Efficiency of lower threshold criteria for the diagnosis of gestational diabetes

To determine the incidence of adverse outcome in normal untreated gravidas with minimal hyperglycemia, classified as having gestational diabetes mellitus (GDM) by threshold criteria lower than current standards; to determine how efficient the different criteria are in identifying infants at risk for morbidity; and to explore the pathophysiology of minimal hyperglycemia using the glucose tolerance test (GTT) periodicity concept. Seven hundred eight subjects considered nondiabetic by current ACOG criteria were reclassified by the criteria of Coustan (fasting 95, 1 hour 180, 2 hours 155, and 5 hours 140 mg/dL), Sacks (96,172,152, and 131 mg/dL), or Langer (at least one abnormal ACOG value). Glucose tolerance test periodicity, the incidence of large for gestational age (LGA) neonates, and macrosomia were then determined for each gravida diagnosed as having GDM by these criteria. Both Coustan and Langer criteria identified a significantly greater incidence of LGA infants compared with non-GDM (23.6 and 25.3%, respectively, versus 14%, P < .05), and identified them as efficiently as current criteria, approximately one LGA infant for every four GDM subjects treated. The incidence of LGA did not differ between the Sacks GDM and non-GDM groups. Glucose tolerance test periodicity for newly diagnosed GDM gravidas was significantly longer than non-GDM for Coustan and Langer criteria (3.9 and 4.06 versus 3.3 hours, P < .01) but not for the Sacks criteria. Using lower threshold criteria to diagnose GDM identified morbidity at an incidence and efficiency comparable to current standards. These newly diagnosed GDM gravidas had abnormal GTT characteristics, with each group exceeding the 3.5-hour GTT periodicity limit previously found for nondiabetic gravidas. Sacks's conversion of existing standards may be too low to efficiently identify pregnant subjects at risk for increased morbidity.