Disability patterns in multiple sclerosis: A meta-analysis on RAW and PIRA in the real-world context.

Objective: To summarize the current evidence on relapse-associated worsening (RAW) and progression independent of relapse activity (PIRA) through a quantitative synthesis of real-world studies.

Methods: Scientific databases were searched to identify suitable articles. Random-effects meta-analyses, subgroup analyses and meta-regression models were ran to provide pooled estimates of RAW and PIRA events and to identify their potential moderators (PROSPERO registration: CRD42024503895).

Results: Eighteen articles met the eligibility criteria, with a pooled sample size of 52,667 patients (93% relapsing-remitting, 6% clinically isolated syndrome and 1% progressive) followed for 2.4 to 12.1 years, yielding to 415,825 patient-years. Pooled event rates for RAW and PIRA were 1.6 (95 confidence interval (CI) = 1.1-2.1) and 3.1 (95% CI = 2.3-3.9) per 100 patient-years, respectively. Less RAW events were found in cohorts including patients with progressive course (β = -0.069, p = 0.006) and under high-efficacy disease-modifying treatments (DMTs) (β = -0.031, p = 0.007), while PIRA events were directly related to older age (β = 0.397, p = 0.027). In addition, we found significant differences in PIRA event rates according to the criteria adopted to define confirmed disability accrual (p < 0.05).

Discussion: PIRA accounts for most events causing disability accumulation in the real-world setting, even at the earlier disease stages, whereas RAW represents a less frequent phenomenon, likely due to effective treatments. The detection and statistical analysis of PIRA outcomes pose challenges, raising the risk of erroneous inference. When interpreting our findings, caution is needed given the wide heterogeneity of included studies.