Screening for interference in immunoassays

Background: The presence of interfering substances in patient samples submitted for immunoassay cannot be reliably anticipated. We therefore evaluated three interference screening techniques and estimated the prevalence of interfering substances as defined by positive outcomes with these protocols. Methods: We evaluated 160 samples for the presence of substances that may interfere with four immunoassays (40 samples for each): thyroid-stimulating hormone, prostate-specific antigen, beta-human chorionic gonadotropin, and cortisol. Interference was defined by nonlinear responses with serial dilution, discrepant results after pretreatment with heterophile blocking reagent (HBR), and positive reactions on a mouse-antibody-negative control reaction (Tandem ICON(R) InummoConcentration HCG). Criteria for declaring significant discrepant results were based on a Z-score computed using the assay CV. The McNemar test was used to compare the prevalence of discrepancies across the three screening techniques. The association between type of immunoassay and prevalence of discrepant results was determined by a modified Pearson chi(2) statistic. Results: Five of the 160 samples [3.1%; 95% confidence interval (CI), 1.0-7.1%] screened positive with the ICON. Seventy-two of the 148 samples with informative serial dilutions (48.6%; 95% CI, 40.4-57.0%) had at least one discrepant result at higher dilutions. After pretreatment with HBR, 53 of the 140 samples (38%; 95% CI, 29.8-46%) were discrepant. Only 48 of the 140 samples With informative measurements for all three screening techniques (34%; 95% CI, 26-43%) were negative by all three. The prevalence of positive screens varied significantly by type of immunoassay (P < 0.0001) for both HBR and serial dilution. Only 3% (0.8-7%) of the samples tested with HBR showed a change from normal to abnormal or the reverse after treatment. Conclusions: Introducing a protocol based on any of these three techniques into the immunochemistry laboratory to prescreen for interfering substances is riot warranted. The evaluation of specimens for the presence of interfering anti-animal antibodies should be reserved for cases in which clinical history or suspicious results indicate the need. (C) 2003 American Association for Clinical Chemistry.