Objective. To assess the utility of B-type natriuretic peptide (BNP) and C-terminal-pro-endothelin-1 (CTproET-1) to predict a severely impaired peak oxygen consumption (peak VO 2 , < 14 mL kg )1 min )1 ) in patients referred for cardiopulmonary exercise testing.Design. Cross-sectional study.Setting. Tertiary care center.Methods. Peak VO 2 , BNP and CT-proET-1 were assessed in 141 consecutive patients referred for cardiopulmonary exercise testing.Results. B-type natriuretic peptide [median (interquartile range) 48 (38-319) vs. 33 (15-86) pg mL )1 ; P = 0.002] and vs. 60 (52-74) pmol L )1 ; P < 0.001] were higher in patients with a peak VO 2 < 14 mL kg )1 min )1 (n = 30) than in those with a peak VO 2 ‡ 14 mL kg )1 min )1 (n = 111). CT-pro-ET-1 had a higher area under the receiver-operator-characteristics curve (AUC) to predict a peak VO 2 < 14 mL kg )1 min )1 than BNP (0.79 vs. 0.68; P = 0.04). The optimal BNP cut-off of 37.2 pg mL )1 had a sensitivity of 80% and a specificity of 56%. The optimal CT-proET-1 cut-off of 74.4 pmol L )1 had a sensitivity of 80% and specificity of 76%. A five-item score composed of body mass index, diabetes, forced expiratory volume within the first second, alveolo-arterial oxygen pressure difference, and BNP had an AUC of 0.88 to predict a peak VO 2 < 14 mL kg )1 min )1 . Adding CT-proET-1 to the score resulted in an AUC of 0.92.Conclusions. C-terminal-pro-endothelin-1 is superior to BNP for the prediction of a peak VO 2 < 14 mL kg )1 min )1 in patients referred for CPET. A score incorporating body mass index, diabetes status, spirometry, blood gases, BNP and CT-proET-1 improves the prediction of a peak VO 2 < 14 mL kg )1 min )1 based on single biomarkers.