Broadening the clinical concept of progression independent of relapses in multiple sclerosis through multidimensional patient-reported and self-administered outcomes

Abstract

Identifying progression independent of relapse activity (PIRA), the main mechanism of disability accumulation in multiple sclerosis (MS), has important prognostic implications. Its detection using the Expanded Disability Status Scale (EDSS) is constrained by limited sensitivity and availability. Multidimensional patient-reported/-administered outcomes (PROs) have greater sensitivity than the EDSS and can be more easily obtained in the clinic, but have never been used to define PIRA. We aimed to investigate PIRA through PROs and explore, through MRI, its pathological correlates and main predictors.

All MS patients with ≥3 standardized clinical assessments and >24 weeks follow-up from the multicenter MS PATHS cohort were eligible for inclusion. Progression was defined on the patient-determined disease steps (PDDS), a self-reported measure equivalent to the EDSS, and the self-administered walking speed (WST), manual dexterity (MDT), and processing speed (PST) tests. PIRA was defined as 24-week confirmed progression in the absence of self-reported relapses in the previous 12 months. T2 lesion volume (T2LV) and brain parenchymal fraction (BPF) were obtained from routine 3T MRI scans. We built adjusted linear mixed-effects and Cox proportional hazards models.

We included 10,832 patients (73% female; age 47·6 years; disease duration 14·9 years). Over a mean period of 2·9 years, 2,357 patients (26%) presented PIRA on any outcome, of which 1,216 (52%) were captured by the PDDS (‘PIRA-PDDS’). Of the remaining 48%, 617 (26%) were captured by the WST or MDT (‘PIRA-motor’) and 524 (22%) were captured only by the PST (‘PIRA-cognition’). Compared to no-progressors, PIRA-PDDS and PIRA-cognition had significantly accelerated atrophy (p≤·023). PIRA-motor and PIRA-cognition had accelerated T2LV increase (p≤·042). PIRA was predicted by older age at first attack (p<·05), longer disease duration (p<·05), and lower BPF (p<·001) at baseline.