Objective: PD may be caused by genetic susceptibility to neurotoxins. CYP2D6 is a candidate gene for PD because it regulates drug and toxin metabolism, but association studies have been inconsistent. The aim of this study was to test if the CYP2D6*4 allele (poor metabolizer phenotype) is associated with earlier age at onset. Methods: Five hundred seventy-six patients with PD and 247 subjects without PD were studied using standard diagnostic, genotyping, and statistical techniques. Results: Surprisingly, mean onset age was significantly later in *4-positive patients. Frequency of *4 was significantly higher in late-onset PD than early-onset PD. When early- and late-onset PD were analyzed separately, *4 had no effect on onset age; hence, the association with delayed onset was likely an artifact of an elevated *4 frequency in late-onset PD. Contrary to a common assumption that CYP2D6 frequencies do not change with age, *4 frequency rose significantly with advancing age, both in patients with PD (from 0.16 at mean age of 56.5 years to 0.21 at mean age of 72) and subjects without PD (from 0.09 at mean age of 45.5 years to 0.21 at mean age of 72). *4 Frequencies in patients with early- and late-onset PD, although different from each other, were in agreement with similarly aged subjects without PD, suggesting the elevated *4 frequency in late-onset PD was likely an age effect, unrelated to PD. Conclusion: The CYP2D6*4 allele is not associated with earlier PD onset. *4 May be associated with survival. Inconsistent results from allelic association studies may have been due to an unrecognized age effect.