Exercise oscillatory breathing and NT-proBNP levels in stable heart failure provide the strongest prediction of cardiac outcome when combining biomarkers with cardiopulmonary exercise testing.

Background: N-Terminal pro-brain natriuretic peptide (NT-proBNP) and cardiopulmonary exercise testing (CPET)-derived variables are gold standards for assessing prognosis in heart failure (HF) patients. We sought to refine cardiac events prediction by performing a combined analysis of NT-proBNP with markers of exercise ventilatory efficiency.

Methods And Results: A total of 260 stable HF patients underwent measurements of plasma NT-proBNP levels before, at peak exercise, and at 1-minute CPET recovery phase along with peak oxygen uptake (VO(2)), ventilation to CO(2) production (VE/VCO(2)) slope, and exercise periodic breathing (EPB) determinations. After a median follow-up period of 20.6 months, there were 54 cardiac-related deaths. Univariate analysis including NT-proBNP at rest, at peak exercise, and at 1 minute recovery, peak VO(2), VE/VCO(2) slope, and EPB showed NT-proBNP to be the strongest independent predictor with equivalent performance for rest, peak, and recovery levels. Thus, only NT-proBNP at rest was considered (Harrel C 0.783, 95% confidence interval [CI] 0.722-0.844) with VE/VCO(2) slope (Harrel C 0.720, 95% CI 0.646-0.794), EPB (Harrel C 0.685, 95% CI 0.619-0.751), and peak VO(2) (Harrel C 0.618, 95% CI 0.533-0.704). With bivariate stepwise analyses, NT-proBNP along with EPB emerged as the strongest prognosticators (Harrel C 0.800, 95% CI 0.737-0.862).

Conclusions: In the refinement for robust outcome predictors in HF patients, NT-proBNP levels together with EPB led to the most powerful definition. VE/VCO(2) slope and peak VO(2) did not provide any prognostic adjunct. A biomarker/CPET approach seems very promising to warrant the continuous implementation in the prognostic work-up of HF patients.