The coefficient (inverse time constant) from the log-linear regression provides only an estimate of the ratio of the effective fresh gas flow rate (QFG) and breathing system volume (Vds); calculation of the total mass of residual sevoflurane contained in the Fabius breathing circuit and ventilator requires a determination of the actual value of Vds (total mass = initial concentration (C0) · Vds, analogous to total dose = initial concentration X volume of distribution). Although the oxygen flow rate and minute volume are known, the actual effective fresh gas flow into the circuit is not simply the delivered oxygen flow and cannot be determined from first principles. We attempted to estimate the mass of sevoflurane retained in the breathing system as follows.
Estimate of Actual Circuit Volume
Making the assumptions that QFG was the same during the bulk (early) clearance and trace (late) clearance phases and that Vds during the early phase was equal to actual breathing system volume (VCIR), it is possible to estimate QFG, and thus Vds during the late phase, provided that an estimate of actual VCIR could be obtained. After equilibrating the breathing circuit and test lung with 5% sevoflurane in 100% oxygen for 10 min, 100 mL of gas from the circuit was drawn into two 50 mL-syringes attached to the circuit Y-piece via a four-way stop-cock. Sevoflurane delivery was then discontinued and the system flushed with 100% oxygen until the residual sevoflurane concentration (SAM module) was 0% for 5 min. Oxygen delivery was then discontinued and the breathing system and ventilator continued to run in “closed circuit” mode. After 5 min in closed circuit mode with a measured sevoflurane concentration of 0%, the reserved 100 mL of 5% sevoflurane (5 mL of sevoflurane vapor) was returned to the system. Using the subsequent concentration of sevoflurane measured in the breathing system (0.2%–0.3%), it was possible to calculate the effective circuit volume into which the returned sevoflurane was diluted (VCIR = 1700–2500 mL).† As the nominal volume of the breathing system reported by the manufacturer is 2800 mL, exclusive of the breathing bag and not accounting for the volume displaced by the CO2 absorbent,5 an estimate of 2000 mL for VCIR seems reasonable.
Estimate of Effective Fresh Gas Flow
Based on the value of QFG/Vds from the early phase clearance experiment (0.035 s−1), QFG can then be estimated to be 2000 mL·0.035 s−1 = 70 mL/s (4200 mL/min).
Estimate of Effective Breathing System Volume
Using the estimate of QFG/Vds from the late phase measurements (0.024 min−1), Vds is estimated to be 160 L (4200 mL · min−1/0.024 min−1), which is obviously much greater than the actual system volume. This can be considered to be analogous to the case of highly lipid-soluble drugs, where the calculated volume of distribution may be much greater than the actual patient volume.
Estimate of Retained Mass of Sevoflurane
Given an initial late phase sevoflurane concentration of 60 ppm and Vds = 160 L, the actual total mass of sevoflurane retained by the Fabius breathing system and ventilator is estimated to be approximately 10 mL of vapor (approximately 100 mg).‡
Clearance of residual trace concentrations of sevoflurane required approximately 6 times as long for the Dräger Fabius anesthesia machine as for the Dräger Narkomed GS anesthesia machine. This difference was due to a combination of a higher initial C0 (61 vs 29 ppm) and a much longer t½ (29 vs 7 min) for the Fabius machine. The time required for the Fabius machine to achieve a residual sevoflurane concentration <5 ppm (105 min) was even longer than that previously reported for the Dräger Primus (Apollo) anesthesia machine (65 min)3; however, this difference may have been due to the higher concentration of sevoflurane used in the preparation procedure in our study (3% vs 2.5%).
There are several possible reasons for the slow clearance of trace sevoflurane from the Dräger Fabius anesthesia machine compared to older, conventional anesthesia machines. The first of these relates to the design of the Fabius breathing system and the operation of the piston ventilator. The unique operational features of the Fabius breathing system are similar to those of the Dräger Primus (Apollo) platform, which have been well described previously.3,4 To briefly recap (Fig. 8), during inspiration, fresh gas is diverted via the fresh gas decoupling valve through the CO2 absorber to either the breathing bag or the scavenging system; the delivered Vt therefore consists of gas from the inspiratory limb of the breathing circuit and the ventilator. During exhalation, the ventilator piston refills with a combination of exhaled gas from the patient, fresh gas, and gas from the breathing bag. When configured with minimal leak, as in our basic experimental setup, the system is extremely efficient in recirculating exhaled gases. With the exception of that drawn into the piston during its return stroke, most fresh gas delivered into the system is simply lost out of the scavenging system, either via fresh gas decoupling or during the expiratory phase after the piston has refilled. We considered the possibility that the observed delayed clearance of trace sevoflurane might have been due to this very efficiency and, therefore, intentionally decreased system efficiency in Experiment 1B. However, despite introducing a large leak and the resultant increased entrainment of fresh gas, clearance of trace sevoflurane was not accelerated.
A second unique feature of the Dräger Fabius anesthesia machine is the piston ventilator itself. With the exception of the positive and negative pressure relief valves, which open only under extraordinary circumstances, the piston itself is a dead end; the ventilator spill valve is located in the breathing system, where it is configured as an open bypass of the adjustable pressure limiting valve (Fig. 8). This is in contrast to older, bellows-type ventilators, where excess gas return to the bellows is vented via a spill valve located within the ventilator. We therefore considered the possibility that our observations were due to stagnant gas in the ventilator piston which was not being turned over. To increase ventilator washout, additional fresh gas was delivered directly into the ventilator drive hose or the ventilator itself in Experiments 2B and 2C. Although there was a modest acceleration of trace sevoflurane clearance under these conditions, discontinuation of the additional fresh gas flow resulted in a return of the residual concentrations to the level expected in the absence of the additional fresh gas flow. The most likely explanation for this observation is that the additional fresh gas served simply to “dilute” the concentration of sevoflurane; the actual residual concentration within the components of the breathing system was unchanged and the measured concentration reverted to its true value when the diluting gas was eliminated.
The third possible explanation for delayed clearance of sevoflurane from the Dräger Fabius anesthesia machine is that it is more soluble in the synthetic components of the Fabius breathing system.6–8 This is certainly compatible with our observations, in that the C0 for the Narkomed GS was less than that for the Fabius (29 vs 61 ppm). The longer clearance half-time seen in the Fabius implies that diffusion of sevoflurane out of its synthetic components is slower than that from the components of the Narkomed GS anesthesia machine, and is probably the rate-limiting step in the clearance process. It seems that components in both the ventilator and the breathing system are implicated, although the effect of the ventilator may be greater. Exclusion of the piston ventilator from the breathing system in Experiment 3 did decrease C0, compatible with a greater residual mass of sevoflurane in the ventilator; however, the actual rate constant observed in Experiment 3 was only about 25% greater than that in Experiment 1, suggesting that diffusion out of the components of the breathing system was only marginally faster than out of the components of the ventilator.
If we assume that diffusion into the synthetic components of the breathing system proceeds at the same rate as diffusion out, the observation that the C0 of sevoflurane was 60 ppm, rather than 3%, suggests that the synthetic components of the breathing system had become saturated with sevoflurane during the 2 h preparation period. Examination of Prinzhausen et al. results reveals concentrations in the 50–100 ppm range at 5 min, compatible with saturation of the components of the Dräger Primus (Apollo) anesthesia machine at a similar concentration of sevoflurane.3 Based on the effective Vds of 160 L estimated above (Results section) and the small actual volume of the synthetic components of the breathing system, the synthetic/gas partition coefficient of sevoflurane must be extremely high for the components of the Fabius breathing system.
Limitations of our data include failure to perform multiple replications of each experiment, failure to perform observations in all experiments to the achievement and maintenance of a residual concentration <5 ppm, and examination of only one halogenated anesthetic (sevoflurane). Although the limited number of runs for each experiment (one, two, or three) certainly makes it impossible to provide a meaningful confidence interval for our observations, the extreme reproducibility of the observations and regression coefficients of determination approaching unity suggest that our results can be viewed with confidence. Sevoflurane is the primary halogenated anesthetic used in our practice and was therefore chosen for our experimental protocol; similar results for isoflurane and sevoflurane were seen in the report from Prinzhausen et al.3, suggesting that the effect is independent of the anesthetic chosen. In addition to the above, the use of two different analytic techniques makes it difficult to produce an integrated model of both early and late phase clearance; future studies would benefit from identification of a single analytical method able to accurately measure instantaneous concentrations during both phases.
The minimum concentration of halogenated anesthetic posing a risk to the MH-susceptible patient is not known. The concentration targets set for our experiments (<10 ppm and <5 ppm) were based on previous descriptions of procedures for flushing anesthesia machines for the MH patient.1–4,9–11 It has been shown that MH-susceptible swine do not trigger in the presence of 10 ppm halogenated anesthetic. §To our knowledge, there are no reports of spontaneous triggering of MH in susceptible humans in the presence of trace halogenated anesthetic concentrations within the National Institute of Occupational Safety and Health limits (2 ppm).∥ Any attempt to pursue definitive human studies to address this question will suffer from obvious ethical limitations.
In the face of these observations and uncertainty regarding the maximum safe exposure limit for halogenated anesthetics for the MH-susceptible patient, several approaches can be considered when anesthetizing an MH-susceptible patient with the Dräger Fabius anesthesia machine. In the absence of evidence supporting the safety of exposure to residual sevoflurane concentrations >10 ppm, we do not believe that a brief flush to remove bulk concentrations of sevoflurane is a tenable solution. However, the time required to perform an adequate flush is likely to prove disruptive to a busy operating room schedule. Use of dedicated “clean” anesthesia machines for MH-susceptible patients entails significant monetary and manpower costs. Crawford et al. have advocated replacing the components of the Apollo breathing system contaminated with halogenated anesthetics with “clean” components before anesthetizing MH-susceptible patients.4 In contrast to the Apollo, where the breathing system can be replaced en bloc, it is our impression that replacement of the Fabius breathing system is a task best left to trained biomedical engineering support personnel.
The ability of activated charcoal to remove in excess of 90% of incident halogenated anesthetics from the inspired gases offers a fifth alternative. Interposition of an activated charcoal filter between the breathing system and the patient can reduce inspired sevoflurane (and presumably other halogenated anesthetic concentrations) to an acceptable level within 5 min. Although there is a monetary cost associated with the use of activated charcoal filters, that cost is more than offset by recovery of potential surgical time (opportunity cost) which would otherwise be lost during switching or flushing anesthesia machines. Our acquisition cost for the QED device ($39) represents about 2 min of operating room time as billed to the patient. Based on our results, a single QED device contains sufficient activated charcoal for at least a 6 h anesthetic. We are unable to estimate the maximum capacity of the QED device; however, after 6 h at a fresh gas flow of 2 L/min, residual sevoflurane concentrations in the Dräger Fabius anesthesia machine were <10 ppm. Therefore, even if the activated charcoal in the QED became completely saturated with halogenated anesthetic at some point after 6 h, the background concentration of halogenated anesthetic in the breathing system should be within the putative safe range.
We have therefore adopted a process referred to as the Five-Five-Five Flush. The Dräger Fabius anesthesia machine is first prepped by removing the vaporizers and replacing the CO2 absorbent. A clean circuit, including a QED on the inspiratory port and a 2 L artificial lung, is then installed and the system is flushed for 5 min with the QED in the “Off” position (Oxygen 10 L/min, Vt = 600 mL, rate = 10/min, I:E 1:2). The QED is then turned “On” and the flush is continued for a second 5 min. The machine is then ready to deliver anesthesia to the MH-susceptible patient, provided that the fresh gas flow is maintained at ≥ 10 L/min for the first 5 min of the anesthetic. After that, fresh gas flow can be adjusted as clinically indicated, maintaining a minimum of 2 L/min of fresh gas throughout the anesthetic. The QED remains on the inspiratory port in the “On” position until the conclusion of the anesthetic and is then discarded.
In conclusion, we have demonstrated that achieving residual sevoflurane concentrations in the <5–10 ppm range in the Dräger Fabius anesthesia machine using conventional flush procedures can require almost 2 h. We were unable to identify any method of significantly accelerating this process, suggesting that the slow clearance of sevoflurane from this machine is due to the materials used in its construction, rather than to its unique breathing system and ventilator design. Additional research will be required to identify the specific components and materials which are responsible for the slow clearance of halogenated anesthetics from the Dräger Fabius and Primus (Apollo) anesthesia machines. A breathing system adjunct containing activated charcoal reduced residual sevoflurane concentrations to <5 ppm within 10 min and maintained those concentrations for at least 6 h. Breathing system adjuncts containing activated charcoal offer a means of rapidly transforming an anesthesia machine contaminated with halogenated anesthetics into a “clean” machine for use in the MH-susceptible patient, and are an attractive alternative to the delay and inconvenience associated with conventional flushing procedures or the use of dedicated “clean” anesthesia machines.
The authors wish to thank Josh Harney of the Children's Hospital Medical Center department of Occupational Safety and Environmental Health for the loan of the Miran SapphIRe XL ambient air analyzer, and instruction in its use and calibration.
For purposes of analysis, the breathing circuit/ ventilator was treated as a single compartment (volume = Vds) containing halogenated anesthetic (initial concentration = C0) flushed with a continuous flow of fresh gas (flow rate = QFG; concentration = CFG); at equilibrium, the concentration in the compartment is equal to that in the fresh gas (CFG). The concentration of halogenated anesthetic in the compartment as a function of time is described by the following differential equation:
A general solution to Eq. 1 is:
For initial conditions C0 > 0 and CFG = 0 (washout), Eq. 2 reduces to:
For initial conditions C0 = 0 and CFG > 0 (wash-in), Eq. 2 reduces to:
These equations are analogous to first-order, single compartment pharmacokinetic models, with QFG and Vds being analogous to clearance and volume of distribution, respectively. The ratio QFG/Vds can be considered as the rate constant, or the inverse of the time constant, for equilibration of the system. The half-time (t½) of wash-in or washout can be calculated analogously to that of half-life for first-order kinetics:
Equilibration to 95% of CFG requires slightly more than four half-times, or three time constants.
1. Petroz GC, Lerman J. Preparation of the Siemens KION anesthetic machine for patients susceptible to malignant hyperthermia. Anesthesiology 2002;96:941–6
2. Schönell LH, Sims C, Bulsara M. Preparing a new generation anaesthetic machine for patients susceptible to malignant hyperthermia. Anaesth Intensive Care 2003;31:58–62
3. Prinzhausen H, Crawford MW, O'Rourke J, Petroz GC. Preparation of the Dräger Primus anesthetic machine for malignant hyperthermia-susceptible patients. Can J Anaesth 2006;53: 885–90
4. Crawford MW, Prinzhausen H, Petroz GC. Accelerating the washout of inhalational anesthetics from the Dräger Primus anesthetic workstation. Anesthesiology 2007;106:289–94
5. Drager Medical. Technical Data Fabius GS Operator's Instruction Manual. Chapter 12. Telford, PA: Drager Medical, 2004:142
6. Eger EI, Larson CP, Severinghaus JW. The solubility of halothane in rubber, soda lime and various plastics. Anesthesiology 1962;23:356–9
7. Lowe HI, Titel JH, Hagler KJ. Absorption of anesthetics by conductive rubber in breathing circuits. Anesthesiology 1971;34:283–9
8. Targ AG, Yasuda N, Eger EI. Solubility of I-653, sevoflurane, isoflurane, and halothane in plastics and rubber composing a conventional anesthetic circuit. Anesth Analg 1989;69: 218–25
9. Beebe JJ, Sessler DI. Preparation of anesthesia machines for patients susceptible to malignant hyperthermia. Anesthesiology 1988;69:395–400
10. Ritchie PA, Cheshire MA, Pearce NH. Decontamination of halothane from anesthetic machines achieved by continuous flushing with oxygen. Br J Anaesth 1988;60:859–63
11. McGraw TT, Keon TP. Malignant hyperthermia and the clean machine. Can J Anaesth 1989;36:530–2
*http://medical.mhaus.org/index.cfm/fuseaction/Content.Display/PagePK/MedicalFAQs.cfm; last accessed March 21, 2008.
†The concentration of sevoflurane (CSEVO) in the breathing system after the addition of 5 mL of sevoflurane vapor (VSEVO) is:
Rearranging and substituting the experimental data,
‡The mass of sevoflurane is the concentration multiplied by the effective volume (160 L · 60 × 10−6 = 9.6 × 10−3 L = 9.6 mL). Based on the molar gas volume at standard temperature and pressure (22.4 L/mole), 9.6 × 10−3 liters of sevoflurane vapor represents 4.3 × 10−4 moles, or 0.086 gm (molecular weight of sevoflurane 200 gm/mole).
§Maccani RM, Wedel DJ, Kor TM, Joyner MJ, Johnson MF, Hall BA. The effect of trace halothane exposure on triggering malignant hyperthermia in susceptible swine. Anesth Analg 1996;82:S287.
∥http://www.osha.gov/pls/oshaweb/owadisp.show_document?_p_table=FACT_SHEETS&p_id=128; last accessed March 21, 2008.© 2008 International Anesthesia Research Society