BACKGROUND: Canagliflozin reduces cardiovascular events including hospitalization for heart failure (HHF) in patients with type 2 diabetes and cardiovascular risk. Elevated amino-terminal pro-B-type natriuretic peptide (NT-proBNP) concentrations are associated with HF diagnosis and predict cardiovascular risk. OBJECTIVES: The purpose of this study was to measure NT-proBNP in CANVAS (Canagliflozin Cardiovascular Assessment Study) participants. METHODS: Associations between baseline NT-proBNP and cardiovascular, renal, and mortality outcomes and intervention-associated changes were determined. RESULTS: Of the 4,330 participants in the CANVAS trial, NT-proBNP was measured in 3,587, 2,918, and 995 participants at baseline, 1 year, and 6 years, respectively. The median baseline NT-proBNP concentration was 91 pg/ml, and 39.3% had NT-proBNP ≥125 pg/ml. NT-proBNP was higher in those with investigator-reported HF (13% of participants at baseline) versus those without (187 pg/ml vs. 81 pg/ml), with substantial overlap between groups. By 1 year, NT-proBNP increased with placebo, whereas canagliflozin reduced NT-proBNP by 11% (geometric mean ratio for canagliflozin vs. placebo = 0.89 [95% confidence interval (CI): 0.84 to 0.94]; p < 0.001). Lower NT-proBNP with canagliflozin was also observed at 6 years (p = 0.004). In adjusted models, baseline NT-proBNP ≥125 pg/ml was prognostic for incident HHF (hazard ratio [HR]: 5.40; 95% CI: 2.67 to 10.9), HHF/cardiovascular death (HR: 3.52; 95% CI: 2.38 to 5.20), and all-cause death (HR: 2.53; 95% CI: 1.78 to 3.61). Mediation analyses suggested that 10.4% of the effects of canagliflozin on HHF were reflected in NT-proBNP lowering. CONCLUSIONS: A substantial percentage of patients in the CANVAS trial had elevated NT-proBNP values. Canagliflozin reduced NT-proBNP concentrations versus placebo; however, reduction in NT-proBNP explained only a small proportion of the benefit of canagliflozin on HF events. (CANVAS [CANagliflozin cardioVascular Assessment Study]; NCT01032629).