To determine if ventricular cerebrospinal fluid (vCSF) alpha-tocopherol levels in Parkinson's disease (PD) patients can be increased by oral alpha-tocopherol supplementation and whether vCSF levels are linearly related to plasma alpha-tocopherol levels. In spite of its putative neuroprotective properties, alpha-tocopherol has failed to alter PD clinical progression. However, the ability of supplemental alpha-tocopherol to affect brain or vCSF levels has never been assessed in humans nor has a dose response curve for alpha-tocopherol in vCSF been established. Five PD patients with Ommaya catheters received oral dl-alpha-tocopherol over 5 months. Each patient ingested alpha-tocopherol daily with monthly dosage increases (400, 800, 1,600, 3,200, 4,000 IU/day). Plasma and vCSF samples were obtained at baseline and at the end of each month. Alpha-tocopherol levels were determined in triplicate by high-pressure liquid chromatography with fluorometric and electrochemical detection. At baseline, endogenous alpha-tocopherol was detected in plasma and vCSF, with a greater than one-hundred-fold difference between the fluid compartments (mean plasma level 18.76 microM/l (SD +/- 4.69) versus mean CSF level 0.114 microM/l (SD +/- 0.084). A clear dose-response curve occurred in plasma, with statistically significant increases over baseline developing even with 400 IU/d. With higher doses, a significant increase continued without evidence of saturation. However, there was no significant increase in vCSF alpha-tocopherol levels at any dose, including the supraclinical (4,000 IU/d). There was no correlation between plasma and vCSF alpha-tocopherol levels. Oral alpha-tocopherol supplementation, even at supraclinical doses, fails to increase vCSF alpha-tocopherol levels. This lack of change may be due to limited passage across the blood-brain barrier or very rapid alpha-tocopherol metabolism. All prior negative studies on efficacy of alpha-tocopherol in PD may need reevaluation in light of these pharmacologic data.