IS THE INVERSE ASSOCIATION BETWEEN PARKINSON’S DISEASE AND SMOKING DUE TO CONFOUNDING?

ABSTRACT Parkinson’s disease (PD) is inversely associated with smoking. Whether this association is due to a causal relationship or to confounding by a covariate of smoking is still debated. This study included populations of the United States and ten U.S. states from 2004 to 2016. The ten U.S. states included the five states with the highest PD incidence rates in 2017 (Vermont, Maine, Kansas, Missouri, Colorado) and the five states with the lowest PD incidence rates in 2017 (Arkansas, Mississippi, South Dakota, Nebraska, Alabama). The study used scatter plots to explore the association between PD incidence and smoking and the association between PD incidence and a covariate of smoking, lower endoscopy utilization. The results indicate that there is a weaker association between PD and smoking when compared to the association between PD and lower endoscopy utilization. For PD incidence verses current smoking, the coefficient of determination (R 2 ) for the United States was 0.813 and ranged from a low of 0.000 for Arkansas to 0.798 for Colorado (average R 2 for the ten states was 0.391). For PD incidence verses lower endoscopy utilization, the R 2 for the United States was 0.995 and ranged from a low of 0.871 for South Dakota to 0.997 for Colorado (average R 2 for the ten states was 0.959). The results suggest that the inverse association between PD incidence and smoking is confounded by the direct association between PD and lower endoscopy utilization. Further investigation of a possible relationship between PD incidence and lower endoscopy utilization is warranted.