KDOQI Update 2000


VII. Clinical Outcome Goals for Adequate Peritoneal Dialysis


Throughout these Guidelines, the Work Group has focused on patient outcomes. Improving patient outcomes is the primary objective of the KDOQI (Kidney Disease Outcomes Quality Initiative). The Work Group realizes that definitions of goals regarding patient outcomes are needed. As stated in the Introduction to these guidelines, the goals are integral to the definitions of adequate, optimal, and effective dialysis.


Measurement of PD Patient Survival (Opinion)

Survival of PD patients should be quantitated serially as an outcome measure.

Rationale Patient survival is an objective outcome that is dependent upon many variables, some controllable and some uncontrollable. Suboptimal doses of delivered dialysis will adversely affect patient survival in adults (data unavailable in children), as discussed in detail in Guideline 15: Weekly Dose of CAPD. A primary goal of ESRD therapy is to prolong life while minimizing uremic symptoms. United States Renal Data Systems (USRDS) data from hemodialysis patients have demonstrated an association between low Kt/V and increased incidence of death from coronary artery disease, other cardiac disease, cerebrovascular accidents, and other conditions.93,94 There is also evidence that underdialysis adversely affects mortality in PD patients with ischemic cardiac disease or left ventricular dysfunction.16,95 Thus, patient survival is a measure of renal replacement program effectiveness. Case mix must be factored into survival analysis, however. The USRDS case mix analysis is an excellent starting point, addressing age, race, primary renal disease, and presence or absence of diabetes. The USRDS is attempting to refine case mix further by addressing other underlying comorbidities. At dialysis centers with a small number of patients, survival may need to be evaluated over many years to obtain a reliable estimate.


Measurement of PD Technique Survival (Opinion)

PD technique survival, both dependent and independent of peritonitis, should be quantitated serially in PD patients as an outcome measure.

Rationale It is common for ESRD patients to change renal replacement therapy modalities during the course of their treatment. Reasons for transfer include: complications of the therapy, inability to perform the therapy (lack of suitable access, no partner to do self care, medical contraindication), and patient request/lifestyle issues. Patient case mix, geographical location, and experience with PD in complicated cases are factors affecting transfer. For some patients, for optimal outcome, it may be medically appropriate to transfer from PD to HD; this does not imply failure of the therapy or the dialysis facility.

Peritonitis remains the primary cause of transfer from PD. It is acknowledged that at times peritonitis is the "precipitating" event for transfer, while the real underlying reason is patient burnout, noncompliance, inadequate dialysis, a request based on lifestyle, or an underlying exit site infection. Overall peritonitis rates can be influenced by the center. An association between malnutrition and frequency of peritonitis has been reported.96 Inadequate dialysis may lead to inadequate dietary protein intake and malnutrition as described in detail in Section IV: Assessment of Nutritional Status as it Relates to Peritoneal Dialysis. The relationship between solute clearance and frequency or severity of peritonitis has not been adequately studied. For example, it is not clear if inadequate dialysis or malnutrition directly predispose the patient to peritonitis. However, peritonitis is an important outcome and its frequency and severity is an index of the overall suitability of PD. It is, therefore, the opinion of the Work Group that peritonitis should be monitored.

Inadequate dialysis is directly responsible for at least 10% of transfer to HD.93 There is an association between PD technique failure and total solute clearance,16,97 and one possible reason for poor technique survival rates may be underlying inadequate dialysis. While the CANUSA study found a relationship between creatinine clearance and PD technique survival, the investigators suspected this was more related to RKF than delivered dose of PD. In the CANUSA study, technique survival was approximately 75% at 2 years in North America.16 However, peritonitis was not analyzed as a variable in the multivariate analysis. No difference was found in technique survival between patients from Canada and the United States. Average Kt/Vurea started at 2.38 and decreased to 1.99 over 2 years. The risk of technique failure increased by 5% for every 5-L/week decrease in creatinine clearance. These findings corroborated data from Tattersall et al,97 who found that patients with a lower Kt/Vurea had a lower rate of technique survival. The lower rate of technique survival was related to underdialysis, not to peritonitis or hernia development. The nutritional parameter of nPCR has been shown to be predictive of CAPD technique failure in a multivariable analysis.98

It is a common perception that patients transferring from any renal replacement therapy are at increased risk for death in the immediate posttransfer period. However, indirect evidence from the USRDS does not support that perception.99 Early demise following transfer may be due to a variety of reasons, including the underlying cause of transfer, inadequate dialysis on PD, and malnutrition. Therefore, the Work Group recommends that technique survival be measured at each PD facility. Admittedly, there may be center-specific differences for rates of transfer and that at centers where there are small numbers of PD patients, these rates may need to be evaluated over many years. It is recommended that these rates be evaluated both dependent and independent of peritonitis, although, in one study (CANUSA), there was no significant difference between the two evaluations.16

PD technique survival is dependent upon many factors including infections, patient motivation, ultrafiltration (transport characteristics), and total solute clearance. Thus, PD technique survival is not a simple outcome measure for the adequacy of PD. Nonetheless, centers should strive to achieve the goal of a 75% 2-year technique survival rate (the rate noted in CANUSA). Case mix must be factored into survival statistics.


Measurement of Hospitalizations (Opinion)

ESRD-related and ESRD-unrelated hospitalizations (admissions/year, hospitalized days/year) in PD patients should be quantitated as an outcome measure.

Rationale Hospitalization is an indicator of the overall effectiveness of treatment of chronic conditions and therefore constitutes an important outcome for dialysis patients. The number of admissions per year and total number of hospital days per year are two separate, but related, serial measures of outcome. Hospitalizations of PD patients can be related or unrelated to ESRD. An association between low creatinine clearance and increased overall hospitalization rate has been reported.16 Based on this evidence, the Work Group recommends that cause, frequency, and length of hospitalizations of PD patients be monitored. Categorizing hospitalizations according to whether they are related to ESRD or not offers certain advantages for analysis and, therefore, it is the opinion of the Work Group that this type of stratified analysis should be performed.

According to the USRDS, PD patients are hospitalized an average of 1.8 times per year.100 The CANUSA study found, by multifactorial analysis, an association between prolonged hospitalization and low creatinine clearance.16 Some admissions are specific to PD, ie, not seen with other dialysis therapy, such as elective abdominal wall herniography. As larger instilled volumes are administered to maintain target doses of dialysis, there is an increased risk of leaks and hernia formation, both of which can lead to hospitalization. Admissions unrelated to ESRD are important indicators of morbidity (cardiac disease, infections, etc) in PD patients. Since hospitalization data are important outcome parameters for all dialysis patients and can reflect solute clearance, the Work Group recommends that they should be monitored.

Although it is uncertain whether inadequate dialysis is directly related to an increased risk of peritonitis and catheter infections, inadequate dialysis is related to uremic symptoms such as nausea, vomiting, and gastrointestinal bleeding. It is acknowledged that some centers may treat all episodes of peritonitis in the hospital, while others only admit those with severe or refractory peritonitis. Therefore, in addition to tracking hospitalization rates, centers should also monitor reasons for hospitalization (related versus unrelated to ESRD, and specific reason for admission). The use of ICD-9 (International Classification of Diseases, 9th revision) or similar codes may be valuable in this process. The USRDS is attempting to categorize causes of hospitalizations as infectious, cardiovascular, dialysis access-related, and all other.

Hospitalizations from causes unrelated to ESRD may be related to inadequacy of PD. As discussed in Guideline 25: Measurement of PD Patient Survival, disease-specific (particularly cardiac) mortality is related to the dialysis dose for both HD and PD. Therefore, although studies of disease-specific hospitalizations and their relation to the dose of PD have not been reported, it is reasonable to monitor this relationship. Each outcome measure should be adjusted as well as possible for case mix. The USRDS is attempting to do so with Standardized Hospitalization Rates (SHRs).


Measurement of Patient-Based Assessment of Quality of Life (Opinion)

Patient-based assessment of quality of life (QOL) in PD patients should be evaluated serially as an outcome measure.

A patient-based quality of life instrument should have both generic and disease/treatment-specific measures of health-related quality of life and should be shown to be valid, reliable, and responsive prior to use. Once such an instrument is available, it should be administered at initiation of dialysis and at intervals determined to be appropriate by its validation studies.

Rationale Quality of life (QOL) can be assessed with generic or disease-specific measures. Many quality of life measures have been used in dialysis patients. However, fewer measures have been used for peritoneal dialysis than for hemodialysis patients.101 Measures used in peritoneal dialysis patients and reported in the literature include101:

• Medical Outcomes Study Short Form 36 (SF-36).

• Sickness Impact Profile (SIP).

• Index of Well Being, Index of Overall Life Satisfaction.

• Index of Psychological Affect.

• General Health Questionnaire.

• Simmons Self Esteem Scale.

• Profile of Mood States.

• Multidimensional Health Locus of Control.

• Modality Specific Stresses Scale.

• General Treatment Stress Scale.

• Global Illness Stress on Self and Others, Global Adjustment to Illness Scale.

• Quality of Life (QL 100 mm) Analogue Scale.

• Dialysis Relationship Quality Scale.

• Social Leisure Activities Index, Social Support Satisfaction Scale.

• General Well Being Index.

• Index of General Affect, Overall Life Satisfaction.

• Katz Activities of Daily Living.

• Time Tradeoff Measures.

Unfortunately, many of these instruments do not have published data indicating that reliability (test-retest, inter-rater), validity (content, construct, internal consistency), and responsiveness-to-change have been rigorously tested.100 For this reason, no particular instrument can be strongly recommended over another. Furthermore, many instruments developed for research purposes may be burdensome for patients or facilities, eg, require interviewer assistance or have complicated scoring algorithms. Nevertheless, generic and disease-specific measures hold promise as useful clinical tools.102

A popular generic measure used in peritoneal dialysis patients is the Medical Outcomes Study Short Form-36 (SF-36). Promising self-administered instruments used in peritoneal dialysis patients include the CHOICE Health Experience Questionnaire103,104 and the Kidney Disease Quality of Life (KDQOL).105

The Work Group recommends that each facility keep abreast of future developments regarding these instruments. As experience increases and one or more instruments are clearly established as useful in PD patients, standardized QOL measurement should be integrated into the routine care and evaluation of patients, programs, and facilities.


Measurement of School Attendance, Growth, and Developmental Progress in Pediatric PD Patients (Opinion)

School attendance (in the absence of other comorbidities precluding school attendance), growth, and developmental progress should be measured serially in pediatric PD patients.

Rationale The ability of pediatric PD patients to attend school is an important measure of the success of PD.

Underdialysis may affect the cognitive development and statural growth of children. However, the exact relationship between dialysis dose and normal growth and development in children is not clear. Nonetheless, the Work Group believes that cognitive development and statural growth should be monitored serially in children and charted in relation to patient age.


Measurement of Albumin Concentration in PD Patients (Opinion)

A stable or rising serum albumin concentration that is greater than or equal to the lower limit of normal for each laboratory should be used as an outcome goal.

Rationale In PD patients, as with HD patients, there is strong evidence to suggest that a low serum albumin level is associated with an increased risk of technique failure and death.16,50,106,107 Patients with the highest serum albumin levels have the lowest risk of death. In the CANUSA study, a difference of 0.1 g/dL serum albumin concentration was associated with a 5% change in the risk of technique failure, a 5% change in days hospitalized, and a 6% change in the risk of death.16 Therefore, the Work Group recommends monitoring serum albumin concentration in PD patients because of its association with important outcomes.

Although the significance of serum albumin as a predictor of outcomes in adults is undisputed, its relationship to overall nutrition and, to a larger extent, to the levels of urea or creatinine clearance is unclear. That albumin synthesis depends on dietary protein intake is well known. However, catabolic illness can reduce albumin synthesis, and probably increase albumin degradation, even when dietary protein intake is not low. Observations in PD patients have provided indirect support for this effect of catabolic illness. Although serum albumin concentration is an important predictor of outcome,50,106 it was not found to be significant in another study when comorbid conditions were entered as covariates in their model.107 In this last study, the presence of comorbid conditions was associated with low serum albumin.107 Several cross-sectional studies have identified a positive correlation between serum albumin concentration and solute clearance.51,108,109 However, urea and creatinine clearance were not identified as predictors of serum albumin by multivariate analysis.56,110 Age, comorbid factors (diabetes), and peritoneal solute transport were the major predictors of serum albumin in these multifactorial analyses.

Normal serum albumin concentrations vary by laboratory methodology; hence local standards should be used. If the serum albumin level is below normal for the laboratory, but is increasing, this suggests that the patient is anabolic and is increasing protein stores. Conversely, a low albumin or decreasing albumin level is likely to be associated with malnutrition or decreasing protein stores. Although there are no published data specifically addressing this point, it is the Work Group’s opinion that a patient whose serum albumin has decreased 0.1 g/dL/month from a baseline of 4.0 g/dL to 3.7 g/dL may be at higher risk than a patient with a stable serum albumin concentration of 3.7 g/dL.

Taken together, the data discussed above suggest to the Work Group that:

• Serum albumin concentration should be monitored on a regular basis and a stable or rising value is desirable. It should be measured at least every 4 months.

• Serum albumin levels should be evaluated in the context of the patient’s overall clinical status including comorbid diseases, peritoneal transport type, delivered dose of PD, and quality-of-life issues.

• The highest albumin level possible should be the goal for each patient.

It is the Work Group’s opinion that an optimal serum albumin level can be obtained by adequate nutrition monitored frequently by the renal dietitian, prevention and treatment of catabolic illness, and maintenance of Kt/Vurea and creatinine clearance at or above the levels recommended in Section V: Adequate Dose of Peritoneal Dialysis.

In summary, low serum albumin is a strong predictor of mortality and morbidity in PD patients. Therefore, serum albumin is an important outcome measure in PD patients and should be monitored, although an association between serum albumin and urea or creatinine clearance has not been convincingly shown. Efforts to maintain serum albumin in the normal range should include adequate nutrition, adequate clearances, and prevention and treatment of catabolic illness.


Measurement of Hemoglobin/Hematocrit in PD Patients

Providers should strive to achieve a hemoglobin level of 11 to 12 g/dL or a hematocrit of 33% to 36% in 75% of PD patients.

Rationale See NKF-KDOQI’s Clinical Practice Guidelines for the Treatment of Anemia of Chronic Renal Failure.


Measurement of Normalized PNA in PD Patients (Opinion)

Providers should strive to achieve a normalized PNA (nPNA) of greater than or equal to 0.9 g/kg/day in PD patients.

See Guideline 16 of the NKF/KDOQI Clinical Practice Guidelines for Nutrition, which recommends a dietary protein intake of 1.2 to 1.3 g/kg body weight/day in clinically stable chronic peritoneal dialysis patients. Such a diet should lead to an nPNA equal to or greater than 0.9 g/kg/day.

Rationale The role of PNA is discussed in Guideline 12: Assessment of Nutritional Status.

Maintenance of positive nitrogen balance and the prevention of underlying malnutrition is important because of the documented detrimental impact of hypoalbuminemia and low SGA scores on patient survival.16,80,106,110 From nitrogen balance studies, Blumenkrantz et al111 estimated that PD patients need to ingest at least 1.2 g/kg/d of protein to maintain positive nitrogen balance.109 This is higher than the recommended daily protein intake for healthy individuals, but not surprising due to the significant amount of protein known to be lost in dialysate. Despite these recommendations from balance studies, Bergstrom et al53 and Nolph et al51 have reported that many patients are in positive nitrogen balance with protein intakes of 0.9 to 1.0 g/kg/d. Cross-sectional studies would suggest that in the absence of significant comorbid diseases, patients with PD doses in the range of Kt/Vurea of 2.0 spontaneously ingest at least 0.9 g/kg/d of protein.52,112,113 Total solute clearance and nPCR* are strongly correlated in cross-sectional studies.57,114 However, it has been suggested that this is due in part to mathematical coupling of data.55,115 Three studies have investigated the effect of an increased Kt/Vurea on nutritional status (nPCR and serum albumin concentration) in a limited number of subjects.56,58,116 While nPCR increased as Kt/Vurea increased, an increase in serum albumin concentration did not occur. A reasonable conclusion from these data would be that in the absence of significant comorbidity, an increase in delivered dialysis dose should result in a corresponding increase in nPNA. In patients who show signs of malnutrition, their dialysis prescription should be closely evaluated with consideration to increasing their dose of dialysis if significantly below target. This is discussed in Guideline 15: Weekly Dose of CAPD.

No study of PD patients has demonstrated that nPNA is an independent predictor of outcome when a multiple regression model is used. Correlation coefficients relating DPI to nPCR are on the order of 0.6. However, PD patients who are neither anabolic nor catabolic tend to demonstrate higher correlation coefficients between nPNA and DPI.53 Also, there are little data to show a significant relationship between nPCR and serum albumin levels.56,110 Age, peritoneal transport type, presence of diabetes, and other comorbid diseases have a greater effect on albumin than does nPCR.

Despite these concerns, the Work Group recommends that PNA should be monitored. Low nPNA values in nonanabolic PD patients indicate a low DPI regardless of the values of other nutritional indices.54 One should strive to achieve an nPNA of at least 0.9 g/kg/day in adult PD patients. PNA values at this level or higher are likely to be associated with neutral or positive nitrogen balance in the absence of significant comorbidity or dialysate protein losses.

While the recommended dietary allowance for normal children is known, there are no definitive data regarding the real protein needs of children, especially young children on dialysis.69b However, clinical practice suggests that the protein needs of children on PD are greater than the recommended dietary allowance, in part related to dialysate protein losses. Current recommended protein intakes for children receiving PD, referenced for age are as follows117:


Age (Years) Protein Intake* for PD
















*Values are expressed in grams of protein/kg/day.

†Based on growth potential.


Because dialysate protein losses may vary widely in children, individualized recommendations for dietary protein intake may benefit from measurement of dialysate protein losses. An equivalent amount of protein to replace dialytic losses must be added to the recommended daily allowance for normal children.118


The relationship between solute clearance and the frequency and/or severity of peritonitis has not been adequately studied. Specifically, does inadequate PD dose contribute to peritonitis?

Studies to assess whether an increase in nPNA is associated with an increase in serum albumin levels, nutritional status, or improved survival would be valuable.

Studies to further define the relationship between nPCR and outcome are needed. For example, the longitudinal follow-up of nutritional status (determined by a variety of methods) will be more influential in improving understanding of nutritional outcomes. Longitudinal studies should be emphasized over cross-sectional studies.

The relationship between PD dose and outcome parameters in children needs definition. Studies of nutritional interventions are lacking and are encouraged.

*Studies that used the term PCR are cited in this rationale. Since the original study authors used the term PCR, this rationale will use the term PCR when specifically describing results from studies which used that term. However, the Work Group favors the term PNA.




© 2001 National Kidney Foundation, Inc

web version created by cyberNephrologyTM and The Nephron Information Center