Uraemic toxins are defined as deleterious compounds, normally excreted in urine, that accumulate in patients with chronic kidney disease (CKD) as a result of their decreased renal clearance [1]. The classification of uraemic toxins is mainly based on their molecular weight and their protein binding properties, resulting in three classes of compounds: low molecular weight (<500 kDa, e.g. urea), middle molecules (500–60 000 kDa, e.g. β2 microglobulin) and protein-bound uraemic toxins [such as indoxyl sulphate (IS) and p-cresyl sulphate (PCS)]. This latter group is the most extensively studied due to the experimental and clinical evidence of their toxicity and their poor removal by classic extrarenal epuration methods since only the free fraction can diffuse across the dialysis membrane. Although their protein-binding affinity varies from one compound to another (90–95% for IS and PCS but only 56% for hippuric acid), they are poorly removed by most of the current dialysis methods. Moreover, the free fraction decreases during dialysis, further reducing the clearance efficacy [3]. Even with ‘high efficiency’ membranes, the dialytic clearance of protein-bound toxins remains a very small fraction of that of urea [4]. Therefore new and different strategies are needed to enhance the removal of protein-bound uraemic toxins and prevent their accumulation in patients on maintenance haemodialysis. Residual kidney function (RKF) is usually defined in patients receiving haemodialysis or peritoneal dialysis as the daily production of at least 100 mL of urine. RKF often rapidly declines and hence has frequently been considered as negligible in clinical practice. The view that RKF is simply an innocent bystander has recently been challenged. Greater RKF among haemodialysis patients has been associated with better survival, can play a role in the production of erythropoietin and the activation of vitamin D and has been correlated with better nutritional status and mental health [5, 6]. Experimental and clinical studies have highlighted that protein-bound uraemic retention solutes are secreted into the tubular lumen by organic anion transporters (OATs) located on the basolateral membranes of renal proximal tubular cells, while urea is reabsorbed [7]. Recently Jansen et al. [8] demonstrated more specifically that IS may induce the expression and function of OAT1 in renal proximal tubule cells. Binding of IS to the epidermal growth factor receptor leads to downstream activation of the aryl hydrocarbon receptor signalling pathway and induction of OAT1, which facilitates removal of protein-bound toxins. As renal function declines, IS and PCS progressively accumulate in the plasma, achieving concentrations sufficient to promote the secretion by OATs into proximal tubules [9]. A significant proportion of urine produced by haemodialysis patients may be the product of proximal tubular secretion, together with some unreabsorbed glomerular filtrate. It seems reasonable to postulate that RKF reflects the contribution of proximal tubular solute excretion and ensures more efficient removal of protein-bound uraemic solutes than dialysis. Supporting this view, different small cohorts have shown, for example, a higher plasma IS concentration in anuric patients than in patients with RKF [10–13], even if this was not confirmed in all studies [14]. Preserved RKF contributes to the achievement of adequate solute clearance for low molecular weight proteins, and RKF estimated by urea clearance is a strong predictor of serum β2-microglobulin level [15]. Taken together, these data suggest that lower concentrations of protein-bound uraemic toxins could contribute to the better outcomes observed in haemodialysis patients with a sustained RKF. In this issue of Nephrology Dialysis Transplantation, Toth-Manikowski et al. [16] discuss results from the Haemodialysis (HEMO) trial of 1280 patients having a urinary urea clearance <1.5 mL/min. Expanding their previous report [12], the authors reported that patients with preserved RKF at baseline exhibited lower levels for seven of eight major uraemic toxins. For instance, hippuric acid was 24% lower [95% confidence interval (CI) 30.8–16.5], IS was 11% lower (95% CI −16 to −6) and trimethylamine N-oxide (TMAO) was 4% lower (95% CI −6 to −1). In good agreement, preserved RKF (both at baseline and after 1 year) was associated with a 19% lower risk of death [hazard ratio 0.81 (95% CI 0.61–1.08)] during the 1-year follow-up period, providing some convincing information that the reduction of uraemic toxins could account for the reduced mortality. It should be noted that they also found a surprising increase in PCS concentration in subjects with RKF compared with those without RKF. The HEMO study was a large randomized, prospective clinical trial that was initially designed to examine the impact of two haemodialysis parameters, membrane flux and dialysis dose [17]. No statistically significant effect of higher dialyser flux [relative risk (RR) 0.92 (95% CI 0.81–1.06)] or higher urea Kt/V [RR 0.96 (95% CI 0.84–1.10)] were observed on mortality. These data suggest that optimal and intense dialysis only have a limited impact on survival and comorbidities. One explanation could be that classic epuration techniques are unable to remove all toxins, especially those belonging to the protein-bound solutes family. Dialysis is viewed as a substitute for glomerular function, while tubular function is rarely considered. Similarly, RKF is classically defined as glomerular filtrate that escaped tubular reabsorption and is considered as a ‘clearance’ by measurement of urea and/or creatinine clearance in urine. However, such measures may downplay the contribution of residual tubular secretion and be impaired by the difficulty in completely collecting urine samples [18]. With this ‘classical’ view, the daily excretion of 100–250 mL of urine in a dialysis patient is equivalent to the prolongation of haemodialysis treatment by 15–30 min. With this model, it seems unlikely that an increase in uraemic toxins clearance of a few litres per day could, as in this study [16, 17], account for the improved outcome seen in patients with RKF. This indeed suggests that RKF should be included and interpreted as combined glomerular and tubular function. The impact of RKF on uraemic toxins was not included in the original analysis plan of the HEMO study, although the number of included patients is quite large (n = 1280 patients, 34% with RKF and 66% without RKF). The analyses were realized by comparing two arms rather than performing correlation analyses that could minimize sources of bias. The range of RKF was limited since patients with an adjusted residual urea clearance >1.5 mL/min were excluded and we thus have no information about the impact of greater levels of RKF on survival. There were also various modalities of dialysis and the presence or not of native kidney. Moreover, the study included prevalent patients, and the influence of treatments on uraemic toxins concentration before inclusion cannot be excluded. It is interesting to note that anuric patients have been dialysed for a longer time than those patients with RKF (1.8 ± 1.8 versus 4.3 ± 4.7 years). Therefore the body accumulation and serum concentration of toxic molecules and the associated clinical outcome may differ among situations. We have no information about diet composition and the putative use of pre-/probiotics or antibiotics. Besides passive accumulation of uraemic toxins, recent data highlight the fact that intestinal microbiota strongly contribute to the production of protein-bound toxins [19]. TMAO is produced from the breakdown of choline and L-carnitine, which are metabolized by intestinal bacteria into trimethylamine and then converted to TMAO in the liver. PCS originates from the intestinal microbial fermentation of tyrosine into p-cresol with subsequent endogenous sulphation. IS is derived from the fermentation of tryptophan into indole, followed by endogenous oxidation and sulphation [20]. The quantity and the quality of proteins and fibres can strikingly influence uraemic toxins production and intestinal microbiota composition. It should be noted that in this study, normalized protein catabolic rate measured as an index of protein intake was significantly lower in patients with no RKF in favour of reduced protein intake and therefore a less toxin productive intake. Thus this observation does not challenge the final conclusion. Only total uraemic toxins have been evaluated and free fractions were not explored. Of note, uraemic toxins were measured only once at a single time point and their concentrations may have changed over time. Free concentrations of uraemic solutes may be a better indicator of toxicity, as cells are exposed to the free solutes. If free solute levels followed the same pattern as total solute levels, free levels could be influenced by different factors such as the presence of other unmeasured protein-bound solutes or medications (e.g. ibuprofen) [21] that share common albumin binding sites with IS and PCS and could displace these toxins from their binding sites, thereby augmenting their concentrations (Figure 1). It is fortunate that in this trial, serum albumin was similar in both groups (3.6 ± 0.3 g/dL). In the HEMO trial, the authors tried to assess the free fraction of uraemic toxins, but the free fractions for PCS and IS in some samples were greatly in excess compared with those observed in samples obtained locally [22, 23]. Samples were analysed >10 years after blood collection and the stability of the studied solutes remains uncertain. An important point to note is that in the HEMO study, no correlation between total PCS, IS and TMAO and cardiovascular outcomes or death were observed [22], in contrast to many other studies in CKD Stages 1–5 implemented to this aim [24, 25]. Dou et al. [20], in another CKD cohort, also reported that in 120 patients with CKD (73 treated with haemodialysis), higher total PCS and IS were not associated with outcomes [26]. Previous studies in haemodialysis patients showed associations only between free but not total concentrations of uraemic toxins and cardiovascular outcomes [27, 28] that could explain the negative results in the HEMO trial. All these drawbacks may provide a key to understanding why paradoxical and unexpected decreases of PCS in patients without RKF were observed in this study [16]. Also, TMAO was lower in patients with RKF, but no effect modification was seen by RKF, as this could have been due to the population size. Factors that influence the kinetics of protein-bound uraemic toxins concentrations and the role of RKF. Alb., albumin; ADMA, asymmetric dimethylarginine; SDMA, symmetric dimethylarginine. This study [16] confirms that preservation of RKF is one of the most important parameters related to patients’ survival during maintenance dialysis and pinpoints the reasons why. Using urea and/or creatinine clearance as indexes may give misleading information that makes the use of timed urine collection unable to estimated RKF. Additional studies clarifying the biological variability of circulating levels of uraemic toxins and in particular PCS are needed. Unfortunately, strategies available to reduce protein-bound uraemic toxins are currently limited. Adsorbents such as AST-120 failed to improve survival in CKD [29] and although dietary interventions are appealing, intervention trials investigating these therapies remain of poor quality [20]. The identification of molecular targets enhancing OAT1 function has led to a new exciting area to preserve RKF and to increase renal secretion of harmful endogenous waste. Estimating the proximal tubular secretory function could add novel and relevant information to existing measures of kidney function and their association with cardiovascular morbidity and mortality. The present positive findings remind us that it is important as a general principle to maintain RKF by avoiding or reducing nephrotoxic events and control blood pressure. Additional efforts are needed to investigate whether methods that extend RKF in dialysis patients could improve this question of uraemic solutes and toxicity. This study was supported by Hospices Civils de Lyon, University of Lyon and INSERM. None declared. (See related article by Toth-Manikowski et al. Contribution of ‘clinically negligible’ residual kidney function to clearance of uremic solutes.
No takes yet. Share an insight, caveat, or question.
Koppe et al. (2019) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: