Discordant performance of assays for free and total prostate-specific antigen in relation to the early detection of prostate cancer

Objective To assess the value of applying rigid threshold values in interpreting prostate specific antigen (PSA) results, by selecting and comparing five current methods for measuring free and total PSA. Materials and methods Samples taken from an ongoing screening study for prostate cancer (total PSA by Tandem-E assay, 17 334 participants; biopsy criterion a PSA of 3.0 mug/L, 4 464 men) from men with a total PSA of 1.0-6.0 mug/L were measured for free and total PSA using the Access, Immulite, Elecsys and Prostatus analysis kits, in two patient groups, i.e. with prostate cancer or no evidence of disease. Results Both patient groups had equal means for total PSA but not for free PSA. In all, 360 samples from men with cancer and 96 from men with no evidence of disease were analysed. All methods applied to both groups deviated statistically significantly from the Tandem-E result for total PSA, except for the Access kit. There was a close correlation among all the methods (correlation coefficients of 0.89-0.97). There were very discordant results for the combination of the Tandem-E vs Prostatus (8% difference), representing 315 participants at a threshold of 3.0 mug/L. For free PSA (free/total PSA) the situation was worse, with extreme differences of 32% and 36% for both patient groups (Elecsys vs Access). Conclusions Depending on the threshold value applied as an indication for biopsy, when using the total PSA alone or combined with the free/total PSA, care is needed in interpreting patient groups because of the discordance among PSA assays.