Objective: To appraise the existing literature on cancer risk among patients with Parkinson's disease (PD), determine the overall cancer risk ratio among patients with PD, explore reasons for variations in study results, and assess the potential for publication bias.
Methods: Studies reporting cancer risk in patients with PD were identified by searching electronic databases through 18 November 2009 using the terms PARKINSON DISEASE, NEOPLASM, and CANCER. Reviewers individually performed data extraction and scored each study using a quality assessment instrument. Cancer risk in all patients with PD was calculated overall, and after excluding melanoma and other skin cancers. We tested for heterogeneity and publication bias, and stratified for gender, smoking-related versus non-smoking-related cancers, and study quality. We pooled effect sizes using fixed-effects and random-effects models.
Results: We included 29 studies in the overall analysis for a total of 107,598 patients with PD. Compared to controls, the aggregate risk for cancer in patients with PD was 0.73 (95% confidence interval [CI], 0.63-0.83), and after excluding skin tumors, 0.69 (95% CI, 0.62-0.78). These risks varied by gender (males, RR = 0.71, 95% CI, 0.57-0.88; females, RR = 0.82; 95% CI, 0.68-0.98). After strictly excluding skin tumors, both smoking-related (RR = 0.61; 95% CI, 0.58-0.65) and non-smoking-related cancer rates (RR = 0.76; 95% CI, 0.65-0.89) were significantly lower among patients with PD .
Conclusions: Studies on cancer risk among patients with PD collectively show significantly reduced cancer risk ratios. Further research to explain the biological mechanisms, particularly for the association with non-smoking-related cancers, appears warranted.