Effect of cardiac resynchronization therapy in patients with diabetes randomized in EchoCRT

Aims As patients with heart failure (HF) and concomitant diabetes carry a poor prognosis, this post-hoc subgroup analysis aimed to compare the outcomes of patients with and without diabetes randomized in the Echocardiography Guided Cardiac Resynchronization Therapy (EchoCRT) study. Methods and results EchoCRT randomized patients with a QRS duration <130 ms and echocardiographic evidence of left ventricular dyssynchrony to CRT turned on (CRT=ON) vs. off (CRT=OFF) following device implantation. At study entry, 328 patients (40.5%) had diabetes. The primary outcome (all-cause death or first hospitalization for worsening HF) occurred more frequently in patients with than without diabetes (32.6% vs. 23%, P = 0.003). A significant treatment interaction was observed for the primary outcome indicating a higher risk for CRT=ON vs. CRT-OFF in patients without [26.5% vs. 19.8%, hazard ratio (HR) 1.58, 95% confidence interval (CI) 1.08–2.31] vs. with diabetes (31.4% vs. 34%; HR 0.86, 95% CI 0.58–1.27; P for interaction 0.041). This effect was mainly driven by a lower rate in HF hospitalizations, but was only of borderline significance after multivariate adjustment (P = 0.063). The most pronounced effect was observed in patients with non-ischaemic cardiomyopathy, where a significantly reduced risk of reaching the primary endpoint for CRT=ON vs. CRT-OFF was observed in patients with (HR 0.27, P = 0.003) vs. patients without diabetes (HR 1.79, P = 0.038; P for interaction 0.005). No treatment interaction by diabetes diagnosis was found for mortality endpoints. Conclusion In EchoCRT, HF patients with a narrow QRS complex and coexisting diabetes demonstrated a signal for less harm caused by CRT compared with patients without diabetes, which was driven by differences in hospitalizations owing to HF.