OBJECTIVE To study the association of PD with preceding smoking, alcohol, and coffee consumption using a case-control design. METHODS The authors used the medical records linkage system of the Rochester Epidemiology Project to identify 196 subjects who developed PD in Olmsted County, MN, during the years 1976 to 1995. Each incident case was matched by age (+/-1 year) and sex to a general population control subject. The authors reviewed the complete medical records of cases and control subjects to abstract exposure information. RESULTS For coffee consumption, the authors found an OR of 0.35 (95% CI = 0.16 to 0.78, p = 0.01), a dose-effect trend (p = 0.003), and a later age at PD onset in cases who drank coffee compared with those who never did (median 72 versus 64 years; p = 0.0002). The inverse association with coffee remained significant after adjustment for education, smoking, and alcohol drinking and was restricted to PD cases with onset at age <72 years and to men. The OR for cigarette smoking was 0.69 (95% CI = 0.45 to 1.08, p = 0.1). The authors found no association between PD and alcohol consumption. Extreme or unusual behaviors such as tobacco chewing or snuff use and a diagnosis of alcoholism were significantly more common in control subjects than cases. CONCLUSIONS These findings suggest an inverse association between coffee drinking and PD; however, this association does not imply that coffee has a direct protective effect against PD. Alternative explanations for the association should be considered.