A multicentre, prospective, double-blind study comparing the accuracy of autoantibody diagnostic assays in myasthenia gravis: the SCREAM studyResearch in context

Background: Laboratory determination of autoantibodies against acetylcholine receptor (AChR), muscle-specific kinase (MuSK) and other autoantigens have been integrated into the diagnosis of myasthenia gravis (MG). However, evidence supporting the selection of methodologies is lacking. Methods: In this prospective, multicentre cohort study, we recruited patients with suspected MG to evaluate the diagnostic accuracy of cell-based assay (CBA), radioimmunoprecipitation assay (RIPA) and enzyme-linked immunosorbent assay (ELISA) in detecting AChR and MuSK autoantibodies. This study is registered with www.clinicaltrials.gov, number NCT05219097. Findings: 2272 eligible participants were recruited, including 2043 MG, 229 non-MG subjects. AChR antibodies were detected in 1478, 1310, and 1280 out of a total of 2043 MG patients by CBA, RIPA, and ELISA, respectively; sensitivity, 72.3% (95% CI, 70.3–74.3), 64.1% (95% CI, 62.0–66.2), 62.7% (95% CI, 60.5–64.8); specificity, 97.8% (95% CI, 95.0–99.3), 97.8% (95% CI, 95.0–99.3), 94.8% (95% CI, 91.9–97.7). MuSK antibodies were found in 59, 50, and 54 from 2043 MG patients by CBA, RIPA and ELISA, respectively; sensitivity, 2.9% (95% CI, 2.2–3.7), 2.4% (95% CI, 1.8–3.2), 2.6% (95% CI, 2.0–3.4); specificity, 100% (95% CI, 98.4–100), 100% (95% CI, 98.4–100), and 99.1% (95% CI, 96.9–99.9). The area under the curve of AChR antibodies tested by CBA was 0.858, and there were statistical differences with RIPA (0.843; p = 0.03) and ELISA (0.809; p