Background Previous studies have failed to agree on the risk factors for excessive daytime sleepiness in Parkinson's disease and the relative contribution of each as the disease progresses.

Methods We studied two community-based incident Parkinson's disease cohorts. Cohort 1 (n=128) were newly-diagnosed patients and Cohort 2 were assessed at 3.5 years (n=118), 5 years (n=90) and 7 years (n=45) following diagnosis. Using Epworth Sleepiness Scale score as the outcome variable, multivariate linear regression was used to identify risk factors for excessive daytime sleepiness at each assessment.

Results Epworth Sleepiness Scale score was significantly lower at diagnosis (Cohort 1) compared to 3.5, 5 and 7 years following diagnosis (Cohort 2). In Cohort 1, the proportion of patients with excessive daytime sleepiness (Epworth Sleepiness Scale score >10) was 20%. In Cohort 2, this rose to 49%, 53% and 44% at 3.5, 5 and 7 years respectively. In the early stages of Parkinson's disease, a non-tremor dominant motor phenotype, dopamine agonist use and depression were associated with excessive daytime sleepiness, whereas later on, motor disability and impaired quality of life became the most significant factors.

Conclusions This is the first study to map the temporal evolution of excessive daytime sleepiness and its risk factors in an incident Parkinson's disease cohort. We found that hypersomnolence is present throughout the course of the disease and increases with disease progression. The link between non-tremor dominant motor phenotype and excessive daytime sleepiness, but lack of association with cognitive function, implicates brainstem rather than cortex as the principal site of underlying pathology. Dopamine agonists appear to play a contributory role.