- Design
- Systematic review and meta-analysis of observational cohorts
- Population
- 3 cohorts, 669 SGLT2 inhibitor-exposed pregnancies in women with type 2 diabetes
- Primary outcome
- Major congenital malformations versus active comparators
- Effect
- RR 0.94 (95% CI 0.68-1.29) vs insulin; cardiac RR 0.97 (0.64-1.48)
A systematic review pooled three observational cohorts comparing women with type 2 diabetes exposed to an SGLT2 inhibitor around conception or in early pregnancy against women on other glucose-lowering drugs. In all, 669 pregnancies were exposed. The analysis was published in Endocrine on 17 October 2026.
Against insulin, the prespecified comparator, exposure was not associated with major congenital malformations (RR 0.94, 95% CI 0.68-1.29) or cardiac malformations (RR 0.97, 0.64-1.48). Adding a cohort with a sulfonylurea comparator gave much the same answer (RR 0.98, 0.74-1.29). One Korean cohort, however, found that exposure specifically in gestational weeks 4-10 was associated with congenital heart defects (RR 2.79, 1.16-7.06).
This matters because these drugs are now prescribed widely to women of reproductive age for kidney and heart protection, and many keep taking them until a period is missed. The authors rated certainty low to very low: few exposed events, live-birth-only data, varying exposure definitions and residual confounding.
In clinic, this supports a calm conversation with a woman who conceived on an SGLT2 inhibitor - it is not a reason to consider termination on malformation grounds alone. It does not support planned use in pregnancy. Switching to insulin when pregnancy is planned or recognised remains the standard approach, and a detailed fetal cardiac scan is reasonable given the unreplicated timing signal.
- Ask every woman of reproductive age on an SGLT2 inhibitor about pregnancy plans and contraception at each review
- If pregnancy is recognised on an SGLT2 inhibitor, stop it and move to insulin as usual
- Counsel that pooled data show no clear rise in major malformations, but the numbers are small
- Consider a detailed fetal echocardiogram where exposure ran into weeks 4-10
- Record the exposure window precisely - it is the detail future counselling will turn on
Why it matters
Many women conceive before they stop these drugs, and until now counselling rested on almost no human data.
Don't overread it
No clear increase is not proof of safety - certainty was low to very low and a cardiac timing signal is unresolved.
The statistics, in plain English
An RR of 0.94 with a 95% CI of 0.68-1.29 means the data fit anything from a 32% lower to a 29% higher malformation risk. The interval crosses 1.0, so there is no detectable difference, but it is too wide to rule out a meaningful increase. The single-cohort cardiac result (RR 2.79) has an interval that stays above 1.0, but one cohort with few events is a signal to check, not a finding.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for obstetrics & gynaecology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free