Author(s): Esakoff TF, Cheng YW, Caughey AB
Abstract Share this page
Abstract OBJECTIVE: To examine whether screening guidelines for gestational diabetes should be modified based on ethnicity. STUDY DESIGN: This is a retrospective cohort study of 14,565 pregnancies screened for gestational diabetes. The primary outcome of interest was the diagnosis of gestational diabetes based on the results of the 100-g glucose tolerance test. A diagnosis was determined if any 2 of the 4 values (fasting or 1-, 2-, or 3-hour postprandial) were at or above 95, 180, 155, or 140 mg/dL, respectively. Values of the 50-g glucose-loading test were examined from 130 mg/dL to 150 mg/dL. The sensitivity, specificity, false-positive rates, and positive predictive values of various glucose-loading test cut-offs were compared among whites, African Americans, Latinas, and Asians. The sensitivities and specificities were compared using receiver-operator characteristic curves. RESULTS: Sensitivity and specificity of the glucose-loading test differ by ethnicity. Generally, African Americans achieved the highest sensitivities and specificities over the range examined. To achieve a false-positive rate of 10\% for the glucose-loading test, the threshold value would be 133 mg/dL for African Americans, 140 mg/dL for whites, 143 mg/dL for Latinas, and 147 mg/dL for Asians. At the lower glucose-loading test values (130 and 135 mg/dL), Asians exhibited the highest positive predictive values (30.0\% and 34.0\%), whereas at higher values of the glucose-loading test (145 and 150 mg/dL), African Americans had the highest positive predictive values (46.0\% and 51.5\%). Throughout the glucose-loading test range, whites had the lowest positive predictive values (P < .001). CONCLUSIONS: To maximize the sensitivity and minimize the false-positive rate of the glucose-loading test, it may be reasonable to consider varying the threshold based on ethnicity. However, modification of glucose-loading test thresholds based on maternal ethnicity merit further study to determine whether improved perinatal outcomes can be achieved.
This article was published in Am J Obstet Gynecol
and referenced in Journal of Biometrics & Biostatistics