Lethal Air Embolism during Cesarean Delivery for Placenta Previa

A 29-yr-old woman, para 5, gravida 1, spontaneous abortion 3, was scheduled for elective cesarean delivery for placenta previa with the fetus in a transverse lie. She was otherwise healthy and did not smoke. At 37 weeks’ gestation, 3 days before her scheduled surgical date, she presented to the Maternity Care ward with ruptured membranes and painless vaginal bleeding. On arrival, her blood pressure was 117/72 mmHg, and her heart rate was 80 beats/min. The fetal heart rate was 130 beats/min with decreased variability. A large-bore intravenous cannula was started, volume preload was initiated, a bladder catheter was inserted, and blood work was performed. Consent for surgery and anesthesia were obtained, and the patient was transferred to the operating room for urgent cesarean delivery.

The patient and fetus were stable on arrival in the operating room. Oxygen was administered via  nasal prongs at 3 l/min. A pulse oximeter, noninvasive blood pressure device, and electrocardiograph were placed, and a spinal anesthetic was performed with the patient in the left lateral decubitus position. A 27-gauge pencil-point spinal needle was placed atraumatically on the first pass, and 10.5 mg hyperbaric bupivacaine, 7.5%, and 300 μg preservative-free morphine were injected without incident. By this time, the patient had received 1.2 l lactated Ringer's solution and 1 g cefazolin. The patient was positioned supine on the operating table with a 10° tilt to the left. A vertical abdominal incision was made, followed by a transverse lower uterine segment incision. The obstetrician noted multiple varicosities across the lower segment, especially laterally. The placenta was dissected through and a breech extraction performed to deliver a healthy male infant. The umbilical cord was double clamped and divided, and the infant was passed to the neonatal team. Oxytocin, 10 U, was administered intravenously, and the uterus was then exteriorized to facilitate repair. Blood loss was approximately 500 ml, and 1,800 ml lactated Ringer's solution had been infused up to this point. The blood pressure was 90/50 mmHg, and the heart rate was 70 beats/min. Ephedrine, 10 mg, was administered intravenously twice.

Approximately 1 min after delivery, the patient became agitated and lifted her head and shoulders off the operating table. She then fell back to the table, underwent a generalized seizure for approximately 5 s, and lost consciousness. The electrocardiogram showed a sinus tachycardia at 130 beats/min, but readings were not measurable on the pulse oximeter and the noninvasive blood pressure device. At that moment, the patient did not have a carotid pulse. A cardiac arrest code was immediately called, the patient's trachea was intubated, 100% oxygen was administered, and cardiopulmonary resuscitation was initiated. The initial end-tidal carbon dioxide on the capnogram was less than 10 mmHg after intubation. The differential diagnosis included amniotic fluid embolism, air embolism, and anaphylaxis. The surgeon closed the uterus as quickly as possible. Advanced cardiac life support protocols were used during resuscitation. With the exception of a short run of ventricular tachycardia and another of ventricular fibrillation, which responded to electrical defibrillation, the patient remained in pulseless electrical activity throughout the resuscitation efforts. Three physicians unsuccessfully attempted to obtain central venous access. End-tidal carbon dioxide was never above 10 mmHg. Resuscitation was unsuccessful, and 49 min after the cardiac arrest, the patient was pronounced deceased.

Postmortem examination was performed within 3 h of the event. The heart was examined under water, and a large amount of air, in excess of 100 ml, escaped from both the right atrium and the right ventricle, consistent with the diagnosis of massive air embolism. There was no evidence of a patent foramen ovale. The rest of the autopsy was unremarkable.

Despite the common occurrence of VAE, morbidity and mortality are rare unless large volumes of air are entrained. Extrapolations from canine studies suggest 400–500 ml air is required to cause death in the average human. 3Recently, a case was described in which there was an injection of 200 ml air into the right ventricle and pulmonary artery during attempted pneumopericardiography. 4A much lower volume can prove fatal if a right-to-left shunt (cardiac or pulmonary) allows air to bypass the lungs and enter the systemic circulation. A small air embolism entering the cerebral or coronary circulation could then be fatal. In the absence of a right-to-left shunt, death occurs because of obstruction of right ventricular outflow and the pulmonary circulation with global hypoperfusion, resulting in right and left ventricular ischemia and sudden right-sided heart failure. When air embolism is fatal, postmortem examination of the heart must be performed under water to confirm the presence of air in the right ventricle.

The patient presented herein was at higher risk for VAE because of the presence of placenta previa. 5,6The large venous sinuses surrounding the lower segment of the uterus and the need to cut through these sinuses and the placenta to extract the fetus predisposes to air entrainment. Exteriorization of the uterus may have contributed by placing the venous sinuses in the uterus above the level of the patient's heart, 7,8although studies do not consistently support this theory. Placing the operating table 5–10° in head-up tilt may have lessened the risk of air entrainment, 9but this does not guarantee the heart will be above the level of the incision or that VAE will not occur. 10The use of a precordial Doppler monitor may have identified air embolism at an earlier stage before the terminal event. Although the signal from the precordial Doppler monitor is not specific for VAE, 1its use may have provided warning of ongoing air entrainment before cardiac arrest.

After cardiac arrest occurred, therapeutic options were limited because the etiology was uncertain. Differential diagnosis included embolism (air or thrombus or amniotic fluid) or anaphylaxis. The uterus was open and bleeding, making flooding of the surgical field with saline difficult. With the abdomen open, placing the patient in the left lateral decubitus position, as recommended for air embolism, was problematic. Because cardiopulmonary resuscitation was required, the patient had to remain supine during closure. Extreme left tilt and head-down bed position may have helped to trap air in the right ventricle, but this was inadvisable because the surgical field was not covered with saline to prevent further air entrainment. The head-up position may have prevented ongoing air entrainment; however, this was not performed, partly because of ongoing efforts to acquire central venous access. The head-up position is also unfavorable for cerebral perfusion during cardiopulmonary resuscitation.

It is important to identify parturients who may be at high risk for VAE. If cesarean delivery is planned, additional monitoring could include precordial Doppler ultrasonography. Although Doppler signals can come from other sources, such as fetal squames and agitated blood from insertion and removal of intraabdominal packs, 1its use triggers suspicion of air emboli in high-risk cases. This would help to aid in early diagnosis and lead to formal management plans. In addition to left uterine displacement, the authors recommend that patients be placed 5–10° head up, although the validity of this maneuver is unproven. 10Exteriorization of the uterus should be avoided if possible. If Doppler monitoring identifies VAE, then flooding the surgical field with saline should stop air entrainment.

Preoperative central venous line placement has been advocated for high-risk obstetric cases. 11Given that massive VAE is such a rare event, the risk of central venous catheterization would outweigh the benefit. However, when VAE occurs, a central venous catheter should be placed to evacuate as much air as possible.