Hemoglobin A1c in early pregnancy to identify preexisting diabetes mellitus and women at risk of hyperglycemic pregnancy complications.

BACKGROUND:Unrecognized diabetes mellitus during pregnancy could pose serious maternal and neonatal complications. A hemoglobin A1c level of ≥6.5% was used to diagnose both diabetes mellitus in nonpregnant individuals and diabetes in pregnancy. As the hemoglobin A1c level could be influenced by maternal physiological changes, the optimal cutoff in early pregnancy to detect women with diabetes in pregnancy and associated complications remains unclear. OBJECTIVE:This study aimed to evaluate the diagnostic performance of various hemoglobin A1c levels and the optimal hemoglobin A1c cutoff to identify mothers with diabetes in pregnancy diagnosed by the gold standard 75 g oral glucose tolerance test before 24 weeks of gestation. In addition, the pregnancy and neonatal outcomes were compared using the optimal hemoglobin A1c cutoff. STUDY DESIGN:A retrospective cohort study was conducted between 2004 and 2019. Women with at least 1 risk factor of gestational diabetes mellitus received an oral glucose tolerance test before 24 weeks of gestation. Terminology of hyperglycemia first detected during pregnancy by oral glucose tolerance test was classified as either diabetes in pregnancy or gestational diabetes mellitus following the World Health Organization's recommendation. Women who met the diagnostic criteria of diabetes in pregnancy and early-onset gestational diabetes mellitus (ie, before 24 weeks of gestation) and had a paired hemoglobin A1c measurement within 4 weeks of their early oral glucose tolerance test were studied. Sensitivity, specificity, and positive and negative predictive values at various hemoglobin A1c cutoffs were calculated for the detection of diabetes in pregnancy. The optimal hemoglobin A1c level was identified from the constructed receiver operating characteristic curves. Multivariate binary logistic regression analyses were performed to calculate the unadjusted and adjusted odds ratios for pregnancy complications. RESULTS:There were 63,111 deliveries, and 22,949 women underwent an oral glucose tolerance test before 24 weeks of gestation. A total of 157 and 3210 women met the diagnostic criteria of diabetes in pregnancy and early-onset gestational diabetes mellitus using an oral glucose tolerance test, respectively. Only 346 participants had a paired hemoglobin A1c and oral glucose tolerance test measurement (82 cases with diabetes in pregnancy and 264 cases with early-onset gestational diabetes mellitus). The receiver operating characteristic curve identified an optimal hemoglobin A1c cutoff of 5.7% to diagnose diabetes in pregnancy, with a sensitivity of 64.6%, specificity of 81.1%, positive predictive value of 51.5%, and negative predictive value of 88.1%. A hemoglobin A1c cutoff of either 5.9% or 6.5% could miss 47.6% or 73.2% of women with diabetes in pregnancy. In multivariate logistic regression analysis, a hemoglobin A1c level of ≥5.7% increased the risk of maternal insulin use (adjusted odds ratio, 6.69; 95% confidence interval, 3.44-12.99), macrosomia (adjusted odds ratio, 7.43; 95% confidence interval, 1.90-29.00), and shoulder dystocia (adjusted odds ratio, 6.56; 95% confidence interval, 1.161-37.03). CONCLUSION:The optimal hemoglobin A1c cutoff to detect diabetes in pregnancy diagnosed using an oral glucose tolerance test before 24 weeks of gestation was 5.7%, but this cutoff could not reliably identify diabetes in pregnancy owing to the low sensitivity. However, an early hemoglobin A1c level of ≥5.7% indicated increased risks of pregnancy and neonatal complications.