Carbon monoxide, not nicotine, may explain why smokers get less Parkinson disease

Design

nationwide prospective cohort study with registry and insurance linkage, Cox regression, about 12 years of follow-up

Population

512,701 adults aged 30–79 across 10 areas of China; 1,131 incident Parkinson disease cases, 675 among never-smokers

Primary outcome

adjusted hazard ratio for incident Parkinson disease by smoking status and by exhaled carbon monoxide in never-smokers

Effect

regular smoking HR 0.70 (95% CI 0.62–0.79); in never-smokers, exhaled CO 3.0 to <5.0 ppm HR 0.62 (0.53–0.73), ≥11.5 ppm HR 0.65 (0.45–0.92), p for trend <0.001

The inverse association between smoking and Parkinson disease is one of the most reproducible findings in neuroepidemiology and one of the least explained. This analysis of the China Kadoorie Biobank — 512,701 adults aged 30 to 79, about twelve years of follow-up, 1,131 incident Parkinson cases — went after the mechanism by using exhaled carbon monoxide as a measured exposure rather than relying on self-reported smoking.

Regular smoking carried an adjusted hazard ratio of 0.70 (95% CI 0.62–0.79) for Parkinson disease, alongside the expected raised risks of lung cancer, ischaemic heart disease, stroke and death. The informative analysis is in never-smokers: among them, higher exhaled carbon monoxide was associated with lower Parkinson risk in a broadly dose-dependent way — 0.85 at 2 to under 3 ppm, 0.62 at 3 to under 5, 0.68 at 5 to under 11.5, and 0.65 at 11.5 or above, p for trend below 0.001.

What makes this more than another confounded association is the specificity. Exhaled carbon monoxide in never-smokers was not associated with smoking-related diseases, and not associated with other neurodegenerative diseases. If residual smoking exposure were driving it, lung cancer and heart disease should have moved too. The signal is confined to Parkinson disease, which is the pattern a real mechanism produces.

• Nothing here changes advice: smoking causes far more disease than it could conceivably prevent.
• In never-smokers, exhaled carbon monoxide in this Chinese cohort largely reflects household solid-fuel and biomass exposure, which the analysis adjusted for.
• Indian household biomass exposure is comparable and often higher; if the association is causal, it is already part of the Indian Parkinson epidemiology.
• Carbon monoxide is under investigation as a therapeutic in Parkinson disease; this is supporting epidemiology for those trials, not a reason to act.
• The association was stronger in never-smoking women, in whom smoking prevalence in this cohort was 3.3% — the cleanest subgroup available.

Why it matters

A thirty-year epidemiological puzzle now has a candidate molecule, and that molecule is already in clinical trials.

Don't overread it

This was a prospective cohort — it cannot show that carbon monoxide prevents Parkinson disease, only that the two track together.

The statistics, in plain English

Hazard ratios from 0.85 down to 0.62 across rising exposure bands, with a trend p below 0.001, describe a dose-response — usually the strongest observational argument for causation. The specificity check does most of the work: exposures that are confounded by smoking should raise smoking-related disease risk, and this one did not. But exhaled carbon monoxide is a snapshot at recruitment, not lifetime exposure, and a measure taken once will misclassify people whose exposure changed over twelve years, which generally weakens an association rather than creating one.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.