Temporal Bone Fractures Caused by Ballistic Projectiles: A Systematic Review

Patient Demographics

Patient demographics were remarkably similar across most of the studies. In every study, there was an overwhelming majority of young males involved: 84% of the 146 total patients included in the review were male, with an average age of 28.8 years. Unfortunately, demographic information beyond this level of detail is unavailable for most patients due to the use of representative cases for subcategories of injuries, as seen in Hagan et al and Duncan et al.^6,8 Interestingly, Haberkamp et al noted that the right side was injured 75% of the time; however, the laterality of each patient was not noted in any other studies. ¹⁰ Though speculative, this could be related to handedness of the shooter and would be an interesting hypothesis to test in a future study. ¹⁰ Hagan et al excluded shotgun blasts, but other authors chose to include them. ⁸

Hearing Loss

Hearing loss was also frequently reported in the studies included. About 62% of patients exhibited some type of hearing loss, but just 5 of the 6 studies included the hearing status of all patients within the study. Only 48% of the patients had preoperative audiograms. Some studies utilized tuning fork tests, whereas others reported audiometric data obtained prior to surgery. The tuning fork examination can be performed to distinguish conductive and sensorineural hearing loss (SNHL). Obtaining audiograms of patients who are critically ill can be technically difficult from a systems perspective but still valuable for guiding care. Sabin et al described 14 of 22 patients who had audiograms, and 86% of these patients showed SNHL or mixed hearing loss. ¹¹ Furthermore, 18% of the patients included in their review had anacusis, but only half of the other studies reported this finding. Ideally, all future studies would report the presence of subjective hearing loss, hemotympanum, and tuning fork findings. This would help establish the type and severity of hearing loss. Patients should be provided with follow-up information to obtain an audiogram once discharged from the hospital. Audiometric outcomes are used to guide patients to appropriate interventions, such as amplification for SNHL or surgical exploration for persistent conductive hearing loss. ¹² Additionally, documentation of pre- and postoperative audiograms will allow a more robust study of outcomes.

Facial Nerve Injury and Rehabilitation

Facial nerve injury was universally mentioned in every study, with facial nerve weakness reported in 48% of all patients in this review. This rate included total, incomplete, immediate, and delayed paralysis. The inconsistent reporting of onset and severity of facial nerve function makes it difficult to generalize the outcomes across studies. Overall, 92% of reported facial nerve injuries were immediate. Only 1 of 6 studies reported House-Brackmann facial nerve grading scale, with the remaining studies describing facial nerve function as complete or incomplete.^13,14

Transection of the facial nerve can occur because it is tethered within the temporal bone and blast injuries will tear or crush the nerve. In most studies, facial nerve injury was the most common indication for intervention, though patients often had multiple indications for surgery. ¹⁰ The use of corneal reflex or grimace as an extension of the physical examination can elicit facial nerve function in patients who are unable to follow commands. This allows the timely diagnosis and assessment of the severity of facial nerve injury. Immediate and total facial nerve paralysis with concern for facial nerve transection is a clear indication for surgical exploration. Furthermore, most authors considered a comminuted fracture of the facial nerve canal on computed tomography as a consistent indication for exploration and repair. ¹⁰ Two-thirds of the studies reported the location of the facial nerve injury and the corresponding method of surgical intervention, including facial reanimation. The vertical portion of the facial nerve was the most common site of injury and was seen in 48% of patients with confirmed facial nerve injury. The tympanic portion of the nerve was the second-most common site of injury. Many patients had multiple segments of the facial nerve injured, making classification difficult. There were various facial nerve reanimation techniques utilized, but the most frequently used was an interpositional cable graft with the great auricular or sural nerve. When anastomosis was indicated, 59% of grafts involved an interpositional cable graft. The interpositional cable graft is frequently used. Tension-free primary neurorrhaphy is rarely possible because the injury frequently induces thermal injury to the frayed ends of the nerve, which have to be trimmed back to create a healthy anastomosis. ¹³ Alternatively there may be a complete transection of the nerve without thermal injury. Variability in practice setting and geographic location adds to the various possibilities of facial nerve injury. Future studies should specify immediate versus delayed facial weakness. They could characterize the degree of injury by describing function of the individual branches of the nerve. At this time, the House-Brackmann scale remains the standard assessment tool.

CSF Leak

Overall, 12% of patients developed a CSF leak, with a range of 8% to 14%. This was either at the time of presentation or after surgical exploration, but estimates may be inaccurate as some studies did not report CSF leak rate. The leak rate may be underreported as some leaks resolve spontaneously. This makes the reported low incidence of CSF leaks difficult to interpret. Postoperative CSF leak was sometimes included with preoperative CSF leaks, which confounds the study of outcomes. Postoperative CSF leaks could be related to the timing, duration, and extent of the procedure as well as the experience of the surgeon involved. Future studies should separate initial CSF leaks from postoperative CSF leaks and report laboratory confirmation with beta-2 transferrin where possible to facilitate differentiation.

Intracranial Complications

Intracranial injury was present in 43% of patients and included stroke, hematoma, abscess, carotid cavernous fistula, and carotid artery injury. Two of the 6 studies lack detail in the types of intracranial injuries.^8,10 Hooper et al described 2 intracranial injuries from the bullet itself, but no intervention was required. ¹⁵ Duncan et al used representative cases to describe bullets within the temporal lobe and a carotid cavernous fistula, but details were not provided for the other 6 patients with intracranial injuries. ⁶ Shindo et al provided the most detail, including subtypes of intracranial hemorrhage (eg, subdural vs subarachnoid) as well as descriptions of the neurologic complications: cortical injuries, stroke, pneumocephalus, and meningitis. ¹³ Meningitis occurred in 9% of their patients. CSF leak is a well-known risk factor for development of meningitis; however, 1 of the patients in their study developed meningitis despite no evidence of a CSF leak.

Vascular Injury

Vascular injury was reported in every study and occurred in 25% of patients. This is likely because vascular injuries direct the patient’s treatment priorities. In accordance with Advance Trauma Life Support protocol, patients are not evaluated for otologic intervention until hemodynamically stabilized. Vascular injury typically implied intracranial vascular injury, such as jugular foramen or carotid artery injury, but could also imply extracranial carotid artery injury. Because the most recent study in this review was 21 years old, most vascular injuries were discovered by angiography, which was not performed on every patient. ¹⁵ With the evolution of technology, computed tomography angiography is now more commonly utilized. Computed tomography angiography is helpful because it assists modern interventional radiologists or neurosurgeons with planning for embolization if necessary.

Surgical Considerations

Haberkamp et al described 5 major indications for surgical intervention: facial paralysis, vascular injury, CSF leak, external auditory canal destruction or cholesteatoma prevention, and intracranial injury. ¹⁰ Treatment algorithms are difficult to create, as each patient and pattern of injury are unique. Figure 2 and 3 demonstrate the significant comminution seen with these types of injuries. Subsequently, surgical interventions are carefully individualized. The greatest variation among the studies included in this review originates from the type of procedures and the timing in which they were performed. Urgent surgical indications include facial nerve paralysis and CSF leaks. Immediate facial paralysis with confirmation of a bony spicule abutting the facial nerve is a widely accepted indication for immediate exploration and facial nerve reanimation. If a patient has immediate complete facial paralysis, then an electroneuronography will typically be performed at least 72 hours after the injury. If the electroneuronography demonstrates 95% degeneration, then facial nerve decompression is recommended. Electromyography can be used several weeks following a traumatic injury to confirm complete nerve injury when there are no fibrillation potentials or if there is an absence of compound muscle action potentials. Electromyography can also be used to detect polyphasic potentials several months after an injury, which suggests nerve regeneration. ¹⁶ CSF leaks are typically managed first with CSF precautions, and if these fail, then a lumbar drain is placed. If the leak continues despite drain placement, then surgical intervention is recommended during the same admission. CSF leaks are managed aggressively because of the risk of developing meningitis. Vascular injuries are often initially managed by neurosurgery. Another important consideration is the risk of infection and cholesteatoma relating to the entry of skin into the middle ear and mastoid space. Gunshot wounds to the temporal bone universally introduce skin and mucosa into the middle ear and mastoid space, which portends eventual infection and cholesteatoma if not managed properly. These patients will eventually require surgery, but delayed surgical intervention as an outpatient is acceptable and sometimes recommended. Performing delayed outpatient surgery is recommended when patients remain critically ill and the delay will not reduce the risk of a positive outcome. In 1 study, the authors typically performed a canal wall down mastoidectomy, removed all visible bullet fragments, and obliterated the mastoid space to prevent future infection and cholesteatoma formation. ¹¹ Figure 4 demonstrates removal of bullet fragments embedded in a temporal bone. If a nerve anastomosis has been performed, closure of the external auditory canal may be considered, as this procedure will help to protect the nerve graft.

OPEN IN VIEWER

Figure 2.

Axial view of a computed tomography scan demonstrating a highly comminuted temporal bone fracture caused by a gunshot wound.

OPEN IN VIEWER

Figure 3.

Coronal view of a computed tomography scan demonstrating multiple bullet fragments within the temporal bone.

OPEN IN VIEWER

Figure 4.

Bullet fragment embedded in the mastoid cavity of a patient.

This review excluded non–English language articles, which may result in some missed studies. Several studies were excluded because they described patients who immediately died after injury. Numerous single-case reports of gunshot wounds to the temporal bone were excluded as these add unnecessary heterogeneity to the systematic review. Finally, follow-up is incompletely reported among the studies discussed. Selection bias may arise from poor follow-up of patients with positive outcomes, which would cause the results of a retrospective study to appear worse than actual patient outcomes.