Short- and long-term reproducibility of the 1-h 50-g glucose challenge test

Abstract

A study was undertaken to assess the short-term (intrapregnancy) and long-term (interpregnancy) variability of the 1-h 50-g oral glucose challenge test (GCT). Two groups of pregnant women had GCTs in consecutive pregnancies, 1 (n = 77) and 2 (n = 43) years apart. Their results were compared with published results for a group (n = 53) who had GCTs on consecutive days. Robust estimates of the mean error variance (sigma2 error = sigma2 within-individual + sigma2 analytical) were calculated on log10-transformed data and were for the three groups 0.003995, 0.002603, and 0.0026249 (mg/dL)2, respectively. There was no significant difference between the group variances, establishing that the short- and long-term reliability of the GCT is comparable. Sigma2 between-individuals was estimated from the GCT values for 2695 pregnant women tested during the same period and was the main component (67.1%) of the total sample variation (sigma2 between-individuals / sigma2 population). Estimates of the population mean, sigma2 between-individuals, and sigma2 error were used to compute the probability that an observed GCT value had a true value equal to or greater than the consensus threshold of 7.8 mmol/L (140 mg/dL).