PSA predicts development of incident lower urinary tract symptoms: results from the REDUCE study.

Abstract

The relationship between baseline prostate-specific antigen (PSA) and development of lower urinary tract symptoms (LUTS) in asymptomatic and mildly symptomatic men is unclear. We sought to determine if PSA predicts incident LUTS in these men. A post-hoc analysis of the 4-year REDUCE study was performed to assess for incident LUTS in 1534 men with mild to no LUTS at baseline. The primary aim was to determine whether PSA independently predicted incident LUTS after adjusting for the key clinical variables of age, prostate size, and baseline International prostate symptom score (IPSS). Incident LUTS was defined as the first report of medical treatment, surgery, or sustained clinically significant symptoms (two IPSS >14). Cox proportional hazards, cumulative incidence curves, and the log-rank test were used to test our hypothesis. A total of 1534 men with baseline IPSS <8 were included in the study cohort. At baseline, there were 335 men with PSA 2.5-4 ng/mL, 589 with PSA 4.1-6 ng/mL, and 610 with PSA 6-10 ng/mL. During the 4-year study, 196 men progressed to incident LUTS (50.5% medical treatment, 9% surgery, and 40.5% new symptoms). As a continuous variable, higher PSA was associated with increased incident LUTS on univariable (HR 1.09, p = 0.019) and multivariable (HR 1.08, p = 0.040) analysis. Likewise, baseline PSA 6-10 ng/mL was associated with increased incident LUTS vs. PSA 2.5-4 ng/mL in adjusted models (HR 1.68, p = 0.016). This association was also observed in men with PSA 4.1-6 ng/mL vs. PSA 2.5-4 ng/mL (HR 1.60, p = 0.032). Men with mild to no LUTS but increased baseline PSA are at increased risk of developing incident LUTS presumed due to benign prostatic hyperplasia.