Disease progression and treatment response in data-driven subgroups of type 2 diabetes compared to models based on simple clinical features: an evaluation using clinical trial data

Abstract

Background Recent research using data-driven cluster analysis has proposed five subgroups of diabetes with differences in diabetes progression and risk of complications. We aimed to compare the clinical utility of this subgroup-based approach for predicting patient outcomes with an alternative strategy of developing models for each outcome using simple patient characteristics. Methods We identified clusters in the ADOPT (n=4,351) trial cohort using the cluster analysis reported by Ahlqvist and colleagues (Lancet Diabetes Endocrinology 2018;6:361-69). Differences between clusters in glycaemic and renal progression were evaluated, and contrasted with stratification using simple continuous clinical features (respectively, age at diagnosis and baseline renal function). We tested the performance of a strategy of selecting glucose-lowering therapy using clusters with one combining simple clinical features (sex, BMI, age at diagnosis, baseline HbA1c) in an independent trial (RECORD (n=4,447)). Findings Clusters identified in trial data were similar to those described in the original study. Clusters showed differences in glycaemic progression, but a model with age at diagnosis alone explained a similar amount of variation in progression. We found differences in CKD incidence between clusters however baseline eGFR was a better predictor of time to CKD. Clusters differed in glycaemic response, with a particular benefit for cluster 3 (insulin-resistant) with thiazolidinediones and cluster 5 (older) with sulfonylureas. However simple clinical features outperformed clusters to select therapy for individual patients. Interpretation The proposed data-driven clusters differ in diabetes progression and treatment response, but models based on simple continuous clinical features are more useful to stratify patients. This suggests precision medicine in type 2 diabetes is likely to have most clinical utility if based on an approach of using specific phenotypic measures to predict specific outcomes, rather than assigning patients into subgroups.