Meta-analysis finds no clear birth defect signal for GLP-1 drugs taken around conception

A meta-analysis published in the American Journal of Obstetrics & Gynecology found no clear signal linking GLP-1 receptor agonist exposure around the time of conception to major birth defects. The review pooled 22 studies covering 49,395 pregnancies ^[1].

Periconceptional refers to the window around conception. For that exposure period, the researchers reported no statistically significant associations with major congenital malformations, cardiac malformations, preterm delivery, live birth, miscarriage, hypertensive disorders of pregnancy or cesarean delivery ^[1].

One outcome did stand out. The analysis found an association with renal, or kidney, malformations, with an odds ratio of 1.23 ^[1]. An odds ratio of 1.23 indicates roughly 23 percent higher odds of the outcome in the exposed group than in the comparison group. The researchers attributed that result largely to a single imbalanced cohort, meaning one study in which the groups being compared differed in ways that could distort the comparison ^[1]. The analysis reported no statistically significant association for the other outcomes it examined ^[1].

The review pools results across the GLP-1 receptor agonist class. Separate estimates for individual drugs in that class, including semaglutide (sold as Ozempic, Wegovy and Rybelsus), have not been reported ^[1].

What the tirzepatide registry will add

Tirzepatide, the drug in Mounjaro and Zepbound, is covered by a separate pregnancy study that has not reported results. The I8F-MC-B016 Tirzepatide Pregnancy Registry is a multi-country, prospective, observational cohort study ^[2]. It was first published in the European Medicines Agency's catalog of real-world data studies on January 14, 2026, and last updated the same day, and it is listed as planned ^[2].

Its primary outcome is the prevalence of major congenital malformations among people treated with tirzepatide for weight management during pregnancy, compared with two groups: people who took a different weight-management medication during pregnancy, and people with obesity or overweight and at least one weight-related condition who took no weight-management product ^[2]. Secondary outcomes include gestational diabetes, pregnancy-induced hypertension, pre-eclampsia, eclampsia, spontaneous abortion, induced abortion, stillbirth, minor congenital malformations, preterm birth, small for gestational age, postnatal growth deficiency up to one year of age, and infant developmental delay up to one year of age ^[2].

The design compares tirzepatide users with an unexposed group that also has obesity or a weight-related condition, an approach intended to separate effects of the drug from effects of the underlying condition ^[2].

Why it matters for patients

• The pooled results cover exposure around conception, not the whole pregnancy. Whether exposure later in pregnancy is linked to the same outcomes has not been established by this analysis ^[1].
• A finding of no statistically significant association is not the same as proof of no risk. The analysis reported no significant links for the outcomes it examined, but it did not report that any specific risk was ruled out ^[1].
• The kidney malformation result shows how much a single skewed study can move a pooled estimate, which is why the researchers did not treat it as a confirmed signal ^[1].
• Evidence on tirzepatide in pregnancy is not in yet. As of January 14, 2026, the registry was listed as planned, with no results reported and no completion date disclosed ^[2].
• The meta-analysis does not address whether people stop or continue a GLP-1 medication once a pregnancy is confirmed, and it does not offer guidance on that decision ^[1].

What happens next

The tirzepatide pregnancy registry is the main outstanding piece of evidence on this question. Its catalog listing showed no results and no completion date as of January 14, 2026 ^[2]. Whether the meta-analysis authors plan to update their pooled estimates as new studies are published has not been reported ^[1].