Introduction

The basic principle of closed-circuit anesthesia is maintenance of a constant anesthetic state by adding gases and vapors to the breathing circuit at the same rate that the patient’s body removes those same substances. Often, the desired anesthetic state is first established using a high fresh gas flow (FGF) composed of gases, such as oxygen and nitrous oxide or air, and vapors (e.g., isoflurane, sevoflurane, desflurane). Once a steady state is attained, inspired and end-expired gas concentrations or tensions are noted, and FGF is reduced. Throughout this chapter, the words tension and partial pressure are used interchangeably. For gases, they have the same value as concentration. This is discussed under the section Partial Pressure. The circuit and patient gas tensions are maintained constant by adding oxygen, nitrous oxide, and agent vapor to the breathing circuit. In one common approach to closed-circuit maintenance, the amount of gas and vapor administered is determined empirically by titration. Several different titration endpoints can be used. Choices include maintenance of inspired tension, expired tension, and/or estimated anesthetic depth. When titrating against a defined endpoint, drugs can be added in a measured or quantified manner, or they can be added empirically without regard to the total amount administered. In closed-circuit anesthesia techniques, carbon dioxide is removed from exhaled gas, and the remaining exhaled gases and vapors are added to the FGF to produce inhaled gas. One advantage of this technique is that all gases exhaled are already warmed and humidified by the patient and are therefore well suited for rebreathing.¹ Another advantage is that oxygen consumption is monitored by titration. A final advantage is that cost is reduced dramatically.

In closed-circuit and low-flow anesthesia, inhaled gas is formed from two sources, and exhaled gas forms most of what the patient will breathe. In addition to exhaled gas, fresh gas is added in the correct quantity and composition to achieve the inspired gas tensions desired. The same inspired and expired gas tensions are established with a closed circuit as with a semiclosed or open (nonrebreathing) circuit. In this chapter the terms open circuit and nonrebreathing circuit are used interchangeably.

Closed-Circuit Anesthesia

Closed-circuit anesthesia can be viewed in several different ways. From one perspective, it is an anesthetic technique unlike all others. The classical closed-circuit literature describes theory and practice different from other techniques. In the classic closed-circuit approach, once a stable level of anesthesia is established with high-flow oxygen, nitrous oxide, and volatile agent, FGF is reduced to the patient’s predicted oxygen consumption (243 mL/min for a 70-kg adult), predicted nitrous oxide uptake rate (approximately 100 mL/min after the first 30 min), and predicted inhaled agent uptake. Nitrous oxide and agent uptake rates are calculated according to a mathematical formula based on body weight.²

The traditional closed-circuit anesthesia literature uses anesthetic liquid injection or infusion rather than a vaporizer. Liquid agent is administered to the breathing circuit according to a prescribed time regimen.^3-5 Specifically, the drug administration rate is inversely proportional to the square root of time (t):

This empiric relationship was first noted by Severinghaus⁶ for nitrous oxide in 1954 and was popularized by Lowe⁵ beginning in 1972. Severinghaus’s original data demonstrated that nitrous oxide uptake followed this power/function relationship fairly closely in the subjects he studied. Connor and Philip recently revalidated this mathematical relationship numerically⁷ and analytically.⁸

A scientific explanation exists for this curious and unexpected relationship. It has long been known that body tissues perfused with blood of a constant drug concentration or vapor tension have an uptake rate that decreases with time. Specifically, theory and research⁹ have demonstrated that uptake into each tissue takes the form of an exponential:

where t is time and τ is the time constant, or the time required to achieve 63% of the final value in response to a step input; e is the base of natural logarithms (2.7183 and so on), and K is a predictable constant. It can easily be derived that

and

where equals the time required to achieve half the final value in response to a step input.

Total body uptake is the sum of the individual uptakes by each tissue. In this case, uptake is the sum of a group of exponentials of different amplitudes (K_tissue) and time constants (τ_tissue). The sum of these exponentials is approximately equal to the power function, . It must be emphasized that the exponential relationship for a single tissue is derived theoretically and can be demonstrated empirically. But the power/function relationship is an empiric one that approximates the multiple exponential functions that describe the uptake into diverse tissues.^7,8

Partial Pressure, Tension, and Concentration

Throughout this chapter, the words tension and partial pressure are used interchangeably. This physical variable represents the effective pressure exerted by a gas, whether it is in the gas phase alone, in combination with another gas, or dissolved in blood or tissue.

Partial pressure is expressed in percent of 1 atmosphere (1 atm = 760 mm Hg). Expressing partial pressure this way serves several purposes. When partial pressure is expressed as percent of 1 atm, partial pressure and concentration have the same numeric value for gases. For example, 1 vol% isoflurane has an isoflurane tension of 1% of 1 atm, an expression often shortened to 1%. For blood and tissues, partial pressure is also expressed as a percent. By this definition, a tissue anesthetic measure of 1% does not represent 1% concentration; rather, it represents a partial pressure of 1% of 1 atm or 1% times 760 mm Hg, or 7.6 mm Hg in terms of absolute pressure.

Next, when partial pressure is expressed as a percent of one standard atmosphere (i.e., 760 mm Hg), the numbers and concepts work equally well at any atmospheric pressure. This is because the physiologic effects of inhaled anesthetics are the result of partial pressure and not concentration.^7,8 Thus, anesthetic tensions in the apparatus and patient are the variables that best explain kinetics in an understandable way.

When anesthetics are present in liquid or tissue, their concentrations are equal to their partial pressure × tissue/gas solubility × atmospheric pressure, according to Henry’s law. According to Dalton’s law, for multiple gases, this applies to each as if it were alone.

It is believed that the blood leaving each tissue compartment is in equilibrium with the tissues in that compartment, therefore the partial pressure in venous blood is equal to that in the tissues, as Graham’s law states. The blood entering each compartment has the same partial pressure of all gases as the arterial blood has everywhere. In the absence of lung shunt, arterial partial pressure equals alveolar partial pressure. Clinicians measure end-tidal (ET) gas concentrations or partial pressures. End-tidal partial pressure equals alveolar partial pressure when there is no physiologic dead space. After an infinite amount of time, the partial pressure of anesthetic inspired gas, alveolar gas, arterial blood, all tissues, and all venous blood, including mixed venous blood, is equal. At this time, there is no blood uptake of anesthetic at all. Thus the inspired and expired partial pressures are equal as well. Thus, after an infinite amount of time, the partial pressure of anesthetic throughout the system, from vaporizer to all tissues, is the same.

The above paragraph is a simplification that serves well in most situations. Exploring this further, in the presence of lung shunt, arterial tension is closer to mixed venous tension than it would be otherwise. When physiologic dead space is present, some of the gas leaving the lungs is inspired gas, rather than alveolar gas. This makes end-tidal partial pressure closer to inspired partial pressure than it would otherwise be. These are temporary limitations. After an infinite amount of time, these effects disappear as well, and anesthetic tension is equal everywhere. Because the alive body continues to consume oxygen and produce carbon dioxide, this equalization does not occur with these gases.

Lowe Technique

In the classic Lowe² technique for closed-circuit anesthesia, a loading dose of anesthetic vapor is administered to the breathing circuit to bring circuit tension to that desired in the patient’s alveoli. This level is somewhat higher than in the inspired gas. A unit dose for maintenance is then selected according to the patient’s body weight and is proportional to body mass to the three-quarter power (kg^3/4), a relationship first shown by Kleiber¹⁰ that has been further described by Brody.¹¹ Kleiber’s law goes on to state that many parameters of metabolic and circulatory processes—oxygen consumption, carbon dioxide production, cardiac output, and daily liquid requirement—are proportional to body mass raised to the three-quarter power.

In Lowe’s classic closed-circuit anesthesia, unit doses are administered at specific times with intervals between administrations increasing with time. These intervals, in minutes, are the sequence of odd integers {1, 3, 5, 7, 9, …} This results in injections being made at 0, 1, 4, 9, 16, and 25 minutes. This is referred to as the square root of time model, in that these injection times are the series of numbers whose square roots are sequential integers. This strange number manipulation may seem farfetched; however, Connor and Philip^7,8 demonstrated that this relationship approximated the uptake of nitrous oxide shown by Severinghaus⁶ with remarkable accuracy.

Even closed-circuit enthusiasts advise care and careful observation of the patient, as always. Lowe recommended that after 25 minutes, the square root of time method should be modified, and that the intervals between injections should not be lengthened as much (Lowe, personal communication, 1979). This is because after 25 minutes of anesthetic at constant depth, the rate of uptake becomes almost constant. By this time, fast (vessel-rich) tissues have equilibrated with arterial blood. Meanwhile, uptake into muscle and fat are diminishing slowly, more slowly than the square root of time, because of their long time constants.

It must be noted that the original 1972 theory and “cookbook” for closed-circuit anesthesia was written during the era of more primitive measurement and analysis (in 1972, the digital pocket calculator had not yet replaced the analog slide rule). Anesthetic agent analysis was uncommon, and the only clinically available device was the Dräger Narkotest (Dräger Medical, Telford, PA), which consisted of a slowly responding silicone rubber band in a box connected to a mechanical indicator.¹² The “rubber band in a box” stretched in proportion to the potency of the single or combined inhalation anesthetic tensions administered.

With each administration technique—open circuit, semiclosed circuit with high flow, semiclosed circuit with low flow, closed circuit with vaporizer, and closed circuit with liquid injection—each patient’s inspired, end-expired, and alveolar anesthetic tensions are the same and are independent of administration technique. Thus from the patient’s viewpoint, all inhalation anesthetic administration techniques are equivalent. The only difference is in the way the drugs are administered and the resulting waste or lack thereof.

Although the conventional closed-circuit literature used models created by fitting experimental data to particular mathematical formulations, this chapter does not rely on these. Rather it assumes that an anesthetic agent monitor is available and in use, measuring inspired and expired anesthetic concentrations. The dosage administration scheme is adjusted on the basis of a patient’s measured and needed levels of inspired and expired vapor and nitrous oxide. Delivered oxygen flow is adjusted similarly. In the absence of a multigas monitor, closed-circuit anesthesia may still be employed, but greater vigilance and understanding of anesthetic depth and uptake of anesthetic and oxygen are required.

Limitations and Solutions

Truly closed-circuit anesthesia cannot always be performed because of technical limitations of our gas delivery systems or gas monitors. In the 1980s, gas monitoring was often performed with a multiplexed mass spectrometer shared by 16 or 31 ORs, monitoring these rooms in succession. In that situation, sampled gas could not be returned to the patient from whom it was sampled. With the advent and commonality of stand-alone or integrated gas monitors with anesthesia delivery systems, there should be no such impediment. However, sampled gas return requires specific approval and is not available with anesthesia gas monitors by GE Healthcare (Waukesha, WI).

In 1963, Eger¹³ introduced the concept of anesthesia at constant alveolar concentration and defined the term minimum alveolar concentration (MAC). As a foundation for his approach, he used the pharmacokinetic theory explained by Seymour Kety^14,15 in 1950. Eger showed that constant alveolar concentration or tension can be produced and maintained when inspired anesthetic concentration is adjusted properly. To compute the proper adjustment sequence, or continuum, he calculated anesthetic uptake into each organ group and adjusted inspired concentration to maintain alveolar concentration constant. Figure 26-1 shows the inspired concentration required to maintain 0.8% alveolar halothane, equal to 1 MAC.¹³ Inspired halothane concentration is begun at 3.3% and is slowly and continuously reduced to 2% at 5 minutes, 1.5% at 20 minutes, and 1% at 3 hours.

A more realistic clinical endpoint is constant or desired anesthetic tension in the site of interest, the target organ, the brain, and the spinal cord. Producing a stepped or controlled change in brain anesthetic tension from its current level to any desired level is a useful clinical objective. This is equally true for high-flow, low-flow, and closed-circuit anesthesia.

Anesthesia Assessment and Adjustment

To the clinician who has used halothane, Eger’s calculated time course of halothane administration is quite reasonable. It mimics what was administered to many patients. Of course, in today’s clinical practice, the vaporizer is not adjusted according to the clock. Rather, the patient’s depth of anesthesia is assessed by mental integration of many signs. The pupils are observed for dilation or constriction; and blood pressure, heart rate, and possibly processed electroencephalogram (EEG; e.g., bispectral index, patient state index, entropy) and other variables are measured and evaluated. With the convenient availability of anesthetic agent monitors, most clinicians monitor expired anesthetic concentration (tension) as an important additional sign. This represents the anesthetic level in the end-expired gas as an approximation to that in the alveoli as an approximation to that in the blood. In the best of all situations, where end-tidal equals alveolar and arterial tension, the brain delay is still 3 to 5 minutes. This means that the measured end-tidal value predicts what the brain concentration will be in the near future. Likewise, view of the end-tidal value a few minutes previous is a real-time indicator of the brain and spinal cord values.

To achieve the “ideal” anesthetic, vaporizer setting and FGF are adjusted to maintain constant brain anesthetic tension. This is done by allowing a reasonable amount of alveolar overpressure during the period when anesthesia is being deepened (described later in this chapter). Finally, an estimate or measure of depth of anesthesia can be used.

High and Low Flows

Low-Flow Anesthesia

Use of the breathing circuit can be carried one step further. Instead of decreasing the vaporizer setting, FGF can be reduced to achieve the same decreasing inspired tension; Figure 26-10 shows this. FGF begins at 8 L/min while the vaporizer is set to 5% isoflurane (DEL). At the end of 2 minutes, rather than decreasing the vaporizer setting, FGF is decreased to 2 L/min. Then, several minutes later, the vaporizer setting (DEL) is decreased as needed. All adjustments are made while the alveolar anesthetic tension (A, dark green) is observed, and the FGF vaporizer setting is adjusted to achieve 1.1% (1 MAC) expired isoflurane. At the end of 15 minutes with this low-flow anesthetic technique, 1.2 L has been delivered to the breathing circuit, and uptake by the patient’s tissues is 0.6 L. This anesthetic administration is then 33% efficient in terms of use of isoflurane vapor (i.e., 0.4/1.2 = 0.333). In other words, with the low-flow technique, the concentration inspired by the patient is identical to that inspired with the high-flow technique; patient response—end-expired anesthetic tension, alveolar tension, and anesthesia depth—are also unaffected. The only difference between the two techniques is the manner in which the appropriate gas mixture is created.

Liquid Injection

To achieve closed-circuit anesthesia beginning with induction, or to accomplish rapid deepening during anesthesia, a modern concentration-calibrated, direct-reading vaporizer will not suffice. This is because the volume of its output at low FGF is insufficient to produce large or rapid changes in inspired tension. For this reason, liquid anesthetic agent is injected directly into the breathing circuit by many practitioners of the closed-circuit technique. The Gas Man model allows liquid injection to be simulated, while the student observes the resulting anesthetic tensions in the circuit and in the patient. Other practitioners increase FGF whenever they desire to change depth.

Figure 26-12 shows a Gas Man representation of closed-circuit anesthesia administered with 0.5 mL liquid isoflurane injections timed to achieve and maintain 1.1% exhaled anesthetic partial pressure. Each 0.5 mL of liquid isoflurane generates approximately 100 mL of vapor at room temperature. Liquid isoflurane is injected at empiric times with the objective of maintaining 1.1% alveolar partial pressure. Note that although inspired tension changes dramatically, and expired tension varies somewhat, brain tension (VRG on the picture, brown [R] on the graph) is quite stable over time. Alveolar tension can be smoothed by decreasing the liquid bolus volume and increasing the frequency of injection by converting to a continuous infusion of liquid anesthetic into the breathing circuit.

At the end of 30 minutes of empiric intermittent liquid injection, the graph in Figure 26-12 shows that patient uptake is again 0.6 L of isoflurane vapor. To achieve this, 0.9 L of vapor was delivered to the breathing circuit as nine individual 0.5 mL liquid injections. The discrepancy between the 0.9 L delivered and the 0.6 L uptake is anesthetic vapor that remains in the breathing circuit (CKT, 5 L) and the patient’s lungs (ALV, 2 L). This same 0.3 L volume of anesthetic vapor was neglected in the previous descriptions for the sake of clarity. Other than storage in the circuit and lungs (alveoli), the efficiency of agent use is 100% when a closed-circuit technique is used as described.

Figure 26-13 shows a Gas Man representation of closed-circuit anesthesia administered with injections of liquid isoflurane according to Lowe’s square-root-of-time model. Observe the 1.1% VRG (R) isoflurane tension that results. The first injection (0.5 mL) is made at time zero (first DEL spike). Subsequent (0.7 mL) injections are made at 1, 4, and 9 minutes. Note that bouncy but generally constant alveolar anesthetic tension (A, dark green) is achieved.

At the end of 30 minutes of empiric intermittent liquid injection, patient uptake is again 0.6 L of isoflurane vapor (see Fig. 26-13). To achieve this, 0.8 L of vapor was delivered to the breathing circuit as five individual 0.7 mL liquid injections following a loading dose of 0.6 mL. The discrepancy between the 0.8 L delivered and the 0.4 L uptake is anesthetic vapor that remains in the breathing circuit (CKT, 6 L) and the patient’s lungs (ALV, 2 L). This same 0.1 L anesthetic volume was neglected in the previous descriptions for the sake of clarity. Other than this, the efficiency of agent use is 100%. The slight differences in uptake in Figures 26-11, 26-12, and 26-13 are due to the slight differences in anesthetic tension maintained in VRG.

The initial (average) inspired concentration necessary to achieve 1.1% alveolar tension was approximately 2.9%, as in the previous examples. Average inspired tension was allowed to fall to 1.3% at the end of 30 minutes, as before, this time by lengthening the time between liquid injections. Inspired tension varied considerably between injections, but on the average, inspired tension was similar to that for higher flow systems.

The same principles apply to desflurane closed-circuit anesthesia. Figure 26-14 shows a Gas Man simulation of closed-circuit desflurane anesthesia with a 2.0 mL initial injection followed by 1.5 mL liquid desflurane injections at 1, 4, and 9 minutes. Figure 26-15 shows a Gas Man simulation of closed-circuit desflurane anesthesia with a 2.0 mL initial injection followed by 1.5 mL liquid desflurane injections at 1, 4, and 9 minutes. At 12 minutes, 1.0 MAC levels are achieved in alveolar gas and in the vessel-rich tissue group; the vaporizer is set to 16% desflurane, and no further liquid injections are required. The vaporizer setting needs to be reduced beginning at 30 minutes.

Choice of Anesthetic Agent

The choice of anesthetic agent should take into consideration the patient, surgery, and closed-circuit technique. Whenever a rebreathing circuit with carbon dioxide absorption is used, there is some expired admixture to FGF. As FGF is reduced toward that of a closed circuit, expired concentration plays a greater role and fresh-gas concentration plays a lesser role in determining inspired concentration. Inspired concentration is the flow-weighted average of fresh gas and expired gas concentrations. Low-solubility agents provide expired concentrations close to those of inspired concentrations. Thus they require relatively lower delivered (vaporizer) concentrations than agents of higher solubility. If the concentration required exceeds the maximum vaporizer setting, alternative administration techniques are required. Higher FGF can be used, and the closed-circuit technique abandoned, in favor of low, also called minimal, flow—at least until the vaporizer alone will suffice. Alternatively, a closed circuit can be achieved by injecting liquid anesthetic into the breathing circuit.

Only desflurane can consistently be used without the liquid injections required in a closed circuit. Desflurane’s low blood/gas solubility (0.42) and vanishingly low metabolism make it well suited for low-flow and closed-circuit anesthesia. When used this way, it becomes quite economical. The maximum vaporizer setting of 18% exceeds that necessary to maintain anesthesia depth.

Sevoflurane has relatively low blood/gas solubility (0.67) and is used for closed-circuit anesthesia in countries where there is no regulation barring it.²⁶ In the United States, the Food and Drug Administration (FDA) package insert requires an FGF of 2 L/min for cases of a duration greater than 2 MAC × the number of hours (e.g., 1 MAC for 2 hours, 2 MAC for 1 hour). For less than this time × duration product, at least 1 L/min is required.

This FDA restriction is based on the increased heat produced by carbon dioxide absorption and the accompanying temperature rise in the carbon dioxide absorbent. Baralyme was removed from the U.S. market because of its substantial heat production.²⁷ In many countries, a variety of “cool” CO₂ absorbents are available that use lower or absent NaOH and KOH in combination with CaCO₃ or other absorbents.²⁷

The resulting increased temperature increases the degradation of sevoflurane to compound A, which is toxic in swine.²⁷ Sevoflurane’s use in low-flow and closed-circuit situations has not shown any clinical problems.²⁶ The maximum vaporizer setting of 8% is just sufficient to maintain anesthesia after adequate depth is attained.

Isoflurane’s higher blood/gas solubility (1.3) requires high vaporizer settings during maintenance. Indeed, the maximum isoflurane vaporizer setting of 5% is insufficient to maintain 1 MAC anesthesia in a closed circuit for the first hour.

Both isoflurane and desflurane are broken down into compounds, including carbon monoxide, when exposed to high temperatures.²⁷ With the removal of Baralyme from the market, this is no longer a problem.²⁸

Halothane’s high blood/gas solubility (2.4) requires particularly high vaporizer settings during maintenance. The maximum halothane vaporizer setting of 5% is insufficient to maintain 1 MAC anesthesia in closed circuit for the first hour.

The ratio of delivered to alveolar tension for several agents was first explored by Zuntz.⁹ Figure 26-16 shows the delivered/alveolar ratio over time simulated with Gas Man. Note that the higher the blood/gas solubility, the higher the delivered/alveolar tension ratio.

Practical Aspects of Administration

The practical aspects of administration of closed-circuit anesthesia must be known and understood for this technique to be used effectively. Many contemporary anesthesia machines available in the United States are not designed to facilitate or allow closed-circuit techniques and thus need modification before or during use; all GE Healthcare anesthesia machines fall into this category. In the United States, the Dräger Apollo and Fabius can provide fully closed-circuit anesthesia because there is no lower limit to FGF, and return of sampled gas is facilitated by a built-in sampled gas-return port.

In some countries, GE Healthcare and Dräger both provide low-flow anesthesia machines with end-tidal control of agent and end-tidal (GE Aisys) or inspired (Dräger Zeus) control of oxygen. Neither of these provides fully closed-circuit anesthesia.

Likewise, the anesthetic record must be adapted for the closed-circuit technique. It is important for the anesthetic record to capture FGFs and vaporizer settings as well as measured inspired and end-expired concentrations. When low-flow or closed-circuit anesthesia is administered, certain additional variables need to be recorded on the anesthetic record. First, the flows of nitrous oxide, air, and oxygen must be recorded to explain what is really administered to the breathing circuit separate and distinct from the resulting inspired concentrations of oxygen, nitrous oxide, and balance gas, which is predominantly unmeasured nitrogen. Expired tidal volume should be carefully observed and recorded.

With older anesthesia delivery systems, patient ventilation was dependent on both ventilator volume settings and concurrent FGF. Newer anesthesia machines have eliminated this interaction. The vaporizer setting should be recorded, however, because the adjustment is made by the anesthesia care provider, and it influences inspired concentration. Because inspired anesthetic tension is quite different from that delivered from the vaporizer, inspired tension should be measured and recorded to demonstrate that it was known and well controlled. Finally, expired anesthetic tension should be controlled and recorded to assist in patient management. The need to monitor and record expired-agent tension applies irrespective of FGF; it is equally applicable to high-flow techniques. When all three of these variables—vaporizer setting, inspired tension, and expired tension—are recorded, the anesthetic record documents the patient’s receipt of a safe and effective anesthetic. All closed-circuit concentrations and flows should be recorded without reservation. Closed-circuit anesthesia has been shown to be as safe as a high-flow technique.²⁹

Practical administration of closed-circuit anesthesia follows the basic principle of the technique: deliver gases and vapors at a rate equal to tissue uptake while maintaining alveolar tension at the desired level (Figure 26-11). For most adult patients, the following regimens work as a first approximation. As with any clinical technique, changes are made to adapt delivery to each patient, surgery, and anesthetic agent.

Flow Control

When a high-flow technique is used first, FGF is reduced to provide a completely closed circuit. The adjustable pressure-limiting (APL), or pop-off, valve is closed to keep gas from leaving the circuit, should the reservoir bag be used later. On older anesthesia machines, in which tidal volume and FGF were related, set tidal volume was increased to provide an undiminished exhaled tidal volume. With newer anesthesia machines that directly control inspired tidal volume, this adjustment is no longer necessary.

With agent plus nitrous oxide plus oxygen anesthesia in adults, oxygen flow is reduced to 200 mL/min, and nitrous oxide is reduced to 100 mL/min. This approximates patient uptake for these gases. With isoflurane, the vaporizer is set to 3% to provide the approximately 9 mL/min isoflurane vapor required for 0.5 MAC contribution of isoflurane. With desflurane, the vaporizer setting is 5% to provide the approximately 16 mL/min desflurane vapor required for the 0.5 MAC contribution of desflurane.

With agent plus air plus oxygen anesthesia, oxygen flow is reduced to 200 mL/min to approximate oxygen uptake, and air is turned off because there is no nitrogen uptake or elimination when circuit gas concentrations are near that of air. With isoflurane, the vaporizer is set to 5%, which does not quite supply the 18 mL/min required to maintain 1 MAC anesthesia. Either FGF must be increased periodically or liquid isoflurane must be injected occasionally. With desflurane, the vaporizer is set to approximately 14% to supply the 32 mL/min desflurane vapor required to maintain 1 MAC anesthesia.

The ventilator is adjusted so that no gas leaves the breathing circuit at the end of exhalation; the exact technique depends on the ventilator used and is described below. Typically, the ascending (standing) bellows is adjusted to prevent it from rising to the mechanical stop that initiates discharge of excess gas. A mark can be placed on the bellows housing to signify the correct end-expired bellows position. This represents a volume mark. Deviation of the bellows from this volume mark signifies inequality between total volume delivered to the circuit and volume taken up by the patient. This deviation suggests that total FGF should be modified in some way.

Changes in concentrations of gases and vapors disclose what should be changed, the direction of change, and the magnitude of change. A change in measured F_IO₂ suggests that a change should be made in the flow ratio of oxygen to the other gas (nitrous oxide or air). A change in inspired agent tension suggests that a change should be made in the administration of anesthetic agent. This agent administration may have been in the form or vapor or liquid. If administration is in the form of vapor from a vaporizer, the vaporizer setting should be changed. If administration is in the form of periodic liquid injection, this suggests a change in interval of the same unit dose, adjustment of the unit dose using the same interval, or some combination of change of dose and interval. If administration is in the form of a continuous liquid injection into the breathing circuit, the rate of injection is altered.

The quantification of these adjustments is not immediately obvious. When the volume of the ventilator bellows changes, flows of nitrous oxide and oxygen are adjusted in proportion to their concentrations in the breathing circuit, not according to their proportion in fresh gas. For example, if the respiratory gas monitor indicates an inspired concentration of 66% nitrous oxide and 33% oxygen, it is in this ratio that additional gas must be added to maintain the same inspired oxygen concentration. Therefore the adjustments in oxygen flow are accompanied by adjustments in nitrous oxide that in this case happen to be twice as large. That is, when oxygen flow is increased by 100 mL/min, nitrous oxide flow is increased by 200 mL/min.

When a decrease in FiO₂ is noted in the absence of a change in circuit volume, oxygen flow is increased, and nitrous oxide flow is decreased, this time by equal amounts to keep circuit volume constant. Simultaneous changes in FiO₂ and circuit volume require mixed compensation, which is a combination of the two techniques described.

Finally, when inspired or expired agent tension requires an increase, the vaporizer setting is increased, or additional liquid is injected into the breathing circuit. To lighten anesthesia, the vaporizer is turned off, or a liquid injection dose is withheld or delayed. An activated charcoal filter may be used to adsorb volatile agent.^30-35 Charcoal functions best when a pair of filters are placed in the inspiratory and expiratory limbs of the circuit.³⁵ Anesthesia may also be lightened by flushing the breathing circuit briefly with oxygen. When a nitrous oxide plus oxygen plus agent anesthetic is used, the agent dilution should be done with a mixture of oxygen and nitrous oxide, with flows proportional to their circuit concentrations. When an agent plus air plus oxygen anesthetic is administered, air plus oxygen is the appropriate gas for agent dilution; the latter two cannot be done simply with today’s anesthesia delivery systems. In each case the resulting change in inhaled tensions of all gases and vapors should be observed and controlled. Because lightening anesthesia happens most often near the end of anesthesia, increasing inspired oxygen concentration may be desired, and oxygen flush or increased flow can be used. This is especially true for agent plus air plus oxygen anesthesia; however, this might not be appropriate for agent plus nitrous oxide plus oxygen anesthesia.

During the administration of any anesthetic, switching between mechanical and manual ventilation (i.e., controlled and spontaneous, respectively) is often desired. This change influences closed-circuit anesthesia in several ways. First, if the content of the reservoir bag and bellows is different, switching modes will produce a change in gas composition. This usually is not desired. Such a change is averted by emptying whichever gas reservoir is not in use, bellows or reservoir bag, and transferring circuit gases back into the empty reservoir (bellows or bag) when convenient. This is done in a manner that will be described below.

When a change in the volume in the ventilator bellows or reservoir bag is desired, this can be accomplished by switching between bag and ventilator. To do this, the clinician maintains a partially filled reservoir bag throughout the case. Meanwhile, the patient is ventilated mechanically. Whenever it becomes necessary to compensate for accumulated circuit volume change, gas is transferred between bellows and bag. This is done by transferring gas via the patient’s lungs during exhalation. Deft adjustment of the bag/bellows selector lever is required. The same technique is used to fill the out-of-use reservoir bag or ventilator bellows, as introduced above.

With the earlier Ohmeda (GE Healthcare) breathing circuit assembly (Modulus 2 Plus), when the selector valve was moved 5% of the way from “Vent” to “Bag,” the ventilator and bag were connected to each other, and gas could be easily transferred without using the patient’s lungs as a temporary reservoir. With newer anesthesia delivery systems, this is no longer possible. With the GE Aestiva, using the patient’s lungs can successfully act as a transfer volume. This is because the switch to ventilator mode interposes a momentary expiratory pause, during which the patient’s lungs can empty into the ventilator bellows. When transferring gas from the bellows to the reservoir bag, switching the mode from “Vent” to “Bag” at the beginning of exhalation fills the reservoir bag and leaves the ventilator bellows partly empty.

With the GE Advanced Breathing System (ABS) Aespire, Avance, and Aisys workstations, switching to mechanical ventilation mode begins instantly with the inspiratory phase of ventilation. Thus the patient’s first tidal volume may be increased above the set tidal volume under some conditions.

With the Dräger FGF decoupling machines (e.g., Fabius, Apollo, and in some countries, Julian), the reservoir bag is in the expiratory limb of the breathing circuit during mechanical and manual/spontaneous ventilation. Therefore the bag contains gases and vapors of the correct concentration and sufficient volume to switch ventilation modes at any time. In these Dräger anesthesia delivery systems, the clinician must be aware that switching from vent to bag creates a moment in which the reservoir bag positive-pressure relief valve is open. Squeezing the reservoir bag during this period results in lost volume from the circuit to the scavenger interface. With high-flow anesthesia, this is not a problem; but with low-flow or closed-circuit anesthesia, this will leave the reservoir bag empty, and it will not refill. Thus the user must delay squeezing the reservoir bag until the valve closes (about 1 second).

Liquid Agent Injection

Where to inject liquid agent is an interesting question. Early writings advocated injecting into the expiratory limb of the circuit to avoid unmonitored and unknown changes in inspired tension. Careful monitoring of airway anesthetic tension allows liquid injection closed-circuit anesthesia to be performed simply and safely. Figure 26-17 shows the time course of airway anesthetic tension in response to a 1.0 mL injection of liquid isoflurane into the expiratory limb of the breathing circuit connected to a 70-kg patient undergoing surgery. With this agent monitor, concentration rises with each inspiration and falls with each expiration. Thus the peaks and troughs represent inspired and end-tidal anesthetic tensions, respectively. Separate curves connecting all the peaks and all the troughs would follow the “I” and “E” curves of the isoflurane simulations in Figs. 26-12 and 26-13. Note that inspired tension does not rise for 1 minute, and an additional 1.5 minutes elapses before peak inspired tension is reached. At the peak, inspired tension rises by 1.5% from its baseline of 0.5%. Meanwhile, expired tension rises by 0.7% from 0.5% to 1.2%.

With the advent of breath-by-breath monitors of anesthetic agents, it became obvious that injection in the inspiratory limb is possible and advantageous. Figure 26-18 shows the result of a 1.0 mL liquid isoflurane injection into the inspiratory limb of the circuit in the same patient. Inspired tension peaks at 4.4% above baseline after two breaths (15 seconds). Expired tension rises by 1.2% within two breaths and remains constant for about 2 minutes, and 3 minutes after injection, the two tracings that depict injection in the two circuit locations become quite similar. A note of caution is in order here: when injected into either the inspired or the expired limb, liquid anesthetic must never be allowed to reach the patient. Conventional breathing circuits with a dependent section between the carbon dioxide absorber and the patient easily achieve this. Continuous, breath-by-breath gas monitoring helps the clinician ensure that inspired and expired tensions are as desired and expected. Figure 26-19 shows alternate injections of 0.5 mL of liquid isoflurane into the inspiratory and expiratory limbs of the circuit. In each case, injection into the inspiratory limb resulted in an inspired peak within two breaths, whereas injection into the expiratory limb resulted in an inspired peak approximately 90 seconds after injection. Injection into the inspiratory limb provides superior control of inspired and expired anesthetic tension. An exhaled limb injection may be acceptable when constant depth is the goal. Figure 26-20 shows a 6 minute trend graph of inspired and expired desflurane concentration after 1 mL desflurane liquid injections were made into the inspired circuit limb adjacent to the carbon dioxide absorber. Note that inspired tension rises to 3% quickly. This small rise in inspired tension does not generally elicit a sympathetic response.

Many anesthesiologists use a concentration-calibrated vaporizer for closed-circuit anesthesia. Some adjust the vaporizer by intermittently selecting between full on and full off. With low flow through a concentration-calibrated vaporizer, the time constant (τ) for circuit tension change is long, because τ is equal to circuit volume divided by circuit flow in a fully mixed circuit. With a volume of 8 L (6 L circuit + 2 L lung volume or functional residual capacity [FRC]) and a flow of 0.25 to 0.5 L/min, τ is between 32 and 16 minutes. Because of this, even a drastic change in vaporizer setting produces a very slow change in the composition of inspired gas. If faster control is desired, either FGF must be increased or liquid must be injected.

Closed-circuit anesthesia can also be administered via measured-flow vaporizers (e.g., Copper Kettle and Vernitrol). However, because of the potential for error in the calculations required,³⁸ and the fact that these vaporizers are considered obsolete, their use is discouraged (see also Chapter 3).³⁸

The initial administration of inhaled agents can be performed by liquid injection. With desflurane, a convenient liquid volume for adults is 1.0 mL. Figure 26-20 is a screen shot of the 6-minute graph of inspired and expired desflurane concentration. When the same administration sequence is viewed over a 15-minute window, the bumps in inspired and expired concentration are smoothed, and the inspired and expired curves appear just as they would with careful vaporizer use with high or low FGF.

Avoiding Hypoxia

During anesthesia, hypoxia must be avoided, so its likelihood must be reduced to as close to zero as possible. It is crucial to understand how inspired hypoxia can develop during closed-circuit and even low-flow anesthesia.

In a closed circuit, as long as oxygen administration exceeds oxygen uptake, inspired oxygen concentration rises. In steady state, oxygen administration exactly equals oxygen uptake, and FiO₂ remains constant. This is a fundamental principle of closed-circuit anesthesia.

In low-flow non–closed-circuit situations, oxygen delivery in excess of uptake is not sufficient to avoid falling FiO₂ and inspired hypoxia. This is because in all non–closed-circuit situations, gas leaves the breathing circuit, and this gas includes oxygen. Thus a 1 L/min FGF of air inevitably leads to inspired hypoxia in a 70-kg patient. This is because a 70-kg awake patient consumes about 250 mL/min oxygen, and this value falls to about 210 mL/min under anesthesia. With an FGF of 1 L/min of air, 790 mL/min nitrogen and 210 mL/min oxygen are delivered to the breathing circuit. Thus oxygen delivery exactly equals oxygen uptake. But each minute, uptake of oxygen is 210 mL, and uptake of nitrogen is zero. Thus the net flow to the breathing circuit is 790 mL/min nitrogen and 0 mL/min oxygen. Breathing circuit oxygen concentration therefore falls, and it falls linearly, because oxygen is taken up at a constant rate until life ceases or is diminished. Thus in a breathing circuit and patient lungs with a combined volume of 8 L, FiO₂ will fall 2.6%/min ([0.21 L/min]/8 L = 0.026/min). Figure 26-21 shows the graph of FiO₂ and SpO₂ over time in an 84 kg subject breathing FGF at 1 L/min air plus 0.050 L/min oxygen from a semiclosed breathing circuit (Ohmeda Modulus 2+ CD anesthesia machine with a GMS absorber assembly). After 12 minutes, FiO₂ fell to 8% and SpO₂ fell to 82% when the demonstration was terminated.

Anesthesia Machines in the United States

Dräger Apollo, Fabius, and Tiro are all capable of FGF down to 100 mL/min. Apollo has sample gas returned to the breathing circuit via internal connections, so the user need not make any changes. Apollo guides the user to use flows less than 500 mL in excess of patient uptake. The Datex-Ohmeda (GE Healthcare) Aestiva allows FGF down to zero, but there is no connection for sample gas return. The Datex-Ohmeda Aespire and Avance allow FGF as low as 50 mL/min, but there is no connection for sample gas return. The Aisys has 200 mL/min minimum oxygen flow with 50 mL/min increments. Aisys air flow is zero, 100 mL/min, and higher with 50 mL/min increments. There is no connection for sample gas return, and if sample gas is returned from a GE monitor, room airflow of 10 to 40 mL/min accompanies it. This makes Aisys difficult to use for closed-circuit anesthesia. Box 26-1 shows closed-circuit attributes of anesthesia delivery systems available in the United States as of the time of this writing.

1. Bengston J.P., Bengston A., Stenqvist O. The circle system as a humidifier. Br J Anaesth. 1989;63:453–457.

2. Lowe H.J. Dose regulated Penthrane (methoxyflurane) anesthesia. Chicago: Abbott Laboratories; 1972.

3. Lowe H.J., Mackrell T.N., Mostert J.W., et al. Quantitative closed circuit anesthesia. Anesthesiol Rev. 1974;12:16–19.

4. Aldrete J.A., Lowe H.J., Virtue R.W., eds. Low flow and closed system anesthesia. New York: Grune and Stratton, 1979.

5. Lowe H.J., Ernst E.A. The quantitative practice of anesthesia: use of closed circuit. Baltimore: Williams & Wilkins; 1981.

6. Severinghaus J.W. The rate of uptake of nitrous oxide in man. J Clin Invest. 1954;33:1183–1189.

7. Connor C.W., Philip J.H. The Severinghaus square root of time relationship for anesthetic uptake and its implications for the stability of compartmental pharmacokinetics, Physiol Meas. 2008;29:685–701. http://iopscience.iop.org/0967-3334/29 Available online at

8. Connor C.W., Philip J.H. Closed-form solutions for the optimum equivalence of first-order compartmental models and their implications for classical models of closed-circuit anesthesia. Physiol Meas. 2009;30:N11–N21.

9. Zuntz N. Zu Pathogenese und Therapie der durch rasche luft druck anderungen erzeliegten krankheiten. Fortschr Med. 1897;15:632.

10. Kleiber M. Body size and metabolism. Hilgardia. 1932;6:315–353.

11. Brody S. Bioenergetics and growth. New York: Reinhold; 1945.

12. Lowe H.J., Hagler K. Clinical and laboratory evaluation of an expired anesthetic gas monitor (Narko-test). Anesthesiology. 1971;34:378–382.

13. Eger E.I., II., Guadagni N.P. Halothane uptake in man at constant alveolar concentration. Anesthesiology. 1963;24:299–304.

14. Kety S.S. The physiological and physical factors governing the uptake of anesthetic gases by the body. Anesthesiology. 1950;11:517–526.

15. Kety S.S. The theory and application of the exchange of inert gas at the lungs and tissues. Pharmacol Rev. 1951;3:1–41.

16. Philip J.H. Gas Man: an example of goal-oriented computer-assisted teaching which results in learning. Int J Clin Mon. 1986;3:165–173.

17. Garfield J.M., Paskin S., Philip J.H. An evaluation of the effectiveness of a computer simulation of anesthetic uptake and distribution as a teaching tool. Med Educ. 1989;23:457–462.

18. Eger E.I., Shafer S.L. Context-sensitive decrement times for inhaled anesthetics. Anesth Analg. 2005;101:688–696.

19. Bouillon T., Shafer S. (Editorial) Hot air or full steam ahead? An empirical pharmacokinetic model of potent inhaled agents. Br J Anaesth. 2000;84:429–431.

20. Eger E.I., II. Anesthetic uptake and action. Baltimore: Williams & Wilkins; 1974.

21. Mapleson W.W. An electrical analogue for the uptake and exchange of inert gases and other agents. J Appl Physiol. 1963;18:197–204.

22. MacKrell T.N. An electrical teaching model. In: Papper E.M., Kitz R.J., eds. Uptake and distribution of anesthetic agents. New York: McGraw-Hill; 1963:215–223.

23. Leroux J.G., Booij L.H. Model-based administration of inhalation anaesthesia 1. Developing a system model. Br J Anaesth. 2001;86:12–28.

24. Gravenstein J.S. Training devices and simulators. Anesthesiology. 1988;69:295–297.

25. Eger E.I., Bahlman S.H. Is the end-tidal anesthetic partial pressure an accurate measure of the arterial anesthetic partial pressure? Anesthesiology. 1971;35:301–303.

26. Bito H., Ikeda K. Closed-circuit anesthesia with sevoflurane in humans: effects on renal and hepatic function and concentrations of breakdown products with soda lime in the circuit. Anesthesiology. 1994;80:71–76.

27. Kharasch E.D., Powers K.M., Artru A.A. Comparison of Amsorb, sodalime, and Baralyme degradation of volatile anesthetics and formation of carbon monoxide and compound A in swine in vivo. Anesthesiology. 2002;96:173–182.

28. ECRI Health Devices Alerts, October 8, 2004, Vol 28, No. 41.

29. Ernst E.A., MacKrell T.N., Pearson J.D., et al. Patient safety: a comparison of open and closed anesthesia circuits [abstract]. Anesthesiology. 1987;67(3A):A474.

30. Epstein H.G. Removal of ether vapour during anaesthesia. Lancet. 1944:114–116.

31. Hawes D.W., Ross J.A.S., White W.C., et al. Servo-control of closed circuit anaesthesia. Br J Anaesth. 1982;54:229–230.

32. Ernst E.A. Use of charcoal to rapidly decrease depth of anesthesia while maintaining a closed circuit. Anesthesiology. 1982;57:343.

33. Baumgarten R.K. Simple charcoal filter for closed circuit anesthesia. Anesthesiology. 1985;63:125. (letter)

34. Jantzen J.P. More on black and white granules in the closed circuit. Anesthesiology. 1988;69:437–438. (letter)

35. Birgenheier N., Stoker R., Westenskow D., Orr J. Activated charcoal effectively removes inhaled anesthetics from modern anesthesia machines. Anesth Analg. 2011;112:1363–1370.

36. Huffman L.M., Riddle R.T. Mass spectrometer and/or capnograph use during low-flow, closed circuit anesthesia administration [letter]. Anesthesiology. 1987;66:439–440.

37. Mushlin P.S., Mark J.B., Elliott W.R., et al. Inadvertent development of subatmospheric airway pressure during cardiopulmonary bypass. Anesthesiology. 1989;71:459–462.

38. Keenan R.L., Boynan C.P. Cardiac arrest due to anesthesia: a study of arrest incidents and causes. JAMA. 1985;253:2373.

39. Morita S. Inspired gas contamination by non-anesthetic gases during closed-circuit anesthesia. Journal of the Society for Closed Circuit Anesthesia. 1985;2:24–25.

40. Sharp J.H., Trudell J.R., Cohen E.N. Volatile metabolites and decomposition products in man. Anesthesiology. 1979;50:2.

41. Eger E.I., II. Dragons and other scientific hazards. Anesthesiology. 1979;50:1. (editorial)

42. Eger E.I., II. Enflurane: a compendium and reference, ed 3. Madison, WI: Anaquest; 1985.

43. Eger E.I., II. Isoflurane: a compendium and reference, ed 2. Madison, WI: Anaquest; 1985.

44. Eger E.I., II. Stability of I-653 in soda lime. Anesth Analg. 1987;66:983–985.

45. Hanaki C., Fujii K., Morio M., et al. Decomposition of sevoflurane by soda lime. Hiroshima J Med Sci. 1987;36:61–67.

46. Speiss W. To what degree should we be concerned about carbon monoxide accumulation in closed circuit anesthesia? Journal of the Society for Closed Circuit Anesthesia. 1984;1:8.

47. Barton F.L., Nunn J.F. Use of refractometry to determine nitrogen accumulation in closed circuits. Br J Anaesth. 1975;47:346–349.

48. Philip J.H. Nitrogen buildup in a closed circuit [abstract]. Anesthesiology. 1990;73:A465.

49. Bengston J.P., Sonander H., Stenqvist O. Denitrogenation and low flow anesthesia. Journal of the Society for Closed Circuit Anesthesia. 1988;5:8–9.

50. Morita S., Latta W., Hambro K., et al. Accumulation of methane, acetone and nitrogen in the inspired gas during closed-circuit anesthesia. Anesth Analg. 1985;64:343–347.

51. Hendrickx J.F., van Zundert A.A., de Wolf A.M. Influence of the reference gas of paramagnetic oxygen analyzers on nitrogen concentrations during closed-circuit anesthesia. J Clin Monit Comput. 1998;14(6):381–384.

52. Aldrete J.A. Vaporizing adaptors as injection ports for closed anesthesia system. Journal of the Society for Closed Circuit Anesthesia. 1985;2:26–28.

53. Petrella W.K. Enhanced discovery of anesthesia related events: an analysis of 400 consecutive low-flow and closed-circuit cases. Journal of the Society for Closed Circuit Anesthesia. 1989;6:14–15.

54. Hylani M.A. Closed circuit anesthesia. Middle East J Anesthesiol. 1986;8:505–510.

55. Droh R., Spintge R. Closed-circuit system and other innovations in anaesthesia. Berlin: Springer-Verlag; 1986.

56. Rolly G., Versichelen L., Verkaaik, et al. Mass spectrometry analysis of a new closed circuit anesthesia apparatus (PhysioFlex). Journal of the Society for Closed Circuit Anesthesia. 1989;6:5–6.

57. Verkaail A.P., Erdmann W. Respiratory diagnostic possibilities during closed circuit anaesthesia. Acta Anaesthesiol Belg. 1990;41:177–188.