Sex and Sport-Specific Differences on Baseline Concussion Balance Test Performance in Division-I Collegiate Athletes

Abstract

Objective:

This study aimed to identify factors affecting baseline performance in collegiate athletes using the Concussion Balance Test (COBALT^© ).

Design:

Cross-sectional study design.

Setting:

Sports medicine research laboratory.

Participants:

NCAA Division-I collegiate athletes (n = 127; 77 male, 50 female).

Independent Variables:

Sport, sex, history of concussion, and time since last concussion.

Main Outcome Measures:

Postural sway and the number of errors across four COBALT™ conditions.

Results:

Significant differences in postural sway and errors were observed based on sex. Females demonstrated higher postural sway (0.34°/s; p < 0.001) and more errors (1.69 errors; p < 0.001) on Condition 7 compared to males. Concussion history and time since last concussion had no significant effect on postural sway or errors (p > 0.05). Differences between sports were identified, with cheerleaders demonstrating more errors than football players (Conditions 3, 7, 8; errors; p<0.05) and soccer players (Conditions 3, 7, 8; p<0.05), and soccer athletes demonstrating more errors than football players on Condition 7 (1.47 errors; p < 0.05).

Conclusions:

Our findings suggest that the COBALT is a useful tool for measuring balance performance, offering insights into baseline performance that may influence concussion management. Differences in performance based on sex and sport, but not concussion history, were observed, highlighting the importance of considering individual factors when interpreting baseline test results.

Clinical Relevance:

Based on the data presented and results of this investigation, clinicians should consider an athlete’s sex, sport, and concussion history when interpreting COBALT performance at baseline. Further research is needed to explore the impact of these factors on post-injury performance.

INTRODUCTION:

Concussions represent one of the most frequently occurring injuries as a result of participation in competitive and recreational athletes.¹ increased attention has been devoted to exploring clinical evaluation strategies because athletes returning from a concussion are potentially placed at an increased risk of musculoskeletal injury after returning to activity.^2–4 The heightened risk for subsequent injury suggests the current battery of clinical assessments used to gauge return to participation readiness may be unable to detect persistent deficits in coordination, postural stability, or sensory integration.^2,5 These findings highlight the need for additional clinical tests that may be more sensitive to coordination, postural control, and other deficits and can be used for determining an athlete’s potential to return to sport related activities after a concussion.

Balance dysfunction is a prevalent symptom in athletes recovering from a concussion.^6–8 The current standard balance assessment used in clinical management of concussion is the Balance Error Scoring System (BESS). Clinically, the BESS is scored on the accumulation of predefined errors across six conditions. However, recent investigations have demonstrated both practice and ceiling effects of the BESS in post-concussion assessments within 3–10 days following injury.^9–11 More importantly, patients recovering from a concussion frequently demonstrate BESS performance similar to a recovered or healthy state while demonstrating balance deficits in tasks of increased task demand difficulty.¹² These findings suggest the BESS lacks the task demands and sensitivity for identifying subtle residual postural control deficits in patients recovering from a concussion.

The Concussion Balance Test (COBALT^© ) was a recently developed clinical balance assessment specifically for concussion populations.^13,14 The task conditions of the COBALT are designed to challenge sensory integration and reweighting processing underlying postural control. Through the incorporation of dynamic tasks, including head rotations with eyes closed and combined head and upper trunk rotations, the COBALT aims to impart increased challenge, which may more accurately reflects sport environments. Like the BESS, patients are scored on predefined errors which are accumulated across four conditions. However, postural sway is also quantified using an integrated force plate.¹³ This assessment has demonstrated good interrater reliability and validity in uninjured, adolescent athletes. Specifically, across 10 subjects tested under all 8 COBALT conditions, the 4 raters agreed on 93.8% of the trials. There was also a high level of agreement among raters on the counted errors (ICC=0.861; p < 0.001). Additionally, 3 raters evaluated recorded videos twice, one week apart, and showed strong intrarater reliability (κ values of 0.824, 1.00, and 1.00, all with p < 0.001). Moreover, prior investigations have identified deficits in COBALT performance in high school athletes approximately two weeks post-concussion.¹⁴ This early data suggests the COBALT can identify balance impairments in athletes for longer periods of concussion recovery than have been typically documented by the BESS.

It is well documented that balance performance is influenced by sex and sports participation in collegiate athletes on a variety of tasks.^15–19 Prior reports of the BESS have identified a lack of sex, sports participation, and concussion history effects in collegiate athletes.¹⁷ Currently, normative data of COBALT performance only exists in adolescent athletes.¹⁴ Importantly, the effect of sex, sport participation, and concussion history on COBALT performance remains unknown. Moreover, no published data has been reported on baseline COBALT performance in collegiate athletes to confirm and extend findings in varying populations. Understanding how these factors influence an athlete’s performance on this test will improve the ability for clinician using COBALT to interpret results in those recovering from a concussion. Therefore, the purpose of this study was to determine the effect of multiple baseline characteristics including sex, sport of primary participation, and concussion history on COBALT performance. Additionally, the data were used to establish preliminary normative values for the athlete population included in the study.

METHODS:

Participants:

All participants were recruited for enrollment as part of a larger study examining concussion outcomes. The COBALT data utilized for this study was part of a baseline assessment. Participants were included if they were cleared for full participation in all athletic activities, were not being treated for a current musculoskeletal injury, and were not currently in a return to sport protocol for concussion. Participants were recruited and tested over a 12-month period from two large public universities. The three sports included were selected and recruited due to their high incidence of concussion, and the aims of the investigations of concussion outcomes. Prior to testing, each subject provided written informed consent which was approved by the Institutional Review Board.

Procedures:

During a single visit to a sports medicine laboratory, each participant completed a demographic and injury history questionnaire and four conditions of the COBALT. The four baseline conditions of the COBALT (Conditions 3, 4, 7, and 8) ^14,20 were collected as described by prior reports.¹³ The task requirements for each condition are outlined in Table 1. All conditions were completed twice and lasted 20-sec. During each condition, a metronome provided an audible reference for head and torso rotations. Rotation magnitudes were standardized using two vertical lines on the wall positioned in front of the participant.

Table 1.

COBALT Condition Descriptions

Condition	Vision	Surface and Base of Support	Task
3	Eyes closed	Firm surface, feet shoulder width apart	Head oscillations of 60 degrees in yaw plane at 120 BPM
4	Eyes open	Firm surface, feet together	Torso rotations of 60 degrees at 40 BPM with shoulders flexed to 90 degrees and elbows extended
7	Eyes closed	Foam surface, feet shoulder width apart	Head oscillations of 60 degrees in yaw plane at 120 BPM
8	Eyes open	Foam surface, feet together	Torso rotations of 60 degrees at 40 BPM with shoulders flexed to 90 degrees and elbows extended

Abbreviations: BPM (Beats per minute)

During the head rotation tasks (Conditions 3 and 7), participants wore a headlamp and initiated rotation with eyes open to receive visual feedback regarding yaw plan rotation magnitude. Once participants were proficient at rotating their head in the 60° arc to the pace of the metronome set to 120 beats per minute, they were instructed to close their eyes and maintain head rotation while study personnel monitored the magnitude of head rotations by ensuring the light from the headlamp passed the vertical line at each metronome beat.²⁰ Condition 7 was performed similarly to Condition 3 except the subject stood on a foam pad. During Conditions 4 and 8, participants were instructed to hold their arms out in front of their bodies, with hands joint at midline and thumbs pointed towards the ceiling. During each metronome beat at 40 beats per minute, participants were instructed to rotate their upper trunk through the 60° arc until their thumbs passed the vertical line.²⁰ Participants also wore a headlamp for Conditions 4 and 8 as upper trunk and head rotation were yoked during the conditions, and study personnel monitored the magnitude of trunk rotation in reference to the pace of the metronome.²⁰ Condition 4 was similar to Condition 8 except the subject performed the task while on a foam pad.

All COBALT Conditions were completed while standing on a force plate (Bertec Corporation, Columbus, OH). Conditions 7 and 8 were completed on a foam pad with a density of 3.6–4.4 lbs/ft.³ Data was collected through Bertec Balance Advantage software (Bertec Corporation, Columbus, OH) which automatically calculated a “sway score” for each trial. Sway scores (°/sec) were calculated as the mean angular velocity of center of pressure sway with respect to the participant’s center of gravity.¹³ Sway scores for each trial within a condition were averaged, per assessment standards.²⁰ In addition, standard deviation was calculated. In addition to postural sway, errors associated with task quality and loss of stability were subjectively recorded by the study personnel administering the test. Errors during the COBALT included lifting 1 or both hands off the iliac crest, opening eyes during an eyes closed trial, moving feet from appropriate positioning or stepping off the force plate due to a loss of balance, and not keeping head turns or trunk rotation in sync with the metronome for two beats or more. If two or more errors were committed simultaneously, only one error was counted. Prior to completing each trial, the specific task (i.e., head oscillations and torso rotations) were demonstrated for each participant and a brief practice trial was performed. The errors recorded during each trial per condition were summed for reporting.

Statistical Analysis:

Demographic information and COBALT performance were summarized using means and standard deviations as well as median and IQR where appropriate. Data normality was assessed using Kolmogorov-Smirnov tests. The effects of sex, concussion history, and sport on postural sway during Conditions 3 and 4 of the COBALT were assessed using Univariate Analyses of Variance. Analyses of Covariance were used for Conditions 7 and 8 with weight used as covariate to control for foam deformation. Mann-Whitney U tests were used to assess sex differences in COBALT errors. Kruskal-Wallis tests with post-hoc pairwise comparisons were used to determine the effect of concussion history and sport on COBALT errors for all Conditions. Concussion history was assessed as a categorical variable with three levels; no history of concussion, history of concussion within the last year, and history of concussion in greater than one year. This approach was used to best represent the range of post-injury timelines to match with recovery as well as account for any variability in reporting as this measure relied on self-report from participants. Additionally, a preliminary review of the differences between athletes who had suffered one or more concussions found no significant differences between the groups. As a result, these groups were combined for analysis. The level of significance was set a priori at p≤0.05 with Bonferroni corrections for multiple comparisons. All statistics were completed using SPSS version 25 (IBM, Armonk, NH).

RESULTS:

A total of 127 collegiate athletes (77 males, 50 females; age: 19.81 ± 1.40 years; height: 1.75 ± 0.14m; mass: 80.98 ± 26.26kg) participated in the study. Demographic information for the participants is provided in Table 2.

Table 2.

Demographic information of athletes who partook in this study presented as median [IQR].

		All (n=127)	Male (n=77)	Female (n=50)	p-value
Age (years)		20[19-21]	20[19-21]	19[19-20]	0.063
Height (m)		1.75[1.63-1.85]	72.0[71.0-75.0]	63.0[60.0-65.0]	<0.001
Mass (kg)		79.83 [58.06-98.88]	94.35.0[81.64-108.86]	90.26 [49.301-63.19]	<0.001
History of Concussion	<1 year	22	15	7
	>1 year	38	24	16
	Total	62(48.8%)	39(50.6%)	23(46%)
Months Since Last Concussion		23.5[9-48]	12.0[8.5-36.0]	24.0[9.0-48.0]
Sport		Football: n=50(39.4%)	Football: n=50(64.9%)	Women’s Soccer: n=30(60%)
		Men’s Soccer: n=12(9.4%)	Men’s Soccer: n=12(15.6%)	Cheerleading: n=20(40%)
		Women’s Soccer: n=30(23.6%)	Cheerleading: n=15(19.5%)
		Cheerleading: n=35(27.6%)

Performance based on sex during the COBALT™ assessment is outlined in Table 3. Significant differences in postural sway were observed between males and females for Condition 7 (p < 0.001), with females demonstrating higher postural sway values than males. Similarly, significant differences in the number of counted errors for Condition 7 were found between males and females (p < 0.001), with females demonstrating more errors than males.

Table 3.

Differences in COBALT Performance Based on Sex Presented as Mean (SD)

Postural Sway (°/s)	Males	Females	p-value
3	0.37(0.13)	0.34(0.12)	0.303
4	0.74(0.14)	0.75(0.21)	0.651
7	1.16(0.34)	1.50(0.37)	<0.001
8	0.95(0.18)	0.92(0.21)	0.523

The effect of concussion history on COBALT performance is summarized in Table 4. No significant effects of concussion history were found for postural sway or counted errors across any COBALT conditions. Additionally, the time since the last concussion did not produce significant differences in performance for any COBALT condition.

Table 4.

Differences in COBALT Performance Based on Concussion History Presented as Mean (SD)

Postural Sway (°/s)	No Concussion History	Concussion History (<1y)	Concussion History (>1y)	Main Effect for Concussion History
3	0.36(0.12)	0.32(0.10)	0.39(0.13)	0.142
4	0.74(0.17)	0.75(0.14)	0.75(0.19)	0.973
7	1.34(0.40)	1.19(0.39)	1.29(0.35)	0.228
8	0.96(0.21)	0.90(0.16)	0.93(0.18)	0.441

COBALT performance differences between athletes from football, soccer, and cheerleading are presented in Table 5. Significant differences in postural sway were identified across sports for Condition 7 (p < 0.001). Post-hoc analyses revealed significant differences between football and soccer athletes for Condition 7 (p < 0.001), and between football and cheerleading athletes for both Condition 3 (p = 0.04) and Condition 7 (p < 0.001).

Table 5.

Differences in COBALT™ Performance across Sports

Postural Sway	Football	Soccer	Cheer	Main Effect for Sport
3	0.39(0.15)	0.35(0.13)	0.33(0.10)	0.107
4	0.74(0.14)	0.73(0.17)	0.77(0.20)	0.591
7	1.09(0.20)	1.48(0.34)	1.36(0.41)	1.420
8	0.94(0.16)	0.94(0.22)	0.93(0.20)	0.526
Errors
3 (n=43)	0.0[0.0-5.0]	0.0[0.0-4.0]	1.0[0.0-2.0]a, b	0.089
4 (n=3)	0.0[0.0-0.0]	0.0[0.0-0.0]	0.0[0.0-3.0] b	0.054
7 (n=98)	1.0[0.0-7.0]	3.0[0.0-5.0]b	6.0[0.0-6.0]a,b	0.331
8 (n=5)	0.0[0.0-0.0]	0.0[0.0-0.0]	0.0[0.0-0.0]a,b	0.091

^a= significantly greater than soccer

^b= significantly greater than football

Significant differences were also found for counted errors across all COBALT conditions (p = 0.001–0.036). Post-hoc testing indicated that cheerleading athletes made significantly more errors than football players for Condition 3 (p = 0.022), Condition 4 (p = 0.036), Condition 7 (p < 0.001), and Condition 8 (p = 0.006). Additionally, cheerleading athletes made significantly more errors than soccer athletes for Condition 3 (p = 0.002), Condition 7 (p < 0.001), and Condition 8 (p = 0.012). Finally, soccer athletes made significantly more errors than football athletes on Condition 7 (p = 0.008).

DISCUSSION:

This study aimed to examine how baseline characteristics including sex, concussion history, and sport influence performance on the COBALT in collegiate athletes. The results revealed that while postural sway did not differ significantly between sexes across most conditions, females demonstrated higher postural sway than males specifically in Condition 7. While concussion history did not significantly influence performance on any COBALT conditions, significant differences in errors across sports were observed, with cheerleading athletes generally demonstrating more errors than both football and soccer athletes. These findings provide insight into how individual characteristics may influence baseline performance on COBALT and how such factors should be considered when interpreting results, especially in concussion management.

In line with the results, female athletes demonstrated higher postural sway and more errors on Condition 7, compared to males but no other differences were identified. These findings support previous research suggesting that minimal aspects of postural sway demonstrate consistent sex difference.^18,21 Other tests of postural control used in concussion care, such as the BESS, have also demonstrated a lack of sex differences across multiple conditions that challenge sensory integration capacity.²² Similarly, females demonstrated increased errors during Condition 7 of the COBALT. These findings differ from previous studies using the BESS, in which no sex effects were identified.¹⁷ The condition in which sex difference arose provides task demands which challenge sensory integrative capacity, particularly the vestibular system, and rely on the up weighting of somatosensory and vestibular system contributions to maintaining postural control. This may be attributed to the foam conditions which may have posed a greater challenge for females of lighter mass than the males. Specifically, the foam pad used in Condition 7 may have posed more of a challenge to lighter females compared to heavier males, potentially influencing the postural sway results. Further, it is possible that biological sex differences in mass and foot size may have contributed to the difference in performance, especially given the mass differential within the sports included in the analysis (football vs. cheerleading). Increased errors in females could therefore signify a decreased ability to integrate sensory information during tasks which impart concurrent somatosensory (i.e., non-compliant surfaces) and vestibular perturbations. As errors are the primary outcome used in clinical tests of balance in concussion assessments, these findings may have important implication for clinical interpretations of post-concussive examinations.

In this study, concussion history did not have a significant impact on postural sway or errors across any COBALT conditions. While this finding is similar to BESS,²² a retrospective study of COBALT performance in adolescent athletes who had recently been diagnosed with a concussion found significantly greater errors and postural sway during multiple conditions.²⁰ However, our concussion history group was substantially removed from their concussive event (28.43 ± 24.56 months) compared to that of previous studies.²⁰ This time lapse suggests that concussion history may not be a significant factor in baseline COBALT performance in athletes who are well removed from their last injury. This finding is particularly noteworthy, as it implies that clinicians may not need to adjust for concussion history when assessing baseline performance in athletes who are far removed from their last concussion. However, it is important to consider that acute or recent concussions could still influence performance, and further research is needed to investigate how concussion history impacts COBALT performance during recovery and if a dose-relationship exists between the number of previous concussions an athlete has experienced and performance on this assessment.

Performance differences across sports were observed, with cheerleading athletes demonstrating significantly more errors than both football and soccer athletes on all Conditions. While prior normative data of COBALT sway in adolescent athletes did not report sport specific-performance,¹³ sport specific differences in postural control have been noted during dynamic postural tasks¹⁹ and other vestibular assessments with a head shake component,²³ but not in static postural control tests used clinically for concussion.²² The results of the current study contrast with previous literature suggesting that cheerleading may be associated with improved vestibular function due to the aerobic nature of the sport. In this study, cheerleaders exhibited more errors, particularly in tasks that challenge the vestibular and somatosensory systems. Moreover, the majority of significant differences were identified in cheerleading athletes who have demonstrated comparable error rates to non-artistic athletes on the BESS.²⁴ In a previous study investigating the effects of sports participation on objective vestibular assessments, cheerleading athletes achieved higher velocities on a measure of gaze stability than both football and soccer athletes.²³ This may be due to the nature of the COBALT tasks, which emphasize static balance and sensory integration in contrast to the dynamic balance typically required in cheerleading. While cheerleaders may have developed enhanced dynamic postural control, tasks requiring static balance under sensory perturbations may pose a greater challenge. This difference in performance between cheerleaders and athletes from other sports highlights how specific types of balance challenges may engage different sensory systems, depending on an athlete’s sport. Additionally, given the diverse skill sets, roles (e.g., specific positions), and types of sports (e.g., collision, contact, and non-contact), further analysis is needed to better understand how these factors may influence COBALT performance.

When interpreting these results and determining clinical utility and implications, it is important to consider the potential confounding effect of including a sport that did not have female athletes. Given that all football athletes were male, the observed sport differences in errors may be influenced by sex rather than sport itself. For example, significant differences between cheerleading and soccer athletes may be partially explained by sex-based differences in performance, particularly because the football group was entirely male. Future studies should aim to include more balanced groups of male and female athletes across all sports to clarify whether these observed differences are truly sport-specific or if they reflect underlying sex differences.

These results suggest that sex and sport participation can influence errors on certain COBALT Conditions, namely those that challenge sensory integration. These results highlight the importance of considering these factors when interpreting baseline performance, while also highlighting the importance of pre-season baseline testing of these measures. Further, they identify the need for further investigation of post-concussion performance and the potential impact of these factors. However, it is important to note that the number of errors in this study was relatively low, which has the potential to limit the ability to detect clinically significant changes post-concussion. While statistical differences were identified, the small number of errors observed could make it difficult to identify meaningful changes in post-concussion performance. This should be noted and considered when using COBALT to assess balance deficits and recovery of function following concussion.

Several limitations regarding this study warrant being addressed. First, there are potential confounding effects of the solely male sport, with the inclusion of football. This limits the ability to evaluate the differences adequately and fully in performance based on sport. Including a more balanced sample of male to female athletes across various sports would better allow for both sex and sport analyses. In addition, future studies should aim to recruit a more diverse sample that better represents a range of sports participation. Efforts should also be made to ensure a more balanced distribution of athletes across sports to facilitate meaningful comparisons and explore potential differences. Second, the density of the foam pad used in this study differs from those denser foam packaged with the COBALT, and this may affect performance and the generalizability of the results. Further, it is important to consider the length of time and number of uses that the foam pad can retain its optimal density. Consultation with the manufacturers should be considered to ensure the best practice of this assessment. Given the novelty of the condition requirements, it is unknown how repeated testing will influence performance. Therefore, both postural sway and errors may exhibit practice effects when repeatedly assessed. Future studies should investigate the test-retest reliability in this population and ensure standardized equipment is used across all investigations. An additional potential limitation exists in the way concussion history was reported as a categorical variable. This approach was used to best manage the accuracy of specific dates of concussive events as data relied on self-reports of injury history and aimed reduce the risk of recall bias. Future prospective studies that track concussion history over time could provide more accurate data on how concussions impact balance performance over the course of recovery.

While the current investigation found no significant effects of concussion history on baseline performance, further research is warranted to investigate how performance changes over time in athletes with a concussion. Longitudinal studies would also help explore the utility of the COBALT™ for monitoring recovery post-concussion and inform return to play decisions.

CONCLUSION:

This study aimed to identify factors influencing COBALT performance in collegiate athletes based on sex, concussion history, and sport participation. Our findings suggest that postural sway during the COBALT is similar across sexes, concussion history status, and sports in collegiate athletes. However, errors during the COBALT, particularly during conditions perturbing multiple sensory modalities (i.e., Condition 7), demonstrated significant differences between sexes and across sports. Females demonstrated higher postural sway and more errors than males on Condition 7, and cheerleading athletes exhibited more errors than both football and soccer athletes, especially during conditions involving vestibular and somatosensory perturbations. As the COBALT is proposed to be a potentially useful tool for concussion rehabilitation due to its ability to challenge sensory integration, our findings provide salient information regarding performance factors in a key patient population that may influence how clinicians approach using this test. These results underscore the importance of considering sex and sport when interpreting baseline COBALT performance, particularly in concussion management. Given the test’s potential for assessing sensory integration, further research is needed to explore its test-retest reliability and its ability to detect persistent deficits upon return to play in collegiate athletes.