Not Residual Kidney Removal but Peritoneal Clearance Dominates in Uric Acid Control in Peritoneal Dialysis

Background: There’s a paucity of systematic study focusing on clearance of uric acid (UA) in peritoneal dialysis (PD). The aim of this study was to investigate peritoneal UA removal and its influencing factors in PD patients. Methods: This was a cross-sectional study. Patients who performed peritoneal equilibration test (PET) and Kt/V from April 1, 2018 to August 31, 2019 were enrolled. The demographic data, clinical and laboratory parameters including the UA levels in the dialysate, blood and urine samples were collected. Results: Finally, there were 180 prevalent PD patients (52.8% male). Compared with normal serum UA (SUA) group, the hyperuricemia group showed significantly lower peritoneal UA clearance (39.1±6.2 vs. 42.0±8.0 L/week/1.73m2, P=0.008). Peritoneal UA clearance but not residual kidney removal was revealed to be independently associated with continuous [standardized coefficients (β) -0.21, 95% confidence interval (CI) -0.07, -0.006] and categorical SUA level [odds ratio (OR) 0.91, 95% CI 0.84,0.98]. Furthermore, the higher (high or high-average) transporters showed greater peritoneal UA clearance than the lower (low or low-average) transporters (42.0±7.0 vs. 36.4±5.6 L/week/1.73 m2, P<0.001). Among the widely used solute removal indicators, peritoneal creatinine clearance (CCL) performed best to predict higher peritoneal UA clearance in ROC analysis [area under curve (AUC) 0.96, 95% CI 0.93-0.99]. In multiple linear regression, lower albumin level (β -0.06, 95%CI -2.09, -0.19), higher transporters (β 0.06, 95%CI 0.05, 1.69) and greater peritoneal CCL (β 0.95, 95%CI 0.81, 0.93) were independently associated with higher peritoneal

dominated in SUA balance of PD patients. Albumin level, peritoneal transport type and peritoneal CCL were identified as independent determinants of peritoneal UA clearance.

Background
Uric acid (UA, 2,6,8-trihydroxypurine; C5H4N4O3), as the end-products of endogenous and dietary purine metabolism for the lack of uricase, is a weak diprotic acid of two dissociable protons with a pKa1 of 5.4 and pKa2 of 10.3, respectively [1]. At a physiology PH of 7.4, 98% of the UA was demonstrated to exist as monosodium urate in the extracellular milieu [2]. Since UA is poorly soluble in aqueous media and can't freely move through the cytomembrane, it is excreted mainly by means of the UA transporters substantially consisting in the kidney and intestines. It's reported that about 70% of UA is excreted by kidney and 30% by gastrointestinal tract [3,4]. Owing to the leading role of kidney in excreting UA and maintaining the balance of UA in internal environment, almost 90% hyperuricemia is caused by the impairment of renal UA excretion [5]. In the same way, the hyperuricemia was common in patients of chronic kidney disease (CKD) which showed a five-fold increase of prevalence of hyperuricemia than the patients with normal renal function [6].
Peritoneal dialysis (PD), a widespread used dialysis modality, is becoming increasingly important in renal replacement therapy of patients with end-stage renal disease for its distinct strength in economic cost and improvements in techniques and patient survival [7]. The prevalence of hyperuricemia was increased as renal function declined, which ranging from 40-70% in CKD stages 1-5 and was similar in hemodialysis (HD) and PD patients [8][9][10]. The SUA level in PD population was reported at a range of 6.7-8.6 mg/dL (mean 7.6 mg/dL) and increased over the lengthening of dialysis vintage [9,11]. As for the role of SUA in PD patients, our previous study demonstrated that the higher SUA level was independently associated with mortality in PD patients [12,13]. In fact, apart from using UA lowering agents or optimizing the dietary and lifestyle factors [14], the dialysis therapy itself can play a role in SUA control in PD populations [15]. However, whether the peritoneal UA clearance or residual renal function plays a dominant role in SUA control in PD patients is still indeterminate. Since there are little studies referring to UA clearance on PD modality especially excluding the effects of medical history of using UA lowering agents, we aimed to conduct a systematic study to investigate the role of peritoneal UA clearance in SUA control and its relative influencing factors in PD patients.

Study population
This was a single center cross-sectional study. Patients who performed peritoneal equilibration test (PET) and Kt/V test in our PD center from April 1, 2018 to August 31, 2019 were enrolled. The inclusion criteria included prevalent patients older than 18 years old and have initiated PD therapy at least 1 month at the time of performing PET and Kt/V tests. The patients who had taken UA lowering agents within 1 month before the PET and Kt/V tests, transferred from the long-term HD (more than 3 month), failed renal transplant, or with malignant tumors were excluded. All patients enrolled used the standard lactate-glucose peritoneal dialysate (1.5%, 2.5%, or 4.25% dextrose, from Baxter, Guangzhou, China). The relevant clinical parameters were tested in the clinical lab of our hospital by standard methods. All patients had signed informed consent. And the study has conformed to the ethical principles of the Declaration of Helsinki and was approved by the Clinical Research Ethics Committee.

Data collection
The demographic data collected included age, gender, body mass index (BMI), diabetes, cardiovascular disease (CVD), primary kidney disease and so on. Data of the first PET and Kt/V tests during the study period were collected. The PD related information included the dialysis vintage, dialysis dose, average glucose concentration of dialysate, measured glomerular filtration rate (mGFR), Kt/V, weekly total creatinine clearance (CCL), 24h residual urine volume, normalized protein catabolic rate (nPCR) and standard PET data covering the urea, creatinine and UA level in dialysate, blood and urine samples with 2L of 2.5% dextrose dialysate dwelling for 0, 2 or 4 h (0 h was the time point in PET test when all of the 2L dialysate flow into abdominal cavity, and the duration of this process was recorded).
The PD patients was classified into high (H), high average (HA), low average (LA) or low (L) transporters according to the classical Twardowski's criterion [16]. And the clinical parameters included the blood pressure, hemoglobin, neutrophil/lymphocyte, high sensitivity C-reactive protein, alkaline phosphatase, serum albumin, prealbumin, corrected calcium, phosphorus, total cholesterol, triglyceride, serum urea nitrogen, creatinine, SUA and intact parathyroid hormone.
The medication history was also collected according to the follow-up records of patients who would regularly revisit our PD center for condition assessing and therapeutic regimen adjustment about once 1-3 months. CVD was defined as angina, myocardial infarction, congestive heart failure, cerebrovascular events or peripheral vascular disease [17]. The charlson comorbidity score was used to evaluate the comorbidity condition of patients enrolled [18]. Male with SUA over 420 μmol/L or female over 360 μmol/L was identified as hyperuricemia status. The data of mGFR, Kt/V, CCL and nPCR were obtained using the PD Adequest software 2.0 (Baxter, Deerfield, IL). The body surface area (BSA) was calculated on the basis of the widely used DuBois & DuBois formula [19], and the UA clearance was calculated according to the formulate as follows:

Statistical analysis
Patients enrolled were divided into 2 groups according to the peritoneal UA clearance. Data were presented as mean standard deviation for normally distributed continuous variables, median (interquartile range) for quantitative variables of skewed distribution and frequencies and percentages for categorical variables.
Difference between the lower and higher peritoneal UA clearance group was analyzed by independent sample T-test for normally distributed continuous variables, Mann-Whitney U test for quantitative variables of skewed distribution and Chi-squared test for categorical variables. Pearson correlation or Spearman rank correlation test were used to evaluate the correlation between variables of normal or skewed distribution, respectively. Multiple linear regression and binary logistic regression were performed to explore the independent influencing factors of continuous and categorical SUA and peritoneal UA clearance, respectively. After excluding the potential effects of multicollinearity, the variables significant in the univariate analysis (P < 0.05) or possessed clinical correlation were selected into the final model. The performance of different small solute removal indicators for predicting higher peritoneal UA clearance were tested in the area under the ROC curve (AUC) analysis. A two-sided P value < 0.05 was regarded as statistically significant. All of the statistical analyses were conducted in the SPSS software (version 20.0).

Patients' characteristics
As shown in Fig. 1 Peritoneal UA removal and its relationship with SUA in PD patients The distribution of SUA in PD patients was shown in Fig. 2a.  Table 2 and Table 3 both revealed that it was not the residual kidney function of mGFR but the peritoneal UA clearance that was independently associated with SUA in PD patients.  The relationship between peritoneal UA clearance and peritoneal transport characteristics The distribution of the peritoneal UA clearance in different peritoneal transport characteristics was shown in Fig. 3a, which indicated a progressively increase of peritoneal UA clearance with the elevating of the peritoneal transport rate. It was

Independent influencing factors of peritoneal UA clearance
As shown in Table 4 Fig. 4).

Discussion
In this study, we found the inverse association between peritoneal UA clearance and SUA level and further revealed that it was the peritoneal UA removal which significantly played a role in SUA control but not the residual kidney function.
Moreover, we demonstrated that lower albumin level, higher peritoneal transporters and higher peritoneal CCL were independently associated with greater peritoneal UA clearance.
To our knowledge, this is the first systematic study specifically focusing on the UA clearance and its independent influencing factors in PD regimen. As a small molecular solute, the vast majority of UA is in the ionized form and no more than 5% of the circulating UA is albumin bound [4]. Since UA presents as high hydrophilic and has a sieving coefficient of 1.01 which allows it to easily diffuse through the dialysis membrane, it is speculated to be sufficiently cleared by PD therapy to some extent [2,9]. The previous study found that UA clearance was inversely proportional to the dwell time on PD, in which the average UA mass transfer for a dwell time of Few studies have focused on the UA removal in PD, among which the average UA clearance in PD was revealed to be positively proportional to the exchange volume and flow rate [13] while inversely proportional to the dwell time [20]. In this study,

Conclusions
In summary, the UA removal in PD patients was deemed to be more rely on the peritoneal clearance but not the residual kidney function. The peritoneal CCL appears to be optimal indicator for assessing UA removal in PD for its close resemblance of removal process through the dialysis membrane. The lower albumin level, greater peritoneal CCL and higher transporters were revealed to be independently associated with greater peritoneal UA clearance.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
The study has conformed to the ethical principles of the Declaration of Helsinki and was approved by the Clinical Research Ethics Committee of the First Affiliated Hospital of Sun Yat-Sen University. And all patients had signed informed consent.

Consent for publication
Not applicable.

Competing interests
The authors declare no competing interests.

15.
Robson M, Oreopoulos DG, Izatt S, Ogilvie R, Rapoport A, deVeber GA.  The correlation between SUA and dialytic UA clearance. A. The distribution of SUA in PD patie Figure 3 The effects of the peritoneal transport characteristics on peritoneal UA clearance. A. The dis Figure 4 The performance of different small solute removal indicators for predicting higher peritoneal

Results
Patients' characteristics As shown in Fig. 1, there were 180 patients (52.8% of male) participating in the study in total.
Peritoneal UA removal and its relationship with SUA in PD patients The distribution of SUA in PD patients was shown in Fig. 2a. The average mass transfer of urea, creatinine or UA with 2L of 2.5% dextrose dialysate for a dwell time of 0 or 4 hours of all PD patients were described in detail in Fig. 2b. Similar to the small molecular of urea and creatinine, the peritoneal mass transfer of UA was remarkably declined with the lengthening of dwell time. Different from the positive correlation between average UA mass transfer for 4 hours' dwell and SUA (r = 0.55, P < 0.001), the peritoneal UA clearance was found to be inversely correlated to the SUA (r=-0.25, P = 0.001) level ( Fig. 2c and 2d). In comparison to normal SUA group, the hyperuricemia group showed significantly lower peritoneal UA clearance (39.1 ± 6.2 vs. 42.0 ± 8.0 L/week/1.73 m 2 , P = 0.008). The further analysis of multiple linear regression (β -0.21, 95%CI -0.07, -0.006, P = 0.02) and binary logistic regression (OR 0.91, 95%CI 0.84, 0.98, P = 0.02) shown in Table 2 and Table 3 both revealed that it was not the residual kidney function of mGFR but the peritoneal UA clearance that was independently associated with SUA in PD patients. Table 2 Associated factors of SUA in multiple linear regression.  The relationship between peritoneal UA clearance and peritoneal transport characteristics The distribution of the peritoneal UA clearance in different peritoneal transport characteristics was shown in Fig. 3a, which indicated a progressively increase of peritoneal UA clearance with the elevating of the peritoneal transport rate. It was revealed that the higher (H or HA) Independent influencing factors of peritoneal UA clearance As shown in   Fig. 4).

Discussion
In this study, we found the inverse association between peritoneal UA clearance and SUA level and further revealed that it was the peritoneal UA removal which significantly played a role in SUA control but not the residual kidney function. Moreover, we demonstrated that lower albumin level, higher peritoneal transporters and higher peritoneal CCL were independently associated with greater peritoneal UA clearance.
To our knowledge, this is the first systematic study specifically focusing on the UA clearance and its independent influencing factors in PD regimen. As a small molecular solute, the vast majority of UA is in the ionized form and no more than 5% of the circulating UA is albumin bound [4]. Since UA presents as high hydrophilic and has a sieving coefficient of 1.01 which allows it to easily diffuse through the dialysis membrane, it is speculated to be sufficiently cleared by PD therapy to some extent [2,9]. The previous study found that UA [20,22]. Furthermore, the peritoneal albumin loss was demonstrated to be positively associated with peritoneal CCL in a cross-sectional study including 351 PD patients [23]. Therefore, the mechanism of the inverse association between the peritoneal UA clearance and serum albumin level may be elucidated as follows: the greater peritoneal UA clearance itself signifying more removal of albumin from peritoneum caused lower circulating albumin reserves, which was mainly on account of the inadequate albumin synthesis to compensate the peritoneal albumin loss [20]. Therefore, the peritoneal albumin loss should be taken into consideration while optimizing dialysis prescription for efficient solute removal. However, whether there existed a causal relationship between the lower albumin status and greater peritoneal UA clearance or not remained to be explored. Besides, we found that the mGFR was significantly lower in higher peritoneal UA clearance group (2.3 ± 2.3 vs. 3.8 ± 2.4 mL/min/1.73 m 2 , P < 0.001), which seemed to show that the worse residual renal function, the greater peritoneal UA removal. Then in the multiple linear regression model, especially after adjusting the peritoneal CCL, the mGFR was shown to positively associated with the peritoneal UA clearance at a near statistical significance (P = 0.07). It was speculated that that the raw negative correlation between the residual renal function and peritoneal UA removal (r=-0.42, P < 0.001) may be aroused by confounding effects of the relatively larger dialysis dose, glucose concentration and peritoneal CCL in the PD patients with worse renal function compensating for insufficient renal solute removal.
Larger sample size was requisite for clarifying relationship between the residual renal function and peritoneal UA removal in PD more accurately.
There are some limitations in our study. Firstly, the cross-sectional observational study itself figures out merely the association relationship but not causal relationship. Secondly, we haven't explored the effect of different PD modality such as continuous cyclic peritoneal dialysis (CCPD), automated peritoneal dialysis (APD) or different exchange flow rate on the peritoneal UA clearance. Thirdly, the dialysis vintage of patients enrolled were relatively short which may lead to bias for inadequate and unstable dialysis. In spite of the limitation above, this was the first study systematically exploring the independent influencing factors of peritoneal UA clearance in relatively large PD sample. We collected common small solute removal indicators at the same time period for further comparison and analysis which has significant guiding value in optimizing prescription for better UA control in PD patients.
Moreover, we excluded the PD patients with history of taking UA lowering agents to study UA clearance in PD regimen more specifically.

Conclusions
In summary, the UA removal in PD patients was deemed to be more rely on the peritoneal clearance but not the residual kidney function. The peritoneal CCL appears to be optimal indicator for assessing UA removal in PD for its close resemblance of removal process through the dialysis membrane. The lower albumin level, greater peritoneal CCL and higher transporters were revealed to be independently associated with greater peritoneal UA clearance.  Figure 1 The flow chart for enrollment process of the PD patients in the study. HD, hemodialysis; PD, perito The correlation between SUA and dialytic UA clearance. A. The distribution of SUA in PD patients e Figure 3 The effects of the peritoneal transport characteristics on peritoneal UA clearance. A. The distribut Figure 4 The performance of different small solute removal indicators for predicting higher peritoneal UA c