In this pro-con debate, we will analyse the relevance of data supporting the prime role of uraemic toxins versus salt and volume overload as the main driver for the impressive mortality and morbidity characteristic of any end-stage re- nal disease patient. Firstly, we are going to review the circumstantial evidence and associated studies/clinical observational data which only weakly suggest that uraemic solutes are indeed linked to damage to the body and its constituents, especially the car- diovascular (CV) system. At the same time, we will equally thoroughly discuss the 'fact' that decreasing salt and fluid load is essential for any dialysed patient. Evidence in favour of a prominent role for uraemic toxins is appealing from a pathophysiological point of view, but unfortunately is mis- leading in real life practice. Evidence in favour of salt and water reduction is overwhelmingly stronger. This difference in apparent evidence strength notwith- standing the weaker support for uraemic toxins does not take away from clinical importance of minimizing uraemic tox- icity load in the large majority of the dialysis population. filtration (HDF) or more frequent/longer dialysis, was only partially associated with better results. Moreover, in the ma- jority of the positive studies, alternative interpretations of the results is possible, including a contribution of better volume control achievable with these particular dialysis techniques. Some 'observational' studies (2-7) showed a beneficial effect of high-flux versus low-flux synthetic dialysis mem- branes on survival, whereas prospective random controlled trials (RCTs) were inconclusive. In the Haemodialysis (HEMO) study, the largest random- ized clinical trial conducted in haemodialysis (HD) so far, patients were randomized to either a high-dialysis dose (achieved eKt/V ¼ 1.53 � 0.09, sp Kt/V ¼ 1.71 � 0.11) or a standard-dose goal (achieved eKt/V ¼ 1.16 � 0.08, sp Kt/V ¼ 1.32 � 0.09) and to dialysis with either high-flux dialysers (b2-microglobulin clearance of > 20 mL/min) or to low-flux dialysers—(b2-microglobulin clearance < 10 mL/ min). The results showed that a higher HD dose was not associated with a lower mortality risk compared with con- ventional HD dose (8). Thus, while a minimumKt/V ,a ccord- ing to the guidelines, is mandatory, the HEMO study clearly demonstrated that there are no advantages in further increas- ing the urea Kt/V, even in the high-flux membrane group. At the same time, each 10 mg/L increase in b2-M level was associated with an 11% increase in mortality, even after ad- justment for years on dialysis and residual kidney function. These results deserve further specifically designed studies in order to clarify if it is a matter of larger molecule removal or better fluid removal. In the other large trial in the field, the Membrane Perme- ability Outcome study, there was no significant difference in survival between high- and low-flux groups when all patients were included in the analysis (9). However, patients with serum albumin � 40 g/L on enrolment and with diabetes had a significantly better survival in the high-flux arm. High-flux dialysis was associated with a significantly lower accumulation of b2-microglobulin. Again, it is important to clarify if the better outcome was related to larger molecule removal or better fluid removal. Compared with conventional HD, HDF could theoretically offer some advantages: better clearance of middle- and