Earlier drinking linked to higher modeled Parkinson’s risk
A model accounting for time since drinking began put Parkinson’s disease rates in drinkers at 1.61 times those in non-drinkers.
Our read
Populations- Studied:
- Cohort study of 23,475 participants aged ≥60 years
- What changed:
- Cumulative alcohol-attributable risk by age 90 was 32.1% for initiation at age 20 against 4.9% for age 40
- The catch:
- Alcohol cessation was not modeled as a separate transition, risking sick-quitter bias
- In the record:
- The 850th paper on alcohol with results in people; 2,929 on record in all.
Editor of record: Boris Djordjevic · AgingResearch.News
Taiwanese adults in two community screening cohorts—23,475 aged 60 or older in Keelung and 90,129 aged 50 or older in Changhua—were analyzed in an observational study. Questionnaires recorded alcohol use, including the age when drinking began. Linked health records helped identify Parkinson’s disease.
In Keelung, an unadjusted analysis linked drinking to lower disease risk. After adjustment for age, sex, education and smoking, the range of likely values included no difference. A model accounting for time since drinking began instead linked alcohol use to higher modeled lifetime risk, particularly with earlier initiation. The association weakened but remained elevated after accounting for smoking. Researchers applied this model to Changhua without recalculating its parameters. Those projections also put disease rates higher in drinkers than non-drinkers, with a wider gap at older ages.
Why it matters
Parkinson’s disease develops slowly, making the timing of possible risk exposures important to study. The authors suggest that accounting for this delay may inform prevention strategies spanning decades.
Caveats
The model could not separate the independent effects of current age, age when drinking began and time since then. Participants were older Taiwanese screening participants, and the Changhua results were projections using the Keelung model rather than an independently fitted model.
From the paper
- Participants
In total, 23,475 participants aged ≥60 years who underwent screening between 2002 and 2005 were included in the present analysis.
- Limitation
We did not model alcohol cessation as a separate transition, so reverse causation and sick-quitter bias may persist; however, sensitivity analyses excluding recent quitters showed similar qualitative timing patterns.
Quoted word for word from the full text on PubMed Central.
The paper
National Taiwan University
Journal of Parkinson's Disease, May 25, 2026, Open access
Comments
Be the first to comment.