Although residual renal function is routinely incorporated in the prescription of chronic peritoneal dialysis, it is generally ignored in the determination of the dose of maintenance hemodialysis. The major reason cited for this difference in approach is multiple observations of a more rapid decay of residual renal function in the hemodialysis setting. 1–4 Such instability in residual renal function in hemodialysis patients might place them at risk of uremia and, hence, necessitate earlier increases in their dose of dialysis than is required for their peritoneal dialysis counterparts. These difficulties would make such a labor intensive strategy of marginal value in the management of hemodialysis patients.
There are good reasons, however, to consider adjusting the dose of dialysis for residual renal function in selected hemodialysis patients. Substantial and relatively stable levels of residual renal function are seen in many hemodialysis patients, particularly in view of recent trends to initiate hemodialysis earlier 5 and, perhaps, to use biocompatible membranes. 6–9 Integration of endogenous renal function allows for shorter dialysis sessions in these patients, with the inherent potential for enhancing their quality of life. Since early 1999, our dialysis facility has been structuring the hemodialysis prescription around the residual renal urea clearance. We performed a retrospective study in a consecutive series of hemodialysis patients to assess the behavior of residual renal function over time and to determine the actual impact of a strategy that incorporates such function on the dose of dialysis.
Dialysis Unit and Apparatus
Our hospital based dialysis facility has 24 stations and provides roughly 27,000 treatments annually for nearly 200 dialysis patients in the central New York region. During the study period, Althin 1000 machines were in use with hollow fiber dialyzers, incorporating CAHP 210 cellulose acetate membranes (Baxter Health care Corporation, Deerfield, IL). Bicarbonate containing dialysis solution was delivered at 500 ml/min.
Patient Selection and Chart Review
This study was approved by the Institutional Review Board at United Health Services Hospitals. All patients with end-stage renal disease (ESRD) started on maintenance hemodialysis between October 1998 and July 2000 were included in this study. Their records, including dialysis charts, hospital charts, and emergency room records were reviewed to establish the patient’s age, gender, race, causes of renal failure, initial and final dates of dialysis, as well as the development of uremic and congestive complications. Data collection ended October 2000. For each testing period, a new set of urea kinetic modeling data was collected, including weekly residual renal urea clearance (Krt/V urea), dialysis Kt/V, and delivered and prescribed doses of dialysis.
Urea Kinetic Modeling and Dialysis Prescription
Determinations of residual renal function and dialysis Kt/V were carried out in February, June, and October of 1999, but on a quarterly basis since January 2000; the urea reduction ratio (URR) has always been monitored on a monthly basis. Patients undergoing multiple hemodialysis treatments per week are asked to collect their urine from the end of one dialysis treatment to the beginning of the next, whereas patients who dialyze once weekly collect urine for a 24 hour period preceding the index dialysis treatment. Blood is drawn for measurements of blood urea nitrogen (BUN) at the onset and completion of the urine collection, and urea clearance is determined by UV/P, where U is urine urea concentration, V is urine flow rate, and P is the mean value of initial and final BUNs. Urea clearance (ml/min) is converted to a weekly Krt/V urea by dividing the product of the urea clearance and number of minutes per week (10,800) by the mean volume of distribution for urea.
Urea kinetic modeling was performed using the single compartment variable volume model as described by Depner. 10 This model yields the current delivered dose of dialysis and is then used to predict the schedule necessary to achieve a target weekly total Kt/V of 3.6, the current DOQI standard, 11 using a close approximation of nomograms developed by Gotch. 12 This target total weekly Kt/V of 3.6 represents the sum of Kt/V due to hemodialysis and that contributed by residual renal function. For all patients with residual renal function, the above calculations were repeated assuming their Krt/V urea to be zero to determine the required dialysis dose if they had no residual renal function. The difference between the two yields the potential reduction of dialysis time afforded by incorporation of Krt/V urea. Recommended reductions of dialysis time were generally implemented unless the criteria enumerated in Table 1 applied in the involved patient. 13
SPSS software was used for data entry and analysis. Data analyses included paired and non-paired Student’s t and chi-squared tests. Relationships among urine volume, Krt/V urea, and potential and actual reduction in time were determined by linear regression. Whenever appropriate, data in the tables are expressed as mean ± SD, with p < 0.05 considered statistically significant.
During the study period, 69 patients with chronic renal disease, of which 51 patients (74%) had measurable levels of residual renal function, were initiated on maintenance hemodialysis. The demographic characteristics of these 51 patients are shown in Table 2. Patients were followed for a mean duration of 14.4 months. None of the study subjects moved or received a kidney transplant during the study period, but three patients transferred to peritoneal dialysis and 12 patients died.
Figure 1 divides all patients into four ranges of Krt/V urea and displays the numbers of patients in each range in temporal manner with respect to sampling time points. The data show a rapid reduction in Krt/V urea for most patients. However, a significant minority did retain substantial levels of renal function, e.g., 12 patients possessed a Krt/V urea > 0.40 at the fourth sampling time point (approximately 1 year). This subgroup of patients represents approximately 24% of the initial group of 51 patients and 46% of the residual group of 26 patients reaching this time point.
Thirty-one individuals had adjustments in their prescribed dialysis times at least once during the study period, whereas 20 did not for reasons outlined in Table 1. Of the 111 modeling studies carried out in the group of 31 patients, 72 (65%) led to actual reductions in time, but 39 (35%) did not because of the same clinical limitations. Figure 2A reveals the actual reduction in dialysis time across the seven sampling time points for these 31 patients. The average time reduction for all measurements is 70 ± 98 minutes per week. Figure 2B shows time reduction data for a subgroup of 10 patients whose dialysis time was adjusted on the basis of Krt/V urea throughout their first year of dialysis. This particular subgroup of patients continued to dialyze 99 ± 98 minutes per week less than they would have had their renal function been ignored.
Figure 3 shows potential reductions (Figure 3A) as well as actual reductions (Figure 3B) in time versus Krt/V urea. Direct and somewhat predictable relationships were observed between these variables. Regression analyses for patients on dialysis three times per week reveal that each 0.10 increment in Krt/V urea yielded additional potential and actual time reductions of 6 minutes and 4 minutes per treatment, respectively. For the group of 31 patients whose residual renal function was incorporated into their dialysis prescription, the cumulative actual time reduction was 325 hours over the study period or 80 minutes per week per patient.
Figure 4 demonstrates the relationship between urine volume and Krt/V urea. Thirty-six percent of the entire group of patients was on furosemide in doses that ranged from 40 mg/day to 320 mg/day. The extreme polyuric observations were derived from one patient with end-stage renal disease secondary to chronic lithium nephropathy. There was a significant correlation between these variables (r = 0.74;p < 0.01) that allows for Krt/V urea to serve as a mediator for the prediction of the effect of urine volume on time reductions. For example, regression analysis predicts that, on a thrice-weekly schedule, a potential time reduction of 30 minutes per treatment required a Krt/V urea level of 0.48. This, in turn, corresponds with a urine volume of approximately 600 ml per day. Twelve of 20 (60%) such patients with substantial levels of Krt/V urea were on furosemide.
Table 3 shows the initial and final urine volumes, urea kinetic modeling results, and dialysis schedules in patients with residual renal function who did (n = 31) versus did not (n = 20) have their time adjusted on the basis of Krt/V urea. The group of patients whose time was adjusted demonstrated levels of urine output and Krt/V urea that were significantly higher than the group whose time was not adjusted. The time adjusted group displayed lower values for URR and hemodialysis Kt/V, but not significantly so. Within each group, reductions in the urinary parameters and increases in the modeling variables were observed over time. The average duration of individual treatments was similar between the groups, but patients whose time was adjusted on the basis of residual renal function dialyzed significantly less frequently than patients whose time was not adjusted.
Complications attributable to the above dosing strategy were confined to one case. A patient with high levels of Krt/V urea was withdrawn from dialysis therapy soon after being observed to have a value of 2.9 at a time when his arteriovenous graft clotted. Soon thereafter, he developed congestive heart failure and was reinitiated on dialysis. Otherwise, no uremic or congestive symptoms were observed in patients who necessitated changes in the dialysis prescription before routine modeling.
This study describes our experience with incorporation of residual renal function into the maintenance hemodialysis prescription. A major concern in adjusting the dose of hemodialysis for residual renal function has been the rapidity with which such function has been shown to decline in patients on this mode of dialysis. However, Van Stone found that endogenous filtration was still sufficiently well preserved to provide 15% and 30% of the dialysis requirement for 19% and 9% of patients still dialyzing 3 years after initiation of maintenance therapy, respectively. 6 We observed that 44% of 27 patients on hemodialysis for approximately 1 year retained sufficient levels of residual renal function (Krt/V urea > 0.4) to potentially provide for 11–21% of their dialysis requirement. This observation assumes even greater importance in view of the relatively short span of time that patients currently spend on hemodialysis.
Another concern often voiced regarding adjustment of treatment times on the basis of Krt/V urea is the potential for resistance of patients to the need for eventual increases in their dialysis schedule. In fact, over the period of time that we have been using the above method for calculating the dose of dialysis, patient complaints have been few, mild, and manageable. This is likely due to patient education and a developing culture of incremental dialysis established at our facility.
We have quantified the relationship between weekly Krt/V urea and the dialysis schedule when adjusted to a total weekly Kt/V of 3.6. In general, such a strategy is most applicable during the initiation phase of dialysis dependency. Our results showed that 31 patients had sufficient residual renal function to justify lower treatment times for at least one period of dialysis, with 10 patients continually benefiting over 1 year and, in four cases, beyond. In our study, the frequency rather than duration of dialysis treatments were predominantly affected. In view of current demands from regulatory agencies for treatment URR > 65% and Kt/V > 1.2, changes in the hemodialysis schedule would need to be relegated to frequency and not duration of treatments. However, if further studies validate our short-term favorable outcome results and extend them over the longer term, then such agencies should reevaluate their current recommendations.
In whom should adjustment of the dose of hemodialysis be considered? Our study showed a direct relationship between daily urine output and weekly Krt/V urea, which allows for estimation of a minimum urine volume required for its inclusion in the modeling procedure. For example, if one believes that a time adjustment of 30 minutes per treatment is the minimum time one should accept for the practice of integrating native kidney function into the dialysis prescription, then for patients dialyzing three times weekly, it is only worth processing urine from patients who excrete at least 680 ml per day. Patients with potential time reductions should be scrutinized closely for reasons to justify such reductions, i.e., arbitrarily “excessive” initial hemodialysis prescriptions or evidence for improvement in renal function. For those patients in whom reasons for dose reduction are identified, the remainder of the exclusion criteria outlined in Table 1 should be applied to avoid uremic and congestive complications. In our study, such clinical considerations limited dialysis time reductions to approximately two thirds of those recommended on the basis of residual renal function.
Whether or not Kt/V is decided to be a valid and complete measure of the adequacy of dialysis is beyond the scope of this study. 14,15 Nevertheless, Kt/V remains the current yardstick for measuring the adequacy of dialysis. It is standard practice to incorporate residual renal function into the peritoneal dialysis prescription. Our results suggest that doing so in the hemodialysis setting permits some patients to derive significant time reductions for at least 1 year after initiation of therapy. In the short term, we observed only one patient who developed congestive complications as a result of measurement of Krt/V urea. However, more subtle factors could jeopardize the long-term health of patients whose time is reduced on the basis of their residual renal function. The number of patients in our study was small and the follow-up period relatively short, precluding such outcome analysis. There are retrospective data available, however, to suggest that the practice of adjusting hemodialysis time to Krt/V urea might be safe. In 1999, Hanson et al. published results from analysis of the United States Renal Data System showing that patients on twice weekly dialysis had a 24% lower mortality rate than those with less endogenous renal function on a three times weekly regimen, even when adjusted for a large number of covariates. 16 In this regard, it must be noted that residual renal function bequeaths a survival advantage independent of dialysis factors. 17 A long-term prospective outcome study comparing sizable groups of patients with equivalent levels of renal function, comorbidities, and dialysis techniques apart from integration of residual renal function would need to be carried out to definitively settle this question.
In no way should our results be interpreted to suggest that patients on intermittent hemodialysis therapy be on minimum maintenance dialysis treatment regimens. Rather, they indicate that if a specific Kt/V is targeted, then residual renal function may be rather simply integrated into the calculated dose of dialysis. The key requirements are cooperation of dialysis staff and patients, a minimum of quarterly urine collections, software that incorporates Krt/V urea into the dialysis prescription, and cautious interpretation of results that allows for maintenance or augmentation of dialysis time to accommodate adequacy needs beyond Kt/V. Greater flexibility in prescribing “equivalent” hemodialysis doses for patients to be treated on frequent, even daily, schedules may be gained from use of the graph developed by Casino. 18 The value of incorporating residual renal function may be greatest with earlier initiation of and incremental approaches to hemodialysis that are being entertained in the nephrology community. 5,19
The authors thank Karen Burdine and Vicki Fiacco for their very generous help with data collection. This research was performed in the dialysis unit at Binghamton General Hospital, Binghamton, NY, and financially supported by its parent organization, United Health Services Hospitals, Binghamton, NY.
1. Lysaght MJ, Vonesh E, Gotch F, et al: The influence of dialysis treatment modality on the decline in remaining renal function. Trans ASAIO 16: 598–604, 1991.
2. Rottembourg J: Residual renal function and recovery of renal function in patients treated by CAPD. Kidney Int 40 Suppl: S106–S110, 1993.
3. Moist LM, Port FK, Orzol SM, et al: Predictors of loss of residual renal function among new dialysis patients. J Am Soc Nephrol 11: 556–564, 2000.
4. Iest CG, Vanholder RC, Ringoir SM: Loss of residual renal function in patients on regular haemodialysis. Int J Artif Organs 12: 159–164, 1989.
5. Keshaviah P, Emerson P, Nolph K: Timely initiation of dialysis: A urea kinetic approach. Am J Kidney Dis 33: 344–348, 1999.
6. Van Stone JC: The effect of dialyzer membrane and etiology of kidney disease on the preservation of residual renal function in chronic hemodialysis patients. ASAIO J 41: M713–M716, 1995.
7. Hartmann J, Fincke H, Schiffl H: Biocompatible membranes preserve residual renal function in patients undergoing regular hemodialysis. Am J Kidney Dis 30: 366–373, 1997.
8. McCarthy JT, Jenson BM, Squillace DP, Williams AM: Improved preservation of residual renal function in chronic hemodialysis patients using polysulfone dialyzers. Am J Kidney Dis 4: 576–583, 1997.
9. Ackert K, Young BS, Chertow GM: The decline in residual renal function in hemodialysis is slow and age-dependent. J Am Soc Nephrol 12: 445A, 2001.
10. Depner T: Single-compartment model, in Depner TA (ed), Prescribing Hemodialysis: A Guide to Urea Modeling
. Norwell MA: Kluwer Academic Publishers, 1991, pp. 65–89.
11. Hemodialysis Adequacy Work Group I: Hemodialysis dose. NKF-DOQI clinical practice guidelines for hemodialysis adequacy: Update 2000. Am J Kidney Dis 37 (Suppl): S7–S64, 2001.
12. Hemodialysis Adequacy Work Group: Appendix A: Detailed rationale for guidelines 1. NKF-DOQI clinical practice guidelines for hemodialysis adequacy. Am J Kidney Dis 37: S109–S115, 2001.
13. Abuelo JG: Large interdialytic weight gains: Causes, consequences, and corrective measures. Semin Dial 11: 25–32, 1998.
14. Charra B, Depner T, Vanholder R, et al: Is Kt/V a satisfactory measure for dosing the newer dialysis regimens? Semin Dial 14: 8–21, 2001.
15. Bommer J: If you wish to improve adequacy of dialysis, urea kinetics, such as Kt/V, may be the wrong parameter to study. ASAIO J 47: 189–191, 2001.
16. Hanson JA, Hulbert TE, Ojo AO, et al: Prescription of twice-weekly hemodialysis in the USA. Am J Nephrol 19: 625–633, 1999.
17. Shemin D, Bostom AG, Laliberty P, Dworkin LD: Residual renal function and mortality risk in hemodialysis patients. Am J Kidney Dis 38: 85–90, 2001.
18. Casino FG: The EKRc graph: A simple method to estimate the time-averaged urea clearance. Semin Dial 12: 11–14, 1999.
19. Golper T: Incremental dialysis. J Am Soc Nephrol 9: S107–S111, 1998.