Can overt diabetes mellitus be predicted by an early A1C value in gestational diabetics?

OBJECTIVE: To test the hypothesis that a hemoglobin A1C value (A1C) in early pregnancy is predictive of overt diabetes mellitus (DM) postpartum in women with gestational diabetes (GDM). STUDY DESIGN: In this case-control analysis of women with an early pregnancy diagnosis of GDM, we estimated the association between an early pregnancy A1C and subsequent diagnosis of DM. Women with a normal postpartum diabetic screen (controls) were compared against those with confirmed postpartum DM (cases). Ability of A1C levels to predict DM was examined via logistic regression analysis and corresponding receiver operating characteristic values. RESULTS: During the 10-year study period 166 women met the inclusion criteria: 140 (84%) had normal postpartum testing (controls), and 26 (16%) were diagnosed with DM (cases). The mean A1C value was significantly higher among cases than controls (6.7 vs. 5.6, p< 0.0001, SD 1.3–5). Cases had A1Cs ranging from 5.5– 11.7%, while controls had A1Cs ranging from 4.3–7.8%. The best discriminatory cut point for postpartum DM was an A1C > 5.9% (sensitivity 81%, specificity 83%, positive predictive value 47%, negative predictive value 96%). CONCLUSION: Our findings suggest that an elevated early pregnancy A1C may be predictive of overt DM. Larger studies are needed to further validate this association.