Management of labyrinthine fistulae in chronic ear surgery

Brief history 

The initial treatment of a labyrinthine fistula with surgery has been attributed to Nylen1 in an article from 1923. These initial treatments involved removal of the matrix from the fistula site and leaving the site uncovered. This approach was generally against the trends of the time, which were to leave the matrix of the cholesteatoma over the fistula site. With advances in otologic surgery in the 1950s, fenestration operations for otosclerosis became common, instrumentation and magnification improved, and still the debate remained as to whether to remove the matrix or leave it in situ. Although further advances have been made in the understanding of the inner ear, the debate over fistula management is far from settled today. Ample studies provide evidence for both complete matrix removal and matrix in situ techniques.

Histopathology and pathogenesis 

Temporal bone collections have yielded histopathologic information important in understanding the development of labyrinthine fistulae. In a study of temporal bones, it was observed that labyrinthine fistula could be created by both cholesteatoma and chronic granulomatous otitis media without evidence of cholesteatoma.2 Furthermore, fistulae are often present without reactive changes to the inner ear histologically. New bone formation was observed at the margins of the fistulae, indicating the ability of the surrounding bone to repair erosive processes. The source for the new bone formation has been postulated to be the endosteal or periosteal layer of the otic capsule.2

Proposed mechanisms for bone erosion from this study were 2-fold. One mechanism postulates chronic infection leads to resorption osteitis from osteoclastic activity in the inflammatory layers of debris or cholesteatoma. A second mechanism observed was noninflammatory osteolysis. This bone destruction appeared to be secondary to direct pressure or production of osteolytic compounds by the cholesteatoma matrix. Proposed osteolytic mediators include collagenase, cathpepsin, and alterations in pH.2

More recent experiments have shown that bone resorption in the setting of cholesteatoma is caused solely by osteoclastic activity.3 Inflammatory cells and cytokine mediators have been shown to activate osteoclasts in cholesteatoma environments. Multiple compounds have been implicated in osteoclastic activation including interleukins 1 and 6, tumor necrosis factor α and β, nitric oxide, and prostaglandins. The final common pathway for bone destruction involves acidic pH, proteolysis via cathpepsin and matrix metalloproteinase 9, and tartrate-resistant acid phosphatase degradation of collagen.3

Fistula staging 

In an attempt to clarify the accumulating data regarding labyrinthine fistulae, a staging system was proposed by Palva and Johanson4 in 1986, categorizing fistulae by the severity of otic capsule destruction. In the preliminary stage, there is bony erosion of the otic capsule with a blue line of the labyrinth, and no true fistula exists. In stage 1 disease, the endosteal membrane is exposed and remains intact. In stage 2, the membranous semicircular canal is in contact with cholesteatoma matrix, and the perilymphatic space is open. In stage 3 disease, the perilymphatic space is open with direct involvement of the membranous labyrinth.4 This staging system has not been widely used in the literature.

A more common categorization of labyrinthine fistulae is by diameter of the fistula. Reported in millimeters, this information gives a sense of the destructive nature of the disease by reporting the size of the crater it creates in the otic capsule bone. Because the erosive process occurs in a saucerized manner, the diameter of the fistula partially indicates which structures may be involved. For instance, a diameter of less than 2 mm is unlikely to involve the endosteal membrane because the cholesteatoma matrix is still largely supported by the eroded edge of bone.

No published staging method accounts for depth of the erosive crater into the otic capsule or describes staging information for cases in which cholesteatoma matrix is left intact over the fistula site. Unfortunately, the majority of the data reported in the literature has no indication as to the stage of the fistula or the size of the bony defect created. The lack of a widely used staging system contributes to the difficulty in interpreting published results regarding fistula management, limiting definite conclusions, and management guidelines.

The management debate 

When reviewing the literature, there are 2 general approaches to the management of labyrinthine fistulae. One approach advocates complete removal of cholesteatoma matrix over the fistula site with repair of the bony defect. This has been done as both a 1-stage procedure as well as a 2-stage procedure with ultimate removal of the matrix over the fistula at a second-look operation. The rationale for complete removal of cholesteatoma matrix has been to remove a potential nidus for infection and source of continued bone resorption. Theoretically, the potential for subsequent suppurative labyrinthitis is completely removed.

The second approach to fistula management has been to leave a thin layer of the matrix in situ over the fistula site and exteriorize the cavity. The rationale for this approach holds that removing the cholesteatoma sac, alleviating its pressure effect, is enough to stop the bone erosion. Furthermore, advocates of this approach propose there is greater risk of sensorineural hearing loss with complete removal of matrix and an open labyrinth.

The data 

Reviewing the English-language literature for labyrinthine fistulae yielded 25 articles with patient data reported.5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29 Each study was retrospective in design and varied widely in the number of fistulae presented (N = 5-158 fistulae). Not all articles contained hearing outcome data crucial to compare methodologies of fistula management. In the tables are the consolidated available published data to compare the management strategies for labyrinthine fistulae (Tables 1-4).

Table 1. Reported incidence and location of labyrinthine fistulae

	Author	Year	Mastoidectomies	Fistulae	LSC	SSC	PSC	Promontory	Multiple Fistulae
	Ritter	1970	692	50	48	4	1	2	5
	Palva	1971	830	18	16	0	0	2	0
	Gacek	1974	168	14	12	3	0	5	6
	Abramson	1974	212	14	12	0	1	3	2
	Tos	1975	292	14	14	0	0	0	0
	Sheehy	1978	1,024	97	83			3	14
	Pfaltz	1982	222	19	10			7	4
	McCabe	1983	792	79	59	2	6	4	8
	Gormley	1986	684	33	33	3	2	2	7
	Smyth	1987	230	23	22	1	0	1	1
	Sanna	1988	1,226	158	144			14	17
	Palva	1989	453	84	76		2	6	0
	Ostri	1989	234	20	15				5
	Kobayashi	1989		5	5	0	0	0	0
	Glasscock	1990	12	4	4	1	0	0	1
	Parisier	1991	426	41	34	6	0	7	6
	Dornhoffer	1995	1,265	37	32	4	0	2	1
	Pulec	1996	1,233	63	51	6		8	2
	Herzog	1996		17	16	0	0	1	0
	Magliulo	1997	1,205	92	78	11	1	0	9
	Busaba	1999	625	34	36	2	1	4	4
	Manolidis	2000	111	23	21	2	0	2	2
	Gersdorff	2000	769	54	39	12	3	6	11
	SodaMerhy	2000	360	27	27	2		1	3
	Kvestad	2001	631	20	18	2	1	2	3
	Total (%)		13,696	1,018 (7)	884 (87)	61 (6)	18 (2)	81 (8)	111 (11)

Table 2. Reported signs and symptoms of labyrinthine fistulae

	Author	Year	+Cholesteatoma	+Vertigo	+Fistula Test	Facial Nerve Involvement	Preoperative CT Predictive
	Ritter	1970	50/50	38/49	23/35	19/50
	Palva	1971	18/18
	Gacek	1974	14/14			5/14
	Abramson	1974	12/14	1/14	11/14	1/14
	Tos	1975	14/14		10/14
	Sheehy	1978	97/97	64/97	30/46	48/97
	Pfaltz	1982	19/19		11/19
	McCabe	1983			57/71
	Gormley	1986	33/33	21/33	13/24	9/33
	Smyth	1987	23/23	15/23	9/15	4/23
	Sanna	1988	158/158
	Palva	1989	81/84		21/40
	Ostri	1989	20/20	13/20	10/20	11/20
	Kobayashi	1989	5/5	5/5	2/3	1/5	5/5
	Glasscock	1990	4/4				3/4
	Parisier	1991	41/41	30/40	9/40	7/40	8/15
	Dornhoffer	1995	37/37	23/37	12/37
	Pulec	1996	63/63			28/63
	Herzog	1996	17/17	9/17	7/17		3/8
	Magliulo	1997	92/92			3/92
	Busaba	1999	33/34	19/34	14/28		7/18
	Manolidis	2000	23/23			17/23
	Gersdorff	2000	54/54	31/54	5/54	4/54	26/54
	SodaMerhy	2000	27/27	21/27	15/27	3/27	22/24
	Kvestad	2001	17/20	10/20	4/20		11/20
	Total (%)		952/961 (99)	300/470 (64)	263/524 (50)	160/555 (29)	85/148 (57)

Table 3. Reported hearing in cases of labyrinthine fistulae

	Author	Year	Dead Ears Preoperatively	Hearing Improved or Unchanged Postoperatively	Hearing Decreased Postoperatively	Dead Ears Postoperatively
	Ritter	1970	15/50	15/24	9/24	0/24
	Palva	1971	0/18	15/18	1/18	2/18
	Gacek	1974		10/14	2/14	2/14
	Abramson	1974	0/14	12/14	0/14	2/14
	Tos	1975	3/14	10/11	0/11	1/11
	Sheehy	1978	12/97
	Pfaltz	1982	2/19	15/17	0/17	2/17
	McCabe	1983
	Gormley	1986	2/33	30/31	0/31	1/31
	Smyth	1987	1/23	13/20	7/20	0/20
	Sanna	1988	18/158	123/138	7/138	8/138
	Palva	1989	21/84	59/63	0/63	4/63
	Ostri	1989	3/20	13/17	3/17	1/17
	Kobayashi	1989	0/5	5/5	0/5	0/5
	Glasscock	1990		3/4	1/4	0/4
	Parisier	1991	6/41	28/35	7/35	0/35
	Dornhoffer	1995	4/37	26/33	0/33	7/33
	Pulec	1996	32/95	63/63	0/63	0/63
	Herzog	1996	0/17	16/17	1/17	0/17
	Magliulo	1997	5/92	73/87	11/87	3/87
	Busaba	1999	3/34	17/31	6/31	8/31
	Manolidis	2000	3/23
	Gersdorff	2000	5/54	43/49	1/49	5/49
	SodaMerhy	2000	4/27	20/23	3/23	0/23
	Kvestad	2001	4/20	13/16	3/16	0/16
	Total (%)		143/975 (15)	622/730 (85)	62/730 (8)	46/730 (6)

Table 4. Postoperative hearing results in labyrinthine fistula: Matrix removal versus matrix in situ techniques

			Matrix Removed			Matrix in Situ
	Author	Year	Hearing Increased or Unchanged	Hearing Decreased	Dead Ear	Hearing Increased or Unchanged	Hearing Decreased	Dead Ear	Takeback Without Matrix
	Ritter	1970	15	7	0	4	7	2
	Palva	1971	5	1	2	8	2	0
	Gacek	1974	9	1	2	1	1	0
	Abramson	1974	8	0	2	4	0	0
	Tos	1975	10	1
	Sheehy	1978							2/23
	Pfaltz	1982
	McCabe	1983
	Gormley	1986
	Smyth	1987				13	7	1
	Sanna	1988	36	2	2	88	5	1	25/37
	Palva	1989
	Ostri	1989	13	3	1
	Kobayashi	1989	4	0	0	1	0	0
	Glasscock	1990				3	1	0
	Parisier	1991	22	7	0	6	0	0
	Dornhoffer	1995	26	0	7
	Pulec	1996	63
	Herzog	1996	17	0	0
	Magliulo	1997	56	10	2	14	1	1
	Busaba	1999
	Manolidis	2000
	Gersdorff	2000
	SodaMerhy	2000	20	7	0
	Kvestad	2001
	Total (%)		304 (84)	39 (11)	18 (5)	142 (83)	24 (14)	5 (3)	27/60 (45)

Results 

In the 13,696 mastoidectomies reported, 1,018 labyrinthine fistulae were found at the time of surgery for an incidence of 7%. In ears with a labyrinthine fistula, a second fistula was present 11% of the time. The location of the labyrinthine fistula was predominantly of the lateral semicircular canal (87%) followed by the promontory (8%), superior semicircular canal (6%), and the posterior semicircular canal (2%) (Table 1).

Classically, patients with erosion of the labyrinth are described as vertiginous with a positive fistula test. Data on vertigo were reported in half of the articles describing labyrinthine fistulae. Of the articles reporting symptoms of vertigo, 64% of patients with labyrinthine fistula experienced vertigo preoperatively. Of patients with labyrinthine fistulae evaluated with a fistula test, 50% had a positive response. Only 15% of patients underwent preoperative CT imaging. In these patients, preoperative CT images were found to predict labyrinthine fistulae in 57%. Increased use of preoperative CT imaging in more recent studies was a clear trend in this review (Table 2).

Evaluation at the time of surgery revealed nearly all labyrinthine fistulae were attributable to cholesteatomatous disease (99%), with only 4 studies citing noncholesteatomatous processes leading to fistulization.8, 16, 25, 29 In these rare instances, fistulae were attributed to chronic infection without cholesteatoma,8 chronic granulation tissue,8, 25 and iatrogenic fistulization with otologic drilling16 (Table 2).

Advanced cholesteatomatous disease would be expected to jeopardize the bony protection and function of the facial nerve. In the literature reviewed, half of the studies reported facial nerve information gathered at the time of operation. From this subset of cases, 29% of labyrinthine fistula patients also had involvement of the facial nerve, defined as either erosion of the fallopian canal with exposure of the nerve or alteration in preoperative facial function (Table 2).

Hearing results are summarized in Table 3. Fifteen percent of patients with a labyrinthine fistula presented with a dead ear preoperatively. Of those undergoing operation with serviceable hearing, 85% were reported as having unchanged or improved hearing postoperatively. Eight percent of patients experienced decreased hearing postoperatively, with no clear indication of whether this loss was conductive or sensorineural in origin. Six percent of patients with a labyrinthine fistula experienced a dead ear postoperatively.

Detailed hearing data were available in a more limited extent in regards to the method of management of the labyrinthine fistula (Table 4). Hearing data were categorized as either complete matrix removal technique or matrix in situ technique for the management of the labyrinthine fistula. Data were available in approximately half the patients reported in the literature, and no staging information was reported for these patients. Three hundred sixty-one patients were managed with the complete matrix removal technique, and 171 were managed with the matrix in situ technique. The reported hearing results are equivalent for the 2 groups.

For patients managed with complete matrix removal, 84% experienced unchanged or increased hearing postoperatively, 11% were noted to have decreased hearing, and 5% lost hearing completely in the operated ear. In patients managed by maintaining matrix in situ over the fistula, 83% were noted to have unchanged or improved postoperative hearing, 14% experienced decreased hearing, and 3% lost hearing in the operated ear.

The anatomic site of fistulization was significant in the incidence of profound sensorineural hearing loss associated with erosion of the labyrinth. When fistulae were present in the semicircular canals, 10 of 368 (3%) patients experienced a dead ear. When a fistula of the promontory was present, 15 of 43 patients (35%) experienced a dead ear after operation.

In 2 of the reported series, information about staged operations was reported. In these articles, a piece of cholesteatoma matrix left covering the labyrinthine fistula was inspected at a second-look operation. In 45% of these cases, no residual matrix or cholesteatoma was observed, and interval bony repair of the labyrinthine fistula was evident at the second-look procedure10, 15 (Table 4).

Conclusions 

The incidence of labyrinthine fistula is 7% in reported series of mastoidectomies for chronic ear disease. Preoperative vertigo and a positive fistula test are poor predictors of the presence of an underlying fistula, with a sensitivity of 64% and 50%, respectively. In the growing data available on preoperative CT imaging, a fistula is again poorly predicted preoperatively in published series, although no prospective data are available. The sensitivity of the CT images is 57% in the reported series. With these facts in mind, an otolaryngologist must suspect a labyrinthine fistula in each case of chronic ear disease and be prepared to treat it effectively and safely.

Careful treatment of a suspected labyrinthine fistula is essential for hearing preservation. An unrecognized labyrinthine fistula is more likely to result in profound sensorineural hearing loss if it is inadvertently opened without proper preparation and treatment. Intraoperatively, areas of suspected fistulae should be widely exposed with high magnification, brilliant illumination, and meticulous hemostasis. The cholesteatoma sac and debris should be carefully removed to preserve the covering matrix. If the fistula is to be repaired, the covering matrix should be removed only under constant irrigation with immediate repair with previously harvested autogenous tissues. Suction of the open labyrinth is contraindicated.

There are 2 schools of thought in the literature about the management of a labyrinthine fistula. One approach advocates complete matrix removal over the fistula site, often in a staged procedure, with repair of the bony defect using usually autogenous tissues (temporalis fascia, bone chips or pate, fibrin glue, perichondrium). A second approach advocates preserving a small amount of in situ matrix over the fistula site and either externalizing the cholesteatoma or returning to the operating room for a staged procedure with controlled removal in an optimal setting. A review of the hearing results reported in the literature to date indicate that these techniques are likely equal in preserving hearing in patients with labyrinthine violation.

Interpretation of the data on hearing preservation must be considered carefully in light of disease severity. Without adequate staging information, one must presume more advanced disease is treated more often with a matrix in situ technique in attempts to preserve residual sensorineural hearing. Conversely, less destructive disease is more likely to be managed with complete removal techniques when the risk of sensorineural loss is lower. This management dilemma of which technique to use is reflected in the fact that 8 of the 15 reporting authors use both complete removal and in situ techniques for fistula management in the same studies. Therefore, the hearing preservation data in the literature likely represents similar results obtained in differently staged groups with more limited disease managed preferentially by complete matrix removal techniques and more involved disease managed preferentially by matrix in situ techniques.

Although opening the labyrinth and preserving hearing may seem counterintuitive at first glance, there is a growing body of literature to support these claims. Aside from the data reviewed here, there are reports of intentionally opening the bony labyrinth. This has been performed for treatment of benign paroxysmal positional vertigo via either occlusion of the posterior semicircular canal30 or partition with argon laser-induced fibrous scar.31 Also, reports exist of extended neurotologic approaches involving hearing preservation with partial labyrinthine destruction.32, 33, 34, 35, 36 Although the number of patients reported in these articles is relatively small (38 patients reported), the trend for careful manipulation and selective controlled destruction of the bony and membranous labyrinth is gradually increasing. A common theme in these reports with hearing preservation is to open the labyrinth and then seal it rapidly (with bone dust, bone wax, gel foam, fascia) to avoid suctioning or depleting the endolymphatic fluids. Although this has shown some promise, a number of dead ears have resulted from even these selective and careful manipulations (11 dead ears and 6 recoverable sensorineural losses).31, 33, 34, 37

Finally, the high incidence of dead ears in promontory fistulae deserves discussion. The cochlear duct lies in a lateral position in the cochlear turns. Erosion of the bone overlying the cochlea would therefore preferentially expose the spiral ligament and cochlear duct to injury. With this anatomic basis, severe sensorineural hearing loss would occur early in promontory erosion and appear more likely with matrix manipulation. Given the high incidence of dead ears with promontory fistulae, arguments can be made that promontory fistulae should be managed exclusively with the matrix in situ technique to increase the chances of hearing preservation.

Discussion from the audience