Association of preoperative testosterone levels with biochemical failure in men undergoing radical prostatectomy

Background: Low testosterone levels in men treated with radical prostatectomy (RP) for prostate cancer have been associated with worse pathological stage and outcomes. Long-term data correlating preoperative testosterone levels and RP outcomes are presented. Methods: A total of 177 patients were identified who had preoperative testosterone levels and were treated with RP with long-term follow-up. Associations of testosterone levels with tumor stage, Gleason score, and time to biochemical failure (BCF) were assessed using Pearson's chi-square test, Mantel-Haenszel test, Spearman rank correlation, and life-table methods (Kaplan-Meier curves, log-rank test, Cox proportional hazards model) as appropriate. Testosterone was analyzed both as a continuous factor and as low ( < 300 ng/dl) versus normal testosterone levels. Results: The median testosterone level was 423 ng/dl, median age was 65 years, and median follow-up was approximately 15 years. A significant association between testosterone values below 300 ng/dl and positive surgical margin ( p = 0.0389) as well as seminal vesicle invasion ( p = 0.0182) was found. While patients with a testosterone level < 300 ng/dl also showed a higher percentage of BCF over time (45% vs. 29% at 15 years), the difference was not statistically significant ( p = 0.076, log-rank test). When adjusted for known predictors of BCF (Gleason score, pre-RP prostate-specific antigen, seminal vesicle involvement, and positive surgical margin), the effect of testosterone as a continuous predictor was not statistically significant ( p = 0.57). Conclusion: The longest follow-up to date (15 years) correlating preoperative testosterone levels and RP outcomes is presented. With this study, low testosterone levels were significantly associated with a positive surgical margin and seminal vesicle invasion, and slightly higher ( p = 076) BCF rates.