Serum sex steroids depict a nonlinear u-shaped association with high-risk prostate cancer at radical prostatectomy.

Abstract

To assess the association between preoperative serum total testosterone (tT), 17β-estradiol (E(2)), sex hormone-binding globulin (SHBG), and tT-E(2) ratio values with high-risk prostate cancer (as defined by the National Comprehensive Cancer Network practice guidelines) at radical prostatectomy. Serum E(2), tT, and SHBG were dosed the day before surgery (7:00-11:00 am) in a cohort of 724 candidates to radical prostatectomy. Restricted cubic spline functions tested the association between predictors (i.e., model 1: age, body mass index, and serum tT, E(2), and SHBG levels; model 2: tT-E(2) values instead of tT and E(2) levels) and high-risk prostate cancer. Low-, intermediate-, or high-risk prostate cancer was found in 251 (34.7%), 318 (43.9%), and 155 (21.4%) patients, respectively. Patients in the high-risk class showed the lowest tT, E(2), and tT-E(2) ratio values (all P ≤ 0.02). At univariate analysis, only age, tT, E(2), and tT-E(2) ratio values were significantly associated with high-risk prostate cancer (all P ≤ 0.006). At multivariate analyses considering model 1 variables, age (P = 0.03), serum tT (all P < 0.001), and E(2) (all P ≤ 0.01) were associated with high-risk prostate cancer; only tT-E(2) ratios achieved independent predictor status for high-risk prostate cancer (all P < 0.001) when considering model 2. Both the lowest and the highest tT, E(2), and tT-E(2) values depicted a nonlinear U-shaped significant association with high-risk prostate cancer. These data showed that preoperative serum sex steroids are independent predictors of high-risk prostate cancer, depicting a nonlinear U-shaped association.