Comorbidity and age cannot explain variation in life expectancy associated with treatment of non-metastatic prostate cancer.

Abstract

Age and Charlson comorbidity index (CCI) affect life expectancy (LE) and other-cause mortality (OCM) in non-metastatic prostate cancer (nmPCa) patients. We examined their ability to predict OCM in individuals treated with radical prostatectomy (RP), brachytherapy (BT), external beam radiation (EBRT) androgen deprivation (ADT) or observation. We postulated that these variables are not sufficient to explain OCM and LE patterns according to different treatment types. We relied on the SEER-Medicare database from 1991 to 2009. Overall, 283,125 patients with non-metastatic prostate cancer aged ≥66years were treated with RP (15.5%), BT (13.9%), EBRT (21.4%), ADT alone (16.3%) or observation (32.8%). Cumulative incidence of OCM and LE was stratified by treatment type and adjusted for age and CCI. Competing risks regression was also used. OCM rates vary according to treatment, despite age and CCI adjustment. In RP or BT patients, LE exceeds 10years, regardless of age and CCI. Conversely, a 10-year LE is not reached by patients >74years treated with observation or ADT. In OCM competing risks regression, age, CCI and treatment type achieved independent predictor status (all p<0.001). In patients with nmPCa, neither age nor CCI can accurately estimate OCM or LE in excess of 10years. Primary treatment assignment is a strong determinant of OCM and LE, where RP and BT patients enjoy better OCM and LE rates than observation ADT or EBRT patients. In consequence, better clinical tools are needed to accurately assess OCM and LE in those settings.