Introduction: Evidence suggests an inverse correlation between smoking and Parkinson's disease (PD), yet the mechanisms remain unclear. This study examines the changing risk relationship between smoking and PD diagnosis using alcohol consumption, another reward-driven behavior, as a comparative measure.
Methods: A nested case-control study was conducted using the UK Biobank (UKBB) database. Participants in the prediagnostic phase of PD were identified, and self-reported data on tobacco and alcohol use were analyzed employing conditional binary logistic regression. Polynomial and piece-wise regression models were employed to discern shifting associations with PD over time.
Results: Of 502,304 participants (63 ± 5.3 years, 63 % male), 3049 prediagnostic PD cases were identified. Non-smokers had a heightened PD risk, and this association strengthened closer to diagnosis. The odds ratio (OR [95 % CI]) associated with PD in non-smokers was 2.02 [1.07-3.81] 1-4 years before diagnosis, compared to 1.36 [1.02-1.83] at >10-year intervals (linear trend, p = 0.012). The time trajectory of ORs was best depicted by a quadratic function, identifying a shift in risk 7.5 years before diagnosis documentation. Similar patterns emerged among alcohol non-consumers, with an 8.5-year interval inflection point.
Conclusion: This study identified two disparate risk trajectories among non-smokers: an initial low-amplitude increased risk at prolonged prediagnostic intervals possibly related to genetic/personality factors, followed by a sharp escalation in risk association commencing 7-8 years before diagnosis, possibly propelled by reverse causality. Similar trends in alcohol consumption reinforce these conclusions. These findings could suggest that smoking cessation may serve as an early indicator of PD.
Copyright © 2024 Elsevier Ltd. All rights reserved.