You are here

  • Home
  • Online First
  • Sports-related concussion not associated with long-term cognitive or behavioural deficits: the PROTECT-TBI study

Article Text

Article menu
Cognition
Original research
Sports-related concussion not associated with long-term cognitive or behavioural deficits: the PROTECT-TBI study
  1. Matthew Joseph Lennon1,2,
  2. Grant Rigney3,
  3. Byron Creese4,5,
  4. Dag Aarsland6,7,
  5. Adam Hampshire8,
  6. Clive Ballard5,
  7. Anne Corbett5,
  8. Vanessa Raymont2
  1. 1 Faculty of Medicine and Health, University of New South Wales, Sydney, New South Wales, Australia
  2. 2 Department of Psychiatry, University of Oxford, Oxford, UK
  3. 3 Harvard Medical School, Harvard University, Cambridge, Massachusetts, USA
  4. 4 Division of Psychology, Department of Life Sciences, Brunel University London, Uxbridge, Greater London, UK
  5. 5 Department of Health and Community Sciences, College of Medicine and Health, University of Exeter, Exeter, Devon, UK
  6. 6 Department of Old age Psychiatry, IoPPN, King's College London, London, UK
  7. 7 Centre for Age-related research, Stavanger University Hospital, Stavanger, Norway
  8. 8 Department of Brain Sciences, Faculty of Medicine, Imperial College London, London, UK
  1. Correspondence to Dr Vanessa Raymont; Vanessa.raymont{at}psych.ox.ac.uk

Abstract

Background The cognitive effects of sports-related concussion (SRC) have been the subject of vigorous debate but there has been little research into long-term outcomes in non-athlete populations.

Methods This cohort study of UK community-dwelling adults (aged 50–90 years) was conducted between November 2015 and November 2020, with up to 4 years annual follow-up (n=15 214). Lifetime history of concussions was collected at baseline using the Brain Injury Screening Questionnaire. The first analysis grouped participants by type of concussion (no concussion, only SRC, only non-SRC (nSRC), mixed concussions (both SRC and nSRC)) and the second grouped the participants by number (0, 1, 2 or 3+ SRC or nSRC). Mixed models were used to assess the effect of concussion on outcomes including four cognitive domains and one behavioural measure (Mild Behavioural Impairment-C).

Results Analysis of the included participants (24% male, mean age=64) at baseline found that the SRC group had significantly better working memory (B=0.113, 95% CI 0.038, 0.188) and verbal reasoning (B=0.199, 95% CI 0.092, 0.306) compared with those without concussion. Those who had suffered one SRC had significantly better verbal reasoning (B=0.111, 95% CI 0.031, 0.19) and attention (B=0.115, 95% CI 0.028, 0.203) compared with those with no SRC at baseline. Those with 3+ nSRCs had significantly worse processing speed (B=−0.082, 95% CI −0.144 to –0.019) and attention (B=−0.156, 95% CI −0.248 to –0.063). Those with 3+ nSRCs had a significantly worse trajectory of verbal reasoning with increasing age (B=−0.088, 95% CI −0.149 to –0.026).

Conclusions Compared with those reporting no previous concussions, those with SRC had no cognitive or behavioural deficits and seemed to perform better in some tasks. As indicated by previous studies, sports participation may confer long-term cognitive benefits.

  • COGNITION
  • CONCUSSION
  • HEAD INJURY
  • HEALTH ECONOMICS
  • NEUROPSYCHIATRY

Data availability statement

Data are available upon reasonable request. Applications for the PROTECT data can be made through the following website: https://www.exeter.ac.uk/research/dementia-research/research/protect/.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/jnnp-2024-334039

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Recurrent sports-related concussion (SRC) in professional athletes is associated with significantly greater risk of mild cognitive impairment (MCI) and dementia, but long-term cognitive outcomes after SRC in non-professional athletes are not known.

    WHAT THIS STUDY ADDS

    • This UK-based community-based longitudinal cohort study (n=15 214, age range=50–90 years) showed that those with SRC showed no long-term cognitive or behavioural deficits compared with those with no concussions.

    • In fact, they showed better performance in working memory and verbal reasoning at the study baseline.

    • By contrast, those with non-SRC showed deficits in processing speed, attention and the Mild Behavioural Impairment (MBI-Checklist) index.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • This study suggests that the cognitive risks of concussion in sport may not be meaningful in the long term in the non-professional athlete population.

    • These results will help inform physicians and public health authorities when communicating the risks and benefits of community sports to patients and the public.

    Introduction

    Approximately 2% of the UK population present to emergency annually with a head injury1 and it is the leading cause of death in those under 40 years of age. Traumatic brain injury (TBI) can increase dementia risk by 1.5–3 times and estimates indicate that TBI contributes 5–15% of the current dementia burden.2

    TBIs vary in classification from ‘mild’ (a temporary change in mental status or loss of consciousness (LOC) of less than 30 min) to ‘moderate-severe’ (prolonged amnesia or LOC more than 30 min).3 The most frequent causes of mild TBI are assaults, falls and road traffic collisions,1 but among children and adolescents, sports-related mild TBI is the second most common cause, affecting 0.3–0.4% of adolescents annually.4 This type of mild TBI is most commonly referred to as sports-related concussion (SRC) in the literature. The impact of SRC on long-term cognitive outcomes and dementia has been the subject vigorous public debate, having been highlighted by many high-profile cases of professional athletes.

    Among athletes, there seems to be a relationship between repeated SRC and poor cognitive outcomes. In a meta-analysis including 21 studies of athletes (n=790 concussion cases, 2014 controls), ranging from high school to professional, there were substantial deficits in the first few days of injury that all essentially resolved by 7–10 days of recovery.5 In a more recent meta-analysis of 11 studies (n=792), Zhang et al 6 compared retired elite athletes who had suffered SRC years earlier with those who had not suffered concussion. The concussed group demonstrated mild to moderate deficits in verbal memory (Standardized Mean Difference (SMD)=−0.29), delayed recall (SMD=−0.30) and attention (SMD=−0.33). In their seminal study of mid-life retired American football players (n=2552), Guskiewicz et al 7 found that those with recurrent concussion had five times the rate of clinically diagnosed MCI and three times the rate of subjective memory impairment compared with non-concussed retired players. Based on these data, advocates argue that recreational contact sport, particularly in young people, is too risky and should be discouraged or banned. Yet, there is very little research on long-term outcomes of SRC in the non-collegiate or professional athlete population. Professional athletes are disproportionately male8 and comprise only a tiny fraction of the population. They are exposed to a considerably greater number, frequency and severity of concussions, as well as more repetitive subconcussive head impacts compared with the non-athlete population.5 Given the ubiquity of concussions in community sport, it is critical that the long-term cognitive outcomes of these injuries are understood in a community-dwelling population rather than just in professional athletes.

    The extant research focuses largely on cognitive outcomes rather than behavioural changes. Mild behavioural impairment (MBI) is a well-established description9 of late-life onset, sustained behavioural and personality changes that are associated with biomarkers of neurodegenerative disease, worsening cognitive impairment and dementia.9 One recent study by Bray et al 10 examined 124 participants with a self-reported history of concussion, as they progressed from normal cognition to a dementia diagnosis. They found that concussion was significantly associated with a dementia prodrome of greater social inappropriateness compared with controls. Thus, MBI may be a useful adjunct to cognitive measures in a comprehensive assessment of concussion-induced deficits.

    This study examines the associations between SRC, non-sports-related concussion (nSRC) (ie, concussions in contexts other than active sports) and long-term cognitive and behavioural outcomes in a longitudinal cohort of community-dwelling adults. Specifically, it examines whether there are different profiles of cognitive and behavioural deficits for SRC and nSRCs.

    Methods

    Participant population

    The PROTECT study (www.protectstudy.org.uk) is a UK-based longitudinal study of 50–90 year olds.11 For eligibility, participants were required to have access to a computer and all those with previously diagnosed dementia at baseline were excluded. All participants gave informed consent prior to involvement and were assessed at baseline (wave 1) and then had up to 4 years of annual assessments (waves 2–5) between November 2015 and November 2020. A full description of the study can be found in prior publications.11 This study was reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (online supplemental table S1).

    Supplemental material

    Classification of concussion

    The Brain Injury Screening Questionnaire (BISQ)12 was used to collect data on previous concussions/TBIs. It was an optional, self-administered battery within the PROTECT study, and of eligible participants with sufficient data in the PROTECT study, 4379 individuals did not complete the BISQ and 15 764 did. The baseline characteristics of responders and non-responders are compared in online supplemental table S2 to assess for self-selection bias. This questionnaire collects data on lifetime history of head injuries asking specifically, ‘Have you ever had a blow to the head … [in a particular context for example, on a motorcycle/all-terrain vehicle]’. Participants are asked about the context of the injury (sports related, motor vehicle crash, etc), the age of the first/last TBI, the severity of each episode (length of time unconscious/dazed or confused) and the number of injuries.

    Similar to a previous publication from this cohort,13 head injuries were classified using the Mayo TBI Severity Classification System.3 Concussion (mild or symptomatic TBI) was defined as a head injury followed by LOC of less than 30 min or a dazed or confused episode. Moderate-severe TBI was defined as a head injury followed by a LOC of 30 min or longer. Each concussion was classified as either sports related (occurring while biking or playing sports)14 or non-sports related (occurring from another cause) (nSRC). For the head injury questions comprising the BISQ, see online supplemental table S3. As there were only a small number of moderate-severe TBI acquired during sports, it was not possible to analyse them separately, and thus for all analyses, any participants with moderate-severe TBI were removed (total moderate-severe TBI n=510, sports-related moderate-severe TBI n=53), only assessing those with concussion (mild or symptomatic TBI). Online supplemental table S4 compares the characteristics of those with and without moderate-severe TBI .

    In the first analysis, individuals were grouped based on the context in which they had suffered concussion and each of these groups were compared with those who had suffered no concussion in any category:

    1. SRC group—Participants who reported concussion only in the context of sports.

    2. nSRC group—Participants who reported concussion in a context other than sports.

    3. Mixed concussion group—Participants who reported concussion in the context of sports and some other context.

    In the second analysis, we used two variables reflecting the numbers of reported SRC and nSRC, respectively. These two variables were assessed both as categorical variables (groups 0, 1, 2 and 3+) and continuous variables.

    Calculation of cognitive scores

    A full description of the PROTECT study cognitive test batteries can be found in the online supplemental methods. In brief, at each wave, participants were instructed to complete each cognitive test three times at least 12 hours apart within the space of a week. The average score of the repeats was taken as the test score for that wave. Naturally, not all participants completed three repeats and in those who did there were learning effects (ie, improving performance with repetition). Therefore, in all our analyses, the number of test repeats within each wave was included as a covariate.

    We used an orthogonal rotated principal components analysis (PCA) to develop cognitive domain scores using the baseline measures of 9 cognitive outcomes. Four outcomes were taken from the PROTECT Cognitive Test Battery (digit span, paired associates learning, verbal reasoning, self-ordered search) and five were measures taken from the COGTRACK assessment battery (attentional intensity index, sustained attention index, attentional fluctuation index, cognitive reaction time, memory retrieval speed) (see online supplemental methods for details). Each test score was z-transformed and winsorised to between +5 and −5 SD from the mean. The following domain scores were calculated by taking the mean of the z-transformed cognitive tests grouped by the PCA:

    1. Working memory—Digit span, paired associates learning and self-ordered search.

    2. Verbal reasoning—Baddeley’s Grammatical Reasoning Test.

    3. Processing speed—Attentional intensity index, cognitive reaction time and memory retrieval speed.

    4. Attention—Sustained attention index and attentional fluctuation index.

    Normality of the domain scores distribution was examined visually and numerically. If skewness was greater than 1 or less than −1, the score was transformed. The attention domain score was negatively skewed and thus was inverted, log-transformed and re-standardised to achieve a normal distribution.

    Calculation of MBI score

    The Mild Behavioural Impairment-Checklist (MBI-C) questionnaire is a validated 34-item list of yes/no questions grouped into five domains (decreased motivation, emotional dysregulation, impulse dyscontrol, social inappropriateness, and abnormal perception or thought content).15 If rated ‘yes’, participants are asked to rate severity between 1 and 3. Our main outcome was the total MBI-C score (ie, the sum of all the MBI-C items, ranging between 0 and 102). The participant MBI-C score had a count data distribution (ie, non-normal clustering at zero with a substantial positive skew). It could not be adequately transformed, was left as a raw score and non-linear methods (negative binomial) were used in its analysis.

    Statistical analysis

    The analysis plan was prespecified on Open Science Framework (osf.io/nsf4b/). Baseline characteristics of each group were compared with the ‘no concussion’ group using Tukey’s Honestly Significant Difference (HSD) and pairwise χ2 analysis. Model construction was developed using fitting parameters Akaike information criteria and Bayesian information criteria. Given that rates of cognitive decline change with age, rather than using a simple time variable, the study used a grand mean-centred ‘age at each wave’ as the ‘time’ variable. Further, given that cognitive decline with age is non-linear, an age2 (ie, grand mean-centred age squared) term was also included. There was considerable missing baseline covariate data particularly for physical activity and vascular risk factors (online supplemental table S5). The challenge of missing data was managed by running both partially adjusted models, including all participants, and fully adjusted models, including those with complete data, and comparing results. Partially adjusted models controlled for age, age2, sex, education status and number of repeats in the wave and included an interaction between age and the either concussion type or concussion number. Fully adjusted models were run, controlling additionally for smoking, hypertension, stroke, coronary heart disease, diabetes, high cholesterol, history of anxiety, history of mood disorders, history of psychotic disorders, socioeconomic status and current physical activity. For details of the covariate classification, see online supplemental methods. The models specified a random intercept and slope (time varying age variable), while the other terms were treated as fixed effects. Fully adjusted models are the focus of this paper, with partially adjusted model results included in the online supplemental tables S6–S8 and discussed when discrepancies arise between fully and partially adjusted models.

    Between-group differences in covariates were assessed using analysis of variance for continuous variables and χ analysis for categorical variables (see table 1). Linear mixed models were used to assess the effect of concussion type and number on cognitive scores at both baseline and on score trajectories over time. As the behavioural outcome (MBI-C) was non-normal count data, negative binomial mixed models were used to assess the effect of concussion at baseline and with increasing age. The first analysis compared the three aforementioned concussion groups (SRC, nSRC and mixed concussion) to those reporting no concussion. The second analysis assessed the effects of numbers of both lifetime SRC and nSRC numbers both as categorical (groups 1, 2, 3+, comparison group 0) and continuous variables. A number of sensitivity analyses were undertaken. First, additional models were run including a sex interaction term for the main effect and effect over time. Second, to address concerns that the concussions were either too recent to be considered chronic or too remote to be recalled correctly, we ran an analysis restricted to those who had their last concussion more than 3 months ago but less than 20 years ago.

    View this table:
    Table 1

    Summary of study population characteristics in comparing those with no concussion, sports-related concussion, non-sports-related concussion and mixed concussion

    To account for multiple comparisons (five outcomes assessed), a Sidak-corrected p value significance threshold of 0.01 was used. Statistical analyses were performed using R (V.4.3.1) using the ‘lme4’ and ‘NBZIMM’ packages.

    Results

    Participant characteristics

    Overall, there were 15 214 participants between the ages of 50 and 90 at baseline (mean age 62.7 (SD=7.2) (table 1). Of these participants, 24% were male. The average time in the study was 2.8 years (SD 1.5 years). For those who had suffered a concussion, the mean age at their first concussion was 25.5 years (SD 19.9) and the mean time since the last concussion was 29.6 years (SD 20). Those in the SRC group (57.3% male) and mixed concussion group (58% male) were substantially more likely (p<0.001) to be male than those in the no concussion group (19.2% male). Those in the SRC and mixed concussion groups had significantly higher levels of education (p<0.001 and p=0.003) compared with those with no concussion. There was a considerably larger portion of those in the SRC group (p<0.001) reporting highest level of household income and physical activity (table 1).

    Cognitive and behavioural outcomes by concussion groups

    At baseline, participants in the SRC group had significantly better working memory (B=0.113, 95%CI 0.038, 0.188, p=0.003) and verbal reasoning (B=0.199, 95% CI 0.092, 0.306, p<0.001) compared with those who had never suffered concussion in the fully adjusted model (figure 1 and table 2). This analysis controlled for education, socioeconomic status and physical activity, among other covariates. Both the non-sports-related concussion (estimate=−0.299, 95% CI−0.386 to –0.212, p<0.001) and mixed concussion groups (estimate=−0.316, 95% CI −0.487 to –0.144, p<0.001) had significantly worse MBI-C scores compared with the no concussion group in the fully adjusted model. There were no significant differences in the various groups score trajectories with increasing age.

    Figure 1
    Figure 1

    Results from baseline analysis of cognitive and behavioural outcomes. (A) Separating groups by concussion subgroup (sports, non-sports, mixed). The comparison group (dotted line) are those in the no concussion group. (B) Combined analysis of the effect of having 1, 2 or 3+ sports-related concussions or non-sports-related concussions. The comparison group (dotted line) are those with 0 sports-related concussions or non-sports-related concussions. MC, mixed concussion; nSRC, non-sports-related concussion; SRC, sports-related concussion.

    View this table:
    Table 2

    Summary of linear and negative binomial mixed model results examining effect of TBI subtype on cognitive/behavioural scores

    Cognitive/behavioural outcomes and the numbers of concussions

    Those who had had suffered one SRC had significantly better verbal reasoning (B=0.111, 95% CI 0.031, 0.19, p=0.006) and attention (B=0.115, 95% CI 0.028, 0.203, p=0.010) compared with those with no SRC at baseline (figure 1 and table 3). Those with 3+ nSRCs had significantly worse processing speed (B=−0.082, 95% CI −0.144 to –0.019, p=0.010) and attention (B=−0.156, 95% CI −0.248 to –0.063, p=0.001) compared with those with no nSRC. Additionally, in the partially adjusted analysis, those with 3+ nSRC had significantly worse working memory (B=−0.061, 95% CI −0.103 to –0.019, p=0.005), although this was not significant in the fully adjusted analysis. Those with 1 (B=−0.136, 95% CI −0.237 to –0.034, p=0.009), 2 (B=−0.322, 95% CI −0.454 to –0.189, p<0.001) or 3+ (B=−0.558, 95% CI −0.709 to –0.406, p<0.001) nSRCs had significantly worse MBI-C scores when compared with those with no nSRC. Longitudinally, those with those with 3+ nSRCs had a significantly worse trajectory of verbal reasoning with increasing age (B=−0.088, 95% CI −0.149 to –0.026, p=0.005) compared with those without nSRC. Assessing the numbers of concussion as a continuous measure, each additional nSRC was associated with progressively worse attention (B=−0.034, 95% CI −0.051 to –0.016, p<0.001) (table 4). In the partially adjusted model, nSRCs were associated with deficits in processing speed (B=−0.014, 95% CI−0.023 to –0.006, p=0.001), although this was not significant in the fully adjusted model.

    View this table:
    Table 3

    Summary of linear and negative binomial mixed model results examining the effect of numbers of both sports-related concussion and non-sports-related concussion (as categorical variables) on cognitive/behavioural scores

    View this table:
    Table 4

    Summary of linear and negative binomial mixed model results examining the effect of numbers of both sports-related concussion and non-sports-related concussion (as continuous variables) on cognitive/behavioural scores

    Sensitivity analyses

    Females who had 3+ SRC had a significantly worse trajectory of processing speed over time (B=−0.442, 95% CI −0.74 to –0.145, p=0.004) but otherwise there was no significant interactions between sex and any of the concussion type or number variables at baseline or longitudinally (online supplemental tables S9–S11). When restricting to those with their last concussion >3 months and <20 years ago, the results were consistent, except that those in the SRC group and those with 1 SRC did not perform significantly better on any of the cognitive measures (online supplemental tables S12–S14).

    Discussion

    This study illustrated that individuals with SRC manifested no long-term cognitive or behavioural deficits compared with those without concussion. Indeed, those who had suffered SRC had better working memory and verbal reasoning. However, this effect seemed limited to those with a single SRC. Those with two or more SRCs did not perform better on any of the cognitive or behavioural measures. Given our understanding of the pathophysiology of concussion,16 it is clear that there is something other than the head injury itself that underlies the better cognitive outcomes in this group. It has been well demonstrated in the literature that the risks of mid-to-late-life cognitive deficits are modified by physical activity,17 education,18 income,19 cardiac health20 and smoking.21 The SRC group had significantly better health outcomes at baseline for each of those covariates, but when controlling for these covariates, the significant differences remained, suggesting that there are unaccounted explanatory factors. One possibility is that lifetime physical activity22 has a cumulative, greater positive impact on cognition that is not adequately captured by controlling for current physical activity, as in our model. Alternatively, involvement in sport may be associated with greater lifetime social connectivity, which is also known to be associated with lower rates of cognitive decline and dementia.23

    Consideration of putative explanatory factors behind this difference still leaves unanswered why this study’s findings are at odds with much of the SRC literature. In a systematic review, including 46 studies and 13 975 participants, Cunningham et al 24 examined cognitive outcomes for retired athletes. They found that retired athletes with a history of SRC had worsened outcomes in 17 of 31 (55%) of studies examining memory and 6 of 11 (55%) of studies examining executive function. They also found that 28% of studies reported a dose-response relationship, suggestive of a causative link. Our study is distinctive in the SRC literature in several ways that may explain these differences. This study examines a community-dwelling sample rather than professional athletes, and thus, the head injuries are likely less frequent, numerous and severe.

    This study examines mid-to-late-life individuals who often have experienced SRC years ago, whereas most other studies of SRC focus on younger athletes in the immediate period after their head injuries when cognitive effects are likely more salient. Our study uses a behavioural measure, the MBI-C,9 which is known to predict cognitive impairment25–27 and dementia,28–32 and this multipronged approach corroborates the finding that SRC is not associated with poorer long-term outcomes in this population. Interestingly, Deshpande et al 33 (n=3904 men, mean age=64.4) published a large study of community-dwelling individuals who played non-professional high school American Football. They found that previous footballers had no cognitive deficits and better depression scores compared with controls. Taken together, while SRC in professional athletes seems to be associated with cognitive deficits, in the general population, there are no cognitive or behavioural deficits associated with SRC.

    By contrast, nSRCs were associated with worsened MBI-C scores in a dose-dependent manner and those with 3+ nSRCs had significantly worse processing speed and attention. Whereas most of the literature examining repeated concussions focuses on professional sports-related injuries, our study demonstrates that the dose-response effect is seen in non-sports-related contexts. In our previous paper,13 we similarly found that there was a dose-response relationship between cognitive outcomes and repeated TBI. This current study suggests that the nSRC may be the more important driver in this relationship. The mechanism of injury results in differential in velocity, intensity and rotational forces,34 which may underlie the discrepancies between nSRC and SRC outcomes. It is also likely that the sport-related physical, social and economic benefits that may offset the cognitive risks of SRC are not present in the same way for concussions associated with falls, assaults and motor vehicle accidents. Interestingly, this study showed that those with 3+ nSRCs had a worsened decline in verbal reasoning with increasing age. The effect size was small and there have not been similar findings for long-term cognitive decline within this study or in other studies,32 and thus this result should be interpreted with caution. In our sensitivity analysis, we found that females with more than three SRC had a substantially worse decline of processing speed compared with males. This may be a chance finding as no similar results were found for either the baseline analyses or for any of the other outcomes. If corroborated in future studies, this result may indicate that females are at greater risk of long-term cognitive decline from SRC. Both the aforementioned results raise the interesting possibility that even years after an injury, a history of multiple concussions may contribute to accelerated cognitive decline in some domains.

    Limitations

    The critical limitations of this study include its retrospective design, limitations in cognitive domains, participant dropout and unmeasured confounders. The retrospective design, with participants frequently recalling events several decades ago, may have resulted in an under-reporting of concussions and an underestimation of the effect size, particularly given that concussion is linked to memory loss. The study design may have also been affected by selection bias as males and those with cognitive deficits are less likely to participate. Because the retrospective design relies on memory to report TBI, any examination of the relationship between current memory function and concussion is confounded. As such, our study does not explore a critical cognitive domain that is known to be affected by concussion.16 Over the course of the study, participant retention was 45.3%, which is comparable to other longitudinal studies of ageing but naturally risks confounding by survivor bias. This issue was mitigated using the linear mixed model design. Finally, as mentioned previously, unmeasured confounders such as social connectedness and lifetime physical activity may explain some of the results in this study but were not included.

    Conclusion

    To conclude, this study has found that those with SRC show no long-term cognitive or behavioural deficits compared with those without concussion. By contrast, those with nSRC showed deficits in processing speed, attention and the MBI-C, as well as a worsened rate of decline in verbal reasoning. Understanding the benefits of sport relative to the long-term risks of the injuries should inform public discourse around community level sport participation.

    Data availability statement

    Data are available upon reasonable request. Applications for the PROTECT data can be made through the following website: https://www.exeter.ac.uk/research/dementia-research/research/protect/.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Ethics approval was granted by the UK London Bridge National Research Ethics Committee (Ref. 13/LO/1578). Participants gave informed consent to participate in the study before taking part.

    Acknowledgments

    We would like to acknowledge all the staff at the University of Exeter College of Medicine and Health and Kings College London involved in the administration of the PROTECT study as well as the PROTECT participants themselves.

    References

    1. ↵
      2. Lawrence T ,
      3. Helmy A ,
      4. Bouamra O , et al
      . Traumatic brain injury in England and Wales: prospective audit of epidemiology, complications and standardised mortality. BMJ Open 2016;6:e012197. doi:10.1136/bmjopen-2016-012197
    2. ↵
      2. Graham NSN ,
      3. Sharp DJ
      . Understanding neurodegeneration after traumatic brain injury: from mechanisms to clinical trials in dementia. J Neurol Neurosurg Psychiatry 2019;90:1221–33. doi:10.1136/jnnp-2017-317557
      OpenUrlAbstract/FREE Full Text
    3. ↵
      2. Malec JF ,
      3. Brown AW ,
      4. Leibson CL , et al
      . The mayo classification system for traumatic brain injury severity. J Neurotrauma 2007;24:1417–24. doi:10.1089/neu.2006.0245
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      2. Trefan L ,
      3. Houston R ,
      4. Pearson G , et al
      . Epidemiology of children with head injury: a national overview. Arch Dis Child 2016;101:527–32. doi:10.1136/archdischild-2015-308424
      OpenUrlAbstract/FREE Full Text
    5. ↵
      2. Belanger HG ,
      3. Vanderploeg RD
      . The neuropsychological impact of sports-related concussion: a meta-analysis. J Int Neuropsychol Soc 2005;11:345–57. doi:10.1017/s1355617705050411
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      2. Zhang Y ,
      3. Ma Y ,
      4. Chen S , et al
      . Long-Term Cognitive Performance of Retired Athletes with Sport-Related Concussion: A Systematic Review and Meta-Analysis. Brain Sci 2019;9:199. doi:10.3390/brainsci9080199
    7. ↵
      2. Guskiewicz KM ,
      3. Marshall SW ,
      4. Bailes J , et al
      . Association between recurrent concussion and late-life cognitive impairment in retired professional football players. Neurosurgery 2005;57:719–26; . doi:10.1093/neurosurgery/57.4.719
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      2. Rasmussen K ,
      3. Dufur MJ ,
      4. Cope MR , et al
      . Gender Marginalization in Sports Participation through Advertising: The Case of Nike. Int J Environ Res Public Health 2021;18:7759. doi:10.3390/ijerph18157759
    9. ↵
      2. Ismail Z ,
      3. Smith EE ,
      4. Geda Y , et al
      . Neuropsychiatric symptoms as early manifestations of emergent dementia: Provisional diagnostic criteria for mild behavioral impairment. Alzheimers Dement 2016;12:195–202. doi:10.1016/j.jalz.2015.05.017
      OpenUrlCrossRefPubMed
    10. ↵
      2. Bray MJC ,
      3. Bryant BR ,
      4. Esagoff AI , et al
      . Effect of traumatic brain injury on mild behavioral impairment domains prior to all-cause dementia diagnosis and throughout disease progression. Alzheimers Dement (N Y) 2022;8:e12364. doi:10.1002/trc2.12364
    11. ↵
      2. Corbett A ,
      3. Williams G ,
      4. Creese B , et al
      . Cognitive decline in older adults in the UK during and after the COVID-19 pandemic: a longitudinal analysis of PROTECT study data. Lancet Healthy Longev 2023;4:e591–9. doi:10.1016/S2666-7568(23)00187-3
      OpenUrlCrossRefPubMed
    12. ↵
      2. Dams-OʼConnor K ,
      3. Cantor JB ,
      4. Brown M , et al
      . Screening for traumatic brain injury: findings and public health implications. J Head Trauma Rehabil 2014;29:479–89. doi:10.1097/HTR.0000000000000099
      OpenUrlCrossRef
    13. ↵
      2. Lennon MJ ,
      3. Brooker H ,
      4. Creese B , et al
      . Lifetime Traumatic Brain Injury and Cognitive Domain Deficits in Late Life: The PROTECT-TBI Cohort Study. J Neurotrauma 2023;40:1423–35. doi:10.1089/neu.2022.0360
      OpenUrlCrossRef
    14. ↵
      2. McCrory P ,
      3. Feddermann-Demont N ,
      4. Dvořák J , et al
      . What is the definition of sports-related concussion: a systematic review. Br J Sports Med 2017;51:877–87. doi:10.1136/bjsports-2016-097393
      OpenUrlAbstract/FREE Full Text
    15. ↵
      2. Ismail Z ,
      3. Agüera-Ortiz L ,
      4. Brodaty H , et al
      . The Mild Behavioral Impairment Checklist(MBI-C): A Rating Scale forNeuropsychiatric Symptoms inPre-Dementia Populations. J Alzheimers Dis 2017;56:929–38. doi:10.3233/JAD-160979
      OpenUrl
    16. ↵
      2. Bryant AM ,
      3. Rose NB ,
      4. Temkin NR , et al
      . Profiles of Cognitive Functioning at 6 Months After Traumatic Brain Injury Among Patients in Level I Trauma Centers: A TRACK-TBI Study. JAMA Netw Open 2023;6:e2349118. doi:10.1001/jamanetworkopen.2023.49118
    17. ↵
      2. Guure CB ,
      3. Ibrahim NA ,
      4. Adam MB , et al
      . Impact of Physical Activity on Cognitive Decline, Dementia, and Its Subtypes: Meta-Analysis of Prospective Studies. Biomed Res Int 2017;2017:9016924. doi:10.1155/2017/9016924
    18. ↵
      2. Clouston SAP ,
      3. Smith DM ,
      4. Mukherjee S , et al
      . Education and Cognitive Decline: An Integrative Analysis of Global Longitudinal Studies of Cognitive Aging. J Gerontol B Psychol Sci Soc Sci 2020;75:e151–60. doi:10.1093/geronb/gbz053
      OpenUrlPubMed
    19. ↵
      2. Brown MJ ,
      3. Hill NL ,
      4. Haider MR
      . Age and gender disparities in depression and subjective cognitive decline-related outcomes. Aging Ment Health 2022;26:48–55. doi:10.1080/13607863.2020.1861214
      OpenUrl
    20. ↵
      2. Ungvari Z ,
      3. Toth P ,
      4. Tarantini S , et al
      . Hypertension-induced cognitive impairment: from pathophysiology to public health. Nat Rev Nephrol 2021;17:639–54. doi:10.1038/s41581-021-00430-6
      OpenUrl
    21. ↵
      2. Deal JA ,
      3. Power MC ,
      4. Palta P , et al
      . Relationship of Cigarette Smoking and Time of Quitting with Incident Dementia and Cognitive Decline. J Am Geriatr Soc 2020;68:337–45. doi:10.1111/jgs.16228
      OpenUrlCrossRef
    22. ↵
      2. Reas ET ,
      3. Laughlin GA ,
      4. Bergstrom J , et al
      . Lifetime physical activity and late-life cognitive function: the Rancho Bernardo study. Age Ageing 2019;48:241–6. doi:10.1093/ageing/afy188
      OpenUrl
    23. ↵
      2. Samtani S ,
      3. Mahalingam G ,
      4. Lam BCP , et al
      . Associations between social connections and cognition: a global collaborative individual participant data meta-analysis. Lancet Healthy Longev 2022;3:e740–53. doi:10.1016/S2666-7568(22)00199-4
      OpenUrl
    24. ↵
      2. Cunningham J ,
      3. Broglio SP ,
      4. O’Grady M , et al
      . History of Sport-Related Concussion and Long-Term Clinical Cognitive Health Outcomes in Retired Athletes: A Systematic Review. J Athl Train 2020;55:132–58. doi:10.4085/1062-6050-297-18
      OpenUrlCrossRefPubMed
    25. ↵
      2. Creese B ,
      3. Brooker H ,
      4. Ismail Z , et al
      . Mild Behavioral Impairment as a Marker of Cognitive Decline in Cognitively Normal Older Adults. Am J Geriatr Psychiatry 2019;27:823–34. doi:10.1016/j.jagp.2019.01.215
      OpenUrlCrossRef
    26. ↵
      2. Feldman H ,
      3. Scheltens P ,
      4. Scarpini E , et al
      . Behavioral symptoms in mild cognitive impairment. Neurology (ECronicon) 2004;62:1199–201. doi:10.1212/01.wnl.0000118301.92105.ee
      OpenUrl
    27. ↵
      2. Ismail Z ,
      3. Elbayoumi H ,
      4. Fischer CE , et al
      . Prevalence of Depression in Patients With Mild Cognitive Impairment: A Systematic Review and Meta-analysis. JAMA Psychiatry 2017;74:58–67. doi:10.1001/jamapsychiatry.2016.3162
      OpenUrl
    28. ↵
      2. Geda YE ,
      3. Roberts RO ,
      4. Mielke MM , et al
      . Baseline neuropsychiatric symptoms and the risk of incident mild cognitive impairment: a population-based study. Am J Psychiatry 2014;171:572–81. doi:10.1176/appi.ajp.2014.13060821
      OpenUrlCrossRefPubMed
    29. ↵
      2. Kørner A ,
      3. Lopez AG ,
      4. Lauritzen L , et al
      . Acute and transient psychosis in old age and the subsequent risk of dementia: a nationwide register-based study. Geriatr Gerontol Int 2009;9:62–8. doi:10.1111/j.1447-0594.2009.00505.x
      OpenUrlPubMed
    30. ↵
      2. Donovan NJ ,
      3. Amariglio RE ,
      4. Zoller AS , et al
      . Subjective cognitive concerns and neuropsychiatric predictors of progression to the early clinical stages of Alzheimer disease. Am J Geriatr Psychiatry 2014;22:1642–51. doi:10.1016/j.jagp.2014.02.007
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      2. Banks SJ ,
      3. Raman R ,
      4. He F , et al
      . The Alzheimer’s disease cooperative study prevention instrument project: longitudinal outcome of behavioral measures as predictors of cognitive decline. Dement Geriatr Cogn Dis Extra 2014;4:509–16. doi:10.1159/000357775
      OpenUrl
    32. ↵
      2. Taragano FE ,
      3. Allegri RF ,
      4. Krupitzki H , et al
      . Mild behavioral impairment and risk of dementia: a prospective cohort study of 358 patients. J Clin Psychiatry 2009;70:584–92. doi:10.4088/jcp.08m04181
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      2. Deshpande SK ,
      3. Hasegawa RB ,
      4. Rabinowitz AR , et al
      . Association of Playing High School Football With Cognition and Mental Health Later in Life. JAMA Neurol 2017;74:909–18. doi:10.1001/jamaneurol.2017.1317
      OpenUrl
    34. ↵
      2. Rowson S ,
      3. Duma SM ,
      4. Beckwith JG , et al
      . Rotational head kinematics in football impacts: an injury risk function for concussion. Ann Biomed Eng 2012;40:1–13. doi:10.1007/s10439-011-0392-4
      OpenUrlCrossRefPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Contributors MJL contributed to methodology, formal analysis, visualisation, writing the original draft, and review and editing drafts. MJL is the guarantor of the study. GR contributed to methodology, writing the original draft and reviewing and editing drafts. BC contributed to conceptualisation, data curation, methodology, project administration, software, supervision, validation and review and editing drafts. DA contributed to funding acquisition, conceptualisation, data curation, methodology, project administration, resources, and review and editing drafts. AH contributed to funding acquisition, conceptualisation, data curation, assessment software development, methodology, project administration, resources, and review and editing drafts. CB contributed to funding acquisition, conceptualisation, data curation, methodology, project administration, resources, and review and editing drafts. AC contributed to funding acquisition, conceptualisation, data curation, methodology, project administration, resources, and review and editing drafts. VR contributed to methodology, project administration, resources, supervision and review and editing drafts.

    • Funding This paper represents an independent research part funded by the National Institute for Health Research (NIHR) Exeter Biomedical Research Centre (Grant ID NIHR203320) and Maudsley Biomedical Research Centre (Grant ID NIHR203318). This research was also supported by the NIHR Collaboration for Leadership in Applied Health Research and Care South West Peninsula.

    • Competing interests CB collected consulting fees from the following companies: Acadia, AARP, Addex, Biohaven, Eli Lily and Company, Enterin Inc, Exciva, H.Lundbeck A.S, Janssen Pharmaceuticals, Novo Nordisk, Orion Corp., Otsuka America Pharm Inc, Sunovion Pharm Inc, Suven, Roche, Axosome and Biogen. CB is on Advisory Boards for the following companies: Acadia, Roche, Novo-Nordisk, AARP, Biogen and Synexus. CB received an Honorarium from Harvard University for speaking. AH is the owner and director of Future Cognition Ltd, a software development company the consulted on the development of the cognitive assessment software.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.