Preoperative serum testosterone level as an independent predictor of treatment failure following radical prostatectomy.

OBJECTIVES Preoperative low serum testosterone (TS) level has been reported to be associated with adverse pathologic results in patients with clinically localized prostate cancer (pCA) treated with radical prostatectomy (RP). However, prior studies failed to show prognostic impact of preoperative low TS in these patients. The aim of this study was to investigate the relationship between preoperative TS and prostate-specific antigen (PSA) failure in these patients. METHODS Of 304 patients diagnosed with clinically localized pCA who had been treated with RP alone, 272 patients whose preoperative TS level had been measured were eligible for this analysis. Postoperative TS levels were also available in 222 of the 272 patients. Cox proportional hazard model was used to elucidate factors predictive for PSA failure. RESULTS Of the 272 patients 49 had low (< 300 ng/dl) and 223 had normal preoperative TS level. In a stepwise multivariate analysis, preoperative TS (p = 0.021) was an independent and significant predictor of PSA failure along with RP Gleason score (p = 0.006), surgical margin status (p = 0.0001), and PSA (p = 0.0001). Five-year PSA failure-free survival rate of the patients with preoperative low TS (67.8%) was significantly worse than that with normal TS (84.9%) (p=0.035). Serum TS levels increased significantly after RP (p < 0.0001). The increment of TS level in preoperative low TS group was significantly greater than that in preoperative normal TS group (p = 0.0003). CONCLUSIONS The current results demonstrated that preoperative TS level is an independent and significant predictor of PSA failure after RP in patients with clinically localized pCA.