Predictors of subsequent relapse-independent progression after high-efficacy therapy in relapsing multiple sclerosis.

Background and objectives: Progression independent of relapse activity (PIRA) is increasingly recognized as a major driver of disability accumulation in multiple sclerosis (MS). However, the factors associated with PIRA occurring after initiation of high-efficacy therapies (HETs) remain poorly investigated. We aimed to identify clinical predictors of subsequent PIRA following first HET initiation. Methods: We retrospectively analyzed data from two prospectively maintained Italian MS registries. Adults with relapsing MS who initiated a HET and had at least one year of post-treatment follow-up were included. Cox proportional hazards models were used to identify predictors of PIRA after first HET initiation. A primary multivariable model included age at disease onset, sex, baseline Expanded Disability Status Scale (EDSS) score, time from symptom onset to HET initiation, and annualized relapse rate (ARR) before HET exposure. A secondary model additionally adjusted for baseline brain MRI characteristics. Exploratory analyses included risk stratification according to age and treatment timing, strategy-specific analyses, comparison of predictors of PIRA and relapse-associated worsening (RAW), risk of PIRA across first HET classes, and sensitivity analyses. Results: Among 466 patients with relapsing MS (71.5% female, median age at onset 32.2 years), 70 (15.0%) developed PIRA after first HET initiation. In the primary multivariable model, older age at onset (HR 1.07, 95% CI 1.05-1.10; p<0.001) and longer time from symptom onset to HET initiation (HR 1.10, 95% CI 1.05-1.16; p<0.001) independently predicted PIRA. These associations remained consistent in the complete-case MRI-adjusted model including baseline brain MRI characteristics (n=269). Patients with younger age at onset and HET initiation within two years of symptom onset had the lowest risk of PIRA. In exploratory analyses, older age at onset and delayed HET initiation predicted PIRA but not RAW, whereas higher pre-treatment relapse activity predicted RAW but not PIRA. Findings were consistent across strategy-specific, on-treatment, landmark, EDSS re-baselining, and calendar-year-adjusted analyses. Conclusions: Older age at disease onset and delayed HET initiation were the most consistent predictors of PIRA. Our findings support the importance of early HET and suggest that clinically defined PIRA and relapse-associated worsening have partially distinct predictor profiles, while not excluding a contribution of subclinical MRI inflammatory activity to PIRA events.