Surgical complications of cochlear implantation: A report of three cases and their clinical features.

Abstract

The most common complications of cochlear implantation are flap necrosis, compressed or improperly placed electrodes, and facial nerve injury, paresis, or stimulation. We report on three recent complications experienced at our institution: 1) a flap ischemia and partial-thickness necrosis without infection or device exposure, 2) facial nerve stimulation, which was managed with a new programming technique, and 3) a postimplant cholesteatoma in a patient with chronic eustachian tube dysfunction that required reoperation and explantation. We discuss the pathophysiology, likely etiology, and recommendations for management of these complications.

Introduction

Since 1988, several reviews [1-4] of complications associated with cochlear implantation have been published, including a comprehensive report by Hoffman and Cohen in l995. [5] The data for their paper were provided by Cochlear Corp., which markets the Nucleus 22-channel device; Advanced Bionics, which markets the Clarion multichannel implant; and Smith Nephew Richards, which markets the Ineraid prosthesis. Cochlear Corp. provided data on 4,969 implants, Advanced Bionics on 167 patients, and Smith Nephew Richards on 172 patients.

Flap-related problems continue to be the most common complications, although their rates have fallen from 5.44% in 1988 [1] to 2.79% in l995. [5] The incidence of compressed/misplaced electrodes has also declined, from 1.74 to 1.18%, as has the incidence of facial paresis, from 1.74 to 0.56%. [1,5] (The data in this paragraph pertain only to the Nucleus 22-channel device.)

The lowering of complication rates is no doubt the result of better training in cochlear implant surgery and more experience among surgeons. Nevertheless, flap necrosis, which is still the most frequently encountered complication, is feared by all surgeons because it can necessitate explantation. Only a few papers have addressed this complication as a single issue. [6-9] In their 1988 paper, Cohen et al reported that only one patient in a series of 459 implants (0.22%) experienced facial nerve stimulation following electrode activation. [1] Apart from the fact that this problem required explantation, no further details were provided about the anatomic status of the cochlea or the facial nerve and fallopian canal. In 1995, Hoffman and Cohen reported facial nerve stimulation in 101 of 4,969 implants (2.03 %). [5] What were the causes and underlying mechanisms? Only a few papers have addressed them, and in nearly all instances, the mechanisms proposed were speculative. [10-14]

The incidence and management of postimplant acute otitis media, persistent middle ear effusions, and postimplant cholesteatomas have not been substantively discussed in the literature. In 1991, Cohen and Hoffman made brief mention of an unpublished 1989 Cochlear Corp. survey on postimplant otitis media. [2] The parents of 133 children were surveyed, and 17 children (12.8%) were reported to have had at least one episode of acute otitis media. No adverse sequelae were noted. No mention was made of how the survey was conducted. Luntz et al conducted a study of 50 children who had received a cochlear implant and found that acute otitis media developed in only eight of these children (16%), all of whom already had a history of the disease. [15] From data provided to Hoffman and Cohen by Smith Nephew Richards, only one patient out of 172 (0.58%) developed a postimplant secondary cholesteatoma. [5] No mention was made of the patient's age (i.e., adult or child) or how the case was managed. In a review of 50 cochlea r implants published in 1995, Lloyd et al noted a postimplant retraction pocket cholesteatoma in one patient (2%). [16]

From this review, it is clear that although complications associated with cochlear implantation have been catalogued numerically, there is a paucity of detailed reports describing them clinically. Such information would help the surgeon avoid these complications. The purpose of this paper is to describe the clinical features and management of three postimplant complications.

Case reports

Case 1. A 57-year-old woman gave a history of bilateral progressive sensorineural hearing loss, which culminated in deafness within 5 months. She also had a history of severe rheumatoid arthritis, asthma, and anemia of unknown origin. She had undergone multiple fundoplications for weight loss.

The woman underwent a right cochlear implantation with a Nucleus 22-channel device. A postauricular C-shaped incision was made, and the integrity of the superficial temporal artery and the occipital artery was maintained. The procedure was well tolerated, but on the second postoperative day, the patient was noted to have a moderate degree of venous congestion, limited to the C-shaped flap (figure 1A). The area of congestion demarcated over the next 3 days to approximately 2 x 3 cm, and then it progressed to epidermolysis. The patient was discharged home on postoperative day 3 and was prescribed oral antibiotics and local wound care in the form of a 0.9% NaC1 solution wet-to-dry dressing change three times a day. An eschar that formed on the affected skin was left in place as a biologic dressing (figure 1B).

The wound responded well to local care, did not become infected, and gradually re-epithelialized. At no time during the postoperative course was the device exposed or at risk. The wound was judged to be closed and healthy on postoperative day 38.

Case 2. A 73-year-old woman with a history of glaucoma, coronary artery disease, and hypothyroidism reported that she had experienced a gradually worsening right-sided sensorineural hearing loss of unknown etiology over many years. At the time of her visit to our institution, the patient was unaidable. She had already lost the hearing in her left ear following a left radical mastoidectomy at another institution.

The patient therefore underwent a right cochlear implantation with a Nucleus 22-channel device. She was discharged home in good condition on postoperative day 2, and her subsequent wound healing was uneventful. Upon activating the implant, however, she felt a vigorous, painful stimulation of all branches of the ipsilateral facial nerve.

Her initial programming mode was bipolar + 1 (BP+ 1). In the BP+1 configuration, the electrode that is more distal in the apical direction to the active electrode is skipped, and the next apical electrode is used as the indifferent one. As the distance between the active and indifferent electrodes increases, the current density decreases, which lessens the possibility of facial nerve stimulation. Lengthening the distance, therefore, results in a need for fewer electrode pairs and, hence, decreases the frequency specificity.

Following the incident of facial nerve stimulation, the device was reprogrammed for every mode up to BP+5. Still, the facial twitching persisted. We then obtained a computed tomographic (CT) scan of the temporal bone. CT showed that the midbasal electrodes had come into contact with the labyrinthine segment of the facial nerve (figure 2A). We deferred further attempts to reprogram for 2 months in the hope that fibrous tissue would grow between the electrodes and the facial nerve. Unfortunately, even after 3 months, she continued to experience facial twitching in response to auditory stimulation.

At about the same time, Cochlear Corp. informed us of a new programming software it had developed. The Variable Mode Stimulation (DPS version 6.125) device allows the audiologist to specify which of the reference electrodes will be active and which will be indifferent in any electrode pair (figure 2B). This variable mode adds flexibility to the process because it increases the number of active channels. After reprogramming our patient's implant with the variable modes, we were able to widen the stimulation mode on all electrode pairs by changing the indifferent electrode. For example, with the standard BP+5 mode, electrodes 8, 7, and 6 had to be eliminated because of the facial nerve stimulation. However, with the variable mode, we were able to successfully use wider bipolar modes in an attempt to reduce the amount of current below the level of nonauditory sensations. We combined three variable modes to avoid the facial nerve stimulations caused by electrodes 8, 7, and 6 in the example above. These three mod es were BP+7 (created by separating the indifferent electrode from the active electrode by a distance of 7 electrodes), BP+8 (separating the indifferent from the active by a distance of 8 electrodes), and BP+9 (separating the two by 9 electrodes). For electrode 8, the indifferent electrode was set to 16 (BP+7); for electrode 7, the indifferent electrode was set to 16 (BP+8); and for electrode 6, the indifferent electrode was set to 16 (BP+9). We were thus able to reprogram all electrodes without stimulating the facial nerve, while still providing a true auditory sensation (except for the most apical electrodes). We were able to give the patient 14 electrode pairs, compared with none previously.

Immediately after programming in the variable mode, the patient was able to detect speech, discriminate differences in word pairs (as shown by a score of 62% on the Cochlear Corp. vowel identification test), and identify common spondee words in a closed set (score: 74% on the four-choice spondee test). After 3 months of auditory therapy, she was able to identify some words and sentences in an open set. She scored 68% on the monosyllable word test and 70% on the CID everyday sentence test when scored for sentences correct rather than key words correct. However, after some weeks of successful use, the patient began to be troubled again by facial stimulation and had to stop using the implant.

Case 3. We saw a 9-year-old girl who had been born at 26 weeks' gestation and who had been diagnosed as deaf during her stay in the neonatal intensive care unit. She had been on mechanical ventilation in the ICU for 4 months. She had also undergone two cardiothoracic surgical procedures and numerous blood transfusions. During subsequent years, the girl had had several sets of pressure-equalizing (PE) tubes inserted.

Preoperatively, we observed that the middle ear on both sides was well aerated, and her tympanic membranes were in the annular plane. A Nucleus 22-channel device was implanted in the left ear. The skin flap was 6mm thick and was handled gently during surgery. The facial recess was opened with a diamond bur, without causing injury to the annular ligament or the chorda tympani. The incus was disarticulated and removed for access. The posterior canal wall was left intact and not thinned excessively. The procedure was well tolerated and without operative complications. Subsequently, the device was programmed and was functional.

Seven months later the patient returned with severe ear pain and purulent otorrhea on the implanted side. On examination under the microscope, the patient was noted to have an external otitis with a thickened, mobile tympanic membrane. This resolved with systemic antibiotic therapy and the instillation of antibiotic-steroid eardrops. Resolution of the inflammation was confirmed by microscopic examination. The tympanic membrane was intact, mobile, and in the annular plane.

One month later, the patient again returned with a postauricular skin breakdown and exposure of the implanted device. The facioplastic service was consulted, and the problem was managed with a local advancement flap. The patient tolerated the procedure well, and she was discharged home on the second postoperative day. After the wound was completely healed, reactivation of the device failed. CT of the temporal bone showed a discontinuity of the stimulating wire. Clearly, the stimulating wire had been cut during the flap advancement procedure. The patient was returned to the operating room for exploration, and a large cholesteatoma was found in the mastoid segment. The neck of the cholesteatoma was located in the facial recess. Upon viewing the tympanic membrane through the external auditory canal, it appeared to be a posterior-superior retraction pocket cholesteatoma. The cochlear device was explanted, and a radical mastoidectomy was performed. Postoperatively, the patient had no drainage from the operated ea r. Because this patient's hearing had been suboptimal even when the implant was working, and because of cost considerations, her parents elected not to implant the opposite ear, but to continue her development with a hearing aid and sign language.

Discussion

Flap necrosis. As mentioned, flap necrosis is the most common complication of cochlear implant surgery. However, only a few reports have addressed the issue substantively. [6-9] From our review of the literature, we found that in most patients who experienced flap necrosis, it occurred some weeks or months after surgery and resulted in device extrusion. [6,9,17] Necrosis or infection of the skin edge has been described in only one report. [7] It is possible that flap-edge necrosis has not been reported more often because Cohen and Hoffman categorized it as a minor complication. [1] The patient described in case 1 falls into the minor category because no secondary surgical intervention was necessary.

But regardless of whether flap necrosis is a major or minor complication, it is one that is feared because it can necessitate explantation of an expensive device. A good understanding of flap physiology is therefore critical because it will help in 1) preoperatively defining risk factors that have a negative impact on flap performance and 2) improving operative technique. Daniel and Kerrigan have reviewed the pathophysiology of skin flaps in some detail. [18]

Survival of a surgical skin flap depends on anatomic, neural, metabolic, and hormonal factors. During elevation of the flap, cutaneous blood vessels and sympathetic nerves are severed. This results in a significant decrease in blood flow, not only because of the decreased vascular supply, but also because of the release of vasoconstricting neurotransmitters from the severed nerve ends. Glucose, oxygen, and cyclic AMP levels fall dramatically in the ischemic portion of the flap as anaerobic metabolism and the production of lactate and toxic superoxide radicals increase. These effects alone can cause an irreversible ischemia in the distal end of a poorly designed flap within 12 hours. Tissue that survives this initial insult will respond with humoral vasodilators, such as prostaglandins and histamine, and the flow to the ischemic areas will gradually increase.

Metabolic factors in an ischemic surgical skin flap--such as hypercapnia, hypoxia, acidosis, and interstitial potassium--all have a primarily vasodilating effect, and they will reverse the low perfusion state during the following 24 to 48 hours. During this same time frame, there is also a degradation and depletion of sympathetic vasoconstrictors that can result in a gradual increase in flow. In patients whose nutritional status is good, adequate collateral circulation is restored in 4 or 5 days. Factors that can contribute to poor flap performance include diabetes, vitamin deficiencies (A, B complex, and C), anemia, poor nutrition or protein depletion, prolonged hospitalization, steroid or other immunosuppressive therapy, and nicotine addiction.

The flap complication rate of 2.79% compares favorably with published complication rates for other neurosurgical procedures. In 1982, Stem estimated that the scalp wound complication rate following major neurosurgical procedures was 6%. [19] The neurosurgery literature describes a number of measures to avoid scalp flap necrosis. Wilkins writes that the width of the base of a scalp flap should be wider than the length of the flap in order to ensure a good vascular and lymphatic supply. [20] Allen and Johnston believe that 1) optimal closure of a scalp flap requires two layers, 2) that flaps must not be closed under tension, so as to avoid ischemia at the suture line, and 3) that the closure of the galea aponeurotica should be separated from a second suture line through dermis [21] They suggest that minimizing scalp flap necrosis depends on 1) careful handling of the flap intraoperatively (excess manipulation or prolonged angulation after reflection can cause ischemia), 2) not using Raney clips (to avoid skin-edge ischemia), 3) assuring a good vascular supply by closing the wound with minimal tension (with the use of relaxing incisions and flap undercutting if necessary), and 4) avoiding tight dressings postoperatively.

Even when all these recommendations are followed and risk factors are accounted for, the scalp flap used in earlier years for cochlear implantation (C-shaped, postauricular) is at particular risk because underneath it lies a large foreign body. Furthermore, the thickness of the flap cannot be greater than 6 mm; otherwise the external device will not adhere to the skin by magnetic attraction. If the tinning of the flap reaches the level of the hair follicles, the device is at risk for extrusion. [6] The distance of the implanted device from the edge of the incision is also critical.

The patient described in case 1 experienced only epidermolysis of a portion of the operative flap. At no time was the device exposed, infected, or at risk. There were no intraoperative events or factors that would have indicated that her flap was at risk. The C-shaped flap was large, it had a broad base that spared both the occipital and superficial temporal arteries, it did not require thinning, and it was handled gently. No Raney clips were used. The patient responded well despite the fact that she had a number of risk factors for complications: she was a smoker, and she had a known autoimmune condition, anemia, and poor nutritional status as a result of her multiple fundoplications. Furthermore, she was on long-term steroid therapy.

Facial nerve stimulation. Hoffman and Cohen reported 101 cases of facial nerve stimulation (83 adults and 18 children) following 4,969 cochlear implantations, for a complication rate of 2.03%. [5] Our review of the literature revealed that facial nerve stimulation occurred in 45 patients described in eight papers. [10-14,22-24] Of these 45 published cases, six occurred in patients implanted with single-channel devices [10] and 39 occurred with multiple-channel devices of various designs. The presumed etiology of the stimulation is the distribution of current from the electrode array to the facial nerve. Niparko et al have proposed that with single-channel devices, the relatively large current field distribution is the cause of facial nerve stimulation. [10] In contrast, Langman et al published a case study in which a single-channel implant did not cause facial nerve stimulation, but a subsequent 22-channel replacement did. [24]

Four groups of authors have observed a correlation between facial nerve stimulation and otosclerosis. [11-14] Marsh et al proposed that the spongiotic bone of otosclerosis might decrease impedance and result in a current spread from the electrode array to the fallopian canal. [11] Cochlear otosclerosis might be a factor in some cases, but it is unlikely to be the cause in all cases. Certainly, the patient described in case 2 did not have cochlear otosclerosis. Shea and Domico described a patient in whom basal electrodes stimulated the descending mastoid segment of the facial nerve. [12] This stimulation was so painful and intractable that the patient was forced to stop using the device and requested an explantation. During the explantation, the surgeons noted the development of adhesions that represented an apparent inflammatory response to the electrode array; these adhesions were the likely source of current shunt.

The patient in case 2--although she did not have otosclerosis or any other inflammatory condition of the otic capsule--experienced severe, painful facial nerve stimulation upon activating her device. The stimulation was most severe at the midbasal electrodes (8, 7, and 6). Radiographic studies indicated that these three electrodes had come into contact with the labyrinthine segment of the facial nerve. The problem was solved by reprogramming the device in a variable mode and changing the indifferent electrode, using wider bipolar modes to reduce the levels of current. After the device had been reprogrammed, the patient was able to use 14 electrode pairs (only the most apical electrodes could not be programmed). In time, she was able to identify some words and sentences in an open set. However, because of a recurrence of persistent facial nerve stimulation, she was forced to stop using the device.

In most cases of facial nerve stimulation reported in the literature, deprogramming of the offending electrodes was required. Notable exceptions were five cases of spontaneous resolution without treatment, [10,12] two of which occurred in patients with single-channel devices. [10] Langman et al reported a case of stimulation that was successfully treated with injections of botulinum toxin. [24] This strategy had an advantage in that it did not require the deactivation of the electrodes, thereby allowing the patient to experience the full benefits of the cochlear implant.

Postimplantation middle ear disease. Our review of the literature revealed no published reports on the incidence of postimplant middle ear effusions. One unpublished survey of the parents of 133 implanted children conducted by Cochlear Corp. revealed that 17 (12.8%) had suffered at least one episode of acute otitis media. [5] There are only two reports of postimplant cholesteatoma. [5,16] Considering this paucity of data on middle ear disease in postimplant children, we have no guidelines on whether children who are susceptible to effusions and/or cholesteatomas should be implanted.

At the time the patient described in case 3 was implanted, she had been free of middle ear disease. Her cholesteatoma developed during the 9 months between the time the microscopic examination showed an intact tympanic membrane in the annular plane and the time of her return to our institution at the onset of a breakdown of the postauricular skin. Thus, it would be inappropriate to categorize this as an iatrogenic cholesteatoma. The child had a long history of an underaerated middle ear, and she had required multiple sets of PE tubes, indicating a probable eustachian tube dysfunction. Her marginal eustachian tube function had become critically diminished sometime after the microscopic examination, and we believe this led to the formation of the posterior-superior quadrant cholesteatoma. Obviously, this child should have received more frequent otologic examinations. We can probably become more sensitive to the fact that a patient needs closer observation if the preoperative CT shows poor pneumatization of the mastoid or if there is a history of multiple PE tube insertions.

When a cholesteatoma does develop, what should we do with the implant? Should we explant it and implant the opposite ear? Should we reimplant it in the same ear after the disease has been eradicated? The latter option is possible when the middle ear is obliterated and the external auditory canal and eustachian tube are blocked. However, the chance that a dangerous complication will develop many years later is a distinct possibility (as pointed out by Donnelly et a1 [25]), so careful and frequent imaging of the ear is a prerequisite. The parents of our patient did not want any further surgery, so we explanted the device and completed a radical mastoidectomy.

In conclusion, we would like to emphasize three points:

1) A careful review of all risk factors for poor flap performance must be made in every patient who is selected for cochlear implantation.

2) Every case of facial nerve stimulation following the activation of a cochlear device should be investigated with a bone-window CT in the axial and coronal planes to determine the relationship of the electrode array to the facial nerve. Currently, we do not know the exact cause of facial nerve stimulation. If a correlation between facial nerve stimulation and cochlear otosclerosis is to be confirmed, we will need multicenter studies. Meanwhile, the use of the Variable Mode stimulation (DPS version 6.125) device can prevent facial nerve stimulation and at the same time allow for auditory input.

3) We need longitudinal multicenter studies on the incidence of middle ear disease in implanted children. Such investigations could provide us with guidelines for managing children who have known eustachian tube dysfunction. For the time being, however, we feel that all children need frequent and careful followup over a long period of time.

From the Department of Otolaryngology--Head and Neck Surgery, University of Illinois at Chicago.

Reprint requests to Arvind Kumar, MD, Room B-42, Eye & Ear Infirmary, 1855 W. Taylor St., Chicago IL 60612. Phone: [phone]; fax: [phone]; e-mail: [e-mail]

References

(1.) Cohen NL, Hoffman RA, Stroschein M. Medical or surgical complications related to the Nucleus multichannel cochlear implant. Ann Otol Rhinol Laryngol Suppl 1988;135:8-13.

(2.) Cohen NL, Hoffman RA. Complications of cochlear implant surgery in adults and children. Ann Otol Rhinol Laryngol 1991;100:708-11.

(3.) Cohen NL, Hoffman RA. Complications of cochlear implant surgery. In: Eisele DW, ed. Complications in Head and Neck Surgery. St Louis: C.V. Mosby, 1993:722-9.

(4.) Collins MM, Hawthorne MH, el-Hmd K. Cochlear implantation in a district general hospital: Problems and complications in the first five years. J Laryngol Otol 1997;lll:325-32.

(5.) Hoffman RA, Cohen NL. Complications of cochlear implant surgery. Ann Otol Rhinol Laryngol Suppl 1995;166:420-2.

(6.) Haberkamp TJ, Schwaber MK. Management of flap necrosis in cochlear implantation. Ann Otol Rhinol Laryngol 1992;l01:38-4l.

(7.) Wang RC, Parisier SC, Weiss MH, et al. Cochlear implant flap complications. Ann Otol Rhinol Laryngol 1990;99:791-5.

(8.) Schweitzer VG, Burtka MJ. Cochlear implant flap necrosis: Adjunct hyperbaric oxygen therapy for prevention of explantation. Am J Otol 1991;12:71-5.

(9.) Manrique MJ, Paloma V, Cervera-Paz FJ, et al. Treatment of cutaneous ulceration after cochlear implantation surgery. Ann Otol Rhinol Laryngol Suppl 1995;166:422-5.

(10.) Niparko JK, Oviatt DL, Coker NJ, et al. Facial nerve stimulation with cochlear implantation. VA Cooperative Study Group on Cochlear Implantation. Otolaryngol Head Neck Surg 1991;104:826-30.

(11.) Marsh MA, Coker NJ, Jenkins HA. Temporal bone histopathology of a patient with a nucleus 22-channel cochlear implant. Am J Otol 1992;13:241-8.

(12.) Shea JJ, Domico EH. Facial nerve stimulation after successful multichannel cochlear implantation. Am J Otol 1994;15:752-6.

(13.) Muckle RP, Levine SC. Facial nerve stimulation produced by cochlear implants in patients with cochlear otosclerosis. Am J Otol 1994;15:394-8.

(14.) Weber BP, Lenarz T, Battmer RD, et al. Otosclerosis and facial nerve stimulation. Ann Otol Rhinol Laryngol Suppl 1995;166:445-7.

(15.) Luntz M, Hodges AV, Balkany T, et al. Otitis media in children with cochlear implants. Laryngoscope 1996;106:1403-5.

(16.) Lloyd RV, Gibbin KP, O'Donoghue GM. Pediatric cochlear implants. Surgical aspects: The Nottingham pediatric cochlear implant programme. Rev Laryngol Otol Rhinol (Bord) 1995;116:85-7.

(17.) Schwartzman JA. Avoidance of complications with cochlear implants. Ann Otol Rhinol Laryngol Suppl 1995;166:431-2.

(18.) Daniel RK, Kerrigan CL. Principles and physiology of skin flap surgery. In: McCarthy JG, ed. Plastic Surgery. 1st ed. Philadelphia: W.B. Saunders, 1990.

(19.) Stern WE. Preoperative evaluation: Complications, their prevention and treatment. In: Youmans JR, ed. Neurological Surgery. 2nd ed. Philadelphia: W.B. Saunders, 1982:1051-116.

(20.) Wilkins RH. Principles of neurosurgical operative technique. In: Wilkins RH, Rengachary SS, eds. Neurosurgery. 1st ed, New York: McGraw-Hill, 1985:427-38.

(21.) Allen MB Jr., Johnston KW. Preoperative evaluation: Complications, their prevention and treatment. In: Youmans JR, ed. Neurological Surgery. 3rd ed. Philadelphia: W.B. Saunders, 1990:833-99.

(22.) Webb RL, Lehnhardt E, Clark GM, et al. Surgical complications with the cochlear multiple-channel intracochlear implant: Experience at Hannover and Melbourne. Ann Otol Rhinol Laryngol 1991;100:131-6.

(23.) House JR, Luxford WM. Facial nerve injury in cochlear implantation. Otolaryngol Head Neck Surg 1993;109:1078-82.

(24.) Langman AW, Quigley SM, Heffernan JT, Brazil C. Use of botulinum toxin to prevent facial nerve stimulation following cochlear implantation. Ann Otol Rhinol Laryngol Suppl 1995;166:426-8.

(25.) Donnelly MJ, Pyman BC, Clark GM. Chronic middle ear disease and cochlear implantation. Ann Otol Rhinol Laryngol Suppl 1995;166:406-8.