Vol. 93, Issue 1, 51-57, July 2002
Detection of porcine oleic acid-induced acute lung injury
using pulmonary acoustics
Jukka
Räsänen1 and
Noam
Gavriely2
1 Department of Anesthesiology, Mayo Clinic,
Rochester, Minnesota 55905; and 2 Rappaport Faculty
of Medicine, Technion-Israel Institute of Technology, Haifa 32000, Israel
 |
ABSTRACT |
To evaluate the utility of
monitoring the sound-filtering characteristics of the respiratory
system in the assessment of acute lung injury (ALI), we injected a
multifrequency broadband sound signal into the airway of five
anesthetized, intubated pigs, while recording transmitted sound over
the trachea and on the chest wall. Oleic acid injections
effected a severe lung injury predominantly in the dependent lung
regions, increasing venous admixture from 6 ± 1 to 54 ± 8%
(P < 0.05) and reducing dynamic respiratory system compliance from 19 ± 0 to 12 ± 2 ml/cmH2O
(P < 0.05). A two- to fivefold increase in sound
transfer function amplitude was seen in the dependent
(P < 0.05) and lateral (P < 0.05)
lung regions; no change occurred in the nondependent areas. High
within-subject correlations were found between the changes in dependent
lung sound transmission and venous admixture (r = 0.82 ± 0.07; range 0.74-0.90) and dynamic compliance
(r = 
0.87 ± 0.05; 0.80 to 0.93). Our
results indicate that the acoustic changes associated with oleic
acid-induced lung injury allow monitoring of its severity and distribution.
respiratory sounds; transmitted sounds
| |
INTRODUCTION |
AUSCULTATION OF BREATH
SOUNDS has been used in medicine for almost two centuries
(8). The stethoscope remains a useful tool in
evaluating patients with acute and chronic lung disease, and it is the
only technique that will give combined functional and localizing
information about lung pathology at the bedside. Modern tools for
monitoring gas exchange, lung compliance, and airway resistance in a
real-time, continuous fashion are available, but they provide
information about the respiratory system as a whole, not which region
of the lung or which lung is most affected. Whereas the lung parenchyma
can be imaged by using radiographs, computerized tomography, or
magnetic resonance, these techniques cannot be used for continuous or
frequent assessment, and they often require moving the patient to a
specific location. Ultrasound, a common bedside imaging tool for a
variety of organ systems, does not penetrate the gas-containing lung
parenchyma and hence is of limited use in pulmonary disease.
Studies performed in healthy humans and in patients with chronic lung
disease indicate that sound energy introduced into the airway can be
detected on the surface of the chest and that the transfer of such
energy can be quantified in terms of amplitude and wave speed (5,
6, 15). An experimental study by Donnerberg et al.
(3) showed that the amplitude of sound transmission through the respiratory system is greater in congested lungs than in
normal lungs and that this change is proportional to the lung wet-to-dry weight ratio.
In an attempt to further characterize the acoustic properties of the
respiratory system during the development of acute lung injury, we
designed this study to determine whether oleic acid-induced permeability-type pulmonary injury would enhance sound transfer through
the respiratory system in proportion to the severity of injury and
whether or not the gravity-dependent distribution of lung pathology
thus could be detected with this method.
| |
METHODS |
Experimental setup.
After approval by the Institutional Animal Care and Use Committee, six
healthy pigs, weighing 35-45 kg and cared for according to the
current guidelines for the care and use of laboratory animals, were
included in the study. Anesthesia was induced with 2 mg/kg intramuscular tiletamine, 2 mg/kg zolazepam, and 2 mg/kg xylazine and
deepened with 3% halothane before tracheal intubation with a
7.5-mm-internal-diameter tube. The animals were mechanically ventilated
(Ohio Medical Systems, Madison, WI) with 0.5-1.0% isoflurane in
air, by using a tidal volume of 8-10 ml/kg and a ventilator rate
sufficient to maintain arterial blood carbon dioxide tension within 35 to 45 Torr.
Tidal volume (NVM-1, Bear Medical Systems, Riverside, CA), airway
pressure (Pneumogard 1200, Novametrix, Wallingford, CT), and end-tidal
carbon dioxide concentration (Nellcor, Hayward, CA) were monitored at
the proximal end of the tracheal tube. A pulse oximeter probe (Nellcor)
was attached to the animal's upper extremity for monitoring heart rate
and oxyhemoglobin saturation (SpO2). A femoral
artery was cannulated for direct measurement of blood pressure and for
sampling of arterial blood. A pulmonary arterial catheter was inserted
via the right jugular vein for monitoring of pulmonary arterial and
pulmonary capillary occlusion pressures, mixed venous oxyhemoglobin
saturation, cardiac output, and core temperature and for sampling of
mixed venous blood (Oximetrix 3, Abbott, Abbott Park, IL). Before
induction of lung injury, the animals received 1,000 ml of Ringer
lactate in anticipation of fluid shifts caused by oleic acid.
Subsequently, Ringer lactate was infused to maintain pulmonary
arterial occlusion pressure between 9 and 14 mmHg; core temperature was
maintained with a heating blanket.
Sound generation and recording.
Broad-band noise was generated digitally (22,050 Hz, 16 bit, Cool Edit
2000, Syntrillium, Phoenix, AZ) and power amplified (40-W root mean
square, 70-20,000 Hz, MPA-50, Optimus, Fort Worth, TX) to drive a
40-W, 5.25-in. speaker (catalog #40-1031, Radio Shack, Fort Worth,
TX). The speaker was housed in a custom-made acoustically
insulated case that attenuated the generated sound by >40 dB (Fig.
1). The 4-cm frontal opening of the case
was tapered to fit a three-way stopcock that connected the sound source
to the breathing circuit in such a way that the sound pathway was straight and the breathing circuit attached to the T piece at a 90°
angle. The three-way valve allowed the loudspeaker to be excluded from
the circuit while the animal was being mechanically ventilated between
measurements.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 1.
Frequency and power plots of recordings made from inside
and outside the breathing circuit, illustrating the attenuating effect
of the loudspeaker enclosure. The recordings were made at the proximal
end of the tracheal tube and at a distance corresponding to chest
sensor location, respectively.
|
|
The transmitted sound was picked up by an electret microphone
(70-16,000 Hz, catalog #33-3013, Radio Shack) connected to
the speaker outlet, and by seven PPG sensors (Linear 50-2,000 Hz, response range 10-5,000 Hz, PPG-02, Technion, Haifa, Israel), one
placed over the neck and six on the chest wall. The chest sensors were
placed bilaterally over the dependent lung 5 cm laterally from the
spine, on the nondependent lung 5 cm from the sternum, and on the
lateral midlung, halfway between the other two locations. The sensors
were secured to the shaved chest with a circumferential elastic strip.
The signal from each sensor was amplified (blue tube, Presonus, Baton
Rouge, LA) and digitized at 10,000 samples/s into a portable computer
by using a 16-channel, 16-bit analog-to-digital converter board
(PCI-6035E, National Instruments, Austin, TX) inserted into a PCI
expansion system (CB1F, Magma, San Diego, CA). The data acquisition and
the determination of the transfer function magnitude, coherence, and
phase occurred on-line (Trans19 Software, Technion). Each record was
5 s long; the software used multiple (n = 194)
512-point segments with 50% overlap to calculate the average transfer
function for each input-output pathway. The magnitude, coherence, and
phase of the transfer function were displayed on-line and saved as
ASCII files. In addition, the raw signals from all of the sound
channels were recorded for off-line analysis.
The integrated hardware-software system was tested and validated by
verifying that all of the sensors had identical frequency responses
under a load similar to that present during the experiments. Feeding
known input and output files with predetermined transfer characteristics between them was used to check the software.
Experimental protocol.
After instrumentation, a stabilization period of 30 min was allowed
before baseline measurements. To restore lung history, three manual
inflations with a peak airway pressure of 45 cmH2O were
given at 10-min intervals throughout the study. Systemic arterial,
pulmonary arterial, and pulmonary capillary occlusion pressures, heart
rate, and body temperature were recorded, and arterial and mixed venous
blood samples were drawn for measurement of blood gases (IL-482,
Instrumentation Laboratories, Lexington, MA) and oxyhemoglobin
saturation (IL-1620, Instrumentation Laboratories). Cardiac output was
measured in triplicate by thermodilution by using 10 ml of lactated
Ringer solution injected at random moments during the respiratory
cycle. End-expiratory and peak inspiratory airway pressures and exhaled
tidal volume were recorded off of the displays of the respective
monitoring devices. The sound measurements were then made during a 5-s
period of apnea induced by switching off the ventilator.
Baseline measurements were made four times at 20-min intervals.
Thereafter, acute lung injury was induced gradually by repeated injections of oleic acid 0.1 ml into the right atrial port of the
pulmonary arterial catheter at 5- to 10-min intervals. Oleic acid
injections were continued until the arterial oxygen tension-to-inspired oxygen fraction ratio fell consistently <250. Oxygen was added to the
fresh gas flow to maintain SpO2 
90%.
Acoustic and hemodynamic measurements were repeated at 20-min intervals
throughout the development of lung injury. The sequence and timing of
data collection, oleic acid injections, and lung inflations were
maintained unchanged. To verify sensor placement and the presence and
distribution of lung injury, posterior-anterior and lateral chest
radiographs were obtained from one animal before and after lung injury.
After final measurements at stable lung injury, the animals were
killed with a large intravenous dose of pentobarbital. Thereafter,
the lungs were removed, examined, weighed, and inflated to a constant airway pressure of 40 cmH2O for 48 h, after which they
were reweighed.
Data analysis.
Venous admixture was calculated from arterial and mixed venous blood
gas and saturation results with the use of standard formula. Dynamic
compliance was calculated by dividing the exhaled tidal volume by the
airway pressure amplitude.
The frequency distribution of the sound transfer function, amplitude,
and coherence were plotted for each recording. All amplitude plots for
a given sensor location were superimposed, and the plots were examined
for peak amplitude. Amplitude peaks <100 Hz were not considered
because of the interference from heart sounds and because low-frequency
sounds are not likely to travel down to small airways (5,
14). The frequency at which the amplitude was highest and the
coherence of the input and output signals was >0.4 was selected, and
that amplitude is reported as the peak amplitude for a given sensor
location. To eliminate the effect of changing amplification of the
recorded signals between sensors and animals, we also calculated the
fractional change in amplitude relative to the average baseline value
for measurements made during lung injury.
The results are presented as means ± SD. The statistical
significance of the observed changes was evaluated by using Friedman's nonparametric analysis of variance, because deviation from normal distribution and inequality of variances could not be reliably ruled
out (12). The strength of association between changes in
sound transmission amplitude and physiological variables reflecting lung injury was tested with Pearson's linear regression analysis. Individual correlation coefficients, calculated separately for each
subject, are reported, along with overall correlations based on
population averages at each data collection point. These calculations were performed separately for anterior, lateral, and posterior lung
regions averaging data from the corresponding regions of both lungs.
Results were considered statistically significant if the probability of
type 
-error was <5%.
| |
RESULTS |
Physiological measurements.
One of the six animals developed acute pulmonary hypertension and died
early during induction of lung injury. This reduced the number of
animals included in the data analysis to five. Changes in variables
reflecting cardiopulmonary function before and during induction of lung
injury are shown in Table 1. The
hemodynamic response to oleic acid administration was characterized by
a slight increase in heart rate and blood pressure, a dramatic increase in pulmonary arterial pressure, and a 23% fall in cardiac output. Despite generous volume resuscitation, the blood hemoglobin
concentration increased, likely reflecting fluid flux into the
extravascular space of the lungs. Adjustments in ventilator settings
and inspired oxygen fraction maintained ventilation and oxygenation
within target ranges. The anesthesia breathing circuit effected a 4- to
5-cmH2O positive expiratory pressure at the proximal end of the tracheal tube throughout the study, even though external positive end-expiratory pressure was not used. The oleic acid injections resulted in a 48% increase in peak airway pressure, a 39% fall in
respiratory system compliance, and an almost 10-fold increase in venous
admixture. Examination of the lungs at the end of the study revealed in
each animal a severe lung injury with froth in the large airways and
the dependent one-third of the lung parenchyma consolidated or
atelectatic. The isolated lungs were completely expandable with a
pressure of 40 cmH2O; the wet-to-dry lung weight ratio was
4.2 ± 1.2.
View this table:
[in this window]
[in a new window]
|
Table 1.
Changes in variables reflecting cardiopulmonary function in 5 pigs
during induction of acute lung injury with oleic acid
|
|
Sound measurements.
Whereas the characteristics of the signal recorded at the outlet of the
loudspeaker box were similar to the original white noise, considerable
attenuation occurred in the circuit connectors and the tracheal tube. A
sensor placed outside the exposed trachea at the tip of the tracheal
tube recorded two distinct peaks of sound transfer at frequency bands
from 300 to 800 Hz and from 1,400 to 2,500 Hz (Fig.
2). This pattern of sound transmission could be seen consistently when recordings were made in intact animals,
regardless of the position of the sensor. A frequency range from 1,500 to 2,500 Hz contained, without exception, the highest amplitudes and
the largest changes during developing lung injury (Fig.
3). Therefore, in addition to peak
amplitude values, we also report amplitudes averaged over a frequency
band between 1,500 and 2,500 Hz. The coherence averaged over this
frequency band was 0.81 ± 0.04 in the dependent areas of the lung
during injury.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 2.
Frequency distribution of the transfer function amplitude
recorded from outside the exposed trachea at the tip of the tracheal
tube. The system allows transmission at 2 major frequency bands: from
300 to 800 Hz and from 1,400 to 2,500 Hz.
|
|
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 3.
Frequency distribution of the transfer function amplitude recorded
from the dependent region of the right lung in a representative animal
during development of oleic acid-induced lung injury. Lines
representing recordings made at different phases of the injury are
superimposed. The highest amplitudes and the largest increase in
amplitude are seen at a frequency band from 1,500 to 2,500 Hz. The
times are minutes after start of oleic acid injections.
|
|
During the baseline period, the transfer function amplitude averaged
across all sensors was 1.04 (median 0.38, range 0.0014-11.01); the
within-subject standard deviations of baseline measurements averaged
0.21 (median 0.11, range 0.0015-0.97). No statistically significant changes were found in sound transfer function amplitude of
any sensor during the baseline measurement period. The baseline amplitude values for each sensor were averaged for comparison with
recordings made during lung injury.
The average peak absolute transmission amplitudes for each sensor are
shown in Table 2; the average mean
amplitudes calculated for a frequency band from 1,500 to 2,500 Hz are
depicted in Fig. 4 in reference to the
average baseline values. During the development of lung injury, both
the peak and the mean sound transmission amplitudes increased
statistically significantly in the sensors overlying the posterior and
lateral lung fields, whereas no significant change was observed in the
sensors attached to the anterior part of the chest. The average change
from baseline was more than fivefold in the posterior-dependent regions
of both lungs, compared with the two- to threefold slower increase
recorded from the lateral lung fields (Fig. 4). The sound transmission
amplitudes in the anterior, nondependent lung regions stayed unchanged
at baseline level throughout the study.
View this table:
[in this window]
[in a new window]
|
Table 2.
Absolute amplitude of the lung sound transfer function recorded from 6 sensors overlying 3 lung regions bilaterally, before and after
development of oleic acid-induced lung injury
|
|
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 4.
Mean amplitude of the lung sound transfer function
relative to baseline level over a frequency band from 1,500 to 2,500 Hz, recorded with 6 sensors overlying 3 lung regions bilaterally,
during development of oleic acid-induced lung injury. Data points are
average values for all 5 animals. BL, baseline; R, right; L, left. Each
unit on the vertical axis denotes an amplitude twice the baseline
average. Values are means ± SD. For clarity, SD bars are only
shown for 1 region. * P < 0.05.
|
|
High within-subject correlations were observed between sound
transmission amplitude of the posterior (dependent) lung regions and
the concurrent increase in venous admixture (r = 0.82 ± 0.07; range 0.74-0.90) and the decrease in dynamic
respiratory system compliance (r = 0.87 ± 0.05;
range 0.80 to 0.93). Linear regression analysis based on population
averages at each data collection point likewise yielded high
correlations for venous admixture (r = 0.98) and
dynamic compliance (r = 0.98) (Fig.
5). The corresponding overall
correlations were lower but statistically significant for the lateral
lung regions (r = 0.91 and 0.94, respectively, P < 0.05) and not statistically significant for the
anterior areas (r = 0.56 and 0.58, respectively).
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 5.
Association among average peak sound transmission
amplitude of the posterior lung fields, venous admixture
( ), and dynamic respiratory system compliance
( ) during development of oleic acid-induced lung
injury. Data points are averages of all 5 animals at various phases of
the study.
|
|
| |
DISCUSSION |
The results of this study show that the development of acute oleic
acid-induced lung injury alters the sound-filtering characteristics of
the porcine respiratory system and that these changes contain information that can be used to evaluate the extent and distribution of
the injury.
Considering the high incidence of pulmonary disorders in critically ill
patients and the widespread use of the stethoscope, relatively little
is known about changes in pulmonary acoustics in patients with acute
lung injury (8). Studies in healthy humans have
established that the mouth-to-chest transit time of an acoustic wave is
1.5-5 ms, corresponding to a velocity of 60-80 m/s,
considerably slower than the speed of sound in air, and that the
amplitude of sound transfer decreases gradually with increasing frequency (6). Although the exact pathway of sound
propagation through the bronchial tree and lung parenchyma is not
known, low frequencies are believed to exit the central airways and
travel along tissue paths, whereas higher frequencies are transmitted to more peripheral airways (5, 14).
In the past, clinicians have used auscultation of sounds vocalized by
the patient in an attempt to detect regional lung volume loss. Baughman
and Loudon (2) studied frequency spectra of the vocalized
"e" sound transmitted through the chest and documented enhanced
transmission of frequencies from 400 to 1,000 Hz through consolidated
lung. Our results are in agreement with this finding.
Only two experimental studies have investigated changes in pulmonary
acoustics during acute conditions affecting the lung. Donnerberg et al.
(3) measured sound transmission index, defined as the
logarithm of the output-to-input power ratio of sound introduced into
the canine airway, while effecting pulmonary vascular congestion using
a left atrial balloon. They found a direct correlation between the
sound transmission index and the degree of pulmonary congestion. Ploysongsang et al. (11) recorded the power spectra of
natural lung sounds in five dogs, two with interstitial and three with alveolar hydrostatic pulmonary edema from overhydration. They noted a
right shift of the power spectrum, which correlated well with the
degree of hypervolemia and impairment in pulmonary mechanics and gas
exchange. These results show that, at least in hydrostatic pulmonary
edema, acoustic changes in the respiratory system could be used to
quantitate the degree of pulmonary dysfunction. No attempt to localize
the injury was made in these investigations.
The porcine oleic acid model is generally accepted as one resembling
human acute permeability-type lung injury. The oleic acid
administration protocol in this study produced a severe acute lung
injury confirmed with measurements of gas exchange and lung mechanics,
radiography, and postmortem examination. The characteristics of this
injury, including hemodynamic changes and mortality, were comparable to
results previously published by others (7). We also
confirmed a gravity-dependent inhomogeneity of the injury: severe
volume loss in the dependent parts of the lungs, nearly normal-appearing lung in the nondependent areas, and a gradual transition between these extremes, depending on vertical location. The
observed deterioration of gas exchange and lung mechanics appeared to
be caused primarily by alveolar collapse, rather than fluid
extravasation, because the lungs were fully expandable after removal
and their wet-to-dry weight ratio had increased relatively little.
It was apparent from recordings made over the exposed trachea at the
tip of the tracheal tube that considerable filtering of sound generated
by the loudspeaker occurred in the artificial airway (Fig. 2). Whereas
this fact narrows the useful frequency range considerably and may
result in the loss of some information, it did have two favorable
consequences. First, the fact that the frequency pattern detected by
the chest wall sensors was nearly identical to the one measured at the
trachea (Figs. 2 and 3) indicates that the recorded signal was not
contaminated by direct airborne sound transfer between the source and
the sensor. Second, because the 1,500- to 2,500-Hz frequency band that
passed through the breathing circuit was highly responsive to the lung
injury, the sound generation can be focused closer to that band. This
allows more compact sound generation equipment to be used in future
studies. The extent to which variations in the breathing circuit design influence the filtering of generated sound requires further study.
Our experimental setup allowed the output of each sensor to be
amplified independently. At the beginning of each experiment, the gain
of each channel was adjusted to provide a recordable signal; these
amplifier settings were then used for the remainder of the experiment.
Because gain settings were different for each channel in the same
subject and for the same channel in different subjects, the absolute
values of transfer function amplitude between sensors or between
subjects have little meaning in this study. Specifically, the high
absolute average amplitude over the right nondependent lung (Fig.
6) results from such differences in gain settings and does not reflect true lateralization of transmitted sound,
which has been observed in normal humans (9, 13). In
contrast, the magnitude of change in amplitude is comparable both
between sensors and between subjects and represents a real change in
sound transmission characteristics.
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 6.
Peak amplitude of the lung sound transfer function
recorded with 6 sensors overlying 3 lung regions bilaterally, during
development of oleic acid-induced lung injury. Data points are average
values for all 5 animals. * P < 0.05.
|
|
Within-subject baseline variability in both mean and peak sound
transmission amplitude was small compared with the two- to fivefold
average increase in the amplitude during lung injury in the posterior
and lateral lung fields (Fig. 6). Increase in the sound
transmission peak amplitude was evident in the first measurements 20 min after the first oleic acid dose in the sensors overlying the
dependent parts of the lungs. By this time, venous admixture and
respiratory system compliance also showed evidence of lung injury. The
design of the present study does not allow us to evaluate more
precisely the time course of changes in sound transmission, lung
mechanics, and gas exchange. The correlation of sound transmission
amplitude with venous admixture and respiratory system compliance was
remarkably high both within subjects and in the study population as a
whole. The closely associated changes in sound transfer function
amplitude, gas exchange, and lung mechanics may reflect derecruitment
of an increasingly large volume of lung tissue as the injury develops.
This is also seen in the spreading of the sound transmission changes
from sensors overlying the dependent lung to those located laterally
over less dependent areas.
Sound conduction through the respiratory system passes sequentially
through the airways, lung parenchyma, pleural space, and the chest
wall. Pulmonary edema can theoretically affect airway sound conduction
by blocking the airways with fluid, thereby diminishing sound
transmission. This phenomenon was not observed in the present study;
sound conduction actually increased as the injury became more severe.
Pulmonary pathology also may alter the density and air-to-tissue ratio
of the affected parenchyma. Diminution of the spongy characteristics of
the lung reduces the acoustic attenuation (insulation) in the lung,
thus enhancing sound transmission to the pleural surface. In addition,
the increased density of the lung improves the acoustic impedance
matching between the lung parenchyma and the chest wall across the
pleural space. The result of the latter mechanism is increased
transmission and reduced reflection at the interface between the lung
and the chest wall. Additional basic research is needed to determine
the contribution of the various mechanisms and to evaluate the
significance of the specific frequency band observed in the present study.
Studies in patients with acute lung injury over the last decade have
established the inhomogeneous nature of pathological changes in the
lung parenchyma and directed attention to interventions, such as
positioning, to alter their distribution (1, 4, 10).
Although continuous techniques to evaluate global lung mechanics and
gas exchange are routinely available for these patients, information
regarding the distribution of injury can only be obtained intermittently and with considerable cost. We have shown in this experimental study that the pathophysiological changes in oleic acid-induced acute lung injury alter the sound-filtering
characteristics of the respiratory system and that these changes can be
measured and monitored acoustically in a noninvasive manner. Acoustic
monitoring with multiple sensors provides localizing information of
lung injury that cannot be obtained with any other currently available bedside technique.
| |
FOOTNOTES |
Address for reprint requests and other correspondence: J. Räsänen, Dept. of Anesthesiology, Mayo Clinic, MB 2-752, 200 First St. SW, Rochester, MN 55905 (E-mail:
Rasanen.Jukka{at}mayo.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.
First published March 1, 2002;10.1152/japplphysiol.01238.2001
Received 17 December 2001; accepted in final form 25 February 2002.
| |
REFERENCES |
1.
Artigas, A,
Bernard GR,
Carlet J,
Dreyfuss D,
Gattinoni L,
Hudson L,
Lamy M,
Marini JJ,
Matthay MA,
Pinsky MR,
Spragg R,
and
Suter PM.
The American-European Consensus Conference on ARDS. II. Ventilatory, pharmacologic, supportive therapy, study design strategies and issues related to recovery and remodeling.
Intensive Care Med
24:
378-398,
1998[Web of Science][Medline].
2.
Baughman, R,
and
Loudon R.
Sound spectral analysis of voice-transmitted sound.
Am Rev Respir Dis
134:
167-169,
1986[Medline].
3.
Donnerberg, RL,
Druzgalski CK,
Hamlin RL,
Davis GL,
Campbell RM,
and
Rice DA.
Sound transfer function of the congested canine lung.
Br J Dis Chest
74:
23-31,
1980[Web of Science][Medline].
4.
Gattinoni, L,
Pelosi P,
Vitale G,
Pesenti A,
D'Andrea L,
and
Mascheroni D.
Body position changes redistribute lung computed-tomographic density in patients with acute respiratory failure.
Anesthesiology
74:
15-23,
1991[Web of Science][Medline].
5.
Kraman, SS,
and
Bohadana AB.
Transmission to the chest of sound introduced at the mouth.
J Appl Physiol
66:
278-281,
1989[Abstract/Free Full Text].
6.
Mahagnah, M,
and
Gavriely N.
Gas density does not affect pulmonary acoustic transmission in normal men.
J Appl Physiol
78:
928-937,
1995[Abstract/Free Full Text].
7.
Neumann, P,
and
Hedenstierna G.
Ventilation-perfusion distributions in different porcine lung injury models.
Acta Anaesthesiol Scand
45:
78-86,
2001[Web of Science][Medline].
8.
Pasterkamp, H,
Kraman SS,
and
Wodicka GR.
Respiratory sounds. Advances beyond the stethoscope.
Am J Respir Crit Care Med
156:
974-987,
1997[Free Full Text].
9.
Pasterkamp, H,
Patel S,
and
Wodicka GR.
Asymmetry of respiratory sounds and thoracic transmission.
Med Biol Eng Comput
35:
103-106,
1997[Web of Science][Medline].
10.
Pelosi, P,
D'Andrea L,
Vitale G,
Pesenti A,
and
Gattinoni L.
Vertical gradient of regional lung inflation in adult respiratory distress syndrome.
Am J Respir Crit Care Med
149:
8-13,
1994[Abstract].
11.
Ploysongsang, Y,
Michel RP,
Rossi A,
Zocchi L,
Milic-Emili J,
and
Staub NC.
Early detection of pulmonary congestion and edema in dogs by using lung sounds.
J Appl Physiol
66:
2061-2070,
1989[Abstract/Free Full Text].
12.
Sachs, L.
Applied Statistics. New York: Springer Verlag, 1982, p. 549-555.
13.
Wodicka, GR,
DeFrain PD,
and
Kraman SS.
Bilateral asymmetry of respiratory acoustic transmission.
Med Biol Eng Comput
32:
489-494,
1994[Web of Science][Medline].
14.
Wodicka, GR,
Stevens KN,
Golub HL,
Cravalho EG,
and
Shannon DC.
A model of acoustic transmission in the respiratory system.
IEEE Trans Biomed Eng
36:
925-934,
1989[Medline].
15.
Wodicka, GR,
Stevens KN,
Golub HL,
and
Shannon DC.
Spectral characteristics of sound transmission in the human respiratory system.
IEEE Trans Biomed Eng
37:
1130-1135,
1990[Web of Science][Medline].
J APPL PHYSIOL 93(1):51-57
8750-7587/02 $5.00
Copyright © 2002 the American Physiological Society
TOP
ABSTRACT
INTRODUCTION
