History

An 11-month-old 6.4-kg (14-lb) spayed female French Bulldog was referred to the Matthew J. Ryan Veterinary Hospital of the University of Pennsylvania for evaluation of a congenital cleft palate. At the time of birth, the dog appeared to have a midline defect of the hard and soft palates, with direct communication between the oral and nasal cavities. The puppy was adopted, and for the first 7 weeks of the puppy's life the new owner fed mostly liquid and canned food transorally via a red rubber catheter inserted into the pharynx; thereafter, the dog was transitioned to a diet of hard commercial puppy food. However, the owner handfed the dog just 1 kibble at a time, because it was noted that the dog would not chew the kibbles but swallow them whole. Occasionally, it was also noted that the dog would get a kibble stuck in the palate defect, followed by sneezing to dislodge it. Since birth, the dog had experienced several episodes of sudden-onset excitement that progressed to collapse and paddling of all 4 legs but that resolved spontaneously within seconds. The episodes of collapse were investigated by a veterinarian who suspected vasovagal syncopal episodes or portosystemic shunting with hepatoencephalopathy. When the dog was 21 weeks old, a CBC and serum biochemical analysis revealed mild hyperkalemia. Further testing included an ACTH stimulation test, abdominal ultrasonography, and measurement of serum bile acids, thyroid hormone, and cortisol concentrations to rule out hypoadrenocorticism, thyroid dysfunction, and liver disease, but no clinically important abnormalities were found. When the dog was 27 weeks old, further evaluation of liver function was performed by an internal medicine specialist to rule out a portosystemic shunt. However, results of repeated evaluation of liver function tests were within reference limits.

At 24 weeks old, the dog was anesthetized at a different veterinary facility for surgical repair of an avulsion of the left tibial crest. At that time, it was noted that endotracheal intubation was difficult because the dog had a very limited ability to open its mouth. Computed tomography of the skull was performed during the same anesthetic episode, confirming incomplete development and fusion of the soft palate and of the hard palate caudal to the incisive papilla. Additional maxillofacial abnormalities included malformation of the nasal turbinates, malocclusion because of brachycephalism, overcrowding of the mandibular and maxillary teeth, thickened walls of the tympanic bullae, and dysplastic temporomandibular joints with hypoplastic retroarticular processes and obliquely arranged mandibular condyles.

On physical examination at our hospital, the dog was bright and alert. Thoracic auscultation revealed mildly increased respiratory sounds. The patient was not able to open its mouth more than approximately 25 to 30 mm when awake. However, it was noted that a midline palatal defect extended from the incisive papilla caudally. Blood type and crossmatch analyses were performed. Preanesthetic cardiac evaluation including electrocardiography was recommended to investigate the syncopal episodes, but was declined by the owners. Therefore, the dog was premedicated with methadone (1 mg/kg [0.45 mg/lb], IM), and a 20-gauge, 3-cm IV catheter was placed into the right cephalic vein. Forty minutes after methadone premedication, the patient received midazolam (0.2 mg/kg [0.09 mg/lb], IV). General anesthesia was induced 20 minutes after midazolam premedication by IV administration of alfaxalone (diluted with saline [0.9% NaCl] solution to 5 mg/mL). A total dose of 1.1 mg/kg (0.5 mg/lb) was required to achieve a sufficient depth of anesthesia to enable endotracheal intubation. Provision for endoscopically guided endotracheal intubation was available in case of difficult intubation, and a tracheostomy kit was also prepared in case tracheostomy proved necessary. However, endotracheal intubation with a sterile 6.0-mm cuffed endotracheal tube was successful on the initial attempt, with a 3.5F, 56-cm-long polypropylene urethral catheter used as a guide and direct visualization via a laryngoscope. After endotracheal intubation, a 22-gauge, 3-cm arterial catheter was placed in the left metatarsal artery for direct blood pressure measurement. Direct blood pressure, heart rate (by means of a lead II ECG), arterial oxygen saturation (by means of pulse oximetry), respiratory rate, end-tidal carbon dioxide concentration (by means of sidestream capnography), and end-tidal isoflurane concentration were monitored throughout with a multiparameter monitor.^a The Doppler oscillometric^b technique was used as an adjunct method of monitoring heart rate; rectal temperature was measured intermittently with a standard rectal thermometer.

General anesthesia was maintained with administration of isoflurane in oxygen and spontaneous ventilation. Isoflurane concentration was adjusted as indicated depending on the anesthetic requirements of the patient. An IV bolus of fentanyl (5 μg/kg [2.27 μg/lb]) was administered at the start of surgery, followed by a constant rate infusion at 0.1 μg/kg/min (0.045 μg/lb/min) IV via a syringe pump. A dopamine constant rate infusion was administered to effect at doses ranging from 5 to 12 μg/kg/min (2.27 to 5.45 μg/lb/min) to maintain mean arterial blood pressure > 60 mm Hg, throughout. Crystalloid fluid therapy administered IV at a rate of 5 mL/kg/h commenced immediately after induction of anesthesia and continued throughout the procedure.

The patient was repositioned from lateral to dorsal recumbency after transfer to the operating room, 45 minutes after anesthetic induction (9:45 am). Following standard aseptic preparation and draping, a ventral approach with mandibular symphysiotomy was performed to gain access to the hard and soft palate defects (10:10 am). The mandibles were maintained in moderately retracted positions, each secured laterally to a pole with umbilical tape. Detachment of the intermandibular musculature and rostrocaudal incision of the sublingual mucosa bilaterally were then performed to further separate the mandibles (10:40 am). The tongue was retracted caudoventrally and temporarily maintained in place with a stay suture for the duration of the cleft palate repair. Maxillary and caudal inferior alveolar nerve blocks were administered bilaterally with 0.5% bupivacaine (total dose, 0.5 mg/kg [0.23 mg/lb]; 0.3 mL/site of injection), followed by nasal lavage with room temperature water (11:15 am). At 11:20 am, a decrease in heart rate from 120 beats/min (11:05 am) to 85 beats/min (11:20 am) was observed and treated with glycopyrrolate (10 μg/kg [4.54 μg/lb], IV). Mechanical ventilation was started at 11:30 am to maintain normocapnia (end-tidal carbon dioxide concentration, 35 to 45 mm Hg). The hard palate defect was closed with an overlapping flap harvested from one side of the defect (with the supplying major palatine artery and nerve attached) that was stretched and sutured beneath an envelope flap created on the opposite side of the defect (11:50 am). During the hard palate closure (12:10 pm), 3 hours 10 minutes after the start of general anesthesia (heart rate, 85 beats/min; mean arterial blood pressure, 70 mm Hg), the dog exhibited 2 episodes of cardiovascular arrest < 1 minute apart; the duration of each episode was approximately 5 seconds. The ECG read 0 beats/min, arterial pressure waves were absent, and the Doppler oscillometer was silent. The dog recovered spontaneously. At that time, the dog was receiving a 9 μg/kg/min (4.1 μg/lb/min) IV infusion of dopamine and a 0.2 μg/kg/min (0.091 μg/lb/min) IV infusion of fentanyl; the end-tidal isoflurane concentration was 1.0%, and the end-tidal carbon dioxide concentration was 51 mm Hg. The dog was being mechanically ventilated (pressure controlled), with a peak inspiratory pressure of 15 mm Hg and an inspiration-to-expiration (I:E) ratio of 1:2. The patient received another dose of glycopyrrolate (10 μg/kg, IV). Twenty minutes later (12:35 pm), the dog underwent a similar episode during which the heart rate decreased to 20 beats/min and mean arterial blood pressure was < 60 mm Hg. The dog responded to IV administration of atropine (20 μg/kg [9.09 μg/lb]), which was chosen for its rapid onset of action. The soft palate defect was then closed primarily via direct apposition of the edges of the defect in 2 layers (1:05 pm to 1:15 pm). The stay sutures and tape retracting and applying tension to the tongue and mandibles were then released, and an esophagostomy tube was placed on the left side of the neck (1:50 pm). This was followed by reconstruction of the lower jaw (2:20 pm to 3:30 pm). The patient was transported from the operating room to the endoscopy suite where appropriate positioning of the esophagostomy tube was confirmed. Finally, a temporary tracheostomy was performed (3:52 pm). Bradycardia accompanied by hypotension occurred on 3 occasions at approximately 20-minute intervals (at 12:50 pm, 1:15 pm, and 1:40 pm) during surgery and responded to treatment with glycopyrrolate, which was first administered IV, then IM (5 μg/kg) in an attempt to increase the duration of efficacy.

Question

What was the cause of this patient's episodes of intraoperative cardiovascular arrest? Was general anesthesia a contributory factor? Did the surgical procedure contribute?

Answer

In this patient, the sudden development of episodes of bradycardia and asystole was presumed to have occurred secondary to a trigeminovagal reflex that was an apparent consequence of manipulation of branches of the trigeminal nerve during retraction of the mandibles and ventrocaudal retraction of the tongue. The latter maneuvers were performed as part of surgical access required for cleft palate repair.

Discussion

Maxillofacial surgery in human patients represents a risk factor for development of the trigeminovagal reflex.¹ This reflex is elicited when a branch of the trigeminal nerve is stimulated, producing increased vagal activity and bradycardia, and can result in death. The trigeminal nerve is the fifth cranial nerve and has motor and sensory branches. It has been theorized that the trigeminovagal reflex occurs as a result of transmission of signals from sensory afferent nerve endings of the trigeminal nerve to the sensory nucleus, forming the afferent pathway of the reflex arc. After the primary afferent neurons project to the sensory nucleus, they form synapses with interneurons that connect with the efferent pathway in the motor nucleus of the vagus nerve. The efferent limb travels via the vagus nerve to the myocardium. Cardioinhibitory efferent fibers arising from the motor nucleus of the vagus nerve terminate in the myocardium. These vagal stimuli provoke negative chronotropic and inotropic responses. Consequently, the clinical features of the trigeminovagal reflex range from sudden onset of sinus bradycardia, bradycardia progressing to asystole, and asystole with no preceding bradycardia.² Several studies^3,4 in human patients suggest that locoregional anesthesia can decrease both the incidence and severity of the trigeminovagal reflex. In veterinary patients, the most commonly discussed clinical scenario for the trigeminovagal reflex is during ophthalmic surgeries.⁵ However, 1 case report⁶ described a severe vagal episode resulting in asystole in a dog with a zygomatic arch fracture. The bradycardia observed in the patient of the present report may have been caused by stimulation of the mandibular division of the trigeminal nerve as a result of the lateral retraction of the mandibles and ventrocaudal retraction of the tongue, which were necessary to allow access to the surgical site. In this patient, concurrent locoregional anesthesia did not prevent the occurrence of the trigeminovagal reflex. Although it is difficult to relate the vagal responses to specific surgical events, we observed that vagal reflexes occurred when surgical stimuli such as making incisions or suturing soft tissues were absent (eg, during didactic pauses and surgical planning). Once surgical stimulation resumed, it was observed that the sympathetic response was apparently sufficient to counter the parasympathetic drive caused by trigeminal stimulation via the lingual nerve caused by the stay suture on the tongue, the mandibular nerve via the lateral retraction of the mandibles, or both. In human patients, anticholinergics are the treatment of choice for bradycardia resulting from trigeminovagal reflexes. Our patient responded to glycopyrrolate and atropine injections that were administered periodically until the traction on the tongue and mandibles was discontinued.

It seems that, under general anesthesia, the balance between sympathetic and parasympathetic tone is altered in favor of the latter. Further, physiologic safety mechanisms that otherwise protect against vagal reflexes are suppressed. Therefore, patients will be most susceptible to the effects of parasympathetic signaling that follow stimulation of afferent vagal reflex loops. Anesthetic monitoring should include an ECG, Doppler oscillometry, and direct blood pressure measurement to accurately and rapidly diagnose severe bradycardia or asystole. When locoregional analgesia fails to prevent the occurrence of the trigeminovagal reflex, rapid communication between the anesthesiologist and surgical team aimed at discontinuing the stimulation of afferent branches of the reflex is imperative, accompanied by IV treatment with anticholinergic drugs. Moreover, caution should be exercised when excessive manipulation of the mandibles or tongue is necessary, and the duration should be minimized to reduce the risk of the trigeminovagal reflex as a consequence of stimulation of the caudal mandibular and lingual nerves. Further studies in veterinary patients are necessary to more specifically characterize risk factors for the trigeminovagal reflex and to develop optimal treatments.

Acknowledgments

The authors thank Dr. Giacomo Gianotti for assistance with editing this manuscript.