Acute surgical intervention for a depressed skull fracture causing a laceration to the brain parenchyma from a bite wound in a dog

Abstract

A 5-month-old spayed female mixed breed dog was attacked by another dog causing multiple fractures of the left calvarium with a fragment penetrating through the gray matter of the parietal lobe. Surgery was performed to remove the bone fragment. A 6-month follow-up showed dramatic improvement in neurologic status.

Traumatic brain injury (TBI) is defined as an alteration in brain function caused by an external force (1). It is a common presentation in the veterinary emergency room and can be caused by dog bites, motor vehicle accidents, falls from heights, crush injuries, and human assaults (2). Most cases of TBI are treated medically; however, surgical intervention may become necessary. A major indication for surgery in human medicine is a poor response to medical treatment resulting in worsening intracranial pressure (ICP) measurements (3). Since ICP monitoring is rarely performed, reasons for surgery in veterinary medicine are controversial and may include inadequate response to medical therapy, depressed or open skull fractures, ongoing hemorrhage, foreign bodies, or hematomas (2,4). There is little in the veterinary literature about surgical intervention for TBI, and currently it is unknown even in human medicine if there is a positive correlation between surgery and clinical outcome (5). To the authors’ knowledge this is the first veterinary report describing acute surgical intervention (< 24 h) for depressed skull fractures with fragment penetration into the brain parenchyma resulting from a bite wound.

Case description

A 5-month-old 2.0-kg spayed female mixed breed dog was examined 3 h after being attacked by another dog. On presentation to the Veterinary Teaching Hospital emergency service, the dog had evidence of 2 open bite wounds on the dorsal aspect of her head with no other obvious external injuries. The dog had a heart rate of 164 beats/min with occasional ventricular premature complexes recorded on electrocardiogram examination. The dog was hypertensive with a systolic blood pressure measured at 210 mmHg with Doppler. On neurologic examination, the dog had an obtunded mentation and a rotary nystagmus in both eyes. No other cranial nerve deficits were identified. The dog was non-ambulatory tetraparetic with decreased withdrawal reflexes in both thoracic limbs. The modified Glasgow Coma Score (MGCS) was 13 (6). The patient’s injury was neurolocalized to the central vestibular areas (brainstem) due to the mentation change, ambulation deficits, and nystagmus. The decreased withdrawals could be explained by damage to the vestibulospinal tracts causing inhibition to the gray matter interneurons and/or injury to the medial longitudinal fasciculus, affecting the dorsal ventral funiculus of the cervical and cranial thoracic spinal cord segments.

Initial stabilization included intravenous boluses of Lactated Ringer’s solution (Hospira, Lake Forest, Illinois, USA), total dose of 45 mL/kg body weight (BW), mannitol (Nova-Tech, Grand Island, Nebraska, USA), 1 g/kg BW, IV bolus, pain management with fentanyl citrate (Hospira), 6 μg/kg BW bolus then 3 to 5 μg/kg BW/h, IV, and supplemental oxygen via an oxygen chamber (FiO₂ of 40%). The wounds on the head were clipped and cleaned aseptically with chlorhexidine gluconate solution (Hibiclens; Mölnlycke Health Care, Norcross, Georgia, USA) and antibiotics were started (ampicillin/sulbactam; Auromedics Pharma, Dayton, New Jersey), 30 mg/kg BW, IV, q8h.

An initial STAT blood profile (Critical Care Express; Nova Biomedical, Waltham, Massachusetts, USA) did not identify significant abnormalities. Following initial triage, the dog became stuporous. After stabilization treatments, the dog’s mental status changed from stuporous to dull. Using minimal sedation with fentanyl citrate, 5 μg/kg BW per hour, IV, and midazolam (West-Ward, Eatontown, New Jersey, USA), 0.1 mg/kg BW, IV, computed tomography (CT; GE Lightspeed 16-slice; GE Healthcare, Milwaukee, Wisconsin, USA) scan of the skull, cervical spine, and thorax was performed using the VetMouse Trap (Universal Medical Systems, Solon, Ohio, USA). Multiple fractures of the left calvarium involving the left frontal, parietal, and occipital bones were found (Figures 1A, B). The fractures were open with a small gas pocket seen within the calvarium indicating pneumocephaly. One bony fragment was noted to be displaced toward the left lateral ventricle, penetrating through the gray matter of the parietal lobe. The fractures and overall compression of the left hemisphere were also associated with mid-line shift of the falx cerebri, and reduced attenuation of the left cerebral white matter consistent with white matter edema (Figures 2A, B). The patient was hospitalized overnight with supportive care and monitoring.

Figure 1.

Transverse (A) and sagittal (B) CT images of the head showing multiple comminuted fractures of the left calvarium (triangles) and associated fragment penetration of the left parietal lobe (arrow).

Figure 2.

Transverse CT images of the head showing compression of the left hemisphere associated with mid-line shift of the falx cerebri (A) and reduced attenuation of the left cerebral white matter consistent with white matter edema (B, arrow).

During the first 12 h of hospitalization, the dog returned to a stuporous state. The dog developed an intermittent rotary and vertical nystagmus OU, miotic pupils OU, and absent menace OU. The dog continued to be non-ambulatory tetraparetic. Her MGCS at this time decreased to 11. An additional dose of mannitol (0.5 g/kg BW, IV) was administered. Her mentation improved to an obtunded state, with normalization of pupil size and a MGCS of 15 was recorded. Although her mental status improved with the additional dose of mannitol, a craniectomy was advised due to concerns of increased intracranial pressure secondary to the compressive fracture.

Based on the location of the depressed skull fracture on CT and history of a bite injury with concern for bacterial infection, surgical intervention was recommended to remove the bone fragment and lavage the wound. Levetiracetam (Keppra; Hospira), 30 mg/kg BW, IV, q8h, was started before surgery, which was conducted 16 h after presentation to the emergency room. The dog was premedicated with midazolam, 0.2 mg/kg BW, IV, lidocaine (Hospira), 2 mg/kg BW, IV, and fentanyl citrate, 10 μg/kg BW, IV, and induced with propofol (Hospira), 1 mg/kg BW, IV. A continuous rate infusion of propofol, fentanyl citrate, and lidocaine was continued during surgery. Inhalant anesthetics were not used during the procedure.

The patient was positioned in sternal recumbency with the head elevated. The hair over the calvarium was shaved and the area prepared with chlorhexidine gluconate solution. A ~ 4-cm skin incision was made along the midline over the frontoparietal region. The underlying musculature was bruised and edematous. The interscutularis, frontalis, and temporalis muscles were sharply dissected with a #15 blade and reflected away from the skull using a periosteal elevator.

The fractured parietal bone was observed upon exposure of the skull. A depressed fracture fragment in the parietal lobe was penetrating deep into the brain parenchyma. This bone fragment was removed and measured ~ 1.5 cm in diameter. There was a mild amount of hemorrhage and a small area (~ 2 cm) of damaged cerebral cortex from which the bone fragment had been removed. The remaining cortex appeared grossly normal. The surgical site was flushed and lavaged with 0.9% sodium chloride irrigation (Hospira) to minimize contamination from the bite wound. An aerobic culture was negative for growth. There was no evidence of gross hemorrhage or brain swelling prior to closure. A thin piece of gelfoam was placed over the skull defect. The musculature was apposed using 3-0 polydioxanone suture (PDS; Ethicon, Guaynabo, Puerto Rico) in a simple continuous pattern, and the subcutaneous layer was closed in a similar manner. The skin was closed in an intradermal pattern with 4-0 poliglecaprone 25 suture (Monocryl; Ethicon). The approximately 1-cm diameter puncture wound lateral to the surgical incision was debrided with a #10 blade and flushed with sterile saline. The patient was stable under anesthesia and did not experience any significant episodes of hypoventilation (mean ETCO₂ 35 mmHg), hypoxia (mean SpO₂100%), or hypotension (mean systolic blood pressure 112 mmHg). The dog recovered from anesthesia uneventfully.

One day after surgery, the dog’s neurologic status was slightly improved. The dog was able to hold her head up but continued to be intermittently obtunded. The dog had a more significant rotary nystagmus, developed anisocoria, and had a persistent absent menace response in both eyes. The dog continued to be nonambulatory tetraparetic but had improved motor activity within the thoracic limbs. The MGCS was similar to the pre-operative value at 15.

The dog continued to show neurological improvement during the following 5 d of hospitalization. However, her mentation remained obtunded, with absent menace responses in both eyes and vestibular signs (rolling, head tilt to the left, nystagmus). During this time, the dog was transitioned to oral medications which included alprazolam (Sandoz, Princeton, New Jersey, USA), 0.03 mg/kg BW, PO, q8 to 12h, meclizine (Rugby Laboratories, Livonia, Michigan, USA), 3.1 mg/kg BW, PO, q24h, levetiracetam (Keppra; Northstar Rx, Memphis, Tennessee, USA), 30 mg/kg BW, PO, q8h, meloxicam (Metacam; Boehringer Ingelheim Vetmedica, St. Joseph, Missouri, USA), 0.1 mg/kg BW, PO, q24h, amantadine hydrochloride (Pharmaceutical Associates, Greenville, South Carolina, USA), 2 mg/kg BW, PO, q24h, amoxicillin trihydrate/clavulanate potassium (Clavamox; Zoetis, Kalamazoo, Michigan, USA), 15.6 mg/kg BW, PO, q12h, gabapentin (Amneal Pharmaceuticals, Bridgewater, New Jersey, USA), 10 mg/kg BW, PO, q8 to 12h, and tramadol (Amneal Pharmaceuticals of New York, Hauppauge, New York, USA), 3 mg/kg BW, PO, q8 to 12h. The dog was discharged 7 d after surgery to the care of the owners with a MGCS of 16. The dog’s neurological abnormalities at time of discharge included a dull mentation, improved but persistent rotary nystagmus, absent menace response in both eyes, and a head tilt to the left. The dog was unable to ambulate due to the profound vestibular ataxia, but had strong motor activity in all limbs.

The 1 wk recheck from discharge showed slow improvement. The dog continued to be dull, but was more responsive and aware of her surroundings. The nystagmus was no longer present. The dog was also regaining some vision in her left eye and had an intermittent menace response, but continued to have an absent menace response in the right eye. She was experiencing profound vestibular ataxia and tried to roll to the left continuously. She had excellent withdrawals and motor activity in all limbs.

The dog continued to improve, and at her 2 mo recheck she was bright and alert and had a menace response in both eyes. There was a mild left head tilt and she was ambulatory with moderate vestibular ataxia. In addition, according to the owners, she had regained most of her pre-injury personality. At 6 mo, the owners stated that they were very happy with the dog’s continued improvement with only minor vestibular ataxia still present.

Discussion

This case report describes the use of a craniectomy within 24 h of TBI to remove a penetrating bone fragment causing a parenchymal brain lesion. There is a dearth of literature in veterinary medicine on craniectomies in the setting of TBI. There have been case reports of dogs and cats having craniectomies performed; however, these reports describe debridement of an abscess associated with a previous injury or removal of a partial bullet fragment a few days after the incident in a patient with normal mental status (7–11). The present case specifically reports surgery in the acute setting (< 24 h) after a laceration causing a compressive fracture within the brain parenchyma.

Based on the examination, this patient’s injury was neurolocalized to the central vestibular portions of the brain based on her mentation (abnormal level of consciousness), gait abnormalities (tetraparesis and vestibular ataxia), and cranial nerve deficits (nystagmus). The progressive changes on her neurologic examination, specifically the miosis and worsening mentation change, indicated compression of the midbrain or a severe prosencephalic lesion. The progression to a deficit with the menace response most likely indicated a lesion affecting the central visual cortex.

The CT imaging showed evidence of the fracture affecting the skull and causing damage to the cerebral cortex. There are 2 potential explanations for her neurologic clinical symptoms. One possibility is that the cerebral edema also caused secondary edema and swelling leading to increased intracranial pressure causing transtentorial and/or transforaminal herniation, which is likely, given the clinical signs of a Cushing’s reflex, changes in mentation, and pupil size. The brainstem is difficult to accurately image via CT (due to beam hardening artifact). The other possibility is that the traumatic injury caused a lesion in the conscious vestibular pathway. This is less well-described in the canine species and is likely associated with the conscious auditory pathway projecting from the thalamus (via the medial geniculate nucleus), through the internal capsule, to the vestibular cortex in the temporal lobe (12).

There is substantial controversy about the use of medical versus surgical management for stabilization of veterinary patients with TBI after primary injury (which is the acute, direct injury causing mechanical damage). Medical management is aimed at minimizing secondary injury, which occurs minutes to days following the trauma and consists of systemic physiologic insults that may decrease cerebral perfusion pressure, worsen cerebral edema, increase intracranial pressure, and damage the blood-brain barrier (2). Our patient, therefore, received 16 h of medical management with intravenous fluids to prevent hypotension, oxygenation to prevent hypoxia, and mannitol to decrease presumed spikes in intracranial pressure from cerebral edema (13). Seizures have also been reported to worsen secondary injury, and given that this patient had several risk factors for the development of early seizures after TBI (13), the dog was started on prophylactic anti-convulsants (14,15). During this stabilization period, the neurologic status of the patient was assessed serially to determine the efficacy of treatments (including motor activity, brain stem reflexes, and level of consciousness to calculate the MGCS score) (6).

While medical therapy remains the mainstay of TBI in veterinary medicine, surgery may be warranted in cases of failed medical treatment, ongoing hemorrhage, hematomas, foreign bodies, pneumocephalus, depressed skull fractures, midline shift of falx cerebri, and compression of the basilar cisterns (2,4,9,16,17). These recommendations are based on little data, as the only experimental surgical study involved healthy dogs and concluded that ICP decreased after the combination of surgical therapies (including craniectomy and durotomy) and medical therapies (including ICP lowering treatments). This study, however, did not differentiate medical versus surgical management (18). In our patient, although medical management was associated with an overall neurological improvement based on MGCS, surgery was recommended due to the presence of depressed skull fractures, suspected contamination from the bite wound, pneumocephaly, and a midline shift on CT. After surgery, the dog displayed continued evidence of neurological improvement, and at 6 mo after surgery the dog only had mild neurological deficits.

There have been numerous human studies investigating the role of decompressive craniectomies in TBI with varying outcomes. Some of the evidence indicates that they may be useful in patients with severe TBI refractory to medical therapy. A 2012 comprehensive review in human medicine showed that decompressive craniectomies may be associated with reduced ICP values at 24 h and 48 h, although the correlation of ICP with clinical outcome is uncertain (19). In addition, the RESCUEicp study concluded that patients undergoing surgery had lower mortality rates at 6 mo after injury, but higher vegetative states, compared with standard medical therapy (5). The DECRA trial, evaluating early decompressive craniectomy (< 72 h) versus standard medical therapy in patients with severe diffuse TBI and refractory intracranial hypertension, showed that although surgery decreased ICP and length of stay in the intensive care unit, it was associated with more unfavorable outcomes (20). Similarly, a 2016 systematic review concluded that decompressive craniectomy was not associated with a favorable outcome compared to medical therapy (1). Due to a lack of consensus on treatments for TBI in humans, the Brain Trauma Foundation recommendations were modified to include decompressive craniectomies for the management of severe TBI, but recognize that there is insufficient high-quality body of evidence on the topic (21).

Another ongoing debate in both human and veterinary medicine involves the ideal timing of surgical intervention after TBI. Given that there have been several reports of intracranial abscess formation after a bite wound in dogs (7,8,10), it was presumed in this case that early surgical intervention and debridement of an open wound to the brain parenchyma would minimize the risk of bacterial proliferation and decrease patient morbidity. As such, surgery was recommended as soon as the dog was stable (< 24 h after injury), and was associated with a favorable outcome. There is no veterinary evidence to support ideal timing of surgical debridement, and human literature is contradictory on this matter. The time frame to the development of generalized infection is considered to be approximately 6 h, and thus prompt wound management would be ideal (22). Culture results from samples obtained during surgery were negative, although only aerobic cultures were submitted. Aerobic bacteria are most common in nonpurulent dog bite wounds which is the reason, in addition to cost, anaerobic culture was not performed. It is possible that anaerobic bacteria were present (23). One prospective human study showed that there was no detriment in outcome in patients who received early (< 24 h) surgical management with a decompressive craniectomy compared to those who had longer medical management and stabilization (> 24 h) before surgery (17). Another human study, however, concluded that a higher survival rate was found in the group of patients who had late decompressive craniectomies (after > 24 h of stabilization or clinical deterioration), compared to early surgical intervention (< 24 h after injury) (24).

Finally, recommendations about treatment may be centered on certain prognostic indicators associated with TBI. In humans, poor prognostic indicators on advanced imaging include compressed or absent basilar cisterns, the presence of subarachnoid hemorrhage, midline shift, and intracranial hemorrhagic lesions (25). In addition, skull fractures in certain anatomical locations, brain herniation, and an increased size of the intraparenchymal lesion, have also been associated with a poor prognosis in dogs and humans (26–29). In dogs with head trauma (but not specifically TBI), prognosis was also associated linearly with a decreased MGCS, poor perfusion (evidenced by metabolic acidosis and hyperlactatemia) and decreased SpO₂, severe concurrent injuries, or a requirement for hypertonic saline or endotracheal intubation (30). In the present case, the dog displayed many of these negative prognostic indicators on presentation, including intracranial hemorrhage, midline shift, and skull fractures. Despite these, the dog showed some improvement with pre-operative medical management (oxygen, intravenous fluids, and mannitol therapy), and remained stable during the peri- and post-operative periods. Although it is unknown whether medical management would have resulted in the same outcome, this case report may provide evidence to support that the use of early aggressive medical management and surgical intervention can be associated with a positive outcome, even in severely affected dogs.

We describe a canine patient with a compressive skull fracture and penetration into the brain parenchyma after a traumatic injury and subsequent surgical removal of the fragment. After surgery the patient continued to improve, with only mild neurologic abnormalities remaining. Surgical intervention may be indicated in patients with contaminated, compressive, and open fractures into the brain parenchyma. Based on the insufficient data on decompressive craniectomies in dogs with TBI, additional studies are necessary to determine which patients will benefit from medical or surgical therapy, the ideal timing of surgery, and prognostic indicators. Until then, every TBI case should be assessed individually and the clinician should determine whether medical therapy alone is sufficient, or if the patient may benefit from surgical intervention, based on their type and severity of injuries, level of contamination, response to medical management, and ability to withstand anesthesia. As seen in this case report, aggressive therapy may yield a good prognosis despite severe parenchymal lesions of the brain. CVJ