We performed a case-control study to investigate the association of the poor metaboliser genotype of the cytochrome P450 2D6 gene with Parkinson's disease (PD). Genotyping was determined by the polymerase chain reaction followed by digestion with restriction enzymes. The poor metaboliser genotype was more frequent in 112 patients with PD than in 206 matched controls (odds ratio 1.7, 95% CI: 0.94-2.45). A meta-analysis of these results together with ten other published studies gave a pooled odds ratio for the poor metaboliser genotype of 1.47 (95% CI: 1.18-1.96, P=0.01). Thus, the poor metaboliser genotype has a small but highly significant association with PD which would be easily missed in small studies. Research now should focus on the mechanism of this association.