Abstract
BACKGROUND: Androgen deprivation therapy (ADT) remains a primary treatment for localized prostate cancer (PCa) even though there is no evidence that its use is beneficial in the absence of curative treatment. METHODS: Men aged ≥70 years (n = 16,534) diagnosed with localized PCa from 1985 to 2014 and managed either with primary observation or ADT in the absence of curative treatment were included. The cases were identified from the population-based Finnish Cancer Registry. We estimated the standardized mortality ratios (SMR) for overall mortality by treatment group. We determined the relative risk (RR) of PCa-specific mortality (PCSM) and other-cause mortality between the two treatment groups. Survival was determined using the life table method. Two age groups (70-79 years and ≥ 80 years) and three calendar time cohorts (1985-1994, 1995-2004, and 2005-2014) were compared following adjustment of propensity score matching between the treatment groups with four covariates (age, year of diagnosis, educational level, and hospital district). Follow-up continued until death or until December 31, 2015. RESULTS: Patients in the observation group had lower overall SMRs than those in the ADT group in both age cohorts over the entire study period. PCSM was higher in men aged 70-79 years undergoing primary ADT compared to those managed by observation only (RR: 1.70, 95% confidence interval [CI]: 1.29-2.23 [1985-1994]; RR 1.55, 95% CI: 1.35-1.84 [1995-2004]; and RR 2.71, 95% CI: 2.08-3.53 [2005-2014]); p = 0.005 for periodic trend. A similar trend over time was also observed in men aged > 80 years; (p for age-period interaction = 0.237). Overall survival was also higher among men in their 70's managed by observation compared to those undergoing ADT. CONCLUSIONS: Primary ADT within four months period from diagnosis is not associated with improved long-term overall survival or decreased PCSM compared to primary conservative management for men with localized PCa. However, this observational study's conclusions should be weighted with confounding factors related to cancer aggressiveness and comorbidities.