Share this article on:

Absorption of Carbon Dioxide by Dry Soda Lime Decreases Carbon Monoxide Formation from Isoflurane Degradation

Knolle, Erich MD*; Gilly, Hermann PhD*†

doi: 10.1213/00000539-200008000-00041
General Articles

This study was performed to determine whether the absorption of carbon dioxide (CO2) influences the formation of carbon monoxide (CO) from degradation of isoflurane in dry soda lime. Isoflurane (0.5%), CO2 (5%), a combination of the two in oxygen, and pure oxygen were separately passed through samples of 600 g of completely dried soda lime (duration of exposure, 60 min; flow rate, 5 L/min). Downstream of the soda lime, we measured concentrations of CO, isoflurane, and CO2 as well as the gas temperature. CO2 increased the peaks of CO concentration (842 ± 81 vs 738 ± 28 ppm) and shortened the rise time of CO to maximum values (12 ± 2 vs 19 ± 4 min). However, CO2 inhibited total CO formation (99 ± 10 vs 145 ± 6 mL). At the same time, CO2 absorption by the soda lime decreased in the presence of CO formation (from 21.4 ± 0.8 to 19.4 ± 0.9 g). The temperature of the gases increased during the passage of both isoflurane and CO2 (to 32.6 ± 2.0°C and 39.4 ± 4.0°C, respectively), but the largest increase (to 41.5 ± 2.1°C) was recorded when isoflurane and CO2 simultaneously passed through the dry soda lime. We assume that the simultaneous reduction in CO formation and CO2 absorption is caused by the competition for the alkali hydroxides present in most of soda lime brands.

Implications We determined, in vitro, that carbon monoxide (CO) formation from isoflurane by dry soda lime is reduced by carbon dioxide (CO2). We believe that the potential for injury from CO is less in the clinical milieu than suggested by data from experiments without CO2 because of an interdependence between CO formation and CO2 absorption.

*Department of Anesthesiology and General Intensive Care (B), University of Vienna; and †Ludwig Boltzmann Institute for Experimental Anesthesiology and Research in Intensive Care Medicine, Vienna, Austria

April 21, 2000.

Supported, in part, by the Austrian Ministry for Science and Transport.

Address correspondence and reprint requests to Erich Knolle, MD, Department of Anesthesiology and General Intensive Care (B), University of Vienna, Waehringer Guertel 18–20, A-1090 Vienna, Austria. Address e-mail to

When soda lime, used in anesthesiological procedures to absorb carbon dioxide (CO2), contains <4.8% water, carbon monoxide (CO) forms from the degradation of some inhalational anesthetics (1). Several groups found carboxyhemoglobin levels to have increased by more than 30% during routine anesthesia (2–4). CO formation depends on the anesthetic used (desflurane > enflurane > isoflurane; but no formation with halothane or sevoflurane), its concentration, the ambient temperature, and, in particular, the chemical composition of the absorbent (5) and its residual water content (6). Even before it was established that some anesthetics induce CO formation, it was known that isoflurane, enflurane, and halothane are partly lost when passing through dry soda lime, a loss that was then attributed to adsorption (7). At present, the loss of some halogenated anesthetics in dry soda lime is considered to be the consequence of anesthetic degradation, with concurrent formation of CO (6) and trifluoromethane (8) initiated by base-catalyzed deprotonation. When sufficient water is present, the resulting carbanion is reprotonated (9).

Under clinical conditions, CO formation and CO2 absorption take place simultaneously in soda lime. Because the absorption of CO2 produces water and heat (along with carbonic acid) (10), we hypothesized that CO2 would affect the degradation of volatile anesthetics in soda lime, as well as the resulting production of CO.

Thus, the aim of the current study was to determine, in an in vitro setting, how 5% CO2, a clinically typical concentration at the inlet to the CO2 absorber in a rebreathing system, added to isoflurane would modify CO formation when the gases are passed through dried soda lime. In addition, we studied the effect of 5% CO2 on the loss of isoflurane, the time course of the gas temperature, and the change in weight of the soda lime.

Back to Top | Article Outline


Fresh soda lime (Draegersorb 800®; Draeger, Luebeck, Germany) was dried by piping oxygen through it (medicinal grade; humidity <100 ppm) at a constant flow of 10 L/min. After 48 h, the soda lime was weighed every 4 h (PJ3000; Mettler-Toledo, Greifensee, Switzerland). Oxygen flow was discontinued when the incremental weight reduction was <0.05% of the prior soda lime weight. The residual water content comprised <0.5% of the weight, as confirmed when 100 g samples were dried further at 120°C for at least 4 h.

The CO2 absorbent canister of an anesthesia apparatus (Dameca, Copenhagen, Denmark) was filled with 600 ± 0.05 g of dried soda lime, and different gas mixtures at a constant gas flow (5 ± 0.1 L · min 1) were passed through it for a period of 60 min. The humidity of the gases was <100 ppm (Gas Analyzer 1301; Brüel & Kjær, Nærum, Denmark). In experimental Group A (n = 6), the gas mixture consisted of piped oxygen, 0.5% isoflurane added by a vaporizer (Vapor 19.3; Draeger), and 5% CO2 added from a tank. In Group B (n = 7), the gas mixture was of 0.5% isoflurane and oxygen. In Group C (n = 6), the gas mixture was 5% CO2 and oxygen. In Group D (n = 3), the passing gas was pure oxygen.

To control our experiment, we repeated it with 600 ± 0.05 g of fresh soda lime, introducing the identical gas mixtures as in the experimental groups. The identification of Groups Afresh–Dfresh, is parallel to the above for the gas mix, and in each case n = 3.

Inlet oxygen and CO2 flow rates were adjusted with separate mass flow controllers (Mass-Flo® 1259; MKS Instruments, Andover, MA). Gas inflow was monitored continuously (AWM 5101 VN; Honeywell, Minneapolis, MN). The inlet CO2 concentration was verified by sidestream measurement (PM8020; Draeger), and the isoflurane concentration was checked with an anesthetic monitor in the mainstream (IRINA; Draeger).

Downstream of the absorbent, a sidestream analyzer (PM8020, Draeger) measured the isoflurane and the CO2 outlet concentration (resolution 0.1%). Downstream CO concentration was measured in sidestream with a polarographic CO sensor (CO3E-1000; Sensoric, Bonn, Germany) (range, 0–1000 ppm). To test cross-sensitivity to agents other than CO, we performed the following test: 400 ppm CO, in a mixture of 40% nitrogen and oxygen were passed through the sensor at a constant flow of 5 L/min. After 30 min 0.5% isoflurane and then, 5% CO2 was added. Within 60 min the CO signal varied by no more than 14 ppm.

A polarographic oxygen sensor (O2 sensor Code No. 6850645; Draeger) (accuracy ± 3%) and a thermocouple (7563 digital thermometer; Yokogawa, Tokyo, Japan) (accuracy ± 0.1°C) were placed in the mainstream. On completion of the experiments, the soda lime was again weighed.

CO concentration (ppm = 10 6), gas flow (L/min), and temperature (°C) were charted on an eight-channel analog recorder (LR 8100; Yokogawa). CO2 percentage and isoflurane percentage concentrations were recorded digitally at 1-min intervals. We calculated the outlet volumes of CO, isoflurane, and CO2 from outlet gas flow, gas concentration, and duration of exposure:MATH where conci = mean concentration during time interval (Δt = 1 min).



Inlet flow was considered to represent outlet flow, although, in fact, outlet flow must have been less than inlet flow because of the absorption of CO2 and isoflurane in the soda lime. Because we expected both gases to be only partly absorbed and the outflow to increase slightly because of vaporizing water from the CO2 absorption, we regarded the difference between inlet and outlet flow to be smaller than 5%, and therefore, the potential error in calculation of the outlet volumes of gases to be negligible. This was confirmed by additional measurements of outlet flow with the Honeywell-flowmeter in some of the experiments, which, in no case, showed a difference from inlet flow of more than 5%.

The loss of isoflurane and of CO2 in soda lime (caused by chemical degradation and absorption) was calculated from the differences between inlet and outlet amounts of the gases. Values of lost gas amounts were corrected for STPD (standard temperature, pressure, dry) conditions by using the conversion factor for humid gases (11) based on the mean outlet temperatures and on an ambient pressure of 741 torr (mean atmospheric pressure during all of the experiments). The weight of the gas lost in the soda lime was calculated as:MATH where MW = molecular weight.



The weight of water generated from CO2 absorption in the soda lime was calculated as:MATH where MW = molecular weight. The water generated from equimolar deprotonation of isoflurane (9) (totaling approximately one tenth of the weight of isoflurane loss in soda lime) was not considered relevant to our calculations.



Results are presented as mean ± sd. For the comparison of variables of CO production, isoflurane loss, outlet temperature, and changes in soda lime weight between experimental Groups A and B the Mann-Whitney U-test was used. The same test was used to compare Groups A and C with respect to CO2 absorption. Calculated and measured weight changes in the absorbent in Groups A, C, Afresh, and Cfresh, respectively, were compared by using the Wilcoxon’s ranked sum test. For statistical calculations, the standard SPSS software version 7.5.2G (SPSS, Chicago, IL) was used. A two-tailed probability of <0.05 was regarded as statistically significant.

Back to Top | Article Outline


In all of the experiments involving the passage of 0.5% isoflurane through dried soda lime, CO generation was consistently reproducible; its time course depended on the presence or absence of CO2 (Figure 1). CO concentration peaked at a considerably higher level in Group A than in Group B (Table 1). Time to CO peak and to reduction in CO concentration to 50% of maximum were significantly shorter in Group A than in Group B. The calculated total production of CO was significantly smaller in A than in B. When 5% CO2 in oxygen (Group C), or oxygen alone (Group D) were passed through the dried soda lime, CO did not form. In the control procedures using fresh soda lime (Groups Afresh–Dfresh), CO generation was not detected in any of the experiments.

Figure 1

Figure 1

Table 1

Table 1

When both the isoflurane and the CO2 in oxygen were passed through dried soda lime (Group A), the downstream isoflurane concentration increased to the set concentration significantly earlier than in the experiments without CO2 (Group B). Thus, the calculated loss of isoflurane in soda lime was smaller in Group A than in Group B, the difference being statistically significant (Table 1). With fresh soda lime (Groups Afresh and Bfresh), downstream isoflurane concentrations increased to the set concentration within 2 min; the loss of isoflurane was thus minimal.

The maximum outlet temperature in Group A (41.5 ± 2.1°C) was significantly higher than in Group B (32.6 ± 2.0°C). Moreover, the time interval to the maximum temperature was significantly shorter in Group A than in Group B (36 ± 1 vs 48 ± 2 min). In Group C the maximum temperature of 39.4 ± 4.0°C was reached in 47 ± 2 min. When pure oxygen was passed through the dried soda lime (Group D), outlet flow temperatures changed by <0.3°C.

In Group Afresh the temperature rose steadily to 33.2 ± 0.3°C at the end of the procedure. Similarly, in Group Cfresh the outlet temperature increased to 33.8 ± 0.4°C. In the procedures of Groups Bfresh and Dfresh, downstream temperatures hardly differed from those upstream. The reaction of 0.5% isoflurane yielded temperatures no higher than those produced by the absorption of 5% CO2 in either dry or fresh soda lime.

With regard to the weight of the soda lime, we determined an increase in Groups A, B, and C, with the greatest weight increase registered in Group C (Table 2). The passage of pure oxygen had no effect on the weight of the dried soda lime (Group D). In the control experiments, we determined a weight increase only in Groups Afresh and Cfresh. In Groups Bfresh and Dfresh, a loss in weight was recorded. In the procedures involving CO2 (Groups A, C, Afresh, Cfresh), the measured increase in the weight of the soda lime was always lower than the increase calculated to result from the weight of the absorbed CO2 and isoflurane, the difference being statistically significant in Group A.

Table 2

Table 2

Downstream CO2 concentration at the end of the procedure for Group A recovered to 4.5 ± 0.2% after 60 min (Figure 2A). The color of the dry soda lime changed entirely from white to blue, indicating exhaustion of the absorbent. In the procedure for Group C, the CO2 outlet concentration was significantly lower (2.9 ± 0.4%) (Figure 2B), and in these samples, only approximately two-thirds of the soda lime changed color. Accordingly, the calculated CO2 absorption was significantly less in Group A than in Group C (11.9 ± 0.6 L and 12.9 ± 0.4 L, respectively).

Figure 2

Figure 2

In the control procedures, when either 0.5% isoflurane and 5% CO2 (Group Afresh), or 5% CO2 in oxygen (Group Cfresh) were passed through fresh soda lime, the downstream CO2 concentration remained <0.1% for the entire 60 min and no change in color was noted; the delivered amount of CO2 (15 L) was absorbed completely.

Back to Top | Article Outline


In our experiments, we determined the time course of the concentrations of CO formed by the degradation of 0.5% isoflurane in dry soda lime. The course of these concentrations shows that, in the presence of 5% CO2, CO formation is initially higher. At the same time, the total amount of CO formed is lower than without the added CO2, because CO2 shortens the duration of CO production (Figure 1).

Because of the inverse correlation between water content and CO formation in soda lime (10), the lower amount of CO formation in Group A than in Group B may be caused by the production of water secondary to CO2 absorption. The amount of water calculated to be produced in Group A is 8 g (see Table 2) and accounts for approximately 1.3% of weight of the soda lime sample. This water content may be responsible for the markedly lower level of CO production than would be the case with completely dry soda lime (6). The difference between the calculated weight of gas lost in soda lime and the measured weight increase of the soda lime in our experiments may be caused by evaporation and removal of produced water by the carrier gas. Therefore, possibly only part of the water produced contributed to the rehydration of the dry soda lime. Thus, the rehydration effect of CO2 might be too small to explain the decrease in CO formation.

Comparison of the amounts of absorbed CO2 in Group A and Group C experiments shows that the CO2 absorbing capacity of dry soda lime is reduced when CO forms. On the other hand, we have shown that CO formation is reduced during CO2 absorption. This simultaneous reduction in CO formation and CO2 absorption leads us to conclude that the two reactions compete for the same reactant. If both reactions depend on its availability, the two processes should end as soon as the reactant is exhausted. The time courses of the concentration curves for CO and CO2 in Group A show, however, that even on termination of CO formation (after approximately 50 minutes) the absorption capacity of soda lime for CO2 has not been exhausted, because the CO2 continues to be absorbed (up to approximately 50% at that time) (Figures 1A and 2A). Thus, we conclude that CO formation requires the reactant, whereas CO2 absorption continues even in its absence. We believe that this reactant is the alkali hydroxide present in the soda lime. This is supported by results from recent experiments on dried samples of calcium hydroxide; with sodium hydroxide and potassium hydroxide absent, CO production decreased to one tenth (5) or CO formation was not even initiated (12,13) although the CO2 absorbing capacity of the calcium hydroxide was not markedly reduced. The reduction in CO formation in our experiments can be explained by the simultaneous consumption of the alkali hydroxides during CO formation and CO2 absorption.

When the gas mixtures in our experiments contained CO2, the detectable isoflurane loss was lower than in the absence of CO2 and isoflurane concentrations at the outlet reached consistently higher levels and required significantly less time to settle at inlet concentrations (Table 1). Clearly, there was an association between the lower total formation of CO when CO2 was present and the lower total loss of isoflurane.

Both the interaction of isoflurane with dry soda lime and the CO2 absorption in soda lime are exothermic reactions (14,15), which clearly explains the increase in temperature in the outlet flow observed in Groups B and C, as well as in Groups Bfresh and Cfresh. The passage of a mixture of isoflurane and CO2 through dry soda lime (Group A) resulted in slightly higher temperatures, which may have been caused by the initially higher CO formation. However, it cannot be ruled out that the temperature increase, in turn, enhanced CO formation, which is known to be temperature-dependent (6).

In the animal study of Bonome et al. (16), performed under normoventilation and 1 minimum alveolar anesthetic concentration (MAC), maximum CO concentrations were 5500 ± 980 ppm with desflurane, and 800 ppm with isoflurane. Even when each anesthetic was reduced to 0.5 MAC, the concentrations obtained were similar. In this animal study, CO production was doubtlessly affected by exhaled CO2. Because our data (842 ± 81 ppm in Group A) are similar to the results of the Bonome study (16), we assume that our in vitro experiments including CO2 reflect CO production under clinical anesthesia conditions with dry soda lime and 0.5–1 MAC isoflurane. Because the measuring range of our CO sensor is restricted to 0–1000 ppm, we were not able to perform additional experiments by using either larger isoflurane concentrations or the anesthetics enflurane or desflurane, which react more strongly with the dry absorbent (6).

When CO2 was passed through dry soda lime, exhaustion of its capacity to absorb CO2 was noted within the 60-minute exposure period in experimental Groups A and C. Because the entire 15 L of CO2 could not be completely absorbed, the calculated absorption capacity was rated at <2.5 L CO2 per 100 g of soda lime. This corresponds to the results from comparable experiments investigating the absorption capacity of dry soda lime (17). Under optimum humidity conditions, soda lime is known to be at least six times as absorptive as when dry. Thus, there was no sign of exhaustion in our experiments using fresh soda lime (Groups Afresh and Cfresh). As regards the clinical situation, one may assume that soda lime, which has dried unnoticed, is in fact, replaced after short-time use because of its low CO2 absorption capacity (18). This may be the reason for the low number of reports on CO intoxication during anesthetic procedures.

We conclude that low-flow anesthesia with up to 5% CO2 in the gas passing through the absorber will not prevent CO formation from isoflurane in dry soda lime. However, as the presence of CO2 decreases the overall production of CO, the potential for injury from CO is less in the clinical milieu than might be suggested by data from bench models that do not account for the effect of CO2. On the other hand, the peak CO concentrations are higher earlier, suggesting the possibility of a greater danger in the first several minutes of anesthesia. For the timely detection of CO formation in an anesthesia circuit, a CO sensor would be needed, as the major portion of CO is generated before the characteristic changes in temperature and in the capacity of soda lime to absorb CO2.

The authors gratefully acknowledge the expert technical assistance of Mr. Michael Röhrich. We also thank Dr. Friedrich Hammerschmidt, Department of Organic Chemistry, University of Vienna, Vienna, Austria, and Mrs. Jane Neuda for their critical review of the manuscript.

Back to Top | Article Outline


1. Fang ZX, Eger EI. Source of toxic CO explained: -CHF2 anesthetic + dry absorbent. Anesthesia Patient Safety Foundation Newsletter 1994; 9:25–7.
2. Lentz RE. Carbon monoxide poisoning during anesthesia poses puzzles. J Clin Monit 1995; 11:66–7.
3. Moon RE. Cause of CO poisoning: Relation to halogenated agents still not clear. J Clin Monit 1995; 11:67–71.
4. Berry PD, Sessler DI, Larson MD. Severe carbon monoxide poisoning during desflurane anesthesia. Anesthesiology 1999; 90:613–6.
5. 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 dioxide absorption. Anesth Analg 1999; 89:768–73.
6. Fang ZX, Eger EI, Laster MJ, et al. Carbon monoxide production from degradation of desflurane, enflurane, isoflurane, halothane, and sevoflurane by soda lime and baralyme. Anesth Analg 1995; 80:1187–93.
7. Grodin WK, Epstein MA, Epstein RA. Soda lime adsorption of isoflurane and enflurane. Anesthesiology 1985; 62:60–4.
8. Woehlck HJ, Dunning M, Gandhi S, et al. Indirect detection of intraoperative carbon monoxide exposure by mass spectrometry during isoflurane anesthesia. Anesthesiology 1995; 83:213–7.
9. Baxter PJ, Garton K, Kharasch ED. Mechanistic aspects of carbon monoxide formation from volatile anesthetics. Anesthesiology 1998; 89:929–41.
10. Strum DP, Eger EI. The degradation, absorption, and solubility of volatile anesthetics in soda lime depend on water content. Anesth Analg 1994; 78:340–8.
11. Wissenschaftliche Tabellen. Edition Basle. Ciba-Geigy, 1981.
12. Renfrew CW, Murray JM, Fee JH. A new approach to carbon dioxide absorbents. Acta Anaesthesiol Scand 1998; 42(Suppl 112):58–60.
13. Murray JM, Renfrew CW, Bedi A, et al. Amsorb: A new carbon dioxide absorbent for use in anesthetic breathing systems. Anesthesiology 1999; 91:1342–8.
14. Wissing H, Kuhn I, Dudziak R. Heat production from reaction of volatile anesthetics with dried soda lime. Anaesthesist 1997; 46:1064–70.
15. Tsuchiya M, Ueda W. Heat generation as an index of exhaustion of soda lime. Anesth Analg 1989; 68:783–7.
16. Bonome C, Belda J, Alvarez RF, et al. Low-flow anesthesia and reduced animal size increase carboxyhemoglobin levels in swine during desflurane and isoflurane breakdown in dried soda lime. Anesth Analg 1999; 89:909–16.
17. Brown ES, Bakamjian V, Seniff AM. Performance of absorbents: Effects of moisture. Anesthesiology 1959; 20:613–7.
18. Jorgensen B, Jorgensen S. The 600 gram CO2 absorption canister: An experimental study. Acta Anaesthesiol Scand 1977; 21:637–44.
© 2000 International Anesthesia Research Society