Comparison of Fixed and Live Cell-Based Assay for the Detection of AChR and MuSK Antibodies in Myasthenia Gravis.

ABSTRACT

BACKGROUND AND OBJECTIVES: Live cell-based assay (CBA) can detect acetylcholine receptors (AChRs) or muscle-specific tyrosine kinase (MuSK) antibodies (Abs) in a proportion of patients with radioimmunoassay (RIA)-double seronegative myasthenia gravis (dSN-MG). A commercial fixed CBA for AChR and MuSK Abs has recently become available; however, comparative studies on fixed and live CBAs are lacking. In this study, we compared the performance of fixed and live CBAs in patients with RIA-dSN MG and assessed their sensitivity in RIA-positive MG samples and their specificity. METHODS: AChR and MuSK Abs were tested in 292 serum samples from 2 Italian MG referral centers by live and fixed CBAs 192 from patients with MG and 100 from controls. All samples had been previously assessed by RIA 66 were AChR positive, 40 MuSK positive, and 86 dSN. All controls were negative. Two independent raters assessed the CBA results. Fixed and live CBAs were compared with the McNemar test; interrater and interlaboratory agreement were assessed with Cohen's kappa or interclass correlation coefficient (ICC), as appropriate. RESULTS: In 86 RIA-dSN samples, fixed CBA detected Abs in 10 cases (11.6%, 95% CI 5.7-20.3), whereas live CBA detected Abs in 16 (18.6%, 95% CI 11.0-28.5) (p = 0.0143). Of these sera, those positive by fixed CBA were also positive by live CBA. In addition, live CBA could detect MuSK Abs in 4 and AChR Abs in 2 samples that were negative by fixed CBA, providing an 8% (95% CI 2.9-16.6) further increase in the Ab detection rate. These results were confirmed by flow cytometry. In the RIA-positive cohort, the sensitivity for AChR Abs was 98.5% (95% CI 91.9%-99.9%) for fixed CBA and 100% (95% CI 94.6-100) for live CBA (p = 0.1573). For both assays, the sensitivity for MuSK Abs was 100% (95% CI 91.2-100), and the specificity was 100% (95% CI 96.4-100). Interrater agreement was almost perfect for live and fixed CBAs (Cohen's kappa 0.972 and 0.978, respectively), alike interlaboratory agreement. Interrater agreement for the CBA score ranged from good to excellent (ICC 0.832-0.973). DISCUSSION: Fixed CBA represents a valuable alternative to RIA for AChR and MuSK Ab detection in patients with MG and could be considered as a first-step diagnostic test. Live CBA can be useful in the serologic evaluation of RIA- and fixed CBA-negative samples.