Traumatic Brain Injury in a 1-year-old Standard Poodle Anna Sarfaty Advisors: Dr. Jordyn Boesch and Dr. Francesca Di Mauro Senior Seminar Paper Cornell University College of Veterinary Medicine February 15, 2017 Key words: Traumatic brain injury, TBI, anesthesia, post-traumatic seizure 16 Abstract A 1-year-old female spayed Standard Poodle was presented after being hit by a car. Immediately after the accident, the patient had 1 or 2 seizures. The initial assessment of the patient was consistent with traumatic brain injury (TBI); she was stuporous and laterally recumbent, with miotic pupils, no menace response, a decreased gag reflex, and absent proprioception. Deep pain and spinal reflexes were present. Besides an open fracture, the rest of her examination was unremarkable. The patient was stabilized with mannitol, head elevation, intravenous (IV) fluid therapy, and analgesia. She was hospitalized with anti-epileptic medications, broad spectrum antibiotics, anti-emetics, IV fluids, analgesia, continuous electrocardiogram and blood pressure monitoring, and supportive care. Over the next 8 days, the patient’s neurologic status improved steadily. On the 9th day, the patient’s right radius-ulna fracture was repaired. The patient was discharged two days later when she was fully bright and ambulatory. Case History The patient, a 1-year-old female spayed Standard Poodle, was presented to the Cornell University Hospital for Animals (CUHA) Emergency Service in October 2016 for evaluation after being hit by a car the previous evening. The patient was presented to the referring veterinarian immediately after the accident. On presentation, the patient was stuporous with miotic pupils. She also had an open mid-diaphyseal comminuted spiral fracture of her right radius and ulna with craniolateral displacement. The patient received hypertonic saline immediately and was managed overnight with mannitol every 6 hours, cefazolin, and a fentanyl continuous rate infusion (CRI). She had 1 or 2 seizures successfully treated with boluses of midazolam, and was later treated with leveteracitam on the advice of the Cornell Emergency Service. Overnight and throughout the following day, the patient improved slightly with some periods of responsiveness, though she was generally stuporous. Thoracic and abdominal radiographs were performed and were unremarkable. Before being transferred to Cornell in the evening, she received a dose of hydromorphone. Initial Presentation On presentation at CUHA, the patient was stuporous and laterally recumbent. She weighed 27.4 kg. She was tachypneic at 60 breaths per minute, normocardic at 95 beats per minute, and normothermic. The patient had scleral hemorrhage OD and contusions on her buccal mucosa. The dorsal aspect of her right antebrachium was shaved, revealing a 2 cm open fracture wound (grade 21) on the dorsal aspect. No other palpable fractures were noted, and the rest of her general exam was unremarkable. On neurologic exam, the patient had miotic pupils, absent pupillary light reflexes (PLR) and menace responses bilaterally. Her gag reflex was decreased. Palpebral reflexes were present bilaterally. The remainder of her cranial nerves were intact. She had absent placing but positive withdrawals in her non-fractured limbs (right hind, left hind, and left front). Her cutaneous trunci, patellar, and perineal reflexes were intact, and she had proper anal tone. Deep pain was present, and the patient exhibited no pain on spinal or head palpation. Point-of-care bloodwork revealed a packed cell volume (PCV) of 47%, total protein (TP) of 7.6 g/dL, Azostix® of 5-15 mg/dL, and metabolic acidosis with a pH of 7.31 and a base excess of -6.2 mEq/L. The patient was hypernatremic at 153.5 mmol/L and hypokalemic at 3.25 mmol/L. The patient’s blood pressure was high normal at 124 mmHg systolic, 104 mmHg diastolic, and 115 mmHg mean. An electrocardiogram (ECG) revealed a sinus rhythm with occasional ventricular premature contractions. A quick assessment abdominal ultrasound (AFAST) revealed no free fluid. A quick assessment thoracic ultrasound (TFAST) showed no evidence of pneumothorax, or pleural or pericardial fluid. Problem List After initial assessment, the patient’s neurological abnormalities localized to brain disease, with no indication of spinal trauma. Because of the history of trauma, TBI was the primary problem and working diagnosis. Secondary problems included the open fracture of the right radius and ulna, and metabolic acidosis with electrolyte abnormalities. Prognosis The Modified Glasgow Coma Scale (MGCS) has been validated as a prognostic indicator in dogs and remains the best means of determining a prognosis for survival within 48 hours in small animal TBI patients.2 The 18-point scale awards 6 points for motor activity, 6 points for brain stem reflexes, and 6 points for level of consciousness. The patient earned an 11 on her initial presentation: a 6 for motor activity (though she was recumbent, her spinal reflexes were normal), a 3 for her brain stem reflexes (absent oculocephalic reflexes), and a 2 for her stuporous level of consciousness. The patient therefore had a guarded prognosis. Patients with MGCS scores between 3 and 8 have a grave prognosis, 9-14 a guarded prognosis, and 15-18 a good prognosis. Critical Care Summary After the patient’s initial assessment, her head was elevated 15-20 degrees, and she was given mannitol (0.5 mg/kg IV over 20 minutes). She received methadone (0.25 mg/kg IV), ampicillin and sulbactam (30 mg/kg IV), and maropitant (1 mg/kg IV), and she was started on Normosol-R isotonic crystalloid fluids at 90 ml/kg/day. Her fracture site was examined and rebandaged with the referring veterinarian’s spoon splint before being transferred to the intensive care unit (ICU) under the management of the Critical Care Service. In the ICU, the patient was hospitalized with an indwelling urinary catheter, continuous electrocardiogram, IV fluid therapy (Normosol R at 60 ml/kg/day + 30 mEq KCl), a fentanyl continuous rate infusion (CRI, 2 mcg/kg/hour), levetiracitam (30 mg/kg IV q8), ampicillin and sulbactam (22 mg/kg IV q8), and zonisamide (30 mg/kg per rectum q12). Her MGCS was monitored every 4 hours, her blood pressure every 30 minutes, and she was on a continuous seizure watch, with instructions to give levetiracetam (30 mg/kg IV) if noted. At 12 AM, the patient’s MGCS was 12 as her PLRs were now present, though slow. Her pupils were still miotic. At 2 AM, the patient appeared more responsive to stimuli. Within the first 24 hours of hospitalization, her neurologic status improved. The next day, the patient was sedated for full body computed tomography (CT) imaging. A focal area of hypoattenuation was noted in the left lateral cerebrum which did not accumulate contrast, suggesting edema or contusion. A small mass effect, suggesting a hematoma, was noted on the left lateral aspect of the third ventricle. The CT also showed small, non-displaced fractures of her right upper maxilla and right mandibular condyle, and a complicated crown fracture of her right upper third premolar. A nasogastric tube was placed to provide adequate enteral nutrition, and the patient was started on metoclopramide (to treat ileus secondary to opioids) and Clinicare. A liver panel showed elevation in hepatocellular enzymes and total bilirubin, which was suspected to be the result of direct injury to hepatocytes caused by blunt trauma. On the 2nd day of hospitalization, the patient’s neurologic exam did not change. However, she began to have periods of arousal in which she would vocalize, seem distressed, and attempt to move around the cage. She received a CRI of dexmedetomidine (0.5 mcg/kg/hour). The liver panel was repeated and enzyme levels were improved, and her wound was healing well at this point. On the 3rd day, the patient exhibited increasingly frequent episodes of arousal and anxiety. During the episodes, she was able to stand without assistance, but she remained in lateral recumbency for the remainder of the time. Proprioception was present in her left forelimb for the first time. The patient had a good appetite despite experiencing difficulty prehending food. On the 4th day, the patient’s pupils were no longer miotic, though she remained nonvisual. Her MGCS had improved to 13, and she continued to eat well. Dexmedetomidine boluses (1 mcg/kg IV) were effective at treating decreasingly frequent periods of agitation. A lateral splint was placed instead of the original spoon splint on her right front radius ulna fracture. Over the next 4 days, the patient continued to improve. She was fully transitioned to oral medications and received a fentanyl patch for analgesia. By the 4th day, her MGCS was 18. Though her menace response had not returned, the patient appeared able to visually navigate her environment. Throughout her ICU hospitalization, the patient’s PCV, TP, and blood gas/acid-base status were evaluated daily. Her metabolic acidosis remained unchanged, with a base excess ranging from -6 to -9 mEq/L. Orthopedic Surgery and Anesthesia Summary On the 9th day, the patient was prepared for anesthesia and surgical repair of her right radius and ulna fracture. The patient received methadone (0.1 mg/kg IV), maropitant (1 mg/kg IV), sodium citrate (0.2 mL/kg PO), and lidocaine (1 mg/kg IV) for premedication. She was induced with propofol (3 mg/kg IV). She received an ultrasound-guided brachial plexus block with bupivacaine (2 mg/kg), triamcinolone (0.1 mcg/kg), and dexmedetomidine (1 mcg/kg). During anesthesia, she received isoflurane between 1.25% and 1.75%, and her end tidal carbon dioxide was kept between 30 mmHg and 36mmHg. The surgery was technically challenging but proceeded without complication. Pre-surgical repeat radiographs did not reveal any signs of infection at the site. If there were signs of infection, an external fixator would have been the most appropriate means to repair the fracture.1 Since there were no signs of infection, a locking compression plate was used. By this time the open wound on the dorsal aspect had healed, and the approach was made on the lateral aspect. Since the fracture was 10 days old, a large amount of callus had formed at the site of the fracture, which would have resulted in malunion if not surgically corrected. The callus was broken down with manual debridement. The radial fracture was reduced then secured with the locking compression plate. The patient recovered smoothly from anesthesia but remained extremely sedate for approximately 4 hours. After 4 hours, she was alert and able to swallow her oral medications safely. Outcome The patient was discharged 2 days post-operatively when she was eating well, fully bright, and ambulatory. The morning after the surgery, the patient was quiet, alert and responsive with minimal interest in food, so she was hospitalized for 1 more day. The patient knuckled on her right front limb when attempting to bear weight for the first 24 hours after surgery, presumably due to the long-lasting effects of the brachial plexus block. The patient’s metabolic acidosis, characterized by a normal anion gap, persisted throughout hospitalization. Differentials for a normal anion gap metabolic acidosis include a loss of bicarbonate through the gut (through diarrhea) or the kidneys (renal tubular acidosis).3 Neither differential fit the patient as she did not have diarrhea and had no evidence of renal tubular damage on chemistry profile and urinalysis. As the patient was clinically improved, it was decided to discharge her without an explanation for her metabolic acidosis. The patient was discharged with instructions for activity restriction and mentation and seizure monitoring. A standard surgical recheck schedule was discussed, with incision inspection at 2 weeks and recheck radiographs at 8 weeks. The patient was discharged with a month’s supply of levetiracetam (30 mg/kg PO q8) and zonisamide (5 mg/kg PO q12), with the instructions to continue administration until directed otherwise by a veterinarian. The updated prognosis was also discussed fully with the owners. The prognosis for healing of the radius-ulna fracture was very good. Prognosis for TBI was also good based on improvement in MGCS noted within the first 48 hours, with some conditions. Although the patient was able to navigate in the hospital environment, her menace response was still absent. Furthermore, though she did properly posture to defecate and urinate outside, she urinated and defecated in her cage and would not attempt to get away from her urine and feces. The owners were informed that her sight and these learned behaviors might continue to improve, but they might not. The patient might need housebreaking training again. The owners were understanding. A few days later, the owners reported that the patient was bright and properly responsive to her environment. She did appear to mostly have sight, but a few times she had walked into a couch and a door. She mostly urinated and defecated outside, but had also defecated in the house, which was unusual for her. They reported that she was being kept exercise restricted. Approximately 3 weeks later, the patient was presented to the Orthopedic Service for evaluation of progressive right front limb lameness. The owners reported that the patient had appeared fully recovered from her TBI at home. However, she had not been fully weight bearing on the limb since the surgery, and they had noted right antebrachial swelling 3-4 days before the office visit. On examination, the patient had a present but decreased menace response, and otherwise appeared to be neurologically appropriate, though a full neurological exam was not performed. The patient had diffuse right antebrachial swelling and a firm object was palpable on the craniomedial aspect of the antebrachium. Radiographs revealed diffuse soft tissue swelling, but there were no signs of infection or implant failure. The object under the skin appeared to be a subcutaneous bone fragment. The owners declined a recommendation to remove the bone fragment, and the patient was discharged with gabapentin and instructions for strict rest until the 8-week post-operative checkup. The owners were also instructed to continue the anti-epileptic medications until the 8-week post-operative checkup. The owners did not return to Cornell for the 8-week checkup. Approximately 4 months after the injury, the owners report that the patient completely recovered her neurologic function with no residual deficits in her vision or learned behaviors. The owners did retrain the patient for housebreaking with a crate and a strict schedule, and the patient successfully relearned these behaviors. They reported that the patient still had a limp on her right front limb, and a recheck with the Orthopedic Service had been scheduled. Discussion Pathophysiology of TBI A direct impact creates primary and secondary injuries in nervous tissue. Primary injuries are created by physical forces and occur at the time of impact.4 They may be focal or diffuse. Types of primary injuries include skull fracture, hemorrhage within the cranial vault, hematomas (either epidural or subdural), and axon twisting.4,5,6 Neuron cell bodies can survive jostling but once the axon is twisted, the cell cannot survive. Once these injuries occur, they are resistant to current medical therapy. After primary injuries occur, perfusion to the brain is compromised. Cerebral perfusion pressure (CPP), or the pressure of blood flowing to the brain, is the difference between the mean arterial pressure (MAP) and intracranial pressure (ICP) (CPP = MAP- ICP).4,5,6 When primary injuries raise ICP, the brain can compensate initially. The Monroe-Kellie doctrine states that the space within the calvarium, containing brain parenchyma, cerebrospinal fluid, and blood, may be kept at a constant pressure if an increase in one component (such as brain parenchyma due to edema) is followed by a decrease in another component (such as blood).4,5,6 But in many cases of TBI, the brain’s compensatory abilities are overwhelmed, and a pathologic increase in ICP occurs, leading to a decrease in CPP.5 A decrease in CPP, along with the primary injuries, create secondary injuries.4 Locally, neurons release glutamate. Glutamate accumulation leads to a loss of ionic gradients across cells, influx of calcium into cells, and generation of free radicals. Inflammatory mediators, along with the increased intracellular calcium and oxygen free radicals, create an increase in the production of nitric oxide and an increase in coagulation and thrombosis. Nitric oxide also alters blood flow and vascular permeability. The end result of these processes, which potentiate each other, is cytotoxic edema, infarction, further increase in intracranial pressure (ICP), and loss of autoregulation—the critical ability of the brain to maintain cerebral perfusion within a range of systemic cardiovascular states.4,5,6 With secondary injuries and the loss of autoregulation, systemic health—especially its effects on MAP—becomes extremely important for maintaining cerebral perfusion.4,5,6 Unfortunately TBI patients often have compromised cardiovascular and metabolic health, including acid-base disturbances, electrolyte disturbances, and hypo- and hypercapnia, as well as other traumatic injuries. Systemic support of the patient is therefore of utmost importance in managing a TBI. Assessment and Treatment of the TBI Patient In humans, ICP is directly measured in TBI patients, but these monitors are not yet standard in veterinary medicine.7 Neurologic signs and reflexes associated with increased ICP must be carefully monitored in any patient who might have sustained a TBI.4 Ideally, neurologic assessment should occur before the administration of analgesics, especially opioids, which might alter results.4 Neurologic assessment must include brainstem reflexes (PLR, pupil size, and oculocephalic reflex), level of consciousness, and motor activity (including actual ability to ambulate, any paresis or rigidity, and spinal reflexes). As mentioned previously, the MGCS provides a score based on this assessment that can be used to formulate a prognosis.2 Other neurologic abnormalities, including cranial nerve abnormalities, should be noted and repeatedly assessed for progress during treatment. Severely increased ICP can create a physiological response known as the Cushing’s reflex.4,5,6 Increased ICP compresses cerebral arteries and leads to reduced blood flow to brain tissue, which causes an increase in CSF carbon dioxide tension that is then detected by central chemoreceptors in the medulla. These stimulate the vasomotor center in the medulla, which stimulates preganglionic sympathetic neuronal cell bodies in the spinal cord. These neurons cause vasoconstriction in systemic arteries and stimulate the adrenal medulla to release catecholamines that also cause vasoconstriction, raising blood pressure.6 However, this increase in MAP causes the baroreceptors in the aorta and carotid sinus to decrease heart rate in an attempt to keep cardiac output constant. Therefore, bradycardia may be a signal of severely increased ICP; a normal heart rate in a painful or hypovolemic patient may also suggest a Cushing’s response.4,5,6 Systemic parameters must also be assessed. The trauma patient should be examined for additional traumatic injuries. Any systemic abnormalities should be addressed; cardiovascular and respiratory injuries especially can worsen secondary brain injury. Heart rate and blood pressure ideally should be continuously monitored for indications of increased ICP. Temperature regulation may be affected in the TBI patient, and hyperthermia may result. The patient should be cooled if this is noted.4 Treatment for TBI focuses on stabilizing the patient and mitigating the damage caused by secondary injury in order to decrease ICP.4,5,6 Oxygen and IV fluid therapy with isotonic crystalloids should be provided.4,5,6 Full mu agonist opioids are preferred for pain management, as they are reversible, and when given as a CRI, can be titrated to effect. Commonly used methods for decreasing ICP include elevating the head and hyperosmolar therapy.4,5,6 Elevating the head between 15 and 30 degrees reduces cerebral blood volume while still maintaining adequate CPP. Ideally, a board would be used instead of a rolled up towel or pillow that can further traumatize the neck or occlude jugular veins. Hyperosmolar therapy targets intracranial hypertension, thus decreasing ICP.4,5,6 Mannitol, an osmotic diuretic, and hypertonic saline are both equally effective at decreasing ICP as they both are excluded from the brain by the blood brain barrier. Mannitol has additional free radical scavenging abilities which are desirable in a TBI patient; however, mannitol is contraindicated in hypovolemic patients due to its osmotic diuretic effects. Hypertonic saline will cause transient electrolyte abnormalities such as hypernatremia and hyperchloremia; keeping the sodium concentration below 160 mEq/L is advised.4 If the patient does not respond to either mannitol or hypertonic saline, the other agent may be considered as long as it is not contraindicated.3 Posttraumatic seizures can occur in TBI patients.4,8 One retrospective study reported that 10% of TBI patients at a teaching hospital had in-hospital seizures.8 Seizures can increase ICP by increasing cerebral oxygen demand.4,8 In human medicine, data support the use of prophylactic antiepileptic treatment for 7 days post TBI; however, no data support the use of prophylactic antiepileptic medication exists in veterinary medicine.4,7 If seizures are noted, aggressive antiepileptic drug therapy is indicated.4 Once initiated, therapy should be continued at least until clinical signs resolve; no specific recommendations regarding duration of therapy exist. Anesthesia of the TBI Patient Anesthesia is often necessary in TBI patients to enable imaging, surgery, and other interventions. Ideally, patients would not be placed under anesthesia until they are stabilized and do not show signs of increased ICP.9,10 If the patient exhibits signs of increased ICP at the time of anesthesia, measures to reduce ICP should be instituted prior to anesthesia.10 As discussed prior, these may include administration of hyperosmolar therapy and elevation of the head. Lowering the end tidal carbon dioxide partial pressure by 10 mmHg can decrease ICP by 30% in 15 seconds; carbon dioxide should then be kept at a low normal level, as discussed below.10 Total intravenous anesthesia (TIVA) may also be considered, as volatile anesthetics (isoflurane, sevoflurane) have a dose dependent relationship with ICP, with doses higher than 1-1.5 times minimum alveolar concentration raising ICP.9,10 Even if the TBI patient does not currently show signs of increased ICP, anesthetic agents alter blood flow to the brain and thus may raise ICP and potentiate secondary injuries.10 Increasing the amount of blood within the cranium—for example, through vasodilation—will increase ICP when compliance mechanisms are exhausted.5 Therefore, the primary goal in anesthesia of the TBI patient is to maintain adequate cerebral perfusion.9,10 Cardiovascular and respiratory monitoring (electrocardiogram, direct blood pressure, blood gas analysis, etc.) are therefore of special importance in these patients. In healthy brain tissue, hypercapnia leads to dilation of cerebral arterioles, thus increasing ICP.6 For TBI patients, it is recommended to keep carbon dioxide partial pressure at the low end of normal (30-35 mmHg).9,10 Though only strictly necessary to maintain this low normal partial pressure of carbon dioxide for patients with actively increased ICP, there is no harm to err on the side of caution and provide it to every patient at risk for elevated ICP. Aside from using a low dose of a volatile anesthetic (or TIVA), additional measures are recommended to minimize the risk of increasing ICP. Coughing and vomiting can increase ICP; providing systemic lidocaine (1 mg/kg IV) as a premedication can prevent cough, and anti-emetics such as maropitant can prevent vomiting.10,11 Providing adequate analgesia—either given systemically or locally as a nerve block—can also lower the patient’s requirement for anesthetic drugs, limiting their effect on ICP.9,10 As in critical care medicine, full mu opioids are the systemic analgesic of choice for anesthesia as they are reversible, effective, and have minimal cardiovascular effects.4,10 Using a CRI allows titration for frequent assessment of comfort, anesthetic depth, and neurologic status. Some anesthetic drugs also have neuroprotective properties by decreasing the brain’s requirement for oxygen. Propofol is an induction agent which potentiates effects of the inhibitory neurotransmitter GABA on GABA receptors. Propofol decreases cerebral oxygen demand, and also has free radical scavenging properties.10 Lidocaine, a local anesthetic, can prevent cough, blunt cardiovascular changes caused by intubation, decrease cerebral edema, scavange free radicals, and decrease ischemic injury.11 NMDA receptors on nerve cells contribute to ischemic injury; theoretically, their antagonists such as ketamine, amantadine, and xenon may protect against ischemic damage, though these treatments are not yet standard in TBI patients.10 For many years, ketamine was not recommended as an analgesic for TBI patients due to concerns that it increases ICP through sympathetic stimulation; however, recent evidence from the human literature suggests no demonstrable adverse effects in sedated and ventilated patients.4,10,12 Ketamine may increase cerebral oxygen consumption through inhibition of the GABA receptor, so administration along with a medication to potentiate those receptors, such as propofol, is recommended.4,10 References DeCamp, C.E., Johnston, S.A., Dejardin, L.M., and Schaefer, S.L. (2015). Brinker, Piermattei, and Flo’s Handbook of Small Animal Orthopedics and Fracture Repair, 5th Ed. Saunders. Platt S., Radaelli S., and McDonnell J. (2001). The Prognostic Value of the Modified Glasgow Coma Scale in Head Trauma in Dogs. Journal of Veterinary Internal Medicine. 15(6): 581–4. DiBartola, S.P. (2011). Fluid, Electrolyte, and Acid-Base Disorders in Small Animal Practice, 4th Ed. Saunders. DiFazio, J. and Fletcher, D.J. (2013). Updates in the Management of the Small Animal Patient with Neurologic Trauma. Veterinary Clinics: Small Animal Practice. 43: 915-940 Freeman, C. and Platt, S. (2012). Head Trauma. In Platt, S. and Garosi, L., eds. Small Animal Neurological Emergencies. London, UK: Manson Publishing/ The Veterinary Press, 363-382. Sturges, B.K. and LeCouteur, R.A. (2014). Intracranial Hypertension. In Silverstein, D. and Hopper, K., eds. Small Animal Critical Care Medicine. Saunders, 436-442. Carney, N. et al. (2016). Guidelines for the Management of Severe Traumatic Brain Injury, Fourth Edition. Neurosurgery. 0: 1-10 Friedenberg, S.G., Butler, A.L., Wei, L., Moore, S.A., and Cooper, E.S. (2012). Seizures Following Head Trauma in Dogs: 259 cases (1999-2009). Journal of the American Veterinary Medical Association. 241(11): 1479-1483. Otto, K. A. (2015). Physiology, Pathophysiology, and Anesthetic Management of Patients with Neurologic Disease. In Grimm, K.A., Lamont, L.A., and Tranquilli, W.J., eds. Veterinary Anesthesia and Analgesia. Somerset, US: Wiley-Blackwell, 559-583. Armitage-Chan, E.A., Wetmore, L.A., and Chan, D.L. (2007). Anesthetic Management of the Head Trauma Patient. Journal of Veterinary Emergency and Critical Care. 17(1): 5-14 Cassutto, B.H. and R.W. Gfeller. (2003). Use of Intravenous Lidocaine to Prevent Reperfusion Injury and Subsequent Multiple Organ Dysfunction Syndrome. Journal of Veterinary Emergency and Critical Care. 13(3): 137-148 Zeiler, FA, Teitelbaum, J., West, M., and L.M. Gillman. (2014). The Ketamine Effect on ICP in Traumatic Brain Injury. Neurocritical Care. 21:163-173.