JOURNAL OF APPLIED PHYSIOLOGY Vol. 38, No. 5, May 1975. Printed
in U.S.A.
Measurement by electrical
of pleural
effusion
impedance
JOSEPH C. DENNISTON AND LEE E. BAKER Department of Physiology, Baylor College of Medicine, Texas Medical
detecting and localizing pleural effusion (1 I), it gained general clinical acceptance for this purpose. The present study was undertaken to assess, in trolled experimental setting, the electrical impedance nique for detecting, quantitating, and monitoring in pleural fluid volumes in the dog and to provide mental data for the interpretation of impedance measured in the clinical environment.
DENNISTON, JOSEPH C., AND LEE E. BAKER. Measurement of pleural effusion by electrical impedance. J. Appl. Physiol. 38(5) : 8514357. 1975.-Changes in the electrical impedance of the thorax were recorded from various electrode arrays on the thorax during the infusion and withdrawal of saline, plasma, or blood from the right, left, or both hemithoraxes. Impedance changes correlated linearly with the volume of infused fluid. The resistivity of the infused fluid significantlyaffected the sensitivity &?/ml) of the impedance method in detecting specific fluid accumulations. The use of hemithoracic electrode arrays permitted the lacalization of fluid accumulations to a specific hemithorax. Various factors that can alter the magnitude and interpretation of impedance changes are discussed, as they would affect the clinical interpretation of impedance changes, particularly in postoperative patients. thoracic noninvasive
electrical impedance; measurement
pleural
effusion,
Center, Houston, Texas 7702.5
MATERIALS
hemothorax;
THE USE OF ELECTRICAL impedance as a means of m .easuring physiological activity has been investigated for many years. Clinically the technique has been used as a noninvasive measure of stroke volume (3, 12, 14, 18), myocardial contractility (13, Zl), periph era1 vascular disease (5, 16, 22, 26), and tidal volume (2, 8). More recently the impedance technique has been used to measure intrathoracic fluids (15, 17, 23, 24). The need exists for a simple, noninvasive means of continuously monitoring shifts in thoracic fluid volumes of patients following trauma, pulmonary and myocardial infarction, and thoracic surgery. The clinical diagnosis of pleural effusion or hemothorax is based primarily on patient history, physical findings (auscultation and percussion), and roentgenographic evaluation of the thorax. Unfortunately these methods of detection are relatively insensitive. Neither percussion nor auscultation is quantitative in nature, and frequently neither detects effusion. The use of the lateral decubitus position, which allows radiographic demonstration of quantities of less than 100 ml of fluid to be visualized (19), has become a standard technique for determining small amounts of pleural effusion ; however, pleural adhesions may prevent layering of fluid, and the densities of soft tissue may mimic layering of free fluid lateral to the lung, resulting in an incorrect diagnosis (10). ultrasound , whi ch has gained wideAlthou gh reflected spread use in cardiology, has been used w ith success in
AND
has not a contechchanges experichanges
METHODS
Two studies, utilizing 43 healthy mongrel dogs weighing from 15 to 22 kg, were conducted. Preanesthetic medication consisted of atropine sulfate (0.02 mg/kg, im) and a solution containing 0.4 mg/ml of fentanyl and 20 mg/ml of droperidol (0.1 ml/kg) g iven iv. All animals were anesthetized with pentobarbital sodium (15-20 mg/kg, iv), and an oral endotracheal airway was established. Succinylcholine chloride (0.1-0.2 mg/kg, iv) was given as required to maintain short-term respiratory paralysis during pressurecycled controlled ventilation. Routine prestudy included total surface depilation of the neck, thorax, and abdomen to insure good electrical contact of the electrodes; placement of a standard 24 Fr thoracic drainage tube into each hemithorax through a l-in bilateral thoracotomy at the 10th intercostal space to permit the infusion and withdrawal of fluid (saline, plasma, or blood at 37°C) from the thorax; catheterization of the right femoral vein and artery to permit subsequent iv infusions and pressure recordings (catheter-tip pressure transducer, Millar Mikro-Tip), respectively ; and the insertion of stainless steel needle electrodes to record the lead II ECG, Figure 1 shows the basic electrode arrays used to measure thoracic impedance. A constant current (20 kHz or 50 kHz) of approximately 2 mA rms was applied through one pair of electrodes from a battery-powered impedance measuring system (1) that continuously recorded the voltage change (reflecting thoracic impedance change) across other pairs of electrodes. The use of two impedance measuring devices operating at different frequencies permitted measurements to be made from two different electrode arrays simultaneously. After placing the electrodes on the animal, a control value of the basal impedance (Z,>, taken at end expiration, was recorded after a period of 15-30 min during which the electrodes stabilized. Data from the records were entered into a PDP-8S
851
Downloaded from www.physiology.org/journal/jappl at Midwestern Univ Lib (132.174.254.157) on February 13, 2019.
852
J.
‘k
‘k
‘k 1. Four-terminal system for the measurement of thoracic impedance. A; band electrode array. Flexible metal electrodes encircling the neck, thorax, and abdomen were applied as shown. A constant sinusoidal current was applied to electrodes 1 and 4. Voltage reflecting thoracic impedance changes was picked up from electrodes 2 and 3. B: circumferential spot electrode array. Standard disposable EGG-type electrodes were applied to the thorax as shown. A constant sinusoidal current was applied to electrodes 2 and 4, and impedances between other pairs of electrodes were recorded. Thus the designations 2416 (right hemithorax), 2456 (left hemithorax), and 2415 (tranethoracic) indicate current applied to electrodes 2 and 4 and impedance recorded between the other two electrodes. C: circumferential strip electrode array (lo-cm length, electrode tape). The strip electrode designation is identical with that of the circumferential spot electrode array except for the notation “S” (2415-S 2416-S and 2456-S). D: longitudinal spot electrode array. Standard disposable ECG electrodes were applied to the thorax as shown. A constant sinusoidal current was applied to electrodes 1 and 6, and impedance changes between electrodes 2 and 3 (ZI, right hemithorax), 4 and 5 (22, left hemithorax), and 2-4 and 3-5 (23, entire thorax) were recorded. Electrodes 1 and 6 were comprised of a pair of electrodes each, one placed on each side of the neck and abdomen, respectively.
C. DENNISTON
AND
L.
E. BAKER
in detecting, localizing, and quantitating pleural effusion, measurements of thoracic impedance were recorded in two different experimental studies. Study 1. Measurements of thoracic impedance during simulated pleural eJusion using isotonic saline. Pleural effusion was sirnulated by the infusion of 50-ml increments of saline (O-300 ml) into the right, left, or both hemithoraxes while recording impedance from the band, 2415, 2416, 2456, 2415-S, 2416-S 2456-S 21, 22, and 23 electrode arrays (Fig. 1). The sensitivities @/ml or %/ml) of the various electrode arrays were determined. The impedance also was recorded during the withdrawal of saline from the thorax. Study 2. Measurements of thoracic impedance during simulated pleural efusion using isotonic saline or plasma and hemothorax . using blood. This study consisted of two separate experirnents. All animals were heparinized (50 mg, iv), and the left femoral artery was catheterized to permit subsequent bleedings. Thirteen animals were studied. In the first experiment (study ZA), the effects on thoracic impedance from infusing 50-ml increments (O-300 ml) of saline, plasma, or blood into the right hemithorax as recorded from the band, 2415, 21, and 22 electrode arrays were compared. In the second experiment (study 2B) the same comparisons were made, but attempts were made to keep the total body fluid volume constant to simulate more closely pleural effusion and hemothorax. Thus, before each 50-ml infusion of saline, plasma, or blood into the right hemithorax, 50 ml of blood was withdrawn from the left femoral artery. After an experimental run in which one of the three fluids was infused into the thorax, the blood collected (300 ml) was reinfused (iv), and the animals were allowed to return to their control status before the next series of infusions of a different fluid into the right herni thorax.
FIG.
computer from an analog-digital converter (Gerber Scientific Corp.) for analysis. The data were normalized in one of two ways. First, each animal was treated as its own control, and the change in impedance (AZ) with each incremental increase in thoracic fluid was determined. Second, each AZ was expressed as a percent of that animal’s initial impedance (AZ/& X 100). In order to assess the value of the impedance technique
RESULTS
Study 1. Measurements of thoracic impedance during simulated pleural eJusion using isotonic saline. A typical record showing the constant heart rate and blood pressure and the decrease in impedance during the infusion of isotonic saline into the thorax is shown in Fig. 2. The basal irnpedance of individual animals remained constant throughout the control period. However, a wide range of 20 values was recorded with the various electrode arrays and among animals for any given, array. This variation in Zo among animals for the band electrode array is depicted in Fig. 3 where the initial (control) 20 varied between 36.7 and 58.7 Q for individual animals. The wide range of basal X0 values made direct comparisons of measurernents between animals and electrode arrays impractical. Almost all electrode pairs gave results significantly different from each other. While linear regression analyses of group data were not always correlated significantly based on absolute values of normalization of data on the basis of the perimpedance, centage change (AZ/& X 100) altered the magnitude and relationship of the slopes determined for each electrode array. For example, based either on the absolute values of thoracic impedance (Table 1) or on AZ, a slope of -2.1 Q/ 100 ml (band electrode array) and - 1.3 St/ 100 ml
Downloaded from www.physiology.org/journal/jappl at Midwestern Univ Lib (132.174.254.157) on February 13, 2019.
NONINVASIVE
MEASUREMENT
OF
THORACIC
853
FLUIDS
FIG. 2. A typical record of transt hor acic impedance measured using the band electrode array, ECG, and arterial blood pressure during the infusion of 50-ml increments of saline (O-300 ml) into the right hemithorax.
TRANSTtKIRAClC IMPEDANCE
30
100 (ML) 0F SALINE
VOLUME
70
150 INFUSED
1~10
RIGHT
200 ii~~iTt40RA~
A 8
8
50
8 n
1
I
0 8
30-w fo=-0.02IVt r =0.34?
0
0
250
300
46.2
IO--
OJ 0
1
.
50
100
200
.i
L
150
200
t
0 ..
00
*I
A 8
n b
-IO--
0 u 0
&z = -o.o21v+ r = 0.792
0
0
(- 0.5)
3o*v
50 0
1 50
a 100
70
50
1I 250
4 300
200
250
300
0.045V+1.0
X loo=
r = 0.922
i
0
50
no VOLUME
150 (ML)
3. Thoracic impedance measured ray vs. volume (ml) of saline infused into absolute value of thoracic impedance (20, impedance (AZ, Q). C: percent change in 20 X 100, %). V, volume of saline infused; P < 0.001; no. of animals, 17; no. of x, y FIG.
II 200
C AZ/Z,
30
II IS0
using band electrode arthe right hemithorax. A. !2). B: change in thoracic thoracic impedance (AZ/ r, correlation coefficient; pairs, 119.
(2415 electrode array) was calculated ; while, in contrast, with the same data normalized as AZ/& X 100 (Table l), the 24 15 electrode array (9.4 %/ 100 ml) gave a greater relative change in impedance than the band electrode array (4.5 %/ 100 ml). In contrast to the transthoracic electrode arrays (band, 2415, 2415-S, and &), the hemithoracic electrode arrays (2416, 2456, 2416-S, 2456-S, 21, and 2,) permitted localization of accumulated fluid to a specific hemithorax. This is exemplified in Fig. 4, which shows that infusion of saline into the right hemithorax resulted in a slope of - 1.1 Q/100 ml as recorded from the right hemithoracic electrode
250
300
array (2416), and a slope of -0.3 St/ 100 ml during simulelectrode taneous recording from the left hernithoracic array (2456) (Fig. 4, A and B). Similarly slopes of 13.4 %/ 100 ml (2416) and 4.5 %/ 100 ml (2456) were calculated using percent change of impedance (Fig. 4C). These data are summarized in Table 1 for all electrode arrays and infusion sites. Study2. Measurements of thoracic impedance during simulated pleural ejusion using isotonic saline or plasma and hemothorax using blood. In study 2A, measurements of thoracic irnonly from the band and 24 15 pedance were recorded (Table 2) indicate electrode arrays. The data obtained that saline and plasma infusions produced identical changes in thoracic impedance for a given electrode array (i.e., the infusion of either saline or plasma resulted in a slope of - 2.1 Q/ 100 ml as recorded from the band electrode arthe impedance method was less sensitive ray). However, (Q/100 ml) in detecting the infusion of blood into the thorax than it was in detecting infusions of saline or plasma. The greatest difference was noted when using the band electrode array in which slopes of - 2.1 and - 1.2 Q/ 100 ml were calculated for the infusion of saline and blood, respectively, into the right hemithorax. In study 2B, the infusion of saline or plasma produced nearly identical changes in thoracic impedance, and the infusion of blood into the thorax produced a decrease in the impedance sensitivity for all electrode arrays, as it did in study 2A. A comparison of the mean changes in thoracic impedance values for 300 ml of exogenous and endogenous thoracic fluid shifts based on data from the band and 24 15 electrode arrays is presented in Table 3. These data show that, although all exogenous fluid infused into the right hemithorax produced a greater change in impedance for both electrode arrays than occurred with a simulated endogenous fluid shift, only with the band electrode array and blood was a significant difference in thoracic impedance detected between the two experimental conditions. DISCUSSION
The data presented here indicate that changes of impedance can be used to predict specific changes of fluid volume within the thorax in dogs under experimental conditions. Increased volumes of intrathoracic fluid always were associated with linear decreases of impedance. The nearly identical impedance changes obtained during the infusion of saline or plasma indicate that infusion of saline into the thorax provides a reasonable model of developing pleural effusion. These results were not unexpected since the resistivities of saline and plasma are both approximately 60 Q-cm (6). The linear decrease in impedance (2.1 St/ 100
Downloaded from www.physiology.org/journal/jappl at Midwestern Univ Lib (132.174.254.157) on February 13, 2019.
854
J.
1. Summary of hear --~--~-_ --. --.. -- -~- _ - --~_-_-~
TABLE
Electrode
Array
L.
E. BAKER
Infusion
Site
n
N
20 = aV + b
~
AZ/Z0
a
Y
b
X 100 = aV+ a
b Y
%3
IRS IRS IRS IRS IRS IRS IRS IRS IRS IRS
17 11 8 8 6 4 4 4 4 4
119 77 56 56 42 28 28 28 28 28
46.2 12.4 6.8 7.5 8.0 5.0 4.7 19.1 19.8 19.4
-0.021 -0.013 -0.010 -0.003 -0.009 -0.005 -0.002 -0.022 -0.014 -0.018
-0.347* -0.413* -O-488* -0.120* -0.572* -0.547* -0.615* -0.790* -0.468* -0.669*
0.0 4.3 7.9 1.9 4.7 7.3 2.8 2.4 1 .o 1.6
0.045 0.094 0.134 0.045 0.105 0.097 0.046 0.110 0.070 0.090
0.922* 0.852* 0.868* 0.736* 0.834* 0.898* 0.863* 0.923* 0.916* 0.969*
Bands 2415 2416 2456 2415-S 2416-S 2456-S
ILS ILS ILS ILS ILS ILS ILS
13 6 6 6 6 4 4
91 42 42 42 42 28 28
47.3 12.2 6.6 7.6 8.1 5.2 4.7
-0.021 -0.014 -0.006 -0 .OlO -0.006 -0.003 -0.005
-0.332 -0.452 -0.338 -0.365 -0.422 -0.301 -0.410
0.9 4.8 2.0 4.3 3.4 1.4 5.4
0.045 0.105 0.080 0.120 0.071 0.047 0.108
0.927* 0.898* 0.740* 0.895* 0.770* 0.822* 0.932*
Bands 2415 2416 2456
IBS IBS IBS IBS
5 5 5 5
20 20 20 20
48.2 12.7 6.2 6.6
-0.023 -0.016 -0.008 -0.008
-0.464 -0.507 -0.448 -0.556
0.9 2.6 2.5 2.8
0.047 0.115 0.112 0.112
0.912* 0.935* 0.883* 0.877*
21 22
Data determined from volume (ml) of saline n, number of animals; N, IRS, infused right side; VS.
AND
regression data
b
Bands 2415 2416 2456 2415-S 2416-S 2456-S
C. DENNISTON
the absolute values of thoracic impedance (20, sl) and percent change in thoracic impedance (AZ/Z, infused into the right, left, or both hemithoraxes. Saline was infused in increments-of 50 ml from number of x, y pairs; a, slope (Q/ml) ; b, y intercept (fi) ; V, volume of saline infused; r, correlation ILS, infused left side; IBS, infused both sides. * P < 0.05.
ml) recorded from the band electrode array during the infusion of saline into the thorax is in close agreement with the observations of Van De Water et al. (24) of 2.3 Q/100 ml in a single dog, but it is different from the increase of 0.23 Q/ 100 ml of effusion fluid removed during thoracentesis of a human patient. The wide range observed in basal impedance values recorded from any given electrode array probably represents, for the most part, variation in electrode position or location due to the constraints of placing the electrodes at specific anatomic sites in dogs of varying size and conformation. The current density is increased in the vicinity of the current-passing electrodes, particularly the one on the neck, producing a large potential gradient in that region which diminishes down the thorax as the cross-sectional area increases. Thus the closer the detecting electrodes are to the current electrodes, the higher the basal impedance measured. Similarly the closer the detecting electrodes are to each other, the smaller the impedance when the current electrodes are maintained in a fixed position. The range of basal impedance values measured by the various electrode arrays, particularly the band and 2415 electrode arrays, represents differences in current pathways. In considering the basic relationship of R = &/A), and assuming a minimal change in p between the detecting electrodes for either the band or 2415 electrode arrays, the factor L for the 2415 electrode array would be small with respect to A, resulting in a relatively small R and hence the lower basal impedance for the 2415 electrode array. The observation that the band electrode array was always more sensitive (Q/ml) to fluid infusions of the pleural cavity than the 2415 electrode array, on the basis of a change in the absolute value of impedance, can be ex-
X 100, %) 0 to 300 ml; coefficient.
plained by the differences in current pathways between the two arrays. The addition of a highly conductive fluid, such as isotonic saline, in parallel with a relatively high impedance path, as obtained with the band electrode array, would cause a greater decrease in impedance than the same fluid placed in parallel with a lower impedance path, as is recorded with the 2415 electrode array. The greater sensitivity of the 2415 electrode array on the basis of percent impedance change (AZ/& X 100) as compared with the band electrode array is related to the initial impedance value. For example, a band electrode array with an initial impedance of 45 Q would require a change in impedance (AZ) three times as great as that from a 2415 electrode array having an initial impedance of 15 i;t to produce the same percent change in impedance (AZ/& X 100). The 23 electrode array detected a change in impedance ( 1.8 St/ 100 ml) which was not significantly different (P > 0.20) from that of the band electrode array (2.1 Q/100 ml), and on a percent basis detected a significantly (P < 0.001) greater change than the band electrode array did during the infusion of saline into the thorax. Clinically the use of ECG disposable spot electrodes, as used with the 23 electrode array, is aesthetically more acceptable than bands, and their use, in the dog, appears to cause no significant difference from the band array in detecting simulated pleural effusion. The use of the hemithoracic electrode arrays (2415, 2456, 21, and 2,) permits localization of fluid to a single hemithorax. However, it may be necessary to limit the interpretation of results to the analysis of changes in impedance (AZ) rather than percent changes in impedance (AZ/& X 100). Although in this study the greatest percent change in impedance was always associated with the
Downloaded from www.physiology.org/journal/jappl at Midwestern Univ Lib (132.174.254.157) on February 13, 2019.
NONINVASIVE
MEASUREMENT
OF THORACIC
FLUIDS
855 2. hh&!muryof the h?ur reg-ession data --P-P_._ ----.._-- ~
TABLE
Electrode Array
20 (2456)
-0
50
-.
= -0.003vt15
100
200
250
300
-0.003V+
A2e456j=
(-0.l)-
r=0.817
~2~~~~)
= -0.01 IV+ t-0.06)
n
N
b
-~
a
--~-
~-- --
SEE
-
Y
--
P Value
---~-.--
Band Band Band
Blood Saline Plasma
4 4 4
28 28 28
-0.2 -0.5 -0.4
-0.012 -0.021 -0.021
0.39 0.72 0.68
-0.955 -0.947 -0.951
in U.S.A.
Measurement by electrical
of pleural
effusion
impedance
JOSEPH C. DENNISTON AND LEE E. BAKER Department of Physiology, Baylor College of Medicine, Texas Medical
detecting and localizing pleural effusion (1 I), it gained general clinical acceptance for this purpose. The present study was undertaken to assess, in trolled experimental setting, the electrical impedance nique for detecting, quantitating, and monitoring in pleural fluid volumes in the dog and to provide mental data for the interpretation of impedance measured in the clinical environment.
DENNISTON, JOSEPH C., AND LEE E. BAKER. Measurement of pleural effusion by electrical impedance. J. Appl. Physiol. 38(5) : 8514357. 1975.-Changes in the electrical impedance of the thorax were recorded from various electrode arrays on the thorax during the infusion and withdrawal of saline, plasma, or blood from the right, left, or both hemithoraxes. Impedance changes correlated linearly with the volume of infused fluid. The resistivity of the infused fluid significantlyaffected the sensitivity &?/ml) of the impedance method in detecting specific fluid accumulations. The use of hemithoracic electrode arrays permitted the lacalization of fluid accumulations to a specific hemithorax. Various factors that can alter the magnitude and interpretation of impedance changes are discussed, as they would affect the clinical interpretation of impedance changes, particularly in postoperative patients. thoracic noninvasive
electrical impedance; measurement
pleural
effusion,
Center, Houston, Texas 7702.5
MATERIALS
hemothorax;
THE USE OF ELECTRICAL impedance as a means of m .easuring physiological activity has been investigated for many years. Clinically the technique has been used as a noninvasive measure of stroke volume (3, 12, 14, 18), myocardial contractility (13, Zl), periph era1 vascular disease (5, 16, 22, 26), and tidal volume (2, 8). More recently the impedance technique has been used to measure intrathoracic fluids (15, 17, 23, 24). The need exists for a simple, noninvasive means of continuously monitoring shifts in thoracic fluid volumes of patients following trauma, pulmonary and myocardial infarction, and thoracic surgery. The clinical diagnosis of pleural effusion or hemothorax is based primarily on patient history, physical findings (auscultation and percussion), and roentgenographic evaluation of the thorax. Unfortunately these methods of detection are relatively insensitive. Neither percussion nor auscultation is quantitative in nature, and frequently neither detects effusion. The use of the lateral decubitus position, which allows radiographic demonstration of quantities of less than 100 ml of fluid to be visualized (19), has become a standard technique for determining small amounts of pleural effusion ; however, pleural adhesions may prevent layering of fluid, and the densities of soft tissue may mimic layering of free fluid lateral to the lung, resulting in an incorrect diagnosis (10). ultrasound , whi ch has gained wideAlthou gh reflected spread use in cardiology, has been used w ith success in
AND
has not a contechchanges experichanges
METHODS
Two studies, utilizing 43 healthy mongrel dogs weighing from 15 to 22 kg, were conducted. Preanesthetic medication consisted of atropine sulfate (0.02 mg/kg, im) and a solution containing 0.4 mg/ml of fentanyl and 20 mg/ml of droperidol (0.1 ml/kg) g iven iv. All animals were anesthetized with pentobarbital sodium (15-20 mg/kg, iv), and an oral endotracheal airway was established. Succinylcholine chloride (0.1-0.2 mg/kg, iv) was given as required to maintain short-term respiratory paralysis during pressurecycled controlled ventilation. Routine prestudy included total surface depilation of the neck, thorax, and abdomen to insure good electrical contact of the electrodes; placement of a standard 24 Fr thoracic drainage tube into each hemithorax through a l-in bilateral thoracotomy at the 10th intercostal space to permit the infusion and withdrawal of fluid (saline, plasma, or blood at 37°C) from the thorax; catheterization of the right femoral vein and artery to permit subsequent iv infusions and pressure recordings (catheter-tip pressure transducer, Millar Mikro-Tip), respectively ; and the insertion of stainless steel needle electrodes to record the lead II ECG, Figure 1 shows the basic electrode arrays used to measure thoracic impedance. A constant current (20 kHz or 50 kHz) of approximately 2 mA rms was applied through one pair of electrodes from a battery-powered impedance measuring system (1) that continuously recorded the voltage change (reflecting thoracic impedance change) across other pairs of electrodes. The use of two impedance measuring devices operating at different frequencies permitted measurements to be made from two different electrode arrays simultaneously. After placing the electrodes on the animal, a control value of the basal impedance (Z,>, taken at end expiration, was recorded after a period of 15-30 min during which the electrodes stabilized. Data from the records were entered into a PDP-8S
851
Downloaded from www.physiology.org/journal/jappl at Midwestern Univ Lib (132.174.254.157) on February 13, 2019.
852
J.
‘k
‘k
‘k 1. Four-terminal system for the measurement of thoracic impedance. A; band electrode array. Flexible metal electrodes encircling the neck, thorax, and abdomen were applied as shown. A constant sinusoidal current was applied to electrodes 1 and 4. Voltage reflecting thoracic impedance changes was picked up from electrodes 2 and 3. B: circumferential spot electrode array. Standard disposable EGG-type electrodes were applied to the thorax as shown. A constant sinusoidal current was applied to electrodes 2 and 4, and impedances between other pairs of electrodes were recorded. Thus the designations 2416 (right hemithorax), 2456 (left hemithorax), and 2415 (tranethoracic) indicate current applied to electrodes 2 and 4 and impedance recorded between the other two electrodes. C: circumferential strip electrode array (lo-cm length, electrode tape). The strip electrode designation is identical with that of the circumferential spot electrode array except for the notation “S” (2415-S 2416-S and 2456-S). D: longitudinal spot electrode array. Standard disposable ECG electrodes were applied to the thorax as shown. A constant sinusoidal current was applied to electrodes 1 and 6, and impedance changes between electrodes 2 and 3 (ZI, right hemithorax), 4 and 5 (22, left hemithorax), and 2-4 and 3-5 (23, entire thorax) were recorded. Electrodes 1 and 6 were comprised of a pair of electrodes each, one placed on each side of the neck and abdomen, respectively.
C. DENNISTON
AND
L.
E. BAKER
in detecting, localizing, and quantitating pleural effusion, measurements of thoracic impedance were recorded in two different experimental studies. Study 1. Measurements of thoracic impedance during simulated pleural eJusion using isotonic saline. Pleural effusion was sirnulated by the infusion of 50-ml increments of saline (O-300 ml) into the right, left, or both hemithoraxes while recording impedance from the band, 2415, 2416, 2456, 2415-S, 2416-S 2456-S 21, 22, and 23 electrode arrays (Fig. 1). The sensitivities @/ml or %/ml) of the various electrode arrays were determined. The impedance also was recorded during the withdrawal of saline from the thorax. Study 2. Measurements of thoracic impedance during simulated pleural efusion using isotonic saline or plasma and hemothorax . using blood. This study consisted of two separate experirnents. All animals were heparinized (50 mg, iv), and the left femoral artery was catheterized to permit subsequent bleedings. Thirteen animals were studied. In the first experiment (study ZA), the effects on thoracic impedance from infusing 50-ml increments (O-300 ml) of saline, plasma, or blood into the right hemithorax as recorded from the band, 2415, 21, and 22 electrode arrays were compared. In the second experiment (study 2B) the same comparisons were made, but attempts were made to keep the total body fluid volume constant to simulate more closely pleural effusion and hemothorax. Thus, before each 50-ml infusion of saline, plasma, or blood into the right hemithorax, 50 ml of blood was withdrawn from the left femoral artery. After an experimental run in which one of the three fluids was infused into the thorax, the blood collected (300 ml) was reinfused (iv), and the animals were allowed to return to their control status before the next series of infusions of a different fluid into the right herni thorax.
FIG.
computer from an analog-digital converter (Gerber Scientific Corp.) for analysis. The data were normalized in one of two ways. First, each animal was treated as its own control, and the change in impedance (AZ) with each incremental increase in thoracic fluid was determined. Second, each AZ was expressed as a percent of that animal’s initial impedance (AZ/& X 100). In order to assess the value of the impedance technique
RESULTS
Study 1. Measurements of thoracic impedance during simulated pleural eJusion using isotonic saline. A typical record showing the constant heart rate and blood pressure and the decrease in impedance during the infusion of isotonic saline into the thorax is shown in Fig. 2. The basal irnpedance of individual animals remained constant throughout the control period. However, a wide range of 20 values was recorded with the various electrode arrays and among animals for any given, array. This variation in Zo among animals for the band electrode array is depicted in Fig. 3 where the initial (control) 20 varied between 36.7 and 58.7 Q for individual animals. The wide range of basal X0 values made direct comparisons of measurernents between animals and electrode arrays impractical. Almost all electrode pairs gave results significantly different from each other. While linear regression analyses of group data were not always correlated significantly based on absolute values of normalization of data on the basis of the perimpedance, centage change (AZ/& X 100) altered the magnitude and relationship of the slopes determined for each electrode array. For example, based either on the absolute values of thoracic impedance (Table 1) or on AZ, a slope of -2.1 Q/ 100 ml (band electrode array) and - 1.3 St/ 100 ml
Downloaded from www.physiology.org/journal/jappl at Midwestern Univ Lib (132.174.254.157) on February 13, 2019.
NONINVASIVE
MEASUREMENT
OF
THORACIC
853
FLUIDS
FIG. 2. A typical record of transt hor acic impedance measured using the band electrode array, ECG, and arterial blood pressure during the infusion of 50-ml increments of saline (O-300 ml) into the right hemithorax.
TRANSTtKIRAClC IMPEDANCE
30
100 (ML) 0F SALINE
VOLUME
70
150 INFUSED
1~10
RIGHT
200 ii~~iTt40RA~
A 8
8
50
8 n
1
I
0 8
30-w fo=-0.02IVt r =0.34?
0
0
250
300
46.2
IO--
OJ 0
1
.
50
100
200
.i
L
150
200
t
0 ..
00
*I
A 8
n b
-IO--
0 u 0
&z = -o.o21v+ r = 0.792
0
0
(- 0.5)
3o*v
50 0
1 50
a 100
70
50
1I 250
4 300
200
250
300
0.045V+1.0
X loo=
r = 0.922
i
0
50
no VOLUME
150 (ML)
3. Thoracic impedance measured ray vs. volume (ml) of saline infused into absolute value of thoracic impedance (20, impedance (AZ, Q). C: percent change in 20 X 100, %). V, volume of saline infused; P < 0.001; no. of animals, 17; no. of x, y FIG.
II 200
C AZ/Z,
30
II IS0
using band electrode arthe right hemithorax. A. !2). B: change in thoracic thoracic impedance (AZ/ r, correlation coefficient; pairs, 119.
(2415 electrode array) was calculated ; while, in contrast, with the same data normalized as AZ/& X 100 (Table l), the 24 15 electrode array (9.4 %/ 100 ml) gave a greater relative change in impedance than the band electrode array (4.5 %/ 100 ml). In contrast to the transthoracic electrode arrays (band, 2415, 2415-S, and &), the hemithoracic electrode arrays (2416, 2456, 2416-S, 2456-S, 21, and 2,) permitted localization of accumulated fluid to a specific hemithorax. This is exemplified in Fig. 4, which shows that infusion of saline into the right hemithorax resulted in a slope of - 1.1 Q/100 ml as recorded from the right hemithoracic electrode
250
300
array (2416), and a slope of -0.3 St/ 100 ml during simulelectrode taneous recording from the left hernithoracic array (2456) (Fig. 4, A and B). Similarly slopes of 13.4 %/ 100 ml (2416) and 4.5 %/ 100 ml (2456) were calculated using percent change of impedance (Fig. 4C). These data are summarized in Table 1 for all electrode arrays and infusion sites. Study2. Measurements of thoracic impedance during simulated pleural ejusion using isotonic saline or plasma and hemothorax using blood. In study 2A, measurements of thoracic irnonly from the band and 24 15 pedance were recorded (Table 2) indicate electrode arrays. The data obtained that saline and plasma infusions produced identical changes in thoracic impedance for a given electrode array (i.e., the infusion of either saline or plasma resulted in a slope of - 2.1 Q/ 100 ml as recorded from the band electrode arthe impedance method was less sensitive ray). However, (Q/100 ml) in detecting the infusion of blood into the thorax than it was in detecting infusions of saline or plasma. The greatest difference was noted when using the band electrode array in which slopes of - 2.1 and - 1.2 Q/ 100 ml were calculated for the infusion of saline and blood, respectively, into the right hemithorax. In study 2B, the infusion of saline or plasma produced nearly identical changes in thoracic impedance, and the infusion of blood into the thorax produced a decrease in the impedance sensitivity for all electrode arrays, as it did in study 2A. A comparison of the mean changes in thoracic impedance values for 300 ml of exogenous and endogenous thoracic fluid shifts based on data from the band and 24 15 electrode arrays is presented in Table 3. These data show that, although all exogenous fluid infused into the right hemithorax produced a greater change in impedance for both electrode arrays than occurred with a simulated endogenous fluid shift, only with the band electrode array and blood was a significant difference in thoracic impedance detected between the two experimental conditions. DISCUSSION
The data presented here indicate that changes of impedance can be used to predict specific changes of fluid volume within the thorax in dogs under experimental conditions. Increased volumes of intrathoracic fluid always were associated with linear decreases of impedance. The nearly identical impedance changes obtained during the infusion of saline or plasma indicate that infusion of saline into the thorax provides a reasonable model of developing pleural effusion. These results were not unexpected since the resistivities of saline and plasma are both approximately 60 Q-cm (6). The linear decrease in impedance (2.1 St/ 100
Downloaded from www.physiology.org/journal/jappl at Midwestern Univ Lib (132.174.254.157) on February 13, 2019.
854
J.
1. Summary of hear --~--~-_ --. --.. -- -~- _ - --~_-_-~
TABLE
Electrode
Array
L.
E. BAKER
Infusion
Site
n
N
20 = aV + b
~
AZ/Z0
a
Y
b
X 100 = aV+ a
b Y
%3
IRS IRS IRS IRS IRS IRS IRS IRS IRS IRS
17 11 8 8 6 4 4 4 4 4
119 77 56 56 42 28 28 28 28 28
46.2 12.4 6.8 7.5 8.0 5.0 4.7 19.1 19.8 19.4
-0.021 -0.013 -0.010 -0.003 -0.009 -0.005 -0.002 -0.022 -0.014 -0.018
-0.347* -0.413* -O-488* -0.120* -0.572* -0.547* -0.615* -0.790* -0.468* -0.669*
0.0 4.3 7.9 1.9 4.7 7.3 2.8 2.4 1 .o 1.6
0.045 0.094 0.134 0.045 0.105 0.097 0.046 0.110 0.070 0.090
0.922* 0.852* 0.868* 0.736* 0.834* 0.898* 0.863* 0.923* 0.916* 0.969*
Bands 2415 2416 2456 2415-S 2416-S 2456-S
ILS ILS ILS ILS ILS ILS ILS
13 6 6 6 6 4 4
91 42 42 42 42 28 28
47.3 12.2 6.6 7.6 8.1 5.2 4.7
-0.021 -0.014 -0.006 -0 .OlO -0.006 -0.003 -0.005
-0.332 -0.452 -0.338 -0.365 -0.422 -0.301 -0.410
0.9 4.8 2.0 4.3 3.4 1.4 5.4
0.045 0.105 0.080 0.120 0.071 0.047 0.108
0.927* 0.898* 0.740* 0.895* 0.770* 0.822* 0.932*
Bands 2415 2416 2456
IBS IBS IBS IBS
5 5 5 5
20 20 20 20
48.2 12.7 6.2 6.6
-0.023 -0.016 -0.008 -0.008
-0.464 -0.507 -0.448 -0.556
0.9 2.6 2.5 2.8
0.047 0.115 0.112 0.112
0.912* 0.935* 0.883* 0.877*
21 22
Data determined from volume (ml) of saline n, number of animals; N, IRS, infused right side; VS.
AND
regression data
b
Bands 2415 2416 2456 2415-S 2416-S 2456-S
C. DENNISTON
the absolute values of thoracic impedance (20, sl) and percent change in thoracic impedance (AZ/Z, infused into the right, left, or both hemithoraxes. Saline was infused in increments-of 50 ml from number of x, y pairs; a, slope (Q/ml) ; b, y intercept (fi) ; V, volume of saline infused; r, correlation ILS, infused left side; IBS, infused both sides. * P < 0.05.
ml) recorded from the band electrode array during the infusion of saline into the thorax is in close agreement with the observations of Van De Water et al. (24) of 2.3 Q/100 ml in a single dog, but it is different from the increase of 0.23 Q/ 100 ml of effusion fluid removed during thoracentesis of a human patient. The wide range observed in basal impedance values recorded from any given electrode array probably represents, for the most part, variation in electrode position or location due to the constraints of placing the electrodes at specific anatomic sites in dogs of varying size and conformation. The current density is increased in the vicinity of the current-passing electrodes, particularly the one on the neck, producing a large potential gradient in that region which diminishes down the thorax as the cross-sectional area increases. Thus the closer the detecting electrodes are to the current electrodes, the higher the basal impedance measured. Similarly the closer the detecting electrodes are to each other, the smaller the impedance when the current electrodes are maintained in a fixed position. The range of basal impedance values measured by the various electrode arrays, particularly the band and 2415 electrode arrays, represents differences in current pathways. In considering the basic relationship of R = &/A), and assuming a minimal change in p between the detecting electrodes for either the band or 2415 electrode arrays, the factor L for the 2415 electrode array would be small with respect to A, resulting in a relatively small R and hence the lower basal impedance for the 2415 electrode array. The observation that the band electrode array was always more sensitive (Q/ml) to fluid infusions of the pleural cavity than the 2415 electrode array, on the basis of a change in the absolute value of impedance, can be ex-
X 100, %) 0 to 300 ml; coefficient.
plained by the differences in current pathways between the two arrays. The addition of a highly conductive fluid, such as isotonic saline, in parallel with a relatively high impedance path, as obtained with the band electrode array, would cause a greater decrease in impedance than the same fluid placed in parallel with a lower impedance path, as is recorded with the 2415 electrode array. The greater sensitivity of the 2415 electrode array on the basis of percent impedance change (AZ/& X 100) as compared with the band electrode array is related to the initial impedance value. For example, a band electrode array with an initial impedance of 45 Q would require a change in impedance (AZ) three times as great as that from a 2415 electrode array having an initial impedance of 15 i;t to produce the same percent change in impedance (AZ/& X 100). The 23 electrode array detected a change in impedance ( 1.8 St/ 100 ml) which was not significantly different (P > 0.20) from that of the band electrode array (2.1 Q/100 ml), and on a percent basis detected a significantly (P < 0.001) greater change than the band electrode array did during the infusion of saline into the thorax. Clinically the use of ECG disposable spot electrodes, as used with the 23 electrode array, is aesthetically more acceptable than bands, and their use, in the dog, appears to cause no significant difference from the band array in detecting simulated pleural effusion. The use of the hemithoracic electrode arrays (2415, 2456, 21, and 2,) permits localization of fluid to a single hemithorax. However, it may be necessary to limit the interpretation of results to the analysis of changes in impedance (AZ) rather than percent changes in impedance (AZ/& X 100). Although in this study the greatest percent change in impedance was always associated with the
Downloaded from www.physiology.org/journal/jappl at Midwestern Univ Lib (132.174.254.157) on February 13, 2019.
NONINVASIVE
MEASUREMENT
OF THORACIC
FLUIDS
855 2. hh&!muryof the h?ur reg-ession data --P-P_._ ----.._-- ~
TABLE
Electrode Array
20 (2456)
-0
50
-.
= -0.003vt15
100
200
250
300
-0.003V+
A2e456j=
(-0.l)-
r=0.817
~2~~~~)
= -0.01 IV+ t-0.06)
n
N
b
-~
a
--~-
~-- --
SEE
-
Y
--
P Value
---~-.--
Band Band Band
Blood Saline Plasma
4 4 4
28 28 28
-0.2 -0.5 -0.4
-0.012 -0.021 -0.021
0.39 0.72 0.68
-0.955 -0.947 -0.951
