Tobacco smoking and the risk of Parkinson disease: A 65-year follow-up of 30,000 male British doctors

Abstract

Objective: To investigate the causal relevance of current tobacco smoking for the risk of Parkinson disease (PD).

Methods: We compared the risks of death from PD with smoking habits in 30,000 male doctors in the British Doctors cohort study in 1951 and in survivors who had been resurveyed periodically for 5 decades. Cause-specific mortality was monitored for 65 years and included 283 deaths from PD. The relative risks (RRs) of PD (and 95% confidence intervals [CIs]) were estimated using Cox models for smoking habits (smoking status, amount smoked, and years since quitting) at baseline or updated habits at resurvey.

Results: The prevalence of current smoking declined progressively during follow-up from 67% to 8% between 1951 and 1998. The crude rates of PD death were lower in current smokers than in never smokers at baseline (30 vs 46/100,000 persons-years). After adjustment for age at risk, current smokers at baseline had a 30% lower risk of PD (RR 0.71; 95% CI 0.60-0.84), and continuing smokers classified using updated smoking habits at resurvey had a 40% lower risk (RR 0.60; 95% CI 0.46-0.77) of PD compared with never smokers. The risks of PD were inversely associated with the amount of tobacco smoked. The protective effect of current smoking vs never smoking for PD was attenuated by increasing duration since quitting smoking.

Conclusions: In contrast to previous suggestions, the present report demonstrates a causally protective effect of current smoking on the risk of PD, which may provide insights into the etiology of PD.

Figure 1. Smoking prevalence in surviving participants by age and year of survey

Black and white points used only to distinguish between consecutive resurveys. The shaded bar shows the change of smoking prevalence from 1951 to 1998 in doctors aged 65–69 years.

Figure 2. Relative risk of PD by smoking status (never, ex-, and current smoker) at baseline survey in 1951 and at resurveys and by years since quitting smoking

The smoking status at resurveys was updated at each resurvey, with a 10-year time lag to minimize reverse causality bias. CI, group-specific confidence intervals. n, number of deaths from PD. Tests for trend of PD risk across categories of never, ex-, and current smoker. PD = Parkinson disease.

Figure 3. Association between daily amount of tobacco smoked and the risk of PD

CI, group-specific confidence interval. The smoking status at resurveys was updated at each resurvey, with a 10-year time lag to minimize reverse causality bias. Cigarettes per day or equivalent in pipe and cigar tobacco: 1 cigar = 5 cigarettes, 1 small cigar = 3 cigarettes, 1 very small cigar = 1 cigarette, and 1 oz of tobacco per week = 4 cigarettes per day. Tests for trend in PD risk across categories of daily amount of tobacco smoked. Values are plotted at the daily amount smoked in each category, with an offset of 0.5 cigarette per day (updated consumption) or 1 cigarette/ per day (1951 consumption) to avoid overlap. The red and blue lines were the best fits across the data points. PD = Parkinson disease.