Objective: The neurological disability accumulation in patients with relapsing-onset multiple sclerosis (MS) is commonly attributed to relapse-associated worsening (RAW) and progression independent of relapse activity (PIRA). Using a mediation model, this research aimed to investigate and quantify the contributions of RAW and PIRA to the overall disability accrual.Methods: Clinical data, containing Expanded Disability Status Scale (EDSS) scores, duration, attack number, and demographics, were collected from 121 patients with relapsing-onset MS in China and included in the mediation model. Two phases were defined: an early phase, from the clinical onset to EDSS 3, and a later phase, greater than EDSS 3.Results: Clinical attack number partly mediated the relationship between duration and neurological disability (Duration-* Attack-* EDSS score) only in the early phase, with the ratio of indirect (RAW) to total effect of 0.414; while this mediator effect became negligible (<10%) in the later phase, with a predominating direct effect (PIRA). Onset age positively correlated with EDSS scores during the early stage, independent of the clinical attack number (the direct effect was significant, but the indirect effect was not), while this association was insignificant later. Besides, compared to females, male patients appeared to relapse less frequently before reaching EDSS 3 but were vulnerable to an accelerated progression after that.Conclusions: In relapsing-onset MS, PIRA is the major contributor to the irreversible disability accrual throughout the whole disease course, albeit RAW is also partly involved during the early stage. The correlations between the disabled outcome and the onset age or sex vary in different phases.