Fresh gas flow is not the only determinant of volatile agent consumption: a multi-centre study of low-flow anaesthesia{dagger}

J. F. Coetzee*,1 and L. J. Stewart2

1Department of Anesthesiology, Faculty of Medicine, University of Stellenbosch, PO Box 19063, Tygerberg 7505, South Africa. 2Technology and Business Development Group, Medical Research Council of South Africa, PO Box 19070, Tygerberg 7505, South Africa*Corresponding author

*Supplementary material is available to subscribers with the on-line version of the journal at the journal website.

Accepted for publication: August 30, 2001


    Abstract
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Methods. Seven academic centres studied 302 patients, using desflurane, enflurane, halothane, or isoflurane using circle-systems and Dräger Julian anaesthetic machines, with fresh gas flows (V·F) of 3, 1, and 0.5 litre min–1. Volatile agent partial pressures in the breathing system were recorded and agent consumptions measured by weighing.

Results. At these flows, desflurane consumption depended on V·F. In contrast, halothane consumption was not influenced by V·F. Isoflurane and enflurane showed differences in consumption between flows of 0.5 and 3 litre min–1. Stepwise linear regression suggested that besides V·F, other factors influenced consumption of the more soluble agents (sex, age, weight, height, altitude, and temperature). The partial pressure ratios were independent of V·F for desflurane (end-tidal to fresh gas=0.8), but the ratios of the more soluble agents varied with V·F (end-tidal to fresh gas=0.3–0.7).

Conclusions. At V·F that involves significant re-breathing, consumption of soluble agents depends only partially on V·F. These results can be explained using Mapleson’s hydraulic analogue model.

Br J Anaesth 2002; 88: 46–55

Keywords: anaesthesia, audit; anaesthetics, volatile; equipment, anaesthesia machines


    Introduction
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
The rising costs of inhaled anaesthetic drugs, and especially the newly introduced agents, sevoflurane and desflurane, have influenced increasing numbers of anaesthesiologists to use decreased fresh gas flows (V·F) in circle-absorber breathing systems. Studies of the economics of anaesthesia have used mathematical models to predict anaesthetic consumption,13 and it is generally accepted that lower V·F allow less consumption of anaesthetic agent.47 While this principle cannot be refuted when comparing high V·F with low V·F, few studies have determined how much anaesthetic is saved by using low V·F in comparison to moderate V·F in clinical practice.7 8

The purposes of this study were: (i) to determine consumption of volatile anaesthetic agent using circle-absorber systems at moderate, low and minimal V·F (for the purpose of this study moderate fresh gas flow is defined as a flow of 3 litre min–1; low fresh gas flow 1.0 litre min–1 and minimal fresh gas flow 0.5 litre min–1). (ii) To evaluate the differences between vaporizer settings and end-tidal agent partial pressures that are achieved at these flows. Clinicians administered the anaesthetics as they would have done in clinical practice, according to their judgement of patient requirements.


    Methods
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
The study was a multi-centre trial at all seven university Faculties of Medicine in South Africa. Studies used standard anaesthetic machines equipped with carbon dioxide absorber circle breathing systems and vaporizers outside the circuit (Drager, model ‘Julian’, Dragerwerk, Lübeck, Germany). Written informed consent was obtained from 302 adult patients aged 18–65 yr, of ASA I or II status, who were to undergo surgery. Using tables of random numbers, patients were allocated to one of four groups who received halothane or enflurane or isoflurane or desflurane. Each group was randomly subdivided into three subgroups who received one of the following V·F during maintenance of anaesthesia (litre min–1): 3, 1, or 0.5. There were, therefore, a total of 12 groups. Exclusion criteria included significant cardiac, pulmonary, renal, or neurolgical disease, pregnancy, previous adverse response to inhaled agents, patients undergoing regional or total i.v. anaesthesia, surgery with special requirements (e.g. cardiothoracic, neurosurgical), and procedures with an expected duration of less than 30 min.

No pre-medication was given. Monitoring provided by the Julian anaesthetic machine included ECG, pulse oximetry, nasopharyngeal temperature, non-invasive arterial pressure, capnography, inspired and expired gas partial pressure measurement (oxygen, nitrous oxide, volatile agent), airway pressure, tidal volume, ventilatory frequency, and minute volume. Neuromuscular transmission was monitored by train-of-four nerve stimulation.

After pre-oxygenation with 100% oxygen (V·F 6 litre min–1) and establishing an i.v. infusion using local anaesthesia, fentanyl 3.0 µg kg–1 was administered intravenously to suppress the response to tracheal intubation. Induction of anaesthesia was with thiopentone 3–5 mg kg–1 and vecuronium 0.08 mg kg–1 followed by manually controlled ventilation by mask with V·F settings of nitrous oxide:oxygen (3.0:1.5). Anaesthetic vapour was introduced after tracheal intubation using Table 1 as an initial guide to the dosage that could be adjusted according to patient response. These initial vaporizer settings were based upon the following calculation: FEAN = (1.3–FEN2O) x MAC, where FEAN = fractional expired partial pressure of the volatile agent in question, FEN2O = fractional expired concentration of nitrous oxide (1.3 x MAC was assumed to be the volatile agent expired partial pressure sufficient to suppress movement on skin incision in 95% of patients.911 For a partial pressure of 66 kPa for nitrous oxide these initially targeted expired partial pressures (kPa) were: halothane 0.5, isoflurane 0.74, desflurane 3.8, enflurane 1.2.


View this table:
[in this window]
[in a new window]
 
Table 1 Initial settings for fresh gas flow and vaporizers after tracheal intubation. These settings served as an initial guide and could be changed as determined by patient response
 
This initial wash-in period lasted 10 min to allow for initial rapid uptake of nitrous oxide and elimination of nitrogen, after which V·F was reduced with initial vaporizer and nitrous oxide:oxygen settings as shown in Table 1. If the inspired partial oxygen partial pressures decreased to less than 30 kPa during maintenance of anaesthesia, changes were allowed by increasing the oxygen V·F and simultaneously reducing the nitrous oxide V·F as follows: during low flow (1 litre min–1) by 50 ml min–1 for each gas and during minimal flow (0.5 litre min–1) by 100 ml min–1 for each gas.

If depth of anaesthesia had to be changed, the following procedures were followed. If a rapid change was not required, the vaporizer setting was adjusted while maintaining the V·F. To increase inspired partial pressure, ‘overpressure’ was used if deemed necessary and to decrease inspired partial pressure, the vaporizer was temporarily shut off as needed. If a rapid change was required, the V·F was increased to 3 litre min–1 in addition to adjusting the vaporizer setting. During low and minimal V·F the vaporizer was shut off towards the end of the procedure (15–20 min) to allow the inhaled vapour partial pressure to decrease slowly. Additional neuromuscular blocking agent was administered as necessary.

Residual neuromuscular block was antagonized at the start of skin closure using neostigmine (0.04 mg kg–1) mixed with glycopyrrolate (0.006 mg kg–1). The vaporizer and nitrous oxide supply was shut off and the V·F increased to 6 litre min–1 of oxygen after the last skin stitch had been applied. The ventilatory frequency was reduced to allow the PE'CO2 to increase to a maximum of 6.5 kPa to initiate breathing.

The following were automatically recorded to computer disk at 5-s intervals, from the time that pre-oxygenation began until the patient awoke:

•Fresh gas flow

•Oxygen concentration in the fresh gas

•Inspired and end-tidal partial pressures of: volatile agent, nitrous oxide, oxygen, carbon dioxide

•Tidal volume, ventilatory frequency

•Oxygen saturation (pulse oximetry)

•Heart rate (from the pulse oximeter)

•Non-invasive arterial pressure (mean) measurements

•Nasopharyngeal temperature

Times of adjustments to V·F and vaporizer settings were recorded manually.

After completion of the procedure the mass (g) of volatile agent was measured using mass balances capable of measuring to 0.1 g.*

Calculations
The data were imported to a computer spreadsheet (Microsoft Excel 97) for analysis. For each patient the arithmetic mean of the end-tidal anaesthetic partial pressures that had been recorded every 5 s were calculated for the duration of volatile agent administration (MFA). The following anaesthetic agent partial pressure ratios were calculated and plotted at 5 s intervals: end-tidal to inspired (FA/FI); inspired to fresh gas (FI/FD); end-tidal to fresh gas (FA/FD). These ratios were noted at the following times: at the end of the 10-min wash-in period and at a later time during anaesthetic maintenance when these ratios were not changing (i.e. at ‘steady state’). Mean rate of anaesthetic agent consumption was calculated by dividing the total mass used by the time that the anaesthetic vaporizer was switched on.

Data analysis
Statistical analysis was done with Statistical Analysis Support System (SAS), Version 6 Release 12, SAS Institute Inc., BMDP Statistical Software Inc. As the data were not always normally distributed, non-parametric, distribution-free tests were performed. Comparisons between two groups were done using the Mann–Whitney test or where appropriate, the Wilcoxon signed ranks paired test. Multiple group comparisons were done using Kruskal– Wallis one-way analysis of variance followed by post hoc multiple comparisons (Tukey’s test). An alpha-value of 0.05 or less was regarded as indicating statistical significance.

To identify factors that influenced rate of consumption of anaesthetic agent, stepwise multiple regression was performed using the following as independent variables: fresh gas flow, duration of agent administration, the inverse of the square root of time (1/{surd}t), age, mass, height, sex, ASA status, MFA, body temperature change, FA/FD. A backward selection method was used with a threshold F-ratio of 4 for entry-to/removal-from the linear regression model.


    Results
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
Patient details are shown in Table 2. In total, 302 patients participated in the study and their distributions among the 12 groups were similar. There were no differences in age, mass, height, body mass, body mass index, MFA, temperature change, or duration of anaesthetic administration. Distribution of ASA status was equal among the groups and the male:female ratios did not differ. There were equivalent numbers of patients at sea level and at altitude.


View this table:
[in this window]
[in a new window]
 
Table 2 Details of the 302 patients studied, grouped according to fresh gas flow and drug administered. Results are presented as mean (SD). n=number of patients in the group; V·F=fresh gas flow; BMI=body mass index; duration=duration of volatile agent administration
 
Table 3 summarizes the significant differences in rate of agent consumption, and the partial pressure ratios (FA/FD, FI/FD, and FA/FI) that were found between the three V·F groups. Table 4 presents the results in detail. At the flows studied, desflurane consumption rate depended on V·F. In contrast, the halothane consumption rate was not influenced by V·F. Isoflurane and enflurane showed differences in consumption rates when flows of 0.5 and 3 litre min–1 were compared. For these two agents there were no differences between flows of 0.5 vs 1 litre min–1, and between 1 vs 3 litre min–1. An incidental finding was that patients at altitude (>1300 m) required significantly greater end-expired partial pressures of volatile anaesthetic agent (MFA) than those at sea level (Fig. 015F1).


View this table:
[in this window]
[in a new window]
 
Table 3 Summary of statistical differences (P<0.05) in agent consumption and partial pressure ratios between the three fresh gas flow groups. V·F=fresh gas flow. Ratios of volatile agent partial pressures during maintenance of anaesthesia: FA/FI=end-tidal to inspired; FI/FD=inspired to fresh gas flow; FA/FD=end-tidal to fresh gas flow. Numbers refer to fresh gas flow groups: 0.5, 1.0, and 3 litre min–1
 

View this table:
[in this window]
[in a new window]
 
Table 4 Detailed results of measurements for the four volatile agents. Patients are grouped according to administered agent and fresh gas flows. V·F=fresh gas flow; MFA=mean expired partial pressure obtained by averaging the end-tidal partial pressures of anaesthetic agent (kPa) recorded every 5 s. Ratios of volatile agent partial pressures: FA/FI=end-tidal to inspired. FI/FD=inspired to fresh gas flow. FA/FD=end-tidal to fresh gas flow. Change in temperature=difference in body temperature between beginning and end of the procedure. *Significant differences between V·F of 0.5 and 1 litre min–1. +Significant differences between V·F of 0.5 and 3 litre min–1. {dagger}Significant differences between V·F of 1 and 3 litre min–1. {ddagger}Significant differences between ‘wash-in’ values and values achieved during maintenance of stable anaesthesia.
 


View larger version (17K):
[in this window]
[in a new window]
 
Fig 1 Box and whisker plots of mean expired partial pressures measured in patients at sea level and altitude. Mean FA=mean expired partial pressure obtained by averaging the end-tidal partial pressures of anaesthetic agent (kPa) recorded every 5 s. Sea=sea level; Alt=altitude greater than 1300 m. Explanation of the plots: the ends of the ‘box’ illustrate the 25th and 75th percentiles and the horizontal line within the box depicts the median value. The ‘notch’ represents the 95% confidence interval of the median. The ‘whiskers’ illustrate the range.

 
On comparing the partial pressure ratios after the intitial, 10-min, high-V·F wash-in period with those during anaesthetic maintenance, the following were found (Tables 3 and 4). Desflurane: no significant differences in FA/FI, FI/FD, and FA/FD. This indicates that the wash-in period was sufficient to have achieved steady state, allowing reduction of V·F with little or no adjustment to vaporizer settings. For the other three agents, there were small, but statistically significant differences in FA/FI indicating that, the 10-min wash-in phase allowed reduction of V·F, if vaporizer settings were increased (as revealed by increased FI/FD ratios). With regard to the FI/FD ratios, there were differences in the low and minimal flow groups. These differences confirmed that vaporizer settings had to be increased when reducing V·F. FA/FD differences were broadly similar to FI/FD differences with the same interpretation.

Stepwise linear regression
Factors associated with agent consumption were sought by stepwise linear regression modelling. Results for volatile agent total consumption and consumption rate are presented in Tables 5 and 6. V·F and duration of administration were included in the regression models for total consumption for all four agents. In addition, enflurane total consumption was influenced by MFA, FA/FD at steady state, patient height, and change in body temperature. Total halothane consumption was also affected by patient height, sex, age, and ASA status; isoflurane by altitude, temperature change, and body mass.


View this table:
[in this window]
[in a new window]
 
Table 5 Results of stepwise linear regression for consumption (g) of volatile anaesthetic. V·F=fresh gas flow (0.5, 1.0, and 3.0 litre min–1). MFA=mean expired partial pressure obtained by averaging the end-tidal partial pressures of anaesthetic agent (kPa) recorded every 5 s. Duration=duration of volatile anaesthetic agent administration (min). Altitude was coded as 0=sea level, 1=>1300 m. Sex was coded as male=1, female=0. FA/FD=volatile agent ratio of end-tidal to fresh gas flow partial pressures at ‘steady state’. Height=patient height (cm). Temp. change=change in body temperature (°C) between beginning and end of procedure. Weight=body mass (kg)
 

View this table:
[in this window]
[in a new window]
 
Table 6 Results of stepwise linear regression for consumption rate (g min–1) of volatile anaesthetic agents: V·F=fresh gas flow (0.5, 1.0, and 3.0 litre min–1). MFA=mean expired partial pressure obtained by averaging the end-tidal partial pressures of anaesthetic agent (kPa) recorded every 5 s. Duration=duration of volatile anaesthetic agent administration (min). 1/{surd}t=inverse of the square root of time (min–1). Altitude was coded as 0=sea level, 1=>1300 m. Sex was coded as male=1, female=0. FGF, fresh gas flow
 
V·F affected the consumption rate of desflurane, enflurane, and isoflurane, but not halothane. MFA influenced the consumption rates of desflurane and halothane. 1/{surd}t affected the consumption rates of enflurane and isoflurane. The halothane consumption rate was also affected by duration of anaesthesia and isoflurane rate by patients’ sex and altitude.


    Discussion
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
The finding that V·F did not influence halothane consumption is unexpected. Furthermore desflurane was the only agent where usage was affected by V·F settings. Consumption of moderately soluble drugs (enflurane and isoflurane) was less influenced by V·F than in the case of desflurane. There was no clear distinction in the consumption rates of enflurane and isoflurane between V·F of 0.5 and 1.0 litre min–1 and between 1 and 3 litre min–1. In this study, all the anaesthetic machines were identical (Dräger, model ‘Julian’) so that equipment of features (e.g. breathing circuit time-constants, inflow and outflow placements, mode of excess gas venting) could not affect agent consumption and partial pressure dynamics.12 13

Recognizing that all three V·F allowed significant re-cycling of exhaled gases, these apparent disparities can be explained as follows. To maintain a desired end-tidal (alveolar) partial pressure it is necessary to compensate for tissue uptake. For halothane this is considerable14 because its high tissue solubility increases uptake into peripheral organs (e.g. muscle) and furthermore a considerable proportion of halothane undergoes hepatic metabolism.14 15 It is therefore necessary to deliver a minimum amount of drug per unit time to compensate for uptake into the blood. This can be effected by either supplying the inhaled drug using a greater V·F containing a partial pressure that is equal to (or close to) the desired inhaled value, or by using a greater partial pressure at low V·F. Either way, a certain rate of delivery of agent to the breathing system is needed to maintain inhaled partial pressures.

This phenomenon may be explained using Mapleson’s hydraulic, pharmacokinetic model for volatile agent uptake.16 In Figure 015F2, uptake and distribution of volatile anaesthetic vapour is shown by a series of interconnected cylinders where the fluid heights represent vapour partial pressures. Volumes of distribution are depicted by the base areas, amounts of drug by fluid volumes, and drug clearances by the diameters of the various PIPES. Fresh gas partial pressure is shown by the height of the cylinder at the extreme left (FD), which is full at all times and V·F is represented by the diameter of its outlet. Inspired partial pressure is portrayed by the height of fluid in cylinder FI from which excess vapour may spill over during high V·F (analagous to the spill-valve of a ventilator). This spill-over is considerable during high V·F, minimal during low V·F and absent during closed-circuit anaesthesia. Two large peripheral compartments (muscle, fat) drain agents away from the lungs (FA), and viscera (including the brain). To maintain a constant level in the brain, it is necessary to supply drug to FA from FI to prevent the level in FA from falling. Removal of drug from FI can in turn be replaced by supplying drug from FD using a high V·F containing a low partial pressure and this is represented in the upper illustration of Figure 015F2 by a short cylinder, FD delivering fluid to FI via a wide diameter pipe of high conductance. On changing to low flow (thin outlet pipe from FD), the same amount of drug per unit time must be delivered to the breathing system and lungs by an increased pressure in FD. This is shown in the lower illustration in Figure 015F2 by the tall cylinder, FD delivering fluid to FI via a small diameter pipe of low conductance. In this manner the fluid level in FI and, therefore, in FA and the viscera is maintained. The dotted lines in the lower illustration in Figure 015F2 show how the partial pressures in FI and FA decrease if the partial pressure in FD is not increased.



View larger version (24K):
[in this window]
[in a new window]
 
Fig 2 Hydraulic analogue model of uptake and distribution of volatile anaesthetic agent from a circle breathing system (modified from Mapleson16). Agent partial pressures are represented by the heights of fluid in the cylinders. The anaesthetic breathing circuit is the two cylinders on the left (FD and FI) that deliver drug to the central cylinder, the alveoli (FA), from which drug is distributed to peripheral compartments. Vapour is introduced into the system from the cylinder on the far left (FD), which is kept full at all times. The partial pressure of the delivered vapour is represented by the height of cylinder FD and that of the inhaled vapour by the height of fluid in cylinder FI. Excess vapour vented from the breathing system is depicted by spill-over from cylinder FI. Fresh gas flow rate and excess vapour spill-over are represented by the associated arrows. FD=delivered (fresh gas) partial pressure; FI=inspired partial pressure; FA=alveolar partial pressure; FGF=fresh gas flow. The upper illustration depicts vapour delivered at a low partial pressure (short cylinder FD) by a high fresh gas flow rate (thick outlet pipe from FD) with large amounts of excess agent drug spilling out of the breathing system at FI. The lower illustration depicts vapour delivered at a high partial pressure (tall cylinder FD) at a low fresh gas flow rate (narrow outlet pipe from FD) with little excess agent being vented. Inhaled and alveolar partial pressures are thereby maintained at the same levels as in the upper illustration. The dotted lines in the lower illustrate that if the partial pressure of FD is reduced while maintaining a low fresh gas flow, the inspired (and therefore alveolar) partial pressures decrease to below targeted values.

 
In the case of the lower illustration in Figure 015F2 where the spill-over is minimal, changes to the properties of the peripheral cylinders and the PIPES connecting them to FA will have a large influence on ‘drug’ disposition. For example, an increase in uptake by the peripheral cylinders will decrease the level in FA and, therefore, in the viscera (brain). This can be compensated for by increasing flow to FA by increasing the height of the level in FI. This in turn can be brought about by increasing the height of fluid in cylinder FD to increase delivery of drug to FI, with no extra spill-over occurring at FI. It is under such circumstances that ‘patient factors’ become major determinants of drug consumption during low-flow anaesthesia.

We suggest that with regard to halothane in this study, continuing uptake of drug was so rapid that at the three V·F studied, there was little venting (spill-over) of vapour via expired gas to the atmosphere, and patient factors became important determinants of drug consumption. This conclusion is strengthened by the following considerations.

FA/FI was 0.75 at all V·F, indicating that during anaesthetic maintenance, transfer from alveoli to blood was a continuing process. The re-circulation of alveolar gas containing reduced amounts of halothane vapour to the breathing system resulted in decreased FI/FD and FA/FD ratios, and these varied with V·F (Tables 3 and 4).

•At low and minimal flows, halothane FA/FI and FA/FD ratios were small (Table 4) so that the vented gas contained reduced amounts of vapour. (At reduced V·F, gas vented from a circle-system is composed mainly of expired gas).

•No differences in halothane consumption rate were found at the three V·F.

•Stepwise linear regression indicated that although V·F did indeed influence halothane usage, patient descriptors were also strongly associated.

At the other extreme, continuing uptake of desflurane, a much less soluble drug is markedly decreased during anaesthetic maintenance. Therefore, V·F strongly influences desflurane consumption at all flow rates because surplus is lost to the atmosphere. This conclusion is strengthened by the following considerations.

FA/FI was 0.94 at all V·F, indicating that at ‘steady state’, transfer of desflurane from alveoli to blood was small.

FI/FD and FA/FD ratios were high (0.8 and 0.75, respectively), indicating that re-circulated exhaled gas contained concentrations of desflurane vapour that were close to the inspired values. Furthermore, these ratios were uninfluenced by V·F (Tables 3 and 4). These high FA/FI and FA/FD ratios, resulted in vented gas containing significant amounts of vapour even at low and minimal flows.

•Significant differences in desflurane consumption rate were measured at the three V·F.

•Stepwise linear regression revealed that V·F was a major determinant of desflurane consumption but patient descriptors were not associated with drug usage.

The drugs of intermediate solubility, isoflurane and enflurane, lie between the two extremes. Agent consumption was significantly less when comparing V·F of 0.5 and 3.0 litre min–1, but there were no differences in consumption between V·F of 0.5 and 1.0 litre min–1 and between 1.0 and 3.0 litre min–1. Consumption of these two drugs was also, like halothane, associated with certain patient descriptors (Tables 5 and 6). Lowe17 has shown that during totally closed-circuit anaesthesia, volatile agent consumption rate is directly related to the square root of time in accordance with the exponentially decreasing uptake of agent over time. This variable played a role in the regression models for enflurane and isoflurane under conditions of partial re-breathing, which suggests that at these low flows, venting of expired vapour from the circle-system was minimal.

Mathematical models have been useful in predicting theoretical uptake and consumption of various agents for comparative purposes,13 but they assume a constant alveolar partial pressure requirement and fixed patient variables that influence agent uptake (e.g. cardiac output, ventilation). In reality, these factors vary during clinical anaesthesia. For example, alveolar partial pressure requirements may increase during surgical stimulations or decreased by supplementary drugs such as nitrous oxide, opioids, and alpha-receptor agonists. Stepwise linear regression revealed statistically significant associations between volatile agent consumption and various patient descriptors. However, it cannot be construed from our data that these are causative. Our results merely indicate that other variables besides V·F influence agent consumption and prospective studies will be necessary to identify the patient characteristics and surgical conditions that determine drug uptake.

Circle-systems are often used at inappropriately high V·F and repeated recommendations have been made for the reduction of V·F in order to reduce costs. 2 47 1821 This study confirms the well known phenomenon that during low V·F there are discrepancies between vaporizer settings and inspired (or end-tidal) vapour partial pressures22 23 which is more marked with soluble agents and at low flows. Low flow anaesthesia should, therefore, be practised using agent monitors. In developing countries, halothane is the mainstay of anaesthesia and agent monitors are seldom available. This study, suggests that a V·F of less than 3 litre min–1 should not be used without agent monitors. Furthermore, a potential danger of halothane overdose is present when using low V·F without an agent monitor, if the anaesthetist increases V·F and forgets to reduce the vaporizer setting.

In contrast to the more soluble drugs, FA/FD and FI/FD for desflurane were similar, predictable, and independent of V·F (0.7–0.8). Furthermore this study supports previous work that desflurane usage can be reduced when given using minimal V·F.1 6 23 24 Considering the predictablity of desflurane partial pressures, perhaps agent monitoring is not mandatory because the setting of the precision vaporizer can be used to indicate the alveolar partial pressure.1 This is not recommended because under- or overdosage could result if the vaporizer was faulty.

A wash-in period of 10 min at high V·F (4.5 litre min–1) was used. This is less than the usually recommended 15–20 min for minimal flow.25 If this were inadequate, a marked difference between FA/FI at the end of the wash-in period and during anaesthetic maintenance at low V·F would have occurred. No difference was found with desflurane. Differences were found with the other three drugs (Table 4) but these differences were small and were easily compensated for by adjustments to the vaporizers. We conclude that even for the most soluble drug, halothane, a 10-min wash-in period was sufficient. For desflurane, an even shorter wash-in period will suffice 3 14 23 24 26 with even greater cost savings.

Three of the participating hospitals were in coastal cities whereas the remainder were situated inland, more than 1300 m above sea level at reduced atmospheric pressures (Table 7). It was therefore possible to study the influence of altitude on MFA and on agent consumption. On pooling and comparing the patients’ mean end-tidal partial pressures of volatile agents according to whether they were anaesthetized at sea level or altitude, it was found that those who had undergone surgery at altitude required greater end-tidal partial pressures of all four volatile agents. In the present study, anaesthesia was with a combination of approximately 66% nitrous oxide and a volatile agent. Furthermore, anaesthesiologists in the seven different institutions were requested to administer anaesthesia as they normally would in their everyday practices, titrating volatile agent partial pressures according to patient requirements. Our findings therefore possibly reflect the reduced effectiveness of nitrous oxide at altitude. It has been demonstrated that at decreased ambient pressures, nitrous oxide causes less analgesia27 and is a less efficient anaesthetic,28 and nitrous oxide is also probably less effective as an adjuvant. As far as we are aware, this may be the first documented evidence for this hypothesis, considering that the only other anaesthetic adjuvant administered was a small dose of fentanyl.


View this table:
[in this window]
[in a new window]
 
Table 7 Sites of the academic hospitals where the study was conducted, heights above sea level and atmospheric pressures. Atmospheric pressures are the mean of the monthly averages during 1998 (by courtesy of the South African Weather Bureau)
 
We conclude that under clinical circumstances, using moderate and low V·F, consumption of relatively soluble volatile anaesthetic agents depends only partially on V·F and that monitoring agent partial pressures within the breathing system is highly desirable. On the other hand, the FA/FD of desflurane is predictable and independent of V·F. The contribution of nitrous oxide to anaesthesia is weakened at altitude.


    Acknowledgements
 
The roles played by the following persons are gratefully acknowledged. Sonja Swanefelder (MRC) for statistical analysis and advice. Study design, patient care, and data gathering: Professor J. Diedericks (University of Bloemfontein), Professor J. Viljoen, Dr R. Nieuwveld (University of Cape Town), Professor A. Rantloane (Medical University of South Africa), Dr C. Daniel (University of Natal-Kwazulu), Professor P. Fourie (University of Pretoria), and Dr. A. Smit (University of Witswatersrand).


    References
 Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 
1 Weiskopf RB, Eger EI. Comparing the costs of inhaled anesthetics. Anesthesiology 1993; 79: 1413–8[ISI][Medline]

2 Baxter AD. Low and minimal flow inhalational anaesthesia. Can J Anaesth 1997; 44: 643–52[Abstract]

3 Mapleson WW. The theoretical ideal fresh-gas flow sequence at the start of low-flow anaesthesia. Anaesthesia 1998; 53: 264–72[ISI][Medline]

4 Body SC, Fanikos J, DePeiro D, Philip JH, Segal BS. Individualized feedback of volatile agent use reduces fresh gas flow rate, but fails to favorably affect agent choice. Anesthesiology 1999; 90: 1171–5[ISI][Medline]

5 Cotter SM, Petros AJ, Dore CJ, Barber ND, White DC. Low-flow anaesthesia. Practice, cost implications and acceptability. Anaesthesia 1991; 46: 1009–12[ISI][Medline]

6 Eger EI. Economic analysis and pharmaceutical policy: a consideration of the economics of the use of desflurane. Anaesthesia 1995; 50 (Suppl): 45–8[ISI][Medline]

7 Pedersen FM, Nielsen J, Ibsen M, Guldager H. Low-flow isoflurane-nitrous oxide anaesthesia offers substantial economic advantages over high- and medium-flow isoflurane-nitrous oxide anaesthesia. Acta Anaesthesiol Scand 1993; 37: 509–12[ISI][Medline]

8 Logan M. Breathing systems: effect of fresh gas flow rate on enflurane consumption. Br J Anaesth 1994; 73: 775–8[Abstract]

9 de-Jong RH, Eger EI. MAC expanded: AD50 and AD95 values of common inhalation anesthetics in man. Anesthesiology 1975; 42: 384–9[ISI][Medline]

10 Schwilden H, Stoeckel H. Quantitative EEG analysis during anaesthesia with isoflurane in nitrous oxide at 1.3 and 1.5 MAC. Br J Anaesth 1987; 59: 738[Abstract]

11 Stevens WD, Dolan WM, Gibbons RT, et al. Minimum alveolar concentrations (MAC) of isoflurane with and without nitrous oxide in patients of various ages. Anesthesiology 1975; 42: 197–200[ISI][Medline]

12 Harper M, Eger EI. A comparison of the efficiency of three anesthesia circle systems. Anesth Analg 1976; 55: 724–9[Abstract]

13 Zbinden AM, Feigenwinter P, Hutmacher M. Fresh gas utilization of eight circle systems. Br J Anaesth 1991; 67: 492–9[Abstract]

14 Yasuda N, Lockhart SH, Eger EI,II, et al. Kinetics of desflurane, isoflurane and halothane in humans. Anesthesiology 1991; 74: 489[ISI][Medline]

15 Carpenter RL, Eger EI,II, Johnson BH, Unadkat JD, Sheiner LB. Pharmacokinetics of inhaled anesthetics in humans: measure ments during and after simultaneous administration of enflurane, halothane, isoflurane, methoxyflurane and nitrous oxide. Anesth Analg 1986; 65: 575–82[Abstract]

16 Mapleson WW. Pharmacokinetics of inhaled anaesthetics. In: Prys-Roberts C, Hug, CC jr., eds. Pharmacokinetics of Anaesthesia. Oxford: Blackwell Scientific Publishers, 1984; 89–111

17 Lowe HJ, Ernst EA. The Quantitative Practice of Anesthesia: The Use of Closed Circuit. Baltimore: Williams & Wilkins, 1981

18 Bailey CR, Ruggier R, Cashman JN. Anaesthesia: cheap at twice the price? Staff awareness, cost comparisons and recommend ations for economic savings. Anaesthesia 1993; 48: 906–9[ISI][Medline]

19 Becker KE jr, Carrithers J. Practical methods of cost containment in anesthesia and surgery. J Clin Anesth 1994; 6: 388–99[ISI][Medline]

20 McKenzie AJ. Reinforcing a ‘low flow’ anaesthesia policy with feedback can produce a sustained reduction in isoflurane consumption. Anaesth Intens Care 1998; 26: 371–6[ISI][Medline]

21 Watcha MF, White PF. Economics of anesthetic practice. Anesthesiology 1997; 86: 1170–96[ISI][Medline]

22 Baum J. Low Flow Anaesthesia: The Theory and Practice of Low Flow, Minimal Flow and Closed System Anaesthesia, 1st Edn. Oxford: Butterworth-Heinemann, 1996

23 Hargasser S, Hipp R, Breinbauer B, Mielke L, Entholzner E, Rust M. A lower solubility recommends the use of desflurane more than isoflurane, halothane and enflurane under low-flow conditions. J Clin Anesth 1995; 7: 49–53[ISI][Medline]

24 Nel MR, Ooi R, Lee DJ, Soni N. New agents, the circle system and short procedures. Anaesthesia 1997; 52: 364–7[ISI][Medline]

25 Virtue RW. Minimal-flow nitrous oxide anesthesia. Anesthesiology 1974; 40: 196–8[ISI][Medline]

26 Gowrie-Mohan S, Muralitharan V, Lockwood GG. The estimation of inspired desflurane concentration in a low-flow system. Anaesthesia 1996; 51: 904–7[ISI][Medline]

27 James MFM, Manson EDM, Dennett DE. Nitrous oxide analgesia and altitude. Anaesthesia 1982; 37: 285[ISI][Medline]

28 Powell JN, Gingrich TF. Some aspects of nitrous oxide analgesia at an altitude of one mile. Anesth Analg 1969; 48: 680[Medline]