The membrane perspective of uraemic toxins: which ones should, or can, be removed?

Informed decision-making is paramount to the improvement of dialysis therapies and patient outcomes. A cornerstone of delivery of optimal dialysis therapy is to delineate which substances (uraemic retention solutes or ‘uraemic toxins’) contribute to the condition of uraemia in terms of deleterious biochemical effects they may exert. Thereafter, decisions can be made as to which of the accumulated compounds need to be targeted for removal and by which strategies. For haemodialysis (HD), the non-selectivity of membranes is sometimes considered a limitation. Yet, considering that dozens of substances with potential toxicity need to be eliminated, and targeting removal of individual toxins explicitly is not recommended, current dialysis membranes enable elimination of several molecules of a broad size range within a single therapy session. However, because HD solute removal is based on size-exclusion principles, i.e. the size of the substances to be removed relative to the mean size of the ‘pores’ of the membrane, only a limited degree of selectivity of removal is possible. Removal of unwanted substances during HD needs to be weighed against the unavoidable loss of substances that are recognized to be necessary for bodily functions and physiology. In striving to improve the efficiency of HD by increasing the porosity of membranes, there is a greater potential for the loss of substances that are of benefit. Based on this elementary trade-off and availability of recent guidance on the relative toxicity of substances retained in uraemia, we propose a new evidence-linked uraemic toxin elimination (ELUTE) approach whereby only those clusters of substances for which there is a sufficient body of evidence linking them to deleterious biological effects need to be targeted for removal. Our approach involves correlating the physical properties of retention solutes (deemed to express toxicity) with key determinants of membranes and separation processes. Our analysis revealed that in attempting to remove the relatively small number of ‘larger’ substances graded as having only moderate toxicity, uncontrolled (and efficient) removal of several useful compounds would take place simultaneously and may compromise the well-being or outcomes of patients. The bulk of the uraemic toxin load comprises uraemic toxins below <30 000 Da and are adequately removed by standard membranes. Further, removal of a few difficult-to-remove-by-dialysis (protein-bound) compounds that express toxicity cannot be achieved by manipulation of pore size alone. The trade-off between the benefits of effective removal of the bulk of the uraemic toxin load and risks (increased loss of useful substances) associated with targeting the removal of a few larger substances in ‘high-efficiency’ HD treatment strategies needs to be recognized and better understood. The removability during HD of substances, be they toxic, inert or beneficial, needs be revised to establish the pros and cons of current dialytic elimination strategies.