Desflurane, enflurane, and isoflurane contain a difluoromethoxy moiety that can be degraded to carbon monoxide (CO). This process occasionally results in patient exposure to toxic CO concentrations and, in rare instances, severe CO poisoning.6 The true incidence of CO exposure during clinical anesthesia is unknown. Routine detection of intraoperative CO exposure is not currently used, but newer pulse oximeters will permit continuous carboxyhemoglobin evaluations.
CO production is inversely proportional to the water content of CO2 absorbents. Calcium hydroxide and barium hydroxide (Baralyme) and soda lime, the two most frequently used CO2 absorbents, are sold wet (13%–15% water by weight), but wet absorbents may dry with high gas-inflow rates. Higher concentrations of CO are most apt to be present during the first case after a weekend because of drying of CO2 absorbent from a continuous inflow of dry oxygen over the weekend.17
Other factors influence the concentration of CO that may result from anesthetic degradation, including temperature (higher temperature increases CO formation), type of absorbent, choice of anesthetic, and concentration of anesthetic. Strong alkalis, such as potassium and sodium hydroxide, initiate the reaction that forms CO. Baralyme, which contains potassium hydroxide, forms more CO than does soda lime, which contains a combination of both.
In one experiment, Baralyme was exposed to 48 hours of dry gas flowing at 10 L/min. Nine swine were then anesthetized with desflurane. Three of the animals died of cardiac arrest within 20 minutes; the other six were successfully resuscitated with IV epinephrine and discontinuation of desflurane.19 Extremely high concentrations of CO (mean peak concentration, 37,000 ppm) were detected in the circuit within 15 minutes of initiating desflurane anesthesia. All the animals had carboxyhemoglobin concentrations above 80%, with a concentration above 90% in seven of the swine. Lower CO concentrations were detected when the CO2 absorbent was exposed to only 24 hours of dry gas and when soda lime was substituted for Baralyme.
Clinical monitors routinely used in the operating room cannot detect CO. Mass spectrometry (available in some operating rooms) cannot directly detect CO because its molecular weight is equivalent to that of nitrogen, a gas usually present in much greater amounts. In addition, detection of CO by fragmentation products is not possible by mass spectrometry because CO2 is present in greater amounts and has similar fragmentation products. However, the presence of CO should be suspected if the mass spectrometer shows the presence of enflurane when it is not being administered.
Trifluoromethane is produced by degradation of isoflurane and desflurane and is responsible for the false readings for enflurane.56 Simultaneous production of trifluoromethane and CO during chemical decomposition of isoflurane and desflurane allows the false reading of the former as enflurane by mass spectrometry to serve as a gross CO monitor and allows for interventions to prevent further CO production and enhance CO elimination. The overall incidence of CO exposure from anesthetic degradation was six of 1372 (0.44%) first cases of the day in which either isoflurane or desflurane was administered.56 Mass spectrometry is a useful monitor for indirect detection of CO poisoning in the clinical setting.55
Although no case reports document patient morbidity or mortality from intraoperative CO exposures, carboxyhemoglobin reportedly as high as 36% may cause morbidity and mortality in patients with concurrent disease.6 Unfortunately, the diagnosis of CO poisoning during anesthesia is difficult because the main clinical features of toxicity are masked by anesthesia, and no routinely available means can identify CO within the breathing circuit or detect when the CO2 absorbent has been desiccated. Delayed neurologic sequelae from intraoperative CO poisoning are likely missed on the anesthesiologist's postoperative patient evaluation.54
The product labels of desflurane and isoflurane have been altered to include a precaution that the CO2 absorbent should be replaced if a practitioner suspects the absorbent is desiccated. However, the problem associated with this warning is the lack of a reliable method for determining when the absorbent is fully or partially desiccated.
If an anesthetic machine is found with the fresh-gas flow "on" at the beginning of the day, a reasonable practice is to replace the absorbent. Changing from Baralyme to soda lime use also should be considered as a protective measure. Newer CO2 absorbents that are less likely to degrade anesthetics are being evaluated but are not yet available in the United States.
Extreme heat and fires within the anesthesia circuits have been reported in a warning distributed by the manufacturer of sevoflurane. The cases involved desiccated carbon dioxide absorbents, high sevoflurane concentrations, and primarily but not exclusively Baralyme adsorbent. The exact mechanisms have not been entirely elucidated but again point to the need to change absorbent when it is dry or routinely every Monday morning before the first case.