Preoperative serum testosterone and early urinary continence recovery after nerve-sparing radical prostatectomy: a landmark analysis

Urinary continence recovery after nerve-sparing radical prostatectomy is largely predicted by nonmodifiable factors, including age and membranous urethral length. Preoperative serum testosterone is a potentially modifiable biochemical factor, but its association with postoperative continence remains unclear. We investigated this association and its temporal pattern using landmark analysis. We retrospectively analyzed 482 patients with clinically localized prostate cancer who underwent robot-assisted or laparoscopic nerve-sparing radical prostatectomy at a single tertiary center between January 2020 and December 2024. Complete continence was defined as 0 pads/day. The primary outcome was time from catheter removal to first documented complete continence recovery, coded as the event; patients without recovery were censored at their last assessment, with follow-up limited to 52 weeks. Testosterone was categorized as low (< 350 ng/dL) or high (≥ 350 ng/dL). Kaplan–Meier analysis, log-rank tests, and multivariable Cox regression were performed. Week 6 was the landmark separating early (0–6 weeks) and subsequent (> 6 weeks) recovery. Of 482 patients, 350 achieved complete continence and 132 were censored. Older age (> 70 vs ≤ 70 years, HR = 0.67, 95% CI 0.54–0.83), unilateral vs bilateral nerve sparing (HR = 0.66, 95% CI 0.53–0.82), and prostate volume > 50 vs ≤ 50 cm 3 (HR = 0.65, 95% CI 0.51–0.84) were associated with lower recovery rates, whereas high vs low testosterone was associated with a higher recovery rate (HR = 1.46, 95% CI 1.17–1.83); all P < 0.001. In landmark analysis, the high-testosterone group showed faster recovery during 0–6 weeks ( P < 0.001), whereas no significant between-group difference was observed after week 6 among patients still without complete continence and under follow-up at the landmark ( P = 0.65). Higher preoperative testosterone was associated with faster continence recovery, with the between-group difference predominantly observed during the first 6 postoperative weeks. Given the retrospective design and unavailable membranous urethral length, these findings should not be interpreted causally. They support prospective evaluation of perioperative testosterone optimization, including short-term testosterone replacement therapy in carefully selected men with confirmed low testosterone, for early continence recovery.