Desflurane degrades to carbon monoxide (CO) when reacted with desiccated carbon dioxide (CO2) absorbents, as has been reported in various in vitro studies [1-6]. The complete desiccation of absorbents is a rare phenomenon in routine anaesthetic practice. Therefore, the production of CO under clinical hydrated absorbent conditions must be considered. In a clinical low-flow setting, the inspired concentrations of CO were affected by absorbent hydration , personal smoking habits [7,8] and other physiological confounding factors of patients including, haemoglobin catabolism  and CO2 excretion. Increasing oxygen (O2) gas flow rates prevented the accumulation of inspired CO , but increased the production of CO by anaesthetic degradation . The reaction of absorbent with CO2 is exothermic [8,9] and accelerates anaesthetic degradation [1,3]. However, some studies demonstrated that CO2 reduced CO production in completely desiccated absorbents [10,11]. All these factors that affect the production of CO have not been examined simultaneously. This in vitro study was performed to monitor the production of CO by controlling the O2 gas flow rates, CO2 flow and anaesthetic concentrations, using hydrated CO2 absorbents under low-flow anaesthesia clinical conditions to elucidate the competitive interaction of these factors.
Figure 1 depicts the experimental arrangement. A lung model was created by connecting a 2 L latex breathing bag to the Y piece of a circle breathing system attached to a Modulus® CD Anaesthesia System (Ohmeda Inc., Madison, WI, USA). The machine was equipped with two canisters of CO2 absorbent. ‘Exhaled' gases flowed through the upper (upstream) canister to the lower (downstream) canister, in the direction depicted by the arrows in Figure 1. The breathing bag was filled with 99% pure CO2 from a CO2 cylinder through a T-connector to simulate the clinical production of CO2 by adult patients. The CO2 flow rates were set to 200 and 400 mL min−1, to represent the excretion of CO2 by the general population and by those with extremely high body weight. The flow rates were calibrated using a GilibratorTM flow calibrator (Gilian® Instrument Corporation, Clearwater, FL, USA). Based on the temperature and humidity records on the breathing circuits of patients in the authors' earlier work , a respiratory humidifier (MR 600, Fisher & Paykel Ltd., Auckland, New Zealand) was connected to the expiratory limb to increase the temperature and humidity in the tube to match human lungs. The probes of a Testo 400 multi-function-measuring instrument (Testo GmbH & Co., Lenzkirch, Germany) were used to monitor the temperature and relative humidity, as presented in Figure 1. One of the probes was set immediately above the absorbents to measure the relative humidity and temperature in the breathing tube; the other was set inside the absorbents of the upstream canister. The data were recorded continuously during the experiments.
Measurement of carbon monoxide and hydration of absorbent
The concentrations of CO in the breathing tube were continuously measured using an electro-chemical monitoring system as given in the authors' earlier work . The detection limit of the CO monitor (Dräger Pac III CO detection instrument; Dräger Safety Inc., Pittsburgh, PA, USA) was 1 ppm. The concentration of CO was measured every minute. The CO monitor was recalibrated using 100 ppm CO standard gas for span calibration and 99.9% nitrogen gas was used to calibrate a zero reading before every experiment as directed by the manufacturer. Absorbents (75–80% calcium hydroxide (Ca(OH)2) and ∼3.5% sodium hydroxide (NaOH), SofnolimeTM; Molecular Products Ltd., Thaxted, Essex, UK) were sampled from the surfaces of the upstream and downstream canisters to determine the water content. The absorbent sample was initially weighed, and then dried at 125°C for 2 h, before being weighed again . The hydration (%) of a particular absorbent sample was defined as the difference between the two mass measurements:
where Ma = mass after drying and Mb = mass before drying.
The ‘change in hydration' of the absorbents was defined as the difference between the hydration (%) after and that before each experiment. Change in hydration (%) = [hydration (%) (after experiment)] − [hydration (%) (before experiment)]. According to this definition, a positive hydration change indicates the increased hydration of absorbents during the experiment.
The tidal volume and the breathing rate in the anaesthesia machine were set to 700 mL and 10 breaths min−1, respectively. The I : E ratio was set to 1 : 2. The partial pressures of desflurane were 0.5, 1 and 2 MAC (minimum alveolar anaesthetic concentration) (3.6%, 7.3% and 14.5%, respectively). The end-tidal desflurane partial pressures were adjusted using the vapourizer and monitored using a monochromatic infrared monitor (RGM 5250; Ohmeda, Louisville, CO, USA). The fresh gas flow (FGF) rates (oxygen) were set to 0.2, 0.5 and 1 L min−1. Experiments were designed to determine the interactive effects of desflurane concentration (0.5, 1 or 2 MAC), CO2 input (200 or 400 mL min−1) and O2 gas flow rate (0.2, 0.5 or 1 L min−1) on CO concentration in the breathing tube. The hydration and temperature of absorbent were measured in all experiments. Each experiment was run for 2 h and performed four times. Prior to each experiment, the absorbents in the upstream and downstream canisters were replaced and the anaesthesia system was flushed using fresh oxygen until a zero CO reading was observed.
The 2-h time-weighted average concentration of CO (TWA-CO) was the arithmetic mean of 120 min and reported as the mean concentration of CO during each experiment. The time-weighted average temperatures in the breathing tube and absorbents, average hydration change (%) in absorbents and average relative humidity in the breathing tube were calculated for regression analyses. The stepwise multiple linear regression analyses were performed using SPSS 12.0 for Windows (SPSS Inc., Chicago, IL, USA). The adjusted r2 value was reported for each regression model. The U-test was performed to compare the hydration changes (%) between the upstream and downstream canisters in each experiment.
Table 1 presents the means and standard deviations of 2-h TWA-CO in the breathing tube under various test conditions. Desflurane degraded to generate CO in the presence of hydrated absorbent. The maximum mean 2-h TWA-CO concentration was 13.1 ppm and the peak CO concentration was 23 ppm in the presence of 0.5 MAC desflurane, a CO2 flow of 400 mL min−1 and an FGF rate of 0.2 L min−1. Figure 2 depicts that the CO concentration in the breathing tube was steadily produced and accumulated during the 2-h experiments. At CO2 flow rate = 200 mL min−1 and FGF rate = 0.5 and 1 L min−1, the CO concentration in the breathing tube was less than 5 ppm or even approached zero in some particular experiments (Figs 2a, c, e). The dilution effect of FGF was most prominent at an FGF rate of 1 L min−1. A lower FGF rate corresponded to a greater CO concentration at a given MAC of the anaesthetic and a given CO2 flow rate. This trend was consistent with our previous result . Inspired concentrations of CO were not correlated with the MAC of the anaesthetics. For a particular FGF and MAC, CO concentrations were higher at a CO2 flow rate of 400 mL min−1 than at a CO2 flow rate of 200 mL min−1.
Stepwise multiple regression analyses were performed to identify the factors that were significantly associated with CO concentrations (Table 2). The FGF and CO2 flow rates were the only two significant factors. The FGF negatively influenced the concentration of CO while the CO2 flow rate positively influenced the concentration of CO. FGF was the factor that was the most strongly associated with CO productions, explaining 31.7% of its variation. Adding the CO2 flow rate explained percentage a further 9.5% of the variation. The regression model with these two factors explained 41.2% of CO accumulation. Desflurane partial pressures were not a statistically significant variable in the multiple linear regression models.
The changes in the hydration (%) of the absorbents in the upstream and downstream canisters were significant being −4.9% (1st∼3rd quartile = −6.5∼ −3.2) in the upstream canister and 2.9% (1st∼3rd quartile = 1.9∼3.8) in the downstream canister (P < 0.05). Stepwise multi-linear regression was separately performed for upstream and downstream canisters. No statistically significant model could be established to explain the change in hydration in the downstream canister. Table 3 presents the regression results of the upstream canister. The CO2 flow rate negatively influenced the change in hydration (%) of the absorbents, explaining 18.2% of the variation (Table 3). A higher CO2 flow rate corresponded to greater desiccation of the absorbent surface. Neither FGF nor desflurane partial pressures significantly influenced the changes in hydration in this case.
Stepwise multi-linear regression analysis of the temperatures in the upstream canister absorbents and the experimentally controlled factors (desflurane partial pressures, CO2 flow rate and FGF) revealed that only the association with the CO2 flow rate was statistically significant, with an r2 value of 0.54 (Table 3). Increasing the CO2 flow rate increased the absorbent temperature. The difference between the absorbent temperatures at the two CO2 flow rates was 10.1°C (Table 3).
During the experiments, CO was produced by the degradation of desflurane with fresh hydrated soda lime; the peak CO concentration in the breathing tube was 23.0 ppm (Figs 2b, d) in the presence of 0.5 MAC desflurane, a CO2 flow of 400 mL min−1 and an FGF rate of 0.2 L min−1. Many studies have adequately determined how the FGF, CO2 production, anaesthetic concentration and absorbent's hydration and temperature influence CO production separately. In our study, FGF was more strongly associated with CO concentration than with CO2 flow rate (Table 2; r2 = 0.317 for FGF only). There was no significant association with desflurane partial pressure. The absorbent's hydration and temperature also affected the CO production [1,3,7]. However, they fluctuated during clinical anaesthesia and were not able to be controlled practically. Statistically, CO2 was the only contributor to the changes in the absorbent's hydration and temperature in this study. The contribution of CO2 to hydration was 18.2% and to the temperature was 54.0%.
Effect of fresh gas flow
CO concentrations in a circle breathing system result from anaesthetic degradation, and expiration by the patient that is counteracted by fresh gas removal. The anaesthesia system can be treated as a confined space in which CO is a pollutant and through which gas passes to dilute and remove CO.
Equation (1) describes the relationship between the steady-state concentration of contaminant (Css), the rate of contaminant generation (G) and the rate of gas flow through the closed space (Q) . The concentration of contaminant, Css, is inversely related to the gas flow rate, Q, in this environment.
where K = safety/mixing factor.
Referring to Table 2, the FGF diluted and washed out CO. Increasing the FGF by 1 L min−1 reduced the TWA-CO concentration in the breathing system by 6.2 ppm. This is in agreement with previous in vitro studies that also demonstrated increased CO concentration with decreasing FGF in the presence of CO2 [4,14].
Woehlck and colleagues  demonstrated that increasing the FGF rate increased CO production, and delivered more anaesthetic to the CO2 absorbent canister, forming a dry absorbent surface that promoted degradation. During our experiments, the lowest hydration of the soda lime in the upstream canister was 3.1% (data not shown). This low hydration of the absorbents probably favoured ongoing anaesthetic degradation and CO production, increasing variable ‘G' in Equation (1).
The degradation products – CO, H2O and heat – flowed to the downstream canister of soda limes in fresh gas. This drift markedly increased the hydration of the downstream canister of absorbents, resulting in the 7.2% difference between the average changes in the hydration (%) of the upstream and downstream canisters. The hydration (%) in the downstream canister was affected by anaesthetic degradation product – H2O from the upstream canister. The hydration of the absorbents actually fluctuated during the experiment. The determinant factor that caused this high hydration content of this part would be the H2O from the upstream canister. Therefore, none of the experimental independent variable was identified as a statistically significant factor.
The FGF was negatively correlated with CO concentration and acted as the dilutor in this breathing circuit where CO was produced. The FGF did provide a dry soda lime surface that facilitated CO production but this positive effect was overwhelmed by the dilution effect, as the Q in Equation (1), in this bench model setting. This dilution effect was identified as a negative correlation between the FGF rate and the concentration of CO (TWA-CO), as presented in Table 2.
Effect of CO2
Some in vitro studies have not incorporated CO2 effects and could not evaluate the effect of the patient's CO2 excretion on CO production during anaesthesia. As Table 3 indicates, the CO2 flow rate is positively related to the temperature of the upstream canister soda lime. The absorption of CO2 by absorbents is an exothermic reaction . The heat released by CO2 absorption accelerates the degradation of anaesthetic. In this work, increasing the CO2 flow rate increased the temperature of the upstream canister absorbent but reduced the hydration of the absorbent (Table 3). The production of water following CO2 absorption is reportedly responsible for the decrease in CO formation in desiccated absorbents [10,11]. However, the water generated by CO2 absorption was carried to the downstream canister by the FGF and increased the hydration of the absorbent in the downstream canister, such that the average change in hydration in the downstream canister was positive (2.3%) while that of the upstream canister was negative (−4.9%). Accordingly, CO2 negatively affected the change in hydration (%) in the upstream canister.
CO2 affected the production of CO and the temperature and change in hydration of the absorbent based on the regression results. Indeed, CO2 directly influenced the condition of the absorbent, and the temperature and hydration (%) of the upstream canister absorbents then affected the production of CO. CO2 indirectly affected the production of CO. In clinical practice, the CO2 flow in the breathing circuit is related to the body weight of the patient and is classified as an endogenous factor that affects CO production. The findings herein help to elucidate why the body weight of a patient can substantially influence the in-tube concentration of CO during clinical anaesthesia .
Effect of concentration of anaesthetic
Anaesthetic degradation is a first-order reaction in desiccated lime, but is zero-order as hydration increases. When lime supports zero-order degradation kinetics, CO is fixed at a low value that is independent of anaesthetic concentration . This low-flow anaesthetic system design did not begin with completely desiccated soda limes. Hence, CO production was independent of the anaesthetic concentration, which could be predicted from the zero-order reaction kinetics.
One limitation of this study is that the experiment duration did not last for more than 2 h. The concentration of CO cannot be adequately predicted by this study if the duration of low-flow anaesthesia is longer than 2 h. Holak and colleagues  indicated that CO increased with minute ventilation (VE), the VE of different patients, such as children, young adults or adults affect the inspired CO concentration. Because the ventilator was set to deliver 10 breaths min−1 at 700 mL tidal volume for a total of 7-L VE in this study, further study of the correlation between inspired CO and VE under low-flow anaesthesia without drying the CO2 absorbents in advance is needed.
Another limitation of this study is the tested soda lime containing strong base (e.g. 3.5% NaOH), which was capable of initiating the reaction responsible for CO production [5,6,17,18]. However, some CO2 absorbents are NaOH-free and KOH-free, and the anaesthetic degradation under these absorbents was not tested by this study.
This study did not take into account the oxygen consumption as would occur during the clinical anaesthesia. Under such circumstances, at an FGF setting of 200 mL min−1 it may be expected that the concentration of CO would be higher than the concentrations that were reported in this study.
The neurologic dysfunction happened when the COHb% in blood was higher than 30% and the lethal effect appeared if it was higher than 50% . Patients with low haemoglobin quantities (such as anemia) or of small body size (children) might have more severe CO exposure based on the attainment of a higher COHb%. Especially for patients with ischaemic heart disease, even low COHb% can produce morbidity . No such patient was included in our study. To assess the risk of clinical harm that caused by the intraoperative CO exposures, patients conditions such as body weight, smoking habits and disease status should be considered for clinical application.
In summary, the significant factors of the concentration of CO in the breathing tube are FGF and CO2 flow for desflurane. The desflurane concentration was not a statistically significant factor in this bench model. The temperatures and hydration contents of absorbents are actually affected by the CO2 flow, which indeed determines whether the absorbent is in favour for CO production. In the clinical practices, the only controllable factor is the FGF rate to reduce the undesirable CO exposure. The other significant factor, but with less power, is CO2 flow rate, is actually differed for each patient. It is recommended that different anaesthetics and durations should be tested under the same experimental design.
The authors like to thank the National Science Council of the Republic of China, Taiwan, ROC, for financially supporting this research under Contract No. NSC 92-2314-B-002-227.
1. 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
2. Baxter PJ, Garton K, Kharasch ED. Mechanistic aspects of carbon monoxide
formation from volatile anaesthetics. Anesthesiology
3. Baxter PJ, Kharasch ED. Rehydration of desiccated baralyme prevents CO formation from desflurane in an anaesthesia machine. Anesthesiology
4. Knolle E, Heinze G, Gilly H. CO formation in dry soda lime is prolonged at low gas flow. Anesth Analg
5. Knolle E, Heinze G, Gilly H. Small CO formation in absorbents does not correlate with small CO absorption. Anesth Analg
6. Stabernack CR, Brown R, Laster MJ, Dudziak R, Eger EI II. Absorbents differ enormously in their capacity to produce compound A and CO. Anesth Analg
7. Tang CS, Fan SZ, Chan CC. Smoking status and body size increase carbon monoxide
concentrations in the breathing circuit during low-flow anaesthesia. Anesth Analg
8. Woehlck HJ, Connolly LA, Cinquegrani MP, Dunning MB III, Hoffmann RG. Acute smoking increases ST depression in humans during general anaesthesia. Anesth Analg
9. Middleton V, Poznak AV, Artusion JF, Smith SM. Carbon monoxide
accumulation in closed circle anaesthesia system. Anesthesiology
10. Woehlck HJ, Dunning MB III, Raza T, Ruiz F, Bolla B, Zink W. Physical factors affecting the production of carbon monoxide
from anaesthetic breakdown. Anesthesiology
11. Knolle E, Gilly H. Absorption of carbon dioxide by dry soda lime decreases CO formation from isoflurane degradation. Anesth Analg
12. Neumann MA, Laster MJ, Weiskopf RB et al.
The elimination of sodium and potassium hydroxides from desiccated soda lime diminishes degradation of desflurane to carbon monoxide
and sevoflurane to compound A but does not compromise carbon monoxide
absorption. Anesth Analg
13. Industrial Ventilation – A Manual of Recommended Practice, 21st edn. American Conference of Governmental Industrial Hygienists, Committee on Industrial Ventilation, Lansing, NE, USA, 1992.
14. Strum DP. Low-flow anaesthesia: anaesthetic degradation to CO and compound A. Curr Opin Anesthesiol
15. Strum DP, Eger EI II. The degradation, absorption and solubility of volatile anaesthetics in soda lime depend on water content. Anesth Analg
16. Holak EJ, Mei DA, Dunning MB III et al. Carbon monoxide
production from sevoflurane breakdown: modeling of exposure under clinical conditions. Anesth Analg
17. Murray JM, Renfrew CW, Bedi A, Mccrystal CB, Jones DS, Howard Fee JP. A new CO absorbent for use in anaesthetic breathing systems. Anesthesiology
18. Evan DK, Powers KM, Artru AA. Comparison of Amsorb, sodalime and baralyme degradation of volatile anaesthetics and formation of CO and compound A in swine in vivo. Anesthesiology
19. Woehlck HJ. Severe intraoperative CO poisoning (Editorial Views). Anesthesiology
20. Woehlck HJ, Mei D, Dunning MB III, Ruiz F. Mathematical modeling of CO exposures from anesthetic breakdown. Anesthesiology