Heart failure (HF) is characterized by impaired myocardial β-adrenergic signal transduction. Single nucleotide polymorphisms (SNPs) within the β(1)- (Ser49Gly, Arg389Gly) and β(2)-adrenoceptor (Arg16Gly, Gln27Glu, Thr164Ile) have been associated with alterations in adrenoceptor (AR) function sensitivity in vitro and in vivo and possibly contribute to HF progression. The present study evaluated the relation of those SNPs to morbidity and mortality in patients with end-stage HF. A total of 226 patients with end-stage HF (ejection fraction ≤35%) were genotyped for the two β(1)AR SNPs and the three β(2)AR SNPs. Outcome (death, heart transplantation (HTX)) was determined from May 2003 to June 2004. Heart rate, systolic and diastolic blood pressure, and peak oxygen uptake were measured during graded treadmill exercise. Left ventricular end-diastolic and end-systolic diameters, ejection fraction, and fractional shortening at rest were measured using two-dimensional echocardiography. Minor allele frequencies were 0.12 for Gly49 and 0.27 for Gly389 (β(1)AR) and 0.37 for Arg16, 0.43 for Glu27 and 0.01 for Ile164 (β(2)AR). During follow-up, 45 patients died (20%), and 27 patients underwent HTX (12%). No significant differences in the incidence or in the time-to-endpoint of death and HTX between genotypes of the different SNPs within the β(1)- and β(2)AR were detected. However, patients carrying the Arg16-β(2)AR tended to have lower exercise capacity and a higher probability for death/HTX within 45 months (survival proportion 46%) than patients carrying the Gly16Gly-β(2)AR (survival proportion 64%). In conclusion, the Arg16Gly-β(2)AR might impact on exercise capacity and outcome in end-stage heart failure.