ALL modern volatile anesthetics produce carbon monoxide in strongly exothermic reactions with anhydrous carbon dioxide absorbents. Although desflurane produces the most carbon monoxide, the reaction with sevoflurane produces the most heat.1Using real anesthesia machines, sevoflurane, and desiccated Baralyme® (Allied Healthcare Products, Inc., St. Louis, MO), peak absorbent canister temperatures of 100–401°C have been documented.1–3In six clinical cases, sevoflurane reacted with dry soda lime to produce hot canisters and, in two of these cases, 4–9% carboxyhemoglobin.4,5We report a sevoflurane reaction with Baralyme® associated with melted absorbent cartridges, thermochemically-induced acute respiratory distress syndrome, and 29% carboxyhemoglobin.

A 40-yr-old, 109-kg man underwent thyroid lobectomy as the first case on a Monday. The patient had smoked one pack of cigarettes daily for 22 yr. He was otherwise healthy except for mild hypertension and hypercholesterolemia. Medications included metoprolol, naproxen, simvastatin, and chlorthalidone. Physical examination was normal. Preoperative hemoglobin was 16.8 g/dl; hematocrit was 49%.

The operating room was equipped with a Datex Ohmeda Modulus II anesthesia machine (Datex-Ohmeda, Inc., Andover, MA). The Baralyme® carbon dioxide absorbent cartridges had been changed the previous Thursday evening. On Friday, the machine was used for five brief, uneventful anesthetics. The anesthesia machine was turned off over the weekend and turned on Monday morning. The sevoflurane vaporizer was filled appropriately. Oxygen then flowed through the machine at 200 ml/min for 4 h.

Initial room air oxygen saturation was 97%. We administered intravenous midazolam 2 mg. Anesthesia was induced intravenously with fentanyl 100 μg, lidocaine 100 mg, and propofol 200 mg. We ventilated the patient with 100% oxygen and 8% sevoflurane. Intravenous succinylcholine 100 mg facilitated tracheal intubation. We instituted controlled ventilation with a minute volume of 9 l/min (750 ml × 12 breaths/min). Anesthetic maintenance consisted of sevoflurane, nitrous oxide 50–70%, and intermittent intravenous doses of fentanyl 50–100 μg. Total fresh gas flow was 3 l/min for 30 min and 2 l/min thereafter. Nondepolarizing muscle relaxants were not used. Because the patient moved after skin incision, 10 min after anesthetic induction, we administered intravenous propofol 50 mg, intravenous fentanyl 50 μg, and briefly increased the sevoflurane vaporizer setting to 8%. The sevoflurane vaporizer setting (8%) did not match the measured inspired sevoflurane concentration (3%). We believed that this was a calibration error and arranged to have the monitor recalibrated as soon as possible. The monitored gas concentration remained lower than the sevoflurane vaporizer setting for the rest of the anesthetic. Fifty min after induction, the airway circuit inadvertently disconnected. The circuit and tube did not feel hot, the only odor was that of sevoflurane, and the gas was colorless. We reconnected the circuit and continued the anesthetic.

When the surgical drapes were removed, 110 min after anesthetic induction, we found that the carbon dioxide container was extremely hot. We immediately disconnected the circuit inspiratory limb from the carbon dioxide canister and saw billowing white gas. We smelled burning plastic. No flames were seen. The inspiratory limb of the circuit tubing contained a small amount of water. The patient was ventilated with 100% oxygen by Ambu bag. The endotracheal tube did not feel hot or look charred. Breath sounds, initially coarse, cleared with suctioning. Fiberoptic bronchoscopy revealed erythema in the trachea and both mainstem bronchi, copious mucus, and a small amount of superficial bleeding. Arterial blood gas analysis showed: pH 7.36, pCO245 mmHg, pO2208 mmHg, oxyhemoglobin 69%, and carboxyhemoglobin 29%.

The patient was admitted to the intensive care unit and transferred to a tertiary hospital on postoperative day (POD) 1. Fiberoptic bronchoscopy on POD 1 revealed necrotic tracheal mucosa and severe hyperemic bronchitis bilaterally in the proximal airways. Chest radiography showed a moderately extensive left perihilar infiltrate. Hospital-acquired pulmonary infection was diagnosed on POD 4 and treated with antibiotics. The patient met blood gas criteria for acute respiratory distress syndrome (Paco2/Fio2< 200 mmHg without chronic cardiac or pulmonary disease)6from the day of surgery through POD 5. He was extubated on POD 6. The patient did not develop proteinuria or elevated blood creatinine or urea nitrogen levels. Urine output was normal. He did not complain of any visual problems. The patient was discharged home on POD 9. Subsequent recovery has been uneventful.

Immediately after the event, personnel from the Barnes-Jewish St. Peters Hospital biomedical department examined the carbon dioxide canister stack. The Baralyme® granules in the top half of the upper cartridge had turned purple; the rest of the granules were white. The outside plastic canister was almost too hot to touch. A large quantity of water was present in the manifold above the absorbent stack. The plastic mesh on the superior and inferior aspects of the lower Baralyme® cartridge and the inferior aspect of the upper cartridge had melted away (fig. 1). The lower Baralyme® cartridge was partially fused to the outside plastic canister (fig. 2). Personnel from ECRI (formerly the Emergency Care Research Institute, Plymouth Meeting, PA) made the following observations after examining the machine on POD 8. The rubber gasket below the lower canister was burnt over half of its inner circumference and part of its upper surface. The superior mesh of the top cartridge had no apparent heat damage but, in a central, roughly circular area 4 cm in diameter, had many fully and partly occluded openings, which appeared to be from a molding defect. On POD 8, the weight of the affected granules and plastic cartridges was 4.8% less than the weight of two unused Baralyme® cartridges from the same lot.

Fig. 1. Inferior aspects of the lower (right) and upper (left) Baralyme® cartridges. Note the extensive melting of the bottom of the upper cartridge and the top and bottom of the lower cartridge. The rubber gasket of the lower cartridge, seen on the right, melted onto the adherent Baralyme® granules. 

Fig. 1. Inferior aspects of the lower (right) and upper (left) Baralyme® cartridges. Note the extensive melting of the bottom of the upper cartridge and the top and bottom of the lower cartridge. The rubber gasket of the lower cartridge, seen on the right, melted onto the adherent Baralyme® granules. 

Close modal

Fig. 2. Side view of the lower Baralyme® canister. The plastic of the Baralyme® cartridge melted onto and deformed the plastic of the outer canister. The top and bottom of this cartridge melted away. 

Fig. 2. Side view of the lower Baralyme® canister. The plastic of the Baralyme® cartridge melted onto and deformed the plastic of the outer canister. The top and bottom of this cartridge melted away. 

Close modal

This is the most severe patient lung injury and the highest carboxyhemoglobin level that has been reported after an exothermic sevoflurane-absorbent reaction. Our patient had no renal or retinal problems despite exposure to sevoflurane breakdown products. Six previous exothermic sevoflurane-soda lime cases have been reported in the German literature.4,5One patient developed bronchospasm and required ventilation overnight; the other five patients were extubated immediately after surgery.4,5All patients, including ours, recovered fully.

Sevoflurane reacts chemically with desiccated Baralyme® or soda lime to produce carbon monoxide and flammable organic compounds, including methanol and formaldehyde.7The reaction produces heat and heat increases the reaction speed, so the rate of sevoflurane breakdown can accelerate rapidly.7,8Sevoflurane may be so extensively consumed that maintaining anesthesia is difficult.7Compound A is produced but then, to a variable extent, degraded further.8At high temperatures, the flammable metabolites can spontaneously combust.2,7A peak absorbent canister temperature of 120–140°C is generally reached 10–50 min after the start of the reaction, followed by a rapid decrease in canister temperature.1–3However, in four nonhuman trials with real anesthesia machines, the temperature rose rapidly to 330–401°C and parts of the absorbent canister melted.2,3,7The reaction generates water, which can suddenly appear in the circuit.3Extremely dry absorbent darkens rapidly during this reaction; however, there may be little or no color change in partially desiccated absorbent despite significant sevoflurane degradation.7In our case, fumes from the melted Baralyme® cartridges may have affected the severity of the lung injury.

Under experimental conditions, rapid sevoflurane degradation requires desiccated absorbent; therefore, we assume that our Baralyme® was partially dry. We do not know how this dehydration occurred. Possibilities include initially dry Baralyme®, an undocumented period of high gas flow, regional drying resulting from gas channeling fostered by the partially perforated cartridge mesh, or unauthorized anesthesia machine use. Nitrous oxide abuse and machine sabotage have been reported as causes of unexpected absorbent exhaustion.9 

Sevoflurane breakdown is a greater problem with Baralyme® than with soda lime. Fresh soda lime contains 14–18% water, and carbon monoxide is only produced in soda lime with <4.7% water.10In contrast, fresh Baralyme® contains 13–19% water, and Baralyme® containing <9.7% water can produce carbon monoxide.7,10Higher peak absorbent temperatures are reached with Baralyme® than soda lime when dry absorbent is exposed to sevoflurane.3All four of the recent sevoflurane-absorbent anesthesia machine fires investigated by ECRI involve Baralyme® (personal written communication on October 30, 2003 between Barbara Leighton and Albert L. de Richemond, MS, PE, Associate Director, Accident and Forensic Investigation Group, ECRI, Plymouth Meeting, PA).

During sevoflurane degradation, carbon monoxide concentrations increase with larger minute ventilation volumes and lower fresh gas flows.7Large minute volumes increase contact between sevoflurane molecules and absorbent granules. Low fresh gas flows allow recirculation of carbon monoxide. Anemia increases the percent carboxyhemoglobin for a given carbon monoxide exposure.7Our patient’s relative polycythemia protected him from an even higher carboxyhemoglobin concentration.

Abbott Laboratories (Abbott Park, IL), the manufacturer of sevoflurane, sent a letter 6 weeks after our case notifying anesthesia providers of revision of the package insert. The package insert now contains the following paragraph.11 

Replacement of Desiccated CO2 Absorbents:  When a clinician suspects that the CO2absorbent may be desiccated, it should be replaced before administration of sevoflurane. The exothermic reaction that occurs with sevoflurane and CO2absorbents is increased when the CO2absorbent becomes desiccated, such as after an extended period of dry gas flow through the CO2absorbent canisters. Extremely rare cases of spontaneous fire in the respiratory circuit of the anesthesia machine have been reported during sevoflurane use in conjunction with the use of a desiccated CO2absorbent. Rapid changes in the color of some CO2absorbents or an unusually delayed rise in the delivered (inspired) gas concentration of sevoflurane compared with the vaporizer setting may indicate excessive heating of the CO2absorbent canister and chemical breakdown of sevoflurane.

Clinician vigilance and a high index of suspicion will be needed to detect sevoflurane degradation in its early stages. All cases of clinically significant sevoflurane breakdown have involved large discrepancies between the dialed and the monitored agent concentration. Such discrepancies should be investigated. However, running out of agent and monitor dysfunction are more common causes of sevoflurane concentration discrepancy than agent breakdown. In our case, there was no known lengthy period of high flow of dry gas and the absorbent color changed minimally. In the future, consideration should be given to monitoring the temperature of the absorbent canister or to monitoring airway gases for carbon monoxide.

If sevoflurane degradation is suspected, fresh absorbent should be installed quickly. Switching to a different potent agent would decrease potential heat generation but would not eliminate carbon monoxide production. Carboxyhemoglobin concentrations of 8–32% were reported in 29 patients receiving enflurane or isoflurane anesthesia.12Fifteen minutes of desflurane anesthesia with desiccated Baralyme® produced 36% carboxyhemoglobin in one patient.13The Baralyme® was then changed and the carboxyhemoglobin level decreased quickly despite continued desflurane anesthesia. Unlike our case, that patient was extubated after 140 min and discharged home 2 h later.13All potent agents react with dry absorbent.1 

Ideally, an absorbent would not react with sevoflurane or would indicate when it was dry. New absorbents that lack sodium and potassium hydroxide (e.g. , Amsorb® (Armstrong Medical, Ltd., Coleraine, Northern Ireland)) react minimally with potent agents and change color when dehydrated.14,15These absorbents are available in Europe but not in the United States.

In summary, we report acute respiratory distress syndrome after an exothermic sevoflurane-absorbent reaction. In response to this case, the authors’ departments have switched from Baralyme® to soda lime and now routinely replace absorbent cartridges each Monday morning. Absorbents that react minimally with potent agents are available in Europe but not in the United States. Until this situation changes, clinicians must avoid dry absorbents and be prepared to recognize and respond to the early signs of this reaction.

Wissing H, Kuhn I, Warnken U, Dudziak R: Carbon monoxide production from desflurane, enflurane, halothane, isoflurane, and sevoflurane with dry soda lime. Anesthesiology 2001; 95:1205–12
Steffey EP, Laster MJ, Ionescu P, Eger EI II, Gong D, Weiskopf RB: Dehydration of Baralyme® increases compound A resulting from sevoflurane degradation in a standard anesthetic circuit used to anesthetize swine. Anesth Analg 1997; 85:1382–6
Eger EI II, Ionescu P, Laster MJ, Weiskopf RB: Baralyme® dehydration increases and soda lime dehydration decreases the concentration of compound A resulting from sevoflurane degradation in a standard anesthetic circuit. Anesth Analg 1997; 85:892–8
Janshon GP, Dudziak R: Interactions of dry soda lime with enflurane and sevoflurane: Clinical report on two unusual anesthesias [in German]. Anaesthesist 1997; 46:1050–3
Baum J, Sitte T, Strauss JM, Forst H, Zimmermann H, Kugler B: Reaction of sevoflurane with dry absorbents: Thoughts based on recent incidents [in German]. Anaesth Intensivmed 1998; 39:11–6
Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L, Lamy M, Legall JR, Morris A, Spragg R: The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994; 149:818–24
Holak EJ, Mei DA, Dunning MB, Gundamraj R, Noseir R, Zhang L, Woehlck HJ: Carbon monoxide production from sevoflurane breakdown: Modeling of exposures under clinical conditions. Anesth Analg 2003; 96:757–64
Fang ZX, Kandel L, Laster MJ, Ionescu P, Eger EI II: Factors affecting production of compound A from the interaction of sevoflurane with Baralyme® and soda lime. Anesth Analg 1996; 82:775–81
Yudenfreund-Sujka SM: Nitrous oxide abuse presenting as premature exhaustion of Sodasorb (letter). Anesthesiology 1990; 73:580
Fang ZX, Eger EI II, Laster MJ, Chortkoff BS, Kandel L, Ionescu P: Carbon monoxide production from degradation of desflurane, enflurane, isoflurane, halothane, and sevoflurane by soda lime and Baralyme®. Anesth Analg 1995; 80:1187–93
Ultane® (sevoflurane) package insert, Abbott Laboratories, Abbott Park, IL, revised August, 2003
Moon RE: Cause of CO poisoning, relation to halogenated agents still not clear. Anesthesia Patient Safety Foundation Newsletter 1994; 9:13–24
Berry PD, Sessler DI, Larson MD: Severe carbon monoxide poisoning during desflurane anesthesia. Anesthesiology 1999; 90:613–6
Murray JM, Renfrew CW, Bedi A, McCrystal CB, Jones DS, Fee JPH: Amsorb: A new carbon dioxide absorbent for use in anesthetic breathing systems. Anesthesiology 1999; 91:1342–8
Knolle E, Linert W, Gilly H: The color change in CO2absorbents on drying: An in vitro  study using moisture analysis. Anesth Analg 2003; 97:151–5