| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
A/
distributions in the
normal rabbit by micropore membrane inlet mass spectrometry
Department of Anesthesia, University of Pennsylvania, Philadelphia, Pennsylvania 19104-4283
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
We
developed micropore membrane inlet mass spectrometer (MMIMS) probes to
rapidly measure inert-gas partial pressures in small blood
samples. The mass spectrometer output was linearly related to
inert-gas partial pressure (r2 of
0.996-1.000) and was nearly independent of large variations in
inert-gas solubility in liquid samples. We infused six inert gases into
five pentobarbital-anesthetized New Zealand rabbits and used the MMIMS
system to measure inert-gas partial pressures in systemic and pulmonary
arterial blood and in mixed expired gas samples. The retention and
excretion data were transformed into distributions of
ventilation-to-perfusion ratios (
A/
) with the
use of linear regression techniques. Distributions of A/ were unimodal and broad, consistent with
prior reports in the normal rabbit. Total blood sample volume for each
A/ distribution was 4 ml, and analysis time
was 8 min. MMIMS provides a convenient method to perform the multiple
inert-gas elimination technique rapidly and with small blood sample volumes.
inert gases; multiple inert-gas elimination technique; solubility; stirring effect
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
THE DEVELOPMENT OF THE
MULTIPLE inert-gas elimination technique (MIGET) by Wagner and
coworkers in 1974 was a major advance in the study of pulmonary gas
exchange (15, 36, 39). In MIGET, data on the retention
[systemic arterial inert-gas partial pressure divided by mixed venous
inert-gas partial pressure (P
)] and excretion (mixed expired
inert-gas partial pressure divided by P) of six infused inert
gases are transformed into a distribution of ventilation-to-perfusion
ratios (A/) in the lung (36, 38,
40). MIGET has been used to study the mechanisms of impaired gas
exchange in several diseases and animal models of disease, including
pneumonia (11, 35), asthma (19, 29), chronic obstructive pulmonary disease (2, 7, 26), atelectasis (33, 34), and acute respiratory distress syndrome
(6, 44).
Despite its utility in providing unique information about the
mechanisms of pulmonary gas exchange, MIGET has been applied by only a
few research laboratories around the world (2, 7, 10, 11, 17,
26-28, 30, 33, 36, 43). In traditional MIGET, the inert-gas
partial pressures in the blood samples are measured by gas
chromatography (GC), which requires a time-consuming extraction of the
inert gas into a gas phase before the GC analysis. The substantial
analysis time (~3 h for each measurement of the 6 retention and
excretion ratios) and the technical complexity of measuring inert-gas
partial pressures in blood samples by GC have limited the use of MIGET
in research applications and in the real-time clinical care of
critically ill patients. For highest accuracy and reproducibility of
the inert-gas extraction, traditional MIGET by GC also requires ~20
ml of blood sample for a single A/
determination, which has limited the applications of MIGET in pediatric
medicine and in research studies in small animal models.
Membrane inlet mass spectrometry (MIMS) offers an alternative method for the analysis of inert-gas partial pressures in blood samples. In MIMS, a polymer membrane separates an aqueous sample from the vacuum chamber of the mass spectrometer, and volatile substances dissolved in the aqueous sample diffuse across the membrane into the vacuum system for mass spectrometer analysis (3, 18, 41). Prior investigators have applied MIMS to the measurement of inert-gas partial pressures in blood samples for six gases suitable for MIGET, with promising results (23, 24). A major limitation of MIMS, however, has been a dependence of the mass spectrometer signal on the inert-gas solubility in blood samples (23). Dependence of the signal on solubility can be reduced by the use of very thick membranes for MIMS, but thick membranes result in prohibitively slow response times (42). As a result, no practical ways of utilizing MIMS for MIGET have been reported.
Micropore membrane inlet mass spectrometry (MMIMS) provides a unique combination of rapid response speed and minimal dependence of the signal on inert-gas solubility in the aqueous sample (1). In MMIMS, the membrane is confined to a small pore. The membrane is thin, resulting in a fast response time, but the dependence of the signal on solubility is minimized by the three-dimensional diffusion profiles around the pore (1). MMIMS can, therefore, rapidly and directly measure the partial pressures of inert gases dissolved in aqueous samples, eliminating the extraction of inert gases into a gas phase that is required for GC analysis.
We developed MMIMS probes specifically designed for the measurement of
trace concentrations of inert gases in small blood samples, and we
applied this MMIMS technique to determine A/ distributions in normal rabbits by MIGET.
| |
METHODS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
MMIMS.
Pores sealed with a polymer membrane were created in the ends of
stainless steel tubing, as previously described (1).
Briefly, 304 stainless steel tubing (1/16 in. OD, 0.040 in. ID), which was sealed on one end with a hemispherical cap weld of constant wall
thickness throughout the weld (MicroGroup, Medway, MA), was mounted on
a vacuum system. The sealed end of the tubing was filed at an angle of
~30° to the axis of the tubing while the helium leak rate into the
vacuum system was monitored, until a leak was created large enough to
register a helium current in the range of 1.5-14 × 10
10 A on a quadrupole mass spectrometer (UTI 100C, see
below for details, in Faraday cup mode). Although the size of the pores for this MMIMS probe was not directly assessed, the pore diameters for
similar probes with larger helium leak rates have been estimated at 28 µm (1). After an appropriate helium leak rate was
achieved, the pore (or pores, see Ref. 1) in the flat, filed surface at
the tip of the probe was sealed with silicone adhesive sealant (Permatex 66B, Loctite, Cleveland, OH). Silicone was used in this probe, instead of the polytetrafluoroethylene vacuum grease used in
prior MMIMS probes (1), to enhance sensitivity, because silicone polymers can be several hundred times more permeable to the
inert gases of interest than polytetrafluoroethylene-based polymers
(21). After the silicone had cured for 24 h, the
probe was rotated 90°, and the shaving process was repeated. The
probe was then rotated another 90°, and the process was repeated once more. The resulting probe had three flat, filed, pore-containing areas
at its tip and, therefore, approximately three times as much
sensitivity as a probe with a single filed area.
10 Torr. The vacuum system was evacuated by
a turbomolecular pump (Turbovac 150, Leybold-Heraus, Export, PA) backed
by a rotary vane pump (D8A, Leybold-Heraus), and the turbomolecular
pump was connected to the vacuum chamber by a right-angle, high-vacuum isolation valve (model 951-5027, Varian, Lexington, MA). The valve was completely opened during bakeout and between experiments but was
nearly closed during experiments to decrease the pumping speed and
increase sensitivity. Vacuum system pressure during experiments was
7 × 108 Torr.
Inert-gas partial pressures in gas and liquid samples. The six inert gases were chosen to span evenly a large range of solubility in blood and to minimize spectral overlap during the mass spectrometer analysis. The six inert gases were sulfur hexafluoride (SF6; Air Products and Chemicals, Allentown, PA), krypton (Kr; BOC Group, Murray Hill, NJ), desflurane (DES; Ohmeda, Liberty Corner, NJ), enflurane (ENF; Ohmeda), diethyl ether (DEE; Fisher Scientific, Fair Lawn, NJ), and acetone (ACT; EM Science, Cherry Hill, NJ). These inert gases are nearly identical to the series used by Mastenbrook et al. (24), but with DES substituted for freon-12. All inert gases were measured by using single-ion monitoring, with the peak location and switching controlled by a computer and interface (Spectralink, UTI, Sunnyvale, CA) and custom software (QDot, Kirtland, NM). The amu peaks selected for monitoring each gas were SF6 at 127, Kr at 84, DES at 101, ENF at 117, DEE at 59, and ACT at 58 amu.
The time response of the system for each gas was assessed with step changes in inert-gas partial pressures for each gas individually, in both the gas phase and liquid phase. Absence of vacuum system memory effects was confirmed by following the signal after a step decrease in inert-gas partial pressure. Sensitivity of the mass spectrometer signals to changes in flow rate over the probe for liquid samples was assessed by preparing large samples of single inert gases dissolved in water and manually injecting these samples over the probe tip, while varying the flow rates from 0.4 to 10 ml/min. The effects of inert-gas solubility on the mass spectrometer signal for each inert gas were assessed by comparisons of the signal for known inert-gas partial pressures in water, blood, and 20% Intralipid solution (Baxter Healthcare, Deerfield, IL). For SF6, Kr, and DES, gas from a premixed tank with a known partial pressure of the inert gas in nitrogen was equilibrated with 30 ml of the liquid (water, rabbit blood from a pooled sample, or Intralipid) by adding 20 ml of gas in a 50-ml glass syringe, mixing, expelling the gas, and repeating the gas exchanges until the mass spectrometer signal for the liquid did not change with further exchanges. The 30 ml of liquid were then separated into 2- to 3-ml samples in 5-ml glass syringes, and the liquid samples were injected over the MMIMS probe at room temperature and in varied order. For ENF, DEE, and ACT, the inert "gas" was first diluted as a liquid into water, then 0.5 ml of the water and inert gas was added to 30 ml of the test liquid (water, rabbit blood from a pooled sample, or Intralipid), and 20 ml of air were added to the test liquid and equilibrated. The resulting inert-gas partial pressure in the test liquid was measured by head-space gas analysis, with comparison of the head-space gas mass spectrometer signal to the signal for a known inert-gas partial pressure from a premixed tank. The 30 ml of liquid were then separated into 2- to 3-ml samples in 5-ml glass syringes, and the liquid samples were injected over the MMIMS probe at room temperature and in varied order. The ratio of mass spectrometer sensitivity for liquid-phase measurements to the sensitivity for gas-phase measurements was determined for each inert gas by preparing mixtures of ~35 ml of air, 15 ml of water, and a single inert gas in 50-ml glass syringes. The glass syringes were equilibrated by mechanical rotation in a temperature-controlled oven at 38°C for 1 h, and the gas and liquid phases were rapidly separated into 50-ml gastight syringes (Hamilton, Reno, NV) and 5-ml glass syringes, respectively. The gas samples were injected over the MMIMS probe at room temperature, and the water samples were injected over the probe with the sample temperature maintained at 38°C with water-jacketed tubing. Linearity of the mass spectrometer response for the gas phase was tested in 50-ml gastight syringes by serial dilutions of each inert gas in air. Linearity of the mass spectrometer response for the liquid phase was tested by serial dilutions of each inert gas in degassed water (distilled water that had been heated to boiling and then sealed in a syringe and cooled with no exposure to gas). Water samples for the liquid-phase serial dilutions were analyzed at room temperature. Spectral overlap was assessed by flowing each single inert gas diluted in nitrogen over the probe, recording the current at the monitored peak for that gas, and recording the current at each of the other five peaks. The ratios of the current at each of the five peaks to the current at the monitored peak were used to construct the spectral overlap matrix K, which was subsequently used in all measurements for mixtures of the inert gases to convert the six measured currents to the six inert-gas partial pressures. The spectral overlap correction equation was
|
(1) |
|
|
MIGET in normal rabbits.
After Institutional Animal Care and Use Committee approval of the
animal protocol, five female New Zealand White rabbits (3.5-4.3 kg) were anesthetized with 25 mg/kg of ketamine im and 5 mg/kg xylazine
im. A 22-gauge catheter was placed in an ear vein, and normal saline
was infused at 80 ml/h. Anesthesia was maintained with a continuous
infusion of pentobarbital at 12 mg · kg1 · h1 throughout
the experiment. Rectal temperature was measured with a thermocouple
probe and regulated to 38°C with a heating pad and lamp. A 3.5-mm
endotracheal tube was placed via tracheostomy, and ventilation was
controlled at a tidal volume of 25-30 ml and rate of 25 breaths/min (model 667, Harvard Apparatus, South Natick, MA). Grade 5 purity oxygen and grade 5 purity nitrogen (Air Products and Chemicals)
were mixed to provide an inspired O2 fraction of 0.75. Pancuronium (0.3 mg) was administered to facilitate controlled ventilation, and 0.1 mg · kg1 · h1 was infused
continuously throughout the experiment. A 20-gauge catheter was placed
by cutdown in a femoral artery for continuous blood pressure
measurement and arterial blood sampling. A 4-Fr catheter (angiographic
balloon catheter, Arrow International, Reading, PA) was placed in the
right external jugular vein by cutdown and positioned in the pulmonary
artery with pressure waveform guidance. Airway pressures and systemic
and PA blood pressures were transduced and recorded continuously. In
preliminary experiments, analysis of arterial and mixed venous blood
and mixed expired gas with the MMIMS system confirmed that there were
no endogenous or exogenous sources of any inert gases in the absence of
the inert-gas infusate.
) was
then calculated from the ratio of arterial and mixed venous SP
products, with the S factoring out. Next, the data for the ratio of
gas-phase sensitivity to liquid-phase sensitivity were used to convert
the SP product for the mixed expired gas measurements to an equivalent
SP product for the liquid phase. Finally, excretion (mixed expired
inert-gas partial pressure/P) was calculated from the ratio of
the mixed expired equivalent SP to the mixed venous SP, again with the
S factoring out.
In a separate set of experiments, partition coefficients for each of
the inert gases in rabbit blood were measured by using the
double-dilution method (38). First, the validity of the gas extraction/double dilution method by using the MMIMS system for
analysis of gas partial pressures in the equilibrated gas phase was
established by repeated measurements of the inert-gas solubilities in
samples of distilled water. Each of the six inert gases was
individually added to measured volumes of water and equilibrated with a
measured volume of air at 38°C. The gas phase was separated from the
liquid phase, and a second measured volume of air was equilibrated with
the water sample. Initial and final inert-gas partial pressures in the
gas phase were measured with the MMIMS probe. Forty-milliliter blood
samples were then collected from six female New Zealand White rabbits
(3.7-4.9 kg). The inert-gas solubilities were measured for each
rabbit by the double-dilution method, matching the technique
established for the distilled water experiments, with the modification
that a 5% CO2-air mixture was used for the equilibration gas.
Retention and excretion data and measured inert-gas solubilities were
transformed into A/ distributions by using
algorithms derived by Evans and Wagner (9). Input data for
the A/ transformations included the measured
retention data for each set of samples, the measured excretion data for
each set of samples, measured partition coefficients for the inert
gases in rabbit blood, minute ventilation measured for each rabbit,
cardiac output for each set of samples derived from the Fick principle
for each inert gas, with a weighted average as described by Wagner and Lopez (37), and weighting factors based on variance
estimates for each retention measurement, as derived from the data on
repeated measurements from the pooled PA samples.
| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Inert-gas partial pressures in gas and liquid samples. The mass spectrometer signals for SF6, Kr, DES, and ENF all reached at least 98% of their steady-state levels within 30 s after a step change for the liquid and gas samples. The mass spectrometer signals for DEE and ACT reached 90-95% of their steady-state signal within 30 s and then, more slowly, reached at least 98% of their steady-state values within 120 s for the liquid and gas samples. The slow component of the time response for DEE and ACT was also present after a step decrease in gas partial pressure, with complete return to baseline for all gases within 4 min after exposure of the probe tip to zero partial pressures of the inert gases.
Increases in the sample flow above the normal rate of 1.0 ml/min resulted in no detectable change in the mass spectrometer signal for any gas, up to a sample flow of 10 ml/min. Decreases in sample flow to 0.4 ml/min resulted in <9% decrease in signal for all gases. Figure 2 shows the mass spectrometer signal for each inert gas dissolved in water, a pooled blood sample from normal rabbits, and 20% Intralipid solution, at the same inert-gas partial pressure. The signal for SF6 in lipid was significantly different from the signals for water and blood (P = 0.006 by one-way ANOVA with Student-Newman-Keuls post hoc testing), with a ratio for the lipid/water signals of 1.12. The signal for ENF in lipid was significantly different from the signal for blood (P = 0.037 by one-way ANOVA with Student-Newman-Keuls post hoc testing), with a ratio for lipid/blood signals of 1.06. No other differences in signals for water, blood, and lipid were statistically significant.
|
|
|
|
|
MIGET in normal rabbits.
Table 3 presents hemodynamic and arterial
blood-gas data for the rabbit experiments.
|
|
|
1 · atm1) were
SF6, 0.00378 ± 0.00062 (SD); Kr, 0.0509 ± 0.0036; DES, 0.287 ± 0.027; ENF, 0.758 ± 0.071; DEE,
14.6 ± 1.0; and ACT 318 ± 35 (n = 8 measurements). Solubilities of the inert gases in rabbit blood
at 38°C (ml gas at 38°C, in standard atm · ml
blood1 · atm1) were
SF6, 0.00812 ± 0.0010 (SD); Kr, 0.0522 ± 0.0062; DES, 0.623 ± 0.051; ENF, 2.34 ± 0.17; DEE,
11.8 ± 0.87; and ACT 309 ± 38 (n = 6 measurements).
The three sequential sets of A/ distributions
for one of the rabbits are shown in Fig.
7. A summary of the mean
A/ data for each rabbit is presented in Table
4. Fourteen of 15 data sets had a
residual sum of squares (RSS) of <16.8 (30). The first
set in rabbit 4 had a RSS of 42.2, likely related to the
obvious error in mixed expired measurement of DEE (Fig. 5), and this
set was not included in the mean data of Table 4. All 14 of the 15 A/ distributions with acceptable RSS were
unimodal.
|
|
| |
DISCUSSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Applications of MIGET have been primarily limited to research
settings in adult humans and large animals, with very few studies in
small animals and no real-time use in clinical care of patients. Although MIGET provides detailed and unique information on pulmonary A/ distributions and pathophysiology, the
large blood sample volume, long analysis time, and highly technical
labor required for traditional MIGET by GC have restricted the
widespread adoption of the technique. MIMS has been applied before to
MIGET to overcome some of these limitations, with one description of
initially promising results but no further reports of this approach
(23).
One problem common to early MIMS systems was very slow time responses and vacuum system memory effects for certain analytes. Although not specifically reported by Mastenbrook et al. (23), the configuration of their tubing between the membrane and the mass spectrometer inlet, as well as the variability they reported for their ACT measurements, suggests that this was a problem for this early MIMS system. Very slow time responses and vacuum system memory effects are now understood to be the result of tight adsorption of certain analytes to the walls of the vacuum system and subsequent slow release (20). We minimized these effects by keeping the length of stainless steel tubing between the membrane and the ion source as short as possible and by heating the inlet, the vacuum system, and the connection tubing. The small slow component of our system's time response for ACT and DEE, observed with step increases and step decreases in gas partial pressure, was likely due to these effects. Avoidance of long sections of unheated tubing between the membrane and the ion source, however, kept these effects within tolerable limits.
Another important limitation of prior MIMS systems has been the stirring effect, which refers to the difference in signal between stagnant liquid and rapidly flowing liquid (12). Stirring effect is a consequence of resistance to diffusion in the liquid boundary layer near the membrane and is a function of the permeability (product of diffusivity and solubility) of the inert gas in the liquid vs. the inert-gas permeability in the membrane. For a MIMS system with significant stirring effect, the mass spectrometer signal is dependent not only on the inert-gas partial pressure in the liquid, but also on the inert-gas solubility and diffusivity in the liquid sample, as well as the flow field in the vicinity of the membrane. The MIMS system applied by Mastenbrook et al. had significant stirring effect (4), and these investigators successfully overcame the resulting flow dependence by using a reproducible flow field near the membrane (23). This approach, however, did not eliminate the solubility dependence of the mass spectrometer signal.
Stirring effect and its resulting flow dependence and solubility dependence can be eliminated entirely in MIMS systems by use of thick membranes, which increase the membrane diffusional resistance relative to the diffusional resistance of the liquid sample. Thick membranes, however, lead to unacceptably long time responses because of the dependence of membrane time response on the square of membrane thickness (42).
MMIMS offers a solution to this dilemma. In MMIMS, confinement of the membrane to a small pore results in both a rapid time response and minimal stirring effect. The rapid time response is a consequence of the use of very thin membranes in MMIMS. The reductions in stirring effect are thought to be the result of the three-dimensional diffusion profiles in the liquid near the membrane (1). The inert-gas concentration profiles in the liquid around a pore that approximates a point with nearly zero area are expected, on the basis of symmetry, to be hemispherical, whereas the concentration profiles within the membrane are expected to be approximately planar. Consequently, the area available for diffusion in the membrane should be much smaller than the area available for diffusion in the liquid, thereby increasing the resistance to diffusion in the membrane vs. the liquid. The pore size cannot be reduced without limit, because reductions in membrane area also reduce the sensitivity of the system (mass spectrometer current/gas partial pressure). The data in Figs. 1-3 and Table 2, however, suggest that, with a sensitive mass spectrometer, the pore area can be small enough to substantially reduce stirring effect and still have enough sensitivity for accurate measurement. In our system, we restored some sensitivity by the use of multiple pores. For several micropores functioning independently, it is expected that sensitivity will increase linearly with the number of pores, but the system time response and stirring effect for a population of pores should be approximately independent of the number of pores.
The insensitivity of the mass spectrometer signal to changes in sample flow over the probe tip is a consequence of the small stirring effect for the MMIMS probe. Although our system did have some flow dependence at low flows for some of the gases (most noticeably SF6 and ENF), the undetectable changes in signal with increases in flow to rates >1.0 ml/min indicate that sample flows of 1.0 ml/min were adequate to overcome the small amount of stirring effect from the MMIMS probe.
In addition to insensitivity to sample flow, the small stirring effect for the MMIMS probe results in insensitivity to variations in inert-gas solubility in the sample, as shown in Fig. 2. Solubility of inert gases in blood samples is known to vary from subject to subject, due primarily to variations in blood protein content and blood lipid content (22, 38). Therefore, even for a subject in which the inert-gas solubilities are at the lowest values in the distribution of solubilities for a population of subjects, the inert-gas solubilities in water are expected to be lower. Similarly, inert-gas solubilities in 20% Intralipid solution are expected to be larger than any of the highest values in blood encountered in any individual subject. Even for these extremes in values of solubility for water and 20% Intralipid, the variations of mass spectrometer signal with solubility shown in Fig. 1 are <12% for SF6, <6% for ENF, and insignificant for the other gases.
Although the mass spectrometer signal is minimally dependent on the gas solubility in the liquid sample, the signal does depend on whether the sample is gas or liquid (Table 1). The difference in sensitivity for gas-phase measurements vs. liquid-phase measurements may be in part due to temperature dependence of permeation through the membrane and may be due in part to swelling effects in the silicone membrane. Membrane temperature for the blood samples is expected to equal the blood sample temperature at 38°C. The gas samples, however, are injected at room temperature. The membrane temperature for the gas samples in our present system, determined by the balance between heat losses to the gas surrounding the probe and heat conducted along the probe tubing from the heated ion source, is ~30°C. Membrane temperature affects both the solubility and the diffusivity of the inert gases in the membrane (5) and, therefore, is expected to influence sensitivity. Silicone membranes are also known to expand slightly when immersed in aqueous samples (21), a phenomenon known as swelling, and membrane swelling is known to affect the permeability of polymer membranes for most gases (21, 25). Further work will be required to assess the relative contributions of each of these mechanisms.
The MMIMS system rapidly provides direct measurements of inert-gas
partial pressures in small blood samples, with no extraction into a gas
phase, and minimal dependence on inert-gas solubility in blood. In this
report, we have applied MMIMS to MIGET and determination of pulmonary
A/ distributions, but there are other
applications in physiology in which these features could be
advantageous, such as the estimation of tissue blood flow
(16) and determinations of heterogeneity of
perfusion-to-volume ratios in tissue (13).
Any six inert gases can be used for MIGET, provided that the gas solubilities in blood cover a wide range of solubilities and are reasonably evenly distributed within that range of solubilities (40). We used six inert gases similar to the series reported by Mastenbrook et al. (23), and the spectral overlap matrix K in Eq. 1 indicates very little overlap, with the largest being the current at the DES of 101 amu of 10.4% of the ENF current and the current at the DEE of 59 amu of 8.8% of the ACT current.
The variability in the MMIMS method shown in Table 2 is greater than the variability reported for traditional MIGET by GC analysis. Wagner et al. (38), for example, reported coefficients of variation for repeated measurements of inert-gas partial pressures in a pooled blood sample of 6% for SF6 and <3% for the other inert gases in their MIGET series. It is anticipated that the precision of the MMIMS method will continue to improve with further refinements, particularly the use of micromachining techniques to substantially increase the number of pores and more rigorous control of the membrane temperature for the analysis of the mixed expired samples.
The hemodynamic and arterial blood-gas data shown in Table 3 illustrate the stability of the animal preparation. The gradual increase in heart rate, mean arterial pressure, and PA pressure over time may have been a result of progressive sympathetic stimulation from the continuous infusion of pancuronium (8). There was no evidence, as indicated by the pulmonary artery diastolic pressures, of volume overload from the inert-gas infusate in any rabbit. The stability of the hematocrit over time suggests that the amount of blood volume sampled for MIGET in this small animal model was well tolerated. The blood sample volume for this MMIMS system was 2 ml/sample or 4 ml per set of retention and excretion data (2 ml each for the mixed venous and arterial samples). The amount of inert gas sampled into the mass spectrometer is a very small fraction of the inert gases in 2 ml of blood, and it is possible that future improvements in probe temperature control, sample flow control, and the precision of the flow channel at the probe tip in Fig. 1 could lead to even smaller blood sample volumes in future MMIMS systems.
The inspired gases were mixed from purified oxygen and nitrogen to eliminate the trace level Kr that can be commonly found in oxygen tanks. Although it is possible to carry out MIGET accurately in the presence of a stable background of exogenous Kr (30), for this experimental protocol we used purified gases to avoid any corrections for environmental background.
Because the MMIMS technique measures inert-gas partial pressures in
blood samples with minimal dependence on the inert-gas solubility in
blood, the accuracy of the retention and excretion data in Figs. 5 and
6 is unaffected by variations in inert-gas solubilities within and
between subjects. Values for solubility are required, however, for
transforming the retention and excretion data into
A/ distributions. The data in Fig. 7 and
Table 4 were calculated by using measurements of the population means for the inert-gas solubilities in rabbit blood and do not account for
variations in solubilities between individual subjects. The acceptable
RSS in 14 of 15 data sets (30) suggests that this approach
provided reasonable accuracy for recovery of
A/ distributions in normal rabbits. In
general, however, the RSS in Table 4 are larger than values that have
been reported for the traditional GC approach to MIGET in experienced
hands (14). The relatively large RSS in our experiments
could reflect either the variability in the partial pressure
measurements with MMIMS or errors in solubility as a result of
individual variations around the population mean.
The small blood sample volume and rapid analysis time for measurement
of inert-gas partial pressures by MMIMS could be used to the greatest
advantage in studies in which gas exchange and A/ distributions are changing rapidly but
solubility remains constant over the course of an experiment. For
example, MIGET by MMIMS could be used in a small-animal model to
measure sequential changes in A/
distributions in response to therapeutic interventions, and a large
blood sample volume could be collected at the end of the experiment for
measurement of inert-gas solubilities by established methods (31,
38).
The mean second moment of the perfusion distribution (log
SD) for the rabbit A/
distributions was 0.91, comparable to the values of 0.91, 0.85, 0.78, and 0.94 for normal rabbits reported by Lagerstrand and Hedenstierna
(19). These A/ distributions are
broad compared with anesthetized, ventilated normal dogs [log
SD of 0.55 (35), 0.45 (10), and 0.68 (17)] and normal humans [log
SD of 0.43 (36)]. Also, the
ventilation distribution was right-shifted toward high
A/ units with a mean
A/ value for ventilation distribution (mean A/) of 3.77, which is also comparable to the value of 2.65 reported by Lagerstrand and
Hedenstierna (19). These ventilation distributions are
substantially right-shifted compared with anesthetized, ventilated
normal dogs [mean A/ of
1.03 (17)] and normal humans [mean
A/ of
1.10 (36)]. The reasons for comparatively broad
A/ distributions in the rabbit are not clear,
but a broad A/ distribution in a small lung
with limited gravitational variation supports the evidence that factors
other than gravity play a substantial role in
A/ heterogeneity (17).
Right-shifted A/ distributions have been
reported in sheep when tidal volumes are relatively large, which is
thought to be due to overdistention of parts of the lung and reductions
in blood flow to those regions (30). Alternatively, high
A/ regions could be a consequence of exchange of the highly soluble gases in the conducting airways
(32).
In summary, the unique combination of rapid response time and minimal
stirring effect provided by MMIMS resulted in a system for the rapid,
direct measurement of inert-gas partial pressures in small blood
samples, with minimal dependence on the solubility of the inert gas in
blood. The blood-sample volume for a complete set of retention and
excretion data for MIGET was 4-5 ml, and analysis time for a
single A/ determination was ~8 min. This new approach for MIGET has potential to make determinations of A/ distributions more widely available in
research applications in small animal models.
| |
ACKNOWLEDGEMENTS |
|---|
The authors gratefully acknowledge Dr. Peter D. Wagner for
supplying the computer program for the A/ transformations.
| |
FOOTNOTES |
|---|
This study was funded in part by National Heart, Lung, and Blood Institute Grants R41-HL-59052 and HL-09040; by SpectruMedix Corporation, State College, PA; and by the University of Pennsylvania Research Foundation.
Address for reprint requests and other correspondence: J. E. Baumgardner, Dept. of Anesthesia, Univ. of Pennsylvania, 3400 Spruce St., Philadelphia, PA 19104-4283 (E-mail: jbaumgar{at}mail.med.upenn.edu).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 26 April 1999; accepted in final form 30 May 2000.
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
1.
Baumgardner, JE,
Quinn JA,
and
Neufeld GR.
Micropore membrane inlet mass spectrometer probes suitable for measurement of tissue surface gas tensions.
J Mass Spectrom
30:
563-571,
1995.
2.
Beydon, L,
Cinotti L,
Rekik N,
Radermacher P,
Adnot S,
Meignan M,
Harf A,
and
Lemaire F.
Changes in the distribution of ventilation and perfusion associated with separation from mechanical ventilation in patients with obstructive pulmonary disease.
Anesthesiology
75:
730-738,
1991[ISI][Medline].
3.
Bier, ME,
and
Cooks RG.
Membrane interface for selective introduction of volatile compounds directly into the ionization chamber of a mass spectrometer.
Anal Chem
59:
597-601,
1987.
4.
Brantigan, JW,
Gott VL,
Vestal ML,
Fergusson GJ,
and
Johnston WH.
A nonthrombogenic diffusion membrane for continuous in vivo measurement of blood gases by mass spectrometry.
J Appl Physiol
28:
375-377,
1970.
5.
Crank, J.
Diffusion in Polymers. New York: Academic, 1968.
6.
Dantzker, DR,
Brook LJ,
Dehart P,
Lynch JP,
and
Weg JG.
Ventilation-perfusion distributions in the adult respiratory distress syndrome.
Am Rev Respir Dis
120:
1039-1052,
1979[ISI][Medline].
7.
Dantzker, DR,
and
D'Alonzo GE.
The effect of exercise on pulmonary gas exchange in patients with severe chronic obstructive pulmonary disease.
Am Rev Respir Dis
134:
1135-1139,
1986[ISI][Medline].
8.
Docherty, JR,
and
McGrath JC.
Sympathomimetic effects of pancuronium bromide on the cardiovascular system of the pithed rat: a comparison with the effects of drugs blocking the neuronal uptake of noradrenaline.
Br J Pharmacol
64:
589-599,
1978[ISI][Medline].
9.
Evans, JW,
and
Wagner PD.
Limits on A/ distributions from analysis of experimental inert gas elimination.
J Appl Physiol
42:
889-898,
1977
10.
Frans, A,
Clerbaux T,
Willems E,
and
Kreuzer F.
Effect of metabolic acidosis on pulmonary gas exchange of artificially ventilated dogs.
J Appl Physiol
74:
2301-2308,
1993
11.
Gea, J,
Roca J,
Torres A,
Agusti AG,
Wagner PD,
and
Rodriguez-Roisin R.
Mechanisms of abnormal gas exchange in patients with pneumonia.
Anesthesiology
75:
782-789,
1991[ISI][Medline].
12.
Gronlund, J,
Johansen O,
and
Jensen B.
New transcutaneous PO2 probe based on mass spectrometry with low stirring effect.
Med Biol Eng Comput
23:
254-257,
1985[ISI][Medline].
13.
Gronlund, J,
Malvin GM,
and
Hlastala MP.
Estimation of blood flow distribution in skeletal muscle from inert gas washout.
J Appl Physiol
66:
1942-1955,
1989
14.
Hachenberg, T,
Tenling A,
Hansson HE,
Tyden H,
and
Hedenstierna G.
The ventilation-perfusion relation and gas exchange in mitral valve disease and coronary artery disease.
Anesthesiology
86:
809-817,
1997[ISI][Medline].
15.
Hlastala, MP.
Multiple inert gas elimination technique.
J Appl Physiol
56:
1-7,
1984
16.
Hoeft, A,
Sonntag H,
Stephan H,
and
Kettler D.
The influence of anesthesia on myocardial oxygen utilization efficiency in patients undergoing coronary bypass surgery.
Anesth Analg
78:
857-866,
1994
17.
Kallas, HJ,
Domino KB,
Glenny RW,
Anderson EA,
and
Hlastala MP.
Pulmonary blood flow redistribution with low levels of positive end-expiratory pressure.
Anesthesiology
88:
1291-1299,
1998[ISI][Medline].
18.
Kotiaho, T,
Lauritsen FR,
Choudhury TK,
Cooks RG,
and
Tsao GT.
Membrane introduction mass spectrometry.
Anal Chem
63:
875A-883A,
1991.
19.
Lagerstrand, L,
and
Hedenstierna G.
Gas-exchange impairment: its correlation to lung mechanics in acute airway obstruction (studies on a rabbit asthma model).
Clin Physiol
10:
363-380,
1990[ISI][Medline].
20.
Lauritsen, FR.
A new membrane inlet for on-line monitoring of dissolved, volatile organic compounds with mass spectrometry.
Int J Mass Spec Ion Proc
95:
259-268,
1990.
21.
Lebovits A. Permeability of polymers to gases, vapors, and
liquids. Modern Plastics 43: 139-150, 196-213,
1966.
22.
Lerman, J,
Gregory GA,
Willis MM,
and
Eger EI II.
Age and solubility of volatile anesthetics in blood.
Anesthesiology
61:
139-143,
1984[ISI][Medline].
23.
Mastenbrook, SM, Jr,
Dempsey JA,
and
Massaro TA.
Ventilation-perfusion ratio distributions by mass spectrometry with membrane catheters.
J Appl Physiol
53:
770-778,
1982
24.
Mastenbrook, SM, Jr,
Massaro TA,
and
Dempsey JA.
Multiple gas measurements at trace levels with a quadrupole mass spectrometer: gas phase calibration.
J Appl Physiol
44:
634-639,
1978
25.
Michaels, AS,
and
Bixler HJ.
Membrane permeation: theory and practice.
In: Progress in Separation and Purification, edited by Perry E.. New York: Interscience, 1968, p. 143-186.
26.
Moinard, J,
Manier G,
Pillet O,
and
Castaing Y.
Effect of inhaled nitric oxide on hemodynamics and A/ inequalities in patients with chronic obstructive pulmonary disease.
Am J Respir Crit Care Med
149:
1482-1487,
1994[Abstract].
27.
Prediletto, R,
Formichi B,
Begliomini E,
Fornai E,
Viegi G,
Ruschi S,
Paoletti P,
Giannella AN,
Santolicandro A,
and
Giuntini C.
Ventilation-perfusion heterogeneity and gas exchange variables in acute pulmonary embolism evaluated by two different computerized techniques.
Int J Clin Monit Comput
5:
221-227,
1988[ISI][Medline].
28.
Putensen, C,
Rasanen J,
Lopez FA,
and
Downs JB.
Effect of interfacing between spontaneous breathing and mechanical cycles on the ventilation-perfusion distribution in canine lung injury.
Anesthesiology
81:
921-930,
1994[ISI][Medline].
29.
Roca, J,
Ramis LI,
Rodriguez-Roisin R,
Ballester E,
Montserrat JM,
and
Wagner PD.
Serial relationships between ventilation-perfusion inequality and spirometry in acute severe asthma requiring hospitalization.
Am Rev Respir Dis
137:
1055-1061,
1988[ISI][Medline].
30.
Shimazu, T,
Yukioka T,
Ikeuchi H,
Mason AD, Jr,
Wagner PD,
and
Pruitt BA, Jr.
Ventilation-perfusion alterations after smoke inhalation injury in an ovine model.
J Appl Physiol
81:
2250-2259,
1996
31.
Snedden, W,
LeDez K,
and
Manson HJ.
A new method for the measurement of gas solubility.
J Appl Physiol
80:
1371-1378,
1996
32.
Souders, JE,
George SC,
Polissar NL,
Swenson ER,
and
Hlastala MP.
Tracheal gas exchange: perfusion-related differences in inert-gas elimination.
J Appl Physiol
79:
918-928,
1995
33.
Tenling, A,
Hachenberg T,
Tyden H,
Wegenius G,
and
Hedenstierna G.
Atelectasis and gas exchange after cardiac surgery.
Anesthesiology
89:
371-378,
1998[ISI][Medline].
34.
Tokics, L,
Hedenstierna G,
Svensson L,
Brismar B,
Cederlund T,
Lundquist H,
and
Strandberg A.
A/ distribution and correlation to atelectasis in anesthetized paralyzed humans.
J Appl Physiol
81:
1822-1833,
1996
35.
Wagner, PD,
Laravuso RB,
Goldzimmer E,
Naumann PF,
and
West JB.
Distributions of ventilation-perfusion ratios in dogs with normal and abnormal lungs.
J Appl Physiol
38:
1099-1109,
1975
36.
Wagner, PD,
Laravuso RB,
Uhl RR,
and
West JB.
Continuous distributions of ventilation-perfusion ratios in normal subjects breathing air and 100% O2.
J Clin Invest
54:
54-68,
1974.
37.
Wagner, PD,
and
Lopez FA.
Gas chromatography techniques in respiratory physiology.
In: Techniques in the Life Sciences, edited by Otis AB.. County Clare, Ireland: Elsevier, 1984, p. 1-24.
38.
Wagner, PD,
Naumann PF,
and
Laravuso RB.
Simultaneous measurement of eight foreign gases in blood by gas chromatography.
J Appl Physiol
36:
600-605,
1974
39.
Wagner, PD,
and
Rodriguez-Roisin R.
Clinical advances in pulmonary gas exchange.
Am Rev Respir Dis
143:
883-888,
1991[ISI][Medline].
40.
Wagner, PD,
Saltzman HA,
and
West JB.
Measurement of continuous distributions of ventilation-perfusion ratios: theory.
J Appl Physiol
36:
588-599,
1974
41.
Westover, LB,
Tou JC,
and
Mark JH.
Novel mass spectrometric sampling device-hollow fiber probe.
Anal Chem
46:
568-571,
1974.
42.
Woldring, S.
Tutorial: biomedical application of mass spectrometry for monitoring pressures. A technical review.
J Assoc Advan Med Instr
4:
43-56,
1970.
43.
Yamaguchi, K,
Mori M,
Kawai A,
Takasugi T,
Asano K,
Oyamada Y,
Aoki T,
Fujita H,
Suzuki Y,
Yamasawa F,
and
Kawashiro T.
Ventilation-perfusion inequality and diffusion impairment in acutely injured lungs.
Respir Physiol
98:
165-177,
1994[ISI][Medline].
44.
Zavala, E,
Ferrer M,
Polese G,
Masclans JR,
Planas M,
Milic-Emili J,
Rodriguez-Roisin R,
Roca J,
and
Rossi A.
Effect of inverse I:E ratio ventilation on pulmonary gas exchange in acute respiratory distress syndrome.
Anesthesiology
88:
35-42,
1998[ISI][Medline].
This article has been cited by other articles:
|
|
 |
|
  K. Markstaller, H.-U. Kauczor, N. Weiler, J. Karmrodt, M. Doebrich, M. Ferrante, M. Thelen, and B. Eberle Lung density distribution in dynamic CT correlates with oxygenation in ventilated pigs with lavage ARDS{dagger} Br. J. Anaesth., November 1, 2003; 91(5): 699 - 708. [Abstract] [Full Text] [PDF]
|
|
| |||||||||
, 
,