Exploring the performance of outcome measures in MS for predicting cognitive and clinical progression in the following years

Abstract

The demand for better outcome measures in multiple sclerosis (MS) management has been increasingly recognized. Nevertheless, the prognostic impacts of available outcome measures for long-term clinical and especially cognitive disability have not been thoroughly investigated. We, therefore, aimed to explore the sustainability and long-term predictive value of outcome measures in MS. We studied a cohort of 42 relapsing-remitting MS patients and 30 healthy subjects. Evaluations were performed at baseline and after two (Y2) and six years (Y6), and included neurological and neuropsychological evaluation (BRBN), MRI (3T), and quality of life assessment. Combined clinical and cognitive measures were evaluated, such as minimal and no evidence of disease activity (MEDA and NEDA, respectively). We performed logistic regression with bootstrapping and calculated the diagnostic properties to identify patients who reached six-year clinical and/or cognitive worsening. NEDA status was observed in up to 30.8% of patients at Y2, but only in 5% at Y6, and did not preclude cognitive decline (SDMT and BRBN). The absence of MRI activity and MEDA status at Y2 were associated with less EDSS worsening in the following years but without impact on cognition. The absence of deterioration on combined clinical/cognitive measures at Y2 (e.g., T25W+ 9HPT+BRBN) was associated with better outcomes in the following years (clinical and cognitive), with moderate to large effect sizes. For the identification of clinical worsening at Y6, best accuracies were found for MEDA (70.6%), and clinical worsening (71.4%), but only MEDA remained in the final model after multivariable logistic regression analysis (OR=6.81, p=0.017). For combined clinical and cognitive worsening at Y6, only T25W+ 9HPT+BRBN remained in the final model (OR=8.5, p=0.017). Early MS inflammatory disease activity is associated with future clinical disability. Nevertheless, NEDA was difficult to sustain in the long-term and did not preclude cognitive deterioration. Clinical and cognitive measures combined predicted outcomes better than each one isolated. Our data suggest that the evaluation of more than one cognitive domain yields a better predictive outcome measure.