American Football Play and Parkinson Disease Among Men

Key Points Question What is the association between participation in organized American football and odds of having a reported diagnosis of parkinsonism or PD? Findings In this cross-sectional study that leveraged data from the Fox Insight online study, 729 participants with a history of playing organized football had higher odds of having a reported parkinsonism or PD diagnosis compared with participants in other organized sports. Longer duration of play and higher level of football play were associated with higher odds of a reported diagnosis. Meaning The findings suggest that participation in American football might be a risk factor for developing parkinsonism or PD.


Introduction
Identification of risk factors for Parkinson disease (PD) [1][2][3] is essential for early detection and diagnosis.5][6] Genetic and environmental risk factors include family history of PD and exposure to toxins and pesticides. 7,8ere are other proposed etiologies of parkinsonism, such as vascular parkinsonism that is caused by cerebrovascular disease in association with vascular risk factors (eg, diabetes and hypertension). 92][13] Crane et al 11 analyzed TBI data of 7130 participants from 3 large prospective cohort studies-the Adult Changes in Thought (ACT) study, the Religious Orders Study (ROS), and the Memory and Aging Project (MAP)-and found an association between TBI with loss of consciousness and parkinsonism symptoms among participants from the ROS and MAP cohorts.Among 525 participants from the ACT cohort, there was a significant association between TBI with loss of consciousness greater than 1 hour and Lewy body pathology. 11Gardner et al identified associations between remote TBI and PD, 14 including among US Veterans. 15Other research has failed to find an association between TBI and risk of PD 16 or pathologies that underlie PD. 17 Studies on the association between TBI and PD have not accounted for TBI frequency, which is increasingly shown to be a contributor to the development of later-life neurological conditions.
Repetitive head impacts refers to the exposure to repeated hits, blows, or impacts to the head.
Exposure to repetitive head impact is incurred through participation in contact sports, such as American football, boxing, soccer, and ice hockey, as well as military service and other activities.
3][24][25][26][27] There is a doseresponse association between duration of football and chronic traumatic encephalopathy risk. 21[30] However, the literature on age at first exposure to football is mixed. 31,32Exposure to repetitive head impact has also been associated with Lewy body pathology 33 and might be an independent risk factor for parkinsonism and PD.5][36][37] A mortality study 38 found an increased risk of death due to PD in 3439 former National Football League players who played between 1959 and 1988.Interviews with family members of deceased football players revealed Parkinsonlike features to be common. 39,40Yet Tarazi et al 41 did not find differences in motor signs between 45 former Canadian Football League players and 25 control individuals without a history of concussion.
Research on the association between participation in football and parkinsonism and PD is scarce, limited by informant reports, 27,40,[42][43][44] and a lack of an appropriate comparison group.Large studies with appropriate referent groups that test the association between participation in football and parkinsonism or PD are a critical next step.Here, we leveraged 1875 male participants from the Fox Insight study to evaluate the association between participation in football and odds for having a parkinsonism and/or PD diagnosis.

Participants and Study Design
The sample included participants enrolled in the Fox Insight study 45 -a longitudinal online study of people with and without PD operated by The Michael J. Fox Foundation to explore the experiences of individuals with PD.The Fox Insight study is available to participants older than 18 years (including more than 50 000 enrolled to date), and longitudinal and cross-sectional surveys are administered that assess neurological history (including reported diagnoses of parkinsonism and PD), motor and nonmotor symptoms, quality of life, environmental exposures, and other outcomes.In November

JAMA Network Open | Neurology
American Football Participation and Parkinson Disease Among Men

Sport History
7][48] Participants were asked, "Have you ever participated in organized sports, which includes membership on a team with scheduled practices and games (excluding pick-up or neighborhood games)?"If participants selected yes, the BU-RHIEA assesses history of participation in contact sports.For football, participants were asked, "Did you play organized American tackle football, which includes membership on a team with scheduled practices and games?"We focused on football due to the literature that has associated football with long-term neurological outcomes, including pathologies that cause parkinsonism and PD. 38Participation in football confers more homogenous levels of exposure between players and positions compared with other contact sports where there are many more individual differences in repetitive head impact and significant differences by era of play.The BU-RHIEA assessed age at which football was first started, highest level the sport was played (youth, high school, college, professional), and seasons played at each level, among other sport characteristics.Duration of football play was calculated by summing the number of years played professionally and the number of fall seasons played at each nonprofessional level. 21Fall season was selected because it is the primary season that football is played.There were 2 participants who only played football during spring seasons and were set to 0 for number of seasons played.The sample for duration of play were reduced to 689 because participants preferred not to answer or were not sure how many seasons they played.

Comparison Group
The sample only included male participants who responded yes to the initial screening item of general sport history and those who responded either yes or no to the football question of the BU-RHIEA.That is, nonathletes were not included in this study.This was done to facilitate the derivation of comparison groups to football players who are similar in terms of demographic characteristics, body habitus, physical activity, general health, and mental health.Following assessment of football and other contact sports, participants were asked, "Did you play any other organized sports, which includes membership on a team with scheduled practices and games (not including pick-up or neighborhood games)?"If answered affirmatively, the sport played was specified.Up to 3 other sports were assessed.The comparison group for football included participants who played nonfootball sports.

Sociodemographic, Medical, and TBI History
Participants self-reported demographic characteristics (eg, age, sex, self-reported race and ethnicity, education), disease (eg, parkinsonism or PD status and medications), and health characteristics.Race and ethnicity were queried to evaluate the representativeness of the sample.Participants selfreported height and weight, which were used to derive body mass index (BMI, calculated as weight in kilograms divided by height in meters squared).An adaptation of the Ohio State University TBI-Identification Method (OSU-TBI-ID) 50 assessed TBI history; this survey is referred to as Repetitive Head Impact in the Fox DEN.Participants were grouped into those with and without TBI with loss of consciousness. 11

Sample Size Derivation
A total of 7367 individuals responded to the initial screening question that preceded the BU-RHIEA.
Of these, 4015 (54.5%) endorsed participation in organized sports serving as the initial sample.This

Statistical Analysis
Multivariable binary logistic regressions tested the association between participation in football (yes or no) and odds of having a reported parkinsonism or PD diagnosis.Among the entire sample, multivariable logistic regression tested the association between duration of football play and odds of having a parkinsonism or PD diagnosis.Those who did not play football were coded as having zero seasons of play.We also examined duration of football play as a 3-level variable coded as those who played 0 seasons (which included nonfootball players), 1 to 5 seasons, and 5 or more seasons.The reference group were those who had 0 seasons of play.Five seasons of play was selected based on previous research. 21The multivariable logistic regression model for duration of football play was repeated among only football players.Among the football players, we investigated associations between level of football play and age at first exposure to football and odds of having a parkinsonism or PD diagnosis in separate models.24][25]51   b Binary logistic regressions tested for associations between participation in American football and odds of having a reported parkinsonism/Parkinson disease diagnosis.
c Duration played was examined in the entire sample with those who did not play football coded as 0.
e Duration of American football play, highest level played (youth/high school vs college/professional), and age at first exposure to football were examined among football players only.

Discussion
In this convenience sample of participants enriched for parkinsonism or PD from the Fox Insight online study, 45 we evaluated the association between playing football and odds of having a reported parkinsonism or PD diagnosis.There were 729 football players who predominantly played at the amateur level; 1146 men who played nonfootball sports served as the reference group.History of playing football was associated with higher odds of having a reported parkinsonism or PD diagnosis after accounting for age, education level, history of diabetes and heart disease, BMI, TBI with loss of consciousness, and family history of PD.There were dose-response associations between seasons and level of football play, such that more seasons and higher level of football play corresponded to higher odds of having a reported diagnosis of parkinsonism or PD.Age at first exposure to football was not associated with odds of having a reported parkinsonism or PD diagnosis.
Previous studies have found associations between TBI and risk of parkinsonism. 11,14,15,52garding repetitive head impact, parkinsonism and PD have been described in boxers. 26Evidence from clinicopathological studies have identified parkinsonism among football players. 27,33,40In 1 autopsy study among 269 contact sport athletes who participated in football, ice hockey, soccer, rugby, boxing, and martial arts, more years of contact sport play was associated with greater odds of having neocortical Lewy body disease. 33Parkinsonism was associated with Lewy body pathology but not with chronic traumatic encephalopathy stage. 33Prospective studies on the association between football participation and parkinsonism and PD are limited.Tarazi et al 41 did not observe motor abnormalities in 45 former Canadian Football League players.Yet, among 3439 former National Football League players who played at least 5 years between 1959 and 1988, there was a 1.69-fold increase in risk of death due to PD compared to the United States population (not statistically significant). 38Our findings provide additional evidence for an association between football and parkinsonism or PD.
The cause of parkinsonism and PD in former football players is multifactorial and related to factors associated and unassociated with exposure to repetitive head impact.Duration of play, a proxy for exposure to repetitive head impact, has been associated with chronic traumatic encephalopathy status and severity. 21,53We observed associations between longer duration of football play and higher level of football play and greater odds of having a reported diagnosis of parkinsonism or PD.Most of the football players played at the amateur level.Studying individuals with lower levels of football play is of high priority because most people play at the high school and college level.However, current research on the long-term neurological outcomes of football has focused on those who played professionally.
To our knowledge, this is the largest study to describe the association between participation in football and odds of having a reported diagnosis of parkinsonism or PD.Strengths of this study include sample size, focus on amateur athletes, and the comparison group of former athletes.

Limitations
This study has several limitations.The generalizability of the current findings is limited by the homogeneous demographic make-up of the sample, including 98% of participants who reported being White.The study used a convenience sample of participants enriched for having parkinsonism and PD, and 1602 individuals (85.4% of the entire sample) reported a diagnosis of parkinsonism or PD.The Fox Insight online study has minimal eligibility criteria and includes people with and without PD.By nature, people who are at risk or have concerns for PD more likely to participate.The selectivity of the sample thus limits generalizability (ie, external validity).While people likely participate in the Fox Insight Online study for reasons that could be related to having Parkinson disease, it is unlikely that people choose to participate because of their football history.Therefore, it seems unlikely that the observed associations between football history and parkinsonism or PD can be attributed to the selection process.Our other studies 21,53 also show that the strength of associations between football history and neurological outcomes are maintained even under scenarios of intense selection pressure.The study was done entirely online without supervision and relied on self-reported exposures and outcomes.There is potential measurement error related to recall bias as participants were asked to recall historical information and events that could have occurred many years ago.While online studies have been validated, 54,55 the gold standard of parkinsonism and PD diagnosis is an in person examination with motor assessments.Parkinsonism and PD were not separately queried, and we were unable to differentiate if the current associations or lack thereof pertain to parkinsonism or PD.Former football players are more prone to orthopedic injuries, particularly those who played at elite levels. 56While focus on parkinsonism and PD diagnosis attenuates concerns for orthopedic injuries as confounding, it remains a potential cause of motor abnormalities.

Conclusions
In this cross-sectional study, there was an association between participation in American football and higher odds of self-reported parkinsonism or PD.Among football players, odds of having a parkinsonism or PD diagnosis were greater with more seasons and higher level of football play.The findings suggest that American football participation might be a risk factor for developing parkinsonism or PD.Prospective research among community-based samples that objectively evaluate parkinsonism and PD in former American football players across different levels of play will clarify the observed associations.
sample size was reduced to 1875 after exclusion of participants with missing data on the BU-RHIEA yes or no football question (n = 14), missing data on our primary model variables (n = 750), female sex (n = 1320), and discrepant reporting of parkinsonism or PD diagnosis (n = 56).Sample sizes varied by associated athletic variables examined (ie, duration played and level of play) due to missing data.While there was a large sample of female individuals, they were generally part of the nonfootball comparison group and analyses were restricted to men due to insufficient sample of female football players (n = 10 [1.8%]).

Figure . B
Figure.Estimated Marginal Mean Differences in Duration of Football Play by Parkinson Disease or Parkinsonism Status August 11, 2023.doi:10.1001/jamanetworkopen.2023.28644Open Access: This is an open access article distributed under the terms of the CC-BY License.© 2023 Bruce HJ et al.JAMA Network Open.Corresponding Author: Michael L. Alosco, PhD, Boston University Alzheimer's Disease Research Center, Boston University Chronic Traumatic Encephalopathy Center, Department of Neurology, Boston University Chobanian & Avedisian School of Medicine, 72 E Concord St, L-5, Boston, MA 02118 (malosco@bu.edu).Author Affiliations: Boston University Alzheimer's Disease Research Center, Boston University Chronic Traumatic [46][47][48]ston University Repetitive Head Impact Exposure Assessment (BU-RHIEA)[46][47][48]was launched in Fox Insight for data collection on exposure to repetitive head impact from contact sports.It also assessed participation in noncontact sports.It was made available to all Fox Insight participants.Data for this study represent participants who responded to the BU-RHIEA by the time data were obtained from the Fox Insight database (Fox DEN).Data used in the preparation of this manuscript were obtained from the Fox DEN on June 9, 2022.Up-to-date information on the study can be found online.49Participantsprovide informed consent online using the Fox Insight website.

Table 2 )
This was done to exclude participants in the reference group from having many years of repetitive head impacts from other high exposure to repetitive head impact contact sports.Analyses controlled for age at time of the BU-RHIEA, education level, history of diabetes, history of heart disease, BMI, TBI with loss of consciousness, and family history of PD.When assessing age at first exposure, duration of football play was controlled for in all models.A P value less than or equal to .05 defined statistical significance.Analyses were conducted using SPSS version 27 (IBM Corp).Of the participants who endorsed playing organized sports, 729 (38.9%) played football.Other common contact sports included soccer (n = 444 [23.7%]), ice hockey (n = 172 [9.2%]), amateur wrestling (n = 115 [6.1%]), and boxing (n = 92 [4.9%]).Of the nonfootball players, 588 (51.3%) played a noncontact sport and the most frequent included baseball (n = 276 [24.1%]), basketball (n = 204 [17.8%]), and tennis (n = 95 [8.3%]).Of 1875 participants, 1602 (85.4%) reported having a parkinsonism or PD diagnosis.Those who had a reported parkinsonism or PD diagnosis were older, had a lower level of education, had a lower BMI, and were less likely to have a family history of parkinsonism or PD..Most played at the youth or high school level (n = 575 [82.4%]) compared with 118 (16.9%) at the college level and 5 (0.7%) at the professional or semiprofessional level.Of the football players, 648 (88.9%) reported being diagnosed with parkinsonism or PD.Playing football was associated with higher odds of having a reported parkinsonism or PD diagnosis (odds ratio [OR], 1.61; 95% CI, 1.19-2.17)(Table3).Longer duration of football play was associated with higher odds for having a reported parkinsonism or PD diagnosis (OR, 1.12 per season played; ResultsA total of 1875 participants endorsed playing any organized sport (mean [SD] age, 67.69[9.84]years;18[1.0%]AmericanIndianorAlaskaNative, 21 [1.1%]Asian, 10 [0.5%] Black or African American, 1839 [98.1%]White, and 9 [0.5%] other, including Native Hawaiian or Other Pacific Islander and those who preferred not to answer; in terms of ethnicity, 79 [4.2%] were Hispanic and 1796 [95.8%] non-Hispanic) (Table 1).95%CI,1.06-1.19).Having low exposure (n = 435) was associated with a 1.39 (95% CI, 0.98-1.98)increase in odds of having a reported parkinsonism or PD diagnosis, compared with a 2.18 OR (95% CI, 1.36-3.49)forthesubstantialexposuregroup(n = 252).The Figure, A, shows mean differences between those with and without parkinsonism or PD on seasons of football.Among the football players, longer duration of football play (OR, 1.12 per season played; 95% CI, 1.02-1.23)andhigherlevel of play (youth or high school vs college or professional: OR, 2.93; 95% CI, 1.28-6.73)wereassociatedwith higher odds of having a reported parkinsonism or PD diagnosis.The Figure, B, shows mean differences.Among the football players, age at first exposure to football (mean [SD], 12.66[2.39];range, 5-20 years) was not associated with a reported parkinsonism or PD diagnosis (OR, 1.04; 95% CI, 0.93-1.17).

Table 1 .
Sample Characteristics by Reported Parkinsonism or PD Status a a The entire sample of participants were men.bComparison using independent sample t tests.cComparison using χ 2 .dTotal sample size was 1594 for years with PD diagnosis when enrolled in Fox Insight.Sensitivity Analyses:

Exclusion of Nonfootball High Repetitive Head Impact Sports Models
were repeated after excluding participants who participated in boxing, soccer, or ice hockey (eTable in Supplement 1).Of these 1247 men (mean [SD] age, 69.40[8.78]years),545(43.7%)played football (mean [SD] age, 69.58[8.30])and494 of 545 (90.6%) had parkinsonism or PD.Playing football was associated with higher odds of having a reported parkinsonism or PD diagnosis (OR, 1.65; 95% CI, 1.13-2.41).Of the entire sample, longer duration of football play was associated with

Table 2 .
Sample Characteristics by American Football Participation a a The entire sample of participants were men.b Comparison using independent sample t tests.c Comparison using χ 2 .d Total sample size was 1594 for years with PD diagnosis when enrolled in Fox Insight.

Table 3 .
Summary of Multivariable Binary Logistic Regression Models a All models adjusted for age, education level, history of heart disease, history of diabetes, family history of Parkinson disease, traumatic brain injury with loss of consciousness, and body mass index. a