BECAUSE of my research interest in anesthetic breakdown, I am aware of cases that I would call “rumors of injury by carbon monoxide,” which is one of the products produced by the interaction of anesthetics and carbon dioxide absorbants. Some lacked documentation, and others were mired in the litigation process. Consequently, none of these rumors was ever published. Our specialty treated every instance of carbon monoxide poisoning or other kinds of anesthetic breakdown as a mere curiosity because no injury had ever been documented and reported.

In this issue of Anesthesiology, three cases are presented where extreme reactions occurred between sevoflurane and desiccated absorbent, producing irrefutable patient injury in one case,1and explosion and fire in the other two.2,3Fatheree and Leighton1describe a patient whose intraoperative pulmonary injury necessitated a prolonged intensive care unit course. Details of absorbent drying were unknown but fit the historical pattern of Monday mornings and a weekend of gas flow through the absorbent. For years, we obsessed over the theoretical risk of compound A nephrotoxicity, but we now know there is a real risk to the lungs. Although the injury can be mild,4serious injury from anesthetic breakdown is no longer a rumor.

Those with a pyromaniacal bent will be interested in reading Wu et al.’  s dramatic report,2in which an unattended anesthesia machine exploded and ignited, and the report by Castro et al. , in which the explosion occurred during an anesthetic.3In the case by Wu et al. , the machine was located in an induction room, where prolonged gas flows through the absorbent were possible if not routine. Presumably, previously desiccated absorbent combined with approximately 1 hour of sevoflurane and oxygen flow and culminated in the first reported explosion and actual fire outside of a research setting since the days of flammable anesthetics.5The hydration status of the absorbent was also unknown in the report by Castro et al. , but the case was the first of the day, the anesthesia machine had not been turned off the night before, and Baralyme® (Allied Healthcare Products, Inc., St. Louis, MO), relatively high concentrations of sevoflurane, and high-flow oxygen were used.

Intraoperative explosions and fires are not new to anesthesiology. As a resident, one of my faculty mentors described an intraoperative explosion that had fatal consequences for the patient. This, too, was a dramatic story, but I felt safe and uninvolved. It was legitimate for me to be in denial because I would never have the opportunity to use ether or cyclopropane in an operating room. Decades of practice with flammable anesthetics had mandated the use of protective antistatic devices; conductive rubber and floor sweeping chains are anachronistic reminders that still linger in some operating rooms. However, in the 21st century, we must again accept that explosion, fire, and pulmonary injury are no longer rumors. Most of us will never experience (or perhaps recognize) a similar case in our careers, just as few of our forebearers experienced an ether fire or cyclopropane explosion, but we are at risk. Today, our uncertainty resides in whether or not the absorbent is dry. We must accept this risk, and work to reduce it.

These mishaps are rare. It would be unhelpful to overreact with panic or illogical responses, but we should remain sufficiently concerned and educated that we do not cower behind denial, ignorance, or apathy. Fundamental questions remain: Why don’t we see this type of reaction more often? Baralyme® and sevoflurane have been used together for a long time. When desiccated Baralyme® or soda lime are used with sevoflurane, breakdown reactions always occur; however, only a narrow range of conditions can produce these dramatic, incendiary events.5Although comparative studies remain unpublished, it may be more difficult for soda lime to create dangerous situations than Baralyme®, but it is clearly not impossible.4Perhaps a better question is “Do milder forms of these reactions occur frequently and go unrecognized?” Other than the meltdowns, explosions, and fires, the clinical signs of moderate carbon monoxide poisoning may resemble commonly seen perianesthetic problems like confusion, headache, and nausea.6If the meltdown described by Fatheree and Leighton went unnoticed, the pulmonary injury might have been attributed to occult aspiration of gastric acid or some other convenient diagnosis of exclusion.

These cases also suggest that ignition and pulmonary injury may be mutually exclusive events. One may need to distinguish between incipient fires,7with pulmonary injury and extreme heat but no ignition, and the two cases of actual fires in which flames resulted. It is likely that patients absorb flammable and presumably toxic gases, preventing fire at the cost of pulmonary injury.

In all of these cases, Baralyme® was used, and desiccation was likely but unproven. Theoretically, we already know that stopping prolonged dry gas flow should prevent absorbent desiccation. Realistically, we can’t watch the absorbent 24/7, and the possibility of absorbent desiccation before placement in anesthesia machines has never been ruled out. We know that we can avoid these dangerous chemical reactions by using absorbents devoid of strong bases. Unfortunately, recent literature is mixed about the ability of absorbents lacking strong bases to absorb carbon dioxide adequately in all situations.8Strong bases were originally included in absorbents because they enhanced carbon dioxide absorption, but definitive modern comparative studies are lacking. Of course, water inhibits these chemical reactions, but water can evaporate. Nonvolatile inhibitors of anesthetic breakdown exist but have not been evaluated extensively.9 

We need to learn more about the toxic and flammable chemicals produced by sevoflurane breakdown. Is it just fumes from the hot plastic? Or does formaldehyde cause the reported injuries?1,5Anesthetic breakdown monitors can be developed, but will we use them? Carbon monoxide is produced by the breakdown of all modern anesthetics, and we already have carbon monoxide sensors that work in the presence of anesthetics.10,11But until our specialty demands this technology, no company will have incentive to build monitors that incorporates it. Color change occurs during absorbent desiccation,12,13but improvements will be necessary before becoming a clinically useful indicator. As suggested by Fatheree and Leighton, perhaps today’s best monitor, available in every operating room, is to place a temperature probe into the absorbent during every case. The use of a patient monitor for absorbent temperature is demonstrated in figures 1 and 2. Most patient monitors typically work up to 50°C. If the temperature approaches this value,14excessive heating from anesthetic breakdown should be suspected, and appropriate action taken. This useful, low-cost intervention is available in any operating room with nearly every modern patient monitor, regardless of manufacturer.

After a decade of study, the problems of anesthetic breakdown and solutions to these problems are still evolving. These case reports should bring them into our consciousness.

Department of Anesthesiology, Medical College of Wisconsin, Milwaukee, Wisconsin, and Department of Anesthesiology, Froedtert Memorial Lutheran Hospital West, Milwaukee, Wisconsin.

Fatheree RS, Leighton BL: Acute respiratory distress syndrome after an exothermic Baralyme®-sevoflurane reaction. Anesthesiology, 2004; 101:531–3
Wu J, Previte JP, Adler E, Myers T, Ball J, Gunter JB: Spontaneous ignition, explosion, and fire with sevoflurane and barium hydroxide lime. Anesthesiology, 2004; 101:534–7
Castro BA, Freedman LA, Craig WL, Lynch C: Explosion within an Aestiva anesthesia machine: Baralyme®, high fresh gas flows and sevoflurane concentration. Anesthesiology 2004, 101:537–9
Baum J, Sitte T, Straus JM, Forst H, Zimmermann H, Kugler B: Die reakton von sevofluran mit trockenem atemkalk: Überlegungen anläβlich eines aktuellen zwischenfalls. Anästhesiol Intensivmedizin 1998; 39:11–6
Holak EJ, Mei DA, Dunning MB 3rd, 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
Woehlck HJ: Severe intraoperative CO poisoning: Should apathy prevail? Anesthesiology 1999; 90:353–4
ECRI Editorial Staff: Hazard Report: Anesthesia carbon dioxide absorber fires. Health Devices 2003; 32:436–40 (Dec 2003 reprinted update)
Higuchi H, Adachi Y, Arimura S, Kanno M, Satoh T: The carbon dioxide absorption capacity of Amsorb® is half that of soda lime. Anesth Analg 2001; 93:221–5
Woehlck HJ, Dunning III MB: Inhibition of carbon monoxide formation from desflurane breakdown by desiccated barium hydroxide lime (abstract). Anesthesiology 2003; 99:A1238
Dunning M 3rd, Woehlck HJ: Performance of an electrochemical carbon monoxide monitor in the presence of anesthetic gases. J Clin Monit 1997; 13:357–62
Bermudez JA: Investigation of electrochemical carbon monoxide sensor monitoring of anesthetic gas mixtures. Anesthesiology 2003; 99:1233–5
Dunning III MB, Barth CD, Woehlck HJ: Yellow discoloration of barium hydroxide lime indicates desiccation (abstract). Anesthesiology 2003; 99:A626
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
Luttropp HH, Johansson A: Soda lime temperatures during low-flow sevoflurane anaesthesia and differences in dead space. Acta Anaesthesiol Scand 2002; 46:500–5