MORTALITY IN PROSTATE-CANCER

Purpose: We evaluated in patients with prostate cancer whether treatme nt is associated with in decreased mortality and whether the associati on of treatment with mortality varies with calendar time. Materials an d Methods: Using the 146,979 prostate cancer patients from the 1973 to 1990 public use tape of the Surveillance, Epidemiology and End Result s program, we performed survival analysis and multivariate proportiona l hazards modeling to estimate the relative risk of disease specific a nd overall mortality. Results: In men with prostate cancer, advanced a ge, black race, high tumor stage and not having treatment were indepen dently associated with disease specific and overall mortality. The rel ative risk of mortality in treated patients decreased significantly fr om 1973 to 1990. Conclusions: The fact that advanced age and black rac e are associated with disease specific mortality (even when treatment and stage are controlled) is a new observation, which suggests that tu mor biology and/or response to treatment is worse in elderly and black men. Overall, the data are consistent with the hypothesis that treatm ent of prostate cancer is associated with lower disease specific and o verall mortality rates. The decreases in relative mortality in treated patients from 1973 to 1990 indirectly support the theory that changes in patterns of care from 1973 to 1990 had a beneficial effect on mort ality. This observation implies that future studies should account for calendar time when interpreting outcomes data.