- Email: [email protected]
Factors determining the rate of referral, transplantation, and survival on dialysis in women with ESRD
Factors Determining the Rate of Referral, Transplantation, and Survival on Dialysis in Women With ESRD Jerry McCauley, MD, William Irish, MSc, Leonard Thompson, BS, Judy Stevenson, Rhonda Lockett, Roy Bussard, BS, and Margaret Washington, MSW, MPH
MSN,
0 The determinants of referral for transplantation in women have not been well-studied. Similarly, factors determining survival on dialysis and the rate of transplantation in women remain controversial. Women have been reported to have lower rates of transplantation than men, and black women have the lowest rates of all groups. We questioned whether black women were referred at lower rates than whites and if race and other socioeconomic factors predicted referral, rate of renal transplantation, and patient survival on dialysis. All women in Allegheny or Philadelphia counties in Pennsylvania initiating dialysis between January 1, 1990, and December 31, 1992, were eligible for this study. Information was requested by questionnaire from each dialysis unit in these areas. Of the 383 eligible patients, completed questionnaires were obtained for 276 (72%). Ninety-three (54.7%) of the black patients and 57 (53.8%) of the white patients were referred for transplantation (P = 0.8). Declining the transplant option was the most common reason for nonreferral in both races. Patients with high school or greater education were approximately twice as likely to be referred than those with grade school educations (odds ratio [OR], 2.2; P = 0.04). Patients with coexisting illness were 67% (OR, 0.33; P = 0.004) less likely to be referred compared with patients with no other illnesses. Each additional year of age reduced the chances of being referred by 6% (OR, 0.94; P = 0.0001). Homemakers were 83% (P = 0.0008) and others 55% (P = 0.07) less likely to be referred compared with employed patients. Patients in Philadelphia County were 56% less likely to be referred compared with those in Allegheny County (P = 0.024). Race was not significantly associated with referral. Predictors of transplantation included age (RR, 0.96; confidence interval [Cl], 0.93 to 0.99; P = 0.13), white race (RR, 2.2; Cl, 1.3 to 4.0; P = O.OOSS), presence of other illnesses (RR, 0.37; Cl, 0.21 to 0.65; P = 0.0006), and employment status. White homemakers were 86% (RR, 0.14; Cl, 0.03 to 0.6; P = 0.0082) less likely than those with other employment situations to receive a transplant. Factors predicting patient survival on dialysis included race, educational status, and presence of comorbid illnesses. White patients were approximately four times (RR, 3.7; Cl, 1.7 to 8.1; P = 0.002) more likely to die than black patients. Patients with high school or greater education were 56% (RR, 0.44; Cl, 0.2 to 0.92; P = 0.008) less likely to die than those with grade school education alone. Patients with at least one coexisting illness were approximately 1.7 times (RR, 1.68; Cl, 1.1 to 2.4; P = 0.001) more likely to die than those without other illnesses. In summary, race was not a factor in referral for transplantation, but was predictive of transplantation and patient survival on dialysis. Socioeconomic factors such as educational status, age, and employment status were highly predictive of transplantation and long-term survival on hemodialysis. White homemakers unexpectedly received transplants less than any other group of dialysis patients. Further study is needed to determine why these potential transplant patients have declined or deferred transplantation. 0 1997 by the National Kidney Foundation, Inc. INDEX alysis;
WORDS: Referral race; gender.
for renal
transplantation;
renal
transplantation;
I
T IS ACCEPTED that access to renal transplantation should be free of race, gender, or socioeconomic bias. In practice, however, allocation of renal allografts is not uniform. Black patients often are less likely to be enrolled on transplant center waiting lists and, if accepted, wait approximately twice as long for cadaveric renal transplants compared with white patients.lX2 Women receive transplants less often than males regardless of race.3 Race and gender may combine to result in black females having the lowest rate of transplantation of all other groups. The rate of renal transplantation in black females aged 45 to 64 years has been reported to be 50% that of black males and 40% that of white females.3 The reasons for the low rate of transplantation have not been clearly elucidated. A lower rate of referral for transplant evaluation has American Journal of Kidney Diseases, Vol 30, No 6 (December),
patient
survival;
patient
selection;
hemodi-
been suggested as a potential cause. Interestingly, blacks have been reported to have superior survival on dialysis. Some have suggested that healthy black dialysis patients who are good transplant candidates but who are not referred
From End-Stage Renal Disease Network #4; the Department of Medicine and the Transplantation Medicine Division, University of Pittsburgh School of Medicine, Pittsburgh, PA; and Thomas E. Starzl Transplantation Institute, Pittsburgh, PA. Received February 8, 1997; accepted in revised form June 27, 1997. Address reprint requests to Jerry McCauley, MD, Transplantation Medicine Division, University of Pittsburgh School of Medicine, 3601 Fifth Ave, 4th Floor Falk Clinic, Pittsburgh, PA 15213. E-mail: [email protected] 0 1997 by the National Kidney Foundation, Inc. 0272~6386/97/3006-0003$3.00/O 1997:
pp 739-748
739
740
MCCAULEY
for transplantation might be a partial explanation for superior survival on dialysis. This implies a potential bias in selection of black patients for referral. Few studies have examined the factors that might be important in the selection of patients for referral for renal transplantation. Most studies have concentrated on the period after acceptance for transplantation. The low rate of renal transplantation for black women may partially be due to a lower rate of referral for transplantation. The decision to refer patients for transplantation is complex, but must include an assessment of the patient’s compliance, socioeconomic factors that might complicate transplantation, and the patient’s overall medical status. Many of these decisions are purely subjective and therefore potentially open to selection bias. We report the results of a retrospective analysis of potential determinants of referral for transplantation in women and the factors that might be associated with transplantation in this cohort. PATIENTS
AND METHODS
All female patients between 18 and 55 years of age who resided in Allegheny or Philadelphia counties in Pennsylvania who had initiated dialysis between January 1, 1990, and December 31, 1992, were eligible for this study. These counties were chosen because they were the areas in which the largest population of black patients on dialysis were located in EndStage Renal Disease (ESRD) Network #4. Patients also must have been treated on dialysis for at least 91 days to qualify for this study. Based on these criteria, 383 patients were drawn from the master database file of Network #4; 246 were black and 137 were white. Of these, 134 resided in Allegheny County and 249 lived in Philadelphia County. A survey that contained 25 questions was developed by ESRD Network #4 and was sent to each patient’s dialysis unit for completion. Information regarding the patient’s demographic status, details of medical illnesses, and recent hospitalizations were solicited, in addition to other factors. The centers were asked whether the patient was referred for transplantation; if not referred, the reason was requested. If a patient was said to have received a transplant, this was verified by review of the HF-2745 (Transplant Information Form) in the ESRD Network #4 files. Death of a patient was verified by review of the Health Care Financial Administration Death Notification Form. Comorbid illnesses were defined as conditions that were not directly related to ESRD and would potentially be a factor in acceptance of patients for transplantation. Conditions such as anemia or hyperparathyroidism were not considered to be comorbid illnesses. Complications of diabetes such as gastropathy or neuropathy were considered to be comorbid conditions; foot ulcers were not considered to be a significant illness. Hypertension was not listed as a comorbid illness
ET AL
unless it was severe or malignant. Referral for transplantation was defined as the center arranging means of contact with the transplant center by letter or other methods.
Statistical Analysis Continuous variables are presented as the mean ? SD and categorical variables are presented as proportions. The standard two-sample t-test was used to test differences between means, while differences in proportions were tested using Pearson’s chi-square test or Fisher’s exact test if expected frequencies were less than 5. The Wilcoxon rank-sum test, a nonparametric equivalent to the standard two-sample t-test, was used for highly skewed data. Linear logistic regression was used to identify factors associated with the probability of referral for renal transplantation.4 The odds ratio (OR) and 95% confidence intervals (CIs) were calculated as measures of strength of association. An OR greater than 1 suggests an increased risk of referral, whereas an OR less than 1 suggests a decreased risk. Stepwise multivariate linear logistic regression (forward inclusion method) was used to assess simultaneously the effect of each factor and to identify those factors independently associated with referral. The statistical significance levels required for inclusion and exclusion at each stepwise run were set at 0.05 and 0.10, respectively. The cumulative risk of transplantation was calculated using the Kaplan-Meier (product-limit) method. Risk estimates were calculated as 1 - s(t), where s(t) = the cumulative probability of being transplant free at time t. Risk estimates for different groups were compared by the log-rank (MantelCox) test.5.6 Patients who died on dialysis or were alive free of transplantation were censored. Cox’s proportional hazards model was used to compute the relative risk (RR) of transplantation and the 95% CIs. ‘,’ A stepwise multivariate Cox regression analysis (forward inclusion method) was used to assess simultaneously the effect of each factor and to identify those factors independently associated with transplantation. The statistical significance levels required for inclusion and exclusion at each stepwise run were set at 0.05 and 0.10, respectively. Patients not referred for transplantation were excluded from the analysis. All tests were two-tailed. P < 0.05 was considered statistically significant. All analyses were performed using SPSS for Windows software.’
RESULTS
Questionnaires were completed for 276 (72%) of the 383 patients eligible for this study. Of these, 126 (45.7%) were from Allegheny County and 150 (54.3%) from Philadelphia County. One hundred seventy (61.6%) were black and 106 (38.4%) were white. Analysis of patients by county revealed that Philadelphia County was predominantly black (121 patients [71.2%]) and Allegheny County was predominantly white (77 patients [61.1%]) (P < 0.001). The mean age of all patients was 43.6 2 9.7 years. There was no difference in age by race (black 44 t 9.6 years
TRANSPLANTATION Table
IN WOMEN
WITH
of Renal
Failure
1. Causes
by Race
Black Patients (%)
Renal Disease
Hypertension Glomerulonephritis Interstitial nephritis AIDS nephropathy Diabetes Polycystic kidney disease Lupus nephritis Sickle cell disease Hemolytic uremic syndrome Wegener’s disease Toxins Obstruction Congenital anomalies Alport’s disease Analgesic abuse Unknown Unspecified Total
58 31 4 2 48 4 12 2
(34.1) (18.2) (2.4) (1.2) (28.2) (2.4) (7.1) (1.2)
2 (1.2)
1 1 2 3
(0.6) (0.6) (1.2) (1.8)
170 (100.0)
Abbreviation: drome.
AIDS,
acquired
741
ESRD
immunodeficiency
White Patients (%)
12 14 4 1 53 2 5 3 1 1 1 1
(11.3) (13.2) (3.8) (1.0) (50.0) (1.9) (4.7) (2.8) (0.9) (0.9) (0.9) (0.9)
5 (4.7) 3 (2.8) 106 (100.0) syn-
v white 43.1 f 9.9 years; P = 0.5). The causes of renal failure are listed on Table 1. The leading cause for blacks was hypertension (58 patients [34.1%]), which was the third most common diagnosis for whites (12 patients [11.3%]; P = 0.00003). Diabetes was the most common diagnosis in whites (53 patients [50%]) and the second most common in blacks (48 patients [28.2%]; P = 0.00018). Glomerulonephritis accounted for similar proportions in blacks (3 1 patients [18.2%]) and whites (14 patients [13.2%]; P = 0.29). The other causes of renal failure occurred at low frequencies and were not significantly different by race. The cause of renal failure was not reported in six patients, and the etiology of renal failure was unknown in seven additional patients. Referral
for
Transplantation
There was no difference in the rate of referral for transplantation by race. Ninety-three (54.7%) black patients and 57 (53.8%) white females (P = 0.8) were referred. In Allegheny County, 76 (60.3%) patients were referred compared with 74 (49.3%) in Philadelphia County (P = 0.07). There also was no difference in the reasons patients were not referred Univariate
analysis.
based on race (Table 2). Refusal of the option for transplantation by the patients themselves was the most common reason for nonreferral. Thirty-two (49.2%) black patients and 22 (52.4%) white patients refused the transplant option (P = 0.8). Severe medical illnesses were present in seven (10.8%) black patients and 10 (23.8%) white patients (P = 0.07). Active drug abuse was the next most common reason patients were not referred and was only applied to eight (12.3%) black patients (P = 0.019). Three (4.6%) black patients and one (2.4%) white patient were not offered the option of transplantation for unspecified reasons. The proportion referred was significantly different based on educational status. Eleven of 16 (68.8%) college graduates, 25 of 38 (65.8%) patients with some college education, 73 of 115 (63.5%) high school graduates (63.5%), and 32 of 80 (40.7%) grade school graduates were referred for transplantation (P = 0.004). Analysis by educational status and race revealed that three of six (50%) black college graduates were referred compared with eight of 10 (80%) whites (P = 0.3). Seventeen of 25 (68%) black patients with some college education were referred compared with eight of 13 (61.5%) whites (P = 0.7). Forty-six of 67 (68.7%) black high school graduates were referred compared with 27 of 48 (56.3%) whites (P = 0.2). Twenty of 54 (37%) black grade school graduates were referred compared with 12 of 26 (46.2%) whites (P = 0.2). Educational status was not reported in 27 patients. Table
2. Reasons Patients Were for Transplantation
Reason
Black Patients (%)
Not Referred
Uncured malignancy Active infection HIV positive Severe disease Drug abuse Active psychosis Noncompliance Declined transplant option Not offered transplantation
3 1 2 7 8 2 7 32 3
Total
65 (100.0)
Abbreviation:
HIV,
Not
human
(4.6) (1.5) (3.1) (10.8) (12.3) (3.1) (10.8) (49.2) (4.6)
immunodeficiency
Referred
White Patients (%)
3 (7.1) 2 (4.8) IO (23.8)
4 (9.5) 22 (52.4) 1 (2.4) 42 (100.0) virus.
742
MCCAULEY Table
3. Most
Significant Comorbid by Race
Condition
Black Patients (%)
Congestive heart failure Coronary artery disease Cerebrovascular accident Cardiomyopathy Valvular heart disease Diabetic complications Seizures Depression Manic depressive Drug/ETOH abuse Obesity Hypetthyroidism Hypothyroidism Severe hypertension GI bleed Peripheral vascular disease Renal artery stenosis Inflammatory bowel disease Recurrent ileus Colon resection Intra-abdominal abscess Incarcerated inguinal hernia Systemic lupus erythematosus Cholelithiasis Pancreatitis Cancer Pericarditis/effusion Deaf/mute Liver failure Malnutrition AIDS/HIV positive Blind Sickle cell disease Steroid myopathy Septic arthritis Sepsis Pulmonary tuberculosis Bacterial meningitis Pulmonary disease Renal calculi Thrombocytopenia purpura Congenital anomalies
12 7 5 3 1 9 3 2 1 5 5 1 1 2 5 3 1
Total
95 (100.0)
Abbreviations: AIDS, acquired immunodeficiency
(12.6) (7.4) (5.3) (3.2) (1.1) (9.5) (3.2) (2.1) (1.1) (5.3) (5.3) (1.1) (1.1) (2.1) (5.3) (3.2) (1.1)
Disease
White Patients (%)
5 (7.6) 10 (15.2) 3 (4.5) 1 (1.5) 12 (18.2) 2 (3.0)
1 (1.5) 1 (1.5)
1 3 2 2 1
(1.5) (4.5) (3.0) (3.0) (1.5)
1 (1.1) 1 (1.5) 1 (1.1) 3 2 3 3
(3.2) (2.1) (3.2) (3.2)
1 2 3 3 1 1 1 1 1 1 1
(1.1) (2.1) (3.2) (3.2) (1.1) (1.1) (1.1) (1.1) (1.1) (1.1) (1.1)
1 (1.5) 2 (3.0) 1 (1.5) 4 1 2 1
4 (6.1)
2 1 1 1
ETOH, ethanol; immunodeficiency virus.
(6.1) (1.5) (3.0) (1.5)
(3.0) (1.5) (1.5) (1.5)
66 (100.0)
GI, gastrointestinal; syndrome; HIV, human
Referral for transplantation was also influenced by the number of comorbid illnesses. Table 3 illustrates the most significant of the comorbid illnesses for each patient by race. Cardiovascular
ET AL
diseases (coronary artery disease, CVA, cardiomyopathy, etc) were the most common conditions for blacks (28 patients [l&5%]) and whites (19 patients [17.9%]; P = 0.8. Nonrenal diabetic complications were the next most common conditions, and were approximately twice as common in whites compared with blacks (P = 0.10). The other conditions were numerous, but occurred at relatively low frequencies. Referred patients had significantly less comorbid illnesses (0.78 rt 0.95) than those who were not referred (1.19 i: 1.1; P = 0.001). Patients with no significant other illnesses, except renal disease, were referred most frequently. Of 114 patients with no comorbid illness, 75 (65.8%) were referred compared with 75 of 161 (46.5%) with one or more comorbid illness (P = 0.002). When comorbid illness was analyzed by race there was no significant difference in the number of illnesses among blacks (0.96 + 1.1) compared with whites (0.98 t 0.97; P = 0.86). Fifty of 75 (66.7%) black and 25 of 39 (64.1%) white patients with no additional illnesses were referred (P = 0.8). Twenty-two of 50 (44%) black and 20 of 38 (52.6%) white patients with one comorbid illness were referred (P = 0.3) while 13 of 25 (52%) black and nine of 22 (40%) white patients with two comorbid illnesses were referred (P = 0.6). Eight of 17 (47.1%) black and three of five (60%) white patients with three comorbid illnesses were referred (P = 0.9). None of the patients with more than four comorbid illness were referred. Two black and two white patients had four illnesses and one black patient had five illnesses. Of the 276 study patients, 202 (73.2%) were hospitalized within 1 year of the study. No statistically significant difference in referral was observed between those patients who were hospitalized and those who were not (52.5% v 59.5%, respectively; P = 0.30). Referral for transplantation was not associated with marital status (P = 0.62) or living situation (live alone, with family, or with nonrelative); however, current employment outside of the home was associated with referral. Patients employed full-time (42 of 58 patients [72%]) or part-time (11 of 15 patients [73.3%]) were more likely to be referred than disabled patients (22 of 37 [59.5%]), unemployed patients (31 of 64 [48.4%]), and homemakers (18 of 47 [38.3%]; P = 0.021). Patients retired due to age (two of three
TRANSPLANTATION
IN WOMEN
WITH
ESRD
[66.7%]) or illness (six of nine [66.7%]) and those on medical leave (two of three [66.7%]) were referred at equivalent rates (P = 1.0). Employment status was not associated with referral for blacks in contrast to whites (P = 0.49). Twelve of 23 (52.2%) black homemakers, 22 of 45 (48.9%) unemployed patients, and 13 of 20 (65%) disabled patients were referred. Employment status was associated with referral for white women: homemakers (six of 24 [25%]), employed full-time patients (18 of 24 [75%]), employed part-time patients (10 of 11 [90.9%]), and unemployed patients (nine of 19 [47.4%]; P = 0.001). White homemakers were more likely to decline the transplant option (10 of 24 [41%]) compared with employed patients (three of 35 [8.6%]; P = 0.0026) and others (seven of 38 [18.4%]; P = 0.046). Six of 23 (26.1%) black homemakers declined the transplant option compared with seven of 38 (18.4%; P = 0.48) employed patients and 12 of 78 (15.4%; P = 0.24) others. White homemakers were older (49 5 7 years v 46 -t 11 years; P = 0.03) and had more comorbid illnesses (0.92 ? 0.88 illnesses v 0.35 ? 0.57 illnesses; P = 0.045) than black homemakers. M&variate analysis. Multivariate analysis of factors associated with referral for transplantation was conducted by using linear logistic regression. Factors initially considered for the model included age, race, cause of renal failure, presence of comorbid diseases, dialysis method, employment status, educational status, whether hospitalization was needed within 1 year of this study, and the county of residence. The independent factors associated with referral for transplantation were educational status, absence of comorbid illness, age, employment status, and county of residence (Table 4). Patients with high school or greater education were approximately twice as likely to be referred than those with grade school educations (OR, 2.2; CI, 1.2 to 4.7; P = 0.04). Patients with comorbid illness were 67% (OR, 0.33; CI, 0.15 to 0.69; P = 0.004) less likely to be referred compared with patients with no other illnesses. Each additional year of age reduced the chances of being referred by 6% (OR, 0.94; CI, 0.9 to 0.97; P = 0.0001). Homemakers were 83% (OR, 0.17; CI, 0.06 to 0.48; P = 0.0008) and others 55% (OR, 0.45; CI, 0.19 to 1.06; P = 0.07) less likely to be referred com-
743 Table
4. Factors Associated for Transplantation
Variable
Education Grade school High school Comorbid illness No comorbid illness Comorbid illness Age Employment status Employed Homemakers Others County Allegheny Philadelphia
Odds
Ratio (95%
2.2 (1.03
0.33 0.94
With
Referral
Cl)
1 .oo to 4.7)
1 .oo (0.15 to 0.69) (0.9 to 0.97)
Significance
0.04
0.004 0.0001
0.168 0.447
1 .oo (0.06 to 0.48) (0.19 to 1.06)
0.0008 0.07
0.436
1 .oo (0.21 to 0.90)
0.024
pared with employed patients. Patients in Philadelphia County were 56% less likely to be referred compared with those in Allegheny County (OR, 0.44; CI, 0.21 to 0.90; P = 0.024). Race was not significantly associated with referral. Transplantation Univariate analysis. Sixty-two (22.5%) patients received a transplant during the study period (median time to transplantation, 1.87 + 0.09 years; CI, 1.69 to 2.06 years). Thirty-one (18.2%) of the 170 black patients received a transplant compared with 31 (29.2%) of the 106 white patients (P = 0.03). Living-related transplantations were performed in four (4.3%) white women and four (7%) black women (P = 0.47) referred, or 12.9% of the women who received a transplant in both groups. The total time on dialysis before transplantation (time to transplantation) was greater in referred black patients (median, 2.19 + 0.3 years) compared with referred white patients (median, 1.37 -t 0.24 years; P = 0.0014). Figure 1 illustrates the probability of transplantation for black and white patients. The time to transplantation was different based on educational status. High school graduation was the highest educational level attained in 116 (45.8%) patients: 68 (43.6%) blacks and 48 (49.5%) whites were in this group. Grade school education was the next most frequent educational status, 81 (32%) patients: 55 (35.3%) blacks and
744
MCCAULEY
Black I
p=o.o014
0.0 t I
I 0.0
I 0.5
I 1.0
f 1.5
I 2.0
I 2.5
I 3.0
*
Time (years) Fig 1. race.
Cumulative
probability
of transplantation
by
26 (26.8%) whites. Kaplan-Meier rate of transplantation was significantly lower for patients with grade school education compared with the college or greater-educated group (P = 0.03; Fig 2). Analysis of transplantation rate by race and educational status revealed that two of six (33.3%) black compared with seven of 10 (70%) white college graduates received a transplant (P = 0.302). Six of 25 (24%) black and four of 13 (30.8%) white patients with some college received a transplant. Sixteen of 67 (23.9%) black and 16 of 48 (33.3%) white high school graduates received a transplant. Four of 54 (7.4%) black and three of 26 (11.5%) white grade school graduates received a transplant. The total time on dialysis before transplantation was also greater for referred patients with comorbid medical illnesses. The median time to transplantation for patients without a comorbid illness was 1.62 + 0.28 years. For patients with at least one comorbid illness, median time to transplantation was 2.59 -+ 0.24 years (P = 0.01). Black patients without comorbid illnesses received a transplant at a median time of 1.92 ? 0.19 years compared with a median time of 0.99 + 0.14 years in white patients (P = 0.003). Multivariate analysis. A multivariate analysis (Cox regression model) was conducted to determine the factors independently associated with renal transplantation. Factors entered into the model included age, race, educational status, employment status, county of residence, hospitalization, presence of comorbid disease, cause of renal failure, and dialysis modality. The factors found
ET AL
to be predictive of transplantation included age, race, employment status, and presence of other illnesses. For each additional year of age, patients were 4% less likely to receive a transplant (RR, 0.96; CI, 0.93 to 0.99; P = 0.13). White patients were approximately twice as likely to received a transplant than black patients (RR, 2.2; CT, 1.3 to 4.0; P = 0.0053). Patients with comorbid illnesses were 63% less likely to receive a transplant (RR, 0.37; CI, 0.21 to 0.65; P = 0.0006) compared with those without other illnesses. Homemakers were 86% (RR, 0.14; CI, 0.03 to 0.6; P = 0.0082) and those with other employment situations were 32% (RR, 0.67; CI, 0.38 to 1.2; P = 0.18) less likely to receive a transplant than part-time or full-time employed patients. The Kaplan-Meier probabilities for transplantation according to employment status are illustrated in Fig 3. Patient Survival Univariate analysis. Forty-five (16.3%) of the 276 patients died during the study period. Twenty (11.8%) of the black and 25 (23.6%) of the white patients died (P = 0.01). The patients who did not survive had a greater number of comorbid illnesses (1.56 + 1.1 illnesses v 0.85 t 0.98 illnesses; P = 0.001) and were slightly, but not significantly, older (46 + 9 years v 43.2 + 9.8 years; P = 0.06). The Kaplan-Meier probabilities for patient survival are depicted in Fig 4 for all patients. Patient survival was significantly different for the transplant (mean, 2.9 Ifr 0.04 years) and nontransplant (2.4 t 0.07 years)
0.8 p=o.o04 0.6 College I
I
: SameCollage ,___ ___.._.._ I ,: -HighSchool
0.0
1, 0.5
1, 1.0
1, 1.5
1, 2.0
1, 2.8
1 3.0
I
Time (years)
Fig 2. educational
Cumulative status.
probability
of transplantation
by
TRANSPLANTATION
zr n 2 ae .-9 ii -i E
IN WOMEN
WITH
745
ESRD
education were 56% (RR, 0.44; CI, 0.2 to 0.92) less likely to die than those with grade school education alone. Patients with at least one comorbid illness were 1.7 times (RR, 1.68; CI, 1.1 to 2.4) more likely to die than those without other illnesses. Age was not an independent predictor of survival. Similarly, the addition of whether patients received a transplant did not change the model: transplantation was not an independent predictor of patient survival.
0.6 -
0.4 -
0.2 -
5 0.0
t I
I 0.0
I 0.5
I, 1.0
b 1.5
II
2.0
1 2.5
'1,
3.0
DISCUSSION
Time (years) Fig 3. employment
Cumulative status.
probability
of transplantation
by
groups (P = 0.004). When only referred patients were considered, the transplant patients had superior survival (mean, 2.89 + 0.04 years) compared with the nontransplant patients (mean, 2.3 2 0.1 years; P = 0.018). Patient survival was significantly better for black patients (mean, 2.62 2 0.07 years) compared with white patients (mean, 2.27 t 0.12 years; P = 0.006). A similar advantage in patient survival by race could be demonstrated when only nontransplant (P = 0.01) or only referred (P = 0.01) patients were considered. There was no statistically significant difference by race in survival by the log-rank test for patients who were not referred (P = 0.12); however, when using the Breslow test, the difference in survival was significant. Inspection of the Kaplan-Meier curves suggests that the effect of race on survival may be time dependent. The differences in early and late survival were similar, with a large difference occurring between 6 months and 2.5 years. Multivariate analysis. A multivariate analysis (Cox regression model) of patient survival was conducted using the following potential variables for the model: age, race, education, county of residence, whether patients were hospitalized in the past year, number of comorbid illnesses, dialysis modality, employment status, and cause of renal failure. Race (P = O.OOlS), educational status (P = 0.0084), and number of comorbid illnesses (P = 0.011) were the factors found to be independent predictors of patient survival. White patients were approximately four times more likely to die than black patients (RR, 3.7; CI, 1.7 to 8.1). Patients with high school or greater
The present study demonstrated that referral for transplantation is strongly influenced by socioeconomic factors (educational and employment status), overall health (number of illnesses), and preferences of the patients. Race, however, was not an important factor in referral of women for renal transplantation, but was important in determining transplant waiting time and the likelihood of receiving a renal allograft. Race also was an independent predictor of survival on dialysis in both patients referred and not referred for transplantation. Although many studies have examined the rate of renal transplantation and allograft survival after transplantation, few, if any, have examined the determinants of referral for transplantation since the landmark studies of Simmons et al.” Because the decision to refer patients is more subjective than other steps in the transplantation process, it might be more sensitive to selection bias by the dialysis staff. Educational status was an important determinant of referral for trans-
0.9 '5F .2 5
57
0.8 -
'I Black
0.41
1 0.0
1 0.5
1 1.0
Time Fig 4. Kaplan-Meier rates race in patients on dialysis.
I"' 1.5
2.0
1 2.5
11 3.0
(years) of
patient
survival
by
746
plantation, waiting time before transplantation, and patient survival. This was true for black and white patients. Potential reasons for this, which were not addressed in this study, might include bias of the dialysis staff against selecting poorly educated patients for transplantation. Simmons et al described what they termed “middle-class bias” in the selection of dialysis candidates when dialysis was a scarce resource before 1973.” Middle-class patients, who usually are more educated and have more social and economic resources, could be unconsciously favored. The present study was not designed to investigate this possibility. Less-educated patients also might have more comorbid illnesses, which might preclude transplantation. There was not, however, a significant difference in presence or total number of comorbid illnesses by educational status. It is more likely that educational status may be a surrogate of other factors (income, compliance, fear of transplantation) that may explain this phenomenon. Paradoxically, patients with high school or greater educations were more likely to decline the transplant option than those with grade school educations alone. Less-educated patients and blacks, however, were more likely not be referred due to drug abuse or noncompliance. Employment status was an independent predictor of both referral and transplantation. We were surprised to note that white homemakers were less likely to receive a transplant than any other group. This was primarily because they removed themselves from the process by declining the option for transplantation. White homemakers also were older and had more comorbid illnesses than black homemakers. The reasons for this are not obvious from this study, but these patients may have been more highly sensitized due to prior pregnancies or transfusions. The presence of comorbid illnesses was associated with a lower likelihood of referral, transplantation, and patient survival. Patients with other nonrenal illnesses require more extensive evaluations before transplantation and have a greater risk of complications or death in the postoperative period. Since the major reason for nonreferral of patients for transplantation was the patient’s decision to decline the transplant option, such patients may have considered themselves to be too ill to risk transplantation.
MCCAULEY
ET AL
The type of comorbid illness may have an influence on patient survival in dialysis patients. Bloembergen et all2 found that the relative risk of death from myocardial infarction is greater in white than in black patients. Patient survival is inferior for diabetics, both for dialysis patients and after transplantation.13 In the present study, there was no difference in the frequencies of ischemic heart disease or diabetes according to race. The similarity in comorbid illnesses may have contributed to the similarity in referral by race, but does not explain the survival advantage of black patients. The results of previous studies have suggested that white patients on dialysis may be older and more ill than the black patients on dialysis, which could explain the difference in survival. We found no difference in the number of comorbid illnesses or age of white and black patients. Although the number of comorbid illnesses were similar in our study, white patients were more likely to have diabetes, which has been observed to increase the risk of death in dialysis patients.14 The present study also included only women between 18 and 55 years, which is a relatively young population. Bleyer et al” found no beneficial effect of age by race in younger patients; younger black patients actually had inferior survival. Equal risk of death was observed for both races at age 49.7 years. Because older patients were not included in this study, the effect of age on patient survival may have been attenuated. Many studies have reported inferior allograft survival in black patients, in contrast to the superior patient survival of blacks on dialysis. Some have speculated that black patients on dialysis were healthier, perhaps because they were not referred for transplantation as frequently, which would leave healthy potential transplant candidates on dialysis for long periods. Even if black patients were being referred for transplantation, those who were accepted for transplantation but waiting twice as long as whites might artificially yield superior patient survival for blacks remaining on dialysis. Bleyer et all5 attempted to mitigate this potentially confounding factor by inclusion of follow-up after transplantation, which might improve patient survival in whites. Despite this adjustment, patient survival on dialysis remained superior for blacks. We attempted to minimize any potential artifact by examining
TRANSPLANTATION
IN WOMEN
WITH
ESRD
patient survival for black patients who were referred and for those who were never referred. We also removed both white and black transplant patients from the analysis, but found patient survival to be superior in blacks regardless of the grouping. Diabetic black patients also experienced superior survival, although this was only evident from 6 months to 2.5 years. This study suggests that the survival advantage for blacks on dialysis exists for essentially all subgroups of patients. The hypothesis that black patients are younger, have less illness, and may have renal diseases that are associated with better survival has not been supported by our observations. Cacciarreli et alI6 demonstrated that longer time on dialysis may be associated with inferior allograft survival. At 5 years there was a 10% difference in allograft survival in patients who received a transplant less than 6 months after initiating dialysis compared with patients receiving more than 60 months of dialysis. Patients who received a transplant quickly were younger, had lower panel-reactive antibodies, and had less delayed graft function and rejection than those who waited longer. In the present study, black patients waited on dialysis approximately twice as long as whites, which might influence graft survival once they do receive a transplant. Our study suggests that the longer times on dialysis was not related to the preference of black patients to remain on dialysis because there was no difference in the frequency of refusal of the transplant option between races. Previous studies have suggested that black patients wait longer once they are placed on the transplant list, which may be related to the emphasis on matching of our current allocation system. The observation that white female homemakers were referred less than any other group was an unexpected observation. These patients were older and had more comorbid illness than other patients, which may partially explain this finding. Some of these patients may have elected to avoid referral for transplantation due to child care responsibilities or other reasons. Many white homemakers would probably have been accepted for transplantation and ultimately obtain allografts since their medical conditions were not prohibitive. These patients may be delaying referral rather than foregoing transplantation altogether. Once referred and accepted for trans-
747
plantation, white female homemakers also wait longer for transplantation. This group is probably more likely to be multiparous, and therefore may have higher panel-reactive antibodies. We found no other studies that examined the availability of transplantation to women based on their employment status. If this observation can be confirmed, white female homemakers may be the most underserved population in transplantation. Fortunately, this problem may be solved by a modification of patient attitudes toward transplantation or the provision of social services, which would allow these women to leave the home long enough to obtain kidney transplants. More information is needed to determine why these patients limited their access to transplantation. Because this study was conducted in two counties in Pennsylvania, it may not be representative of the nation as a whole. An expanded nationwide study is needed to confirm the findings of this report. In this study, the difference in transplantation rate between black and white females could not be traced to differences in selection for referral. Black patients were as likely to be referred for transplantation as whites. However, once referred and accepted for transplantation, they waited twice as long. From the results of this study, it appears that the most underserved population of women were white female homemakers. Although there were statistically significant differences in factors that might make referral and ultimate transplantation less likely, patient preference was a major contributor to the low transplantation rate in this group. In summary, there was no apparent influence of race in the decision to refer patients for transplantation. Factors that indicated the degree of illness and educational status were more important. Educational status, comorbid illness, and age were important in determining whether patients actually received a transplant. Race was an important factor. Patient survival for blacks was superior to whites in all subgroups, even diabetics and patients who were never referred. Superior patient survival could not be explained by the retention of healthy black patients on dialysis and thereby artificially yielding better results. The determinants of improved patient survival have yet to be identified. Further study of the referral process may provide part of the explanation for the disparity of transplantation rates in black females.
748
MCCAULEY
REFERENCES 1. Scoucie JM, Neylan JF, McClellan W: Race and sex differences in the identification of candidates for renal transplantation. Am J Kidney Dis 19:414-419, 1992 2. Held PJ, Pauley MV, Bovbjerg RR, Newmann J, Salvatierra 0 Jr: Access to kidney transplantation: Has the United States eliminated income and racial differences? Arch Intern Med 148:2594-2600, 1988 3. US Renal Data System: USRDS 1995 Annual Data Report. Bethesda, MD, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 1995, pp 91-108 4. Hosmer DW, Lemeshow S: Applied Logistic Regression. New York, NY, Wiley, 1989 5. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 6. Mantel N: Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 50:163-170, 1966 7. Cox DR: Regression models and life tables (with discussion). J R Stat Sot Series B 34:187-220, 1972 8. Cox DR: Partial likelihood. Biometrika 62:269-276, 1975 9. Norusis M: SPSS for Windows Advanced Statistics Release 6.0. Chicago IL, SPSS Inc, 1993 10. Simmons RG, Marine SK, Simmons RL: Gift of Life:
ET AL
The Effect of Organ Transplantation on Individual, Family, and Societal Dynamics. New Brunswick, NJ, Transaction Books, 1987 Il. Simmons RG, Marine SK, Simmons RL: Transplantation and changing norms: Cultural lag and ethical ambiguities, in Gift of Life: The Effect of Organ Transplantation on Individual, Family, and Societal Dynamics. New Brunswick, NJ, Transaction Books, 1987, pp 12-14 12. Bloembergen WE, Port FK, Mauger EA, Wolfe RA: Causes of death in dialysis patients: Racial and gender differences. J Am Sot Nephrol 5:1231-1242, 1994 13. US Renal Data System: USRDS 1995 Annual Data Report. Bethesda, MD, The National Institute of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 1995, pp 59-77 14. US Renal Data System: USRDS 1995 Annual Data Report. The National Institute of Health. National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD. Am J Kidney Dis 26:S69-S84, 1995 15. Bleyer AJ, Tell GS, Evans GW, Ettinger WH, Burkart JM: Survival of patients undergoing renal replacement therapy in one center with special emphasis on racial differences. Am J Kidney Dis 28:72-81, 1996 16. Cacciarelli TV, Sumrani N, Dibenedetto A, Hong JH, Sommer BG: Influence of length of time on dialysis before transplantation on long term renal allograft outcome. Transplant Proc 25:2474-2476, 1993
MSN,
0 The determinants of referral for transplantation in women have not been well-studied. Similarly, factors determining survival on dialysis and the rate of transplantation in women remain controversial. Women have been reported to have lower rates of transplantation than men, and black women have the lowest rates of all groups. We questioned whether black women were referred at lower rates than whites and if race and other socioeconomic factors predicted referral, rate of renal transplantation, and patient survival on dialysis. All women in Allegheny or Philadelphia counties in Pennsylvania initiating dialysis between January 1, 1990, and December 31, 1992, were eligible for this study. Information was requested by questionnaire from each dialysis unit in these areas. Of the 383 eligible patients, completed questionnaires were obtained for 276 (72%). Ninety-three (54.7%) of the black patients and 57 (53.8%) of the white patients were referred for transplantation (P = 0.8). Declining the transplant option was the most common reason for nonreferral in both races. Patients with high school or greater education were approximately twice as likely to be referred than those with grade school educations (odds ratio [OR], 2.2; P = 0.04). Patients with coexisting illness were 67% (OR, 0.33; P = 0.004) less likely to be referred compared with patients with no other illnesses. Each additional year of age reduced the chances of being referred by 6% (OR, 0.94; P = 0.0001). Homemakers were 83% (P = 0.0008) and others 55% (P = 0.07) less likely to be referred compared with employed patients. Patients in Philadelphia County were 56% less likely to be referred compared with those in Allegheny County (P = 0.024). Race was not significantly associated with referral. Predictors of transplantation included age (RR, 0.96; confidence interval [Cl], 0.93 to 0.99; P = 0.13), white race (RR, 2.2; Cl, 1.3 to 4.0; P = O.OOSS), presence of other illnesses (RR, 0.37; Cl, 0.21 to 0.65; P = 0.0006), and employment status. White homemakers were 86% (RR, 0.14; Cl, 0.03 to 0.6; P = 0.0082) less likely than those with other employment situations to receive a transplant. Factors predicting patient survival on dialysis included race, educational status, and presence of comorbid illnesses. White patients were approximately four times (RR, 3.7; Cl, 1.7 to 8.1; P = 0.002) more likely to die than black patients. Patients with high school or greater education were 56% (RR, 0.44; Cl, 0.2 to 0.92; P = 0.008) less likely to die than those with grade school education alone. Patients with at least one coexisting illness were approximately 1.7 times (RR, 1.68; Cl, 1.1 to 2.4; P = 0.001) more likely to die than those without other illnesses. In summary, race was not a factor in referral for transplantation, but was predictive of transplantation and patient survival on dialysis. Socioeconomic factors such as educational status, age, and employment status were highly predictive of transplantation and long-term survival on hemodialysis. White homemakers unexpectedly received transplants less than any other group of dialysis patients. Further study is needed to determine why these potential transplant patients have declined or deferred transplantation. 0 1997 by the National Kidney Foundation, Inc. INDEX alysis;
WORDS: Referral race; gender.
for renal
transplantation;
renal
transplantation;
I
T IS ACCEPTED that access to renal transplantation should be free of race, gender, or socioeconomic bias. In practice, however, allocation of renal allografts is not uniform. Black patients often are less likely to be enrolled on transplant center waiting lists and, if accepted, wait approximately twice as long for cadaveric renal transplants compared with white patients.lX2 Women receive transplants less often than males regardless of race.3 Race and gender may combine to result in black females having the lowest rate of transplantation of all other groups. The rate of renal transplantation in black females aged 45 to 64 years has been reported to be 50% that of black males and 40% that of white females.3 The reasons for the low rate of transplantation have not been clearly elucidated. A lower rate of referral for transplant evaluation has American Journal of Kidney Diseases, Vol 30, No 6 (December),
patient
survival;
patient
selection;
hemodi-
been suggested as a potential cause. Interestingly, blacks have been reported to have superior survival on dialysis. Some have suggested that healthy black dialysis patients who are good transplant candidates but who are not referred
From End-Stage Renal Disease Network #4; the Department of Medicine and the Transplantation Medicine Division, University of Pittsburgh School of Medicine, Pittsburgh, PA; and Thomas E. Starzl Transplantation Institute, Pittsburgh, PA. Received February 8, 1997; accepted in revised form June 27, 1997. Address reprint requests to Jerry McCauley, MD, Transplantation Medicine Division, University of Pittsburgh School of Medicine, 3601 Fifth Ave, 4th Floor Falk Clinic, Pittsburgh, PA 15213. E-mail: [email protected] 0 1997 by the National Kidney Foundation, Inc. 0272~6386/97/3006-0003$3.00/O 1997:
pp 739-748
739
740
MCCAULEY
for transplantation might be a partial explanation for superior survival on dialysis. This implies a potential bias in selection of black patients for referral. Few studies have examined the factors that might be important in the selection of patients for referral for renal transplantation. Most studies have concentrated on the period after acceptance for transplantation. The low rate of renal transplantation for black women may partially be due to a lower rate of referral for transplantation. The decision to refer patients for transplantation is complex, but must include an assessment of the patient’s compliance, socioeconomic factors that might complicate transplantation, and the patient’s overall medical status. Many of these decisions are purely subjective and therefore potentially open to selection bias. We report the results of a retrospective analysis of potential determinants of referral for transplantation in women and the factors that might be associated with transplantation in this cohort. PATIENTS
AND METHODS
All female patients between 18 and 55 years of age who resided in Allegheny or Philadelphia counties in Pennsylvania who had initiated dialysis between January 1, 1990, and December 31, 1992, were eligible for this study. These counties were chosen because they were the areas in which the largest population of black patients on dialysis were located in EndStage Renal Disease (ESRD) Network #4. Patients also must have been treated on dialysis for at least 91 days to qualify for this study. Based on these criteria, 383 patients were drawn from the master database file of Network #4; 246 were black and 137 were white. Of these, 134 resided in Allegheny County and 249 lived in Philadelphia County. A survey that contained 25 questions was developed by ESRD Network #4 and was sent to each patient’s dialysis unit for completion. Information regarding the patient’s demographic status, details of medical illnesses, and recent hospitalizations were solicited, in addition to other factors. The centers were asked whether the patient was referred for transplantation; if not referred, the reason was requested. If a patient was said to have received a transplant, this was verified by review of the HF-2745 (Transplant Information Form) in the ESRD Network #4 files. Death of a patient was verified by review of the Health Care Financial Administration Death Notification Form. Comorbid illnesses were defined as conditions that were not directly related to ESRD and would potentially be a factor in acceptance of patients for transplantation. Conditions such as anemia or hyperparathyroidism were not considered to be comorbid illnesses. Complications of diabetes such as gastropathy or neuropathy were considered to be comorbid conditions; foot ulcers were not considered to be a significant illness. Hypertension was not listed as a comorbid illness
ET AL
unless it was severe or malignant. Referral for transplantation was defined as the center arranging means of contact with the transplant center by letter or other methods.
Statistical Analysis Continuous variables are presented as the mean ? SD and categorical variables are presented as proportions. The standard two-sample t-test was used to test differences between means, while differences in proportions were tested using Pearson’s chi-square test or Fisher’s exact test if expected frequencies were less than 5. The Wilcoxon rank-sum test, a nonparametric equivalent to the standard two-sample t-test, was used for highly skewed data. Linear logistic regression was used to identify factors associated with the probability of referral for renal transplantation.4 The odds ratio (OR) and 95% confidence intervals (CIs) were calculated as measures of strength of association. An OR greater than 1 suggests an increased risk of referral, whereas an OR less than 1 suggests a decreased risk. Stepwise multivariate linear logistic regression (forward inclusion method) was used to assess simultaneously the effect of each factor and to identify those factors independently associated with referral. The statistical significance levels required for inclusion and exclusion at each stepwise run were set at 0.05 and 0.10, respectively. The cumulative risk of transplantation was calculated using the Kaplan-Meier (product-limit) method. Risk estimates were calculated as 1 - s(t), where s(t) = the cumulative probability of being transplant free at time t. Risk estimates for different groups were compared by the log-rank (MantelCox) test.5.6 Patients who died on dialysis or were alive free of transplantation were censored. Cox’s proportional hazards model was used to compute the relative risk (RR) of transplantation and the 95% CIs. ‘,’ A stepwise multivariate Cox regression analysis (forward inclusion method) was used to assess simultaneously the effect of each factor and to identify those factors independently associated with transplantation. The statistical significance levels required for inclusion and exclusion at each stepwise run were set at 0.05 and 0.10, respectively. Patients not referred for transplantation were excluded from the analysis. All tests were two-tailed. P < 0.05 was considered statistically significant. All analyses were performed using SPSS for Windows software.’
RESULTS
Questionnaires were completed for 276 (72%) of the 383 patients eligible for this study. Of these, 126 (45.7%) were from Allegheny County and 150 (54.3%) from Philadelphia County. One hundred seventy (61.6%) were black and 106 (38.4%) were white. Analysis of patients by county revealed that Philadelphia County was predominantly black (121 patients [71.2%]) and Allegheny County was predominantly white (77 patients [61.1%]) (P < 0.001). The mean age of all patients was 43.6 2 9.7 years. There was no difference in age by race (black 44 t 9.6 years
TRANSPLANTATION Table
IN WOMEN
WITH
of Renal
Failure
1. Causes
by Race
Black Patients (%)
Renal Disease
Hypertension Glomerulonephritis Interstitial nephritis AIDS nephropathy Diabetes Polycystic kidney disease Lupus nephritis Sickle cell disease Hemolytic uremic syndrome Wegener’s disease Toxins Obstruction Congenital anomalies Alport’s disease Analgesic abuse Unknown Unspecified Total
58 31 4 2 48 4 12 2
(34.1) (18.2) (2.4) (1.2) (28.2) (2.4) (7.1) (1.2)
2 (1.2)
1 1 2 3
(0.6) (0.6) (1.2) (1.8)
170 (100.0)
Abbreviation: drome.
AIDS,
acquired
741
ESRD
immunodeficiency
White Patients (%)
12 14 4 1 53 2 5 3 1 1 1 1
(11.3) (13.2) (3.8) (1.0) (50.0) (1.9) (4.7) (2.8) (0.9) (0.9) (0.9) (0.9)
5 (4.7) 3 (2.8) 106 (100.0) syn-
v white 43.1 f 9.9 years; P = 0.5). The causes of renal failure are listed on Table 1. The leading cause for blacks was hypertension (58 patients [34.1%]), which was the third most common diagnosis for whites (12 patients [11.3%]; P = 0.00003). Diabetes was the most common diagnosis in whites (53 patients [50%]) and the second most common in blacks (48 patients [28.2%]; P = 0.00018). Glomerulonephritis accounted for similar proportions in blacks (3 1 patients [18.2%]) and whites (14 patients [13.2%]; P = 0.29). The other causes of renal failure occurred at low frequencies and were not significantly different by race. The cause of renal failure was not reported in six patients, and the etiology of renal failure was unknown in seven additional patients. Referral
for
Transplantation
There was no difference in the rate of referral for transplantation by race. Ninety-three (54.7%) black patients and 57 (53.8%) white females (P = 0.8) were referred. In Allegheny County, 76 (60.3%) patients were referred compared with 74 (49.3%) in Philadelphia County (P = 0.07). There also was no difference in the reasons patients were not referred Univariate
analysis.
based on race (Table 2). Refusal of the option for transplantation by the patients themselves was the most common reason for nonreferral. Thirty-two (49.2%) black patients and 22 (52.4%) white patients refused the transplant option (P = 0.8). Severe medical illnesses were present in seven (10.8%) black patients and 10 (23.8%) white patients (P = 0.07). Active drug abuse was the next most common reason patients were not referred and was only applied to eight (12.3%) black patients (P = 0.019). Three (4.6%) black patients and one (2.4%) white patient were not offered the option of transplantation for unspecified reasons. The proportion referred was significantly different based on educational status. Eleven of 16 (68.8%) college graduates, 25 of 38 (65.8%) patients with some college education, 73 of 115 (63.5%) high school graduates (63.5%), and 32 of 80 (40.7%) grade school graduates were referred for transplantation (P = 0.004). Analysis by educational status and race revealed that three of six (50%) black college graduates were referred compared with eight of 10 (80%) whites (P = 0.3). Seventeen of 25 (68%) black patients with some college education were referred compared with eight of 13 (61.5%) whites (P = 0.7). Forty-six of 67 (68.7%) black high school graduates were referred compared with 27 of 48 (56.3%) whites (P = 0.2). Twenty of 54 (37%) black grade school graduates were referred compared with 12 of 26 (46.2%) whites (P = 0.2). Educational status was not reported in 27 patients. Table
2. Reasons Patients Were for Transplantation
Reason
Black Patients (%)
Not Referred
Uncured malignancy Active infection HIV positive Severe disease Drug abuse Active psychosis Noncompliance Declined transplant option Not offered transplantation
3 1 2 7 8 2 7 32 3
Total
65 (100.0)
Abbreviation:
HIV,
Not
human
(4.6) (1.5) (3.1) (10.8) (12.3) (3.1) (10.8) (49.2) (4.6)
immunodeficiency
Referred
White Patients (%)
3 (7.1) 2 (4.8) IO (23.8)
4 (9.5) 22 (52.4) 1 (2.4) 42 (100.0) virus.
742
MCCAULEY Table
3. Most
Significant Comorbid by Race
Condition
Black Patients (%)
Congestive heart failure Coronary artery disease Cerebrovascular accident Cardiomyopathy Valvular heart disease Diabetic complications Seizures Depression Manic depressive Drug/ETOH abuse Obesity Hypetthyroidism Hypothyroidism Severe hypertension GI bleed Peripheral vascular disease Renal artery stenosis Inflammatory bowel disease Recurrent ileus Colon resection Intra-abdominal abscess Incarcerated inguinal hernia Systemic lupus erythematosus Cholelithiasis Pancreatitis Cancer Pericarditis/effusion Deaf/mute Liver failure Malnutrition AIDS/HIV positive Blind Sickle cell disease Steroid myopathy Septic arthritis Sepsis Pulmonary tuberculosis Bacterial meningitis Pulmonary disease Renal calculi Thrombocytopenia purpura Congenital anomalies
12 7 5 3 1 9 3 2 1 5 5 1 1 2 5 3 1
Total
95 (100.0)
Abbreviations: AIDS, acquired immunodeficiency
(12.6) (7.4) (5.3) (3.2) (1.1) (9.5) (3.2) (2.1) (1.1) (5.3) (5.3) (1.1) (1.1) (2.1) (5.3) (3.2) (1.1)
Disease
White Patients (%)
5 (7.6) 10 (15.2) 3 (4.5) 1 (1.5) 12 (18.2) 2 (3.0)
1 (1.5) 1 (1.5)
1 3 2 2 1
(1.5) (4.5) (3.0) (3.0) (1.5)
1 (1.1) 1 (1.5) 1 (1.1) 3 2 3 3
(3.2) (2.1) (3.2) (3.2)
1 2 3 3 1 1 1 1 1 1 1
(1.1) (2.1) (3.2) (3.2) (1.1) (1.1) (1.1) (1.1) (1.1) (1.1) (1.1)
1 (1.5) 2 (3.0) 1 (1.5) 4 1 2 1
4 (6.1)
2 1 1 1
ETOH, ethanol; immunodeficiency virus.
(6.1) (1.5) (3.0) (1.5)
(3.0) (1.5) (1.5) (1.5)
66 (100.0)
GI, gastrointestinal; syndrome; HIV, human
Referral for transplantation was also influenced by the number of comorbid illnesses. Table 3 illustrates the most significant of the comorbid illnesses for each patient by race. Cardiovascular
ET AL
diseases (coronary artery disease, CVA, cardiomyopathy, etc) were the most common conditions for blacks (28 patients [l&5%]) and whites (19 patients [17.9%]; P = 0.8. Nonrenal diabetic complications were the next most common conditions, and were approximately twice as common in whites compared with blacks (P = 0.10). The other conditions were numerous, but occurred at relatively low frequencies. Referred patients had significantly less comorbid illnesses (0.78 rt 0.95) than those who were not referred (1.19 i: 1.1; P = 0.001). Patients with no significant other illnesses, except renal disease, were referred most frequently. Of 114 patients with no comorbid illness, 75 (65.8%) were referred compared with 75 of 161 (46.5%) with one or more comorbid illness (P = 0.002). When comorbid illness was analyzed by race there was no significant difference in the number of illnesses among blacks (0.96 + 1.1) compared with whites (0.98 t 0.97; P = 0.86). Fifty of 75 (66.7%) black and 25 of 39 (64.1%) white patients with no additional illnesses were referred (P = 0.8). Twenty-two of 50 (44%) black and 20 of 38 (52.6%) white patients with one comorbid illness were referred (P = 0.3) while 13 of 25 (52%) black and nine of 22 (40%) white patients with two comorbid illnesses were referred (P = 0.6). Eight of 17 (47.1%) black and three of five (60%) white patients with three comorbid illnesses were referred (P = 0.9). None of the patients with more than four comorbid illness were referred. Two black and two white patients had four illnesses and one black patient had five illnesses. Of the 276 study patients, 202 (73.2%) were hospitalized within 1 year of the study. No statistically significant difference in referral was observed between those patients who were hospitalized and those who were not (52.5% v 59.5%, respectively; P = 0.30). Referral for transplantation was not associated with marital status (P = 0.62) or living situation (live alone, with family, or with nonrelative); however, current employment outside of the home was associated with referral. Patients employed full-time (42 of 58 patients [72%]) or part-time (11 of 15 patients [73.3%]) were more likely to be referred than disabled patients (22 of 37 [59.5%]), unemployed patients (31 of 64 [48.4%]), and homemakers (18 of 47 [38.3%]; P = 0.021). Patients retired due to age (two of three
TRANSPLANTATION
IN WOMEN
WITH
ESRD
[66.7%]) or illness (six of nine [66.7%]) and those on medical leave (two of three [66.7%]) were referred at equivalent rates (P = 1.0). Employment status was not associated with referral for blacks in contrast to whites (P = 0.49). Twelve of 23 (52.2%) black homemakers, 22 of 45 (48.9%) unemployed patients, and 13 of 20 (65%) disabled patients were referred. Employment status was associated with referral for white women: homemakers (six of 24 [25%]), employed full-time patients (18 of 24 [75%]), employed part-time patients (10 of 11 [90.9%]), and unemployed patients (nine of 19 [47.4%]; P = 0.001). White homemakers were more likely to decline the transplant option (10 of 24 [41%]) compared with employed patients (three of 35 [8.6%]; P = 0.0026) and others (seven of 38 [18.4%]; P = 0.046). Six of 23 (26.1%) black homemakers declined the transplant option compared with seven of 38 (18.4%; P = 0.48) employed patients and 12 of 78 (15.4%; P = 0.24) others. White homemakers were older (49 5 7 years v 46 -t 11 years; P = 0.03) and had more comorbid illnesses (0.92 ? 0.88 illnesses v 0.35 ? 0.57 illnesses; P = 0.045) than black homemakers. M&variate analysis. Multivariate analysis of factors associated with referral for transplantation was conducted by using linear logistic regression. Factors initially considered for the model included age, race, cause of renal failure, presence of comorbid diseases, dialysis method, employment status, educational status, whether hospitalization was needed within 1 year of this study, and the county of residence. The independent factors associated with referral for transplantation were educational status, absence of comorbid illness, age, employment status, and county of residence (Table 4). Patients with high school or greater education were approximately twice as likely to be referred than those with grade school educations (OR, 2.2; CI, 1.2 to 4.7; P = 0.04). Patients with comorbid illness were 67% (OR, 0.33; CI, 0.15 to 0.69; P = 0.004) less likely to be referred compared with patients with no other illnesses. Each additional year of age reduced the chances of being referred by 6% (OR, 0.94; CI, 0.9 to 0.97; P = 0.0001). Homemakers were 83% (OR, 0.17; CI, 0.06 to 0.48; P = 0.0008) and others 55% (OR, 0.45; CI, 0.19 to 1.06; P = 0.07) less likely to be referred com-
743 Table
4. Factors Associated for Transplantation
Variable
Education Grade school High school Comorbid illness No comorbid illness Comorbid illness Age Employment status Employed Homemakers Others County Allegheny Philadelphia
Odds
Ratio (95%
2.2 (1.03
0.33 0.94
With
Referral
Cl)
1 .oo to 4.7)
1 .oo (0.15 to 0.69) (0.9 to 0.97)
Significance
0.04
0.004 0.0001
0.168 0.447
1 .oo (0.06 to 0.48) (0.19 to 1.06)
0.0008 0.07
0.436
1 .oo (0.21 to 0.90)
0.024
pared with employed patients. Patients in Philadelphia County were 56% less likely to be referred compared with those in Allegheny County (OR, 0.44; CI, 0.21 to 0.90; P = 0.024). Race was not significantly associated with referral. Transplantation Univariate analysis. Sixty-two (22.5%) patients received a transplant during the study period (median time to transplantation, 1.87 + 0.09 years; CI, 1.69 to 2.06 years). Thirty-one (18.2%) of the 170 black patients received a transplant compared with 31 (29.2%) of the 106 white patients (P = 0.03). Living-related transplantations were performed in four (4.3%) white women and four (7%) black women (P = 0.47) referred, or 12.9% of the women who received a transplant in both groups. The total time on dialysis before transplantation (time to transplantation) was greater in referred black patients (median, 2.19 + 0.3 years) compared with referred white patients (median, 1.37 -t 0.24 years; P = 0.0014). Figure 1 illustrates the probability of transplantation for black and white patients. The time to transplantation was different based on educational status. High school graduation was the highest educational level attained in 116 (45.8%) patients: 68 (43.6%) blacks and 48 (49.5%) whites were in this group. Grade school education was the next most frequent educational status, 81 (32%) patients: 55 (35.3%) blacks and
744
MCCAULEY
Black I
p=o.o014
0.0 t I
I 0.0
I 0.5
I 1.0
f 1.5
I 2.0
I 2.5
I 3.0
*
Time (years) Fig 1. race.
Cumulative
probability
of transplantation
by
26 (26.8%) whites. Kaplan-Meier rate of transplantation was significantly lower for patients with grade school education compared with the college or greater-educated group (P = 0.03; Fig 2). Analysis of transplantation rate by race and educational status revealed that two of six (33.3%) black compared with seven of 10 (70%) white college graduates received a transplant (P = 0.302). Six of 25 (24%) black and four of 13 (30.8%) white patients with some college received a transplant. Sixteen of 67 (23.9%) black and 16 of 48 (33.3%) white high school graduates received a transplant. Four of 54 (7.4%) black and three of 26 (11.5%) white grade school graduates received a transplant. The total time on dialysis before transplantation was also greater for referred patients with comorbid medical illnesses. The median time to transplantation for patients without a comorbid illness was 1.62 + 0.28 years. For patients with at least one comorbid illness, median time to transplantation was 2.59 -+ 0.24 years (P = 0.01). Black patients without comorbid illnesses received a transplant at a median time of 1.92 ? 0.19 years compared with a median time of 0.99 + 0.14 years in white patients (P = 0.003). Multivariate analysis. A multivariate analysis (Cox regression model) was conducted to determine the factors independently associated with renal transplantation. Factors entered into the model included age, race, educational status, employment status, county of residence, hospitalization, presence of comorbid disease, cause of renal failure, and dialysis modality. The factors found
ET AL
to be predictive of transplantation included age, race, employment status, and presence of other illnesses. For each additional year of age, patients were 4% less likely to receive a transplant (RR, 0.96; CI, 0.93 to 0.99; P = 0.13). White patients were approximately twice as likely to received a transplant than black patients (RR, 2.2; CT, 1.3 to 4.0; P = 0.0053). Patients with comorbid illnesses were 63% less likely to receive a transplant (RR, 0.37; CI, 0.21 to 0.65; P = 0.0006) compared with those without other illnesses. Homemakers were 86% (RR, 0.14; CI, 0.03 to 0.6; P = 0.0082) and those with other employment situations were 32% (RR, 0.67; CI, 0.38 to 1.2; P = 0.18) less likely to receive a transplant than part-time or full-time employed patients. The Kaplan-Meier probabilities for transplantation according to employment status are illustrated in Fig 3. Patient Survival Univariate analysis. Forty-five (16.3%) of the 276 patients died during the study period. Twenty (11.8%) of the black and 25 (23.6%) of the white patients died (P = 0.01). The patients who did not survive had a greater number of comorbid illnesses (1.56 + 1.1 illnesses v 0.85 t 0.98 illnesses; P = 0.001) and were slightly, but not significantly, older (46 + 9 years v 43.2 + 9.8 years; P = 0.06). The Kaplan-Meier probabilities for patient survival are depicted in Fig 4 for all patients. Patient survival was significantly different for the transplant (mean, 2.9 Ifr 0.04 years) and nontransplant (2.4 t 0.07 years)
0.8 p=o.o04 0.6 College I
I
: SameCollage ,___ ___.._.._ I ,: -HighSchool
0.0
1, 0.5
1, 1.0
1, 1.5
1, 2.0
1, 2.8
1 3.0
I
Time (years)
Fig 2. educational
Cumulative status.
probability
of transplantation
by
TRANSPLANTATION
zr n 2 ae .-9 ii -i E
IN WOMEN
WITH
745
ESRD
education were 56% (RR, 0.44; CI, 0.2 to 0.92) less likely to die than those with grade school education alone. Patients with at least one comorbid illness were 1.7 times (RR, 1.68; CI, 1.1 to 2.4) more likely to die than those without other illnesses. Age was not an independent predictor of survival. Similarly, the addition of whether patients received a transplant did not change the model: transplantation was not an independent predictor of patient survival.
0.6 -
0.4 -
0.2 -
5 0.0
t I
I 0.0
I 0.5
I, 1.0
b 1.5
II
2.0
1 2.5
'1,
3.0
DISCUSSION
Time (years) Fig 3. employment
Cumulative status.
probability
of transplantation
by
groups (P = 0.004). When only referred patients were considered, the transplant patients had superior survival (mean, 2.89 + 0.04 years) compared with the nontransplant patients (mean, 2.3 2 0.1 years; P = 0.018). Patient survival was significantly better for black patients (mean, 2.62 2 0.07 years) compared with white patients (mean, 2.27 t 0.12 years; P = 0.006). A similar advantage in patient survival by race could be demonstrated when only nontransplant (P = 0.01) or only referred (P = 0.01) patients were considered. There was no statistically significant difference by race in survival by the log-rank test for patients who were not referred (P = 0.12); however, when using the Breslow test, the difference in survival was significant. Inspection of the Kaplan-Meier curves suggests that the effect of race on survival may be time dependent. The differences in early and late survival were similar, with a large difference occurring between 6 months and 2.5 years. Multivariate analysis. A multivariate analysis (Cox regression model) of patient survival was conducted using the following potential variables for the model: age, race, education, county of residence, whether patients were hospitalized in the past year, number of comorbid illnesses, dialysis modality, employment status, and cause of renal failure. Race (P = O.OOlS), educational status (P = 0.0084), and number of comorbid illnesses (P = 0.011) were the factors found to be independent predictors of patient survival. White patients were approximately four times more likely to die than black patients (RR, 3.7; CI, 1.7 to 8.1). Patients with high school or greater
The present study demonstrated that referral for transplantation is strongly influenced by socioeconomic factors (educational and employment status), overall health (number of illnesses), and preferences of the patients. Race, however, was not an important factor in referral of women for renal transplantation, but was important in determining transplant waiting time and the likelihood of receiving a renal allograft. Race also was an independent predictor of survival on dialysis in both patients referred and not referred for transplantation. Although many studies have examined the rate of renal transplantation and allograft survival after transplantation, few, if any, have examined the determinants of referral for transplantation since the landmark studies of Simmons et al.” Because the decision to refer patients is more subjective than other steps in the transplantation process, it might be more sensitive to selection bias by the dialysis staff. Educational status was an important determinant of referral for trans-
0.9 '5F .2 5
57
0.8 -
'I Black
0.41
1 0.0
1 0.5
1 1.0
Time Fig 4. Kaplan-Meier rates race in patients on dialysis.
I"' 1.5
2.0
1 2.5
11 3.0
(years) of
patient
survival
by
746
plantation, waiting time before transplantation, and patient survival. This was true for black and white patients. Potential reasons for this, which were not addressed in this study, might include bias of the dialysis staff against selecting poorly educated patients for transplantation. Simmons et al described what they termed “middle-class bias” in the selection of dialysis candidates when dialysis was a scarce resource before 1973.” Middle-class patients, who usually are more educated and have more social and economic resources, could be unconsciously favored. The present study was not designed to investigate this possibility. Less-educated patients also might have more comorbid illnesses, which might preclude transplantation. There was not, however, a significant difference in presence or total number of comorbid illnesses by educational status. It is more likely that educational status may be a surrogate of other factors (income, compliance, fear of transplantation) that may explain this phenomenon. Paradoxically, patients with high school or greater educations were more likely to decline the transplant option than those with grade school educations alone. Less-educated patients and blacks, however, were more likely not be referred due to drug abuse or noncompliance. Employment status was an independent predictor of both referral and transplantation. We were surprised to note that white homemakers were less likely to receive a transplant than any other group. This was primarily because they removed themselves from the process by declining the option for transplantation. White homemakers also were older and had more comorbid illnesses than black homemakers. The reasons for this are not obvious from this study, but these patients may have been more highly sensitized due to prior pregnancies or transfusions. The presence of comorbid illnesses was associated with a lower likelihood of referral, transplantation, and patient survival. Patients with other nonrenal illnesses require more extensive evaluations before transplantation and have a greater risk of complications or death in the postoperative period. Since the major reason for nonreferral of patients for transplantation was the patient’s decision to decline the transplant option, such patients may have considered themselves to be too ill to risk transplantation.
MCCAULEY
ET AL
The type of comorbid illness may have an influence on patient survival in dialysis patients. Bloembergen et all2 found that the relative risk of death from myocardial infarction is greater in white than in black patients. Patient survival is inferior for diabetics, both for dialysis patients and after transplantation.13 In the present study, there was no difference in the frequencies of ischemic heart disease or diabetes according to race. The similarity in comorbid illnesses may have contributed to the similarity in referral by race, but does not explain the survival advantage of black patients. The results of previous studies have suggested that white patients on dialysis may be older and more ill than the black patients on dialysis, which could explain the difference in survival. We found no difference in the number of comorbid illnesses or age of white and black patients. Although the number of comorbid illnesses were similar in our study, white patients were more likely to have diabetes, which has been observed to increase the risk of death in dialysis patients.14 The present study also included only women between 18 and 55 years, which is a relatively young population. Bleyer et al” found no beneficial effect of age by race in younger patients; younger black patients actually had inferior survival. Equal risk of death was observed for both races at age 49.7 years. Because older patients were not included in this study, the effect of age on patient survival may have been attenuated. Many studies have reported inferior allograft survival in black patients, in contrast to the superior patient survival of blacks on dialysis. Some have speculated that black patients on dialysis were healthier, perhaps because they were not referred for transplantation as frequently, which would leave healthy potential transplant candidates on dialysis for long periods. Even if black patients were being referred for transplantation, those who were accepted for transplantation but waiting twice as long as whites might artificially yield superior patient survival for blacks remaining on dialysis. Bleyer et all5 attempted to mitigate this potentially confounding factor by inclusion of follow-up after transplantation, which might improve patient survival in whites. Despite this adjustment, patient survival on dialysis remained superior for blacks. We attempted to minimize any potential artifact by examining
TRANSPLANTATION
IN WOMEN
WITH
ESRD
patient survival for black patients who were referred and for those who were never referred. We also removed both white and black transplant patients from the analysis, but found patient survival to be superior in blacks regardless of the grouping. Diabetic black patients also experienced superior survival, although this was only evident from 6 months to 2.5 years. This study suggests that the survival advantage for blacks on dialysis exists for essentially all subgroups of patients. The hypothesis that black patients are younger, have less illness, and may have renal diseases that are associated with better survival has not been supported by our observations. Cacciarreli et alI6 demonstrated that longer time on dialysis may be associated with inferior allograft survival. At 5 years there was a 10% difference in allograft survival in patients who received a transplant less than 6 months after initiating dialysis compared with patients receiving more than 60 months of dialysis. Patients who received a transplant quickly were younger, had lower panel-reactive antibodies, and had less delayed graft function and rejection than those who waited longer. In the present study, black patients waited on dialysis approximately twice as long as whites, which might influence graft survival once they do receive a transplant. Our study suggests that the longer times on dialysis was not related to the preference of black patients to remain on dialysis because there was no difference in the frequency of refusal of the transplant option between races. Previous studies have suggested that black patients wait longer once they are placed on the transplant list, which may be related to the emphasis on matching of our current allocation system. The observation that white female homemakers were referred less than any other group was an unexpected observation. These patients were older and had more comorbid illness than other patients, which may partially explain this finding. Some of these patients may have elected to avoid referral for transplantation due to child care responsibilities or other reasons. Many white homemakers would probably have been accepted for transplantation and ultimately obtain allografts since their medical conditions were not prohibitive. These patients may be delaying referral rather than foregoing transplantation altogether. Once referred and accepted for trans-
747
plantation, white female homemakers also wait longer for transplantation. This group is probably more likely to be multiparous, and therefore may have higher panel-reactive antibodies. We found no other studies that examined the availability of transplantation to women based on their employment status. If this observation can be confirmed, white female homemakers may be the most underserved population in transplantation. Fortunately, this problem may be solved by a modification of patient attitudes toward transplantation or the provision of social services, which would allow these women to leave the home long enough to obtain kidney transplants. More information is needed to determine why these patients limited their access to transplantation. Because this study was conducted in two counties in Pennsylvania, it may not be representative of the nation as a whole. An expanded nationwide study is needed to confirm the findings of this report. In this study, the difference in transplantation rate between black and white females could not be traced to differences in selection for referral. Black patients were as likely to be referred for transplantation as whites. However, once referred and accepted for transplantation, they waited twice as long. From the results of this study, it appears that the most underserved population of women were white female homemakers. Although there were statistically significant differences in factors that might make referral and ultimate transplantation less likely, patient preference was a major contributor to the low transplantation rate in this group. In summary, there was no apparent influence of race in the decision to refer patients for transplantation. Factors that indicated the degree of illness and educational status were more important. Educational status, comorbid illness, and age were important in determining whether patients actually received a transplant. Race was an important factor. Patient survival for blacks was superior to whites in all subgroups, even diabetics and patients who were never referred. Superior patient survival could not be explained by the retention of healthy black patients on dialysis and thereby artificially yielding better results. The determinants of improved patient survival have yet to be identified. Further study of the referral process may provide part of the explanation for the disparity of transplantation rates in black females.
748
MCCAULEY
REFERENCES 1. Scoucie JM, Neylan JF, McClellan W: Race and sex differences in the identification of candidates for renal transplantation. Am J Kidney Dis 19:414-419, 1992 2. Held PJ, Pauley MV, Bovbjerg RR, Newmann J, Salvatierra 0 Jr: Access to kidney transplantation: Has the United States eliminated income and racial differences? Arch Intern Med 148:2594-2600, 1988 3. US Renal Data System: USRDS 1995 Annual Data Report. Bethesda, MD, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 1995, pp 91-108 4. Hosmer DW, Lemeshow S: Applied Logistic Regression. New York, NY, Wiley, 1989 5. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 6. Mantel N: Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 50:163-170, 1966 7. Cox DR: Regression models and life tables (with discussion). J R Stat Sot Series B 34:187-220, 1972 8. Cox DR: Partial likelihood. Biometrika 62:269-276, 1975 9. Norusis M: SPSS for Windows Advanced Statistics Release 6.0. Chicago IL, SPSS Inc, 1993 10. Simmons RG, Marine SK, Simmons RL: Gift of Life:
ET AL
The Effect of Organ Transplantation on Individual, Family, and Societal Dynamics. New Brunswick, NJ, Transaction Books, 1987 Il. Simmons RG, Marine SK, Simmons RL: Transplantation and changing norms: Cultural lag and ethical ambiguities, in Gift of Life: The Effect of Organ Transplantation on Individual, Family, and Societal Dynamics. New Brunswick, NJ, Transaction Books, 1987, pp 12-14 12. Bloembergen WE, Port FK, Mauger EA, Wolfe RA: Causes of death in dialysis patients: Racial and gender differences. J Am Sot Nephrol 5:1231-1242, 1994 13. US Renal Data System: USRDS 1995 Annual Data Report. Bethesda, MD, The National Institute of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 1995, pp 59-77 14. US Renal Data System: USRDS 1995 Annual Data Report. The National Institute of Health. National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD. Am J Kidney Dis 26:S69-S84, 1995 15. Bleyer AJ, Tell GS, Evans GW, Ettinger WH, Burkart JM: Survival of patients undergoing renal replacement therapy in one center with special emphasis on racial differences. Am J Kidney Dis 28:72-81, 1996 16. Cacciarelli TV, Sumrani N, Dibenedetto A, Hong JH, Sommer BG: Influence of length of time on dialysis before transplantation on long term renal allograft outcome. Transplant Proc 25:2474-2476, 1993