A systematic review on the risk of neurodegenerative diseases and neurocognitive disorders in professional and varsity athletes

Previous epidemiological studies reported an association between professional sport and neurodegenerative diseases (NDs) and neurocognitive disorders (NCDs). These include a range of conditions primarily affecting brain and central nervous system (CNS) functioning [1, 2]. NDs are multifactorial, still incurable, and debilitating conditions resulting in the progressive degeneration and/or death of nerve cells [1]. The pathophysiological mechanisms underlying these conditions can differ, causing deficits ranging from memory loss and cognitive impairment to movement and gait deficits and impairments in language and breathing abilities [3]. Amyotrophic lateral sclerosis (ALS) represents the most common motor neuron disease (MND) and described as a rare progressive neuromuscular disease, typically causes the degeneration of both the upper and lower motor neurons but may also include cognitive and behavioral symptoms [4‐6]. Parkinson’s disease (PD) is primarily characterized by a specific set of motor symptoms, including tremor, rigidity, and bradykinesia, and by an impairment of balance and postural reflexes. It can also cause several non-motor symptoms, such as gastrointestinal and urogenital disorders, and psychiatric symptoms, including depression [7]. NCDs is a broad term for defining several different chronic and progressive conditions affecting the ability to think, remember and make decisions. Symptoms may vary according to the type and severity of disease, and include memory loss, attention and communication deficits, and impairments in the ability to reason and solve problems [8, 9]. According to the DSM-5 dementia and mild NCD (MCI) are the main categories of NCDs. As for dementia, some subtypes are included among NDs, as they are caused by neurodegenerative mechanisms, while some types of dementia and other NCDs can be due to other causes, including secondary causes as intoxication or trauma [2]. The most common cause of ND dementia is Alzheimer's disease (AD), which is characterized by memory loss and specific deficits in language, sensory, motor and/or executive functions [8, 10]. Among the risk factors for dementia, the most known are genetic factors and aging. However, gender, educational level, and some health conditions are also included among risk factors [11].

Some studies investigated previous participation in some contact sports as a potential risk factor for neurodegenerative diseases (ND) and neurocognitive disorders (NCD) [12‐14] focusing on pathological changes in former players [15, 16] in particular, soccer [14, 17‐21] and American football players [12, 13]. Some small studies and reports have suggested that sports involving repeated traumatic brain injuries (TBI) may cause a long-term risk of NDs [22]. When considering specific NDs or NCDs, a relatively recent systematic review reported a significantly higher frequency of ALS in former professional and varsity soccer, and American football players [23]. The most frequently investigated potential sport-related risk factors for ALS and for all other NDs and NCDs, include multiple trauma and concussion [24, 25] along with oxidative stress [26] and exposure to toxic substances/chemicals (e.g., pesticides) [27, 28] or drugs/dietary supplements [28]. However, as for ALS, a more recent hypothesis also suggested that former athletes might share a phenotype causing simultaneously a higher athletic predisposition, with lower body mass index (BMI) and cardiovascular risk and higher physical fitness, and a higher risk of ALS [28, 29]. Another relatively recent systematic review, specifically focusing on long-term consequences of repeated concussions, reported that former participants in contact, collision, and combat sports seem to be at an increased risk of cognitive impairment and other mental health problems [30]. The review reported a higher frequency of cognitive decline and some NCDs and NDs in former professional athletes, considering NDs/NCDs along with diseases of the nervous system and sense organs, and other mental health conditions, including suicide and cognitive decline, hypothesizing multiple trauma and head injuries as potential causes of such condition. Multiple concussion and trauma and their potential role in the long-term onset of NDs or NCDs have been widely investigated, and are still a matter of debate, particularly in relation to popular sports such as soccer and rugby [31]. However, even though repeated head trauma has been linked to irreversible cognitive deficits [32‐34] the pathophysiological mechanisms underlying this potential association are still to be clarified.

Traditionally, multiple concussion and head trauma had been known to cause the so-called ‘‘punch-drunk syndrome’’ firstly descripted in 1928 in retired boxers [35] which was associated to a basic lesion due to traumatic multiple hemorrhages, and replaced few years later by the wording ‘‘Dementia pugilistica’’ term coined in 1937 and which took into consideration in its definition the hypothesis of a type of dementia induced by brain trauma [36]. In the following decades, the scientific community focused the attention on the possible association between this type of sport and NDs, suggesting an etiological link between exposure to repeated head trauma and concussion and dementia and/or PD, even hypothesizing synergistic effects between head trauma and apolipoprotein-epsilon four in patients with Alzheimer's disease [37, 38]. In recent times, although the debate is still open, a study investigated the association between the former participation in contact and collision sports such as American football, boxing, soccer, rugby, and hockey, and a higher risk of chronic traumatic encephalopathy (CTE), highlighting a strong dose–response relationship between the duration of American football played and this pathology [39].

Based on all these hypotheses, evidence seems to suggest an association between former participation in professional sports, and mainly in contact and collision sports, with a higher risk of NDs or NCDs. Therefore, the aim of this systematic review (SR) was to gather all available epidemiological evidence on former participation in any type of sport, at a professional and varsity level, as a potential risk factor for NDs or NCDs.

This SR was performed following the methodology of the Cochrane handbook for SRs [40] and reported based on the PRISMA statement for reporting SRs and meta-analyses [41]. Systematic searches for all available literature up to February 1, 2022, were performed on PubMed, the Cochrane Databases (that gives access to: Cochrane Database of Systematic Reviews, Embase, PubMed, CT.gov, ICTRP and CINAHL), and ISI Web of Knowlegde (that gives access to: Science Citation Index, Social Sciences Citation Index, Arts and Humanities Citation Index, Emerging Sources Citation Index, Conference Proceedings Citation Index–Science, Conference Proceedings Citation Index-Social Sciences and Humanities, Book Citation Index–Science, Book Citation Index–Social Sciences and Humanities, Current Chemical Reactions, and Index Chemicus). The following terms were used: ((neurodegenerative* OR neuro-degenerative* OR “amyotrophic lateral sclerosis” OR ALS OR “motor neuron disease” OR “motor neuron diseases” OR “motor-neuron disease” OR “motor-neuron diseases” OR parkinson* OR parkinsonism* OR Alzheimer* OR dementia* OR encephalopath* OR ((cognitive OR neurocognitive OR neuro-cognitive) AND (impairment* OR degenerati* OR disord*))) AND (((professional* OR competitive* OR competition* OR elite* OR varsity* OR college* OR “high school” OR “high-school”) AND (sport* OR player*)) OR athlete* OR rider* OR fighter* OR boxeur* OR boxer* OR runner* OR skier* OR rower* OR “martial art” OR “martial arts” OR “martial artist” OR “martial artists” OR “combat sport” OR “combat sports” OR “players”)). Further relevant articles were also identified by reviewing the references of considered studies. No limitations were applied for date of publication nor study design; only studies published in English were considered. Retrieved literature was screened by two independent reviewers (GB, PP), and studies were selected based on their pertinence and relevance to the topic of the review. The full texts of selected studies were assessed for inclusion by applying the following predefined eligibility criteria. Potential disagreements were resolved by discussion or, where needed, involvement of a third reviewer (EL). All case–control and cohort studies addressing the possible association between any types of sport practiced as a varsity and/or a professional athlete and the risk of NDs and/or NCDs, were included. Since studies on this topic are performed using any type of available data, including data from structured information sources, ranging from death certificates to biological tissue banks, we included all studies reporting a diagnosis of any of the considered condition based on any type of criteria, including clinical criteria and neurobiological criteria. Therefore, all studies reporting by any means (e.g., death certificates, clinical records) a clinical, neurobiological or neuropathological diagnosis of ALS or motor neuron disease (MND), dementia of any type or mild cognitive impairment (MCI), PD, CTE at any state of disease and with any clinical pattern were also included considering the wider diagnostic category. Case reports, case series, non-systematic or narrative reviews, letters, commentaries, and editorials were excluded. Only epidemiological studies (i.e., cohort and case–control) reporting data on the frequency of any of the considered NDs and/or NCDs were included.

For studies analyzing the association between professional and/or varsity sports and ALS, we updated a previously published systematic review [23] therefore only studies published from January 2016 to February 2022 were included.

Studies meeting the eligibility criteria were qualitatively assessed using the Newcastle–Ottawa Scale (NOS) for quality assessment of non-randomized studies [42]. The NOS scale is structured in 8 items assessing the adequacy of 3 major domains: selection of the study group, comparability of the groups, and ascertainment of the exposure of interest for case–control studies or of the outcome of interest for cohort studies. A maximum of 9 stars can be assigned to each assessed study, and specifically a maximum of 4 stars for the Selection domain, 2 stars for the Comparability domain, and 3 stars for either the Outcome or the Exposure domain. The potential presence of other biases and/or methodological flaws was also considered in assessing the included studies. Data were extracted using structured tables. Specific information was gathered on the type of sport considered by each study, along with how the enrolled population of athletes was defined, and on the diagnostic criteria and procedure used to define the identified cases. Moreover, for cohort studies, specific information was also extracted on the source of the reference population. Gathered information also included study design, year of publication, characteristics of the included population, definition of the exposure(s), diagnostic criteria, length of the follow-up, and results for each analysis.

Structured search on bibliographic databases yielded 11,189 articles. Two further studies were identified by reviewing the bibliographic references of included studies [43, 44]. After selection based on the pertinence and relevance to the topic of the review, 11,191 records were excluded, leaving 47 articles meeting the selection criteria. The full texts of the selected articles were assessed for inclusion, and 30 further articles were excluded as 3 were case reports [17, 19, 45], 1 review [46], 1 reported participation in sports as a leisure activity [47] 2 reported data only on an overall measure of physical activity [48, 49] 2 included only athletes who had ≥ 1 head/brain injury [50, 51] 8 did not report any measure of risk of NDs and/or NCDs [20, 52‐58] and 13 were already included in the 2016 review [59‐71].

The flow diagram of the literature is reported in Fig. 1.

A total of 17 articles were included in the qualitative summary [43, 44, 72‐86].

Overall, five cohort studies reported data on the frequency of all NDs/NCDs included in this review in former professional or varsity athletes [74, 77, 79, 82, 84]. Considering each single pathology, 10 cohort studies [82, 84] and 1 case–control study [44] reported data on the frequency of dementia/MCI in former athletes, 8 cohort studies [43, 73‐75, 77, 79, 82, 84], and 1 case–control study [44] reported data on the frequency of PD or Parkinsonism, 1 cohort study [80] reported data on CTE, and 7 new cohort studies [77, 82‐86] and 1 new case–control study [81] reported data on the frequency of ALS or MND. The summary of the main characteristics, quality assessment and results from the included studies were reported according to each considered outcome in Tables 1, 2, 3, 4, and 5.

Overall, the present systematic review reported a higher frequency of NDs/NCDs in former professional athletes. Specifically, published studies showed an increased frequency of any type of NDs/NCDs in former soccer players [82, 84] and American football players [74] even when compared to other non-contact sport athletes [79]. When considering specific NDs and NCDs, data from the update of the previous SRs confirmed a higher frequency of ALS or MND in former soccer players [82‐85]. Gathered data also reported a significantly higher risk of dementia and AD in former soccer players [82, 84] and a higher risk of MCI in former American football players, mainly associated to a higher risk of concussion [72]. Moreover, data also showed a significantly higher risk of PD in former soccer and American football players [74, 75, 79, 82, 84] and a significantly higher risk of CTE in former boxers and in former American football players [80].

Some studies reported either non-significant results or no increase in risk in some categories of athletes. However, most of these studies were of low or relatively low methodological quality and included very small or relatively small samples defining cases based on self-reports or non-standardized/validated tools. Moreover, some studies only compared different types of sports (i.e., contact vs non-contact, football vs non-football). The only large and high-quality study [44] reporting no difference in the mortality rate for any of the considered NDs and NCDs considered as exposed participants classified as “athlete/dancer” based on the occupational census, which is a wider definition that could lead to potentially underestimating results. Most of the included studies attempted to take into consideration potential confounding factors by either matching non exposed subjects or controls by at least the main characteristics (e.g., gender, age, ethnicity), with some of them also stratifying for potential interacting factors such as length of career and type of sport or position/role played. However, the factors considered for adjustment where heterogeneous across some of the included studies, and most studies only include male professional players due to the availability of data. This further affected the direct comparison of data. Previous reviews suggested that the higher risk of NDs observed in some specific categories of former athletes could be sport specific. However, no studies on sports other than soccer, American football and marginally boxing, were available. Therefore, conclusion can only be based on available studies, as no data are currently available on different sports. Further large studies should be performed on other sports, mainly in other contact sports (e.g., boxing, martial arts) and sports requiring running/speed roles (e.g., rugby, hokey, etc.), should be carried out, as these types of sports resulted to be associated with the highest risk of NDs and NCDs.

As for ALS and MND, our results confirmed the higher risk observed in formed soccer and American football player reported in the previously published SR [23]. Included studies reported an increased risk of ALS in soccer and American football players also confirmed an earlier onset of ALS in soccer players, and a higher frequency of ALS in players with a longer career, in specific positions or roles, and playing in main leagues [82‐84, 86]. Therefore, based on results from the previous SR and the present update, soccer and American football players seem to be at risk of a specific, more severe, type of ALS or MND, with earlier onset (< 45 years) and shorter survival.

Our data also showed an increased risk of dementia, AD and MCI in former soccer and American football players. This increase in the frequency of these conditions was linked, in some studies, to an increased exposure to brain injuries and concussions, specifically in former American football players. In this last category of athletes, a higher frequency of CTE was also observed. As expected, the risk of CTE resulted the highest in former boxers, who were among the first categories of athletes in whom the formerly called “punch drunk syndrome” and “dementia pugilistica” were investigated [99, 100].

Moreover, a higher risk of PD and Parkinsonism was observed in former soccer and American football players, which was not linked to specific roles or positions. This increased frequency of PD, as for dementia and cognitive decline, was also generally linked to an increased exposure to brain injury and trauma.

On this basis, our results seem to suggest that competitive athletes, and specifically those participating in contact sports, are exposed to a higher risk of trauma, concussion, and subsequent brain injuries, thus leading to a higher risk of developing cognitive outcomes or PD/Parkinsonism later in life. However, professional athletes appear to also be at a higher risk of what seems to be a specific form of ALS/MND. Based on our data and data from our previous SR, in fact, the type of ALS/MND observed in former soccer players seem to have a much earlier, bulbar onset, and a much shorter survival. This increased risk has been linked to several possible genetic and environmental risk factors. Several gene variants are associated with a susceptibility to ALS/MND, and when considering environmental risk factors, the exposure to heavy metals, trace elements, solvents, and other volatile organic compounds (e.g., pesticides and radiation) or toxins (e.g., cyanobacteria), has been linked to a higher risk of developing this condition. As for some life-style factors, the widespread use or abuse of branched-chain amino acids in professional athletes has been hypothesized to be potentially linked to a higher risk of ALS/MND [28]. This led to the specific hypothesis that professional athletes might share a specific phenotype linked to both a better predisposition to high-level sport and a risk of developing ALS/MND. However, all these genetic, environmental, and lifestyle risk factor are shared risk factors for all the considered NDs and NCDs. The exposure to some pesticides, including organophosphorus compounds and carbamates, in fact, has been reported to be associated not only to an increased risk of ALS, but also to a higher risk of AD and PD [101]. Several other compounds, such as solvents and neurotoxic agents, have also been proven to be shared risk factors for ALS and other NDs and NCDs, in particular PD [102]. Moreover, several polymorphisms and gene variants have been reported to be associated to a higher frequency of specific NDs, and some have also been investigated as involved in the epigenetic mechanisms leading to the onset of the disease [103‐106]. The exposure to several lifestyle factors, including smoking, physical activity, and dietary habits along with the use/abuse of some dietary supplements, have also been proven to have a role in the etiological mechanisms of some NDs and NCDs. The evident interlinking and interaction among these factors, and their being shared risk factors for several NDs and NCDs underline the need to start considering neurodegeneration as a process starting with an early, subclinical, phase, which might be common to all NDs. The neurodegenerative process in NDs, in fact, has been proven to start several decades before the onset of symptoms [107]. This long latent period, along with the shared association between several risk factors and some specific NDs suggests that this early process might be common to different NDs and only in a subsequent phase cause specific symptoms according to a subjective susceptibility and specific interactions between risk factors [3, 108]. Thus, based on this hypothesis, research should focus on investigating the interacting role of genetic, environmental, and lifestyle risk factors in this early phase of neurodegeneration to explore the possibility of defining potential preventive strategies [106]. Moreover, former professional athletes should be further investigated, as these risk factors and their interactions, as well as the risk of some specific NDs, appear to be significantly higher in this population. Studies on this population should also aim at clarifying whether the risk might be specific to some type of sports or roles, and whether the clinical characteristics of the NDs are typical of this population.

Our SR has some limitations. The main limitation is the relatively small number of included studies. Studies on this topic are relatively few and are limited to a very small number of sports, mainly soccer and American football. This limits the generalizability of results and the reliability of results. The small number of published studies might be due to the difficulty of investigating a relatively rare exposure as having been a professional athlete, and the subsequent difficulty of enrolling large samples of professional athletes. Moreover, when considering ALS, the difficulty of carrying out large studies might also be due to the rarity of the disease, which would require very large samples. Furthermore, NDs and NCDs are diseases whose typical onset is later in life, thus longitudinal studies require covering very long observation periods. Therefore, prospective studies are almost unfeasible, and retrospective studies might be very difficult to carry out in absence of effective and efficient informative systems and databases. However, the use of data from information systems, such as death certificates and hospital records, largely underestimates the frequency of NDs and NCDs, thus limiting the quality of studies, as does the use, inevitable in these systems, of different diagnostic criteria. A further limitation, due to the above-mentioned difficulties, is that the considered cohorts of athletes should be intended as mostly the same cohorts observed in different time periods. This prevents the comparison and the cumulative analysis of data, allowing only to observe a possible trend in the risk of NDs and NCDs over time within the same cohort. Further high-quality studies including athletes participating in different sports would therefore be extremely useful to assess whether the observed higher risk might be sport-specific or shared with other types of sport, and to further investigate possible characteristic or risk factors suggesting specific pathophysiological mechanisms explaining the observed increase in the risk of NDs and NCDs.