Assessing the value of clinical prognosticators of developing relapsing myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD)

Abstract

Introduction 40-60% of people experiencing a first clinical attack of myelin oligodendrocyte glycoprotein antibody associated disease (MOGAD) will develop relapsing disease (RMOGAD). The accurate risk prediction of R-MOGAD is challenging. Objective To study the value of a combined predictive model for developing R-MOGAD using clinical features described by prior research studies. Methods A combined multivariate binary logistic regression model was applied on a retrospective clinical dataset of patients with MOGAD from the Nottingham NMOSD/ MOGAD Clinic, with further validation in an independent dataset of MOGAD patients from the NMOSD Specialist Service, Walton Centre Liverpool. Results In the observational dataset (n=34), two MOG-IgG positive tests >= one year apart, age, optic neuritis and early relapses were not significant predictors of developing RMOGAD. Treatment with steroids >=10mg >= three months at disease onset was a significant negative predictor (p=0.006) of R-MOGAD, with 75% sensitivity (95%CI:48%- 93%) and 83% specificity (95%CI:59%-96%). In the independent dataset (n=68) used for validation, this predictor had a sensitivity of 55% (95%CI:36%-72%) and a specificity of 77% (95% CI:60%-90%). Conclusion Our predictive model suggests that the treatment with steroids >=10mg >=three months after the inaugural attack is a significant negative predictor of developing R-MOGAD.