Meta-analysis of 49,395 pregnancies finds no clear malformation signal for GLP-1 drugs

Researchers pooling 22 studies covering 49,395 pregnancies reported no statistically significant association between periconceptional exposure to GLP-1 receptor agonists and major congenital malformations, according to a systematic review and meta-analysis published in the American Journal of Obstetrics & Gynecology ^[1]. "Periconceptional" means exposure around the time of conception, including the weeks just before and after a pregnancy begins.

The analysis also found no significant association with cardiac malformations, preterm delivery, live birth, miscarriage, hypertensive disorders of pregnancy, or cesarean delivery ^[1]. Those outcomes cover most of the questions that come up when someone taking a GLP-1 medication for type 2 diabetes or weight loss becomes pregnant unexpectedly.

What the numbers showed

One finding stood out from the otherwise null results: a small association with renal (kidney) malformations, with an odds ratio of 1.23 ^[1]. An odds ratio of 1.0 means no difference between exposed and unexposed pregnancies, so 1.23 suggests roughly 23% higher odds. The authors attributed that signal largely to a single cohort in which the exposed and comparison groups were imbalanced ^[1]. In plain terms, when one study's groups differ in ways the researchers could not fully account for — such as underlying diabetes, obesity severity, or other medications — a difference can show up that is not caused by the drug itself.

Meta-analyses like this one combine observational data rather than randomized trials. Pregnant people are not enrolled in GLP-1 drug trials, so nearly all human safety information comes from registries, insurance claims databases, and health records of people who were taking these medicines when they conceived. That design can detect large risks reasonably well but is less reliable for small ones, and it depends on how completely each source database captures birth defects.

The published summary does not break out results by individual molecule, so it is not clear from what is available here how much of the pooled data reflects semaglutide (Ozempic, Wegovy, Rybelsus) versus other GLP-1 drugs such as liraglutide, dulaglutide or exenatide ^[1]. Details on how long after conception exposure continued in each study are also not specified in the available text.

The tirzepatide gap

Tirzepatide (Mounjaro, Zepbound) has its own dedicated pregnancy registry, and it has not reported results. The I8F-MC-B016 Tirzepatide Pregnancy Registry is listed in the joint HMA-EMA catalogue of real-world data studies with the status "Planned," first published and last updated on January 14, 2026 ^[2].

The registry is designed as a multi-country, prospective, observational cohort study comparing three groups: people treated with tirzepatide for weight management during pregnancy, people treated with other weight-management drugs that do not have GLP-1 receptor agonist activity, and people with obesity or overweight plus at least one weight-related condition who take no weight-management drug during pregnancy ^[2]. Its primary objective is to compare the overall prevalence of major congenital malformations across those three cohorts ^[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].

Why it matters for patients

GLP-1 medications are widely used by adults in their reproductive years, and unplanned pregnancies happen. Until now, the human data on what exposure around conception means for a baby has been thin and scattered across small studies. Pooling nearly 50,000 pregnancies gives clinicians and patients a larger body of evidence to discuss, and the headline result is the absence of a clear malformation signal ^[1].

That is not the same as proof of safety. A null result in observational data means no risk was detected at the size and quality of the studies included, not that no risk exists. The renal malformation finding shows how a single imbalanced dataset can move a pooled estimate ^[1]. And because tirzepatide's own registry has not reported, there is no molecule-specific answer yet for the two tirzepatide brands ^[2].

What happens next

The tirzepatide registry remains in planned status as of its January 14, 2026 catalogue entry, and no results timeline appears in that listing ^[2]. Anyone weighing pregnancy plans alongside a GLP-1 prescription would need to take questions about timing, contraception and stopping or continuing treatment to their own clinician; those decisions are outside what these two sources address.