Smoking reduces PSA accuracy for detection of prostate cancer: results from an Italian cross-sectional study.

Background: The aim of our study is to evaluate the diagnostic accuracy of prostate-specific antigen (PSA) according to the smoking status in a cohort of European men undergoing prostate biopsy.

Methods: From 2008 onwards, we consecutively enrolled, at a single institution in Italy, men undergoing 12 core transrectal ultrasound-guided prostate needle biopsy. Demographic, clinical and histopathological data were collected. We excluded men who had PSA >30 ng/mL. Patients were classified in three groups: non-smokers, smokers and former smokers. Receiver-operator characteristics (ROC) curve analysis were used to compare predictive properties of PSA across smoking categories for the final histopathological diagnosis of prostate cancer.

Results: 872 patients were enrolled. with a median age and PSA of 67 years (IQR: 61/74) and 6.2 ng/mL (IQR: 4.4/9.7) respectively.402/872 patients (46%) were non-smokers; 151 of 872 were smokers (17%) and 319 of 872 were former smokers (36%); 374 of 872 (43%) had cancer on biopsy. PSA accuracy for smokers (AUC= 0.47, P=0.490) was lower when compared to non-smokers (AUC=0.59, P=0.003) and former smokers (AUC=0.59, P=0.005).

Conclusions: In a cohort of Italian men undergoing prostate biopsy, the performance accuracy of PSA, as a predictor of prostate cancer, is lower in smokers. Although the molecular link behind our findings is still unknown, our study firstly showed that the PSA accuracy for PCa diagnosis is significantly influential by smoking.