclampsia and gestational hypertension [s1]. Pooled odds ratios were calculated using a random-effects model, and risk of bias was assessed with the ROBINS-I tool, a standard framework for evaluating bias in observational studies [s1].

What it found

Of 75 records initially identified, only 3 retrospective cohort studies met inclusion criteria, covering 10,880 total pregnancies — 4,942 exposed to a GLP-1 receptor agonist and 5,938 unexposed [s1]. All three studies were conducted in the United States between 2014 and 2025, and collectively covered exposure to semaglutide, liraglutide, dulaglutide, tirzepatide, exenatide, lixisenatide, and albiglutide [s1]. The three studies disagreed with each other on direction: two found lower hypertensive disorder risk among GLP-1-exposed pregnancies, while one found higher risk [s1]. In the pooled analysis, GLP-1 receptor agonist exposure showed an odds ratio of 0.91 for hypertensive disorders of pregnancy (95% CI 0.57–1.47) — a result with no statistical significance, since the confidence interval spans well below and above 1.0 [s1].

Why "no significant association" here means "we don't know" more than it means "it's safe"

A pooled odds ratio of 0.91 with a wide, non-significant confidence interval (0.57 to 1.47) is a meaningfully different, weaker kind of finding than a well-powered study showing no effect with a tight confidence interval clustered near 1.0. This result reflects genuine uncertainty from thin evidence — just three studies, with disagreeing individual results — rather than robust reassurance that early GLP-1 exposure carries no elevated hypertensive pregnancy risk. The confidence interval's range, from a 43% risk reduction at one end to a 47% risk increase at the other, is wide enough to be compatible with a real protective effect, a real harmful effect, or no effect at all.

Why only three studies exist for a question this clinically relevant

GLP-1 receptor agonists are recommended to be discontinued before conception, and pregnancy itself is generally an exclusion criterion in the drugs' clinical trials — meaning almost all available data on in-pregnancy exposure comes from unplanned, real-world scenarios rather than from any study specifically designed around pregnancy safety. That scarcity is itself informative: with GLP-1 prescriptions expanding rapidly among reproductive-age women for both diabetes and weight management, and pregnancies not always planned, the population of women becoming pregnant during unrecognized or recent GLP-1 use is real and likely growing — yet the evidence base to counsel them remains limited to three retrospective cohorts.

What this doesn't establish

This meta-analysis pools observational, retrospective cohort data — not randomized evidence, which would be ethically difficult to generate for a pregnancy-exposure question like this one. The included studies varied in which specific GLP-1 drugs were captured and over what time periods, and the meta-analysis doesn't report how precisely "periconceptional or first-trimester exposure" was defined or measured across the three source studies, which could itself introduce inconsistency. Confounding by indication is a real concern here too: women who continued a GLP-1 drug into early pregnancy may differ systematically — in underlying diabetes severity, obesity, or other health factors — from those who didn't, in ways that could independently affect hypertensive pregnancy risk.

What the authors say is needed

The study's authors state plainly that "this available evidence is limited," and explicitly call for large prospective studies to properly establish whether an association exists [s1] — a direct acknowledgment that this meta-analysis, despite pooling the best currently available evidence, cannot answer the underlying clinical question with any confidence.

What to watch

Whether larger, prospective pregnancy registries or cohort studies, potentially aided by the drugs' rapidly expanding use, generate more definitive data on early-pregnancy GLP-1 exposure and hypertensive pregnancy outcomes. This article is not medical advice; current guidance recommends discontinuing GLP-1 receptor agonists before conception, and any pregnancy-related medication questions should be directed to a treating clinician.

Sources

1. Hypertensive disorders of pregnancy and GLP-1 receptor agonist timing: a systematic review and meta-analysis — Endocrine, 31 July 2026