Abstract
CONTEXT: Type 2 diabetes (T2D) remains a significant public health problem, and current approaches to risk reduction fail to adequately prevent T2D in all individuals. OBJECTIVE: The purpose of this study was to apply clustering methods that include metabolic risk factors and body composition measures to identify and characterize prediabetes phenotypes and their relationships with treatment arm and incident T2D. DESIGN: Secondary analysis of the Diabetes Prevention Program clinical trial. SETTING: Previously completed Diabetes Prevention Program trial. PATIENTS OR OTHER PARTICIPANTS: Subset of participants (n = 994) with body composition measures. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Unsupervised k-means clustering analysis was applied to derive the optimal number of clusters of participants based on common clinical risk factors alone or common risk factors plus more comprehensive measures of glucose tolerance and body composition. RESULTS: Five clusters were derived from both the common clinical characteristics and the addition of comprehensive measures of glucose tolerance and body composition. Within each modeling approach, participants showed significantly different levels of individual risk factors. The clinical only model showed higher accuracy for time to T2D; however, the more comprehensive models further differentiated an overweight phenotype by overall metabolic health. For both models, the greatest differentiation in determining time to T2D was in the metformin arm of the trial. CONCLUSION: Data-driven clustering of patients with prediabetes allows for identification of prediabetes phenotypes at greater risk for disease progression and responses to risk reduction interventions. Further investigation into phenotypic differences in treatment response could enable better personalization of prediabetes and T2D prevention and treatment choices.