Survival of patients on renal-replacement therapy (RRT) is no longer improving. Increasingly, such patients are older and have co-morbid conditions affecting organs other than the kidney. In a retrospective study, we calculated actuarial survival of 375 patients starting RRT during a 6 1/2 year period at renal units in Aberdeen and Dundee, UK, after stratification of patients into three risk groups (low, medium, and high) based predominantly on co-morbidity and to a lesser extent on age. 2-year survival differed significantly between low, medium, and high risk groups both before (86%, 60%, and 35%, respectively; p < 0.002 for all comparisons) and after (90%, 70%, 46%; p < 0.004 for all comparisons) excluding early deaths (within 90 days of starting RRT). Overall survival was 61% in Aberdeen and 68% in Dundee (p = 0.04), but 73% and 74%, respectively, when deaths in the first 90 days were excluded (p = 0.73). We conclude that RRT is a highly successful treatment (86% 2-year survival) for patients aged under 70 with no co-morbid conditions (low-risk group); that coexisting non-renal disease has an important influence on survival of patients on RRT; and that risk stratification and analysis of data including and excluding early deaths should allow more valid comparison of data from different centres.