Inter-Laboratory Validation of Nodal/Paranodal Antibody Testing.

BACKGROUND AND AIMS

Reliable detection of antibodies against nodal targets is vital for the diagnosis of autoimmune nodopathies. The performance characteristics of recently developed in-house assays are unknown. We compared testing at four centres.

METHODS

Each submitted 29-40 serum samples to a coordinating centre from one of three groups: (1) autoimmune nodopathy patients, with positive nodal/paranodal antibodies; (2) seronegative patients with other inflammatory neuropathies, and (3) healthy individuals or those with other neurological diseases. The coordinating centre recoded all samples and returned 160 identical aliquots to each testing centre for blinded testing. Once data from all centres had been received by the coordinating centre, unblinded results were returned for analysis. Sensitivity was defined by the proportion of group 1 samples returned as positive. Accuracy was defined as 0.075(sensitivity) + 0.925(specificity).

RESULTS

Centres performed various combinations of ELISA, cell-based (CBAs) and teased-nerve fibre assays. All labs produced highly accurate results (96%-100%) and concordance for the overall result across at least 3 or all 4 test centres was observed for 98% and 89% of the samples respectively. However, 10/30 individual assays (6/14 CBAs and 4/16 ELISAs) were less than 90% sensitive. Only 3 assays had more than 1 false positive result (2 ELISAs and 1 CBA). Combining different assay modalities to produce an overall result did not improve accuracy. Inter-laboratory consistency in the determination of antibody subclasses was poor.

INTERPRETATION

Although most samples were correctly categorised in all 4 centres, the use of a specific test modality or multiple tests did not guarantee accuracy. Early and repeated interlaboratory testing with sharing of samples is important to understand test performance and reproducibility, identify areas for improvement and maintain consistency. To aid this, we provide detailed methods for the best performing tests. Further standardisation of antibody subclass determination is required.