Age and sex specific mortality in amyotrophic lateral sclerosis (ALS)

Abstract

Background: Surprisingly few risk/protective factors have been consistently identified in Parkinson's disease (PD), so far. Prospective cohort studies are the optimal study design in observational epidemiology for causal inference; they are recall biasfree and the potential for selection bias is minimised if the follow-up is complete. Aim of this research is to identify PD cases within the EPIC cohort. Methods: A total of 299, 568 individuals recruited across 16 EPIC centres were included in the study. Case ascertainment procedures were articulated in 3 phases: phase I - possible cases identification via record linkage and/or self-report; phase II - case validation and collection of additional clinical information; phase III - across centre validation. Results: A total of 735 PD case were expected on the basis of the age distribution of the populations. During phase I, 1, 349 possible cases were identified. Of these, 555 cases were reviewed in 10 out of 16 centres. Among reviewed cases, 355 PD cases were identified, in addition to 16 multiple system atrophy, 9 progressive supranuclear palsy, 12 vascular parkinsonisms, 16 Lewy body dementia, 15 essential tremor, and 4 PD/essential tremor cases. All cases were either labelled as definite, very likely, probable or possible on the basis of the quality and amount of clinical information available, and the degree of confidence of the reviewing neurologist. The algorithm will be further revised in phase III. Conclusion(s): The present study represents a proof-of-principle of methods for ascertaining PD cases in large populationbased studies.