Design: Systematic review and meta-analysis of eight observational and randomised studies
Population: 186,598 pregnancies, 47,159 exposed to a GLP-1 receptor agonist before or early in gestation
Primary outcome: Gestational diabetes, preterm birth, pre-eclampsia, hypertensive disorders
Effect: Gestational diabetes OR 0.99 (0.61 to 1.61); preterm birth OR 1.01 (0.76 to 1.33)

With more women of reproductive age taking GLP-1 receptor agonists for diabetes and weight, many conceive while still on treatment. A meta-analysis in Obstetrics & Gynecology pooled eight studies covering 186,598 pregnancies, 47,159 of them exposed before or early in gestation.

Exposure was not associated with gestational diabetes (OR 0.99, 95% CI 0.61 to 1.61), preterm birth (OR 1.01, 0.76 to 1.33), pre-eclampsia (OR 1.05, 0.60 to 1.84) or hypertensive disorders (OR 0.79, 0.34 to 1.83). A leave-one-out sensitivity analysis hinted at less gestational diabetes, which the authors treat as exploratory.

This is reassuring for the woman who discovers she is pregnant while on one of these drugs. It addresses maternal complications only, not fetal anomalies, and it does not change the advice to stop the drug before a planned pregnancy.

Ask every woman of reproductive age on a GLP-1 receptor agonist about pregnancy plans and contraception.
For an unplanned pregnancy on treatment, stop the drug and offer reassurance that pooled data show no excess of maternal complications.
Keep the usual pre-pregnancy advice to stop the drug ahead of a planned conception, as product labels advise.
Do not treat the hinted reduction in gestational diabetes as a reason to continue treatment into pregnancy.

Why it matters

Clinicians counselling a woman exposed in early pregnancy now have pooled maternal-outcome data instead of case series.

The statistics, in plain English

Odds ratios close to 1 with intervals that cross 1 mean no clear difference either way. The wide intervals for pre-eclampsia and hypertensive disorders mean a modest increase or decrease cannot be ruled out. The gestational diabetes signal appeared only after removing one study at a time, which is a weaker form of evidence than the main analysis.