Universal early pregnancy glycosylated hemoglobin A1c as an adjunct to Carpenter-Coustan screening: an observational cohort study

Abstract

Early pregnancy screening for preexisting and gestational diabetes mellitus is widely recommended, but the details of screening (eg, targeted vs universal screening, criteria to identify women requiring early screening, specific screening strategy) remain controversial and poorly defined. The objective of the study was to determine the utility of universal early glycosylated hemoglobin A1c obtained at the first prenatal visit in diagnosing preexisting diabetes and high-risk gestational diabetes mellitus (early glycosylated hemoglobin A1c, 5.9-6.4%). We further sought to determine whether early glycosylated hemoglobin A1c could replace routine Carpenter-Coustan testing and to determine the correlation between early glycosylated hemoglobin A1c and maternal and neonatal morbidity and mortality. This was an observational cohort study of women delivering from May 2016 to July 2017 (14 months) at a single county teaching hospital. Multiple gestations and second deliveries during the study interval were excluded. Women with an early glycosylated hemoglobin A1c of ≥ 6.5% were diagnosed with preexisting diabetes. Women with early glycosylated hemoglobin A1c of 5.9-6.4% underwent immediate 3 hour glucose tolerance testing, which if abnormal diagnosed gestational diabetes mellitus and if normal was repeated at 24-28 weeks. Women with early glycosylated hemoglobin A1c <5.9% underwent routine Carpenter-Coustan screening at 24-28 weeks. Receiver-operator curve methodology was used to evaluate the diagnostic properties of early glycosylated hemoglobin A1c for gestational diabetes mellitus. The correlation between early glycosylated hemoglobin A1c and composite measures of maternal and neonatal morbidity and mortality were calculated. A total of 4144 deliveries remained after exclusions. Median gestational age at early glycosylated hemoglobin A1c draw was 9 weeks (interquartile range, 7-12). Early glycosylated hemoglobin A1c diagnosed 26 women with preexisting diabetes (0.8% of all patients, 37.7% of all preexisting diabetes). A total of 41.9% of 93 women with early glycosylated hemoglobin A1c of 5.9-6.4% had an early diagnosis of gestational diabetes mellitus, accounting for 25.8% of total gestational diabetes mellitus cases. Based on receiver-operator curve analysis, no early glycosylated hemoglobin A1c cutoff had sufficient sensitivity and positive predictive value to diagnose gestational diabetes mellitus. An early glycosylated hemoglobin A1c ≤5.0% (29.2% of patients) had a 98% negative predictive value for gestational diabetes mellitus, suggesting women with an early glycosylated hemoglobin A1c ≤5.0% in a similar-risk population could potentially forego further testing. The per-patient incremental cost for the glycosylated hemoglobin A1c was $3.72. Early glycosylated hemoglobin A1c correlates with maternal and neonatal morbidity and mortality but cannot entirely replace routine Carpenter-Coustan testing because of poor sensitivity. Rather, its use as an adjunct to Carpenter-Coustan testing, with reflex to early 3 hour glucose tolerance testing for those with values 5.9-6.4%, is an inexpensive and simple method that identifies women with preexisting diabetes and high-risk gestational diabetes mellitus early in pregnancy, allowing early intervention and the prospect of improved outcomes.