I. Initiation of Dialysis


Two clinical guidelines for when to initiate dialysis are provided because there appear to be two independent predictors of clinical outcome. The first guideline is based on the level of kidney function (as measured by Krt/Vurea per week); the second is based on nutritional indices and is located in the K/DOQI Clinical Practice Guidelines in Chronic Renal Failure (Guideline 27).2

Although less than 1% of American dialysis patients begin dialysis with a serum creatinine concentration <8.0 mg/dL or a CCr >10 mL/min, approximately 60% suffer from nausea/vomiting at the time of dialysis initiation.3 Thus, the likelihood of malnutrition in this population is high. Evidence from the Modification of Diet in Renal Disease (MDRD) study4 and a recent large Australian study5 clearly show that when the glomerular filtration rate (GFR) decreases to 25 to 50 mL/min, patients adapt by reducing their protein intake. Protein intake continues to decline as kidney disease progresses to kidney failure, administratively termed end-stage renal disease (ESRD). These observations have been corroborated in a prospective study.6 As kidney function deteriorates, protein and energy intake decreases, leading to changes in body weight, fat mass, serum albumin, and transferrin concentrations. Earlier initiation of dialysis may prevent or perhaps even reverse this deterioration in nutritional status. Increased serum albumin concentration has been shown to parallel the increase in Kt/Vurea for HD patients.7 In addition, an editorial review of cited data8 suggests that albumin level at initiation of dialysis is predictive of survival.9

Adverse clinical and economic consequences of failure to properly manage patients with progressive chronic kidney disease as they approach ESRD and become dialysis dependent were first described in Britain.10 These observations have now been corroborated in other regions of Britain, the United States, and France.11-14 Specifically, costs, hospitalization, and morbidities decrease if attention is paid to nutrition, acid-base status, hypocalcemia, hyperphosphatemia, anemia, hypertension, volume status, and dialysis access (vascular or peritoneal). Hence, it is likely that delays in referral for initiation of dialysis result in unnecessary morbidity and potentially higher costs as well. While seeing a nephrologist does not guarantee that patients will be adequately prepared and referred for dialysis, it dramatically increases the likelihood that this will occur.13

The initiation of dialysis guidelines as described in this section are based on adult data. No such data exist yet in children.


When to Initiate Dialysis–Kt/Vurea Criterion (Opinion)

Unless certain conditions are met, patients should be advised to initiate some form of dialysis when the weekly renal Kt/Vurea (Krt/Vurea) falls below 2.0. The conditions that may indicate dialysis is not yet necessary even though the weekly Krt/Vurea is less than 2.0 are:

1. Stable or increased edema-free body weight. Supportive objective parameters for adequate nutrition include a lean body mass >63%, subjective global assessment score indicative of adequate nutrition (see Guideline 12: Nutritional Status Assessment, and Appendix B: Detailed Rationale for Guideline 2) and a serum albumin concentration in excess of the lower limit for the lab, and stable or rising; and

2. Nutritional indications for the initiation of renal replacement therapy are detailed in Guideline 27 of the NKF-K/DOQI Clinical Practice Guidelines on Nutrition, part of which is reproduced as Guideline 2 of the PD Adequacy Guidelines.

3. Complete absence of clinical signs or symptoms attributable to uremia.

A weekly Krt/Vurea of 2.0 approximates a kidney urea clearance of 7 mL/min and a kidney creatinine clearance that varies between 9 to 14 mL/min/1.73 m2. Urea clearance should be normalized to total body water (V) and creatinine clearance should be expressed per 1.73 m2 of body surface area. The GFR, which is estimated by the arithmetic mean of the urea and creatinine clearances, will be approximately 10.5 mL/min/1.73 m2 when the Krt/Vurea is about 2.0.

Rationale A detailed rationale is described in Appendix A. The following is a summary.

It is paradoxical that nephrologists have focused on optimizing urea clearance once patients are started on dialysis, but have accepted much lower levels of kidney urea clearance during the pre-dialysis phase of patient management. For example, a weekly total (residual renal plus peritoneal dialysis) Kt/Vurea (Kprt/Vurea) of 2.0 or higher is associated with improved outcomes in patients on PD (see Guideline 15: Weekly Dose of CAPD), yet dialysis is usually not initiated until weekly Krt/Vurea falls to the range of 0.71 to 1.3. There is no definitive direct proof for the belief that a given level of urea clearance by the kidney is associated with better control of uremia than PD with this same urea clearance. In fact, recent studies suggest that the relationship between protein intake and weekly Kt/Vurea is nearly identical in patients with chronic renal failure not yet on dialysis and in patients on PD. Thus, until proven otherwise, residual kidney and peritoneal clearances of small solutes should be considered equivalent.

Once Krt/Vurea falls below 2.0 per week, patients should be considered at increased risk for malnutrition and uremic complications. With further decreases in Krt/Vurea in the absence of renal replacement therapy, the risk increases. Dialysis, or some form of renal replacement therapy, should be strongly considered when Krt/Vurea falls below 2.0 (or CCr falls in the range of 9 to 14 mL/min/1.73 m2) and definitely implemented if:

1. Despite vigorous attempts to optimize protein and energy intake, any of the following nutritional indicators show evidence of deterioration: (a) more than a 6% involuntary reduction in edema-free usual body weight (%UBW) or to less than 90% of standard body weight (NHANES II) in less than 6 months; (b) a reduction in serum albumin by greater than or equal to 0.3 g/dL and to less than 4.0 g/dL (see Nutrition Guideline 3), in the absence of acute infection or inflammation, confirmed by repeat laboratory testing; or (c) a deterioration in SGA by one category (ie, normal, mild moderate, severe; see Nutrition Guideline 9 and Nutrition Appendix VI).

If PD is initiated, the Kpt/Vurea could be increased incrementally so the combined weekly value of Krt/Vurea + Kpt/Vurea (Kprt/Vurea or total Kt/Vurea) does not fall below the target level of 2.0. With the incremental initiation approach frequent measurement of residual kidney function (RKF) will be necessary to assure that total delivered solute removal does not drop below targets (see Guidelines 3: Frequency of Delivered PD Dose and Total Solute Clearance Measurement Within Six Months of Initiation, and Guideline 5: Frequency of Measurement of Kt/Vurea, Total CCr, PNA, and Total Creatinine Appearance). Alternatively, the initiation of a "full dose" of PD may be offered (equivalent of four 2-L exchanges per day, which may yield a weekly Kpt/Vurea of 1.5 to 2.0, depending on transport characteristics, ultrafiltration, and body size). With initiation of "full dose" PD, frequency of measurement of RRF can be less intense.

The Work Group strongly supports the opinion that the PD outcome data for a weekly Kt/Vurea of 2.0 are so compelling that using the same figure for initiation of dialysis justifies the small risks of performing peritoneal dialysis. Those risks include infections and the possibility that increasing the length of time on PD contributes to eventual patient "burn-out." If a patient is suspected to be at high risk for these complications, PD may not be the best choice for renal replacement therapy. The Work Group acknowledges that the risks of early initiation of PD are not clearly known, but that the risks of late initiation are known and are unacceptable. Furthermore, not knowing which initiation strategy (incremental versus full therapy initiation) is better, the Work Group recommends that either approach be used to reach or exceed targets.

Compared to CAPD, it is more complex to calculate the incremental dose of hemodialysis (HD) that would be needed such that the total continuous delivered weekly Kt/Vurea would be greater than 2.0. However, it can be estimated using the fundamental assumption underlying CAPD, that at the same protein catabolic rate, continuous renal replacement therapy must keep the steady state BUN equal to the average pre-hemodialysis BUN. (See Appendix G for further discussion of this assumption.) If weekly Krt/Vurea is 1.6 , for example, a one time per week HD treatment must deliver an equilibrated (double-pool) Kt/Vurea of 2.0 to achieve a total continuous weekly Kt/Vurea equivalent to 2.0. This is quite difficult to achieve, so two HD treatments per week may be more realistic for this level of RRF. If weekly Krt/Vurea is 0.5, two HD treatments must each deliver an equilibrated (double-pool) Kt/Vurea of 2.0 to achieve a total continuous weekly Kt/Vurea equivalent to 2.0. This is also quite difficult to achieve, so three HD treatments per week may be more realistic for this level of RKF. More technical details about intermittent HD are described in Appendix A, including the role of biocompatible membranes to help preserve RKF.

It is a general consensus that patients with diabetes should initiate dialysis at levels of RKF higher than in patients with causes of ESRD other than diabetes. That practice is not altered by this guideline.

The Work Group also recognizes that for many clinicians, initiating dialysis based on Kt/Vurea is a new concept. Therefore, we have attempted to equate this to the traditional measure of urea clearance, CCr, and GFR (estimated by the arithmetic mean of urea and creatinine clearance).

The Work Group recognizes that the patient will play a major role in accepting the initiation of dialysis based on a certain "laboratory value." It is the responsibility of the care providers to make clear to the patient the rationale for initiating dialysis when the above conditions become applicable. In particular, the nephrologist must explain to the patient the risk of malnutrition with delayed initiation of dialysis and the strong inferential evidence that survival might be improved with an earlier start of dialysis. Thus, appropriate patient education regarding an informed decision about dialysis is necessary. Medical conditions that may explain why dialysis is not being initiated when weekly Krt/Vurea is less than 2.0 need to be documented. These conditions are described above.

Some individuals have expressed concern that this guideline will run afoul of the Health Care Financing Administration (HCFA) regulations regarding the initiation of dialysis (eg, form 2728, ESRD Medicare Medical Evidence Report). The leadership of the NKF-K/DOQI is working with HCFA to ensure that this will not be the case.


Indications for Renal Replacement Therapy

In patients with chronic kidney failure (eg, GFR < 15 to 20 mL/min) who are not undergoing maintenance dialysis, if protein-energy malnutrition (PEM) develops or persists despite vigorous attempts to optimize protein and energy intake and there is no apparent cause for malnutrition other than low nutrient intake, initiation of maintenance dialysis or a renal transplant is recommended. (Opinion)

Note: This is Guideline 27 of the K/DOQI Nutrition Guidelines, reproduced here without the specific reference citations included. See the Nutrition Guidelines2 for these details. This Guideline was written by members of both the PD Adequacy and Nutrition Work Groups.

Rationale It is well documented that mortality and morbidity are increased in individuals with ESRD who begin dialysis therapy with overt evidence of PEM. Accumulating evidence also indicates that initiation of dialysis more in line with current NKF-K/DOQI practice guidelines (ie, GFR ~10.5 mL/min) results in improved patient outcomes compared with when dialysis is delayed until the GFR is <5 mL/min and symptomatic uremia and associated medical complications are present. Furthermore, there is evidence that initiating maintenance dialysis under these circumstances, and when there has been nutritional deterioration, results in an improvement in nutritional indices. There is no evidence that earlier initiation of dialysis leads to improved nutritional status among patients without overt uremia. Moreover, it has not been established that improved nutritional status at the initiation of dialysis directly leads to improved survival or fewer dialysis-related complications. Despite the lack of evidence from controlled clinical trials, interventions that maintain or improve nutritional status before the requirement for renal replacement therapy are likely to result in improved long-term survival.

There is ample evidence that the survival of patients with ESRD is closely associated with their nutritional status (Guidelines 3 through 6, 8, 18, and 23). These findings have been demonstrated not only in large, diverse populations of prevalent maintenance dialysis (MD) patients, but also in patients commencing MD therapy. Hypertension, pre-existing cardiac disease, and low serum albumin concentrations were independently associated with diminished long-term survival in 683 ESRD patients who started dialysis during 1970 through 1989. In 1,982 hemodialysis (HD) patients, a low serum albumin concentration at the initiation of dialysis was associated with a significant increase in the relative risk of death. A direct relation between serum albumin and survival and an independent association between modified SGA and survival was observed in 680 incident CPD patients. In contrast, in one study no significant associations were found between serum albumin, creatinine, and urea concentrations and survival in incident HD patients. The sample size in the latter study was relatively small (n = 139), and 94% of the study sample were black (83%) or Hispanic (11%). No studies have specifically examined the relations among other nutritional indicators (eg, %SBW, PNA, and DEXA) and survival in incident HD or peritoneal dialysis patients.

Low-protein (eg, 0.60 g protein/kg/d), high energy (35 kcal/kg/d) diets may retard the rate of progression of chronic kidney disease (CKD) and should maintain patients with chronic renal disease in good nutritional status (Guidelines 24 and 25). However, it is recognized that such low-protein diets may not maintain adequate nutritional status in all patients, particularly if an adequate energy intake is not maintained (Guideline 25). Furthermore, there is evidence that the spontaneous intake of protein and energy, and other indicators of nutritional status, tend to diminish in patients with progressive CKD who are consuming unregulated diets. Therefore, patients with CKD need to undergo nutritional assessment at frequent intervals so that any deterioration in nutritional status can be detected early (Guidelines 23 and 26 and Appendix IV). The plan of care and nutritional interventions outlined in Guideline 18 for the nutritional management of the dialysis patient is also appropriate for patients with progressive CRI.

Because of the association between PEM and poor outcome, it is recommended that MD be initiated or kidney transplantation performed in patients with advanced CKD (ie, GFR <20 mL/min) if there is evidence of deteriorating nutritional status or frank PEM, no other apparent cause for the malnutrition, and efforts to correct the nutritional deterioration or PEM are unsuccessful, despite the absence of other traditional indications for dialysis (eg, pericarditis or hyperkalemia). Although the following criteria are not considered rigid or definitive, initiation of renal replacement therapy should be considered if, despite vigorous attempts to optimize protein and energy intake, any of the following nutritional indicators show evidence of deterioration: (1) more than a 6% involuntary reduction in edema-free usual body weight (%UBW) or to less than 90% of standard body weight (NHANES II) in less than 6 months; (2) a reduction in serum albumin by greater than or equal to 0.3 g/dL and to less than 4.0 g/dL (Guideline 3), in the absence of acute infection or inflammation, confirmed by repeat laboratory testing; or (3) a deterioration in SGA by one category (ie, normal, mild, moderate, or severe; Guideline 9 and Appendix VI).


1. Studies to assess the optimal timing and indications for commencing renal replacement therapy are needed.

2. Serial evaluations of nutritional status in the course of these studies will help to determine whether initiation of dialysis indeed improves nutritional status.

3. Studies should be conducted to determine whether any GFR level can be used to indicate when maintenance dialysis should be initiated.

4. Whether earlier initiation of renal replacement therapy can prevent the development or worsening of PEM and its attendant complications needs to be evaluated in a controlled study.





© 2001 National Kidney Foundation, Inc

web version created by cyberNephrologyTM and The Nephron Information Center