Survival is the ultimate outcome measure in renal replacement therapy (RRT) and may be used to compare performance among centres. Such comparison, however, is meaningless if the influences of comorbidity, age and early deaths are not considered. We therefore studied survival rates on RRT in seven centres in Europe after taking into account the influence of age, early deaths, primary renal diagnoses, and comorbidity. A retrospective survival analysis was carried out on 1407 patients who commenced RRT in seven centres across five European countries during a 7-year period. Patients were stratified into low-, medium- and high-risk groups based mainly on comorbidity and to a lesser extent on age at commencement of RRT. Kaplan-Meier survival and Cox's proportional hazards model were used to compare survival. Before risk stratification overall 2-year survival across the seven centres ranged from 60.2 to 85.3% (69.3-89.9%) after excluding early deaths) masking a range of survivals of 27.4% for the high-risk group with the worst survival to 100% in the low-risk group with the best survival. After excluding early deaths 2-year survival in the low risk groups (n=622) was greater than 90% in all centres. Multivariate analysis showed that the mortality risk increased four fold from low- to medium- and a further 1.6-fold from medium- to high-risk group. However, despite this adjustment for comorbidity and age there still remained a significant difference in survival among some centres, i.e. a 'centre effect' which ranked the centres. Risk stratification diminishes the variance in survival between centres but a centre effect remains despite adjusting for age and comorbidity. Multicentre prospective studies are urgently required to identify the reasons for this apparent centre effect.