Peritoneal Solute Transport Rate as an Independent Risk Factor for Total and Cardiovascular Mortality in a Population of Peritoneal Dialysis Patients

Advances in Peritoneal Dialysis, Vol. 30, 2014 Katarzyna Janda, 1 Marcin Krzanowski, 1 Paulina Dumnicka, 2 Beata Kuśnierz Cabala, 3 Przemysław Miarka, 1 Władysław Sułowicz 1 Peritoneal Solute Transport Rate as an Independent Risk Factor for Total and Cardiovascular Mortality in a Population of Peritoneal Dialysis Patients The aim of the present study was to assess the influence of peritoneal permeability expressed as the dialysate-to-plasma ratio of creatinine (D/P Cr) on total and cardiovascular (CV) mortality in a population of peritoneal dialysis (PD) patients during a 6-year observation period. The study recruited 55 patients (mean age: 53 years) treated with PD for a median of 24 months. Hematology parameters and serum albumin were determined using routine methods. Tumor necrosis factor α (TNF-α) and transforming growth factor β1 (TGF-β1) were determined by high-sensitivity ELISA. Peritoneal transport characteristics were identified using D/P Cr reference values after a peritoneal equilibration test. During the 6-year observation period, 22 patients (40%) died, mostly from CV complications (77% of deaths). In multiple Cox regression, D/P Cr and dialysate volume at PD initiation predicted total [hazard ratio (HR): 1.57; p = 0.02; and HR: 1.20; p = 0.04 respectively] and CV mortality (HR: 1.65; p = 0.02; and HR: 1.23; p = 0.05 respectively) independent of age, dialysis therapy duration, serum albumin concentration, dialysis adequacy measures, TGF-β1, and TNF-α. Additionally, TNF-α was independently associated with all-cause and CV mortality, and albumin, with all-cause mortality. Baseline D/P Cr was a strong independent marker of survival in PD patients. Baseline D/P Cr and From: 1 Chair and Department of Nephrology, 2 Department of Medical Diagnostics, and 3 Chair of Clinical Biochemistry, Jagiellonian University, Collegium Medicum, Cracow, Poland. dialysate volume were independent risk factors for total and CV mortality in the PD population and could be significant for assessing CV risk in this population. Key words Cardiovascular mortality, peritoneal solute transport rate, total mortality Introduction Cardiovascular (CV) diseases are the most frequent cause of comorbidity and mortality in patients with end-stage renal disease. Traditional risk factors for CV disease such as advanced age, diabetes mellitus, fluid volume overload, hypertension, and hyperlipidemia frequently coexist in patients with chronic kidney disease. The various dialysis modalities, such as continuous ambulatory peritoneal dialysis (CAPD), automated peritoneal dialysis (APD), and hemodialysis have different effects on fluid volume control. Automated peritoneal dialysis (PD) shares some characteristics with CAPD, and with the shorter dwell times used in APD, higher ultrafiltration volumes can be obtained (1). Peritoneal transport characteristics play an important role in determining morbidity, mortality, and management in PD patients. Small-solute peritoneal transport is assessed by the peritoneal equilibration test (2). Peritoneal transport varies from one patient to another and can change over time in the same patient. Patients with high small-solute peritoneal transport have an increased risk of morbidity and mortality despite their more rapid diffusive clearance of urea and creatinine. In patients with high peritoneal small-solute transport, the increased risk of mortality can potentially be mitigated by combining

16 PSTR as a Predictor for Mortality in PD Patients optimization of the short dwell times of APD with icodextrin use, rather than by using CAPD (3). Increases in the peritoneal solute transport rate (PSTR) are associated with an increased risk of mortality and a tendency to increased technique failure (4 7). Development of a high PSTR has been attributed to numerous factors, including hypoalbuminemia, comorbid disease, duration of PD (with exposure to high glucose), number and severity of peritonitis episodes, loss of residual renal function (RRF), and bioincompatible dialysis fluid (8 10). The aim of the present study was to assess the influence of peritoneal permeability expressed as the dialysate-to-plasma ratio of creatinine (D/P Cr) on total and CV mortality in a population of PD patients during a 6-year observation period. Methods The study group consisted of 55 patients (25 women, 30 men; mean age: 53 ± 13 years) being treated with PD (27 on APD, 28 on CAPD). Median duration of dialysis treatment was 24 months (range: 4 100 months). On the day of blood collection, serum albumin was measured using an automated clinical chemistry analyzer (Modular P: Roche Diagnostics, Mannheim, Germany), and hematology parameters were assessed using an automated hematology analyzer (Sysmex XE 2100: Sysmex Corporation, Kobe, Japan). Sera samples for other laboratory tests were aliquoted and stored at more than 70 C. Measurements of tumor necrosis factor α (TNF-α) and transforming growth factor β1 (TGF-β1) were performed using highsensitivity ELISA kits (R&D Systems, Minneapolis, MN, U.S.A.). References ranges for the tests (Table I) represent the values set by the manufacturers. Peritoneal transport characteristics were identified using D/P Cr reference values after a peritoneal equilibration test (2,11). In addition, the PD dose was assessed using weekly Kt/V, weekly creatinine clearance (CCr), RRF, and dialysate volume. Body mass index was calculated using the Quetelet index. Data on mortality were collected over a 6-year (72-month) period. All deaths occurred in hospital, and the causes of death were determined according to disease history. The study was approved by the Bioethics Committee of Jagiellonian University, and all patients provided informed consent for their participation. Statistical analysis Number and proportion of patients in a group are reported for categorical variables, and mean ± standard deviation or median with lower and upper quartiles are reported for continuous variables, according to distribution. The Shapiro Wilks test was used to assess normality. Spearman coefficients were used to assess simple correlations (at least 1 variable being non-normally distributed). Survival curves were computed by the Kaplan Meier method. Unadjusted and adjusted hazard ratios (HRs) for all-cause and cardiovascular mortality were estimated by Cox proportional regression and are reported with 95% confidence intervals (CIs). All tests were two-sided, and results at p 0.05 were considered significant. The Statistica 9.0 statistical software application (StatSoft, Tulsa, OK, U.S.A.) was used for all computations. Results Table I presents clinical and biochemical characteristics and dialysis adequacy parameters for the study group. During the 6-year observation period, 22 patients died (17 from CV causes, 3 from encapsulating peritoneal sclerosis, and 2 from infections), 19 patients (35%) were transferred to hemodialysis, and 11 (20%) underwent renal transplantation. Median overall survival was 37 months. In univariate analysis, the D/P Cr was significantly positively associated with all-cause mortality (HR per 0.1 increase: 1.49; 95% CI: 1.08 to 2.04; p = 0.01) and approached statistical significance for mortality from CV causes (HR per 0.1 increase: 1.43; 95% CI: 0.99 to 2.04; p = 0.054). Figure 1 presents cumulative survival for patients having high-average and high (HA/H) peritoneal permeability (D/P Cr 0.65) compared with those having low-average and low (LA/L) peritoneal permeability (D/P Cr < 0.65). Significant correlations with mortality were also shown for dialysate volume (HR for all-cause mortality: 1.21; 95% CI: 1.04 to 1.41; p = 0.02; HR for CV mortality: 1.21; 95% CI: 1.02 to 1.45; p = 0.03). Another significant predictor of all-cause and CV mortality was TNF-α (HR: 1.57; 95% CI: 1.14 to 2.18; p = 0.006; HR: 1.70; 95% CI: 1.19 to 2.44; p = 0.004 respectively). Albumin was associated with overall survival (HR: 0.89; 95% CI: 0.81 to 0.99; p = 0.03). Using multivariable analysis, we studied associations with survival in the context of dialysis adequacy

Janda et al. 17 table i Characteristics of the study group Characteristic Value [n (%)] Mean or median IQR Min Max Reference Patients 55 Age (years) 53±13 19 75 Sex Men 30 (55) Women 25 (45) Dialysis duration (months) 24 15 51 4 100 Hypertension 47 (85) Ischemic heart disease 21 (38) Observation period (months) 20 6 45 1 72 All-cause mortality 22 (40) Cardiovascular mortality 17 (31) Body mass index (kg/m 2 ) 25.3±4.1 17.2 34.2 20 25 Hemoglobin (g/dl) 11.9±1.6 7.7 15.3 Women: 11.0 15.0 Men: 12.0 17.0 White blood cells ( 10 3 /μl) 7.48±2.43 2.77 14.33 4.0 10.0 Albumin (g/l) 37.9±4.9 20.0 47.0 35 50 TNF-α (pg/ml) 2.89 2.26 3.54 1.37 7.91 0 4.71 TGF-β1 (ng/ml) 4.22 3.22 5.63 1.92 20.06 0.903 1.654 Weekly Kt/V 2.22 1.93 2.63 1.34 5.26 >1.7 Weekly CCr (L/1.73 m 2 ) 68.0 58.8 95.7 43.2 213.8 >45 RRF (ml/min/1.73 m 2 ) 1.72 0.04 4.62 0 11.33 D/P creatinine 0.63±0.13 0.38 0.91 0.34 1.03 Dialysate volume (dl/24 h) 8.0 8.0 10.0 6.0 15.5 IQR = interquartile range; Min = minimum; Max = maximum; TNF-α = tumor necrosis factor α; TGF-β1 = transforming growth factor β1; CCr = creatinine clearance; RRF = residual renal function. (weekly Kt/V and CCr), RRF, TGF-β1, age, and dialysis duration. However, the correlation of D/P Cr with dialysate volume was so highly significant (R = 0.45, p = 0.0005) that, to avoid redundancy, we constructed separate multiple Cox regression models including either D/P Cr or dialysate volume. In multivariable Cox regression, D/P Cr was an independent predictor of total and CV mortality (Table II). Similarly, dialysate volume was shown to predict mortality independently of the other studied variables (Table III). Additionally, proinflammatory TNF-α was independently associated with total and CV mortality (Tables II and III). figure 1 Kaplan Meier survival curves for all-cause mortality in patients with a dialysate-to-plasma creatinine ratio of 0.65 or more (solid line) and less than 0.65 (dashed line), log-rank p = 0.08. Discussion In PD, the PSTR is a major determinant of patient survival. Our study shows that in the population of PD patients, D/P Cr and dialysate volume both

18 PSTR as a Predictor for Mortality in PD Patients table ii Cox regression models including dialysate-to-plasma (D/P) creatinine Independent variable All-cause mortality Cardiovascular mortality HR 95% CI p Value HR 95% CI p Value D/P creatinine Change of 0.1 1.57 1.08 to 2.28 0.02 1.65 1.07 to 2.53 0.02 Albumin Grams per liter 0.89 0.80 to 0.98 0.02 0.93 0.81 to 1.06 0.3 TNF-α Picograms per milliliter 1.83 1.16 to 2.88 0.009 1.87 1.14 to 3.07 0.01 TGF-β1 Nanograms per milliliter 0.98 0.84 to 1.14 0.8 0.96 0.81 to 1.16 0.7 Weekly Kt/V 1.39 0.38 to 5.13 0.6 1.51 0.33 to 6.91 0.6 Weekly CCr Liters per 1.73 m 2 0.98 0.95 to 1.02 0.4 0.99 0.96 to 1.03 0.6 RRF Milliliters per minute per 1.73 m 2 1.10 0.77 to 1.56 0.6 1.01 0.70 to 1.45 0.9 Dialysis duration Months 1.00 0.97 to 1.02 0.7 0.99 0.96 to 1.01 0.3 Age Years 1.00 0.96 to 1.05 0.9 1.01 0.96 to 1.07 0.6 HR = hazard ratio; CI = confidence interval; TNF-α = tumor necrosis factor α; TGF-β1 = transforming growth factor β1; CCr = creatinine clearance; RRF = residual renal function. table iii Cox regression models including dialysate volume Independent variable All-cause mortality Cardiovascular mortality HR 95% CI p Value HR 95% CI p Value Dialysate volume Deciliters in 24 hours 1.20 1.003 to 1.44 0.04 1.23 1.001 to 1.52 0.05 Albumin Grams per liter 0.86 0.78 to 0.95 0.004 0.89 0.78 to 1.02 0.08 TNF-α Picograms per milliliter 1.72 1.07 to 2.74 0.02 1.80 1.08 to 3.00 0.02 TGF-β1 Nanograms per milliliter 0.99 0.83 to 1.19 0.9 0.98 0.80 to 1.21 0.8 Weekly Kt/V 1.04 0.27 to 3.96 0.9 1.21 0.25 to 5.65 0.8 Weekly CCr Liters per 1.73 m 2 0.99 0.96 to 1.03 0.7 1.00 0.97 to 1.03 0.8 RRF Milliliters per minute per 1.73 m 2 1.09 0.78 to 1.54 0.6 0.98 0.69 to 1.39 0.9 Dialysis duration Months 1.00 0.98 to 1.02 0.7 1.00 0.97 to 1.02 0.7 Age Years 1.01 0.96 to 1.06 0.7 1.02 0.97 to 1.08 0.4 HR = hazard ratio; CI = confidence interval; TNF-α = tumor necrosis factor α; TGF-β1 = transforming growth factor β1; CCr = creatinine clearance; RRF = residual renal function. predicted CV and all-cause mortality independently of inflammatory cytokines, albumin level, dialysis adequacy (weekly Kt/V and CCr), RRF, age, and dialysis therapy duration. Authors from the Karolinska Institutet indicated that intraperitoneal and systemic inflammation increase in PD patients during the first year of therapy. Intraperitoneal and systemic inflammation might be interrelated, and the interleukin 6 (IL-6) system might be a link. Particularly in the early phase of PD treatment, IL-6 is associated with PSTR, and smalland large-solute transport are linked. Inflammation might be responsible for the development of a high PSTR, which could potentially be a reason for the high mortality in patients with a high PSTR (12). Higher solute transport at the start of PD might be associated with the risk of peritonitis. Gołembiewska et al. (13) showed that solute transport measured as D/P Cr during the peritoneal equilibration test was significantly higher in the group of patients who had experienced a peritonitis episode than in the group of peritonitis-free patients. Matsuo et al. (14) evaluated risk factors and mortality in PD patients. In their group of 98 patients, the leading causes of death were CV and infectious disease. Patients who died were significantly older and more frequently had a history of CV disease. Their serum albumin was significantly lower, and their D/P Cr was higher. The authors postulated that diabetic nephropathy, a history of CV disease,

Janda et al. 19 and higher peritoneal permeability at PD initiation are independent risk factors for death in patients starting PD. In another study, Rodrigues et al. (15) revealed that fast peritoneal transport at baseline was not associated with markers of systemic inflammation, nor was it predictive of worse patient survival in incident PD patients. In a population with preserved RRF and an absence of serious baseline comorbidity, it was not predictive of a worse prognosis. Fast transporters did not present higher levels of C-reactive protein (CRP) or serum IL-6. Patients with more than 2 comorbidities had lower levels of plasma albumin, significantly higher median levels of serum IL-6, and a greater intima media thickness. Multivariate analysis confirmed that baseline peritoneal transport was not a significant determinant of patient survival, but comorbidity score remained significant. In our multivariable Cox regression (adjusted for all studied variables and for age and dialysis therapy duration), D/P Cr at PD initiation was an independent predictor of total and CV mortality. In dialyzed patients (CAPD or hemodialysis), the concentrations of inflammatory factors, especially high-sensitivity CRP and TNF-α, are increased. In a study by Kir et al. (16), the highest increases in TNF-α were observed in patients on CAPD. In our multivariable analysis, an increased TNF-α concentration at baseline was independently associated with increased all-cause and CV mortality. Peritoneal transport status is one of the main determinants of dialysis adequacy and of dialysisrelated complications in chronic kidney disease patients who undergo PD (17). Zhe et al. (18) showed that carotid femoral pulse wave velocity, established as a CV risk factor, was positively associated with patient age, time on PD, diabetes status, D/P Cr, pulse pressure, and the ratio of extracellular water to total body water. In multivariate regression analysis, pulse wave velocity was independently determined by the ratio of extracellular water to total body water, pulse pressure, age, and D/P Cr, suggesting that greater aortic stiffness and an increased rate of peritoneal small-solute transport might be linked. In another study, Sezer et al. (19) investigated the relationship between peritoneal transport characteristics and known promoters of atherosclerosis in PD patients. After 36 months, those authors observed that, compared with LA/L transport, HA/H transport was associated with lower albumin, higher CRP, and a higher need for recombinant human erythropoietin. During follow-up, 28 of 84 patients showed an atherosclerosis-related event (myocardial infarction, coronary artery disease diagnosis by angiography or myocardial scintigraphy, cerebrovascular accident, or development of clinically evident peripheral arterial disease). Of those 28 patients, 22 were in the HA/H group (constituting 43.1% of that group), and only 6 were in the LA/L group (constituting 18.1% of the group, p < 0.01). Re-analysis of the 18 patients with atherosclerosis-related events and high CRP levels (>10 mg/l) showed that 15 were in the HA/H group, and 3 were in the LA/L group. Of patients with an atherosclerotic event, 68% belonging to the HA/H group and 50% belonging to the LA/L group also had chronic inflammation (p < 0.001). A Pearson correlation analysis showed that D/P Cr was positively correlated with 36-month mean CRP concentration and negatively correlated with 36-month mean serum albumin. A high-transport peritoneal membrane characteristic was thus shown to be a risk factor for an inflammatory state in patients with end-stage renal disease. Compared with their low-transport counterparts, high-transport patients were at an increased risk of atherosclerosis through chronic inflammation (19). Sawai et al. (20) investigated the relationships of local peritoneal inflammation, angiogenesis, and systemic inflammation with baseline permeability. Peritoneal biopsy specimens from 42 pre-dialysis uremic patients and 11 control subjects were investigated. Compared with control peritoneum, predialysis uremic peritoneum showed infiltration by CD68+ macrophages and mast cells. Baseline D/P Cr correlated with the density of CD68+ macrophages, IL-6 positive cells, CD31-positive blood vessels, and serum albumin. On multiple linear regression analysis, the number of CD68+ macrophages in peritoneum was an independent predictor of baseline peritoneal permeability. Our study shows that PSTR is an important determinant of survival in PD patients. Additionally, high PSTR at PD initiation is associated with an increased risk for total and CV mortality. Conclusions Baseline D/P Cr represents a strong independent marker of survival of PD patients. Baseline D/P Cr and dialysate volume are independent risk factors for

20 PSTR as a Predictor for Mortality in PD Patients total mortality and for CV mortality in the PD population and can be significant markers in the assessment of CV risk in that population. Acknowledgments This work was supported by a department grant (no. K/ZDS/000597) and was previously presented as a poster during the 49th ERA-EDTA Congress; Paris, France; May 24 27, 2012. Disclosures The authors have no financial conflicts of interest to declare. References 1 Paniagua R, Ventura MD, Avila Díaz M, et al. NTproBNP, fluid volume overload and dialysis modality are independent predictors of mortality in ESRD patients. Nephrol Dial Transplant 2010;25:551 7. 2 Twardowski ZJ. Clinical value of standardized equilibration tests in CAPD patients. Blood Purif 1989;7:95 108. 3 La Milia V, Limardo M, Cavalli A, Crepaldi M, Locatelli F. Transport of peritoneal membrane assessed before and after the start of peritoneal dialysis. Nephrol Dial Transplant 2009;24:2894 8. 4 Churchill DN, Thorpe KE, Nolph KD, Keshaviah PR, Oreopoulos DG, Pagé D. Increased peritoneal membrane transport is associated with decreased patient and technique survival for continuous peritoneal dialysis patients. The Canada USA (CANUSA) Peritoneal Dialysis Study Group. J Am Soc Nephrol 1998;9:1285 92. 5 Brimble KS, Walker M, Margetts PJ, Kundhal KK, Rabbat CG. Meta-analysis: peritoneal membrane transport, mortality, and technique failure in peritoneal dialysis. J Am Soc Nephrol 2006;17:2591 8. 6 Rumpsfeld M, McDonald SP, Johnson DW. Higher peritoneal transport status is associated with higher mortality and technique failure in the Australian and New Zealand peritoneal dialysis patient populations. J Am Soc Nephrol 2006;17:271 8. 7 Cueto Manzano AM. Rapid solute transport in the peritoneum: physiologic and clinical consequences. Perit Dial Int 2009;29(suppl 2):S90 5. 8 Margetts PJ, McMullin JP, Rabbat CG, Churchill DN. Peritoneal membrane transport and hypoalbuminemia: cause or effect? Perit Dial Int 2000;20:14 18. 9 Davies SJ, Phillips L, Naish PF, Russell GI. Quantifying comorbidity in peritoneal dialysis patients and its relationship to other predictors of survival. Nephrol Dial Transplant 2002;17:1085 92. 10 Oh KH, Moon JY, Oh J, et al. Baseline peritoneal solute transport rate is not associated with markers of systemic inflammation or comorbidity in incident Korean peritoneal dialysis patients. Nephrol Dial Transplant 2008;23:2356 64. 11 Krediet RT, Struijk DG. Peritoneal dialysis membrane evaluation in clinical practice. Contrib Nephrol 2012;178:232 7. 12 Pecoits Filho R, Carvalho MJ, Stenvinkel P, Lindholm B, Heimbürger O. Systemic and intraperitoneal interleukin-6 system during the first year of peritoneal dialysis. Perit Dial Int 2006;26:53 63. 13 Gołembiewska E, Safranow K, Kabat Koperska J, Ciechanowski K, Romanowski M. Solute transport at the start of peritoneal dialysis and the risk of peritonitis. Adv Clin Exp Med 2013;22:77 83. 14 Matsuo N, Maruyama Y, Terawaki H, et al. Risk factors for death in patients starting PD for their first renal replacement therapy [Japanese]. Nihon Jinzo Gakkai Shi 2009;51:38 43. 15 Rodrigues AS, Almeida M, Fonseca I, et al. Peritoneal fast transport in incident peritoneal dialysis patients is not consistently associated with systemic inflammation. Nephrol Dial Transplant 2006;21:763 9. 16 Kir HM, Eraldemir C, Dervisoglu E, Caglayan C, Kalender B. Effects of chronic kidney disease and type of dialysis on serum levels of adiponectin, TNFalpha and high sensitive C-reactive protein. Clin Lab 2012;58:495 500. 17 Kim YL. Update on mechanisms of ultrafiltration failure. Perit Dial Int 2009;29(suppl 2):123 7. 18 Zhe XW, Tian XK, Chen W, et al. Association between arterial stiffness and peritoneal small solute transport rate. Artif Organs 2008;32:416 19. 19 Sezer S, Tutal E, Arat Z, et al. Peritoneal transport status influence on atherosclerosis/inflammation in CAPD patients. J Ren Nutr 2005;15:427 34. 20 Sawai A, Ito Y, Mizuno M, et al. Peritoneal macrophage infiltration is correlated with baseline peritoneal solute transport rate in peritoneal dialysis patients. Nephrol Dial Transplant 2011;26:2322 32. Corresponding author: Katarzyna Janda, md phd, Department of Nephrology, Jagiellonian University, Collegium Medicum, Kopernika 15C Street, Cracow 31-501 Poland. E-mail: kasiajanda@op.pl