Complications of Surgery for Chronic Otitis Media

Preoperative Counseling

Each patient should be informed of the indications, alternatives, and potential complications of surgery. We describe the diagnosis and procedure in layman’s terms. We explain the probability of success in simple ratios, such as “8 or 9 times out of 10 we close the perforation with a gain in hearing.” We then explain the potential risk of partial or total sensorineural hearing loss, dizziness, facial weakness, and other complications of mastoid surgery. We have used a handout that follows the example outlined by Sheehy to facilitate the discussion. With practice, this can be done in a way that does not sound negative. To do so, the patient must understand the goals of the surgery and how they will be accomplished. The principal goal of chronic otitis media surgery is to eradicate infection. The correction of hearing loss is secondary and is often delayed until a second procedure.

The preoperative counseling depends on the findings of a proper history and physical examination. A history of vertigo or sensorineural hearing loss is important. The condition of the tympanic membrane should be noted because central perforations are rarely associated with cholesteatoma, whereas marginal perforations or retractions are more problematic. We carefully inspect the retraction pockets of the pars flaccida or the posterosuperior pars tensa. A hearing loss of greater than 30 to 35 dB implies erosion of the ossicular chain, but hearing is often normal with posterosuperior quadrant cholesteatomas. The patient should be forewarned that sound transmission is occurring through the mass, and that hearing is likely to decrease after the initial surgery.

Labyrinthine Fistula Secondary to Chronic Otitis Media

Labyrinthine fistulas are exposures or (in a minority) violations of the endosteum of the labyrinth. They are the most common complication of cholesteatoma and traditionally have been present in 10% of cases, but this has decreased in some series to 3% to 7%. The most common location is the lateral semicircular canal in all published series, with isolated lateral semicircular canal fistulas in approximately three-quarters of cases. The remainder of fistulas in most series involved the lateral canal plus one or more other sites.

Labyrinthine fistulas are a serious operative hazard and their appropriate management begins with a heightened sense of suspicion. One should consider the possibility in every case. This has always relied on the preoperative signs and symptoms, but that has given way to high-resolution CT, which has been reported to have a sensitivity and specificity of nearly 100%. ^, That is not to say that scans are perfect, or always available, so we do not think that we should readily forget the hard-earned lessons of our predecessors.

Those with a long history of chronic otitis media are more likely to have a fistula. Sheehy and Brackmann, reporting on 97 cases, noted that over 50% had a history of chronic otitis media for 20 years or more. Ritter, with 50 cases, also noted that many of his patients had lifelong otorrhea, often dating back to childhood. It follows that fistulas are less common in children but can still occur.

Vestibular symptoms are common with labyrinthine fistulas, which makes sense, although they are not universal. Sheehy and Brackmann found that almost two-thirds were dizzy, and that dizziness was constant in 12%. Ritter said that 76% of his patients complained of vertigo, which was usually brief, lasting seconds to minutes. McCabe reported vestibular symptoms in 90% of 70 cases, the highest prevalence of any larger series. In addition, there was a high prevalence of positive fistula tests (72%). He gave an outstanding description of the test along with the anticipated eye movements for different fistula locations ( Table 20.1 ).

With a positive fistula test of a postampullary lateral canal fistula, air compression into the ear displaces the endolymph and cupula toward the vestibule. This stimulates the crista ampullaris and causes conjugate deviation of the eyes toward the opposite ear, and fast-beat nystagmus toward the tested ear with sustained pressure. If the pressure is pulsed, there is still deviation to the opposite ear, but the eyes drift back to the midline with pressure release. If the fistula is located at the junction of the lateral canal ampulla and vestibule (preampullary region), positive pressure displaces the cupula away from the vestibule with deviation of the eyes toward the diseased ear. Positive pressure on a fistula of the vestibule moves the horizontal and vertical canal ampullae, causing horizontal-torsional eye movements toward the diseased ear.

Sensorineural hearing loss is another sign of a fistula. There appears to be a binary distribution between mild and profound hearing loss when hearing loss is present. A recent series reported that 47% had a bone conduction threshold of 11 to 20 dB, while 24% had anacusis. This is strikingly similar to Ritter’s findings in 1970, and other classic series report preoperative anacusis in 12% to 15%. ^{, ,} This includes a series of 97 fistulas by Sheehy et al., which remains the largest to date. This literature reports that sensorineural levels are more often diminished with extensive fistulas, especially at sites other than the lateral canal. However, extensive labyrinthine destruction does not invariably cause hearing loss. Bumsted et al. reported four cases of extensive labyrinthine destruction in which the vestibular function was gone, but the hearing remained after surgery. They suggested that the inflammatory response sealed the cochlea and protected it. There is at least one report of a labyrinthine fistula with preoperative anacusis and a large hearing gain after surgery. The hearing loss was considered to be from a serous labyrinthitis.

Preoperative facial weakness, although rare, is more common with fistulas than without: 4% versus 1%. Facial nerve dehiscences are present in a quarter to half of cases, ^{, ,} so there is an increased risk of intraoperative nerve injury with fistulas.

Comparisons of the outcomes require consistent reporting of fistula size. Most case series appear to use the convention used by Gacek. He classified fistulas as either small or large, with 2 mm as the delineation. He admitted that this was somewhat arbitrary, but he posited that the bony margins could support the cholesteatoma matrix when fistulas were smaller than 2 mm, and thus sustain more surgical maneuvering. Sanna et al. classified fistulas as small (0.5 to 1 mm), medium (1 to 2 mm), and large (>2 mm). Alternatively, they may be staged by the degree of labyrinth involvement, as proposed by Milewski and Dornhoffer: type I, exposure of the endosteum; type II, violation of the endosteum (perilymphatic exposure); and type III, disturbance of the membranous labyrinth.

Intraoperative Management of Labyrinthine Fistulas

There are several strategies for the management of fistulas. The most conservative approach is to leave the matrix down and exteriorize the fistula via a canal wall-down mastoidectomy. Others have advocated for a canal wall-down procedure, but to minimize postoperative vestibular symptoms will remove the matrix and repair (or even plug) the fistula. ^, Otherwise, the matrix is removed and the fistula is repaired via a canal wall up mastoidectomy, in either one or two stages (4 to 6 months later), ^{, ,} depending on the extent of disease.

The suitability of leaving cholesteatoma matrix over a fistula versus complete removal has been controversial since the 1950s, ^, and there is apparently no end in sight as there are no prospective studies, and a recent meta-analysis pooled 28 studies and found no significant difference in the sensorineural hearing outcomes between these strategies. However, this also shows that the differences are not so stark, and it has been suggested that these lesions are not as hazardous as previously considered. Our perspective is that surgeons experienced enough to publish large series generally achieve good outcomes when they treat fistulas with respect. This review also revealed a trend toward matrix removal over the past two decades. We have used all of the above strategies, but have found that other factors (i.e., a contracted mastoid or defect in the external auditory canal wall) often determine the appropriateness of an open cavity, and the presence of a fistula is one of several variables that should be considered.

The successful management of a fistula requires coupling good judgment to good technique and there is some consensus of practice:

• • Open and evacuate any cholesteatoma sac in the mastoid ( Fig. 20.1 ); if unsure, carefully palpate the medial wall to detect any bony erosion, especially on the dome of the lateral canal.

    

  • Leave the matrix over any fistula to protect it, even if planning removal, and focus on the rest of the disease. Commence removal immediately before closing. When exposed, quickly cover the fistula with a tissue graft ( Fig. 20.2 ). The risk of sensorineural hearing loss is much higher if the fistula is not covered.

    

  • Matrix removal is safe with lateral canal fistulas less than 2 mm in diameter, but only an experienced surgeon should attempt removal when a fistula is larger.

  • The dissection should be stopped if the matrix appears adherent to the membranous labyrinth.

  • Strong caution is advised whenever fistulas are extensive, multiple, or involve the vestibule or cochlea ( Fig. 20.3 ), and some authors believe that these should always be left alone. ^,

    

  • Leave the matrix alone in an acutely inflamed or better-hearing ear.

Hydroxyapatite bone cement is a relatively new way to resurface fistulas. The matrix is elevated off the fistula, after which fascia and cement are placed over it. Once set, another layer of fascia is placed over the cement. We have found that this technique reduces vestibular symptoms with barometric changes, although there are no reports of outcomes so far. For similar reasons, some authors reduce the barometric symptoms by mastoid cavity obliteration with bone pâté, or by plugging the canal (although this will obviously cause a canal paresis).

Postoperative hearing results have been variable and most classic case series cannot be compared due to nonstandardized reporting. However, a recent meta-analysis of 617 subjects from 28 articles gives insight. When hearing preservation was defined as a change in the bone conduction pure-tone average of <10 dB, this was achieved in 86% of the matrix removal group and 95% of the matrix exteriorization group, although this was not statistically significant. It also did not matter if the procedures were staged. Meanwhile, hearing was preserved in 90% of fistulas ≥2 mm, 77% of cochlear fistulas, and 50% of multisite fistulas. None of these were significant differences, and were likely a type II error due to small numbers. A subsequent series also had difficulty correlating hearing preservation with the fistula extent. This is at least partially because of survival bias, as most extensive fistulas cause preoperative deafness. Furthermore, the loss of hearing during surgery seems to relate more to the presence of a violation of the membranous labyrinth rather than to the size of the fistula. Even then, when hearing remains despite an extensive fistula or outright labyrinthine invasion, partial labyrinthectomy with hearing preservation is possible. ^,

Cochlear promontory fistulas are rare and represent approximately 1% of fistulas. Despite the 77% hearing preservation rate reported above, they are notorious for poor outcomes and have been one of the few absolute indications for radical mastoidectomy. Sheehy and Brackmann had a 56% incidence of severe or total sensorineural hearing loss when they attempted matrix removal from “extensive” fistulas. Gacek reported a partial to profound hearing loss in three cases in which he removed the matrix from cochlear fistulas. He believed that the membranous semicircular canals and their ampullae have thicker, more rigid walls than do Reissner’s membrane or the basilar membrane, and are less likely to rupture with manipulation. In addition, the cochlear duct (scala media) is in the outermost portion of the bony cochlea, so it is more closely in contact with the overlying matrix than the membranous labyrinth is elsewhere.

Iatrogenic Labyrinthine Fistula

The three sites at the most risk for iatrogenic fistula are the lateral semicircular canal, promontory, and oval window. Palva et al. reported iatrogenic injury to the lateral canal in the absence of a cholesteatoma fistula in 0.1% of chronic otitis media cases (2 of 2192). A review by Canalis et al. reported a 0.08% incidence of this mishap. They observed that there was an acute, moderately severe sensorineural loss that ultimately had a favorable outcome, and hearing returned close to preoperative levels after 3 to 6 weeks. A permanent, mild, high-frequency loss was common, but the speech discrimination scores returned to normal, and tinnitus was rare. Vertigo and nystagmus occur as anticipated, but commonly last beyond the usual period of central compensation. Patients sometimes have unsteadiness, positional vertigo, and spontaneous nystagmus 1 to 2 years after surgery. This is consistent with a partial labyrinthine injury along with continued neurosensory activity. To summarize, an iatrogenic breach of a semicircular canal is treacherous, but the hearing is not always lost. However, one must recognize the injury and seal the opening.

Accidental opening of the inner ear is more common at the oval window. Palva et al. had a 1.4% incidence. Of their 12 cases, two had complete avulsion of the footplate, three had dislocation, and the rest had fractures without dislocation. There were no drastic hearing losses, but three had depressed bone conduction thresholds several months later. In a 5-year follow-up report, they opened the oval window once in an additional 1362 cases, for an overall incidence of 0.5%.

The relative retention of hearing in most iatrogenic fistulas at the oval window supports the findings of an earlier report by Weichselbaumer. He had 27 cases of accidental fenestration of the oval window, but no dead ears. Only two had new sensorineural losses, each by 20 dB. Likewise, Sheehy and Brackmann, in their paper on labyrinthine fistulas from cholesteatoma, reported accidentally opening the oval window in 11 cases. These events did not cause sensorineural impairment.

We offer several guidelines to protect the oval window, as follows:

• • Remove cholesteatoma by dissection parallel to the stapedius tendon to steady the stapes ( Fig. 20.4 ). Manipulations of the stapes in a superoinferior direction or depressing it can disrupt the annular ligament ( Fig. 20.5 ).

    
    

  • Be cautious about removing any granulation tissue or inflamed mucosa from around the stapes. This area naturally heals when the rest of the middle ear is disease-free and sealed by tympanoplasty. When needed, a laser is useful for the atraumatic removal of tissue from the stapes and oval window. ^,

  • If the dissection becomes difficult, leave the cholesteatoma, reconstruct the tympanic membrane, and perform a second stage operation in 6 to 9 months. Any inflammation should be resolved and there usually is an epithelial pearl that is easily removed. In addition, there is less risk of inner ear contamination and hearing loss if the labyrinth is breached during the second stage.

  • Immediately cover any opening with tissue, avoiding any suction. Palva et al. recommend delaying ossicular reconstruction until the area heals.