Diabetes and Parkinson's: Mendelian randomisation

The research record

This is the project's file for this landmark, as written and checked. Sources are listed at the end; anything not yet confirmed is marked "Not yet verified".

Who

Huw Chohan (first author), Konstantin Senkevich, Ziv Gan-Or and Alastair Noyce (last author) with colleagues at Queen Mary University of London (Wolfson Institute), McGill University and the US National Institute on Aging. Movement Disorders 36(6):1420–1429, 2021. The Europe PMC record lists funding from the NIHR, Barts Charity, the Michael J. Fox Foundation, Parkinson Canada, the Fonds de recherche du Québec and the NIH intramural programme.

What happened

Studies of people with diabetes (landmark 216) cannot rule out confounding: diabetes might simply go with other risk factors. Mendelian randomisation uses gene variants that raise diabetes risk. Because genes are assigned at conception, an effect on Parkinson's is less likely to come from lifestyle or illness.

• Observational meta-analysis (6 databases searched). Type 2 diabetes was linked to higher Parkinson's risk (odds ratio 1.21, 95% CI 1.07–1.36), and with less certainty to faster motor progression (standardised mean difference 0.55) and cognitive decline (−0.92).
• Mendelian randomisation. Using genome-wide association data, diabetes raised Parkinson's risk (inverse-variance weighted OR 1.08, 95% CI 1.02–1.14; p = 0.010) and motor progression (OR 1.10, 95% CI 1.01–1.20; p = 0.032). It found no effect on cognitive progression.

Why it moved us toward a cure

It is the only genetic test of whether the diabetes link in landmark 216 could be causal. A causal link supports trying diabetes drugs such as GLP-1 agonists (landmark 20). The effect is small, and these trials have not yet shown slowing.

Outcome

Advanced as supporting evidence. An odds ratio of 1.08 means about 8% higher odds, not a strong driver. Mendelian randomisation can be biased by genes that act on both conditions. Related work by Senkevich et al. (Mov Disord 2023;38:1350) looked at type 1 diabetes, not type 2, and found a possible protective effect (odds ratio 0.97). No correction or retraction is recorded for the 2021 paper.

Sources

• Chohan et al., Mov Disord 2021, DOI 10.1002/mds.28551, PMID 33682937 (abstract read via Europe PMC)
• Senkevich et al., Mov Disord 2023, DOI 10.1002/mds.29424 (abstract read via Europe PMC, PMID 37148456)
• Not yet verified: the pleiotropy checks, sample sizes and number of gene variants used (full text not read).