Serum PSA as a Predictor of Testosterone Deficiency

Abstract

Introduction The relationship between serum prostate-specific antigen (PSA) and testosterone (T) levels is still controversial. According to the “saturation hypothesis,” a significant relationship is apparent only in the low T range. Aim To verify whether, in a large sample of male subjects seeking medical care for sexual dysfunction (SD), PSA might represent a reliable marker of T levels. Methods A consecutive series of 3,156 patients attending our unit for SD was studied. Among them, only subjects without history of prostate disease and with PSA levels <4 ng/mL (N = 2,967) were analyzed. Main Outcome Measures Several hormonal and biochemical parameters were studied, along with structured interview on erectile dysfunction (SIEDY), ANDROTEST, and PsychoANDROTEST. Results Receiver operating characteristic curve analysis for predicting severe hypogonadism (T < 8 nmol/L) showed an accuracy of PSA = 0.612 ± 0.022 (P < 0.0001), with the best sensitivity and specificity at PSA < 0.65 ng/mL (65.2% and 55.5%, respectively). In the entire cohort, 254 subjects (8.6%) showed T < 8 nmol/L and, among them, more than half (N = 141, 4.8%) had PSA < 0.65 ng/mL. After adjusting for age, low PSA was associated with hypogonadism-related features (i.e., delayed puberty, lower testis volume) and associated conditions, such as metabolic syndrome (hazard ratio [HR] = 1.506 [1.241–1.827]; P < 0.0001), type 2 diabetes (HR = 2.044 [1.675–2.494]; P < 0.0001), and cardiovascular diseases (HR = 1.275 [1.006–1.617]; P = 0.045). Furthermore, low PSA was associated with impaired sex- and sleep-related erections. The association between low PSA and hypogonadal symptoms and signs as well as with metabolic syndrome was retained even after adjusting for T levels. Sensitivity and positive predictive values of low PSA increased, whereas specificity and negative predictive value decreased as a function of age. Conclusions PSA is a marker of T concentrations and it may represent a new tool in confirming hypogonadism. The determination of PSA levels might give insights not only on the circulating levels of total T but also on its active fractions.

Testosterone deficiency in the aging male

“…More recently, the 'saturation model' emerged [Morgentaler and Traish, 2009], establishing that prostate tissue, benign or malignant, is only sensitive to serum levels of testosterone below a certain threshold, above which prostate tissue behaves irrespective of androgen levels. Other data have confirmed these findings and propose a threshold level of approximately 8.6 mmol/l [Rastrelli et al 2013]. Several studies also support the use of testosterone in men 'at risk' for prostate cancer without increasing their baseline odds of prostate cancer development [Roddam et al 2008;Muller et al 2012;Cui et al 2014].…”

Ursodeoxycholic acid alleviates cholestasis-induced histophysiological alterations in the male reproductive system of bile duct-ligated rats

“…PSA is a serine protease produced from the prostate gland under androgenic control, and has been postulated as an effective predictor of prostate volume [49], a marker of testestoterone concentrations, and a valuable tool in confirming hypogonadism [50]. The present study revealed that BDL rats had smaller prostate glands and lower serum PSA levels in comparison to sham ones.…”

Testosterone and glucose metabolism in men: current concepts and controversies

“…More recently, prostate-specific antigen (PSA) has been proposed as a tissue marker of testosterone deficiency: in a large cohort of O3000 men, a PSA of !0.65 mg/l predicted a total testosterone of !8 nmol/l with modest sensitivity (65%) and specificity (56%). A low PSA was associated with delayed puberty, lower testicular volume and interestingly, with the presence of the metabolic syndrome (hazard ratio 1.5 (1.2-1.8), P!0.001) and type 2 diabetes (hazard ratio 2.0 (1.7-2.5), P!0.001) (Rastrelli et al 2013). In a cohort of 138 men with type 2 diabetes, free testosterone levels were inversely associated with bone density at some but not all sites (Dhindsa et al 2007).…”