Low-density lipoprotein cholesterol (LDL-C) levels are suggested to be associated inversely with Parkinson's disease (PD). To test the hypothesis that LDL-C levels may increase PD risk, we studied a prospective cohort of 3,233 men (Honolulu-Asia Aging Study) for whom the LDL-C from fasting lipid profiles was obtained during 1991 to 1993. The cohort was followed longitudinally until 2001 for incident Parkinson's cases. During follow-up, 41 men developed PD (18.4/10,000 person-years). Although the incidence of PD increased with decreasing LDL-C in a dose-dependent manner, the association was only significant for men aged 71 to 75 years. In the latter group, risk of PD declined from 38.5/10,000 person-years in men with LDL-C levels <80 mg/dl to less than 9/10,000 person-years for concentrations that were > or =140 mg/dl. After adjustment for age, smoking, coffee intake, and other factors, the relative odds of PD for men at the 80th versus the 20th percentile of LDL-C (135 vs. 85 mg/dl) was 0.4 (95% confidence interval: 0.2, 0.9). This prospective study supports the hypothesis that low LDL-C is associated with an increased risk of PD. Although confirmation is required, the underlying mechanisms may be useful in understanding key aspects of PD.