Key result
Heart failure phenotype in T2D and ASCVD linked to ~165% higher risk of major cardiovascular events.
Why the study?
Does sitagliptin have a differential treatment effect on cardiovascular outcomes across different phenotypic clusters of patients with T2DM and ASCVD?
Observational (n=14,671)
Does sitagliptin have a differential treatment effect on cardiovascular outcomes across different phenotypic clusters of patients with T2DM and ASCVD?
Answers are grounded in this paper and cite their sources.
Machine learning identified five distinct clinical phenotypes in patients with T2DM and ASCVD with varying cardiovascular risk, but the cardiovascular safety of sitagliptin remained consistent across all clusters.
Introduction: Phenotypic heterogeneity among patients with type 2 diabetes (T2D) and atherosclerotic cardiovascular disease (ASCVD) is ill defined although likely there are multiple subtypes. Purpose: We aimed to use a data agnostic machine learning algorithm to identify potentially different phenotypes among patients with T2DM and ASCVD. Methods: We used data from the 14,671 patients with T2DM and ASCVD enrolled in TECOS, a cardiovascular (CV) safety outcome trial comparing sitagliptin vs placebo with median 3.0 years follow-up. Hierarchical clustering of 40 potentially prognostic baseline variables was conducted. The primary composite outcome (CV death, nonfatal MI/stroke or unstable angina hospitalization) across these clusters was then assessed using Cox proportional models. We also examined whether a differential treatment effect of sitagliptin across the clusters affected clinical outcomes. Results: Five distinct patient clusters were identified. Cluster I included older men with a high prevalence of prior coronary artery disease. Cluster II primarily included women with non-coronary ASCVD. Cluster III comprised Asian patients with a low BMI. Cluster IV included younger males with a high BMI. Cluster V consisted of patients with heart failure. The primary composite outcome occurred in 11.9%, 10.6%, 8.7%, 11.0%, and 16.7% of patients in clusters I to V respectively. The CV risk for the highest vs lowest risk clusters (cluster V vs III) was statistically significant (HR 2.65; p≤0.001; Figure). No heterogeneity of sitagliptin vs. placebo on CV risk across the clusters was evident (interaction P value = 0.54).
No takes yet. Share an insight, caveat, or question.
Sharma et al. (2018) conducted an observational in Type 2 diabetes and atherosclerotic cardiovascular disease (n=14,671). Highest risk phenotypic cluster (Cluster V: patients with heart failure) vs. Lowest risk phenotypic cluster (Cluster III: Asian patients with low BMI) was evaluated on Composite of CV death, nonfatal MI/stroke or unstable angina hospitalization (HR 2.65, p=≤0.001).
Synapse has enriched 10 related papers. Consider them for comparative context: