Abstract
Patients with long-standing complaints (lasting longer than 3 months) experience their self-reported disability to a greater extent than patients with new onset pathology (illness duration of one month and less). Moreover, patients suffering from continuous complaints have a larger DHI score than patients with shorter symptom duration. The first effect (new onset vs. long-standing pathology) is primarily caused by emotional factors, the latter effect (symptom duration) is attributable to functional and physical factors, not to emotional aspects. Patients with daily and weekly complaints have larger DHI scores than patients who reported only one episode. Female patients reported larger DHI scores than males. We found no effect of age, diagnostic group (no diagnosis, episodic, acute or chronic vestibular syndrome) or reported symptoms on the DHI scores.
Introduction
Due to the aging population, more and more patients suffer from balance problems and seek professional help for their complaints [8, 14]. In a Dutch population, dizziness caused work absenteeism in more than half of the patients referred to a tertiary referral center and 12% was completely disabled [18]. Since the development of the Dizziness Handicap Inventory (DHI) by Jacobson and Newman in 1990, it is possible to quantify and evaluate the self-perceived disability of patients with complaints of dizziness or imbalance [6]. The DHI is a widely used clinical questionnaire due to its cross-cultural adaptation into 16 languages and various countries [7, 13]. Besides a total disability score, it identifies emotional, functional and physical aspects of the perceived disability (related to balance disorders) by giving subscores for each of the above-mentioned aspects. In patients with pathologies characterized by attacks, timing of the assessment should be taken into account since the DHI score is directly affected by the period of the evaluation in their disease course [13]. Aside from its important role in defining the subjective impairment, the DHI can also be used to evaluate the success of balance disorder treatment [1, 15]. Moreover, Whitney et al. found that patients with a severely elevated score (>60) are at increased risk for falling [21].
The purpose of the present retrospective study, performed in our outpatient vestibular clinic in a patient database of 568 patients, was to correlate the DHI scores with demographic, symptomatic and diagnostic parameters. In a retrospective study design, it is not always possible to make a clear and correct diagnosis. Therefore, we divided the patients in 3 groups (patients with acute, episodic and chronic vestibular syndromes) according to the ICD-11 classification and we investigated the differences in DHI outcome between these groups. The following research questions were put under investigation: Does the DHI score differ with gender? Does the DHI score differ with age? Does the DHI score differ between patients with acute, episodic and chronic vestibular syndrome? Is there a difference in DHI score with respect to the length of illness? Is the DHI score influenced by the reported symptoms? What is the effect of symptom occurrence (frequency) on the DHI score? What is the effect of symptom duration on the DHI score?
Methods
Study design and subjects
The current study has a retrospective design. The database was analyzed anonymously and in accordance with the declaration of Helsinki Helsinki on ethical principles for medical research involving human subjects. Data was collected from 568 patients who visited our tertiary referral center between December 2010 and December 2013 with general or specific complaints of dizziness or imbalance. All subjects had at least one consultation in the clinic and they had filled out the Dizziness Handicap Inventory prior or during their visit. For patients who visited the clinic more than once during this period, only data from the first visit was collected.
Our patient group consisted of 359 women with a mean age of 54.3 years (standard deviation (SD) = 16.1 years; range 12.4–90.4 years) and 209 men with a mean age of 54.9 years (SD = 16.1 years; range 15.3–88.9 years).
DHI questionnaire
The DHI is a self-reporting questionnaire that consists of 25 questions with respect to the experienced disability due to vestibular complaints. The possible responses to each question are ‘yes’, ‘sometimes’ or ‘no’. These answers correlate with a specific score (4, 2 or 0 points respectively). The total score (DHI total) is 100 and the higher the score, the larger the perceived disability. Besides the total DHI score, there are 3 subscores that evaluate the functional (36 points, DHIfunc), physical (28 points, DHIphys) and emotional (36 points, DHIemo) aspects of the perceived disability. Patients can be categorized in 3 categories based on the DHI total score: mild (0–29), moderate (30–59) and severe (60–100) [21].
Clinical investigation and vestibular testing
Besides history taking and clinical examination (Dix Hallpike maneuver, screening for spontaneous nystagmus, head impulse test, headshake testing, oculomotor testing, standing and walking balance tests), patients were submitted to an audio-vestibular test battery (pure tone audiometry, collic vestibular evoked myogenic potentials and videonystagmography with sinusoidal rotary testing and caloric investigation).
Statistical analysis
Statistical analysis was performed with IBM® SPSS® Statistics, version 20 (IBM, Armonk, NY, U.S.A.). For each analysis, a significance level of 5% was adopted. To assess whether DHI scores differed between men and women (research question 1), between acute and chronic patients (research question 4), independent samples T-tests were used. The difference in DHI scores between acute, episodic and chronic vestibular syndromes (research question 3) was investigated by means of the non-parametric Kruskal-Wallis test. The age effect (research question 2), the effect of symptom frequency (research question 6) and symptom duration (research question 7) was investigated by means of one-way analysis of variance (ANOVA). We applied a multinomial logistic regression model (stepwise method – forward entry) in order to find out if the amount of self-reported disability could be predicted based on the reported symptoms and comorbitities (research question 5).
Effect sizes are calculated using Cohen’s d effect size index. Effect sizes of 0.2 are considered as small, while effect sizes of 0.5 and 0.8 constitutes medium and large effects respectively [16].
Results
Gender
38.7% of patients had a mild DHI total score (0–29), 41.4% moderate (30–59) and 19.9% had severe impairment (60–100).
On average, men scored significantly (t(566) = 3.19; p = 0.002) lower on the total DHI score (mean (M) = 34.4, Standard Deviation (SD) = 20.6), than women (M = 40.4, SD = 22.2). The effect size (d) was 0.29, which represents only a small effect. On each DHI subscale, men scored significantly lower than women (DHIemo: t(566) = 2.1, p = 0.035; DHIfunc: t(566) = 2.3, p = 0.024; DHIphys: t(566) = 4.2, p < 0.001). The effect sizes were small for DHIemo and DHIfunc (d = 0.18 and d = 0.20 respectively) and between small and medium (d = 0.37) for DHIphys. The mean values, standard deviations, p-values and effect sizes are summarized for each DHI score in Table 1.
Age
The total population (n = 568) was grouped by the decade of their age. Only one patient was over 90 years of age and therefore excluded from the analysis. There was no significant effect of age (decade) on the total DHI score (F(7,559) = 0.89, p = 0.51).
Diagnostic groups
For this analysis, only the patients of one ENT-specialist (n = 301) were selected in order to prevent bias. Patients with uncommon diagnoses (occurring less than 5 times in the study cohort) and those for whom there were insufficient data to ensure an accurate diagnosis were excluded from the analysis (n = 37). The diagnosis persistent postural-perceptual dizziness (PPPD) was only made in retrospect for those patients with no other specific diagnosis and with chronic (3 months and more), daily and continuous (or symptoms lasting from 1 hour to 1 day) dizziness or instability. All PPPD-patients (n = 23) were categorized in the ‘chronic vestibular syndrome group’
Patients who could not be categorized in a specific vestibular diagnosis (based on the current diagnostic criteria) were combined into the category ‘no diagnosis’ (n = 14). Besides the ‘no diagnosis group’, the study population was divided in 3 other groups based on the occurrency of their symptoms and diagnosis: 1) ‘acute vestibular syndrome’ (vestibular neuritis, n = 8), 2) ‘episodic vestibular syndrome’ (BPPV, n = 77; Ménière’s disease, n = 64; vestibular migraine, n = 38; BPPV and Ménière’s disease, n = 17; Ménière’s disease and vestibular migraine, n = 10; BPPV and vestibular migraine, n = 6) and 3) ‘chronic vestibular syndrome’ (PPPD, n = 23; chronic central pathology, n = 7). An overview of the diagnosed pathologies and diagnostic groups can be found in Table 2.
There was no significant difference (χ(2) (3) = 1.97, p = 0.58) between DHI-scores of the different diagnostic groups‘.
Length of illness
We could retrieve the information of the parameter ‘time since symptom onset’ from 456 patients. 102 patients (22.4%) visited the hospital within the first month after the symptoms started. This group was defined as the patients with a ‘new onset’ of their pathology. Another 15.1% (n = 69 patients) visited the clinic between 1 and 3 months after time of symptom onset. They were referred to the group ‘intermediate length of illness’. The remaining patients (n = 285; 62.5%) had ‘long-standing’ vestibular complaints, lasting 3 months or more (range: 3 months – 44 years).
On average, the ‘new onset group’ (M = 35.5, SD = 22.1) had a lower total DHI score (t(385) = –2.41, p = 0.016) than the ‘long-standing group’ (M = 41.5, SD = 21.2). The effect size was small (d = 0.28). The emotional subscale was also lower for ‘new onset’ patients (DHIemo: t(385) = –3.33, p = 0.001), while there were no differences for the functional and physical subscales (DHIfunc: t(385) = –1.66, p = 0.098; DHIphys: t(385) = –1.18, p = 0.24). The effect size was between small and medium (d = 0,38) for DHIemo. The mean values, standard deviations, p-values and effect sizes are summarized for each DHI score in Table 3.
Symptoms
The presence of following symptoms was obtained for further analysis: vertigo, dizziness, unsteadiness (while standing), unsteadiness (while walking), nausea/vomiting, tinnitus, aural fullness/pressure, headache/migraine and ‘other neurological complaints’ (symptoms like diploplia, scotoma and paraesthesia were all categorized as ‘other neurological complaints’). For this analysis, only the patients of one ENT-specialist (n = 301) were selected. Due to the retrospective design of this study, the authors were unable to determine if each symptom was explicitly inquired. For this reason and in order to prevent bias, we consider the non-reporting of a symptom as a missing value (and not as an absence of this specific symptom). An overview of the valid and missing data is summarized in Table 4.
The most prevalent symptom was dizziness (80.1%), followed by unsteadiness while standing (71.8%), and unsteadiness while walking (61.8%). Vertigo, nausea/vomiting and headache/migraine were symptoms in nearly half of the patients (53.5%, 55.8% and 51.5 % respectively), followed by tinnitus (42.5%), aural fullness/pressure (33.9%) and other neurological complaints (22.3%).
The multinomial logistic regression analyses revealed that only ‘unsteadiness while walking’ had a significant effect (Chi Square = 15.8, p < 0.001) on the outcome of the total DHI score. The pseudo R-Square values, Cox & Snell and Nagelkerke, were very small, 0.088 and 0.099 respectively. This indicates that the regression model based on the parameter ‘unsteadiness while walking’ has very low predictive capabilities with respect to the outcome variable ‘DHI total’.
Symptom occurrence (frequency)
From 370 patients, we obtained information with respect to the occurrence frequency of symptoms/complaints. More than 50% (n = 205; 55.4%) reported daily complaints, followed by weekly episodes (n = 60; 16.2%), 2–12 episodes/year (n = 46; 12.4%) and a single episode (n = 34; 9.2%). Only a small part of our population reported a monthly (n = 14; 3.8%) and yearly (n = 11; 3%) episode frequency.
There was a significant effect of the symptom frequency on DHI total (F(5,364) = 7.60, p < 0.001), DHIemo (F(5,364) = 4.39, p = 0.001), DHIfunc (F(5,364) = 5.12, p < 0.001) and DHIphys (F(5,364) = 8.40, p < 0.001). Post hoc Bonferroni tests revealed that patients who reported a single symptom episode had significantly lower scores for DHI total (p < 0.001), DHIemo (p < 0.001), DHIfunc (p < 0.001) and DHIphys (p < 0.001) than patients with daily symptoms. Moreover, patients with a single symptom episode also had lower scores for DHI total (p = 0.011) and DHIphys (p < 0.022) than patients with weekly symptoms. Effect sizes varied from (almost) medium to large. The mean values, standard deviations, p-values and effect sizes of the significant different post hoc comparisons are summarized in Table 5.
Symptom duration
From 394 patients, we obtained information with respect to symptom duration (nearly 40% of them had a symptom duration of ‘seconds to 5 minutes’ (n = 155; 39.3%), followed ‘1 hour to 1 day’ (n = 118; 29.9%), ‘continuous’ (n = 55; 14%), ‘5 minutes to 1 hour’ (n = 44; 11.2%) and ‘several days’ (n = 22; 5.6%). There was a significant effect of symptom duration on DHI total (F(4,389) = 4.89, p = 0.001), DHIemo (F(4,389) = 3.3, p = 0.011), DHIfunc (F(4,389) = 6.3, p < 0.001) and DHIphys (F(4,389) = 3.1, p < 0.015). Post hoc Bonferroni tests revealed that patients with ‘continuous’ symptoms had significantly higher DHI total scores than patients with symptoms that lasted for ‘seconds to 5 minutes’ (p = 0.010), ‘5 minutes to 1 hour’ (p = 0.032) and several days (p = 0.011). The DHIfunc score was significantly higher for the patient group with ‘continuous’ symptoms in comparison with the following groups: ‘seconds to 5 minutes’ (p = 0.002), ‘5 minutes to 1 hour’ (p = 0.019) and ‘several days’ (p = 0.011). The patients that belonged to the group ‘1 hour to 1 day’ had also significantly higher DHIfunc scores in comparison with the ‘seconds to 5 minutes’ group (p = 0.020). The DHIphys score was significantly higher for the patients with ‘continuous’ symptoms in comparison with patients that experienced symptoms for ‘several days’ (p = 0.024). Although the one-way ANOVA demonstrated a significant effect of symptom duration on DHIemo, there were no significant post hoc differences. Effect sizes varied from between small and medium to large. The mean values, standard deviations, p-values and effect sizes of the significant different post hoc comparisons are summarized in Table 6.
Discussion
The DHI is the most widely used questionnaire for the evaluation of the self-reported disability in patients with balance problems [13]. In this study, we investigated the relationship between the subjective impairments of patients with balance problems (given by the DHI score) and demographic, symptomatic and diagnostic parameters. We observed a total of 61.3% of patients with a moderate or severe disability. This is approximately in accordance with Ten Voorde et al., who reported that 69% of their patients (n total = 2242) reported a moderate or severe disability (based on the total DHI score) during follow up in their tertiary referral center for balance problems [17].
We found that men significantly reported lower DHI scores than women, with the largest effect size for the DHI physical score in comparison with the DHI total, emotional and functional score. This is in agreement with Ten Voorde et al., although they did not mention any effect sizes [17]. We did not find any age effects, which is in accordance with Jacobson and Newman (1990), Handa et al. (2005) and Loughran et al. (2006), while Ten Voorde et al. found a difference between young patients (< 20 years, mean DHI total = 33.7; SD = 17) and senior patients (> 80 years, mean DHI total = 45.2; SD = 17.7), although this difference was not considered as significant [4, 17]. Whitney et al. revealed a significantly higher DHIphys score in older adults, however they made a comparison between only two age groups (age 20–40 years vs. age 60–80 years) and not between age decades [21].
A total of 301 patients of one ENT-specialist were selected in order to investigate the difference in DHI score between several diagnostic groups: no diagnosis, acute, episodic and chronic vestibular syndromes. We found no significant statistical differences between the different groups. The results of a study of Jacobson and Calder (2000) pointed out that vestibular illnesses leading to a bilateral vestibular loss result in higher DHI total and DHI physical scores than diseases with a unilateral vestibular loss [5]. This was not confirmed in a similar study of Gill-Body et al. (2000), who reported similar levels of disability (based on the DHI score) for patients with unilateral and bilateral vestibular hypofunction. In our patient database, we did not make a distinction between unilateral and bilateral vestibular function loss. Ten Voorde et al. demonstrated that vestibular patients diagnosed with hyperventilation had a higher perceived disability than patients with BPPV, recurrent vestibulopathy, no diagnosis and migraine [17]. We did not include hyperventilation as a separate diagnosis and therefore can not compare our findings with Ten Voorde et al. [17]. In our department, a first screening for hyperventilation (as the primary cause or secondary to a vestibular problem) is being done by the self-administered Nijmegen hyperventilation questionnaire that has to be completed by patients seen during their vestibular diagnostic examination [19]. When a patient has a positive outcome on this questionnaire and when we believe that hyperventilation plays an important role in the pathology, they are referred to other specialists for extensive examination and treatment. Handa et al. (2005), who compared only 2 diagnostic categories in their study, reported that patients with Meniere’s disease had significantly higher DHI scores than patients with BPPV, regardless of their period of vertigo attacks [4]. In our analysis, both patient groups were combined in the ‘episodic vestibular syndrome group’ but when comparing the DHI scores of patients with BPPV and Meniere’s disease we found no significant differences. A long term follow up of patients with vestibular neuritis revealed that the DHI total and physical subscores were larger than in healthy subjects and that this difference was due to the vestibular disorder and not to psychological factors [10]. This finding could not be confirmed by our study as we only included patients with vestibular complaints. Patients with acoustic neuroma had significant lower DHI scores than patients with other pathologies in a study of Vereeck et al. [20]. This pathology was not included in our patient population.
One important limitation of our study was that the diagnosis of PPPD was done in a retrospective approach, based on the following criteria: chronic (3 months and more), daily and continuous complaints of dizziness or instability and exclusion of another specific diagnosis. This means that the available clinical information might not be enough to ensure that all the diagnostic criteria for PPPD were met. Therefore, we had to make the compromise that the diagnosis of PPPD was the best approximation possible and for the same reason we were not able to acknowledge the co-existence of PPPD with other (episodic) vestibular syndromes in this study. For the above mentioned reasons, we included patients with PPPD (with no other comorbid syndromes) in the chronic vestibular syndrome category.
Regardless of underlying pathology, we found that patients with long-standing complaints (lasting 3 months and more) have a higher DHI total and emotional score than patients with new onset pathology (less than 1 month). The effect size was higher for the emotional subscore than for the DHI total score. Mbongo et al. compared 3 unilateral complete vestibular loss patients according to their post-lesion stage: 1-2 months, 4–7 months and 1 year and older. The only difference found was the 1-year-and-older group having a significantly higher DHI physical score than other groups [11]. Interestingly, Gomez-Alvarez and Jauregui-Renaud followed 10 patients with an acute peripheral vestibular loss and reported a significant decrease in DHI scores from 54.2 (SD = 22.8) in the acute stage to 40.2 (SD = 8.6). This decrease was probably due to the vestibular rehabilitation programme (Cawthorne & Cooksey exercises) that their subjects had to go through during follow up [3].
We did not find any significant effect of the reported symptoms on the DHI outcome. A multinomial logistic regression analysis revealed a significant effect of the symptom ‘unsteadiness while walking’, but the pseudo R-square values were too small to acknowledge the predictive capabilities of this symptom on the DHI score.
We found an effect of the occurrence rate of symptoms (frequency) and the duration of symptoms on the DHI scores. Patients with a single episode of balance complaints reported significantly lower scores than patients with daily and weekly episodes. The effect sizes were all medium to large. Moreover, patients with continuous complaints have significantly larger scores than patients who reported a symptom duration of seconds to 5 minutes, 5 minutes to 1 hour and several days. The DHI physical score was also significantly larger for patients with continuous symptoms in comparison with patients who reported symptoms for several days. The DHI functional score was larger in patients who reported complaints for 1 hour to 1 day in comparison with seconds to minutes, although the effect size was smaller in comparison with the previous mentioned significant effects. Strikingly, there were no effects on DHI emotional score, indicating that emotional factors don’t play a role in this matter. Jacobson and Newman (1990) reported that the DHI total, emotional and functional scores (but not the physical subscore) were significantly larger for patients who report more frequent attacks of dizziness or unsteadiness and that patients with 12 or fewer episodes of dizziness per year can feel particularly disabled. They described a wide range in DHI total score (10–94) for patients with continuous dizziness complaints, which means that these patients do not always perceive themselves as severely disabled [6].
Conclusions
Patients with long-standing complaints experience their self-reported disability to a greater extent than patients with new onset pathology. Moreover, patients suffering from continuous complaints have a larger self-reported disability than patients with a shorter duration of symptoms. While the first effect (new onset vs. long-standing pathology) is primarily caused by emotional factors, the latter effect (symptom duration) is attributable to functional and physical factors, not to emotional aspects of balance problems. Not surprisingly, patients with daily and weekly complaints have larger DHI scores than patients who reported only one episode. Female patients reported larger DHI scores than males. We found no effect of age, diagnostic group or reported symptoms on the DHI scores.
This indicates that the information retrieved from the DHI questionnaire is complementary to the information obtained from clinical investigation and diagnostic tests and therefore is an essential tool in a vestibular clinic.