The arterial ketone body ratio (AKBR), an established clinical tool that reflects hepatic mitochondrial oxidation-reduction potential, predicts the outcome of patients with shock and multiple organ failure and the postoperative outcome in patients who have undergone major liver or heart surgery. The purpose of this study was to determine the prognostic significance of AKBR in patients with acute heart failure. The subjects of this study were 52 patients with acute heart failure. The following parameters were analyzed after Cox univariate hazard analysis was performed: AKBR, plasma norepinephrine, left ventricular ejection fraction, cardiac index, pulmonary arterial wedge pressure, sex, age, human atrial natriuretic peptide, endothelin-1, and cholesterol. The follow-up period was 30 weeks with cardiac death as the end point. Stepwise multivariate proportional hazard analysis revealed that AKBR was the most significant predictor of death, followed by norepinephrine and human atrial natriuretic peptide. Curve-fitting analysis revealed that the relationship between log (norepinephrine) and AKBR could best be described by two distinct lines, with their intersection at AKBR = 0.7 and norepinephrine = 418. With these results we conducted Kaplan-Meier analysis for AKBR > or = 0.7 and AKBR <0.7. The survival rate in patients with AKBR > or = 0.7 was 100%, whereas that in patients with AKBR <0.7 was 15% (p < 0.0001, log-rank analysis). These results indicate that AKBR is a novel independent predictor of death in heart failure.