| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
Laboratories of
1
Toxicology and
2
Medical Biochemistry and Clinical Analysis, Ghent University, Harelbekestraat 72, B-9000 Ghent, Belgium.
3
Department of Anesthesia, University Hospital, Ghent
University, De Pintelaan 185, B-9000 Ghent, Belgium.
a Author for correspondence. Fax 32-9-264-81-97; e-mail
Andre.DeLeenheer{at}rug.ac.be.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Methods: Compound A concentrations were measured by fully automated capillary gas chromatography–mass spectrometry with cryofocusing. Calibrators of compound A in the vapor phase were prepared from liquid volumetric dilutions of stock solutions of compound A and sevoflurane in ethyl acetate. 1,1,1-Trifluoro-2-iodoethane was chosen as an internal standard. The resulting quantitative method was fully validated.
Results: A linear response over a clinically useful concentration interval (0.3–75 µL/L) was obtained. Specificity, sensitivity, and accuracy conformed with current analytical requirements. The CVs were 4.1–10%, the limit of detection was 0.1 µL/L, and the limit of quantification was 0.3 µL/L. Analytical recoveries were 100.6% ± 10.1%, 102.5% ± 7.3%, and 99.0% ± 4.1% at 0.5, 10, and 75 µL/L, respectively. The method described was used to determine compound A concentrations during simulated closed-circuit conditions. Some of the resulting data are included, illustrating the practical applicability of the proposed analytical approach.
Conclusions: A simple, fully automated, and reliable quantitative analytical method for determination of compound A in air was developed. A solution was established for sampling, calibration, and chromatographic separation of volatiles in an area complicated by limited availability of sample volume and low concentrations of the analyte.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Nevertheless, sevoflurane is not the most ideal anesthetic because it is metabolized in vivo (although moderately) and it is chemically unstable in the presence of alkalis, such as soda lime and Baralyme,® which are CO2 adsorbents typically used in circle-breathing systems in low-flow and closed-circuit anesthesia. Both processes produce potentially toxic byproducts (1)(2)(3)(4).
Sevoflurane is partially metabolized (2–5%) by the liver to produce inorganic fluoride ions and hexafluoropropanol (5)(6)(7). Of more concern, however, is its interaction with CO2 adsorbents, which generates several degradation products (1)(8). Five byproducts were previously identified in vitro, and they were designated as compounds A, B, C, D, and E (9). Only compound A [and to a lesser extent compound B (10)(11)] is produced under conditions likely to be encountered clinically. The quantitative formation of compound A seems dependent on multiple factors, among which are the identity and brand of CO2 adsorbent (12)(13)(14), the number of previous uses (15), water content (16)(17)(18), temperature of the adsorbent during sevoflurane exposure (13)(18), total fresh-gas flow rate (13)(18), anesthetic concentration (13)(18), and duration of anesthesia (14)(19).
Compound A, an olefinic ether (fluoromethyl-1,1,3,3,3-pentafluoro-2-propenyl ether), has a dose-dependent nephrotoxicity in rats [median lethal concentration (LC50) after 3 h of exposure, 400 µL/L] (20), which primarily involves renal, hepatic, and cerebral damage (21)(22). Whether the compound A formed during anesthetic use produces similar toxicity in humans is still the subject of scientific debate. The body of experimental and clinical evidence to date indicates that the amount of compound A generated under usual clinical conditions is substantially lower than the concentration that elicits acute toxicity in animals (2)(3), suggesting that sevoflurane administration is safe. Nevertheless, some reservations still exist among the anesthesia community concerning sevoflurane application in low-flow and closed-circuit anesthesia. Many have cautioned against it (23)(24).
Low-flow and closed-circuit anesthesia offer potential economic and environmental benefits (25). However, in low-flow and closed-circuit anesthesia, reinhalation of expired gases occurs, including the CO2 produced by the patient. A fundamental requirement is the correct functioning of a canister filled with a CO2 adsorbent, which chemically binds the exhaled CO2 and thus prevents its reuptake by the patient. One serious drawback is the potential for accumulation of dangerous trace gases (e.g., methane and acetone) and volatile anesthetic breakdown products (e.g., compound A) (10)(25).
In our department, closed-circuit anesthesia is performed mainly using the Physioflex® apparatus, a computer-controlled, valveless, closed-circuit system with a built-in blower (70 L/min) that rotates the gases unidirectionally in the circuit. Four membrane chambers are also present for ventilation purposes and/or sensing respiratory movements of the patient. The anesthetic is injected electronically, taking into account the continually measured gas and anesthetic concentrations (by built-in infrared spectrometry) and coupled-feedback information. The device complies, to a great extent, with the requirements for safe performance in present day anesthesia (26). Because of the precise metering of anesthetic into the system and the high circular flow, we hypothesized that compound A formation, even in closed-circuit conditions, would be negligible to absent, thereby lending credence to the hypothesis of safe sevoflurane administration. Confirming this involves measuring low compound A concentrations during in vitro and in vivo sevoflurane anesthesia in closed-circuit conditions. A validated, quantitative assay for the determination of compound A in limited gas volume samples would be necessary in this regard.
A limited number of analytical procedures to measure low concentrations of compound A (a volatile liquid, as is sevoflurane) in the vapor phase have been described previously (27). Logically, all of the procedures are based on gas chromatographic separation. In most cases, sensitivity considerations have required the use of packed gas chromatographic columns, permitting the injection of large gas volumes (10)(16)(18)(27). However, we found that state-of-the-art capillary columns, which enable better separation, combined with sensitive and selective mass spectrometric detection were necessary when analyzing the low compound A concentrations that we expected in our Physioflex system. Although capillary gas chromatography (GC),1 in combination with flame ionization detection (11)(17) or mass spectrometry (MS) (12) has been mentioned in the literature, an analytical approach that fully takes into consideration both analytical requirements (sample volumes and sensitivity) and chromatographic or technical limitations has not been fully elaborated and described.
This report describes a novel, sensitive, and automated chromatographic assay for the quantitative determination of trace amounts of compound A in limited-volume, vapor-phase samples using existing headspace injection technology and MS detection. In addition, quantitative data from an experimental setup designed to study compound A formation under simulated clinical conditions are given.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
reagents
Sevoflurane and compound A were both supplied by Abbott
Laboratories. 1,1,1-Trifluoro-2-iodoethane (99% pure) and ethyl
acetate (HPLC grade) were both obtained from Aldrich. All other
reagents and products used were of analytical grade and purchased from
Aldrich.
calibrators
Calibrators were produced from gravimetrically prepared liquid
stock solutions of sevoflurane and compound A in ethyl acetate. The
stock concentration of sevoflurane was 1352 g/L, and that of
compound A was 
7.6 g/L. Liquid dilutions of the latter were prepared
volumetrically by means of a Hamilton Digital Diluter (Bonaduz),
producing working dilutions of 5700, 3800, 1900, 760, 380, 76, and 38
mg/L in ethyl acetate. Each working calibration solution was
prepared by mixing volumetrically equal parts of the stock sevoflurane
solution and the appropriate compound A working dilution. This
procedure produced liquid calibration solutions containing 676 g/L
sevoflurane and 2850–19 mg/L compound A. The liquid dilutions and
solutions were stored in well-filled, airtight receptacles (2-mL glass
vials) at -18 °C and were used no longer than 1 month.
Before every procedure, the dilutions and solutions were mixed
thoroughly and allowed to reach room temperature. In addition, a
calibrator containing pure ethyl acetate in place of compound A
solution was prepared in the same manner.
The transition to the vapor phase was then achieved. For calculation of the corresponding vapor-phase concentrations, we made use of the ideal gas law (PV = nRT, where T = 293.15 K) and an average gas bulb volume of 135 mL (see comment on gas-bulb volume in Results and Discussion).
A precise amount of each calibration solution (26.56 µL) was injected into separate, calibrated, and evacuated 125-mL gas-sampling bulbs (Alltech) using a calibrated, gastight syringe (1725RN Hamilton; Bonaduz), and the solution was allowed to evaporate completely. After adjustment to atmospheric pressure and equilibration for 10 min, which promoted adequate diffusion of the gases, 2 mL of gas was sampled using a 2.5-mL gastight, sample-lock syringe (1002SL Hamilton; Bonaduz) and carefully transferred into a sealed 5-mL glass headspace vial. An overdraw-and-compress technique was applied, whereby an excess of gas was drawn into the syringe (2.5 mL), the syringe was locked, and the plunger was pushed down to precisely 2 mL. The syringe was then briefly unlocked, releasing the excess gas. This technique facilitated the adjustment of calibrators (and samples) to atmospheric pressure, irrespective of whether they had previously been at slightly negative or positive pressures. For practical reasons, all handling was performed at room temperature and manipulation by hand of the glass recipients containing the solutions was avoided as much as possible.
As is customary in anesthesiology, all concentrations used throughout were in the vapor phase and were expressed in volume/volume ratios, corresponding with the final component concentration in the vapor phase as present in the particular gas bulb.
Before injection, 0.5 µL of a liquid solution containing
1,1,1-trifluoro-2-iodoethane (500 mg/L) in ethyl acetate was added
as an internal standard (IS) to each vial using a very precise
Hamilton CR-700 constant-rate syringe (Bonaduz).
experimental setup
Our experimental setup consisted of a closed-circle breathing
system, including a pair of artificial rubber lungs, connected to a
computer-controlled Physioflex anesthesia apparatus, in which a
canister filled with fresh CO2 adsorbent [in
this case soda lime (Sodasorb; Grace)], and a charcoal canister were
incorporated (system A). This Physioflex system was compared with a
conventional closed-circuit setup (system B). To construct system B,
two unidirectional valves were added to system A. During the system B
experiments, neither the built-in blower nor the charcoal canister was
used. The ventilation rate was 10 breaths/min. The circuit was
operated with a constant sevoflurane concentration of 2.1% (21
mL/L) in 100% oxygen for 120 min, after which anesthetic
administration was halted and, only in system A, circulating gases and
vapors were immediately guided through the canister containing
charcoal. As a result, the anesthetic concentration dropped
drastically. Our goal was the evaluation of compound A formation under
various clinical conditions. Therefore, several experiments, each
consisting of 10 experimental runs, were conducted using different
operational conditions. ANOVA and t-tests were performed on
the compound A concentrations measured and canister temperature
readings.
samples
Gas samples (2 mL) for compound A analysis were withdrawn, in
duplicate, from the inspiratory and the expiratory limbs of the circuit
after 5 and 15 min and every 15 min during the remainder of the
experimental run. At the end of each experimental run, two additional
gas samples (at 5 and 10 min after suspension of sevoflurane
administration) were taken. Gas sampling was performed at
circuit-sampling ports fitted with three-way valves by means of 2.5-mL
gastight sample-lock syringes. An overdraw-and-compress technique was
again applied before the samples were transferred into headspace vials.
As with the calibrators, 0.5 µL of liquid solution of
1,1,1-trifluoro-2-iodoethane (IS) was injected into each vial before
analysis.
headspace sampling and injection
The critical step of injecting, in a reproducible way, a large
volume of gas onto a capillary column was addressed using the
technological features of a headspace sampler, which allowed full
automation of the method.
Before injection, every vial was preheated for 10 min at 38 °C to ensure minimal influence of environmental temperature fluctuations. After pressurization with carrier gas and equilibration (1 min), 1 mL of the vapor phase was introduced into the injector liner by means of a 1-mL gastight syringe (Gerstel) at a computer-controlled rate of 5 mL/min.
To enhance sensitivity, cryofocusing in the injector liner was applied. To that end, the deactivated liner was packed with Tenax® TA (20/35 mesh) obtained from Alltech. The programmed temperature vaporizer injector was operated under the following conditions in all runs: during the precooling step, the cold trap was cooled down to -80 °C by liquid N2 departing from a Dewar vessel. The liner was maintained at that temperature until 0.30 min after the injection was completed, thus ensuring adequate analyte adsorption and concentration. Subsequently, the cold trap was flash heated up to 250 °C and held at that temperature for 10 min, providing flash desorption and transfer of the analyte onto the capillary column. During this desorption phase, the injector split ratio was 0.7:1, avoiding pressure effects in the injector liner and, thus, minimal sample mass loss. At 0.75 min after the injection was completed, the split ratio was increased to 40:1 to flush the liner and avoid peak tailing and carryover.
chromatographic conditions
Isothermal chromatographic separation was achieved on a
wall-coated, open-tubule, fused-silica CP-Select 624 custom-made
capillary column [41 m x 0.25 mm (i.d.); 2.1 µm film
thickness; Chrompack]. Helium was used as the carrier gas at a flow
rate of 1.0 mL/min. Oven temperature was 38 °C for a runtime of
17.50 min, and the MS transfer line temperature was maintained at
100 °C. The mass detector was operated in the scan mode (mass range,
m/z 40–300) with the m/z scan rate set at 1.46
scans/s. The compounds were identified by comparing retention times
(tRs) and mass spectra of the samples
with those obtained for calibrators.
quantification and method evaluation
For quantification, mass fragmentograms of m/z 128
(compound A) and m/z 210 (IS) were constructed and
integrated. As can be seen in Fig. 2
, both ions were characterized by a
high abundance as well as a high m/z value, promoting assay
sensitivity and selectivity.
|
Linearity.
Calibration curves were prepared as described
previously for 0.5–75 µL/L compound A in air. For each curve, seven
different concentrations, mostly in the lower part of the curve, were
analyzed. Each calibration was completed by a 0 µL/L added compound A
calibrator. For calibration purposes, the peak-area ratio (compound A
vs IS) was regressed against the accurate, recalculated vapor-phase
concentrations of the calibrators (e.g., depending on calibrated gas
bulb volume). In an effort to account for data heteroscedasticity,
weighed linear regression (weighing factor, 1/y) was applied
to calculate the calibration curve.
Precision.
The precision of the method was evaluated by
analyzing calibrators at three different concentrations (0.5, 10, and
75 µL/L), on the same day (6 replicates; within-day reproducibility),
and on separate days (10 replicates; total reproducibility). Peak-area
ratios were transformed to concentration using calibration curves
prepared and analyzed simultaneously.
Analytical recovery.
No certified quality-control samples
exist. To monitor the overall analytical recovery, three positive
control samples were prepared. The quantitative results at those
concentrations (0.5, 10, and 75 µL/L) were then related to the
supplemented concentrations.
Limits of detection and quantification.
The limit of detection
(LOD) was determined by injecting decreasing concentrations of compound
A in air. The LOD was defined as the lowest concentration that produced
a signal-to-noise ratio 
3. The limit of quantification (LOQ) was
established at a signal-to-noise ratio of 10 (28),
conditional to quantification at that level with an error of <15%.
Specificity.
To monitor the selectivity of the method, several
inhalation anesthetics that have been applied in clinical practice, as
well as many other volatile organic compounds, were chromatographed at
concentrations in large excess of the highest calibration point.
Solutions of these compounds were prepared in ethyl acetate, and 0.5
µL was added into a glass headspace vial. Retention and mass-spectral
characteristics were examined and compared with those of compound A and
the IS.
safety considerations
To minimize laboratory environmental contamination and to protect
the operator, a charcoal trap was attached to the split vent of the GC.
All experiments were carried out in a well-ventilated room and,
whenever possible or needed, under a fume hood. Halogenated volatiles
were meticulously collected in separate disposal canisters. Beyond
this, the method demands no additional safety precautions other than
those used when working with volatile organic solvents.
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
depicts the chromatograms of the GC-MS analysis of a calibrator
containing 1 µL/L compound A in air and a sample obtained during an
in vitro experiment at 90 min after the start of the sevoflurane
administration. As can be seen, the compound A peak is well separated
from the sevoflurane peak, is characterized by an excellent peak shape,
and is devoid of any interference. The
tRs of compound A, sevoflurane, and
1,1,1-trifluoro-2-iodoethane are 5.8, 7.9, and 14.0 min, respectively.
|
The need for complete chromatographic resolution of a mixture of sevoflurane and compound A required the use of a thick-film capillary column. In view of the highly volatile and semipolar nature of both components, a liquid 6% (60 g/L) cyanopropylphenylmethylsilicone stationary phase was preferred, as used by Ruzicka et al. (11) and recommended in EPA method 524.2 (29). With the commercially available columns of this type and range of film thickness, resolution and retention were still unsatisfactory. As a result, a custom-made capillary column (CP-select 624; see Materials and Methods) was purchased.
Generally in chromatography, demands concerning the speed of analysis
are increasing. In our assay, the oven temperature was isothermal at
38 °C, promoting fast analytical runs and thus a shorter overall
batch analysis time. Stability data for compound A in the vapor phase
are shown in Fig. 3
. Furthermore, an oven temperature of 38 °C
ensured adequate retention and could be kept stable, regardless of
environmental conditions. The run time was 17.50 min, allowing elution
of a large peak of ethyl acetate, the solvent used in the liquid
dilutions for preparation of the calibrators, after the GC run but
during the intermediate preheating time for the next vial.
|
Injection technique.
There are two major factors to be
considered when conducting capillary GC analysis of gases:
(a) samples usually contain the key analyte at a very low
concentration, requiring the injection of a large volume of gas in a
reproducible way to place enough analyte mass on the capillary column
and thus reach the intended sensitivity; and (b)
chromatographic or instrumental restrictions (e.g., very small column
internal diameter) prevent direct loading of such gas volumes. Our
particular application also restricts the sample volume to be withdrawn
from the anesthesia circuit. Withdrawing large sample volumes does not
substantially affect total circuit volume (4 L), but it does distort
the similarity between experiments and routine clinical practice. These
small sample volumes preclude the use of standard loop-type gas
samplers needing large samples to flush the loop homogeneously. To
circumvent these limitations, the technological features of a headspace
sampler were applied in combination with cryofocusing.
A cryotrap containing an adsorbent, located in the injector liner, was used for enrichment and preconcentration before separation on the analytical column. Additionally, a minimization of injection bandwidth was achieved. During the injection, the analyte was introduced into the precooled CIS (-80 °C) and adsorbed onto the Tenax, and most of the air was purged out of the system (at a split ratio of 7:1). During the flash-heating phase, which transfers the volatiles onto the column, the split ratio was kept minimal to increase sensitivity as much as possible.
Detection.
Data generated during in vitro experiments studying
compound A formation under true closed-circuit conditions revealed
compound A concentrations in the low µL/L range or lower. Previously
cited methods, using packed columns
(10)(16)(18)(27) and/or
flame ionization (11)(17) detection systems,
report detection limits approaching 1 µL/L. The greater aim of our
work was to establish operational conditions to eliminate compound A
formation as much as possible. Because low to nonexistent compound A
concentrations were expected, it was deemed necessary to use MS
detection, which provides high selectivity as well as the vital
sensitivity. Because data are collected in full scan mode, sensitivity
can be enhanced by switching to selected-ion mode.
In addition to the chromatographic retention data, the full-scan
MS detection offered us unambiguous confirmation of peak
identities. The electron-impact mass spectra of
compound A and sevoflurane, represented in Fig. 2
, were characterized by prominent peaks at m/z values
of 69, 128, 161, and 180 (M+) and m/z
values of 69, 131, 181, and 199 (M+),
respectively. The high-abundance ion at m/z 128 was selected
as the target ion for quantitative purposes, using reconstructed mass
chromatograms.
Calibration.
To prepare the calibrators for the calibration
curves, we used liquid dilutions with ethyl acetate as a solvent.
Compared with compound A and sevoflurane, ethyl acetate is
characterized by a higher boiling point; therefore, it elutes later
(tR = 19.51 min), causing no
interference whatsoever in regard to the compounds of interest.
Furthermore, ethyl acetate is minimally toxic and readily available in
a very high degree of purity and at an acceptable price. Other solvents
(e.g., methanol) were screened, but none was as satisfactory.
We preferred to depart from gravimetrically prepared liquid dilutions of the analyte. In contrast to gases (27), liquids can be more precisely prepared, are easily miscible, and are more practical to work with. Only during the last stage in the preparation of calibrators was the transition to the gas phase achieved.
The gas-sampling bulbs we used were labeled as 125 mL, but their true internal volume varied by up to 15% (the smallest volume was 135 mL). To construct a correct calibration curve, the exact concentration of each calibrator in µL/L (volume/volume ratios) was recalculated accordingly, departing from the calibrated volume of the gas bulb used and the exact amount of compound A present in the particular calibration solution. To make the experimental conditions resemble the clinical conditions as closely as possible and to avoid major discrepancies between calibrators and samples (which contained a large concentration of the actual sevoflurane anesthetic), a fixed amount of sevoflurane was added to each calibrator. Trapping (on Tenax) and chromatographic distortions because of the large amount of sevoflurane present were considered critical in this respect. The amount of sevoflurane added to the calibrators was set at 1.6% (16 mL/L), slightly lower than the minimum concentration in the circuit. Several considerations guided this choice. Mainly because of the technological features of the equipment, the sevoflurane concentration during anesthesia fluctuates constantly, according to the needs of the (artificial) patient. For example, if a setpoint of 2% (20 mL/L) sevoflurane is dialed on the anesthetic machine, the actual concentration, after an initializing period, usually is 1.8–2.4% (18–24 mL/L). Furthermore, the lower end of the concentration range was chosen to extend the linear dynamic range as far as possible, thus avoiding detector saturation, and also because of the already inevitable presence of traces of compound A in the sevoflurane calibrator. It is a known fact that during sevoflurane synthesis, compound A is formed as a byproduct, albeit in very small amounts (30). Consequently, a "base" amount of compound A was present in every calibrator, independent of the amount added to reach the intended concentration of the calibrator.
IS.
To enhance the precision of the analysis, the use of an IS
was important. Several compounds were screened based on their chemical
structure, volatility, and retention profile.
1,1,1-Trifluoro-2-iodoethane was chosen because of its chemical
structure (halogenated and thus xenobiotic),
tR, and typical mass spectrum. Indeed,
Fig. 2
indicates that the molecular ion (m/z 210) is the
dominant ion in the spectrum, offering a higher selectivity compared
with 1-chlorobutane (27) and ethanol (17), ISs
used in the literature. In addition, unlike 1-chlorobutane, the
stability of 1,1,1-trifluoro-2-iodoethane in the vapor phase at room
temperature was satisfactorily assessed over a period of 48 h.
Stability data for the IS are shown in Fig. 3
. On the basis of our investigations, adding a small amount of
liquid IS (0.5 µL), which partitions itself between the liquid and
the vapor phases within the headspace vial, dramatically improves the
reproducibility and accuracy of the method, while not affecting peak
area or sensitivity to any extent.
method validation
Linearity.
The calibration curves were linear over a specified
range [0.3–75 µL/L (volume/volume ratios)]. An average correlation
coefficient (r2) of 0.996 was obtained
for the relationship between the peak-area ratio (compound
A/1,1,1-trifluoro-2-iodoethane) and the corresponding calibration
concentrations ( Sy|x = 0.067 ± 0.25,
mean ± SD; n = 10). Because a small amount of compound A
was always present (even in the zero calibrator), the
y-intercept for each calibration curve was always different
from zero: y-intercept, 0.018 ± 0.003; SE of the
intercept, 0.007 ± 0.0027 (both mean ± SD; n = 10).
The regression slope characteristics were as follows: slope, 0.048
± 0.003; SE of the slope, 0.001 ± 0.00047 (both mean ± SD;
n = 10), providing a sufficiently low CV (6.9%). The method
showed good linearity over the specified concentration range, which
largely covered the in vitro and expected in vivo concentrations of
compound A.
Precision.
Table 1
presents the within-day (n = 6) and total (n = 10)
reproducibility data obtained for the different concentrations tested.
The CVs were 4.1–10%. This indicates good to excellent
reproducibility over the studied concentration interval.
|
Analytical recovery.
For validation purposes, the analytical
recovery for compound A was also determined at three separate
concentrations. At 0.5 and 10 µL/L, values of 100.6% ± 10.1% and
102.5% ± 7.3%, respectively, were found, whereas at 75 µL/L, the
analytical recovery was 99.0% ± 4.1% (mean ± SD; n = 10).
LOD and LOQ.
The LOD, as defined in Methods and
Materials, was 0.1 µL/L compound A in air. This validation
parameter was not discussed (10)(14), nor was
the criterion met
(11)(12)(17)(27) in
previously published studies. The LOQ was 0.3 µL/L compound A, which
was quantitatively measurable with acceptable reproducibility (total CV
<15%). In view of the importance of these assay performance
characteristics, it must be stressed that, in assessing both LOD and
LOQ, not only was the signal-to-noise ratio [3 for LOD, 10 for LOQ
(28)] applied but also more stringent spectral criteria. In
doing so, the double identification potential
(tR and mass spectrum) was completely
preserved even in the lowest range.
Specificity.
Because volatile organic compounds are
present not only in the global atmosphere, but also, in higher
concentrations, in the typical laboratory environment, assay
selectivity is paramount. Through the use of MS, selectivity is
guaranteed not only from a chromatographic standpoint, but also through
the selectivity of MS ion extraction, the mass fragmentograms allowing
quantification of the compound of interest.
Table 2
shows the tRs of a list of
volatile organic compounds and other compounds that were considered as
potential interferences. This is a volatility-limited list, which
includes other inhalational anesthetics, frequently used solvents, and
other products that were screened as possible ISs.
|
analysis of in vitro samples
Substantiation of the usefulness of our method was achieved by
analyzing samples in 20 in vitro runs during 120 min, studying compound
A formation in two types of closed-circuit anesthesia systems. In
system A, the Physioflex was used, and in system B, our setup was a
common closed circuit, in which liquid sevoflurane was injected using
valves and using neither the charcoal canister nor the incorporated fan
driving the breathing gases at a flow rate of 70 L/min. As mentioned,
samples were taken from both the inspiratory and expiratory
limbs, in duplicate, before anesthesia was administered, at 5
and 15 min after anesthesia was administered, and every 15 min
thereafter until 120 min. Sevoflurane administration was then halted.
Additional samples were taken at 5 and 10 min after sevoflurane was
stopped. The temperature of the CO2 adsorbent was
measured at the same time points. Fig. 4
shows inspiratory compound A concentrations and canister
temperatures in the two systems. Data from those samples are
represented in Table 3
. We found compound A concentrations of 4.1–14.3 µL/L.
Compound A concentrations were significantly lower with the Physioflex
(system A) compared with the setup with two unidirectional valves
(system B). These results have been reported previously
(31). Furthermore, system B was characterized by higher soda
lime temperatures, suggesting that the (high) 70 L/min flow in the
Physioflex induces heat dissipation throughout the anesthesia system
and hence lower heat- and alkali-mediated formation of compound A,
explaining the lower compound A concentrations, as hypothesized.
Further in vitro and in vivo experiments are now being conducted using
the presented method in a clinical setting.
|
|
In conclusion, a simple, automated, and reliable quantitative analytical method for determination of compound A in air was established, using capillary GC separation followed by MS detection. Cryofocusing was successfully applied to reconcile the large injected-gas volumes with the small column internal diameter. We overcame the intricacies of gas analysis with respect to sampling, calibration, and chromatographic separation in a difficult application field, complicated by the limited sample volumes available, the low concentrations, the unfavorable ratio of sample number to available analysis time, and the high quantitative standards imposed by the clinical application field. Moreover, this was achieved without having to use dedicated gas analysis instrumentation. In the method presented, data were collected in full scan mode. This allows further enhancement of sensitivity by switching to selected-ion monitoring mode. This might be very useful in view of our intentions to conduct further in vivo investigations that focus on determining compound A uptake by the patient through measurement of blood concentrations. Extremely low concentrations are expected in this respect. A robust, routinely applicable method was achieved as evidenced by its daily application in clinical research (31)(32)(33), the results of which substantiated our hypothesis of reduced compound A formation in Physioflex-based, closed-circuit anesthesia systems using sevoflurane.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  M.-P. L. A. Bouche, L. F. M. Versichelen, M. M. R. F. Struys, J. F. P. Van Bocxlaer, A. P. De Leenheer, E. P. Mortier, and G. Rolly No Compound A Formation with Superia(R) During Minimal-Flow Sevoflurane Anesthesia: A Comparison with Sofnolime(R) Anesth. Analg., December 1, 2002; 95(6): 1680 - 1685. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and the IS,
1,1,1-trifluoro-2-iodoethane (
), in vapor phase over a period of
48 h.