Pervasive Failed Rehabilitation in Center-Based Maintenance Hemodialysis Patients

Pervasive Failed Rehabilitation in Center-Based Maintenance Hemodialysis Patients Onyekachi Ifudu, MD, Henry Paul, MD, Joan D. Mayers, RN, Linda S. Cohen, MPH, William F. Brezsnyak, Allen I. Herman, Morrell M. Avram, MD, and Eli A. Friedman, MD • At its inception in 1972, the end-stage renal disease (ESRD) program was conceived with a set of assumptions about cost, rate of growth, and treatment outcomes in its client population. Despite the potential to correct anemia with recombinant erythropoietin (EPO) introduced in 1987 and improved survival, the level of physical activity among some segments of the hemodialysis population remains suboptimal. This study was undertaken, among other reasons, to identify correlates of poor functional status as measured by a modified Kamofsky scale. Using a modified Kamofsky scale, we measured the functional status of 430 patients who had been treated by hemodialysis for at least 1 year and some of whom were also receiving concomitant treatment with EPO. Patients studied were randomly selected from eight dialysis units in urban New York and suburban New Jersey. A Kamofsky score of less than 70 indicated frank disability-the subject was unable to perform routine living chores without assistance. In addition, current vocational activity was ascertained, and comorbid conditions were quantified. The necessity for wheelchair dependence was noted for each patient. The mean age (±SD) of the study population was 56 ± 14 years (range, 21 to 92 years). Subjects had been on maintenance hemodialysis for 4.09 ± 3.8 years (range, 1 to 23 years). The study group included 215 men and 215 women, of whom 65% were black, 27% white, 6% Hispanic, and 2% Asian; 36.5% had diabetes mellitus. Although 376 members (87%) of the study group were under treatment with EPO, the mean hematocrit of the study population was only 29% ± 4.5%. As rated by the Kamofsky scale, 154 (36%) of 430 patients were unable to perform routine living chores without assistance. Dependence on a wheelchair was reported by 73 members (17%) of the study group, and severe debility resulted in 36 (8.4%) patients requiring a home attendant. The comorbidity index of patients who scored less than 70 on the Kamofsky scale was 3 ± 1.6, which indicated a substantive prevalence of comorbid conditions as compared with 0.9 ± 0.7 for those who scored at least 70 on the Kamofsky scale (P < 0.001). Only 43 (10%) patients were employed outside the home, although 147 members (33%) of the entire group were younger than age SO. Analysis of covariance showed that race (P < 0.001), gender (P < 0.002), age (P < 0.0001), and diabetes (P < 0.001) all independently affected functional status. Tukey's multiple comparison test (adjusted least-squares mean ± SE) showed that blacks had a higher score on the Kamofsky scale (74.4 ± 0.96) than did Hispanics (63.8 ± 2.8, P < 0.001), and whites also scored higher (71.5 ± 1.5) than Hispanics (P < 0.04). Men (72.3 ± 1.3) scored higher than women (67.6 ± 1.34, P < 0.002) on the Kamofsky scale, and nondiabetic patients (72.7 ± 1.3) scored higher than diabetic patients (67.2 ± 1.5, P < 0.001). Kamofsky scores decreased with advancing age (P < 0.0001). Of the laboratory variables measured, only serum creatinine concentration correlated significantly with Kamofsky scores (r = 0.39, P < 0.001). Indeed, 155 (56%) of the 276 patients who scored at least 70 on the modified Kamofsky scale had a serum creatinine concentration exceeding 12.5 mg/dL, as contrasted with 38 (25%) of 154 patients who scored less than 70 on the modified Kamofsky scale (P < 0.00001). Although the age-adjusted mean ± SE serum creatinine level was higher in blacks (13 ± 0.2 mg/dL) than in whites (10.5 ± 0.3 mg/dL), Kamofsky scale scores were equivalent for both races (P < 0.232). We conclude that a significant proportion of hemodialysis patients are functionally debilitated. Diabetics, women, Hispanics, the elderly, and patients with low serum creatinine concentrations were at greater risk for poor functional status than their respective counterparts. Also, the rate of employment of hemodialysiS patients remains very low. Treatment with EPO did not necessarily result in rehabilitation. © 1994 by the National Kidney Foundation, Inc. INDEX WORDS: Hemodialysis; rehabilitation; erythropoietin; employment; functional; creatinine; end-stage renal disease; comorbidity; diabetes mellitus.

E

ND-STAGE renal disease (ESRD) is unique in being the only major chronic illness for

From the SUNY Health Science Center at Brooklyn; The Nephrology Foundation of Brooklyn; and The Long Island College Hospital. Brooklyn. NY. Received July 29.1993; accepted in revisedform September 28.1993. Address reprint requests to Onyekachi Ifudu. MD. Renal Diseases Division. Box 52. Department of Medicine. SUNY Health Science Center at Brooklyn. 450 Clarkson Ave. Brooklyn. NY 11203. © 1994 by the National Kidney Foundation. Inc. 0272-6386/94/2303-0009$3.00/0 394

which the cost of care is funded by the government with the acceptance criterion being solely diagnosis. I At its inception in 1972, few demographers or policy makers anticipated that the ESRD program would grow to its current level in terms of number of patients under treatment or cost. 2-4 In the 1970s and 1980s, a series of studies on the rehabilitation of hemodialysis patients detected both suboptimal physical activity and a low employment rate.5•6 Employment was used in prior surveys as a major index of rehabilitation primarily because healthcare policy makers were initially convinced that a great pro-

American Journal of Kidney Diseases, Vol 23. No 3 (March), 1994: pp 394-400

395

REHABILITATION IN HEMODIALYSIS

portion of hemodialysis patients would be vocationally rehabilitated, thereby contributing to the tax base. I Although several studies have found that the latter objective has not been achieved, other investigators have found improved "quality of life," especially when recombinant erythropoietin (EPO) was added to the treatment regimen. 7 •8 In the current atmosphere of impending healthcare "reform" and fiscal austerity, we reexplored the functional and vocational rehabilitation of maintenance hemodialysis patients by means of a multicenter survey of a large number of patients, most of whom were under treatment with EPO. In addition, we sought to identify demographic and clinical variables that correlate with functional status. SUBJECTS AND METHODS

Subjects Data were collected from 430 randomly selected maintenance hemodialysis patients in six ambulatory hemodialysis units in Brooklyn (two hospital-based units, four freestanding not-for-profit units) and two ambulatory hemodialysis units in suburban New Jersey (one hospital-based, one freestanding). Patients were invited to participate in the study in consecutive order if they were present for dialysis treatment on the day of the surveyors' visit. Of 440 patients asked to participate, only 10 (2.3%) declined. We interviewed only patients who had been on maintenance hemodialysis for at least 12 months. A cohort of 20 consecutively available patients was reinterviewed by a second investigator to validate the reproducibility of the scoring system. The survey was conducted by two physicians and two nurses. To minimize interinvestigator variability, all four surveyors were carefully briefed before proceeding with data collection to ensure consistent phrasing of questions. All interviews were performed on site within the various dialysis units. Consent for the study was obtained from the appropriate board at each institution. Information collected from each subject included age, gender, race, etiology of ESRD, years on hemodialysis, highest educational level achieved, type of health insurance, prior kidney transplants, coresidents, recombinant EPO therapy, and vocational status (employed outside the home, full-time student, or full-time homemaker). The most recent predialysis blood chemistry tests including serum creatinine, serum albumin, and hematocrit were reviewed.

Objective Parameters Functional status. We used a modified Kamofsky activity scale9 to assess the level of physical activity. Due to noted pitfalls in the Kamofsky scale, IO we increased the number of steps to 14 different levels of activity ranging from less than 30 (hospitalized, progressive fatal process) to 96 and greater (normal function, no disability), narrowing the range at each level to minimize observer variation (Table I). A score ofless

Table 1. Modified Karnofsky Scale Activity

Normal function, no disability Minor signs and symptoms, full activity Usual activities with effort Independent, most out-of-home activities Independent, limited to home Needs assistance with errands Needs assistance with meal preparation Needs assistance with bathing/ dressing Home attendant, not totally disabled Disabled, living at home Nursing home for chronic care Hospitalized, fair condition Hospitalized, poor condition Hospitalized, progressive fatal process

Score

96-100 91-95 81-94 76-80 70-75 65-69

.. Disabled

60-64 55-59 50-54 45-49 40-44 35-39 30-34 <30

than 70 indicated that the subject was unable to perform routine living chores without assistance-a marker of disability. Comorbidity index. The presence of significant medical conditions or disability associated with major organ systems was documented. Each of 15 medical conditions was assigned the number I, and a numerical comorbidity index was generated by totaling the medical conditions present in each patient. The presence of anyone of the following was noted: cancer, use of a walking aide, congestive heart failure, amputation, angina, blindness, obstructive airway disease, stroke, wheelchair use, arthritis, bone disease, endocrine disease, bowel disease, neuropsychiatric disease, and blood disease excluding anemia correctable with EPO therapy. For example, if a patient had an amputation and also used a wheelchair or walking aid, it counted as two comorbid conditions. Similarly, the presence of angina and congestive heart failure in a patient would count as two comorbid conditions. The maximum possible comorbidity index for each subject was 15.

Statistical Analysis Comparison of groups for statistical significance was performed using a nonpaired Student's t test. Fisher's Exact Test was used for comparison of within-group changes. Using an analysis of covariance in which the Kamofsky score was the outcome, we examined the relationships of age, gender, ethnicity, and diabetes mellitus. Individual comparisons were adjusted for multiple comparison bias by Tukey's Honest Difference method. II Computations were performed on a 386 PC using Systat (IBM, Armonk, Ny).12 Unless otherwise indicated, all plus-minus values are the mean ± standard deviation. All reported P values are two-tailed.

IFUDU ET AL

396

RESULTS

General A total of 430 patients were studied. The mean age of the study group was 56 ± 14 years (range, 21 to 92 years), as shown in Table 2. There were 215 men and 215 women, and the racial distribution was 280 (65%) blacks, 114 (27%) whites, 27 (6%) Hispanics, and nine (2%) Asians. Renal diagnoses in the study subjects were as follows: hypertension, 162 (37.7%); diabetes mellitus, 157 (36.5%); chronic glomerulonephritis, 21 (4.9%); adult polycystic kidney disease, 15 (3.5%); systemic lupus erythematosus, 14 (3.3%); heroin nephropathy, nine (2%); obstructive nephropathy, eight (1.9%); drug toxicity, five (1.2%); congenital kidney disease, four (0.9%); sickle cell disease, one (0.2%); and unknown, 34 (7.9%). The mean duration of maintenance hemodialysis was 4.09 ± 3.8 years (range, 1 to 23 years). Educational histories of the group showed that 72 (17%) had a college education, 237 (55%) had completed high school, and 121 (28%) had less than a high school education. Medical insurance coverage was provided only by Medicaid for 84 patients, only by Medicare for 65 patients, by both Medicare and Medicaid for 132 patients, by private insurance for 16 patients, and by Medicare and private insurance for 111 patients; 22 patients had all three coverages. Only 26 patients (6%) had received a kidney transplant. One patient had two prior kidney transplants, and another had three prior kidney transplants. The majority of patients, 417 (97%), lived at home, and 13 resided in a chronic care facility. Of 417 patients living at home, 318 (76%) lived with their family or a friend, 63 (15%) lived alone, and 36 (9%) lived with a home atteridant. EPO was regularly administered to 376 (87%) of 430 patients. As determined on the day of survey, predialysis laboratory test results were as follows: serum creatinine, 12.2 ± 3.9 mg/dL; serum albumin, 3.7 ± 0.4 g/dL; and hematocrit, 29% ±4.5%. Vocational Status Only three subjects were full-time students. A total of 70 (16%) patients were full-time homemakers, and 43 (10%) patients were employed outside the home. College-educated patients were most likely to be employed outside the home.

Table 2. Demographic Profile and Laboratory Data of Subjects (n = 430) Age (yr) Mean Range Gender (M/F) Diabetics Nondiabetics Mean duration of MD (yr) Race Black White Hispanic Asian Level of education Below high school High school College Serum creatinine (mg/dL) Hematocrit (%) EPO-treated NoEPO Serum albumin (g/dL)

56 ± 14 21-92 215/215 157 (36.5%) 273 (63.5%) 4.09 ± 3.8 280 (65%) 114 (27%) 27(6%) 9(2%) 121 237 72 12.2

(28%) (55%) (17%) ± 3.9

28.5 ± 4.1 31.8 ± 4.9 3.7 ± 0.4

Abbreviation: MD, maintenance hemodialysis.

Whereas 21 (29%) of72 college-educated subjects were employed outside the home, only 16 (7%) of237 with a high school education and six (5%) of 121 subjects with less than a high school education were employed outside the home (P < 0.0001). Furthermore, of 193 patients aged less than 65 years who were deemed physically capable of working (ie, modified Karnofsky score ;;::: 76 = independent, participating in most outof-home activities), only 26 (13%) were employed outside the home. The explanation for unemployment proffered by most of the latter group was fear of losing all or part of health insurance benefits or a disability rating upon reporting employment. An unquantified subset of the purportedly unemployed group admitted to intermittent compensated work "off the books." Functional Status A surprisingly large fraction of the study population was disabled; 154 (36%) scored less than 70 on the modified Karnofsky scale, meaning that they were unable to perform routine living chores without assistance (Table 3). Dependence on a wheelchair was reported by 73 (17%) subjects of the study group. Diabetic patients had lower scores than their nondiabetic counterparts (Table

REHABILITATION IN HEMODIALYSIS

397

Table 3. Functional and Vocational Status of Patients (n = 430)

Table 4. Disability of Diabetic Subjects v Nondiabetic Subjects Diabetics (n = 157)

No. of Patients

Unable to perform routine living chores without assistance (ie, modified Karnofsky score < 70) Using a wheelchair Living with home attendant Employed outside the home Full-time homemaker Full-time student

154(36%) 73 (17%) 36 (8.4%) 43 (10%) 70 (16%) 3(0.7%)

4). Thus, 86 (55%) of 157 diabetic patients were unable to perform routine living chores without assistance, compared with 68 (25%) of 273 nondiabetic patients (P < 0.0001). Also, 45 (29%) of 157 diabetic patients were wheelchair-bound compared with 28 (10%) of nondiabetic patient~ who were dependent on a wheelchair (P < 0.001). However, when stratified into age groups as shown in Table 5, there is a paucity of diabetic patients below age 50 years, and the diabetic patients who were significantly more functionally impaired than their nondiabetic counterparts were those in the groups aged 50 to 69 years. Among those aged 70 years and older, nondiabetics had inferior scores to those of diabetics. This surprising finding perhaps reflects the fact that the diabetic survivors in this age range represent the strongest and best-conditioned diabetic patients. Correlates of Physical Activity

Analysis of covariance showed that race (P < 0.001), gender (P < 0.002), age (P < 0.0001), and diabetes (P < 0.001) all independently affected functional status. Tukey's multiple comparison test (adjusted least-squares mean ± SE) showed that blacks had a higher score on the Karnofsky scale (74.4 ± 0.96) than Hispanics (63.8 ± 2.8, P < 0.001), and whites also scored higher (71.5 ± 1.5) than Hispanics (P < 0.04). Men (72.3 ± 1.3) scored higher than women (67.6 ± 1.34, P < 0.002) on the Karnofsky scale, and nondiabetic patients (72.7 ± 1.3) had a higher score than diabetic patients (67.2 ± 1.5, P < 0.001). Karnofsky scores declined with advancing age (P < 0.0001). Ofthe major laboratory variables measured (serum creatinine, serum albumin, and hematocrit), only serum creatinine

Mean age (yr) Patients unable to perform routine living chores without assistance Patients using a wheelchair Mean comorbidity index

61.3±10.4

Nondiabetics (n = 273)

54

± 16

P

0.001

86 (55%)

68 (25%)

0.0001

45 (29%)

28 (10%)

0.001

2.4 ±

1.4 ± 1.4

0.0001

1.7

concentration correlated significantly with Karnofsky scores (r = 0.39, P < 0.001). Indeed, 155 (56%) of 276 patients who scored at least 70 on the modified Karnofsky scale had a serum creatinine concentration exceeding 12.5 mg/dL, compared with only 38 (25%) of 154 patients who scored less than 70 on the modified Karnofsky scale (P < 0.00001). Although the age-adjusted mean ± SE serum creatinine concentration was higher in blacks (13 ± 0.2 mg/dL) than in whites (10.5 ± 0.3 mg/dL), their mean scores on the Karnofsky scale were equivalent (blacks, 74.4 ± 0.96; whites, 71.5 ± 1.5; P < 0.232). Only 75 (17.4%) of 430 patients had a serum albumin concentration greater than 4 g/dL. Twenty (13%) of 154 patients with Karnofsky scores less than 70 had a serum albumin concentration greater than 4 g/dL, compared with 55 (20%) of276 subjects scoring at least 70 on the Karnofsky scale (P < 0.08).

Table 5. Percentage of Patients in Each Age Group Who Were Unable to Perform Routine Living Chores Without Assistance (ie, modified Karnofsky scores < 70)

Age Group (yr)

Diabetics (n = 157)

Nondiabetics (n = 273)

P

20-29 (n = 0 D, 14 ND) 30-39 (n = 5 D, 50 ND) 40-49 (n = 12 D, 66 ND) 50-59 (n = 47 D, 41 ND) 60-69 (n = 59 D, 64 ND) 70+ (n = 34 D, 38 ND)

40% 8.3% 40.4% 67.8% 70.6%

28.6% 8% 6% 9.8% 26.6% 92%

0.27 0.7 0.002 0.0001 0.02

Abbreviations: D, diabetic; ND, nondiabetic.

IFUDU ET AL

398

Comorbidity Index

The mean comorbidity index for patients who scored less than 70 on the Kamofsky scale was 3 ± 1.6, versus 0.9 ± 0.7 (P < 0.001) for patients who scored more than 70. As examples, a blind, lower-limb-amputee diabetic patient with heart disease would score a 4, whereas a patient with arthritis might score a 1. Comorbid medical conditions were more prevalent in diabetic subjects, resulting in a mean comorbidity index of2.4 ± 1.7, compared with a comorbidity index of 1.4 ± 1.4 in nondiabetic subjects (P < 0.0001; Table 4). EPO

Three hundred seventy-six (87%) of 430 subjects were receiving EPO, and their hematocrit was 28.5% ± 4.1 %. The mean dose of EPO administered to each patient was 4,260 U intravenously (range, 2,000 to 10,000 U), usually on a thrice-weekly schedule. The stated target hematocrit was 35%, but only 24 (6.4%) ofEPO-treated patients achieved this objective. The mean hematocrit in patients not treated with EPO was 3l.8% ± 4.9%, whereas 14 (26%) in the group attained a hematocrit of 35%. Anemia alone does not explain our finding of poor rehabilitation, because among the 38 patients with hematocrit values of at least 35%, 11 (29%) had Kamofsky scores less than 70, indicating the need for assistance with routine daily activities. Reproducibility of Results

Scores by a "blind" second observer were nearly identical to those obtained by the original investigator. Thus, mean first and second Karnofsky scores on the 20 subjects compared were 79 ± 12 and 77 ± 10, respectively, and first and second morbidity index ratings on the same patients were l.6 ± 1.1 and 1.65 ± l.1. DISCUSSION

Our main finding is that poor rehabilitation is pervasive in maintenance hemodialysis. Based on the Kamofsky score, 36% of the patients were disabled (required assistance with routine daily chores), whereas 8.4% were so impaired that they needed a home attendant, and 17% were wheelchair-bound. Diabetic patients had inferior rehabilitation compared with nondiabetic patients

as measured by the modified Kamofsky scale. Although diabetics and the elderly on hemodialysis have always had a worse outcome than their respective counterparts in most reports of functional rehabilitation 5,13,14-a reflection of the burden of comorbid medical conditions among both groups I 5, 16_t he finding ofa poorer outcome in Hispanic patients and in women is new and not easily explained. One possible explanation is that the excess debility in women was conferred by diabetes mellitus, since more women (92 [43%] of 215) carried the diagnosis of diabetes than men (65 [30%] of 215, P < 0.009). Whether the higher rate of diabetes mellitus in women than in men in our study subjects is enough to explain the gender disparity in functional status is unclear. It is noteworthy that a recent report on dialysis vascular access morbidity in the United States l ? found that women were at greater risk than men for vascular access complications-a problem that potentially leads to inadequate dialysis, and perhaps poor functional status and excess mortality. Furthermore, the 3-percentage point advantage in I-year survival that women on hemodialysis enjoyed over their male counterparts during the early and mid-1980s,2 has virtually disappeared in recent years. IS Our results show that only 10% of ambulatory hemodialysis patients surveyed were employed outside the home, whereas 16% were full-time homemakers. This represents a sharp decline from studies in the early 1980s that found 25% ofESRD patients to be employed. 5 Viewed from a different perspective, of the 193 patients aged less than 65 who were deemed functionally able (modified Kamofsky score ~ 76), only 26 (13%) were employed outside the home. One common explanation for unemployment in this group was anxiety over forfeiture of medical benefits and disability rating should they become employed. The evident failure ofEPO either to increase the mean hematocrit to at least 35% or to achieve broad rehabilitation must be viewed as a failure in application of the drug. Why EPO treatment failed to attain targeted goals of 35% hematocrit in the present series requires an exploration of dose range, iron studies, and blood loss during dialysis as possible mechanisms. We do not doubt that properly treated anemic hemodialysis patients with adequate iron stores will have their hematocrit increased to normal or above with

REHABILITATION IN HEMODIALYSIS

EPO. 19 Whatever the mechanism underlying an insufficient response to EPO, the fact stands that in our practice a major medical advance was not fully realized. The problem with EPO response was not limited to one dialysis unit. The range of mean hematocrit values in the eight dialysis units surveyed was 27% ± 4.4% to 33% ± 4.5%. Of interest and a possible clue to the overall poor response rate to EPO was the fact that the highest hematocrit value (33% ± 4.5%) was derived in a suburban dialysis unit, made up mainly of middle-class patients. We are surprised that serum creatinine concentration correlated significantly with functional status and serum albumin concentration did not, since both indices had been shown to be predictive of early death in hemodialysis patients when their levels are low. 2o We did not measure the adequacy of dialysis in this study, but note an overall mean serum albumin level of 3.7 ± 0.4 g/dL, which is less than the desired 4-g/dL level of well-dialyzed patients, as shown by Lowrie and Lew. 20 Also, since serum creatinine concentration has been shown to correlate directly with treatment duration,20 and since we demonstrated a significant correlation between serum creatinine concentration and functional status, we speculate that functional status, as measured on a modified Karnofsky scale, most likely correlates directly with dialysis treatment time. Whether the level of rehabilitation in our study population would be enhanced by longer dialysis treatment is conjecture at present. The reasons for our findings of poor functional status and low rate of employment are uncertain. In the two decades since the initiation of the Medicare ESRD program, profound changes have occurred both in uremia therapy and in the patient population accepted for dialysis. The influence of these changes on our findings is unclear because, although when examined individually the impact of changes such as the use of EPO l9 can be easily discerned, the outcome of all the variables in concert is not measurable or predictable. On the upside is the use of EPO, improvement in dialyzer and dialysate technology, and the influence of ESRD patient advocacy groupS.2,19 There have also been improvements in drug therapy for some severe components of the uremic syndrome, such as renal osteodystro-

399

phy, that respond to calcitriot,Z' and the avoidance of aluminum-containing antacids as phosphate binders. 22 Additionally, nephrologists have acquired extensive expertise in the management of dialysis patients. However, on the downside of attaining rehabilitation is the reality that newly admitted maintenance hemodialysis patients have been progressively older each year, with a mean age in 1991 of 56.6 years. 2Furthermore, new starts now tend to be sicker, since the prevalence of comorbid medical conditions is higher in older patients than in their younger counterparts. 2,16 Also to the point is the increasing proportion of diabetic patients who are more likely to deteriorate at a more rapid rate than nondiabetic patients. 2 Of great concern is the possible effect of economically forced shortened dialysis prescriptions on patient morbidity and mortality.20 Any impact of proposed healthcare reform on the ESRD program is unknown, but our findings are instructive and may be valuable in devising guidelines for governmental funding of life-prolonging therapy in the chronically ill. Patients who have chronic illness are under major physical and psychological burdens, and a realistic approach should be applied to the goals of therapy and the potential for both functional and vocational rehabilitation. Physicians, in their desire to be advocates for their patients, need not be driven to assure third-party payers as to future rehabilitation of the patient before medical coverage is granted. This contention is pertinent, since patients entering the ESRD program are now older and it is projected that by the year 2000 the mean age of new dialysis patients will be 60 years.2 Second, there is a need to incorporate into healthcare reform measures to ensure that ESRD patients or patients with other chronic illnesses who are physically fit and decide to work do not lose part or all of their medical benefits. In the meantime, efforts must be devoted toward maximizing the use of available resources to achieve excellent functional rehabilitation in ESRD patients, because life is much more than survival. The value of an extra hour or 2 of hemodialysis per week is a worthy subject for prospective study in this and similar dialysis populations. Whether our findings are representative of the American dialysis population is unclear. We recognize that the high proportion of inner-

400

IFUDU ET AL

city, underprivileged, health-deprived patients in our population may contribute to an outcome not representative of the nation as a whole.

REFERENCES 1. Gutman RA: High-cost life prolongation: The National Kidney Dialysis and Transplantation Study. Ann Intern Med 108:898-899, 1988 2. US Renal Data System: USRDS 1990 Annual Report. Bethesda, MD, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Disease, 1990 3. Friedman EA, Delano BG, Butt KMH: Pragmatic realities in uremia therapy. N Engl J Med 298:368-371, 1978 4. Iglehart JK: The American health care system-The End-Stage Renal Disease Program. N Engl J Med 328:366371,1993 5. Gutman RA, Stead WW, Robinson RR: Physical activity and employment status of patients on maintenance hemodialysis. N Engl J Med 304:309-313, 1981 6. Evans RW, Manninen DL, Garrison LP, Hart G, Blagg BR, Gutman RA, Hull AR, Lowrie EG: The quality of life of patients with end-stage renal disease. N Engl J Med 312: 553-559, 1985 7. Carlson DM, Johnson WJ, Kjellstrand CM: Functional status of patients with end-stage renal disease. Mayo Clin Proc 62:338-344, 1987 8. Evans RW, Rader B, Manninen DL, and the Cooperative Multicenter EPO Clinical Trial Group. The quality of life of hemodialysis recipients treated with recombinant human erythropoietin. JAMA 263:825-830, 1990 9. Karnofsky DA, Burchenal JH: The clinical evaluation of chemotherapeutic agents in cancer, in Macleod CM (ed): Evaluation of Chemotherapeutic Agents. New York, NY, Columbia University Press, 1949, pp 191-205 10. Hutchinson T A, Boyd NF, Feinstein AR: Scientific problems in clinical scales, as demonstrated in the Karnofsky Index of Performance status. J Chronic Dis 32:661-666, 1979

II. Kleinbaum DG, Kupper LL, Muller KE: Applied Regression Analysis and Other Multivariate Methods. Boston, MA, PWS-KENT, 1988, pp 365-368 12. Steinberg D, Colla P: A Supplementary Module for Systat. Evanston, IL, Systat Inc, 1991 13. Ifudu 0, Mayers J, Matthews J, Tan CC, Cambridge A, Friedman EA: Rehabilitation of elderly patients on hemodialysis, in Nephrology and Urology in the Aged Patient. Dordrecht, The Netherlands, Kluwer Academic, 1993, pp 277291 14. Bonney S, Finkelstein FO, Lytton B, Schiff M, Steele TE: The treatment of end-stage renal failure in a defined geographic area. Arch Intern Med 138:1510-1513,1978 15. Julius M, Hawthorne Y, Carpentier-Alting P, Kneisley J, Wolfe RA, Port FK: Independence in activities of daily living in end-stage renal disease patients: Biomedical and demographic correlates. Am J Kidney Dis 13:61-69, 1989 16. Shapiro FL, Umen AJ: Risk factors in hemodialysis patient survival. Am Assoc ArtifIntern Organs J 6:176-184, 1983 17. Feldman HI, Held PJ, Hutchinson JT, StC!iber E, Hartigan MF, Berlin JA: Hemodialysis vascular access morbidity in the United States. Kidney Int 43:1091-1096,1993 18. US Renal Data System: USRDS 1993 Annual Data Report. Bethesda, MD, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 1993 19. Eschbach JW, Egrie JC, Downing MR, Browne JK, Adamson JW: Correction of anemia of end-stage renal disease with recombinant human erythropoietin. N Eng! J Med 316: 73-78, 1987 20. Lowrie EG, Lew NL: Death risk in hemodialysis patients: The predictive value of commonly measured variables and an evaluation of death rate differences between facilities. Am J Kidney Dis 15:458-482,1990 21. Delmez JA, Slatopolsky E: Recent advances in the pathogenesis and therapy of uremic secondary hyperparathyroidism. J Clin Endocrinol Metab 72:735-739, 1991 22. Alfrey AC, leGendre GR, Kaehny WD: Dialysis encephalopathy syndrome: Possible aluminum intoxication. N Eng! J Med 294:184-188, 1976