All currently volatile anesthetics are degraded by carbon dioxide absorbents to compounds that are toxic. Desflurane, enflurane, and isoflurane are degraded to carbon monoxide (CO), which is neurotoxic and cardiotoxic. A 1-h exposure to 1,000 ppm CO will produce approximately 25% carboxyhemoglobin, sufficient to cause severe neuropsychiatric impairment, whereas 67% carboxyhemoglobin causes death.  The current Environmental Protection Agency limit for a 1-h exposure is 35 ppm. Signs and symptoms of CO toxicity are masked during and after anesthesia, and carboxyhemoglobin cannot be detected by pulse oximetry.
In the past decade, case reports, abstracts, and newsletter articles have described various aspects of CO formation and poisoning. A gradual pattern emerged, in which most CO cases involved the first patient anesthetized on a Monday morning using an anesthetic machine idled for 2 days, the carbon dioxide absorbent was implicated in CO formation, and high inspired CO concentrations were associated with absorbent that had been used for a long time. This pattern was partially rationalized by the observation that relatively dry absorbent was required for CO formation,  which suggested a scenario whereby gas flowing through an anesthesia machine over a weekend could dry the absorbent and produce CO poisoning on Monday morning. Unfortunately, most of these reports never appeared as manuscripts in the peer-reviewed literature and received little attention. More importantly, several questions were left unanswered: What is the clinical incidence of CO exposure? What were the clinical effects of such exposure? Under what conditions might it occur? Is CO exposure a real concern or a “tempest in a teapot?” In this issue of Anesthesiology are two investigations that address these issues.
Frink et al. studied the sequelae of absorbent drying in an anesthesia machine under plausible conditions of clinical use: oxygen (5–10 l/min) for 24–48 h through a circle system containing Baralyme [registered sign] or soda lime with the reservoir bag in place or removed, followed by desflurane (1 minimum alveolar concentration) anesthesia. This is a Laboratory Report only because swine, not humans, were anesthetized. The results are startling and disconcerting. After drying Baralyme [registered sign] at 10 l/min for 48 h without a reservoir bag, desflurane anesthesia caused extremely high inspired CO concentrations (37,000 ppm), and all animals had carboxyhemoglobin concentrations of more than 80%. Three pigs died, and the remaining animals remained hypotensive despite resuscitation with epinephrine. Clearly, physiologically devastating and unacceptable CO exposures can occur. Frink et al. also defined conditions in which significant CO formation did not occur (reservoir bag in place, lower gas flows).
Woehlck et al. studied the incidence of CO exposure in a teaching hospital and tested the hypothesis that educating operating room support staff (anesthesia technicians and housekeepers) to intervene to prevent absorbent drying could reduce CO exposures. Despite educating all anesthesia personnel about the factors predisposing to CO formation (specifically absorbent drying), the incidence of CO exposure was 1 in 200 first cases. After instructing OR support personnel to turn off anesthesia machines at the end of the day and to replace absorbent if gas was found flowing in the morning, the incidence was reduced to 1 in 2,000 cases. That operating room support personnel can contribute to patient safety is novel. Woehlck et al. also identified a previously unreported factor that promotes CO formation.
These two reports serve notice that clinical CO formation is not uncommon and can have serious sequelae. They further define and refine conditions under which CO formation occurs and can be avoided. Most importantly, they provide straightforward and readily implemented ways in which it can be prevented, both by anesthesia practitioners and nonpractitioners. In addition to those guidelines promulgated by the Food and Drug Administration to reduce CO exposure,  we now have additional ways to protect patients.
Evan D. Kharasch, M.D., Ph.D.
Professor; Departments of Anesthesiology and Medicinal Chemistry (Adjunct); University of Washington; Box 356540; Seattle, Washington 98195