Type 2 diabetes raises Parkinson's risk

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

Qun Xu, Honglei Chen and colleagues at the US National Institute of Environmental Health Sciences (Diabetes Care 34:910–915, 2011; NIH intramural funding). Esther De Pablo-Fernandez, Raph Goldacre, Alastair Noyce, Thomas Warner and colleagues, Oxford and UCL (Neurology 91:e139–e142, 2018; Barts Charity and Medical Research Council funding listed).

What happened

• Xu 2011. A prospective study of 288,662 NIH-AARP participants, using diabetes reported in 1995–96; 1,565 developed Parkinson's after 1995. Parkinson's was about 40% more likely in people with diabetes (odds ratio 1.41, 95% CI 1.20–1.66). The excess was mostly in those with diabetes for over 10 years (OR 1.75).
• De Pablo-Fernandez 2018. Linked English hospital and death records, 1999–2011. About 2.0 million people admitted to hospital with type 2 diabetes versus about 6.2 million in the comparison group. Parkinson's was more common afterwards (hazard ratio 1.32), more so with complicated diabetes (1.49) and in people diagnosed at 25–44 (3.81).

Why it moved us toward a cure

It gave a population-level reason to look at insulin signalling and GLP-1 drugs (diabetes drugs) in Parkinson's, leading to the exenatide pilot (landmark 101); the 2011 paper is not cited as the trigger in the sources I read, so this link is the project's inference. Suggested new parent link: 216 should be a parent of 101.

Outcome

Advanced as a rationale, but an association is not proof of cause. The authors suggest shared genes or disease pathways. The results are not uniform: a US cohort of 147,096 people (Palacios et al., Mov Disord 2011; 656 cases) found no link (relative risk 0.88, 0.62–1.25). The GLP-1 trials that followed have not yet yielded a treatment: exenatide Phase 3 (landmark 142) and NLY01 (landmark 143) did not show slowing, and lixisenatide showed a small motor benefit in Phase 2 (landmark 141). Hospital-record studies may miss mild cases. No correction or retraction was found for either main paper.

Sources

• Xu et al., Diabetes Care 2011;34:910–915, DOI 10.2337/dc10-1922, PMID 21378214 (abstract read via Europe PMC)
• De Pablo-Fernandez et al., Neurology 2018;91:e139–e142, DOI 10.1212/WNL.0000000000005771, PMID 29898968 (abstract read via Europe PMC)
• Palacios et al., Mov Disord 2011;26:2253–2259, DOI 10.1002/mds.23855 (the null cohort)
• Not yet verified: meta-analyses pooling the cohorts (not searched); that landmark 101's design cites this work; PubPeer not checked.