Vestibular autorotation testing in patients with benign paroxysmal positional vertigo

Abstract

OBJECTIVES:

The current gold standard for diagnosis of benign paroxysmal positional vertigo (BPPV) is the Dix-Hallpike maneuver. However, because of fatigability, the Dix-Hallpike is often falsely normal. The objective of this study was to evaluate the utility of vestibular autorotation testing in the diagnosis of BPPV.

METHODS:

The charts of 210 patients at a tertiary referral center for vertiginous disorders were reviewed. All patients underwent clinical evaluation, Dix-Hallpike testing, audiometry, electronystag-mography, and vestibular autorotation testing. The vestibular autorotation results of patients with BPPV were compared with the findings in patients with non-BPPV vestibular disorders. The sensitivity and specificity of vestibular autorotation testing in diagnosing BPPV were calculated.

RESULTS:

Ninety-one patients (42.9%) had BPPV, 76 patients (36.2%) had vertigo of uncertain cause, 28 (13.3%) had unilateral vestibular hypofunction, 9 patients (4.3%) had Meniere's disease, and 2 patients (1.0%) had perilymphatic fistula. Patients with BPPV were 3.32 times more likely to have a normal horizontal gain (95% CI = 1.54–7.19). A normal horizontal gain is 85% sensitive but only 36% specific for BPPV. Patients with BPPV were 1.9 times more likely to have vertical phase lead (95% CI = 0.953.93). Patients with BPPV were 2.20 times more likely to have both normal horizontal gain and vertical phase lead (95% CI = 1.03–4.69) The sensitivity of the combination of normal horizontal gain and vertical phase lead on vestibular autorotation testing is 87% specific but only 25% sensitive in the diagnosis of BPPV.

CONCLUSION:

A normal horizontal gain or vertical phase lead on vestibular autorotation testing in a vertiginous patient is suggestive of but not exclusive to a diagnosis of BPPV. The combination of a normal horizontal gain and vertical phase lead on vestibular autorotation testing is highly suggestive of the diagnosis of BPPV. Adjuvant use of these parameters in vestibular autorotation testing may prove to be helpful in the diagnosis of BPPV.

Benign paroxysmal positional vertigo (BPPV) is a potentially debilitating but easily treatable disorder. It is responsible for up to 25% of all cases of vertigo and is the principal cause of vertigo in the elderly. 1 - 3 First described by Barany 4 in 1921, the disorder is characterized by vertigo and nystagmus that are precipitated by head tilt toward the affected ear. Classically, the rotary nystagmus beats toward the affected ear, begins after a several-second latency, has a duration of less than 1 minute, and is fatigable with repetitious head tilts. Although most patients with BPPV will experience the classic symptoms, atypical presentations are frequently encountered. These symptoms can occur in the normal, head upright position and may be difficult to differentiate from other peripheral or central vestibular disorders.

The cause of BPPV has been ascribed to the accumulation of dense particles in dependent portions of the semicircular canals. These particles, or otoconia, move with gravitational and positional forces and stimulate the cupula, resulting in the transmission of aberrant signals to the central nervous system. 5

The gold standard of diagnosing BPPV is the Dix-Hallpike test. 6 The maneuver is performed by rapidly placing the patient in the lateral supine position with the head hanging below the level of the horizontal plane. The result is considered positive if the position elicits rotary nystagmus and vertigo that last for less than 1 minute, are fatigable, and are associated with a several-second latency.

The peripheral abnormality in BPPV and its relationship to certain head movements suggest that the vestibular-ocular reflex (VOR) may assist in the diagnosis of BPPV. The VOR functions to maintain the image of a distant object on the retina during head rotation, such as occurs during locomotion or other activities of daily life. To accomplish this aim, the VOR generates compensatory eye movements equal in velocity and opposite in direction to the motions of the head. Traditionally, the VOR is evaluated by caloric testing. Calorics, however, are limited in that they test only the horizontal semicircular canals in a nonphysiologic frequency (<0.1 Hz).

Table 1.

Distribution of cohort diagnoses (N = 210)

Diagnosis	n (%)
BPPV	91 (42.9)
Dysequilibrium of unknown cause	76 (36.2)
Unilateral vestibular hypofunction	28 (13.3)
Bilateral vestibular hypofunction	9 (4.3)
Meniere's disease	4 (1.4)
Perilymphatic fistula	2 (1)

Table 2.

Association between age, race, sex, and BPPV (N = 210)

	BPPV	Control ∗	P value
Mean age (y)	60.5 ± 14.1	53.7 ± 18.28	<0.05
Sex
Male	32	50	>0.05
Female	59	69
Race
White	85	105	>0.05
Nonwhite	6	14

∗Non-BPPV vertiginous patients.

Table 3.

Sensitivity and specificity of VAT results in diagnosing BPPV

VAT result	Sensitivity %	Specificity %
Normal horizontal gain	85	36
Vertical phase lead	34	79
Vertical phase lead and normal horizontal gain	25	87

The vestibular autorotational test (VAT) was first described by O'Leary et al 7 in 1988. This test involves the use of a portable computer connected to a head strap containing a velocity sensor. Head oscillations in the physiologic range of 2 to 6 Hz are used to test both horizontal and vertical VOR responses. Eye movements are recorded by electro-oculography. Parameters obtained from autorotation testing include gain and phase. Gain is defined as eye velocity divided by head velocity amplitude. Decreased gain indicates eye movement velocity less than the velocity of head movement in a given plane. Phase is defined as the time lag in degrees of eye velocity relative to head velocity. Expressed as phase lead or phase lag, lag is expressed as the time lag in degrees of the eye peak response relative to the preceding head peak response. VAT is an attractive addition to the balance specialist's armamentarium because it assesses both horizontal and vertical responses in the physiologic range, is easily administered, and is well tolerated by most patients. The purpose of this investigation was to assess the utility of VAT in diagnosing patients with BPPV.

METHODS AND MATERIAL

The charts of 210 patients at a tertiary referral center for vestibular disorders were retrospectively reviewed. These charts represented a convenience sample of all vertiginous patients archived in the clinic's medical record database. All patients underwent pure-tone audiometry, electronystagmography (ENG), and autorotation testing. Electro-oculography-assisted Dix-Hallpike testing was conducted on each patient during the ENG. Charts were randomly selected from two groups of vertiginous patients. The first group consisted of patients with a diagnosis of BPPV. The diagnosis of BPPV was confirmed by examination with Fresnel lenses and electro-oculography-assisted Dix-Hallpike testing during ENG. All patients with BPPV had a positive Dix-Hallpike test; any patient with a history of positional vertigo and a negative Dix-Hallpike test was excluded from this investigation. The second (control) group of subjects was taken from a pool of patients with non-BPPV vestibular disorders. Patients with BPPV were excluded if they had a negative Dix-Hallpike test, a greater than 11% difference between right and left speech discrimination scores, a 15-dB difference between right and left pure-tone averages, or a medical history of middle or inner ear disease or otologic surgery. An abnormal VAT was defined as 2 adjacent recordings outside of 2 SDs. The vestibular autorotation parameters of patients with BPPV were compared with those of patients with non-BPPV vestibular disorders. With the electro-oculography-assisted Dix-Hallpike test used as the gold standard, the sensitivity of VAT in diagnosing vertiginous patients with BPPV was calculated. Student t tests, χ² analyses, and conditional logistic regression were used to compare audiometric and vestibular test results as well as demographic factors between groups.

VAT PROTOCOL

The patient is first fitted with two electro-oculographic electrodes at each of the outer canthi for horizontal eye movement and with two electro-oculographic electrodes above and below one eye for vertical eye movement. One electrode is placed on the forehead for reference. The adjustable headband housing the rotational velocity sensor is then securely fitted to the patient. The patient is instructed to stare at a wall target and move his or her head in time with a computer-generated tone. Smooth head movements are performed throughout 1 to 6 Hz. Three horizontal and 3 vertical “runs” are carried out for each patient. The portable computer, which is attached to the headband, stores, averages, analyzes, and plots the data as gain/phase versus frequency (2–6 Hz) for vertical and horizontal channels.

Fig 1.

Typical VAT recording for subject with BPPV. Solid triangles represent frequency-specific results for test subject with 2 SD error bars; clear circles represent normative data (± 2 SDs). A, Normal horizontal gain; B, horizontal phase; C, vertical gain; and D, high-frequency phase lead outside of 2 SDs for frequencies 4.25–5.50 Hz.

RESULTS

Ninety-one patients (42.9%) had BPPV, 76 patients (36.2%) had vertigo of unknown cause, 28 patients (13.3%) had unilateral vestibular hypofunction, 9 patients (4.3%) had bilateral vestibular hypofunction, 4 patients (1.4%) had Meniere's disease, and 2 patients (1.0%) had perilymphatic fistula. Table 1 displays these results.

The mean age of the patients in the cohort was 57 years. One hundred twenty-seven patients (61%) were female; 188 patients (91%) were white. The median pure-tone average was 15 dB, and the median speech discrimination score was 96%. The median duration of vertigo was 9 months. Patients with BPPV were significantly older than non-BPPV vertiginous patients. Sex and race were not associated with a diagnosis of BPPV (Table 2). Because of physical limitations, 3 patients were unable to complete the horizontal and 2 patients the vertical VAT study.

Abnormalities in vertical gain and horizontal phase had no association with BPPV. Patients with BPPV were 3.32 times more likely than non-BPPV vertiginous patients to have a normal horizontal gain (95% CI = 1.54–7.19). A normal horizontal gain in vertiginous patients is 85% sensitive but only 36% specific in diagnosing BPPV. Patients with BPPV were 1.9 times more likely than non-BPPV vertiginous patients to have vertical phase lead (95% CI = 0.95–3.83). Vertical phase lead is 79% specific but only 34% sensitive in diagnosing BPPV. Patients with BPPV were 2.20 times more likely to have both a normal horizontal gain and vertical phase lead than were non-BPPV vertiginous patients (95% CI = 1.03–4.69). The combination of normal horizontal gain and vertical phase lead on VAT is 87% specific but only 25% sensitive in diagnosing BPPV (Table 3).

DISCUSSION

The most important function of the vestibular system is stabilization of vision during head movements. This is accomplished by the VOR. Pathologic stimulation of the VOR leads to the physical finding of nystagmus and the concomitant symptom of vertigo. Unlike caloric testing, which stimulates the VOR at ultra low frequencies, VAT assesses the VOR over the physiologic range of 2 to 6 Hz. Similarly, standard caloric testing stimulates only the horizontal semicircular canal in a non-physiologic manner, whereas VAT tests both the vertical and horizontal canals during active head movements. The use of VAT may therefore reveal abnormalities not identified by means of standard vestibular testing.

The etiologic classification of vertigo is often difficult. Atypical presentations, especially at referral centers for vertiginous patients, are often the norm. Such a patient may, in fact, have more than one vestibular disorder. The vestibular findings in these patients may be unpredictable. A patient with Meniere's syndrome and BPPV, for example, may have a normal or equivocal Dix-Hallpike maneuver. The addition of VAT to the vestibular test battery may shed some light on these difficult diagnostic challenges. In the present cohort, vertiginous patients with BPPV were 3.32 times more likely to have a normal horizontal gain than non-BPPV vertiginous patients (95% CI = 1.54–7.19). Although patients with BPPV were 1.9 times more likely to have vertical phase lead than non-BPPV vertiginous patients, these results were only of marginal significance (95% CI = 0.95–3.83). The combination of normal horizontal gain and vertical phase lead, however, was a statistically significant predictor of BPPV (relative risk = 2.20; 95% CI = 1.03–4.69). A normal horizontal gain is 85% sensitive in diagnosing BPPV. Although its specificity is unreliable, an abnormal horizontal gain would help rule out a diagnosis of BPPV. The addition of a finding of vertical phase lead in a patient with normal horizontal gain is highly specific, however (87%). Although the sensitivity of these findings is unreliable, the combination of vertical phase lead and a normal horizontal gain would be highly suggestive of a diagnosis of BPPV. Fig 1 displays a typical VAT result of a patient with BPPV in our cohort.

Through our selection criteria for this study, we attempted to eliminate patients with mixed vestibulopathies–ie, with BPPV combined with a disorder such as Meniere's disease or vestibular neuritis. A normal horizontal gain may be a sensitive indicator of BPPV because of the constructs of our selection criteria. When disorders that create horizontal VOR abnormalities are eliminated from the vertiginous patient population, we are left with a group of patients who have vertical VOR abnormalities. Because BPPV most commonly affects the posterior semicircular canal, a vertical VOR abnormality would be expected. At the same time, if mixed vestibulopathies have both horizontal and vertical VOR abnormalities, the sensitivity of a normal horizontal gain would be lowered by the inclusion of any such patients. The combination of a normal horizontal gain and vertical phase lead was found to be very specific but poorly sensitive in our patient population. We suspect that any test for BPPV will have less than satisfactory sensitivity because of the phenomenon of fatigability found with BPPV.

In 1994, Corvera-Behar et al 8 compared VAT findings in patients with BPPV with VAT findings in non-vertiginous control subjects. They found that the mean horizontal phase was smaller in patients with BPPV than in controls. They found no significant difference in mean horizontal or vertical gain between the groups. This compares favorably with our results, inasmuch as a normal horizontal gain in a vertiginous patient is actually predictive of a diagnosis of BPPV in our cohort. Although the mean vertical phase of patients with BPPV in the investigation by Corvera-Behar et al 8 appears to be lower than that in control subjects for most frequencies, these results were not statistically significant. That study, however, used mean group data and is subject to the effects of outlying data; in contrast, our investigation compared individual data with standardized norms and is not affected by data at either extreme. This allows a more reliable assessment of individual patient responses and may explain the discrepancy in results between the two studies.

Several caveats must be considered before these findings are generalized. Data in this retrospective analysis were collected from charts on patients at a tertiary referral center for vertigo. The vestibular characteristics of these patients may differ substantially from those of patients in another setting. The sensitivity and specificity of VAT in diagnosing BPPV were calculated with the Dix-Hallpike maneuver used as the gold standard. The Dix-Hallpike test is very specific; however, its sensitivity is unknown. Because of fatigability, many cases of BPPV can defy diagnosis. VAT may help identify these individuals. A prospective investigation using a combination of history, physical examination, and diagnostic testing is needed to evaluate this notion.

CONCLUSIONS

In this cohort, patients with BPPV were significantly more likely to have a normal horizontal gain as well as phase lead in the vertical channel. A normal horizontal gain in a vertiginous patient is 85% sensitive in diagnosing BPPV. An abnormal horizontal gain would suggest a vestibular disorder other than BPPV, although such a finding does not rule out concomitant BPPV. The combination of a normal horizontal gain and vertical phase lead in a vertiginous patient would, however, be highly suggestive of a diagnosis of BPPV (specificity, 87%). Additional research is necessary to evaluate the VAT findings in patients with positional vertigo and normal Dix-Hallpike testing.

References

1. Epley

. Particle repositioning for benign paroxysmal positional vertigo. Otolaryngol Clin N A 1996; 29: 323–31.

2. Brandt

. Positional and positioning vertigo. J Neurol Sci 1990; 95: 3–28.

3. Hughes

Proctor

. Benign paroxysmal positional vertigo. Laryngoscope 1997; 107: 607–13.

4. Barany

. Diagnonose von Krankheitserscheinungen im Bereiche des Otolithenapparates. Acta Otolaryngol 1921; 2: 434–7.

5. Moriarty

Rutka

Hawke

. The incidence and distribution of cupular deposits in the labyrinth. Laryngoscope 1992; 102: 56–9.

6. Dix

Hallpike

. The pathology, symptomatology and diagnosis of certain common disorders of the vestibular system. Ann Otol Rhinol Laryngol 1952;61L:987–1016.

7. o'leary

Davis

Kevorkian

. Dynamic analysis of age related responses of the vestibulo-ocular reflex using sweep frequency active head movements. Adv Otorhinolaryngol 1988; 4: 179.

8. Corvera-Behar

Corvera

Hernandez

. Benign positional vertigo produces a specific pattern of abnormalities with high frequency vestibulo-ocular reflex testing. Ear Nose Throat J 1994; 73: 768–71.