Esophageal perforation is one of the most severe injuries of the digestive tract and is considered a true medical emergency. When a perforation occurs, gastrointestinal contents including saliva, retained gastric contents, bile, and acid enter the mediastinum. These events can lead to serious complications such as mediastinitis, abscess, empyema, meningitis, sepsis, and death (1).

The ventral approach to the cervical spine has gained wide acceptance since the original technique was popularized in the late 1950s and early 1960s (2, 3 and 4). Current indications for the surgical approach include trauma, neoplasm, instability, and spondylotic myelopathy.

In general, the approach has been shown to be a relatively safe way to access the cervical spine. Complications of ventral cervical spine surgery are rare and can be divided into three categories: (a) neurologic injuries, (b) spine stabilization issues (i.e., bone graft or instrumentation complications), and (c) soft tissue injuries.

There are anatomical considerations that render the esophagus vulnerable to injury especially during anterior cervical surgery. Decreased esophageal wall thickness and increased size of peristaltic waves at the caudal cervical spine levels lend to susceptibility (5). Weakness of the lateral wall of pyriform sinuses at the level of the sixth cervical vertebrae has also been described as a predisposing factor for esophageal injury (1). Proposed mechanisms of injury include (a) trapping or pinching of the esophageal wall against the vertebral bodies, (b) stretching during distraction of the spine with anterior longitudinal ligament disruption, (c) direct contusion or laceration by fracture fragments or osteophytes, and (d) early versus delayed perforation from surgical intervention by bone fragments or instrumentation (6).

The first use of a ventral cervical plate was described by Bohler in 1967 for traumatic instability (7). Instrumentation has become widely accepted for enhancing stability, improving fusion rates, and maintaining sagittal correction obtained during surgery (8, 9, 10, 11 and 12). Both constrained and nonconstrained systems of anterior cervical instrumentation are currently available. The failure rate in nonconstrained systems has been reported to be significantly higher (13). Typically, failure is due to screw back out, which can result in graft migration or esophageal erosion. Cloward (14) reported the first case of transesophageal migration in 1971. In 2003, Riew et al. (9) described a case of plate failure and graft migration that lead to airway impairment and eventual death. Constrained systems have also been reported to have screw failure related to inadequate purchase, placement into the disk, or pseudarthrosis (13,15,16). In these situations, plate lift-off may result in esophageal injury.

Blunt injury to the cervical spine particularly in hyperextension exposes the pharynx and proximal esophagus to increased risk of trauma due to their proximity to the vertebral bodies. Makoyo (17) described a cervical fracture dislocation causing cervical esophageal rupture in 1979. The incidence of esophageal perforation secondary to trauma is unknown but has been reported to place patients at risk for mediastinitis. Morrison (18), Spendler and Benfield (20), and Stringer et al. (19) each reported a case of esophageal rupture after a hyperextension mechanism to the cervical spine. Two of these patients went on to develop mediastinitis with abscess formation (18, 19 and 20). Several authors described cervical fracture dislocations leading to mediastinitis as well (21, 22 and 23). Tomaszek and Rosner (24) and Krespi et al. (25) discussed five cases of cervical spine fractures that led to esophageal rupture causing mediastinitis.

The true incidence of esophageal perforation following ventral cervical spine surgery is unknown and likely underreported (5). Cloward (3) presented the common occurrence of transient dysphagia in his series of over 200 cases of anterior spine surgery without incidence of perforation. In a larger series by Tew and Mayfield (26), 1 case in 500 had an esophageal perforation. Barber (27) reviewed different series in the literature and found no reports of pharyngoesophageal injuries in 700 anterior cervical fusions. Van Berge Henegouwen et al. (28) and Eleraky et al. (29) each found 3 cases of esophageal perforation in 441 and 185 cases of ventral spinal surgery, respectively.

In a survey by Newhouse et al., an incidence of 0.25% esophageal perforation following over 10,000 cases of ventral cervical procedures was reported. Approximately one-third of esophageal perforations were recognized at the time of surgery and related to sharp or motorized instruments (5). Orlando et al. (1) reported an incidence of 0.4% of 1,075 patients undergoing ventral spine surgery presenting with esophageal perforation over a 38-year period. In his review of the literature, Navarro et al. (30) described an incidence reported between 0.02% and 1.49% of esophageal perforation after anterior spine surgery. Espersen et al. (31) reported an incidence of 1.8% in 1,106 patients treated with various graft types using the Cloward technique. A deep space infection occurred in 0.27% of patients (31). Blunt dissection and careful placement of retractors while exposing the cervical spine may help to avoid iatrogenic injury during surgery. However, there are rare occasions on which an esophageal perforation can present after in a delayed fashion after surgery.

Asymptomatic oral extrusion of screws as a late complication has been anecdotally reported in the literature (13,32, 33, 34 and 35). Two-thirds of the esophageal perforations discovered in the review by Newhouse were after surgery, and 40% of these were due to the use of hardware. The most common levels involved were C5-C6 when instrumentation was used (5). Interestingly, a cadaveric study of seven specimens showed significantly increased intraesophageal pressures at C5-C6 compared to C3-C4 when anterior cervical plating was used (36). One review of the literature showed that the presentation of clinical symptoms varied significantly averaging 2.5 years (37).

Whitehill et al. was the first to describe a case of delayed cervical abscess formation 10 weeks after spinal fusion. In this case, the cause was migration and protrusion of the bone graft into the esophageal wall (38). Kuriloff et al. (6) reported two cases of delayed abscess formation following anterior cervical surgery in two trauma cases. Bohlman (39) also reported on a case of esophageal perforation, which led to a mediastinitis and eventual death.

Although relatively rare, a high index of suspicion for esophageal perforation should be maintained in patients following cervical spine trauma and in postoperative patients with unexplained symptoms. The most common symptoms include fever, dysphagia, neck swelling/discharge, pneumonia, odynophagia, hoarseness, and breathing difficulty (37). Patients who develop the more severe complication of mediastinitis can present with fever, tachycardia, and tachypnea. On examination, mediastinal crunch sounds (Hamman’s sign) may be heard. Although symptoms in the early postoperative period may be observed in the attentive clinician, it should be noted that patients may present in a delayed fashion. Nourbakhsh and Garges described the initial presentation of clinical symptoms and found that it varied significantly, averaging 2.5 years. In his review, 25% of patients presented at 3 months, and 46% of cases at 1-year follow-up (37).

When clinical exam elevates suspicion, diagnostic laboratory studies and imaging should be done to confirm diagnosis. Previous literature on lumbar postoperative infections has given insight into laboratory diagnostic evaluation. Although nonspecific, routine analysis of differentiated white blood cell (WBC) count, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) studies may help to establish a diagnosis. The WBC may be increased but in general is nonspecific. Although typically elevated in the postoperative period, the CRP level peaks after 2 to 3 days and normalizes within between 5 and 14 days. The ESR usually peaks at 5 days and typically declines irregularly (40). Due to the nonspecificity and high false-negative values of laboratory data, further diagnostic workup is recommended.