Esophageal perforation from anterior cervical screw migration

Abstract

Background

Esophageal perforation from anterior cervical instrumentation migration is an uncommon but potentially highly morbid or even fatal complication. Early recognition and aggressive investigation and treatment are essential to ensure good outcome.

Case Description

A 58-year-old man underwent C6 vertebrectomy and C5-7 interbody fusion with a cage and anterior cervical plate. After surgery he developed fever and recurrence of his symptoms and deficits, but was managed expectantly. He was then referred to the author's institution. A barium swallow demonstrated an esophageal fistula (a Gastrograffin swallow was falsely negative) caused by a migrated screw; serial radiographs confirmed its passage through the gastrointestinal tract. Revision surgery was required to repair the perforation and reconstruct the cervicothoracic spine. Intraoperative esophageal injection of methylene blue was helpful in demonstrating the site of leakage. Despite a prolonged postoperative course complicated by pulmonary embolus, the patient recovered with minimal residual deficit, and continues to do well 2 years later.

Conclusions

A high index of suspicion followed by aggressive investigation are crucial in the setting of unexpected neck pain, new neurologic deficit, fever, or swallowing difficulties in the early postoperative period after anterior cervical spine instrumentation. If esophageal perforation is suspected, a barium swallow is recommended over Gastrograffin, which, although less irritating to the surrounding tissues, may be falsely negative. Intraoperative methylene blue injection into the esophageal lumen is useful in identifying the site of perforation.

Introduction

The use of anterior cervical plates in cervical spine surgery is now commonplace. Current-generation plating systems are very safe and user-friendly, but their use is not without complication. Screw/plate loosening and migration may have clinical sequela ranging from completely benign to potentially fatal complications such as esophageal perforation, deep neck space infection, mediastinitis, and airway obstruction [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12]. Careful attention to surgical technique and appropriate postoperative vigilance, should one suspect instrumentation failure, can reduce the risk of serious complication. In this report, a case of esophageal perforation and deep neck space infection from anterior cervical plate screw migration and instrumentation loosening is described. The screw passed through the gastrointestinal tract uneventfully, but management of the resultant esophageal, cervical spine, and medical complications proved challenging. After a protracted 9-month illness the patient finally made an excellent recovery. Complication avoidance and management in relation to ventral cervical instrumentation are discussed with particular reference to esophageal perforation.

A 58-year-old right-handed police officer presented to another hospital with symptomatic cervical spondylosis (neck and left arm pain and numbness) and a preoperative MRI of the cervical spine demonstrating prominent left-sided C5-6 and C6-7 causing spinal cord and nerve root compression at both levels and at the C6 vertebral body level (Fig. 1). He underwent C6 vertebrectomy and C5-7 fusion with interbody cage and anterior cervical plate fixation (using a plate with a locking mechanism). No intraoperative difficulties were reported. After surgery he experienced resolution of his preoperative arm pain and was discharged uneventfully. Immediate postoperative cervical radiographs showed slight prominence of one of the caudal screws. One week later, while turning in bed, he felt a “crunch” in his neck and developed a recurrence of arm pain and numbness. Cervical radiographs revealed further backing out of the loose screw (Fig. 2A). The patient was advised by his surgeon to wear a stiff collar. Two weeks later he attended the local emergency department with a high fever, and underwent abdominal x-rays, abdominal ultrasound, and chest x-ray, all of which had negative findings. A cervical spine x-ray failed to visualize the caudal aspect of the plate. He was treated with intravenous antibiotics, initially as an inpatient, and then on an outpatient basis, but soon returned to the hospital with chest pain; a cardiac workup was negative. Cervical x-rays were repeated, but again did not visualize the caudal aspect of the plate. He was again discharged home. Four weeks later (9 weeks postoperatively), his fever returned and he complained of neck pain and dysphagia. Repeat cervical spine x-rays revealed that one of the caudal screws had backed out of the plate completely but was still visible (Fig. 2B). One week later (10 weeks postoperatively) the screw was no longer visible in the neck (Fig. 2C), but was subsequently documented to pass through the gastrointestinal tract and out of the body (Fig. 2D and E). The patient was then transferred to the author's facility. On arrival, he was complaining of chills, dysphagia, odynophagia, neck and upper extremity pain, numbness, and paresthesias. On examination, he was afebrile, had moderate anterior neck swelling and tenderness, and numbness and weakness in the C7, C8, and T1 distributions bilaterally. Laboratory investigations were consistent with infection (WBC, 10.0; ESR, 90 [0-15 mm/h]; and CRP, 5.9 [0-3.3 mg/L]). Barium swallow demonstrated what appeared to be a blind pharyngeal pouch, suggesting a healed fistula (Fig. 3).

The neck incision was reopened to expose the anterior cervical plate and prevertebral space. Intraoperative findings consisted of copious inflammatory and granulation tissue but no pus, a missing caudal screw and loose anterior plate, and an interbody cage that had telescoped into the body of the C7. A careful inspection of the pharynx/esophagus failed to demonstrate any perforation. A methylene blue injection via a nasogastric tube was attempted by the anesthetist but unsuccessful due to failure to pass the tube despite several attempts. After removing all instrumentation and performing a C7 vertebrectomy, reconstruction was achieved from C5-T1 using allograft fibula and a cervical plate. The wound was closed primarily over a drain. Supplemental C5 through T1 lateral mass/pedicle screw dorsal stabilization was also performed. Wound cultures yielded mixed aerobic and anaerobic growth. The patient was started on intravenous clindamycin and gatifloxacin. He was allowed clear fluids by mouth 3 days after surgery. On the fourth postoperative day, there was visible air and fluid drainage from the incision along with subcutaneous swelling and crepitus. A Gastrograffin swallow was negative, but a barium swallow (Fig. 4) and CT of the neck (Fig. 5) demonstrated leak of contrast material to the level of the anterior plate and along a sinus tract to skin. There was no evidence of mediastinitis.

After first passing a nasogastric tube to the midcervical level, the neck incision was reopened. A methylene blue injection via the tube revealed the site of esophageal perforation and a primary repair was performed. The instrumentation was not removed and the wound was closed over deep and superficial drains after copious irrigation. Intraoperative cultures yielded Pseudomonas aeruginosa and Candida albicans, and therapy was initiated with intravenous tazocin and fluconazole for 6 weeks. The patient's recovery was complicated by mild posterior neck wound dehiscence requiring packing. Total parenteral nutrition was started and a gastrojejunostomy tube inserted. Both neck drains were removed 2 weeks later. The patient was placed in a Philadelphia collar for 10 weeks. At 10 weeks of follow-up, cervical flexion-extension x-rays demonstrated stability and the collar was removed. A repeat barium swallow failed to demonstrate any evidence of esophageal leak and oral intake was allowed. He was readmitted 2 months later with bilateral leg deep vein thrombophlebitis and pulmonary emboli requiring anticoagulation.

Four months postoperatively the patient's complaints of dysphagia and odynophagia had resolved and all hematologic indices had normalized. Cervical radiographs revealed the instrumentation to be in good position with spinal alignment maintained. Two years after surgery, minimal bilateral C8 weakness remains, but otherwise the patient had recovered completely, and investigations have demonstrated a solid fusion construct.