JNEPHROL 2013; 26 ( Suppl 21): S83-S95

BEST PRACTICE

DOI: 10.5301/JN.2013.11635

Prescription in peritoneal dialysis Gianpaolo Amici1

Nephrology and Dialysis Unit, San Daniele del Friuli Hospital, ASS4 Medio Friuli, Udine - Italy 2 Nephrology and Dialysis Unit, ASL 12, Biella - Italy 3 Nephrology and Dialysis Unit, A.O. Spedali Civili di Brescia, Brescia - Italy 4 Department of Nephrology and Dialysis, Bari Hospital, Bari - Italy 5 Nephrology and Dialysis Unit, Dell’Angelo-Mestre Hospital, Venezia - Italy 6 Department of Nephrology and Dialysis, San G. Bosco Hospital, Torino - Italy 7 Department of Nephrology and Dialysis, A. Manzoni Hospital, Lecco - Italy 8 Nephrology and Dialysis Unit, ASL Cuneo 2 Alba-Bra, Alba - Italy 9 Nephrology and Dialysis Unit, A. Muri Hospital, Jesi Italy 10 Nephrology and Dialysis Unit, Camposampiero Hospital, Camposampiero, Padova - Italy 1

Reviewers: Roberto Bergia2, Giovanni Cancarini3, Roberto Corciulo4, Mariano Feriani5, Gian Maria Iadarola6, Vincenzo La Milia7, Luigi Manili3, Loris Neri8, Roberto Russo4, Massimo Sandrini3, Stefano Santarelli9, Giusto Viglino8, Giovambattista Virga10

Table of Contents 1. Introduction 2. The prescription of CAPD 2.1 Transport, prescription and clinical outcome in CAPD 2.2 CAPD and adequacy 2.3 CAPD and intraperitoneal pressure 2.4 CAPD over time 3. Prescription of APD 3.1 Continuous cycling peritoneal dialysis 3.2 Tidal 3.3 Fill Volume, IPP and APD 3.4 APD and adequacy 4. Alternative and biocompatible solutions 4.1 Neutral-pH, low glucose degradation product, bicarbonate-buffered fluids for multicompartment peritoneal dialysis 4.2 Alternative osmotic agents to glucose: icodextrin 4.3 Alternative osmotic agents to glucose: amino acids 4.4 Plastifying agents

5. Computerized kinetic models 5.1 Pd Adequest 5.2 Personal dialysis capacity 5.3 PatientOnLine 5.4 Quality of adequacy and kinetics data 5.5 Modeling errors 6. Incremental peritoneal dialysis (IPD) 6.1 Analysis of the literature 6.2 Principles of incremental prescription 7. Summary of good practice recommendations 7.1 CAPD 7.2 APD 7.3 Fluids 7.4 Kinetic models 7.5 Incremental IPD 8. References

Key words: Peritoneal Dialisis, Prescription, CAPD, APD, Biocompatible Solutions, Kinetic Models, Incremental Dialisis

© 2013 Società Italiana di Nefrologia - ISSN 1121-8428

S83

Best Practice on: Prescription in peritoneal dialysis

1. Introduction The prescription of peritoneal dialysis (PD) has evolved through the years into a series of therapeutic options that have considerably enriched it. Scientific contributions regarding manual (continuous ambulatory peritoneal dialysis [CAPD]) and automated (APD) PD, the peritoneal equilibrium test (PET), computerized models, intraperitoneal pressure, the various different fluids and incremental PD (IPD) make the objective of individualized prescriptions feasible.

2. The prescription of capd CAPD as specified by Moncrief and Popovich (1) indicated that the drainage of 10 L of dialysate daily was sufficient to purify the blood of a patient to equilibrium. The schedule included five 2-L exchanges with achievement of complete equilibrium for urea or 4 hypertonic 2-L exchanges with ultrafiltration (UF) of at least another 2 L. On the other hand, urea dialysate to plasma concentration (D/P) of 24-hour collections is, on average, 0.90 for all types of transport and must therefore be taken into consideration for adequacy calculations and prescriptions (2). Based on their studies of the interrelation among body size, intraperitoneal volume and mass transfer area coefficient (MTAC), Keshaviah et al (3) suggest that the increase in volume ranging from 0.5 to 2 L always involves a linear increase in MTAC, and further increases, in searching for peak MTAC, yield different results according to body size, suggesting a 2.5-L load for a surface of 1.7 m2 and a 3-3.5-L load for body sizes >2.0 m2. Therefore, new prescription criteria have been proposed according to body size.

2.1 Transport, prescription and clinical outcome in CAPD In a study by Johnson et al (4), performed on a large database, the survival of slow transporters at PET was longer with CAPD than the survival of rapid transporters, whose survival was significantly longer with APD. This finding should be assessed, taking into consideration the fact that rapid transporters have a higher risk of death in general and, in particular, while on CAPD, but do not have an increase in risk when they are treated with APD and icodextrin or when they are switched to hemodialysis. These risk data were collected in various case series and in systematic reviews, so they are to be S84

Fig. 1 - Criteria for the prescription of dialysis volume in CAPD according to the various authors cited. BSA = body surface area; CAPD = continuous ambulatory peritoneal dialysis; UF = ultrafiltration; Vin = initial volume; Vmax = maximum volume; IPP = intraperitoneal pressure.

considered evidence-based information, not an opinion. CAPD in general is not considered adequate treatment for rapid transporters with modest drainage and no residual renal function (RRF), because of the cardiovascular risk associated with poor UF. Therefore, the assessment of individual transport by PET is essential, as it is a diagnostic criterion for the application of the appropriate dialysis schedule. PET is also essential for monitoring, because the characteristics of the peritoneal membrane tend to change over time, increasing transport. A Cochrane (5) systematic review comparing CAPD and APD shows that there is no evidence of longer survival with one method vs. the other, but it does not take the issue of an appropriate prescription for patient transport features into consideration, so the outcome of the comparison in these terms is not meaningful.

2.2 CAPD and adequacy In 1994, Tattersall et al (6), in a trial addressing the adequacy of CAPD in patients who had developed anuria, calculated that 140 ml/kg body weight of dialysate at equilibrium are required to achieve minimally adequate Kt/V (9,800 mL for 70 kg). Other studies report prescription volumes ranging between 125 and 145 ml/kg according to whether one considers infusion or drainage

© 2013 Società Italiana di Nefrologia - ISSN 1121-8428

JNEPHROL 2013; 26 ( Suppl 21): S83-S95

2.3 CAPD and intraperitoneal pressure

Fig. 2 - Indications for the prescription of continuous ambulatory peritoneal dialysis (CAPD) by peritoneal transport (PET). Without PET data, residual renal function (RRF), creatinine clearance (CRCL) and ultrafiltration are considered in order to guide the prescription. APD = automated peritoneal dialysis; D/P CR = dialysate to plasma creatinine concentration; HD = hemodialysis.

with UF (7, 8) (Fig. 1). Based on the mean values obtained from clinical trials in the event of prescription of CAPD with 4 exchanges a day, in patients without RRF and with a prescription of 125 ml/kg, 1.7 Kt/V is obtained with increasing body weights and increasing loading volumes (60 kg 1.9 L x 4; 70 kg 2.2 L x 4; 80 kg 2.5 L x 4; and 90 kg 2.8 L x 4). These mean data show that the prescription of CAPD in patients with anuria becomes difficult when body weight is high. In heavy (90 kg), slow peritoneal transporters (dialysate to plasma creatinine concentration [D/P Cr] at 4 hours 1.78 m2 BSA. In view of mean pCRCL (p = peritoneal) amounting to 54 vs. 63 L per week/1.73 and, respectively pKt/V week of 1.80 vs 2.27, the results of the trial in terms of absence of outcome differences with the various treatments do not encourage the prescription of high volumes in the attempt to achieve higher purification levels in practice.

An alternative to the option of prescribing an increase in filling of the peritoneal cavity is the measurement of hydrostatic intraperitoneal pressure (IPP) proposed by Durand et al (10). IPP values range between 5 and 15 cm H2O with 2 L in the abdomen vs. mean pressure of 17 cm H2O in peritoneal capillaries. Its value varies according to the loaded volume, body size, parietal resistance features and internal organ volume. IPP is a prescription parameter in PD, as it influences UF and lymphatic and extralymphatic absorption, which, when IPP exceeds capillary pressure, become independent of fluid tonicity. Durand et al proposes maximum abdominal filling up to 1,500 ml/m2 BSA (2,595 mL per 1.73 m2) when IPP 2.0 m2 (3). However, the increase in peritoneal cavity filling requires IPP verification, which is a useful prescription parameter as it influences respiratory vital capacity, UF, fluid absorption and, specifically for APD, quality of sleep. Therefore, abdominal filling amounting to 1,500 ml/m2 BSA (2,595 mL per 1.73 m2) is proposed only when IPP is lower than 18 S86

cm H2O; in the event of higher values, the volume should be reduced (28). Recently also modulation of various volumes, timings and fluids during APD has been proposed with interesting results (29).

3.4 APD and adequacy In 1996, Diaz-Buxo set forth the concept of using the whole day with the following CCPD scheme: 3 L x 3 nocturnal cycles + 2 diurnal dwelling periods of 2 L + UF for a total drainage of 15 L, obtaining 2.03 Kt/V and creatinine clearance (CRCL) amounting to 51.14 L a week. The study was subsequently confirmed in a larger number of patients (18, 19). Durand et al reported, with the prescription of tidal for 8 hours and half at night, 50% of the load of 2 L, that when D/P Cr at 4 hours is lower than 0.65 it becomes impossible to achieve CRCL of 50 L/week per 1.73 m2 BSA even using 35 L of dialysate per session. On the contrary, in slow transporters, clearance worsens as the number of liters prescribed increases (22). The same authors, however, report that from 30 to 50 ml/kg per hour an increase in Kt/V from 1.73 to 2.15 is achieved, and in another study, an increase from 17 L of continuous tidal peritoneal dialysis (CTPD) to 32 L resulted in an increase in Kt/V from 2.09 to 2.52 (23-25). Therefore different removal of toxins with a different molecular weight is achieved. The European APD Outcome Study (EAPOS) recorded good survival of the technique (62%) and of the patient (78%) with target CRCL in the proximity of 60 L a week. The schedules that enable these results are (a) in slow transporters, CCPD 3 cycles of 2 L for 9-10 hours/night + 2 dwell periods of 2 L during the day; (b) in mean transporters, CCPD 3-4 cycles of 2, 2.5 or 3 L for 9-10 hours/ night + 2 dwell periods during the day of 2-2.5 L (according to body size); (c) in rapid transporters, CCPD 4-5 cycles of 2-2.5 L for 8 hours/night + 2 dwell periods during the day of 2-2.5 L (according to body size), one of which with icodextrin. The prescription was defined with PD Adequest and a clinical discussion for a total effective daily drainage of 15.716.7 L and effective CRCL of 63.4-68.8 l/week per 1.73 m2 (17). In the same study, for equilibrium of the patient with anuria, in terms of salt and water, optimal UF must be greater than 750 mL (17, 30). Consequently, the literature regarding APD is divided between low flow (7-12 l/night) and high flow (18-27 l/night), and, to summarize, the best results are achieved only by considering APD a therapy that must be individualized according to patient transport characteristics, after having verified clearance (Fig. 3). Rapid transporters have a higher risk of death, especially if they are on CAPD, but they do not have an increase in risk if they are treated with APD and icodextrin, or are on hemodialysis. These risk

© 2013 Società Italiana di Nefrologia - ISSN 1121-8428

JNEPHROL 2013; 26 ( Suppl 21): S83-S95

Fig. 3 - Indications for the prescription of automated peritoneal dialysis (APD) by peritoneal transport (PET). Without PET data, residual renal function (RRF), creatinine clearance (CRCL) and ultrafiltration are considered in order to guide the prescription. CCPD = continuous cycling peritoneal dialysis; CTPD = continuous tidal peritoneal dialysis; D/P CR = dialysate to plasma creatinine concentration; IPP = intraperitoneal pressure.

data were collected in various case series and in systematic reviews, so they are to be considered evidence-based information, not an opinion (31). Therefore, the assessment of individual transport by PET is essential, as it is a criterion for the application of the appropriate dialysis schedule. Fluid with icodextrin together with APD is the optimal schedule for the survival of rapid transporters (32).

4. Alternative and biocompatible fluids The peritoneal membrane changes over time, developing histopathological and functional alterations (33, 34). Chronic inflammation is the main cause of the aging of the membrane, and PD fluids, with their components, are believed to be responsible for this problem. It is therefore useful to consider a strategy for prescriptions that takes advantage of fluids that are more biocompatible than the ones that are currently available.

4.1 N  eutral pH, low glucose degradation product, bicarbonate-buffered fluids for multicompartment peritoneal dialysis The double compartment in the bags was designed to reduce the glucose degradation products (GDPs) in the

fluid, sterilizing glucose in a concentrated, no acid and salt-devoid fluid (first compartment), subsequently to be combined with a solution containing mineral salts and buffer (second compartment) so as to increase the pH of the mixed product up to the physiological pH range. GDPs are considered precursors of the advanced glycosylation end products (AGEs) responsible for proinflammatory and prooxidative effects in all tissues, including the kidneys and the peritoneum (35). The multicompartment fluids contain about 10 times less GDPs than the single compartment bags (35, 36). For a practical definition of low level of GDPs, the color of the fluid, the presence of multiple compartments that separate glucose from salts and buffers, and low pH in the glucose compartment (pH 1.8-3.5) have to be taken into consideration (35). The GDP values that distinguish biocompatible solutions from standard solutions according to the presence of the well known 3,4-dideoxyglucosone-3-ene (3,4-DGE) compound are 0.2-11 vs. 13-19 microM/L. Another reason for the double compartment is the introduction of bicarbonate as the physiological buffer that is the ideal replacement for lactate. The bicarbonate in the presence of calcium salts tends to form insoluble deposits so it has to be kept separate. At present there are different types of commercial fluids for PD: standard 1 compartment PVC bags with lactate, pH 5.0-5.4 and high content in GDPs; 3-compartment PVC bags with lactate at pH 6.5 and low content in GDPs; trilaminar 2-compartment plastic (non-PVC) bags with a combination of bicarbonate and lactate at pH 7.3, and average content in GDPs; 2-compartment polyolefin bags with lactate buffer at pH 7.0 and low content in GDPs; 2 compartment polyolefin bags with pure bicarbonate buffer, devoid of lactate at pH 7.4 and with low content in GDPs (Fig. 4). The literature available so far shows improvement in indirect indices of peritoneal toxicity using bicarbonate and bicarbonate lactate fluids, with low content in GDPs and neutral pH (37-41). Retroactive studies in large cohorts show an advantage in terms of survival of the patient and of the technique both with low-GDP and neutral pH fluids and with bicarbonate fluids (37, 42), and another study shows less decline in RRF with low-GDP fluids (43) (Fig. 4).

4.2 A  lternative osmotic agents to glucose: icodextrin The use of icodextrin has been one of the most important steps forward regarding PD in the last few years (44). It is a fluid that does not contain glucose. It contains a mixture of glucose polymers (maltodextrins) with various molecu-

© 2013 Società Italiana di Nefrologia - ISSN 1121-8428

S87

Best Practice on: Prescription in peritoneal dialysis

ers (49, 50). The intensive use of larger quantities of icodextrin is still to be considered experimental (51).

4.3 Alternative osmotic agents to glucose: amino acids

Fig. 4 - Schematic summary of the different commercial solutions according to their characteristics. GDPs = glucose degradation products; PD = peritoneal dialysis.

lar weights, which generates UF with a colloido-osmotic mechanism that is similar to that of long-dwelling 3.86% glucose. The single-compartment fluid with PVC bag has a 7.5% icodextrin concentration, pH 5.0 and low GDP content (Fig. 4). The production of UF is slower than with glucose, but lasts longer and does not involve the aquaporin system. This fluid has considerable advantages during long dwell periods, it enables UF production even in the presence of peritonitis or UF deficiency, and reduces calorie absorption (44). Icodextrin has some disadvantages, such as absorption of the polymer with increases in blood maltose, which restricts its use to only 2 L a day (44); interference with blood glucose measurement with some reactive strips and with amylase determination (45); and it can induce an increase in transaminases and alkaline phosphatase and a reduction in sodium concentrations (46). Finally it is associated with skin hypersensitivity (47). The dextrin mixture may vary according to the manufacturing process and may be subject to bacterial contamination with the presence of proinflammatory antigen particles even after heat sterilization (48). There are precise principles for the prescription of icodextrin: it must be used for long dwell periods (not less than 8 hours) because it induces UF progressively over time (44); it can be used as a bag for CAPD during the night and a bag for APD during the day; it is able to prolong the duration of dialysis in patients with anuria; it increases UF, the removal of sodium and purification in all types of transporters and improves the management of rapid transportS88

Another alternative osmotic agent to glucose is a mixture of amino acids at 1.1% concentration, which offers biocompatibility and a contribution to nutrition (Fig. 4). The amino acids, in view of their modest molecular weight, are very active osmotically, but only for short dwell periods, and are rapidly absorbed in considerable quantities (52). This characteristic, together with their acidic nature, restricts their prescription to only 2 L a day because of metabolic acidosis and increases in blood urea nitrogen. The fluid is proposed also for APD, to be mixed and absorbed together with glucose, with positive effects for protein anabolism (53-56).

4.4 Plastifying agents Phthalates added to the PVC of the bags are believed to be responsible for chronic peritoneal toxicity. Many studies have been published that have shown that there is a relationship between phthalates dissolved in the peritoneal dialysis fluid and inhibition of leukocytes, release of IL-1 from monocytes and increase in apoptosis (57, 58). There are no studies in humans comparing PVC bags with and without phthalates, in terms of long-term damage (59). In any case non-PVC bags devoid of plastifying agents are currently available.

5. Computerized kinetic models Prescriptions can be facilitated by the use of computer programs that are able to predict the results that can be obtained with the various dialysis schedules, based on mathematical simulations. These programs use kinetic models of peritoneal transport. There have been many contributions to peritoneal transport models, but 3 software programs are available and can be used: PD Adequest, Synergy PDC and PatientOnLine.

5.1 PD Adequest This program uses the Pyle-Popovich model with homogeneous pores (60, 61). The data required by the model, besides demographic data, are daily clearances, PET and the exchange during the previous night. Calculations related to measurements are total weekly Kt/V and normal-

© 2013 Società Italiana di Nefrologia - ISSN 1121-8428

JNEPHROL 2013; 26 ( Suppl 21): S83-S95

ized CRCL (l/1.73 m2), normalized protein catabolic rate (g/kg per day), glomerular filtration rate (GFR; mean renal clearance of urea and creatinine, ml/min) and PET classification D/P Cr according to Twardowski. The program provides all the D/Ps of urea, glucose and UF; MTAC related to urea, creatinine and glucose; fluid absorption and hydraulic permeability. The program simulates the results of dialysis prescription mathematically according to 2 methods: Regimen, in which the modeling process starts from a prescription regimen and the program provides the results for the proposed prescription. The results are expressed in terms of weekly CRCL, Kt/V, UF and glucose reabsorption in CAPD and APD with any volume or type of fluid. For the second method, called Optimise, the modeling process starts from a nutrition, adequacy, UF or lifestyle target. The program proposes 4 different prescription schedules that may correspond to the target. The results are expressed in terms of dialysis volume during the day and night, concentrations and clearance (Fig. 5).

Fig. 5 - Comparison of the characteristics of softwares for the calculation of adequacy and the mathematical simulation of peritoneal dialysis. For abbreviations see text.

5.2 Personal dialysis capacity The program Personal Dialysis Capacity (PDC) uses a “3-pore” capillary peritoneal transport model that is able to describe the properties of the membrane (62, 63). It describes the flow of water, crystalloids and proteins through 3 types of theoretical cylindrical pores in the endothelium: small pores of 40-50 Å, large pores of 250-300 Å and ultrasmall pores dedicated only to the transport of water of 2-4 Å. The 3 types of pores then generate widespread resistance that is added. The main parameter that characterizes peritoneal transport is the ratio between total pore area and diffusion distance (A0/ Δx), other parameters of the model are the absorption of fluid (JvR) and protein loss (JvL). The parameters required by the program are obtained through a CAPD schedule with 5 dwell periods carried out at home, annotating the time; concentration, weight and sampling related to urea, creatinine, glucose and albumin; 2 blood samples representative of the day; and collection of urine for residual renal function. Alternatively, a CCPD schedule with 2 diurnal dwell periods is able to provide similar data. Calculations of the model are thus based on the performance of dwell periods of different duration. The program provides the possible outcomes for various applied dialysis schedules through simulations. The reference values of A0/Δx per 1.73 m2 are 23,600 cm (17,200-30,000), corresponding by and large to the PET transport categories (Fig. 5).

5.3 PatientOnLine The software PatientOnLine for calculations and simulations was developed referring to the model of Garred et al, based on homogeneous pores (64). The peritoneal MTAC is calculated and expressed as Pt50 (time to achieve D/P 0.5) starting from a test that can be standard PET or quality assurance exchange–peritoneal function testing (QA-PFT). The key component of the model is urea kinetics, but assessments of creatinine kinetics are included, together with an estimate of lean body mass. The software accepts different types of input: collection of exchanges and 24-hour urine + 1 exchange in compartment (QA); modified peritoneal function test (MPFT) with 24-hour collection with separate bags and different dwell times, a test proposed by Gotch et al (65); simplified PFT (sPFT): exchange + urine QA; PET; 24-hour dialysate collection + urine; and PET + 24-hour collection. Obviously a test that collects as much information as possible is recommended for the construction of an accurate model. Measurements of urea, creatinine, glucose and proteins are required for all samples. The software simulates PD mathematically according to 2 methods: it provides the results of defined dialysis schedules (“prescription”) or it provides dialysis schedules for adequacy or lifestyle objectives (“targets”) (Fig. 5).

© 2013 Società Italiana di Nefrologia - ISSN 1121-8428

S89

Best Practice on: Prescription in peritoneal dialysis

5.4 Quality of adequacy and kinetics data Random and systematic errors occur in all physical and biological measurements, and these may be with regard to volume, weight or biochemical analyses. These errors occur in clearances and in PETs. The results of the simulations performed by kinetic software programs are influenced by these errors, sometimes with amplification effects. Consequently, it is necessary to assess results critically. It is well known that the measurement of creatinine by the Jaffé kinetic method is subject to interference by the glucose contained in the PD fluid and requires correction, just as glucose measurement at high concentrations may be considerably underestimated. Therefore, the correct assessment of PET requires critical review of the data that are not compatible with physiology or the conditions of the patient. The classification of patients using D/P Cr by PET has a Gaussian distribution: the average transport categories are the most common (68%), the extreme categories (rapid 16% and slow 16%) are less common. Other methodological problems encountered during the performance of the test are due to the method and volume of sampling, at time zero with the calculation of residual volume and exact weighing of the infused and drained volume to exclude bag overfill. The patients must be clinically stable during the tests. Patients with acute conditions, hyperhydration, dehydration, peritonitis or sepsis are to be excluded. Total daily UF must be positive. If it is nil or negative collections are to be repeated, because they are not compatible with steady state. Patients in poor nutritional conditions will have overestimated clearance data because of normalization, and it is more correct to normalize by reffering to ideal body weight, providing dialysis objectives that are suitable for the regaining of normal nutritional status.

5.5 Modeling errors PD Adequest underwent 2 multicenter clinical validations managed directly by its creator in a total of 215 patients, and the percentage error was ascertained: weekly Kt/V 0.1% ± 10.0%, weekly CRCL −0.5% ± 12.2% and net daily UF −13.6% ± 69.6% (66, 67). An independent and multicenter validation in 78 patients showed a larger percentage error: weekly Kt/V 0.2% ± 11.9%, weekly CRCL 5.7% ± 13.1% and net daily UF −7.2% ± 271.0%, and also reported the influence of the type of PET on modeling errors (68). The verification principle used was based on the comparison between the results of an actually implemented dialysis schedule and modeling of the same schedule with the software. Synergy-PDC has some important typical S90

features for the validation process. Indeed, the test lasts 24 hours for the acquisition of baseline data and changes the schedule that is usually followed by the patients substantially. The only independent clinical multicenter validation study in 336 patients showed that the standard deviation of the differences between model and true data was 0.5 ml/min for UNCL (Urea Nitrogen Clearance), 0.55 ml/min for CRCL and 282 mL for 24-hour UF (69). It also found a trend toward underestimation of UNCL and overestimation of CRCL, which corresponds to what was also found with other software, as well as the marked variability of UF. There are no independent validation studies regarding the PatientOnLine software, but validation tests were carried out by its creator in 108 patients and showed a 95% confidence interval for differences in absolute terms of ± 0.30 units for weekly peritoneal Kt/V and in relative terms of 20% only for peritoneal urea Kt/V (65) (Fig. 5).

6. Incremental peritoneal dialysis The basic concept of incremental peritoneal dialysis (IPD) is that PD can be prescribed gradually, in parallel and with integration of RRF, for the achievement of a conventional PD adequacy target. This enables treatment with a low dialysis dose, which is then adjusted over time as RRF deteriorates. At present, it still is up for debate whether treatment with PD should be carried out with a full or an incremental dose.

6.1 Analysis of the literature The experience reported consists mainly in pilot studies and illustrates selection criteria, type of treatment and follow-up. De Vecchi et al (70) reported 4 years of follow-up of 25 patients who underwent CAPD with 1 dwell period at night or 2 dwell periods during the day with initial RRF, who were measured with urinary CRCL higher than 6 ml/ min. Results were positive in terms of dialysis adequacy, quality of life and low number of complications related to PD. Time on IPD was about 1 year (10.6 ± 8.9 months), and over 2 years, RRF declined from 6.2 ml/min to 4.2 ml/ min on average with more or less constant urinary volume throughout the study (from 1,328 to 1,283 ml/day). Viglino et al and Neri et al (71, 72) reported the effects of IPD on RRF and on the adequacy of treatment in patients with GFR 5 ml/min, 3 dwell periods for