BRain health and healthy AgeINg in retired rugby union players, the BRAIN Study: study protocol for an observational study in the UK

Introduction Relatively little is known about the long-term health of former elite rugby players, or former sportspeople more generally. As well as the potential benefits of being former elite sportspersons, there may be potential health risks from exposures occurring during an individual’s playing career, as well as following retirement. Each contact sport has vastly different playing dynamics, therefore exposing its players to different types of potential traumas. Current evidence suggests that these are not necessarily comparable in terms of pathophysiology, and their potential long-term adverse effects might also differ. There is currently limited but increasing evidence that poorer age-related and neurological health exists among former professional sportsmen exposed to repetitive concussions; however the evidence is limited on rugby union players, specifically. Methods and analysis We present the protocol for a cross-sectional study to assess the association between self-reported history of concussion during a playing career, and subsequent measures of healthy ageing and neurological and cognitive impairment. We are recruiting a sample of approximately 200 retired rugby players (former Oxford and Cambridge University rugby players and members of the England Rugby International Club) aged 50 years or more, and collecting a number of general and neurological health-related outcome measures though validated assessments. Biomarkers of neurodegeneration (neurofilaments and tau) will be also be measured. Although the study is focusing on rugby union players specifically, the general study design and the methods for assessing neurological health are likely to be relevant to other studies of former elite sportspersons. Ethics and dissemination The study has been approved by the Ethical Committee of London School of Hygiene and Tropical Medicine (reference: 11634-2). It is intended that results of this study will be published in peer-reviewed medical journals, communicated to participants, the general public and all relevant stakeholders.

there are now plausible mechanisms for these effects, and a recognition that these problems do not just occur in former boxers, but in a variety of sports involving repeated concussions (2), and possibly also in sports in which low-level head trauma is common (3). These neurodegenerative effects potentially include increased risks of impaired cognitive function and dementia (4-7), Parkinson's disease (PD) (8)(9)(10)(11)(12)(13), and amyotrophic lateral sclerosis (ALS) (3,(14)(15)(16)(17)(18)(19). The term chronic traumatic encephalopathy (CTE) was introduced as a clinical-pathological construct for the neurodegeneration associated with American football and wrestling (20) (see Table 1). Seminal work by Ann McKee (21) investigating of CTE led to important considerations on the overlapping of neurodegenerative lesions in this condition (22).
However, each contact sport has vastly different playing dynamics therefore exposing its players to different types of potential traumas. Current evidence suggests that these are not necessarily comparable in terms of pathophysiology, and hence in terms of their potential long-term adverse effects on health. There is currently limited but increasing evidence that poorer general and neurological health exists among professional sportsmen exposed to repetitive concussions; however, there is little evidence from rugby union players (23)(24)(25).
Decq et al (23) investigated retired French-speaking high-level sportsmen, aged 45 to 65 years, who had played sports for at least 10 years. Mild cognitive disorder (measured as score at the modified Telephone Interview for Cognitive Status (TICS-m) <30) was lower in players of other sports (40.4%) than in former rugby players (56.6%) (p=0.005). However, after adjustment for smoking and higher education, no association was observed between TICS score and number of reported concussions (23).
Furthermore, in New Zealand, 366 former players were tested on their engagement in sport, general health, sports injuries and concussion history, and demographic information. Cognitive functioning was assessed using the online CNS Vital Signs neuropsychological test battery. The elite-rugby group performed worse on tests of complex attention, processing speed, executive functioning, and cognitive flexibility than the non-contact-sport group, and worse than the community-rugby group on complex attention. Former players who recalled one or more concussions had worse scores on cognitive flexibility, executive functioning, and complex attention than players who did not recall experiencing a concussion (24). A recent Scottish study assessed 52 former Scottish International rugby players and 29 controls. Players performed worst on a test of verbal learning and of fine co-ordination of the dominant suggests that the Mayo Clinic TBI Classification system should cover both of them (33). A glossary of definitions used in this protocol is found in Table 1, adapted from Jordan (2014) (34).
For the purpose of this study, we adapted the NIH definition of concussion. Participants will be asked to report their previous concussions according with the following definition: Concussion is defined as an alteration in brain function, caused by an external force. Symptoms include: • A decreased level / loss of consciousness • Memory Loss (before or after the injury) • Weakness • Temporary Paralysis

Impaired cognitive function and dementia in sportspersons
While the neurological and cognitive effects of acute traumatic brain injuries (TBI) have been extensively studied (4)(5)(6)(7), the association between TBI and delayed sequelae have been less studied because of the variable latency period before overt neurologic dysfunction, and the difficulty of retrospectively recalling relevant events (recall bias), in particular in presence of memory impairment. The neurocognitive effects of repetitive mild head injury in sport were initially recognised in boxers, a syndrome that was distinct from the clinical and pathological sequelae of single-incident severe TBI (35) (see also Table 1).
The clinical syndrome of dementia pugilistica (punch-drunk syndrome) is associated with prominent tauopathy, with typical neurofibrillary tangles and neuropil threads, distributed in patches throughout the neocortex (36). Neuropathological case series have demonstrated primarily tau-related pathology in the brains of individuals who have suffered from a clinical syndrome encompassing dementia and movement disorders after repeated head trauma (37). Also, post-mortem examination of the brains of several professional American football players and wrestlers has revealed the pathological underpinnings for the cognitive and neuropsychiatric decline seen in these men in later life (20).
Although cognitive decline in long time professional American football players has been noted for some years, the first autopsy report from such a player appeared in the literature only recently (38). In all cases, cognitive decline began years after retirement from the game.
Cognitive function is an ideal outcome measure for this type of study as being a continuous measure maximizes power to detect an association; for using neurodegenerative disease as outcome measures a substantial larger sample would be needed. Cognitive function as measured with neurocognitive tests is also an aspecific measure, does not imply a clinical diagnosis, and it can be the prodrome manifestation for a number of different neurodegenerative diseases.
However, to date, no biomarker of long-term effects of concussion has been identified, although recent studies on cerebrospinal fluid (CSF) from sportsmen with post-concussion syndrome suggest that a subset of these have biomarker signs of ongoing axonal injury and microglial activation (44)(45)(46).
Of the biomarkers measurable in serum or other body fluids, those more likely to be seen a long time after the concussion episode, are those related to the axonal injury and more in generally neuroinflammation, that is neurofilament and tau protein (43). Neurofilaments (NF) co-assemble from protein subunits to form NFs defined as light (light (NF-L), medium (NF-M), or heavy (NF-H) according to their relative molecular weights). NFs are one of the key structural elements of neurons, providing mechanical stability and determining axonal diameter. NFs are of particular interest because they are structural elements and therefore they might be more susceptible to mechanical deformation of the inflammation, and metabolic pathways are among the most important ones (58). The identified single nucleotide polymorphisms are usually located in the non-coding regions and their functionality remains to be determined, although they presumably influence gene expression. Similarly, since the discovery of mutations in SOD1 gene, which account for ~2% of ALS cases, increasing efforts have been made to understand the genetic component of ALS risk. Specifically mutation in the chromosome 9 open reading frame 72 (C9orf72) were associated to both ALS and Fronto-Temporal Dementia (FTD) (59).

Rationale of the BRAIN Study
Recently, World Rugby (formally known as the International Rugby Board), have developed a process to support team doctors in the recognition, assessment, and subsequent management of elite adult players who sustained a potential concussion. This process includes the development of a multimodal assessment: the Head Injury Assessment, formerly the Pitch Side Concussion Assessment (PSCA) tool (60).
However, there is little clear evidence, about the possible long-term effects of concussion in rugby union players and it is not clear if the findings from other sports are directly generalizable to rugby players, per se. Many sports involve concussion or repetitive low-level head trauma, but it can be argued that each sport should be viewed differently depending on the unique technical and physiological profile that a player is exposed to over the course of a career (61,62). Thus, to determine the risk of long-term adverse health effects of playing rugby, in addition to the other research that is ongoing in different contact sports, specific studies of rugby players are needed. The BRAIN Study builds upon a recent questionnaire-based study conducted by members of the Arthritis Research UK sport, exercise and osteoarthritis centre, University of Oxford, which has assessed the general and musculoskeletal health of former elite rugby players. Rugby players from the Oxford and Cambridge University Football Clubs (Oxbridge Blues), and members of the England Rugby Internationals Club (ERIC) -an honorary association of all England and ex-England players -were recruited. Self-reported information on demographic factors, playing history (including head trauma), past medical history (including dementia, depression, and memory impairment), and perceived health were collected, in addition to detailed information on musculoskeletal health and pain. A total of 320 participants have been surveyed in this study to date, and 205 of those aged 50 years or older have agreed to be re-contacted for further testing/questionnaires. The 205 participants aged 50+ years who agreed to be re-contacted will be invited to take part to the BRAIN Study.

Aim and objectives
The general aim of the BRAIN study is to investigate the possible associations between concussion in rugby and ageing -including physical and cognitive capabilities -as well intermediate neurological and musculoskeletal endpoints among former rugby players.
In order to achieve these aims the following objectives will be pursued: 1. Investigate the associations between self-reported concussion history and ageing, measured as physical and cognitive capabilities. This will be achieved by using the following outcome  5. Investigate the long-term musculoskeletal health outcomes of rugby players with particular emphasis on hip, knee, and hand osteoarthritis allowing changes over time (from the previous study to the current study) to be assessed In addition, we will build a multimedia database (data from tests and questionnaire plus video-recorded neurological examination plus biobank of blood samples) that will serve as baseline for further tests and further follow-up if this is considered appropriate in the future

Study design
Participants will be invited to study clinics in London, Manchester, or Bath. For those who would prefer, a home visit will be arranged. Participants will be invited to bring with them any medication they are taking regularly for a more accurate recording of all medications.
All participants will be administered a Core Module interview including questions on lifestyle factors, potential confounders, and extensive information on the five domains of physical and cognitive ability, neurological examination, intermediate neurological outcomes and musculoskeletal health.
The lifestyle questionnaire will complement that already administered by our colleagues in Oxford, and will include questions on potential confounders of the association between concussion and physical and cognitive capability, and intermediate neurological outcomes. Participants will also be asked to donate a blood sample (a normal blood sample will be collected by participants seen in one of the clinics, a dry spot sample plus saliva swab for DNA will be collected from participants seen at home). At the end of the Core Module, all participants will be also asked to undergo some additional tests (the Additional Optional Module) providing they have sufficient time (Figure 1). At the end of the interview, all participants will be asked some final questions on their concussion history. This will ensure that the interviewer is blind to the participant's concussion history during the entire duration of the interview and this will also act as validation check on data previously collected. . This Additional Optional Module should take no longer than 1 hour and half to be completed ( Figure 1 and Error! Reference source not found.).

General Questionnaires
Questionnaires will be used to collect relevant information from the participants.
Lifestyle and confounder questionnaireall participants will be asked a number of questions on their lifestyle in order to collect information on possible confounders for the main analysis (smoking, alcohol, coffee, drugs, past medical history, sleep quality).
Concussion history questionnaireat the end of the interview, participants will be re-asked some detailed information regarding their history of head trauma and concussion when playing.
Hand pain -A questionnaire on hand pain to identify possible hand osteoarthritis will be used using a hand mannequin (additional optional module).

Physical ability assessment
The methodology used in the 1946 birth cohort (63) will be leveraged to assess physical and cognitive capabilities in this study. This will enhance comparability of results and facilitate future collaborations.
The most commonly used objective measures of physical capability for assessing healthy ageing, are tests of grip strength (64,65), walking speed (64,66,67), chair rises (68) and standing balance (64); these aim to assess physical functioning, including the capacity to undertake the physical tasks of daily living (69). There is robust evidence that higher scores on these measures are associated with lower rates of mortality, and there is more limited evidence of lower risks of morbidity, and of age-related patterns of change (69). Height and weightheight and weight will be collected in order to calculate the body mass index (BMI) according to the formula BMI = weight (Kg) / height (m) 2 .
Grip strength -JAMAR Hydraulic Hand Dynamometer is the most widely used instrument with established test-retest, inter-rater and intra-rater reliability (70).
Chair rise -The 30-Second Chair Stand Test will be used to assess lower body strength (71). This test provides a reasonably reliable and valid indicator of lower body strength in generally active, communitydwelling older adults. Test-retest intra-class correlations of 0.84 for men and 0.92 for women, utilizing one-way analysis of variance procedures appropriate for a single trial, together with a non-significant change in scores from Day 1 testing to Day 2, indicate that the 30-s chair stand has good stability reliability (71). Moreover, a moderately high association between chair-stand performance and maximum weight-adjusted leg-press performance for both men and women (r = 0.78 and 0.71, respectively) supports the criterion-related validity of the chair stand as a measure of lower body strength (71). As expected, chair-stand performance decreased significantly across age groups in decades -from the 60s to the 70s to the 80s (p < 0.01) and was highly statistically significantly lower for low-active participants than for high-active participants (p < 0.0001) (71).
Walking speed -To assess normal comfortable walking speed and maximum walking speed, a Timed 10-Meter Walk Test will be used. This test has been validated in 230 healthy volunteers (66). Mean comfortable gait speed ranged from 127.2 cm/s for women in their seventies to 146.2 cm/s for men in their forties. Mean maximum gait speed ranged from 174.9 cm/s for women in their seventies to 253-3 cm/s for men in their twenties. Both gait speed measures were reliable (correlation coefficients > 0.903) and correlated significantly with age (r>-0.210), height (r>0.220) and the strengths of four measured lower extremity muscle actions (r = 0.190-0.500). These normative values give clinicians a reference against which patient performance can be compared in a variety of settings (66).
A photograph of the hand will be taken at the end of this section. QuIKS -The Questionnaire to Identify Knee Symptoms (QuIKS) is aimed at early osteoarthritis, and understanding symptomology and potential adaptation to activity before osteoarthritis has been diagnosed. It has been produced in adults aged 40-65 with evidence of ongoing knee problems and recommended for use in studies exploring early osteoarthritis (74) (Additional optional module).

Cognitive ability assessment
Also for the cognitive capability assessment, the methodology used in the 1946 birth cohort will be followed (63). Tests and questionnaires will be described according their belonging to the Core Module or the Additional Optional Module, and by domain ( Figure 1). (75)) -The MMSE is a widely used 30-point screening tool for cognitive impairment within clinical practice, assessing multiple cognitive domains including: i) orientation to time and place (10 points); ii) registration (3 points), which involves repeating back to the examiner three items; iii) attention +/-calculation (5 points), which involves the participant spelling "WORLD" backwards or performing a serial subtraction task; iv) recall (3 points), which involves recalling the originally registered three items two minutes later; v) language (1 point) which involves a two item naming tasks; vi) repetition (1 point); vii) reading (1 point), which involves following a written command;

Mini-Mental State Examination (MMSE
viii) writing (1 point), which involves writing a grammatically correct sentence; ix) visuospatial function (1 point), tested by drawing interlocking pentagons; and x) following a 3-stage command (3 points), which assesses language function, executive function and praxis.
Logical Memory from the Wechsler Memory Scale-Revised (WMS-R (76)) -The Logical Memory test assesses free recall of a short story. The participant is asked to recall the story immediately and after a 20minute delay.

12-item Face-Name Associative Memory Exam (FNAME-12A) -
The FNAME-12A is a modified version of the 16-item Face-Name Associative Memory Exam (FNAME-16). The FNAME-12A has fewer stimuli and additional learning trials which have been found to be well tolerated by those with Mild Cognitive Impairment, whilst remaining challenging in cognitively normal older adults (77). The FNAME-12A has demonstrated psychometric equivalence with the FNAME-16, which has been shown to be related to beta-amyloid burden in cognitively normal elderly people (78). The FNAME-12A requires the participant to learn 12 face-name and face-occupation pairs. Participants are given two exposures to all 12 facename/occupation pairs. After each exposure and following a 5 minute delay they are asked for name and occupation associated with each face. After a 30 minute delay they are shown 3 faces and asked to identify the face that they recognise and give the name and occupation. Given multiple options to pick from, they are then asked to select the name and/or occupation associated with the face. (79,80) -The task-set shifting and response inhibition task examines the relationship between executive tasks of task-switching/preparation time. In the arrow only condition, participants are shown the cue 'arrow'. Following a short delay they must respond to the direction of the arrow ('right' or 'left'). In the word only condition, participants are shown the cue 'word'. Following a short delay they must respond to the direction of the word ('right' or 'left'). There is a switching condition in which the participant is shown the cue 'arrow' or 'word'. Following a delay both a combined arrow and word stimulus appears. The stimulus is either congruent (left arrow and left word), or incongruent (left arrow and right word). Trials in the switching task are categorised into switch and non-switch. In a non-switch trial the cue is the same as for the immediately preceding trial. In a switch trial the cue differs from the immediately preceding trial.

Digit-Symbol Substitution Test, from the Wechsler Adult Intelligence Scale-Revised (WAIS-R (81)) -The
Digit-Symbol Substitution test explores attention and psychomotor speed. Participants are given a code  (84). The NART comprises a list of 50 words printed in order of increasing difficulty. The words are relatively short in order to avoid the possible adverse effects of stimulus complexity on the reading of dementing subjects, and they are all 'irregular' with respect to the common rules of pronunciation in order to minimise the possibility of reading by phonemic decoding rather than word recognition (84). (85) -This is a circle tracing task which includes both direct and indirect visual feedback conditions. Continuous performance measures are provided including accuracy, speed and speed of error detection and correction. The test has revealed changes in speed and accuracy in Huntington Disease mutation carriers more than 10 years before expected age-of-onset (Additional optional module).

Matrix Reasoning from the Wechsler Abbreviated Scale of Intelligence (WASI (86)) -The Matrix
Reasoning test assesses nonverbal reasoning. The participant is shown a matrix of geometric shapes with a section missing. They are required to select the option that completes the matrix (Additional optional module). (87) -Participants are given a computerised letter-search task and are required to make a rapid decision as to whether the target letter 'X' or 'N' has appeared in the search display (in either low or high load conditions). On some of the trials, a task-irrelevant distractor (a cartoon character) appears on the outside of the search display. The task evaluates the extent to which attention is captured and captivated by the distractor (Additional optional module).

Neurological clinical assessment
The neurological assessment of participants is needed for detecting any subtle neurological sign, and/or to set a baseline for the absence of one or more signs. A video-recorded standard neurological examination will be included as part of the Core Module. An additional test of self-reported impairment mainly due to movement disorders will be administered as part of the Additional Optional Module coordination, balance, ocular movements, cranial nerves, gait, and repeated movements. The examination includes also a test of knee functionality (88).
Information on potential movement disorders will be collected using the unified Parkinson's Disease Rating Scale (PDDRS) part II. This includes self-reported ratings on several motor domains including speech, saliva and drooling, chewing and swallowing, eating, dressing, hygiene, handwriting, hobbies, turning in bed, tremor, standing up, walking and balance, and freezing (Additional optional module).

Blood and saliva sample collection, and bio-banking
Participants seen in the research clinics will be asked to donate a blood sample for testing for tau protein, neurofilament levels, full blood count analysis and to obtain data for a Genome Wide Association Study (GWAS). Samples of serum, plasma, and DNA will be collected and processed according to a pre-specified protocol and stored in a freezer at -80 °C at the Blizard Institute (Queen Mary, University of London). All participants will also asked to donate a blood sample through a dried blood spot test and a saliva sample (in order to get DNA if it is not possible to do a phlebotomy).
Samples will be tested to measure the expression of NF-L, a Nf-subunit showing consistent and reproducible measurements in plasma/serum using a newly developed sensitive methodology exploiting single molecule trapping and measurement by ELISA (SIMOA) (93)(94)(95). The use of both NFH and NFL allows for the determination of different markers whose dynamic of release into bio fluids following injury vary according to their particular chemical structure. NfL has shown good levels of linearity on dilution experiments suggesting that confounders like aggregation or immune response may not interfere with its measurements. NfH different phosphorylation states in blood may in turn provide more information on the systemic biological changes induced by trauma. The relative low abundance of these proteins in blood, in absence of CSF, which is normally more enriched of by-products of neuronal destruction, will not represent a major obstacle as analytical sensitivity has evolved with the development of novel assays which cover the lower end of these protein dynamic range. SIMOA, the proposed methodology for the target biomarkers, is now emerging as the core technique for the  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  GWAS will be conducted at UCL in the lab led by Prof John Hardy.

Multimedia database set up and management
A multimedia database will be set up as part of this project. The final dataset will include data collected through the questionnaire and the tests (see Figure 1), and the video-recorded neurological examination. This data will be also linked to the stored blood samples through a unique identifier.
The data will be stored in mirror, anonymized datasets at LSHTM and QMUL on secure servers, access will be regulated by the study PIs who are the ultimate responsible for maintaining confidentiality and data protection. Only the researcher(s) employed on this study will have access to subject identifiable information. Identifiable information (name, DOB, and address) will be removed prior to analysis and subjects will be identified by pseudoanonymised codes. Personal data and key codes for pseudoanonymisation will be stored in a database with restricted access. Before conducting the statistical analyses, a number of checks will be run to ensure high quality of data.

Statistical analysis
Standard descriptive statistical methods (geometric means, geometric standard deviations, and 95% Confidence Intervals (95% C.I.) will be used to summarise the rugby history and lifestyle exposure data.
Analyses linking history of head trauma with neurological symptoms (obtained by questionnaire and neurobehavioural tests) will be then performed. Parameters of physical and cognitive capability alongside with clinical neurological intermediated endpoints will initially be dichotomised (yes/no symptoms), therefore analysed with prevalence odds ratios using logistic regression (96). We will also analyse continuous outcome measures using multiple linear regression. In addition to analysing individual symptoms, we will assess associations with domains of symptoms (i.e. physical and cognitive capability). For this purpose we will use cut-points previously published in the international literature (97) particularly related to the 1946 Birth Cohort whose testing protocol was used. For those analyses involving repeated measurements (neurobehavioural testing) assessing acute and chronic effects we will use multilevel models. Analyses will be adjusted for age, sex, and alcohol consumption.
Finally, we will compare neurological outcomes assessed by questionnaire with those assessed by computer-administered tests using Kappa-statistics (and in case of continuous outcomes linear regression analyses). We will also compare the strengths of the associations with exposure between both outcome measures.

Study size and power
We will invite about 205 participants to participate, half of whom have been exposed to concussion during their career (M Cross, personal communication). We conservatively assume that about 150 will participate. Based on previous studies, the standard deviations of the psychometric tests are in the range of 8-15% of the absolute value; assuming a conservative figure of 15% overall, the study will have more than 95% power to detect a 10% difference, and 80% power to detect a 7% difference, in psychometric test scores, between exposed (to concussion) and non-exposed participants.

Limitations
The main limitation of this study is the cross-sectional design. Information on exposure (history of concussion) and health outcomes is collected at one point in time making data subject to recall bias.
Participants with a worst health might be prone to recollect more precisely the number of concussions they suffered if they blame them for their current poor health status.
However, this type of study design will provide us with meaningful data on the burden of ill-health and neurological ill-health among retired rugby players in only a two-year time frame. Depending on results (and availability of additional funding), this study could inform a better, more detailed, but more timeconsuming cohort study aimed at assess the association between concussion in rugby players and health outcomes prospectively.
One of the direct consequences of the cross-sectional design is also an intrinsic risk of selection bias.
Rugby players who have developed serious neurodegenerative conditions (or have died from them) in the meantime are less likely (or not able) to participate in the study thus limiting our sample to a subsample of healthier retired rugby players. However, this won't impede investigating the association between exposure to concussion and general and neurological health, despite the fact that a smaller range of health outcomes will diminish the power for detecting an association as significant.
Nonetheless, the power of this study is calculated on the basis of continuous outcome measures (i.e. neurocognitive tests or grip strength measurement) making this study adequately powered to detect the planned associations based on the expected number of recruited participants.
Also, it is worth noting that results from this study will be not immediately apply to current rugby players as in the last 30 years or so as playing rules and conditions have been changed to increase player safety.

Ethics and dissemination
Ethical approval has been sought from the London School of Hygiene and Tropical Medicine (LSHTM) Ethics Committee. The study in general is focussing on subclinical problems (e.g. slightly lower than average scores in cognitive function tests) which would not usually require any medical attention. If any problems are identified we will (with the permission of the study participant) refer them to their GP. The consent form includes the following statement:

I wish to inform my GP that I am taking part to this study I wish the study team to inform my GP of any unusual test result
This study will provide a unique snapshot of physical and neurological health of retired rugby players who are now aged 50 years or more. This is valuable and precious information per se, potentially comparable with other occupational cohorts or the general population (accepting the possible bias discussed previously in comparing these samples). Moreover, within this study, data will be analysed in as a function of the history of traumatic brain injury and concussion during their careers so in order to assess associations between history of injuries and health-related outcomes.  A novel aspect of this project is the molecular component: clinical and neurocognitive findings will be particularly reinforced by the measurement of blood biomarkers. These will give a quantitative measurement of active neurodegeneration (NfL) and of possible CTE (tau) to be analysed conjunctly with the health-related outcomes and personal history of concussion.
Also, relying on a number characteristics of the rugby playing history collected (i.e. length of playing at elite level, role played, numbers of game played, age when started playing) it will be possible to identify potential categories of players at potentially increased risk of health related outcomes.

Chronic traumatic brain injury (TBI)
A spectrum of disorders associated with long-term consequences of single or repetitive TBI

Abstract
Introduction -Relatively little is known about the long-term health of former elite rugby players, or former sportspeople more generally. As well as the potential benefits of being former professional sportspersons, there may be potential health risks from exposures occurring during an individual's playing career, as well as following retirement. Each contact sport has vastly different playing dynamics, therefore exposing its players to different types of potential traumas. Current evidence suggests that these are not necessarily comparable in terms of pathophysiology, and their potential long-term adverse effects might also differ. There is currently limited but increasing evidence that poorer age-related and neurological health exists among former professional sportsmen exposed to repetitive concussions; however the evidence is limited on rugby union players, specifically.
Methods and analysis -We present the protocol for a cross-sectional study to assess the association between self-reported history of concussion during the playing career, and subsequent measures of healthy ageing and subtle neurological and cognitive impairment. We are recruiting a sample of approximately 200 retired rugby players (former Oxford and Cambridge University rugby players and members of the England Rugby International Club) aged 50 years or more, and collecting a number of general and neurological health-related outcome measures though validated assessments. Biomarkers of neurodegeneration (neurofilaments and tau) will be also be measured. Although the study is focusing on rugby union players specifically, the general study design and the methods for assessing neurological health are likely to be relevant to other studies of former professional sportspersons.
Ethics and dissemination -The study has been approved by the Ethical Committee of London School of Hygiene and Tropical Medicine (LSHTM) (reference: 11634 2). It is intended that results of this study will be published in peer reviewed medical journals, communicated to participants, the general public, and all relevant stakeholders.

Introduction
The evidence relating to head trauma in sport and the subsequent risks of neurological disease have previously reviewed, and it has been established that sports involving repeated head trauma may have an increased risks of neurodegenerative disease in the long-term (1). Furthermore, there are now plausible mechanisms for these effects, and a recognition that these problems do not just occur in former boxers, but in a variety of sports involving repeated concussions (2), and possibly also in sports in which low-level head trauma is common (3). These neurodegenerative effects include potentially increased risks of impaired cognitive function and dementia (4-7), Parkinson's disease (PD) (8)(9)(10)(11)(12)(13), and amyotrophic lateral sclerosis (ALS) (3,(14)(15)(16)(17)(18)(19). The term chronic traumatic encephalopathy (CTE) was introduced as a clinical-pathological construct for the neurodegeneration associated with American football and wrestling (20) (see Table 1). Seminal work by Ann McKee (21) investigating CTE has led to important deliberations on the overlapping neurodegenerative lesions in this condition (22). However, each contact sport has vastly different playing dynamics therefore exposing its players to different types of potential traumas. Current evidence suggests that these are not necessarily comparable in terms of pathophysiology, and hence in terms of their potential long-term adverse effects on health. There is currently limited but increasing evidence that poorer general and neurological health exists among professional sportsmen exposed to repetitive concussions; however, there is little evidence from rugby union players (23-25).
Decq et al (23)  in former rugby players (56.6%) (p=0.005). However, after adjustment for smoking and higher education, no association was observed between TICS-m score and number of reported concussions (23). Furthermore, in New Zealand, 366 former players were tested on their engagement in sport, general health, sports injuries and concussion history, and demographic information. Cognitive functioning was assessed using the online CNS Vital Signs neuropsychological test battery. The elite rugby group performed worse on tests of complex attention, processing speed, executive functioning, and cognitive flexibility than the non-contact-sport group, and worse than the community rugby group on complex attention. Former players who recalled one or more concussions had worse scores on cognitive flexibility, executive functioning, and complex attention than players who did not recall experiencing a concussion (24). A recent Scottish study assessed 52 former Scottish international rugby players and 29 controls. Players performed worse on a test of verbal learning, and of fine co-ordination  (25).
Ultimately, establishing the extent of these potential issues will require long-term prospective studies involving the repeated measurement of various to head trauma exposures and repeated tests of neurological health in large numbers of current and former players. Furthermore, identifying the risks, if any, of severe neurological problems, such as PD and ALS, will require large numbers of participants and long periods of follow-up. However, a first step in this process is to conduct cross-sectional studies in former players to assess whether there is an association between the history of concussion during their rugby career, and subsequent measures of healthy ageing and subtle neurological and cognitive impairment. In particular, such cross-sectional studies can be conducted in a relatively short time and with 'standard' measures of cognitive function since impaired cognitive function is an important health outcome in itself, and may be a precursor of more serious long-term neurological effects (26).
Furthermore, cross-sectional studies of this type can also be conducted in current players, and thus form the basis for prospective studies measuring changes in neurological health over time.

Definitions of trauma and concussion
Traumatic brain injury (TBI) is usually classified as mild, moderate, or severe, on the basis of the initial Glasgow Coma Scale (GCS) (27) recorded in the Emergency Departments including the duration of any loss of consciousness; and duration of post-traumatic amnesia (i.e., loss of memory of events after the injury) (28). Chronic TBI represents a spectrum of disorders associated with long-term consequences after single or repetitive TBI (29). Conversely, there remains no consensus on the definition of concussion. The 2012 Zurich Consensus Statement on Concussion in Sport proposed that concussion and mild TBI should be viewed as distinct entities (30). The group defined concussion as a "complex pathophysiological process affecting the brain", and despite allowing for the presence of neuropathological damage, they postulated that concussive symptoms largely reflected a functional disturbance, typically resolving spontaneously with no imaging abnormality. In contrast, recent guidelines from the American Academy of Neurology for sports concussion in 2013, do not separate concussion from mild TBI, defining concussion as "a clinical syndrome of biomechanically induced alteration of brain function, typically affecting memory and orientation, which may involve loss of consciousness" (31). A recent report (32) examines why having two different pathological entities might be unhelpful and suggests that the Mayo Clinic TBI Classification system should cover both definitions (33). A glossary of definitions used in this protocol is found in Table 1, adapted from Jordan (2014) (34).
For the purpose of this study, we adapted the NIH definition of concussion (35). Participants will be asked to report their previous concussions according with the following definition: Concussion is defined as an alteration in brain function, caused by an external force. Symptoms include: • A decreased level / loss of consciousness

Impaired cognitive function and dementia in sportspersons
While the neurological and cognitive effects of acute traumatic brain injuries (TBI) have been extensively studied (4-7), the association between TBI and delayed sequelae have been less studied because of the variable latency period before overt neurologic dysfunction, and the difficulty of retrospectively recalling relevant events (recall bias), in particular in the presence of memory impairment. The neurocognitive effects of repetitive mild head injury in sport were initially observed in boxers, a syndrome that was distinct from the clinical and pathological sequelae of single-incident severe TBI (36) (see also Table 1).
The clinical syndrome of dementia pugilistica (punch-drunk syndrome) is associated with prominent  (38). Also, post-mortem examination of the brains of several professional American football players and wrestlers has revealed the pathological underpinnings of cognitive and neuropsychiatric decline seen in these men in later life (20). Although cognitive decline in professional American football players has been suggested for some years, the first autopsy report from such a player was published only recently (39). In all cases, cognitive decline began years after retirement from formal sport participation.
Cognitive function is an ideal outcome measure for this type of study, as being a continuous measure maximizes power to identify an association; and using neurodegenerative disease as an outcome measure would require a substantially larger sample size. Cognitive function, as measured with neurocognitive tests, is also an aspecific measure, which does not imply a clinical diagnosis, and it can be the prodrome manifestation for a number of different neurodegenerative diseases.

Biomarkers
Research in the field of TBI biomarkers has increased exponentially over the last 20 years (29,(40)(41)(42), with studies assessing biomarkers that could provide diagnostic and prognostic, as well as monitoring information. S100β, glial fibrillary acidic protein (GFAP), neuron-specific enolase (NSE), tau, neurofilament light protein (NF-L), amyloid beta, brain-derived neurotrophic factor (BDNF), serum creatinine kinase (CK), and heart-type fatty acid binding protein (hFABP), prolactin, cortisol, and albumin are all biomarkers more frequently studied in relation to sport-related concussion (43,44). However, to date, no biomarker for the long-term effects of concussion has been identified, although recent studies on cerebrospinal fluid (CSF) from sportsmen with post-concussion syndrome suggest that a subset of these have biomarker signs of ongoing axonal injury and microglial activation (45)(46)(47).
Of the biomarkers measurable in serum or other body fluids, those more likely to be detectable for longer time periods after the concussion episode, are those related to the axonal injury and more in generally neuroinflammation, that is neurofilament and tau protein (44). Neurofilaments (NF) coassemble from protein subunits to form NFs defined as light (light (NF-L), medium (NF-M), or heavy (NF-H) according to their relative molecular weights). NFs are one of the key structural elements of neurons,  (51). A study compared serum NF-L, a biological marker of head trauma, in American football athletes and non-contact sport athletes and examined changes over the course of a season. Results suggest that a season of collegiate American football is associated with elevations in serum NF-L, which is indicative of axonal injury, resultant of head impacts (52). NF-L has also been associated with head trauma severity detected by CT scan, immediately after the episode (53) and with concussive and subconcussive head impacts in boxing (47). Importantly, whilst plasma tau concentration increases and disappears rapidly (within hours to a few days) following concussion (54), NF-L has a prolonged increased levels lasting for many weeks (47,55,56).
Phosphorylation of tau is a normal event in healthy neurons, but hyperphosphorylation and aggregation into neurofibrillary tangles is a characteristic of Alzheimer's disease and CTE (57). Tau concentrations correlate with lesion size and outcome in severe TBI when measured in the ventricular CSF, while remaining unchanged in TBI and other forms of acute brain injury (44). Studies on mild TBI show increased CSF concentrations of both t tau and NF L, although the increase in CSF NF-L is greater than the increase in t-tau, suggesting that head impacts have a greater effect on long, large-calibre axons that extend subcortically than on short, nonmyelinated axons in the cortex (44). After concussion, serum tau concentrations were found to be increased, but this increase did not correlate with severity of trauma and lesion load as measured using CT scan signs (44).

Genetics
Few hypotheses on the potential role of genetics in modulating the possible association between concussion and cognitive decline have been formulated. Specific genes might be associated with an increased risk of concussion via worse attention or executive function, or more vulnerable brain anatomy (58), or via personality trait (59). Conversely, a number of genes are involved in modulating the risk of developing dementia and other neurodegenerative diseases, irrespective of concussion.
Dementia risk is higher amongst carriers of the epsilon4 allele at the Apolipoprotein E (APOE) gene (60).

Rationale of the BRAIN Study
Recently, World Rugby (formally known as the International Rugby Board), has developed a process to support team clinicians in the recognition, assessment, and subsequent management of elite adult players who have sustained a potential concussion. This process includes the development of a -were recruited. Self-reported demographic factors, playing history (including head trauma), past medical history (including dementia, depression, and memory impairment), and perceived health were collected, in addition to detailed information on musculoskeletal health and pain.
A total of 319 participants have participated in this study to date, and 205 of those aged 50 years or older have agreed to be re-contacted for further studies. The 205 participants aged 50+ years who agreed to be re-contacted will be invited to take part to the BRAIN Study.

Aim and objectives
The overall aim of the BRAIN study is to investigate the possible associations between concussion in rugby and ageing, including physical and cognitive capabilities, as well intermediate neurological and musculoskeletal endpoints among former rugby players.
In order to achieve these aims the following objectives will be pursued: 1. To investigate the associations between self-reported concussion history and ageing, measured as physical and cognitive capabilities. This will be achieved by using the following outcome  In addition, we will develop a multimedia database (data from tests and questionnaire plus videorecorded neurological examination plus biobank of blood samples) that will serve as baseline for further tests and further follow-up.

Study design
Participants will be invited to study clinics in London, Manchester, or Bath. A home visit can also be arranged by request. Participants will be invited to bring with them any medication they are taking regularly for an accurate recording of current medication usage.
All participants will be administered a Core Module interview including questions on lifestyle factors, potential confounders, and extensive information on the five domains of physical and cognitive ability, neurological examination, intermediate neurological outcomes and musculoskeletal health.
The lifestyle questionnaire will complement that already administered to participants of the Oxford-base cross-sectional study, and will include questions on potential confounders of the association between concussion and physical and cognitive capability, and intermediate neurological outcomes. Participants will also be asked to donate a blood sample (a normal blood sample will be collected by participants seen in one of the clinics, a dry spot sample plus saliva swab for DNA will be collected from participants seen at home). At the end of the Core Module, all participants will be also asked to undergo some additional tests (the Additional Optional Module) providing they have sufficient time (Figure 1). At the end of the interview, all participants will be asked some final questions on their concussion history. This will ensure that the interviewer is blind to the participant's concussion history during the entire duration of the interview, and will also act as validation to confirm data previously collected. Participants who disclose information about their concussion history during the interview, will be noted in order to undertake a sensitivity analysis excluding them. . This Additional Optional Module should take no longer than 1 hour and half to be completed ( Figure 1). The full test/interview including core and additional module should not take more than 3 hours and 15 minutes to complete.

General Questionnaires
Questionnaires will be used to collect relevant information on: Lifestyle and confoundersall participants will be asked a number of questions on their lifestyle in order to collect information on possible confounders for the main analysis (smoking, alcohol, coffee, drugs, past medical history, sleep quality).
Concussion historyat the end of the interview, participants will be re-asked some detailed information regarding their history of head trauma and concussion whilst playing.
Musculoskeletal hand pain -A questionnaire on hand pain to identify possible hand osteoarthritis will be utilized involving a hand mannequin (additional optional module).

Physical ability assessment
The methodology used in the 1946 birth cohort (67) will be leveraged to assess physical and cognitive capabilities in this study. This will enhance comparability of results and facilitate future collaborations.
The most commonly used objective measures of physical capability for assessing healthy ageing, are  (68); these aim to assess physical functioning, including the capacity to undertake the physical tasks of daily living (73). There is robust evidence that higher scores on these measures are associated with lower rates of mortality, and there is more limited evidence of lower risks of morbidity, and of age-related patterns of change (73).
Height and weightheight and weight will be collected in order to calculate the body mass index (BMI) according to the formula BMI = weight (Kg) / height (m) 2 .
Grip strength -JAMAR Hydraulic Hand Dynamometer is the most widely used instrument with established test-retest, inter-rater and intra-rater reliability (74).
Chair rise -The 30-Second Chair Stand Test will be used to assess lower body strength (75). This test provides a reasonably reliable and valid indicator of lower body strength in generally active, communitydwelling older adults. Test-retest intra-class correlations of 0.84 for men and 0.92 for women, utilizing one-way analysis of variance procedures appropriate for a single trial, together with a non-significant change in scores from Day 1 testing to Day 2, indicate that the 30-s chair stand has good stability reliability (75). Moreover, a moderately high association between chair-stand performance and maximum weight-adjusted leg-press performance for both men and women (r = 0.78 and 0.71, respectively) supports the criterion-related validity of the chair stand as a measure of lower body strength (75). As expected, chair-stand performance decreased significantly across age groups in decades -from the 60s to the 70s to the 80s (p < 0.01) and was highly statistically significantly lower for low-active participants than for high-active participants (p < 0.0001) (75). A photograph of the hands will be taken at the end of this section to assess inflammation, finger-nodes, and other rheumatological and osteoarthritis-related changes at the hand.

Cognitive ability assessment
The methodology used in the 1946 birth cohort will also be followed for the assessment of cognitive capability (67). Tests and questionnaires will be described according their belonging to the Core Module or the Additional Optional Module, and by domain ( Figure 1). (79)) -The MMSE is a widely used 30-point screening tool for cognitive impairment within clinical practice, assessing multiple cognitive domains including orientation to time and place; registration; attention +/-calculation; recall; language; repetition; reading; writing; visuospatial function; and executive function and praxis.

Logical Memory from the Wechsler Memory Scale-Revised (WMS-R (80)) -The Logical Memory test
assesses free recall of a short story. The participant is asked to recall the story immediately and after a 20minute delay.

12-item Face-Name Associative Memory Exam (FNAME-12A) -
The FNAME-12A is a modified version of the 16-item Face-Name Associative Memory Exam (FNAME-16). The FNAME-12A has fewer stimuli and  14 additional learning trials which have been found to be well tolerated by those with Mild Cognitive Impairment, whilst remaining challenging in cognitively normal older adults (81). The FNAME-12A has demonstrated psychometric equivalence with the FNAME-16, which has been shown to be related to beta-amyloid burden in cognitively normal elderly people (82). The FNAME-12A requires the participant to learn 12 face-name and face-occupation pairs. Participants are given two exposures to all 12 facename/occupation pairs. After each exposure, and following a 5-minute delay, participants are asked for the name and occupation associated with each face. After a 30-minute delay they are shown 3 faces and asked to identify the face that they recognise and give the name and occupation. Given multiple options to choose from, they are then asked to select the name and/or occupation associated with the face. (83,84) -The task-set shifting and response inhibition task examines the relationship between executive tasks of task-switching/preparation time. In the arrow only condition, participants are shown the cue 'arrow'. Following a short delay they must respond to the direction of the arrow ('right' or 'left'). In the word only condition, participants are shown the cue 'word'. Following a short delay they must respond to the direction of the word ('right' or 'left'). There is a switching condition in which the participant is shown the cue 'arrow' or 'word'. Following a delay, both a combined arrow and word stimulus appears. The stimulus is either congruent (left arrow and left word), or incongruent (left arrow and right word). Trials in the switching task are categorised into switch and non-switch. In a non-switch trial the cue is the same as for the immediately preceding trial. In a switch trial the cue differs from the immediately preceding trial.

Digit-Symbol Substitution Test, from the Wechsler Adult Intelligence Scale-Revised (WAIS-R (85)) -The
Digit-Symbol Substitution test explores attention and psychomotor speed. Participants are given a code table displaying digits (from 1 to 9); each digit is paired with a symbol. The participant is required to complete in blank squares with the corresponding symbol for each digit as shown in the code table.
They are given 90 seconds to fill in as many squares as possible. (86,87) -This test requires the participant to view one or three fractal objects, presented simultaneously in random locations on the screen. The participant is asked to remember both the objects and their location. After a delay of 1 or 4 seconds they have to make a forced choice between one of the displayed fractals (the target) and a 'dummy' fractal. Participants are required to touch the object they think has been previously presented and 'drag' it on the touch screen to its remembered, original location. The binding of such featured information has been shown to be vulnerable in asymptomatic FAD mutation carriers (87).   (89) -This is a circle tracing task which includes both direct and indirect visual feedback conditions. Continuous performance measures are provided including accuracy, speed, and speed of error detection and correction. The test has revealed changes in speed and accuracy in Huntington Disease mutation carriers more than 10 years before expected age-of-onset (Additional optional module).

Matrix Reasoning from the Wechsler Abbreviated Scale of Intelligence (WASI (90)) -The Matrix
Reasoning test assesses nonverbal reasoning. The participant is shown a matrix of geometric shapes with a section missing. They are required to select the option that completes the matrix (Additional optional module).

Irrelevant Distractor Paradigm
(91) -Participants are given a computerised letter-search task and are required to make a rapid decision as to whether the target letter 'X' or 'N' has appeared in the search display (in either low or high load conditions). On some of the trials, a task-irrelevant distractor (a cartoon character) appears on the outside of the search display. The task evaluates the extent to which attention is captured and captivated by the distractor (Additional optional module).

Neurological clinical assessment
The neurological assessment of participants is needed for detecting any subtle neurological sign, and to establish a baseline for the absence of one or more signs. A video-recorded standard neurological examination will be included as part of the Core Module. An additional test of self-reported impairment, mainly due to movement disorders, will be administered as part of the Additional Optional Module ( Figure 1).
A standard video recording of each participant accessing the study at one of the clinics will be performed and evaluated by a clinical neurologist. The assessment includes examination of strength, Theexamination includes also a test of knee functionality (92).
Information on potential movement disorders will be collected using the unified Parkinson's Disease Rating Scale (PDDRS) part II. This includes self-reported ratings on several motor domains, including speech, saliva and drooling, chewing and swallowing, eating, dressing, hygiene, handwriting, hobbies, turning in bed, tremor, standing up, walking and balance, and freezing (Additional optional module).

Intermediate neurological endpoints
In addition to neurological signs and symptoms, some intermediate outcomes have been recognised as part of the complex clinical picture of movement disorders, manifesting before the onset of the movement impairment itself. These will be investigated with three tests described below (Figure 1).

Blood and saliva sample collection, and bio-banking
Participants seen in the research clinics will be asked to donate a blood sample for testing for tau protein, neurofilament (Nf) levels, full blood count analysis, and to obtain data for a GWAS. Samples of serum, plasma, and DNA will be collected and processed according to a pre-specified protocol and stored in a freezer at -80 °C at the Blizard Institute (Queen Mary, University of London). All participants will also asked to donate a blood sample through a dried blood spot test and a saliva sample (in order to obtain DNA where phlebotomy is not possible).
Samples will be tested to measure the expression of NfL (Neurofilament light) , a Nf-subunit showing consistent and reproducible measurements in plasma/serum using a newly developed sensitive methodology exploiting single molecule trapping and measurement by ELISA (SIMOA) (97-99). The use of both NfL and NfH (heavy) allows for the determination of different markers whose dynamic of release into bio fluids following injury vary according to their particular chemical structure. NfL has shown good levels of linearity on dilution experiments, suggesting that confounders like aggregation or immune response may not interfere with its measurements. NfH different phosphorylation states in blood may in turn provide more information on the systemic biological changes induced by trauma. The relative low abundance of these proteins in blood, in absence of CSF, which is normally more enriched in byproducts of neuronal destruction, will not represent a major obstacle, as analytical sensitivity has evolved with the development of novel assays which cover the lower end of these protein dynamic range. SIMOA, the proposed methodology for the target biomarkers, is now emerging as the core GWAS will be conducted at UCL in the lab led by Prof John Hardy.

Multimedia database set up and management
A multimedia database will be established as part of this project. The final dataset will include questionnaire and tests data (see Figure 1), in addition to the video-recorded neurological examination.
This data will be also linked to the stored blood samples through a unique identifier.
The data will be stored in mirror, anonymised datasets at LSHTM and QMUL on secure servers, and access will be regulated by the study PIs who are ultimately responsible for maintaining confidentiality and data protection. Only the researcher(s) employed on this study will have access to subject identifiable information. Identifiable information (name, date of birth, and address) will be removed prior to analysis, and subjects will be identified by pseudoanonymised codes. Personal data and key codes for pseudoanonymisation will be stored in a database with restricted access. Before conducting the statistical analyses, checks will undertaken to ensure data quality.

Statistical analysis
Descriptive statistical methods (geometric means, geometric standard deviations, and 95% Confidence Intervals (95% C.I.) will be used to summarise the rugby history and lifestyle exposure data. Analyses linking a history of head trauma with neurological symptoms (obtained by questionnaire and neurobehavioural tests) will then be performed. Parameters of physical and cognitive capability, alongside clinical neurological intermediate endpoints, will initially be dichotomised (yes/no symptoms), and analysed with prevalence odds ratios using logistic regression (100). We will also analyse continuous outcome measures using multiple linear regression. In addition to analysing individual symptoms, we will assess associations with domains of symptoms (i.e. physical and cognitive capability).
For this purpose we will use cut-points previously published in the literature (101), particularly related to the 1946 Birth Cohort whose testing protocol was used. For analyses involving repeated measurements (neurobehavioural testing) assessing acute and chronic effects, we will use multilevel models. Analyses will be adjusted for age, sex, and alcohol consumption.
Finally, we will compare neurological outcomes assessed by questionnaire with those assessed by computer-administered tests using Kappa-statistics (and in case of continuous outcomes linear regression analyses). We will also compare the strengths of the associations with exposure between both outcome measures.

Study size and power
We aim to invite 205 participants to participate, approximately half of whom have been exposed to concussion during their career (M Cross, personal communication). We conservatively estimate that about 150 former players will participate. Based on previous studies, the standard deviations of the psychometric tests are in the range of 8-15% of the absolute value; assuming a conservative figure of 15% overall, the study will have more than 95% power to detect a 10% difference, and 80% power to detect a 7% difference, in psychometric test scores, between exposed (to concussion) and non-exposed participants.

Limitations
The main limitation of this study is the cross-sectional design. Data on lifetime exposure (history of concussion) and health outcomes are collected at one time point, making data subject to recall bias.
Participants with a poorer health, or sentiment of discontent with their health status, may overestimate concussive exposure, if they attribute their poor health to it.
However, this cross-sectional study design will provide meaningful data on the burden of ill health and neurological health among retired rugby players, in a relatively short two-year time frame. Depending on results (and availability of additional funding), this study could inform a, more detailed, and time intensive cohort study, aimed at assessing the association between concussion in rugby players and health outcomes prospectively.
A direct consequence of the cross-sectional design is an intrinsic risk of selection bias. Rugby players who have developed serious neurodegenerative conditions (or have died from them) in the meantime are less likely (or not able) to participate in the study thus limiting our sample to a subsample of healthier retired rugby players. On the other hand, it is possible that former players who experience some cognitive symptoms will be more keen to participate in the study, to seek reassurance and be examined in more detail.
Irrespective of these considerations, our investigation of the association between the exposure to sporting concussion and later-life general and neurological health outcomes remains valid, despite a smaller range of health outcomes diminishing the power for detecting an association as significant.
Nonetheless, the power calculation for this study is calculated on the basis of continuous outcome measures (i.e. neurocognitive tests or grip strength measurement) making this study adequately powered to detect the potential associations, based on the expected number of recruited participants.
It is worth noting that results from this study will be not immediately apply to current rugby players, due to this cohort's former playing status, history of amateur playing exposure and alterations to rules and conditions, aimed at increasing plaer safety over recent years, which will have altered player's overall exposure to concussive events..

Ethics and dissemination
Ethical approval has been granted from the London School of Hygiene and Tropical Medicine (LSHTM) Ethics Committee. The study in general is focusing on subclinical problems (e.g. slightly lower than average scores in cognitive function tests) which would not usually require any medical attention. If any problems are identified, we will (with the permission of the study participant) refer participants to their primary care physician. The consent form includes the following statement:

I wish to inform my GP that I am taking part to this study I wish the study team to inform my GP of any unusual test result
This study will provide unique insight into the physical and neurological health of retired rugby players who are now aged 50 years or more. This is valuable information per se, potentially comparable with other occupational cohorts or the general population (accepting the potential for bias discussed previously in comparing these samples). Moreover, within this study, data will be analysed as a function of the history of traumatic brain injury and concussion during their careers, and therefore assess associations between the history of reported concussive injuries and health-related outcomes.  A novel aspect of this project is the molecular component: clinical and neurocognitive findings will be particularly reinforced by the measurement of blood biomarkers. These will give a quantitative measurement of active neurodegeneration (NfL) and of possible CTE (tau), to be analysed conjunctly with the health-related outcomes and personal history of concussion.
Also, relying on a number characteristics of the rugby playing history collected (i.e. length of playing at elite level, position played, numbers of game played, age when started playing), it will be possible to identify potential categories of players at potentially increased risks of health-related outcomes.
This study may identify a small deviation from normality in any of the continuous measures collected via cognitive and other tests (i.e. grip strength, memory function, walking speed, etc). This will be mainly investigated in association with a history of frequent concussions, or other head traumas. The investigators will interpret the results based on the continuum of distribution of the test outcomes, which will not have specific clinical translation. However, it should be emphasised that although this study is unlikely to identify patterns suggesting a clinical disease in any of the participants, if these are identified, then the relevant participant(s) will be referred to appropriate medical services.
These data will be disseminated initially to players and sports governing bodies, in addition to through peer reviewed publication and conference presentation, in order to begin to establish an evidence basis for the association between exposure to sports-related concussion, and potential later life cognitive and neurological impairment.

Abstract
Introduction -Relatively little is known about the long-term health of former elite rugby players, or former sportspeople more generally. As well as the potential benefits of being former professional sportspersons, there may be potential health risks from exposures occurring during an individual's playing career, as well as following retirement. Each contact sport has vastly different playing dynamics, therefore exposing its players to different types of potential traumas. Current evidence suggests that these are not necessarily comparable in terms of pathophysiology, and their potential long-term adverse effects might also differ. There is currently limited but increasing evidence that poorer age-related and neurological health exists among former professional sportsmen exposed to repetitive concussions; however the evidence is limited on rugby union players, specifically.
Methods and analysis -We present the protocol for a cross-sectional study to assess the association between self-reported history of concussion during the playing career, and subsequent measures of healthy ageing and subtle neurological and cognitive impairment. We are recruiting a sample of approximately 200 retired rugby players (former Oxford and Cambridge University rugby players and members of the England Rugby International Club) aged 50 years or more, and collecting a number of general and neurological health-related outcome measures though validated assessments. Biomarkers of neurodegeneration (neurofilaments and tau) will be also be measured. Although the study is focusing on rugby union players specifically, the general study design and the methods for assessing neurological health are likely to be relevant to other studies of former professional sportspersons.
Ethics and dissemination -The study has been approved by the Ethical Committee of London School of Hygiene and Tropical Medicine (LSHTM) (reference: 11634 2). It is intended that results of this study will be published in peer reviewed medical journals, communicated to participants, the general public, and all relevant stakeholders.

Strengths and limitations
• The study will provide meaningful data on the burden of health and neurological health among retired rugby players (in only two-year's time) • The cross-sectional design does not allow preventing from the potential for recall bias and selection bias • Results will not be immediately generalizable to current players as in the last 30 years playing rules and conditions are changed, increasing the player safety

Introduction
The evidence relating to head trauma in sport and the subsequent risks of neurological disease have previously reviewed, and it has been established that sports involving repeated head trauma may have an increased risks of neurodegenerative disease in the long-term (1). Furthermore, there are now plausible mechanisms for these effects, and a recognition that these problems do not just occur in former boxers, but in a variety of sports involving repeated concussions (2), and possibly also in sports in which low-level head trauma is common (3). These neurodegenerative effects include potentially increased risks of impaired cognitive function and dementia (4-7), Parkinson's disease (PD) (8)(9)(10)(11)(12)(13), and amyotrophic lateral sclerosis (ALS) (3,(14)(15)(16)(17)(18)(19). The term chronic traumatic encephalopathy (CTE) was introduced as a clinical-pathological construct for the neurodegeneration associated with American football and wrestling (20) (see Table 1).
Each contact sport has vastly different playing dynamics therefore exposing its players to different types of potential traumas. Current evidence suggests that these are not necessarily comparable in terms of pathophysiology, and hence in terms of their potential long-term adverse effects on health. There is currently limited but increasing evidence that poorer general and neurological health exists among professional sportsmen exposed to repetitive concussions; however, there is little evidence from rugby union players (21-23).
Decq et al (21) investigated retired French-speaking high-level sportsmen, aged 45 to 65 years, who had played sports for at least 10 years. Mild cognitive disorder was lower in players of other sports (40.4%) than in former rugby players (56.6%) (p=0.005). However, after adjustment for smoking and higher education, no association was observed between cognitive function and number of reported concussions (21). In New Zealand, 366 former players were tested on their engagement in sport, general health, sports injuries and concussion history, demographic information and cognitive functioning. The elite rugby group performed worse on tests of complex attention, processing speed, executive functioning, and cognitive flexibility than the non-contact-sport group, and worse than the community rugby group on complex attention. Former players who recalled one or more concussions had worse scores on cognitive flexibility, executive functioning, and complex attention than players who did not recall experiencing a concussion (22). A recent Scottish study assessed 52 former Scottish international rugby players and 29 controls. Players performed worse on a test of verbal learning, and of fine coordination of the dominant hand; however, no statistically significant differences were observed on other cognitive tests. Additionally, no significant association was found between the number of concussions and cognitive test performance (23).
Ultimately, establishing the extent of these potential issues will require long-term prospective studies involving the repeated measurement to head trauma exposures and repeated tests of neurological health in large numbers of current and former players. However, a first step in this process is to conduct cross-sectional studies in former players to assess whether there is an association between the history of concussion during their rugby career, and subsequent measures of healthy ageing and subtle neurological and cognitive impairment. In particular, such cross-sectional studies can be conducted in a relatively short time and with 'standard' measures of cognitive function since impaired cognitive function is an important health outcome in itself, and may be a precursor of more serious long-term neurological effects (24).

Definitions of trauma and concussion
Traumatic brain injury (TBI) is usually classified as mild, moderate, or severe, on the basis of the initial Glasgow Coma Scale (GCS) (25) recorded in the Emergency Departments including the duration of any loss of consciousness; and duration of post-traumatic amnesia (i.e., loss of memory of events after the injury) (26). Chronic TBI represents a spectrum of disorders associated with long-term consequences after single or repetitive TBI (27). Conversely, there remains no consensus on the definition of concussion. The 2012 Zurich Consensus Statement on Concussion in Sport proposed that concussion and mild TBI should be viewed as distinct entities (28). The group defined concussion as a "complex pathophysiological process affecting the brain", and despite allowing for the presence of neuropathological damage, they postulated that concussive symptoms largely reflected a functional disturbance, typically resolving spontaneously with no imaging abnormality. In contrast, recent guidelines from the American Academy of Neurology for sports concussion in 2013, do not separate concussion from mild TBI, defining concussion as "a clinical syndrome of biomechanically induced alteration of brain function, typically affecting memory and orientation, which may involve loss of consciousness" (29). A recent report (30) examines why having two different pathological entities might be unhelpful and suggests that the Mayo Clinic TBI Classification system should cover both definitions (31). A glossary of definitions used in this protocol is found in Table 1, adapted from Jordan  For the purpose of this study, we adapted the NIH definition of concussion (33). Participants will be asked to report their previous concussions according with the following definition: Concussion is defined as an alteration in brain function, caused by an external force. Symptoms include: • A decreased level / loss of consciousness • Memory Loss (before or after the injury) • Weakness • Temporary Paralysis

Biomarkers
Research in the field of TBI biomarkers has increased exponentially over the last 20 years (27,(34)(35)(36), with studies assessing biomarkers that could provide diagnostic and prognostic, as well as monitoring information (37,38). However, to date, no biomarker for the long-term effects of concussion has been identified, although recent studies on cerebrospinal fluid (CSF) from sportsmen with post-concussion syndrome suggest that a subset of these have biomarker signs of ongoing axonal injury and microglial activation (39)(40)(41).
Of the biomarkers measurable in serum or other body fluids, those more likely to be detectable for longer time periods after the concussion episode, are those related to the axonal injury and more in generally neuroinflammation, that is neurofilament and tau protein (38). Neurofilaments (NF) coassemble from protein subunits to form NFs defined as light (light (NF-L), medium (NF-M), or heavy (NF-H) according to their relative molecular weights). NFs are one of the key structural elements of neurons,  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  in experimental models of TBI that produce cortical contusion in combination with selective hippocampal neuronal death (45). A study compared serum NF-L, a biological marker of head trauma, in American football athletes and non-contact sport athletes and examined changes over the course of a season. Results suggest that a season of collegiate American football is associated with elevations in serum NF-L, which is indicative of axonal injury, resultant of head impacts (46). NF-L has also been associated with head trauma severity detected by CT scan, immediately after the episode (47) and with concussive and subconcussive head impacts in boxing (41). Importantly, whilst plasma tau concentration increases and disappears rapidly (within hours to a few days) following concussion (48), NF-L has a prolonged increased levels lasting for many weeks (41,49,50).
Phosphorylation of tau is a normal event in healthy neurons, but hyperphosphorylation and aggregation into neurofibrillary tangles is a characteristic of Alzheimer's disease and CTE (51). Tau concentrations correlate with lesion size and outcome in severe TBI when measured in the ventricular CSF, while remaining unchanged in TBI and other forms of acute brain injury (38). Studies on mild TBI show increased CSF concentrations of both t tau and NF L, although the increase in CSF NF-L is greater than the increase in t-tau, suggesting that head impacts have a greater effect on long, large-calibre axons that extend subcortically than on short, nonmyelinated axons in the cortex (38). After concussion, serum tau concentrations were found to be increased, but this increase did not correlate with severity of trauma and lesion load as measured using CT scan signs (38).

Genetics
Few hypotheses on the potential role of genetics in modulating the possible association between concussion and cognitive decline have been formulated. Specific genes might be associated with an increased risk of concussion via worse attention or executive function, or more vulnerable brain anatomy (52), or via personality trait (53). Conversely, a number of genes are involved in modulating the risk of developing dementia and other neurodegenerative diseases, irrespective of concussion.
Dementia risk is higher amongst carriers of the epsilon4 allele at the Apolipoprotein E (APOE) gene (54). A total of 319 participants have participated in this study to date, and 205 of those aged 50 years or older have agreed to be re-contacted for further studies. The 205 participants aged 50+ years who agreed to be re-contacted will be invited to take part to the BRAIN Study.

Aim and objectives
The overall aim of the BRAIN study is to investigate the possible associations between concussion in rugby and ageing, including physical and cognitive capabilities, as well intermediate neurological and musculoskeletal endpoints among former rugby players.
In order to achieve these aims the following objectives will be pursued: 1. To investigate the associations between self-reported concussion history and ageing, measured as physical and cognitive capabilities. This will be achieved by using the following outcome 5. Investigate the long-term musculoskeletal health outcomes of rugby players with particular emphasis on hip, knee, and hand osteoarthritis allowing changes over time (from the previous study to the current study) to be assessed In addition, we will develop a multimedia database (data from tests and questionnaire plus videorecorded neurological examination plus biobank of blood samples) that will serve as baseline for further tests and further follow-up.

Study design
Participants will be invited to study clinics in London, Manchester, or Bath. A home visit can also be arranged by request. Participants will be invited to bring with them any medication they are taking regularly for an accurate recording of current medication usage.
All participants will be administered a Core Module interview including questions on lifestyle factors, potential confounders, and extensive information on the five domains of physical and cognitive ability, neurological examination, intermediate neurological outcomes and musculoskeletal health.
The lifestyle questionnaire will complement that already administered to participants of the Oxford-base cross-sectional study, and will include questions on potential confounders of the association between concussion and physical and cognitive capability, and intermediate neurological outcomes. Participants will also be asked to donate a blood sample (a normal blood sample will be collected by participants seen in one of the clinics, a dry spot sample plus saliva swab for DNA will be collected from participants seen at home). At the end of the Core Module, all participants will be also asked to undergo some additional tests (the Additional Optional Module) providing they have sufficient time (Figure 1). At the end of the interview, all participants will be asked some final questions on their concussion history. This will ensure that the interviewer is blind to the participant's concussion history during the entire duration of the interview, and will also act as validation to confirm data previously collected. Participants who disclose information about their concussion history during the interview, will be noted in order to undertake a sensitivity analysis excluding them. . This Additional Optional Module should take no longer than 1 hour and half to be completed ( Figure 1). The full test/interview including core and additional module should not take more than 3 hours and 15 minutes to complete.

General Questionnaires
Questionnaires will be used to collect relevant information on: Lifestyle and confoundersall participants will be asked a number of questions on their lifestyle in order to collect information on possible confounders for the main analysis (smoking, alcohol, coffee, drugs, past medical history, sleep quality).
Concussion historyat the end of the interview, participants will be re-asked some detailed information regarding their history of head trauma and concussion whilst playing.
Musculoskeletal hand pain -A questionnaire on hand pain to identify possible hand osteoarthritis will be utilized involving a hand mannequin (additional optional module).

Physical ability assessment
The methodology used in the 1946 birth cohort (59) will be leveraged to assess physical and cognitive capabilities in this study. This will enhance comparability of results and facilitate future collaborations.
The most commonly used objective measures of physical capability for assessing healthy ageing, are tests of grip strength (60,61), walking speed (60,62,63), chair rises (64) and standing balance (60); these aim to assess physical functioning, including the capacity to undertake the physical tasks of daily living (65). There is robust evidence that higher scores on these measures are associated with lower rates of mortality, and there is more limited evidence of lower risks of morbidity, and of age-related patterns of change (65). Grip strength -JAMAR Hydraulic Hand Dynamometer is the most widely used instrument with established test-retest, inter-rater and intra-rater reliability (66).
Chair rise -The 30-Second Chair Stand Test will be used to assess lower body strength (67). This test provides a reasonably reliable and valid indicator of lower body strength in generally active, communitydwelling older adults. Test-retest intra-class correlations of 0.84 for men and 0.92 for women, utilizing one-way analysis of variance procedures appropriate for a single trial, together with a non-significant change in scores from Day 1 testing to Day 2, indicate that the 30-s chair stand has good stability reliability (67). Moreover, a moderately high association between chair-stand performance and maximum weight-adjusted leg-press performance for both men and women (r = 0.78 and 0.71, respectively) supports the criterion-related validity of the chair stand as a measure of lower body strength (67). As expected, chair-stand performance decreased significantly across age groups in decades -from the 60s to the 70s to the 80s (p < 0.01) and was highly statistically significantly lower for low-active participants than for high-active participants (p < 0.0001) (67).
Walking speed -To assess normal comfortable walking speed and maximum walking speed, a Timed 10-Meter Walk Test will be used. This test has been validated in 230 healthy volunteers (62). Mean comfortable gait speed ranged from 127.2 cm/s for women in their seventies to 146.2 cm/s for men in their forties. Mean maximum gait speed ranged from 174.9 cm/s for women in their seventies to 253-3 cm/s for men in their twenties. Both gait speed measures were reliable (correlation coefficients > 0.903) and correlated significantly with age (r>-0.210), height (r>0.220) and the strengths of four measured lower extremity muscle actions (r = 0.190-0.500). These normative values give clinicians a reference against which patient performance can be compared in a variety of settings (62).
A photograph of the hands will be taken at the end of this section to assess inflammation, finger-nodes, and other rheumatological and osteoarthritis-related changes at the hand.   13 and function in sport and recreation. It has been used in subjects with hip disability with or without hip osteoarthritis (OA) (68) (Additional optional module).

KOOS -
The Knee Injury and Osteoarthritis Outcome Score (KOOS) examines pain frequency and severity during functional activities, and symptoms such as the severity of knee stiffness, the presence of swelling, grinding or clicking, difficulty in ADL, with sport and recreation, and knee-related quality of life.
It is intended for use in young and middle-aged populations with post-traumatic OA, in addition to those with injuries who may go on to develop secondary OA (69) (Additional optional module).
QuIKS -The Questionnaire to Identify Knee Symptoms (QuIKS) is used for early osteoarthritis, and understanding symptomology and potential adaptation to activity before osteoarthritis has been clinically diagnosed. It has been developed in adults aged 40-65 with evidence of ongoing knee problems and recommended for use in studies exploring early osteoarthritis (70) (Additional optional module).

Cognitive ability assessment
The methodology used in the 1946 birth cohort will also be followed for the assessment of cognitive capability (59). Tests and questionnaires will be described according their belonging to the Core Module or the Additional Optional Module, and by domain ( Figure 1).

12-item Face-Name Associative Memory Exam (FNAME-12A) -
The FNAME-12A is a modified version of the 16-item Face-Name Associative Memory Exam (FNAME-16). The FNAME-12A has fewer stimuli and additional learning trials which have been found to be well tolerated by those with Mild Cognitive Impairment, whilst remaining challenging in cognitively normal older adults (73). The FNAME-12A has demonstrated psychometric equivalence with the FNAME-16, which has been shown to be related to beta-amyloid burden in cognitively normal elderly people (74). The FNAME-12A requires the participant  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  to learn 12 face-name and face-occupation pairs. Participants are given two exposures to all 12 facename/occupation pairs. After each exposure, and following a 5-minute delay, participants are asked for the name and occupation associated with each face. After a 30-minute delay they are shown 3 faces and asked to identify the face that they recognise and give the name and occupation. Given multiple options to choose from, they are then asked to select the name and/or occupation associated with the face. (75,76) -The task-set shifting and response inhibition task examines the relationship between executive tasks of task-switching/preparation time. In the arrow only condition, participants are shown the cue 'arrow'. Following a short delay they must respond to the direction of the arrow ('right' or 'left'). In the word only condition, participants are shown the cue 'word'. Following a short delay they must respond to the direction of the word ('right' or 'left'). There is a switching condition in which the participant is shown the cue 'arrow' or 'word'. Following a delay, both a combined arrow and word stimulus appears. The stimulus is either congruent (left arrow and left word), or incongruent (left arrow and right word). Trials in the switching task are categorised into switch and non-switch. In a non-switch trial the cue is the same as for the immediately preceding trial. In a switch trial the cue differs from the immediately preceding trial.

Digit-Symbol Substitution Test, from the Wechsler Adult Intelligence Scale-Revised (WAIS-R (77)) -The
Digit-Symbol Substitution test explores attention and psychomotor speed. Participants are given a code table displaying digits (from 1 to 9); each digit is paired with a symbol. The participant is required to complete in blank squares with the corresponding symbol for each digit as shown in the code table.
They are given 90 seconds to fill in as many squares as possible.
Visual Short-term Memory Binding (78,79) -This test requires the participant to view one or three fractal objects, presented simultaneously in random locations on the screen. The participant is asked to remember both the objects and their location. After a delay of 1 or 4 seconds they have to make a forced choice between one of the displayed fractals (the target) and a 'dummy' fractal. Participants are required to touch the object they think has been previously presented and 'drag' it on the touch screen to its remembered, original location. The binding of such featured information has been shown to be vulnerable in asymptomatic FAD mutation carriers (79).

National Adult Reading Test (NART) -
The NART was specially designed to provide a means of estimating the premorbid intelligence levels of adults suspected of suffering from intellectual deterioration (80). The NART comprises a list of 50 words printed in order of increasing difficulty. The  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  words are relatively short in order to avoid the possible adverse effects of stimulus complexity on the reading of dementing subjects, and they are all 'irregular' with respect to the common rules of pronunciation in order to minimise the possibility of reading by phonemic decoding rather than word recognition (80). (81) -This is a circle tracing task which includes both direct and indirect visual feedback conditions. Continuous performance measures are provided including accuracy, speed, and speed of error detection and correction. The test has revealed changes in speed and accuracy in Huntington Disease mutation carriers more than 10 years before expected age-of-onset (Additional optional module).

Matrix Reasoning from the Wechsler Abbreviated Scale of Intelligence (WASI (82)) -The Matrix
Reasoning test assesses nonverbal reasoning. The participant is shown a matrix of geometric shapes with a section missing. They are required to select the option that completes the matrix (Additional optional module). (83) -Participants are given a computerised letter-search task and are required to make a rapid decision as to whether the target letter 'X' or 'N' has appeared in the search display (in either low or high load conditions). On some of the trials, a task-irrelevant distractor (a cartoon character) appears on the outside of the search display. The task evaluates the extent to which attention is captured and captivated by the distractor (Additional optional module).

Neurological clinical assessment
The neurological assessment of participants is needed for detecting any subtle neurological sign, and to establish a baseline for the absence of one or more signs. A video-recorded standard neurological examination will be included as part of the Core Module. An additional test of self-reported impairment, mainly due to movement disorders, will be administered as part of the Additional Optional Module ( Figure 1).
A standard video recording of each participant accessing the study at one of the clinics will be performed and evaluated by a clinical neurologist. The assessment includes examination of strength, coordination, balance, ocular movements, cranial nerves, gait, and repeated movements.

Blood and saliva sample collection, and bio-banking
Participants seen in the research clinics will be asked to donate a blood sample for testing for tau protein, neurofilament (Nf) levels, full blood count analysis, and to obtain data for a GWAS. Samples of serum, plasma, and DNA will be collected and processed according to a pre-specified protocol and stored in a freezer at -80 °C at the Blizard Institute (Queen Mary, University of London). All participants will also asked to donate a blood sample through a dried blood spot test and a saliva sample (in order to obtain DNA where phlebotomy is not possible).
Samples will be tested to measure the expression of NfL (Neurofilament light) , a Nf-subunit showing consistent and reproducible measurements in plasma/serum using a newly developed sensitive methodology exploiting single molecule trapping and measurement by ELISA (SIMOA) (89)(90)(91). The use of both NfL and NfH (heavy) allows for the determination of different markers whose dynamic of release into bio fluids following injury vary according to their particular chemical structure. NfL has shown good levels of linearity on dilution experiments, suggesting that confounders like aggregation or immune response may not interfere with its measurements. NfH different phosphorylation states in blood may in turn provide more information on the systemic biological changes induced by trauma. The relative low abundance of these proteins in blood, in absence of CSF, which is normally more enriched in byproducts of neuronal destruction, will not represent a major obstacle, as analytical sensitivity has evolved with the development of novel assays which cover the lower end of these protein dynamic range. SIMOA, the proposed methodology for the target biomarkers, is now emerging as the core technique for the measurement of structural proteins, down to the fempto-molar end of the spectrum.

Multimedia database set up and management
A multimedia database will be established as part of this project. The final dataset will include questionnaire and tests data (see Figure 1), in addition to the video-recorded neurological examination.
This data will be also linked to the stored blood samples through a unique identifier.
The data will be stored in mirror, anonymised datasets at LSHTM and QMUL on secure servers, and access will be regulated by the study PIs who are ultimately responsible for maintaining confidentiality and data protection. Only the researcher(s) employed on this study will have access to subject identifiable information. Identifiable information (name, date of birth, and address) will be removed prior to analysis, and subjects will be identified by pseudoanonymised codes. Personal data and key codes for pseudoanonymisation will be stored in a database with restricted access. Before conducting the statistical analyses, checks will undertaken to ensure data quality.

Statistical analysis
Descriptive statistical methods (geometric means, geometric standard deviations, and 95% Confidence Intervals (95% C.I.) will be used to summarise the rugby history and lifestyle exposure data. Analyses linking a history of head trauma with neurological symptoms (obtained by questionnaire and neurobehavioural tests) will then be performed. Parameters of physical and cognitive capability, alongside clinical neurological intermediate endpoints, will initially be dichotomised (yes/no symptoms), and analysed with prevalence odds ratios using logistic regression (92). We will also analyse continuous outcome measures using multiple linear regression. In addition to analysing individual symptoms, we will assess associations with domains of symptoms (i.e. physical and cognitive capability). For this purpose we will use cut-points previously published in the literature (93), particularly related to the 1946 Birth Cohort whose testing protocol was used. For analyses involving repeated measurements (neurobehavioural testing) assessing acute and chronic effects, we will use multilevel models. Analyses will be adjusted for age, sex, and alcohol consumption.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  Finally, we will compare neurological outcomes assessed by questionnaire with those assessed by computer-administered tests using Kappa-statistics (and in case of continuous outcomes linear regression analyses). We will also compare the strengths of the associations with exposure between both outcome measures.

Study size and power
We aim to invite 205 participants to participate, approximately half of whom have been exposed to concussion during their career (M Cross, personal communication). We conservatively estimate that about 150 former players will participate. Based on previous studies, the standard deviations of the psychometric tests are in the range of 8-15% of the absolute value; assuming a conservative figure of 15% overall, the study will have more than 95% power to detect a 10% difference, and 80% power to detect a 7% difference, in psychometric test scores, between exposed (to concussion) and non-exposed participants.

Limitations
The main limitation of this study is the cross-sectional design. Data on lifetime exposure (history of concussion) and health outcomes are collected at one time point, making data subject to recall bias.
Participants with a poorer health, or sentiment of discontent with their health status, may overestimate concussive exposure, if they attribute their poor health to it. However, this cross-sectional study design will provide meaningful data on the burden of ill health and neurological health among retired rugby players, in a relatively short two-year time frame. Depending on results (and availability of additional funding), this study could inform a, more detailed, and time intensive cohort study, aimed at assessing the association between concussion in rugby players and health outcomes prospectively.
A direct consequence of the cross-sectional design is an intrinsic risk of selection bias. Rugby players who have developed serious neurodegenerative conditions (or have died from them) in the meantime are less likely (or not able) to participate in the study thus limiting our sample to a subsample of healthier retired rugby players. On the other hand, it is possible that former players who experience some cognitive symptoms will be more keen to participate in the study, to seek reassurance and be examined in more detail.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  Irrespective of these considerations, our investigation of the association between the exposure to sporting concussion and later-life general and neurological health outcomes remains valid, despite a smaller range of health outcomes diminishing the power for detecting an association as significant.
Nonetheless, the power calculation for this study is calculated on the basis of continuous outcome measures (i.e. neurocognitive tests or grip strength measurement) making this study adequately powered to detect the potential associations, based on the expected number of recruited participants.
It is worth noting that results from this study will be not immediately apply to current rugby players, due to this cohort's former playing status, history of amateur playing exposure and alterations to rules and conditions, aimed at increasing plaer safety over recent years, which will have altered player's overall exposure to concussive events..

Ethics and dissemination
Ethical approval has been granted from the London School of Hygiene and Tropical Medicine (LSHTM) Ethics Committee. The study in general is focusing on subclinical problems (e.g. slightly lower than average scores in cognitive function tests) which would not usually require any medical attention. If any problems are identified, we will (with the permission of the study participant) refer participants to their primary care physician. The consent form includes the following statement:

I wish to inform my GP that I am taking part to this study I wish the study team to inform my GP of any unusual test result
This study will provide unique insight into the physical and neurological health of retired rugby players who are now aged 50 years or more. This is valuable information per se, potentially comparable with other occupational cohorts or the general population (accepting the potential for bias discussed previously in comparing these samples). Moreover, within this study, data will be analysed as a function of the history of traumatic brain injury and concussion during their careers, and therefore assess associations between the history of reported concussive injuries and health-related outcomes.
Importantly, the neurological examination and collection of intermediate neurological outcomes, allows the detection of subtle preclinical changes that might be associated with future risk of developing a neurodegenerative condition. Studying to what extent these changes are also associated with history of  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  A novel aspect of this project is the molecular component: clinical and neurocognitive findings will be particularly reinforced by the measurement of blood biomarkers. These will give a quantitative measurement of active neurodegeneration (NfL) and of possible CTE (tau), to be analysed conjunctly with the health-related outcomes and personal history of concussion.
Also, relying on a number characteristics of the rugby playing history collected (i.e. length of playing at elite level, position played, numbers of game played, age when started playing), it will be possible to identify potential categories of players at potentially increased risks of health-related outcomes.
This study may identify a small deviation from normality in any of the continuous measures collected via cognitive and other tests (i.e. grip strength, memory function, walking speed, etc). This will be mainly investigated in association with a history of frequent concussions, or other head traumas. The investigators will interpret the results based on the continuum of distribution of the test outcomes, which will not have specific clinical translation. However, it should be emphasised that although this study is unlikely to identify patterns suggesting a clinical disease in any of the participants, if these are identified, then the relevant participant(s) will be referred to appropriate medical services.

Competing interest statement
None