The reasons for the differences in emphasis on striatonigral or olivopontocerebellar involvement in multiple system atrophy (MSA) remain to be determined. Semi-quantitative pathological analyses carried out in the United Kingdom and Japan demonstrated that olivopontocerebellar-predominant pathology was more frequent in Japanese MSA than British MSA. This observation provides evidence for a difference in phenotype distribution between British and Japanese patients with definite MSA. Studies of the natural history and epidemiology of MSA carried out in various populations have revealed that the relative prevalences of clinical subtypes of MSA probably differ among populations; the majority of MSA patients diagnosed in Europe have predominant parkinsonism (MSA-P), while the majority of MSA patients diagnosed in Asia have predominant cerebellar ataxia (MSA-C). Although potential drawbacks to the published frequencies of clinical subtypes and pathological subtypes should be considered because of selection biases, the difference demonstrated in pathological subtype is also consistent with the differences in clinical subtype of MSA demonstrated between Europe and Asia. Modest alterations in susceptibility factors may contribute to the difference in MSA phenotype distribution between populations. Synergistic interactions between genetic risk variants and environmental toxins responsible for parkinsonism or cerebellar dysfunction should therefore be explored. Further investigations are needed to determine the environmental, genetic, and epigenetic factors that account for the differences in clinicopathological phenotype of MSA among different populations.