Tumor factors only partly explain bladder cancer disparities

Rochester, NY-In patients with bladder cancer, tumor characteristics at presentation are less favorable in women and African-Americans versus Caucasian men, according to a study by researchers at the University of Rochester, New York. However, these features make different contributions to gender- and race-related disparities in bladder cancer mortality, which suggests that different solutions are likely needed to improve bladder cancer survival outcomes for different demographic groups, said senior author Edward Messing, MD, professor of urology and oncology at the University of Rochester.

In order to investigate how much age and tumor characteristics could explain the gender- and race-related variability in bladder cancer mortality, Dr. Messing and colleagues analyzed data from the Surveillance, Epidemiology, and End Results (SEER) database. Information on patient demographics, cancer presentations, and outcomes were extracted for patients diagnosed with bladder cancer between 1990 and 2003. More than 101,000 patients were included in the study.

With patients divided into four groups by gender and race, analyses of hazards rates for bladder cancer-specific mortality reaffirmed the importance of both of these demographic features as prognostic variables, with the outlook for survival being poorest in African-American women. In addition, data on patient age and distribution of tumor characteristics corroborated previous reports that women and African-American patients of both genders present at a significantly higher age and with higher proportions of higher-stage tumors and nonurothelial carcinomas than Caucasian men do.

"From earlier epidemiology studies, we've long known that bladder cancer occurs most often in Caucasian males, but has much higher mortality rates in women than in men and in African-Americans of both genders compared with Caucasians," Dr. Messing said. "Understanding the underlying causes for these differences could suggest strategies for prevention and improving outcomes.

"While there are probably multiple factors accounting for the gender- and race-related disparities in death, our study focused on the role of tumor characteristics. The results surprised me because my pre-existing bias was that delay in diagnosis, evidenced by advanced stage at diagnosis, would emerge as the most significant contributor to the excess mortality hazard in both demographic groups. The findings support conducting educational programs aimed at the public and medical providers to improve early diagnosis of bladder cancer, but suggest these efforts would likely have the greatest impact on reducing disease mortality in African-Americans."

The results of the study showed that excess hazard for death from bladder cancer during the first 3 to 4 years of follow-up was about 80% higher among women versus men, >100% higher comparing African-American women and Caucasian women, and 40% to 73% higher in African-American men versus Caucasian men. The proportion of more aggressive cancer types, ie, nonurothelial carcinoma, was twice as high in women versus men for both races. However, these cancers only accounted for a small minority of all bladder cancers, even among African-American women, in whom their incidence was highest (10.5%).

Muscle-invasive disease was present in 25% of Caucasian women, 22% of Caucasian men, 43% of African-American women, and 30% of African-American men. Statistically significant differences were present for comparisons by gender in both racial groups and by race in men and women. In patients with nonmuscle-invasive disease, high-grade tumors were present most often in African-American women (36%). That incidence was significantly higher than among Caucasian women (27%) and African-American men (31%), whereas Caucasian men presented significantly more often with high-grade disease than Caucasian women (31% vs. 27%).

While previous studies have sought to explain disparities in bladder cancer mortality between demographic groups, the University of Rochester study is distinguished by its use of the SEER database and incorporation of tumor characteristics as covariates in a complex regression model. The SEER registry contains about 25% of all bladder cancer cases recorded in the U.S., Dr. Messing noted.

"This national database has some limitations in that all factors that might influence cancer survival are not ascertainable. In particular, information on treatment is limited. In addition, there may be inaccuracies and inconsistencies in the recorded data in terms of pathology and cause of death. However, because SEER provides such an enormous and broad sample, potential biases relating to these latter considerations are likely to be equally distributed among the comparator groups," he explained.

Other explanatory variables

Sex and racial disparities in the selection and efficacy of treatments might also account for increased mortality, which suggests a role for educating providers to achieve optimal delivery of appropriate treatment.

"Aggressive treatment is needed and often effective for aggressive forms of bladder cancer, but it is often withheld until too long," Dr. Messing said.

However, there also appears to be a clear need for further research to explore gender-related differences in tumor mechanisms. Dr. Messing and colleagues have been interested in understanding possible hormonally mediated differences in bladder cancer biology that might explain the increased risk of this malignancy in men and the higher mortality in women.

Results of the study were published in Cancer (2009; 115:68-74).