This is a prospective follow-up of 4,693 women in the Hyperglycemia and Adverse Pregnancy Outcome Follow-Up Study, assessed 10 to 14 years after delivery. Gestational diabetes was classified by the underlying physiology rather than by severity: insulin deficient (3.0%), insulin resistant (9.0%), mixed defect (1.6%) and unclassified (0.7%). The main outcome was prediabetes or diabetes, with hypertension, dyslipidaemia, metabolic syndrome and predicted cardiovascular disease as secondary outcomes.
Compared with women who had no gestational diabetes, every subtype except unclassified carried a higher risk of prediabetes or diabetes, with adjusted risk ratios between 2.14 and 2.88. The same subtypes were also associated with metabolic syndrome and higher predicted cardiovascular risk. Only the insulin-resistant subtype was associated with dyslipidaemia. The finding that matters is the one that is not a difference: subtyping did not pick out a group who could be safely let go. The insulin-deficient woman, who is often thinner and easier to reassure at the postnatal visit, carried the same order of risk as the insulin-resistant one.
So the action is the pathway, not the phenotype: previous gestational diabetes -> postpartum glucose check at 6 to 12 weeks -> annual glucose and lipid review -> weight and lifestyle support -> early detection of dysglycaemia. This is where the gap is widest in Indian practice. Gestational diabetes is common here, onset of type 2 diabetes is younger, body mass index is lower at any given level of risk, and postpartum follow-up attendance is poor. A woman with what looked like mild gestational diabetes at 28 is a woman with a two- to three-fold diabetes risk at 40, and she will usually only be seen again if someone books it now.
- Enter the gestational diabetes diagnosis in the woman's own long-term problem list, not only in the obstetric record.
- Book the 6-to-12-week postpartum oral glucose tolerance test before discharge rather than at the postnatal visit.
- Set an annual recall for fasting glucose or HbA1c plus a lipid profile, however mild the pregnancy hyperglycaemia looked.
- Check blood pressure and waist circumference at the same review - metabolic syndrome and raised predicted cardiovascular risk ran across subtypes.
- Ask about plans for another pregnancy; pre-pregnancy glucose testing is easier than a repeat diagnosis at 24 weeks.
The statistics, in plain English
The adjusted risk ratios of 2.14 to 2.88 look like a range of different risks, but they are better read as one shared signal: with only about 140 insulin-deficient and 75 mixed-defect women in the cohort, the confidence intervals around those estimates are wide and overlap each other, so this is roughly doubled to tripled risk across the board rather than a ranking of subtypes. The unclassified group, at 0.7% of the cohort, was too small to show anything at all, and its lack of association should be read as no information rather than as reassurance. Note also that 'predicted cardiovascular disease' is a calculated risk score, not counted heart attacks and strokes. And this is observational: the subtypes were not assigned, so what is shown is that these women differ, not that the subtype caused the later diabetes.
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 diabetes & endocrinology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free