Modelling the natural history of primary progressive multiple sclerosis

Background: A minority of patients with multiple sclerosis (MS) have primary progressive disease (PPMS). Treatment options are currently limited, but as prospects for interventional studies become more realistic, understanding contemporary outcome data will be key to successful trial design. Methods: 234 PPMS patients were identified from a population-based cohort of 2131 (11.0%) and mean follow-up of 13.1 years. Time to established disability endpoints was compared with patients with relapsing-onset MS (ROMS) using survival analysis, and Cox regression employed to explore factors contributing to disability accumulation. Results: were used to create predictive power models for clinical trials in PPMS. Results Time to fixed disability milestones was shorter than in ROMS (Expanded Disability Status Scale (EDSS) 4.0:8.1 vs 17.1 years, p<0.001; EDSS 6.0: 9.6 vs 22.1 years, p<0.001; EDSS 8.0: 20.7 vs 39.7 years, p<0.001), but there were no differences in age-related disability. Age and cerebellar symptoms at onset affected rate of progression. Modelling of these data indicated that trials employing EDSS change of 1.0 as the primary outcome measure would be powered to detect a 20% difference in progression using 600 patients with initial EDSS of 4.0 per trial arm, or 400 patients with initial EDSS of 5.0 per arm. However, trials including patients with fixed EDSS of ≤6.0 will be underpowered even with large numbers or prolonged duration. Conclusions: Disability progression in PPMS is variable and influenced by age at onset. Although progression is more rapid, age-related disability milestones are identical to relapsing-onset disease. These data offer a contemporary paradigm for clinical trial design in progressive MS.