Share this article on:

Increased Carbon Monoxide Concentration in Exhaled Air After Surgery and Anesthesia

Hayashi, Masao MD*; Takahashi, Toru MD†; Morimatsu, Hiroshi MD†; Fujii, Hiromi MD†; Taga, Naoyuki MD†; Mizobuchi, Satoshi MD†; Matsumi, Masaki MD†; Katayama, Hiroshi MD†; Yokoyama, Masataka MD†; Taniguchi, Masahiro MD*; Morita, Kiyoshi MD†

Section Editor(s): Barker, Steven J. Section Editor

doi: 10.1213/01.ANE.0000123821.51802.F3

Heme oxygenase-1 (HO-1) is induced by oxidative stress and is thought to confer protection against oxidative tissue injuries. HO-1 catalyzes the conversion of the heme moiety of hemeproteins, such as hemoglobin, myoglobin, and cytochrome P450, to biliverdin, liberating carbon monoxide (CO) in the process. CO reacts with hemoglobin to form carboxyhemoglobin. In this study, to examine the effect of anesthesia and/or surgery on endogenous CO production, we measured the amount of exhaled CO and the arterial carboxyhemoglobin concentration of patients who underwent surgery under general or spinal anesthesia. Both CO and carboxyhemoglobin concentrations were significantly larger on the day after surgery than during the preoperative period (P <0.05) and in the recovery room (P < 0.05), regardless of anesthesia. However, neither index differed between general and spinal anesthesia. These results suggest that oxidative stress caused by anesthesia and/or surgery may induce HO-1, which catalyzes heme to produce CO, leading to increased exhaled CO concentration.

IMPLICATIONS: We measured exhaled carbon monoxide (CO) concentrations and arterial carboxyhemoglobin concentrations in patients who underwent surgery under general or spinal anesthesia. Both indices were significantly larger than the preoperative value on the day after surgery and in the recovery room, regardless of anesthesia, suggesting that endogenous CO production is increased by generalized oxidative stress.

* Department of Anesthesiology, National Okayama Medical Center, Tamasu, Okayama, Japan; and † Department of Anesthesiology and Resuscitology, Okayama University Medical School, Okayama, Japan

Supported in part by a Grant-in-Aid for Scientific Research (15659427) from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Accepted for publication February 5, 2004.

Address correspondence and reprint requests to Toru Takahashi, MD, Department of Anesthesiology and Resuscitology, Okayama University Medical School, 700-8558, Japan. Address e-mail to

Humans mainly produce endogenous carbon monoxide (CO) through heme catabolism (1). Heme oxygenase-1 (HO-1), the rate-limiting enzyme in heme catabolism, is induced not only by its substrate heme, but also by various types of oxidative stress (2–5). CO synthesized through the HO reaction diffuses out of cells, enters the blood to form carboxy-hemoglobin, and is transported to the lungs, where it is excreted in ambient air (6). Exhaled CO concentrations are increased in patients with inflammation of the respiratory tract (7,8) and in those with systemic inflammation (9). However, the effect of surgery and/or anesthesia on exhaled CO concentrations has not yet been investigated. CO is generated by the interaction of volatile anesthetics with soda lime (10,11). With this point in mind, we examined the amount of exhaled CO and the arterial carboxyhemoglobin concentrations of patients who underwent surgery under general anesthesia with propofol and fentanyl under an oxygen/nitrous oxide mixture or under spinal anesthesia.

Back to Top | Article Outline


The IRB approved the study protocol, and we obtained written, informed consent from each patient. A total of 46 patients aged from 15 to 65 yr were included in this study. We studied 25 consecutive ASA status I–II patients who underwent general anesthesia and 21 consecutive ASA status I–II patients who underwent spinal anesthesia (Table 1). We excluded current smokers. Patients were premedicated with oral midazolam 0.1 mg/kg. General anesthesia was induced with propofol 1.5–2 mg/kg and fentanyl 2–4 μg/kg with a 50% oxygen/50% nitrous oxide mixture. Vecuronium (0.1 mg/kg) facilitated orotracheal intubation. During surgery, anesthesia was maintained with propofol and fentanyl under the oxygen/nitrous oxide mixture. A semiclosed circuit equipped with soda lime was used for all general anesthesia. No volatile anesthetics were used, because soda lime can degrade volatile anesthetics to CO (10,11). After recovery from anesthesia, all patients were tracheally extubated and sent to the recovery room according to the general practices of our institute. Oxygen (4.0 L/min) was administered to all patients under observation in the recovery room. Patients were returned to the ward 2 or 3 h later. If needed, diclofenac or indomethacin suppositories were administered to control pain at the discretion of attending anesthesiologists. For the spinal anesthesia group, bupivacaine (0.5%; 2–2.5 mL) was administered for all spinal anesthesia. Except for oxygen administration, the postoperative courses and pain control policies were identical to those of the general anesthesia group.

Exhaled CO concentrations were measured by using a portable breath CO monitor (EC50 MICRO III Smokerlyzer™; Bedfont Technical Instruments Ltd., Sittingbourne, UK) as described by Jarvis et al. (12). Before the start of the study, the analyzer, with a sensitivity of 1 ppm, was calibrated with a mixture of 50 ppm CO in nitrogen gas. Patients were asked to exhale fully, inhale deeply, and hold their breath for 20 s before exhaling rapidly into a disposable mouthpiece. This procedure was repeated 3 times, with 1 min of normal breathing between repetitions, and the mean value was used for analysis. We measured the CO concentrations exhaled by all patients (1) at preoperative ward rounds (on the day before surgery) (2), in the recovery room (immediately after surgery), and (3) at postoperative ward rounds (the day after surgery). At the same time, we sampled arterial blood from the femoral artery and determined the carboxyhemoglobin concentration by using a cooximeter blood gas analyzer (Rapidlab 860™; Bayer, Boston, MA).

Summary descriptive statistics are presented as medians with an interquartile range. Continuous variables were compared by using the Mann-Whitney U-test or Wilcoxon’s signed rank test, as appropriate. Categorical variables were compared by using Fisher’s exact test. Friedman’s test (nonparametric analysis of variance) determined significant changes in the CO and carboxyhemoglobin concentrations over time. Such changes were compared post hoc by using Wilcoxon’s signed rank test adjusted for multiple comparisons. The StatView™ (Abacus Concepts Inc., Berkeley, CA) package performed all statistical analyses. P < 0.05 was considered statistically significant.

Back to Top | Article Outline


Although the median age of the general and spinal anesthesia groups did not significantly differ, the former contained fewer women than the latter (Table 1). The durations of the operation and the anesthesia were significantly longer in the group given general, rather than spinal, anesthesia (Table 1).

We first examined differences between the medians of the exhaled CO concentrations. The preoperative exhaled CO concentrations were extremely small (Fig. 1). In contrast, the concentrations on the day after surgery were significantly larger than the preoperative value and the recovery room value for both groups, although the preoperative concentration and that in the recovery room did not significantly differ (Fig. 1). Comparing the general anesthesia group with the spinal anesthesia group at the same time point, we could not find any differences in exhaled CO concentrations for the three measurement points (Fig. 1).

Next we examined differences between medians of the arterial carboxyhemoglobin concentrations. The arterial carboxyhemoglobin concentration on the day after surgery was also significantly larger than that before the operation and in the recovery room in both groups (Fig. 2). The arterial carboxyhemoglobin concentration of the general anesthesia group in the recovery room was significantly smaller than the preoperative value. In contrast, these two time points did not differ in the spinal anesthesia group (Fig. 2). The arterial carboxyhemoglobin concentration was significantly smaller in the general anesthesia group in the recovery room, although values before surgery and on the day after surgery did not significantly differ.

Finally, we examined individual differences in preoperative and postoperative CO and carboxyhemoglobin concentrations. Among 25 general anesthesia patients, 18 (72%) patients had increased CO concentrations on the first postoperative day (median increase, 1.5 ppm; inter-quartile range, 0–2.0 ppm; P = 0.0003) (Fig. 3A). Similarly, 11 (44%) patients had increased postoperative carboxyhemoglobin concentrations (median increase, 0.1%; interquartile range, 0%–0.4%; P = 0.05) (Fig. 3B). Among21 spinal anesthesia patients, 15 (71%) had increased CO concentrations on the first postoperative day (median increase, 1.0 ppm; interquartile range, 0–1.0 ppm; P = 0.0008) (Fig. 4A). Again, 17 (81%) patients had increased postoperative carboxyhemoglobin concentrations (median increase, 0.5%; interquartile range, 0.08%–0.53%; P = 0.0015) (Fig. 4B).

Back to Top | Article Outline


We demonstrated for the first time that exhaled CO and arterial carboxyhemoglobin concentrations were increased on the day after surgery in patients who underwent anesthesia and surgery as compared with preoperative values, although neither index differed between those who underwent general or spinal anesthesia. Exhaled CO levels are increased not only during inflammation of the respiratory tract (7,8), but also during systemic inflammatory conditions such as sepsis (9). Tracheal intubation and artificial ventilation associated with general anesthesia may cause an inflammatory response leading to oxidative stress, which elicits HO-1 induction in the respiratory tract. However, this study found that exhaled CO concentrations were increased on the day after surgery regardless of anesthesia. Consistent with the increase in exhaled CO concentration, arterial carboxyhemoglobin levels also increased in both groups. Although the carboxyhemoglobin concentration was decreased in the recovery room after general anesthesia, this may be attributable to the large concentration of inspired oxygen during surgery and in the recovery room. Endogenous CO is mainly derived from heme catabolism in the HO reaction (1). We demonstrated that HO-1 is induced in the lung as well as in the kidney, the liver, the intestine, and the brain in a rat model of sepsis (13–15). Taken together, we speculate that systemic inflammation caused by surgery and/or anesthesia elicits oxidative stress, which induces systemic HO-1 induction and leads to the production of endogenous CO. Synthesized CO forms carboxyhemoglobin, which is transported to the lung, from which it is excreted, leading to an increased CO concentration in exhaled air.

Excessive systemic inflammation of sepsis leads to a series of complex cascades, ultimately resulting in multiple organ dysfunction syndrome (16), the most common cause of death in intensive care units (17). In septic organ damage, reactive oxygen species are thought to play an important role as an end-effector molecule (18,19). Very recently, it has been proposed that antioxidants, such as N-acetylcysteine, might be a new modality for the treatment of severe sepsis (20–22). However, there is no easily measured marker to monitor antioxidant therapy. Because HO-1 is induced by oxidative stress in sepsis (4,13–15), exhaled CO concentration may, in the future, act as a clinical indicator to evaluate the efficacy of the antioxidant therapy for sepsis.

In summary, our findings suggest that exhaled CO concentration may be a novel noninvasive marker of oxidative stress elicited not only by surgery and/or anesthesia, but also by sepsis and systemic inflammation.

We are grateful to Dr. Shigeru Sassa (The Rockefeller University, NY), for his helpful suggestions and kind support throughout this work. We are also grateful to Dr. Reiko Akagi (Okayama Prefectural University, Soja, Japan) for her encouragement to perform this work.

Back to Top | Article Outline


1. Marilena G. New physiological importance of two classic residual products: carbon monoxide and bilirubin. Biochem Mol Med 1997;61:136–42.
2. Shibahara S. Regulation of heme oxygenase gene expression. Semin Hematol 1988;25:370–6.
3. Takahashi T, Akagi R, Shimizu H, et al. Heme oxygenase-1: a major player in the defense against the oxidative tissue injury. In: Abraham NG, Alam J, Nath KA, eds. Heme oxygenase in biology and medicine. New York: Kluwer Academic/Plenum, 2002:387–98.
4. Otterbein LE, Choi AM. Heme oxygenase: colors of defense against cellular stress. Am J Physiol Lung Cell Mol Physiol 2000;279:L1029–37.
5. Akagi R, Takahashi T, Sassa S. Fundamental role of heme oxygenase in the protection against ischemic acute renal failure. Jpn J Pharmacol 2002;88:127–32.
6. Coburn RF, Forster RE, Kane PB. Considerations of the physiological variables that determine the blood carboxyhemoglobin concentration in man. J Clin Invest 1965;44:1899–910.
7. Horvath I, MacNee W, Kelly FJ, et al. “Haemoxygenase-1 induction and exhaled markers of oxidative stress in lung diseases,” summary of the ERS Research Seminar in Budapest, Hungary, September, 1999. Eur Respir J 2001;18:420–30.
8. Yamaya M, Sekizawa K, Ishizuka S, et al. Increased carbon monoxide in exhaled air of subjects with upper respiratory tract infections. Am J Respir Crit Care Med 1998;158:311–4.
9. Zegdi R, Perrin D, Burdin M, et al. Increased endogenous carbon monoxide production in severe sepsis. Intensive Care Med 2002;28:793–6.
10. Baum J, Sachs G, Driesch C vd, Stanke HG. Carbon monoxide generation in carbon dioxide absorbents. Anesth Analg 1995;81:144–6.
11. Fang ZX, Eger EI II, 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.
12. Jarvis MJ, Russell MA, Saloojee Y. Expired air carbon monoxide: a simple breath test of tobacco smoke intake. BMJ 1980;281:484–5.
13. Suzuki T, Takahashi T, Yamasaki A, et al. Tissue-specific gene expression of heme oxygenase-1 (HO-1) and non-specific delta-aminolevulinate synthase (ALAS-N) in a rat model of septic multiple organ dysfunction syndrome. Biochem Pharmacol 2000;60:275–83.
14. Fujiwara T, Takahashi T, Suzuki T, et al. Differential induction of brain heme oxygenase-1 and heat shock protein 70 mRNA in sepsis. Res Commun Mol Pathol Pharmacol 1999;105:55–66.
15. Fujii H, Takahashi T, Nakahira K, et al. Protective role of heme oxygenase-1 in the intestinal tissue injury in an experimental model of sepsis. Crit Care Med 2003;31:893–902.
16. Cohen J. The immunopathogenesis of sepsis. Nature 2002;420:885–91.
17. Johnson D, Mayers I. Multiple organ dysfunction syndrome: a narrative review. Can J Anaesth 2001;48:502–9.
18. Macdonald J, Galley HF, Webster NR. Oxidative stress and gene expression in sepsis. Br J Anaesth 2003;90:221–32.
19. Motoyama T, Okamoto K, Kukita I, et al. Possible role of increased oxidant stress in multiple organ failure after systemic inflammatory response syndrome. Crit Care Med 2003;31:1048–52.
20. Paterson RL, Galley HF, Webster NR. The effect of N-acetyl-cysteine on nuclear factor-kappa B activation, interleukin-6, interleukin-8, and intercellular adhesion molecule-1 expression in patients with sepsis. Crit Care Med 2003;31:2574–8.
21. Albuszies G, Bruckner UB. Antioxidant therapy in sepsis. Intensive Care Med 2003;29:1632–6.
22. Pinsky MR. Antioxidant therapy for severe sepsis: promise and perspective. Crit Care Med 2003;31:2697–8.
© 2004 International Anesthesia Research Society