Sorptive Loss of Volatile and Gaseous Anesthetics from In Vitro Drug Application Systems

Suzuki, Takahiro MD; Uchida, Ichiro MD, PhD; Mashimo, Takashi MD, PhD

Section Editor(s): Barker, Steven J.

Anesthesia & Analgesia:
doi: 10.1213/01.ANE.0000142112.30807.5C
Technology, Computing, and Simulation: Research Report

In in vitro pharmacological experiments, determination of effective concentration values for various anesthetics depends on understanding the exact concentration of the drugs dissolved in physiological solutions. Actual anesthetic concentration may differ from expectations because of drug adsorption, absorption or other loss, especially in tubing. We tested the hypothesis that delivered concentrations of anesthetics decrease when solutions pass through laboratory tubing and investigated such loss by measuring the entering and exiting dissolved concentrations of two volatile (sevoflurane and isoflurane) and two gaseous (nitrous oxide and xenon) anesthetics. We tested solutions passed through tubes (1 m × 2 mm ID × 4 mm OD) made of five different materials (glass, Teflon, polyethylene (PE), polyvinyl chloride (PVC), and silicon rubber). Exiting concentrations of anesthetics were significantly reduced when they were passed through PVC (>33%) and silicon (>43%) tubes. There were no decreases in anesthetic concentrations with glass, Teflon, or PE tubes. When sevoflurane solution flowed through PVC and silicon tubes, it took 20 and 30 min, respectively, after start of flow until the anesthetic loss became negligible. These results indicate that frequently used PVC and silicon tubes, whereas flexible and easy to handle, have serious drawbacks when used in inhaled anesthetic pharmacology experiments.

In Brief

IMPLICATIONS: In in vitro experimental settings, the delivered anesthetic concentration in solutions can decrease to the extent that one cannot determine exact effective concentration values. We observed the dissolved anesthetic concentrations decreased after passing through polyvinyl chloride and silicon tubes but not through tubes made of glass, Teflon, or polyethylene.

Author Information

Department of Anesthesiology, Osaka University Medical School, Japan

Presented, in part, at the Sixth International Conference on Molecular and Basic Mechanisms of Anesthesia, Bonn, Germany, 28–30 June 2001.

Supported, in part, by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan, and a research grant from Maruishi Pharmaceutical Co, Ltd, of Japan.

Accepted for publication July 28, 2004.

Address correspondence and reprint requests to Takahiro Suzuki, MD, Department of Anesthesiology, Osaka University Medical School, 2-2 Yamadaoka, Suita, Osaka, Japan 565-0871. Address e-mail to

Article Outline

A number of reports on the electrophysiological action of inhaled anesthetics, including gaseous anesthetics, such as xenon (Xe) and nitrous oxide (N2O), have been published. Some of these protocols have used bubbling systems to dissolve inhaled anesthetics in physiologic solutions and have measured the anesthetic concentration in the bubbled solution (1–3). In a previous in vitro electrophysiological study (3), we noticed discrepancies in the concentrations of anesthetics before and after passage through tubing. Some anesthetic loss occurred from solutions flowing through tubes. Although physicians are aware that some anesthetics are soluble in rubber and various plastics (4) and that anesthetics are liable to be lost to adsorption during flow through tubing (5), there have been no reports comparing the effects of tubing materials used in laboratory inhaled anesthetic application systems. It is possible that the anesthetic concentration, during administration, may vary according to the tubing material used and that this variation may influence pharmacological evaluation, including the determination of effective concentration (EC) values.

Here we report how tubing material affects delivered anesthetic concentrations in solution. We evaluated the loss of two volatile anesthetics (sevoflurane and isoflurane) and two gaseous anesthetics (Xe and N2O) in an in vitro experimental system made of glass, Teflon, polyethylene (PE), polyvinyl chloride (PVC), and silicon tubes, which are frequently used for administering anesthetics in the laboratory.

Back to Top | Article Outline


The following anesthetics were used: isoflurane (Abbott Laboratories Ltd, Chicago, IL), sevoflurane (Maruishi Pharmaceutical Co, Ltd, Osaka, Japan), N2O (Teisan Ltd, Tokyo, Japan), and Xe (99.995%) (AirWater Co Ltd, Wakayama, Japan).

The following tubes (each 1 m × 2 mm ID × 4 mm OD) were evaluated: hard glass (Minamirika Glass Co, Ltd, Osaka, Japan), Teflon (Chukoh Chemical Industries, Ltd, Osaka, Japan), PE (Yamaichi Chemical Co, Ltd, Osaka, Japan), PVC As One Co, Osaka, Japan), and silicon (Tigers Polymer Co, Osaka, Japan).

Anesthetic concentrations in solutions were measured by gas chromatography (GC) (Trace™ GC2000; ThermoQuest, CE Instruments, Austin, TX)/mass spectrometry (MS) (GCQ™ plus; ThermoQuest). A sample introduction into the GC was done by headspace sampling using an automatic gas sampler (Combi Pal; CTC Analytics AG, Zwingen, Switzerland). Mass/charge values (m/z) of 51 for isoflurane and sevoflurane, 30 for N2O, and 129 for Xe were used for MS.

A reservoir consisting of a hard glass bottle filled with 500 mL of distilled water was incubated in a water bath (Thermo Minder Mini-80; Taitec Corp, Tokyo, Japan) at 25°C (D641; Takara Thermistor Instruments Co Ltd, Kanagawa, Japan) and continuously bubbled with anesthetic gases. The bubbled gas flow rate was set at 500 mL/min. For volatile anesthetics, air was passed through calibrated vaporizers: Sevotec 3 for sevoflurane (Ohmeda, Steeton, West Yorkshire, United Kingdom) and Forawick for isoflurane (Murako Medical, Tokyo, Japan). The concentrations of sevoflurane and isoflurane were set at 3%, whereas N2O and Xe were set at 100%. The concentrations of the gas mixtures were monitored using a Capnomac Ultima (Datex Instrumentarium Corp, Helsinki, Finland) and Xe meter (Riken, Tokyo, Japan). The solution was drawn by suction with a roller-pump (MasterFlex, Cole-Parmer Instrument Co, Vernon Hills, IL) passed through the test tubing and returned to the reservoir bottle. The flow rate of the solutions was adjusted to 20 mL/min. Twenty minutes after the start of bubbling, before sampling, the anesthetic solutions were allowed to flow through the tubes for 1 min for the initial study and 0, 1, 2, 3, 5, 10, 20, 30, 60, 90, 120, and 180 min for the next study. Control samples were obtained directly from the reservoir bottle, and the test samples were drawn from a three-way metal stopcock placed immediately distal to the test tubing. Fifty-microliter samples of the solutions were drawn into a 50-μL gas tight syringe (Gastight #1805; Hamilton, Reno, NV) and introduced into 20-mL glass vials (Headspace Vial; National Scientific Company Ltd, Jeddah, Kingdom of Saudi Arabia) sealed by an aluminum cap with a Teflon sheet and a rubber septum (Magnetic Cap, National Scientific Company). The vials were then set in the headspace portion of the GC/MS system. To extract most of the volatile and gaseous anesthetics from the liquid phase, the head space temperature was set at 70°C for 15 min, which is sufficiently in excess of the boiling points of sevoflurane (58.5°C) (6) and isoflurane (48.5°C) (7). After that, 1 mL of the gas phase was, using automatic gas sampling equipment, drawn from the headspace of the incubated vial and injected into the injection port of the GC/MS system.

To evaluate loss of the anesthetic after passage through different tubing materials, the relative anesthetic concentration (R) was calculated using the following equation: R = AUCTube/AUCReservoir, where AUCTube is the area under the curve of the peak spectrum obtained by GC/MS for the sample drawn after passage through the tube, and AUCReservoir represents the equivalent data for the sample obtained directly from the reservoir bottle.

To determine statistically significant differences in results for different tubing materials, we used analyses of variance. A probability value of P < 0.05 was considered statistically significant. All data were expressed as mean ± sem.

Back to Top | Article Outline


For each anesthetic, AUCReservoir rapidly increased after bubbling started, reached a plateau within 5 min, and remained stable thereafter (data not shown).

(Figure 1 shows R 1 min after passing through various tubing materials. No loss was apparent when the anesthetic solutions passed through glass, Teflon, or PE tubes. With PVC, there was no significant change in R for Xe, but R significantly decreased to 0.750 ± 0.023 for sevoflurane, 0.670 ± 0.017 for isoflurane, and 0.885 ± 0.003 for N2O. With a silicon tube, R significantly decreased to 0.570 ± 0.020 for sevoflurane, 0.633 ± 0.014 for isoflurane, 0.687 ± 0.013 for N2O, and 0.715 ± 0.019 for Xe. With silicon tubing, the decrease in R for sevoflurane and N2O was more than with PVC tubing.

Because anesthetic concentration significantly decreased 1 min after start of flow at a flow rate of 20 mL/min, we examined how long it would take until the anesthetic concentration reached a plateau. (Figure 2 shows time course change of sevoflurane concentration 0, 1, 2, 3, 5, 10, 20, 30, 60, 90, 120, and 180 min after passing through PVC and silicon tubes (n = 3). The smallest anesthetic concentration was observed at the start of flow (0 min), and then anesthetic concentration gradually increased until 20 min after the start of flow for PVC tube and 30 min for silicon tube.

Back to Top | Article Outline


The possibility of adsorption of many drugs, including anesthetics, by plastic materials has been a source of concern in both basic experimental and clinical settings (8–14). Clinically, anesthetic uptake has been documented in anesthetic circuits composed of rubber, PVC, PE, and polypropylene (8,11,12). In in vitro experimental systems, it is quite likely that the most frequently used plastic tubing material in the experimental apparatus has been PVC. Because some drugs are prone to be adsorbed by PVC, which reduces their delivered concentration, the use of PVC may compromise results when it is used to simulate the administration route for various drugs in solution (4,9,11,15,16). This loss of anesthetic in the application system might be a contributing factor to discrepant EC values found in past experiments. To investigate this issue, we measured and compared the anesthetic concentrations after anesthetic solutions had passed through various tubing materials. There was no reduction of anesthetic concentration when tubes made of glass, Teflon, and PE were used. By contrast, significant declines were found when PVC and, especially, silicon tubes were used.

PVC tubing reduces the concentration of various drugs by adsorption. For example, significant loss of insulin and nitroglycerine via PVC tubes are well known (13,14). Whereas the concentration of lidocaine incubated in a PVC tube continues to decrease for over one hour, there is no similar reduction with a Teflon tube (15). In anesthetic circuits, PVC tubing adsorbed significantly more isoflurane than PE and polypropylene tubing (8). However, in the experiment examining adsorption of liquid sevoflurane and isoflurane in PE and polypropylene tubing, no adsorption was observed during seven days, but some adsorption was observed in PE after 250 days (17). Adsorption after 250 days is, however, insignificant for clinical and in vitro experimental settings. The factor determining the potency of sorption is reported to be polymer density (18): Our findings that the use of higher density PE results in no loss of anesthetic, whereas the use of less dense PVC results in significant loss is in accord with this suggestion. Moreover, lipophilic drugs easily adsorb to PVC and silicon (4), and the more hydrophobic a drug is, the more it is likely to be adsorbed by hydrophobic polymers (19). Drug uptake by polymers is related to the polymer-water partition coefficient of a drug and to its octanol-water partition coefficient (19,20). Therefore, we conclude that PVC is not a suitable conduit in investigations involving inhaled anesthetics. Although adsorption of anesthetics does not occur with glass, because glass is rigid and fragile, it is not an ideal material for tubing in drug application systems. Teflon and PE do not reduce the delivered concentration of anesthetics and have physical properties that make them more amenable for anesthetic application systems.

The first experiment observed up to 43% anesthetic loss in the experimental apparatus and conditions of our protocol. Because among the variables that have been reported to determine the extent and rapidity of adsorption are the solution partition coefficient of the tubing material, the ratio of tubing surface area to solution volume, and flow rate (5), it is possible that complete anesthetic loss may also occur. The degree of adsorption should be strongly dependent on the conditions of delivery. This idea led us to the next experiment. (Figure 2 shows that the adsorption rate is initially rapid but that it gradually decreases for both PVC and silicon tubing. Equilibrium was reached at 20 minutes for PVC tubing and at 30 minutes for silicon tubing in the present study. This result indicates that although the adsorption rate is rapid during the initial period of flow, the loss of anesthetic concentration can be much smaller, or even none at all, if we pause for enough time (e.g., longer than 30 minutes in this study) before sampling. However, immediate sampling or a shorter pause will result in a larger loss. Presumably, had we tested tubes made from adsorptive materials that were longer or of smaller diameter, or used a slower flow rate, the anesthetic concentration after passage through the tube would have been less.

When extracellular solutions that have been freshly bubbled by inhaled anesthetics are introduced to a specimen or a sample, the apparatus used in laboratory investigations of anesthetic action usually involves the passage of anesthetic through long tubes. The longer the contact of the anesthetic solution with the surface of adsorptive tubing, the more likely it is that loss of anesthetic concentration from solution will occur. Moreover, especially in the settings of patch-clamp techniques, it is not easy to precisely determine the anesthetic concentration that is actually administered to a specimen after passage through long tubes, and it is not practical to keep anesthetic solutions flowing for more than 30 minutes before the administration to a sample until anesthetic adsorption reaches equilibrium. These considerations show how important it is to administer anesthetics through inert tubing materials.

In conclusion, using GC/MS, we examined how anesthetic concentration was affected by passing anesthetic solutions through different tubing materials. Our findings show that Teflon and PE are more suitable tubing materials for administering inhaled anesthetics than PVC and silicon. When designing in vitro anesthetic application systems, care should be taken to avoid loss of anesthetic concentration. When determining the EC values of anesthetics during in vitro pharmacological experiments, administration of precise concentrations of anesthetic solutions is a prerequisite.

The authors would like to thank Kazuro Nakano and Yoshiteru Sakamoto, technicians at the Central Laboratory for Research and Education of Osaka University Medical School, for their technical assistance.

Back to Top | Article Outline


1. Yamakura T, Lewohl JM, Harris RA. Differential effects of general anesthetics on G protein-coupled inwardly rectifying and other potassium channels. Anesthesiology 2001;95:144–53.
2. Huneke R, Jungling E, Skasa M, et al. Effects of the anesthetic gases xenon, halothane, and isoflurane on calcium and potassium currents in human atrial cardiomyocytes. Anesthesiology 2001;95:999–1006.
3. Suzuki T, Koyama H, Sugimoto M, et al. The diverse actions of volatile and gaseous anesthetics on human-cloned 5-hydroxytryptamine3 receptors expressed in Xenopus oocytes. Anesthesiology 2002;96:699–704.
4. Eger EI 2nd, Larson CP Jr, Severinghaus JW. The solubility of halothane in rubber, soda lime and various plastics. Anesthesiology 1962;23:356–9.
5. Roberts MS. Modeling solute sorption into plastic tubing during organ perfusion and intravenous infusions. J Pharm Sci 1996;85:655–65.
6. Budavari S. The Merck index. 12th ed. Whitehouse Station, NJ: Merck & Co, Inc, 1996:883.
7. Budavari S. The Merck index. 12th ed. Whitehouse Station, NJ: Merck & Co, Inc, 1996:1457.
8. Smith C, Flynn C, Wardall G, Broome IJ. Leakage and absorption of isoflurane by different types of anaesthetic circuit and monitoring tubing. Anaesthesia 2002;57:686–9.
9. Takakura K, Muramatsu I, Miyamoto E, Fukuda S. Adsorption of lidocaine into a plastic infusion balloon. Anesth Analg 2000;91:192–4.
10. Barann M, Friederich P, Retzmann K, et al. Loss of propofol during in vitro experiments. Anesthesiology 2000;93:310–1.
11. Targ AG, Yasuda N, Eger EI 2nd. Solubility of I-653, sevoflurane, isoflurane, and halothane in plastics and rubber composing a conventional anesthetic circuit. Anesth Analg 1989;69:218–25.
12. Gilly H, Weindlmayr-Goettel M, Koberl G, Steinbereithner K. Anaesthetic uptake and washout characteristics of patient circuit tubing with special regard to current decontamination techniques. Acta Anaesthesiol Scand 1992;36:621–7.
13. Hans P, Paris P, Mathot F. Intravenous nitroglycerin perfusion techniques: clinical implications. Intensive Care Med 1982;8:93–5.
14. Schildt B, Ahlgren T, Berghem L, Wendt Y. Adsorption of insulin by infusion materials. Acta Anaesthesiol Scand 1978;22:556–62.
15. Unger JK, Kuehlein G, Schroers A, et al. Adsorption of xenobiotics to plastic tubing incorporated into dynamic in vitro systems used in pharmacological research: limits and progress. Biomaterials 2001;22:2031–7.
16. Titel JH, Lowe HJ. Rubber-gas partition coefficients. Anesthesiology 1968;29:1215–6.
17. Renfrew CW, Murray JM, Fee JP. A qualitative investigation into the physical stability of polypropylene and polyethylene in liquid isoflurane and sevoflurane. Anaesthesia 2000;55:793–7.
18. Pikal MJ, Bibler DA, Rutherford B. Polymer sorption of nitroglycerin and stability of molded nitroglycerin tablets in unit-dose packaging. J Pharm Sci 1977;66:1293–7.
19. Roberts MS, Kowaluk EA, Polack AE. Prediction of solute sorption by polyvinyl chloride plastic infusion bags. J Pharm Sci 1991;80:449–55.
20. Smith JC, Davies MC, Melia CD, et al. Uptake of drugs by catheters: the influence of the drug molecule on sorption by polyurethane catheters. Biomaterials 1996;17:1469–72.
© 2005 International Anesthesia Research Society