Brain pathology and symptoms linked to concussion history: beyond chronic traumatic encephalopathy

Abstract

Repeated head trauma acquired through sports injuries has been associated with the development of long-term disabling symptoms that negatively impact the quality of life. In this retrospective case series, 52 male former professional athletes involved in contact sports and with a history of multiple concussions were evaluated for chronic clinical symptoms and post-mortem neuropathological diagnoses. The clinical symptoms of 19 cases were examined in greater detail for symptom type, severity and duration. Information on neurological, psychiatric and physical symptoms, substance use profiles and concussion histories was obtained from the athletes’ next of kin and assessed in relation to post-mortem neuropathological diagnoses. Cases were categorized into three different neuropathological groups: no major neuropathological findings, the presence of only chronic traumatic encephalopathy (CTE) and the diagnosis(es) of other neurodegenerative diseases. Age at death and the presence of DNA damage in the post-mortem brains were analysed for correlation with the clinical symptoms. In this case series, 14/52 (26.9%) cases (mean age 48.2 ± 11.4) had neuropathological evidence of low-stage/low-burden CTE. A total of 11/52 (21.2%) cases (mean age 38.7 ± 12.7) presented a similar profile and severity of behavioural symptoms to those with CTE, despite the lack of significant post-mortem neuropathological findings. A total of 27/52 (51.9%) cases (mean age 75.5 ± 8.7) presented with complex post-mortem neurodegenerative diagnoses, including Alzheimer’s disease and other mixed pathologies, and clinical symptoms associated with language, memory and sensory dysfunction. The presence of DNA damage in the brain was found in all neuropathological groups, predominantly in the ependymal lining of ventricles, and phosphorylated histone H2AX staining was correlated with higher age at death (r = 0.59) and symptoms of language dysfunction (r = 0.56). Findings from our case series suggest that post-concussive symptoms are not driven by CTE. Our findings show that proteinopathies alone may not account for the complexity of the clinical manifestations and suggest the possibility of other drivers, such as DNA damage, as potentially useful markers of brain trauma. Broadening the search for biological markers that reflect the effects of brain injury, even when proteinopathy is not observed, and taking a symptom-driven approach are therefore advised.

Fifty-two professional male athletes with history of repeated mild traumatic brain injury were evaluated for clinical symptoms and post-mortem neuropathology. Despite experiencing chronic clinical symptoms, many of the athletes lacked a neurodegenerative diagnosis, like chronic traumatic encephalopathy, suggesting markers beyond proteinopathy are needed for a holistic examination of neurotrauma.

Graphical Abstract

Graphical Abstract.

Introduction

Traumatic brain injury (TBI) is a leading cause of death and disability worldwide, affecting an estimated 69 million individuals each year.¹ Mild TBIs (mTBIs), or concussions (terms used interchangeably in this article though not strictly synonymous), are common in athletes involved in contact sports such as football, hockey, mixed martial arts, professional biking and skiing.² Symptoms after a concussion are heterogeneous and can include sensory, cognitive and neuropsychiatric complaints such as headaches, dizziness, poor concentration and anxiety.³ The symptoms are typically resolved within several weeks to months. While these symptoms are transient for some, in 10–20% of concussed individuals, symptoms can persist or emerge months or years after injury.⁴ Such individuals commonly experience chronic headaches and pains, long-term cognitive impairments, mood disorders, irritability, fatigue, insomnia, visuospatial and motor dysfunction.⁵ History of mTBI has also been shown to correlate with the future development of neurodegenerative diseases and psychiatric illnesses, although a causative relationship has not been established.^6,7 At present, treatment of persistent symptoms is limited. The underlying mechanisms that drive long-term clinical presentations are currently unknown, although efforts have been made to attribute the symptoms to pathological entities such as chronic traumatic encephalopathy (CTE).⁸

CTE is described as a neurodegenerative condition associated with a history of repetitive mTBIs (rmTBIs). However, the mechanism by which this may occur is speculative, and the true prevalence of CTE is unknown, as the diagnosis can only be made by post-mortem neuropathological analysis. The current consensus criteria for a CTE diagnosis require hyperphosphorylated tau (p-tau) deposits in neurons at perivascular sites in the depths of cortical sulci.⁸ However, even for the present criteria, unblinding the raters to the clinical symptoms was necessary to reach a consensus in cases with high CTE. Moreover, experts in the field have found astroglial tau depositions provide higher specificity for a CTE diagnosis,^9,10 demonstrating that the criterion is still evolving and is yet to be universally agreed upon. As such, controversy remains whether neuronal tau is a suitable marker for capturing the clinicopathologic features observed in rmTBI. Some tau deposition is associated with normal aging¹¹ and may even have a protective role against neuronal apoptosis.¹² Further, CTE is often presented with other pathologies such as accumulation of amyloid-beta, TAR DNA-binding protein 43 (TDP-43), brain atrophy, axon degeneration and neuroinflammation that have also been linked to a history of TBI.^13-17 In our previous work, we have identified the presence of DNA damage around the ventricular zone and in the astrocytes and oligodendrocytes in grey and white matters, as well as evidence of inefficient DNA repair and cellular senescence in athletes with a history of rmTBI, including the cases with no neurodegenerative tauopathy.¹⁶ With several concurrent molecular cascades often present, it remains unclear whether a neuropathological diagnosis of CTE can explain the behavioural symptoms experienced in some individuals with rmTBI. Moreover, CTE can be difficult to distinguish from other neurodegenerative diseases,^14,18 specifically progressive tauopathies that exhibit age-related accumulation of p-tau.¹⁹ In a post-mortem study, symptomatic professional athletes often had either CTE pathology mixed with another neurodegenerative disease or neurodegenerative pathology unrelated to CTE.²⁰ Finally, with substantial coverage from the media, the general public’s opinion has preceded the science behind CTE, giving the diagnosis a universality that is yet to be verified. The lack of a diagnostic biomarker for chronic clinical symptoms often causes frustration in patients, prompting them and their families to seek an explanation, many of which prematurely attribute the clinical symptoms to CTE. As such, delineating clinical and pathologic entities in CTE remains a challenge.

In this retrospective study, we aimed to investigate the relationship between clinical symptoms and neuropathologic diagnoses of 52 male athletes involved in contact sports with a history of rmTBI. Here, exposure of interest is indicated as rmTBI, for a more precise definition of TBI exposure as opposed to repetitive head impacts (RHI). rmTBI is also referenced as repeated concussions when referring to the case series, as the term ‘concussion’ was most frequently communicated by the next of kin. Reports included the athletes’ most prominent behavioural symptoms, psychiatric presentations, presence of systemic diseases and use of substances. Their clinical profiles were evaluated within three categories of the athletes’ post-mortem neurodegenerative diagnoses: no diagnosis, CTE or complex and/or co-morbid neurodegenerative diagnoses. Detailed questionnaires obtained from 19 cases’ next of kin on their cognitive, behavioural and constitutional symptoms were used to explore the differences between clinical dynamics accompanied by the neurodegenerative diagnoses. Comparing the symptom profiles between cases with no diagnoses, CTE, and other complex neurodegenerative diseases revealed similarities among the former two groups, whereas more complex neurodegenerative disease diagnoses were accompanied by a distinct set of symptoms. Additionally, in this study, we examined DNA damage in the context of clinical symptoms post-rmTBI and whether it may serve as an alternative biological marker. DNA fragmentations, single-strand and double-strand DNA breaks (DSBs) have been shown to persist following TBI in mice, due to disruptions to the major DNA damage response pathways.^21,22 Previously, we have identified DSBs, evidence of a deficient DNA repair response, and cellular senescence in the post-mortem brains of male athletes with rmTBI history.¹⁶ In this series, DNA damage was present in most cases with symptoms, including ones with no pathology, and DNA damage burden correlated with language dysfunction. Therefore, we suggest that CTE pathology alone may not drive the clinical symptoms associated with brain trauma, and that other molecular changes may coincide or better align with the development of clinical symptoms.

Materials and methods

Human brain sample acquisition

This study was approved by the Ethics Review Board at the Hospital for Sick Children (REBs #1000059400 and #1000067940), and informed consent for study participation was provided by the next of kin. Former male professional athletes with concussion history were selected from our brain bank, consisting of individuals with TBI history from sports-related causes.

Neuropathology diagnosis and histological quantification

Details on neuropathological diagnoses and histological quantification are described in a previous study.¹⁶ Briefly, formalin-fixed paraffin-embedded (FFPE) tissue sections of post-mortem brains underwent a full neuropathological assessment according to the National Institute on Aging Association guidelines.²³ Neuropathological findings included CTE, Alzheimer’s disease, Parkinson’s disease, aging-related tau astrogliopathy, primary age-related tauopathy, corticobasal degeneration, frontotemporal lobar dementia, argyrophilic grain disease, amyotrophic lateral sclerosis and Pick’s disease. For this study, cases were grouped into three neuropathological categories: cases with no evidence of proteinopathy or other major neuropathological changes (‘None’), cases presenting with CTE (‘CTE’) only, and cases with other neurodegenerative diagnoses such as Alzheimer’s disease, complex tauopathy including aging-related tau astrogliopathy or mixed tauopathy with features of both CTE and other neurodegenerative pathology (‘NDG’). Staining of FFPE brain sections with phosphorylated H2AX (γH2AX, Ser139) and the method of double-stranded DNA damage staging are described in detail in a previous study.¹⁶ Briefly, Stage 1 was defined as cases with DNA damage limited to the ependymal lining of the lateral ventricle, Stage 2 was defined as cases with DNA damage to the ependymal lining, subpial astrocytes and grey matter peri-neural satellite cells and Stage 3, in addition to the cell types involved in previous stages, included cases with DNA damage in oligodendrocytes in white matter. Cases negative for γH2AX were defined as Stage 0.

Clinical evaluation

General clinical symptoms prior to death were noted for all athletes at the time of brain donation by means of retrospective interviews with next of kin. The next of kin was further contacted to complete a detailed symptom questionnaire on athletes’ behaviour throughout their life and prior to death, educational and occupational background information and concussion and medical histories. Of the 52 cases, 19 questionnaires were retrieved and assessed with the cases’ neuropathological diagnoses. The questionnaire was constructed in-house and consisted of background information, concussion history and questions on eight clinical symptom classes. The background information surveyed athletic career, educational history, medical history (including the presence of neurological or psychiatric systemic diseases) and information on substance use. For concussion history, the number of concussions sustained, visits to the emergency room and hospitalizations, number of losses of consciousness (LOC), age at first concussion and details on the most recent concussion were surveyed. Clinical symptoms were categorized into eight classes: memory, executive function, behaviour/emotional well-being, language, visuospatial, motor, sensory and constitutional function. Details on symptoms reported for each class are in Table 1. The next of kin indicated whether the athlete exhibited a symptom with a yes/no or ‘I don’t know’. If selected ‘yes’, the start time of the symptoms was inquired ranging from ‘the last 3 months’ to ‘longer than 10 years’. The severity of the symptoms was indicated with either ‘mild’, ‘moderate’ or ‘severe’, and additional comments were provided where necessary. For clinical symptom quantification, the absence of a symptom was scored as 0, and its presence was scored as 1 for ‘mild’, 2 for ‘moderate’ and 3 for ‘severe’. Mean composite score for each symptom class was calculated for every case and analysed per neuropathology group.

Table 1.

Symptoms evaluated for each clinical symptom class

Clinical symptom class	Symptoms
Executive dysfunction	Difficulty planning, multi-tasking, problem-solving and concentrating; mental rigidity, impulsivity, poor judgement.
Memory	Difficulty with memory, remembering recent and remote events.
Behaviour/emotional well-being	More/less emotional, change in personality, apathy, disinhibition, repetitive behaviour, agitation, violence, depression, anxiety, hallucinations, impaired hygiene, change in eating habits, restlessness, trouble sleeping.
Constitutional	Light-headedness, headaches, vertigo, fatigue, muscle/joint aches, nausea, urinary/bowel incontinence.
Motor	Impaired balance, falls, tremor, muscle weakness, involuntary movements, difficulty using hands/feet/utensils, changed handwriting.
Language	Difficulty finding words, poor articulation, stutter, incorrect word usage, more/less speech output; impaired language/word comprehension, reading, writing, spelling.
Visuospatial	Lost in familiar scene, difficulty seeing and recognizing faces, impaired object perception.
Sensory	Change in vision, somatic sensation, numbness, tingling, pains.

Statistical analysis

Age at death, years of exposure, γH2AX reactivity scores and clinical symptom scores between neuropathology groups were evaluated with Kruskal–Wallis (KW) tests followed by Dunn’s multiple comparisons tests. Age at death for cases segregated by the presence of a CTE diagnosis was evaluated with a Mann–Whitney U test. Mean and standard deviation (SD) values are presented in the text along with statistics. Spearman’s correlation [correlation coefficient (r), 95% confidence interval (95CI)] was used to statistically assess the relationship between age at death, γH2AX reactivity and symptom scores. Statistics and figures were generated using GraphPad Prism 9. Alpha level was set to 0.05.

Results

Neuropathological findings in male athletes with history of rmTBI

Fifty-two male athletes with rmTBI were assessed for post-mortem neuropathological diagnoses (Table 2). Based on the given diagnoses, cases were subsequently divided into three neuropathological categories. Cases were assigned to the ‘None’ category (n = 11) if they did not present with any neurodegenerative disease or abnormal protein accumulation. In this category, one case showed evidence of diffuse axonal injury (DAI) and transverse myelitis, one with peri-vascular haemosiderin, and one case showed cerebral oedema, hypoxic–ischaemic injury and haemorrhagic infarct. Next, cases were assigned to the ‘CTE’ category (n = 14) if they presented with CTE in accordance with the 2021 Second NINDS/NIBIB Consensus Meeting criteria.⁸ A case in this group showed evidence of brain atrophy, one case exhibited microinfarcts, microhaemorrhages and global hypoxia, one case of DAI, one of argyrophilic grain disease and état criblé, one case of white matter microbleeding and axonal damage, one with axonal injury and one case with hypertensive blood vessel changes. Finally, cases diagnosed with a neurodegenerative diagnosis besides CTE, including complex and co-morbid proteinopathy diagnoses, were categorized into group ‘NDG’ (n = 27). In this group, 19/27 (70.4%) cases presented with Alzheimer’s disease: 8/19 (42.1%) in Braak Stage VI, 5/19 (26.3%) in Stage V, 4/19 (21.1%) in Stages III and IV and 2/19 (10.5%) in Stage II. A total of 5/27 (18.5%) cases showed CTE in addition to other pathologies, totalling 19/52 (36.5%) cases of CTE in the entire series. Other diagnoses in the ‘NDG’ group included 4/27 (14.8%) with Parkinson’s disease, 4/27 (14.8%) with aging-related tau astrogliopathy, 3/27 (11.1%) with frontotemporal lobar dementia, 2/27 (7.4%) with progressive supranuclear palsy, 2/27 (7.4%) with progressive Lewy body dementia and one case each of amyotrophic lateral sclerosis, corticobasal degeneration and Pick’s disease. There were 6/27 (22.2%) cases of vasculopathy, 5/27 (18.5%) with white matter infarctions, 4/27 (14.8%) cases of amyloid angiopathy, 2/27 (7.4%) presented with TDP-43-positive inclusions and one case each of cavernous periventricular haemangioma, hepatic encephalopathy and cavum septum pellucidum.

Table 2.

Clinicopathological presentations of 52 cases of male athletes with history of rmTBI

Neuropathology, n	Nonea (n = 11)	CTEb (n = 14)	NDGc (n = 27)
Other neuropathological observations	DAI, cerebral oedema, ischaemic and haemorrhagic infarctions, microinfarctions	Atrophy, DAI, argyrophilic grain disease, global hypoxia, microinfarctions and microhaemorrhage, white matter microbleeding, axonal damage	Alzheimer’s disease (several VI/VI), Parkinson’s disease, ‘high’ CTE, aging-related tau astrogliopathy, progressive supranuclear palsy, frontotemporal lobar dementia, corticobasal degeneration, progressive Lewy body dementia, Pick’s disease, white matter and vascular infarctions, hepatic encephalopathy, amyloid angiopathy
Mean age at death (SD)	38.7 (12.7)	48.2 (11.4)	75.5 (8.7)
Mode of exposure to TBI, n
Football	3	4	19
Hockey	6	4	4
Other (ski, cycling, bull riding, mixed martial arts, military)	2	4	2
Unknown		2	2
Mean years of exposure (SD, n)	16 (3.6, 3)	19.6 (6.7, 8)	15.3 (4.0, 15)
Cause of death, n
Suicide	4	7	2
Acute physical trauma/accidental	2 (1 overdose)	2 (acute vascular injury, overdose)
Stroke, cerebral vasculitis, cardiac arrest	2 (heart failure, stroke)	1 (cardiac arrest)	1 (head carcinoma)
Dementia associated			20 1 Parkinson’s disease, 1 amyotrophic lateral sclerosis related
Complications with existing condition	1 (transverse myelitis)	1 (metastatic lung carcinoma)	1 (organ failure, peritonitis, cirrhosis)
Unknown	2	3	1
Clinical symptoms reported, n
Depression/anxiety	7	4	4
Mood disorder		5	1
Behavioural changes/cognitive dysfunction	2	1	3
Dementia	1	1	20
Other	Falls, domestic violence, decreased physical tolerance	Irritability, psychotic episodes, fatigue, headaches	Visuospatial and recall decline, Parkinson’s disease and amyotrophic lateral sclerosis symptoms, aphasia, panic attacks, aggression, anger

The number of cases is mutually exclusive, except for the clinical symptoms section, where cases with dementia correspond to some cases with other symptoms. CTE, chronic traumatic encephalopathy; DAI, diffuse axonal injury; NDG, neurodegenerative disease; rmTBI, repeated mild traumatic brain injury.

^aNone—no presentation of CTE or other major neurodegenerative diseases.

^bCTE—CTE as per 2021 NINDS/NIBIB consensus meeting criteria.

^cNDG—complex/multiple diagnoses of neurodegenerative diseases.

Ages at death (KW statistic = 36.13, P < 0.001) for cases grouped into ‘None’ (mean 38.7, 12.7 SD) and ‘CTE’ (mean 48.2, 11.4 SD) were similar (Dunn’s test P = 0.95), whereas the ‘NDG’ group’s age at death was significantly higher than both groups (mean 75.5, 8.7 SD, Dunn’s tests P < 0.001). Cases segregated by the presence of CTE (n = 19, mean 56.4, 17.3 SD) or not (n = 33, mean 62.7, 20.0 SD) did not differ in age significantly (U = 248, P = 0.22). Most athletes played American football (26/52, 50%) or hockey (14/52, 26.9%) professionally while some athletes participated in fighting sports, BMX biking or bull riding or served in the military in addition to sports history. The average length of athletic careers was similar for all groups (KW = 2.91, P = 0.23) ranging from 15.3 to 19.6 years (3.6–6.7 SD). The cause of death was mostly attributed to suicide or accidental death (cases of physical trauma or overdose) for the ‘None’ (6/11, 54.5%) and ‘CTE’ (9/14, 64.3%) groups. One case in the ‘None’ group had a fatal stroke (1/11, 9.1%), and one case in the ‘None’ (1/11, 9.1%) and one case in the ‘CTE’ (1/14, 7.1%) groups died of cardiac arrest. The cause of death in the ‘NDG’ group, however, was predominantly complications with dementia (20/27, 74.1%) or other progressive neurodegenerative diseases such as amyotrophic lateral sclerosis (1/27, 3.7%) or Parkinson’s disease (1/27, 3.7%). In this group, there were two cases of suicide (2/27, 7.4%) and one case of death by age-related diseases such as organ failure (1/27, 3.7%).

Retrospective interviews with the next of kin

Information regarding clinical symptoms was obtained from retrospective interviews with the next of kin at the time of brain donation (Table 2). The clinical symptoms reported primarily entailed psychological and behavioural changes. The most reported symptoms across the entire case series were mood disorders (21/52, 40.4%), namely depression and anxiety. Segregated by neuropathology group, mood symptoms were reported frequently in both the ‘None’ (7/11, 63.6%) and ‘CTE’ (9/14, 64.3%) groups and in several ‘NDG’ cases (5/27, 18.5%). Two cases of behavioural changes (emotional turmoil, onset of depression) were reported for the ‘None’ group (2/11, 18.2%), and one case in the ‘CTE’ group experienced psychotic episodes (1/14, 7.1%). The ‘NDG’ group’s symptoms were largely attributed to dementia (20/27, 74.1%), as opposed to a single case in each ‘None’ (1/11, 9.1%) and ‘CTE’ (1/14, 7.1%) groups, and affected memory, language and cognitive function. Four cases experienced symptoms of Parkinson’s disease (4/27, 14.8%) and one amyotrophic lateral sclerosis (1/27, 3.7%).

Systemic disease, psychiatric symptoms and substance use profiles

Along with the clinical presentation, the presence of systemic disease in the athletes, their history with substance use and psychiatric symptoms were obtained from the next of kin interviews (Table 3). Despite the reports acquired from a comparable number of cases across groups (11–17 cases per group), cases with no diagnosis or with CTE were less frequently reported to have any systemic disease. Several were reported to have sleep apnoea (‘None’ 2/11, 18.2%; ‘CTE’ 1/12, 8.3%), hypertension (‘None’ 1/11, 9.1%; ‘CTE’ 2/12, 16.7%), heart-related and circulatory diseases (‘None’ 3/11, 27.3%; ‘CTE’ 3/12, 25%) and various cancers (‘CTE’ 3/12, 25%). On the other hand, many cases in the ‘NDG’ group with complex diagnoses, and notably higher in age, presented several systemic diseases including hypertension (6/17, 35.3%), prostate, bone and lung cancer (5/17, 29.4%) and circulatory diseases such as atrial fibrillation (AFib) and cardiomyopathy (7/17, 41.2% with 2/7 AFib). Several cases of diabetes (3/17, 17.6%) and one case each of epilepsy (1/17, 5.9%) and dysdiadochokinaesia (1/17, 5.9%) were reported. Substance use was frequent across the entire case series. Alcohol was the predominant substance used in the ‘NDG’ group (6/17, 35.3%). Some cases in the ‘CTE’ group (3/12, 25%) and the ‘None’ group (1/11, 9.1%) consumed alcohol daily. Some recreational drug use was observed in the ‘None’ (2/11, 18.2%) and in ‘CTE’ (3/12, 25%) groups and one case in the ‘NDG’. Additionally, two cases in the ‘None’ group (2/11, 18.2%) used prescribed opiates daily. The most frequently reported psychiatric symptoms across the entire case series were depression and anxiety. These two disorders were most frequently reported in the ‘None’ group (7/11, 63.6%), and additionally, this group had reported cases with a history of panic attacks (2/11, 18.2%). Depression and anxiety (4/12, 33.3%) and unspecified mood disorders (4/12, 33.3%) were also abundant in the ‘CTE’ group. Within the ‘NDG’ group, several cases of depression and anxiety (4/17, 23.5%), some cases of obsessivecompulsive disorder (2/17, 11.8%), a case with history of psychotic episodes (1/17, 5.9%) and a case of paranoia (1/17, 5.9%) were noted.

Table 3.

Reported systemic disease, substance use and psychiatric symptoms of athletes with rmTBI history

Neuropathology, n	None (n = 11/11)a	CTE (n = 12/14)	NDG (n = 17/27)
Systemic disease
Hypertension	1	2	6
Sleep apnoea	2	1	2
Cancer		3	5
Heart and circulatory disease	3 (1 AFib)	3	7 (2 AFib)
Diabetes			3
Other	1	1	3 (1 epilepsy, 1 dysdiadochokinaesia)
Substance use
Alcohol (daily)	1	3	6
Cigarettes	1
Recreational drug use	2	3	1
Opiates	2
Psychiatric symptoms/episodes
Depression	7	4	3
Anxiety	2	2	2
OCD	1		2
Psychotic episodes		1	1
Unspecified mood disorder		4	1
Other	2 (panic attacks)	1 (bipolar)	1 (paranoia)

AFib, atrial fibrillation; OCD, obsessive–compulsive disorder.

^a n ratio is number of cases represented in this report against the total number of cases per neuropathology group. Cases are not mutually exclusive, i.e. some cases had multiple systemic diseases, used several substances or presented with several clinical symptoms.

Evaluation of concussion history and clinical symptoms with detailed questionnaires

Of the 52 cases, 19 next of kin (5 for ‘None’ cases, 4 for ‘CTE’ and 10 for ‘NDG’) provided details on the athletes’ concussion history and clinical symptoms over the last several years to a decade of their lives through a detailed questionnaire. Due to the lower age at death for the ‘None’ and ‘CTE’ groups, the next of kin was most likely able to recall the athletes’ concussion history with greater detail, due to known biases in retrospective recall.²⁴ 3/5 (60%) of the ‘None’ and 3/4 (75%) of the ‘CTE’ cases sustained a total of over seven concussions with their first concussion sustained in teenage years: two at 13 and one at 16 years old in ‘None’ and one at 15, one at 16 and one at unspecified teenage years in ‘CTE’ groups. The same three cases in the ‘None’ group had also multiple visits to the emergency room and four to five concussions that resulted in LOC. The mean number of concussions was unknown in most ‘NDG’ cases (5/10, 50%) or recalled by the next of kin as having multiple concussions (4/10, 40%).

Clinical symptoms were grouped into eight classes: executive, behavioural, constitutional, memory, visuospatial, sensory, motor and language-associated symptoms (Table 1). The next of kin indicated whether the athletes experienced a variety of symptoms per class, along with their severity, and this information was used to generate a semi-quantitative scale (0–3, 3 being most severe and 0 being no such symptom) per case (Fig. 1). Cases in the ‘None’ group (n = 5) experienced executive, behavioural and constitutional symptoms most severely. Executive symptoms included difficulty planning, poor concentration and mental rigidity (mean = 1.77, SD = 0.74). Behavioural symptoms often developed as depression or anxiety, changes in personality, agitation and disinhibition (mean = 1.16, SD = 0.52). These two symptom classes developed less than a year pre-mortem in 1/5 (20%) cases, 5 years pre-mortem in 2/5 (40%) and 10 years pre-mortem in 2/5 (40%). Many cases experienced constitutional symptoms of chronic headaches and general fatigue (mean = 1.38, SD = 0.66). For 4/5 (80%) cases, the next of kin observed the onset of executive, behavioural and constitutional symptoms around the same time. One case started to show symptoms of executive dysfunction about 10 years pre-mortem, followed by behavioural and constitutional changes within a year before passing. Other symptoms such as complications with visuospatial (mean = 0.45, SD = 0.74), sensory (mean = 0.98, SD = 0.57) and language (mean = 0.22, SD = 0.26) functions were seen infrequently in this group.

Figure 1.

Comprehensive clinical symptom analysis by severity in 19 male athletes with a rmTBI history. The next of kin reported the clinical symptoms of athletes with history of rmTBIs (n = 19) based on eight symptom classes: memory, executive dysfunction, behaviour and emotional well-being, language, visuospatial, motor, sensory and constitutional symptoms. A variety of symptoms were evaluated per class and analysed based on the presence and severity (0–3 semi-quantitative scale). Clinical presentations were averaged based on the athletes’ neuropathological diagnoses (‘None’—no diagnosis, CTE diagnosis and ‘NDG’—complex neurodegenerative diagnosis) and tested for significant differences within each symptom class via KW tests followed by Dunn’s multiple comparisons tests. The ‘None’ group scored high in severity for executive and constitutional symptoms. The ‘NDG’ group scored highly on executive and memory-associated symptoms. The ‘CTE’ group scored the lowest in severity for most categories compared to the other two groups. Symptoms pertaining to memory (Dunn’s test P = 0.043), language (P = 0.005), motor (P = 0.015) and visuospatial (P = 0.015) functions differentiated the ‘NDG’ group from the ‘CTE’ group. ‘NDG’ group scored higher for symptoms of language dysfunction than the ‘None’ group (P = 0.026). *P < 0.05, **P < 0.01.

Cases in the ‘CTE’ group (n = 4) exhibited a symptom repertoire similar to the ‘None’ group. The predominant symptoms were emotional/behavioural (mean = 1.13, SD = 0.46), executive (mean = 0.75, SD = 0.87) and constitutional (mean = 0.42, SD = 0.50) symptoms, which included symptoms of depression and anxiety, irritability, headaches, fatigue and insomnia. For all symptom classes, the ‘CTE’ group on average scored lower in severity than the ‘None’ group (Fig. 1). In the ‘CTE’ group, the next of kin identified executive and behavioural symptoms consistently 10 years pre-mortem in 3/4 (75%) cases, whereas symptom duration was more variable in the ‘None’ group and ranged between 5 and 10 years. Although the number of questionnaires returned for these groups was limited and therefore discretion must be used when drawing conclusions, the results obtained from retrospective interviews with the next of kin in a larger series of 52 cases exhibited the same pattern; clinical presentations were highly alike between those diagnosed with CTE and those with no neuropathological diagnoses.

Finally, symptoms described for cases in the ‘NDG’ group (n = 10) mainly constituted executive (mean = 1.67, SD = 0.79), memory (mean = 1.93, SD = 0.72), motor (mean = 0.94, SD = 0.58) and language (mean = 1.23, SD = 0.64) dysfunction. ‘NDG’ group scored significantly higher for language dysfunction (KW statistic = 12.98, P < 0.001) than both ‘None’ and ‘CTE’ groups (Dunn’s tests P = 0.026 and P = 0.005, respectively). Additionally, ‘NDG’ group scored higher than the ‘CTE’ group for memory (Dunn’s test P = 0.043), motor (Dunn’s test P = 0.015) and visuospatial dysfunction (Dunn’s test P = 0.015). This group had a higher age at death (mean = 78.4, SD = 5.6), and most cases in this series developed symptoms of dementia (9/10, 90%). The most common symptoms in this group included difficulty with memory, disorganization, impaired perception, trouble finding words, reduced verbal communication, tremors and falls. The onset of executive (mean = 8.36 years) and memory dysfunction (mean = 8.6 years) occurred on average several years sooner than language (mean = 5.6 years) and motor dysfunction (mean = 5.57 years). This group scored lowest in severity for behavioural symptoms (mean = 0.89, SD = 0.58), and indeed in the entire ‘NDG’ group, 8/27 (29.6%) cases were reported to exhibit psychiatric issues.

Age and DNA damage in relation to clinical questionnaires

For the 19 cases for whom clinical questionnaires were obtained, mean age at death and the presence of DNA damage via γH2AX staining were evaluated in relation to their clinical symptoms (Fig. 2 and Table 4). As seen in the larger series, age at death was significantly lower in the cases with no diagnosis (mean = 39.8, SD = 14.8; Dunn’s test P = 0.003) or only a CTE diagnosis (mean = 47.5, SD = 8.0; Dunn’s test P = 0.032) compared to those that presented complex neurodegenerative diseases (mean = 78.4, SD = 6.0; KW statistic = 13.75, P < 0.001; Fig. 2A). Age at death (n = 19) positively correlated with symptoms of memory (r = 0.60, 95CI = 0.18–0.83, P = 0.007), language (r = 0.74, 95CI = 0.42–0.90, P = 0.0003) and visuospatial dysfunction (r = 0.63, 95CI = 0.23–0.85, P = 0.004; Fig. 2D and Table 4). In the ‘None’ and ‘CTE’ groups (n = 9), age negatively correlated with memory symptoms (r = −0.71, P = 0.04). In the ‘NDG’ group alone (n = 10), age positively correlated with memory (r = 0.77, P = 0.012) and visuospatial (r = 0.73, P = 0.02) symptoms.

Figure 2.

Correlation between age at death, presence of γH2AX post-mortem and clinical symptoms in 19 male athletes with rmTBI history. (A) Age at death was significantly higher for cases with a complex neurodegenerative disease diagnosis (‘NDG’, n = 10, mean age = 78.4; KW statistic = 13.75, P < 0.001) than for cases with no diagnosis (‘None’, n = 5, mean age = 39.8, Dunn’s test P = 0.003) and a CTE diagnosis (‘CTE’, n = 4, mean age = 47.5, Dunn’s test P = 0.032). (B) Post-mortem brains were stained for DNA damage via γH2AX and semi-quantitatively staged. Cases with no diagnosis presented minimal presence of γH2AX. ‘CTE’ cases varied in γH2AX reactivity, and ‘NDG’ cases mostly presented with Stages II and III of γH2AX and significantly differed from the ‘None’ group (Dunn’s test P = 0.007). Kruskal–Wallis tests followed by Dunn’s multiple comparisons tests were used to test significant differences in age at death (A) and γH2AX reactivity (B). (C) Behavioural symptoms segregated by γH2AX staging showed all groups exhibiting clinical symptoms, scoring highest in executive and memory symptoms. Cases with no γH2AX reactivity also scored high in constitutional symptoms, cases in Stage I scored high in behavioural symptoms and cases in Stage III scored high in language dysfunction. SD, standard deviation. (D) Spearman’s correlations visualized as scatter plots revealed a significant correlation (r) between age at death (n = 19) and clinical symptoms in relation to memory (0.60), language (0.74) and visuospatial (0.63) capabilities. In ‘None’ and ‘CTE’ groups (n = 9), age at death negatively correlated with memory (−0.71), and in the ‘NDG’ group (n = 10), it positively correlated with memory (0.77) and visuospatial (0.73) dysfunctions. Presence of γH2AX (n = 19) correlated with age at death (0.59) and language dysfunction (0.56). In the ‘None’ and ‘CTE’ groups (n = 9), γH2AX negatively correlated with language (−0.78) and constitutional (−0.78) symptoms. *P < 0.05, **P < 0.01.

Table 4.

Spearman’s correlation between age at death, γH2AX reactivity and symptom scores in 19 male athletes with rmTBI history

All n = 19	Executive	Memory	Behaviour	Constitutional	Motor	Language	Visuospatial	Sensory
$\mathbf{0.59}\left[\begin{array}{c}\mathrm{Age}\\ \gamma \mathrm{H}2\mathrm{AX}\end{array}\right]$	0.20	0.60	−0.19	−0.01	0.41	0.74	0.63	0.12
	−0.18	0.31	−0.24	−0.21	0.19	0.56	0.35	−0.12
None and CTE n = 9	Executive	Memory	Behaviour	Constitutional	Motor	Language	Visuospatial	Sensory
$0.32\left[\begin{array}{c}\mathrm{Age}\\ \gamma \mathrm{H}2\mathrm{AX}\end{array}\right]$	−0.35	−0.71	−0.32	−0.23	−0.05	0.15	0.07	0.14
	−0.68	−0.67	0.00	−0.78	−0.64	−0.78	−0.73	−0.58
NDG n = 10	Executive	Memory	Behaviour	Constitutional	Motor	Language	Visuospatial	Sensory
$-0.23\left[\begin{array}{c}\mathrm{Age}\\ \gamma \mathrm{H}2\mathrm{AX}\end{array}\right]$	0.33	0.77	0.33	0.21	0.20	0.07	0.73	0.37
	−0.19	−0.04	−0.04	0.02	0.24	0.61	0.23	−0.03

Statistically significant correlation in bold.

Histological analyses of brain tissue with γH2AX in the 19 cases for whom clinical questionnaires were obtained (Fig. 2B) showed DNA damage in various cell types, including peri-neuronal glial cells, ependymal and subependymal cells, subpial astrocytes and oligodendrocytes. The presence of DNA damage was semi-quantitatively staged based on which cell types stained for γH2AX as described in the ‘Materials and methods’ section.¹⁶ In this group of 19, 3/19 (15.8%) were γH2AX negative, 5/19 (26.3%) were Stage I, 5/19 (26.3%) were Stage II and 6/19 (31.6%) were Stage III. 2/5 (40%) ‘None’ and 1/10 (10%) ‘NDG’ cases were γH2AX negative, 3/5 (60%) ‘None’ and 2/4 (50%) ‘CTE’ were Stage I, 2/4 (50%) ‘CTE’ and 3/10 (30%) ‘NDG’ were Stage II and 6/10 (60%) ‘NDG’ were Stage III. Cases in the ‘NDG’ group presented significantly higher stages of γH2AX than cases in the ‘None’ group (KW statistic = 9.55, P = 0.003, Dunn’s test P = 0.007; Fig. 2B). Clinical symptoms for cases that were γH2AX negative (n = 3) ranged in severity across all categories with executive dysfunction scoring the highest (mean = 2.24, SD = 0.72; Fig. 2C). Cases in Stage I (n = 5) scored consistently low in severity for language (mean = 0.07, SD = 0.10) and visuospatial (mean = 0.05, SD = 0.11) symptoms and highest in executive dysfunction (mean = 1.53, SD = 0.92). Cases in Stage II (n = 5) averaged evenly across symptoms (mean = 0.48–1.39) with the highest severity for memory (mean = 1.39, SD = 1.34). Symptom averages in cases at Stage III (n = 6) were overall higher compared to Stages I and II (mean = 0.53–1.89), with memory (mean = 1.89, SD = 0.50), language (mean = 1.58, SD = 0.55) and executive dysfunction (mean = 1.58, SD = 0.84) scoring the highest. Notably, γH2AX reactivity (n = 19) positively correlated with age (r = 0.59, 95CI = 0.17–0.83, P = 0.008) as well as language dysfunction (r = 0.56, 95CI = 0.13–0.82, P = 0.012; Fig. 2D and Table 4). γH2AX reactivity assessed for cases in the ‘None’ and ‘CTE’ groups (n = 9) showed negative correlation with language (r = −0.78, P = 0.027) and constitutional (r = −0.78, P = 0.016) symptoms. In the ‘NDG’ group alone (n = 10), γH2AX reactivity did not exhibit a significant correlation with clinical symptoms.

Discussion

In this study, we evaluated clinical symptoms accompanied by neuropathological post-mortem diagnoses of male athletes with a history of rmTBI. Although this study examines a small case series, we provide justification for exploring pathophysiological changes outside of CTE and other neurodegenerative diseases that may contribute to clinical manifestations.

This case series of athletes with rmTBI history demonstrated substantial behavioural changes even in those without substantial post-mortem pathological findings. Almost a quarter of our cases did not present with CTE in the brain, nor were they diagnosed with a neurodegenerative disease despite almost all of these cases experiencing neuropsychological symptoms. These athletes, categorized into the ‘None’ group in this paper, most often experienced mood disorders, particularly depression and anxiety, and some experienced agitation and violent behaviour. Many athletes with only CTE were found to experience similar clinical symptoms to those without pathology. Interestingly, in the subset of cases for whom detailed clinical symptoms were obtained, some cases with a CTE diagnosis even exhibited milder symptoms compared to those with no diagnosis or a complex neurodegenerative diagnosis. Notably, these findings are of a small group and would require extensive corroboration. Nevertheless, this warrants investigation outside of a CTE diagnosis or other neurodegenerative diseases, as these diagnoses do not fully capture the scope of behavioural changes, and other pathophysiological mechanisms may be at play.²⁵ Expanding on this, we conducted a staged γH2AX staining on the post-mortem samples to elucidate a potential relationship between DNA damage burden and clinical symptoms. Cases with a neurodegenerative diagnosis were more likely to show evidence of DNA damage that spanned across multiple brain regions and cell types, and γH2AX reactivity positively correlated with an increase in age and language dysfunction. Progressive accumulation of DNA lesions and insufficient DNA repair mechanisms have been implicated in cognitive decline and several age-associated neurodegenerative disorders such as Alzheimer’s disease, Parkinson’s disease and amyotrophic lateral sclerosis,^26,27 speculated to contribute to the pathogenesis by inducing accelerated cellular senescence, neuronal degeneration and neuroinflammation.²⁸ Interestingly, most of the younger cases with limited pathology were also γH2AX reactive. However, in the younger sample, γH2AX reactivity negatively correlated with symptoms of language and constitutional dysfunction. Most clinical symptoms, specifically memory, motor, language and visuospatial, progressively worsened from Stage I to Stage III of γH2AX reactivity, whereas the few cases that were γH2AX negative had exacerbated symptoms comparable to Stage III. Although these findings are of a small sample and their interpretation is not definitive, they suggest that in cases with γH2AX reactivity, DNA damage burden may be a promising behavioural marker. However, young cases lacking γH2AX reactivity require a more extensive biomarker search that may lie outside of tauopathy and DNA damage, pointing to the complexity and heterogeneity of molecular determinants of clinical symptoms. Some studies have introduced γH2AX reference controls^16,29; however, further studies are required for the assessment of DNA damage and repair in the context of brain trauma. Nonetheless, here we observed a discernible presence of double-stranded DNA damage in symptomatic cases of brain trauma akin to tauopathy manifestation. As such, DNA damage may serve as an alternative marker for symptomatic rmTBI cases that do not present tauopathy.

At present, the extent of pathobiological mechanisms that perpetuate clinical symptoms post-rmTBI is poorly understood, and the relevance of CTE is still debated.^13,30-32 Diagnostic criteria for traumatic encephalopathy syndrome (TES) were developed to establish a clinical correlate to CTE³³ that requires a history of significant exposure to mTBI and progressive cognitive or neurobehavioural dysfunction unattributed to other neurodegenerative diseases. Ascribing TES to CTE has been critiqued for neglecting other forms of chronic TBI that do not receive a CTE diagnosis.³⁴ Moreover, establishing TES criteria specific enough for a CTE diagnosis has not been successful.³⁵ Studies have found that within the general US population, those with chronic pain, mood disorders or those who have experienced suicidality were highly likely to meet the TES criteria.^36,37 While the link between the clinical symptoms and the full pathophysiological picture remains unclear, we also urge to refrain from a rudimentary clinical representation of CTE.^38-40 A recent longitudinal study has evaluated a series of rmTBI cases with neuropsychological symptoms in accordance with the TES criteria against their post-mortem neuropathological manifestations.¹⁵ Cognitive, memory and executive function decline were observed fulfilling the TES criteria; however, additional neuropathological changes with or without a CTE diagnosis were common. Atrophy and decreased white matter integrity in the limbic system and the medial temporal lobe and hypometabolism in the thalamus were persistent. Moreover, co-morbidity with limbic TDP-43 proteinopathy and hippocampal sclerosis was frequently observed with the CTE cases. This study further signifies the necessity for broadening the diagnostic markers of post-rmTBI symptoms beyond CTE and TES.

Since the original report of post-mortem findings in an American football player,⁴¹ CTE has received explosive attention from the media. Although public awareness has grown on potential pathophysiological outcomes of repeated concussions, and policies regarding return to play were put into place, the prevalence of CTE and its true relation to clinical symptoms remain speculative in scientific research.^42,43 Current consensus on the CTE diagnosis has no clinical correlate. Moreover, other factors in one’s life history such as developmental environment, pre-existing disorders and mental health issues, adjustment to retirement, surgeries, systemic diseases and substance use may be co-factors of these symptoms.⁴⁴ Two NINDS/NIBIB consensus meetings have been held to establish the neuropathological diagnostic criteria for CTE; however, disagreements within the neuropathologist community are still prevalent.⁴⁵ Cases of ‘pure’ CTE are scarce and often show very focal pathology or are present with other forms of degeneration such as axonal injury and DNA damage/cellular senescence,¹⁶ beta-amyloid plaques, TDP-43 and astroglial scarring. The minimum threshold for CTE diagnosis is the presence of a single pathognomonic lesion in the cortex.⁴⁶ A diagnosis based on a single lesion is not sufficient in associating pathology with clinical presentation, as it is implausible for one focal lesion to catalyze a systemic neurocognitive deficit. In comparison, in Alzheimer’s disease Braak staging, the disease is often clinically silent until the later stages where tau accumulation is more severe than ‘low’ CTE.²³ Although the CTE consensus is not meant for a clinical diagnosis but rather a preliminary step for researchers,⁴⁷ the set criteria risk the production of future research that overestimates the relevance of CTE and distracts from the discovery of molecular biomarkers that better correlate to clinical symptoms and may even precede pathology. Moreover, CTE tauopathy can be difficult to distinguish from other proteinopathies,^48,49 many of which are co-morbid with CTE in older patients as seen in our cases as well. In such co-morbid cases where other pathologies with similar clinical presentations as CTE are present, it is difficult to attribute clinical presentations to a CTE diagnosis with certainty.^30,50

In this case series of 52 athletes with a history of rmTBI, we have observed a relationship between the presence of a neurodegenerative diagnosis, DNA damage in the brain and impaired memory, language and visuospatial function. As expected, complex neurodegenerative diagnoses were more frequent in brains with higher age at death, due to the progressive nature of diseases such as Alzheimer’s disease, Parkinson’s disease and corticobasal degeneration. In our series, a CTE diagnosis was observed in those with an average age at death of mid-50s. Whether CTE is a progressive neurodegenerative disease has been disputed. Studies have demonstrated an age-dependent severity of CTE, p-tau distribution and density⁵¹ and the presence of higher stages of CTE with co-morbid neurodegenerative diseases at an older age at death.⁵² Other groups have found CTE to not be inexorably progressive,^19,45 and it may reveal a stable disease state where its severity varies based on an individual’s susceptibility. In advanced stages, neuronal tau near the vasculature is difficult to distinguish from other tauopathies, especially in an abundance of neurofibrillary tangles. In our case series, CTE was confirmed in a wide range of ages, from 17 to 70 years old at diagnosis, and the propensity of its diagnosis did not increase with age. As such, our findings were not able to confirm the hypothesized progressive nature of CTE, and clinical symptoms of memory, language and motor dysfunction were more likely to be attributed to other neurodegenerative diagnoses such as Alzheimer’s disease. In contrast, the presence of DNA damage was found to be correlated with age in our case series, along with symptoms of language dysfunction. Moreover, several symptomatic cases with no major pathological findings exhibited DNA damage burden. Whether accumulation of DNA damage is a consequence of rmTBI is to be determined. In our previous work, we found compelling evidence in human post-mortem brains^16,53 and in mice^54,55 for accumulation of DNA damage as a possible marker of rmTBI and a potential contributing factor to symptoms.^54,55 As such, DNA damage markers may make a valuable addition to the diagnostic criteria of TBI pathology, serving as an example of molecular events outside of proteinopathies that justify a shift in TBI research focus.

Correlational studies on clinical presentations of post-mortem neuropathological diagnoses inherently carry many limitations and require a heedful approach. The lack of a complementary control group and a controlled environment and the complexity of personal history can make it difficult to attribute symptomology to pathology. Survey-based studies relying on informants’ knowledge are prone to recall bias and subjectivity.⁵⁶ Referral bias for symptomatic individuals is particularly of issue in concussion studies, as study participation is influenced by the inclination to attain a reason for the patient’s change in behaviour. Significant attention from the media, the public and athletic entities has provoked disputes regarding CTE prevalence in the population of contact sports athletes and its correlation to chronic symptoms.⁴³ Longitudinal studies are underway to elucidate these gaps in knowledge.⁵² We caution against making a causative association between the reported clinical symptoms and the athletes’ concussion histories, as many other factors including developmental and environmental factors, substance use and family medical history can be significant contributors.¹⁹ We acknowledge the smaller scale of the study and put forth our findings as a premise to explore other pathobiological mechanisms beyond tauopathy and CTE.

Conclusion

Association of the athletes’ chronic clinical symptoms and their post-mortem neuropathological presentations is a challenging endeavour. Currently, prevention and treatment of chronic clinical symptoms associated with rmTBI are multifaceted, albeit non-specific, to the possible underlying pathology. Prevention revolves around the implementation of sports policies, such as return to play or penalization of athletes who display unwarranted aggression towards other players.⁵⁷ Different therapeutic approaches including medication for ongoing somatic, emotional, cognitive and sleep-associated symptoms are the main mode of treatment to date.⁵⁸ However, opiate reliance and addiction in former athletes have raised concern as to whether non-specific short-term relief of symptoms is a double-edged sword.⁵⁹ Symptomology developed from repeated concussive episodes is likely non-specific to a single underlying pathology. As such, diagnosis with CTE does not explain the long-term behavioural changes observed in the athletes. Based on our work, we propose the accumulation of DNA damage as a promising addition to a more holistic diagnostic approach.¹⁵ It has become abundantly clear that the direction we take to investigate underlying pathobiological mechanisms in concussion research requires re-evaluation.

Funding

This study was funded by the Canadian Institute of Health Research (CIHR) grant number 6210100803.

Competing interests

The authors report no competing interests.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article. To protect the privacy and identity of the participants, data availability outside of what is presented in this article is limited. Any additional information that does not compromise participants’ identity may be available from the corresponding author upon reasonable request.