Superior Canal Dehiscence (SCD): Symptoms and Diagnosis.

Timothy C. Hain, MD   Marcello Cherchi, M.D. Page last modified: December 29, 2019

See also: SCD (overview) SCD treatment. SCD references

Symptoms of an inner ear dehiscence syndrome


Usually there is an unsteadiness which increases with activity and which is relieved by rest.

Some people with fistulas find that their symptoms get worse with coughing, sneezing, or blowing their noses, as well as with exertion and activity. This sort of symptom goes under the general name of "Valsalva induced dizziness", and it can also be associated with other medical conditions in entirely different categories --for example, the Chiari malformation, and a heart condition called "IHSS".   Oddly, a recent report suggests that the Chiari is far more common in SCD (Kuhn and Clenney, 2010) than the normal population. We think that this report is likely due to sampling bias (i.e. this isn't true).

Movie of nystagmus elicited by Valsalva in person with Superior Canal Dehiscence (51 meg)

Pressure sensitivity:

The changes in air pressure that occur in the middle ear (for example, when your ears "pop" in an airplane) normally do not affect your inner ear. When a fistula is present, such as in SCD, changes in middle ear pressure will directly affect the inner ear, stimulating the balance and/or hearing structures within and causing typical symptoms. Pressure sensitivity due to SCD generally causes much stronger nystagmus than pressure sensitivity in persons with round or oval window fistulae, presumably because the pressure stimulus is directly applied to a single semicircular canal in SCD rather than disturbing the inner ear in a less direct way. There are also a very few other conditions that can also cause pressure sensitivity such as Meniere's disease and vestibular fibrosis.

One would think that persons both on CPAP and SCD would have dizziness from the interaction between the two conditions. We do not know of any confirmation of this as yet.

Supplemental material

Sound sensitivity

In superior canal dehiscence or in persons with fenestrations, it is not unusual to notice that use of ones own voice or a musical instrument will cause dizziness (this is called the "Tullio's phenomenon").

Movie of nystagmus elicited by sound. For other examples see the page on Tullio's

There are also patients who can indicate that their voice sounds louder than normal to them. This is a form of "autophony". More commonly autophony is caused by a patulous eustachian tube, which is another subject entirely. In eustachian tube malfunction, such as the patulous ET, the voice is "boomy", as if in a barrel. This is due to a longer resonant cavity in the middle ear.

Persons with unilateral SCD may have a positive "hum" test. When they hum a pitch, it is louder on one side.

Rare patients with SCD can "hear their eyes move". Or hear themselves blink. (Bertholon et al, 2017) If the patients noticed this, without reading about it on the internet, it is very specific. Of course, people are suggestible and if they develop this after reading Dr. Google, it is much less reliable. Occasional people can "hear their eyes move" after acoustic neuroma surgery, presumably due to abberant regeneration.

Other auditory symptoms

Some people experience ringing or fullness in the ears, and many notice a hearing loss. According to Yuen et al (2009), 85% of persons with SCD have auditory symptoms including autophony (40%), hyperacusis to bodily sounds (65%), hearing loss (40%), aural pressure (45%), and tinnitus (35%). What is missing in this report is a comparison to a control group -- our experience with SCD does not bear out Yuen's observations. We think that the main presenting symptom of SCD is pressure or sound sensitivity. We don't find that these other symptoms or signs are generally troublesome.

Clearly there are some patients with hearing loss - -this is puzzling as the damage to the ear in SCD is nowhere near the cochlea.  Perhaps the difficulty in SCD is that the pressure fluctuates too widely because the inner ear is directly connected to spinal fluid pressure through the opening.  This might be a similar mechanism to the "enlarged vestibular aqueduct" syndrome.

Some patients with SCD experience pulsatile tinnitus. This appears to be a variant of autophony.

How does the doctor know if I have superior canal dehiscence?

Dehiscence, being a bone defect, is nearly always diagnosed using a high resolution temporal bone CT scan. Of course, CT scans involve radiation, and radiation is a little damaging, and best avoided when feasible. Other tests, not involving X-rays, may provide a very good clue that an temporal bone CT is indicated. We think that generally either the valsalva test or a large amplitude of the oVEMP tests should be positive as an indication to do a temporal bone CT. We don't think that these CTs should be done in everybody.

Tests that may be helpful in the office (Valsalva is the best) are as follows:

Laboratory tests that may be helpful (VEMP is most useful) are the following:

Of the office based tests, the Valsalva is the best as it is specific, though insensitive. Also rather sensitive is asking patients if they can "hear" a tuning fork (128 hZ) applied to their wrist. The latter method is of course vulnerable to suggestibility. Patients who can "hear their eyes move" always have SCD.


Valsalva test:

In SCD, positive pressure or Valsalva against pinched nostrils produces downbeating nystagmus, with a torsional fast phase consistent with stimulation of the affected ear (CCW for right ear, CW for left ear). See example below. Negative pressure or Valsalva against a closed glottis may produce upbeating nystagmus and nystagmus beating with the torsional fast phase in the opposite direction (CW for right ear, CCW for left ear). We ourselves prefer the Valsalva against a closed glottis.

Practically, we don't think that you can do this test without magnification -- i.e. a video-frenzel system with a good enough focus that you can see torsion.

Another method is to use an examining microscope focused on the sclera. We are less enthused about technique as it is very hard to keep the sclera in view while the patient is undergoing a maneuver. Also, the light can be uncomfortable.

For those familiar with posterior canal BPPV, the vector relationships between vertical and torsional components is reversed so that the upbeating nystagmus beats away from the "bad" ear, and downbeating, towards the "good" ear. More commonly, however, no nystagmus at all is produced by either maneuver. In persons with lateral canal fistulae (which are rare and usually confined to persons with cholesteatoma or after fenestration surgery), horizontal nystagmus can be produced (see example below). In persons with window fistulae, generally very little nystagmus is produced by Valsalva or for that matter, any maneuver.

nystagmus elicited by Valsalva in person with L Superior Canal Dehiscence (51 meg)

R SCD nystagmus elicited by Valsalva in person with R Superior Canal Dehiscence -- figure 2b (2 meg)

Fenestration Supplemental material on the site DVD: Movie of nystagmus elicited by Valsalva in person with fenestration

Case example: In the man shown in figures 3 and 4, 10 seconds of straining produced a very powerful torsional nystagmus (and a lot of dizziness).

Other office based tests:

Our current feeling is that these tests are much lower yield than the Valsalva.

A fistula test , which entails making a sensitive recording of eye movements while pressurizing each ear canal with a rubber bulb, is occasionally helpful. A positive test is good grounds for a temporal-bone CT. Fistula tests are little used because they are difficult to do and insensitive. Fistula tests are often not available or even thought of. However, if a patient complains of dizziness during tympanometry, this is a clue that the patient has a positive pressure test.

A strong nystagmus (vertical and rotatory) may be produced by pressure in the external ear canal. However, we do not think that this is very sensitive. It is very specific

Asking patients if they can "hear" a 128 hz tuning fork on their wrist, is often positive, but of course this is vulnerable to suggestibility.

Upbeating nystagmus provoked by vibration over the mastoid of person with left sided SCD.  Image courtesy of Dr. Dario Yacovino.

Vibration can occasionally produce nystagmus over the defective ear.  An example of this is shown above.  Again, our impression is that this is VERY insensitive. It is also nonspecific as there are far more patients without any SCD that will have vertical nystagmus.

Simple observation of the patient's eyes with appropriate equipment (such as video frenzel goggle) may also provide the diagnosis, as in some cases, there is a pulse-synchronous oscillation (Rambold, 2001; Hain et al, 2008), see videos below and case 2. This rare sign requires either use of an ophthalmoscope or video frensel goggles to see it. One also has to think of it (: this is usually the hard part) The main confounding possibility is oculopalatal myoclonus, which causes a similar but non-pulse synchronous oscillation.

SCD pulse synchronous nystagmus movie Supplemental material on the site DVD: ----Pulse synchronous nystagmus in SCD

SCD pulse synchronous nystagmus movieSupplemental material on the site DVD: ----Pulse synchronous nystagmus in R SCD -- figure 2b

Figure 3: Conductive hyperacusis in patient with L SCD. VEMP testing was much stronger on the left side, which is the one with the air-bone gap. From this, the audiologist concluded that the patient had SCD, and she was right !


scd audio
Figure 3b: Conductive hyperacusis in patient with bilateral SCD. This can easily be missed as audiologists just assume that one can't hear better than '0' dB, and don't always test the patient thoroughly. In other words, thresholds can be better than 0.


Audiograms may show bone conduction better than air. We think this is method of detecting SCD insensitive because we think that audiologists tend to see what they expect to see, and they just don't expect this. It can work when the audiologist is used to diagnosing SCD.

Laboratory based tests:

VEMPS (Vestibular evoked myogenic responses)

Figure 5 left: cVEMP obtained in an individual shown in figure 3, who has left sided superior canal dehiscence, using a Bio-Logic Navigator Pro. The left side is much larger than the right.

Right: Threshold VEMP in same person, showing lower threshold on the left side.

VEMPs are very useful in dehiscence syndromes because they quantify sound sensitivity. There are two general flavors -- cVEMPs (cervical) and oVEMPs (ocular). Both are useful for diagnosis. These sound evoked vestibulocollic evoked potentials have been described as useful in diagnosing Tullio's phenomenon (sound induced dizziness) from superior canal dehiscence (Brantberg et al, 1999; Watson et al, 2000). The side with the larger cVEMP (figure 5 left) or lower threshold (figure 5 right) is the abnormal side. A threshold at or lower than 65 is very suggestive of SCD.

cVEMPs (and here we mean threshold cVEMPs) are not always positive. In other words, it is very clear that one may have SCD on X-ray, and a normal VEMP. The lack of sensitivity probably is due to a mixture of "autoroofing" of SCD by the dura, and the usual decline in VEMPs with age or other ear disorders.

On the other hand, threshold VEMPs are fairly specific. We have rarely encountered a person with a positive threshold VEMP that did not have SCD. The exceptions are generally young women, who tend to have very large VEMPs.

oVEMPs are a more recent development. oVEMPs can be far more obviously positive than cVEMPs, because the potential on the symptomatic side can be 10 times larger -- this is not possible with a cVEMP.

We currently think the oVEMP is the most sensitive laboratory test for SCD. An amplitude > 20 is very suggestive.

Audiometry is generally done as a preliminary test, and an alert audiologist who knows about SCD may make the diagnosis on the spot. In patients with SCD (see figure 3), audiometry may show bone conduction scores better than air (conductive hyperacusis). This is not universal -- but occurs in roughly 40% (Yuen et al, 2009). If there is a simultaneous sensorineural hearing loss in SCD, the overall picture may mimic the conductive hearing loss pattern of otosclerosis (Mikulec et al, 2004). However, as VEMP's are present in SCD, but absent in conductive hearing loss, it is easy to tell these two apart.

Also, tympanometry may induce dizziness, which may lead to the diagnosis.

ENG testing often shows a minor reduction in responses on the dehiscence side. Also a downbeating nystagmus may be seen on positional testing, which resembles that of anterior canal BPPV. Most of the time though, ENG testing is not diagnostic.

An "ECochG", or electrocochleography may be of help also, although only in rare instances. The main role of ECochG is to diagnose Meniere's disease, which is a common alternative source of pressure sensitivity. ECochG is technically challenging and it may be difficult to locate a laboratory that does it well. We would not do this test at all if the VEMP is abnormal -- we would go right to the CT.

Figure 4: Coronal CT scan of the temporal bone clearly showing missing bone at the top of the left anterior (superior) semicircular bone. Also the tegmen is dehiscent.

R SCD movie of X-ray of SCD (contributed by Dr. Dario Yacovino) (4 meg)

A CT scan of the temporal bone should generally be obtained in persons with sound or pressure sensitivity. CT of the temporal bone is supposedly very accurate in identifying canal fistulae (Fuse et al, 1996), although as there is really no other good way to identify canal fistulae, it is hard to be sure that it is picking them all up. As SCD is a type of canal fistula and it is moderately common, the main reason for this procedure is to check for SCD.

Before beginning this discussion - -note that most temporal bone CT scans are "high radiation" procedures because enough Xray energy must be used to "see" into a very hard bone (temporal bone). All Xrays increase cancer risk. Accordingly, CT scans should not be done to "screen" for SCD. Safe tests such as the VEMP should be used first. CT scans also should not be done in a "successive approximation" mode -- i.e. you don't start with a poor scan, and then get a good one. If you are going to do it - - do it right the first time. Be extra cautious in scanning children, as radiation exposure increases cancer risk, and children have a long life expectancy (Tunkel et al, 2012).

An exception to the general rule that temporal bone CT scans are "high radiation" are "cone beam CT scans". These scanners have far lower radiation. They are commonly found in dentist offices, but so far, have not been favored in hospital settings. We hope that eventually all CT scans for SCD are done with this new, lower radiation technology.

CT should be done of the temporal bone with at 0.6 mm resolution or better (lower is better). It may be impossible to get a CT scan with a resolution < 0.6 mm. This is generally OK, but don't accept lower resolution (i.e. 1 or 1.25 mm is usually not good enough).

Conventional CT scans of the brain are nearly always useless to diagnose SCD as their cut resolution is 8-10mm -- this is almost as big as the entire inner ear ! There is also a trade-off between radiation and resolution. One might argue that the tiny lesions that can be discovered with 0.1 mm cuts are not worth the radiation load. This issue is presently unclear but perhaps it is true. Although resolution is not quite as good as with direct coronal, the radiation load is half compared to the protocol that combines direct axials and coronals, and we think that this is a reasonable compromise.


Figure 5. Temporal bone CT scan with images taken in plane of superior canal. There is a wide area of dehiscence seen at the top.


Reformatted sagittal views are essential for the situation where there is no direct coronal. Coronal reformats are a bare minimum. Better are oblique reformats parallel to and perpendicular to the plane of the superior semicircular canals. Direct coronal views are better for diagnosis, but the more views you do, the more radiation. Until cone-beam CT comes into general use, right now we think it is best to stick with reformatted axials.

Because the posterior canal on each side is perpendicular to the ipsilateral superior canal, oblique reformats allow the clinician to check for dehiscence in all canals.

In our opinion, this is what your CT-temporal bone prescription looking for dehiscence should say:

CT scan of the temporal bone, with high resolution (0.6 mm or less). Direct axial, with reformatted coronal and oblique views parallel to and perpendicular to the plane of the Superior Semicircular Canals.

DO NOT PROCEED If you cannot perform a CT scan of the temporal bone with a resolution of 0.6 mm or less. In this case, send the patient home so that an adequate scan can be done elsewhere.



A properly done MRI can rule out SCD, as well as point strongly towards SCD (Browaeys et al, 2013). The method is to use a newer 3T scanner, and obtain T2-Coronal views. On T2, the fluid filled semicircular canals stand out from the bone, and one can usually see that the canal goes "right to the dura" -- in other words, there is no dark bone between the loop of the canal and the brain. The image above shows a man with SCD identified from his very large VEMP on the right side. His CT-temporal bone showed SCD.

An example of a coronal with no SCD is here (T2, 4 mm cut size): Note that there is space between the loop of the canal and the brain. Thus no SCD.

Coronal no SCD

The huge advantage of using MRI is that there is no radiation. Temporal bone CT scans need a lot of X-ray radiation to see through the hard bone of the temporal area. This problem may eventually be handled through use of the newer cone-beam CTs.

On the other hand, MRI is more expensive and slower. We think MRI is underutilized, as most patients get MRI's for other purposes. We suggest that the proper MRI should be a 3T (high field) MRI, with T2 direct coronal cuts, with resolution ideally of 3 mm. Browaeys et al (2013) used higher resolution Fiesta imaging in a lesser field strength scanner (1.5T). We are not sure which is better, but we think the 3T T2 images are very reasonable.

An MRI is also useful to exclude potentially confounding entities such as cholesteatoma or tumor. MRI is not as good a test for dehiscence as a temporal bone CT because it doesn't show the bone and the resolution is not as good as a temporal bone CT scan. However, MRI is the best way of showing other possibly confounding problems such as acoustic tumors, cholesteatoma, or multiple sclerosis plaques. Note that "basic" CT scans such as are done in emergency rooms are always useless for diagnosis of SCD.

Emerging tests for SCD:

Click-evoked VOR -- never became popular

This test was described by Halmagyi and others (2003). Event triggered averaging is used to detect electro-oculographic responses to loud clicks -- intensities ranging from 80 to 110 Db. 128 clicks were delivered at a rate of 5/s from 60 to 110 db, in 10 db steps. Normal subjects have no response or a very low amplitude response of < 0.25 deg at 110. The latency was 8 msec. This test is not generally available, and since there have been no adopters since 2003, this is probably not going to be pursued.

MRI (3T 4 hrs post contrast) -- irrational.

Sone et al (2015) reported that of 5 patients with SCD, 4 of them showed "severe hydrops" on their MRI. We find this finding difficult to understand, but bears repeating. We think the 3T MRI method is "bleeding edge" with respect to vestibular diagnosis. In other words, don't make any serious decisions based on 3T MRI findings of hydrops.

See also: SCD (overview) SCD diagnosis. SCD treatment. SCD references