Metformin did not prevent gestational diabetes in a pooled analysis of 2,297 pregnancies

Metformin taken during pregnancy did not reduce the odds of developing gestational diabetes, in a pooled analysis of individual data from seven randomised trials published in NEJM Evidence on August 25 [s1]. The drug did have measurable effects elsewhere — a slightly longer pregnancy, fewer preterm births — but the headline hypothesis that it prevents gestational diabetes was not supported [s1].

That is a useful negative result in a field where metformin is sometimes proposed as a low-cost way to head off a common complication of pregnancy.

The question

Gestational diabetes — high blood glucose first recognised in pregnancy — raises the risk of a large baby, delivery complications and later type 2 diabetes. Metformin lowers blood glucose and is widely used; the open question is whether giving it pre-emptively to women at high metabolic risk, before diabetes develops, prevents the condition and its downstream harms [s1].

Individual trials have been too small and too varied to settle that. The analysis pooled the raw participant-level data rather than published summaries, which allows consistent adjustment and a cleaner answer than combining headline figures.

What was pooled

The authors searched three databases through May 9, 2025 for double-blind, placebo-controlled randomised trials of metformin in pregnancies without pre-existing diabetes [s1]. Ten trials (N=2,695) met the inclusion criteria; seven provided individual participant data (n=2,485, 92.2% of the available data) [s1]. After harmonisation, 2,297 pregnancies were analysed — 1,159 randomised to metformin and 1,138 to placebo — in mixed-effects models adjusted for maternal age, body-mass index, gestational age at the start of treatment and baseline blood glucose [s1].

The primary result: no effect on gestational diabetes

Metformin was not associated with a reduction in gestational diabetes on the main definitions [s1]:

• World Health Organization 1999 criteria: odds ratio 1.04 (95% CI, 0.80 to 1.36); adjusted odds ratio 1.00 (95% CI, 0.71 to 1.41)
• NICE 2015 criteria: odds ratio 0.98 (95% CI, 0.76 to 1.27); adjusted odds ratio 1.00 (95% CI, 0.71 to 1.41)

The one exception was a marginal signal on the International Association of Diabetes and Pregnancy Study Groups thresholds, where the adjusted odds ratio was 0.71 (95% CI, 0.52 to 0.98) [s1]. A single borderline result across several definitions is the kind of finding that is easy to over-read; treated on its own it would not support a claim of prevention.

Metformin was associated with a marginally lower fasting blood-glucose level — a mean difference of −0.06 mmol/L (95% CI, −0.10 to −0.01) — with no apparent difference in the two-hour post-load value [s1].

The secondary findings that did move

Where metformin did show effects was downstream of glucose. It was associated with [s1]:

• Longer gestation: a mean difference of 0.30 weeks (95% CI, 0.06 to 0.54)
• Lower odds of preterm birth: adjusted odds ratio 0.64 (95% CI, 0.47 to 0.89)
• Larger neonatal head circumference: a mean difference of 2.43 percentile points (95% CI, 0.13 to 4.72)

Gastrointestinal side effects were more commonly reported with metformin [s1]. The editorialists, Gardner and Meek, frame the pattern in their title as a possible role "beyond gestational diabetes prevention" — that is, the value, if any, may lie in the pregnancy-duration findings rather than in stopping diabetes [s2].

How to read it

The primary endpoint was negative, and that is the main message. Across the two standard definitions, adjusted and unadjusted, the odds ratio sat at or near 1.00 [s1]. Metformin did not prevent gestational diabetes in this population.

The secondary findings are associations, not proof of benefit. A longer gestation and fewer preterm births are clinically meaningful directions, but they are secondary outcomes in an analysis whose primary question was negative, and they carry their own uncertainty. The head-circumference difference is small and of unclear significance.

This concerns women without diabetes. The analysis is about prevention in higher-risk pregnancies, not about treating established gestational or type 2 diabetes, where metformin's role is different and better established.

Pooled data inherit their trials' differences. Definitions of gestational diabetes, timing of treatment and populations varied across the seven contributing trials, which is part of why the individual-participant approach was used — but it does not erase that heterogeneity.

The site's earlier coverage of lifestyle interventions to prevent gestational diabetes examined a parallel question with its own mixed answer. A drug that lowers glucose but does not prevent the diagnosis is a reminder that surrogate effects and clinical prevention are not the same thing.

What to watch

Whether the preterm-birth and gestation findings are confirmed in trials designed to test them directly, and whether any subgroup — defined by baseline glucose or body-mass index — derives a prevention benefit that the pooled average hides.

This article describes research findings for informational purposes only. It is not medical advice and not a recommendation about any medication in pregnancy.

Sources

• [s1] Mousa A, Løvvik TS, Carlsen SM, et al. Metformin in High Metabolic Risk Pregnancies — A Patient-Level Meta-Analysis. NEJM Evidence, published online 2026-08-25.
• [s2] Gardner B, Meek CL. Metformin in Pregnancy — A Role beyond Gestational Diabetes Prevention? NEJM Evidence, published online 2026-08-25.