Survival experience of black patients and white patients with bladder carcinoma


  • George R. Prout Jr. M.D.,

    1. Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, Maryland
    Search for more papers by this author
    • George R. Prout, Jr. is a consultant to the NCI Division of Cancer Control and Population Sciences.

  • Margaret N. Wesley Ph.D.,

    1. Information Management Services, Inc., Silver Spring, Maryland
    Search for more papers by this author
  • Peter G. McCarron M.D., M.P.H.,

    1. Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, Maryland
    Search for more papers by this author
  • Vivien W. Chen Ph.D.,

    1. Louisiana State University Health Sciences Center, Epidemiology Program, Louisiana Tumor Registry, New Orleans, Louisiana
    Search for more papers by this author
  • Raymond S. Greenberg M.D., Ph.D.,

    1. Medical University of South Carolina, Charleston, South Carolina
    Search for more papers by this author
  • Robert M. Mayberry M.P.H., Ph.D.,

    1. Program for Healthcare Effectiveness Research, Morehouse School of Medicine, Atlanta, Georgia
    Search for more papers by this author
  • Brenda K. Edwards Ph.D.

    Corresponding author
    1. Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, Maryland
    • Division of Cancer Control and Population Sciences, National Cancer Institute, National Institutes of Health, 6116 Executive Boulevard, Bethesda, MD 20892-8315
    Search for more papers by this author
    • Fax: (301) 480-4077

  • Senior staff at the NCI provided overall scientific direction and program management of the multicenter collaborative study conducted under contracts. They have shared responsibility for the design, conduct, interpretation, and analysis of study results as well as participating as coauthors of the submitted article based on established policy and procedures for governance of this multiinvestigator study. Review and clearance of the article was provided by NCI scientific staff other than the coauthors.



Blacks are less likely than whites to develop bladder carcinoma. However, once they are diagnosed, black patients experience poorer survival. The authors investigated which factors were related to survival differences in black patients and white patients with bladder carcinoma stratified by extent of disease.


A population-based cohort of black patients with bladder carcinoma and a random sample of frequency-matched white patients with bladder carcinoma, stratified by age and gender, were identified through cancer registry systems in Atlanta, New Orleans, and San Francisco/Oakland. Patients had no previous cancer history and were ages 20–79 years at the time they were diagnosed with bladder carcinoma in 1985–1987. Medical records were reviewed at initial diagnosis, and 77% of patients were interviewed. Tumor grade, T classification, and other variables, including age, socioeconomic position, symptom duration, smoking history, and comorbidities, were recorded. Survival of black patients and white patients by extent of disease was modeled using Cox regression analysis.


A greater proportion of black patients had histologic types of tumors that were associated with poorer survival. Among those with pure urothelial carcinoma, black patients had greater extent of disease at the time of diagnosis. Within specific extent-of-disease categories, there was some evidence of poorer survival for black patients with T2 tumors and strong evidence of poorer survival among those with T3 tumors compared with white patients. Black patients with muscle-invasive carcinoma who died within 6 months of diagnosis tended to present with life-threatening symptoms. Black patients and white patients did not differ with respect to diagnostic tests performed or therapy given.


Black patients with bladder carcinoma had poorer survival due to greater extent of disease at diagnosis and a higher proportion of more aggressive histologies compared with white patients. Within urothelial carcinomas, by extent of disease (clinical/pathologic stage) these black/white survival differences were limited to patients with muscle invasion (T2 and T3 tumors). Cancer 2004. Published 2003 by the American Cancer Society.

The National Cancer Institute Black/White Cancer Survival Study (BWCSS) was instituted to elucidate the causes of observed cancer survival differences between blacks and whites.1 For bladder carcinoma, 1 of 4 disease sites studied, the 5-year relative survival rates over the period 1986–1988 were 51.7% for black women and 67.6% for black men, compared with 76.2% for white women and 82.3% for white men.2 More recently, from 1992 to 1999, there has been improvement in survival rates for black women (52.9%), black men (70.1%), white women (77.0%), and white men (84.6%).2 The survival differences are contrary to differences in incidence, with white males having approximately twice the incidence of black males.3

Black patients are reported to have greater extent of disease at the time of diagnosis than white patients2 and also are more likely to have aggressive cell types, such as squamous cell carcinoma and adenocarcinoma.4, 5 Patients with these nonurothelial carcinomas often present with higher stage of disease and have worse survival.4

There is evidence that individuals with low socioeconomic status have more advanced disease at diagnosis.4, 5 A large study of bladder carcinoma survival found that histologic type, stage, and grade influenced black/white survival differences.6 In addition, black patients received radiation and chemotherapy more often than white patients.6 A later study found that, after controlling for gender, age, tumor histology, and stage at diagnosis, black patients were more likely than white patients to receive no treatment.7

To date, racial differences in survival by stage of bladder carcinoma have not been examined. In this article, we report, in more detail than was possible previously, on the factors that influence the observed black/white differences in survival in patients with bladder carcinoma, concentrating on the analysis of survival differences by extent of disease.


Full details of the study design are available elsewhere.1 Briefly, patients ages 20–79 years with histologic evidence of bladder carcinoma who were diagnosed between January 1, 1985, and December 31, 1987, and who had no previous history of cancer were identified through population-based tumor registries in Atlanta, New Orleans, and San Francisco/Oakland. All eligible black patients were selected. White patients were sampled randomly to frequency match the expected distribution of black patients within age groups (ages 20–49 years, 50–64 years, and 65–79 years), gender, and registry. Due to lower than expected accrual among blacks, the sampling fraction of whites was increased later, with the goal of increasing the statistical power to detect differences between blacks and whites.

Abstractors accessed the patients' medical records and obtained copies of operative and pathology reports. Diagnostic slides were reviewed by two central pathologists.4 Patients who consented were interviewed by trained study personnel. All sources of submitted data on 566 eligible patients were reviewed by one of the authors, a urologist (G.R.P.). Symptoms, tumor grade, and extent of disease were assigned in a uniform manner. Eleven patients were excluded because the central pathology review did not identify tumor in the submitted slide. Two additional patients lacked information regarding survival. The remaining 553 patients were included in this analysis. Extent of disease and tumor grade was based on the best available data using a modification of the American Joint Committee on Cancer (AJCC) TNM classification system.8 Information concerning diagnostic tests and treatment up to 4 months after diagnosis was obtained from the hospital medical record as well as from physicians' offices.

Patients from the Atlanta and San Francisco/Oakland metropolitan areas were followed for survival and cause of death through Surveillance, Epidemiology, and End Results Program (SEER) reporting requirements. Patients from New Orleans were followed through the Louisiana Tumor Registry.

Statistical Analysis

Determinants of survival were examined by stage of disease. We previously reported on the factors associated with extent of disease at diagnosis.4 Survival curves were estimated using the Kaplan–Meier method9 and were compared using the log-rank test.10 Cox regression analysis11 was used to model survival. The endpoint for survival analysis was death from any cause. Patients were followed to December 31, 1997. Ninety-eight percent of white patients and 96% of black patients were followed to death or for at least 10 years. The proportional hazards models were stratified by age group (ages 20–49 years, 50–64 years, and 65–79 years), gender, and metropolitan area (Atlanta, New Orleans, or San Francisco/Oakland). For T2 and T3 tumors, some covariate and/or stratification levels were combined due to small cell sizes. Differences in categoric variables between black patients and white patients were analyzed using the chi-square test. Analyses were performed using SAS software (SAS Institute, Cary, NC).

Interview status, tumor grade, presence of comorbidities, symptom status, time from first symptom to first medical consultation, time from first symptom to initial anticancer therapy, smoking history, gender, income, health insurance, usual source of care, poverty index,12 occupation,13 education level, and marital status were tested for their influence on overall survival and on the black/white difference. Factors that were unknown were included as a separate category for each covariate.


Details of baseline characteristics are presented elsewhere.4 In brief, black patients were more likely to have higher grade tumors, lower socioeconomic position, and fewer years of education and were more likely to be current smokers and to have comorbidities compared with white patients. Symptom duration was longer in black patients, who also were less likely than white patients to have medical insurance and, furthermore, were more likely either to have no usual source of care or to have public sources of care. Overall, black patients had poorer all-cause survival compared with white patients (P = 0.003). The median survival for white patients was 125 months (95% confidence interval [95% CI], 105–140) versus 54 months for black patients (95% CI, 38–90).

The majority of patients (97% of whites, 82% of blacks) had pure urothelial carcinomas, whereas most squamous cell carcinomas and adenocarcinomas (Table 1) occurred in black patients. Patients with these tumors tended to have more advanced disease and poorer survival compared with patients who had pure urothelial tumors.

Table 1. Distribution of Tumor Histologies by Race
Tumor histologyNo. of patients (%)
White (n = 369)Black (n = 184)
Pure urothelial carcinoma353 (96.7)151 (82.1)
Urothelial carcinoma with other cell types  4 (1.1) 12 (6.5)
Squamous with or without neuroendocrine carcinoma  7 (1.9)  8 (4.3)
Adenocarcinoma  2 (0.5) 10 (5.4)
Other histologies  
 Undifferentiated  1 (0.3)  3 (1.6)
 Anaplastic  1 (0.3)  0 (0.0)
 Small cell  1 (0.3)  0 (0.0)

Urothelial Tumors

Table 2 shows that black patients had greater extent of disease than white patients (P < 0.0001). Black patients and white patients with nonmuscle-invasive (Ta, T1) urothelial carcinomas survived equally well. At 10 years, the estimated survival among patients with Ta tumors was 62% for white patients and 70% for black patients; whereas, among patients with T1 tumors, > 50% of both black and white patients were alive at 10 years. When the tumor was both muscle invasive and extended to the pelvic lymph nodes and to distant sites outside the pelvis (T4, N+, or M+), survival was very poor, with a median survival of < 18 months, but was similar with regard to race. Only among those with T2 and T3 tumors were black patients found to have poorer survival compared with white patients (Figs. 1, 2).

Table 2. Distribution of Disease T Classification by Race in Patients with Pure Urothelial Carcinoma
StageNo. of patients (%)
White (n = 353)Black (n = 151)
  • a

    Includes four patients (three white patients and one black patient) for whom it could not be determined whether the tumor was Ta or T1.

Tis  3 (1) 3 (2)
Taa195 (55)57 (38)
T1 89 (25)33 (22)
T2 27 (8)19 (13)
T3 17 (5)16 (11)
T4, N0 or Nx, M0 or Mx  7 (2) 5 (3)
N+, any T, M0 or Mx  6 (2) 6 (4)
M+, any T, any N  5 (1) 7 (5)
Unknown  4 (1) 5 (3)
Figure 1.

Overall survival in patients with T2 urothelial cell carcinoma by race (P = 0.0638; log-rank test).

Figure 2.

Overall survival in patients with T3 urothelial cell carcinoma by race (P < 0.001; log-rank test).

Table 3 provides an overview of 5-year survival estimates by selected variables for patients with nonmuscle-invasive and muscle-invasive tumors. Increasing age generally was associated with worse survival, and patients with high-grade disease also had poorer survival. There was evidence that patients who were interviewed had better survival compared with patients who were not interviewed. In general, patients with higher socioeconomic position, as measured by income, poverty index, occupation, and education, had better survival compared with patients from lower socioeconomic strata. There were no consistent associations between survival and smoking status, symptom duration, time to diagnosis, or marital status.

Table 3. Association between Prognostic Variables and 5-Year Survival, with 95% Confidence Intervals, for Nonmuscle-Invasive and Muscle-Invasive Tumors in Black Patients and White Patients
VariableNonmuscle-invasive tumors (Ta, T1)Muscle-invasive tumors (T2, T3)
White patientsBlack patientsWhite patientsBlack patients
No.5-yr survival (%)95% CINo.5-yr survival (%)95% CINo.5-yr survival (%)95% CINo.5-yr survival (%)95% CI
  1. 95% CI: 95% confidence interval.

Age at diagnosis (yrs)            
Tumor grade            
 Grade 1958578–92298673–99010
 Grade 21468478–90497866–8968354–1004250–67
 Grade 3435843–73125022–78386347–7830206–34
Time to diagnosis (mos)            
 < 1908678–93278977–100108055–1008250–55
 1 to < 3517158–83147148–956330–719220–49
 3 to < 6207556–9477138–10087545–10020
 ≥ 6509284–100247962–9597851–1008250–55
 < $10K277458–91276345–814501–9916254–46
 $10K to < $20K437967–91208569–100107042–9820
 $20K to < $35K528170–91138565–10077138–10030
 ≥ $35K739084–97129174–10087545–1001100
Poverty index            
 > 400%978982–95158667–100147148–952500–100
Marital status            
 Never married168162–10068354–1002010

Tis Tumors

All 6 patients in the group with Tis tumors had Grade 3 disease. Two patients died of bladder carcinoma, and the remaining four patients died from other chronic conditions.

Ta/T1 Tumors

There were 284 white patients (42% female) and 90 black patients (24% female) with Ta/T1 disease. There was no evidence that white patients and black patients differed with regard to their access to diagnostic tests. Among white patients, 73% received a chest X-ray, 16% received computed tomography (CT) scans, and 7% received bone scans, compared with 73%, 18%, and 9%, respectively, among black patients.

A transurethral resection of bladder tumor (TURB) was undergone by 96% of white patients and by 98% of black patients. Sixteen percent of white patients and 20% of black patients also received radiation or chemotherapy. Seven white patients and one black patient underwent total or partial cystectomies. The remaining patients underwent biopsies with or without radiation or chemotherapy.

There were 134 deaths among the white patients, 56 of which (42%) occurred within 5 years of diagnosis, including 14 deaths due to bladder carcinoma, 13 deaths due to other malignancies, and the majority of the remaining deaths due to cardiorespiratory diseases. Among patients who died after 5 years, 8 patients died of bladder carcinoma, 21 patients died of other malignancies, and the remaining patients died mainly of cardiorespiratory and other chronic illnesses. Of 38 deaths among black patients, 21 deaths occurred within 5 years of diagnosis. Four patients died of bladder carcinoma, and the remaining deaths were due to other malignancies or chronic diseases. During the next 5 years, there was 1 death due to bladder carcinoma.

Table 4 shows the association between risk of death and prognostic variables for patients with Ta/T1 tumors. There was no difference in survival between black patients and white patients stratified by age, gender, or metropolitan area (hazard ratio, 0.95; 95% CI, 0.65–1.38). Increasing age was associated with a greater risk of death; compared with the reference age group of youngest patients (ages 20–49 years), the hazard ratio in patients ages 50–64 years was 2.34 (95% CI, 0.97–5.66), and the hazard ratio in the oldest age group (ages 65–79 years) was 9.65 (95% CI, 4.26–21.90).

Table 4. Ta/T1 Hazard Ratios and 95% Confidence Intervals for Associations between Single Covariates and Risk of Death in Black Patients and White Patients with Urothelial Carcinoma when Added to a Model Containing the Design Variablesa
CovariateHazard ratio (95% confidence interval)
White and black patientsWhite patientsBlack patients
  • a

    Design variables: race, metropolitan area, gender, and age group.

  • b

    Reference group.

Tumor grade   
 Grade 3b1.001.001.00
 Grade 10.43 (0.26–0.71)0.43 (0.26–0.71)0.43 (0.13–1.45)
 Grade 20.66 (0.44–1.01)0.68 (0.44–1.01)0.53 (0.20–1.40)
 No0.20 (0.11–0.34)0.19 (0.10–0.37)0.31 (0.09–1.13)
 Unknown0.56 (0.29–1.05)0.67 (0.35–1.27)
Total time to diagnosis (mos)   
 1–31.90 (1.21–2.99)2.18 (1.30–3.64)1.28 (0.38–4.30)
 3–62.04 (1.21–3.71)2.55 (1.31–4.93)1.12 (0.22–5.61)
 > 61.14 (0.69–1.88)0.95 (0.51–1.75)1.87 (0.69–5.10)
 Unknown1.75 (1.15–2.66)1.74 (1.08–2.82)1.93 (0.73–5.12)
 < $10K1.001.001.00
 $10–20K0.54 (0.33–0.89)0.58 (0.32–1.08)0.32 (0.11–0.91)
 $20–35K0.45 (0.26–0.78)0.41 (0.21–0.79)0.37 (0.10–1.41)
 > $35K0.46 (0.27–0.81)0.44 (0.23–0.83)0.44 (0.09–2.16)
 Unknown0.78 (0.50–1.21)0.71 (0.41–1.24)0.70 (0.27–1.82)
Poverty index   
 < 125%
 125–200%0.83 (0.48–1.47)0.86 (0.41–1.79)0.51 (0.15–1.70)
 201–300%0.43 (0.22–0.85)0.53 (0.23–1.22)0.22 (0.05–1.07)
 301–400%0.43 (0.21–0.89)0.47 (0.20–1.09)0.03 (0.0–2.26)
 > 400%0.51 (0.30–0.86)0.49 (0.25–0.96)0.51 (0.15–1.75)
 Unknown0.82 (0.51–1.32)0.80 (0.42–1.51)0.75 (0.29–1.96)
Occupational class   
 Technical/sales1.11 (0.65–1.91)1.05 (0.59–1.87)2.60 (0.25–27.03)
 Skilled1.49 (0.88–2.52)1.64 (0.93–2.90)1.69 (0.19–14.87)
 Unskilled2.19 (1.28–3.74)1.45 (0.77–2.73)8.25 (1.04–65.65)
 Homemaker2.90 (1.04–8.07)2.69 (0.95–7.63)
 Unknown1.90 (1.14–3.15)1.66 (0.96–2.86)6.10 (0.71–52.12)
Marital status   
 Widowed1.59 (1.06–2.39)1.94 (1.20–3.14)1.27 (0.50–3.25)
 Divorced/separated1.07 (0.63–1.81)1.39 (0.74–2.63)0.75 (0.29–1.96)
 Never1.58 (0.74–3.40)1.95 (0.78–4.90)1.41 (0.29–7.01)
 Unknown1.23 (0.16–9.71)1.46 (0.17–12.28)

Adding single covariates to the model with race altered the hazard ratio for race by ≤ 10%. A lower risk of death was observed in patients with no comorbidities, a shorter duration from first symptom to diagnosis, income > $10,000, professional/management occupational level, and higher poverty indices. There was some evidence that patients who never married or who were widowed had a greater risk of death compared with married patients. Other covariates were not associated significantly with survival.

When black patients and white patients were modeled separately, similar patterns emerged. When the above analyses described above were repeated for patients with Ta tumors separately from patients with T1 tumors, the results were similar.

T2 Tumors

There were 27 white patients (48% female) and 19 black patients (42% female) with T2 disease. The median age at diagnosis was 67 years for whites and 68 years for blacks. There were no asymptomatic patients. All patients had either Grade 2 or Grade 3 tumors.

There were no differences in the proportions of black and white patients who received chest X-rays (79% in blacks vs. 78% in whites), CT scans (53% vs. 52%, respectively), and bone scans (37% vs. 26%, respectively). Total or radical cystectomies were undergone by 8 white patients (30%) and by 2 black patients (11%; P = 0.16). Of the 17 white patients (63%) who underwent TURB, 8 patients also received chemotherapy or radiation therapy. Among black patients, 13 patients (68%) underwent TURB; 7 of those patients also received chemotherapy or radiotherapy. Overall, 15 white patients (56%) and 9 black patients (47%) underwent pelvic surgery and/or received irradiation. The median age for black patients who did not undergo pelvic surgery or receive irradiation was 73 years; black patients who received these treatments had a median age of 66.5 years. In white patients, the corresponding median ages were 68 years and 67 years, respectively. There was some evidence that survival for black patients was poorer compared with survival for white patients (P = 0.06) (Fig. 1). The hazard ratio for black patients compared with white patients was 1.64 (95% CI, 0.72–3.73). There was no statistically significant difference in survival (P = 0.95) between the 25 men (17 deaths) and the 21 women.

There were 16 deaths among white patients with T2 tumors. Eight deaths were due to bladder carcinoma, 5 of which occurred within 5 years after diagnosis. Other causes of death included second malignancies, embolism/infarction of the extremities, chronic obstructive pulmonary disease, suicide, and psychosis. Among black patients, there were 15 deaths, 8 due to bladder carcinoma, and 7 of those deaths occurred in the first 5 years after diagnosis. Other causes of death included second malignancies, cardiorespiratory diseases, senility, intestinal obstruction, and pleural effusion.

When they were stratified according to metropolitan area alone, patients who had no comorbidities had better survival compared with patients who had comorbidities (hazard ratio, 0.29; 95% CI, 0.11–0.77). However, when the analysis also was stratified according to age group (age ≥ 65 years vs. age < 65 years) and gender, comorbidity was not significant. Patients in the combined category of public care or no usual source of care had better survival compared with patients who had private sources of care (hazard ratio, 0.16; 95% CI, 0.04–0.64). Patients who were widowed had a higher hazard ratio compared with married patients (hazard ratio, 7.09; 95% CI, 1.13–44.38). Other covariates were not associated significantly with survival.

T3 Tumors

There were 17 white patients (59% female) and 16 black patients (44% female) with T3 disease. All patients were symptomatic at the time of diagnosis. Chest X-rays were reported in 82% of white patients and in 100% of black patients, CT scans were reported in 65% of white patients and in 75% of black patients, and bone scans were reported in 41% of white patients and in 38% of black patients. Six black patients and 5 white patients underwent total or radical cystectomies. One-half of the black patients underwent TURB, one black patient underwent a partial cystectomy, and one black patient underwent a biopsy only. White patients who did not undergo total or radical cystectomy underwent either TURB or partial cystectomy in equal numbers. Five white patients and 5 black patients received either chemotherapy or radiotherapy in addition to surgery.

There were 13 deaths among white patients, including 7 deaths due to bladder carcinoma; and the remaining deaths were due to other conditions, including cardiorespiratory disease, leukemia, and Alzheimer disease. Whereas no white patient died within 6 months of diagnosis, 5 black patients died within this period. Three of those black patients presented with hydronephrosis; treatments included biopsy only, TURB, and radical cystectomy. One patient with significant comorbidities underwent TURB. The remaining patient (comorbidity unknown) underwent a radical cystectomy. All 16 black patients died within 5 years of diagnosis, and 14 of those deaths were due to bladder carcinoma.

The survival curves shown in Figure 2 indicate a large disparity between blacks and whites (P ≤ 0.0001). There were no patients in the group ages 20–49 years, and there was only 1 black patient in each metropolitan area in the group ages 50–64 years. When patients were stratified according to metropolitan area and gender, black patients were at much higher risk of death compared with white patients (hazard ratio, 8.48; 95% CI, 2.40–29.95). Lack of an interview was a significant predictor of survival. Patients with no comorbid conditions lived longer than patients with at least 1 comorbid condition (hazard ratio, 0.17; 95% CI, 0.03–0.96). Occupational class and income were not modeled, because there was little overlap between blacks and whites. All white patients had privately funded health insurance. Similar to the patients with T2 disease, survival was comparable between men and women (15 deaths occurring in 16 men, and 14 deaths occurring in 17 women; P = 0.63).

T4, N+, or M+ Tumors

There were 18 white patients (44% female) and 18 black patients (17% female) with T4, regional lymph node positive, or metastatic bladder carcinoma. Patients of both races with T4, N+, or M+ disease had similarly poor survival (P = 0.31). Eleven white patients and 16 black patients died of bladder carcinoma. The remaining deaths were due to other malignancies, cardiorespiratory disease, or septicemia.


To our knowledge, this is the first study to examine the survival of black and white patients with bladder carcinoma by extent of disease. There was no difference in survival between black and white patients with either early (Ta/T1) or highly advanced (T4, N+, or M+) disease. However, black patients with muscle-invasive (T2/T3) disease had poorer survival compared with white patients; these survival differences were moderate for patients with T2 tumors and substantial for patients with T3 tumors.

The majority of patients with bladder carcinoma have low-grade, low-stage urothelial carcinomas; in the current study, 60% of black patients and 80% of white patients presented with Ta or T1 disease. In the current study, black patients were more likely to have invasive carcinoma (≥ T1) at diagnosis and, among patients with invasive carcinoma, were more likely to have muscle-invasive carcinoma.4 These findings are consistent with reports for other tumor sites in the BWCSS,14 in which extent of disease at diagnosis was the primary determinant of survival. Three possible explanations for these survival differences are considered: 1) underdiagnosis or uncertainty of stage, 2) unequal access to care or inadequate treatment, and 3) impact of comorbidity.

Underdiagnosis and Uncertainty in Staging

Earlier reports suggested that the higher proportion of black patients with greater extent of disease at diagnosis may have been due to underdiagnosis of noninvasive bladder carcinomas in blacks.6, 15 Differences in survival in the current study occurred only in patients with T2/T3 disease. It is unlikely that any under-diagnosis of early tumors would influence our findings. Second, black patients with T2 and T3 disease may have had greater extent of disease than reported. However, we found no racial differences in diagnostic work-ups. A further possible explanation is that the AJCC TNM staging scheme is relatively crude and leads to variability within classification of stage. Black patients may have had more extensive disease within the categories of T2 and T3, yet this did not appear to be true for patients with T1 disease.


Appropriate therapy varies widely with the histology and extent of disease of bladder tumors. For patients with low-grade, low-stage disease, undergoing transurethral resection may be adequate.16 Muscle-invasive disease calls for more aggressive surgery, with the possible inclusion of radiotherapy and/or systemic chemotherapy.17 It has been hypothesized that black/white survival differences observed in the SEER data base occurred because black patients were less likely to receive surgery compared with white patients.7 In the current study of a carefully staged sample of patients, although there were only 17 white patients and 16 black patients with T3 disease, there did not appear to be racial disparities in therapy.


We found that comorbidity was a significant predictor of survival in groups with muscle-invasive tumors and nonmuscle-invasive tumors. Some information on comorbidity can be gleaned from death certificates, but it is uncommon to find details about deaths from causes other than bladder carcinoma in reports of the results of cystectomy, radiotherapy, or chemotherapy unless the complications were related to therapy. Malmstrom and colleagues18 noted that, of 124 deaths in 220 patients (median age, 68 years), 64% were from bladder carcinoma, and 34% were from other causes. In the current study, we found that a large proportion of deaths were from causes unrelated to bladder carcinoma. In particular, cardiorespiratory death was common, indicating that many patients had other serious health conditions.

Most bladder carcinoma occurs in older people, who are likely to have other medical conditions that compromise their ability to tolerate therapy and their overall prognosis.19 Reportedly, radical cystectomy for muscle-invasive bladder carcinoma in the elderly has been successful, with acceptable morbidity and mortality20; however, that success was seen in carefully selected patients, making it difficult to determine the applicability of such results to the general population. Patients in the current study with T2 or T3 tumors who were not treated presented with serious comorbid conditions that likely would impair their ability to tolerate major surgery, radiotherapy, and chemotherapy. Other studies also have reported that the presence of comorbid conditions in patients who were eligible for cystectomy may have resulted in less aggressive and potentially curative treatments being offered. Population-based studies from Sweden21 and Norway22 have reported rates of radical cystectomy far less than expected under national guidelines.

The literature addressing comorbidity in patients with bladder carcinoma, 50% of whom are age > 70 years at the time they are diagnosed, is limited.23 Despite this, urologists and radiation and medical oncologists have designed modifications that are tolerated better compared with standard therapy by patients with significant comorbid conditions.24, 25 In particular, radical radiotherapy in patients age > 70 years with muscle-invasive carcinoma results in a reported 50% complete response rate, with a 30% 5-year survival rate.25

Strengths and Limitations

The strengths of the BWCSS include the population-based sample, relatively large sample size, and collection of data in three different urban areas. In addition, a standardized review of clinical and pathologic material and a central review of slides were conducted to improve the reliability and consistency of data on extent of disease.

Several limitations of the study should be mentioned. Medical care could not be evaluated adequately, because only broad categorizations of chemotherapy or radiotherapy were documented, and the study was limited to treatment for bladder carcinoma. A small number of participants were deceased or were too ill, resulting in missing interview information, a problem that was particularly acute for indicators of socioeconomic position and access to care. Beyond the 4-month window after diagnosis, there was no follow-up information available other than date and cause of death.


The current study has highlighted the differences in survival between black patients and white patients with bladder carcinoma. Because black patients are more likely to present with more aggressive, nonpure-urothelial tumors compared with white patients and among patients with pure urothelial tumors, they present with greater extent of disease. Survival differences between black patients and white patients were confined to those with T2 and T3 tumors. Neither underdiagnosis nor treatment differences explained these differentials. Future studies aimed at improving the understanding of racial differences in bladder carcinoma survival should concentrate on the muscle-invasive tumor types, evaluate carefully the extent of disease and comorbidity, collect data on treatment and monitoring beyond the immediate postdiagnosis period, and evaluate the medical care environment beyond the confines of bladder carcinoma.


The authors acknowledge the assistance of Dr. Charles C. Brown for his input on early versions of the article and Ms. Terri Harshman for preparation and editing of the final article.