Karwath A, Bunting KV, Gill SK, et al. - In patients with heart failure and reduced left ventricular ejection fraction (LVEF), prognostic response from β blockers can be distinguished using an artificial intelligence-based clustering approach. Patients in sinus rhythm with suboptimal efficacy as well as a cluster of patients with atrial fibrillation where β blockers did lower mortality were included in this.

• On pooled individual patient data from nine double-blind, randomized, placebo-controlled trials of β blockers, researchers applied neural network-based variational autoencoders and hierarchical clustering.

• Included were 15,659 patients with heart failure and LVEF of less than 50%.

• In sinus rhythm (n = 12,822), β blockers were consistently linked with overall mortality benefit in most clusters; no significant efficacy was identified in one cluster of older patients with less severe symptoms.

• In atrial fibrillation (n = 2,837), overall neutral effect of β blockers vs placebo was consistent in four of five clusters.

• A statistically significant reduction in mortality was observed with β blockers in one cluster of younger atrial fibrillation patients at lower mortality risk but similar LVEF to average.