Risk of prostate cancer mortality in men with a history of prior cancer.

Objectives: To describe outcomes of patients with prostate cancer diagnosed after another malignancy and identify factors associated with prostate cancer death in this population, as little is known about the clinical significance of prostate cancer as a subsequent malignancy.

Patients And Methods: We studied 18 225 men diagnosed with prostate cancer after another malignancy from 1973 to 2006. We compared demographic and clinical variables, and the proportion of death from prostate cancer vs prior malignancy with t-test and chi-squared analyses. Fine and Gray's regression was used to consider the effect of treatment on prostate cancer death. We then studied a second cohort of 88 013 men with prostate cancer as a first or second malignancy to describe current diagnostic and treatment patterns.

Results: One in seven men died from prostate cancer in our first cohort. More died from prostate cancer following colorectal cancer (16.8% vs 13.7%), melanoma (13.4% vs 7.56%), and oral cancer (19.1% vs 4.04%), but fewer following bladder cancer, kidney cancer, lung cancer, leukaemia and non-Hodgkin's lymphoma (all P < 0.001). Prostate cancer treatment was associated with a nearly 50% lower risk of death when high-grade or high-stage (adjusted hazard ratio 0.55, 95% confidence interval [CI] 0.47-0.64). Patients who died from prostate cancer had higher grade and stage disease, and received less treatment than patients who died from prior malignancy. The second cohort showed subsequent prostate cancer had more high-risk disease (36.3% vs 22.2%, P < 0.001) and less prostate cancer treatment (adjusted odds ratio 0.872, 95% CI 0.818-0.930) than primary prostate cancer.

Conclusions: Prostate cancer remains a significant cause of mortality when diagnosed as a subsequent cancer. These results suggest prostate cancer treatment should be seriously considered in patients with prior malignancies, especially those with high-grade or locally advanced prostate cancer.