Abstract
The aim of this study is to assess the ability of serum prostate-specific antigen (PSA) to predict prostate volume (PV) and lower urinary tract symptoms (LUTS) represented by the international prostate symptom score (IPSS). From January 2001 to December 2011, data were collected from men who first enrolled in the Korean Prostate Health Council Screening Program. Patients with a serum PSA level of >10 ng ml(-1) or age <40 years were excluded. Accordingly, a total of 34 857 men were included in our study, and serum PSA, PV and the IPSS were estimated in all patients. Linear and age-adjusted multivariate logistic analyses were used to assess the potential association between PSA and PV or IPSS. The predictive value of PSA for estimating PV and IPSS was assessed based on the receiver operating characteristics-derived area under the curve (AUC). The mean PV was 29.9 ml, mean PSA level was 1.49 ng ml(-1) and mean IPSS was 15.4. A significant relationship was shown between PSA and PV, and the IPSS and PSA were also significantly correlated after adjusting by age. The AUCs of PSA for predicting PV >20 ml, >25 ml and >35 ml were 0.722, 0.728 and 0.779, respectively. The AUCs of PSA for predicting IPSS >7, >13 and >19 were 0.548, 0.536 and 0.537, respectively. Serum PSA was a strong predictor of PV in a community-based cohort in a large-scale screening study. Although PSA was also significantly correlated with IPSS, predictive values of PSA for IPSS above the cutoff levels were not excellent. Further investigations are required to elucidate the exact interactions between PSA and LUTS and between PSA and PV in prospective controlled studies. Such studies may suggest how PSA can be used to clinically predict PV and the IPSS.