Kahkoska AR, Geybels MS, Klein KR, et al. - Given that unsupervised clustering methods and data from Swedish individuals with recent‐onset diabetes were used to identify populations at differential risk of complications, researchers validated these clusters in three global populations with long‐standing type 2 diabetes (T2D) at high cardiovascular risk and tested for differences in the risk of major diabetes complications and survival endpoints. Participants were assigned from recent global outcomes trials (DEVOTE [n = 7,637], LEADER [n = 9,340] and SUSTAIN‐6 [n = 3,297]) to the previously defined clusters according to age at diabetes diagnosis, baseline glycated haemoglobin and BMI. Using Kaplan–Meier analysis and log‐rank tests, outcomes were assessed. Findings suggested that previously identified clusters can be replicated in three long-standing T2D geographically diverse cohorts and are correlated with cluster-specific risk profiles for additional clinical and survival outcomes to further validate the clustering methodology. External validity and stability of clusters across cohorts provides a basis for future work on optimizing the clustering approach to yielding maximum predictive T2D subgroups that may benefit from subtype-specific treatment paradigms.