Potential sources of bias in the studies of risk and protection factors associated with Parkinson’s Disease

Abstract

Introduction: Numerous epidemiological studies indicate that Parkinson’s disease (PD) results from a complex interaction between several factors, some of which confer risk and others protection. However, when the methodological design of each study is thoroughly analyzed, it is found that many did not control for various confounding factors, opening the possibility that the identified association (or absence of it) was in fact a methodological artifact.
Objective: To highlight some of the confounding factors that should be controlled to minimize sources of bias in epidemiological studies of risk and protection factors for PD.
Development: A description is made of the possible sources of bias in the studies of factors that have been associated with an increased risk of developing PD (exposure to pesticides, traumatic brain injuries, having relatives in the first and second degree with PD, having suffered from depression, having a history of chronic constipation), as well as factors that seem to confer protection (consumption of tobacco, caffeine and non­steroidal analgesic drugs).
Conclusions: It is necessary to reach a consensus on the methodology of epidemiological studies of risk/protection factors associated with PD. To reduce bias, the following measures are recommended: (1) apply criteria that increase the certainty that family members with PD were correctly diagnosed, (2) quantitatively estimate the intensity and duration of the factors under study, (3) reduce bias due to reverse causality, excluding from the analysis factors that could occur as a consequence of the prodromal symptoms of PD, which conservatively begin 10 years before diagnosis.