Two phenotypes of metabolic cardiomyopathy: Data-driven cluster analysis of non-invasively determined cardiovascular phenotypical traits

OBJECTIVE: To evaluate if metabolic cardiomyopathy comprises different phenotypes. METHODS: Data-driven cluster analysis with hierarchical clustering followed by gap and silhouette width analysis and clustering by k-means was conducted on cardiovascular variables in a cohort of 192 patients with type 2 diabetes (T2D). RESULTS: Two distinct clusters were identified. 59 % of patients with DM2 had small left ventricles and atria with normal mass, high heart rate with small stroke volumes, and only moderately impaired myocardial perfusion reserve (3.32 ± 1.20 vs. 5.07 ± 1.51 ml/min/g in normal controls, p < 0.01). With equal e/e and extracellular volumes, other 41 % of patients with T2D had eccentric hypertrophic large left ventricles with dilated left atria, high stroke volumes, and a mean 6 mmHg higher mean arterial blood pressure and mean 15 mmHg higher pulse pressure, lower heart rates and lower myocardial perfusion reserve (2.60 ± 0.87 mL/min/g, P < 0.01 vs. other groups). CONCLUSIONS: Two distinct types of cardiomyopathies were identified. The majority exhibited small left hearts with only moderate impairment in myocardial perfusion ratio. However, when stiff conductance arteries were present, cardiomyopathy with larger and eccentrically hypertrophic left ventricles and markedly reduced myocardial perfusion ratio was observed. Future research should seek to determine if these two phenotypes carry independent prognostic implications and should be treated differently.