Variation of free and total prostate-specific antigen levels: the effect on the percent free/total prostate-specific antigen.

OBJECTIVE

Recent studies have shown that calculations of the percent free/total prostate-specific antigen (PSA) improves the specificity of PSA testing. Characterizing the variability of free PSA and total PSA is necessary to evaluate the utility of an isolated free/total PSA measurement. We investigated the total variation of free and total PSA levels to determine how the percent free/total PSA was affected.

DESIGN

Serum was obtained from nine urological patients on 5 different days over a 2-week period. Free and total PSA levels were measured on the day of collection. The total variation expressed in terms of percent coefficient of variation (%CV) was calculated, and the biological variation was derived taking analytical variation into consideration.

SETTING

Patients were from Seattle (Wash) Urological Associates, and samples were processed at the Dynacare Laboratory of Pathology, Seattle, Wash.

PATIENTS

Nine men (aged 48 to 69 years) were evaluated; three had been diagnosed with prostate cancer, three with benign prostatic hyperplasia, one with chronic prostatitis, one with high-grade prostatic intraepithelial neoplasia, and one was clinically normal.

MAIN OUTCOME MEASURES

Total variation for free, total, and percent free/total PSA.

RESULTS

The average total variation was 13.9% CV, 7.5% CV, and 10.6% CV for free, total, and percent free/ total PSA, respectively. Biological variation was derived to be 13.0% CV, 5.6% CV, and 8.0% CV for free, total, and percent free/total PSA, respectively.

CONCLUSIONS

When applied, these results suggest that there are significant random changes in the numerator and denominator of the free PSA-total PSA ratio that could result in clinical misinterpretation. Clinicians must be aware that free PSA and total PSA levels will fluctuate owing to nonpathologic variation.