OBJECTIVE: To examine whether a single repeat prostate-specific antigen (PSA) helps discriminate cancer from non-cancer-related PSA elevation. METHODS: Men aged 50-70 yr (n=54,087) in a multicentre randomised controlled trial comparing treatments for localised prostate cancer were tested. A total of 4102 (7.6%) with an initial PSA in the range of 3-19.9 ng/ml had repeat measurement (median interval: 50 d) followed by prostate biopsy. The decision to biopsy was based on the first PSA level. The outcome was the presence of prostate cancer on biopsy. RESULTS: Men with a 20% drop in PSA had a lower risk of cancer (odds ratio [OR]=0.43; 95% confidence interval [CI], 0.35-0.52; p<0.001) and high-grade cancer (OR=0.29; 95%CI, 0.19-0.44; p<0.001) compared to the rest of the cohort. The effect of percentage reduction was greater in men aged < or =60 yr than in those >60 yr. (OR for any cancer=1.6; 95%CI, 1.0-2.4; p=0.05; OR for high-grade cancer=2.9; 95%CI, 1.2-6.7; p=0.014). This equated to a risk reduction of high-grade cancer from 4% to 0.5%, 6% to 2%, and 15% to 2% in men < or =60 yr with an initial PSA of 3.0-3.99, 4.0-5.99, and > or =6 ng/ml, respectively. No level of repeat PSA confidently predicted absence of cancer. CONCLUSION: Following an initial PSA of 3.0-19.99 ng/ml in men aged 50-70 yr, repeat PSA within 7 wk allows more accurate risk prediction that may assist in the decision-making as to whether or not to proceed with prostate biopsy.
777 - 784
Aged, Biopsy, Humans, Logistic Models, Male, Middle Aged, Predictive Value of Tests, Prostate-Specific Antigen, Prostatic Neoplasms, ROC Curve, Risk Assessment