Environmental risk factors and Parkinson's disease: An umbrella review of meta-analyses

•We assessed the strength of the evidence linking 75 environmental risk factors with Parkinson's disease.•Physical activity and constipation presented Class I evidence for an association.•Head injury, anxiety or depression, beta-blockers, smoking, and serum uric acid presented Class II evidence for an association.•The majority of the examined meta-analyses had large or very large heterogeneity, many had signals of bias, and confounding or reverse causation was a common explanation.•Data from more prospective studies and investigation of sources of heterogeneity are needed to better understand the association between the risk factors and PD.

BackgroundParkinson's disease is a neurological disorder with complex pathogenesis implicating both environmental and genetic factors. We aimed to summarise the environmental risk factors that have been studied for potential association with Parkinson's disease, assess the presence of diverse biases, and identify the risk factors with the strongest support.MethodsWe searched PubMed from inception to September 18, 2015, to identify systematic reviews and meta-analyses of observational studies that examined associations between environmental factors and Parkinson's disease. For each meta-analysis we estimated the summary effect size by random-effects and fixed-effects models, the 95% confidence interval and the 95% prediction interval. We estimated the between-study heterogeneity expressed by I2, evidence of small-study effects and evidence of excess significance bias.ResultsOverall, 75 unique meta-analyses on different risk factors for Parkinson's disease were examined, covering diverse biomarkers, dietary factors, drugs, medical history or comorbid diseases, exposure to toxic environmental agents and habits. 21 of 75 meta-analyses had results that were significant at p < 0.001 by random-effects. Evidence for an association was convincing (more than 1000 cases, p < 10−6 by random-effects, not large heterogeneity, 95% prediction interval excluding the null value and absence of hints for small-study effects and excess significance bias) for constipation, and physical activity.ConclusionMany environmental factors have substantial evidence of association with Parkinson's disease, but several, perhaps most, of them may reflect reverse causation, residual confounding, information bias, sponsor conflicts or other caveats.