History

A 9-month-old 6.2-kg (13.6-lb) sexually intact male Cavalier King Charles Spaniel with pulmonary valve stenosis was referred for cardiac consultation regarding treatment with balloon valvuloplasty. The owners reported that the dog had no clinical signs of heart disease but a history of a grade 5/6 heart murmur since birth. Echocardiography performed by the referring veterinary cardiologist when the dog was 2 months old revealed severe pulmonary valve stenosis, and atenolol (1 mg/kg [0.45 mg/lb], PO, q 12 h) was prescribed at that time. Transthoracic echocardiography was repeated at our facility by a cardiology resident (AVC) under the supervision of a board-certified cardiologist (MSK) for surgical planning purposes prior to balloon valvuloplasty. Findings included severe pulmonary valve stenosis, mild pulmonary valve insufficiency, dilation of the pulmonary artery distal to the pulmonary valve, trace tricuspid valve regurgitation, moderate right ventricular concentric hypertrophy with flattening of the interventricular septum (suggestive of increased right ventricular pressure), and right atrial dilation.

The day before balloon valvuloplasty, physical examination of the dog revealed a regularly irregular heart rhythm that varied with the phase of respiration. The dog had a heart rate of 110 beats/min (reference range, 80 to 160 beats/min¹) and a grade 5/6 left basilar systolic heart murmur with a palpable thrill. The lungs sounded clear on auscultation, and femoral pulses were strong and synchronous bilaterally. Other than an approximately 1-cm-diameter umbilical hernia, the remainder of the physical examination was unremarkable. Results of a CBC and serum biochemical analyses indicated mild leukocytosis (15,120 WBCs/μL; reference range, 5,700 to 14,200 WBCs/μL) with neutrophilia (11,190 neutrophils/μL; reference range, 2,700 to 9,400 neutrophils/μL) and eosinopenia (0 eosinophils/μL; reference range, 100 to 2,100 eosinophils/μL), consistent with a stress leukogram. The dog's PCV and serum total protein concentration were within the reference limits. Findings on a 6-lead ECG indicated a respiratory sinus arrhythmia, a mean heart rate of 80 beats/min, and right ventricular hypertrophy but no evidence of atrial enlargement.

Food, but not water, was withheld from the dog for 8 hours overnight before the anesthetic event; however, atenolol (1 mg/kg, PO) was continued as prescribed and administered the morning of surgery. A 20-gauge, 1.16-inch, over-the-needle polyurethane catheter^a was placed in the left cephalic vein, and a 22-gauge, 1-inch, over-the-needle polyurethane catheter^a was placed in the left saphenous vein. Administration of lactated Ringer solution (3 mL/kg/h [1.36 mL/lb/h], IV) was initiated. Hair over the right and left precordial areas was clipped, and the dog was instrumented with pediatric electrode pacing pads connected to a biphasic defibrillator.^b After preoxygenation by face mask for 5 minutes with 100% oxygen, the dog received fentanyl (5.0 μg/kg [2.27 μg/lb], IV), and general anesthesia was induced with propofol (2.5 mg/kg [1.13 mg/lb], IV) and midazolam hydrochloride (0.3 mg/kg [0.136 mg/lb], IV). The dog was orotracheally intubated with a 7-mm-internal-diameter cuffed endotracheal tube that was then connected to a universal F rebreathing circuit. Anesthesia was maintained with 0.6% to 0.85% isoflurane (vaporizer settings) in O₂ and a constant rate infusion (CRI) of remifentanil (0.1 to 0.4 μg/kg/min [0.04 to 0.18 μg/lb/min], IV) and lidocaine (50 μg/kg/min [22.7 μg/lb/min], IV). Cefazolin (22 mg/kg [10 mg/lb], IV, before surgery and q 90 min during surgery) was also administered.

The dog breathed spontaneously and was positioned in left lateral recumbency. A 24-gauge, 0.75-inch, over-the-needle polyurethane catheter^a was placed in the left dorsal pedal artery and, with non-compliant, saline-filled tubing, was connected to a transducer calibrated at the level of the right atrium for direct blood pressure measurements. Heart rate and rhythm (measured with a lead II ECG), peripheral capillary O₂ saturation of hemoglobin (Spo₂; measured with pulse oximetry), end-tidal partial pressure of carbon dioxide (Petco₂) and respiratory rate (measured with sidestream capnography), and end-tidal isoflurane concentration (measured with infrared spectroscopy) were monitored continuously with a multiparameter monitor^c throughout the anesthetic procedure and recorded every 5 minutes. The dog's rectal temperature was monitored and recorded every 15 minutes.

Transesophageal echocardiography was performed, and findings were similar to those of trans-thoracic echocardiography performed during the cardiac consultation. During transesophageal echocardiography, the dog's mean arterial pressure (MAP) decreased from 72 to 58 mm Hg (reference range, 80 to 120 mm Hg¹). Therefore, a bolus of lactated Ringer solution (5 mL/kg [2.27 mL/lb], IV) was administered over 15 minutes. The MAP only increased to 65 mm Hg; thus, a second bolus (5 mL/kg) was administered, and a CRI of dopamine (3 to 5 μg/kg/min [1.36 to 2.27 μg/lb/min], IV) was started. Afterward, the MAP increased to > 70 mm Hg.

A right jugular venous access was used for balloon valvuloplasty. With onset of ballooning (balloon inflation over approx 5 seconds and then deflation over approx 5 seconds, or as quickly as possible), the dog's MAP decreased to 50 mm Hg. Therefore, the CRI of dopamine was increased to 12 μg/kg/min (5.45 μg/lb/min), IV. The first 2 balloon inflations were unsuccessful because the balloon moved distal to the pulmonary valve and into the pulmonary artery. However, even in the absence of balloon inflation, the dog's hypotension persisted; therefore, norepinephrine (0.05 to 0.1 μg/kg/min [0.023 to 0.045 μg/lb/min], IV) was administered. Immediately following the third inflation, which was deemed successful, Spo₂ suddenly decreased from 98% to 77%. Analysis of arterial blood gas revealed a mixed acid-base disturbance and hypoxemia (Table 1).

Question

What were the most likely causes of rapid O₂ desaturation of hemoglobin and hypoxemia that occurred during balloon valvuloplasty in the dog of the present report?

Answer

There were 5 potential causes of hypoxemia: low fraction of inspired O₂ (Fio₂), hypoventilation, diffusion impairment, ventilation-to-perfusion ratio (V/Q) mismatch, and right-to-left cardiac shunt.

The dog's Fio₂ was 0.97 when the sudden drop in Spo₂ occurred; therefore, a low Fio₂ as the underlying cause was ruled out. Mechanical ventilation was initiated, and an alveolar recruitment maneuver that consisted of insufflating the lungs to an airway pressure of 20 cm H₂O for 20 seconds to re-expand the entire lung and increase Spo₂ was attempted. However, the maneuver was unsuccessful, and Spo₂ further dropped to 60%. A second arterial blood gas analysis revealed a continued mixed acid-base disturbance, with improved blood pH but worsened oxygenation (Table 1).

Balloon valvuloplasty was completed, and remifentanil, lidocaine, and isoflurane were discontinued. The dog had stable blood pressure and was weaned from all the pressure-supportive drugs. Five minutes after anesthetic agents were discontinued, the dog began swallowing and chewing and was extubated; however, the dog's movements at that point precluded measurement of Spo₂. The dog received flumazenil (0.01 mg/kg [0.005 mg/lb], IV) and was transferred to the intensive care unit for continued recovery and monitoring.

In the intensive care unit, the dog had a heart rate of 130 beats/min and a respiratory rate of 30 breaths/min, with mildly increased respiratory effort but no crackles or wheezes detected. The dog's Spo₂ ranged between 91% and 95% (measured with a portable pulse oximeter^d) while it received supplemental O₂ (4 L/min) by face mask. After approximately 15 minutes, supplemental O₂ was discontinued, and the dog was allowed to breathe room air. After the dog had been breathing room air for approximately 15 minutes, arterial blood gas analysis was repeated and revealed improved but not fully satisfactory oxygenation (Table 1). The dog was placed in an O₂ cage set at 40% O₂ (Fio₂, 0.40) overnight, which resulted in neither substantial nor consistent improvement in the dog's Spo₂ (range, 70% to 88%).

The following day, the dog had 2 neurologic episodes that consisted of increased muscle tone, opisthotonos, and signs of dull mentation during excitement or stress. Neurologic examination revealed no obvious deficits, and the episodes were attributed to transient cerebral hypoxia. Thoracic radiography was performed, and findings included dilation of the main pulmonary artery and enlargement of the right side of the heart as expected. Further, the pulmonary parenchyma appeared radiographically normal, with no evidence of an abnormality that may have contributed to poor oxygenation of the dog. Results of postoperative echocardiography indicated a successful outcome for balloon valvuloplasty. In addition, echocardiography was then performed while 3 mL of saline (0.9% NaCl) solution agitated with room air to contain microbubbles was injected through a cephalic vein catheter to identify the possible presence of a right-to-left intracardiac shunt. During this agitated-saline echocardiography, microbubbles were visualized passing from the right atrium into the left atrium and then the left ventricle, diagnostic for a right-to-left atrial-level shunt, such as a patent foramen ovale or atrial septal defect.²

Given the findings on agitated-saline echocardiography combined with the ineffectiveness of positive pressure ventilation and supplemental O₂ for the dog of the present report, the primary differential diagnosis for O₂ desaturation of hemoglobin and hypoxemia that occurred during balloon valvuloplasty was a right-to-left atrial-level shunt. In this dog, we suspected that the balloon catheter and guidewire may have been pulled rostrally during the successful balloon inflation and that this movement caused a tear in the tricuspid valve. The increased volume of tricuspid regurgitation caused a sudden increase in right atrial pressure and shunting of blood from right to left across the previously unidentified atrial-level shunt, explaining the rapid decline in systemic O₂ saturation.

Aside from the 2 neurologic episodes during hospitalization the day after valvuloplasty, the dog showed no abnormal clinical signs and was discharged 3 days later. Atenolol was discontinued, and the owner was instructed to restrict the dog's exercise because exercise could exacerbate hypoxemia by decreasing systemic peripheral vascular resistance and increasing right-to-left shunting. The owner was also advised to monitor for collapse, neurologic episodes, or signs of exercise intolerance.

A recheck examination was performed 4 months after balloon valvuloplasty. The dog's Spo₂ while breathing room air was 95%. Echocardiography revealed mild pulmonary valve stenosis, subjectively improved right ventricular hypertrophy, and similar to slightly improved tricuspid valve regurgitation. It was hypothesized that right atrial remodeling (eccentric hypertrophy) after balloon valvuloplasty caused a reduction in right atrial pressure and therefore less right-to-left shunting across the atrial-level shunt.

Discussion

With exclusion of other causes, we suspected a right-to-left cardiac shunt as the main underlying cause of hypoxemia in the dog of the present report. Nonetheless, after the rapid decrease in Spo₂, the finding of a high arterial partial pressure of CO₂ (Paco₂) indicated that some degree of hypoventilation was present. Hypoventilation increases the alveolar partial pressure of CO₂ (Paco₂) and Paco₂ and, because of a lower alveolar partial pressure of O₂ (Pao₂), also decreases the arterial pressure of O₂ (Pao₂).^3,4 Increasing the Fio₂ is effective in treating hypoxemia secondary to hypoventilation.³ However, the Fio₂ was 0.97 in the dog of the present report. The initiation of mechanical ventilation was considered effective in treating the dog's hypoventilation as indicated by the decrease in Paco₂ on the second arterial blood gas analysis. However, hypoxemia did not improve, suggesting that another mechanism was responsible for the low Pao₂.

Diffusion limitations develop when an impairment of the blood-gas barrier is present and O₂ transport is subsequently reduced. Causes include lung diseases, such as fibrosis, inflammation, or thickening of the alveolar-capillary interface. At rest, the pulmonary capillary transit time for RBCs is usually long enough to complete gas exchange, rarely resulting in hypoxemia from diffusion impairment. During exercise, conversely, the capillary transit time is shortened because of a rise in cardiac output, and hypoxemia can develop.^3–5 This problem rarely occurs during anesthesia in animals without concomitant respiratory diseases.³ One of the most common causes of hypoxemia during anesthesia is V/Q inequality or mismatch.⁶ Physiologic heterogeneity between ventilation and perfusion occurs in the different areas of the lungs owing to gravity and differences in intrapleural pressures. During anesthesia, additional mechanisms have the potential to exacerbate this imbalance and decrease the lung's ability to transfer O₂ and remove CO₂. High V/Q develops when ventilation exceeds perfusion, and causes include low cardiac output, hypovolemia, or pulmonary thromboembolism. In these conditions, the total O₂ uptake from the lung is decreased, and hypoxemia can develop.⁷ Poor perfusion to the lung increases the V/Q by increasing the physiologic dead space. Although anatomic dead space does not usually increase when appropriate endotracheal tube lengths and anesthesia circuits are used, an increase in physiologic dead space (the sum of the anatomic and alveolar dead space) is mainly from an increase of the alveolar component.⁸ Because a reduction in pulmonary blood flow results in reduced removal of CO₂ by the lungs and in turn a low Petco₂ while the Paco₂ remains constant, an increase in the Paco₂-Petco₂ gradient is used as an index of alveolar dead space.⁹ The opposite situation, a low V/Q, is defined as an excess of perfusion in proportion to ventilation.^3–5 The decrease in ventilation has the potential to cause hypoxemia by decreasing the alveolar O₂ concentration and subsequently the Sao₂. In humans and dogs, the most common cause of a low V/Q during anesthesia is atelectasis.^10–12 A common feature of low and high V/Q imbalances is an improvement of hypoxemia in response to treatment with O₂.⁴ In most cases, treating the underlying cause is the most effective intervention in resolving hypoxemia.

Cardiac output was not measured in the dog of the present report. However, results of direct blood pressure measurements indicated that the dog was normotensive during anesthesia when treated with dopamine and norepinephrine. Therefore, cardiac output was not considered to contribute to the dog's hypoxemia.

After the desaturation episode, the dog in the present report had a Paco₂-Petco₂ gradient between 8 and 12 mm Hg, suggestive that a high V/Q mismatch was not the most likely cause for the dog's hypoxemia, whereas a Paco₂-Petco₂ gradient > 15 mm Hg would have been consistent with a high V/Q mismatch being a likely cause for hypoxemia.¹³ An alveolar recruitment maneuver can be effective in treating a low V/Q mismatch caused by atelectasis.¹⁴ However, the alveolar recruitment maneuver did not improve oxygenation in the dog of the present report, a finding that suggested that V/Q mismatch was likely not the cause of hypoxemia.

A right-to-left shunt is a condition in which blood from the right side of the heart enters the left side without passing through the lungs for gas exchange.^3,4 Shunting represents the extreme degree of a low V/Q mismatch caused by pathological conditions, such as arteriovenous communications or intrapulmonary shunting. Because of the inability to improve the Pao₂ in unventilated alveoli, a unique feature of patients with low V/Q mismatch caused by a right-to-left shunt is their poor response to treatment with O₂. For this reason, a failure to improve Pao₂ when 100% O₂ is inspired can be considered a useful diagnostic result in the discrimination of causes of hypoxemia.³ Because of the stimulation of the respiratory center by chemoreceptors, hypercapnia is uncommon in a shunt scenario until the shunt fraction reaches 50%.^1,15 Additionally, the Pao₂:Fio₂ ratio can be considered an estimate of shunt fraction in that a Pao₂:Fio₂ ratio < 200 mm Hg indicates that the shunt fraction is > 20%, whereas a Pao₂:Fio₂ ratio > 200 indicates that the shunt fraction is < 20%.⁵ The dog of the present report had a Pao₂:Fio₂ ratio that ranged between 55 and 74 mm Hg, which suggested that the dog had a shunt fraction > 20%. Further, the absence of any response by the dog in the present report to the alveolar recruitment maneuver and treatment with supplemental O₂, combined with findings on thoracic radiography, supported a suspected right-to-left shunt. Therefore, agitated-saline echocardiography was performed, and findings confirmed the presence of a right-to-left atrial-level shunt.

A patent foramen ovale has been described as a cause of right-to-left shunting in human patients with pulmonic stenosis.¹⁶ The prevalence of patent foramen ovale in dogs with pulmonic stenosis has been reported to be as high as 39% (12/31).¹⁷ Shortly after birth, circulation to the lungs increases, causing left atrial pressure to rise above right atrial pressure. In dogs with pulmonic stenosis, however, abnormally high right atrial pressure may lead to failure of the valve of the foramen ovale to seal against the septum secundum, resulting in persistent patency of the foramen ovale.

Findings in the dog of the present report provided a reminder that patients could have multiple congenital cardiac defects. We recommend that before patients with congenital cardiac defects undergo general anesthesia, they first undergo full cardiac evaluation, including agitated-saline echocardiography to evaluate for a preexisting atrial-level shunt, and have their Spo₂ and PCV evaluated so that potential shunt-related complications (eg, hypoxemia) may be more readily identified and addressed.