During nitrous oxide (N2O) elimination, arterial oxygen tension (PaO2) decreases because of the phenomenon commonly called diffusive hypoxia. The authors questioned whether similar effects occur during xenon elimination.


Nineteen anesthetized and paralyzed pigs were mechanically ventilated randomly for 30 min using inspiratory gas mixtures of 30% oxygen and either 70% N2O or xenon. The inspiratory gas was replaced by a mixture of 70% nitrogen and 30% oxygen. PaO2 and carbon dioxide tensions were recorded continuously using an indwelling arterial sensor.


The PaO2 decreased from 119+/-10 mm Hg to 102+/-12 mm Hg (mean+/-SD) during N2O washout (P<0.01) and from 116+/-9 mm Hg to 110+/-8 mm Hg during xenon elimination (P<0.01), with a significant difference (P<0.01) between baseline and minimum PaO2 values (deltaPaO2, 17+/-6 mm Hg during N2O washout and 6+/-3 mm Hg during xenon washout). The PaCO2 value also decreased (from 39.3+/-6.3 mm Hg to 37.6+/-5.8 mm Hg) during N2O washout (P<0.01) and during xenon elimination (from 35.4+/-1.6 mm Hg to 34.9+/-1.6 mm Hg; P< 0.01). The deltaPaCO2 was 1.7+/-0.9 mm Hg in the N2O group and 0.5+/-0.3 mm Hg in the xenon group (P<0.01).


Diffusive hypoxia is unlikely to occur during recovery from xenon anesthesia, probably because of the low blood solubility of this gas.

BECAUSE the exchange of inert gases across the alveolar-capillary membrane can be described by Equation 1the transfer ratio of two different inert gases x and y (x/y) equilibrating simultaneously in the lung is expressed as Equation 2where M = transfer rate, Pa= arterial and P (v)= venous partial pressure, Q = blood flow, and [small lambda, Greek]= the blood-gas partition coefficient.

Therefore, if one gas is washed out and the other is washed in, the inert gas volumes moving between the capillary blood and the alveolar space will differ with varying gas solubilities, which leads to dilution or concentration of the alveolar gas. Diffusive hypoxia, which occurs during recovery from nitrous oxide (N2O) anesthesia, is a manifestation of this phenomenon: Because the amount of the higher soluble N2O ([small lambda, Greek]N(2) O = 0.47) released into the alveolus exceeds that of nitrogen simultaneously taken up ([small lambda, Greek]N(2)= 0.015), alveolar gas is diluted, and the pressures of oxygen (PaO(2) and carbon dioxide PaCO(2) in arterial gas decrease. Using xenon instead of N2O, similar but smaller effects on blood gases may be expected because [small lambda, Greek]xenon(= 0.115)1is closer to [small lambda, Greek]nitrogen. Therefore, the aim of the current study was to compare the time course of PaO(2) and PaCO(2) during N2O and xenon washout, respectively, in anesthetized and ventilated pigs to assess the likelihood of diffusive hypoxia after xenon anesthesia. To compare the measured values with theoretical assumptions, the expected time course of PaO(2) and PaCO(2) during N2O and xenon elimination was derived using a computer simulation.

Materials and Methods

Animal Preparation

The study design was approved by the Federal Animal Care Committee. Nineteen pigs of either sex (body weight, 41 +/− 5 kg; mean +/− SD) randomly assigned to receive N2O (n = 10) or xenon (n = 9) were anesthetized with pentobarbital sodium (Nembutal, Sanofi-Wintrop, Munich, Germany; a 15-mg/kg induction dose, followed by a continuous infusion of 6-12 mg [middle dot] kg-1[middle dot] h-1). In previous experiments, this pentobarbital infusion rate allowed full anesthesia without the need for additional muscle relaxation. [2]In addition, intermittent 0.3-mg intravenous boluses of buphrenorphine (Temgesic; Boehringer-Mannheim, Mannheim, Germany) were added every fourth hour and before any surgical or noxious stimuli. The depth of anesthesia was assessed using hemodynamic variables and continuous electroencepholographic monitoring (Neurotrac; Interspec, Cronshohocken, PA). The spectral edge frequency was always less than 15 Hz, and the median power frequency was 5-10 Hz. During the study period, the animals were paralyzed with alcuronium dichloride (Alloferin; Hoffmann-LaRoche, Basel, Switzerland; 0.25 mg/kg initial dose followed by a continuous application of 0.25 mg [middle dot] kg-1[middle dot] h-1) to ensure constant flow ventilation. Immediately after anesthesia was induced, the pigs were intubated with a cuffed endotracheal tube (8.5 mm ID), and their lungs were mechanically ventilated in the supine position with a standard semiclosed circuit anesthesia machine (Cicero; Dragerwerk AG, Lubeck, Germany) modified for xenon application. Ventilatory settings throughout the experiment were tidal volume (VT)= 12 to 14 ml/kg (adjusted to achieve a PaCO(2) of 37-43 mmHg), ventilator frequency (f)= 12/min, inspiratory time (Ti)= 1.5 s, inspiration hold = 1 s, expiratory time (Te)= 2.5 s, and positive end-expiratory pressure = 5 cm water. The inspiratory oxygen fraction (FIO(2)) of 0.3 was maintained regardless of the gas mixture used (N2O and oxygen, xenon and oxygen, or nitrogen and oxygen) and measured continuously in the inspiratory limb of the ventilator circuit by the machine-integrated oxygen monitor (a fuel-cell sensor with an accuracy of +/− 2%) calibrated before each experiment. Fresh gas supply of the anesthesia circuit was set to 50% of the minute ventilation and doubled during the measurement periods to minimize the delay in inspiratory gas composition change caused by the volume of the ventilator circuit.

A central venous catheter and a thermistor-tipped pulmonary artery flotation catheter (model 93A 754 7F; Baxter Healthcare, Irvine, CA) were placed through the right jugularis externa vein into the superior vena cava and into a pulmonary artery, respectively, for drug infusion, monitoring, and data sampling of hemodynamics and gas exchange. A 4-French catheter was inserted into a femoral artery for arterial blood sampling and placement of the Paratrend 7 sensor (Biomedical Sensors Ltd., High Wycombe, UK) used for continuous PaO(2) and PaCO(2) registration. This sensor measures pH and PaCO(2) by an optical method and PaO(2) by the electrochemical method using a Clark-type sensor. Automated in vitro calibration (using three different gas mixtures with 2% carbon dioxide + 15% oxygen in nitrogen, 5% carbon dioxide + 15% oxygen in nitrogen, and 10% carbon dioxide + 15% oxygen in nitrogen) was performed by the system before each insertion. The accuracy of the Paratrend 7 sensor for PaO(2) is +/− 5% at values less than 120 mmHg and +/− 10% at values > 120 mmHg and +/− 3 mmHg for PaCO(2). Immediately after placement of the sensor, arterial blood gas was sampled and analyzed for PaO(2) and PaCO(2) with an IL 1306 blood gas analyzer (Instrumentation Laboratory, Lexington, MA) for additional in vivo calibration of the Paratrend 7 system.

Study Design and Measurements

After the surgical instrumentation and a 1-h recovery period, the fresh gas flow was augmented to equal minute ventilation, and the inspiratory nitrogen-oxygen for 30 min to ensure a complete nitrogen washout. Then we switched back to nitrogen-oxygen to wash out N2O or xenon while maintaining an FIO(2) of 30% and a fresh gas supply equal to 1-min ventilation. Recording of arterial blood gases at a registration frequency of 1 value/min began approximately 15 min before starting the washout to document plateau partial pressures for each gas as reference values for the subsequent washout phase and continued during a period of 45 min. Immediately before and 15 min after starting the washout, arterial and mixed venous blood samples were obtained to measure arterial and mixed venous total hemoglobin concentration and oxygen saturation (Co-Oximeter IL 282 calibrated for pig blood; Instrumentation Laboratory). To analyze putative differences in anesthesia depth during the wash-in and washout of N2O and xenon, which might have affected systemic hemodynamics and consequently gas exchange, heart rate, mean arterial pressure, central venous pressure, mean pulmonary artery pressure, and pulmonary artery occlusion pressure were measured simultaneously with cardiac output. The cardiac output values recorded are expressed as the mean of triplicate injections of 10 ml ice-cold saline randomly spread over the respiratory cycle. [3] 

Theoretical Calculations

Theoretical calculations were made using a computer model described in more detail in appendix 1. Briefly, the model calculates the alveolar partial pressures of each gas before and after inspiration and subsequently the gas uptake into or release from the blood, assuming a complete partial pressure equilibration for all gases between the alveolar space and the capillary blood. This program allowed us to mimic the washout experiment by theoretically calculating the time course of PaO(2) and Pa (CO)(2) during elimination of N2O and xenon, respectively, assuming a given set of values for tidal volume; respiratory rate; fresh gas supply; functional residual capacity; and oxygen, carbon dioxide, and inert gas partial pressures in the inspiratory and alveolar gas mixture, blood, and peripheral tissue. All theoretical calculations presented in this study were performed assuming the same conditions as during the experiments; that is, the inspiratory gas composition was 30% oxygen and 70% nitrogen, and the N (2) O or xenon wash-in was assumed to be complete before starting the simulated washout. To mimic the gradual change of inspiratory gas composition imposed by the semiclosed anesthesia circuit during the initial phase of washout, the inspired fractions of N2O and xenon, respectively, were assumed to be replaced stepwise by nitrogen within the first 20 breaths of washout.

Statistical Analysis

All data are presented as the mean +/− SD, which were calculated each minute for PaO(2) and PaCO(2) in the two groups. Baseline and minimum values of each group were compared using the Wilcoxon signed rank test. Subsequently, we compared the minimum PaO(2) and PaCO(2) values and the differences between the baseline and minimum PaO(2) and PaCO(2) of both groups (Delta PaO(2) and Delta PaCO(2)) using the Mann-Whitney U test. A P value < 0.05 was considered significant.



Replacing the anesthetic gas with nitrogen did not significantly affect systemic or pulmonary hemodynamics, regardless of the inhaled gas.

Blood Gases

All data presented in this section are mean values +/− SD. Figure 1shows the PaO(2) values. The PaO(2) value decreased from 119 +/− 10 mmHg to a minimum of 102 +/− 12 mmHg after 4 min of N2O washout (P < 0.01) and from 116 +/− 9 mmHg to a minimum of 110 +/− 8 mmHg (P < 0.01) after 4 min of xenon washout. The difference between baseline and minimum PaO(2)(Delta PaO(2), 17 +/− 6 mmHg in the N2O group and 6 +/− 3 mmHg in the xenon group) was significantly less in the xenon group (P < 0.01). In contrast, the minimum PaO(2) values did not differ significantly between the xenon and the N2O group. At the end of the washout period, PaO(2) was 107 +/− 12 mmHg in the N2O group and was 111 +/− 7 mmHg in the xenon group. The PaO(2) data obtained by the computer simulation are presented in the lower panel of Figure 1. The calculated baseline PaO(2) values were 119 and 115 mmHg, and the minimum values were 100 and 108 mmHg for N2O and xenon washout, respectively.

Figure 1. (Upper) The time course of arterial oxygen tension (PaO(2)) during nitrous oxide (closed triangles) and xenon (closed circles) washout. (Lower) The predicted time course of PaO(2) as calculated by the computer simulation. All variables requested by the program to perform the calculations (ventilatory pattern; functional residual capacity; initial alveolar, arterial, and peripheral tissue partial pressures of oxygen, carbon dioxide, nitrogen, and anesthetic gas; peripheral tissue volume and perfusion; and cardiac output) were set to identical values for both gases, except for the blood-gas partition coefficients, which were set to 0.47 for nitrous oxide and 0.115 for xenon. Solubility in the peripheral tissue compartment was assumed to be equal to the blood-gas partition coefficient.

Figure 1. (Upper) The time course of arterial oxygen tension (PaO(2)) during nitrous oxide (closed triangles) and xenon (closed circles) washout. (Lower) The predicted time course of PaO(2) as calculated by the computer simulation. All variables requested by the program to perform the calculations (ventilatory pattern; functional residual capacity; initial alveolar, arterial, and peripheral tissue partial pressures of oxygen, carbon dioxide, nitrogen, and anesthetic gas; peripheral tissue volume and perfusion; and cardiac output) were set to identical values for both gases, except for the blood-gas partition coefficients, which were set to 0.47 for nitrous oxide and 0.115 for xenon. Solubility in the peripheral tissue compartment was assumed to be equal to the blood-gas partition coefficient.

The PaCO(2) value decreased from 39.3 +/− 6.3 mmHg to 37.6 +/− 5.8 mmHg after 7 min of N2O washout (P < 0.01) and from 35.4 +/− 1.6 mmHg to 34.9 +/− 1.6 mmHg after 8 min of xenon washout (P < 0.01), with a significantly higher Delta PaCO(2) value in the N2O group (1.7 +/− 0.9 mmHg in the N2O group and 0.5 +/− 0.3 in the xenon group; P < 0.01). The PaCO(2) values calculated by the computer program were 43.0 and 42.4 mmHg as baseline values, and 40.2 and 41.5 mmHg as minimum values for N2O and xenon washout, respectively.


The purpose of the current study was to analyze the effect of xenon elimination on PaO(2) and PaCO(2) to evaluate the risk of diffusive hypoxia when using xenon instead of N2O as an inhalational anesthetic agent. The primary result is that changes in PaO(2) and Pa (CO)(2) occurred after washout of both gases but were approximately three times greater after N2O.

When compared with previous data of gas exchange during N2O elimination, we found Delta PaO(2) values that were similar to those reported by Sheffer et al. [4](Delta PaO(2)= 15.1 mmHg) and Sugioka et al. [5](16.6 mmHg). Changes in PaCO(2) during N2O elimination were measured previously by Sheffer et al. [4](maximum Delta PaCO(2)= 7.8 mmHg). The marked difference between the results of these two groups was probably caused by the fact that the former measured the Pa (CO)(2) in spontaneously breathing patients in whom minute ventilation increased during the phase of N2O elimination, whereas the latter studied patients during controlled ventilation. Variable minute ventilation, of course, modulates the effects on PaCO(2) during N2O washout. In contrast, the Delta PaCO(2) value of 2 mmHg measured by Rackow et al. [6]using constant minute ventilation corresponds with our data (peak Delta PaCO(2)= 2 mmHg).

Although rules for inert gas elimination and their effect on arterial oxygenation have been derived previously from theoretical models published by Farhi and Olszowka [7]and Scrimshire and Tomlin, [8]until now, the effects on PaO(2) and PaCO(2) have been analyzed in vivo only during N2O elimination. Farhi and Olszowka [7]concluded from their theoretical model that PaO(2) changes during washout or wash-in of inert gases replacing nitrogen are directly related to their blood solubility. Their calculations, when applied to xenon, support our results because the ratio [small lambda, Greek]N(2) O/[small lambda, Greek]xenon= 4.09. The importance of inert gas blood solubility for the estimation of alveolar gas dilution is further underscored by our theoretical calculations, which (assuming [small lambda, Greek]N(2) O = 0.47 and [small lambda, Greek]xenon= 0.115) yield PaO(2) and PaCO(2) and Delta PaO(2) and Delta PaCO(2) values that correspond fairly well with the measured data. Thus, our results further confirm the theoretical considerations that inert gas exchange is largely determined by blood solubility for xenon as well.

In principle, three major properties of inert gases influence alveolar- capillary gas exchange and intrapulmonary gas transport: density, molecular weight, and blood solubility. In contrast to N2O, not only blood solubility of xenon, but also gas density and molecular weight, significantly differ when compared with nitrogen. Therefore, predictions regarding xenon washout are less certain than for N2O washout, because intrapulmonary gas transport, which is largely governed by convective and diffusive forces, [9]is affected by both the molecular weight and the density. [10-12]Our experimental data suggest that PaO(2) and PaCO(2) values during xenon washout can be predicted largely by a rate of xenon elimination determined based on its blood solubility alone, as is true for N (2) O. Further mechanisms related to intrapulmonary gas mixing seemingly are not important.

When data regarding xenon and N2O elimination are evaluated, it is noteworthy that the total amount of the two gases dissolved in the body differs because of the different tissue solubilities. [13]In particular, the lipid solubility of xenon is greater than that of N2O, and, therefore, the total body content of xenon might indeed be greater than that of N2O when compared with calculations based on the blood [small lambda, Greek] values alone. Regardless of the total amount dissolved in the peripheral tissue, however, only that fraction of the two gases dissolved in blood and transported to the lung by blood flow can be eliminated by ventilation and thereby will affect alveolar gas composition. Consequently, elimination is limited by blood solubility ([small lambda, Greek]) and by perfusion, even when assuming a substantial amount of xenon dissolved in lipid-rich tissues. Therefore, the PaO(2) and PaCO(2) values that we measured in our experiment depended almost completely on the blood [small lambda, Greek] values of xenon and N2O, although the total body stores of the two gases did not precisely reflect the difference in blood solubility when the washout was begun. Finally, the potential effects of hemodynamics on gas exchange can be considered of minor relevance for our results, because the observed differences in carbon dioxide and mixed venous oxygen saturation were not significant.

In conclusion, our data show that diffusive hypoxia is unlikely to occur during recovery from xenon anesthesia; this finding corresponds with previous and with our actual theoretical model. As for N2O, the effects of xenon elimination on arterial and alveolar oxygen and carbon dioxide tensions can be explained by its blood solubility.

Appendix 1

Model Calculations

Breath-by-breath calculations of the alveolar and capillary gas composition and the arterial and mixed venous blood partial pressures were performed by following these steps.

1. definition of initial values for alveolar, arterial, and mixed venous partial pressures for oxygen, carbon dioxide, nitrogen, and for any additional inert gas, inspiratory gas composition, functional residual capacity (assumed to be equal in both compartments), tidal volume, respiratory rate, blood volume, and cardiac output

2. calculation of the alveolar volume of each gas x by Equation 3where fAx = fractional concentration of gas x obtained by PAx/(PB- PH(2)O) and functional residual capacity = functional residual capacity.

3. calculation of inspired volume for each gas by Equation 4where VT= tidal volume and fix = inspiratory concentration of gas x.

4. calculation of the resulting end-inspiratory alveolar volume of each gas by Equation 5and of the end-inspiratory partial pressure by Equation 6where VT= tidal volume, PB= barometric pressure, and PH(2)O = water vapor pressure

5. alveolar-capillary gas exchange: The quantity of a gas (M) contained in the alveolus (MAx) and that dissolved in blood (MBx) can be calculated by Equation 7and Equation 8where [small beta, Greek]G= capacitance coefficient of a gas in a gaseous medium (= 1.16 ml (STPD) x LBTPS-1 x mmHg-1at 37 [degree sign]C [14]), Pvx = mixed venous partial pressure of gas x, VB= capillary blood volume (= blood flow x duration of a respiratory cycle) and [small beta, Greek]B= capacitance coefficient of gas x in blood). The sum of both quantities (MTOT) after equilibrium is then given by Equation 9Equation 7, Equation 8, Equation 9can be combined to yield Equation 10where PEQ= equilibrium partial pressure and Q = alveolar perfusion per unit of time. The volume of gas being transported across the alveolar-capillary membrane (Mx) can be calculated as Equation 11

Equilibrium oxygen and carbon dioxide tensions were calculated by an iterative procedure that increased incrementally the capillary oxygen pressure and decreased the capillary carbon dioxide pressure starting from the mixed venous point and then calculated the increase in oxygen content and the decrease in carbon dioxide content based on previously published algorithms. [15,16]After each step, PaO(2) and PaCO(2) were recalculated to account for the amount of uptake and release of the two gases, and the procedure was repeated until the point was reached where Pa (O)(2) and PaCO(2) were equal to PCCO(2) and PcCO(2)(capillary pressure of oxygen and carbon dioxide). The partial pressure equilibrium between peripheral tissue and blood was calculated at the end of each cycle using equations that correspond to those used to calculate gas exchange at the alveolar level.


Goto T, Suwa K, Uezono S, Ichinose F, Uchiyama M, Morita S: The blood-gas partition coefficient of xenon may be lower than generally accepted. Br J Anaesth 1998; 80:255-6
Santak B, Radermacher P, Iber T, Adler J, Wachter U, Vassilev D, Georgieff M, Vogt J: In vivo quantification of endotoxin-induced nitric oxide production in pigs from Na15NO3-infusion. Br J Pharmacol 1997;122:1605-10
Jansen JR, Schreuder JJ, Settels JJ, Kloek JJ, Versprille A: An adequate strategy for the thermodilution technique in patients during mechanical ventilation. Intens Care Med 1990; 16:422-5
Sheffer L, Steffenson JL, Birch AA: Nitrous-oxide-induced diffusion hypoxia in patients breathing spontaneously. Anesthesiology 1972; 37:436-9
Sugioka K, Cattermole RW, Sebel PS: Arterial oxygen tensions measured continuously in patients breathing 21% oxygen and nitrous oxide or air. Br J Anaesth 1987; 59:1548-53
Rackow H, Salanitre E, Frumin MJ: Dilution of alveolar gases during nitrous oxide excretion in man. J Appl Physiol 1961; 16:723-8
Farhi LE, Olszowka AJ: Analysis of alveolar gas exchange in the presence of soluble inert gases. Respir Physiol 1968; 5-53-67
Scrimshire DA, Tomlin PJ: Gas exchange during initial stages of N2O uptake and elimination in a lung model. J Appl Physiol 1973; 34:775-89
Piiper J, Scheid P: Model for capillary-alveolar equilibration with special reference to O2uptake in hypoxia. Respir Physiol 1981; 46:193-208
Schumacker PT, Samsel RW, Sznajder JI, Wood LD, Solway J: Gas density dependence of regional VA/V and VA/Q inequality during constant-flow ventilation. J Appl Physiol 1989; 66:1722-9
Scheid P, Hlastala MP, Piiper J: Inert gas elimination from lungs with stratified inhomogeneity: Theory. Respir Physiol 1981; 44:299-309
Robertson H, Whitehead J, Hlastala MP: Diffusion-related differences in elimination of inert gases from the lung. J Appl Physiol 1986; 61:1162-72
Giunta F, Natale G, Del Turco M, Del Tacca M: Xenon: A review of IST anaesthetic and pharmacological properties. Appl Cardiopulmonary Pathophysiology 1996; 6:95-103
Piiper J, Dejours P, Haab P, Rahn H: Concepts and basic quantities in gas exchange physiology. Respir Physiol 1971; 13:292-304
Winslow RM, Samaja M, Winslow NJ, Rossi-Bernardi L, Shrager RI: Simulation of continuous blood O2equilibrium curve over physiological pH, DPG, and PaCO(2) range. J Appl Physiol 1983; 54:524-9
Kelman GR: Digital computer procedure for the conversion of P (CO)(2) into blood CO2content. Respir Physiol 1967; 3:111-5