To the Editor:-Frink et al. showed that extremely high carboxyhemoglobin (COHb) concentrations can be produced in 25-kg pigs by the inhalation of desflurane through a breathing circuit with dried CO sub 2 absorbents, which produce carbon monoxide (CO) through anesthetic breakdown. These experiments were performed at 40% inspired oxygen, although in clinical practice, inspired oxygen concentrations can vary from less than 25% to nearly 100%, impacting the resultant severity of CO poisoning as measured by COHb concentrations.
We attempt to exemplify the clinical effect of various inspired oxygen concentrations during the rapidly changing CO concentrations produced by desflurane breakdown by using mathematical modeling to extrapolate the data of Frink et al. We interpolated parts per million CO from the graphic data presented in figures 3 and 4 of the study by Frink et al. and applied these data to the equation developed by Coburn et al., (CFK equation) to determine the COHb concentration. Good experimental fit of this equation has been documented by Peterson et al. We used this equation to calculate the accumulation of COHb as a function of the absorption and elimination of CO through respiration using Microsoft Excel (Microsoft Corporation) for the MacIntosh (Apple Computer, Inc., Cupertino, CA). The implementation of the CFK equation required that assumptions be made to complete the respiratory, profile, as shown in Table 1. We used the CFK equation to calculate the COHb concentration by an iterative method using CO concentrations interpolated from the graphic data of Frink et al. To validate the appropriateness of using the CFK equation, we graphically show the correlation of the calculated COHb data to the measured experimental data of Frink et al. in Figure 1. This results in an r2value of 0.997 for the descending data points and an overall r2value of 0.967. Several possible explanations exist for the slight deviation of the experimental and calculated data during the ascending phase of the curve. It is possible that physiologic parameters (e.g., cardiac output) were severely compromised during this phase of the experiment, and differences reflected changes in circulation in the experimental animals from those for which the CFK equation was developed. Despite these limitations, the CFK equation provides excellent predictions under these extreme conditions.
In Figure 2, we show the results of mathematical modeling using the CFK equation to extrapolate the CO concentration data to different inspired oxygen concentrations (FIO2). These COHb concentrations would be predicted to result from exposure to the CO concentrations interpolated from the data of Frink et al. and also at 100% and 25% oxygen. Predicted results at 100% oxygen yielded smaller COHb concentrations, possibly accounting for the low incidence of CO poisoning detected by clinical signs. Conversely, the predicted results at 25% oxygen suggest that CO toxicity would be enhanced under these conditions. The routine use of an increased FIO2may reduce CO poisoning during anesthesia because of the competitive binding of CO and oxygen, but CO poisoning occurring with low FIO2may be more severe.
Harvey J. Woehlck, M.D.
Associate Professor of Anesthesiology
Marshall B. Dunning III, Ph.D.
Assistant Professor of Medicine; Medical College of Wisconsin; Milwaukee, Wisconsin 53226
(Accepted for publication December 2, 1997.)