Dementia in former amateur and professional contact sports participants: population-based cohort study, systematic review, and meta-analysis.

Background: Although there is growing evidence that former professional athletes from sports characterised by repetitive head impact subsequently experience an elevated risk of dementia, the occurrence of this disorder in retired amateurs, who represent a larger population, is uncertain. The present meta-analysis integrates new results from individual-participant analyses of a cohort study of former amateur contact sports participants into a systematic review of existing studies of retired professionals and amateurs. Methods: The cohort study comprised 2005 male retired amateur athletes who had competed internationally for Finland (1920-1965) and a general population comparison group of 1386 age-equivalent men. Dementia occurrence was ascertained from linked national mortality and hospital records. For the PROSPERO-registered (CRD42022352780) systematic review, we searched PubMed and Embase from their inception to April 2023, including cohort studies published in English that reported standard estimates of association and variance. Study-specific estimates were aggregated using random-effect meta-analysis. An adapted Cochrane Risk of Bias Tool was used to assess study quality. Findings: In the cohort study, up to 46 years of health surveillance of 3391 men gave rise to 406 dementia cases (265 Alzheimer's disease). After adjustment for covariates, former boxers experienced elevated rates of dementia (hazard ratio: 3.60 [95% CI 2.46, 5.28]) and Alzheimer's disease (4.10 [2.55, 6.61]) relative to general population controls. Associations were of lower magnitude in retired wrestlers (dementia: 1.51 [0.98, 2.34]; Alzheimer's disease: 2.11 [1.28, 3.48]) and soccer players (dementia: 1.55 [1.00, 2.41]; Alzheimer's disease: 2.07 [1.23, 3.46]), with some estimates including unity. The systematic review identified 827 potentially eligible published articles, of which 9 met our inclusion criteria. These few retrieved studies all sampled men and the majority were of moderate quality. In sport-specific analyses according to playing level, there was a marked difference in dementia rates in onetime professional American football players (2 studies; summary risk ratio: 2.96 [95% CI 1.66, 5.30]) relative to amateurs in whom there was no suggestion of an association (2 studies; 0.90 [0.52, 1.56]). For soccer players, while dementia occurrence was raised in both erstwhile professionals (2 studies; 3.61 [2.92, 4.45]) and amateurs (1 study; 1.60 [1.11, 2.30]) there was again a suggestion of a risk differential. The only studies of boxers comprised former amateurs in whom there was a tripling in the rates of dementia (2 studies; 3.14 [95% CI 1.72, 5.74]) and Alzheimer's disease (2 studies; 3.07 [1.01, 9.38]) at follow-up compared to controls. Interpretation: Based on a small number of studies exclusively sampling men, former amateur participants in soccer, boxing, and wrestling appeared to experience an elevated risk of dementia relative to the general population. Where data allowed comparison, there was a suggestion that risks were greater amongst retired professionals relative to amateurs in the sports of soccer and American football. Whether these findings are generalisable to the contact sports not featured, and to women, warrants examination. Funding: This work was unfunded.