D.J. Myall, T.L. Pitcher, J.F. Pearson, J.C. Dalrymple-Alford, T.J. Anderson, M.R. MacAskill

Traditionally the risk of Parkinson's has been considered to increase monotonically with age, although there is evidence that prevalence and incidence may decrease in the oldest old. To examine this further we estimated the national prevalence and incidence of Parkinson's in New Zealand, using drug-tracing methods, to examine the relationship of Parkinson's with sex and age up to 100+. Information on Parkinson's-related medications was extracted from the national pharmaceutical database of community-dispensed medications from 2005 to 2014. Diagnoses for a large subset of individuals were independently determined through national mortality and hospital admissions datasets. We used a Bayesian model, accommodating diagnostic uncertainty and bias, to estimate the number of people with Parkinson's. The 2013 prevalence of Parkinson's in New Zealand was 191 per 100 000 population (95% uncertainty interval 187-194) with age-standardized prevalence rates higher for males (ratio 1.7:1). Incidence was 29 per 100 000 person-years (95% uncertainty interval 29-30), also higher in males (ratio 1.7:1). Incidence and prevalence by age increased exponentially until 75 years, peaked at 85 years, and then dropped sharply. The prevalence of Parkinson's in New Zealand is expected to double over a 25-year period but then increase at a slower rate due to the drop-off in prevalence and incidence in the oldest old. The findings suggest that Parkinson's disease is not an aging-dependent but an age-dependent disorder.