- Design
- Systematic review and meta-analysis of six cohort studies
- Population
- 2,166 women with normal mid-pregnancy glucose screening retested after 28 weeks for suspected large fetus or polyhydramnios
- Primary outcome
- Pooled detection rate of late-onset gestational diabetes and perinatal outcomes
- Effect
- Detection 15.0% (95% CI 9.9 to 21.0%); neonatal hypoglycaemia OR 1.82 (1.18 to 2.81); caesarean OR 2.00 (1.47 to 2.72)
A systematic review pooled six cohort studies with 2,166 women who had a normal mid-pregnancy glucose screen and a repeat oral glucose tolerance test after 28 weeks because scan findings suggested a large-for-gestational-age fetus or polyhydramnios. Quality was assessed with the Newcastle-Ottawa Scale.
The pooled detection rate of gestational diabetes was 15.0% (95% CI 9.9 to 21.0%), with high heterogeneity (I² 91%). It was 20.8% (17.4 to 24.6%) when a large-for-gestational-age fetus was suspected and 4.8% (2.0 to 10.8%) for isolated polyhydramnios.
Women with late-onset gestational diabetes had higher odds of neonatal hypoglycaemia (OR 1.82, 1.18 to 2.81), caesarean delivery overall (OR 2.00, 1.47 to 2.72), elective caesarean for suspected macrosomia (OR 3.37, 2.01 to 5.64), emergency caesarean (OR 1.64, 1.05 to 2.57) and induction (OR 2.27, 1.32 to 3.89). Birthweight-defined macrosomia and large-for-gestational-age at delivery were not significantly raised. The high elective caesarean odds may partly reflect how clinicians respond to a diagnosis, not only biology.
- Consider a repeat oral glucose tolerance test after 28 weeks when scan suggests a large-for-gestational-age fetus after a normal screen.
- Expect a lower yield with isolated polyhydramnios; look for other causes alongside testing.
- If late-onset diabetes is found, plan neonatal glucose monitoring.
- Counsel carefully so a late diagnosis does not by itself lead to an elective caesarean for presumed macrosomia.
- Follow your local guideline; this evidence does not change recommendations on routine screening.
Why it matters
A normal early screen is not the end of the question once growth runs ahead, and a late diagnosis alters delivery planning.
Don't overread it
Six observational studies with marked heterogeneity do not show that retesting improves outcomes, and routine universal third-trimester testing was not studied.
The statistics, in plain English
An I² of 91% means the studies disagreed far more than chance would explain, so 15% is an average rather than a number to expect in any one unit. Odds ratios above 1 whose confidence intervals exclude 1 indicate a real association in these data, but the studies were observational and the groups differed in more than glucose.
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