Chapter 27 Perilymphatic Fistula

Perilymphatic fistula (PLF) is a condition resulting from an abnormal connection between inner ear perilymph and the middle ear space. Its etiology has been described as being congenital, acquired, or idiopathic. Congenital middle or inner ear defects are believed to play a role in the development of PLF. PLFs exist in association with stapedectomies and other invasive procedures of the cochlea. Likewise, severe head injury, abdominal blows, and rapid shifts in environmental pressure are accepted causes of PLF. Acute and chronic mastoiditis with erosion into the labyrinth, and chronic granulomatous diseases, such as syphilis and tuberculosis, have also led to the development of PLF. In the absence of a clear antecedent traumatic event, idiopathic PLF remains a diagnostic and treatment challenge.

Iatrogenic poststapedectomy PLFs were first described in the mid-1960s. Steffen and colleagues ¹ reported findings of gross perilymph flow at the oval window in poststapedectomy patients with hearing loss, tinnitus, and vertigo. Fee² reported three patients who presented with vertigo, fluctuating hearing loss, and tinnitus, who also had known or suspected recent head trauma. Intraoperative findings showed perilymph leak at the oval window. Repair of the leak resulted in significant improvement in symptoms.

In 1971, Goodhill ³ coined the terms implosive (Valsalva-induced) and explosive (increased intracranial pressure) to describe pressure changes that can result in PLF. In reference to implosive, Goodhill stated that increased negative pressure from the tubotympanic region can be directed via the ossicles to the perilymphatic space. These external forces can result in a tear in the oval or round window membranes, driving air into the inner ear and displacing perilymph into the middle ear. The explosive route results from increased intracranial pressure possibly secondary to a severe blow to the head or abdomen. Increased intra-abdominal pressure transmits via the vertebral veins to increase the cerebrospinal pressure with resultant increase in intracranial pressure. Increased intracranial pressure can be transmitted to the inner ear via the cochlear aqueduct or internal auditory canal, potentially causing subsequent rupture, from inside out, of the oval or round window, or both.³

Homeostasis of the pressure differentials between endolymph and perilymph is maintained by patency of the endolymphatic duct and sac. If this pressure/volume balance is disrupted, as might occur in the case of a PLF, endolymphatic hydrops may result in damage to the hearing apparatus. Animal models with experimental PLF have shown hydrops, usually resolving within 3 weeks.^4,⁵ The cochlear aqueduct provides an important communication between the subarachnoid and perilymphatic space. The only outlet from increased intracochlear pressure is via a tear of the oval or round window membranes. Tears of the oval window annular ligament occur most frequently anteroinferiorly, rarely superiorly around the stapes footplate. The round window membrane tears less frequently.⁶ Tears of the round window membrane have occurred when its position was 45 degrees to the promontory, and there was little or no overhanging promontory, allowing direct visualization of the round window membrane transtympanically (Fig. 27-1B).^7,⁸

FIGURE 27-1 A, Right middle ear: surgeon’s view with tympanomeatal flap and drum folded forward. Sites of oval window fistula are numbered in order of occurrence (1 and 2 most common). B, Abnormal round window (RW) membrane at 45 degree angle to promontory and no overhanging lip. C, Normal RW membrane hidden from view in depths of niche, with lip of bone overhang. Mucosal folds frequently appear to seal niche partially, with a hole in the center. EAC, external auditory canal; PLF, perilymphatic fistula; TM, tympanic membrane.

The cribriform areas at the depths of the internal auditory canal are another area of potential transmission of increased cerebrospinal fluid pressure to the perilymphatic space. Rarely, these sites of communication are widely open, allowing direct connection of cerebrospinal fluid to perilymph, such as in Mondini’s malformation. Cerebrospinal fluid gushers can result in these patients after any invasive operative procedure of the cochlea.

Microfissures of the otic capsule in the area of the oval window, round window niche, and posterior canal ampulla have been previously theorized to lead to PLF. This theory was based on temporal bone studies by Kohut and colleagues,⁹ who noted “loose fibrous composition in the fissula ante fenestra” in patients with PLF. A follow-up study by El Shazly and Linthicum¹⁰ showed that these microfissures are commonly present in temporal bones. No association between the presence of microfissures and sudden sensorineural hearing loss was found, and no evidence of PLF was noted.

Animal studies have tried to examine the pathophysiologic mechanism contributing to the hearing loss noted in PLF patients. Weisskopf and colleagues ¹¹ generated perforations in the round window membrane of guinea pigs. The resultant hearing loss recovered with time, and the authors concluded that perforations alone did not explain the degree of hearing loss present in PLF. Some research has shown pneumolabyrinth to cause sizable auditory threshold shifts, with removal of the air bubble from the cochlea resulting in resolution of hearing loss.¹²

Early stapedectomy procedures were more likely to result in PLF than methods practiced today. The entire stapes and footplate were often removed and replaced with a pointed polyethylene strut or wire loop prosthesis placed atop a thin piece of absorbable gelatin sponge (Gelfoam) or tissue seal. The wire could migrate, and the thin tissue membrane could rupture, both increasing chances of PLF. The current small fenestra, small piston, tissue seal techniques have dramatically reduced the incidence of PLF.¹³^–¹⁷ Lesinski¹⁸ prospectively reviewed 19 cases of revision stapedectomy performed for symptoms of possible oval window fistula. The 14 of 19 patients who were noted to have an active PLF had undergone prior complete stapedectomy with migration of the prosthesis eccentrically, close to the edge of the oval window.

Invasive procedures of the stapes footplate, including the Fick or Cody tack procedure for Meniere’s disease (sacculotomy) were ultimately abandoned because of the high incidence of sensorineural hearing loss. Later exploration of many of these ears revealed a persistent PLF.¹⁹ Likewise, the 20% sensorineural hearing loss associated with cochleostomy or cochleosacculotomy, as it was originally described, was probably related to a persistent round window PLF.¹⁹

As awareness of PLFs increased, the number of diagnoses and surgical explorations for PLF did as well. Lower rates of positive exploration and of successful repair were being reported, however. In many cases of re-exploration for possible persistent PLF, prior operative records had described the presence of a hole in the round window membrane. Re-exploration findings showed a normal-appearing round window membrane deep in the niche (Fig. 27-1C). It is likely the prior surgeon had sealed the mucosal folds surrounding the round window niche.

Similarities in symptoms may have led to the misdiagnosis of Meniere’s disease or superior semicircular canal dehiscence as PLF. Attempts at identifying predictive tests, such as glycerin challenge testing or electrocochleography, to differentiate Meniere’s disease from PLF have been unsuccessful.²⁰^–²³ The classic symptoms of Meniere’s disease are episodic vertigo, fluctuant hearing loss, tinnitus, and, in 80%, a sense of fullness in the involved ear. Patching of windows in these cases with no visible leak would give the same result as allowing time to pass. In Minor’s²⁴ review of 65 patients with superior semicircular canal dehiscence, 54 (83%) had vestibular symptoms elicited by loud sounds, and 44 (67%) had pressure-induced (sneezing, coughing, and straining) symptoms. In addition, a 10 dB or greater conductive hearing loss was present in 70%. (See Chapter 42 for further discussion.)

Previous studies of children with sensorineural hearing loss of unknown etiology showed a rate of 6% to 16% positive PLFs on exploration.²⁵^–²⁷ In the largest prospective series by Reilly²⁵ of 244 children with sensorineural hearing loss, 17% underwent exploration, all of whom had prior abnormal computed tomography (CT) scan findings. Of the children who underwent exploration, 26% had active congenital PLF. A more recently published retrospective series reported 64% of ears with suspected PLF to have a leak confirmed visually.²⁸ Middle ear malformations have been noted in 81% of PLF patients at the time of surgery, most commonly an anomalous stapes.²⁹

PATIENT SELECTION

Patient presentation is variable with PLF. Symptoms have been reported to be present from days to decades. The typical patient with a PLF presents with a sudden onset of hearing loss or mild vertigo or dysequilibrium, or both hearing loss and vertigo, associated with a traumatic event. Trauma includes invasive middle or inner ear surgical procedures, abdominal or head blows, blast injuries, or severe changes in environmental pressure, particularly in the presence of an upper respiratory infection or an acute allergic attack. In some series, one third to one half of patients recall no potential triggering event.^30,³¹

In cases of surgically proven PLF, patients present with some degree of vestibular and auditory disturbances. Patients reported vestibular complaints including episodic or positional vertigo, dysequilibrium, lightheadedness, and motion intolerance 46% to 91% of the time.³¹^–³⁵ A chief complaint of hearing loss was present in 28% to 93% of patients.³¹^–³⁵ Although only 28% reported noting hearing loss in the Iowa series, 54% had abnormal audiograms.³² Hearing losses are most commonly reported to be sudden or rapidly progressive; however, in some series they were also noted to be fluctuant, raising debate as to whether the disease was actually Meniere’s disease.

Tinnitus has been reported as a chief complaint in 25% to 76% of patients with PLFs, always in combination with vestibular or auditory symptoms or both.^31,^32,^35,³⁶ In some series, no patients reported tinnitus as a presenting symptom.³⁴

PREOPERATIVE EVALUATION

Vestibular signs are commonly present in PLF. Positional nystagmus may have a very short or no latency and relatively long duration. Repeated testing shows nystagmus with minimal or no fatigue, no direction reversal when changing from inducing position to sitting position, and less violent than seen with benign paroxysmal positional vertigo. The nystagmus is rarely rotatory. From most to least frequent, the nystagmus may be horizontal, diagonal, or vertical.^27,³⁷ The direction of the nystagmus is of no diagnostic value in determining which ear is involved.

Hearing loss is typically sensorineural, without a specific pattern. Conductive hearing loss is rarely observed. This hearing loss has been attributed to a slipped stapes prosthesis or possible presence of air within the labyrinth. The speech reception threshold is usually worse than anticipated from the pure tone average, and the discrimination score is usually lower than expected.³⁸

Most series report usage of a fistula test by applying positive or negative pressure to an intact tympanic membrane with a pneumatic otoscope. Resultant nystagmus indicates a positive test and is present only about 25% of the time.⁴⁵ A sensation of dysequilibrium suggests a positive test. Modifications, such as administration of the test with the patient standing and eyes closed^30,³⁹ or on a moving posture platform (platform fistula test), have been used to try to improve test sensitivity. Electronystagmography has also been performed concurrently. Several studies noted no significant difference with electronystagmography-enhanced fistula testing.^34,⁴¹ High positive exploration rates have been reported in cases using the moving platform fistula test; however, these findings have not been supported by others.^42,⁴³

The Quix test involves having the patient stand erect with feet together, eyes closed, and arms outstretched. The examiner looks for deviation of the arms to the side of the lesion. The result is positive in only about 20% of cases.⁴⁴

The eyes-closed-turning test has shown high sensitivity and specificity for PLF. It is performed by having the patient walk in a straight line with eyes closed. The examiner taps the subject’s shoulder, indicating to the patient to turn 180 degrees either right or left and stop in a position of attention with the eyes still closed (Fig. 27-2). A positive test is readily recognized by the patient’s swaying or having a tendency to lose balance when he or she has turned to the side of the lesion.^45,⁴⁶