Influence of left ventricular ejection fraction on the reduction in N-terminal pro-brain natriuretic peptide by canagliflozin in patients with heart failure and type 2 diabetes: A sub analysis of the CANDLE trial

Abstract

AimTo investigate the effect of left ventricular ejection fraction (LVEF) on the behavior of N-terminal pro-brain natriuretic peptide (NT-proBNP) levels in patients with heart failure and type 2 diabetes mellitus with the use of canagliflozin compared to glimepiride. MethodsPatients (n = 233) from the CANDLE trial were randomly assigned to either the add-on canagliflozin (n = 113) or glimepiride treatment groups (n = 120). The patients were followed-up for 24 weeks. The NT-proBNP levels were measured at baseline and after 24 weeks. The LVEF was determined at baseline. ResultsThere was a significant relationship between the baseline NT-proBNP level (X1) and the change in NT-proBNP levels from baseline to 24 weeks (Y) in the canagliflozin group (Y = −0.533 × X1 + 178; r = −0.860, p < 0.001). However, this relationship was not observed in the glimepiride group (p = 0.428). The baseline LVEF (X2) correlated with Y with a marginal significance in the canagliflozin group (Y = 7.72 × X2–549; r = 0.192, p = 0.054), but no relationship was observed in the glimepiride group. In the canagliflozin group, bivariate regression analysis showed a significant correlation between Y, X1, and X2; Y = −0.567 × X1–6.04 × X2 + 542 (R = 0.871, p < 0.001). The partial regression coefficients of X1 (p < 0.001) and X2 (p = 0.006) significantly explained the variance in Y. The correlation coefficient for X2 was negative. There was a significant relationship between the logarithmically transformed NT-proBNP [ln(NT-proBNP)] at baseline (X1′) and the change in ln(NT-proBNP) values from baseline to 24 weeks (Y′), a surrogate of the rate of change in NT-proBNP levels, in the canagliflozin group (Y′ = −0.18 × X1′ + 0.93; r = 0.450, p = 0.001). ConclusionsThe baseline NT-proBNP level significantly affected the extent and the rate of its decrease by canagliflozin. The reduction in NT-proBNP levels by canagliflozin was prominent in patients with a higher LVEF at baseline. However, its confounding effect of LVEF on canagliflozin treatment was not recognized without adjusting for the NT-proBNP level at baseline.