Purpose

In the last decade numerous groups have shown that low levels of pretreatment serum total testosterone consistently predict more aggressive disease, worse prognosis and worse treatment response in patients with metastatic prostate cancer. Prior studies have not demonstrated this same correlation in patients with known localized disease. We rigorously tested pretreatment total testosterone levels as a potential staging and prognostic marker in a large cohort of 879 patients with localized cancer treated with radical prostatectomy. Materials and Methods

We retrospectively reviewed the clinical records of 879 patients treated with radical prostatectomy between January 1, 1986 and June 30, 2002 from 9 hospital sites. Nonparametric tests were used to compare the relationship of pretreatment testosterone to other variables. Multivariate logistic regression analysis was used to assess clinical predictors of extraprostatic disease. Kaplan-Meier survival methods and Cox regressionanalysis were used to assess predictors of biochemical recurrence. Results

Patients with nonorgan confined prostate cancer (pT3–T4) showed significantly lower pretreatment total testosterone levels than those with organ confined cancer (pT1–T2) (nonparametric p =0.041). In multivariate analysis pretreatment total testosterone emerged as a significant independent predictor of extraprostatic disease (p =0.046). Total testosterone was not a significant predictor of biochemical (prostate specific antigen) recurrence (p =0.467). Conclusions

Pretreatment total testosterone was an independent predictor of extraprostatic disease in patients with localized prostate cancer. As testosterone decreases patients have an increased likelihood of nonorgan confined disease. Low testosterone was not predictive of biochemical recurrence, although trends observed dictate study in larger cohorts with mature followup.