FRI615 The 1-Hour OGTT Plasma Glucose As A Biomarker Of β-cell Function

Abstract Disclosure: B.T. Legvold: None. A. Zhang: None. K.M. Utzschneider: None. M.K. Rhee: None. L.R. Staimez: None. K.A. Easley: None. M.M. Van Greevenbroek: None. C.J. Van Der Kallen: None. C.G. Schalkwijk: None. C.D. Stehouwer: None. L.S. Phillips: None. Development of diabetes (DM) is attributed to reduced β-cell function, but we lack a measure of β-cell function that is a good predictor of progression to diabetes, and sufficiently convenient and inexpensive for use in large studies. To determine if the 1-hour plasma glucose in a 75g oral glucose tolerance test (1hrOGTT) might serve this purpose, we examined (i) correlations with insulin secretion (insSEC) vs. action (insACT), (ii) prediction of progression from nonDM to DM, and (iii) sample size needed for a prospective trial. The Cohort study of Diabetes and Atherosclerosis Maastricht (CoDAM) included 362 participants with nonDM at baseline who were restudied after 7 years. Incident DM (n=54) was based on American Diabetes Association OGTT cutpoints; insACT included 1/[fasting ins] (I), HOMA2-S (H), and the modified Matsuda Index (M); and insSEC as OGTT Δins/Δglucose0-30min (ISI) and ΔCpep/Δglucose0-30 (CpSI). Since β-cell function reflects insulin secretion relative to insulin action in each individual, we also assessed the disposition index (DI) as insSEC*insACT. (i) Spearman correlations: The 1hrOGTT was correlated with both the 0hrOGTT (0.54) and 2hrOGTT (0.64); weakly with insACT (I, H, and M as 0.31, 0.33, and 0.53, respectively); more strongly with insSEC (ISI and CpSI as 0.54 and 0.67, respectively); and most strongly with DI (ISI*M, CpSI*M, ISI*I, and CpSI*I as −0.89, −0.83, −0.77, and −0.73, respectively); correlations with DI including H were all weaker. (ii) Prediction of progression from nonDM to DM: The area under the receiver operating characteristic curve (ROC) for the 1hrOGTT was 0.84±0.02 (SE), while the ROCs with DI ISI*M, CpSI*M, ISI*I, CpSI*I, ISI*H, and CpSI*H were 0.86, 0.84, 0.85, 0.84, 0.85, and 0.85, respectively. There was no difference between the ROC with the 1hrOGTT and that of the DI models (all p >0.15), while the ROC with the 1hrOGTT was significantly higher than that with most other metrics. (iii) Sample size needed for use in a trial: A side-by-side, two-arm prospective study, with α=0.05 and 80% power to detect a 20% change in β-cell function, would require 26 subjects with the 1hrOGTT, vs. 228 subjects with DI ISI*M. The lower sample size needed reflects lower underlying variance: the coefficient of variation (CV) was 26.7% with the 1hrOGTT, while the CVs for DI ISI*M, CpSI*M, ISI*I, CpSI*I, ISI*H, and CpSI*H were 176.7%, 180.0%, 150.0%, 155.4%, 107.2%, and 90.0%, respectively. Conclusion: The 1-hour 75g OGTT plasma glucose is strongly correlated with the glucose disposition index, and comparable to the disposition index in predicting progression of nonDM to DM, but would require fewer subjects for statistical power in a prospective trial. Since the 1-hour OGTT glucose appears to provide performance similar or superior to more complex and labor-intensive methods, at a greatly reduced cost, consideration should be given to its use as a measure of β-cell function. Presentation: Friday, June 16, 2023