Abstract
Background:
Sports participation benefits children but increases the risk of mild traumatic brain injury (mTBI) and orthopedic injury (OI). This study examines risks of mTBI vs. OI associated with specific sports and benefits of sports participation.
Method:
This is a cross-sectional study analyzing baseline data from the Adolescent Brain Cognitive Development (ABCD) Study, with a sample of 11,055 children aged 9–10. Generalized linear mixed-effects models were used to examine whether the risks of mTBI and OI differed across individual sports.
Results:
Compared to children who had not participated in climbing, those who participated had a higher risk of mTBI than OI (ratio of odds ratio = 1.881, p = 0.013). Sports participation was associated with better behavioral/emotional outcomes, with stronger benefits for mTBI children compared to those with no injury (p = 0.043), but no significant difference between mTBI and OI groups.
Conclusion:
Different sports have distinct risks for mTBI and OI in children. Behavioral benefits of sports were more pronounced for children with mTBI than for uninjured children but similar between mTBI and OI groups. While causal connections cannot be established with the current study design, these findings suggest the need for sport-specific and injury-specific strategies to mitigate risks and maximize benefits of youth sports.
Keywords: Mild traumatic brain injury, orthopedic injury, sports, behavior, neurocognition
Introduction
Participation in sports offers many psychological and social benefits for youth, such as preventing obesity, improving self-esteem and social interactions, and enhancing health-related quality of life (1-5). In the United States (U.S.), approximately 30–45 million youth 6 participate in organized or recreational sports. However, sports participation is not without risk. The intrinsic risks of athletic activities expose youth to injuries including head injuries (4-7). Approximately 1.1 to 1.9 million sports- and recreation-related concussions (SRRCs) occur each year among children under 18 in the U.S. (8).
Concussions in children or pediatric traumatic brain injury (TBI) are a significant public health concern in the U.S. In 2013 alone, among children under the age of 15, TBI was responsible for approximately 1,500 deaths 18,000 hospitalizations, and over 640,000 emergency department (ED) visits (9). The majority of pediatric TBIs fall under the category of mild traumatic brain injury (mTBI), defined by the American Congress of Rehabilitation Medicine as a traumatically induced disruption of brain function including at least one of the following: (1) loss of consciousness (LOC) lasting less than 30 minutes, (2) posttraumatic amnesia lasting less than 24 hours, (3) altered mental state, (4) focal neurological deficit (10-12). Mild TBI can have potential long-term effects on children’s mental health, behavioral development, psychiatric outcomes, and academic performance (13-16).
Given the widespread involvement of children in sports, understanding the relationship between sports participation and injury risk is essential for developing effective prevention strategies and promoting safe youth sports participation. Certain sports, particularly sports that heavily involving play–play contact and play-surface contact such as American football, soccer, hockey, rugby, and lacrosse, have been identified as having a higher risk for concussions in high school athletes (17,18) or among children aged 18 or less (7). However, as we noted in our previous study (5), research specifically focusing on prepubertal children (such as ages 9–10) is still needed, as their sports participation tends to be less intense than that of adolescents, and their ongoing development in coordination, balance, and risk awareness may lead to different injury patterns.
The Adolescent Brain Cognitive Development (ABCD) study is an ongoing, multisite, and largest long-term study of children’s health and brain development in the U.S. The study enrolled over 11,000 children and followed them annually recording their biopsychosocial data (19). Our previous cross-sectional study using the baseline visit data from the ABCD study found that some individual sports were associated with risks of mTBI in children aged 9–10, and these associations differed by sex (5). Boys who played tennis had a significantly higher risk of mTBI than girls who played tennis, and boys who participated in swimming/water polo had a significantly lower risk of mTBI than girls who participated in swimming/water polo (5).
While our previous findings highlight the sport-specific risks of mTBI, concussion is just one type of sports-related traumatic injury. A large proportion of sports-related traumatic injuries result in fractures (orthopedic injury; OI), joint dislocations, muscle strains, and ligament strains (20,21). Fractures have been identified as one of the major injuries in youth American football (22), climbing (finger growth plate injury) (23), and basketball (24). It remains unclear whether participation in certain sports poses a higher or lower risk of mTBI compared to OI. Understanding these differences is important for identifying which sports carry a higher/lower risk of head injury versus broken bones and informing targeted prevention strategies. The primary aim of this paper is to examine the association between sports participation and the risks of mTBI versus OI, focusing on sport-specific differences in the risk of mTBI versus the risk of OI. It is important to note that in the ABCD study, the timing of injury is not recorded. As such, we cannot determine whether the reported injuries occurred during sports participation or outside sports. Therefore, the current study evaluates injury risk among children who participated in sports versus those who did not.
In addition, our previous study indicated that regardless of children’s history of mTBI, playing sports was associated with better behavioral and neurocognitive functioning, and fewer depressive symptoms (5). To better understand the unique effect of sports associated with mTBI, it is essential to note that OI serves as an important comparison group in addition to an uninjured group in mTBI studies. The use of an OI control group is necessary as comparing children with mTBI to a healthy control group does not capture both the experience of sustaining an injury, regardless of the impact of mTBI, and the unique personality and behavioral characteristics that might predispose a child to injury (25,26). Thus, our secondary objective is to investigate the benefits of sports participation among children aged 9–10 who had a history of mTBI versus who had a history of OI or whom were uninjured.
Materials and methods
ABCD study design
The ABCD study (https://nda.nih.gov/abcd) is a large-scale, longitudinal study of adolescent development. From 2016 to 2018, a cohort of 11,878 children aged 9–10 were enrolled from 21 U.S. sites and have since been assessed annually or biannually through biological and behavioral evaluations, neuroimaging and neurocognitive testing, and youth and parent assessments on mental health, substance use, physical health, and cultural and environmental factors (19). The current investigation is cross-sectional, leveraging baseline data from ABCD study release 5.1, including 11,868 participants. The University of California, San Diego Institutional Review Board approved all aspects of this study for the ABCD consortium. Parental consent and adolescent assent were obtained prior to participating in the study.
Sports participation
At the baseline visit, caregivers reported their child’s sports participation by the question: ‘Please indicate whether your child has ever participated in any of the following sports and activities continuously for 4 months or more.’ A total of 23 distinct sports were assessed (listed in Figure 1).
Figure 1.
Association between individual sports and mTBI versus OI. Odds ratio (95% confidence interval) from multi-predictor logistic regression, including individual sport participation, adjusting for demographic variables (age, sex, race/ethnicity, parental income, parental education, CBCL internalizing, and externalizing T-scores). The analysis sample included only children who had participated in at least one sport. Each individual sport was fitted using a multi-predictor logistic regression model adjusting for all the age, sex, race/ethnicity, parental education, parental income, and internalizing and externalizing behaviors, for mTBI or OI as the outcome separately. The 95% confidence intervals and p-values were computed by likelihood ratio tests. P-values were adjusted for multiple comparisons using the Benjamini–Hochberg (31) procedure, applied separately within sports of different prevalence levels at <1%, 1%–10%, >10%. Significant p-values were highlighted: *< .05; **< .01; ***< .001. bthe sport had a significantly higher risk for both mTBI and OI; tthe sport had a significantly higher risk for mTBI; othe sport had a significantly higher risk for OI.
Mild traumatic brain injury (mTBI)
Parent-reported head injuries were assessed through a modified Ohio State Traumatic Brain Injury Screen-Short Form (27,28). Caregivers answered nine yes/no questions regarding situations their children might have encountered that could result in a concussion/TBI. Follow-up questions assessed loss of consciousness, alteration of consciousness, amnesia, and the child’s age at the time of injury. Based on this information, TBI was categorized into five categories: Improbable TBI (no TBI or TBI without LOC or memory loss); Possible mild TBI (TBI without LOC but memory loss); Probable mild TBI (TBI with LOC ≤30 min); Moderate TBI (TBI with LOC 30 min – 24 hours); Severe TBI (TBI with LOC ≥24 hours). Possible and mild TBI were combined to form the mTBI group for this analysis (5). Children with moderate-to-severe TBI (n = 7) were excluded in all analyses. Our definition of mTBI was limited by the categories available in the data. If there was no loss of consciousness, memory loss was required for an mTBI classification. This criterion likely led to an underestimation of mTBI incidence, particularly in cases where other concussion symptoms, such as nausea, vomiting, double vision, and photophobia, were present.
Orthopedic injury (OI)
Parents reported their child’s orthopedic injury history by answering the question ‘Has she/he ever been to a doctor for broken bones?’ from the medical history questionnaire.
Demographics and psychosocial measures
Demographic information included age at baseline, sex assigned at birth, race/ethnicity, highest level of parental income, and highest level of parental education. We used ABCD’s range of income and education and both were treated as continuous variables (5). Income ranges from 1 to 10, where 1 = Less than $5,000; 2=$5,000 through $11,999; 3=$12,000 through $15,999; 4=$16,000 through $24,999; 5=$25,000 through $34,999; 6=$35,000 through $49,999; 7=$50,000 through $74,999; 8 = $75,000 through $99,999; 9=$100,000 through $199,999; 10=$200,000 and greater. Education ranges from 0 to 21, where 0 = Never attended/Kindergarten; 1 through 12 corresponds to 1st through 12th grade; 13 = High school graduate; 14 = GED or equivalent Diploma; 15 = Some college; 16 = Associate’s degree: Occupational; 17 = Associate’s degree: Academic Program; 18 = Bachelor’s degree; 19 = Master’s degree; 20 = Professional School degree; 21 = Doctoral degree.
Physical health
Children’s weight, height, and waist circumference were collected at their baseline visit. Body mass index (BMI) was defined as weight/height2.
Behavioral and emotional assessments
The Child Behavior Checklist (CBCL), completed by caregivers, evaluates children’s emotional and behavioral problems (higher scores indicate worse behaviors). Internalizing T-scores (CBCL-I), externalizing T-scores (CBCL-E), and CBCL DSM-5 Depression Scale T-scores were used in this analysis (29).
Neurocognitive performance
The NIH Toolbox Cognitive Battery assessment (https://nihtoolbox.org/) was administered to evaluate children’s neurocognitive performance. The uncorrected composite scores of crystallized cognition and fluid cognition were used in the analysis (30). Crystallized cognition reflects acquired knowledge and past learning experiences, while fluid cognition measures the problem-solving skills and ability to process new information (30).
Statistical analyses
We reported descriptive statistics of demographic variables and the prevalence of mTBI and OI, by sports participation. Group differences (participated in at least one sport versus never participated in sports) of demographic variables were tested using Pearson’s chi-squared test for categorical variables and Welch Two Sample t-test for continuous variables.
Primary analysis
As in our previous paper, we only included children who participated in at least one sport in our primary analysis of sports risks associated with mTBI versus OI in order to focus our analysis on a more comparable population (5). We fitted multi-predictor logistic regression models separately for mTBI or OI as the outcomes, where predictors were each of the sports individually (participation yes/no), adjusting for age, race/ethnicity, parental income, parental education, and internalizing and externalizing behaviors. The p-values and confidence intervals (CI) were computed by likelihood ratio tests. P-values were adjusted for multiple comparisons using the Benjamini–Hochberg (31) procedure, applied separately within sports of different prevalence levels at <1%, 1%–10%, >10%.
Generalized linear mixed-effects models with random intercepts were used to test whether the risks of sports for mTBI versus OI were different. In this model, each subject contributed to two rows of data, one for mTBI and one for OI, with a binary outcome indicating whether they were injured or not. Random intercept was used to account for repeated measure on the same individual. To illustrate the data structure, we provided an example dataset in Table S1, supplement file. The main predictors in this model included injury type (mTBI or OI), participation in individual sports, and sport-by-injury-type interaction terms. Since there were a total of 23 sports, to determine which sport-by-injury-type interaction term to include in our final model, we first fitted partially adjusted models with each model including one sport, the sport-by-injury-type interaction term, and all confounders as predictors. We then fitted a fully adjusted model with all sports as separate predictors, the sport-by-injury-type interaction terms (only those with a p-value <0.05 from the partially adjusted models were included here), and all confounders. Backward model selection was performed on the individual sports variables at p < 0.20 using the likelihood ratio test (32).
Secondary analysis
We performed principal component analysis (PCA) on three health-related domains – physical health, behavior/emotion, and neurocognition, and used the first principal component (PC1) from each domain as the outcome variables in our secondary analysis on benefits of sports participation. Specifically, physical health was represented by the PC1 of BMI and waist circumference. Behavior/emotion was represented by the PC1 of the CBCL internalizing T-score, CBCL externalizing T-score, and CBCL DSM-5 depression scale. Neurocognition was represented by the PC1 of the NIH Toolbox fluid cognition uncorrected scores and NIH Toolbox crystallized cognition uncorrected scores. This approach was consistent with our previous study (5). All children (including those who participated and never participated in sports) were categorized into three injury categories: mTBI, OI, and non-injury (NI). There were no children who had a history of both mTBI and OI. We fitted linear regression models with the three PC1 as the outcomes. The main predictors were the three-level categorical variables indicating the injury history (mTBI, OI, or NI), participation in sport (yes/no), and a sports-by-injury-category interaction term, adjusting for age, race/ethnicity, parental income, and parental education.
Only children with no missing covariate information were included in all analyses. All analyses were done using R version 4.3.1 and Rstudio.
Results
Descriptive statistics
The main analysis sample included 11,055 children with complete covariate information, from the 11,868 total sample size. Among the 11,055 children, 9,423 (80%) had participated in any sports, with 378 (4.0%) having a history of mTBI and 1,375 (15%) having a history of OI (Table 1). The rate of OI history was significantly higher among children playing sports, 15% versus 9.3%, p < 0.001 (Table 1). Of children playing sports, there were 47% female, 57% White, 12% Black, 19% Hispanic, 2.1% Asian, and 10% other races; the average age was 9.93, standard deviation (SD) = 0.62; the average parental education was 17.48 (SD = 2.39), equivalent to an Associate degree; the average parental income was 6.91 (SD = 2.29), where a seven on income was equivalent to an annual income of $50,000 through $74,999; the average CBCL internalizing and externalizing T-score was 48.34 (SD = 10.46), 45.39 (SD = 10.07), respectively. As we noted in our previous paper, children who had never played sports had significantly lower parental income and education, as well as higher internalizing and externalizing T-scores (5).
Table 1.
Sample demographics of ABCD baseline data.
| Participated in at least one sporta N = 9,423b |
Never participated in sports N = 1,632 |
P-valuee | |
|---|---|---|---|
| Mild traumatic brain injury | 378 (4.0%) | 50 (3.1%) | 0.078 |
| Orthopedic injury | 1,375 (15%) | 152 (9.3%) | <0.001*** |
| Sex | 0.005** | ||
| Female | 4,460 (47%) | 835 (51%) | |
| Male | 4,963 (53%) | 797 (49%) | |
| Race | <0.001*** | ||
| White | 5,415 (57%) | 543 (33%) | |
| Black | 1,100 (12%) | 459 (28%) | |
| Hispanic | 1,750 (19%) | 419 (26%) | |
| Asian | 195 (2.1%) | 24 (1.5%) | |
| Other | 963 (10%) | 187 (11%) | |
| Age | 9.93 (0.62) | 9.85 (0.62) | <0.001*** |
| Parental education c | 17.48 (2.39) | 15.46 (2.83) | <0.001*** |
| Parental income d | 6.91 (2.29) | 4.89 (2.45) | <0.001*** |
| CBCL internalizing (T-scores) | 48.34 (10.46) | 49.19 (11.15) | 0.004** |
| CBCL externalizing (T-scores) | 45.39 (10.07) | 47.17 (11.13) | <0.001*** |
Only these children who participated in at least one sport were included in the main analysis. Children who never participated in sports were included in the analysis of the benefits of sports.
N = sample size of baseline complete data; n (%); Mean (SD).
Education ranges from 0 to 21, where 0 = Never attended/Kindergarten; 1 through 12 corresponds to 1st through 12th grade; 13 = High school graduate; 14 = GED or equivalen3t Diploma; 15 = Some college; 16 = Associate degree: Occupational; 17 = Associate degree: Academic Program; 18 = Bachelor’s degree (ex. BA; 19 = Master’s degree (ex. MA; 20 = Professional School degree (ex. MD; 21 = Doctoral degree (ex. PhD.
Income ranges from 1 to 10, where 1 = Less than $5,000; 2=$5,000 through $11,999; 3=$12,000 through $15,999; 4=$16,000 through $24,999; 5=$25,000 through $34,999; 6=$35,000 through $49,999; 7=$50,000 through $74,999; 8= $75,000 through $99,999; 9=$100,000 through $199,999; 10=$200,000 and greater.
Pearson’s chi-squared test for categorical variables; Welch Two Sample t-test for continuous variables.
Risks of individual sports in mTBI versus OI
This analysis sample included 9,423 children aged 9–10 who had participated in at least one sport. The risks of mTBI/OI are summarized in Figure 1, which presents adjusted odds ratios and 95% confidence intervals in parentheses for each sport in relation to mTBI and OI. Adjusting for confounders and multiple comparisons, children who played soccer had a higher risk of mTBI, adjusted odds ratio (aOR) = 1.462 (1.179, 1.817), p = 0.004, and a higher risk of OI, aOR = 1.155 (1.026,1.299), p = 0.045, compared to children who had never played soccer (95% confidence interval in parentheses; Figure 1). Children who played American football had a higher risk of mTBI compared to children who have never played American football, aOR = 1.471 (1.102, 1.935), p = 0.038. Children who played baseball or basketball had a higher risk of OI compared to children who had never played baseball or basketball, aOR = 1.203 (1.066, 1.358), p = 0.013, and aOR = 1.204 (1.064, 1.361), p = 0.013, respectively.
The estimates from the partially adjusted generalized linear mixed-effects models are presented in Table S2, supplement file. Among all sports examined, only climbing showed a significant interaction with injury type in the partially adjusted models. Therefore, in the final model (Table 2), we included only the climbing-by-injury-type interaction term. All other sports were included as covariates to account for participation in different sports, without interaction terms. Based on the final model, compared to children who had never participated in climbing, children who had participated in climbing had a significantly higher risk of mTBI than OI, the ratio of odds ratio (ROR) = 1.881 (1.140, 3.103), p = 0.013 (Table 2).
Table 2.
The risk of individual sports on mTBI compared to OI.
| Odds Ratio (95% CI) | p-value | |
|---|---|---|
| Injury type (mTBI):Climbinga | 1.881 (1.140, 3.103) | 0.013* |
| Injury type (mTBI) | 0.233 (0.207, 0.264) | <0.001*** |
| American football | 1.163 (0.993, 1.363) | 0.061 |
| Soccer | 1.191 (1.071, 1.324) | 0.001** |
| Baseball | 1.123 (1.002, 1.258) | 0.045* |
| Basketball | 1.095 (0.976, 1.229) | 0.122 |
| Ice Skating | 0.873 (0.712, 1.071) | 0.194 |
| Climbing | 1.672 (1.071, 2.512) | 0.025* |
| Horse/Polo | 1.218 (0.972, 1.528) | 0.087 |
| Climbing | 0.932 (0.694, 1.252) | 0.64 |
| Volleyball | 1.234 (0.944, 1.613) | 0.125 |
| Age | 1.199 (1.106, 1.300) | <0.001*** |
| Sex (Male) | 0.972 (0.866, 1.09) | 0.625 |
| Race (Black) | 0.446 (0.355, 0.561) | <0.001*** |
| Race (Hispanic) | 0.851 (0.734, 0.987) | 0.033* |
| Race (Asian) | 0.644 (0.427, 0.971) | 0.036* |
| Race (Other) | 0.846 (0.711, 1.006) | 0.058 |
| Parental education | 1.018 (0.989, 1.048) | 0.228 |
| Parental income | 1.021 (0.99, 1.053) | 0.188 |
| CBCL internalizing T-score | 1.010 (1.004, 1.016) | 0.001** |
| CBCL externalizing T-score | 1.002 (0.996, 1.008) | 0.555 |
Odds ratio (95% confidence interval) from a multi-predictor generalized linear mixed-effects models, including injury type (mTBI/OI), all the individual sports at p < 0.20 from backward model selection, their interaction terms, and demographic variables.
The climbing-by-injury-type interaction term was included in this model because this term was significant from the partially adjusted models. All other sport-byinjury-type terms were not significant from partially adjusted models, so they were included as covariates to account for participation in different sports, without interaction terms, see supplement file Table S2 for all partially adjusted models. The analysis sample included only children who had participated in at least one sport. The odds ratios were estimated from a generalized linear mixed-effects model with random intercepts. The binary outcome was injured or not. Each subject had two rows of data, one indicating the injury history of mTBI and the other indicating the injury history of OI. The predictors included injury type (mTBI or OI) and a climbing-by-injury-type interaction, adjusting for all confounders and individual sports at p < 0.20 from a backward model selection using likelihood ratio tests.
Significant p-values were highlighted: *<.05; **< .01; ***< .001.
Benefits of playing sports among mTBI, OI, and NI children
This analysis sample included all children. The PCA loadings and proportion of variance explained were presented in Table S3-S5, supplement file. Based on the PCA loadings, higher physical health PC1 scores are associated with lower BMI and waist circumference. Higher behavior/emotion PC1 scores correspond to higher CBCL scores, indicating more behavioral and emotional problems. Higher neurocognition PC1 scores correspond to lower NIH Toolbox scores (5).
Figure 2 shows the distribution of physical health, behavior/emotion, and neurocognition PC scores across six groups defined by injury status (mTBI, OI, or NI) and sports participation (yes/no). In the physical health panel, PC scores appear similar across all groups. In the behavior/emotion panel, children with mTBI who did not participate in any sport exhibited worse behavioral/emotional problems than other groups. In the neurocognition panel, children who played sports, regardless of injury status, had better neurocognition PC scores. The estimated effects of sports and injury on the three domains from linear regression models were presented in Table 3, where complete model estimates can be found in Table S6-S8, supplement file.
Figure 2.
Boxplot of physical health, behavior/emotion, and neurocognition by sports/injury groups: no sports and mTBI (NSTBI), play sports and mTBI (PSTBI), no sports and OI (NSOI), play sports and OI (PSOI), no sports and no injury (NSU), play sports and no injury (PSU). For this boxplot, all children were categorized into six mutually exclusive groups based on their participation in sports and injury history: never played sports and had a history of mTBI (NSTBI), played sports and had a history of mTBI (PSTBI); never played sports and had a history of mTBI (NSOI), played sports and had a history of mTBI (PSOI); never played sports and uninjured with mTBI or OI (NSU), and played sports and uninjured with mTBI or OI (PSU). There were no children who had a history of both mTBI and OI. The outcomes of physical health, behavior/emotion, and neurocognition were derived from the first principal component (PC1) of principal component analyses (PCA) conducted on relevant measures within each domain. Specifically, physical health was derived from the PC1 of BMI and waist circumference; behavior/emotion was derived from the PC1 of CBCL internalizing T-score, CBCL externalizing T-score, and CBCL DSM-5 depression scale; neurocognition was derived from the PC1 of NIH Toolbox fluid cognition uncorrected scores and NIH Toolbox crystallized cognition uncorrected scores.
Table 3.
Effects of sports participation and child’s injury history on physical health, behavior/emotion, and neurocognition.
| Outcomes Predictors |
Physical Health |
Behavior/Emotion |
Neurocognition |
|||
|---|---|---|---|---|---|---|
| Estimatesa (95% CI) | p-value | Estimates (95% CI) | p-value | Estimates (95% CI) | p-value | |
| NI vs. mTBIb | 0.14 (−0.23, 0.5) | 0.468 | −1.01 (−1.44, −0.59) | <0.001*** | −0.12 (−0.4, 0.16) | 0.404 |
| OI vs. mTBI | −0.06 (−0.47, 0.36) | 0.795 | −0.95 (−1.43, −0.47) | <0.001*** | −0.25 (−0.57, 0.07) | 0.119 |
| Sportsc vs. No sports, mTBI | 0.14 (−0.24, 0.53) | 0.468 | −0.55 (−0.99, −0.11) | 0.014* | −0.26 (−0.55, 0.04) | 0.088 |
| Sports vs. No sports, NI vs. mTBId | −0.13 (−0.52, 0.26) | 0.507 | 0.45 (0.01, 0.90) | 0.048* | 0.12 (−0.18, 0.42) | 0.432 |
| Sports vs. No sports, OI vs. mTBId | −0.01 (−0.46, 0.43) | 0.947 | 0.42 (−0.08, 0.93) | 0.102 | 0.26 (−0.08, 0.6) | 0.132 |
Differences in means (95% confidence intervals) were estimated from multi-predictor linear regression, including group (mTBI/OI/NI controls), sports participation, their interaction, adjusting for demographic variables (age, sex, race/ethnicity, parental income, and parental education). The outcomes of physical health, behavior/emotion, and neurocognition were derived from the first principal component (PC1) of principal component analyses (PCA) conducted on relevant measures within each domain. Specifically, physical health was derived from the PC1 of BMI and waist circumference; behavior/emotion was derived from the PC1 of CBCL internalizing T-score, CBCL externalizing T-score, and CBCL DSM-5 depression scale; neurocognition was derived from the PC1 of NIH Toolbox fluid cognition uncorrected scores and NIH Toolbox crystallized cognition uncorrected scores. See supplement file Table S3-S5 for PCA loadings and proportion of variance explained.
Estimated from linear regression models adjusting for age, sex, race/ethnicity, parental income, and parental education. The analysis sample included all children. See supplement file Table S6-S8 for complete model estimates.
The injury type categorical variable had three levels: OI means children who had a history of orthopedic injury; mTBI (the reference level) means children who had a history of mild traumatic brain injury; NI means children who had no history of OI or mTBI.
Children having participated in at least one sport (yes/no).
This is an interaction coefficient, corresponding to a difference in differences of means.
Significant p-values were highlighted: *< .05; **< .01; ***< .001.
Physical health
There were no significant differences in physical health associated with sports participation and injury history (Figure 2, physical health). Among children not playing sports, the physical health PC1 scores of NI vs. mTBI, estimated mean differences (Δ) = 0.14 (−0.23, 0.5), p = 0.469, and OI vs. mTBI, Δ = −0.06 (−0.47, 0.36), p = 0.795, did not show significant differences (Table 3). The interaction effects between sports and injury groups were also not significant, p = 0.507 for NI vs. mTBI and p = 0.947 for OI vs. mTBI.
Behavior/Emotion
Among children who did not play sports, those with a non-injury history (NSU/light gray bar in Figure 2) or with an OI history (NSOI/light orange bar in Figure 2) had lower PC1 scores than those who had a history of mTBI (NSTBI/light blue bar in Figure 2), Δ = −1.01 (−1.44, −0.59), p < 0.001, and −0.95 (−1.43, −0.47), p < 0.001, respectively (Table 3). Among children with a history of mTBI, those who had played sports had significantly lower PC1 scores, Δ = −0.55 (−0.99, −0.11), p = 0.014 (Table 3; PSTBI/dark blue bar versus NSTBI/light blue bar in Figure 2, behavior/emotion). The difference in PC1 scores associated with playing sports was significantly smaller for the children with non-injury history, compared to the children with a mTBI history, Δ = 0.45 (0.01, 0.90), p = 0.048 (Table 3; NSTBI/light blue bar versus PSTBI/dark blue bar compared to NSU/light gray bar versus PSU/dark gray bar in Figure 2, behavior/emotion).
Neurocognition
There were no significant differences in neurocognition associated with sports participation and injury history (Figure 2, neurocognition). The interaction effects between sports and injury groups were not significant, p = 0.432 for NI vs. mTBI and p = 0.132 for OI vs. mTBI.
Discussion
Youth sports participation provides numerous benefits but also leads to risks of injury, such as brain injury or fractures. While prior research found that certain contact and collision sports are associated with higher concussion risks in adolescent and high school athletes, less is known about injury risks in prepubertal children (ages 9–10), whose developing status and risk-awareness may contribute to different injury patterns. Additionally, while mTBI in sports has been widely studied, fractures and other orthopedic injuries also account for a large proportion of sports-related traumatic injuries, yet the risks of mTBI versus OI across different sports remain unclear.
Using the baseline ABCD data, we previously found participation in soccer or climbing associated with higher mTBI risks among children aged 9–10 (5). In this paper, we extended our study to compare children injured with OI versus uninjured with OI, where we found that participating in soccer, baseball, or basketball was associated with a higher OI risk, compared to not participating in these sports. These findings on OI were consistent with some existing literature on children and adolescent sports-related fractures, where basketball was among the most common sports associated with fractures (24,33). We used generalized linear mixed-effects models to assess whether specific sports were associated with a different risk of mTBI versus OI. Children who had participated in climbing had a significantly higher risk of mTBI than OI. These findings contrast some existing literature where fractures and sprains and strains were the most common injuries in climbing, but studies on mTBI and climbing were limited, especially among youth (34,35). Our results implied that sports-related injury prevention among children should emphasize sport-specific strategies according to the distinct risks of mTBI and OI. It is also important to note that the cross-sectional nature of the data did not contain information of whether injuries occurred during sports participation. Therefore, while we report significant association between sports and injury type, these associations did not necessarily mean causality.
Sports participation was associated with improved neurocognitive performance (higher NIH Toolbox fluid composite scores) and fewer behavioral and emotional problems (lower CBCL internalizing, externalizing, and total problem T-scores), regardless of children’s mTBI history (5). In this study, we extended the PC analysis to add OI as a control group for mTBI, and found that the benefits of playing sports versus not playing sports in behavior/emotion were significantly larger for the children with mTBI history, compared to the children with no injury history. However, there were no significant differences in the benefits of sports between children with mTBI and children with OI, suggesting that the benefits of playing sports may be similar for both injury groups. These findings highlight the importance of using OI as a control group in mTBI studies (25,26), as it allows for an investigation of the effects of sports participation while accounting for the general influences of sustaining a traumatic injury.
Limitations
As with our previous analyses, this analysis used the ABCD baseline data, so it was retrospective and only applied to children aged 9–10. In this sample, children who had played sports showed different sociodemographic characteristics compared to children who had never played sports. For the primary analysis, we excluded children not playing sports to create a more comparable population and controlled for socioeconomic factors in all analyses. A further limitation is that there were no specific data on the time of injury or the cause of injury. In addition, injury history, the severity of injury, and sports participation were collected from parent reports, which may introduce recall bias and fail to differentiate levels of sports participation. Because of the reliance on parent report of LOC, concussions only with cognitive or vestibular symptoms may have been missed, which likely resulted in an underestimation of mTBI incidence. Therefore, it is important to note that our findings reflect the association between sports participation and injury history, rather than evidence that playing sports caused injury. Nevertheless, this cross-sectional study used data from the largest cohort study of children’s health and development in the U.S., allowing us to examine the patterns of mTBI and OI across a wide range of sports. Our methods enable us to quantify the sports-specific injury associations by comparing children who played a specific sport to those who did not, while adjusting for sociodemographic confounders. This analytical approach provides useful perspective on potential sport-specific differences in injury risk among children.
Conclusion
Different types of sports were associated with risk for mTBI and risk for OI in children aged 9–10: children who played soccer had higher mTBI and OI risks compared to those who did not; children who had participated in basketball or baseball had higher OI risks. Moreover, compared to children who had never participated in climbing, those who climbed were exposed to higher mTBI risks than OI risks. The benefits of sports participation in children’s behavior/emotion were larger among children with mTBI compared to children with no injury history, but no significant differences in benefits provided by sports were found between children with mTBI and children with OI. Our results imply sports-specific and injury-specific strategies to reduce risks of mTBI and OI in youth sports. The results do not necessarily apply to every individual child, and individual risk factors such as family history of psychological or behavioral issues are important to consider when deciding the risks and benefits of children’s sports participation.
Supplementary Material
Supplemental data for this article can be accessed online at https://doi.org/10.1080/02699052.2025.2553324
Funding
This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) [RO1HD105338].
Footnotes
Disclosure statement
Dr. Max provides expert testimony in cases of traumatic brain injury on an ad hoc basis for plaintiffs and defendants on a more or less equal ratio. This activity constitutes approximately 5% of his professional activities. No competing financial interests exist for the other authors.
References
- 1.Eime RM, Young JA, Harvey JT, Charity MJ, Payne WR. A systematic review of the psychological and social benefits of participation in sport for children and adolescents: informing development of a conceptual model of health through sport. Int J Behav Nutr Phys Act. 2013;10(1):98. doi: 10.1186/1479-5868-10-98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Vella SA, Cliff DP, Magee CA, Okely AD. Sports participation and parent-reported health-related quality of life in children: longitudinal associations. J Pediatr. 2014;164(6):1469–74. doi: 10.1016/j.jpeds.2014.01.071. [DOI] [PubMed] [Google Scholar]
- 3.Lee JE, Pope Z, Gao Z. The role of youth sports in promoting children’s physical activity and preventing pediatric obesity: a systematic review. Behav Med. 2018;44(1):62–76. doi: 10.1080/08964289.2016.1193462. [DOI] [PubMed] [Google Scholar]
- 4.Merkel DL. Youth sport: positive and negative impact on young athletes. Open Access J Sports Med. 2013;4:151–60. doi: 10.2147/OAJSM.S33556. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Meng W, Vaida F, Dennis EL, Wilde EA, Jacobus J, Yang X, Cheng M, Troyer EA, Delfel EL, Abildskov T, et al. Boys versus girls youth sports participation risk for mild traumatic brain injury and behavioral, physical health, and neurocognitive outcomes. J Head Trauma Rehabil. 2025. doi: 10.1097/HTR.0000000000001065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Brenner JS. Medicine atCos, fitness. Overuse injuries, overtraining, and burnout in child and adolescent athletes. Pediatrics. 2007;119(6):1242–45. doi: 10.1542/peds.2007-0887. [DOI] [PubMed] [Google Scholar]
- 7.Pfister T, Pfister K, Hagel B, Ghali WA, Ronksley PE. The incidence of concussion in youth sports: a systematic review and meta-analysis. Br J Sports Med. 2016;50(5):292–97. doi: 10.1136/bjsports-2015-094978. [DOI] [PubMed] [Google Scholar]
- 8.Bryan MA, Rowhani-Rahbar A, Comstock RD, Rivara F. Sportsand recreation-related concussions in US youth. Pediatrics. 2016;138(1). doi: 10.1542/peds.2015-4635. [DOI] [PubMed] [Google Scholar]
- 9.Taylor CA, Bell JM, Breiding MJ, Xu L. Traumatic brain injury-related emergency department visits, hospitalizations, and deaths - United States, 2007 and 2013. MMWR Surveill Summ. 2017;66(9):1–16. doi: 10.15585/mmwr.ss6609a1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Silverberg ND, Iverson GL. Expert panel survey to update the American Congress of rehabilitation Medicine definition of mild traumatic brain injury. Arch Phys Med Rehabil. 2021;102(1):76–86. doi: 10.1016/j.apmr.2020.08.022. [DOI] [PubMed] [Google Scholar]
- 11.Lumba-Brown A, Yeates KO, Sarmiento K, Breiding MJ, Haegerich TM, Gioia GA, Turner M, Benzel EC, Suskauer SJ, Giza CC, et al. Centers for disease control and prevention guideline on the diagnosis and management of mild traumatic brain injury among children. JAMA pediatr. 2018;172(11):e182853. doi: 10.1001/jamapediatrics.2018.2853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kay T, Harrington DE, Adams R, Anderson T, Berrol S, Cicerone K, et al. Definition of mild traumatic brain injury. J Head Trauma Rehabil. 1993;8(3):86–87. doi: 10.1097/00001199-199309000-00009. [DOI] [Google Scholar]
- 13.Gagner C, Landry-Roy C, Bernier A, Gravel J, Beauchamp MH. Behavioral consequences of mild traumatic brain injury in preschoolers. Psychol Med. 2018;48(9):1551–59. doi: 10.1017/S0033291717003221. [DOI] [PubMed] [Google Scholar]
- 14.Max JE, Judd N, Bigler ED, Wilde EA, Patterson JE, Edwards TM, Calahorra A, De La Garza BG, Vaida F. Three-month psychiatric outcome of pediatric mild traumatic brain injury: a controlled study. J Neurotrauma. 2021;38(23):3341–51. doi: 10.1089/neu.2021.0324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Loher S, Fatzer ST, Roebers CM. Executive functions after pediatric mild traumatic brain injury: a prospective short-term longitudinal study. Appl Neuropsychol Child. 2014;3(2):103–14. doi: 10.1080/21622965.2012.716752. [DOI] [PubMed] [Google Scholar]
- 16.Rivara FP, Tennyson R, Mills B, Browd SR, Emery CA, Gioia G, Giza CC, Herring S, Janz KF, LaBella C, et al. Consensus statement on sports-related concussions in youth sports using a modified Delphi approach. JAMA pediatr. 2020;174(1):79–85. doi: 10.1001/jamapediatrics.2019.4006. [DOI] [PubMed] [Google Scholar]
- 17.Marar M, McIlvain NM, Fields SK, Comstock RD. Epidemiology of concussions among United States high school athletes in 20 sports. Am J Sports Med. 2012;40(4):747–55. doi: 10.1177/0363546511435626. [DOI] [PubMed] [Google Scholar]
- 18.O’Connor KL, Baker MM, Dalton SL, Dompier TP, Broglio SP, Kerr ZY. Epidemiology of sport-related concussions in high School athletes: National athletic treatment, injury and outcomes network (NATION), 2011-2012 through 2013-2014. J Athl train. 2017;52(3):175–85. doi: 10.4085/1062-6050-52.1.15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Barch DM, Albaugh MD, Avenevoli S, Chang L, Clark DB, Glantz MD, et al. Demographic, physical and mental health assessments in the adolescent brain and cognitive development study: rationale and description. Dev Cogn NeuroSci. 2018;32:55–66. doi: 10.1016/j.dcn.2017.10.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ukogu C, Patterson D, Sarosi A, Colvin AC. Epidemiology of youth sports injury: a review of demographic and sports-related risk factors for injury. Ann Jt. 2017;2:79–79. doi: 10.21037/aoj.2017.12.04. [DOI] [Google Scholar]
- 21.Lynch KR, Kemper HCG, Turi-Lynch B, Agostinete RR, Ito IH, Luiz-De-Marco R, Rodrigues-Junior MA, Fernandes RA. Impact sports and bone fractures among adolescents. J Sports Sci. 2017;35(24):2421–26. doi: 10.1080/02640414.2016.1272708. [DOI] [PubMed] [Google Scholar]
- 22.Volpi P, Pozzoni R, Galli M. The major traumas in youth football. Knee surgery, sports traumatology. Knee Surg. Sports Traumatol. Arthrosc 2003;11(6):399–402. doi: 10.1007/s00167-003-0343-1. [DOI] [PubMed] [Google Scholar]
- 23.Meyers RN, Hobbs SL, Howell DR, Provance AJ. Are adolescent climbers aware of the most common youth climbing injury safe training practices? Int J Environ Res Pub Health And Public Health. 2020;17(3):812. doi: 10.3390/ijerph17030812. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zynda AJ, Wagner KJ, Liu J, Chung JS, Miller SM, Wilson PL, Ellis HB. Epidemiology of pediatric basketball injuries presenting to emergency departments: sex- and age-based patterns. Orthop J Sports Med. 2022;10(1):23259671211066503. doi: 10.1177/23259671211066503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wilde EA, Ware AL, Li X, Wu TC, McCauley SR, Barnes A, et al. Orthopedic injured versus uninjured comparison groups for Neuroimaging research in mild traumatic brain injury. J Neurotrauma. 2019;36(2):239–49. doi: 10.1089/neu.2017.5513. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Emery CA, Barlow KM, Brooks BL, Max JE, Villavicencio-Requis A, Gnanakumar V, Robertson HL, Schneider K, Yeates KO. A systematic review of psychiatric, psychological, and behavioural outcomes following mild traumatic brain injury in children and adolescents. Can J Psychiatry. 2016;61(5):259–69. doi: 10.1177/0706743716643741. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Corrigan JD, Bogner J. Initial reliability and validity of the Ohio state University TBI identification method. J Head Trauma Rehabil. 2007;22(6):318–29. doi: 10.1097/01.HTR.0000300227.67748.77. [DOI] [PubMed] [Google Scholar]
- 28.Bogner JA, Whiteneck GG, MacDonald J, Juengst SB, Brown AW, Philippus AM, Marwitz JH, Lengenfelder J, Mellick D, Arenth P, et al. Test-retest reliability of traumatic brain injury outcome measures: a traumatic brain injury model systems study. J Head Trauma Rehabil. 2017;32(5):E1–e16. doi: 10.1097/HTR.0000000000000291. [DOI] [PubMed] [Google Scholar]
- 29.Achenbach TM. Achenbach System of empirically based assessment (ASEBA). In: Kreutzer J, DeLuca J Caplan B, editors. Encyclopedia of clinical neuropsychology. Cham: Springer International Publishing; 2017. p. 1–7. [Google Scholar]
- 30.Akshoomoff N, Beaumont JL, Bauer PJ, Dikmen SS, Gershon RC, Mungas D, Slotkin J, Tulsky D, Weintraub S, Zelazo PD, et al. VIII. NIH Toolbox cognition Battery (CB): composite scores of crystallized, fluid, and overall cognition. Monogr Soc Res Child Dev. 2013;78(4):119–32. doi: 10.1111/mono.12038. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B Stat Methodol. 1995;57(1):289–300. doi: 10.1111/j.2517-6161.1995.tb02031.x. [DOI] [Google Scholar]
- 32.Vittinghoff E, Glidden DV, Shiboski SC, McCulloch CE. Regression methods in biostatistics: linear, logistic, survival, and repeated measures models. (NY), NY (US): Springer Publishing Co; 2005. [Google Scholar]
- 33.Wang H, Liu H, Wu J, Li C, Zhou Y, Liu J, Ou L, Xiang L. Age, gender, and etiology differences of sports-related fractures in children and adolescents: a retrospective observational study. Med (Baltim). 2019;98(4):e13961. doi: 10.1097/MD.0000000000013961. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Woollings KY, McKay CD, Kang J, Meeuwisse WH, Emery CA. Incidence, mechanism and risk factors for injury in youth rock climbers. Br J Sports Med. 2015;49(1):44–50. doi: 10.1136/bjsports-2014-094067. [DOI] [PubMed] [Google Scholar]
- 35.Buzzacott P, Schöffl I, Chimiak J, Schöffl V. Rock climbing injuries treated in US emergency departments, 2008-2016. Wildern Environ Med. 2019;30(2):121–28. doi: 10.1016/j.wem.2018.11.009. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.