History

A 12-year-old 450-kg (990-lb) crossbreed mare was referred to the Equine Teaching Hospital of the Universitat Autònoma de Barcelona for elective bilateral ovariectomy by means of left flank laparotomy. The mare was used as a jump mare for semen collection. On physical examination, there were no abnormal findings; heart rate was 48 beats/min, respiratory rate was 20 breaths/min, and rectal temperature was 37.6°C (98.4°F). Results of a preoperative CBC were within reference limits. An IV catheter was placed in the left jugular vein, and 1 hour before surgery, sodium penicillin (22,000 U/kg [10,000 U/lb], IV) and phenylbutazone (2.2 mg/kg [1 mg/lb], IV) were administered. The mare was restrained in standing stocks with the head cross-tied and prepared for surgery. Detomidine (10 μg/kg [4.5 μg/lb], IV) and butorphanol (20 μg/kg [9.1 μg/lb], IV) were administered for sedation. Lactated Ringer solution was infused at a rate of 5 mL/kg/h (2.3 mL/lb/h), IV, throughout the procedure. After 15 minutes, the extent of sedation was considered inadequate; therefore, additional IV doses of detomidine (5 μg/kg) and butorphanol (10 μg/kg) were administered. Thereafter, a continuous rate infusion (CRI) was initiated with an electronic syringe pump to provide detomidine at a rate of 10 μg/kg/h and butorphanol at a rate of 24 μg/kg/h (10.9 μg/lb/h). Heart rate, respiratory rate, and pulse quality were monitored throughout with auscultation, observation of thoracic excursions, and frequent palpation of a facial artery. Locoregional analgesia was provided by means of an inverted-L block with local infiltration of 2% mepivacaine solution (3.5 mg/kg [1.6 mg/lb]; total volume, 80 mL) in the left paralumbar fossa, and a left flank laparotomy with a modified grid approach was performed.¹ The left ovary was exteriorized, and the pedicle (mesovarium) was infiltrated with 10 mL of 2% lidocaine solution (0.4 mg/kg [0.18 mg/lb]) and then excised with a chain écraseur. Right ovariectomy was performed in a similar manner; however, because the right ovary could not be exteriorized via the left flank laparotomy approach, infiltration of the ovarian pedicle with 2% lidocaine was performed with an 18-gauge, 25-cm spinal needle. After bilateral ovariectomy was completed, the left ovarian stump was checked for bleeding by means of direct visualization; the right ovarian stump was checked by means of careful palpation followed by examination of the glove for the presence of blood. Eighty minutes after the CRI was started, the mare was responsive to surgical stimuli; therefore, additional doses of detomidine (5 μg/kg, IV) and butorphanol (10 μg/kg, IV) were administered. On completion of surgery, the laparotomy incision was sutured routinely in 3 layers. The total procedure time, including preparation of the surgical field, was 110 minutes. Although both heart rate and respiratory rate (40 beats/min and 18 breaths/min, respectively) decreased after initial administration of detomidine and butorphanol, these vital signs remained stable (28 to 32 beats/min and 6 to 9 breaths/min, respectively) thereafter for the 95-minute duration of the CRI. On completion of surgery, administration of the CRI and IV fluids was discontinued and the patient did not receive any other drugs. The mare appeared to recover and was walked back to the stall without apparent complications.

Fifteen minutes later, the mare was alert but began to have muscle tremors and was noted to have pale mucous membranes. Within a few minutes, clinical signs progressed to tachypnea, dyspnea with strong inspiratory efforts, and cyanosis. The mare became distressed, collapsed to lateral recumbency, and lost consciousness. Oxygen was administered at a flow rate of 15 L/min via a nasal tube while an emergency tracheotomy was performed. A peripheral pulse was not palpable; therefore, CPR was initiated following current guidelines² (60 chest compressions/min, 6 breaths/min, and epinephrine [5 μg/kg, IV]). Resuscitation maneuvers were sustained for 10 minutes, but were unsuccessful, and the mare died. A necropsy revealed no signs of abdominal hemorrhage or other complications related to the ovariectomy procedure. The histopathologic findings consisted of congestion and interstitial hemorrhage of laryngeal mucous membranes, pulmonary congestion, and alveolar edema.

Question

What was the likely cause of the sudden cardiopulmonary arrest and death that occurred immediately after apparently uncomplicated bilateral ovariectomy performed with standing sedation and locoregional analgesia in this mare?

Answer

In this patient, the sudden onset of clinical signs and the necropsy findings suggested either an adverse drug reaction or postural laryngeal edema resulting in acute severe upper airway obstruction as the most likely cause of death.

Discussion

Ovariectomy in mares is often performed with the patient standing, with a combination of sedation and locoregional analgesia. Many surgeons may prefer to perform surgery with the horse standing because of improved access to the surgical site and improved patient safety, compared with the use of general anesthesia.³ Current data on morbidity and mortality rates for horses undergoing surgery while standing are limited,⁴ with hemorrhage and shock the most commonly reported postoperative complications reported for mares undergoing ovariectomy.^5,6 In a retrospective study⁷ of 157 mares, the authors reported an overall postoperative morbidity rate of 10.8% (n = 17, including incisional complications, dehiscence, and colic) with no deaths. However, not all procedures were performed with the patient standing.

For the mare of the present report, the results of necropsy excluded hemorrhage related to the surgical procedure as the cause of death. Evidence suggestive of pulmonary embolism, a complication reported in human patients after gynecologic surgery,⁸ was also not found at necropsy; however, it could not be completely ruled out in this case. Adverse drug reactions (ADRs) in horses are considered rare.⁹ Depending on the underlying mechanism involved, ADRs may be classified as allergic (type I IgE–mediated hypersensitivity), nonallergic (idiosyncratic toxicosis that is not dose dependent), or intrinsic (ie, dose-dependent drug toxicosis resulting from overdose or inadvertent IV administration). Drugs most commonly associated with ADRs in horses include antimicrobials, steroidal anti-inflammatory drugs, NSAIDs, anesthetics, sedatives, and anthelmintics.¹⁰ Mild adverse reactions to anesthetics, sedatives, and analgesics have been reported, with signs including urticaria, excitement, or local reaction at the site of injection; fatal adverse reactions have been reported after administration of antimicrobials.^11,12 Anaphylaxis is a life-threatening, acute allergic reaction that is characterized by clinical signs that progress rapidly and can affect most organ systems. An anaphylactic reaction occurs within a few minutes after exposure to an allergen and can produce bronchospasm, asphyxia, dyspnea, and cyanosis (secondary to angioedema, laryngeal or pharyngeal edema, and bronchoconstriction); cardiovascular collapse; and, in severe cases, death.¹¹ Acute nonallergic drug hypersensitivities, also referred to as anaphylactoid reactions, manifest similar life-threatening clinical signs but are not immune mediated.¹³ Therefore, anaphylactoid reactions (vs anaphylaxis) can occur after a single exposure in sensitized patients; both conditions require the same treatment, with rapid intervention imperative. In patients affected by acute anaphylactic or anaphylactoid reactions, the lungs are primarily affected, rapidly leading to respiratory distress and death without treatment.¹⁴

Prior medical history was unknown for the mare of the present report. Although the clinical signs were suggestive of an acute anaphylactic reaction, an idiosyncratic adverse reaction to one of the drugs administered or a possible drug interaction could not be ruled out. Clinical signs of adverse reactions to drugs administered IV would be expected to be evident within seconds to a few minutes, but may be delayed for hours with other routes of administration.⁹ The time elapsed between IV drug administration and the development of clinical signs in this patient initially led us to rule out an adverse reaction to penicillin, phenylbutazone, or the initial doses of detomidine or butorphanol.

Drugs administered by a non-IV route that may have contributed, in view of the time course of events for the mare of this report, included the local anesthetics used as part of the locoregional analgesic regimen. The rate of absorption and time to peak plasma concentration of local anesthetics is affected by the vascular anatomy at the injection site. Sellers et al¹⁵ reported that the peak plasma concentration of lidocaine was detected 50 minutes after administration of an inverted-L nerve block in cows. Furthermore, in human patients, the plasma concentration of mepivacaine remained high for 120 minutes after administration of several common regional nerve blocks.¹⁶ Anaphylactic reactions to amide local anesthetics are extremely rare, but adverse reactions to these drugs are commonly reported in human patients, with most events occurring within the first hour after administration.¹⁷ Most of these reactions are reportedly not true anaphylactic reactions, but rather involve adverse reactions related to systemic absorption of the local anesthetic, effects of large doses, inadvertent intravascular injection, or pseudoallergic reactions.¹⁸

Muscle twitching and tremors are typically the first signs of an adverse reaction to a local anesthetic. Increasing doses may produce seizures, followed by unconsciousness, coma, and respiratory arrest in dogs.¹⁹ Horses are reportedly²⁰ more likely to develop CNS signs as a manifestation of lidocaine toxicosis, compared with other species. In horses, seizures may occur at lidocaine plasma concentrations of 6 μg/mL,²⁰ whereas in dogs, seizures occurred when lidocaine plasma concentrations reached 8.2 μg/mL.²¹ It has been reported that adult horses (body weight, 500 kg [1,100 lb]) can safely tolerate the administration of 250 mL of 2% lidocaine HCl solution for locoregional analgesia, which is equivalent to a dose of 10 mg/kg.²² The total doses of 2% lidocaine and 2% mepivacaine administered to the mare of the present report were 1.6 mg/kg (0.72 mg/lb) and 3.5 mg/kg, respectively, which are below the maximum recommended doses of both drugs (6 mg/kg [2.7 mg/lb]) in horses.²³ In horses, it has been reported²⁴ that signs of toxicosis such as muscle tremors can occur within a few minutes after IV administration of lidocaine at a dose of 1.5 mg/kg (0.68 mg/lb), with a serum lidocaine concentration of 3.5 μg/mL. In the mare of the present report, we cannot rule out delayed systemic absorption of local anesthetic eventually resulting in a toxic serum concentration. Moreover, the specific dose of local anesthetic that can result in toxicosis is also dependent on patient factors such as acid-base status, concomitant disease, and any concurrent condition or drug causing a decrease in cardiac output (eg, α₂-adrenergic receptor agonists), which might potentially decrease hepatic clearance.²⁵ The hepatic clearance of amide local anesthetics depends on the activity of hepatic enzymes (ie, cytochrome P450) and on hepatic blood flow. Concurrent conditions, such as liver disease, could alter the pharmacokinetics and distribution of local anesthetics,²⁶ increasing the likelihood of an adverse reaction. In the patient described in the present report, a preoperative serum biochemical analysis was not performed. No clinical signs of liver disease were observed. However, altered hepatic clearance can occur when multiple drugs metabolized by the liver are administered to a patient; this can particularly affect drugs with a high hepatic extraction ratio (eg, lidocaine).²⁷ When α₂-adrenergic receptor agonists and other highly protein-bound drugs (eg, phenylbutazone) are administered IV at the same time, competition for protein-binding sites may dramatically increase circulating drug concentrations. Therefore, we suggest that an adverse reaction as a result of high plasma concentrations of local anesthetic and other drugs administered concurrently to the mare of this report cannot be ruled out.

Detomidine and butorphanol are drugs that are commonly administered for sedation of horses undergoing surgery while standing. The sedative and analgesic effects of these drugs are achieved either by administration of a single initial bolus with subsequent doses given as needed during the procedure or by administration of a single bolus followed by a CRI.^28,29 In the mare of the present report, detomidine and butorphanol were administered by means of a CRI in an attempt to provide a constant level of sedation throughout surgery. Nonetheless, additional boluses were required to achieve an adequate level of sedation to complete the procedure. Cardiorespiratory depression is a common adverse effect of α₂-adrenergic receptor agonists in horses. Sedation with detomidine is reportedly accompanied by dose-dependent decreases in respiratory rate and tidal volume, with pulmonary redistribution of blood flow that results in impaired pulmonary gas exchange and hypercapnia.³⁰ The additional administration of butorphanol suppresses ventilation, exacerbating hypercapnia.³¹ The latter increases cerebral blood flow, increasing drug delivery to the brain. This potentially decreases plasma protein binding of local anesthetics and increases free drug availability.³² We suggest that such changes could have contributed to the suspected adverse effects of the local anesthetics administered to the patient of this report.

Although preoperative cardiac auscultation of the mare described in the present report did not reveal any abnormalities, an ECG was not performed; as such, an underlying arrhythmia could not be excluded in this mare. Furthermore, an ECG was not monitored during the surgery; heart rate was monitored via auscultation, and peripheral pulses were assessed. Others have recommended an ECG to monitor for cardiac arrhythmias secondary to cardiopulmonary depression in horses during surgical procedures performed with standing sedation.³³ Correct positioning of the head must also be monitored. Overextension of the neck together with a lowered head position and relaxation of the muscles in the larynx and nares because of the effect of α₂-adrenergic receptor agonists administered for sedation may contribute to laryngeal injury. To avoid this, in the present case, the patient's head was cross-tied at a height level with the withers, aimed at minimizing the possibility for compression of the upper airway. Although increased inspiratory efforts were not observed during the procedure, increased respiratory effort when the mare was walked to the stall could have exacerbated a partial upper airway obstruction. This would also be consistent with histopathologic findings affecting the laryngeal mucous membranes observed at necropsy. Supplemental oxygen administration would have been indicated if a decrease in arterial oxygen saturation had been noted, but pulse oximetry is difficult to perform with horses standing, because typical placement of the oximeter probe on the tongue is not an option. However, attempts can be made to place the probe on the lip or inside a nostril. Unfortunately, transcutaneous pulse oximetry, which is commonly used in human patients in intensive care units (forehead probe), is currently not available for conscious horses. Nonetheless, we emphasize that careful and thorough monitoring during and after surgical procedures performed with the patient standing could aid in earlier detection of complications, allowing for rapid intervention with appropriate targeted treatments that should improve outcome.