VHIT conflict with other vestibular tests
This material is intended for clinicians and vestibular scientists. See also: HIT test
Timothy C. Hain, MD • Page last modified: October 1, 2022
The VHIT is an instrumented technique used to diagnose reduction in vestibular function in one ear vs. the other, or bilateral loss. An examiner abruptly accelerates and then decellerates the head, moving the head in rapidly at high speed and then stopping it. The VHIT is currently popular, somewhat less expensive than alternatives, and the more enthusiastic users claim that it can replace the older repertoire of vestibular testing, such as for example, rotatory chair testing and caloric testing (Morrison et al, 2021). If this conjecture were true, then the VHIT should supply the same data as the other two. It doesn't.
Some very enthusiastic articles about VHIT have been published. MacDougall, Weber, McGarvie and Halmagyi (2009), from the Australian group that commercialized a variant of the VHIT, studied 8 patients with vestibular lesions, and reported that it was 100% sensitive and 100% specific. The elusive perfect vestibular test !
On the other hand, Mezzalira, R., et al. (2017) commented "Within our sample, the caloric test was more sensitive for vestibular dysfunction. Therefore, the video head impulse test is not a suitable screening tool of the vestibular system in patients with chronic complaints. "
This page provides some examples and discussion considering why the VHIT test may conflict with other vestibular tests (usually being normal).
Weber et al (2009) reported that VHIT "detects gentamicin ototoxicity". The implication is that VHIT detects bilateral vestibular paresis. Here we mean to point out that this is sometimes untrue.
r-chair VHIT conflict Case 1: Highly abnormal rotatory chair with decreased gain, increased phase and asymmetry -- same person as right Normal VHIT in same person as on left side
r-chair VHIT conflict2 Case 2: Highly abnormal rotatory chair with decreased gain, increased phase and asymmetry -- same person as right Extremely abnormal VHIT in same person as on left side
The two cases above illustrate that results of the VHIT and rotatory chair can be wildly different. This suggests that there is a lot of "play" in the high-frequency VHIT. Much more than in the better established Rotatory chair test. Of course, a gold standard is not available here. We think that when there is doubt, the rotatory chair test should win.
What might be going on here ?
Perhaps explaining case 1, VHIT only detects rather severe unilateral or bilateral vestibular loss. Normal persons can compensate and raise their high frequency gain, sacrificing their low frequencies. As the VHIT does not monitor low frequency responses, someone with as much as a 50% loss of vestibular function, equally distributed on both ears, could go entirely undetected by VHIT. Thus the VHIT would be a bad test to monitor early aminoglycoside ototoxicity. This is illustrated by the first case above.
In case 2, the VHIT appears to be worse than the rotatory chair. It is clear that there were very good head thrusts, and also a large number of covert saccades. Thus this seems to be a situation where the high-frequencies are "out", and the low-frequencies are still present. This particular lady had migraine as well. Perhaps her pathway to compensation involved covert saccades rather than increasing high-frequency VOR gain.
In other words, one might conjecture that people can take different pathways to compensation after a bilateral loss. Some increase the high-frequency gain at the expense of the low-frequency gain. Others use covert saccades instead, and do not adjust their high-frequency gain upward. If this is true, VHIT is shown to be at best, just a partial measure of vestibular function. Rotatory chair covers more ground.
VHIT is also reported to be insensitive to Meniere's, not performing as well as caloric testing (Bell et al, 2015). VHIT and calorics may be sensitive to different conditions.
McGarvie et al (2015) discussed that calorics are more sensitive than VHITs to Meniere's. In particular, they said "There was a clear dissociation: patients with MD had a small or absent response to caloric stimulation of their affected ear, whilst their response to vHIT was in the normal range. " Choi et al (2017) also pointed out the discrepency. Fukushima et al (2018) compared VHIT to caloric tests in patients with MRI documented endolymphatic hydrops (EH). They reported "The difference in the vestibular EH between the presence and absence of CP was not significant if assessed using the vHIT (P = .5591), but it was statistically different if assessed using the caloric test (P = .0467)." Or in other words, calorics were sensitive to endolymphatic hydrops, but VHIT was not. Lee et al (2017) reported that during acute attacks of Meniere's, caloric responses are generally weaker while VHITs are variable. They concluded "These dissociations suggest a frequency-dependent impairment of canal function or mechanical property of the endolymphatic hydrops during the attacks of MD."
In aggregate, these data appear pretty convincing. So obviously, we have two different measures of the VOR here - -caloric and VHIT. One is sensitive to Meniere's (caloric), one is not (VHIT). The obvious inference is that they are measuring different things. The combination of a high and low frequency test - -VHIT and calorics - -may be useful for diagnosis of Meniere's. It also follows that the VHIT cannot replace caloric testing. One might also argue that the information in calorics is contained within the low-frequency response for rotatory chair testing, and by the same token, one would expect that Meniere's patients might have reduced rotatory chair time constants but normal VHITs. Somewhat more simply, one might also expect a reduced gain-TC product in Meniere's, accompanied by a normal VHIT.
Shaw and Raghaven (2018) had a different take in their discussion of two cases with congenital abnormalities of the semicircular canals. They suggested that because the VHIT was normal, that their two cases where the caloric was more sensitive led them to -- "questions the functional relevance of canal paresis values calculated from caloric test results." This line of thought is hard for us to follow. We would not assume the VHIT is either functionally relevant or better than caloric, given that when someone is having an attack of Meniere's disease vertigo, it would be hard to argue that the normal VHIT is more functionally relevant.
FIgure 2 from McCaslin et al, showing the correlation (or lack thereof) between caloric testing and VHIT.
McCaslin and associates (2014) reported on the correlation between caloric testing and VHIT. It is not tight -- reflecting the immense variability of both of these tests. They suggested that the VHIT and caloric tests should be viewed as complementary, and that neither one replaces the other.
r-chair VHIT conflict 3 Case 3. Near complete loss on rotatory chair testing, and complete loss on caloric. Much better performance on VHIT.
Case 3 had bilateral sequential vestibular neuritis. Both the caloric test and Rotatory Chair test document severe weakness. The VHIT test however, produces a gain of about 0.66, which falsely suggests that the patient has nearly intact vestibular function. This is almost certainly due to relative preservation of high frequency gain. The VHIT is wrong.
r-chair VHIT conflict -- VHIT worse than R-chair and HSN/Vibration.
Case 4. Near complete loss on rotatory chair testing, and complete loss on caloric. HSN is weak, and vibration nystagmus is absent. VHIT shows very severe L loss, more so than would seem possible with unilateral loss.
Case 4 had an apparent left sided vestibular neuritis, but with some symptoms as well on the right. In particular, a weak BPPV was seen on the right. Vibration and HSN nystagmus are not in keeping with 100% loss as suggested by VHIT and ENG. This may be a situation where there is a partial bilateral, and the patient has chosen to "optimize" for good VOR gain to the right, and just given up on left. It is not entirely clear what is going on here, but perhaps different patients can choose different compensation strategies.
Janky and Patterson (2020) acknowledge that there can be conflicts in the vHIT and rotatory chair test, and reported a "misclassification rate of 9% overall, and 23% in CI patients. These authors considered the rotatory chair as "the gold standard,", and did not consider the possibilty that both could be accurate at the same time. Other authors have pointed out that there can be a discrepency between VHIT and R-chair, especially in Meniere's patients.
VHIT normal, Rotatory chair (or perhaps caloric test) abnormal.
This is a fairly common pattern because the VHIT measures high frequencies ALONE, the rotatory chair measures high and low frequencies, and the caloric low frequencies alone. Generally speaking, the low frequencies are more sensitive than the highs, and thus the VHIT can fail to diagnose more easily than the rotatory chair.
In partial bilateral loss, the rotatory chair will be abnormal at low frequencies, but the VHIT can wrongly suggest that vestibular function is normal. This is due to plasticity increaseing the VOR gain, but leaving the time constant reduced. Thus the VHIT creates false NEGATIVES. Note that when patients are on benzodiazepines, the low frequency information of the rotatory chair test is often incorrect.
The rotatory chair is somewhat dependent on having working saccades, while the VHIT is not. While the explanation is complicated, the rotatory chair can be abnormal in persons with no saccadic eye movements, but the VHIT can be normal. Here, the VHIT is (probably) correct. So in persons with saccadic problems (usually rare cerebellar patients or PSP), the Rchair can be false POSITIVE. This is a very extremely rare situation.
Meniere's disease (also see above)
As discussed above, Patients with Meniere's disease are reported to have reduced caloric responses but normal VHIT testing, due to hydrops (Choi et al, 2017; McGarvie, et al. 2015). We agree with this observation from our own clinical data. Furthermore, the extent of hydrops correlates with the extent of caloric weakness, in patients with normal VHIT tests. This observation certainly needs confirmation. If it is true that hydrops selectivel reduces caloric responses (which are an example of low-frequency responses), one would think that this would also extend to rotatory chair tests. Whether or not this is hydrops or just plasticity (i.e. improved high frequency responses, reduced time constant) due to central processes remains to be established. Caloric tests are problematic because they have a high false positive rate, as they are very dependent on technique.
Rotatory chair normal, VHIT abnormal.
This is usually an VHIT artifact -- i.e. Rotatory chair is correct. The second example above illustrates a situation where the VHIT suggests complete loss but the rotatory chair, partial loss. The rotatory chair has a lot more data to work with than the VHIT, and the VHIT is vulnerable to technical artifact, especially goggle slippage, poor head movement technique, failure to calibrate, looking at the wrong place.
That being said, there are times when the VHIT shows a clear unilateral reduction, and the rotatory chair gain and TC are normal. This mainly occurs when patients are taking benzodiazepines such as Xanax, as these medications increase the time constant of the VOR.
VENG normal, VHIT abnormal, Rotatory chair also abnormal.
We have encountered this pattern, and have no good explanation. One would think that it might reflect rapidly fluctuating vestibular function, but we are not aware of a disorder that can go from normal to unilateral loss in a day.
References: We have linked journal names that are from less reputable sources to our reference quality page.