Background: Meta-analyses have reported lower cancer incidence in patients with Parkinson's disease (PD) compared to the general population but with considerable data heterogeneity.

Objective: To explore how the validity of the PD diagnoses is related to the association with cancer.

Methods: We conducted a systematic review and meta-analysis in which studies were stratified into groups based on the diagnostic validity of Parkinson's disease. Studies investigating mortality data and those examining cancer risk within certain genetic subgroups of PD were excluded.

Results: Thirty-four articles encompassing 533,102 patients with PD from 11 countries met the inclusion criteria. Stratified analyses revealed no association between PD and overall cancer risk preceding or following the PD diagnosis in studies using validated PD data. Studies utilizing less robust PD identification methods, the majority of which were cohort studies, demonstrated a neutral or decreased cancer risk among PD patients. In the studies with the most rigorous PD validation organ-specific analyses showed an increased risk of cutaneous melanoma but no decreased risk in any type of cancer. The positive association between PD and melanoma was more pronounced in the studies with more robust PD diagnosis validity.

Conclusions: The reported associations between PD and cancer are substantially influenced by the quality of PD data. Future investigations should concentrate on organ-specific cancers, instead of pooling cancers together, and use only PD cohorts with validated diagnosis.