Secondary Logo

Journal Logo

Administration of Oxygen Before Tracheal Extubation Worsens Gas Exchange After General Anesthesia in a Pig Model

Loeckinger, Alexander, MD; Kleinsasser, Axel, MD; Keller, Christian, MD; Schaefer, Andreas, MD; Kolbitsch, Christian, MD; Lindner, Karl H., MD; Benzer, Arnulf, MD

doi: 10.1097/00000539-200212000-00057

Administration of 100% oxygen before tracheal extubation is common clinical practice. We determined the effect of this technique on postoperative gas exchange in a porcine model using the multiple inert gas elimination technique. After general anesthesia with mechanical ventilation for a period of 30 min (inspiratory fraction of oxygen of 0.3), anesthesia was discontinued, and the pigs were randomized to an inspiratory fraction of oxygen of 0.3 or 1.0 until they could be safely extubated. Thirty minutes after extubation while breathing air, blood flow to poorly ventilated units had significantly increased in pigs that had been administered 100% oxygen as compared with those receiving 30% oxygen (17% ± 15% versus 7% ± 5%;P = 0.009). We conclude that exposure to 100% oxygen before extubation may cause an undesirable alteration in gas exchange.

Department of Anesthesiology, Critical Care and Emergency Medicine, The Leopold-Franzens University Innsbruck, Austria

April 5, 2002.

Address correspondence and reprint requests to Alexander Loeckinger, MD, Department of Anesthesiology, Critical Care and Emergency Medicine, The Leopold-Franzens University Innsbruck, Anichstrasse 35, 6020 Innsbruck, Austria. Address e-mail to

General anesthesia can adversely affect pulmonary function and gas exchange (1). Atelectasis, which causes pulmonary shunting, is a major cause of this phenomenon (2). The additional administration of oxygen further worsens gas exchange because of an increase in poorly ventilated lung units and the collapse of poorly aerated regions (3,4). This is because in poorly aerated lung units during oxygen breathing, more alveolar gas is absorbed than delivered by inspiratory ventilation. As a consequence, smaller oxygen concentrations may be better with the administration of 100% oxygen limited to patients obviously at risk for hypoxia (5).

Emergence from anesthesia, during which large concentrations of oxygen are still administered before extubation, can be seen as the opposite of anesthesia induction. The effect of this routine practice on gas exchange after extubation has not yet been studied.

The aim of this study was to evaluate pulmonary function after extubation during oxygen breathing. Therefore, we tested the hypothesis that extubation during oxygen breathing alters the distribution of pulmonary perfusion in comparison with extubation breathing 30% oxygen balanced with nitrogen.

Back to Top | Article Outline


This investigation was approved by the Animal Care Committee of the Austrian Ministry of Science. Twenty 12- to 16-wk-old domestic pigs of either sex weighing 35–38 kg were studied during and after general anesthesia. Baseline measurements were taken after 30 min of anesthesia and mechanical ventilation at an inspiratory fraction of oxygen (Fio2) of 0.3. Anesthesia was then discontinued and, depending on the randomization, pigs were administered 100% oxygen (Fio2 = 1.0) or 30% oxygen in air (Fio2 = 0.3) until extubation. When the pig regained sufficient spontaneous respiration, a second measurement was taken, and the pig was subsequently extubated. Thirty minutes after extubation and breathing room air, a third measurement was taken.

Anesthesia was induced with ketamine IM (15 mg/kg) and maintained by continuous infusion of propofol (6 mg · kg−1 · h−1) and remifentanil (0.015 mg · kg−1 · h−1). Lactated Ringer’s solution (4 mL · kg−1 · h−1) was administered continuously throughout the study period. Tracheas were intubated with the pigs in the supine position using endotracheal tubes (inside diameter, 7.5 mm). Pigs were then turned into the prone position, and the lungs were ventilated using an Evita-4 intensive care unit ventilator (Dräger Medical International, Lübeck, Germany) in volume-controlled mode with a tidal volume of 10 mL/kg. Positive end-expiratory pressure was set to 8.1 cm H2O (continuous positive-pressure ventilation). During the preparatory phase, respiratory rate was adjusted to maintain a Paco2 between 35 and 40 mm Hg. Body temperature was maintained between 38.0°C and 39.0°C using an electric heating blanket. When the readings from the ventilator’s flowmeter indicated spontaneous ventilatory effort >3 L/min, continuous positive airway pressure was started at the identical positive end-expiratory pressure value as during continuous positive-pressure ventilation (8.1 cm H2O). When the inspiratory pressure effort during the first tenth of a second of a single inspiration (p 0.1) (6), as measured with the ventilator, indicated a value more than 3.0 mbar, a second measurement was taken, and the pig was subsequently extubated. Thirty minutes after extubation and breathing room air, a third measurement was taken.

Venous catheters were inserted percutaneously into auricular veins for infusion of inert gas and continuous infusion of propofol. A 7F thermistor-tipped flow-directed pulmonary artery catheter was inserted directly into the right internal jugular vein and advanced into a main pulmonary artery using direct-pressure monitoring. This permitted measurements of cardiac output, pulmonary arterial blood pressure, pulmonary arterial occlusion pressure, mixed venous blood sampling, and collection of inert gas blood samples. The left femoral artery was cannulated for measurement of systemic arterial blood pressure, arterial blood gas sampling, and collection of inert gas blood samples. Pressures were measured by use of a intensive care unit-monitor (Servomed, Hellige GmbH, Freiburg, Germany) and standard pressure transducers. Transducers had been zeroed to the level of the right atrium. Measurements were taken immediately before each collection of blood and expired samples for inert gas determination. Cardiac output was measured using the thermodilution technique, and the mean of three subsequent determined values was recorded.

Distributions of ventilation and perfusion were determined by using the multiple inert gas elimination technique (MIGET) in the usual fashion (7–10). This technique’s main goals are calculations of the amounts of blood flow or ventilation to lung areas-perfusion ratio (VA/Q) with zero ventilation (shunt, VA/Q <0.005), limited ventilation (low VA/Q, VA/Q >0.005 through 0.1), and normal ventilation (normal VA/Q, VA/Q >0.1 through 10). Ventilation of unperfused lung areas (alveolar dead space, VA/Q >100) may equally be calculated. The sd of the distribution of perfusion (Log sdQ) reflects the width of this distribution of perfusion (10). The MIGET is based on the assumptions that ventilation and blood flow were continuous and not tidal or pulsatile, respectively, and that all calculated lung units were in a steady state. To perform the MIGET, a mixture of six inert gases, including sulfur hexafluoride, ethane, cyclopropane, halothane, diethyl ether, and acetone, was dissolved in saline and infused via a peripheral vein at a rate of 3 mL/min. These substances have different blood-gas solubilities and, accordingly, a different retention-excretion behavior on the passage through the lung. The infusion was started at least 30 min before the first set of measurements. Ten-milliliter blood samples were collected in heparinized glass syringes from the pulmonary and the femoral artery. Thirty-milliliter mixed expired gas samples were obtained from a heated mixing chamber into warmed gas-tight glass syringes. All samples were kept at a temperature of 38°C and immediately prepared for analysis. Inert gas extraction was performed as described by Wagner et al. (7,8). Concentrations of inert gases were measured by using gas chromatography. The gas chromatograph was equipped with packed columns, a flame ionization detector, and an electron capture detector. The device was exclusively set up and used for the technique. VA/Q distributions were computed from inert gas data by using the 50-compartment model of Evans and Wagner (9). The basis of this calculation is to compare the ratios of retention and excretion of the individual gases in mixed venous and arterial blood as well as in mixed expired air. Also, many other variables are incorporated in this calculation ranging from cardiac output and respiratory minute volume to room and animal temperature and barometric pressure. For examining the irreducible variability (noise) in the data, the residual sum of squares (RSS) was used as an indicator of the goodness of fit of the model to the data (10).

Arterial and mixed venous samples (2 mL) were collected immediately after the collection of each set of inert gas arterial and mixed venous samples and immediately analyzed using a Ciba Corning 806 blood gas analyzer (Ciba Corning, Medfield, MA). All values were corrected to body temperature.

Statistical analysis of hemodynamic, blood gas, inert gas, and respiratory variables was performed using an analysis of variance for repeated measurements. A two-sided test was used. Significant differences were further post hoc examined using the Newman-Keuls test. Data were also tested for normal distribution. All values are presented as mean ± sd. A P value smaller than 0.05 was considered significant.

Back to Top | Article Outline


Pigs administered 100% oxygen after termination of anesthesia could be extubated after 21.3 ± 2.4 min compared with 19.7 ± 3.1 min in the group breathing at a Fio2 of 0.3.

Thirty minutes after extubation, blood flow to lung units with a low VA/Q ratio was significantly larger in pigs breathing 100% oxygen before extubation when compared with pigs receiving 30% oxygen in air (17% ± 15% versus 7% ± 5%;P = 0.009).

Log sdQ, which represents the width of the distribution of perfusion, increased in pigs treated with 100% oxygen (P = 0.0003), indicating that at the time of the second and the third measurement, the distribution of perfusion included blood flow to units with a low VA/Q ratio. A corresponding decrease in blood flow to units with a normal VA/Q ratio was found at the measurement after extubation (P = 0.0002). The distribution of perfusion was left-shifted towards lower VA/Q ratios in pigs extubated on oxygen (Fig. 1). An indication of acceptable quality of the VA/Q distribution is a RSS of 5.3 or less in half of the experimental runs (50th percentile) or 10.6 or less in 90% of the experimental runs (90th percentile) (9). In the present study, 96% of all RSS were less than the 5.3 limit, and no value was above 10.6.

Figure 1

Figure 1

No significant differences in Pao2, alveolar-arterial difference in oxygen tension, and inert gas shunt were found between the two groups at 30 min after extuba- tion.

Time from discontinuing anesthesia to extubation, as well as the other respiratory and circulatory variables including respiratory minute-volume calculated using Fick’s law after tracheal extubation (10), were comparable in both groups.

Back to Top | Article Outline


This study investigated whether the inspiratory gas mixture during emergence of anesthesia altered postoperative pulmonary gas exchange. In pigs treated with a Fio2 = 1.0, the distribution of pulmonary perfusion was left-shifted and became wider when compared with pigs treated with a Fio2 = 0.3 during the period of emergence. Differences in Pao2 did not reach significance.

Insertion and removal of artificial airways are critical moments in terms of hypoxia. Also, general anesthesia per se depresses the oxi-hemoglobin saturation (1). Accordingly, Fio2 = 1.0 during the induction and emergence of anesthesia had been generally considered to be a cornerstone of safety in anesthesia (11). However, Rothen et al. (5) reported that oxygen breathing plays a key role in the formation of atelectasis during anesthesia. This was followed by vigorous debates in the pulmonary community and by the general recommendation that preoxygenation before the induction should only be applied in selected cases (5,12). Despite these findings, the administration of 100% oxygen at the end of general anesthesia, as a measure of precaution to prevent unexpected hypoxemia during the period of extubation, has still not been investigated.

After general anesthesia, four classic factors may contribute to arterial hypoxemia: a decrease in Pao2 may be based on hypoventilation, right-to-left shunt, pulmonary diffusion impairment, or VA/Q mismatch. In our experiment, extubation was performed only after detecting a p.01 > 3 mbar (6), excluding hypoventilation as a possible cause. Also, neither significant increases in right-to-left shunt nor any diffusion limitation was found. Only VA/Q mismatching was left to cause a low Pao2 in our experiment, namely an increase in Log sdQ, a left-shifted distribution of perfusion (decrease in mean of Q), and an increased blood flow to lung areas with a low VA/Q ratio (Table 1).

Table 1

Table 1

At baseline, the examined intubated, ventilated, and supine pigs had a Log sdQ of 0.48. This indicates an unimpaired pulmonary gas exchange at this time point or, put differently, healthy lungs because awake, spontaneously breathing pigs had a Log sdQ of 0.4 (13). Whereas differences in arterial blood gas tensions did not reach significance in our pigs, a different situation arises in animals with a preexisting gas exchange disorder.

In patients with an already increased blood flow to poorly ventilated lung units, as frequently found in chronic obstructive pulmonary disease (14), Log sdQ is increased to 0.74, and further increases in mismatch are likely during oxygen breathing. Low VA/Q alveoli become unstable at larger oxygen concentrations (4,15). In these patients, an already existing population of low VA/Q alveoli is prone to become unstable during oxygen breathing and turn into shunting units. In this particular scenario, oxygen breathing during emergence might mandate supplementary oxygen after general anesthesia—a therapeutic surcharge on the application of Fio2 = 1.0 during emergence?

Several limitations of this study should be noted. First, the absence of collateral ventilation in the pig may have supported the persistence of low VA/Q units in our experiment because without collateral ventilation, low VA/Q units cannot be ventilated by neighboring ones. Second, the last time point was scheduled at 30 minutes after extubation, and we cannot present data on what happened thereafter. However, at the time when this study was designed, the sustained effect in healthy pigs could not be expected. Third, the time span of approximately 20 minutes between termination of anesthesia and extubation seems inadequately long. This apparently resulted from species-specific pharmacokinetics of the otherwise fast-acting drugs applied. However, this time span does not devaluate our findings because Rothen et al. (5) found oxygen-induced atelectasis to appear as early as 3.0 to 3.5 minutes after switching to oxygen. These atelectases did not change considerably thereafter.

In conclusion, oxygen breathing during general anesthesia before extubation results in the development of low VA/Q lung units after extubation in this pig model. According to the findings of our study, the recommendation of Rothen et al. (5) for the anesthesia induction phase, namely to set the oxygen concentration as large as required but as small as possible, may also be a useful guideline for the period of emergence.

Back to Top | Article Outline


1. Nunn JF, Bergman NA, Coleman AJ. Factors influencing the arterial oxygen tension during anaesthesia with artificial ventilation. Br J Anaesth 1965; 37: 898–914.
2. Tokics L, Hedenstierna G, Strandberg A, et al. Lung collapse and gas exchange during general anesthesia: effects of spontaneous breathing, muscle paralysis, and positive end-expiratory pressure. Anesthesiology 1987; 66: 157–67.
3. Reber A, Engberg G, Wegenius G, Hedenstierna G. Lung aeration: the effect of pre-oxygenation and hyperoxygenation during total intravenous anaesthesia. Anaesthesia 1996; 51: 733–7.
4. Wagner PD, Laravuso RB, Uhl RR, West JB. Continuous distributions of ventilation-perfusion ratios in normal animals breathing air and 100 per cent O2. J Clin Invest 1974; 54: 54–68.
5. Rothen HU, Sporre B, Engberg G, et al. Prevention of atelectasis during general anaesthesia. Lancet 1995; 345: 1387–91.
6. Kryger MH, Yacoub O, Anthonisen NR. Effect of inspiratory resistance of occlusion pressure in hypoxia and hypercapnia. Respir Physiol 1975; 24: 241–8.
7. Wagner PD, Naumann PF, Laravuso RB. Simultaneous measurement by eight foreign gases in blood using gas chromatography. J Appl Physiol 1974; 36: 600–5.
8. Wagner PD, Saltzman HA, West JB. Measurement of continuous distributions of ventilation-perfusion ratios: theory. J Appl Physiol 1974; 36: 588–99.
9. Evans JW, Wagner PD. Limits on VA/Q distributions from analysis of experimental inert gas elimination. J Appl Physiol 1977; 42: 889–98.
10. Roca J, Wagner PD. Contribution of multiple inert gas elimination technique to pulmonary medicine. I. Principles and information content of the multiple inert gas elimination technique. Thorax 1994; 49: 815–24.
11. Editorial. Preoxygenation: physiology and practice. Lancet 1992;339:31–2.
12. Rothen HU, Sporre B, Engberg G, et al. Atelectasis and pulmonary shunting during induction of general anaesthesia: can they be avoided? Acta Anaesthesiol Scand 1996; 40: 524–9.
13. Hopkins SR, Stary CM, Falor E, et al. Pulmonary gas exchange during exercise in pigs. J Appl Physiol 1999; 86: 93–100.
14. Wagner PD, Hedenstierna G, Bylin G. Ventilation-perfusion inequality in chronic asthma. Am Rev Respir Dis 1987; 136: 605–12.
15. West JB. New advances in pulmonary gas exchange. Anesth Analg 1975; 5: 409–18.
© 2002 International Anesthesia Research Society