Novel OGTT metrics do not provide incremental predictive value beyond conventional glycemic criteria for remission of prediabetes in individuals with impaired fasting glucose: results from the PREVIEW trial

While conventional metrics of glycemia (fasting and 2-h glucose and HbA1c) define prediabetes and are associated with its progression, they do not capture the full heterogeneity of individual glucose metabolism. We aimed to examine whether novel OGTT-derived metrics provide incremental predictive value beyond conventional metrics for prediabetes remission during long-term lifestyle intervention. This post-hoc analysis used longitudinal data from PREVIEW, a 3-year, multinational, randomized trial, which assessed lifestyle interventions for diabetes prevention in adults with overweight/obesity and predominantly impaired fasting glucose (IFG). Participants underwent 8-week weight loss followed by 148-week lifestyle-based weight maintenance. Conventional metrics of glycemia included fasting and 2-h glucose and HbA1c. Novel OGTT metrics included 1-h glucose, time to glucose peak, maximum glucose excursion, peak glucose, area under the curve during OGTT, and curve shape. Outcomes were prediabetes remission at 1, 2, and 3 years. Restricted cubic spline analysis and modified Poisson regression models examined the associations of baseline conventional glycaemic and novel OGTT metrics with remission. Incremental predictive value of individual OGTT metrics was assessed using Likelihood ratio tests and optimism-adjusted area under the receiver operating characteristic curve (AUROC) and net reclassification improvement (NRI) via bootstrapping. 1187 participants (262 remitters and 925 non-remitters at 1 year) were included. There was a non-linear association between baseline fasting glucose and prediabetes remission at 1 year (P = 0.008), while inverse linear associations were observed at 2 and 3 years, independent of 2-h glucose and HbA1c. Higher baseline HbA1c, but not 2-h glucose, was associated with lower likelihood of remission at 1, 2, and 3 years, regardless of fasting and 2-h glucose. After adjustment for conventional metrics, there were no significant associations of baseline novel OGTT metrics with prediabetes remission. Incremental value analysis showed that adding individual novel OGTT metrics did not yield significant improvements in model fit or discrimination (P > 0.05 for Likelihood ratio tests and tests for ΔAUROC and NRI). In long-term lifestyle intervention among individuals who predominantly had IFG, individual novel OGTT metrics do not provide incremental predictive value for prediabetes remission beyond conventional metrics of glycemia. Trial registration Clinical trial reg. NCT01777893, clinicaltrials.gov.