Distinguishing early-onset GDM from undiagnosed pregestational DM?

Takeaway

• Most women diagnosed with diabetes prior to 24 gestational weeks had gestational diabetes mellitus (GDM) that resolved after delivery.
• HbA1c was superior to oral glucose tolerance testing in distinguishing undiagnosed pregestational diabetes from GDM at <24 weeks of gestation.
• Definitive diagnosis should be done after delivery.

Why this matters

• Pregnancies complicated by pregestational diabetes are associated with more foetal and maternal complications than are GDM pregnancies.

Study design

• Population-based retrospective cohort study of 285 women with diabetes diagnosed during pregnancy at <24 weeks who delivered during 2005-2018.
• Funding: None.

Key results

• 11 (5%) women were diagnosed with diabetes within 6-12 weeks postdelivery.
• The most accurate marker for type 2 diabetes (T2D) was HbA1c, with a cutoff of 5.8% predicting T2D with:
  • 89% sensitivity,
  • 86% specificity,
  • 99.4% negative predictive value (NPV), and
  • 23% positive predictive value (PPV).
• A cutoff of 6.5% predicted T2D with:
  • 56% sensitivity,
  • 98.5% specificity,
  • 98% NPV, and
  • 62% PPV.
• Using HbA1c <6.5% as the cut-off will detect 73% of women with T2D vs 91% using 5.8%; i.e., a 6.5% cutoff will miss 18% of women with T2D.
• HbA1c was the only independent predictor for T2D: OR, 19 (95% CI, 3.2-109).

Limitations

• Retrospective design.
• T2D evaluated after pregnancy.