Intermittent stimulation of the left vagus nerve in the neck is an adjunctive treatment for patients with medically refractory epilepsy (16). The system consists of a pacemaker-like generator, placed in an infraclavicular pocket, and stimulating electrodes, placed around the left vagus nerve. The device delivers intermittent electrical stimulation to the cervical vagus nerve trunk. The nerve transmits rostral impulses to exert widespread effects on neuronal excitability throughout the nervous system (1).

The design of the coil-like stimulating electrodes provides a reliable interface between electrodes and nerve. This can be disadvantageous when discontinuation of therapy is desirable. Such situations include (a) lack of efficacy of treatment or even worsening of seizure situation, (b) lead failure, (c) choice of the patient, or (d) wound infection.

Whereas the successful removal of stimulating electrodes in situations (a) through (c) was reported (7,8), the most appropriate treatment in case of wound infection is not clear. Antibiotic therapy, optimized according to results of sensitivity testing, is mandatory. Removal of the foreign body (i.e., generator and stimulating electrodes) is desirable and a time-honored surgical principle. The generator can be removed easily. Concerns exist regarding the safe removal of the stimulating electrodes. The vagus nerve and the walls of adjacent major blood vessels might be either encased by fibrosis, making dissection difficult, or, even worse, might be extremely vulnerable to surgical trauma because of the ongoing inflammation.

We report the case of a patient implanted with a vagus nerve stimulator who had a Staphylococcus aureus infection deep under the superficial cervical fascia. The patient underwent successful open wound treatment without removal of the stimulating electrodes.

CASE REPORT

Go to:

Choose Top of page CASE REPORT << DISCUSSION Acknowledgment: REFERENCES

A 35-year-old man with drug-resistant epilepsy with complex partial seizures with secondary generalization since age 4 years was evaluated within our comprehensive epilepsy surgery program. His epileptic syndrome was considered to be not amenable to resective surgery.

On March 12, 1999, a Neurocybernetic Prosthesis (NCP) System (Cyberonics Inc., Webster, TX) was implanted following standard implantation directions (1,3,9). In brief, with the patient under general anesthesia, the left vagus nerve was exposed within the carotid sheath after a transverse skin incision across the anterior sternocleidomastoid border. The electrode was fitted around the nerve. Then, starting from a second infraclavicular incision, a subcutaneous pocket was fashioned. The electrode cables were tunneled from the neck incision to the thoracic wound and connected with the stimulator. After checking the impedance and a test stimulation, the stimulator was fixed to the pectoralis fascia. All wounds were rinsed with bacitracin-containing antibiotic solution and closed in layers. Duration of the intervention was 80 minutes. Perioperative prophylactic antibiotics were not given. The patient was observed overnight and then sent home. Sutures were removed by the family physician on postoperative day (POD) 8. Wound problems were not reported.

Five weeks later, on April 21, 1999, the patient presented again with a reddish and swollen cervical wound. History revealed that swelling had started about 1 week earlier, with local hyperemia following somewhat later. Beside the local symptoms, the patient was well. Fever and leukocytosis were absent. C-reactive protein (CRP) was 1.84 mg/dL (normal values, 0.000.70 mg/dL). No seizures had occurred in the meantime.

The wound was opened the same day with the patient under general anesthesia. After removal of the partly absorbed platysma sutures, pus escaped under considerable pressure. Pus was present deep under the cervical fascia, but the carotid sheath itself was embedded in dense scarred tissue that did not show signs of inflammation. Cultures were taken, and the wound was rinsed several times with bacitracin-containing antibiotic solution. The leads were dissected free, following the cables into the depth and toward the generator until the surrounding tissue looked inflammation free.

The original plan to follow the leads down to the nerve and remove the electrodes was abandoned for fear of damaging the vagus nerve or one of the major cervical vessels within the scarred tissue. Wide dissection of the cervical area to identify the structures in an unscarred area and follow them into the scar seemed inappropriate. The wound cavity, with electrodes entering and leaving, was thus packed with sterile gauze impregnated with 3% iodoform paste and covered with a sterile dressing. Endovenous antibiotic therapy was started with fosfomycin, 4 g t.i.d., and cefmenoxim, 2 g t.i.d.

Microbiologic examination revealed infection with Staphylococcus aureus, which was treated adequately with the chosen antibiotics. The wound dressing was changed daily by removing the gauze, rinsing the wound with 3% hydrogen peroxide solution, followed by rinses with 5% saline and reapplication of a fresh gauze packing. No fever was observed during the postoperative course; maximal elevation of CRP was 8.28 mg/dL on POD 4.

On POD 13, granulation tissue was present throughout the wound, fibrin deposits were minimal, and wound cultures were sterile. Signs of inflammation around the wound had disappeared. CRP had returned to 1.71 mg/dL. The wound was closed with a few nonabsorbable sutures. Systemic antibiotic therapy was terminated 1 day later.

On May 17, 1999, wound revision under local anesthesia with excision and new adaptation of the wound margins was necessary because the retracted wound margins were not well adapted. The patient, somewhat noncompliant, continued to irritate the wound while shaving. Local and systemic signs of infection were absent.

The further clinical course was uneventful. Wound healing occurred by third intention. Systemic and local signs of infection remained absent. The NCP remained active during the whole period. A definitive statement on changes in the seizure situation of the patient is not yet possible. Follow-up is now at 1 year.

DISCUSSION

Go to:

Choose Top of page CASE REPORT DISCUSSION << Acknowledgment: REFERENCES

Surgical wound infection requires antibiotic therapy and sometimes open wound debridement. Foreign bodies should be removed. A surgical dilemma arises if (a) the foreign body is thought to be vital or at least functionally indispensible, (b) removal can be associated with severe morbidity, and (c) sometimes when the foreign body is very expensive.

We report the case of a patient in whom a local infection with S. aureus around the electrodes of a vagus nerve stimulator was treated successfully without removal of the device.

The literature contains very few reports on infection in connection with implantation of the NCP. In one of 114 patients included in the E03 study, a local infection was treated with antibiotics without further complications (10). In three of 199 implanted patients reported in the E05 study, the device had to be removed because of infection. One of these patients was reimplanted later (11). How the leads were dealt with is not clear from this published information. Wheless et al. (12), after analyzing complications associated with 51 patients implanted with the NCP, described three cases with local skin infection that resolved with oral antibiotics and two cases of generator infection that required removal and subsequent reimplantation of the generator. The handling of the electrodes is not clear from this abstracted information.

The fact that the electrodes can be removed without damage to the nerve, at least in a situation in which local inflammatory changes are absent, has been known since Uthmann (8) reported successful revision operations with unwinding of damaged electrodes from the nerve in three patients. Recently, Espinosa (7) reported the successful removal of the entire NCP in seven of 10 patients (in the three remaining patients, new electrodes were placed cephalad to the old electrodes). Wound infection was not among the causes for electrode removal in these patients. We were hesitant to expose the vagus nerve within the carotid sheath, first because this would have meant a large dissection of vital structures in a probably contaminated field, and second because those parts of the lead entering the scar around the carotid sheath itself seemed not to be surrounded by inflamed tissue.

Treatment of local infection without foreign body removal has been reported in neurosurgical patients with infected cerebrospinal shunts (see 13 for review) and after spinal instrumentation (14). Directions for the treatment of our patient were derived empirically from other cases of deep wound infection. Endovenous antibiotic therapy is mandatory, in this case with a third-generation cephalosporin and with fosfomycin, a drug belonging to a new class of antibiotics. The wound must be opened, debrided, and packed with gauze containing 3% iodoform, a mild antiseptic (15). Daily rinses are performed with an aqueous solution of 3% hydrogen peroxide and with saline. Hydrogen peroxide, an oxidizing agent, is used as an antiseptic. Additionally, the mechanical effect of effervescence is useful for wound cleaning (15). Saline is believed to stimulate the formation of granulation tissue. Wound margins can be reapproximated with sutures when cultures are sterile and fibrin coverings within the wound have disappeared, usually 714 days later.

Some limitations should be noted: first, we describe our experience in a single case. Second, treatment of deep wound infection followed guidelines that, although time honored within our department, have never undergone strong scientific evaluation. We report these techniques, but general recommendation is inappropriate for these two reasons. Third, long hospitalization was necessary to obtain the desired results. We are convinced that antibiotics alone would not have done the job. We are not sure whether removal and subsequent reimplantation of a new NCP would have been the more cost-effective way to treat this complication.

Wound infection will remain a major cause of morbidity in patients implanted with foreign bodies. The occurrence of infection in connection with vagus nerve stimulation is expected to increase with the increasing number of NCPs implanted worldwide. Only detailed reporting of such cases will make it possible to formulate more general recommendations on the best treatment in this situation.

Acknowledgment:

Go to:

Choose Top of page CASE REPORT DISCUSSION Acknowledgment: << REFERENCES

We acknowledge the help of E. Semenitz, Pharm. D., with pharmaceutical details. We thank Ms. M. Margreiter for proofreading the English text.