Serum Biomarker Profiles Discriminate AQP4 Seropositive and Double Seronegative Neuromyelitis Optica Spectrum Disorder
Sara Carta,Alessandro Dinoto,Marco Capobianco,Paola Valentino,Francesca Montarolo,Arianna Sala,Markus Reindl,Marianna Lo Re,Vanessa Chiodega,Pierre Branger,Bertrand Audoin,Jennifer Aboab,Caroline Papeix,Nicolas Collongues,Philippe Kerschen,Helene Zephir,Alain Créange,Bertrand Bourre,Kathrin Schanda,Eoin P Flanagan,Vyanka Redenbaugh,Javier Villacieros-Álvarez,Georgina Arrambide,Alvaro Cobo-Calvo,Sergio Ferrari,Romain Marignier,Sara Mariotto
DOI: https://doi.org/10.1212/NXI.0000000000200188
Abstract:Background and objectives: Glial fibrillary acidic protein (GFAP) and neurofilament light chain (NfL) serum levels are useful to define disease activity in different neurologic conditions. These biomarkers are increased in patients with aquaporin-4 antibody-positive NMOSD (AQP4+NMOSD) during clinical attacks suggesting a concomitant axonal and glial damage. However, there are contradictory results in double seronegative NMOSD (DS-NMOSD). The aim of this study was to characterize the neuronal, axonal, and glial damage of DS-NMOSD in comparison with AQP4+NMOSD. Methods: Patients with DS-NMOSD (i.e., for AQP4 and myelin oligodendrocyte glycoprotein antibodies-MOG-Abs) and age-matched AQP4+NMOSD diagnosed according to the latest diagnostic criteria and with available serum samples obtained within 3 months from onset/relapse were retrospectively enrolled from 14 international centers. Clinical and radiologic data were collected. Serum NfL, GFAP, tau, and UCH-L1 levels were determined using an ultrasensitive paramagnetic bead-based ELISA (SIMOA). Statistical analysis was performed using nonparametric tests and receiver-operating characteristic (ROC) curve analysis. Results: We included 25 patients with AQP4+NMOSD and 26 with DS-NMOSD. The median age at disease onset (p = 0.611) and female sex predominance (p = 0.072) were similar in the 2 groups. The most common syndromes at sampling in both AQP4+NMOSD and DS-NMOSD were myelitis (56% vs 38.5%) and optic neuritis (34.6% vs 32%), with no statistical differences (p = 0.716). Median EDSS at sampling was 3.2 (interquartile range [IQR] 2-7.7) in the AQP4+NMOSD group and 4 (IQR [3-6]) in the DS-NMOSD group (p = 0.974). Serum GFAP, tau, and UCH-L1 levels were higher in patients with AQP4+NMOSD compared with those with DS-NMOSD (median 308.3 vs 103.4 pg/mL p = 0.001; median 1.2 vs 0.5 pg/mL, p = 0.001; and median 61.4 vs 35 pg/mL, p = 0.006, respectively). The ROC curve analysis showed that GFAP, tau, and UCH-L1, but not NfL, values were able to discriminate between AQP4+ and DS-NMOSD (area under the curve (AUC) tau: 0.782, p = 0.001, AUC GFAP: 0.762, p = 0.001, AUC UCH-L1: 0.723, p = 0.006). NfL levels were associated with EDSS at nadir only in patients with AQP4+NMOSD. Discussion: Serum GFAP, tau, and UCH-L1 levels discriminate between AQP4+NMOSD and DS-NMOSD. The different biomarker profile of AQP4+NMOSD vs DS-NMOSD suggests heterogeneity of diseases within the latter category and provides useful data to improve our understanding of this disease.