Physiologic (intraindividual) variation of serum prostate-specific antigen in 814 men from a screening population

Objectives. We measured the intraindividual variation of prostate-specific antigen (PSA) in the serum of healthy men screened for prostate cancer. Methods. We used a fully automated PSA assay system (ACS: 180 assay) to evaluate a screening population of 814 men (mean age, 63.3 years; range, 50 to 79 years) without documented prostate cancer or prostate surgery. A second blood sample was drawn 15 to 183 days after the first specimen (mean, 80 days). Results. In the ACS PSA ranges of 0 to 7.2 ng/mL, 7.3 to 17.9 ng/mL, and 18.0 ng/mL or greater (0 to 4 ng/mL, 4 to 10 ng/mL, and 10.0 ng/mL or greater by the Tandem-R assay), the mean coefficient of variation of the first and second blood drawn was 20%, 12%, and 10%, respectively. In 435 men whose first blood samples were measured twice for PSA difference (interassay or run-to-run variation), the intraindividual variation in the range of 0 to 7.2 ng/mL was significantly larger than the interassay variation, which was also true for the 7.3 to 17.9 ng/mL range. In the range of 0 to 7.2 ng/mL, 251 of 695 (36%) showed a 20% or greater relative increase and 69 of 695 (10%) showed a 1.3 ng/mL (0.75 ng/mL by the Tandem-R assay) or greater absolute increase of PSA at the second blood sample. Conclusions. We conclude that in the low ranges of PSA concentrations, one should consider the possibility of substantial intraindividual variation when interpreting serial PSA measurements.