Abstract

We describe a case of the sudden onset of cardiovascular collapse during emergence from anesthesia resulting from a massive venous air embolism, which was detected by transesophageal echocardiography. We present this case to remind anesthesiologists to be aware of the risk of a sudden return of air trapped in the venous system during emergence from anesthesia. The air is freed because the sympathetic tone is increased, muscle-pumping power is regained, ventilation shifts from positive-pressure to negative-pressure spontaneous ventilation, and the patient is repositioned after surgery.

Keywords

anesthesia recovery period: emergenceembolism, air;

1. Introduction

Massive venous air embolism (VAE) is a major intraoperative catastrophe. Most intraoperative VAEs occur when the surgeons are performing surgical manipulation.¹⁻³ We present a case of sudden-onset cardiovascular collapse as a result of massive VAE while the patient was emerging from anesthesia. Transesophageal echocardiography helped confirm the diagnosis. The possible causes and potential risks of massive air embolism during emergence from anesthesia are discussed.

2. Case Report

A 56-year-old man, with a height of 173 cm and weight of 70 kg, suffering from a right tibia-fibula open fracture (Gustilo type II) from a traffic accident, was scheduled to undergo an open reduction, with internal intramedullary nail fixation, and wound debridement. The patient was still conscious after the accident. Upon arrival in the operating room, the patient’s blood pressure was 140/76 mmHg, heart rate was 100 beats/min, and respiratory rate was 22 breaths/min. The breathing sounds were clear bilaterally. Past medical history revealed that the patient was a victim of fatty liver and had abused alcohol for more than 20 years. Laboratory data were within normal limits, except for an elevated blood glucose level of 242.3 mg/dL and elevated liver enzymes (glutamic-pyruvic transaminase, 98.2 IU/L; glutamic-oxaloacetic transaminase, 125 IU/L). Electrocardiography showed a normal tracing, and chest X-ray disclosed borderline cardiomegaly with no active lung lesion. American Society of Anesthesiology physical status was evaluated as class IIIe.

The patient was given 150 μg fentanyl, 250 mg thiopentone, 80 mg succinylcholine, and 25 mg atracurium in sequence to induce anesthesia and to facilitate tracheal intubation. Anesthesia was maintained with 1.5% isoflurane in oxygen and nitrous oxide (O2 flow, 1.0 L/min; N2O flow, 1.0 L/ min). The anesthetic ventilator was set at volumecontrol mode with a tidal volume of 600 mL, a respiratory rate of 10 breaths/min, and an inspiratory/ expiratory time ratio of 1:2. The patient’s airway pressure was normal (18−22 mmHg).

With the patient in the supine position, a tourniquet was attached to the right thigh, which was then inflated to a pressure of 350 mmHg. Through an infrapatellar incision, an unreamed tibial nail was introduced. Finally, the wound was irrigated with normal saline, debrided, and sutured. Tourniquet pressure was released after 104 minutes in use. During surgery, the patient’s vital signs were stable. Ten minutes after the tourniquet was released, the patient’s blood pressure suddenly dropped to 80/50 mmHg, but neither his oxygen saturation nor his end-tidal CO₂ were affected. His blood pressure returned to normal 5 minutes later. However, his leg was pale, pedal pulses were not palpable, and a weak posterior tibial Doppler signal was observed. For fear of an anterior tibial artery and saphenous vein injury, the operation was suspended and the patient was immediately sent for emergency angiography, which revealed a torn anterior tibial artery. A cardiovascular surgeon was consulted for repair of the injured vessel.

The anterior tibial artery was bypassed to the peroneal trunk with a 6 mm × 5 cm Gore-Tex graft, and the saphenous vein was ligated. The entire operation lasted 5 hours. The patient’s blood loss was approximately 250 mL and his urine output was 580 mL. He was given 2450 mL of crystalloid during the operation.

Intravenous neostigmine (2 mg) and atropine (1 mg) were given to reverse the neuromuscular blockade. The patient was then transferred from the surgical table to the transporting gurney. The patient had occasional premature atrial contractions and premature ventricular beats, and then his oxygen saturation dropped steeply. His electrocardiograph (ECG) changed from sinus rhythm to broad QRS and complex tachycardia, and then to pulseless ventricular tachycardia. Resuscitation was initiated with epinephrine (1 mg i.v.), and cardiac massage.

The patient’s sudden cardiovascular collapse post-surgery, as contemplated, might have been caused either by hypovolemia, acute myocardial infarction, tension pneumothorax, cardiac tamponade, or embolism. Skin turgor was normal, and the patient had neither skin wheals nor conjunctival and mucosal edema. Before this episode, ECG monitoring showed a normal sinus rhythm without any sign of ischemic changes. The patient’s intraoperative systolic blood pressure was stable at 100−120 mmHg. There seemed to be no hypovolemia, anaphylactic shock, or myocardial infarction. In addition, through a stethoscope, the breathing sounds of both lungs were clearly audible; therefore, tension pneumothorax was ruled out.

Substantial hydration was started and a transesophageal echocardiography probe was inserted to search for the underlying cause of the patient’s shock. Massive echogenic air bubbles were found in the right atrium, the right ventricle, and the pulmonary arteries, and the volume of the left heart had severely shrunk (Figure 1). We carefully checked the intravenous lines but no disconnection was found. We asked the surgeons to recheck and repack the surgical dressing. The rush of air bubbles lasted more than 20 minutes and then gradually disappeared.

A central venous catheter was inserted through the patient’s right internal jugular vein, but no air bubbles could be aspirated out from the catheter. Ventricular cardiac rhythm resumed after a 2-hour cardiac massage. A monophasic damped sinusoidal waveform defibrillation was tried three times at 100, 200, and 300 joules escalation. Normal sinus rhythm returned and blood pressure was detectable again. Arterial blood gas (ABG) analysis showed a pH of 6.84, PCO₂ of 95 mmHg, PaO₂ of 35 mmHg, base excess of −14.8 mmol/L, and hematocrit (Hct) of 25%. Sodium bicarbonate was given to correct the metabolic acidosis, and two units of packed red blood cells were transfused. Before sending the patient to the intensive care unit (ICU), repeat ABG data revealed a pH of 7.3, PCO₂ of 35 mmHg, PaO₂ of 384 mmHg, Na⁺ of 149.8 mmol/L, K⁺ of 4.02 mmol/L, Ca²⁺ of 1.23 nmol/L, and Hct of 28%. Inotropic support with epinephrine (16 μg/kg/min) and dopamine (5 μg/kg/min) was also initiated. Upon transfer to the ICU, blood pressure was 131/57 mmHg and heart rate was 73 beats/min. A repeated transthoracic echocardiography was performed and showed concentric left ventricular hypertrophy, no chamber dilatation, good left ventricular performance, and only mild mitral regurgitation. The chest X-ray taken in the supine position showed diffuse infiltration over bilateral lung fields, suggestive of pulmonary edema. In the ICU, the patient’s vital signs were maintained with intravenous inotropic agents and 100% oxygen. Unfortunately, a stress ulcer with coffee-ground drainage from a nasogastric tube was found on the second postoperative day, and the patient was given a blood transfusion. The general condition of the patient did not improve; rather, it became worse with disseminated intravascular coagulation setting in. Because of the development of acute renal failure with hyperkalemia (K⁺ of 6.61 mmol/L), he underwent hemodialysis. A sudden episode of ventricular fibrillation which quickly advanced to asystole occurred on the fourth day of the ICU stay, and cardiopulmonary resuscitation was attempted again. However, the patient soon died of multiple organ failure.

Download full-size image

Figure 1 Transesophageal echocardiography during resuscitation of the patient. (A) There were massive air bubbles in the right atrium and ventricle, and the interatrial membrane deviated to the left side. (B) Another view shows massive air bubbles in the right atrium, the right ventricle, and the pulmonary artery.

3. Discussion

We present a case of a rare fatal venous air embolism, detected by transesophageal echocardiography, during emergence from anesthesia. The air bubbles might have been entrapped in the venous system because of open wounds with injured veins immediately after the accident, during surgery, or during wound irrigation. Emergence from anesthesia might carry risks that contributed to a sudden venous return with the trapped air.

When the patient arrived in the operating room, his respiratory rate was 22 breaths/min (tachypnea), which was interpreted as a consequence of nervousness or pain and was, therefore, neglected. In addition, the injury to the saphenous vein had not been diagnosed before the wound was irrigated with normal saline during debridement. Irrigation has been considered a risk for VAE in some case reports.⁴ However, the accumulated air bubbles remained static in the small venules of the right lower leg before the tourniquet was released. After the interlocking nail had been fixed, releasing the tourniquet on the right thigh would have encouraged the air bubbles to move into larger blood vessels, which is why the patient’s vital signs showed a transient change at that point.

Why did the patient have serious complications when he emerged from anesthesia rather than during the arterial bypass and saphenous vein ligation surgery, when the tourniquet was released? We speculated that certain factors increased the number of air bubbles spreading during emergence from anesthesia. First, the patient’s effective blood circulation was blocked by a large number of air bubbles: the “air-lock” effect. Sufficient venous movement requires higher blood pressure. During surgery, the systolic blood pressure was controlled at about 100−120 mmHg, which rendered the air bubbles in the veins static. When the patient emerged from anesthesia, the termination of inhalation anesthetic raised the sympathetic tone, hence elevating the blood pressure, which forced the static air to move forward. Second, during surgical anesthesia, muscle contractions were blocked by a muscle relaxant, but while emerging from anesthesia, the patient’s returning muscle power caused the muscles to contract, which pumped the entrapped air bubbles along the route of venous return and, finally, to the pulmonary circulation. Third, after the patient had been weaned from the anesthetic ventilator, negative-pressure spontaneous breathing was regained, which decreased the intrathoracic and right atrial pressures and further increased the amount of air entering the pulmonary system.⁵ Finally, moving the patient from the surgical table to the trolley also facilitated the spread of air bubbles from their entrapped sites. Too much air returning to the pulmonary circulation suddenly caused a severe obstruction of right ventricular outflow with progression to cardiovascular collapse.

Transesophageal echocardiography helped us find the air in the right side of the heart immediately; therefore, we were able to diagnose the patient’s major problem and decide upon the next step. Transesophageal echocardiography and Doppler are still the most sensitive tools for diagnosing VAE, although other methods are used: end-tidal carbon dioxide concentration in the expired air, pulmonary artery pressure, a central venous catheter, a right atrial catheter, and an esophageal stethoscope.^6,7 Although physiologically insignificant episodes of air entry, crystallized mannitol solutions, or rapid injections of aqueous solutions may mimic intracardiac air in echocardiography, none of these were applicable in this patient, because he had no central venous catheter and had not been given mannitol during the transesophageal echocardiography examination.

The factors that determine the morbidity of an episode of VAE include the rate of air entrainment, the volume of air entrained,⁸ and the position of the patient at the time of the embolism.⁹ The minimum volume of air lethal to human beings has not been established, but estimates of 200−300 mL of air have been reported to be lethal.¹⁰

Strategies for preventing VAE focus on decreasing the pressure gradient between the surgical site and the right atrium. This includes hydrating the patient to keep an adequate level of fluid. The surgeon should be alert and try to cauterize and tie the vessels around the surgical wound. Maintaining high levels of positive end-expiratory pressure is a controversial procedure for decreasing the pressure gradient. VAE may occur even after the postoperative release of positive end-expiratory pressure.¹¹

There are still no efficacious treatments for VAE. To inform the surgeon about the highly suspected VAE and to stop the air from entering the pulmonary system were the steps we had to follow. Flooding the field with saline should submerge the area of air entry. If nitrous oxide is used, it should be discontinued because it may expand the patient’s volume of aspired air. In this case, we used 50% oxygen and 50% nitrous oxide to deliver the inhalation agent. Nitrous oxide may augment the volume if there are small air bubbles in the bloodstream.¹²

The left lateral decubitus position has been advised to release right ventricular outflow obstruction, and aspiration of air using a central venous catheter should be attempted.¹³ If hemodynamic instability progresses, inotropic agents such as epinephrine are recommended. One hundred percent oxygen should be given to provide adequate oxygenation.¹⁴

A large amount of air may obstruct the microvascular circulation in the lungs and lead to the release of endogenous vasoactive substances, such as tumor necrosis factor-α and interleukin 1β, and proinflammatory cytokines which cause a significant systemic inflammatory response^15,16 and induce acute lung injury,¹⁷ both of which can compromise the function of organ systems and lead to multiple organ dysfunction syndrome (MODS).¹⁸ This might be why the patient’s condition gradually worsened despite successful resuscitation in the operating room. The patient soon developed pulmonary edema and acute renal failure, and died from MODS on postoperative day 4.

In summary, anesthesiologists should be alert to the possibility that severe venous air embolism can occur not only during surgical manipulation, but also during emergence from anesthesia, because the physiological changes during this period carry some risk of freeing trapped air from injured peripheral vessels into the pulmonary circulation and inducing a lethal air embolism.

Acknowledgments

We thank Bill Franke for editorial assistance.