We report 14 retired footballers who developed dementia in later life. They were all skilled headers with half playing in positions that required frequent heading of the ball. Their playing career spanned two to three decades and most of them started training and heading the ball during childhood. The six post-mortem cases showed mixed pathologies, including criteria-defined CTE in four cases, AD and TDP-43 pathologies in six, CAA in five, and hippocampal sclerosis in two, and others had vascular pathology (N = 1), CBD (N = 1), and LBD (N = 1), and all would have contributed synergistically to the clinical manifestations. The frequency of CTE pathology in four of six cases in the present series represents a significant excess when compared with the 12% average background rate of incidental CTE pathology in elderly individuals with or without neurodegenerative disorders in our QSBB survey [25]. We hypothesize that CTE and, probably, AD and TDP-43 pathologies in these retired footballers are related to their past prolonged exposure to repetitive subconcussive head impacts from heading and head-to-player collisions.

5 13 14 31 4 Case report No. of cases Age at death Years playing football Football playing position; level History of concussion Alcohol or substance abuse Age at symptom onset Disease duration Presenting symptoms Behavioural and mood changes Other clinical features Final clinical diagnosis Septal abnormalities Neuropathological diagnoses McKee et al. [31] 1 29 26 NK but headed the ball frequently since age 5; semi-professional NK – 27 2 Limb weakness – – ALS NK CTE MND TDP-43 Hales et al. [14] 1 80 16 Forward/striker; professional NK NK 70 10 Executive dysfunction and memory loss Irritable, obsessive Parkinsonism AD NK CTE AD (mild) TDP-43 HS Grinberg et al. [13] 1 83 21 Centre-back; professional – – 67 16 Memory loss Shouting spells Slow and abnormal gait AD C CTE AD (Intermediate) TDP-43 HS SVD (Mild) Bieniek et al. [4] 1 73 NK NK; amateur NK Alcohol 66 7 Memory loss NK NK AD NK AD Ling et al. (present series) 6 (PM cases) 72 (mean) 25.7 (mean) Centre-back or centre-forward (N = 3) but all headed the ball frequently; 5 professional, 1 committed amateur In 5 footballers (only 1 episode in each) – 64.3 (mean) 7.7 (mean) Memory loss and/or behavioural changes Yes (all) Parkinsonism (3) AD/FTD ± PD F (6) C (1) CTE (N = 4) AD (N = 6) TDP-43 (N = 6) CAA (N = 5) HS (N = 2) SVD (N = 1) CBD (N = 1) LBD (N = 1) Our four CTE cases add to the four case reports of footballers in the literature, three of whom had histological evidence of CTE (Table). Of the three cases with histological evidence of CTE, two presented with an AD-like dementia and died in their early 80s [] and another 29-year-old had amyotrophic lateral sclerosis (ALS) [], substantiating the notion that playing football is a risk factor for CTE. The fourth footballer was an amateur and had AD pathology [].

There was no histological evidence of the previous diffuse axonal injury typically observed in acute traumatic brain injury, whereas all six of our post-mortem cases had septal fenestration and one also had cavum septi pellucidi. The finding of septal fenestration is supportive of their past histories of chronic repetitive head impacts from playing football [29]. The rate of septal abnormalities in our footballers is greater than the non-boxer general population, in whom septal fenestration and fenestration with cavum septi pellucidi were 6 and 3%, respectively, in autopsy [9]; while these macroscopic features were found in all 11 professional boxers in the Corsellis series, except in one case in whom cavum septi pellucidi was not observed, but ‘the septum was nevertheless fenestrated to destruction’ [9]. Concussion rate was limited in six of our 14 cases to one episode during their careers, which is a typical finding in professional footballers [19, 36] with one study reporting 74 episodes of concussion in 39 out of 72 male professional footballers [3]. Other potential repetitive head impacts outside the football field, including amateur boxing (N = 2), seizures (N = 2), and postural instability with frequent falls in later life (N = 4, Table 2), would have also contributed to the risk of CTE. The notion that prolonged exposure to repetitive subconcussive head impacts can lead to CTE is supported by the large Boston series, in which 16% of contact sport athletes and military veterans had CTE pathology but reported no concussion [32, 38].

Antemortem prediction of CTE pathology is difficult as all cases in the present series had a clinical presentation resembling either AD or FTD with the majority presenting in their 50s and 60s. In hindsight, motor impairments and early behavioural changes may serve as clinical pointers in the three cases with advanced CTE pathology (Cases 2, 5, and 6). Professional footballers also have an increased risk of developing ALS, but it was not a feature in our cases [8]. Mixed pathologies are the rule rather than exception in older individuals with dementia [1, 22]. For example, CTE, AD, TDP-43, and hippocampal sclerosis were also observed in the other two case reports of retired professional footballers [13, 14] (Table 5). We assume that the majority of our cases in the clinical group would have had mixed pathologies, including CTE. In CTE, AD and TDP-43 pathologies are common concomitant findings (Tables 3, 5) [32], and are increasingly considered as part of the CTE pathological entity, especially in older individuals [30] with the likelihood of Aβ deposition increased by APOE4 allele status [39]. The family history of dementia noted in our two oldest post-mortem cases (Cases 2, 6) may support a genetic predisposition. Unfortunately, frozen tissue and DNA were not available for genetic analysis in our cases.

This clinico-pathological series started as a surveillance that had spanned three decades, which was initiated and undertaken by a consultant psychiatrist (DDRW) with an interest in understanding the potential link between playing football and long-term neurodegenerative consequences. This descriptive study has a small number of cases without detailed psychometric, neuroimaging, or genetic data, yet its prospective collection of demographic data, playing and concussion history from close relatives, and regular surveillance from out-patient follow-up minimise case selection and recall bias in contrast to other post-mortem CTE series which relied on retrospective data collection only. Football is the most popular sport worldwide with over 250 million players at all levels. Although this study does not provide a firm causal relationship between CTE and exposure to repetitive head impacts from playing football, our findings support the pressing need to instigating large-scale studies to identify at risk groups of footballers, including age of exposure [37], which will justify for the implementation of protective strategies and education of current players. The significance of heading and the weight of the football remain elusive since the threshold of the impact force required to trigger the pathological process of CTE is currently unknown and repetitive head impacts in football are not limited to heading of the ball as footballers are also frequently exposed to head-to-player collisions [5]. All our cases were exposed to the heavier leather football used before the 1980s, which weighed 450 g and in wet condition became 25% heavier with the corresponding increased impact on contact with the head [2]. Nevertheless, lighter balls travel faster and may result in the same net force on head impact [19]. The assumption is that the lighter synthetic ball may also put modern footballers at risk of subconcussion and CTE is supported by the radiological findings of abnormal white matter microstructure in young footballers who headed the ball frequently [26] and the post-mortem finding of CTE in a 29-year-old footballer with ALS [31].

Future prospective longitudinal studies with radiological (including tau and amyloid PET scans and diffusion-tensor MRI), psychometric, biochemical, CSF, and genetic data in contemporary professional footballers with control group (e.g., athletes without increased risk of repetitive head impact), objective quantitative measure of head impact force, and clinical and pathological follow-up are required to confirm the potential causal relationship between CTE and exposure to repetitive head impacts from playing football.