This material is intended for clinicians and vestibular scientists.
The HIT test is a bedside technique used to diagnose reduction in vestibular function in one ear vs. the other.
It describes the result of having an examiner abruptly accelerate and then decelerate the head, moving the head in rapidly at high speed and then stopping it. The term "HIT" is generally used for a sudden rotational movement about the up-down (vertical) axis of the head. A similar "impulse" can also be delivered about a linear axis -- this is generally called the "head heave test", with the word "heave" denoting that the movement is lateral along the interaural axis. In both cases, the trajectory of motion consists of something that looks a bit like a "bell curve" of rotational velocity in one case, and of linear velocity in the other.
Actually the name "impulse" is not the most accurate -- while it makes for a convenient and euphonious "HIT" description. The term impulse in engineering means a force (i.e. an acceleration) that acts for a very short time. To be accurate from the engineering perspective, we would need to only consider the acceleration, and not the de cell era ti on force.
The HIT test, and its instrumented cousin the "VHIT" test, is one of several bedside methods offering similar information concerning how well the vestibular ocular reflex compensates for a change in head position. The table below provides our opinions regarding their general clinical characteristics.
|Test||Sensitive||Specific||Vulnerability to bias||Durability||Clinical Utility||Safety|
|Vibration Test||Middle||Middle||Low||very high||High||High|
|VHIT test||Middle||Low||Low||Probably very high||Medium||Medium|
Practically in our practice, we do the "HIT" test in roughly 30% of new dizzy patients. We do spontaneous, vibration and Head-shaking in 100%. Why the difference in frequency of use ? Basically, we don't think it is necessary to move the neck that quickly when the other tests work so well. Our practice differs from the opinion of Brandt and Strupp who state that the head-impulse test is the most relevant bedside test for the vestibular ocular reflex. This is somewhat of an apples/oranges comparison - - in our dizzy practice, we think the most useful tests are spontaneous, vibration and head-shaking. Brandt and Strupp think the HIT test is the most "relevant" to the VOR. Not at all the same. There is a lot of dizziness that doesn't involve a diminished VOR.
The HIT, like head-shaking nystagmus, is a method of measuring directional asymmetries in vestibular responses. Gain is usually expressed in terms of peak eye-velocity/head-velocity, but with the HIT/VHIT tests, the situation is not as clear as it appears that the output depends not only on velocity but also on acceleration. The dependence of the output on aspects of the input, tells us that the response is nonlinear. This is not unexpected as of course, we are attempting to measure an asymmetry, which is a nonlinearity by itself.
The HIT and VHIT test cannot be explained by invoking the idea that there is saturation of velocity dependent firing in vestibular afferents. Recent models of the HIT/VHIT have required greater complexity and nonlinear conjectures about interactions between head velocity and head acceleration.
Perhaps because it is so simple to do and takes no equipment, the HIT test has been the subject of an immense number of recent papers. As of late 2014, a search brought up about 200 papers, with the earliest occurring in 1997. We find it puzzling that there are some many papers about HIT and relatively fewer about the more useful vibration and head-shaking nystagmus tests.
What is the HIT test good for ?
The HIT is reported to be very sensitive to complete vestibular loss but insensitive to mild or moderate vestibular loss (Beynon et al, 1998; Hamid 2005; Harvey et al, 1997). Hamid said that more than a 50% canal paresis is needed for a HIT test to be positive. This makes its sensitivity similar to vibration and head-shaking nystagmus. In other words, the HIT test is not positive in most cases of vestibular neuritis, because most cases are not accompanied by a canal paresis of 50%.
In our clinic setting, we think the HIT test is poorer than the vibration test for localization in all comers, as it can fail (especially in compensated patients) and is vulnerable to bias. However, HIT is a very good way of detecting uncompensated unilateral vestibular loss, because the HIT is more specific than vibration and head-shaking. When one has a positive HIT, head-shaking nystagmus, and vibration induced nystagmus all at the same time, one can be quite certain that there is a highly significant unilateral vestibular weakness.
So what common sense and literature data says, is that the HIT test has a small useful role to contribute in vestibular diagnosis. It does not replace ENG (which is sensitive to 35% weakness), or vibration (which works 20 years later), or Rotatory chair (which checks the low frequencies). However, it is a very quick way to detect a unilateral loss.
The HIT test is fast -- it just takes a few head turns to decide whether it is positive or not. This, as well as the lack of any equipment needed has led to its use in the emergency department. In essence, the idea is that if the HIT is positive, it is probably an inner ear problem, and if it is negative, probably not. This probably does work most of the time.
Recently Chen et al (2014) reported that the HIT test is indeed positive in unilateral vestibular loss, but only sometimes positive in cerebellar strokes. They called the strokes where the test failed, and the patients looked like peripheral lesions, "pAICA" strokes, and the strokes where the test succeeded and the patients looked like normals or mild bilaterals "cAICA" strokes (AICA is the anterior inferior cerebellar artery, from which the labyrinthine artery originates). Their study showed that this general idea works most of the time.
One would think that migraine patients would be normal and have normal HIT tests. So far, as of 2014, no study has been made of this.
One would think that Meniere's patients would also have largely normal HIT tests, as it is rare for patients with Meniere's to have more than 50% weakness. Zuleta-Santos et al (2014) studied 36 patients and did indeed find a rather diverse set of results, and noted that the more tests that were done, the more abnormalities were found (as any sensible person should expect).
Lee et al (2019) pointed out that patients with acute vertigo in the context of Meniere's often have normal HIT tests, and in particular, don't often have strokes. This means that the utility of the "HINTs" test, to differentiate vertigo from stroke from vestibular neuritis, is lowered because of the possibility of the patient having a hair-cell disorder (with a normal HIT test), rather than a nerve disorder.
HIT (head impulse test) -- rotational version.
Click here to see movie of HIT test in a person with a vestibular nerve section (on the left side). While there is little question that this person has a very significant vestibular lesion, as caloric testing documented only a 4 deg/sec response in the left ear (right ear had 40), we find it hard to appreciate the 10:1 difference in vestibular responsiveness from the HIT recording.
The patient's head is rapidly rotated by the examiner (abruptly and with high acceleration) about 20 degrees to the right or left. The patient is told to fixate on the examiners nose. After the head stops, the examiner watches the patient's eye to see whether or not a refixation saccade is needed to get the patient's eye on the examiners nose. A reliable and significant refixation saccade is judged as positive. On the graphic to the above, on the bottom, the up going (rightward) eye movements are in the "good" direction, and the down going (leftward) eye movements are in the "bad direction". This is for a patient who just had a vestibular nerve section. (graphics from Dr. Dario Yacovino). The bedside test requires a subjective judgement on the part of the examiner, who also controls the stimulus. This is not a good situation.
The HIT test also is vulnerable to prediction as patients who know which way their head will be turned, can generate "covert" saccades. (Weber et al, 2008). Another name for these saccades are "vestibular catch-up saccades". (Tian and Crane, 2000). These are shown above. A discussion of covert saccades is found here. The vestibular clinician wants to minimize covert saccades. This is best done by making the head movements unpredictable. It is not a good idea to warn the patient that you are going to rapidly rotate their head, or to let them know which way you are going to turn it. If you do, they may be able to "fool" you by making a predictive saccade as you turn their head. Mantokoudis et al (2015) pointed out that inward head impulses are more predictable and elicit more "covert" saccades.
The HIT test requires rapid head movements, which may cause neck pain in persons with arthritis. So far, we are unaware of any reports of carotid or vertebral dissection due to the HIT, but we would expect that it would have a similar prevalence of vascular compromise as do similar chiropractic maneuvers. It seems to us that it is only a matter of time till someone reports a stroke in temporal proximity to a HIT test or VHIT test. Heart block has been rarely reported after the HIT (Ullman and Edlow 2010). This is difficult to comprehend.
The HHT is so far not been used often at the bedside. The head is moved linearly along the interaural axis (Kessler et al, 2007). The heave is about 5-6 cm in excursion. There are some practical problems with this procedure as it is difficult to confine the head to a single direction, and also (perhaps), some safety concerns incurred by very rapid acceleration/acceleration movements similar to those that occur from chiropractors or in whiplash injuries.
Like the HIT, the head-heave test depends on a subjective judgement of the examiner, and is likely affected by prediction (although the test is very little studied).
Allowing the patient to rotate their own head (active head rotation) is not as sensitive as the passive HIT as patients can use prediction to normalize their performance (Black et al. 2005).
Similarly, motorized devices can be used to automate the test (Aalto et al, 2002). This is impractical for clinical use.
The "Video HIT" or VHIT device is available in several configurations. More about the VHIT is here.
While the author of this page has never seen any damage result from the HIT or VHIT test, simple logic would suggest that a test that involves rapid head rotation could occasionally injure the neck or the vertebral arteries. Similar considerations apply to chiropractic treatment, and certainly there are many examples of neck pain, and a few cases of vertebral artery dissection after similar chiropractic "snapping" procedures. Of course, these risks have not stopped chiropractors from treating neck pain, and we don't think that they should stop vestibular clinicians either.
That being said, we do not think HIT is appropriate for "all comers" -- i.e. as a screening test for vestibular disorders. We think that this role is far better filled by the bedside Frenzel goggle test.
The HIT appears to be most useful as part of a rapid battery of bedside tests (as described above). (Mandala et al. 2008). It provides unique information about compensation (covert saccades), and ultimately might end up being primarily useful to vestibular physical therapists.
With the HIT or VHIT, by combining several quick clinical tests, the "dizzy doctor" can quickly determine the status of the inner ear. We would not advise practicing without a bedside video Frenzel however, so the VHIT device does NOT substitute for a method of monitoring eye movements in the dark.
We do not advise reliance on the HIT test to detect unilateral loss at the bedside, because of covert saccades. If it is positive - -it was productive. If negative, you still don't know for sure, but combined with the vibration/HSN test -- you should be OK. The VHIT probably is sufficient to detect unilateral loss, which comprises a small but important fraction of all dizzy patients.