Prostate specific antigen variation in patients without clinically evidentprostate cancer

Purpose: We address long-term within individual variation of serum prostate specific antigen (PSA) in men without clinical or biopsy evidence of prost ate cancer. Materials and Methods: We studied 943 men from a prostate cancer screening program with 2 PSA (PSA1 and PSA2) measurements available. A third PSA (PSA S) was obtained from 571 men. Only participants with no clinical evidence o f cancer were included in the study. Within individual PSA variability was calculated based on indexes of percent coefficient of variation, ratio diff erence and PSA velocity. The relationship among these indexes, interval bet ween measurements and number of PSA samples was assessed. Results: Mean interval was 670.4 days between PSA1 and PSA2, and 801.8 days between PSA2 and PSA3 (p <0.001). Mean coefficient of variation was 18% af ter 2 and 15.7% after 3 PSA measurements. Mean ratio differences were -0.04 7 ng./ml. for 2 and 0.033 ng./ml. for 3 samples. Mean PSA velocity was -0.1 28 ng./ml. per year for 2 and -0.055 ng./ml. per year for 3 samples, with 9 5% confidence intervals of 0.634 and 0.315, respectively. Variability was h igher if only 2 PSA measurements were done (p <0.001). No clear relationshi p was found between individual variability and interval between measurement s. Conclusions: PSA velocity is within normal limits in almost all men (more t han 95%) without clinically relevant prostate cancer. PSA individual variab ility is not fully dependent on the time between measurements when interval s are long, and can be substantially decreased with a third PSA sample.