Timothy C. Hain, MD Page last modified: October 3, 2017
Common surgical approaches to the vestibular nerve (middle fossa approach is not shown here).
(Image from Jackler R, Atlas of Neurotology and Skull Base Surgery (Mosby 1996, First edition, with permission).
The first VNS was performed by Performed by Perry in 1904, in an attempt to control symptoms of Meniere's disease. The method was the "middle fossa", which is to go through the side of the skull. In 1912 Frazier used the suboccipital approach (back of the skull). In 1936, McKenzie performed a selective vestibular nerve section. In 1961, House popularized the middle fossa approach again. Pulec, Fisch, Garcia-Ibanez and Aristegui proposed technical modifications. In 1971, Pulec first described the retrolabyrinthine approach and its use for the removal of acoustic neuromas. In 1980, Silverstein and Norell popularized this approach for vestibular nerve section. In 1987, Silverstein adopted the retrosigmoid approach. In 1990 Vernick reported on two cases where the translab approach was used for VNS.
All of these procedures are rapidly falling from favor for Meniere's, because a newer procedure -- transtympanic gentamicin, is replacing them in most centers. Nevertheless, they are still useful in individuals in whom gentamicin fails, or in whom the site of lesion is between the inner ear and brainstem (i.e. nerve lesions).
The usual reason for a vestibular nerve section is control of intractable unilateral Meniere's disease, where there is some hearing (better than 80 DB PTA). Considering the relative safety and efficacy of gentamicin treatment, this indication is very weak for VNS in Meniere's. We don't think it should be done until gentamicin has failed.
On extremely rare occasions VNS is performed for other reasons than Meniere's such as recurrent or very refractory vestibular neuritis. Pappas and Pappas (1997) reported using VNS to treat Meniere's and vestibular neuritis. In our practice, we have seen two cases of refractory perilymph fistula (PLF) improve dramatically and go back to work after vestibular nerve sections. This is a big deal !
Relative contraindications are bilateral vestibular disease, old age, poor medical condition, or CNS involvement. Any disorder that can be treated medically, should be so treated, before an irreversible procedure such as VNS is undertaken. As it seems likely that hair cell regeneration technology will eventually be available, another reason NOT to do a VNS is to preserve the connection between the inner ear and brain. Hair cell regeneration would not be likely to work after a VNS.
Where there is hearing to preserve, the main surgical alternative to vestibular nerve section is transtympanic gentamicin (TTG) treatment. This would be an irrational procedure however if the site of lesion were known to be in the vestibular nerve itself (such as in vestibular neuritis). Hillman et al (2004) suggested that VNS is superior to TTG. We disagree with them, because we think that TTG has been refined and is superior because of low risk and high efficacy.
If hearing is not an issue then a labyrinthectomy can be done (for Meniere's or PLF). Results of labyrinthectomy are very similar to vestibular nerve section (Eisenman et al, 2001). In general TTG treatment seems more reasonable than vestibular nerve section or labyrinthectomy in most cases because of the lack of major complications and also because of the lack of a need for general anesthesia. There are only rare exceptions to this general rule.
The vestibular nerve may be sectioned (cut or severed) via the middle fossa, retrolabyrinthine and retrosigmoid approaches, with similar efficacy. Surgical treatment is traditionally felt to be indicated when the patient is incapacitated with unilateral Meniere's disease and quality of life is affected. Historically 20% of patients eventually have had surgery done (Silverstein and Rosenberg, 1992), but with the advent of transtympanic gentamicin treatment, operative treatments are being replaced by outpatient procedures.
On the diagram below, one can visualize the retrolabyrinthine approach, which entails an attempt to cut the nerve from on the right side of this diagram, going through the mastoid sinus and into the cranial cavity to get access. The picture on the right was taken by Dr. Alan Micco, and shows the actual appearance of the surgery.
The middle fossa approach is similar, but access to the cranial cavity is obtained with the assistence of a neurosurgeon, from a slightly higher location (Glasscock et al, 1977). The middle fossa approach is somewhat more dangerous and requires more expensive equipment (e.g. stealth CT). It does provide better visualization however. It is sometimes used in people who fail the retrolabyrinthine approach, as it provides better visualization (Green et al, 1992).
In the retrosigmoid approach, access to the nerve is obtained by going through the posterior (back of head) part of the skull. According to Silverstein (2002), the combined RRVN -- a variant retrosigmoid approach, is the most effective at this writing.
|Vestibular nerve section -- operative field (courtesy of Dr. Alan Micco)|
After a VNS, there should be no vestibular responses from the sectioned side. In particular, there should be no caloric response, and no VEMP. Rotational tests and posturography tests are not very sensitive to VNS. Follow this link to see the expected test results for VNS.
Hearing is usually slightly impaired on the operated side, because there are some hearing fibers that go through the vestibular nerve. Colleti et al (2007) reported that hearing deteriorated (on average) from 48.5db down to 50.3 db down in their group of patients. This suggests that if one starts with a moderate hearing loss, it is little changed by surgery. To us, this drop seems unreasonably small considering our clinical experience. It would seem likely to us, though, that if one had "more to lose" -- i.e. started with better hearing, there would be a greater functional drop.
These authors also reported that gentamicin treatment was accompanied by greater deterioration (50.1 to 74.7) than VNS. However, these results could be caused by use of a high dose gentamicin protocol, which presently seems imprudent. Thus, we presently are dubious that VNS has better hearing results than low-dose gentamicin.
Given that there was vestibular function prior to the surgery (which is usually the case, as otherwise why bother), there should be vertigo immediately post surgery. This nearly always gradually abates. The duration of the post-operative vertigo is variable -- those who have already lost most of their vestibular function due to the disorder that prompted the VNS, may only have a short period of dizziness and vertigo post surgery. Those who had good vestibular function on the side of VNS prior to surgery, experience a bigger change after the nerve is cut, and may have post-operative dizziness and vertigo for a longer period.
Although 95 percent of patients are satisfied with the procedure immediately after their surgery, there are substantial risk of a serious complication in all of the described surgical procedures. All available procedures -- retrolabyrinthine, retrosigmoid, and middle-fossa have significant risks. For this reason, transtympanic gentamicin treatment is gradually replacing all of these operations.
As mentioned above, there is an expectation of acute vertigo following surgery as well as a modest decline of hearing on the operated side. These are not really complications -- which implies that they are unusual, but rather are expected consequences. Over the long term, following VNS one should expect a permanent modest reduction in balance. In theory, as one has suddenly lost 50% of one's vestibular system, persons with VNS may become less steady at an accelerated rate as they age.
Failure of procedure: Thedinger and Thedinger (1998) found that 20% of 142 VNS failed to relieve vertigo. Reasons included incomplete section, poor compensation, new vestibular disease in the opposite ear, and various other causes.
Regarding incomplete section, although from the name of the procedure one would expect that the entire vestibular nerve would be cut, in reality this is not always possible. According to Eisenmen (2001) there is evidence for retained vestibular function in about half of patients following nerve section or labyrinthectomy. We have not encountered this ourselves however. Some of the fibers of the vestibular nerve run very close to the cochlear (hearing nerve), and because of this they may be spared. Saccule derived nerve fibers may be purposefully spared because they tend to run close to the cochlear nerve (Silverstein et al, 1994).
Head impulse testing in persons with failed VNS often reveal residual function in the posterior canal (Lehnen et al, 2004), which is innervated by the inferior vestibular nerve, which runs with the saccule fibers. However, we are somewhat dubious that HIT testing can be trusted about any axis than horizontal. Sometimes there is an attempt to cut these fibers at another site with a singular neurectomy. Singular neurectomies, however, are somewhat difficult and unreliable even in very experienced hands.
Neuromas may form from the cut stub of the nerve, and nerve may regrow in a few individuals (Pulec and Patterson, 1997).
CSF leak. About 10% of cases with retrolabyrinthine nerve sections develop a cerebrospinal fluid (CSF) leak, which is treated with continuous lumbar drainage for several days.
Facial weakness. The facial nerve runs adjacent to the vestibular nerve, and it can be damaged by procedures that intend to damage the vestibular nerve. This complication is rare in recent times due to better surgical technique and the availability of facial nerve monitors.
Headache. Postoperative headaches are common in the retrosigmoid approach, and about 25% of patients undergoing this procedure there are severe headaches requiring medication 2 years later.
Other: The middle-fossa approach is essentially a neurosurgical approach. The skull must be opened and the brain retracted. Neurosurgical procedures intrinsicially have considerably more risk than those where the brain is not exposed. Our view is that middle fossa nerve sections are rarely indicated, as transtympanic gentamicin treatment has similar effectiveness with much less risk.