Obesity and stillbirth run in opposite directions once diabetes is in the picture

Design

retrospective population-based cohort study using national birth and fetal death certificates, piecewise additive mixed models

Population

6923146 singleton live and stillbirths in the United States between 20 and 43 weeks' gestation, 2022 to 2023

Primary outcome

gestational-age-specific stillbirth risk across pre-pregnancy body mass index categories, stratified by pre-pregnancy diabetes

Effect

stillbirth 16.6 vs 4.4 per 1000 with and without pre-pregnancy diabetes; at 31 weeks, BMI 40 vs 20 adjusted HR 0.68 (95% CI 0.54 to 0.85) with diabetes and 1.22 (1.13 to 1.33) without

This nationwide cohort used US live birth and fetal death certificates from 2022 to 2023, covering 6923146 singleton pregnancies between 20 and 43 weeks, and modelled gestational-age-specific stillbirth risk across pre-pregnancy body mass index categories, separately for women with and without pre-pregnancy diabetes.

Overall stillbirth was 4.4 per 1000 total births without pre-pregnancy diabetes and 16.6 per 1000 with it, and rates rose with increasing body mass index. The unexpected finding is in the interaction. At 31 weeks, comparing a body mass index of 40 with 20, the adjusted hazard ratio was 1.22 (95% CI 1.13 to 1.33) in pregnancies without diabetes - and 0.68 (0.54 to 0.85) in pregnancies with it. The associations differed by diabetes status at every gestational week examined. Absolute risk remained highest in women with diabetes and class III obesity.

The two results are not in conflict once absolute and relative risk are held apart. Women with pre-pregnancy diabetes are at high stillbirth risk whatever their weight, so the relative contribution of obesity within that group is small and can even invert; women without diabetes see obesity emerge as an independent gradient. What follows for practice is the authors' own conclusion, which is about timing: the optimal gestational age for delivery in an obese pregnancy depends on what else is present, and a policy keyed to body mass index alone will be wrong for one of these two groups.

• Decide delivery timing on the combination of body mass index and diabetes status, not on body mass index alone.
• Keep absolute risk in front of you: 16.6 per 1000 with pre-pregnancy diabetes against 4.4 without.
• Do not read the hazard ratio of 0.68 as obesity being protective in diabetes - it describes a relative comparison within an already high-risk group.
• This uses birth and fetal death certificate data, so ascertainment of pre-pregnancy diabetes is imperfect and gestational diabetes is not separated out.
• Where pre-pregnancy diabetes is present, the surveillance and timing conversation should start well before term.

The statistics, in plain English

The inversion of the hazard ratio - 1.22 without diabetes and 0.68 with it, at the same gestational week and the same body mass index comparison - is a textbook example of why a relative measure needs its baseline. Within a group whose absolute stillbirth risk is already nearly four times higher, the marginal contribution of weight is small, and residual confounding or differences in surveillance intensity can easily flip its sign. This is registry data from birth and death certificates, where diabetes status is recorded by checkbox and misclassification is well documented, so the interaction is a strong signal about heterogeneity rather than a precise estimate of it.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.