Acute pneumolabyrinth: a rare complication after cochlear implantation in a patient with obstructive sleep apnoea on CPAP therapy

Background

Pneumolabyrinth is a condition characterised by the presence of air in the perilymphatic space of the inner ear. This condition, first described by Mafee et al in 1984, is a radiological finding characterised by the presence of air bubbles localised indifferently at the level of the cochlea, utricle, saccule or semicircular canals.1 In most cases, pneumolabyrinth results from the formation of a perilymphatic fistula due to a traumatic event such as fractures of the temporal bone and barotrauma, or as an iatrogenic complication after ossiculoplasty, stapes surgery or cochlear implantation in patients with inner ear malformations. This condition is instead rarely described in subjects not affected by cochlear malformations.2 We report the case of a pneumolabyrinth affecting the left vestibule developed as acute complication after left cochlear implantation in a patient with bilateral severe-profound sensorineural hearing loss and obstructive sleep apnoea (OSA) on continuous positive airway pressure (CPAP) therapy.

Case presentation

A middle-aged male patient presented with congenital severe-profound sensorineural hearing loss on both sides to receive a cochlear implantation to restore hearing. The patient was treated with hearing aids in childhood. Its medical history reported high blood pressure, atherosclerosis, a previous episode of bulbar ischaemia and OSA on CPAP therapy. The patient did not report any traumatic events, use of ototoxic drugs or family history of hearing loss. Physical examination was normal in all the districts, and no signs of malformations of the inner or middle ear were detected at MRI and CT scans. Preoperative diagnostics were performed which did not reveal any contraindications to surgery. After receiving informed consent, the patient underwent left cochlear implantation. The surgical approach was performed by a posterior tympanotomy with insertion of the electrode through a cochleostomy obtained by means of a 1 mm burr. After electrode insertion, the cochleostomy was obliterated with autologous muscle tissue. CPAP therapy was administered during the first night of hospitalisation, as per protocol followed by the patient before the surgery. The day after surgery, the patient reported onset of left high-pitched tinnitus, severe rotatory vertigo, nausea and vomiting.

Investigations

The clinical vestibular examination included: searching for spontaneous nystagmus with video-Frenzel, Dix-Hallpike test and supine roll test to exclude positional vertigo; head shaking test (HST), head impulse test (HIT), skew test, Romberg test and Unterberger-Fukuda test. The examination showed a spontaneous right-beating horizontal (grade 3) nystagmus with HIT, HST, Romberg test and Unterberger-Fukuda test, revealing a left vestibular deficit. The skew test was negative. To confirm the hypothesis of a dislocation of the cochlear electrode from the scala tympani to the scala vestibuli, a cone-beam CT scan of the left ear was performed. Through the study of the images in axial, coronal and sagittal projection and through the 3D reconstructions of the inner ear, it was possible to detect the correct positioning of the cochlear electrode in the tympanic scale with an insertion angle of about 400°. The CT scan allowed to exclude a dislocation of the electrode. The evaluation of the posterior labyrinth led to the identification of a pneumolabyrinth characterised by the presence of multiple air bubbles located mainly at the utricular level and resulting as the main cause of the symptoms complained by the patient (figure 1).

Figure 1

Coronal and axial cone-beam CT images.

Treatment

The patient was treated with conservative care including bed rest, antibiotics (ceftriaxone 2 g/day, for 1 week), corticosteroids (prednisone 25 mg, 2×1/day in the first 3 days, then tapered over 10 days) and vestibular suppression (cinnarizine 20 mg + dimenhydrinate 40 mg 2×1/day for 10 days). After 3 days of therapy, the patient reported a reduction of symptoms, and therefore, 5 days after the surgery, he was discharged from the hospital with the recommendation to continue bed rest and medical therapy, as prescribed.

Outcome and follow-up

A follow-up was carried out to understand the evolution of the vestibular deficit. One month after surgery, the patient reported resolution of symptoms with disappearance of the tinnitus and persistence of mild dizziness. The evaluation was performed by means of video HIT (vHIT) and by measuring the cervical vestibular evoked myogenic potentials (c-VEMPs).

We used the Eye-See-Cam system (Interacoustics, Middelfart, Denmark) to record vHIT. The patient was seated 1.5 m in front of a target and was asked to continue watching it as his head was passively rotated by the examiner. The head was exposed to passive high acceleration and low-amplitude rotations in the planes of the horizontal semicircular canals. The patient’s eye movements were evaluated using video-oculography, while his head movements were recorded using inertial sensors. At least 15 valid head impulses were recorded for each horizontal semicircular canal. The vestibulo-ocular reflex gains during the vHIT (eye velocity/head velocity) were automatically measured using a software that computed the slope of the regression between head and eye velocity. Abnormal functioning of the canal was taken into consideration in case of gain asymmetry >5% and mean gain in vHIT <0.8 for the lateral canals with detection of catch-up (corrective) saccades.3 The examination detected a gain of 0.76 on the right side and 0.02 on the left side with the presence of overt and covert saccades on the left and asymmetry of 82%, as shown in figure 2.

Figure 2

Video head impulse test.

c-VEMPs were recorded with Socrates (Hedera Biomedics S.r.l., Taranto, Italy) with biofeedback of the muscle contraction on the monitor. Biofeedback was used to reduce the variability in case of difference in muscle contraction between the two sides and was achieved through a digital LED and displayed waveform rectification for a realistic comparison between the ipsilateral and contralateral amplitudes.

The negative electrode was placed at the middle third of the sternocleidomastoid muscle ipsilateral to the acoustic stimuli, the positive electrode was placed on the middle third of the clavicle and the ground electrode on the forehead; the patient was lying in a supine position with the head rotated contralateral to the acoustic stimuli, and the recording started by asking the patient to flex the head to the sternum by maintaining the position during the examination. The acoustic stimuli were air conducted by earphone inserts as tone burst 500 Hz, 130 dB sound pressure level and repeated 200 times. The presence and repeatability of the curve as well as amplitude and latency of the p13–n23 complex were evaluated; the asymmetry between the two sides above 33% was considered as pathological.4 As shown in figure 3, it was not possible to identify p13 and n23 waves on the left side. The results of the diagnostic tests therefore confirmed the presence of a left vestibular hypofunction.

Figure 3

Cervical vestibular evoked myogenic potentials.

The patient was evaluated again at 3 and 6 months. During both the evaluations, the patient reported the persistence of mild dizziness. The vHIT test and c-VEMPs test confirmed the persistence of a left vestibular hypofunction. At the end of each evaluation, the patient was advised to undertake vestibular rehabilitation.

From an auditory rehabilitation point of view, the activation and fitting of the cochlear implant took place regularly according to the Italian guidelines.5