Research Article: Plasma p-Cresol Lowering Effect of Sevelamer in Peritoneal Dialysis Patients: Evidence from a Cross-Sectional Observational Study

Date Published: August 28, 2013

Publisher: Public Library of Science

Author(s): Bruna Guida, Mauro Cataldi, Eleonora Riccio, Lucia Grumetto, Andrea Pota, Silvio Borrelli, Andrea Memoli, Francesco Barbato, Gennaro Argentino, Giuliana Salerno, Bruno Memoli, Rosa Maria Affonso Moysés.


p-Cresol is a by-product of the metabolism of aromatic aminoacid operated by resident intestinal bacteria. In patients with chronic kidney disease, the accumulation of p-cresol and of its metabolite p-cresyl-sulphate causes endothelial dysfunction and ultimately increases the cardiovascular risk of these patients. Therapeutic strategies to reduce plasma p-cresol levels are highly demanded but not available yet. Because it has been reported that the phosphate binder sevelamer sequesters p-cresol in vitro we hypothesized that it could do so also in peritoneal dialysis patients. To explore this hypothesis we measured total cresol plasma concentrations in 57 patients with end-stage renal disease on peritoneal dialysis, 29 receiving sevelamer for the treatment of hyperphosphatemia and 28 patients not assuming this drug. Among the patients not assuming sevelamer, 16 were treated with lanthanum whereas the remaining 12 received no drug because they were not hyperphosphatemic. Patients receiving sevelamer had plasma p-cresol and serum high sensitivity C-reactive protein concentrations significantly lower than those receiving lanthanum or no drug. Conversely, no difference was observed among the different groups either in residual glomerular filtration rate, total weekly dialysis dose, total clearance, urine volume, protein catabolic rate, serum albumin or serum phosphate levels. Multiple linear regression analysis showed that none of these variables predicted plasma p-cresol concentrations that, instead, negatively correlated with the use of sevelamer. These results suggest that sevelamer could be an effective strategy to lower p-cresol circulating levels in peritoneal dialysis patients in which it could also favorably affect cardiovascular risk because of its anti-inflammatory effect.

Partial Text

Uremic toxins are a heterogeneous group of compounds that accumulate in the plasma of patients with chronic kidney disease (CKD). More than 90 different uremic toxins have been identified up to day; based on their molecular weight and their affinity for plasma proteins, they can be classified in three different groups: small water soluble molecules not significantly bound to plasma proteins, small molecules significantly bound to plasma proteins and middle/large proteins [1]. The great interest that has been accruing on uremic toxins over the years derives from experimental evidence suggesting that some of them may have a causative role in the development of long-term complications of CKD and, in particular, of cardiovascular disorders [2], which are major cause of death in this disease [3]. Recent evidence points to p-cresol as one of the uremic toxins more directly implicated in the pathogenesis of CKD complications. This aromatic compound is generated in the gut by the degradation of tyrosine and phenylalanine operated by resident intestinal flora [4]–[5]. After absorption, p-cresol is converted into its conjugates p-cresylglucuronide and p-cresylsulfate. The latter, which represents more than 95% of circulating p-cresol, is responsible for the majority of p-cresol toxic effects [6]. The plasma concentrations of p-cresol and p-cresylsufate, which belong to the subgroup of small molecules significantly bound to plasma proteins, are strongly related to cardiovascular risk in CKD [6]–[9] and are predictive of mortality in these patients [10]. This is consistent with a number of studies in vitro that clearly showed that p-cresol and its derivative p-cresylsulphate are toxic for endothelial cells and can cause endothelial dysfunction [11]–[13]. Intense efforts are currently directed to identify the best therapeutic strategy to lower uremic toxins in CKD patients because it has been shown that this can lead to a significant improvement in their survival [14]. Unfortunately, dialysis seems to be effective only in removing small water soluble uremic toxins whereas those significantly bound to plasma proteins are significantly retained despite the dialysis treatment [15]. Specifically, p-cresol and its sulphate derivative are extremely difficult to dialyze [16]. An interesting alternative approach to lower the plasma concentrations of p-cresol is directed to lowering its intestinal absorption [17]. The rationale behind this strategy is that all circulating p-cresol is derived from that produced by bacteria in the gut because this compound cannot be generated by the metabolism of aromatic aminoacids by human cells [17]. Studies in vitro showed that the non-calcium non-aluminum containing phosphate binder sevelamer hydrochloride (Sev), which is largely used to treat hyperphosphatemia in end stage renal disease (ESRD) [18]–[19], also binds p-cresol [20]. This evidence suggested that this orally administered phosphate binder could lower p-cresol concentrations in human patients with CKD by preventing its intestinal absorption. Contrarily to these expectations, Brandeburg et al. (2010) [21] reported that p-cresol plasma concentrations were significantly higher at the end of an 8 week treatment with Sev than before it was started and, importantly, that they returned at their basal levels when the treatment with this drug was stopped. However, this remains the only study that explored the effects of Sev on p-cresol in hemodialysis patients. In addition, the impact of the treatment with this drug on p-cresol levels in peritoneal dialysis (PD) patients has never been investigated. Considering this lack of information, in the present cross-sectional observational study, performed on a cohort of 57 patients with ESRD treated with PD, we compared p-cresol plasma concentrations in patients assuming Sev for the treatment of ESRD-induced hyperphosphatemia and in those not treated with this drug.

57 ESRD patients receiving PD and attending the Division of Nephrology of the Federico II University of Naples as outpatients were recruited for the study. Fifteen of them were women (26%) and the remaining 42 men (74%). At the time of study, patients had a mean age of 59.7±14.5 years. Forty-one patients were on continuous ambulatory peritoneal dialysis and 16 on automated peritoneal dialysis. Based on whether they assumed or not phosphate binders and on which binder they assumed, the patients were stratified in three groups: no binder (n = 12), lanthanum (n = 16) and Sev (n = 29). 20 patients (Sev  = 10, Lanthanum  = 5, No binder  = 5) were treated with calcitriol (0.25 µg every other day–0.5 µg/day) and 23 (Sev  = 13, Lanthanum  = 5, No binder  = 5) with paricalcitol (1 µg every other day–1 µg/day). Because the therapy with hypophosphatemic drugs was individually tailored to achieve target plasma PO4 concentrations, Sev and lanthanum were administered at different dosages in different subjects (dosage ranges: 1600–14400 and 750–3000 mg/die for Sev and lanthanum, respectively). There was no significant difference among the different groups neither in mean age at the time of the study, nor in mean body weight, or relative percentage of the two sexes. Also peritoneal dialysis vintage (i.e. the length of time on dialysis in months) was similar in the three groups. Its mean in the whole patient population was 25.4±22.1 months and in all cases it was longer than six months. Six patients in each of the three patient groups were diabetic. No difference was observed among Sev , lanthanum and no binder groups in mean percentage of HbA1c . It was <7% in both groups suggesting that a good glycemic control was obtained both in all patient groups [24]–[25]. All data we reported so far suggest that Sev-treated, lanthanum-treated and no binder patients are very similar in their demographic, clinical and laboratory profile (Table 1 and Table 2). Nevertheless, significant differences emerged when we compared total p-cresol plasma concentrations. Plasma levels of this uremic toxin were significantly lower in Sev than lanthanum or no-binder groups [median and IQR: 3.3 (1.4–6.9) vs 7.9 (4.1–9.8) and 9.2 (4.3–15.9) in Sev, lanthanum and no binder groups, respectively; H = 9.6, p<0.008] (Fig. 1 and Table 2). In addition, in Sev-treated patients plasma p-cresol concentration was linearly related to the dose of the PO4 binder assumed by the patient being higher Sev doses associated to lower concentrations of this uremic toxin (r = −0.319; P = 0.025) (Fig. 2). Another relevant difference was observed in hs-CRP concentrations that were significantly lower in Sev than in lanthanum or no-binder groups (median and IQR: 3.8 (1.2–6.6) vs 6.3 (2.6–10.0) and 5.9 (3.4–8.4) in Sev, lanthanum and no binder groups, respectively; H = 10.2, p<0.006) (Table 2). No significant difference was observed neither in total creatinine clearance, weekly Kt/V, rGFR and urine volume suggesting that residual renal function and dialysis efficiency were similar in these three groups (Table 1). Moreover, also serum albumin concentrations were not significantly different among the three groups suggesting that the differences in the plasma concentration of p-cresol, a uremic toxin that circulates largely bound to serum albumin, could not be explained by a lower protein-bound fraction (Table 2). Considering that Sev therapy was started because of concurrent hyperphosphatemia and that the main pharmacological effect of Sev is to lower PO4, we compared PO4 circulating levels in the three groups (Table 2). No difference among groups was found (Table 2) suggesting that the treatment with the PO4 binder was effective in normalizing PO4 profile. The main finding of the present study is that in PD patients the concomitant use of Sev for hyperphosphatemia is associated with lower plasma p-cresol concentrations. This was suggested by the significantly lower plasma p-cresol concentrations in patients assuming Sev as compared with those assuming no binder or lanthanum and was confirmed by multiple linear regression analysis.   Source: