Influence of Diagnostic Criteria on the Incidence of Gestational Diabetes and Perinatal Morbidity

Abstract

Objective. —To determine the incidence of gestational diabetes and its associated maternal and infant morbidity by two sets of 3-hour glucose tolerance test criteria, those recommended by the National Diabetes Data Group or the lower, modified criteria of Carpenter and Coustan. Design. —Prospective, observational outcome cohort study. Setting. —Prepaid health maintenance organization. Subjects. —A total of 2019 pregnant women preregistered in a health maintenance organization were screened for a plasma glucose of 7.77 mmol/L (140 mg/ dL) or greater, 1 hour after ingestion of a 50-g glucose drink administered after an overnight fast. Positive subjects received a 3-hour glucose tolerance test interpreted by the two criteria. Data are presented for 521 randomly selected negative screenees; 264 positive screen, negative glucose tolerance test subjects; and 101 subjects with gestational diabetes mellitus. Outcome Measures. —Maternal risk factors for diabetes, infant birth weight corrected for gestational age (birth-weight ratio), umbilical cord serum insulin levels, and 33 maternal and infant perinatal morbidities assessed from chart review. Results. —Gestational diabetes incidence was 5.0% overall based on the modified criteria and 3.2% by the recommended criteria. Maternal age and family history of diabetes were higher in both modified and recommended groups. Birth-weight ratio was 1.05 in negative screenees, 1.09 in gestational diabetes overall (P<.05 when only diet-treated subjects were considered), and 1.11 in modified and 1.08 in recommended criteria groups. Cord serum insulin levels in infants of gestational diabetic mothers by both criteria were 40% above those of negative screenees (P<.001). The average percentage incidence of 33 possible perinatal morbidities was 41% higher in gestational diabetic pregnancies by the modified criteria. The cumulative number of morbidities was higher in both modified and recommended criteria groups (P<.01 in both instances). In contrast, subjects having a positive glucose screening test but a negative glucose tolerance test by modified criteria had none of the characteristics of gestational diabetes with the single exception of greater age. Conclusions. —Fifty percent more cases of gestational diabetes were identified using the more inclusive, modified criteria. These cases had as much excess in maternal diabetes risk factors, infant macrosomia, and cord hyperinsulinemia and nearly as much increase in perinatal morbidity as subjects diagnosed by the recommended criteria. The incidence and perinatal impact of gestational diabetes may be greater than previously appreciated. The modified criteria deserve wider verification and use. (JAMA. 1993;269:609-615)