Search

Objective: To determine the reproducibility of the oral glucose tolerance test (OGTT). Design: A retrospective study of 205 women who underwent screening for gestational diabetes. Setting: A university teaching hospital in a South African city. Sample: Women who had an abnormal OGTT had the test repeated during the same pregnancy without any diabetic or dietary advice given in the intervening time period. Methods: Women with two OGTTs in the same pregnancy had a proforma completed at the time. Completed proformas were filed and reviewed for the purpose of this study. Main outcome measure: The κ statistic was used for estimating the agreement between repeated tests using the same nominal or dichotomous scale. Results: The OGTT was repeated during the index pregnancy in 205 women and in a subset of 76 women within 17 days. The κ statistic was 0.269 for 205 women and 0.212 for 76 women for the fasting glucose value (fair strength of agreement). The κ statistic for the 2-h glucose value was 0.157 for 205 patients and 0.174 for 76 patients (slight strength of agreement). The overall OGGT classification produced κ statistics of 0.167 and 0.150 for the whole group and the 76 patients, respectively. Conclusion: The reproducibility was better with the fasting glucose and less with the 2-h result and the overall OGGT classification. Caution needs to be exercised when interpreting the single positive result of an OGTT in pregnant women. [ABSTRACT FROM AUTHOR]