Abstract
Introduction
Children and adults with concussion often report dizziness and balance problems [3, 21]. The prevalence of dizziness after concussion is 28 to 79%, [8, 29] and persistent dizziness at follow-up periods lasting at least 3 weeks has ranged from 16 to 43% [20, 39]. The origins of post-concussion dizziness have been attributed to traumatic injury to the peripheral vestibular end-organs as well as central vestibular structures [10]. Vestibular test abnormalities occur in 50–70% of individuals with concussion who also report dizziness [11, 42]. Davies et al. [11] reported that 71% had abnormal semicircular canal function and 14% central findings. Further, vestibular and ocular-motor impairment have been reported in 50% of persons post-concussion with dizziness and visual instability [33] with dizziness identified as a predictor of prolonged recovery [9, 21]. The extant literature primarily focuses on adult cohorts with no studies investigating this phenomenology in adolescents who have experienced a sports related concussion. Accordingly, it is not clear if children have a peripheral vestibular disorder (PVD) after a concussion [3].
The vestibulo-ocular reflex (VOR) provides gaze stabilization of images on the retina during head movement [27]. With a normal VOR, a person can maintain fixation by moving the eyes with an equal velocity, but in the opposite direction to the head motion. If a corrective saccade is made after the head movement terminates and occurs in the direction of the VOR and can be seen by the examiner, it is called an overt saccade [5, 27]. If a saccade is made during the head movement, it is usually undetected and when in the direction of a deficient VOR is called a covert saccade [25, 27]. Newly developed technology has enabled clinicians to record both the covert and overt saccades [25].
The video head impulse test (vHIT) is a portable device using high speed cameras and inertial sensors to measure horizontal, vertical and torsional eye movements in response to head impulses. Consequently, the VOR gain (ratio of eye velocity to head velocity) can be calculated, most commonly for the horizontal semicircular canal [32]. As a test of the peripheral vestibular system, [35] the vHIT has been validated by comparing it to scleral search coils on normal subjects and patients r = 0.9 [24]. High cost, poor tolerance and damage to the eye can occur during scleral coil testing. Alternatively, the vHIT device costs less and is time efficient [24] and has been shown to be well-tolerated in children over the age of three [18]. The vHIT has higher sensitivity and specificity compared to the traditional bedside head impulse test [25]. In children, Hulse et al reported that 76% of the children tested (n = 55) could exhibit reproducible results with the vHIT [18].
The gain obtained from vHIT has been compared with responses during caloric testing. One of the original vHIT studies correctly distinguished eight subjects with caloric reduction due to peripheral vestibular disease from eight control subjects, without any false positives or negatives, using a gain below 0.68 as the threshold for abnormality [25]. Others have reported a sensitivity of 67% and specificity of 100% in children with a cutoff score for abnormal gains below 0.7 [17]. In a comparison with positive air calorics, only 29% of the subjects also had an abnormal vHIT response [6]. A caloric asymmetry of 40% or greater appears to optimize the ability of the vHIT to determine an abnormal result [30]. It is important to note that while caloric test functionally assesses very low frequency components of the VOR (less than 0.003 Hz), the head impulse test assesses high frequency components (greater than 1 Hz). In two case studies of persons with vestibular neuritis, VOR gain changes over time demonstrated normal gains in one subject and covert saccades over time in the second subject. Both reported “good” recovery, suggesting that the vHIT may be able to demonstrate two different types of compensation [28]. The ability to utilize saccadic eye movements during active head movement while stabilizing gaze appears to aid in functional recovery [23].
Post blast injury, Pan et al. [37] reported that none of their subjects had positive head impulse test results (n = 14) 7 months to 7 years post injury. Little is known as to whether vestibular signs and symptoms following non-blast concussive injuries are due to a peripheral vestibular disorder or due to central vestibular dysfunction. Therefore, the primary purpose of this study was to determine if peripheral vestibular disorders are present in patients after concussion. A secondary aim of the study was to determine if there is a relationship between physical performance and self-report measures of vestibular function when compared to vHIT scores.
Methods
Subjects
Children and adults with a diagnosis of concussion seen at the University of Pittsburgh Medical Center Concussion program within a minimum of one week after injury were recruited. Subjects received the diagnosis of concussion based on their reported signs and symptoms plus based on a neurocognitive testing battery administered by trained neuropsychologists who specialize in concussion management. All subjects provided informed consent and the Biomedical Institutional Review Board (IRB) of the University of Pittsburgh approved the study.
Twenty-nine youths under the age of 21 (17 female; mean age 15 y: SD 2.6 y; range 10–20 y) and 27 adults (14 female; mean age 32 y; SD 11.2 y; range 21–68 y) who were being treated for concussion participated in the study. All were undergoing active physical therapy intervention at the time of enrollment. On average, the youths were tested approximately 4 months post-concussion and the adults were tested approximately 6 months post-concussion (p = 0.29). Patients who were non-ambulatory, had significant neck range of motion limitations, or who had current orthopedic or gross neurologic injuries were excluded from the study. All subjects were screened for cervical pain prior to participation. If a patient answered affirmatively, the participant rated his/her pain on an 11 point Likert scale (0 (no pain) –10 (extreme pain)). Subjects with pain scores above five were excluded from the study in order to avoid increased neck pain.
Procedures
The vHIT system consists of a laptop computer with software (Interacoustics AS, Denmark) connected to a high speed recording camera and inertial measurement unit (IMU). The IMU records head velocity. The camera records eye movement via reflective glass. The software tracks movement of the pupil and calculates eye velocity. All subjects sat 1.5 meters in front of a wall and the target location varied depending on the subject’s height. The vHIT was calibrated according to the manufacturer’s specifications. The investigator performed quick horizontal head impulses of approximately 20 degrees with high velocity (150–300 degrees/s) to activate the horizontal semicircular canal and superior vestibular nerve [35]. All subjects rated their headache, dizziness, nausea, and fogginess symptoms before and after the head impulses using a 0–10 Likert scale.
Horizontal impulses were delivered to the left and right in random order while the subject kept his/her eyes on the target until five acceptable impulses satisfying the velocity and acceleration ranges were recorded in each direction. We used the default analysis procedures provided by the EyeSeeCam software (Interacoustics, vHIT EyeSeeCam, Item no. 8102676- Rev 1-ver 05/2013). First valid trials are selected based on minimum head impulse velocity and stationary eye velocity at onset of head movement. Traces with artifact such as eye blinks are removed automatically. We used the estimate of gain based on the linear regression of eye velocity to head velocity during the valid trials [36]. A separate gain was obtained for each direction. Asymmetry in gain was calculated as the absolute left and right difference divided by the sum of the left and right gains. In a healthy sample, the VOR gain of the horizontal canal is approximately equal to one and a VOR gain under 0.8 is considered pathological for identifying a peripheral vestibular disorder [26, 31]. With a vHIT gain of <0.7, sensitivity was 67% and specificity was 100% (positive predictive value: 100%; negative predictive value: 91%) for identifying later canal involvement in children using a vHIT Impulse system [17].
Youths and adults completed questionnaires including: 1) the Dizziness Handicap Inventory (DHI), [19] 2) the Vestibular Activities and Participation questionnaire (VAP) [1, 2] modified with the addition of eight environmental items from the International Classification of Functioning, Disability ad Health (ICF), 3) a verbal analog scale of‘ pre- and post-vHIT sensations (nausea, headache, mental fogginess and dizziness) that were also rated on the same 11-point Likert scale scale, [33] and 4) the Pediatric Vestibular Symptom Questionnaire [38]. Although the Pediatric Vestibular Symptom Questionnaire was designed for children, adults were asked to complete the questionnaire.
The Dizziness Handicap Inventory (DHI) was developed to provide a self-report of disability from dizziness [19]. The DHI is a reliable and validated questionnaire (Cronbach α 0.95) which can be used to quantify disruptions in quality of life due to dizziness influences [41].
The Vestibular Activities and Participation questionnaire (VAP) is a self-report measure to evaluate limitation in activities and participation [1, 41]. In addition, eight environmental items were included that were selected from a consensus conference related to the development of the World Health Organization (WHO) balance and dizziness core set [16]. For the total VAP score, we calculated the percentage of the 39 activity and participation items that subjects had at least mild difficulty with.
A verbal analog scale was used to report symptoms of fogginess, nausea, dizziness or headache based on an 11 point Likert scale used in the development of the Vestibular and Ocular Motor Screening tool for concussion [33]. Zero indicates no symptoms and 10 suggests the worst symptoms that they could imagine on the verbal analog scale.
The Pediatric Vestibular Symptom Questionnaire has been shown to discriminate between youths with vestibular symptoms, including youths with concussion and healthy youths [38]. Youths are asked to respond to how frequently over the last month they have experienced certain sensations with the following response options: most of the time, some of the time, almost never, never, or don’t know.
The Dynamic Gait Index (DGI) and gait speed were recorded. Normal DGI scores and gait speeds have been reported for youths between the ages of 14 and 17 [4]. Gait speed was recorded over a 6.1 m distance with a standard stop watch.
Data analysis
Data analyses were performed using the IBM SPSS Statistics 23 package. Independent t-tests were utilized to determine the effect of age and gender on vHIT, self-report (DHI, VAP, PVSQ), and gait (DGI, gait speed) measures. In instances where data were not normally distributed, the nonparametric Mann-Whitney test was used to examine the effect of age and gender. Paired t-tests were used to investigate changes in symptoms before and after the performance of the vHIT. In order to examine associations between vHIT gains, self-report and gait measures, we calculated nonparametric correlations (Spearman rho), corrected for multiple comparisons using the method of Benjamini and Hochberg [7].
Results
vHIT
A representative depiction of the vHIT scores obtained from the right and left head impulses is provided in the Fig. 1. Note that the data cluster around a VOR gain score of 1 with the mean of the child data of 1.05 and the adult VOR gains at 1.06 for the left head impulse. All of the subjects had a vHIT gain above 0.74. The mean gain asymmetry was 0.05 (SD 0.04). No difference was observed between right and left vHIT gains; consequently the average gain was computed and used for all following analyses. Significant differences in gain between youths and adults, or between males and females, were not found (p > 0.05). Headaches, dizziness, and nausea, but not fogginess, were significantly greater post-vHIT relative to pre-vHIT (Table 1).
Self-report measures
There was no effect of gender on any self-report measure. Adults had significantly worse DHI scores compared with youths (Table 2, p = 0.013). Adults had greater limitations in activities and restrictions in participation due to their concussion, as measured by the VAP (p = 0.006). In addition, adults had some difficulty with 55% of the activity and participation items on the VAP compared with only 35% of the items for youths. For the environmental ICF items, 47% of adults and 36% of the youths reported that their symptoms were made worse by environmental triggers. There was not a significant difference in PVSQ scores between adults and youths. An examination of the most severe symptoms reported by the youths were headache, spinning, lightheaded and blurriness (Table 3).
Gait measures
Gait speed did not differ between youths and adults (Table 2). The mean values indicate that both groups had normal gait speed. However, the youths performed significantly better on the DGI, with a mean score that was two points better (p = 0.002).
Relationship among vHIT self-report, and gait measures
We assessed the association amongst the measures for the entire sample, and also for each age group. The average vHIT gain was not significantly correlated with any of the self-report measures. All of the self-report measures (DHI, VAP, environmental items, and PVSQ) were significantly correlated in the entire sample (Table 4), as well as in both age groups. The correlations revealed that greater symptoms were associated with greater dizziness handicap, more limitations in activity and participation, and being affected by more environmental items. In the entire sample and both age groups, faster gait speed was associated with better DGI performance. A significant association between DGI and all of the self-report measures was observed in the entire sample. This association was primarily driven by the significant correlation between the DGI and self-report measures in the adult sample, but not in the child sample, which had limited variability in DGI scores.
Discussion
The vHIT did not identify peripheral vestibular hypofunction in any of the subject’s post- concussion. There were also no differences in VOR gain in adults versus youths. Hamilton et al. [17] reported that horizontal semicircular canal vHIT gain scores do not change as a result of ageing in groups between the ages of 11–14 and 15–19 years suggesting that it was appropriate to group all of the youths in our sample, with an age range from 10–20 years, into one group. All of our youths were “normal” based on these vHIT criteria. In our sample, fifty-three out of the 56 subjects were referred by physical therapists who were currently treating the subjects for dizziness dysfunction, with the remaining 3 referred by a neuropsychologist. Consequently, we expected some positive findings for PVD in the children and adults tested post-concussion. In persons post blast injury, Scherer et al. [40] reported that passive yaw angular VOR (aVOR) gains were similar in groups with and without dizziness. However, during active head impulse testing with a scleral search coil, the aVOR gains were lower in the post blast group that reported dizziness. Our vHIT impulse was applied passively, suggesting agreement with Scherer et al’s findings [40].
Davies and Luxon [10] reported that persons with mild TBI have vestibular dysfunction with 88% of their subjects (n = 100) having at least one audiologic or vestibular abnormality at a tertiary care center. Their medico-legal subjects had symptoms present for 34 months versus their other group of subjects who had a mean symptom duration of 22 months. Our mean symptom duration was a median of 51 days (range: 11–803 days), suggesting that the Davies and Luxon group was more chronic. All of their subjects reported dizziness as did the subjects in the current study.
Others have previously reported abnormal vestibular lab findings in 32–70% of adults with mild to moderate head injury [14, 44]. Recently in 42 children post-concussion seen at a tertiary care clinic, abnormalities in computerized dynamic visual acuity testing, sensory organization testing, and rotational chair testing were noted [46]. Their length of time since concussion was a mean of 26 weeks with a standard deviation of 20 weeks, also suggesting that Zhou et al’s subjects were more chronic than our cohort. Our subjects were recruited from an out-patient concussion center rather than a tertiary referral center, which could have affected our findings.
Corwin et al. [9] reported that children with vestibular deficits (abnormalities with VOR testing or abnormal tandem gait) took longer to recover after a concussion than those without vestibular symptoms. Others have also reported that those children with vestibular symptoms take longer to recover, [12, 43] suggesting that the development of reliable and valid tests of vestibular function have value.
The vHIT test resulted in increased headache, dizziness, and nausea symptoms post testing in our subjects. The investigators did not determine the duration of the subjects symptoms post testing. Although there was an increase in symptoms after vHIT testing in the study participants, others have reported that caloric testing is not “pleasant” and the vHIT may be an alternative to the vertigo experienced with caloric testing [17].
Adults reported problems with more items on the VAP than younger subjects (55% vs 35%). The mean total score for all subjects on the VAP was 0.86, which is a lower score than reported by others [2, 13]. In persons with vestibular and balance disorders, Friscia et al. [13] reported VAP scores of 1.2 and Alghwiri et al. of 1.4 [34]. The adults with concussion in the current study approached the mean scores of the vestibular sample with a score of 1.15. The youths with concussion had less activity limitations and participation restrictions than the adults with vestibular disorders with a mean score of 0.59.
The children had better DHI and DGI scores than the adults. Our findings agree with those of Alsalaheen et al. [3] who previously reported that older adults post-concussion had worse DHI scores than those under 18 years of age. No comparable literature was found for the use of the DGI. Alsalaheen et al. [3] used the Functional Gait Assessment (FGA) [45] in their study of older and younger persons post-concussion. The FGA includes many of the same items as the DGI and Alsahaheen et al. [4] reported that children’s scores were better on the FGA than adults post-concussion, which is consistent with the current findings.
The PVSQ mean score in our sample of youths was 1.16, with the mean age of our youths at 15 years. Pavlou et al. [38] had a mean PVSQ score of 1.2 in her children with vestibular complaints and a mean age of 13.6. Heathy children had a mean PVSQ score of 0.3. Thus the youths with concussion had a mean PVSQ that was only slightly less than Pavlou et al’s sample of children with either vestibular disorders or concussion [38].
Our goal was to determine the incidence of vestibulo-ocular reflex dysfunction in persons post-concussion. Fifty-three out of our 56 subjects were referred to vHIT testing by physical therapists who were treating the subjects for dizziness, with the remaining three referred by a neuropsychologist. None of the 56 had vestibular laboratory testing, which is not commonly conducted in children and adults’ post-concussion, and a limitation of our study. Generally, only severely involved children or adults are referred to a neurologist who specializes in otology in our setting. Our findings are in agreement with Pan et al [37]. who also reported normal vHIT findings in persons post blast, suggesting that their symptoms of dizziness may be related to a central vestibular processing issue.
Overall, older adults post-concussion were more symptomatic than the children tested. It is not clear why this occurred as there was no difference in the time since concussion in the two groups. It has been suggested in the rat model that brain-derived neurotrophic factor (BDNF) was increased if exercise was delayed post-concussion [15]. Adults may have had work related responsibilities that caused them to return to work more quickly after concussion with resultant increase in symptoms.
Conclusion
Persons with concussion in our sample had complaints of dizziness, yet had normal VOR gains on the vHIT suggesting normal horizontal semicircular canal functioning. This study suggests that the cause of dizziness post-concussion may not emanate from the horizontal semicircular canal. Older adults reported more activity and participation limitations than the younger subjects with concussion. Further investigation is warranted to determine the cause of dizziness complaints post-concussion.