Systematic Differences Between Total and Free Prostate-Specific Antigen Immunoassays: Comparison Using Passing and Bablok Regression

Recent studies have shown that there are systematic differences among total and free prostate-specificantigen (PSA) immunoassays. In this study we analyzedintermethod differences in total PSA (tPSA) and free PSA(fPSA) measurement using ARCHITECT i2000SR (Abbott Diagnostics) and COBAS E601 (Roche Diagnostics). A number of 160 blood samples were tested for tPSA and 50 samples for fPSA (selecting only sampleswith tPSA: 4.1–10.0 μg/L). Passing–Bablok regression analysis was used to compare the two analytical methods fortPSA, fPSA and percentage of fPSA (%fPSA). A strong correlation was noticed between ARCHITECT i2000SR and COBAS E601 for tPSA, fPSA and %fPSA (r between 0.94 and 0.99). Concentrations of tPSA and fPSA measured by COBAS E601 were higher thanthose measured by ARCHITECT i2000SR with a bias of 0.8 μg/L for tPSA and 0.14 μg/L for fPSA. Analyzing therelative difference between methods for fPSA and %fPSA, COBAS E601 exceed a 10% relative difference limit. Our study confirms that there are differences in measured concentrations of tPSA and fPSA byvarious commercial methods. Because clinical judgment on subsequent diagnostic procedures, such as prostatebiopsy, is based on tPSA and fPSA results, tests harmonization should be a priority.