Background Recent evidence suggests that high adiponectin levels serve as an independent predictor of mortality in chronic heart failure (CHF) patients. We aimed to assess the prognostic importance of adiponectin in CHF towards heart failure-related hospital admissions and mortality, in relation to other clinical, laboratory and exercise data. Methods Seventy-three CHF patients were recruited from the Heart Failure Clinic of the Antwerp University Hospital and followed for a median of 7 (range 1.59.1) years. Study endpoint was the combined occurrence of heart failure-related hospitalizations and all-cause death. At baseline patients underwent clinical assessment, echocardiography and cardiopulmonary exercise testing. Circulating concentrations of adiponectin, NT-proBNP and lipoproteins were measured. After follow-up the hazard ratio (HR) of adiponectin for outcome was estimated using multivariable Cox proportional hazard regression analysis. Results During follow-up, 14 (19%) patients died and 46 (63%) were admitted for CHF deterioration. The unadjusted hazard for poor outcome was higher in patients with adiponectin values above the 75th percentile (15.2 mg/L) (P = 0.031). Adiponectin remained independently predictive [HR (95% CI) 2.47 (1.215.03), P = 0.013], when controlling for well-established predictors of mortality/morbidity in CHF. Additional correction for BMI, NT-proBNP, VO2 peak, HDL and triglycerides did not affect the HR estimate. After adjusting for beta-blocker intake the association between adiponectin and poor outcome was no longer significant. Conclusions High adiponectin levels predict poor outcome in CHF patients independently of well-established and novel prognostic factors, but this prognostic value is significantly affected by beta-blocker treatment.