Accumulation of foreign gases during closed-system anaesthesia

British Journal of Anaesthesia 1996; 76: 668 672 Accumulation of foreign gases during closed-system anaesthesia L. VERSICHELEN, G. ROLLY AND H. VERMEULEN Summary In a previous study, accumulation of methane was found at the end of closed-system ventilation. As on-line analysis of gas concentrations is now available, we examined the progressive increase in concentrations of methane, carbon monoxide and acetone during modern, closed-system conditions, and their influence on infrared halothane analysis, in 26 non-pregnant, gynaecological patients. A computer-controlled closed-system anaesthesia apparatus (PhysioFlex) was used for ventilation during total i.v. anaesthesia (excluding nitrous oxide or potent inhalation anaesthetics) for gynaecological laparoscopy. Methane, carbon monoxide and acetone concentrations were analysed every 15 min in a photoacoustic infrared monitor and halothane concentrations by built-in infrared spectrometry. Mean methane concentrations increased progressively after 105 min to 941 (SD 1094) ppm, but concentrations of carbon monoxide and acetone did not increase significantly. In 18 patients, the infrared measurement falsely indicated 0.79 (0.52) % halothane after 60 min, but no reading appeared in the other eight patients. We conclude that methane accumulated progressively under strict closed-system conditions in higher concentrations than reported previously. In two-thirds of patients it induced false halothane readings. (Br. J. Anaesth. 1996; 76: 668 672) Key words Equipment, breathing systems. Gases non-anaesthetic, methane. Gases non-anaesthetic, carbon monoxide. Gases non-anaesthetic, acetone. Measurement techniques, spectrometry. Modern anaesthetic machines permit low-flow or even totally closed-system anaesthesia. It is known that foreign gases accumulate during closed-system anaesthesia, and that methane in particular can be detected in moderate concentrations [1]. Recently, however, we found that methane was present at the end of closed-system ventilation in much greater concentrations than expected [2]. In animals, during total i.v. anaesthesia, expired false halothane concentrations of up to 0.8 % have been recorded during rebreathing [3]. As herbivores partly expel methane via the lungs [4], it might be that methane affects the infrared measurement of halothane. In human, we unexpectedly found that closed-system ventilation during i.v. anaesthesia was associated sometimes with falsely recorded halothane concentrations [2]. As methods for on-line analysis of gases have become available recently, we have examined the progressive increases in concentrations of methane, carbon monoxide and acetone under modern closedsystem conditions, and also the influence of methane on infrared analysis of halothane. Patients and methods Since 1989, we have routinely used a computercontrolled anaesthetic apparatus, PhysioFlex (Physio, The Netherlands) in clinical practice. The apparatus has been described in detail previously [5]. In brief, it is a closed system with a volume of 3.5 litre, wherein small amounts of oxygen, nitrous oxide (or air) are given by built-in, computercontrolled injection; excess gases are evacuated if necessary (fig. 1). The inspiratory oxygen concentration is measured continuously with a paramagnetic oxygen analyser, whereas concentrations of nitrous oxide, carbon dioxide and volatile anaesthetics are measured with a built-in, non-agentspecific infrared spectrometer (Andros, Berkeley, CA, USA). Measurement of halothane, enflurane or isoflurane can be set at an infrared wavelength of 3.3 m. If no particular anaesthetic agent is selected for measurement, halothane sensitivity is chosen automatically in the apparatus used by us. If so chosen, liquid anaesthetic is injected by a computercontrolled syringe. Four-membrane chambers are built into the breathing system for the purpose of controlled ventilation. The apparatus can be used for inhalation anaesthesia or for oxygen air ventilation, as in this study. Recently, apparatus became available for measuring trace concentrations of gases. Using the Brüel and Kjaer multi-gas monitor 1302 (Denmark), trace concentrations are determined by a photoacoustic infrared method. At chosen moments, samples of 140 ml of breathing gases can be obtained for analysis, which is performed with dedicated, narrowband optical filters. In this study, concentrations of methane were measured at a wavelength of 3.6 m, while carbon monoxide and acetone were analysed at 4.7 and 8.5 m, respectively. To avoid interference L. VERSICHELEN, MD, G. ROLLY, MD, PHD, H. VERMEULEN, MD, Department of Anaesthesia, University Hospital, De Pintelaan, 185, B-9000 Ghent, Belgium. Accepted for publication: September 20, 1995. Correspondence to G.R.

Foreign gases and closed-system anaesthesia 669 Figure 1 Direction of gases flowing at a flow rate of 70 litre min 1 in the circle is shown by the four arrows. No unidirectional valves are present. Inflow of oxygen (O2), nitrous oxide (N2O) or compressed air, and eventual injection of liquid halothane (Hal.), enflurane (Enfl.) or isoflurane (Iso.) is shown in the upper part. Excess gases are evacuated by vacuum. The four-membrane chambers are shown to the left. Windows showing carbon dioxide (CO2), N2O and volatile anaesthetics, and oxygen concentrations are shown. At the bottom, the Guardian unit is depicted. at these specific wavelengths, the concomitant use of nitrous oxide and potent inhalation anaesthetics was avoided, and the PhysioFlex apparatus was flushed carefully with oxygen before each use to eliminate any trace of these agents. After 5 min of closed-system use, and every 15 min thereafter, 140-ml samples of gas were obtained directly at the Y-piece of the breathing circuit via a polytetrafluoroethylene tube and fed into the measuring chamber of the multi-gas monitor, while controlled ventilation was maintained continuously with the PhysioFlex apparatus. Although only oxygen and air were used, and no inhalation anaesthetic was given or selected for measurement, the analyser for halothane is always operational and readings are shown at the appropriate window on the screen. This non-invasive assessment was approved by the Ethics Committee of the hospital. Total i.v. anaesthesia (TIVA), with oxygen air ventilation, is used routinely in our department for gynaecological laparoscopy with insufflation of carbon dioxide. TIVA is performed with propofol 1 2.5 mg kg 1, followed by continuous administration at a rate of 10 mg kg 1 h 1, and fentanyl (intermittent 0.1-mg boluses as required clinically). Vecuronium 0.1 mg kg 1 is used for intubation and surgical neuromuscular block. Controlled ventilation with an FI O of 0.4 is used, with minute volumes to 2 maintain end-tidal PCO 2 at 4.0 4.6 kpa. Statistical analysis of the data was done using the following tests: Mann Whitney U-test, Wilcoxon signed rank test and Student s t test, where appropriate, for comparison of variables at different times within a group and at similar times between groups; ANOVA, Wilcoxon test and Friedman test, where appropriate, for detection of variation in time within a group; and two-way ANOVA for comparison between groups. Results We studied 26 ASA I or II, non-pregnant women undergoing routine gynaecological laparoscopy. Mean age, weight and height are shown in table 1, together with the occurrence of false halothane readings. Anaesthesia lasted 60 235 min. A progressive increase in methane concentration was found with time (P 0.002; ANOVA); concentrations of carbon monoxide did not change significantly with time (table 2). A reading for inhalation anaesthetic concentration was present in 18 patients; it did not appear, or occasionally gave the lowest reading possible on the screen, in eight patients. For analysis of the results, patients were therefore allocated to one of two groups according to the presence (group A) or absence (group B) of readings for inhalation anaesthetic values. Concentrations of methane in the 18 patients in group A were significantly higher than in the eight Table 1 Patient characteristics in relation to false halothane readings (mean (SD or range)) Presence of halothane reading (n 18) Absence of halothane reading (n 8) Total group Age (yr) 42.7 (29 78) 39.0 (28 77) 41.6 (28 78) Weight (kg) 62.8 (11.4) 67.6 (10.3) 64.2 (11.0) Height (cm) 165.1 (7.6) 167.0 (9.0) 165.7 (7.9)

670 British Journal of Anaesthesia Table 2 Mean (SD) concentrations of methane, carbon monoxide (CO) and acetone measured at progressive times in the total group of 26 patients (overall statistical difference with time was P 0.002 for methane; ns for CO) 5 min 15 min 30 min 45 min 60 min 75 min (n 20) 90 min (n 13) 105 min (n 11) Methane (ppm) 240 (211) 374 (253) 510 (398) 594 (513) 671 (582) 789 (746) 802 (971) 941 (1094) CO (ppm) 26.0 (30.5) 21.9 (28.6) 20.7 (25.7) 20.4 (23.4) 20.4 (21.8) 21.0 (22.8) 23.7 (27.1) 25.8 (28.3) Acetone (ppm) 31.6 (28.7) 35.8 (35.1) 38.7 (39.9) 38.9 (42.3) 39.4 (42.7) 56.6 (97.0) 69.4 (110.8) 40.2 (33.4) Table 3 Methane and false halothane concentrations in group A (n 18) and methane concentrations in group B (n 8) (mean (SD)). Results up to 60 min only were available from every patient. *P 0.05, **P 0.01 vs 5 min; P 0.05, P 0.01 vs 15 min; P 0.01 vs 30 min 5 min 15 min 30 min 45 min 60 min Group A Methane (ppm) 295 (233) 458 (249)* 664 (385)** 793 (499)** 899 (564)** Halothane (%) 0.14 (0.07) 0.27 (0.19)** 0.46 (0.36)** 0.64 (0.44)** 0.79 (0.52)** Group B Methane (ppm) 116 (52) 185 (140) 162 (74) 145 (52) 158 (22) Table 4 Mean (SD) concentrations of acetone and carbon monoxide (CO) in groups A (n 18) and B (n 8); (no significant difference overall for acetone and CO, both with time and between groups) 5 min 15 min 30 min 45 min 60 min Group A Acetone (ppm) 36.8 (31.1) 39.7 (38.3) 42.5 (42.1) 41.6 (42.8) 41.5 (44.8) CO (ppm) 30.1 (35.6) 25.5 (33.8) 23.8 (30.5) 23.4 (27.7) 23.3 (25.8) Group B Acetone (ppm) 20.0 (19.5) 27.1 (26.5) 30.1 (35.2) 32.9 (43.3) 34.7 (40.0) CO (ppm) 16.9 (9.9) 13.7 (6.1) 13.8 (5.0) 13.5 (3.9) 13.9 (3.5) patients in group B (table 3). In group A, a progressive increase in methane concentration was noticed (P 0.05 or P 0.01; Wilcoxon signed rank and Friedman s test), as was an increase in the reading for inhalation anaesthetic (P 0.01; Wilcoxon signed rank and Friedman s test). A significant (P 0.015) correlation was found between methane (X) and the anaesthetic reading (Y): halothane % (8.5 ε methane (ppm) ε 10 4 ) 0.069 (r 2 86.10 %). In group B, there was no statistically significant difference in methane concentration at different times. Mean values for both carbon monoxide and acetone were higher in group A than group B (table 4). Although no significant changes with time were found for carbon monoxide, concentrations of acetone increased marginally with time. Discussion Methane is present in the alimentary tract of animals and humans. It is produced primarily in the distal colon by fermentation of carbohydrates under strict anaerobic conditions by Methanobacterium ruminatium. Methane is liberated by reduction of carbon dioxide with hydrogen [6]. Intestinal gas in humans contains up to 26 % methane [7]. The main part is excreted in flatus, but a smaller part is taken up in blood. Because of the insolubility of methane in blood relative to air, 90 95 % of this gas is cleared in a single passage through the lungs. There are large individual differences: approximately two-thirds of subjects had methane breath concentrations of less than 1 ppm, and one-third concentrations ranging from 1 to 70 ppm [8]; 80 % of patients with cancer of the large bowel had detectable levels of methane in their breath [9] and 83 % of patients with aorto-iliac disease [10]. Also, patients with extensive ulcerative colitis and colonic polyposis produce high concentrations of methane [11]. Using modern, true closed-system conditions, we have shown that methane not only accumulated in concentrations 10 times higher than reported previously [1], but that the extent of the accumulation was also highly individual. Indeed, two-thirds of our patients had high initial values that increased further with time, whereas one-third had lower values that only moderately and non-significantly increased with time. In one patient a concentration of 4130 ppm was noted, which is more than 15 times the maximum value of 229 ppm found after 72 min by Morita, Latta and Snider [1]. An interesting aspect is that, in our sample of patients, two-thirds were high methane exhalers, whereas in the apparently normal population, twothirds had only traces of methane excretion [8]. Our patients were women, undergoing laparoscopy for gynaecological reasons. None had ulcerative colitis, aorto-iliac disease or cancer of the large bowel. Even the low methane exhalers had pronounced concentrations of gas under closed-system conditions. We cannot explain this but we speculate that laparoscopic pneumoperitoneum induced with carbon dioxide might be a contributing factor. Unfortunately, for technical and practical reasons, baseline measurements of pre-anaesthetic methane excretion

Foreign gases and closed-system anaesthesia 671 could not be made in most of our patients. Allocation of our patients to two groups was made solely on the presence or absence of false halothane readings on the PhysioFlex apparatus. Retrospective analysis of the findings for methane shows that, on average, values greater than 150 175 ppm influenced the infrared measurement. However, in three patients, even lower values induced halothane readings, whereas in four patients, values up to 389 ppm did not consistently activate the infrared sensor. We therefore speculate that methane was not the only factor disturbing the measurement of inhalation concentration. It has been suggested previously that other hydrocarbons such as acetone might interfere with this [3]. Nevertheless, in our high methane exhalers, there was a positive correlation between methane concentration and the false halothane concentration on the screen. The highest recorded value was 3.7 %, coinciding with a methane concentration of 4130 ppm. Acetone produced in the liver during metabolism of free fatty acids is excreted partly by the lungs [12]; during fasting there is increased pulmonary excretion [12]. In our study acetone concentrations were, on average, double those found by Morita, Latta and Snider [1], and tended to increase with time, although this was not significant. It has been reported that flushing the system with fresh gas fails to decrease the acetone concentration [1]; this aspect was not analysed here. Carbon monoxide in alveolar air may be exogenous or endogenous. Cigarette smoke and accidental inhalation of carbon monoxide are examples of the first, whereas endogenously formed carbon monoxide is a by-product of haemoglobin catabolism. Accumulation of carbon monoxide during closedsystem anaesthesia was studied as long ago as 1965; values of 0 210 ppm were reported [13], and twothirds of patients showed an increase in carbon monoxide concentration with time, whereas onethird had no increase or a decrease. In another report, it was stated that concentrations of carbon monoxide decreased in similar circumstances [14]. In this study, values of carbon monoxide of 7.5 164 ppm were found. This last, single, extremely high value influenced greatly the mean in the first group (A). There was no statistically significant difference in carbon monoxide concentrations either between the two groups (A and B) or with time. In three patients an increase and in eight a decrease was seen with closed-system anaesthesia, whereas in 15 patients there was no pronounced change. The clinical implications of progressively increasing concentrations of methane and acetone are not clear. It has been suggested that with concomitant intake of alcohol, they could reach flammable levels [1]. However, personnel in submarines are often exposed to methane and acetone for months without adverse effects. The concentration limits of the US Navy for acetone and methane for 24 h are 2000 and 5000 ppm, respectively, and it has been argued that concerns about these trace contaminants are unfounded [15]. Although concentrations of acetone measured here were lower than those accepted as toxic, in one patient concentrations of methane were close to toxic levels. However, we never flushed the system deliberately, except before use, and therefore methane concentrations were increased by the chosen method. In normal practice it is customary to flush the system at least every 2 h [14]. Concentrations of methane should be significantly lower during everyday clinical practice with closed-system anaesthesia. The erroneous influence on the infrared measurement set at 3.3 m, as used in most inhalation anaesthetic monitors, is disturbing. Our results showed an impressive effect, with a linear relation between methane and the falsely recorded halothane concentration. Blindly following the monitored halothane concentration during closed-system or minimal-flow anaesthesia would be hazardous. The false halothane concentration incites lowering of the vaporizer output, with the inherent danger of too light anaesthesia and the possibility of awareness. In the PhysioFlex apparatus, automatic injection of liquid anaesthetic is computer-controlled in a closed loop according to the measured infrared concentrations. Some years ago the use of a feedback loop for halothane was already being questioned, as too low halothane values were recorded by mass spectrometry [16]. Techniques in which potent inhalation anaesthetics are analysed at higher infrared wavelengths (10 13 m) are not influenced by the presence of methane. These should be preferred during closedsystem anaesthesia, or alternatively other techniques, such as mass spectrometry or Raman scattering. Recently, for the PhysioFlex apparatus, an alternative, second method for measuring inhalation anaesthetics (Guardian Unit: piëzo-electrical sensor) has also been built in, the difference between measurements on the two devices providing evidence for the presence of foreign gases such as methane. In the most recent version of the PhysioFlex apparatus, the infrared readout is switched off automatically when no potent inhalation anaesthetic has been selected and used. References 1. Morita S, Latta W, Snider M. Accumulation of methane, acetone and nitrogen in the inspired gas during closed-circuit anaesthesia. Anesthesia and Analgesia 1985; 64: 343 347. 2. Rolly G, Versichelen L, Mortier E. Methane accumulation during closed circuit anesthesia. Anesthesia and Analgesia 1994; 79: 545 547. 3. Taylor PM. Interference with the Datex Normac anaesthetic agent monitor for halothane in horses and sheep. Journal of the Association of Veterinary Anaesthesia 1990; 17: 32 34. 4. Christensen K, Thorbek G. Methane excretion in the growing pig. British Journal of Nutrition 1987, 57: 355 361. 5. Rendell-Baker L. Future directions in anesthesia apparatus. In: Ehrenwerth J, Eisenkraft JB, eds. Anesthesia Equipment Principles and Applications. Chicago: Mosby-Yearbook Inc, 1993; 691 694. 6. Stadtman TC. Methane fermentation. Annual Review of Microbiology 1967; 21: 121 142. 7. Levitt MD, Bond JH. Volume, composition and source of intestinal gas. Gastroenterology 1970; 59: 921 929. 8. Bond JH, Engel RR, Levitt MD. Factors influencing pulmonary methane excretion in man. Journal of Experimental Medicine 1971; 133: 572 588.

672 British Journal of Anaesthesia 9. Haines A, Metz G, Dilawari J, Blendis L, Wiggins H. Breath methane in patients with cancer of the large bowel. Lancet 1977; ii: 481 483. 10. McKay LF, Brydon WG, Eastwood MA, Honsley E. The influence of peripheral vascular disease on methanogesis in man. Atherosclerosis 1983; 47: 77 81. 11. Piqué JM, Pallares M, Cuso E, Vilar-Bonet J, Gassull MA. Methane production and colon cancer. Gastroenterology 1984; 87: 601 605. 12. Crofford OB, Mallard RE, Winton RE, Rogers NL, Jackson JC, Keller U. Acetone in breath and blood. Transactions of the American Clinical and Climatological Association 1977; 88: 128 139. 13. Middleton V, Van Poznak, A, Artusio JF, Smith SM. Carbon monoxide accumulation in closed circle anesthesia systems. Anesthesiology 1965; 26: 715 719. 14. Lowe HJ, Ernst EA. The Quantitative Practice of Anesthesia Use of Closed Circuit. Baltimore/London: Williams and Wilkins, 1981; 13. 15. Baumgarten RK, Reynolds WJ. Much ado about nothing: Trace gaseous metabolites in the closed circuit. Anesthesia and Analgesia 1985; 64: 1029 1030. 16. Rolly G, Versichelen L, Verkaaik A, Erdmann W, Soens E. Mass spectrometric analysis of a new closed-circuit anaesthesia apparatus (PhysioFlex R ). European Journal of Anaesthesia 1990; 7: 333 334.