The association between concussion history and increased symptom severity reporting is independent of common medical comorbidities, personality factors, and sleep quality in collegiate athletes

Abstract

Objective:

We investigated the degree to which the association between history of concussion with psychological distress and general symptom severity is independent of several factors commonly associated with elevated symptom severity. We also examined whether symptom severity endorsement was associated with concussion injury specifically, or response to injury in general.

Setting:

Academic Medical Center

Participants:

Collegiate athletes (N=106; age M=21.37 ± 1.69 years; 33 female) were enrolled based on strict medical/psychiatric exclusion criteria.

Design:

Cross-sectional single-visit study. Comprehensive assessment, including semi-structured interviews to retrospectively diagnose the number of previous concussions, were completed. Single-predictor and step-wise regression models were fit to examine the predictive value of prior concussion and orthopedic injuries on symptom severity, both individually and controlling for confounding factors.

Main Outcome Measures:

Psychological distress was operationalized as Brief Symptom Inventory-18 Global Severity Index (BSI-GSI) ratings; concussion-related symptom severity was measured using the Sport Concussion Assessment Tool.

Results:

Controlling for baseline factors associated with the symptom outcomes (agreeableness, neuroticism, negative emotionality, and sleep quality), concussion history was significantly associated with psychological distress (B=1.25[.55]; p=.025, ΔR²=.034) and concussion-like symptom severity (B=.22[.08]; p=.005, ΔR²=.064), and accounted for a statistically significant amount of unique variance in symptom outcomes. Orthopedic injury history was not individually predictive of psychological distress (B= −.06[.53]; p=.905) or general symptom severity (B=.06[.08]; p=.427) and did not explain the relationship between concussion history and symptom outcomes.

Conclusions:

Concussion history is associated with subtle elevations in symptom severity in collegiate-aged, athletes; this relationship is independent of medical, lifestyle (i.e., sleep), and personality factors. Furthermore, this relationship is associated with brain injury (i.e., concussion), and is not a general response to injury history.

Introduction

Changes in neurobehavioral function later in life have been observed among older, former athletes with a history of cumulative concussion.^1–4 Growing evidence indicates neurobehavioral abnormalities associated with cumulative concussion can be detected in younger, active athletes. Across several independent samples of high school and/or collegiate athletes, higher levels of psychological distress or concussion-related symptoms were endorsed by athletes with a greater history of concussion.^5–11 Achieving a better understanding of the association between concussion history and symptom severity at younger ages will aid in further identification of the potential pathogenesis and progression of neurobehavioral changes throughout the lifespan within these samples.

While associations between greater concussion history and symptom severity have been reported in several samples of younger athletes, the relationship between the two may be confounded by several factors. These factors include comorbid conditions (e.g., sleep disorders), neurodevelopmental disorders (e.g., attention-deficit/hyperactivity disorder; ADHD), and psychiatric conditions (e.g., depression).^12–17 Additionally, sleep quality has been observed as directly influencing symptoms in athletes at baseline assessment.^12,18–20 Greater symptoms have also been reported as being associated with demographic (i.e., race and female sex) and personal factors (i.e., socioeconomic status; SES) in athletes.^6,17,21–23 Personality traits and dispositions, such as negative emotionality (NEM), have also been documented as being moderately correlated with measures of general distress in athletes at baseline and could theoretically influence other self-report measures, such as retrospectively assessed concussion history. Establishing elevated symptom severity endorsement as being distinctly associated with greater concussion history and not driven by the above factors would support the notion that adverse long-term outcomes reported in those with a history of concussion potentially reflect gradual processes that have been ongoing for many years prior to the manifestation of clinical impairment. Additionally, this will help inform clinicians when considering potential etiologies influencing symptom severity endorsement in young athletes.

There are parallel questions regarding whether increases in symptom reporting and distress observed in athletes with multiple prior concussions are due specifically to brain injury (i.e., concussion), or represent a general, non-specific response to cumulative physical injury of any kind. Higher prevalence of psychological distress has been observed beyond one year of injury among orthopedic injury patients.^24,25 Discerning elevations in symptom severity as being associated with prior concussion distinctly, rather than general history of all-cause injuries, is essential in advancing our knowledge of the early chronic sequelae associated with prior concussion.

Due to the above comorbidities and factors, achieving a better understanding of the association between prior concussion and elevated symptom severity in young, active athletes requires the completion of studies specifically designed to answer this and related questions in a samples of healthy (i.e., no confounding comorbidities) collegiate-aged athletes.⁹ The aims of the current study were to: 1) examine whether the association between history of concussion with psychological distress and general symptom severity remains after controlling for psychosocial factors, demographics, sleep quality, and personality traits that could jointly influence reported concussion history and psychological distress; 2) investigate whether greater distress and general symptom severity is associated with orthopedic injury history in a manner similar to prior concussion. As an exploratory aim, we investigated the associations of individual symptoms with concussion history and confounding factors to determine whether there are distinct symptoms associated with concussion history.

Material and Methods

Enrollment occurred between December 2018 and March 2020 as part of an ongoing project assessing the chronic effects of concussions and contact sport exposure (Project ChronEx). Inclusion criteria included sport participation within the last 12 months and being between the ages of 18 and 26 years. Strict exclusion criteria controlling for conditions identified as influencing symptom endorsement in athletes included: 1) self-reported diagnosis of neurodevelopmental (including ADHD), psychiatric, neuroendocrine, severe neurological, or severe autoimmune diseases; 2) prior moderate or severe traumatic brain injury; 3) problematic alcohol or substance use which was screened with the CAGE Questions Adapted to Include Drug Use screener (CAGE-AID; ≥2 items); 4) use of neuroactive medication; 5) concussion within 6 months or incomplete recovery from previous concussion; 6) contraindication to parent study procedures. The study met Institutional Review Board at the Medical College of Wisconsin and participants provided written informed consent.

Semi-structured interviews for prior concussion

Comprehensive information was collected to retrospectively diagnose and quantify prior concussions through a semi-structured interview (described elsewhere).⁹ Collected information for each injury included presence of post-traumatic amnesia, retrograde amnesia, altered mental status, witnessed or suspected loss of consciousness, and a symptom checklist to indicate the presence of symptoms (i.e., SCAT-5 symptom checklist). Altered mental status was operationally defined and described to participants as feeling “confused, dazed, stunned, see stars, or feel like you got your bell rung.” The number of prior concussions for each participant was determined based on retrospective diagnosis consistent with the American Congress of Rehabilitation Medicine (ACRM) criteria.²⁶ Given that prior studies have demonstrated a dose response relationship with prior concussion and symptom severity endorsement,^1,6,8,11 concussion history reflected the number of prior concussions reported by the participant and binned into four primary categories including 0, 1, 2, and 3+ based on the distribution of the sample (i.e., comparable size of concussion history groups). The determination to create the 3+ concussion group was based on the fact that this was the most parsimonious way to test the linear association of cumulative concussion history with symptom severity while trying to balance distribution across the sample. Specifically, only 6.6% of the sample reported more than 3 concussions. Capping prior concussion at 3 or more allowed for the formation of prior concussion groups that were broadly comparable in size (1 prior concussion =18%; 2 prior concussions =15.1%; 3+ prior concussions = 14.1%).

Confounding factors

Sleep quality was assessed using the Pittsburgh Sleep Quality Inventory (PSQI).²⁷ The PSQI generates a total sleep quality score that is derived from self-reported aspects of sleep quality over the previous month. Socioeconomic status (SES) was estimated using the modified Hollingshead four-factor index.²⁸ Demographic factors of interest included self-reported race and sex (for non-standardized SCAT scores). Due to the limited distribution of the sample, race was binarized as White and non-White.

Personality dimensions were measured using an abbreviated version of the 35-item Multidimensional Personality Questionnaire (MPQ) developed for a large epidemiological study.^29,30 The measure assessed the two higher-order personality dimensions of negative emotionality (NEM) and positive emotionality (PEM). NEM reflects a tendency toward aspects of unpleasant mood states and consists of lower-order traits including aggression, alienation, and stress reaction. PEM represents a general temperament toward characteristics such as enthusiasm and positive engagement, and consists of lower-order factors of achievement, social closeness, social potency, and wellbeing. Personality factors were also measured using the Big Five Inventory (BFI-10), reflecting distinct dimensions of personality including extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience.³¹ MPQ NEM, MPQ PEM, and the five dimensions of the BFI were examined as the personality-related independent variables of interest.

Orthopedic injury history

Prior orthopedic injuries were inquired about during the comprehensive interview. Participants were asked about the number of prior injuries sustained in and out of sport. Examples of orthopedic injuries include fractures, ligamentous tear or sprain, joint dislocations, and muscle/tendon injuries. Based on the sample distribution and to make prior orthopedic injury history as comparable to concussion history as possible, prior orthopedic injury history was binned into the following categories; 0, 1, 2, and 3+.

Outcome measures of distress and general symptoms severity

A comprehensive clinical outcome assessment was performed to examine the associations between prior concussion and clinical function, which have been previously reported.⁹ Psychological distress was measured using the Brief Symptom Inventory-18.³² Analyses of the BSI-18 emphasized The Global Severity Index BSI-GSI, a composite index that comprises three dimensions of psychological distress: anxiety, depression, and somatic symptoms. Given sex-related effects in symptom severity among injured athletes, raw scores on the BSI-GSI were converted to sex-based standardized T-scores derived from normative data collected in the general population (normative M=50, SD=10; higher scores reflect more psychological distress).²² The Sport Concussion Assessment Tool-5^th edition (SCAT) symptom checklist is a commonly used measure to assess common symptoms following sport-related concussion and as a means to track injury recovery.³³ The SCAT consists of 22-items associated with sport-related concussion that capture physical, cognitive, and mood symptoms. Due to the high positive skew and tendency toward lower scores, SCAT symptom severity scores were log transformed to better approximate a normal distribution.

Statistical analysis

Analyses were conducted using IBM SPSS Statistics, Version 24. First, in order to identify potential factors that could confound the association between concussion history and symptom severity endorsement, single-predictor linear models were fit to assess the unadjusted bivariate association between individual prospective confounding variables described above with the outcomes of psychological distress and concussion-like symptom severity. Second, following identification of potential confounds from single predictor models, step-wise linear regression models were then performed with potentially confounding factors (defined as variables significantly associated with the outcomes) entered in block 1, and concussion history entered as the second block entry. This approach allowed for investigation of the unique predictive value of concussion history, above and beyond confounding factors. Specifically, the incremental predictive value added by concussion history was quantified as the increase in variance explained in the outcome when adding prior concussion to the model (i.e., change in R²) in block 2. Multicollinearity of independent variables within step-wise model was assessed using the variance inflation factor (i.e., <5 deemed as acceptable).³⁴ To examine the association between concussion history and symptom severity outcomes as being independent from general injury history, step-wise linear regression models involved step 1 entry of confound factors identified in the above described single predictor analyses and orthopedic history, with secondary entry of concussion history. This allowed for the investigation of the unique associations of concussion history as independent from orthopedic history, above and beyond confounding factors.

As an exploratory aim, associations between concussion history and individual symptom item ratings on the SCAT and BSI were assessed through Spearman’s Rho (ρ) correlations. Similar Spearman’s Rho correlations were performed for confounding factors significantly associated with overall symptom severity scores on the SCAT and BSI. Statistical significance for all analyses was evaluated at the 0.05 level.

Results

Concussion history, confounding factors, and psychological distress

Demographic, injury, and symptom severity information for participants (N=106) is presented in Table 1. Greater concussion history was significantly associated with elevated psychological distress (BSI-GSI scores; B[SE]; B=1.67[.61]; p=.007; η_p²=.067). Single-predictor linear regression models indicated that greater psychological distress was associated with lower personality tendency toward agreeableness (B=−3.49[1.10]; p=.002; η_p²=.089), greater neuroticism (B=2.51[.76]; p=.001; η_p²=.096), and greater NEM (B=4.59[.84]; p<.001; η_p²=.224), as well as greater sleep disturbance (B=.93[.25]; p<.001; η_p²=.117; Table 2). Factors such as SES, race, premorbid intellectual function, and other personality factors were not associated with elevated psychological distress, ps >.05. Controlling for agreeableness, neuroticism, NEM, and sleep disturbance, history of concussion significantly predicted psychological distress and accounted for a statistically significant amount of unique variance in psychological distress (B=1.25[.55]; p=.025, ΔR²=.028; Table 3).

Table 1:

Sample characteristics

Demographic/History	Total
Total No.	106
Sex (No. Male)	73
Age M (SD)	21.37 (1.69)
Socioeconomic Status M (SD)	50.08 (9.48)
PSQI Score M (SD)	4.15 (2.65)
WTAR Std. Score M (SD)	108.03 (10.13)
Race (No.)
Black or African-American	8
White	97
Not Reported	1
Ethnicity (No.)
Hispanic or Latino	8
Not Hispanic or Latino	97
Unknown	1
Concussion History (No.)
0	55
1	20
2	16
3+	15
Orthopedic Injury History (No.)
0	22
1	25
2	20
3+	39
Personality Measures
BFI Extraversion M (SD)	3.58 (.89)
BFI Agreeableness M (SD)	4.22 (.61)
BFI Conscientiousness M (SD)	4.01 (.75)
BFI Neuroticism M (SD)	2.22 (.89)
BFI Openness M (SD)	3.5 (.77)
MPQ PEM M (SD)	.03 (.57)
MPQ NEM M (SD)	−.02 (.74)
Clinical Outcome Measures
BSI-GSI T-score M (SD)	41.13 (7.19)
SCAT median (IQR)	1 (0–4)

No. = number; M = mean; SD = standard deviation; IQR = interquartile range; WTAR = Wechsler’s Test of Adult Reading; PSQI = Pittsburgh Sleep Quality Inventory; BFI= Big Five Inventory; MPQ= Multidimensional Personality Questionnaire; BSI-GSI = Brief Symptom Inventory-General Symptom Index; SCAT= Sport Concussion Assessment Tool; Socioeconomic Status based upon Hollingshead Four-Factor Index²⁸

Table 2.

Single-predictor linear regression models for confounding factor selection

	BSI-GSI Total Score			SCAT Symptom Severity
Factor	B (SE)	p-value	ηp2	B (SE)	p-value	ηp2
Race	1.12 (2.66)	.676	0.002	0.09 (0.35)	0.790	0.001
Sex	0.98 (1.51)	.516	0.004	−0.13 (0.20)	0.531	0.004
SES	0.03 (0.07)	.696	0.001	<0.01 (0.01)	0.801	0.001
PSQI Score	0.93 (0.25)	<.001	0.117	0.09 (0.03)	0.009	0.064
BFI Extraversion	−0.80(0.79)	.309	0.010	−0.19 (0.10)	0.065	0.032
BFI Agreeable.	−3.49 (1.10)	.002	0.089	−0.42 (0.15)	0.005	0.075
BFI Conscient.	−1.69 (0.92)	.069	0.031	−0.24 (0.12)	0.053	0.035
BFI Neuroticism	2.51 (0.76)	.001	0.096	0.23 (0.10)	0.025	0.048
BFI Openness	−0.65 (0.91)	.478	0.005	−0.16 (0.12)	0.174	0.018
MPQ PEM	−1.62 (1.22)	.188	0.017	−0.32 (0.16)	0.044	0.038
MPQ NEM	4.59 (0.84)	<.001	0.224	0.27 (0.12)	0.027	0.046

BSI-GSI = Brief Symptom Inventory-General Symptom Index; SCAT= Sport Concussion Assessment Tool (log transformed); B=unstandardized beta; SE=standard error; η_p² = partial eta-squared; PSQI = Pittsburgh Sleep Quality Inventory; MPQ= Multidimensional Personality Questionnaire; BFI = Big Five Inventory, MPQ = Multidimensional Personality Questionnaire; Agreeable. = Agreeableness; Conscient. = Conscientiousness; PEM = positive emotionality; NEM = negative emotionality; SES = socioeconomic status.

Table 3.

Step-wise linear regression model of confound factors and concussion history with psychological distress

	Model F(df)	R2	B (SE)	β (95% CI)	p-value
Step 1	11.45(101)	0.285
BFI Agreeableness			−1.63 (1.05)	−0.14 (−3.71, 0.45)	0.123
BFI Neuroticism			1.74 (0.69)	0.22 (0.38, 3.11)	0.013
MPQ NEM			3.12 (0.93)	0.32 (1.27, 4.96)	0.001
PSQI Score			0.42 (0.24)	0.15 (−0.07, 0.91)	0.092
	Model F(df)	R2*	B (SE)	β (95% CI)	p-value
Step 2	10.58(100)	0.313
BFI Agreeableness			−1.44 (1.03)	−0.12 (−3.48, 0.60)	0.165
BFI Neuroticism			1.96 (0.68)	0.24 (0.61, 3.31)	0.005
MPQ NEM			3.05 (0.91)	0.31 (1.24, 4.85)	0.001
PSQI Score			0.29 (0.25)	0.11 (−0.20, 0.78)	0.242
Concussion History			1.25 (0.55)	0.19 (0.16, 2.33)	0.025

B=unstandardized beta; SE=standard error; β = standardized beta coefficient; CI=Confidence interval; PSQI = Pittsburgh Sleep Quality Inventory; MPQ= Multidimensional Personality Questionnaire; PEM = positive emotionality; NEM = negative emotionality

^*= change in Adjusted R² statistically significant

Concussion history, confounding factors, and concussion symptom severity

Similar to psychological distress, greater concussion history was significantly associated with higher levels of log-transformed concussion-like symptom severity (B=.26[.08]; p=.001; η_p²=.096). Independent single-predictor linear regression models indicated that concussion-like symptom severity was associated with lower trait agreeableness (B=−.42[.15]; p=.005; η_p²=.075), higher trait neuroticism (B=.23[.10]; p=.025; η_p²=.048), higher trait NEM (B=.27[.12]; p=.027; η_p²=.046), and more sleep disturbance (B=.93[.25]; p<.001; η_p²=.064; Table 2). In addition to the above factors that were associated with psychological distress, PEM was also inversely associated with concussion-like symptom severity (B=−.32[.16]; p=.044; η_p²=.038). Controlling for agreeableness, neuroticism, PEM, NEM and sleep disturbance, history of concussion was significantly associated with elevated concussion symptom severity and accounted for a statistically significant increase in the variance explained in log-transformed levels of concussion-like symptom severity (B=.22[.08]; p=.005, ΔR²=.059; Table 4).

Table 4.

Step-wise linear regression model of confounding factors and concussion history predicting concussion-like symptom severity

	Model F(df)	R2	B (SE)	β (95% CI)	p-value
Step 1	4.15(100)	0.130
BFI Agreeableness			−0.33(0.15)	−0.21(−0.63, -0.02)	0.035
BFI Neuroticism			0.18(0.10)	0.17(−0.02, 0.37)	0.079
MPQ NEM			0.04(0.13)	0.03(−0.23. 0.31)	0.774
MPQ PEM			−0.26(1.5)	−0.16(−0.56, 0.04)	0.094
PSQI Score			0.06(0.04)	0.16(−0.01, 0.13)	0.110
	Model F(df)	R2*	B (SE)	β (95% CI)	p-value
Step 2	5.09(99)	0.189
BFI Agreeableness			−0.29(0.15)	−0.19(−0.58, -0.01)	0.05
BFI Neuroticism			0.22(0.10)	0.20(0.02, 0.41)	0.029
MPQ NEM			0.03(0.13)	0.02(−0.23, 0.28)	0.838
MPQ PEM			−0.24(0.15)	−0.15(−0.53, 0.05)	0.104
PSQI Score			0.03(0.04)	0.10(−0.04, .010)	0.329
Concussion History			0.22(0.08)	0.27(0.07, 0.38)	0.005

B=unstandardized beta; SE=standard error; = standardized beta coefficient; CI=Confidence interval; PSQI = Pittsburgh Sleep Quality Inventory; MPQ= Multidimensional Personality Questionnaire; PEM = positive emotionality; NEM = negative emotionality

^*= change in Adjusted R² statistically significant

Orthopedic injury history, confounding factors, and symptom severity

Orthopedic injury history was not significantly associated with either psychological distress or concussion-like symptoms after controlling for the above confounding factors (psychological distress DV: B=−.06[.54]; p=.905; concussion-like symptoms DV: B=.06[.08]; p=.427; Supplemental Tables 1 and 2). Controlling for confounding factors and orthopedic injury history, concussion history remained an independent significant predictor of psychological distress symptom endorsement, B=1.26[.55]; p=.024, ΔR²=.029. Controlling for identified covariates and orthopedic injury history, concussion history was also significantly associated with log-transformed levels of concussion-like symptom severity, B=.22[.08]; p=.006, ΔR²=.057.

Exploratory individual item analyses

Concussion history was significantly, but mildly (i.e., small effects), associated with a number of individual items from the BSI-18 (Figure 1), the strongest being “loss of interest,” ρ=.27, p=.005. Interestingly, although concussion history was independently associated with BSI-GSI total score above and beyond the other confounding factors within the above-mentioned model, there were no individual items that were uniquely associated with prior concussion, distinguishing endorsement patterns from other confounding factors (i.e., all items significantly correlated with concussion history were also significantly correlated with confounding factors as well). Significant, but small correlations were observed between concussion history and 12 of 22 SCAT items, ρ_s =.20-.28, ps <.05 (Figure 2). A number of individual items were significantly associated with concussion, but not other confounding factors, including headache, feeling like in a fog, drowsiness, and irritability.

Figure 1.

Heat map representing Spearman’s correlation of concussion history and confounding factors with individual items on the Brief Symptom Inventory. Y-axis represents content area of individual BSI-18 items. Hx = history; NEM = negative emotionality; r(s) = Spearman’s rho

Figure 2.

Heat map representing Spearman’s correlation of concussion history and confounding factors with individual items on the Sport Concussion Assessment Tool. Hx = history; PEM = positive emotionality; NEM = negative emotionality; r(s) = Spearman’s rho

Discussion

A number of independent studies have reported a significant association between concussion history and symptom severity in young (e.g., high school and college) athletes. However, the degree to which these associations are attributable to potential confounding factors or are unique to brain injury (vs. general injury) history was unclear. The current findings demonstrated that concussion history independently accounted for a significant amount of variance in psychological distress and concussion-like symptom severity in collegiate athletes, even after controlling for a multitude of potentially confounding variables such as sleep and personality tendencies associated with greater symptom reporting. Additionally, elevated symptom severity was observed among those with a greater history of concussion, but not in those with more extensive general orthopedic injury history. Taken together, these results suggest that neurobehavioral changes associated with prior concussion, reported in older adults, can be detected in a more subtle (i.e., lower in severity, generally below the threshold of clinical significance, and not disruptive to daily function) manner in active younger athletes, and that these elevations in symptoms are distinctly associated with concussion, and not general injury or other related factors.

Concussion history, psychological distress, and concussion-like symptoms

Consistent with numerous prior studies across all levels of development within active athletes, including youth, high school, and collegiate levels, we found a significant positive association between concussion history and both psychological distress and concussion-like symptom severity.^5–10 The presence of multiple comorbidities (e.g., depression) and factors (e.g., decreased sleep quality) known to influence symptoms has often obscured the association between concussion history and symptom severity in athletes with remote concussion. For example, one alternative explanation for the previously reported associations between the number of self-reported prior concussions and symptom severity is that they reflect underlying health or health behavior differences (i.e., sleep). The current study does not support this alternative explanation, but strengthens the notion that changes in neurobehavioral function (i.e., elevated symptom severity) associated with greater concussion history are associated with repeat concussion and can be detected as subtle changes in adolescent and young adult athletes. Additionally, the possibility that a positive reporting bias (individuals more likely to report symptoms are more likely to report prior injuries due to a positive reporting bias) underlies the association of concussion history and symptom severity is reduced in the current study through use of assessment of diverse personality traits likely to predict reporting bias, as well as a comprehensive semi-structured interview for retrospective diagnosis of prior concussions, as opposed to grossly inquiring about number of previous concussions.

The independent influence of concussion history on symptom severity manifested more distinctly on individual concussion-like items of the SCAT, as compared to items of the BSI-GSI. Specifically, items on the BSI-18 that were significantly associated with concussion history were also significantly associated with personality and sleep-related factors. Given that these factors independently contributed to unique variance on the overall BSI-GSI scores, significant items overlapping between concussion history and confounding factors (e.g., loss of interest in activities enjoyed, feeling tense, etc.) may represent general areas of functional vulnerability susceptible to detrimental processes. This is reinforced by the fact that the amount of variance accounted for in psychological distress by concussion history was virtually unchanged from single-predictor to step-wise regression models that included additional confounding factors.

Brain versus orthopedic injury history and symptom severity

Within the current study, greater history of concussive injury, but not orthopedic injury history, was significantly associated with psychological distress and concussion-like symptom severity. This suggests that elevations in symptom ratings are more likely to reflect brain injury-specific sequelae, as opposed to general response to cumulative all-cause injury or some other third variable that increasing symptom and injury reporting generally. While the causal mechanism between cumulative concussion and elevated symptom reporting requires further study, cumulative concussion is known to affect neural correlates that underlie many of the subjective symptoms, including limbic, medial, frontal, and multiple subcortical regions.^5,35–38 Furthermore, many of the same neuroanatomical regions have been identified as correlating with general symptom endorsement and, in some cases, the underlying neural substrate of subjective symptom endorsement, independent of physical changes or experience.^39–41 Future studies are required to further refine the association between neuroanatomical differences and specific symptom endorsement patterns, as well as the way in which other factors associated with greater concussion history and symptom reporting, such as inflammation, influence these associations.^42–45

Study strengths and limitations

The current study was designed and conducted to exclusively examine the potential chronic or early long-term effects of remote concussion and employed strict enrollment criteria of other potentially confounding conditions associated with neurobehavioral function (ADHD, depression, concussion within the last 6-months, etc.). This presented the opportunity to exclusively investigate the effects of prior concussion on neurobehavioral function and decreases potential for spurious associations influenced by confounding factors. Additionally, measurement and consideration of personality differences in relation to concussion history and symptom severity is a novel study feature which has yet to be considered in prior investigations of potential chronic or early long-term effects of remote concussion. The current study reduced the potential for reporting bias influencing the positive association between prior concussion and symptom severity by employing a rigorous standardized interview for retrospective identification and characterization of concussion history based on established criteria.

Nevertheless, the current study does have limitations. While race was considered as a potential factor influencing the association between concussion history and symptom severity, the overall sample is limited in ethnoracial representation and the generalizability of these findings to non-White individuals is unknown. Relatedly, generalizability of the findings to developmental groups outside collegiate-age should be undertaken with caution. Replication of the current findings should include larger, more diverse samples across the lifespan. Though rigorous methodological and statistical control was employed, given the cross-sectional nature of the study, we cannot infer the temporal relationship between repeat concussion and symptom severity.

Conclusion

Taken together, current findings suggest that subtle neurobehavioral abnormalities (i.e., greater symptom severity and distress) associated with greater concussion history can be detected in young, active athletes and that this relationship is independent of medical, lifestyle (i.e., sleep), and personality factors. These results also suggest that symptom severity in young athletes can be due to a range of factors, and it is critical for clinicians to consider the etiology of symptoms as part of treatment planning and intervention. Further, the findings indicate that the association between greater concussion history and symptom severity is more likely to represent specific brain-injury related changes, as opposed to reflecting general injury history. Future studies are required to investigate the way in which these elevations in symptom severity associated with prior concussion in younger active athletes may place individuals at greater risk for adverse outcomes later in life.