Uremic toxin removal based on diffusion and/or convection allows eliminating solutes with negative metabolic impact. Uremic solutes can be classified as small and water-soluble compounds, larger "middle" molecules, or protein bound solutes. The question arises whether more removal of each of these solute classes affects patient survival. Kt/V of urea is currently used as a surrogate for small water-soluble solute removal. There is ample evidence that Kt/V and survival are correlated, but the threshold Kt/V remains a matter of debate. Probably, the actually proposed threshold of 1.2 is too low. This impact of Kt/V is in contradiction with the low toxicity of urea and points to a role for other water-soluble solutes, e.g., potassium. More removal of middle molecules results in a lower morbidity and also in a lower mortality. In addition, a relationship has been demonstrated between the use of membranes with large pore size and a decrease of inflammatory status, by itself an important factor related to mortality. One of the problems is that large pore membranes are at the same time more biocompatible and reflect more dialysate impurities, compared to many small pore membranes, whereas they also reflect more dialysate impurities. It remains uncertain which one of these factors, if any, has a predominant effect. Recent studies point to a separate effect of pore size but await confirmation. Protein bound toxins inhibit several biochemical functions. Their removal pattern is totally different from that of classical markers such as urea. In analogy with drugs, it is essentially the free unbound fraction that exerts biological action; this free fraction is inversely related to serum albumin, another inflammatory marker related to survival. In a final section of this presentation, attention will be drawn to the relationship in uremic patients between inflammation, malnutrition, cardiovascular disease, and mortality, and some of the potential culprits are discussed. Virtually all of these molecules have a high molecular weight or are protein bound. It is concluded that both small and middle molecule removal have an impact on survival, so that more than urea removal alone should be pursued.