Urotensin II (UTN) is a vasoactive substance that may induce vasoconstriction or vasodilatation. Although this peptide is seen as a vasculotoxic substance, to date there is no prospective study examining the relationship between UTN and hard end points like cardiovascular (CV) events. UTN is much increased in end-stage renal disease (ESRD) and this disease may represent a useful natural model to explore the relationship between UTN and CV outcomes. In this study, we analysed the relationship between plasma UTN and incident CV events (fatal and non-fatal) in a cohort of 191 haemodialysis patients followed up for an average time of 3.6 years (range 0.07-5.8 years). Plasma UTN in haemodialysis patients (median: 6.5 ng/ml) was twice higher than in healthy subjects (median: 3.3 ng/ml). During the follow-up period, 94 patients died and 88 had incident fatal and non-fatal CV events. UTN was significantly lower in patients with incident CV events (median: 5.3 ng/ml) than in events-free patients (median: 7.1 ng/ml), and in a Kaplan-Meier analysis, high UTN was strongly and inversely associated with incident CV events (P<0.001). Multivariate Cox's regression analysis fully confirmed plasma UTN as an inverse predictor of adverse CV outcomes, and in this analysis, UTN resulted to be the third factor in rank, after age and diabetes, explaining the incidence of CV events. UTN is an inverse predictor of CV outcomes in ESRD. Our data suggest that UTN should not be necessarily seen as a vasculotoxic peptide in haemodialysis patients.