Abstract
Neuromyelitis optica spectrum disorder is a rare autoimmune inflammatory demyelinating disease that must be differentiated from multiple sclerosis. The impact of misclassification on these patients in Taiwan remains unclear. We conducted a hospital-based retrospective cohort study of neuromyelitis optica spectrum disorder patients using the Chang Gung Research Database from 2005 to 2021. This cohort included diagnostic serostatus data from anti-Aquaporin 4 antibody testing, providing insights into incidence, demographics, annual relapse rate, and initial misclassification as multiple sclerosis. To extend these findings, we applied the same methodology to a nationwide neuromyelitis optica spectrum disorder cohort (2006-2020) using Taiwan's National Health Insurance Research Database, which does not include aquaporin 4 serostatus information; thus, the data made it impossible to distinguish seropositive from seronegative nationwide neuromyelitis optica spectrum disorder cases. Misclassification was calculated as the ratio of nationwide neuromyelitis optica spectrum disorder patients initially diagnosed with multiple sclerosis (MS) to the total nationwide neuromyelitis optica spectrum disorder cases in each cohort. In the Chang Gung Research Database cohort, we identified 193 seropositive nationwide neuromyelitis optica spectrum disorder patients, including four (2.1%) paediatric cases. Hospital incidence increased from 0.08 to 1.19 per 100 000 persons (2006-2021). Comorbidities were present in 32% of adults, with Sjögren's syndrome as the most common (21.2%). Median annualized relapse rate in adults was 0.49 (range, 0.26-0.85) and was higher in the first 3 years post-diagnosis (0.67 versus 0.00; P < 0.01). The National Health Insurance Research Database cohort included 1892 neuromyelitis optica spectrum disorder patients, 92 (4.9%) paediatric. Incidence in adults rose from 0.25 to 0.84 per 100 000 persons, with a prevalence of 8.01 per 100 000 in 2020. Median annualized relapse rate was 0.36 (range, 0.07-60.83) for adults and 0.44 (range, 0.07-6.40) for paediatric patients. Misclassification as multiple sclerosis occurred in 55.0% Chang Gung Research Database and 63.5% National Health Insurance Research Database cohort. Misclassified patients were more likely to be younger, female, experience delayed antibody testing, have frequent hospitalizations, and suffer more relapses. Both hospital-based and nationwide cohorts revealed increasing neuromyelitis optica spectrum disorder incidence and prevalence in Taiwan. This early-phase elevation in relapse activity highlights the critical therapeutic window where intervention may most effectively prevent long-term disability. Greater clinical awareness and close follow-up may improve neuromyelitis optica spectrum disorder diagnosis and management.