Background: The Expanded Disability Status Scale (EDSS) is widely used to rate multiple sclerosis (MS) disability, but lack of disease duration information limits utility in assessing severity. EDSS ranking at specific disease durations was used to devise the MS Severity Score, which is gaining popularity for predicting outcomes. As this requires validation in longitudinal cohorts, we aimed to assess the utility of EDSS ranking as a predictor of 5-year outcome in the MSBase Registry. Methods: Rank stability of EDSS over time was examined in the MSBase Registry, a large multicentre MS cohort. Scores were ranked for 5-year intervals, and correlation of rank across intervals was assessed using Spearman's rank correlation. EDSS progression outcomes at 10 years were disaggregated by 5-year EDSS scores. Results: Correlation coefficients for EDSS rank over 5-year intervals increased with MS duration: years 1-6=0.55, years 4-9=0.74, years 7-12=0.80 and years 10-15=0.83. EDSS progression risk at 10 years after onset was highly dependent on EDSS at 5 years; one-point progression risk was greater for EDSS score of >2 than ≤2. Two-point progression was uncommon for EDSS score of <2 and more common at EDSS score of 4. Conclusions: EDSS rank stability increases with disease duration, probably due to reduced relapses and less random variation in later disease. After 4 years duration, EDSS rank was highly predictive of EDSS rank 5 years later. Risk of progression by 10 years was highly dependent on EDSS score at 5 years duration. We confirm the utility of EDSS ranking to predict 5-year outcome in individuals 4 years after disease onset.