GLP-1 exposure around conception: no consistent pattern of harm across 22 studies

Design

Systematic review and meta-analysis of cohort and observational studies, plus narrative synthesis of case reports

Population

22 studies; 49,395 human pregnancies with periconceptional GLP-1 receptor agonist exposure

Primary outcome

Fetal, pregnancy, obstetric and delivery outcomes, including malformations

Effect

No significant association for most outcomes; renal malformations OR 1.23 (95% CI 1.09 to 1.39)

A systematic review and meta-analysis pooled 22 studies covering 49,395 human pregnancies with periconceptional exposure to a GLP-1 receptor agonist. The meta-analysis proper used 10 cohort and observational studies with unexposed comparators, and case reports and series were summarised narratively.

There was no statistically significant association with major congenital malformations, cardiac malformations, gestational age or weight, preterm delivery, live birth, spontaneous pregnancy loss, hypertensive disorders or caesarean delivery. One signal appeared: renal malformations, odds ratio 1.23 (95% CI 1.09 to 1.39). The authors attribute it largely to one big cohort with substantial baseline imbalances, and heterogeneity was moderate to high for several outcomes.

The authors frame the findings as cautious reassurance after inadvertent exposure, not support for using these drugs in pregnancy. For a woman who conceives on a GLP-1 agonist, the practical step is to stop it, review the indication and glycaemic plan, and offer counselling and detailed anomaly scanning at the usual points.

• Ask about GLP-1 agonist use at preconception and booking visits, including for weight loss.
• Advise stopping the drug once pregnancy is confirmed and plan alternative glycaemic management where relevant.
• Counsel that current human data do not show a consistent pattern of major harm, but are observational.
• The renal malformation signal is small and likely confounded; it is worth noting at the anomaly scan, not a cause for alarm.
• Document exposure timing and drug so that follow-up data can be linked.

Why it matters

Clinicians are increasingly seeing women who conceive on these drugs, and counselling needs numbers rather than guesses.

Don't overread it

These are observational data and confounding by maternal disease severity is likely; they do not show the drugs are safe in pregnancy.

The statistics, in plain English

An odds ratio of 1.23 with a confidence interval of 1.09 to 1.39 means a statistically detectable but small excess, and it came mainly from one cohort whose groups differed at baseline. Moderate to high heterogeneity means the studies disagreed more than chance would explain, so pooled estimates are less certain than they look.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.