Table 3.

Summary of results from studies reporting data on the frequency of MCI and dementias

Characteristics					Quality assessment			Summary of findings
								N. of subjects		Results
Fist author, year	Sport	Source of cases	Reference population	Exposure	Selection	Comparability	Outcome/exposure	n. cases among the exposed/total exposed	n. cases among the reference group/total reference	Data ORs/HRs (95%CIs) or SIRs/PRs/SMRs/SPMRs
Cohort studies
Russell 2021 [84]	Soccer	Hospital records, prescriptions and death certificates with ICD-9 and ICD-10 codes	Random sample of matched individuals (3:1 ratio) from the community health index database	Having been a soccer player in the Scottish League as documented in the archives of the Scottish Football Museum and individual league clubs	***	**	**	338/7676 players	312/23028 non-players	frequency of dementia (not otherwise specified) diagnosis HR 3.59 (95% CI 2.93–4.39) p < 0.001
Mackay 2020 [82]	Soccer	Electronic health records to obtain data on death certification and medications typically prescribed for the treatment of dementia ICD-9 and ICD-10 codes that were used to identify outcomes as death with neurodegenerative disease listed as the primary or a contributory cause	Controls from general population matched to players based on sex, age, and degree of social deprivation	Former professional soccer player born before January 1, 1977	**	**	**	Dementia 180/7676 players 178 non-players AD 64/7676 players non-AD dementia 121/7676 players	Dementia 178/23,028 non-players AD 47/23,018 non-players Non-AD dementia 133/23,018 non-players	dementia HR for Death (95% CI): 3.87 (2.86–5.24) p: < 0.001 AD HR for Death (95% CI): 5.07 (2.92–8.82) p: < 0.001 non-AD dementia HR for Death (95% CI): 3.48 (2.42–5.00) p: < 0.001
Nguyen 2019 [79]	American football baseball	Death certificates from 1979 to 2013 classified with ICD underlying (U) and underlying or contributing (U&C)	MLB cohort: from LBD of all players that appeared in ≥ 1 game at a professional level in 1871–2006	NFL cohort: from NIOSH consisting in players who played ≥ 1 season in 1959–1988	**	**	**	NFL U: 6/3419 players U&C: 16/3419 players	MLB U 7/2708 players U&C 10/2708 players	Dementia/AD HR (95%CI) U: 1.28 (0.41–4.07) U&C: 2.26 (0.99–5.17)
Willer 2018 [78]	American football hockey	Comprehensive neuro-cognitive assessment	Age-matched noncontact sport athlete controls	Retired professional hockey and football athletes (average age 56 years)	**	*	*	8/21 contact athletes	3/21 non-contact athletes	MCI frequency contact sport vs noncontact sport 8/21 vs 3/21 p = 0.08
Janssen 2017 [77]	American football basketball swimming wrestling	3-step procedure: screening of medical index through REP tool, review of diagnoses, diagnosis confirmed by senior neurologist	Varsity athletes (football vs non-football) enrolled in the Mayo and Rochester High Schools in 1956 to 1970	Varsity activity in football, basketball, swimming, or wrestling documented in the yearbooks	**	*	**	7/296 football players	5/190 non-football athletes	> 1 year of playing football: 4.99 (1.62–11.65) non-football: 3.24 (0.88–8.28) IRR: 1.54 (0.50–3.60), p = 0.26
Vann Jones 2013 [76]	Soccer	Test Your Memory, self-administered questionnaire	Frequency also compared to data from a large UK-based MCI prevalence study	Having been a soccer player from Former Player Associations (FPAs) of 4 professional football clubs in the UK	**	*	*	MCI—possible dementia 10/92 soccer players	–	Overall MCI/dementia 10/92 MCI: 8/92 dementia: 2/92 No significant difference in cognitive impairment prevalence between ex-professional footballers and a large sample of men in Wales
Lehman 2012 [74]	American football	Death certificates with ICD code for AD in effect at death	US male mortality rates (1960–2007) for 119 cause of death categories	Having played in the National Football League for at least 5 seasons between 1959 and 1988	****	**	*	U: 2/3439 players C: 7/3439	–	AD SMR-U: 1.80 (0.22–6.50) SMR-U&C: 3.86 (1.55–7.95)
Savica 2012 [75]	American football	Medical records confirmed by a neurologist	Having been a band, glee club, or choir member as for the yearbooks of the only 2 high schools in Rochester in 1946–1956 Expected incidence calculated based on previously published rates for that area	Having been a football player as for the yearbooks of the only 2 high schools in Rochester in 1946–1956	***	**	**	13/438players	2/140 non-players	Dementia HR (95%CI) 1.58 (0.36–7.01) (p = 0.55) SIR (95%CI) observed vs expected players: 0.72 (0.38–1.23) non-players: 0.47 (0.05–1.68)
Guskiewicz 2005 [72]	American football	AD diagnosed by clinicians MCI diagnosed by clinician, self-reported or reported by informant	General US Population (AD)	Members of the National Football League Retired Player’s Association	**	*	*	AD 33/2252 MCI diagnosed 22/758 MCI reported by informant 77/758	–	AD 33/2252 overall age-adjusted prevalence ratio 1.37 (95% CI 0.98–1.56) higher in (≤ 70 years) 22 diagnosed MCI and 77 memory impairment
Schulte 1996 [43]	Athletes (any)	Death certificates with ICD-9 code for NDs	Deaths for all causes 1982–1991, and proportion of NDs deaths in all occupations	occupation and industry coded as “writers, artists, entertainers, and athletes” according to the 1980 Bureau of the Census classification	****	**	*	265 deaths (white female writers, artists, entertainers, or athletes)/130420 total deaths	–	AD PMR: 135 (95% CI 111–162)
CASE–CONTROL STUDIES
Park 2005 [44]	Athletes (any) dancers	Death certificates with ICD-9 code for NDs	Deaths for all other causes (excluding CNS neurologic, degenerative, and neoplasms of brain, lymphatic, or hematopoietic tissue)	Certificates, coded according to the 1980 Bureau of Census (BOC) system	***	**	**	Pre-senile dementia 27,374/2,614,346 AD 47,783/2,614,346 (total n° of deaths 2,614,346)		MOR (95%CI) presenile dementia: 0.93 (0.46–1.66) AD: 1.46 (0.96–2.10)

MLB US Major League Baseball, LBD Lahman Baseball Database, HR hazard ratio, IRR incidence rate ratio, SMR standardized mortality ratio, SIR standardized incidence ratio, PMR proportionate mortality ratio, MOR mortality odds ratio, CI confidence interval, U underlying, U&C underlying and contributing