Performance evaluation of a novel chemiluminescence assay for detection of anti-GBM antibodies: an international multicenter study

Background. Autoantibodies to the non-collagen region (NC1) of the alpha-3 subunit of collagen IV represent a serological hallmark in the diagnosis of Goodpasture's syndrome (GPS). The objective of our study was to carefully analyze the performance characteristics of a novel anti-glomerular basement membrane (GBM) chemiluminescence immunoassay (CIA). Methods. Sera from patients with GPS (n = 90) were collected from four clinical centers. Samples from different disease groups (n = 397) and healthy individuals (n = 400) were used as controls. All samples were tested for anti-GBM antibodies by a rapid, random access CIA (QUANTA Flash (TM) GBM). Most of the samples were also tested using other methods including different commercial anti-GBM IgG assays and research assays for anti-GBM IgA and IgM. Results. The sensitivity and specificity of the novel CIA was 95.6% [95% confidence interval (CI) 89.0-98.8%] and 99.6% (95% CI 98.9-99.9%), respectively. Receiver operating characteristic analysis showed good discrimination between GPS patients and controls. The area under the curve was 0.98 (CI 0.96-1.0). The three anti-GBM antibody-positive samples from the control group were from two healthy individuals and one human immunodeficiency virus (HIV)-infected patient. All three individuals had low levels of anti-GBM antibodies [20, 24 and 25 chemiluminescent unit (CU), cutoff 20 CU]. When the results of the new CIA were compared to other methods, good agreement was observed: 95.8% (kappa = 0.92) versus EliA (TM) GBM, 97.4% (kappa = 0.95) versus both BINDAZYME (TM) Anti-GBM and QUANTA Lite (R) GBM. Anti-GBM IgA was detectable in low concentrations in patients with GPS and was associated with anti-GBM IgG but was less useful in discriminating GPS patients and controls. No discrimination was found for anti-GBM IgM. Conclusion. The novel QUANTA Flash (TM) GBM CIA demonstrated good sensitivity and specificity and had good agreement with other methods. Our data confirm that similar to 5% of patients with GPS do not have detectable levels of anti-GBM antibodies.