Baseline predictors of cross-sectional and longitudinal performance on the symbol digit modalities test in individuals with multiple sclerosis

Background: Cognitive impairment occurs frequently in persons with multiple sclerosis (PwMS) at some point in the course of the disease. However, not all PwMS develop cognitive difficulties suggesting a role for important moderating factors. We examined baseline predictors of cross-sectional and longitudinal change in cognitive performance in PwMS. Methods: 680 PwMS enrolled in the Comprehensive Longitudinal Investigation of Multiple Sclerosis at the Brigham and Women's Hospital who completed the Symbol Digit Modalities Test (SDMT), a brief measure of speed of information processing, at least twice during a 10-year period were identified. Potential baseline demographic (age, education, and sex), clinical (disability, disease duration, and disease category), and patientreported outcome (PRO) (fatigue, depression, and quality of life) predictors were examined in cross-sectional analyses using linear regression and in longitudinal analyses using linear mixed effects models. Results: In cross-sectional analyses, age, disease duration, and disability each showed associations with SDMT. Group differences were observed between females and males, subjects with and without college degrees, and subjects with relapsing and progressive MS. All PRO measures showed associations with SDMT, and the strongest association was with fatigue. In the longitudinal model, increased baseline age and increased baseline disability were each associated with a greater decline in SDMT performance. None of the baseline PROs were associated with longitudinal change in SDMT. Conclusion: We observed strong associations between baseline demographic, clinical, and PRO measures and concurrent SDMT, but more limited associations between these measures and longitudinal change in SDMT.