Although most patients with multiple sclerosis (MS) present with relapsing disease (RRMS), 10–20% have progressive disease from onset (PPMS). Robust prospective natural history data on PPMS remain scarce. Establishing a contemporary natural history of PPMS is important for future development of therapeutic interventions. 2009 patients were recruited to a population-based epidemiological study. Clinical data were collected prospectively. 207 (10.3%) had PPMS and 1802 (89.7%) RRMS. PPMS patients with superimposed relapses were compared to those with none. Survival analysis was used to compare time to EDSS 4, 6 and 8. Mixed linear modelling was used to assess the factors important in determining change in EDSS over time. F : M ratio was 1.07 in PPMS and 2.55 in RRMS (p=0.003). Mean age at onset was significantly higher in PPMS (p<0.001) and time to EDSS 4, 6 and 8 was significantly shorter (p<0.001). Within the PPMS cohort, patients with superimposed relapses were younger at disease onset (p=0.008) but had an identical disability trajectory. This was confirmed with mixed modelling demonstrating more rapid accumulation of disability with increasing age at onset. We have described the natural history of PPMS in a population-based cohort. Age at onset is the most important factor in determining rate of disability accumulation. No evidence was found for a difference between PPMS with and without superimposed relapses. Focus on older age groups may be useful in understanding the pathophysiology of progression and development of new therapeutic approaches.