Resource limitations and strategies for the treatment of uremia. A dialysis unit in the Himalayan foothills.

India has made great strides in health care since gaining independence in 1947. Much still needs to be done. Scarce health care resources are directed at priorities that include infant and maternal mortality, immunizations, malnutrition, communicable disease prevention, and access to protected water and sanitation. There is, therefore, no government reimbursement for treatment of patients with ESRD (estimated incidence of 100 per million population). Transplantation, the modality of choice in India, benefits only 2-3% of ESRD patients. Hemodialysis is primarily used short-term for pretransplant stabilization. A very small minority of patients is placed on maintenance hemodialysis or CAPD. The annual cost of renal replacement therapies is more than 10 times the per capita gross national product. Financial constraints pose ethical problems for the nephrologist related to adequate prescriptions and compliance. Preventing the progression of kidney disease and reducing the cost of disposables through indigenous manufacture are initiatives that need to be explored.

Imprecision of the hemodialysis dose when measured directly from urea removal. Hemodialysis Study Group.

BACKGROUND: The postdialysis blood urea nitrogen (BUN; Ct) is a pivotal parameter for assessing hemodialysis adequacy by conventional blood-side methods, but Ct is relatively unstable because of hemodialysis-induced disequilibrium. The uncertainty associated with this method is potentially reduced or eliminated by measuring urea removed on the dialysate side, a more direct approach that can determine adequacy from the fraction of urea removed and by substituting an estimate of the equilibrated postdialysis BUN (Ceq) for Ct. For a patient with a known urea volume (V), Ceq, the equilibrated Kt/V (eKt/V), and the solute removal index (SRI) can be calculated from the predialysis BUN (C0), total urea nitrogen removed (A), and V from simple mass balance calculations (dialysate/volume method). However, a theoretical error analysis showed that relatively small errors in A, C0, or V are magnified when SRI or eKt/V is calculated using this method, especially at higher eKt/V values (for example, if eKt/V = 1.4 per dialysis, a 7% dialysate collection error causes a 20% error in eKt/V). METHODS: During three to four baseline dialyses in each of 39 patients enrolled in the pilot phase of the HEMO Study, "A" was measured using an instrument that sampled dialysate frequently (Biostat), and V was calculated from A, C0, and Ceq (median CV for V = 5.6%). The mean V was then applied to the dialysate/volume method to estimate eKt/V and SRI during two to five subsequent dialyses per patient (comparison dialyses). The accuracy and precision of these estimates were assessed by comparing them with eKt/V and SRI derived from a direct measurement of Ceq drawn 30 minutes after dialysis (reference method), from mathematical curve-fitting of sequential dialysate urea concentrations (dialysate curve-fit method), and from another blood-side method that estimates eKt/V from single pool Kt/V and the fractional rate of solute removal (rate method): eKt/V = spKt/V - 0.6.K/V + 0.03. RESULTS: During 128 comparison dialyses, median absolute errors for calculated eKt/V compared with the reference method were 0.169, 0.061, and 0.071 for the dialysate/volume method, the rate method, and the dialysate curve-fitting method, respectively. The corresponding correlation coefficients were 0.47, 0.88, and 0.81. For SRI, median absolute errors were 0.044, 0.018, and 0.027, and the correlation coefficients were 0.54, 0.85, and 0.74 for the three methods. CONCLUSIONS: The precision of eKt/V and SRI measurements was significantly lower for the dialysate/volume method compared with the blood-side methods. Inclusion of the dialysate curve analysis provided by the Biostat restored precision to the dialysate method to a level comparable to that of the blood-side methods. New techniques employing dialysate urea analysis should include a concentration profile to avoid these inherent methodological errors and assure the accuracy of eKt/V and SRI.

Timely initiation of dialysis: a urea kinetic approach.

The traditional approach of initiating dialysis when the patient begins to manifest uremic symptoms may result in the development of significant malnutrition with detrimental effects on subsequent morbidity and mortality. The recently issued Dialysis Outcome Quality Initiative guidelines suggest that dialysis be initiated when the Kt/V from residual renal function decreases to less than 2.0. We have used the urea kinetic model to show how dialytic dose can be titrated to compensate for declining renal function while maintaining a constant total dose of delivered therapy (Kt/V = 2.0). For hemodialysis (HD), we show that initiating dialysis with once-weekly therapy may be a viable option only for a few months, being replaced by twice-weekly and subsequently with the more typical regimen of thrice-weekly HD. We recommend that the patient be directly initiated with twice-weekly HD to minimize wide swings in the serum concentrations of small-molecular-weight solutes. With continuous ambulatory peritoneal dialysis (CAPD), a hypothetical average-sized patient with high-average transport can be maintained for approximately 8 months with a single 2.5-L nocturnal exchange and from 8 to 17 months with two nocturnal exchanges of 2.5 L each. The use of nocturnal exchanges allows more normal daytime activities and is less intrusive on patient lifestyle. We have shown that both HD and CAPD regimens can be successfully adjusted to achieve a constant total Kt/V of 2.0 for 5 or more years, although CAPD may provide a smoother transition from no dialysis to a complete 10-L regimen.

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.

The objective of this study was to evaluate the association of peritoneal membrane transport with technique and patient survival. In the Canada-USA prospective cohort study of adequacy of continuous ambulatory peritoneal dialysis (CAPD), a peritoneal equilibrium test (PET) was performed approximately 1 mo after initiation of dialysis; patients were defined as high (H), high average (HA), low average (LA), and low (L) transporters. The Cox proportional hazards method evaluated the association of technique and patient survival with independent variables (demographic and clinical variables, nutrition, adequacy, and transport status). Among 606 patients evaluated by PET, there were 41 L, 192 LA, 280 HA, and 93 H. The 2-yr technique survival probabilities were 94, 76, 72, and 68% for L, LA, HA, and H, respectively (P = 0.04). The 2-yr patient survival probabilities were 91, 80, 72, and 71% for L, LA, HA, and H, respectively (P = 0.11). The 2-yr probabilities of both patient and technique survival were 86, 61, 52, and 48% for L, LA, HA, and H, respectively (P = 0.006). The relative risk of either technique failure or death, compared to L, was 2.54 for LA, 3.39 for HA, and 4.00 for H. The mean drain volumes (liters) in the PET were 2.53, 2.45, 2.33, and 2.16 for L, LA, HA, and H, respectively (P < 0.001). After 1 mo CAPD treatment, the mean 24-h drain volumes (liters) were 9.38, 8.93, 8.59, and 8.22 for L, LA, HA, and H, respectively (P < 0.001); the mean 24-h peritoneal albumin losses (g) were 3.1, 3.9, 4.3, and 5.6 for L, LA, HA, and H, respectively (P < 0.001). The mean serum albumin values (g/L) were 37.8, 36.2, 33.8, and 32.8 for L, LA, HA, and H, respectively (P < 0.001). Among CAPD patients, higher peritoneal transport is associated with increased risk of either technique failure or death. The decreased drain volume, increased albumin loss, and decreased serum albumin concentration suggest volume overload and malnutrition as mechanisms. Use of nocturnal cycling peritoneal dialysis should be considered in H and HA transporters.

Comparison of methods to predict equilibrated Kt/V in the HEMO Pilot Study.

The ongoing HEMO Study, a National Institutes of Health (NIH) sponsored multicenter trial to test the effects of dialysis dosage and membrane flux on morbidity and mortality, was preceded by a Pilot Study (called the MMHD Pilot Study) designed to test the reliability of methods for quantifying hemodialysis. Dialysis dose was defined by the fractional urea clearance per dialysis determined by the predialysis BUN and the equilibrated postdialysis BUN after urea rebound is completed (eKt/V). In the Pilot Study the blood side standard for eKt/V was calculated from the predialysis, postdialysis, and 30-minute postdialysis BUN. Four techniques of approximating eKt/V that eliminated the requirement for the 30-minute postdialysis sample were also evaluated. The first adjusted the single compartment Kt/V using a linear equation with slope based on the relative rate of solute removal (K/V) to predict eKt/V (rate method). The second and third techniques used equations or mathematical curve fitting algorithms to fit data that included one or more samples drawn during dialysis (intradialysis methods). The fourth technique (dialysate-side) predicted eKt/V from an analysis of the time-dependent profile of dialysate urea nitrogen concentrations (BioStat method; Baxter Healthcare, Inc., Round Lake, IL, USA). The Pilot Study demonstrated the feasibility of conventional and high dose targets of about 1.0 and 1.4 for eKt/V. Based on the blood side standard method, the mean +/- SD eKt/V for patients randomized to these targets was 1.14 +/- 0.11 and 1.52 +/- 0.15 (N = 19 and 16 patients, respectively). Single-pool Kt/Vs were about 0.2 Kt/V units higher. Results were similar when eKt/V was based on dialysate side measurements: 1.10 +/- 0.11 and 1.50 +/- 0.11. The approximations of eKt/V by the three blood side methods that eliminated the delayed 30-minute post-dialysis sample correlated well with eKt/V from the standard blood side method: r = 0.78 and 0.76 for the single-sample (Smye) and multiple-sample intradialysis methods (N = 295 and 229 sessions, respectively) and 0.85 for the rate method (N = 295). The median absolute difference between eKt/V computed using the standard blood side method and eKt/V from the four other methods ranged from 0.064 to 0.097, with the smallest difference (and hence best accuracy) for the rate method. The results suggest that, in a dialysis patient population selected for ability to achieve an equilibrated Kt/V of about 1.45 in less than a 4.5 hour period, use of the pre and postdialysis samples and a kinetically derived rate equation gives reasonably good prediction of equilibrated Kt/V. Addition of one or more intradialytic samples does not appear to increase accuracy of predicting the equilibrated Kt/V in the majority of patients. A method based on dialysate urea analysis and curve-fitting yields results for equilibrated Kt/V that are similar to those obtained using exclusively blood-based techniques of kinetic modeling.

Lower probability of patient survival with continuous peritoneal dialysis in the United States compared with Canada. Canada-USA (CANUSA) Peritoneal Dialysis Study Group.

In a prospective cohort study of 680 incident continuous peritoneal dialysis (PD) patients in North America, dialysis in the United States compared with Canada was associated with a relative risk (RR) of death of 1.93 (95% confidence interval [CI], 1.14 to 3.28). The 2-yr survival probability was 79.7% in Canada and 63.2% in the United States. This difference was not explained by race, age, gender, functional status, insulin-dependent diabetes mellitus, history of cardiovascular disease (CVD), nutritional status, or adequacy of dialysis. Other potential explanatory variables were further evaluated. These included severity of CVD, residual renal function, race, differential transfer to hemodialysis or transplantation, patient compliance, modality selection bias, and incidence of endstage renal disease requiring dialysis. Cardiovascular morbidity and peritonitis probabilities were compared. The CVD severity index was not different between countries; the RR risk associated with dialysis in the United States remained high at 1.87 (95% CI, 1.09 to 3.19). Residual renal function at initiation of dialysis was not different between countries. The 2-yr survival for Caucasians was 77% in Canada and 55% in the United States. There was no difference in the probability of transfer to hemodialysis or transplantation. The RR of a nonfatal cardiovascular event in the United States compared with Canada was 1.80 (95% CI, 1.21 to 2.67). There was no difference in time to first peritonitis. The observed to predicted creatinine ratio, as an estimate of compliance, was 1.13 in Canada and 1.00 in the United States. The prevalence of PD in the study centers was 48% in Canada and 22% in the United States. The incidence of new dialysis patients in 1992 was 100/million population in Canada compared with 211/ million in the United States. The survival difference is not explained by age, gender, insulin-dependent diabetes mellitus, nutritional status, or adequacy of dialysis. Neither is it explained by race, severity of CVD, transfer to hemodialysis, transplantation, or an estimate of compliance. The lower proportion of patients receiving PD in the United States may represent a selection bias of uncertain direction. The higher acceptance rate for dialysis in the United States may explain, in part, the greater cardiovascular morbidity and the decreased survival observed.

Hemodialyzer mass transfer-area coefficients for urea increase at high dialysate flow rates. The Hemodialysis (HEMO) Study.

The dialyzer mass transfer-area coefficient (KoA) for area is an important determinant of urea removal during hemodialysis and is considered to be constant for a given dialyzer. We determined urea clearance for 22 different models of commercial hollow fiber dialyzers (N = approximately 5/model, total N = 107) in vitro at 37 degrees C for three countercurrent blood (Qb) and dialysate (Qd) flow rate combinations. A standard bicarbonate dialysis solution was used in both the blood and dialysate flow pathways, and clearances were calculated from urea concentrations in the input and output flows on both the blood and dialysate sides. Urea KoA values, calculated from the mean of the blood and dialysate side clearances, varied between 520 and 1230 ml/min depending on the dialyzer model, but the effect of blood and dialysate flow rate on urea KoA was similar for each. Urea KoA did not change (690 +/- 160 vs. 680 +/- 140 ml/min, P = NS) when Qh increased from 306 +/- 7 to 459 +/- 10 ml/min at a nominal Qd of 500 ml/min. When Qd increased from 504 +/- 6 to 819 +/- 8 ml/min at a nominal Qh of 450 ml/min, however, urea KoA increased (P < 0.001) by 14 +/- 7% (range 3 to 33%, depending on the dialyzer model) to 780 +/- 150 ml/min. These data demonstrate that increasing nominal Qd from 500 to 800 ml/min alters the mass transfer characteristics of hollow fiber hemodialyzers and results in a larger increase in area clearance than predicted assuming a constant KoA.

Effect of low-frequency ultrasound on peritoneal transport in rabbits.

It has recently been suggested that sonophoresis, or the application of ultrasound (US) in the kilohertz range, could enhance peritoneal mass transport. To examine this hypothesis, six nephrectomized rabbits were exposed to ultrasound while under isoflurane anesthesia. An additional five also had bilateral nephrectomies and were used as a control group. Each group underwent four exchanges of 90 minutes duration with 1.5% dextrose while anesthetized. Dialysate samples were taken at 0, 30, 60, and 90 minutes and assayed for urea, creatinine, glucose, and protein. Blood samples were taken pre- and postexchange. In the US group, 20 kHz ultrasound was applied during exchanges 2 and 3 at 47.5 W and 95 W, respectively, using a Virsonic 475 cell disrupter acoustically coupled to the abdomen through a water column and gel-coated PVC membrane. Results were analyzed by calculating the mass transfer area coefficient (MTAC) and 90-minute D/P values for each exchange. No significant differences were observed in the absolute means of either parameter between the control and US groups. However, when exchanges 2 to 4 were normalized with respect to exchange 1, the resulting urea D/P means were less for the US exchanges compared to the control (p < 0.05). This suggests a possible decrease in transport through US application.

Multicenter clinical validation of an on-line monitor of dialysis adequacy.

Quantitation of hemodialysis by measuring changes in blood solute concentration requires careful timing when taking the postdialysis blood sample to avoid errors from postdialysis rebound and from recirculation of blood through the access device. It also requires complex mathematical interpretation to account for solute disequilibrium in the patient. To circumvent these problems, hemodialysis can be quantified and its adequacy assessed by direct measurement of the urea removed in the dialysate. Because total dialysate collection is impractical, an automated method was developed for measuring dialysate urea-nitrogen concentrations at frequent intervals during treatment. A multicenter clinical trial of the dialysate monitoring device, the Biostat 1000 (Baxter Healthcare Corporation, McGaw Park, IL) was conducted to validate the measurements of urea removed, the delivered dialysis dose (Kt/V), and net protein catabolism (PCR). The results were compared with a total dialysate collection in each patient. During 29 dialyses in 29 patients from three centers, the paired analysis of urea removed, as estimated by the dialysate monitor compared with the total dialysate collection, showed no significant difference (14.7 +/- 4.7 g versus 14.8 +/- 5.1 g). Similarly, measurements of Kt/V and PCR showed no significant difference (1.30 +/- 0.18 versus 1.28 +/- 0.19, respectively, for Kt/V and 42.3 +/- 15.7 g/day versus 52.2 +/- 17.4 g/day for PCR). When blood-side measurements during the same dialyses were analyzed with a single-compartment, variable-volume model of urea kinetics, Kt/V was consistently overestimated (1.49 +/- 0.29/dialysis, P < 0.001), most likely because of failure to consider urea disequilibrium. Because urea disequilibrium is difficult to quantitate during each treatment, dialysate measurements have obvious advantages. The dialysate monitor eliminated errors from dialysate bacterial contamination, simplified dialysate measurements, and proved to be a reliable method for quantifying and assuring dialysis adequacy.

A proposed glossary for dialysis kinetics.

Quantification of the dialysis dose and assessment of nutritional status and response to nutritional therapy have become standard parts of the management of the chronic dialysis patient. Although advances in these areas have led to a more rational basis for therapy, certain misconceptions and points of confusion appear to have occurred. Recognizing the importance of a standard nomenclature to the development of concepts and the communication of research findings, we have attempted to compile a list of terms that are commonly used in the field of dialysis. New terms have been proposed for current ones that do not seem adequate. In addition, we have discussed potential methodologies for obtaining more accurate data for dialysis kinetics and for precise monitoring of nutritional intake and status. It is hoped that this glossary will stimulate discussion that will lead to refinements in terminology and concepts that will, in turn, improve research and practice in nephrology. It is anticipated that many of these definitions and recommendations will be modified or superseded as the management of patients with renal failure continues to advance.

Lean body mass estimation by creatinine kinetics, bioimpedance, and dual energy x-ray absorptiometry in patients on continuous ambulatory peritoneal dialysis.

Lean body mass (LBM), which is fat free body mass, can be used as an index of nutritional status. We evaluated three techniques for LBM estimation, including dual energy x-ray absorptiometry (DEXA), creatinine kinetics (CrKin), and bioimpedance (BI) in 10 patients on continuous ambulatory peritoneal dialysis (CAPD). Two different formulae were applied for BI LBM estimation, Segal (S) and Deurenberg (D). Mean values (+/- SEM) of LBM estimated were 48.2 +/- 3.6, 46.12 +/- 2.87, 43.32 +/- 3.87, and 41.27 +/- 4.26 by DEXA, BI-S, BI-D, and CrKin, respectively. LBM by CrKin was significantly lower than that by DEXA and BI-S values. There was no statistically significant difference between DEXA and BI-S values. Statistically significant correlations were found between LBM values by all methods. Particularly strong correlations were found between DEXA versus BI-S (r = 0.976) and BI-S versus BI-D (r = 0.98). Because clinical assessment of hydration status is inaccurate, and both BI and DEXA measure excess extracellular water in LBM, falling muscle mass may be missed by these techniques. The CrKin technique for estimating LBM at normal body fluid volumes (dry weight) may be a better index of nutritional status in patients on CAPD because this may truly reflect the dry LBM and changes in muscle mass. Both DEXA and BI include excess body water in LBM and may mask malnutrition in the presence of subclinical or clinical overhydration, which is common in patients on peritoneal dialysis.

A new approach to optimizing urea clearances in hemodialysis and continuous ambulatory peritoneal dialysis.

Recent studies suggest that the relationship of the net normalized protein catabolic rate (which is the normalized protein equivalent of nitrogen appearance [nPNA]) to the weekly clearance of urea normalized to total body water (Kt/V urea) in patients on continuous ambulatory peritoneal dialysis (CAPD) is curvilinear, rather than linear, as has been thought. The authors have reexamined the relationship of nPNA to weekly Kt/V urea in a CAPD population by cross-sectional analysis to see if the curvilinear definition of the relationship is as good as or better than the usual linear description. They also examined this relationship in the hemodialysis populations at the Dialysis Clinics Inc. in Columbia, Missouri, and in the Renal Kidney Disease Program in Minneapolis, Minnesota. It seems obvious that there should be a plateau of nPNA in each therapy because extension of linear regressions would predict protein intakes of normal individuals exceeding 8 g/kg/body weight/day. The authors compared their findings to other published results. Intuitively and analytically, the curvilinear relationships seem likely. The authors observed that the nPNA plateau is achieved at lower Kt/V in patients on CAPD than in those on hemodialysis, which is compatible with the peak concentration hypothesis. Asymptotes for CAPD and hemodialysis are similar. Weekly Kt/V urea requirements to achieve nPNA values at 95% of the asymptote are greater than those usually delivered. However, such nearly complete elimination of uremic appetite suppression may not be practical or necessary for achieving acceptable nutritional status and long-term survival in most patients. Optimum therapy may be well above adequate therapy relative to minimizing appetite suppression by uremia.

Pitfalls of in vivo dialyzer clearance measurement.

Dialyzer small-molecule clearance measurements are commonly made to help identify the cause of inadequate dialysis prescriptions, to determine the efficacy of reuse procedures, or to choose between different types of dialyzers. Clearance measurements can be blood-side- or dialysate-side-based. While blood-side clearance measurement is the classical technique, it suffers from several serious flaws that decrease its accuracy. Chief among these are the inability to accurately measure the blood flow rate and the difficulty in accounting for the presence of nonaqueous components in the blood. Using a dialysate-based clearance measurement technique overcomes these problems for most solutes, provided appropriate guidelines are followed. This article reviews the theory behind both blood- and dialysate-side techniques as well as discussing the practical application of that theory to clearance measurement.

On-line monitoring of the delivery of the hemodialysis prescription.

The BioStat 1000 is a new device which employs dialysate-based urea kinetics to calculate the dose of dialysis (Kt/V) based on a two-pool model and protein catabolic rate (PCR). Previous methods relying on blood sampling techniques were subject to error and difficult to implement. This paper describes the features of the Biostat and the results of the first clinical validation study with an early prototype. The BioStat was found to compare favorably with the reference method of direct dialysate quantification (mDDQ) which had been modified to obtain a "two-pool" Kt/V. In 31 patients no significant difference was found between mean Kt/V from the mDDQ and the mean Kt/V from the BioStat (1.35 +/- 0.33 versus 1.38 +/- 0.36, respectively). The PCR was also not significantly different (53.4 +/- 18.5 g/day versus 51.8 +/- 16 g/day, respectively). The BioStat was demonstrated to be a convenient method producing reliable results.

The solute removal index--a unified basis for comparing disparate therapies.

The removal-based SRI can be applied to the CAPD setting, and this new index of dialysis dose is numerically equal to the KT/V for a continuous therapy like CAPD. This is not true for hemodialysis where the SRI is numerically lower than the KT/V. If therapy prescriptions for CAPD, APD, and HD are all adjusted to provide the same removal at the same predialysis BUN, then the values of SRI will be the same in all three modalities despite differences in the frequency and duration of the therapy modalities. This is a major advantage of the SRI over KT/V in that it provides a unified, intuitive basis for comparing disparate therapies. The peak concentration hypothesis is implied in this approach because the therapies have equivalent SRI when there is matching of the peak predialysis BUN of APD and HD with the steady-state BUN of CAPD. However, if KT/V is used as the basis of comparison for different modalities, the weekly KT/V has to be adjusted with a scaling factor that varies with the frequency of dialysis. A different scaling factor will apply for daily APD versus that for thrice-weekly HD. The requirement of a scaling factor and the variability of this factor with therapy frequency complicate therapy comparisons based on KT/V. This complication is avoided with the SRI. Further, as detailed in a prior publication (1), the SRI has several advantages over the KT/V index for hemodialysis therapy, which include the avoidance of clearance corrections for access recirculation, cardiopulmonary recirculation, and compartmental disequilibrium.(ABSTRACT TRUNCATED AT 250 WORDS)

Relationship between body size, fill volume, and mass transfer area coefficient in peritoneal dialysis.

A peritoneal dialysate fill volume of 2 L has become the standard of clinical practice, but the relationships between body size, fill volume, and mass transfer area coefficient (KoA) have not been well established. These relationships were studied in 10 stable peritoneal dialysis patients who underwent six peritoneal equilibration studies (2 h each) at fill volumes of 0.5, 1, 1.5, 2, 2.5, and 3 L. The concentration-time profiles for urea, creatinine, and glucose were measured at each fill volume, and residual volumes were calculated from the preceding dwell period. A modified Henderson equation was used to calculate the KoA for the three solutes as a function of fill volume. By normalizing the KoA for each solute to the value at 2 L, the data for all three solutes collapsed onto the same trend line when plotting the normalized KoA versus dialysate volume. Between 0.5- and 2-L fill volumes, the average normalized KoA increases in an almost linear fashion, its value almost doubling over this range. Between 2- and 3-L fill volumes, there is less than a 10% change in the normalized KoA. However, fill volumes for peak urea KoA were found to increase with increasing body surface area (R = 0.76), being around 2.5 L for an average-sized patient and increasing to between 3 and 3.5 L for body surface areas > 2 m2. To maximize solute transport, these relationships between body size, volume, and KoA should be considered when choosing fill volumes for continuous ambulatory peritoneal dialysis and automated peritoneal dialysis and when deciding reserve and tidal volumes for tidal peritoneal dialysis.