|
A.I. RIVKIND*, A. ZVULUNOV#, A. J. SCHWARTZ†, P. REISSMAN* and H.
BELZBERG**
*Department of Surgery/Trauma Unit and †Radiol Hadassah University Medical Center, Jerusalem, Israel, #Department of
Internal Medicine, Yosephtal Medical Center, Eilat, Israel, **Division of Trauma, Department of Surgery, LAC+USC Medical
Center, Los Angeles, California U.S.
Injuries to the eosophagus are notoriously difficult to diagnose pre-operatively. Patients with such injuries usually will not have pre-operative signs and symptoms to suggest the presence of this type of injury. These injuries require a high index of suspicion, appreciation of the presence of injuries to adjacent structures, and an understanding that the clinical and radiological findings may evolve over a period of time. We describe a child with a rare presentation of an acute traumatic esophageal spinal fistula due to a bullet wound. This complicated injury required a variety of diagnostic modalities, including contrast radiography, multiple computerised tomography (CT) scans and operative assessments to make the definitive diagnosis.
Keywords: Fistula, neck injury, oesophagus, perforation, spinal cord
J.R.Coll.Surg.Edinb., 46, April 2001, 113-116
A 7-year-old child was transferred from an outlying community hospital to Hadassah Medical Center, a Level I trauma centre, with a gunshot wound to the neck. The wound was caused by a .22 calibre bullet from a Baretta semi-automatic handgun (muzzle velocity 300-350 m/sec) fired accidentally by the patient’s sibling from a distance of approximately 2 m. The 5 mm entrance wound was located immediately below the thyroid cartilage, 1 cm right of the midline. The 1 cm exit wound was through the right trapezius muscle at the level of the first thoracic vertebra. At the outlying hospital, admission blood pressure was 100/50 mmHg, pulse was 100/min and respiration was 16/min, regular and without stridor. There was minimal bleeding from both the entrance and exit wounds. A very small, non-expanding hematoma was noted at the entrance site.
Neurologically, the patient manifested mild right-sided weakness that was greater in the lower extremity than the upper limb. Deep tendon reflexes were symmetric; however, abdominal and cremasteric reflexes on the right side were absent, and priapism was noted. There was a left-sided sensory deficit below the level of the C-4 dermatome. The examination was consistent with an incomplete Brown-Sequard syndrome. The patient maintained a position of right upper extremity and left lower limb extension. The only additional physical finding was a right-sided Horner’s sign. Of specific interest was the absence of any airway compromise.
Initial radiograph examinations revealed a normal chest and cervical spine.
At the outlying hospital, after haemodynamic stability was confirmed, and cervical immobilization with a Philadelphia collar was established, an attempt was made to reverse the neurological deficit with hyperbaric oxygen therapy. Despite a transient improvement, there was no significant difference in the neurological function, when the patient was transferred to and assessed at Hadassah Medical Centre, (12 hours post-trauma).
On arrival, the physical examination was as above, and cervical stabilization was continued. The following laboratory examinations were obtained: white blood cell count of 15,600/mm3, haemoglobin of 11.3 gm/dl, platelets of 190,000/cc, normal serum electrolytes and urea and blood glucose level. Arterial blood gases revealed a pH of 7.38; paO2 of 79, paCO2 of 31, O2 saturation was 96%. A repeat chest radiograph revealed evidence of aspiration in the left hemithorax.
A gastrografin (meglumine diatrozoate) swallow was performed, which revealed extravasation of contrast from the eosophagus at the C7/T1 level (Figure 1). Due to the need for better definition of the leak before emergency surgical repair, a barium swallow was performed. With this examination a hang-up and extravasation was noted, with a tract extending towards the spine (Figure 2). Further studies included a CT scan of the neck and chest using 5 mm cuts; these were interpreted as normal. Based on these findings, exploration of the neck was performed.
Figure 1: Gastrografin swallow image revealing leakage of contrast medium from the eosophagus into the spinal column at the level of C7 to T1
Figure 2: Barium swallow (oblique view) demonstrating leakage from the eosophagus at the level of C7 and T1
The patient was taken to the operating room and given a general anaesthetic with endotracheal intubation. Direct laryngoscopy and fibre-optic bronchoscopy were normal. Flexible fibre-optic eosophagoscopy revealed a left posterior perforation of the eosophagus 14 cm from the incisors. Exploration of the neck was performed through an incision that included the entrance wound. The vascular structures in the region of the gunshot tract were identified and found to be intact. The eosophagus was identified and an attempt was made to close the perforation primarily. During the procedure, an accumulation of clear fluid in the pre-vertebral plane was noted. A sample of the fluid was obtained for laboratory analysis, which revealed a glucose level of 7 mmol/l and total protein, 974 mg/l (normal for cerebrospinal fluid being 200/650 mg/l and normal for blood 60/80 g/l).
A muscle patch was placed between the pre-vertebral space and the eosophagus in an attempt to isolate the presumed dural tear from the esophageal injury. A Penrose drain was placed in situ, local debridement was performed, and the wound was approximated. Skeletal traction was implemented using a halo vest with a cranial ring. A lumbar drain was inserted to reduce cerebrospinal fluid pressure and encourage closure of the dural defect. Amikacin and cefotaxime were given immediately preoperatively and changed to metronidazole postoperatively. Immediately postoperatively, a repeat CT scan with 2.5 mm cuts, as opposed to the initial CT scan with 5 mm cuts, was obtained to identify the source of the presumed cerebrospinal fluid leak.
The repeat CT scan revealed fracture of the lamina of C7 as well as contrast material within the spinal canal (Figure 3). A fracture of the inferior aspect of C7 was noted, with fracture fragments in the right lateral aspect of the spinal canal with an extension of fragments behind the right lateral lamina. Small gas bubbles were noted in the canal at the C2 and C3 regions (Figure 4). The CT scan suggested traumatic eosophageal injury with paravertebral and para-eosophageal air. The repeat CT scan revealed barium extending through the fracture site, with spread into the thorax (Figure 5).
Figure 3: CT scan demonstrating contrast material within the spinal canal
Figure 4: CT scan demonstrating small gas bubbles within the spinal canal at the level of C2 and C3
Figure 5: CT scan demonstrating barium extending through the fracture site with spread into the thorax sac
On the second post-operative day, the patient developed a leakage of saliva through the drain site, which ultimately resolved spontaneously on the 15th post-operative day. The lumbar drain was removed on the 17th postoperative day. The remainder of the post-operative course was uneventful, and there was no evidence of infection either locally or in daily cerebrospinal fluid cultures. On the 20th post-operative day, the patient was discharged to a rehabilitation centre without significant change in neurological status. On follow-up examination 6 years after the injury, the patient was doing well and performing at the appropriate level for his age in school. However, there remained a paretic gait disturbance with mildly abnormal movement of the left leg.
Penetrating injuries to the neck are a diagnostic and therapeutic challenge. The majority of these injuries can be managed without surgical intervention, but meticulous attention to physical signs and clinical symptoms is crucial.1,2,3 Previous authors have strongly recommended early mandatory routine exploration of penetrating cervical trauma, especially in the ‘zone 2’ region.4 These recommendations were based on a 1956 review of 274 patients by Fogelman and Stewart (1956). Patients who underwent early exploration had an operative mortality of 6%; those explored late or not at all had a mortality of 35%. More recently, Demetriades et al (1996) and colleagues have demonstrated that conservative non-operative management of penetrating cervical injuries can be performed safely if there are no clinical findings suggestive of vascular or aerodigestive injury.5 This approach is consistent with the findings of Golueke et al (1984) who prospectively randomized 160 patients with penetrating neck injuries and found no statistical difference in length of stay, morbidity or mortality between those patients who underwent routine or selective exploration of their injuries.6 They concluded that surgeons should base their treatment on their own experience and the particular circumstances surrounding the patient.
While the evidence for non-operative management of penetrating cervical injury is compelling, the extent of the diagnostic evaluation remains a crucial decision. McConnell and Trunkey (1994), in a meta-analysis of penetrating neck trauma reports, found that the incidence of isolated eosophageal injury was 6.6%, isolated spinal cord damage was 3%, and the overall incidence of injury in the neck was proportional to the volume occupied by the organ.7 In the presence of symptoms or signs of aerodigestive or vascular injury, radiological or ultrasound imaging is clearly indicated. In addition to the signs and symptoms routinely assessed, the mechanism of injury must be considered. Gunshot wounds often do not follow a predictable trajectory, and may travel in unexpected patterns, especially if they impact upon bony structures.8-11 In addition to the path of the bullet, two additional factors must be considered: the blast effect and the possibility of either bullet or bone fragments causing additional injury.
In this case, there were no signs or symptoms attributable to aerodigestive injury, but there was a significant injury to the eosophagus. Despite the absence of a direct aerodigestive finding, the presence of an acute neurologic deficit consistent with a penetrating spinal cord injury led to aggressive evaluation of adjacent structures. When a Brown-Sequard pattern of spinal cord injury is present, the possibility of penetration of the spinal canal is high. When there is evidence of penetration of the spinal canal, the possibility of an injury to the eosophagus should be considered due to its proximity to the spinal column. In this case, the routine examination of the spinal column (plain radiograph and 5 mm cut CT scan) failed to demonstrate any abnormality. Notwithstanding these negative screening examinations, the possibility of penetration of the canal led to more extensive evaluation of adjacent structures and, ultimately, these studies documented the injury to the eosophagus, the spinal fractures, and the dural violation of the spinal canal.
The significance of the eosophageal injury in this setting is magnified by the presence of a dural interruption. In most cases, small perforations of the cervical eosophagus do not lead to catastrophic consequences.12 However; the risk of infection of the central nervous system in undiagnosed cervical eosophageal injuries is real. The combination of eosophageal and dural injury can lead to neurologic complications including meningitis and quadroparesis.13-15
While this injury has not been previously reported, we believe that the risk of complications from such an injury is high. In this case, the eosophagus was primarily repaired and a muscle flap was interposed between the eosophagus and the dural defect. We postulate that despite a leak of cerebrospinal fluid through the dural injury, the combination of the primary repair of the eosophagus, the muscle flap and lumbar drainage may have contributed to the successful avoidance of meningitis and consequent neurological deterioration.
In conclusion, penetrating injuries to the neck when caused by gunshot wounds may be more extensive than suspected on routine clinical and radiological examinations. Evidence of injury to the spinal column should increase the index of suspicion of injuries to adjacent structures. Eosophagography and fine-cut CT scans should be considered when there is evidence that the gunshot wound has involved the spinal column.
Copyright date: 5th November 2000
Correspondence: Howard Belzberg, M.D., Department of Surgery, LAC+USC Medical Center, 1200 N. State St., Room 9900, Los Angeles, California 90033 USA
©2001 The Royal College of Surgeons of Edinburgh, J.R.Coll.Surg.Edinb.