| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Division of Pulmonary and Critical Care Medicine, Department of Medicine, and Departments of Atmospheric Sciences, Environmental Health, and Statistics, University of Washington, Seattle, Washington 98195-6522
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
ACKNOWLEDGEMENTS
FOOTNOTES
REFERENCES
ABSTRACT 
Robertson, H. Thomas, Robb W. Glenny, Derek Stanford, Lynn
M. McInnes, Daniel L. Luchtel, and David Covert. High-resolution maps of regional ventilation utilizing inhaled fluorescent
microspheres. J. Appl. Physiol. 82(3):
943-953, 1997.
The regional deposition of an inhaled aerosol of
1.0-µm diameter fluorescent microspheres (FMS) was used to produce
high-resolution maps of regional ventilation. Five anesthetized, prone,
mechanically ventilated pigs received two 10-min inhalations of pairs
of different FMS labels, accompanied by intravenous injection of
15.0-µm radioactive microspheres. The lungs were air dried and cut
into 1.9-cm3 pieces, with notation
of the spatial coordinates for each piece. After measurement of
radioactive energy peaks, the tissue samples were soaked in
2-ethoxyethyl acetate, and fluorescent emission peaks were recorded for
the wavelengths specific to each fluorescence label. The correlation of
fluorescence activity between simultaneously administered inhaled FMS
ranged from 0.98 to 0.99. The mean coefficient of variation for
ventilation for all 10 trials (47.9 ± 8.1%) was similar to that
for perfusion (46.2 ± 6.3%). No physiologically significant
gravitational gradient of ventilation or perfusion was present in the
prone animals. The strongest predictor of the magnitude of regional
ventilation among all animals was regional perfusion
(r = 0.77 ± 0.13).
pulmonary aerosol deposition; ventilation heterogeneity; ventilation-perfusion heterogeneity; gravitational gradient; radial
gradient; pigs
INTRODUCTION THE OVERALL EFFICIENCY of pulmonary gas exchange is
determined by the matching of ventilation to pulmonary blood flow
within the lung. The established method to characterize this match of gas and blood flow is the multiple inert-gas-elimination technique (MIGET) (23). Although this technique has provided important insight
into lung function, it cannot provide spatial information on the
distribution of ventilation-perfusion heterogeneity within the lung.
Information on the regional distribution of ventilation and perfusion
has been obtained from radionuclide labels marking blood flow and
regional ventilation, with recent studies utilizing either
single-photon emission tomography or positron emission tomography (PET). Although a number of different investigational approaches have been applied to the characterization of
ventilation-perfusion heterogeneity by application of radionuclide
scans, all of these methods confront two limitations. First, the
ventilation and blood flow markers cannot be administered
simultaneously. The temporal shifts observed in pulmonary blood flow
distribution (9) raise a concern for error introduced by any sequential
imaging procedure. Second, the best resolution attainable even for PET
images is limited by the distance a positron can travel in lung before
encountering an electron. This artifact ensures that the resolving
power of PET in the lung can never surpass 1-2
cm3, which will minimize the
estimate of spatial heterogeneity. Spatial heterogeneity of perfusion
in the lung has fractal characteristics (4, 13, 16) that
dictate that the extent of heterogeneity measured is dependent on
resolution of the data set. The purpose of the present study was to
establish a high-resolution measurement of regional ventilation that
could be measured simultaneously with regional perfusion at the same
level of resolution, thereby permitting acquisition of accurate spatial
information on the distribution of ventilation-perfusion heterogeneity.
Radionuclide scans of
99mTc-labeled 0.01-µm aerosol
deposited in the lung correlate well with scans of regional ventilation
measured with radioactive gases at the level of resolution attained
with clinical radionuclide scanners (1). It is not established, however, that an aerosol deposition method faithfully reflects regional
ventilation at higher degrees of spatial resolution. At the alveolar
level, gases are distributed primarily by diffusion, whereas the
deposition of aerosols in the 1.0-µm range is determined by inertial
and gravitational influences (3). Both modeling studies (5, 6, 20) and
studies of radioactive aerosol deposition in isolated lungs (21) have
suggested that at normal respiratory rates, as long as the inhaled
particles maintain an aerodynamic diameter of <2.0-µm, only a minor
fraction of the particles will deposit in conducting airways. With
appropriate choices of ventilatory parameters and aerosol
characteristics, the deposition of aerosol can provide a sufficient
estimate of regional ventilation for a lung volume of sufficient size
to ensure that convective flow is the predominant factor influencing
movement of both gases and aerosols to that volume.
Zeltner et al. (26) exposed spontaneously breathing hamsters to an
aerosol of 0.9-µm fluorescent microspheres (FMS). Subsequent microscopic examination of the hamster lungs revealed that 95% of the
FMS were deposited in gas-exchanging regions of the lungs. From the
work by Zeltner et al. (26) and previous theoretical studies, it
appeared plausible that pulmonary deposition of FMS aerosols could be
used to characterize overall ventilation heterogeneity at a high level
of resolution. An advantage afforded by FMS aerosols in contrast to the
radionuclide labels is that at least eight different fluorescence
labels can be incorporated into uniformly sized aerosol particles (10),
providing the opportunity for sequential measurements of ventilation
distribution.
The fractal properties of perfusion distribution in the lung (12, 13)
raise the question of whether this same property will be present in the
ventilation distribution. High-resolution studies of pulmonary
parenchymal movement during tidal ventilation (14, 18) provided
evidence that pulmonary ventilation may possess the same characteristic
of scale-dependent heterogeneity as that observed for pulmonary blood
flow. However, the relationship of high-resolution measurements of
parenchymal movement to alveolar ventilation is indirect; simultaneous
measurements of regional pulmonary blood flow were not included in
these studies. Wilson and Beck (25) provided an estimate of ventilation
heterogeneity based on measurements of pulmonary blood flow from
systematically sampled small pieces of lung combined with
ventilation-perfusion heterogeneity estimates made by MIGET (22).
Although an estimate of ventilation heterogeneity was obtained, this
approach required an assumed correlation between regional ventilation
and perfusion.
Simultaneous comparison of the heterogeneity of regional ventilation to
the heterogeneity of regional perfusion has not been reported at the
high level of resolution attained in recent studies of perfusion
heterogeneity in the lung. A preliminary report on experiments
utilizing goat lungs with simultaneously administered intravascular and
aerosol radionuclide markers, cut and analyzed at comparable high
resolution, has demonstrated high correlation between ventilation and
perfusion (15). Although the abstract itself did not specifically
compare the observed heterogeneity of ventilation with that of
perfusion, the information was available at the time of presentation.
The obvious expectation of simultaneous high-resolution measurements of
ventilation and perfusion would be that heterogeneity of ventilation
would equal that of perfusion, with a high correlation between the two
measurements. Another equally plausible hypothesis, consistent with the
observed homogeneity of overall pulmonary gas exchange, would be a
smaller degree of regional ventilation heterogeneity in comparison with
perfusion heterogeneity. By this hypothesis, the alveolar gas
composition within adjacent regions would be homogenized by both
reinspired dead space gas and by collateral ventilation, so that the
end-capillary gas concentrations would be more similar than would
otherwise be predicted from the heterogeneous perfusion. Hence a lower
overall heterogeneity of regional ventilation would be predicted.
Simultaneous measurement of ventilation and perfusion in the lung at
the same level of resolution is needed to test these two
hypotheses.
This study introduces a method of inhaled FMS administration that
provides high-resolution maps of regional ventilation; it also
demonstrates that adequate fluorescence signals can be attained to
provide data comparable with data obtained from pulmonary perfusion studies using FMS. FMS measurements of regional distribution of pulmonary ventilation are paired with simultaneous measurements of
regional blood flow to define overall ventilation-perfusion heterogeneity and to determine spatial distribution for the two parameters.
METHODS
For each experimental measurement, an aerosol consisting of a mixture of two FMS fluorescent labels was administered via a positive-pressure ventilator system for a period of 10 min. Radioactive microspheres (RMS) of 15-µm diameter were simultaneously injected intravenously in multiple, small, evenly spaced increments during the same 10-min period. The RMS were vortexed, sonicated, and diluted with saline to a volume of 10 ml. During the measurement period, the RMS were kept in suspension in the saline injectate by mixing between two syringes attached to the venous stopcock. A minimum of 2.3 × 106 microspheres of each RMS label were administered. Thirty minutes later, after the aerosol generation system had been flushed clean of the first two FMS labels and baseline physiological measurements were repeated, a different pair of FMS labels was inhaled for 10 min with simultaneous injection of a different RMS. Immediately after the second run, the pigs were intravenously injected with 10,000 U of heparin and exsanguinated under deep pentobarbital anesthesia. The animals were turned to a supine position, a thoracotomy was performed, and the pulmonary artery and left atria were cannulated. The lung vasculature was flushed with a solution of 2% dextran at a perfusion pressure of 10 cmH2O, and the lungs were removed from the thorax and cut free from cardiac structures. No edema fluid was observed in the airways of the excised lungs. The lungs were inflated to a pressure of 18 Torr with dry heated air, each lobe was glued to adjacent lobes to maintain the appropriate anatomic configuration during drying, and the lungs were punctured 20-30 times with an 18-gauge needle to facilitate rapid, even drying. The lungs were air-dried inflated at a pressure of 18 Torr for 3 days. Lung preparation and sample measurement. After the 3-day drying period, the lungs were coated with 5 - 10 mm of polyurethane foam (Kwik Foam, DAP, Dayton OH) and dried for an additional day. They were then suspended in a box with the dorsal surface parallel to the box walls and encased in a block of rapidly setting isocyanate foam (Polyfoam A & B, Tacoma Fiberglass Products, Tacoma, WA), as previously described, to provide a fixed coordinate system for cutting (11). The lungs were cut along orthogonal coordinates in 1.9-cm3 cubes, any attached foam was removed, and the pieces were weighed. Pieces weighing <8.0 mg were discarded (mean piece weight = 41.3 + 24.6 mg). Assigned spatial coordinates and lobe location of each piece were recorded. Conducting airway within each piece was graded on a scale of 0 to 5 (0 = no visible airways; 5 = conducting airway >75% of total piece volume). Each lung piece was placed in a well counter (Packard Autogamma 5000 series, Downers Grove, IL) to measure the activity of the radioactive labels. Each piece was then soaked in 1.5 ml of 2-ethoxyethyl acetate (Cellosolve, Aldrich Chemical, Milwaukee, WI) for 48 h to dissolve the FMS and extract the fluorescent marker. The extract was transferred to a cuvette, and the fluorescence peak of each of the FMS was measured on a fluorescence spectrophotometer (Perkin-Elmer luminescence spectrometer LS 50B, Norwalk, CT), as previously described (8, 10). Sections for evaluation with fluorescence microscopy were obtained from each animal, selecting three to five pieces of dried lung tissue that had airway scores of 3 or 4 from caudal lobe (these sections were not soaked in Cellosolve and were not included in the statistical analysis). Sections of 50-, 75-, or 100-µm thickness were cut with a Vibrotome (Technical Products International, St. Louis, MO) and mounted dry on slides under coverslips. Data reduction and statistical analysis. Pieces containing only trachea were excluded, leaving 580-794 pieces from each pig. Measurements of radioactivity at each energy window for each piece were corrected for decay time and spillover, as described previously (11). Overlap corrections were not required for the four fluorescent labels used in these studies (10), but a small fluorescence blank signal was subtracted from each FMS reading, based on the fluorescence signal obtained for that wavelength from a cuvette filled with solvent alone. Each FMS aerosol data set with paired, simultaneously administered fluorescent labels was examined by plotting the difference between the two regional ventilation estimates and their mean value. Pieces containing FMS pairs with discrepancies greater than three SDs from zero were identified and excluded from analysis; between 7 and 13 pieces were thereby excluded from each of the 10 runs performed on the five pigs. The discordant signals invariably had one very high fluorescence signal compared with the three other ventilation and two blood flow measurements made on that piece. Whenever large discrepancies were identified in paired FMS measurements, repeat measurements of all four fluorescence emission windows yielded similar values. No obvious pattern explained the large discordance between fluorescence signals in those few pieces with respect to location, piece size, piece blood flow, or the proportion of airway in the piece. Preliminary studies of unlabeled lung pieces that were prepared by the same extraction, drying, and cutting procedures described above and then suspended in solvent revealed a variable, low-level fluorescence signal present primarily in the blue-green and yellow-green emission spectra. For the animals reported in this study, if the mean FMS fluorescence signal per piece for those labels was less than four times the fluorescence signal of unlabeled lung, those data were not used for the comparison of two simultaneous measurements of minute ventilation. This criterion resulted in exclusion from analysis of either blue green or yellow green from four of the 10 runs performed to estimate the error of the aerosol method. Three of the four runs that were excluded from the paired-data analysis were performed on the unmodified ventilator system. All 10 runs with paired-FMS administration had at least one fluorescence signal that met the criterion for adequate signal intensity. For consistency, all data analysis of spatial determinants of ventilation used only the red and orange signals, because these signals were adequate for all of the experiments. Mean values for radioactivity and fluorescence for each of the RMS and FMS markers and for mean piece weight were calculated for each animal. Fluorescence and radionuclide signals from each lung piece were normalized by mean signal (flow normalized) and then by mean weight (flow and weight normalized). The latter normalization procedure described both ventilation and perfusion in terms of flows per unit alveolar tissue mass, with dry, blood-free, piece weight chosen as the best available estimate of alveolar mass. The agreement between paired FMS aerosols was described by a Pearson correlation coefficient. The comparison between mean ventilation and perfusion estimates over time and comparison of ventilation and perfusion at different times were described by weighted correlation coefficients (5), using the lung piece weight as the weighting factor. The spatial coordinates for the pieces from each lung were adjusted so that the dorsal-caudal tip of the caudal lobe was set to zero, and the dorsal surface of the caudal lobe was given a negative 15° slope to best approximate the normal orientation of the lung in a prone pig (19). Gravitational gradients were then calculated incorporating the distance from the zero set point of that adjusted vertical axis of the lung, using the flow- and weight-normalized values measured for the labels in each lung piece. Radial gradients were estimated as the distance from the center of mass for each lobe, again using the flow- and weight-normalized values for each label. The gravitational and radial gradients were calculated by using weighted linear regression (7), where the weight of the lung piece was the weighting factor. Weighted correlation and regression assigned each weight- and flow-normalized lung piece an appropriate influence in the overall determination of model parameters. To analyze spatial trends in ventilation and flow distribution within the lungs, both gravitational and radial gradients were calculated for each animal, with 95% confidence intervals for these slopes estimated for all animals by using a one-sample t-test at each time. The indexes of dispersion of ventilation and perfusion on the log
A/
scale (described as log SD of alveolar ventilation or log SD of perfusion) were obtained by first calculating a perfusion- and a
ventilation-weighted mean log
A/
and then using those mean values respectively for the calculation of
the ventilation-weighted SD or the perfusion-weighted SD. Unlike this
calculation when utilized in the MIGET, where blood flow and
ventilation calculations produce 50 evenly spaced
A/
compartments (23), our data sets were individual measurements of
ventilation and perfusion for each cut piece. It was therefore
necessary to incorporate a weighting factor so that the
relative contribution of each piece to the ventilation or perfusion
dispersion would be appropriately represented.
RESULTS
Correlation between simultaneously administered fluorescence labels. Each administered aerosol marker contained a mixture of two different fluorescent labels. Thus relative discrepancies between fluorescence signals in any lung piece served to estimate experimental error of aerosol administration. As noted previously, each animal had between 7 and 13 pieces where the fluorescence signal for one of the labels was very high, producing a between-FMS label difference of greater than three SDs from the mean difference between paired labels. Attempts were made to characterize further those few pieces excluded from analysis for this discrepancy. No consistent characteristics could be identified with respect to piece location, piece size, presence of pleural surface, or conducting airway. With one exception, all pieces identified for exclusion based on one paired discrepancy had a normal concurrence of labels for the other aerosol pair. In addition to those rare large discrepancies, another potential source of error was identified. Previous fluorescence measurements of unlabeled lung pieces revealed that a small, nonspecific fluorescence signal was produced secondary to light scattering from fragments of lung in the sample. Because the extent of this artifact from any given piece was not predictable, only those runs where both FMS markers had high mean fluorescence signals were accepted for this error analysis. In the six aerosol runs with adequate fluorescence signals from both inhaled markers, the correlation between the two labels ranged between 0.98 and 0.99 (intercept range from
0.06 to 0.09 and slope range from 0.91 to 1.07). The best correlations (Fig. 3)
were obtained from the higher deposition of fluorescence reached in the
final three animals after modification of the ventilator system to
contain the aerosol within a circuit containing only conductive tubing
and an anesthesia bag (Fig. 1). These comparisons demonstrate that with
acceptably high fluorescence signals, nonspecific fluorescence does not
appreciably influence the measurements.
Estimates of aerosol deposition on conducting airways. A potential source of error in the FMS estimates of regional ventilation may arise from significant deposition of FMS on conducting airways. Evidence for this artifact was sought by two approaches. First, thick sections were cut from three to five representative dried lung pieces from each animal and examined with a fluorescence microscope. The initial sections cut at arbitrary intervals revealed little deposition of FMS on conducting airways and showed extensive heterogeneous deposition of FMS in the alveolar spaces (Fig. 4). However, when serial transverse sections of several airways were cut, FMS deposition was found at airway bifurcations. The heterogeneity of deposition in the airways made it difficult to quantitatively estimate the number of particles in conducting airways relative to the density of deposition in alveolar spaces.
The second approach compared the fluorescence signals as a function of the estimated airway content of each lung piece. A plot of per-piece ventilation and blood flow measurements vs. airway content score (Fig. 5, A and B) shows a decrease in both measurements as the proportion of conducting airway in the sample increased. The pieces with airway scores of 0 were generally peripheral incomplete cubes (mean weight 22-28 mg) that contained pleural surface. Thus, because of the pleural weight, the weight normalization procedure that we employed most likely underestimated the true flow per alveolus value that the weight normalization was meant to represent. Similarly, the normalization process probably underestimates the true flow per alveolus value for the pieces with airway scores of 4 and 5 because of the weight of the conducting airways. Nevertheless, the mean fluorescence signal in pieces including primarily conducting airway (scores 4 and 5, mean weight 116-150 mg) was 39 ± 7% of those pieces containing no appreciable conducting airway content (scores 0 and 1, mean weight 32-40 mg), which at least demonstrates no gross effect of airway content on fluorescence deposition. In addition, pieces with airway scores of 4 and 5 constituted <2% of the total sample and were not concentrated at the extremes of radial or gravitational distances. Thus even a moderate excess deposition of FMS in the large airways would minimally influence the overall data analysis. Similar trends for relationship between airway score and blood flow were also noted (Fig. 5), again suggesting that the trends in ventilation distribution were not distorted by disproportionate fluorescence deposition in airways. Although this analysis excludes gross deposition of FMS on conducting airways, it is not sensitive enough to detect a small excess of FMS deposition in those lung pieces containing a large fraction of conducting airways.
Physiological effects of aerosol administration. During each aerosol administration period, arterial blood gases, pulmonary arterial pressure, systemic arterial pressure, peak airway pressure, and cardiac output were monitored. Comparison of baseline measurements with those taken after the first aerosol administration (Table 1) shows no consistent trends in the measured parameters. Exercising the precautions described in METHODS to generate a sterile, pyrogen-free aerosol, no changes were observed in these parameters for four of the five pigs. During preliminary experiments, pulmonary hypertension and hypoxemia were noted after administration of an aerosol that was subsequently found to have bacterial contamination. The one animal with abnormal gas exchange (pig 4) showed a marked redistribution of pulmonary blood flow during the second experiment, with relatively little change in ventilation distribution. No evidence of aerosol contamination was found for this animal.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||
Topography of aerosol deposition. The caudal lobes showed a trend toward higher weight-normalized ventilation, but this did not attain statistical significance. Gravitational gradients for ventilation in lungs of prone pigs similarly failed to show a statistically significant effect between all animals and times. Table 4 presents the slopes of weighted least squares fits of regional ventilation to distance from the highest point on the surface of the prone lung for the five animals at the two times; a positive slope represents greater ventilation in the more dependent regions of the lung. The 95% confidence intervals for the slopes at both measurement times include positive and negative values, failing to reject the null hypothesis that the gravitational ventilation gradient was zero. Table 5 presents fits of the data to radial distance from the center of the lobes, where a negative coefficient indicates lower flows distant from the center of the lobes. Although the magnitude of the mean slope for a radial gradient was as small as that obtained for the gravitational gradient, the effect was more consistent, with 95% confidence limits not including zero, and P values of <0.005 for both runs. The magnitude of both gradients was small compared with the overall degree of ventilation heterogeneity observed in these lungs. The r2 for the single predictor model of the relationship between regional ventilation and radial distance from the center of the lobe averaged 0.039; <4% of the variability of regional ventilation in prone animals could be attributed to a radial-distribution effect.
|
||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||
A/
units; the blood-gas measurements and pulmonary vascular pressures did
not differ from the other animals. Pig
4 was the one animal with an abnormal
A-aDO2
during both measurements. Figure
7 utilizes a simple scatterplot to
illustrate the correlation of ventilation and pulmonary blood flow for
pig 5, run
1. Figure 8,
A and
B, presents the same data showing the
individual piece ventilation and blood flow as a function of the piece
A/
ratio, a format similar to that used for presentation of results
obtained by the MIGET (23). In this format, it is apparent that the
pieces with high values of either ventilation or blood flow are
centered over the mean
A/
ratio and that relatively little ventilation or blood flow is directed
to those pieces away from that mean. The relative breadth of the
ventilation and perfusion peaks on the log
A/
scale is one index of
A/
heterogeneity that has been used as a simple description of
gas-exchange estimates of heterogeneity obtained by MIGET. Table 6
gives calculations for all runs of the log SD of ventilation and log SD
of perfusion, the two indexes of gas-exchange heterogeneity used to
describe the peak breadths. Note that pig
1 has a large SD of
A, as would be
expected with the presence of high
A/
units, and that pig 4, with the
widest
A-aDO2 has the
largest SD of . With the exception of those
measurements, the values obtained for SD of
A
and SD of are slightly below the range obtained by
MIGET measurements from normal upright human lungs (24). The
sensitivity of the combined aerosol deposition and intravascular
microsphere-injection method for estimating regional ventilation
appears adequate to describe the range of
A/
heterogeneity observed in normal lungs by MIGET.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1 · unit
tissue wt1) calculated by
multiplying cardiac output by fractional normalized flow) and
ventilation
(ml · min1 · unit
tissue wt1) calculated by
multiplying total ventilation minus anatomic dead-space ventilation by
fractional normalized ventilation (animal
5, run 1).
A/.
B: plot of blood flow data (
) from
Fig. 7 vs. log10
A/.
DISCUSSION
The application of FMS aerosol deposition as a marker for high-resolution maps of ventilation distribution offers a number of significant advantages. First, the FMS in the aerosol are monodisperse, of known size, and have no detectable concentrations of large coagulated or clumped particles. A few large particles within an aerosol could substantially misrepresent heterogeneity of ventilation when small lung pieces are analyzed. For example, a single 15-µm particle among 1-µm particles would produce a signal >3,000 times larger than anticipated. Furthermore, a large particle would more likely deposit on a conducting airway, frustrating the intent to produce a map of regional ventilation. Hence the infrequent generation of larger particles that occurs with clinical aerosol-generation devices can produce an artifact unacceptable for the goals of the studies described here. We have attempted to systematically assess the risk of generating of large particle aggregates within our aerosol. The microspheres were microscopically examined before administration to assure uniform diameters, and on-line measurements of aerosol diameter and concentration were taken during administration of the FMS labels to the animals. To ascertain that large FMS aggregates were not contained in the aerosol administered to the animals, we examined electron-microscopic images of aerosol collected onto substrates. Finally, we have surveyed thick microscopic sections of dried lung samples from the experimental animals to detect any appreciable deposition of aerosol on conducting airways and found none except at airway bifurcations. Deposition of 1.0-µm FMS at airway bifurcations was also found by Pinkerton et al. (17) in their studies of aerosol deposition in rat lung. On the basis of our estimates of the relative proportion of conducting airway in lung parenchymal pieces, those pieces with airway scores of 4 and 5 had ~50 and 25% of their total volume as alveolar parenchyma and had ~40% of the fluorescence intensity of pieces without large conducting airways. Therefore, it appears unlikely that the extent of fluorescence deposition on the conducting airways would significantly influence our conclusion that, at the scale of resolution attained in this study, fluorescence intensity is related to alveolar ventilation.
The intensity of fluorescence deposited per piece is important to this technique. The background fluorescence from solvent in cuvettes containing no lung tissue ranged from 0.1 to 3.4% of the mean signal obtained from the pieces with FMS labels, a minor and easily corrected effect. However, the fluorescence signal obtained from unlabeled lung tissue can potentially cause significant artifact if the overall FMS signal is low. The additional background fluorescence originating from lung tissue samples is greatest on the blue side of the emission spectrum and decreases progressively with the longer wavelengths. The artifact arises when suspended particulate matter from the lung sample is present in the fluorescence cuvette, causing a fluorescence signal of variable intensity by scattering of the excitation signal from the fluorimeter. This particulate effect was detected by comparing readings of cleanly cut, unlabeled lung pieces with pieces that were deliberately crushed. The mean signal from cuvettes with very high particulate concentration was eight times greater than the mean signal from cleanly cut pieces in solvent, which was comparable to the signal from solvent alone. Although the high readings from crushed pieces decreased as the suspension was allowed to settle over several minutes, they failed to come within the range of the cleanly cut pieces. This signal artifact cannot necessarily be detected by comparing simultaneously administered fluorescencent labels, because the particulate effect proportionately increases the signal at all wavelengths in unlabeled pieces.
We resolved this artifact in those FMS emitting at shorter wavelengths by increasing the concentration of FMS and by modifications of the aerosol-delivery system. The final three animals had blue-green and yellow-green FMS signals 10-25 times greater than the magnitude of mean unlabeled tissue fluorescence. This artifact for orange and red FMS labels was much smaller; we found mean red and orange signals 10 times greater than unlabeled tissue in the first three pigs, and >20 times greater than unlabeled tissue in the last two animals. We conclude that appropriate care in cutting and handling of dried lung pieces, the administration of high aerosol FMS loads, and the use of longer wavelength labels can minimize the scattering artifact to 5% or less of the mean fluorescence signal.
We have no explanation for those few high-signal pieces that we excluded from this analysis. Although the fluorescence measurement for those high-signal pieces was reproducible, the lower (presumably normal) signal of the pair always matched the other pair of ventilation measurements. Part of the explanation may be related to the particulate scattering effect described above, which is much more marked with shorter wavelength colors. Another contributing possibility would be the presence of a rare large particle within the aerosol. As noted above, a single large particle has the potential to distort an estimate of regional ventilation when a mean deposition of 2,000-10,000 1.0-µm FMS is present in each lung piece. Whereas larger spheres would most likely deposit in the apparatus or upper airways, some could reach peripheral lung regions. Direct microscopic examination of many fields of our FMS mixture before aerosolization did reveal rare 15- to 30-µm spheres, but passing this mixture through a 2-µm filter before administration to the final two animals did not eliminate the rare aberrant signals. In any event, the simultaneous administration of two FMS labels permits detection of this artifact.
The density of FMS deposition in lung parenchyma was influenced by 1) aerosol density delivered by the apparatus; 2) loss of aerosol in the ventilator system before delivery to the animal; 3) duration of administration; and 4) the selected ventilatory parameters. In preliminary experiments, we noted that a 1-s inspiratory hold increased total deposition to 50%, but inspiratory hold was not used in the experiments presented here. Valberg et al. (21) have observed that a high respiratory rate (65-90 breaths/min) will alter distribution of the aerosol administered to an isolated lung, with proportional increases in deposition in the conducting airways compared with parenchyma. Thus the relatively large tidal volumes we chose for our animals permitted slower respiratory rates and allowed more time for gravitationally mediated deposition of the FMS in lung parenchyma. Aerosol density at the generator and administration time remained constant throughout all studies. For the last three animals reported here, the substitution of conductive tubing for rubber tubing in the delivery system and change in the ventilator configuration doubled the concentration of aerosol delivered. The final configuration of the generator-ventilator system with the selected ventilatory parameters and labeling times produced a mean fluorescence signal 10-25 times greater than the fluorescence signal from unlabeled lung. Because the protocol for these experiments involved administering two aerosol colors, the concentration of each individual color was reduced. If the yellow-green and blue-green labels had been administered undiluted by other colors, they would have provided satisfactory signal levels even with the original ventilator system.
The weighted regression used to investigate spatial ventilation and blood flow gradients is a refinement of our previous approach to the description of blood flow heterogeneity (11). Our present approach is based on the rationale that the most appropriate representation of ventilation and perfusion heterogeneity reflects the contribution of each piece to the total measured ventilation or blood flow. The simple analytic approach employed earlier assigned equal statistical weight to each piece. Because of our rectilinear cutting protocol, the volume of most pieces containing lung surface is smaller than the central pieces, but the choice to normalize both ventilation and perfusion by piece weight assigned those surface pieces greater statistical weight than was merited. The present analytic refinement decreases the significance of both gravitational and radial gradients, but these effects in the prone position are already so small that they are not useful to predict ventilation or blood flow distribution.
Although these results are promising, additional validation remains to be performed. The concept of alveolar ventilation arose from gas-exchange models and is difficult to apply to studies of regional function, as regional alveolar ventilation does not have an accepted "gold standard" technique, such as injected microspheres for regional perfusion. The most important limitation of aerosol deposition as an estimate of regional ventilation is related to the scale of the measurement taken. At the scale used in this study (1.9-cm3 per piece), ventilation to that volume is dependent primarily on convective gas movement, and thus 1.0-µm aerosol deposition should appropriately represent regional ventilation. However at a microscopic scale, the model studies of both Tsuda et al. (20) and Darquenne and Paiva (5, 6) have demonstrated that the deposition of 1.0-µm aerosol is dependent on gravitational settling and inertial impaction, two mechanisms that are irrelevant to alveolar ventilation. At the microscopic level of scale, 1.0-µm aerosol deposition measurements cannot be interpreted as representing alveolar ventilation. Examination of the histological samples (Fig. 4) confirms that the FMS deposition at the alveolar level is highly irregular, with suggestion of deposition around alveolar ducts and deposition in a gravitational direction. Systematic histological studies of deposition patterns will be required to quantitatively assess the fractional distribution of FMS on conducting airways and other deposition patterns. The FMS-derived regional ventilation distributions should also be compared with in vivo radionuclide-imaging techniques to ascertain agreement at least at a segmental and subsegmental level of resolution. Finally, studies need to be performed to compare regional ventilation distribution patterns observed with the large tidal volumes used here with more physiological breathing patterns.
The findings presented here indicate a high degree of ventilation heterogeneity at the scale of measurement attained, comparable in extent to the measured perfusion heterogeneity and to the extent of perfusion heterogeneity described in previous studies of pulmonary blood flow taken at this level of resolution (2, 4, 11, 16). Because ventilation and perfusion were measured simultaneously in this study, the high degree of correlation predicted for ventilation with regional blood flow was directly demonstrated. In pigs, it appears that at the measurement resolution attained in these studies, ventilation heterogeneity is comparable to perfusion heterogeneity. This may not be true for dogs or other animals that have a high degree of collateral ventilation within the lung. The finding of physiologically insignificant radial and gravitational gradients of both ventilation and perfusion in prone pigs parallels previous measurements of regional blood flow in dog lungs made at this same scale (11) in which regional effects of radial gradients explained <11% and gravitational gradients explained <3% of the total spatial heterogeneity of blood flow. The most striking observation of this study is that regional perfusion is the strongest statistical predictor of regional ventilation. Whereas the normal gas exchange manifested by these animals mandates a high correlation between the two variables, our findings emphasize the close spatial matching of ventilation to perfusion, regardless of whether this match arises out of anatomic relations between airway and vascular structure and/or the tightly coordinated regulation of vascular and bronchiolar tone.
ACKNOWLEDGEMENTS
The authors acknowledge the excellent assistance of D. An, Dr. S. Bernard, and P. Campbell in completion of these studies and thank Molecular Probes for donation of the fluorescent microspheres.
FOOTNOTES
Address for reprint requests: H. T. Robertson, Div. of Pulmonary and Critical Care Medicine, Box 356522, Univ. of Washington, Seattle, WA 98195-6522 (E-mail: tomrobt{at}u.washington.edu).
Received 30 April 1996; accepted in final form 28 October 1996.
REFERENCES
| 1. | Amis, T. C., A. B. H. Crawford, A. Davison, and L. A. Engel. Distribution of inhaled 99mtechnetium-labelled ultrafine carbon particle aerosol (Technegas) in human lungs. Eur. Respir. J. 3: 679-685, 1990. [Abstract] |
| 2. |
Beck, K. C.,
and
K. Rehder.
Differences in regional vascular conductances in isolated dog lungs.
J. Appl. Physiol.
61:
530-538,
1986.
|
| 3. | Brain, J. D., and P. A. Valberg. Deposition of aerosol in the respiratory tract. Am. Rev. Respir. Dis. 120: 1325-1373, 1979. [Medline] |
| 4. |
Caruthers, S. D.,
and
T. R. Harris.
Effects of pulmonary blood flow on the fractal nature of flow heterogeneity in sheep lungs.
J. Appl. Physiol.
77:
1474-1479,
1994.
|
| 5. |
Darquenne, C.,
and
M. Paiva.
One-dimensional simulation of aerosol transport and deposition in the human lung.
J. Appl. Physiol.
77:
2889-2898,
1994.
|
| 6. |
Darquenne, C.,
and
M. Paiva.
Two- and three-dimensional simulations of aerosol transport and deposition in alveolar zone of human lung.
J. Appl. Physiol.
80:
1401-1414,
1996.
|
| 7. | Draper, N., and H. Smith. Applied Regression Analysis (2nd ed.). New York: Wiley, 1981. (Wiley Ser. in Probability and Mathematical Statistics) |
| 8. | FMRC. FMS-Manual. Manual for Using Fluorescent Microspheres to Measure Regional Organ Perfusion. (WWW:http://fmrc.pulmcc.washington.edu/fmrc/fmrc.html), 1996. |
| 9. |
Glenny, R. G.,
N. L. Polissar,
S. McKinney,
and
H. T. Robertson.
Temporal heterogeneity of regional pulmonary perfusion is spatially clustered.
J. Appl. Physiol.
79:
986-1001,
1995.
|
| 10. |
Glenny, R. W.,
S. Bernard,
and
M. Brinkley.
Validation of fluorescent-labeled microspheres for measurement of regional organ perfusion.
J. Appl. Physiol.
74:
2585-2597,
1993.
|
| 11. |
Glenny, R. W.,
W. J. E. Lamm,
R. K. Albert,
and
H. T. Robertson.
Gravity is a minor determinant of pulmonary blood flow distribution.
J. Appl. Physiol.
71:
620-629,
1991.
|
| 12. |
Glenny, R. W.,
and
H. T. Robertson.
Fractal modeling of pulmonary blood flow heterogeneity.
J. Appl. Physiol.
70:
1024-1030,
1991.
|
| 13. |
Glenny, R. W.,
and
H. T. Robertson.
Fractal properties of pulmonary blood flow: characterization of spatial heterogeneity.
J. Appl. Physiol.
69:
532-545,
1990.
|
| 14. |
Hubmayr, R. D.,
B. J. Walters,
P. A. Chevalier,
J. R. Rodarte,
and
L. E. Olson.
Topographical distribution of regional lung volume in anesthetized dogs.
J. Appl. Physiol.
54:
1048-1056,
1983.
|
| 15. |
Nicolaysen, G.,
C. Müller,
and
J. Kramer-Johanson.
Regional /-ratio in goats shows large non-gravitational heterogeneity and increases down the lung (Abstract).
FASEB J.
6:
A1477,
1992.
|
| 16. |
Parker, J. C.,
J. L. Ardell,
C. R. Hamm,
S. A. Barman,
and
P. J. Coker.
Regional pulmonary blood flow during rest, tilt, and exercise in unanesthetized dogs.
J. Appl. Physiol.
78:
838-846,
1995.
|
| 17. | Pinkerton, K. E., J. T. Gallen, R. R. Mercer, V. C. Wong, C. G. Plopper, and B. K. Tarkington. Aerosolized fluorescent microspheres detected in the lung using confocal scanning laser microscopy. Microsc. Res. Tech. 38: 437-443, 1993. |
| 18. |
Rodarte, J. R.,
M. Chaniotakis,
and
T. A. Wilson.
Variability of parenchymal expansion measured by computed tomography.
J. Appl. Physiol.
67:
226-231,
1989.
|
| 19. | Schummer, A., R. Nickel, and W. O. Sack. The Viscera of the Domestic Mammals. New York: Springer-Verlag, 1979. |
| 20. |
Tsuda, A.,
J. P. Butler,
and
J. J. Fredberg.
Effects of alveolated duct structure on aerosol kinetics. II. Gravitational sedimentation and inertial impaction.
J. Appl. Physiol.
76:
2510-2516,
1994.
|
| 21. |
Valberg, P. A.,
J. D. Brain,
S. L. Sneddon,
and
S. R. LeMott.
Breathing patterns influence aerosol deposition sites in excised dog lungs.
J. Appl. Physiol.
53:
824-837,
1982.
|
| 22. |
Wagner, P. D.,
P. F. Naumann,
and
R. B. Laravuso.
Simultaneous measurement of eight foreign gases in blood by gas chromatography.
J. Appl. Physiol.
36:
600-605,
1974.
|
| 23. |
Wagner, P. D.,
H. A. Saltzman,
and
J. B. West.
Measurement of continuous distributions of ventilation-perfusion ratios: theory.
J. Appl. Physiol.
36:
588-599,
1974.
|
| 24. | West, J. B., and P. D. Wagner. Pulmonary gas exchange. In: Bioengineering Aspects of the Lung, edited by J. B. West. New York: Dekker, 1977, vol. 3, p. 361-457. |
| 25. |
Wilson, T. A.,
and
K. C. Beck.
Contributions of ventilation and perfusion inhomogeneities to the A/ distribution.
J. Appl. Physiol.
72:
2298-2304,
1992.
|
| 26. |
Zeltner, T. B.,
T. D. Sweeney,
W. A. Skornik,
H. A. Feldman,
and
J. D. Brain.
Retention and clearance of 0.9-µm particles inhaled by hamsters during rest or exercise.
J. Appl. Physiol.
70:
1137-1145,
1991.
|
This article has been cited by other articles:
|
|
 |
|
  A. Kozian, T. Schilling, F. Freden, E. Maripuu, C. Rocken, C. Strang, T. Hachenberg, and G. Hedenstierna One-lung ventilation induces hyperperfusion and alveolar damage in the ventilated lung: an experimental study Br. J. Anaesth., April 1, 2008; 100(4): 549 - 559. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. T. Robertson, B. Neradilek, N. L. Polissar, and R. W. Glenny Sporadic coordinated shifts of regional ventilation and perfusion in juvenile pigs with normal gas exchange J. Physiol., September 1, 2007; 583(2): 743 - 752. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Chon, K. C. Beck, B. A. Simon, H. Shikata, O. I. Saba, and E. A. Hoffman Effect of low-xenon and krypton supplementation on signal/noise of regional CT-based ventilation measurements J Appl Physiol, April 1, 2007; 102(4): 1535 - 1544. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Galvin, G. B. Drummond, and M. Nirmalan Distribution of blood flow and ventilation in the lung: gravity is not the only factor Br. J. Anaesth., April 1, 2007; 98(4): 420 - 428. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. T. Robertson and M. P. Hlastala Microsphere maps of regional blood flow and regional ventilation J Appl Physiol, March 1, 2007; 102(3): 1265 - 1272. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Y. C. Tsang, W. J. E. Lamm, B. Neradilek, N. L. Polissar, and M. P. Hlastala Endothelin receptor blockade does not improve hypoxemia following acute pulmonary thromboembolism J Appl Physiol, February 1, 2007; 102(2): 762 - 771. [Abstract] |
