The epidemiology of multiple system atrophy (MSA) is scarcely known, and risk factors have not been definitely identified. We investigated the effect of family history for neurodegenerative diseases and environmental factors on MSA risk in a multicentric case-control study. A total of 73 MSA patients (42 men, 31 women; age, 64.3 +/- 8.1 years; disease duration, 4.8 +/- 3.9 years), 146 hospital controls (84 men, 62 women; age, 64.9 +/- 8.4 years), and 73 population controls (42 men, 31 women; age, 63.7 +/- 8.9 years) matched for sex, age (+/-3 years), and province of residence were enrolled consecutively at seven neurological centers from 1 January 1994 to 31 July 1998. The following variables were investigated: family history of neurodegenerative diseases, education, smoking habits, hobbies, and occupational history. Occupational history of farming was significantly more frequent among MSA cases than controls (OR adj = 2.52; 95% CI, 1.25 to 5.07, MSA vs. hospital controls; OR adj = 4.53; 95% CI, 1.68 to12.2, MSA cases vs. population controls). A dose-response analysis for years of farming corroborated this association. We recently found that smoking is significantly less frequent among MSA cases than controls (Vanacore et al.  Neurology 54:114-119). Here, we report that the effects of farming and smoking on MSA risk do not interact. Our results suggest that occupational history of farming is a risk factor for MSA. Smoking and farming seem to influence MSA risk independently. Further epidemiological studies might provide clues on the etiopathogenesis of MSA.