Postconcussion Dizziness, Sleep Quality, and Postural Instability: A Cross-Sectional Investigation

Participants and Study Design

We performed a cross-sectional study of pediatric and adolescent participants who were evaluated at Children's Hospital Colorado Sports Medicine Center, Aurora. We recruited, enrolled, and assessed participants with concussion within 14 days of injury. Concussion was diagnosed in these participants by one of the Sports Medicine Center's board-certified sports medicine physicians in line with current consensus definitions,¹ independent of the study. Control participants were recruited to complete the assessments during a sport preparticipation physical evaluation performed by the Sports Medicine Center's medical team.

Patients with concussion were included if they still had symptoms at the time of enrollment, defined as a Post-Concussion Symptom Inventory (PCSI) score >9, to ensure they had not recovered before assessment.⁸ This cutoff threshold was selected because concussion symptoms are nonspecific and many high school athletes without a recent concussion history commonly report concussion symptoms at baseline (eg, headache, fatigue, difficulty concentrating).¹⁸ Athletes were included as healthy, uninjured control participants if they were cleared during their sport preparticipation physical evaluation for full sport participation by the Sports Medicine Center's physicians at the time of enrollment, confirming their healthy status. All participants were 12 to 18 years of age. We excluded athletes if they had a coexisting lower extremity injury affecting balance at the time of assessment, a concussion within the year before the study, a preexisting learning disability or a documented structural brain injury or had sustained their concussion during a high-velocity mechanism (eg, motor vehicle accident). All participants and parents or guardians, if participants were <18 years of age, provided written informed assent and consent, respectively, and the study was approved by the Colorado Multiple Institutional Review Board.

Symptom Ratings

Patients completed symptom ratings using the PCSI.¹⁹ The PCSI was validated in a youth population and demonstrated strong test-retest reliability (intraclass correlation coefficient [ICC] = 0.79) and internal consistency (r = 0.79–0.90).¹⁹ At the time of the assessment, participants rated 21 concussion symptoms at 2 time points: preinjury severity (asked as Before the Injury/Preinjury) and current severity (asked as Current Symptoms Yesterday and Today). Responses for each symptom ranged from 0 (not a problem) to 6 (severe problem), with a score of 3 indicating the symptom was a moderate problem. We calculated total symptom severity as the total PCSI score (current symptoms) excluding dizziness, as dizziness was used for our primary grouping variable.

Grouping Variable

We categorized participants into 3 groups: concussion, dizzy; concussion, not dizzy; and controls. To account for preinjury dizziness, we based our categorization on the difference between preinjury and postinjury (current) symptom severity ratings for the dizziness item on the PCSI. Patients with concussion who had a change of ≥3 points between the preinjury and current dizziness ratings were categorized as dizzy, while those with a change of ≤2 points were categorized as not dizzy. This threshold was selected because a 3-point change indicated that dizziness was a moderate to severe problem, consistent with previous research in which participants with concussion were grouped into no/mild or moderate/severe dizziness categories.⁸ A change of ≥3 points also suggested likely clinical importance in the individual change in preinjury to postinjury dizziness.

Outcome Variables

Our primary outcome variable of interest was the Pittsburgh Sleep Quality Index (PSQI) score. The PSQI is a questionnaire used to assess sleep quality and disturbance over the past month.^12,20–22 Test-retest reliability for insomnia is 0.87.²⁰ The score range for the PSQI is 0 to 21, and a score >5 has 98.7% sensitivity and 84.4% specificity for identifying clinically important poor sleep quality.²⁰ For initial assessment within 14 days of injury for participants with concussion, the timeframe of the PSQI was modified to reflect sleep quality since the time of injury to the first clinical visit (versus over the past month as in the original timeframe of the PSQI instrument). Control participants were asked to rate their sleep quality over the past month. The PSQI consists of 19 self-rated questions assessing 7 components of sleep: subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbances, sleep medication usage, and daytime dysfunction.²²

Secondary outcome variables of interest were single-task tandem-gait time, dual-task tandem-gait time, cognitive accuracy during dual-task tandem gait, and modified Balance Error Scoring System (mBESS) errors. Single-task (ICC = 0.86) and dual-task (ICC = 0.84) tandem gait demonstrated high test-retest reliability in youth and adolescent athletes for measuring motor and cognitive function.²³ Participants completed single-task and dual-task tandem gait consistent with earlier research.^24,25 During both tandem-gait testing conditions, individuals received standardized instructions to walk using an alternating heel-toe pattern along a 3-m line, make a 180° turn beyond the end of the line, and return to the starting position using the same heel-toe gait pattern.^24,25 Each person completed 3 timed trials per condition, and the time was averaged across the 3 trials. Instructions for dual-task tandem-gait trials were the same as for the single-task trial with the addition of a simultaneous cognitive task. For each trial, participants were provided with a different cognitive task: (1) spelling a 5-letter word backward, (2) performing serial subtraction by 7 from a randomly selected 2-digit number, and (3) reciting the months in reverse.^24,25 Similar to the single-task condition, the primary outcome for dual-task conditions was time to completion, averaged across the 3 trials. During dual-task trials, we also recorded cognitive accuracy for each task, calculated as the number of correct responses divided by total responses.

We assessed static postural stability using the mBESS.^2,26 This balance assessment consists of 3 static conditions tested on a solid surface: double-limb stance, nondominant single-limb stance, and tandem-limb stance. For each testing position, participants were instructed to place their hands on their hips and maintain balance for 20 seconds with eyes closed. The number of balance errors for each 20-second trial was the score for the trial. Errors included taking the hands off the hips, opening the eyes, falling out of position, flexing or abducting the hip >30°, lifting the forefoot or heel off the floor, or remaining out of the testing position for >5 seconds.¹⁷ The total score was the total number of errors across the 3 conditions, with a higher score indicating worse balance.²⁶ In a systematic review of the BESS, Bell et al²⁶ observed moderate to good reliability, although scores may normalize within 1 week of concussion.²⁷

Statistical Analysis

For purpose 1, sleep quality, we compared descriptive information between concussion and control groups using the 1-way analysis of variance (ANOVA), χ² test, or Fisher exact test (if cell sizes within the group were <5). Variables that demonstrated a difference between groups (P < .05) were covariates in subsequent regression models. To assess differences between groups on sleep quality, we performed ANOVA. The α level was set at .05. If we observed a main effect, we conducted post hoc Tukey testing to identify between-groups differences. We then calculated a multiple linear regression model to assess the association between postconcussion dizziness and sleep quality while controlling for potential confounding variables (age, height, total symptom severity, concussion history, and migraine history). For purpose 2, postural stability, we repeated the same process using ANOVA and multiple linear regression for the outcome variables of single-task tandem gait, dual-task tandem gait, cognitive accuracy during dual-task tandem gait, and mBESS errors. Statistical analyses were performed using Stata (version 15; StataCorp), and all tests were 2 sided.

Dizziness and Sleep Quality

On univariable examination, both concussion groups had worse PSQI scores than the control group (Figure 1). The dizzy group had worse PSQI scores (9.6 ± 3.7) than the not-dizzy group (7.2 ± 3.5) and the control group (4.3 ± 2.6; P < .001). After adjusting for age, height, total symptom severity, concussion history, and migraine history, we observed that the strongest association was between total symptoms and sleep quality, and dizziness was no longer associated with sleep quality (Table 2).

View Figure

• Download as Powerpoint
• Download Figure

Table 2 Multiple Linear Regression Results for Each Outcome Measure^a

View Fullscreen

Dizziness and Functional Outcomes

In the univariable analysis, we noted that the dizzy group had slower single-task tandem-gait times than the other 2 groups and the not-dizzy group was slower than the control group (dizzy = 27.2 ± 11.7 seconds versus not dizzy = 21.2 ± 6.3 seconds versus control = 14.7 ± 3.6 seconds, P < .001; Figure 2A). We found the same pattern of differences among the dizzy, not-dizzy, and control groups for dual-task tandem gait (dizzy = 38.4 ± 16.2 seconds versus not dizzy = 29.9 ± 7.2 seconds versus control = 21.6 ± 7.5 seconds, P < .001; Figure 2B). After adjusting for age, height, total symptom severity, concussion history, and migraine history, we determined that single-task and dual-task tandem-gait times were still associated with concussion (for both the dizzy and not-dizzy groups; Table 2). No differences were identified among any of the groups for cognitive accuracy during dual-task tandem gait (Figure 2C). In the univariable analysis of mBESS scores, the dizzy and not-dizzy groups each demonstrated more errors than the control group (dizzy = 9.8 ± 5.1 versus not dizzy = 6.9 ± 5.8 versus control = 3.8 ± 3.5, P < .001), but no difference existed between the dizzy and not-dizzy groups (Figure 2D). After adjusting for potential confounding variables (age, height, total symptom severity, concussion history, and migraine history), we observed that mBESS errors were not associated with group status and total symptom severity was the only variable associated with mBESS errors (Table 2).

View Figure

• Download as Powerpoint
• Download Figure

Limitations

Our study had limitations that should be considered when interpreting the results. First, the cross-sectional study design constrained our ability to determine causal relationships. Although we demonstrated an association between dizziness and tandem gait, prospective research is needed to understand the direction of this relationship. Second, we did not have a rating of preinjury sleep quality, and sleep impairments before concussion could have contributed to poor postconcussion sleep quality. Additionally, given the observational nature of our study, we did not control for the use of sleep aids or medications preconcussion or postconcussion; thus, the use of sleep aids may have influenced our findings. Third, we recruited participants from a specialty sports medicine clinic, and they may have represented a more symptomatic sample than those presenting to other settings for concussion care. This may have limited the generalizability of our results. Future authors should evaluate the underlying dysfunction contributing to dizziness and postural instability, as this may advance the treatment of individuals experiencing these symptoms postconcussion.