Article Text
Abstract
Objectives To examine the relationships among self-reported sport-related concussion (SRC) history and current health-promoting behaviours (exercise frequency, diet quality and sleep duration) with self-reported measures of brain health (cognitive function, symptoms of depression and anxiety and emotional–behavioural dyscontrol) in former NFL players.
Methods In this cross-sectional study, a questionnaire was sent to former NFL players. Respondents reported SRC history (categorical: 0; 1–2; 3–5; 6–9; 10+ concussions), number of moderate-to-vigorous aerobic and resistance exercise sessions per week, diet quality (Rapid Eating Assessment for Participants—Shortened) and average nightly sleep duration. Outcomes were Patient-Reported Outcomes Measurement Information System Cognitive Function, Depression, and Anxiety, and Neuro-QoL Emotional-Behavioral Dyscontrol domain T-scores. Multivariable linear regression models were fit for each outcome with SRC history, exercise frequency, diet quality and sleep duration as explanatory variables alongside select covariates.
Results Multivariable regression models (n=1784) explained approximately 33%–38% of the variance in each outcome. For all outcomes, SRC history (0.144≤|β|≤0.217) was associated with poorer functioning, while exercise frequency (0.064≤|β|≤0.088) and diet quality (0.057≤|β|≤0.086) were associated with better functioning. Sleeping under 6 hours per night (0.061≤|β|≤0.093) was associated with worse depressive symptoms, anxiety and emotional–behavioural dyscontrol.
Conclusion Several variables appear to be associated with mood and perceived cognitive function in former NFL players. SRC history is non-modifiable in former athletes; however, the effects of increasing postplaying career exercise frequency, making dietary improvements, and obtaining adequate sleep represent important potential opportunities for preventative and therapeutic interventions.
- concussion
- american football
- diet
- exercise
- Sleep
Data availability statement
The data from this survey study are available upon reasonable request. Inquiries may be sent to Dr ZYK.
Statistics from Altmetric.com
Sport-related concussion (SRC) is a common injury in collision sports, including American football,1 2 yet the neurobiological understanding of the long-term effects of SRC on brain health is unknown.3 Studies of former professional and collegiate American football players have suggested that repetitive SRCs might negatively affect long-term cognitive function,4–13 mood-related symptoms and disorders (eg, depression)4 10 14–19 and executive functioning (eg, disinhibition).4 14 20–25 A limitation of previous research is the lack of consideration of factors that influence the relationship between SRC history and brain health. Specifically, studies of health-promoting factors related to neurobehavioural outcomes are sparse in this population.26 27 Although personal history of SRCs is not modifiable, there may be modifiable factors associated with long-term health that could be leveraged to improve function and quality of life.
Much of the growing body of research in former collision-sport athletes examines factors relating to poor functioning (eg, mood-related symptoms) and risk of neurodegenerative disease (eg, mild cognitive impairment). Interrogation of protective factors that positively affect long-term health and quality of life (ie, exercise, diet and sleep behaviours) may address key knowledge gaps and open avenues for promoting brain health in these individuals. Specifically in former American football players, self-reported physical inactivity, changes in appetite, lower diet quality, sleep disturbance and fatigue have all been associated with self-reported cognitive difficulties and emotional distress.10 28–35 Beyond former athletes, exercise, diet and sleep are potentially modifiable health-promoting behaviours that have previously been shown to improve cognitive function,26 36–41 mood-related symptoms26 40 42 and general health-related quality of life in older individuals.26 43 44 Moderate-to-vigorous exercise performed two or more times per week over multiple years has been linked to the attenuation of cognitive decline in typical ageing, preservation of brain structure and slowing of neurodegenerative disease progression in older adults.36 45 Further, atypical nighttime sleep (ie, less than 6 or greater than 9 hours per night) has been associated with increased brain atrophy, worse cognitive function and a faster cognitive decline with ageing.37 38 46 47 Thus, efforts to modify exercise, diet and sleep may be preventative and therapeutic targets for clinical intervention in former athletes.
The present study examined the associations of self-reported SRC history and health-promoting behaviours (exercise frequency, diet quality and sleep duration) with self-reported measures of cognition and mood (cognitive function, symptoms of depression and anxiety, and emotional–behavioural dyscontrol) in former NFL players. We hypothesised that: (1) more self-reported lifetime SRCs would be associated with worse cognitive function, greater mood-related symptom burden and more emotional–behavioural dyscontrol; and (2) more frequent exercise, better diet quality and adequate sleep (eg, between 6 and 9 hours per night)46 47 would be associated with better cognitive function, lower mood-related symptom burden and less emotional–behavioural dyscontrol.
Methods
This cross-sectional study surveyed former NFL players using a self-administered questionnaire that was provided either online (Qualtrics, SAP America, Newton Square, Pennsylvania, USA) or via paper hardcopy. The study was approved by the Institutional Review Board at the University of North Carolina at Chapel Hill. All participants provided written informed consent prior to participation. This was required in order to gain access to the electronic survey; or it was completed on the first page of the hardcopy survey, prior to completing any demographic or study-related items.
Participants
Former NFL players were contacted between 1 January 2019 and 15 February 2020 through email and hard-copy letters mailed to home addresses through existing relationships with the NFL Alumni Association, as well as individual teams and players (figure 1). Participants were required to have played at least one season in the NFL to be eligible for the study.
Data collection
The current study was part of an ongoing project entitled, Neurologic Function across the Lifespan: A Prospective, LONGitudinal, and Translational Study for Former National Football League Players (NFL-LONG). The questionnaire captured demographic information, football playing history, medical history, concussion history, musculoskeletal injury history, health status, nutrition, substance use and current health-related habits (table 1).
Patient involvement
The comprehensive questionnaire was a modified version of a previously administered questionnaire—enhanced with input from epidemiologists, athletic trainers, neuropsychologists, physicians and other sports medicine professionals.8 15 Additionally, five former NFL players provided critical feedback to improve the questionnaire’s clarity and flow.
Outcome measures of self-reported functioning
Main outcome measures were the PROMIS Adult Cognitive Function Short Form-4a (COG),48 the four-item Depressive Symptoms (DEP) and Anxiety (ANX) scales from the PROMIS-29 Profile V.2.049 and the Emotional and Behavioral Dyscontrol section of the Neuro-QoL battery (EBD).50 51 These measures were developed with support from the National Institutes of Health in the USA using classical test and item-response theories to concisely and precisely measure their intended constructs.48–51 All measures consisted of items that were rated on a 5-point Likert Scale (table 2). Scores for individual items were summed and transformed to T-scores for each domain such that a score of 50 represented the US population normative mean, with a SD of 10. Higher scores indicated a more adaptive outcome in the COG domain and less adaptive outcomes for DEP, ANX and EBD.
SRC history
Participants were prompted to report all concussions that they believed they had sustained while playing football, including instances that were not medically diagnosed. A definition of concussion, used in prior research, was provided to participants52: ‘A concussion typically occurs from a blow to the head and is followed by a variety of symptoms that may include any of the following: headache, dizziness, loss of balance, blurred vision, ‘seeing stars’, feeling in a fog, or slowed down, memory problems, poor concentration, nausea, or throwing-up. Getting ‘knocked out’ or being unconscious does NOT always occur with a concussion’. This definition was selected to be consistent with prior studies that included many of the same participants and, while not exactly the same, the authors felt it was not disparate from the most recent international consensus definition from the concussion in sport group.53 Respondents separately reported the number of SRCs they suspected they sustained on an ordinal scale (ie, 0, 1, 2, …, ‘10 or more’) for each of the following categories: prior to high school, during high school, during college, as a professional football player and during ‘other’ (eg, military-related) football. Self-reported SRC was collapsed into five categories based on the sums reported across all levels: 0; 1–2; 3–5; 6–9; and 10+ lifetime SRCs.
Current health-promoting behaviours (exercise frequency, diet quality, sleep duration)
Current exercise participation was simply reported as the typical number of weekly sessions of moderate-to-vigorous aerobic exercise (defined as exercise that increases breathing rate and heart rate (eg, jogging, cycling)) or resistance exercise (defined as resistance training or weight lifting). No minimum session duration was indicated. Exercise frequency was the sum of moderate-to-vigorous exercise sessions per week, plus the number of resistance exercise sessions per week.
Current diet quality was assessed with the Rapid Eating Assessment for Participants—Shortened Version (REAPS).54 The REAPS captures the types and frequency of food consumption (eg, whole grains, sweets), as well as the preparation of meals (eg, restaurant vs home). It is a brief, 16-item instrument that has shown moderate correlations with the Block 1998 Food Frequency Questionnaire regarding intake of specific food groups.54 For the current study, diet quality was calculated as the sum score of each of the 13 scored REAPS items, ranging from 13 (worst) to 39 (best).
Current, typical sleep duration was recorded via reports of typical times going to sleep and waking up on weekdays and weeknights, separately. The sleep duration variable was calculated as the average number of hours slept per night weighted by weekday and weekend nights. Sleep duration was then categorised into three groups based on the WHO’s Study on Global Ageing and Adult Health46 47 : ‘Low’ (<6 hours per night), ‘Intermediate’ (6–9 hours per night), ‘High’ (>9 hours per night). Low and High sleep groups were dummy-coded so that each group was compared with the Intermediate group as the referent category (eg,<6 hours per night vs 6–9 hours per night).
Covariates
Covariates were selected a priori based on previous literature.8 15 31 55–60 Demographic covariates included: age; body mass index; race/ethnicity (dichotomous; white/non-Hispanic vs person of colour); current marital status (dichotomous; married/cohabitating vs not cohabitating); highest education attained (ordinal; less than a bachelor’s degree, bachelor’s degree, postgraduate degree); and total years of football played (inclusive of prehigh school, high school, college and professional settings). Lifetime physician diagnosis of osteoarthritis or degenerative arthritis (dichotomous; yes/no) was also included as it has been previously shown to be associated with lower quality of life and mood-related symptoms in former elite athletes.55 Behavioural covariates included: current frequency of alcohol intake (ordinal; ‘never’, ‘Monthly or less’, ‘2–4 times a month’, ‘2–3 times a week’ or ‘4 or more times a week’) and illicit drug usage over the prior 3 months (dichotomous; yes/no). Finally, typical experience of pain over the past 7 days was reported on an 11-point Likert Scale (0=no pain to 10=worst pain) as part of the PROMIS-29 Profile V.2.0.49
Data processing and statistical analyses
All analyses were performed with SPSS V.25.0 (Armonk, New York, USA). Separate multivariable linear regression models were fit for COG, DEP, ANX and EBD. Missingness observed within the analysis data set was determined not to be missing completely at random, and consequently, multiple imputation was used while fitting these models. Details on this process are included in the supplement. To fit the regression models using pooled data, standardised values for each predictor and outcome (Z-scores with a mean of 0 and SD of 1) were employed. Normality and heteroscedasticity (homogeneity of variance) of residuals were visually inspected with P–P and spread-level plots, respectively, for each model and variance inflation factors were examined to probe multicollinearity in order to ensure that none of the variables yielded values above the recommended cut-off point of 10.61 Adjusted R 2 and standardised beta coefficients (β) for each predictor were calculated for each model. Standardised β values were considered statistically significant if their 95% CIs did not include zero. Self-reported SRC history and health-promoting behaviours (exercise frequency, diet quality and sleep duration) were included for each outcome alongside all the selected covariates described above. Analyses were additionally performed both with and without participants who reported dementia-related diagnoses (eg, Alzheimer’s disease); however, the results were nearly identical, and these models were therefore not considered further in the present study (online supplemental table 1). Additionally, the summed relative statistical effects (|β|) of exercise frequency, diet quality and Low Sleep were aggregated into a single ‘Health-Promoting Habits’ effect for comparison with the statistical effects of self-reported SRC on each outcome.
Supplemental material
Results
A total of 1784 participants responded to our survey and were subsequently included in this study (figure 1). Most respondents reported that they sustained more than three lifetime SRCs with one-quarter of the sample reporting 10 or more. Approximately one-half of former players indicated that difficulties with cognition occurred ‘often’ or ‘very often.’ Overall, self-reported COG, DEP, ANX and EBD approximated population norms (table 2).48 49 62
Multivariable linear regressions
Overall, the multivariable regression models explained approximately 33%–38% of the variance (ie, adjusted R 2) associated with each outcome (table 3). Following covariate adjustment, a greater history of self-reported SRC was significantly associated with worse self-reported COG, DEP, ANX and EBD (table 3 and figure 2). More frequent exercise and better diet quality were significantly associated with better outcomes in each domain and sleeping less than 6 hours per night (compared with 6–9 hours) was associated with worse DEP, ANX and EBD, but not COG. Sleeping more than 9 hours per night was not significantly associated with any outcome when compared with Intermediate sleep (6–9 hours). Additionally, the statistical effect size (β) of ‘Health-Promoting Habits’ was slightly smaller than that of self-reported SRC on COG, and was relatively larger than self-reported SRC on the DEP, ANX and EBD outcomes (figure 2). Corresponding β values for covariates in these regression models are presented in online supplemental table 2.
Discussion
This study, with a large sample of former NFL players, represents an important step toward identifying modifiable, health-promoting factors associated with daily functioning in former athletes. Explanatory variables individually had small to moderate statistical effects on self-reported functioning, and the combination of these variables, alongside important covariates, explained a substantial proportion of variance (ie, adjusted R 2) for each outcome. Reporting a higher number of lifetime SRCs was significantly associated with worse cognitive function, depressive and anxiety symptoms, and emotional–behavioural dyscontrol. Conversely, significant relationships between health-promoting behaviours and each of these outcomes were identified such that more frequent exercise, better diet quality and sleeping six or more hours per night were each related to better self-reported functioning. The combined magnitude of beneficial effects (β values) observed for these behaviours was relatively similar to or greater than the deleterious effects of increasing SRC history group (figure 2). These findings suggest that even minor changes to health-promoting behaviours may improve cognitive and mental health. As such, these modifiable behaviours are potential targets for therapeutic interventions designed to improve the health and quality of life of former players.
Self-reported concussion history
The observation that self-reported SRC history was related to each outcome is similar to previous work with self-reported cognitive function and mood-related symptoms.8 10 11 14 15 For instance, Guskiewicz et al 8 reported worse SF-36 Mental Component Scale scores in former NFL players with three or more prior concussions when compared with those with fewer. Further, lifetime prevalence of depression diagnosis along with odds of self-reporting moderate or severe symptoms of depression appears significantly higher in former football players reporting three or more prior concussions when compared with those reporting fewer.10 15 Unlike these studies, SRC history was grouped into five categories in the current study—three of these categories included participants that would have been sorted into the ‘3 or more’ category in prior studies,8 10 15 allowing for more granularity in interpreting the possible effects of multiple prior SRCs. Still, the current findings that self-reporting more SRCs was significantly associated with worse self-reported cognitive and mood-related functioning is consistent with this prior work (table 3, online supplemental table 2).
Health-promoting behaviours
A unique component of this study with former NFL players was the investigation of current exercise, diet and sleep habits as possible promoters of health-related functions. The extant literature regarding the benefits of regular exercise on cognitive function, mood-related symptoms and brain health is robust—but less available in former American football players.26 28 33 36 42 45 63 Previous survey studies with former NFL players have shown that self-reported cognition-related problems and moderate to severe depressive symptoms were associated with low levels of physical activity and loss of fitness.28 33 It is unclear, however, whether the depressive symptoms endorsed by players in those studies (as well as the current study) were due to the psychosocial experience during and prior to transition from a sport they had played for years, decreased physical activity, changes in self-esteem or identity, an underlying neurobiological process related to depression symptoms, or some complex mixture thereof. In this study, participation in moderate-to-vigorous aerobic exercise and/or resistance training was significantly associated with better scores on measures of cognition, depressive symptoms, anxiety and behavioural dyscontrol. Compared with the observed deleterious statistical effect (β value) of increasing one concussion history category (eg, 1–2 lifetime SRCs to 3–5 SRCs) for each outcome, the present findings suggest that two or three additional exercise sessions per week may have a favourable effect with comparable magnitude.
Better diet quality (higher REAPS score) was significantly associated with better COG, DEP, ANX and EBD outcomes. Statistical interpretation of this measure in the current study infers that higher quality food choices may result in measurable improvements in daily function. For example, to improve REAPS score by one point, the unit of measure in this study, one could increase the frequency that they consume more than two servings of vegetables in a day, or decrease their frequency of eating fried foods.54 In a related study with former football players, higher fat and cholesterol intake were significantly correlated with greater self-reported cognitive difficulties, suggesting that diet choices are associated with daily functioning.26 Like exercise, dietary changes may have a dose–response relationship with self-reported functioning, for better or worse, warranting further longitudinal investigation with more sophisticated assessments of dietary intake behaviours. Furthermore, dietary recommendations are tangible interventions with the potential to improve daily functioning in this population.
The relationships between typical nightly sleep duration and the outcomes of interest were consistent with previous work that identified durations of sleep of greater than 6 hours per night as being preferable to fewer hours.46 47 Sleeping less than 6 hours per night has been related to worse cognitive functioning and both structural and functional brain health (eg, increased atrophy and functional connectivity changes over time, respectively) across the lifespan.37 38 46 47 64 In former American football players, those reporting ‘trouble sleeping’ have exhibited higher odds of reporting moderate to severe depression symptoms.33 In the present study, being in the ‘low sleep’ group was significantly associated with worse DEP, ANX and EBD, but not COG (the upper limit of the 95% CI for β barely crossed zero). It’s important to consider that sleep duration may not equate to quality of sleep, and a more comprehensive assessment of sleep is needed to fully understand the relationships among sleep and self-reported function, as well as to identify targets for specific sleep-related interventions. Although sleep quality and napping habits were not included in the present study, our results suggest that sleeping at least 6 hours per night is preferable.
Considering the beneficial effects suggested by the data in this study, former players and healthcare providers should discuss the potential for incorporating exercise-related, diet-related and sleep-related behaviours as preventative and therapeutic agents for health-related function. Future research should examine how these behaviours relate to self-reported and objectively measured markers of health through longitudinal investigations with former athletes.
Limitations
First, it should be acknowledged that the response rate for this survey was relatively low. However, an exact number could not be calculated as the authors did not have access to all the necessary data, such as invalid email addresses. Relatedly, we cannot rule out the possibility that the sample of survey respondents was biased in some way (eg, representing a cohort of individuals that are more eager to participate in research studies because of current health-related issues) and that participants who chose not to respond to this survey may not be adequately represented by those who did participate in this study. Second, some of the self-reported predictors and dependent outcomes are limited as they were not corroborated with objective assessments. For example, cognitive functioning was self-reported and was not assessed via neuropsychological testing, and medical chart reviews were not performed to corroborate diagnoses (eg, SRC). As such, we cannot objectively confirm what respondents reported. In addition, respondents were asked to report concussion that occurred years, even decades, earlier, while aware of any potential difficulties they now have, making these responses subject to recall bias. It has been speculated that recalling SRC history from years or decades past may change over time and in relation to changes in health status or understanding of concussion.65 Additionally, it is impossible to discern the directionality or causal nature of relationships observed in a cross-sectional study such as this one. Longitudinal investigations are imperative to determine the natural history and modifiers of health and function over time in former athletes. Finally, predictor and covariate selection for the current study was based on a-priori prioritisation of factors that were suspected to have a meaningful influence on COG, DEP, ANX and EBD, but these variables do not represent an exhaustive list and are consequently vulnerable to residual confounding.
Conclusion
Several variables appear to be associated with the health status of former NFL players. Although SRC history is non-modifiable in former NFL players, the aggregate statistical effects of increasing exercise frequency, making dietary improvements and obtaining adequate sleep duration could be sufficient to impart benefits to self-reported functioning and should be explored as therapeutic targets for health-related outcomes in former players.
What are the new findings?
Sport-related concussion history, exercise frequency, diet quality and sleep duration appear to have measurable individual and aggregate associations with self-reported measures of cognitive, mood-related and behavioural functioning.
How might it impact on clinical practice in the near future?
Although concussion history is unchangeable in former athletes, exercise, diet and sleep are modifiable health-promoting behaviours that warrant future investigation as therapeutic and preventative clinical interventions to improve health and quality of life in former American football players.
Data availability statement
The data from this survey study are available upon reasonable request. Inquiries may be sent to Dr ZYK.
Acknowledgments
The authors are grateful to Mrs Candice Goerger, Ms Hope Campbell, and Ms. Caprice Hunt for their invaluable efforts in the administration, coordination, and execution of the NFL-LONG study. Special thanks are in order for members of the NFL Alumni Association and Legends community for their assistance in recruitment of study participants. Further, we would like to thank each of the clinicians, researchers, educators, and former players who helped us adapt and develop the current general health survey.
References
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
Twitter @SammoWalton, @BenjaminBrett1
Contributors All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Specific author contributions are as follows: SRW, ZYK, BLB and AC: creation and distribution of the study survey, management of study database, established research hypotheses, planned analyses of the data, interpretation of study findings, manuscript preparation and revision, final approval of the submitted manuscript. JDDF, AES-R, LS, RJE, MM, WPM III and KMG: creation and distribution of the study survey, interpretation of study findings, manuscript preparation and revision, final approval of the submitted manuscript. Other contributors to the project are as follows (included in acknowledgements): Mrs CG, Project Administrator: Assisted in the creation and distribution of the survey instrument as well as collection and management of the survey responses. Additionally, she was responsible for overseeing the project administration and execution. Ms HC, Project Coordinator (Medical College of Wisconsin): Assisted in the creation and distribution of the survey instrument as well as collection and management of the survey responses. Additionally, she assisted with oversight of the project administration and execution. Ms. Caprice Hunt, Project Coordinator (University of North Carolina at Chapel Hill): Assisted in the creation and distribution of the survey, oversight of the project administration, and execution of the study.
Funding Funding for this research project comes from the National Football League and Boston Children’s Hospital.
Competing interests BLB acknowledges support from the National Institute of Neurological Disorders and Stroke under the National Institutes of Health under the award NO L301L30NS113158-01. RJE is a paid consultant for the NHL and co-chair of the NHL/NHLPA Concussion Subcommittee. He is also a paid consultant for Major League Soccer and Princeton University Athletic Medicine and occasionally provides expert testimony in matters related to MTBI and sports concussion. WPM receives royalties from 1) ABC-Clio publishing for the sale of his books, Kids, Sports, and Concussion: A guide for coaches and parents, and Concussions; 2) Springer International for the book Head and Neck Injuries in Young Athlete and 3) Wolters Kluwer for working as an author for UpToDate. His research is funded, in part, by philanthropic support from the National Hockey League Alumni Association through the Corey C. Griffin Pro-Am Tournament and a grant from the National Football League.
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.