Abstract

The association between head injury and Parkinson's disease remains uncertain due to potential problems of reverse causation and recall bias. Few studies have compared young onset (YOPD) and later onset (LOPD) cases as both types of bias would be less severe with YOPD cases. We undertook a community-based case-control study (426 cases, 405 controls). Cases and controls were similar for age, rurality and ethnicity, but controls were more likely to be female (p<0.001). YOPD was defined as age at onset <45 and LOPD ≥45 years. We ignored a head injury within 1 year of disease onset (reverse causation) and only accepted a serious injury resulting in loss of consciousness or A&E attendance or hospital admission to diminish recall bias. Head injury was associated with a twofold increased risk (age adjusted odds ratio 2.1 95% CI 1.5 to 3.0) which was hardly altered (OR 2.3) after adjustment sex, education, smoking, alcohol and rurality. Similar sized multivariable estimates were found with YOPD (OR 2.4) and LOPD (2.2) though only the latter reached conventional levels of statistical significance, due to greater power. The consistency of findings for both YOPD and LOPD suggests that the role of head injury requires further attention.