Geographic Clusters of Incident Parkinson Disease in US Medicare Beneficiaries
Irene Faust1, Brad Racette1, Brittany Krzyzanowski1, Susan Nielsen2
1Barrow Neurological Institute, 2Washington University in St. Louis
Objective:
To identify geographic clusters of incident Parkinson disease (PD).
Background:
Twin studies indicate environmental exposures contribute to PD pathogenesis, but few candidate risk factors exist. Cluster analyses can inform hypothesis generation. A nationwide, agnostic search for PD clusters accounting for multiple comparisons has not been done in the U.S. previously.
Design/Methods:
In the Multiple Air Pollutants in PD (MAP-PD) study we conducted a nationwide population-based geographic cluster analysis of 21.5 million Medicare beneficiaries in 2009. We identified 89,790 incident PD cases. We assessed clustering using a two-year exposure lag using residential zip codes in 2007. While stratifying by race/ethnicity (White, Black, Hispanic, Native American, and Asian/Pacific Islander/other), we used SaTScan to conduct 20 primary nationwide cluster analyses for each of the five groups, repeating analyses while allowing no more than 1%, 5%, 10%, or 25% of the underlying population to contribute to a cluster. We also conducted eight sex-specific analyses among White beneficiaries and 86 race/ethnicity-stratified state-specific analyses.
Results:
Across all 28 nationwide cluster analyses, we identified 185 significant clusters: 152 White (60 overall, 45 male, 47 female), 14 Black, 14 Hispanic, and 5 Asian/other. The median relative risk in the cluster vs. the rest of the U.S. in the most restrictive analysis was 1.46 White, 4.31 Black, 2.61 Hispanic, and 2.48 Asian. The mean cluster radius was 58.5 km White, 9.5 km Black, 7.4 km Hispanic, and 7.2 km Asian. Overall, clusters were more common in the eastern vs. western U.S. and accordingly, state-specific analyses often were required to reveal clusters in the western U.S.
Conclusions:
Targeted studies focusing on environmental exposures more common in the eastern U.S. may provide insight into the pathogenesis of PD. Both regional and nationwide studies might be informative.