Women who were taking a GLP-1 weight-loss or diabetes drug around the time they conceived did not go on to have more pregnancy complications than women who were not. Across nearly 187,000 pregnancies, the drugs did not move the odds of gestational diabetes, premature birth, preeclampsia, or high blood pressure in pregnancy in any direction that reached statistical significance.
That is the pooled result of a meta-analysis published July 2 in Obstetrics and Gynecology ↗, and it addresses a question that has become common precisely because the drugs work. GLP-1 receptor agonists include semaglutide ↗, sold as Ozempic and Wegovy, along with liraglutide (Victoza, Saxenda) and dulaglutide (Trulicity). All carry a label telling patients to stop before pregnancy, because there is almost no trial data in pregnant women. Drug companies excluded them from the studies. But the drugs improve fertility in some women who had trouble conceiving, and they can blunt the reliability of oral birth control. So unplanned exposures in the first weeks of pregnancy keep happening, often before a woman knows she is pregnant.
What the pooled data show
The authors combined eight studies covering 186,598 pregnancies, of which 47,159 involved exposure to a GLP-1 drug before or during early pregnancy. None of the four maternal outcomes they pooled showed a significant difference. Gestational diabetes came out at an odds ratio of 0.99, essentially no change. Preterm birth was 1.01, preeclampsia 1.05, and hypertensive disorders of pregnancy 0.79, each with a confidence interval wide enough to cross the no-difference line. An odds ratio near 1.0 means the exposed and unexposed groups had roughly the same rate; the wide ranges mean the data cannot rule out modest effects in either direction.
One exploratory result is worth flagging without overselling it. When the authors repeated the gestational diabetes analysis while dropping one study at a time, a standard check for whether a single dataset is driving the result, the pooled odds ratio drifted to 0.81 (confidence interval 0.67 to 0.98), which would suggest the drugs slightly lower the risk of gestational diabetes. That is plausible on the biology, since these are diabetes drugs, but it is a hypothesis pulled from a sensitivity test, not a headline finding. The authors label it hypothesis-generating and so should anyone reading it.
The caveats that matter
This is reassurance with limits. The eight studies varied in quality on the Newcastle-Ottawa scale, a standard scoring tool for observational research. Statistical heterogeneity was high, driven by differences in how each study defined exposure, chose its comparison group, and adjusted for confounders. Pooling messy studies produces a number, but the number carries the mess. The analysis also covers maternal complications, not the fetal and newborn outcomes that a prospective parent would ask about next, and it captures exposure around conception rather than sustained use through pregnancy, which no one is recommending.
What it does is narrow the uncertainty at the front end. The realistic clinical situation is a woman on a GLP-1 drug who discovers she is pregnant and stops, and worries about what the early exposure did. This meta-analysis, the largest look at that scenario so far, finds no signal of raised maternal risk. GLP-1 receptor agonists remain the most-tracked drug class ↗ in the metabolic space, and the accidental-exposure question is one the trials were designed to avoid answering. Real-world data is filling the gap the trials left, and for maternal complications the early news is quiet.