A 10-year-old, 23.6-kg female patient presented for posterior spinal fusion, at levels C5 to L4, for the correction of severe kyphoscoliosis. A medical history of meningitis as a baby was complicated by recurrent spinal abscesses requiring drainage. Subsequently, the patient required multiple spine procedures including laminectomies and insertion of growth rods. She was developmentally normal and had no neuromuscular disorder. The only abnormality on neurologic examination was a minor left-sided weakness from a spinal cord injury that occurred during one of her previous surgical procedures. Three months before the surgery, her growth rods had been removed and a halo brace fitted for traction. She was on no regular medications and had no known drug allergies.
After induction of general anesthesia, the patient was turned prone with the patient’s halo connected directly to the Jackson spinal table. Motor-evoked potentials and somatosensory-evoked potentials were normal at baseline and remained normal throughout the procedure. The surgical repair took approximately 7 hours. Blood loss was minimal, and the patient remained hemodynamically stable throughout. Just before closure the surgeons applied Tisseel (Baxter, Westlake Village, CA), a fibrin sealant, to parts of the wound to aid hemostasis and then attempted to place an epidural at the T4/T5 level. The spinal anatomy at this level was altered severely, making direct placement of the catheter impossible. Approximately 5 attempts were made to locate the epidural space with a Tuohy needle and a loss of resistance (LOR) to air technique. Each attempt used a 10-mL syringe that was half-filled with air. Immediately after this, there was a sudden loss of cardiac output, evidenced by a loss of arterial blood pressure and end-tidal carbon dioxide (Etco2). The electrocardiogram showed normal sinus rhythm.
The patient was resuscitated according to the pediatric advanced life support guidelines. The resuscitation was made difficult because of the patient’s prone position and open back wound. The wound was closed rapidly with staples and the halo disconnected from the traction table to allow the patient to be tuned supine, while resuscitation efforts continued. Spontaneous circulation returned after 7 minutes during which time 3 arrest doses of epinephrine were administered.
Clinical examination, mobile chest radiography, and transthoracic echocardiography failed to provide a definitive diagnosis for the arrest. The patient was commenced on low-dose dopamine, for blood pressure support, and transferred intubated and ventilated to the intensive care unit. At this stage, the working diagnosis was anaphylaxis caused by Tisseel and blood was taken for mast cell tryptase assays.
On the first postoperative day, she had a repeat transthoracic echocardiography that demonstrated apical ballooning of the left ventricle that the cardiologists described as being consistent with Takotsubo cardiomyopathy. The inotropic support begun in the operating theater was able to be weaned by the fourth postoperative day. During this time, a computed tomography scan of head revealed ischemia in the right middle cerebral artery territory. Opinions from both neurology and radiology suggested that this was consistent with an embolic event and there was no evidence of watershed infarction that might be expected after a period of decreased cerebral perfusion. During the sequential echocardiograms performed in the 4 days after the event, the apical dyskinesia and ballooning resolved and a bubble test failed to demonstrate any evidence of intracardiac shunt.
On the fourth postoperative day, she returned to theater for completion of her surgery, which was uneventful, and she was extubated later that day. Mast cell tryptase results from immediately after the event and 4 hours following were within normal limits. She was discharged from intensive care unit 2 days later, 6 days after the original surgery. During the period after extubation, it became apparent that the child had marked left upper limb weakness that was not present before the surgery, consistent with the right-sided infarction seen on computed tomography scan. Over the course of the following 2 weeks, however, these new neurologic symptoms returned to baseline.
Consent for Publication
Written consent was provided by the patient’s family to publish this report.
Venous air embolism (VAE) occurs when there is an opening in a vein, and there is a pressure gradient that allows gas to enter the vein. Prolonged prone position, disruption to tissues from trauma or surgery, and low venous pressure are well-recognized risk factors for developing VAE.1
Sutherland and Winter2 described 2 fatal cases of VAE that occurred during posterior spinal fusions in children. The first case described an 8-year-old child who had an asystolic arrest while positioned prone during surgery. During the resuscitation, a thoracotomy was performed and air was felt within the heart. In the second case, a 12-year-old child having spinal surgery, prolonged resuscitation attempts were unsuccessful, and at autopsy air was found in both the coronary and cerebral circulation.
In addition to trauma during surgery, it also is possible for air to enter the circulation during epidural placement. Children have smaller epidural spaces than adults with a rich and dilated plexus of epidural veins. According to Schwartz and Eisenkraft3, both these factors make injury to vessels, with the relatively large Tuohy needle, more likely in children. Complications from pediatric epidurals, however, have been found to occur at a similar rate to adult epidurals.4 One technique for identifying the epidural space involves a LOR to an air-filled syringe, which can lead to an unknown quantity of air entering the circulation. The air may reach the circulation by direct intravenous injection, or alternatively the sudden increase in epidural pressure may disrupt epidural venules and allow air to enter.3
There have been multiple cases reported in the literature of VAE secondary to LOR to air. A review of more than 24,000 epidurals, caudals, and spinals in children during a 10-year period in France and Belgium found 5 serious events, including 3 deaths. All of these cases involved a LOR to air technique and probable VAE.5
Schwartz and Eisenkraft3 reported a case of a 9-month-old infant who had a probable VAE secondary to air in the epidural space. Three milliliters of air was injected into the space through a Tuohy needle during epidural placement, and there was a subsequent decrease in Etco2, blood pressure, and oxygen saturations.3 Guinard et al6 described a case of VAE from air being injected into the caudal space of a 2-year-old child evidenced by a subsequent decrease in Etco2.
After multiple reports of VAE in children secondary to LOR to air techniques, it was recommended in a 1993 editorial by Sethna and Berde7 to only use saline for LOR techniques in children. Despite this recommendation, expert opinion 10 years later still recommended that air was the best medium for LOR techniques in infants and neonates.8 In 2005, Ames et al9 found that despite expert opinion most pediatric anesthetists in Canada used an LOR to saline technique and the authors concluded that using LOR to air was less safe due to the potential for complications, including VAE.
Various other complications have been attributed to LOR to air techniques. These include incomplete analgesia, pneumocephalus, nerve root compression, and subcutaneous emphysema. Most of the complications from LOR to air have been described in case reports, retrospective reviews, observational studies, and surveys.10 Very few randomized, controlled trials have been performed comparing LOR with air versus saline. A Cochrane review in 2014 examined 7 randomized, and quasi-randomized, controlled trials and found no difference in complications between the 2 techniques, although the authors point out that the results were based on low-quality evidence.11
In our case, the placement of the epidural was difficult and required multiple attempts by the surgeon. A number of attempts were made because it was felt that epidural analgesia would provide the best chance of a quick and comfortable recovery after such an extensive repair. Usual practice involves placing the epidural catheters under direct vision; however, in this case, the anatomy was altered significantly, making this technique impossible. Some spinal surgeons prefer a LOR to air technique because it is the technique that is most familiar to them and some believe it gives a better feel than saline. Repeated injections of air may have resulted in a significant volume of air entering the circulation.
In children, the volume of air required to cause morbidity is relatively low. In a review of fatal embolic events in children, Byard12 found that 2 mL of air in the cerebral circulation and 0.5 to 1 mL of air in the pulmonary vein was enough to cause death. Our patient had a sudden loss of cardiac output after the epidural attempts that we believe was caused by air trapping in the right side of the heart and pulmonary artery. The patient received 3 arrest doses of epinephrine in a 6-minute period; however, spontaneous circulation did not return until she was emergently turned supine, allowing for more effective chest compressions.
Our patient had significant neurologic sequelae from the arrest. The neurologic symptoms observed in the postoperative course were most likely due to air in the cerebral circulation. Cerebral air embolism nearly always is associated with an intracardiac shunt; however, it is possible for air to travel in a retrograde manner from the epidural veins to the cerebral circulation bypassing the pulmonary circulation. A case of fatal cerebral air embolism, in an 11-year-old girl undergoing scoliosis surgery, has been reported in which autopsy showed no cardiac defects.1 In our case, however, the neurologic findings and computed tomography brain findings were consistent with right middle cerebral artery territory ischemia that is the most common site of arterial embolic infarcts.
Given our patient had no evidence of intracardiac shunt, it raises the question of how air emboli could enter the arterial circulation. Thackray et al13 published the case of an adult man who had air accidentally enter his circulation via an open right internal jugular catheter. Within 5 minutes a transesophageal echocardiogram was performed and air was seen in the left atrium, left ventricle, and aortic root. Air was seen flowing in from the pulmonary vein, and no air was seen in the right side of the heart. There was no intracardiac shunt. The authors discuss the concept of the pulmonary vasculature acting as a filter for microbubbles of air; however, if the amount of air is large enough, then this filtration threshold is exceeded and air may reach the systemic circulation. In animal models, it has been found that infusions of air greater than 0.35 mL/kg/min exceed the filtration capacity of the lungs and microbubbles appear in the left side of the heart. These microbubbles may coalesce and form larger, more clinically significant bubbles that can cause cerebral embolic events. In our patient, a bolus of 8.3 mL would be enough to exceed the filtration capacity of the lungs and cause a paradoxical embolus.
In conclusion, this case demonstrates that it is possible for air to be introduced to the circulation when a LOR to air technique is used for epidural placement and that children only require small amounts of air to cause serious clinical events. It also confirms that paradoxical air embolus can occur in the absence of an intracardiac shunt if the filtration capacity of the lungs is exceeded. We would urge all practitioners to employ a LOR to saline technique when placing epidurals in children.
The authors thank the patient and her family for providing consent to publish this case report.
Name: Paul F. Lee-Archer, MBBS, FANZCA.
Contribution: This author helped review the literature, and prepare and revise the manuscript.
Name: Brett Chaseling, MBBS, FANZCA.
Contribution: This author helped research the case and revise the manuscript.
This manuscript was handled by: Mark C. Phillips, MD.
1. De-Giorgio F, Arena V, Turturro F, d’Aloja E, Miscusi M. 11 year-old girl undergoing scoliosis surgery. Brain Pathol2012;22:571574.
2. Sutherland RW, Winter RJ. Two cases of fatal air embolism in children undergoing scoliosis surgery. Acta Anaesthesiol Scand1997;41:10731076.
3. Schwartz N, Eisenkraft JB. Probable venous air embolism during epidural placement in an infant. Anesth Analg1993;76:11361138.
4. Llewellyn N, Moriarty A. The national pediatric epidural audit.Paediatr Anaesth. 200717520533.
5. Flandin-Bléty C, Barrier G. Accidents following extradural analgesia in children. The results of a retrospective study. Paediatr Anaesth1995;5:4146.
6. Guinard JP, Borboen M. Probable venous air embolism during caudal anesthesia in a child. Anesth Analg1993;76:11341135.
7. Sethna NF, Berde CB. Venous air embolism during identification of the epidural space in children. Anesth Analg1993;76:925927.
8. Bissonnette B, Dalens D. Pediatric Anesthesia. Principles and Practice. 2002.New York: McGraw-Hill.
9. Ames WA, Hayes JA, Pétroz GC, Roy WL. Loss of resistance to normal saline is preferred to identify the epidural space: a survey of Canadian pediatric anesthesiologists.Can J Anaesth. 200552607612.
10. Shenouda PE, Cunningham BJ. Assessing the superiority of saline versus air for use in the epidural loss of resistance technique: a literature review. Reg Anesth Pain Med2003;28:4853.
11. Antibas PL, do Nascimento Junior P, Braz LG, Vitor Pereira Doles J, Módolo NS, El Dib R. Air versus saline in the loss of resistance technique for identification of the epidural space. Cochrane Database Syst Rev. 2014;7:CD008938.
12. Byard RW. Fatal embolic events in childhood. J Forensic Leg Med2013;20:15.
13. Thackray NM, Murphy PM, McLean RF, deLacy JL. Venous air embolism accompanied by echocardiographic evidence of transpulmonary air passage. Crit Care Med1996;24:359361.