IntroductionVestibular rehabilitation (VR) is a specific physical therapy program designed to reduce dizziness and improve balance in patients with peripheral vestibular dysfunction. To determine the effectiveness of VR we are currently conducting a double blind controlled study of patients with Bilateral Vestibular Hypofunction (BVH). A number of measures are used to quantify the baseline condition of these patients and to measure the changes in function resulting from VR. The purpose of this paper is to explore the meaningfulness of quantitative functional assessment measures of balance impaired patients obtained in different laboratories and to illustrate how a gait laboratory can support and supplement clinical and more traditional balance laboratory quantitative assessment.MethodologyWe evaluated 19 patients (7 males and 12 females, mean age 69.6+/-12.3 years) with BVH as diagnosed by electronystagmography and sinusoidal vertical axis rotation (SVAR) tests. The results of initial tests only are reported here. Clinical tests included timed static balance tests (newspc), the number of successful tandem gait steps eyes open (dtgeo), self-report of dizziness and disequilibrium frequency (sensdiz and sensdis) and intensity (intdiz and intdis) and self-report of perception of handicap using the Dizziness Handicap Inventory (DHI) (dhisum). Laboratory tests included SVAR test gains at 0.05Hz (grot1), 0.1Hz (grot2) and 0.2Hz (grot3) and the combined results (postsum) and combined results of test conditions 4, 5 and 6 (sum) from Equitest T M posturography conducted at the Jenks Vestibular Laboratory of the Mass. Eye and Ear Infirmary. The subjects were also evaluated at the MGH Biomotion Laboratory where phase plane analyses of CG and CoP motion were used to evaluate static balance control with feet wide (fwcg and fwcp), feet together eyes open (eocg and eocp) and eye closed (eccg and eccp) and semitandem (stcg and stcp). Dynamic balance control was evaluated during free speed (g) and paced (h) gait. Gait parameters analyzed included double support time (gdstm and hdstm), free speed average velocity (gcgavx), and maximum moment arm (gmomax and hmomax). Where repeat tests were performed within a given test session the best performance score was used for analysis. Twenty-five parameters were evaluated. Pearson correlations were used to evaluate the relationship between these quantitative measures. A 0.05 confidence level was used to determine statistical significance. Correlations between closely related parameters such as the various SVAR and phase plane parameters were expected and not considered of interest.ResultsThe DHI (dhisum), which is fairly widely reported in clinical studies, was not significantly correlated with any objective measure. Dhisum correlated only with intdiz, another self reported subjective measure. The other self reported clinical measures were significantly correlated with only 1 to 3 objective parameters. Newspc was correlated with 6 objective measures including stcp and several of the gait parameters as well as dtgeo. Dtgeo was correlated with 8 objective measures and sensdiz and intdiz. Interestingly the SVAR scores, direct measures of vestibular function, showed very little correlation with objective measures of activity of daily living (ADL) performance. Sum correlated with newspc and dtgeo, but not with the phase plane measures of static balance control. Sum also correlated with intdis as did postsum. Postsum correlated with 6 objective parameters including 4 of the phase plane parameters but not newspc. Among the phase plane parameters, fwcg and fwcp parameters correlated only with other phase plane parameters. The other phase plane parameters correlated with one to three objective parameters frequently including postsum. They correlated with dtgeo more often than they correlated with newspc. Among the gait parameters gcgavx correlated with 8 objective parameters, 3 other gait lab parameters and dtgeo, 3 phase plane parameters and newspc. Gmomax and hmomax correlated with 4 to 5 objective parameters each including newspc and dtgeo as well as gcgavx and one or two phase plane parameters each. Hmomax also correlated with sensdis. Double support times did not show correlation with other parameters except that gdstm was correlated with gcgavx.DiscussionThe lack of correlation of self reported measures such as the DHI with each other and with objective measures is indicative of the limitation of this type of measure. Patients' self-perception of impairment level or symptom severity is referenced only to their own experience, which varies widely between individuals. The DHI may be more useful as a measure of change of status where the patient is asked to evaluate their current condition compared to their condition at a previous time. The purpose for including these self reported measures in the initial evaluation was to provide a baseline for later comparisons. SVAR gains are widely accepted as an accurate measure of vestibular system function. The lack of correlation between SVAR scores and objective measures of locomotor ADL performance highlights the fact that there are means of compensating for impaired or absent vestibular function. Compensation is one basis for VR. Patients' locomotor ADL function is influenced by their ability to employ these compensatory mechanisms as well as by the severity of the vestibular pathology. The finding that the overall posturography score is more informative than the test 4, 5 and 6 results seem surprising as BVH subjects have far more difficulty with test 4, 5 and 6 than with test 1, 2 and 3. Failure on 4, 5 and 6 is so common that the score is highly skewed toward 0. Hence, it is difficult to correlate the results with parameters that have a more normal distribution. Changes in the test 4, 5, and 6 results are expected to be more useful in detecting change in function, but were not explored in the present study. The limited correlation between phase plane measures of standing balance control and other objective functional measures indicates that quiet standing performance is not a good predictor of dynamic balance control. Parameters associated with semitandem stance were most likely to be correlated with other objective measures. We are investigating why BVH subjects have such great difficulty with this posture. Gait parameters do seem to be relatively good predictors of overall balance control in that they have fairly good correlations with other objective measures. This is to be expected as dynamic balance control during gait is a relatively severe challenge. It has been pointed out that the frequency content of head movements during gait is higher than that seen in laboratory measures such as the SVAR. Hence gaze stabilization during gait can be more difficult than during vestibular lab tests, further stressing patients' compensating mechanisms. It should be noted that the most informative parameter seems to a relatively simple one, average velocity. Moment arm, and other more sophisticated gait parameters may be useful for determining specific gait stability dysfunction, but are not more globally informative than average velocity. Further the dtgeo measure obtained in a clinical setting seems to be as telling as any measure obtained in a gait laboratory. To enhance the usefulness of the gait laboratory in assessing BVH patients whose impairment is purely neurological, with no direct biomechanical or neuromuscular consequence, we will need to develop global functional performance measures rather than focus on biomechanical parameters. We may also want to measure the performance of functions which are directly affected by the pathology such as gaze stabilization during gait.