A hospital-based cross-sectional study to develop an estimation formula for 2-h post-challenge plasma glucose for screening impaired glucose tolerance

Abstract

Aims: To create and validate an estimation formula for 2-h post-challenge plasma glucose (2-hPG) as an alternative to oral glucose tolerance test (OGTT) for impaired glucose tolerance (IGT) screening.

Methods: 380 Japanese subjects (57.6% males, aged 58.5 (14.0); mean (SD) years) undergoing OGTT were included in this hospital-based cross-sectional study mainly at Kyoto University Hospital between 2000 and 2011. We determined the main predictive variables of 2-hPG from clinical variables and separated the subjects randomly into two groups: a derivation group to construct an estimation formula of 2-hPG on the basis of predictive variables and a validation group to evaluate the accuracy of the formula.

Results: Fasting plasma glucose (FPG) and hemoglobin A1c (HbA1c) were highly correlated with 2-hPG measured by OGTT. Multiple linear regression analysis showed that estimated 2-hPG (e2-hPG) was calculated by the formula: e2-hPG = 1.66 × FPG (mmol/l) + 1.63 × HbA1c (%)-10.11 (R(2), coefficient of determination=60.2%). When the cut-off value was set to the diagnostic criteria of IGT, 7.8 mmol/l of e2-hPG, sensitivity, specificity, and negative predictive value (NPV) were 83.3%, 44.1%, and 74.3%, respectively. When the cut-off value was set lower (7.2 mmol/l), these values were 94.4%, 30.5%, and 85.7%, respectively. The area under the receiver operating characteristic (ROC) curve was 0.68.

Conclusions: This high-sensitive estimation formula may be a useful alternative to OGTT for IGT screening. For the levels ≤ 7.2 mmol/l, this formula may also be useful in cross-sectional study to identify people whose glucose tolerance is normal.

Keywords: IGT; OGTT; Screening.