YPAS: Yale Physical Activity Survey; FFQ: Food Frequency
Questionnaire; CTE; Chronic Traumatic Encephalopathy;
NFL: National Football League; NHL: National Hockey League;
DHA:Docosahexaenoic Acid; EPA: Eicosapentaenoic Acid.
The prevalence of cognitive impairment is suspected to be high
in the retired contact sport athlete population [1]. Mild traumatic
brain injury (mTBI), commonly referred to as concussion in
the sports literature, is a common injury that National Football
League (NFL) and National Hockey League (NHL) athletes
suffer from repeatedly throughout their professional careers. It is thought that repeated mTBI throughout life may lead to
the development of Chronic Traumatic Encephalopathy (CTE),
which is a progressive neurodegenerative disorder [2]. There
has been much research done on CTE and post-mortem analysis
shows that 80% to 99% of contact sport athletes, mainly former
NFL and NHL players, had observable tau deposition which was
characteristic of CTE [3,4]. However, this tau deposition has not
been linked to the suspected clinical manifestations of CTE and
there is much controversy regarding the subject [5,6].
There has been some research on other possible causes of
neurodegeneration, including cardiovascular risk and cognitive
reserve, [7-9] in the retired contact sport population, but little
attention has been given to their nutritional and general lifestyle
habits. Lack of micro- and macro-nutrients in the diet and the
decline of regular physical activity has been linked to neurodegradation
and early aging of the brain [10,11]. Nutrition is
a fundamental component of an athlete’s training regimen;
however, news articles suggest many current professional football
players are not aware of their specific dietary recommendations.
Athletes just starting out in their professional careers lack
proper dietary and life style education for post-professional
sport life and may develop improper habits during retirement.
The purpose of this study is to determine nutritional and lifestyle
habits of retired professional contact sport athletes and compare
with age-matched, retired non-contact sport athletic controls.
This study furthermore aims to provide a descriptive report of
any differences in nutritional intake between contact and noncontact
sport athletes after retirement and suggest a diet rich in micro-nutrients which may have a role in brain health after
mTBI. We hypothesize that retired contact-sports athletes would
have lower energy expenditure, higher calorie intake, and lower
micronutrients than retired non-contact sport athletes.
This IRB-approved case control study was completed as a
part of the Healthy Aging Mind Project [12] at the University at
Buffalo, SUNY.
Participants
Participants were invited to complete a series of
neurocognitive tests, a blood draw and questionnaires about
their lifestyle. Retired professional contact sport athletes (CS)
were contacted and recruited through NFL and NHL alumni
associations. Criteria for inclusion in this group included (1)
played a professional contact sport for two or more seasons; (2)
currently retired from competitive sports; and (3) between the
ages of 36-72 years. Criteria for exclusion included (1) unwilling
or medically unable to receive an MRI, (2) sustained a concussion
within the past two years, (3) history of moderate to severe brain
injury (4) history of cerebrovascular event that could lead to
hypoxia, and (5) history of a learning disability. Non-contact sport
athletic controls (NS) were recruited by contacting athletic clubs
for runners, swimmers or tri-athletes with rosters that included
older athletes. Criteria necessary for inclusion were (1) athletes
who participated in local or national competitive individual noncontact
sport such as running, cycling, or swimming when they
were younger; (2) currently retired from competitive sports;
(3) between the ages of 36-72 years; and (4) no history of selfreported
or documented concussions. Exclusion criteria were the
same as the retired professional contact sport athlete group.
Outcome measures
Demographics:A standard questionnaire was used to obtain
demographics and medical, vascular risks factors, and sport
history. History of alcohol abuse, drug abuse, hypertension,
diabetes and hypocholesteremia were self-reported. Venous
blood sample was collected by phlebotomist at first visit, and
sent to in-site laboratory for analysis. High density lipoproteins
(HDL) and low density lipoproteins (LDL) were measured using
standard laboratory protocols and cut-off values were 40-60 mg/
dL and 0-100 mg/dL respectively [13]. Resting blood pressure
after 2-minute sitting (using Welch Allyn Connex ProBP 3400
Digital Blood Pressure Device) was measured to calculate
Atherosclerotic Cardiovascular Disease (ASCVD) 10-year risk
using the Framingham Heart Study equation. The ASCVD risk
calculator uses a high blood pressure cut-off value of ≥ 140/90
mmHg and cardiovascular event risk of 7.5% or higher is
considered as elevated risk [14].
Mild cognitive impairment (MCI): MCI was diagnosed by
a neuropsychologist using the Comprehensive Criteria, which
is that at least two performances within a cognitive domain
fall below the established cutoff in order for that domain to
contribute to the MCI classification [6,15].
IQ estimation: IQ was estimated using Wide Range Achievement Test-4 (WRAT-4) which was performed by a
neuropsychologist.
Estimated energy expenditure (EEE): Yale Physical Activity
Survey (YPAS), [16] a self-reported questionnaire, was used
to assess the subjects’ EEE. It includes seasonal modifications
that take in to account the time of year that the activity is being
completed and has documented repeatability and validity [17].
The YPAS asks subjects to recall various activities that they
regularly perform and creates a score based on the intensity,
season, and duration of the activity, and is measured by estimated
calories burned while doing various tasks. The calories of each
physical activity are added to the Basal Metabolic Rate (BMR,
calculated by Harris-Benedict equation) to obtain the EEE. This
questionnaire was completed at home with instructions.
Food frequency questionnaire (FFQ): FFQ was used to
estimate nutritional intake [18]. The FFQ comprises a detailed
self-report daily, weekly, and monthly nutrition intake which is
used to obtain a year-long recall of diets and has been validated
for reproducibility in a variety of populations [19,20]. Nutritionist
Pro software (Nutritionist ProTM, Axxya Systems, USA) is then used
to calculate the macro- and micro-nutrient intake. Micronutrients
are divided by the daily recommended intake values to calculate
Percent Daily Recommended Intake (%DRI). This questionnaire
was completed at home with instructions.
Statistical analysis
Based on computed power analysis, a total of 20 participants
in each group was required in order to achieve a power of 0.80
with a one sided test at level 0.05. A series of t-tests (equal varices)
and Chi-squared tests were used to assess group-wise differences
in demographics, estimated IQ, vascular risk factors, and sport
history. P-values less than 0.05 were considered significant. Nonparametric
t-test was used to examine group wise differences
in each EEE, macronutrient (grams) and micronutrient (%DRI).
Bonferroni correction was used to account for multiple pair
wise comparisons in micronutrients and a p-value of 0.001
was considered significant (0.05/27). Logistic regression was
performed to see if there was any association between diagnosis
of MCI (binary response) and age, BMI, calorie intake, EEE and
hours spent performing physical activity. Statistical analysis was
performed using SPSS Version 24 (Armonk, NY: IBM Corp).
A total of 27 retired professional football and hockey players
took part in the study. Five participants withdrew from the study
before completing all the procedures and one was excluded due
to a serious brain injury from a motor vehicle accident which
was only revealed during imaging. Hence, 21 retired NFL and
NHL athletes made up the CS group. All participants correctly
filled out the YPAS, but only 12 out of 21 participants correctly
filled out the FFQ so demographics and EEE are calculated
from 21 participants and nutritional intake is calculated on 12
participants. Completion rates of the FFQ were significantly
different, but there were no statistically significant differences in
demographics, vascular risk factors, or physical activity between
those participants who completed the FFQ correctly and those who did not. A total of 24 non-contact sport retired athletes
signed up to participate but three withdrew from the study before
completing all the procedures. All participants correctly filled out
the YPAS and FFQ, hence 21 participants made up the NS group.
CS were significantly more overweight than NS (30.1 ± 3.5 vs.
24.5 ± 2.5 kg/m2, p < 0.001), had lesser education (p = 0.024),
lesser estimated IQ (49.29 ± 6.76 vs. 57.57 ± 8.82, p = 0.01), and
had lower HDL cholesterol (42.6 ± 8 vs. 49.8 ± 11, p = 0.017).
More CS met the criteria for MCI than NS (8 vs 3, p = 0.083), but
this did not reach significance. None of the other variables were
significantly different. Demographics, estimated IQ, incidence of
MCI, and vascular risk factors are presented in Table 1.
CS had a significantly higher BMR (1960.8 ± 217 vs. 1628.2
± 177 kCal/day, p < 0.001) than NS. There was no significant
difference in EEE between CS and NS (2229.7 ± 371 vs. 2123.7
± 272 kCal/day, p = 0.297), however, CS spent significantly less
time performing physical activities than NS (22.5 ± 18.7 vs. 51.1
± 15.0 hours/week, p < 0.001). Individual domain scores from the
YPAS are presented in Figure 1. Vigor (p < 0.001), walking (p =
0.009), and sitting (p = 0.021) scores were significantly different
between groups.
There were no significant differences between the CS and
NS’s estimated macronutrient intake. Additionally, there were no
statistically significant differences in the different fats, i.e. oleic,
linoleic, linolinic, eicosapentaenoic (EPA) and docosahexaenoic
(DHA) acid. Riboflavin (p = 0.047), biotin (p = 0.012), folate
(p = 0.02), vitamin D (p = 0.035), copper (p = 0.019), selenium
(p = 0.037), potassium (p = 0.02), phosphorous (p = 0.043),
manganese (p = 0.002), and fiber (p = 0.02) were lower in CS
than NS, but were not significant after Bonferroni correction.
Figure 2 shows the macronutrient intake and Figure 3 shows
the micronutrient %DRI of the 12 CS and 21 NS participants who
correctly completed the FFQ.
On regression analysis, diagnosis of MCI was not significantly
associated with age (p = 0.199), calorie intake (p = 0.604), BMI (p
= 0.067), EEE (p = 0.669), or hours spent doing physical activity
(p = 0.069).
The purpose of the Healthy Aging Mind study was to
extensively evaluate the cognitive, behavioral, and lifestyle
characteristics of athletes who had professional careers playing contact sports that may have left them vulnerable to CTE and
compare them to athletes who competed in individual, noncontact
sports. The research and media attention on CTE have
given the impression that former contact-sport athletes will
experience early onset dementia marked by cognitive impairment
from receiving repetitive concussive or sub-concussive blows
to the head [21]. Although the CS group scored significantly
lower on estimated IQ, this can be explained by the difference in
education which has been shown to have a directly proportional
relationship with IQ [22]. Based on demographics, the control
NS group was considered healthier due to lower BMI and higher
HDL, which are known risk factors for early cognitive decline. An
investigation in 2012 [9] studied 38 over-weight and 38 healthyweight
retired NFL players (mean age = 57 years, range 25 -
82) and showed that players with higher BMI had significantly
more cognitive decline which suggests it is an independent risk factor in retired contact sport athletes. Increased HDL has also
been shown to be correlated with increased cognitive function in
healthy and pathological conditions, [23] as well as greater grey
matter volume [24]. The CS group had a higher incidence of MCI
(38% vs 14%) that did not reach significance. This could be due
to our small sample size because our power calculation did not
include detection of rare events; however, the control group was
much superior in physical and cognitive health so we expected
this to reach significance. Regression analysis between MCI and
main outcome variables (EEE, BMI, calorie intake and hours
spent doing physical activity) had interesting results. Although
the results were not statistically significant, both BMI and hours
spent performing physical activity, were approaching significance
(p = 0.067 and 0.069 respectively) indicating there may be some
correlation. Further research needs to be performed with larger
sample sizes. Detailed neurocognitive testing was performed on all participants as part of the extended study protocol; these
results are presented in a separate paper [6].
The YPAS provided mixed results for physical activity when
we compared the two groups. BMR is the amount of calories the
body burns at rest and EEE is the amount of calories it burns
when exercise is taken into account. CS spent significantly less
time performing physical activities, i.e. were more sedentary,
and had significantly less vigor and walking index scores, but
had a much higher BMR due to their high BMIs. This led to nonsignificant
differences in EEE between the two groups. Estimated
average calorie intake per day can also be obtained from the
FFQ and although the results from the two questionnaires were
comparable, we presented the values from the YPAS instead of
the FFQ because a large proportion of the CS did not correctly
complete the FFQ. Still, sedentary life style irrespective of BMR
has been associated with a variety of neurocognitive diseases.
They have been linked to brain atrophy, [25] neuroinflammation,
[26] and vulnerability to trauma and disease, [27] whereas
regular exercise has been linked to several benefits [28]. The
mechanism of action is suspected to be due to induction of
factors that promote neural growth and repair. Brain derived
neurotropic factor (BDNF) has been associated with these
effects and has been shown to increase hippocampal volume and
improve spatial memory [29]. Animal and human studies confirm
that exercise increases BDNF levels and brain function as early
as 5-6 weeks after initiation of aerobic training [30,31]. The
rapid benefit effect of exercise on neuroplasticity suggests that
improved neuronal function rather than reduced cerebrovascular
disease risk is the cause for functional improvement.
While analyzing the FFQ, it was evident that several members of the CS group did not fill out the FFQ correctly, leaving only
12 sets of FFQ’s to analyze against all 21 of the NS group. The
main reason for the forms being incorrect was that the CS group
checked the different food items instead of writing the amount
they consumed. These forms were completed at home and
returned to the research center when they returned. Follow-up
was attempted to obtain completed FFQ but majority did not
want to return. The reason for this drastic difference is uncertain,
both groups received the same instructions and there were
no significant differences in demographics, except education,
between those who completed them correctly and those who did
not. There were no significant differences in macro- or micronutrient
intake, but several brain healthy nutrients (copper,
selenium, folate, manganese, vitamin D) were lower in CS than
NS. In past studies, low plasma copper levels have been linked
to cognitive decline [32]. This indicates a potential correlation
between copper intake and cognitive decline, as well as a need
for further research on the topic. Recently, researchers have
theorized that selenium may play a critical role in preserving
the cognition of individuals consuming a high fish diet [33].
Although there is little data on selenium’s effect on brain health,
if administered intravenously it may help improve neurologic
function after suffering a TBI [34]. Foods high in folic acid, which
gets converted to folate in the body, are essential components of
a healthy diet, and may have an effect on brain health. Numerous
studies have shown that lower rates of folate intake correlate to
faster cognitive decline [35]. Manganese is transported to the
brain through the blood-brain barrier, and when levels are within
normative values, may influence synaptic neurotransmission,
making it a potential brain-healthy nutrient [36]. In low levels,
manganese has proven to lead to neuro-deficiencies, while at high levels it may act as a neurotoxin, indicating that close maintenance
of manganese is important in preventing its negative effects more
than improving cognition [37]. There is evidence that low levels of
vitamin D is related to lower levels of cognition, yet there is little
to no evidence demonstrating that supplementation of vitamin D
will improve cognition [38]. In the past, studies have noted the
correlation between decreased cognition and less sun abundant
months (a major source of vitamin D) [39]. However, determining
the effects of dietary vitamin D intake is made difficult due to
varying outdoor or sedentary lifestyles in participants.
There are several vitamins and minerals found in the diet
that have been coined as “brain healthy” foods, and regular
consumption of foods containing these nutrients has been linked
to improved cognition and overall brain health. Some of these
brain healthy nutrients which were not different between the
two groups are vitamin C and E, omega-3 fatty acids, zinc, niacin,
and pyridoxine. Vitamin C is essential for brain function due to its
ability to form ascorbic acid, an antioxidant, and prevent cognitive
decline. A review from 2014 found a direct effect of vitamin C
deficiency on brain development following traumatic injury [40].
Vitamin E has been linked to protecting against oxidative stress in
the brain. While past study results vary in their legitimization of
vitamin E as a tool in fighting cognitive decline, its anti-oxidative
effects seem promising [41]. Little is known about the role of
DHA and EPA on the brain in response to mild head trauma, but
there has been substantial evidence demonstrating the positive
effects of these nutrients in stroke patients, demonstrating that
it may have beneficial effects on brain health [42,43]. They have
also been shown to prevent the loss of gray matter volume in
patients suffering from schizophrenia [44]. Zinc is furthermore
useful in the brain because it is needed to metabolize EPA and
DHA. A 2005 study found that Zinc supplementation may benefit
adolescent cognition; however no such studies have been
performed in reference to traumatic brain injury [45]. Niacin
therapy has also been proven beneficial; it has improved brain
plasticity in post-stroke mice, suggesting that it may be a worthy
tool in treating traumatic brain injuries [46]. Pyridoxine is also
suspected to be a brain healthy nutrient. A study in 2014 found
an inverse correlation between complete vitamin B intake and
several cognitive conditions, such as mild cognitive impairment,
suggesting a need for further investigation into individual subsets
of vitamin B [47]. A suggested brain healthy diet to consume
after brain injury, including common sources, is presented in
Supplementary File 1.
This study has a number of limitations. The greatest limitation
is the unequal sample size for the nutrient intake comparisons.
The overall sample size is also small and there is no obvious
means for determining the representativeness of the sample to
the population of retired professional athletes who played contact
sports. We recognize that studies that report no significant
differences between groups are often underpowered to make
such claims, but our control population was much more superior
(retired non-contact master athletes who continue to remain
active to some extent, lower BMI, more educated and higher
estimated IQ). We hypothesized that these former athletes would
be significantly different in lifestyle choices when compared with a superior control group who not only played in their youth, but
also continued to stay active since they were members of cycling
and running clubs. We were surprised to find so few major
differences in food intake or energy expenditure. Still, our sample
size is small and the study results need to be replicated with
longitudinal assessment. We did not collect information about
cause of retirement or if the athletes continued to physically train
after retirement, also, the average age of retirement is different
for different types of sports, which can confound our results. The
body composition between contact sport athletes (e.g. football
players) is different from non-contact sports (e.g. runners) which
may have contributed to our difference in BMI, however, we
found significant differences in hours spent performing physical
activities (i.e. chores, work, recreational activities) which should
not be affected by body composition. Additionally, all of the forms
used were self-reported, therefore subject to possible recall bias.
Retired contact sport athletes were significantly more
overweight than age-matched, retired non-contact sport athlete
controls and spent significantly less time performing physical
activities, which makes them at risk for early neurocognitive
decline. However, there was no significant difference between
their daily estimated energy expenditure, which can be attributed
to their higher BMI. Retired contact sport athletes consumed
lesser brain healthy nutrients than healthy controls, but there
was no significant evidence that they were lacking in any major
macro- or micro-nutrients in their diets. Longitudinal research
with larger sample sizes are required to further assess the risk of
nutrition and lifestyle choices on early neurocognitive decline in
retired contact sport athletes.
Authors’ role
MH, KD, JL, PH and BW designed research. MH, IB, JL and BW
collected data. KD, MH and BW analyzed the data. MH, KD IB and
JL interpreted the data and prepared the manuscript, all authors
approved the final version of the paper.
Conflict of Interest (COI) Statement
The authors report no conflict of interest.
Funding
We wish to thank the following organizations for financial
support: The Robert Rich Family Foundation, Program for
Understanding Childhood Concussion and Stroke, Buffalo Bills
(Ralph Wilson) Team Physician Fund and the Buffalo Sabres
Foundation. Research reported in this publication was supported
by the National Institute of Neurological Disorders and Stroke
of the National Institutes of Health under award number
1R01NS094444 and by the National Center for Advancing
Translational Sciences of the National Institutes of Health under
award number UL1TR001412 to the University at Buffalo. The
content is solely the responsibility of the authors and does not
necessarily represent the official views of the National Institutes
of Health.