RECENTLY, warnings have been published by the manufacturer of sevoflurane and in the Anesthesia Patient Safety Foundation Newsletter warning about the potential risks for fires resulting from an exothermic reaction between sevoflurane, oxygen, and a desiccated Baralyme® (Allied Healthcare Products, Inc., St. Louis, MO) carbon dioxide absorbent.‖ We report an incident in which the heat and the gases generated resulted in an explosion within an anesthesia machine.
A 3-yr-old male with a history of chronic cough, dysphagia, and recurrent pneumonias was scheduled for rigid bronchoscopy, esophagoscopy, and bilateral inguinal hernia repair. A mask induction was performed with a combination of 8% sevoflurane, nitrous oxide (6 l/min), and oxygen (4 l/min). While a peripheral intravenous catheter was inserted the sevoflurane concentration was decreased to 6%, nitrous oxide was discontinued, and oxygen flow was maintained at 6 l/min. Direct laryngoscopy was performed, and the vocal cords were topicalized with 4% lidocaine. The airway was then turned over to the surgeon. The bronchoscope was introduced into the trachea and the anesthesia circuit was connected to the side port of the bronchoscope. Ventilation was assisted and the inspired sevoflurane concentration was maintained at 8% with an end-tidal concentration of 6–7%. Oxygen flow was maintained at 5–6 l/min.
Approximately 3 min after the introduction of the bronchoscope and 15 min after induction of anesthesia, a loud explosion was heard in the vicinity of the anesthesia machine. A quick inspection revealed that the front and side panel of the ventilator housing was blown out, the expiratory valve was destroyed, the automatic pressure limiting valve was completely dislodged from its seat, and the scavenger tubing was blown off the unit (fig. 1). In addition, the lower canister of carbon dioxide absorbent (Baralyme®) was charred, the drain plug of the canister was bright red, and multiple sparks were seen in the canister (fig. 2). The anesthesia circuit and bag were completely intact and not warm to touch.
The anesthesia circuit was immediately disconnected from the bronchoscope and the patient was ventilated with an anesthesia bag connected to a wall oxygen source. The patient remained hemodynamically stable at this point with saturations of 98–100%. Anesthesia to permit continuation of the bronchoscopy employed a propofol infusion. There was no evidence of thermal injury or barotrauma in the airway. The anesthesia machine was removed form the operating room for further inspection.
The anesthesia machine was subsequently disassembled by technicians of Datex-Ohmeda (Datex-Ohmeda, Inc., Andover, MA) with our Clinical Engineering Department to ascertain the degree and pattern of damage.
Recent warnings from both Abbott Laboratories (Abbott Park, IL) and the Anesthesia Patient Safety Foundation have raised awareness of the potential for severe heat buildup in the anesthetic circuit, particularly in the carbon dioxide absorber canister, when sevoflurane is used in association with desiccated carbon dioxide absorbant.1,2
The anesthesia machine in use was a Datex-Ohmeda Aestiva™ Model 5. The procedure was the first case of the day and a complete check of the machine was performed before inducing anesthesia. The machine apparently had not been turned off the day before, but all fresh gas flows and vaporizers were off. The carbon dioxide absorbent in use was Baralyme®. The absorbent was not discolored but there was no way to determine if the absorbent was desiccated. High fresh gas flows were used during induction. There was no discrepancy between the sevoflurane vaporizer setting and the inspired sevoflurane concentration, induction of anesthesia was not prolonged, and adequate anesthesia was achieved. There was no warning of the problem before the explosion. Fortunately, there was no apparent patient injury.
For the explosion to occur there must have been combustible vapor present in addition to oxygen. Although sevoflurane may be combustible in oxygen at high concentrations, it is more likely that breakdown products such as methanol and formaldehyde that have been observed under laboratory conditions were responsible.2In addition, the plastic mesh base of the lower canister was completely melted and charred. It is possible that plasticizing agents present in the materials may have been vaporized by the excessive heat and constituted a combustible material. Carbon monoxide is also a breakdown product of this reaction; however, there was nothing to suggest that it was present in this case. Judging by the location of the observed sparks, charring, and melting, the primary site of the exothermic reaction was at the base of the canister. In the Aestiva machine, the expiratory gas enters at the top of the absorbent canister. It seems likely that the expiratory carbon dioxide and moisture from the patient may have been sufficient to hydrate the Baralyme® granules in the upper layers sufficiently to prevent reaction with sevoflurane but not in the lower canister.
Perhaps most noteworthy, the melted plastic bottom of the lower plastic canister totally occluded outflow port at the bottom of the canister base. This occlusion may have protected the patient from the explosive pressure wave being propagated down the inspiratory limb of the circuit. Instead, the explosive force proceeded backwards through the canister into the breathing circuit module and the expiratory valve. This was responsible for lifting the circuit module off the base, shattering the expiratory valve disc, dislodging the gas sampling module, and blowing the automatic pressure limiting valve out of its seat. It is unclear if the occlusion of the outflow port of the carbon dioxide contributed to the explosion by restricting the gas flow at the bottom of the canister such that there was an accumulation of combustible gases.
In retrospect, this may not have been the first incident in which an exothermic reaction was observed in our facility. Approximately 18 months before this incident, the plastic casing of the carbon dioxide absorber of a Dräger 2B anesthesia machine (Dräger Medical, Inc., Telford, PA) was discovered to be melted into a peculiar convex shape, as if the desiccant had become extremely hot and resulted in the expansion of the casing. This particular machine is located in the induction room in our magnetic resonance imaging suite and is infrequently used. In addition, episodes of measured sevoflurane concentrations far lower than set on the vaporizer combined with overheating of an absorbent canister was recently reported in another minimally used machine.
Because of this event, we have elected to change our carbon dioxide absorbent in our main operating rooms from Baralyme® to soda lime and to use Amsorb® (Armstrong Medical Ltd., Coleraine, Northern Ireland) in our machines that are utilized only intermittently. Amsorb® is an effective carbon dioxide absorbant that remains hydrated and does not contain the strong bases necessary to produce carbon monoxide in clinically significant amounts or react exothermically with sevoflurane.3