Background and purposeLevodopa treatment in Parkinson's disease (PD) causes motor fluctuations and dyskinesias, but few data describe their development or severity in unselected incident cohorts.MethodsDemographic, clinical, treatment, smoking, caffeine and alcohol data from 183 people with PD were gathered from the Parkinsonism Incidence in Northeast Scotland (PINE) study, a community‐based, incident cohort. With Kaplan–Meier survival analysis and Cox regression modelling the development, and severity, of dyskinesias and motor fluctuations and which factors independently influenced their onset were assessed.ResultsAfter a mean follow‐up of 59 months, 39 patients (21.3%) developed motor fluctuations and 52 (28.4%) developed dyskinesias. Kaplan–Meier estimates of the probability of motor fluctuations and dyskinesias after 5 years of dopaminergic treatment were 29.2% [95% confidence interval (CI) 21.5%–38.8%] and 37.0% (95% CI 28.5%–47.1%) respectively. 19.8% developed motor fluctuations requiring treatment changes but only 4.0% (95% CI 1.5%–10.4%) developed dyskinesias requiring treatment changes by 5 years. Cumulative levodopa dose [hazard ratio (HR) 1.38 (95% CI 1.19–1.60)], female sex [HR 2.41 (1.19–4.89)] and younger age at diagnosis [HR 1.08 (1.04–1.11)] were independently associated with development of motor fluctuations. Cumulative levodopa dose [HR 1.23 (1.08–1.40)] and female sex [HR 2.51 (1.40–4.51)] were independently associated with dyskinesias. In exploratory analyses, moderate caffeine exposure was associated with fewer motor fluctuations, longer symptom duration with more dyskinesias, and tremor at diagnosis with higher rates of both complications.ConclusionsIn this community‐based incident PD cohort, severe dyskinesias were rare. Cumulative levodopa dose was the strongest predictor of both dyskinesias and motor fluctuations.