Determination of prostate-specific antigen in serum by immunoradiometric assay

Abstract

Abstract A better understanding is needed of the role of pre-analytical factors if prostate-specific antigen (PSA) is to be reliably used as a tumor marker. Reports on the analytical performance of TANDEM-R PSA (Hybritech, Inc., San Diego, CA) differ considerably with respect to detection limit and imprecision, differences that might be due (e.g.) to the use of different control matrices, to between-batch variations in reagent composition, or to nonrobustness of the assay. During nine months we determined PSA in 986 serum samples from 728 male urology patients and 54 samples from women (25 assay runs). As controls we used two serum pools with low PSA concentration and two widely used commercial controls. The within-assay CV for patients' samples was similar to that found with the commercial controls: 2.6% to 3.4% in the upper part of the normal reference interval. Precision was worse at lower concentrations (CVs 5-10% at about 0.5 micrograms/L). Imprecision tended to be higher at the end of runs. Assay drift for 100-tube runs was -4%. PSA was stable at -20 degrees C during six months. Neither the polyester polymer in SST tubes nor a hemolysate had any detectable effect on PSA values. Clinical analysis of the first 322 patients and all patients with PSA less than or equal to 0.20 micrograms/L highlighted the requirements for strict adherence to sampling instructions and to stringent quality control also at low analyte concentrations (analyte-free sera and sera with PSA concentrations 0.2-0.5 micrograms/L). Values with TANDEM-R PSA and IRMA-Count PSA (Diagnostic Products Corp., Los Angeles, CA) correlated well with no difference in detection limit or with samples from women. Within-assay precision was better with IRMA-Count PSA in the upper part of the normal reference interval and above. The designs of the two assays were compared in a format that is generally applicable for immunoassay kits (NORDKEM kit group, unpublished), and subjective impressions were recorded.