PND18 - PROGRESSION OF DISABILITY IN PRIMARY PROGRESSIVE MULTIPLE SCLEROSIS: A RETROSPECTIVE COHORT STUDY USING DATA FROM THE MSBASE REGISTRY AND A CONTEXTUALISATION WITH AN EXISTING NATURAL HISTORY DATASET

We explored the natural history of primary progressive multiple sclerosis (PPMS) in the MSBase registry. We estimated the transition probabilities between Expanded Disability Status Scale (EDSS) states for ocrelizumab’s PPMS cost-utility model. We contextualized results with published data from relapsing MS (RMS). Retrospective cohort study including adult patients with PPMS from MSBase. We used the Palace et al. (2014) approach for a continuous time multi-state model. EDSS states were grouped and we considered the longest untreated period of patients with at least two EDSS measurements. Deaths not due to MS were censored. Unadjusted and adjusted models were computed and selection made based on Akaike information criterion (AIC). In parallel, we conducted a literature review of natural history of RMS. We extracted 1,079 PPMS patients from MSBase on 10-Dec-2016. Mean (SD) age was 51.0 (10.2) years, follow-up was 7.3 (4.7) years and 55% were female. Mean (SD) time since onset of MS symptoms was 8.2 (7.6) years and median score at first eligible EDSS was 4. We produced transition matrices containing the probabilities of patients moving between EDSS states over one year and contrasted results with the published RMS data by Palace et al. (e.g. EDSS 2-3, 5-6 and 8-9 transitions were 0.30, 0.38 and 0.06 in MSBase PPMS, vs. 0.17/0.17, 0.27/0.28, 0.06/0.02 in Palace et al. RMS for age at onset <28years and ≥28years, respectively). Our analysis is in line with our understanding from the review: PPMS patients experience faster disability progression than patients with RMS. Comparing the probability of progression between MSBase PPMS natural history data and a Palace et al. RMS analysis, we see that PPMS patients experience generally a higher probability of progressing across EDSS states. Further analyses are warranted to confirm our findings, which arose from comparison of two different databases.