Background: A delayed onset of treatment effect, termed therapeutic lag, may influence the assessment of treatment response in some patient subgroups. Objectives: The objective of this study is to explore the associations of patient and disease characteristics with therapeutic lag on relapses and disability accumulation. Methods: Data from MSBase, a multinational multiple sclerosis (MS) registry, and OFSEP, the French MS registry, were used. Patients diagnosed with MS, minimum 1 year of exposure to MS treatment and 3 years of pre-treatment follow-up, were included in the analysis. Studied outcomes were incidence of relapses and disability accumulation. Therapeutic lag was calculated using an objective, validated method in subgroups stratified by patient and disease characteristics. Therapeutic lag under specific circumstances was then estimated in subgroups defined by combinations of clinical and demographic determinants. Results: High baseline disability scores, annualised relapse rate (ARR) ⩾ 1 and male sex were associated with longer therapeutic lag on disability progression in sufficiently populated groups: females with expanded disability status scale (EDSS) < 6 and ARR < 1 had mean lag of 26.6 weeks (95% CI = 18.2–34.9), males with EDSS < 6 and ARR < 1 31.0 weeks (95% CI = 25.3–36.8), females with EDSS < 6 and ARR ⩾ 1 44.8 weeks (95% CI = 24.5–65.1), and females with EDSS ⩾ 6 and ARR < 1 54.3 weeks (95% CI = 47.2–61.5). Conclusions: Pre-treatment EDSS and ARR are the most important determinants of therapeutic lag.