Patients with chronic kidney disease (CKD) who live up to renal replacement therapy (RRT) are a selected group of patients with a high mortality risk. The aim of this paper is to contribute a potential epidemiological explanation as to why therapeutic interventions - targeting specific causes of death - of which the effectiveness has been shown in the general population may not have a similar impact in a highly selected population like RRT patients. In this perspective, selection processes over the course of renal disease progression as well as the potential 'dilution' of an effect in the presence of highly increased mortality from other causes need attention. We suggest that the results from well-conducted high-quality studies in incident RRT patients without or with only very limited in- and exclusion criteria are likely the ones best qualified to be extrapolated to other RRT populations.