menu Contact Us Dizzy Patients Health Care Providers Research BPPV DVD Tai Chi DVD Understanding Dizziness Acknowledgements Disclaimer Quoting

Hyperventilation Induced Nystagmus

Timothy C. Hain, MD Page last modified: July 21, 2016

This page is written for clinicians.

Hyperventilation (HVT) is thought to induce symptoms through it's effect on lowering the concentration of carbon dioxide in the blood. Most authors report nystagmus, but there are also reports that it increases postural sway in both normal subjects and persons with bilateral vestibular loss. (Sakellari, V. and A. M. Bronstein, 1997). At present, it is thought that HVT is an indicator of damage to the vestibular nerve or central vestibular pathways.

The early literature suggested that this was a sign of psychogenic (psychiatric) disturbance (Drachman and Hart, 1972), but later workers using better technology to monitor eye movements suggest that nystagmus induced by hyperventilation is a good sign of vestibular disease (see review that follows). Some workers have suggested that hyperventilation increases the gain of vestibular responses, unmasking asymmetries ( Lempert and Brevern, 1996; Theunissen et al, 1986). This conclusion, if indeed the case, may be relevant to findings in persons with unilateral vestibular loss.

The main problem with this test is false positives. If a test can be positive in normal persons as well as patients with disease, because there are generally more normal persons presenting for testing than patients with disease, the number of false positives can greatly exceed the true positives.

The conclusion of this review is that hyperventilation is a useful test for vestibular disorders, when combined with other evidence to increase the specificity of the test. See our paper written on the subjective of maneuvers -- Cherchi/Hain 2010, for more.

What hyperventilation induced nystagmus (HVIN) looks like.

In normal persons, HVIN is a low amplitude downbeating nystagmus. An example is shown below. This is not found in "100%', but roughly 50% of normals. The other 50% either have nothing, or a minor horizontal nystagmus. Almost nobody gets upbeating nystagmus or torsional nystagmus after hyperventilation (if they are normal), although this can happen rarely in persons with 8th nerve injury (such as acoustic neuromas). We don't know of any formal studies about this, but we have an immense experience as we have done this test on new dizzy patients at least 3 times every day for many years.

Typical appearance of HVT induced nystagmus in normal persons.


In persons with vestibular lesions,

Rather than being purely downbeating, in persons with substantial vestibular lesions (here we are talking about 50%), HVIN usually beats toward the side of lesion (Minor et al., 1999), as opposed to head-shaking nystagmus or vibration induced nystagmus, which initially beats away from the lesion. This has been suggested in the past as evidence that the mechanism for HVT nystagmus is to improve conduction in a partially demyelinated vestibular nerve.

This mechanism would predict the opposite direction for patients who have nystagmus after a complete vestibular lesion, as there is no longer a mechanism for improving vestibular function.


Hyperventilation induced nystagmus (slow phase velocity)


Methodology of the hyperventilation test

Most investigators simply ask patients to breath deeply and rapidly for 30 breaths, or for a given duration (i.e. 30 seconds to 3 minutes). This method clearly is somewhat difficult to quantify as some patients will avoid deeply breathing (perhaps out of anxiety), while others are "good" hyperventilators, and one may have to remind them to stop. There is a tiny risk of triggering a seizure, as for example, in persons with "petit mal" epilepsy, or passing out. For this reason, we suggest that the examiner stand close to the subject and be ready to assist.

The way that I test for hyperventilation induced nystagmus in my dizzy clinic. There is also a MOVIE: hvt small.avi (15 meg) Hyperventilation induced nystagmus in person with left sided acoustic neuroma. This is a rather long (1.5 minute) clip. Many other movies can be found on the site DVD:

You should expect to see downbeating nystagmus in most people after a vigorous 30 breath hyperventilation. If you don't see it at least 50% of the time, you are doing the test wrong (i.e. without moving enough air). The test means nothing if it is of a pretend hyperventilation.

I use a video-ENG device (as shown at the top of this page), which allows me to see the eye without allowing vision of the person who is being tested. After watching for spontaneous nystagmus in darkness for about 10 seconds, I hyperventilate the individual for 30 deep breaths, in complete darkness. They should be fast and deep. I steady the patient by holding on to one of their shoulders or the top of their head while doing this as HVT often induces some ataxia. I observe (and record to DVD) the nystagmus during the whole procedure. Generally this test is preceded by the vibration test. A "significant" change in nystagmus -- generally a reversal in direction, is counted as a positive HVIN. No change, minor downbeating nystagmus, or accentuation of already present nystagmus, is counted as a negative HVIN.  Downbeating nystagmus in particular suggests that the HVT was done properly and vigorously.

There have been few attempts to do this in a more rigorous way. Wilson and Kim (1981) studied ventilation testing with electronystagmography, and compared hyperventilation (which lowers PCO2), as well as oxygen (which increases PCO2). In a study of two groups totaling over 700 patients, they found that oxygen testing is only occasionally positive (2.5%) and is of little practical value. Hyperventilation testing was positive 8% of the time but rarely alone, and had no localizing value. When hyperventilation and oxygen tests are both positive, the nystagmus may be in opposite directions.

In 6 of 18 patients with acoustic neurinomas, hyperventilation produced transitory direction-changing nystagmus. We have never observed this ourselves although we have tested many patients with acoustics. We have occasionally seen biphasic nystagmus in MS.

Eye movements (nystagmus) may be recorded with many devices ranging from direct observation, Frenzel goggles, video systems, EOG, and scleral eye coil. As might be expected, the most sensitive method is recording in darkness using a low-noise method such as video or scleral eye coil. The least sensitive method is use of unaided observation or optical Frenzels, or EOG. Little has been published concerning use of video-eye movement recordings. This methodology would be expected to be useful.

Evidence for the utility of hyperventilation in vestibular disorders.

Many studies have documented the utility of hyperventilation in vestibular disorders as well as a relative paucity of nystagmus in persons who are otherwise normal.

In the author's clinical practice, HVT is done on nearly all new patients. Here are his observations:

Normal persons.

Bance et al. (1998) found HVT nystagmus in 3.5% of normal persons. Monday and Tetreault (1980) found no effect on caloric nystagmus after 90 seconds of hyperventilation. These studies suggest that the rate of false-positives is low. Of course, the frequency of false positives will depend greatly on the method of recording (see above).

In our clinical practice, we routinely hyperventilate new patients for 30 deep breaths. Roughly half of all patients have a weak downbeating nystagmus. We think that this is simply normal. Obviously, it is different than the findings of Bance et al. We think that Bance et al used older technology and just missed it.

Persons with acoustic neuromas. There have been many papers on this subject. These observations generally confirm our own clinical impression. Acoustic neuroma is simply an example of a vestibular problem that involves the 8th nerve. Other conditions -- meningioma's, any problem that affects the nerve, would be expected to have similar findings.

A potential exception to this general rule are acoustics that do damage through increased perilymph protein. This mechanism has recently been discussed as a mechanism for hearing loss in NF2. One would expect here that while head-shaking nystagmus might be positive, HVT should be negative, because there is no nerve damage.

We have observed a few patients with acoustics (prior to surgery) that have strong torsional HVT. Because torsional HVT is never encountered in normal persons, this is a rare but very sensitive clue that there is a damaged vestibular nerve (and maybe an acoustic neuroma).

Post Gamma Knife:

We have noticed (but not published) that patients with tumors of the 8th nerve who have been treated with Gamma Knife nearly always have HVT. An very strong example of HVT induced nystagmus is shown here in this movie. The nystagmus beats toward the side of radiation. It seems to resolve after 2 years.

Persons with other disorders.

Treatment of Hyperventilation induced Nystagmus

Nothing has been written concerning treatment of HVIN. If HVIN is due to nerve irritibility, nerve stabilization agents should work. As HVIN is triggered by hypocapnea and associated metabolic alkalosis, methods of avoiding this should work. As HVIN is a nystagmus, drugs that suppress nystagmus might work. As HVIN is probably associated with partially damaged nerves, vestibular nerve section (i.e. taking out an acoustic neuroma) should work too.

Methods of avoiding hypocapnea include "breathing into a paper bag", and drugs that induce acidosis such as acetazolamide.

None of these treatments has been assessed in a quantitative fashion, and their utility is therefore simply a matter of conjecture as well as anecdotal evidence.

Bottom line for hyperventilation induced nystagmus:

Hyperventilation is a useful test in diagnosing disorders of the vestibular nerve. In our opinion, HVIN is mainly useful when it is found in persons with no other signs of vestibular disorder, and also a known acoustic neuroma or the "quick spin" symptom (which is suggestive of vestibular paroxysmia). In this context, it induceds a nystagmus that beats away from the side of lesion. The main problem with the test is it's positivity in otherwise normal persons. Because of this, it should be combined with other tests, such as vibration testing that induces a nystagmus in the opposited direction, to improve the specificity.

Research project opportunities having to do with hyperventilation induced nystagmus (as of end of 2014)

As noted above, HVT is a mildly useful dizziness test which has not been studied to a great extent, perhaps because it has been incorrectly assigned as a test for anxiety, by past studies using less sensitive methodology than we now have available. There were 50 papers in the literature accessed by Pubmed as of 2013, searching on hyperventilation and nystagmus.

It would be interesting to know many things -- how sensitive is the response to the number of breaths ? How specific is the combination of HVT with another test (such as spontaneous nystagmus, head-shaking or vibration ?) -- especially the direction changing nystagmus. One might guess that hyperventilation done first might increase the percentage of "abnormal" findings in vibration or head shaking.

Some patients get nystagmus with just 10 deep breaths. Are these "supersensitive" patients part of a clinical group ? Migraine ? Metabolic alkalosis ? Borderline hyperventilators ? We would guess that migraine patients might develop HVT with less stimulus.

Does the test work for vestibular neuritis ? (we don't think it does). Can the unsteadiness induced by hyperventilation in various clinical contexts be quantified ?(perhaps use a posturography system). Is it meaningful ? (we don't think it is meaningful, as postural stability is too noisy)

How do drugs affect HVIN ? One would think that nerve stabilizers such as anticonvulants and carbonic anhydrase inhibitors such as acetazolamide might change the amplitude of HVT nystagmus. (we think that oxcarbamazine may reduce the amplitude)

Does HVT differ from HSN in patients with acoustic neuromas and hearing loss associated with increased perilymph protein ? One would expect HVT in nerve damage, but not in increased protein. One would expect HSN in both.


Copyright August 3, 2016 , Timothy C. Hain, M.D. All rights reserved. Last saved on August 3, 2016