Partitioning diaphragm
of inspiratory pressure and intercostal/accessory
swings between muscles
PETER T. MACKLEM, D. GROSS, A. GRASSINO, AND C. ROUSSOS Meakins-Christie Laboratories, McGill University Clinic, Royal Victoria Hospital, Montreal H3A 2B4 Quebec, Canada
D. GROSS, A. GRASSINO, AND C. pressure swings between diaphragm and intercostal/accessory muscles. J. Appl. Physiol. : Respirat. Environ. Exercise Physiol. 44(2): 200-208, 1978. -We tested the hypothesis that the inspiratory pressure swings across the rib-cage pathway are the sum of transdiaphragmatic pressure (Pdi) and the pressures developed by the intercostal/accessory muscles (Pit). If correct, Pit can only contribute to lowering pleural pressure (Ppl), to the extent that it lowers abdominal pressure (Pab). To test this we measured Pab and Ppl during Mueller maneuvers. Contraction of external intercostal muscles was inferred from surface EMG recordings. We found evidence of external intercostal contraction during Mueller maneuvers in which APab = 0. Because there was no outward displacement of the rib cage, Pit must have contributed to APpl, as did Pdi. Under these conditions the total pressure developed by the inspiratory muscles across the rib-cage pathway was less than Pdi + Pit. Therefore, we rejected the hypothesis. A plot of Pab vs. Ppl during relaxation allows partitioning of the diaphragmatic and intercostal/accessory muscle contributions to inspiratory pressure swings. The analysis indicates that the diaphragm can act both as a fixator, preventing transmission of Ppl to the abdomen and as an agonist. When abdominal muscles remain relaxed it only assumes the latter role to the extent that Pab increases, MACKLEM, PETER ROUSSOS. Partitioning
T.,
of inspiratory
respiratory muscles; respiratory mechanics; EM G; fixators; agonists; pleural pressure; abdomi .nal pressure
AN UNDERSTANDING OF the mechanical link between the diaphragm and rib cage is necessary in order to determine the relative contributions of the diaphragm and intercostal/accessory muscles to respiratory pressure swings. The situation is relatively simple when either one or the other is the only muscle contracting. When the intercostal/accessory muscles contract alone, transdiaphragmatic pressure (Pdi) remains zero at least initially, so that abdominal pressure (Pab) remains equal to pleural pressure (Ppl) and falls as inspiration proceeds. As a result the abdomen is displaced inward and upward into the thorax causing the diaphragm to lengthen. Eventually the diaphragm will develop passive tension and as a result, Pdi will rise. In this paper we are concerned with events before passive tension causes a rise in Pdi. Goldman and Mead have suggested that in upright humans when the diaphragm is the only muscle cun-
tracting, the increase in Pab that results, drives the relaxed rib cage (5). Indeed they showed that the relationship between Pab and rib cage volume was identical during quiet breathing to the relationship obtained during relaxation and during abdominal compression in relaxed subjects, whereas the relationship between Ppl and rib-cage volume was different during the three conditions. Thus, they suggested that when the diaphragm is the only muscle contracting, it shortens, producing an increase in Pab which displaces both the relaxed abdomen and rib cage, and a fall in Ppl which inflates the lungs. If, in one instance the diaphragm lengthens, whereas in the other it shortens, it must be possible to inspire in such a way that the diaphragm contracts isometrically. Under these circumstances the diaphragm has presumably performed no external work, and all the work performed and the pressures required to inflate the lung and displace the chest wall must have been accomplished by the intercostal/accessory muscles. Under what circumstances does the diaphragm contract isometrically? If the abdominal muscles remain relaxed then abdominal displacement and/or Pab can be used as an approximate indicator of diaphragmatic length. Grassino has shown that when the diaphragm contracts isometrically, changes in abdominal dimensions are small (6). Therefore, as a useftil approximation we assume that when abdominal muscles remain relaxed and APab = 0 during an inspiration, the diaphragm has contracted quasi-isometrically and performed no external work. The condition of abdominal relaxation pertains to the rest of our theoretical development and is assumed unless specifically stated otherwise. Thus when the diaphragm contracts quasi-isometrically and APab = 0, APdi = -APpl. When the diaphragm produces no external work the pressure changes generated by the intercostal/accessory muscles (APic) must equal the sum of the pressures required to displace the lung (APL) and chest wall (APw) so that APic = APw + APL. When mouth pressure remains atmospheric APL = -APpl = APdi. Thus, under these circumstances, APpl has been achieved by both the diaphragm and intercostallaccessory muscles and this pressure change is shared by both. However, only the intercostal/accessory muscles shorten and the work of inflating the lung is performed by these muscles, and not by the diaphragm. This analysis departs in important ways from that
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
PARTITIONING
OF
RESPIRATORY
201
PRESSURES
developed by Goldman et al. (4). They suggested that during inspirations when APab = 0, the diaphragm performs the work of inflating the lung, and the intercostal/accessory muscles, the work of displacing the chest wall, in contrast to our approach in which we attribute all the work to the intercostal/accessory muscles and none to the diaphragm. Although their approach is appealing, it fails to explain how an isometrically contracting muscle can perform external work. In dealing with this issue, they state that “* . . it would appear from our analysis . that) the diaphragm can perform mechanical work without shortening substantially. We note here that ‘substantially’ is a key word (we do not propose that a truly isometric contraction can perform external mechanical work) . . . .” It can certainly be argued (as they do) that an inspiration performed without abdominal displacement (APab = 0, when the abdominal muscles remain relaxed) does not represent a truly isometric contraction. We have assumed that it does because there is substantial evidence that abdominal displacements are an approximate indicator of diaphragmatic length (6). It is for this reason that we assume that diaphragmatic contraction is quasi-isometric when APab = 0 and that to a useful approximation the diaphragm has performed no external work. However, the analysis of Goldman et al. (4) attributes positive external work to the diaphragm whenever Pdi > 0. But it is certainly possible for an inspiration to occur with a substantial fall in Pab, an inward abdominal displacement, and a small increase in Pdi. During such an inspiration, the diaphragm must be lengthening as it contracts. Thus, there is an even greater paradox in their analysis - the diaphragm is interpreted as doing inspiratory work on the lung as it is lengthening, In the model of Goldman et al., the diaphragm and intercostal/accessory muscles are arranged serially in terms of the pressures developed across the pathway from the mouth to the rib cage (4, 7). Starting at the mouth, and proceeding in order, this pathway consists of the lungs, the diaphragm, the intercostal/accessory muscles and the rib cage. Ppl is the pressure between the lungs and diaphragm, and Pab is the pressure between the diaphragm and intercostal/accessory muscles. The total pressure change developed across this pathway is the sum of the pressures developed by the two muscle groups, i.e., Pdi + Pit. When mouth pressure is atmospheric and APab = 0, APdi = -APpl and, as explained above, they attribute all the change in Ppl to the diaphragm and all the pressures necessary to displace the chest wall to the intercostal/accessory muscles. If this view is correct, during such a breath the intercostal/accessory muscles have not contributed to the transpulmonary pressure swings. Indeed, it is implicit in their analysis that the primary effect of intercostal/accessory muscle contraction is a lowering of Pab and that this change in pressure must be transmitted through the diaphragm in order to influence pleural pressure. This prediction of the Goldman and Mead model is intuitively unlikely. On anatomical consideral
l
tions alone, the inspiratory intercostals are situated closer to the pleural surface than they are to the abdomen. It seems unlikely that they must lower Pab before they can lower Ppl. This is supported by experiments in animals, in which intercostal muscle contraction causes greater pressure swings over the costal pleural surface than over the diaphragm (l), and in humans they cause greater pressure swings in the upper esophagus than in the lower (2). These data indicate that intercostal muscle contraction produces pressure changes primarily at sites close to where these muscles are situated, i.e., the pleural surface. We suggest that under the particular circumstances when the diaphragm contracts quasi-isometrically the pressures developed across the lung are shared by both the diaphragm and intercostal/accessory muscles and that therefore the sum of APdi and APic are greater than the total respiratory pressure swings developed across the rib-cage pathway. We further predict that intercostal/accessory muscle contraction can contribute to pleural pressure changes without Pab changing in a negative direction. This prediction can be tested experimentally. It is one of the purposes of this paper to present the results of this test. We show that APpl during Mueller maneuvers when APab = 0 is shared by both diaphragm and intercostal/accessory muscles in keeping with our analysis. We then develop the analysis further in order to partition the diaphragmatic and intercostal/accessory muscle contributions to respiratory pressure swings and conclude that the diaphragm can function either as an agonist or a fixator. When acting agonistically, the diaphragm shortens and the resulting increase in Pab drives the rib cage. When fixating, the diaphragm prevents the transmission of Ppl to the abdomen and thus prevents paradoxical movement of this compartment during inspiration. However, the work of inflating both the lung and the chest wall under these circumstances is performed entirely by the intercostal/accessory muscles. RATIONALE
AND
METHODS
During Mueller maneuvers the volume of the respiratory system and thus the chest wall (VW) remains almost constant. Thus, any changes in rib cage volume (AVrc) that occur must be equal and opposite to changes in abdominal volume (AVab) because AVw = AVrc + AVab = 0. If the performance of the Mueller maneuver was accomplished with diaphragm alone (all other muscles remaining relaxed) and if Pab drives both the relaxed rib cage and the abdomen, there could be no change in Pab. If Pab were to change, AVrc and AVab would have the same sign and reciprocal changes are impossible. Thus, according to Goldman and Mead’s approach, when the diaphragm is the only muscle contracting during the Mueller maneuver, Pab cannot change. On the other hand, if the relaxed rib cage is driven by Ppl, the Mueller maneuver performed by the diaphragm alone would again result in no change in VW, but there would be equal and opposite changes in Vrc
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
202
MACKLEM,
and Vab because the decrease in Ppl with diaphragmatic contraction would deflate the rib cage, whereas the increase in Pab would inflate the abdomen. Thus AVab = APab Cab = - AVrc = - APpl Crc l
they did not recruit the maneuvers.
GROSS,
GRASSINO,
other expiratory
AND
ROUSSOS
muscles during
RESULTS
l
where Cab and Crc are the compliance of abdomen and rib cage respectively. To prevent the inward displacement of the rib cage it would be necessary to recruit the intercostal/accessory muscles. If this were accomplished just to the extent that AVrc = 0, then AVab = 0 = APab. In this instance, the Mueller maneuver performed at constant Pab would result in intercostal/ accessory muscle recruitment whereas if Goldman and Mead’s approach is correct such recruitment would not occur. Furthermore, because AVrc = 0 the pressure change across the rib cage must be zero and any intercostal/accessory muscle contraction that occurred must have lowered Ppl without lowering Pab. We studied three normal subjects who were highly trained to perform respiratory maneuvers. We measured inspiratory intercostal and diaphragmatic EMG with surface electrodes placed in the 2nd intercostal space parasternally and in the 6th and 7th intercostal space in the anterior axillary line, respectively. Although the latter electrodes presumably detect electrical activity originating from muscles other than the diaphragm, the activity detected during the Mueller maneuver with Pdi = 0 was very small (Fig. 3) but much larger when Pdi >O (Figs. 2 and 4). Gastric and esophageal pressures were used as indices of Pab and Ppl, respectively, and were measured with balloon-catheter systems in the stomach and midesophagus (balloon length 5 cm connected to polyethylene tubing 100 cm long). The gastric balloon contained 1 ml air whereas the esophageal balloon contained 0.5 ml. Abdominal and rib cage dimensions were measured with two pairs of magnetometers placed on the midline anteriorly and posteriorly, one at a level half way between the angle of Louis and the xiphoid, the other 2 cm above the umbilicus. HP267B transducers were used to measure mouth pressure and Pab relative to atmospheric pressure, PL (mouth pressure relative to Ppl) and Pdi (Pab - Ppl). Mouth pressure was measured with a catheter between the mouthpiece and one of the transducers. All signals were recorded on an eight-channel, directwriting Hewlett Packard oscillograph. The subjects were instructed to perform inspiratory efforts against a closed airway, producing a gradual fall in mouth pressure, while keeping Pab constant. To assist in this, Pab was displayed to them on an oscilloscope. The signal from the EMG electrode in the 2nd intercostal space was used to detect inspiratory intercostal muscle recruitment. The subjects were then asked to perform the Mueller maneuver solely with their intercostal/accessory muscles producing an inspiratory effort accompanied by no change in Pdi. Finally, they were asked to perform the Mueller maneuver solely with the diaphragm (as judged by the absence of an inspiratory intercostal EMG signal). In all maneuvers the subjects were asked to keep their abdominal muscles relaxed and it was assumed that
During the Mueller maneuver in which the subjects attempted to keep Pab constant, inspiratory intercostal electrical activity was always detected. Its amplitude increased as mouth pressure became more negative as shown in Fig. 1. In Fig. 1 Pab remained constant until mouth pressure fell to -8 cmH,O after which it increased. In spite of the increase in Pab intercostal electrical activity continued to increase. A similar maneuver performed by another subject is shown in Fig. 2. There was no abdominal motion indicating that abdominal muscle tone did not change and that shape did not change. Pdi and the amplitude of diaphragmatic electrical activity increased as mouth pressure decreased. In this example EMG activity from the electrodes in the 2nd intercostal space was not detected until mouth pressure fell by about 5-8 cmH&)O. AIthough intercostal electrical activity was detected during all Mueller maneuvers in which APab = 0 and there was no abdominal motion (at least three such tracings were obtained in each subject), the threshold at which intercostal electrical activity was first detected was quite variable. That this is due to insensitivity of surface electrodes is suggested by additional experiments (C. ROUSSOS, J* P. Derenne, L. Delhez, and P. T. Macklem, unpublished observations) in which needle electrodes placed in the external intercostal muscle in the 2nd intercostal space detected strong electrical activity immediately upon initiating a Mueller maneuver with constant abdominal pressure. The rise in transpulmonary pressure during the Mueller maneuver in Fig. 2 is presumably an artifact possibly explicable by an increasing thoracic blood volume causing the heart to impinge upon the esophagus. Fig. 3 shows that during the Mueller maneuver in which Pdi remained close to zero, Pab fell, the abdomen was displaced inward, the rib cage outward and there was electrical activity of the inspiratory intercostals. Some electrical activity was also recorded from the electrode in the 7th intercostal space in the anterior axillary line probably because Pdi increased slightly. During the Mueller maneuver in which the subjects attempted to contract only the diaphragm, (Fig. 4) there was very little electrical activity recorded from the 2nd intercostal space, marked activity recorded from the 7th intercostal space, an increase in Pab, an outward displacement of the abdomen, and a reciprocal inward displacement of the rib cage. At least three similar tracings were obtained in all three subjects. It is unlikely that shape changes of abdomen and rib cage materially influenced these results, because during similar isovolume maneuvers, Konno and Mead found that each compartment behaved with essentially a single degree of freedom (8). DISCUSSION
Interpretation
of results.
We interpret
these results
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
PARTITIONING
OF
RESPIRATORY
PRESSURES
EMG OF INTERCOSTALS I
,
8/ s
I
J’
:
EMG OF DIAPHRAGM
Pab cm Ii*0
FIG. 1. Mueller maneuver performed while trying to maintain abdominal pressure (Pab) constant. Pdi = transdiaphragmatic pressure; Ppl = pleural pressure; Pm = mouth pressure. Second marker between Ppl and Pm tracings. For further explanation see text.
Pdi cm H,O
EMG
IC
EMG :,
_‘(
,,,
,’ :
:,
,, .’ ‘\ .,, ‘(
I ( \
r
” !‘:)I
1 ..’
,i)
i,
:, _I
.-
\
.._- li-~
30
Pab cm H,O
z
m\
oj -10
-20 pdi cmH,O
PL cm ii,0
Pm cm H,O
‘i
’ “_____y__----
25 0 -25I 10
\
. -
,: 1;; -40 1 -, .,,” _ -20 -30 -40 /
FIG. 2. Another example of Mueller maneuver performed at constant Pab. Abdominal (A-P AB) dimensions are also shown. Paper speed 25 mm/s. For further explanation see text.
4
d-k/
i
yv-*L-”
r r-
- CII-
- ._1___^“1_
,
I
A-P Abdominal
as indicating that the Mueller maneuver performed purely with intercostal/accessory muscles produced changes in chest wall configuration that were the opposite to those in which the subjects attempted to produce the maneuver solely with the diaphragm. In the former instance, Pab fell, the abdomen was displaced inward and the rib cage outward. In the latter instance Pab increased, the abdomen was displaced outward, and the rib cage inward. This is in keeping with the notion
--L.
that Ppl drives the relaxed rib cage during the Mueller maneuver and contrary to the hypothesis that Pab drives the relaxed rib cage. If the latter pertained, the Mueller maneuver performed solely with the diaphragm would not have resulted in any change in configuration of the chest wall and Pab would not have changed. We were concerned that the electrical signals we picked up in the second intercostal space might have
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
204
MACKLEM,
GROSS,
GRASSINO,
AND
ROUSSOS
Pm cm Hz0
Pab cm H,O
Pdi cm H,O
PL cm YO
A-P
AB I
A-P
RC I
EMG
IC
EMG Diaphragm FIG. 3. Mueller maneuver performed primarily with intercostal/ accessory muscles. PL = transpulmonary pressure; A-P RC indicates rib cage dimensions; arrows indicate direction of increase in dimensions. Otherwise, symbols are same as in Fig. 2. Second marker between A-P RC and EMG IC tracings.
resulted from contraction of either the platysma or pectoralis major muscles. However, voluntary contraction of the platysma produced no signals in the second intercostal space. Voluntary contraction of the pectoralis major did, but we were unable to detect contraction of this muscle by electrodes placed over its main body during Mueller maneuvers. Thus, the electrical activity recorded from the 2nd intercostal space as shown in Fig. 2 indicates that in order to keep Pab constant during the Mueller maneuver, it was necessary to recruit the inspiratory intercostals. Because there was no displacement of the rib cage, the pressure developed by the inspiratory intercostals did not act on the rib cage. It must therefore have contributed to the fall in Ppl. This occurred even though Pab did not become negative (in fact it became slightly more positive when mouth pressure fell lower than -8 cmH,O) indicating that the primary effects of inspiratory intercostal contraction are not on Pab but on the pleural surface. Conceivably Pab could remain constant with contraction of the abdominal muscles simultaneously with the inspiratory intercostals but this would have resulted in an inward displacement of the abdomen which was not observed. We conclude that Pab does not have to become negative in order for intercostal/accessory muscle con-
FIG. 4. resulting of AB and as in Fig.
Mueller maneuver performed primarily with diaphragm, in an immediate increase in Pab an outward displacement reciprocal inward displacement of RC. Symbols are same 3. Second marker between A-P RC and EMG IC tracings.
traction to lower Ppl. Furthermore, the tracings in Figs. 1 and 2 indicate that both the diaphragm and intercostal/accessory muscles are responsible for the changes in Ppl and that under conditions of quasiisometric diaphragmatic contraction the pressure produced across the lung by the two sets of muscles are developed in parallel. Each muscle group develops the total pleural surface pressure change that is generated so that the total pressure across the rib-cage pathway is not the sum of the pressures developed by the diaphragm and the other inspiratory muscles. Clearly, the analysis proposed by Goldman and Mead (3-5, 7) does not account for our results. This does not mean that the hypothesis is wrong that Pab drives the relaxed rib cage in normal upright man when the diaphragm is the only contracting muscle and the respiratory system is not constrained by a Mueller maneuver to have equal and opposite changes in rib cage and abdominal volume. Indeed, it is established that normal upright subjects breathe along the relaxation line for rib-cage and abdominal motion (8) under circumstances when the respiratory system is by definition, not relaxed. Furthermore, as pointed out earlier, the relationship between Pab and Vab and Vrc respectively, remains identical during quiet breathing in upright man when the system is not relaxed to what it is during relaxation (5). Goldman and Meads hypothesis that the diaphragm displaces the relaxed rib cage and abdomen by changing abdominal pressure (5) provides an eminently reasonable and satisfying explanation of these apparently paradoxical observations. Apart from the particular constraints of the Mueller
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
PARTITIONING
OF
RESPIRATORY
205
PRESSURES
maneuver, our data does not invalidate their hypothesis and indeed we know of no other data that does. What we do not accept is the explicit extension of this hypothesis to the claim that the total pressures developed across the rib cage pathway are the sum of Pdi and Pit. Nor do we accept the implicit concept that when the abdominal muscles remain relaxed, the change in abdominal pressure must be negative in order for intercostal accessory muscle contraction to influence transpulmonary pressure. We suggest that whenever the intercostal/accessory muscles and the diaphragm contract simultaneously, the diaphragm must, first of all develop the same changes in Ppl as the intercostal/accessory muscles do, in order to prevent its fibers from lengthening. It is only when APdi > -APpl so that APab is positive that the diaphragm produces positive external work. If these ideas are correct, the sum of the pressures developed across the rib cage pathway are always less than Pdi + Pit whenever the two muscles contract simultaneously. Because this is so, intercostal/accessory muscle contraction can contribute to transpulmonary pressure swings when APab 3 0. The Pub, Pel diagram. The Goldman and Mead hypothesis that, Pab drives the relaxed rib cage (5) is particularly important in relation to the present work, because it allows measurement of the separate contributions of the diaphragm and intercostal/accessory muscles to the respiratory pressure swings. When the diaphragm is the only muscle contracting, the increase in Pab displaces both the abdomen and the rib cage and the fall in Ppl inflates the lung. If this is performed sufficiently slowly so that flow-resistive pressure losses can be neglected, it follows that the relationship between Pab and Ppl is unique when the diaphragm is the only muscle contracting. When flow-resistive losses can be neglected, lung elastic recoil pressure (Pel) equals - Ppl, so that the relationship between Pel and Pab is also unique. The unique relationship between Pel and Pab when the diaphragm is the only muscle contracting is shown in the four-quadrant diagram of Fig. 5, The upper right quadrant is the unique relationship between Vrc and Pab during relaxation. According to Goldman and Mead’s hypothesis, when the diaphragm is the only muscle contracting, the relationship between Vrc and Pab is identical to relaxation because Pab drives the relaxed rib cage (5). The upper left quadrant is the relationship between Vrc and lung volume, VL. This is also the same during relaxation and when the diaphragm is the only muscle contracting and is similarly unique. The lower left quadrant displays another unique relationship, the relationship betwen VL and Pel. The lower right quadrant completes the diagram and is the Pab, Pel relationship. Because all three other curves are unique, the Pab, Pel curve is also unique. Because all three other curves pertain both during relaxation and during inspirations when the diaphragm is the only muscle contracting, the Pab, Pel curve is also identical under these circumstances. This property of the curve allows its measurement with little difficulty. One only needs to plot Pab vs. Pel
during relaxation at different lung volumes. An example obtained in a normal subject is shown in Fig. 6 in which Pel is plotted above the origin rather than below. During quiet breathing when this subject was seated the relationship between Pel and Pab remained on this line. This suggests that the diaphragm was the only muscle contracting and confirms the earlier findings of Goldman and Mead that the relationship between Vrc and Pab is identical during relaxation to what it is in subjects seated and breathing quietly (5). If when the diaphragm is the only muscle contracting, Pel and Pab move along the relaxation line, then during an inspiration when other muscles are recruited, the relationship will be displaced off the relaxation line. Thus such displacements denote recruitment of other muscles and can be used to quantify the pressures produced by these muscles.
FIG. 5. Four-quadrant diagram illustrating unique relationship between Pab and lung elastic recoil pressure (Pel) during relaxation and during inspirations in which diaphragm is the only respiratory muscle contracting. Upper right quadrant is relationship between rib cage volume (Vrc) and Pab during relaxation. Upper left quadrant is relationship between Vrc and lung volume (VL). Lower left quadrant is static pressure volume curve of lung (PeI = elastic recoil pressure) and lower right quadrant is relationship between Pel and Pab. For further explanation see text.
zo-
Pd cm H,O Pdi =
Pab
cm H,O
FIG. 6. Pab, Pel diagram. Solid line pertains during relaxation and duri .ng inspirations when diaphragm is the only muscle contracting. Dashed line is the locus of points where Pdi = 0. For further explanation see text.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
206 In Fig. 6 the dashed line gives the locus of points where Pdi = 0. FRC is given by point A. An inspiration that proceeds along the relaxation line from A to B indicates that the diaphragm is the only muscle contracting. It produces an increase in Pel given by CB which acts to inflate the lung. It produces an increase in Pab given by AC which acts to displace both the abdomen and the rib cage. The Pdi during such a breath is given by AC + CB. When the intercostal/accessory muscles are the only ones contracting, Pdi remains zero and the relationship between Pab and Pel follows the dashed line to point D (provided that no passive tension is produced in the diaphragm). The change in Pel is given by ED. The pressures required to displace the chest wall are more complex. When the rib cage and abdomen are displaced along their reIaxation characteristics a pressure equal to AC is required to increase the volume of the chest wall by an amount equal to the increase in VL produced by an increase in Pel given by CB. However, when Pdi remains zero, Pab decreases, and the abdomen is sucked inward. Thus the rib cage volume must be increased to a substantially greater extent than when the diaphragm alone contracts. The pressures required to do this are given very approximately by EC. EC is not ?he exact. pressure, because by definition the rib cage and abdomen have departed from their relaxation characteristics. As a result the pressures required to displace the chest wall are somewhat greater than EC (3, 4). A third type of inspiration is illustrated by a breath from A to F. During such a breath Pab remains constant and for reasons stated above, the diaphragm has contracted quasi-isometrically, When this occurs Pdi = -Ppl = AF. Because the diaphragm has performed no external work the work of inflating the lung must have been accomplished by the intercostal/accessory muscles. Thus, they have produced an elastic recoil pressure change given by AF. In order to accomplish all the work they have also increased the volume of the chest wall by the same amount as the lungs. However, as in the previous example the chest wall has departed from its relaxation characteristic. Therefore, the pressures required to displace the chest wall are solnewhat greater than FB. However, they are substantially less than those required when the intercostal/accessory muscles were the only ones contracting. Thus when it contracts quasi-isometrically, the diaphragm prevents the transmission of Ppl to the abdomen. As a result the abdomen is not displaced inward and the increase in Vrc and Pit necessary to accommodate the increase in VL is proportionately reduced. It is apparent that under these circumstances the diaphragm is acting as a fixator and preventing the intercostal/ accessory muscles from performing inefficient work to decrease Vab during inspiration. A fourth type of breath is illustrated in Fig. 7 in which Pel and Pab have followed a pathway described by AG. A breath such as this can be thought -of as occurring in two stages, the first from A to H and the second from H to G. The first stage is along the relaxation line and therefore is accomplished by the diaphragm alone. The diaphragm is acting as an ago-
MACKLEM,
GROSS,
GRASSINO,
AND
ROW&OS
Pel cm H,O
-15
-IO
-5
0 Pab
7. Pab, Pel diagram. 6. For further explanation
FIG.
Fig.
+5
t IO
cmH20
Solid and dashed see text.
lines
are same
as in
nist. During the second stage Pab remains constant and thus, for reasons stated earlier the diaphragm remains at constant length and performs no further work. In this instance it is acting as a fixator. In reality, of course, both stages occur simultaneously. It is apparent that the diaphragm can both fixate and agonize at the same time, but it only performs positive external work to the extent that it contracts agonistitally. In this breath Pab has increased by AJ. This change is sufficient to increase lung volume and Pel by JH. Thus to the extent that Pab increases, the diaphragm contracts agonistically and produces external work. The pressure change AJ has produced external vwork on the chest wall whereas JH has produced external work on the lung. The total Pdi during the breath is given by GJ + AJ. The pressure HG developed by the diaphragm has, however, nut produced any external work and the work of inflating the lung from a Pel of H to a Pel of G must have been accomplished. by the intercostal/accessory muscles. The chest wall Imust have increased by the same volume as the lung and the pressure developed by the intercostal/accessory Imuscles to accomplish this is given approximately by GK. Again the pressure is somewhat greater than this because of departures from the relaxation characteristic. In acting simultaneously as a fixator and as an agonist the diaphragm has produced a fixating pressure given by HG, an agonizing pressure on the lung by HJ, and an agonizing pressure on the chest wall by AJ. It is apparent that an infinite variety of inspirations are possible in which the diaphragm and intercostal/ accessory muscles contract simultaneously. In all conditions except the particular case when A Pab = 0, the diaphragm performs external work. When A Pab > 0 the diaphragm shortens and to the extent that Pab increases it produces positive external work. When A Pab < 0 the diaphragm lengthens and to the extent that Pab decreases it produces negative external work. Furthermore, during all possible inspirations in which the diaphragm and intercostal/accessory muscles contract simultaneously the diaphragm acts in part, as a fixator and assists the intercostal/accessory muscles by
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
PARTITIONING
OF
RESPIRATORY
PRESSURES
207
reducing or eliminating the tendency to suck the abdomen into the chest. Limitations of the Pub, PeL diugram. The equation of motion of the lung states that A PL = ~/CL. V, + RL* v whereas that of the chest wall states that A Pw = l/ Cw *VT + Rw .v where Pw = pressure driving the chest wall, CI, and Cw are the compliance of lung and wall respectively; RL and Rw are the resistance of lung and wall respectively. Thus
ence between esophageal and mouth pressure) against Pab. This will quantify the contribution of each muscle to the lung volume changes but it will not reveal directly how the two muscle groups contribute to the pressure which overcomes the load. For a load placed at the mouth, this pressure is given by the difference between mouth pressure and atmosphere. In the case of an elastic load, the solution is simple. Each muscle contributes work to overcoming an elastic load at any instant in time that is in direct proportion to its agonisAPL VT/CL Jr RL. v ----tic contribution to lung elastic recoil pressure at that (1 ~ APw = VrlCw + Rwvv instant. Again it is possible for the diaphragm to pressure equal to the Under static conditions, and if Pab is the pressure generate a transdiaphragmatic sum of lung elastic recoil pressure and the pressure driving the relaxed wall, Eq. 1 simplifies to across the elastic load at the mouth, yet if Pab dues not A P&l Cw increase, one cannot attribute the work of overcoming -=(2) the load to the diaphragm. A Pab CL In the case of flow-resistive loads the problem is Thus, the slope of the Palo, Pel diagram is the ratio of more complex. Each muscle contributes work to overchest wall compliance to lung compliance. Under dy- coming the resistive load which is in direct proportion namic conditions, the slope of the Pab, PL relationship to the rate of change of its agonistic contribution to is given by Eq- 1, If it is the same as under static lung elastic recoil pressure at that instant. In order to conditions, then, solve for this, one needs to determine the agonistic contributions of each group to lung elastic recoil presVT/CL + RL-T~ ~cw - ------" sure as a function of time, which may be accomplished CL vT/cW + Rw* v if one has tracings of the pressure swings on a strip and chart as well as an LTI(*-Y plot. The rate of change of the muscle contribution to elastic recoil pressure can then RL-CL = RwXw be determined graphically. Thus, the Pab, PL relationship when the diaphragm is The influence of abdominal muscle contraction. To the only muscle contracting, is unique under dynamic this point our analysis of the diaphragmatic and interconditions, if and only if the time constant of the lung costal/accessory miscie contributions to the respiratory equals that of the chest wall. Strictly speaking the pressure swings has assumed that the abdominal musPab, PL diagram can only be used to partition the cles remain relaxed. Grimby, Goldman, and Mead have respiratory pressure swings under static conditions. shown that under particular circumstances the abdomHowever, in normal subjects the time constants of lung inal muscles have an inspiratory action and that they and chest wall may be sufficiently similar so that perform a useful function by displacing the diaphragm during quiet breathing and moderate increases in ven- into the chest so that it becomes more efficient as a tilation, the diagram may still prove useful. pressure generator (7). Clearly, if our approach is to A crucial feature of our analysis is the assumption have much practical value, abdominal muscle. contracthat an isometrically contracting diaphragm can per- tion will have to be taken into account. form no external work. This may not necessarily be the When the abdominal muscles contract, Vab decreases case. If the pressure difference across the diaphragm and the relationship between Vrc and VL is no longer can be approximated by the LaPlace Law so that Pdi = unique. Because it is no longer unique, the relationship T/R where T is the tension in the diaphragm and R, its between Pab and Pel is also no longer unique. Theoretradius of curvature, an increase in Pdi when A Pab = 0 ically the solution to this problem is straightforward. might be accompanied by an increase R with a propor- For any given value of Vab there is still a unique tionately greater increase in T. Thus, it is conceivable relationship between Vrc and VL and therefore between that diaphragmatic flattening during an isometric con- Pab and Pel. By measuring the relationships between traction might perform positive external work. How- Vrc and Vab at different lung volumes as described by ever, to the extent that the diaphragm produces no net Konno and Mead (S), a series of isopleths relating Vrc force to displace the chest wall when A Pab = 0, it to VL at different constant values of Vab may be cannot have performed external work on the chest wall. constructed. Using the four-quadrant diagram in Fig. Under these circumstances it is difficult to see how it 5, these yield a series of Pab, Pel isopleths fur constant can perform external work on the lung. Nevertheless, values of Vab. Only one needs to monitor Vab (which one of the tasks of future research in this field will be can be accomplished by measuring its AP dimensions to define the exact changes in Pel and Pab when with magnetometers) to determine which isopleth perdiaphragmatic contraction performs no external work. tains at that instant and by comparing the measured Loading of the respiratory system. Under circumPab and Pel with that predicted from the appropriate stances such as resistive or elastic loading of the respi- isopleth, the diaphragmatic and intercostal/accessory ratory system at the mouth one can still-partition t-he muscle contributions can be determined as described respiratory pressure swings by plotting PT, (the differabove. Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
208
MACKLEM,
Of course this would be exceedingly tedious to do manually. We are currently attempting to do it electritally by describing the isopleths by form fitting equations, monitoring Vab as well as Pab and PL continuously, and determining digitally, on-line the instantaneous displacements from the appropriate isopleth and thus the diaphragmatic and intercostal/accessory muscle contributions to the respiratory pressure swings for
GROSS,
any degree of abdominal successful, the approach be more practical. This search Council. Received
research Council
GRASSING,
ROUSSOS
muscle contraction. If we are described in this paper should
was supported by a grant of Canada. P. T. Macklem
for publication
AND
10 February
from the Medical Reis an Associate of the
1977.
REFERENCES E., G. SANT’AMBROGIO, AND E. AGOSTONL Effect of diaphragm activity or paralysis on distribution of pleural pressure. J. Appl. Physiol. 37: 311-315, 1974. DOSMAN, J., A. GRASSINO, P. T. MACKLEM, AND L. ENGEL. Factors influencing the esophageal pressure gradient in upright man (Abstract), Physiologist 18: 194, 1975. GOLDMAN, M. D. Mechanical coupling of the diaphragm and rib cage. In: Loaded Breathing, edited by L. D. Pengelly, A. S. Rebuck, and E. J. M. Campbell. Edinburgh and London: Churchill Livingstone, 1974, p. 50-63. GOLDMAN, M. D., G. GRIMBY, AND 3. MEAD. Mechanical work of breathing derived from rib cage and abdominal V-P partitioning J. Appl. Physiol. 4: 752-563, 1976. GOLDMAN, M. D., AND J. MEAD. Mechanical interaction between the diaphragm and rib cage. J. AppZ. PhysioZ. 35: 197-204, 1973.
1. D’ANGELO,
2.
3.
4.
5.
6. GRASSINO, A. Influence of chest wall configuration on the static and dynamic characteristics of the contracting diaphragm. In: Loaded Breathing, edited by L. D. Pengelly, A. S. Rebuck, and E. J. M. Campbell. Edinburgh and London: Churchill Livingstone, 1974, p. 64-72. 7. GRIMBY, G., M. GOLDMAN, AND J. MEAD. Respiratory muscle action inferred from rib cage and abdominal V-P partitioning. J. Appl. Physiol. 41: 739-751, 1976. 8. KONNO, K., AND J. MEAD. Measurement of the separate volume changes of the rib cage and abdomen during breathing. J. Appl. Physiol. 22: 407-422, 1967. 9. MEAD, J., N. PETERSON, G. GRIMBY, AND J. MEAD. Pulmonary ventilation measured from body surface movements. Science 156: 1383-1384, 1967.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
of inspiratory pressure and intercostal/accessory
swings between muscles
PETER T. MACKLEM, D. GROSS, A. GRASSINO, AND C. ROUSSOS Meakins-Christie Laboratories, McGill University Clinic, Royal Victoria Hospital, Montreal H3A 2B4 Quebec, Canada
D. GROSS, A. GRASSINO, AND C. pressure swings between diaphragm and intercostal/accessory muscles. J. Appl. Physiol. : Respirat. Environ. Exercise Physiol. 44(2): 200-208, 1978. -We tested the hypothesis that the inspiratory pressure swings across the rib-cage pathway are the sum of transdiaphragmatic pressure (Pdi) and the pressures developed by the intercostal/accessory muscles (Pit). If correct, Pit can only contribute to lowering pleural pressure (Ppl), to the extent that it lowers abdominal pressure (Pab). To test this we measured Pab and Ppl during Mueller maneuvers. Contraction of external intercostal muscles was inferred from surface EMG recordings. We found evidence of external intercostal contraction during Mueller maneuvers in which APab = 0. Because there was no outward displacement of the rib cage, Pit must have contributed to APpl, as did Pdi. Under these conditions the total pressure developed by the inspiratory muscles across the rib-cage pathway was less than Pdi + Pit. Therefore, we rejected the hypothesis. A plot of Pab vs. Ppl during relaxation allows partitioning of the diaphragmatic and intercostal/accessory muscle contributions to inspiratory pressure swings. The analysis indicates that the diaphragm can act both as a fixator, preventing transmission of Ppl to the abdomen and as an agonist. When abdominal muscles remain relaxed it only assumes the latter role to the extent that Pab increases, MACKLEM, PETER ROUSSOS. Partitioning
T.,
of inspiratory
respiratory muscles; respiratory mechanics; EM G; fixators; agonists; pleural pressure; abdomi .nal pressure
AN UNDERSTANDING OF the mechanical link between the diaphragm and rib cage is necessary in order to determine the relative contributions of the diaphragm and intercostal/accessory muscles to respiratory pressure swings. The situation is relatively simple when either one or the other is the only muscle contracting. When the intercostal/accessory muscles contract alone, transdiaphragmatic pressure (Pdi) remains zero at least initially, so that abdominal pressure (Pab) remains equal to pleural pressure (Ppl) and falls as inspiration proceeds. As a result the abdomen is displaced inward and upward into the thorax causing the diaphragm to lengthen. Eventually the diaphragm will develop passive tension and as a result, Pdi will rise. In this paper we are concerned with events before passive tension causes a rise in Pdi. Goldman and Mead have suggested that in upright humans when the diaphragm is the only muscle cun-
tracting, the increase in Pab that results, drives the relaxed rib cage (5). Indeed they showed that the relationship between Pab and rib cage volume was identical during quiet breathing to the relationship obtained during relaxation and during abdominal compression in relaxed subjects, whereas the relationship between Ppl and rib-cage volume was different during the three conditions. Thus, they suggested that when the diaphragm is the only muscle contracting, it shortens, producing an increase in Pab which displaces both the relaxed abdomen and rib cage, and a fall in Ppl which inflates the lungs. If, in one instance the diaphragm lengthens, whereas in the other it shortens, it must be possible to inspire in such a way that the diaphragm contracts isometrically. Under these circumstances the diaphragm has presumably performed no external work, and all the work performed and the pressures required to inflate the lung and displace the chest wall must have been accomplished by the intercostal/accessory muscles. Under what circumstances does the diaphragm contract isometrically? If the abdominal muscles remain relaxed then abdominal displacement and/or Pab can be used as an approximate indicator of diaphragmatic length. Grassino has shown that when the diaphragm contracts isometrically, changes in abdominal dimensions are small (6). Therefore, as a useftil approximation we assume that when abdominal muscles remain relaxed and APab = 0 during an inspiration, the diaphragm has contracted quasi-isometrically and performed no external work. The condition of abdominal relaxation pertains to the rest of our theoretical development and is assumed unless specifically stated otherwise. Thus when the diaphragm contracts quasi-isometrically and APab = 0, APdi = -APpl. When the diaphragm produces no external work the pressure changes generated by the intercostal/accessory muscles (APic) must equal the sum of the pressures required to displace the lung (APL) and chest wall (APw) so that APic = APw + APL. When mouth pressure remains atmospheric APL = -APpl = APdi. Thus, under these circumstances, APpl has been achieved by both the diaphragm and intercostallaccessory muscles and this pressure change is shared by both. However, only the intercostal/accessory muscles shorten and the work of inflating the lung is performed by these muscles, and not by the diaphragm. This analysis departs in important ways from that
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
PARTITIONING
OF
RESPIRATORY
201
PRESSURES
developed by Goldman et al. (4). They suggested that during inspirations when APab = 0, the diaphragm performs the work of inflating the lung, and the intercostal/accessory muscles, the work of displacing the chest wall, in contrast to our approach in which we attribute all the work to the intercostal/accessory muscles and none to the diaphragm. Although their approach is appealing, it fails to explain how an isometrically contracting muscle can perform external work. In dealing with this issue, they state that “* . . it would appear from our analysis . that) the diaphragm can perform mechanical work without shortening substantially. We note here that ‘substantially’ is a key word (we do not propose that a truly isometric contraction can perform external mechanical work) . . . .” It can certainly be argued (as they do) that an inspiration performed without abdominal displacement (APab = 0, when the abdominal muscles remain relaxed) does not represent a truly isometric contraction. We have assumed that it does because there is substantial evidence that abdominal displacements are an approximate indicator of diaphragmatic length (6). It is for this reason that we assume that diaphragmatic contraction is quasi-isometric when APab = 0 and that to a useful approximation the diaphragm has performed no external work. However, the analysis of Goldman et al. (4) attributes positive external work to the diaphragm whenever Pdi > 0. But it is certainly possible for an inspiration to occur with a substantial fall in Pab, an inward abdominal displacement, and a small increase in Pdi. During such an inspiration, the diaphragm must be lengthening as it contracts. Thus, there is an even greater paradox in their analysis - the diaphragm is interpreted as doing inspiratory work on the lung as it is lengthening, In the model of Goldman et al., the diaphragm and intercostal/accessory muscles are arranged serially in terms of the pressures developed across the pathway from the mouth to the rib cage (4, 7). Starting at the mouth, and proceeding in order, this pathway consists of the lungs, the diaphragm, the intercostal/accessory muscles and the rib cage. Ppl is the pressure between the lungs and diaphragm, and Pab is the pressure between the diaphragm and intercostal/accessory muscles. The total pressure change developed across this pathway is the sum of the pressures developed by the two muscle groups, i.e., Pdi + Pit. When mouth pressure is atmospheric and APab = 0, APdi = -APpl and, as explained above, they attribute all the change in Ppl to the diaphragm and all the pressures necessary to displace the chest wall to the intercostal/accessory muscles. If this view is correct, during such a breath the intercostal/accessory muscles have not contributed to the transpulmonary pressure swings. Indeed, it is implicit in their analysis that the primary effect of intercostal/accessory muscle contraction is a lowering of Pab and that this change in pressure must be transmitted through the diaphragm in order to influence pleural pressure. This prediction of the Goldman and Mead model is intuitively unlikely. On anatomical consideral
l
tions alone, the inspiratory intercostals are situated closer to the pleural surface than they are to the abdomen. It seems unlikely that they must lower Pab before they can lower Ppl. This is supported by experiments in animals, in which intercostal muscle contraction causes greater pressure swings over the costal pleural surface than over the diaphragm (l), and in humans they cause greater pressure swings in the upper esophagus than in the lower (2). These data indicate that intercostal muscle contraction produces pressure changes primarily at sites close to where these muscles are situated, i.e., the pleural surface. We suggest that under the particular circumstances when the diaphragm contracts quasi-isometrically the pressures developed across the lung are shared by both the diaphragm and intercostal/accessory muscles and that therefore the sum of APdi and APic are greater than the total respiratory pressure swings developed across the rib-cage pathway. We further predict that intercostal/accessory muscle contraction can contribute to pleural pressure changes without Pab changing in a negative direction. This prediction can be tested experimentally. It is one of the purposes of this paper to present the results of this test. We show that APpl during Mueller maneuvers when APab = 0 is shared by both diaphragm and intercostal/accessory muscles in keeping with our analysis. We then develop the analysis further in order to partition the diaphragmatic and intercostal/accessory muscle contributions to respiratory pressure swings and conclude that the diaphragm can function either as an agonist or a fixator. When acting agonistically, the diaphragm shortens and the resulting increase in Pab drives the rib cage. When fixating, the diaphragm prevents the transmission of Ppl to the abdomen and thus prevents paradoxical movement of this compartment during inspiration. However, the work of inflating both the lung and the chest wall under these circumstances is performed entirely by the intercostal/accessory muscles. RATIONALE
AND
METHODS
During Mueller maneuvers the volume of the respiratory system and thus the chest wall (VW) remains almost constant. Thus, any changes in rib cage volume (AVrc) that occur must be equal and opposite to changes in abdominal volume (AVab) because AVw = AVrc + AVab = 0. If the performance of the Mueller maneuver was accomplished with diaphragm alone (all other muscles remaining relaxed) and if Pab drives both the relaxed rib cage and the abdomen, there could be no change in Pab. If Pab were to change, AVrc and AVab would have the same sign and reciprocal changes are impossible. Thus, according to Goldman and Mead’s approach, when the diaphragm is the only muscle contracting during the Mueller maneuver, Pab cannot change. On the other hand, if the relaxed rib cage is driven by Ppl, the Mueller maneuver performed by the diaphragm alone would again result in no change in VW, but there would be equal and opposite changes in Vrc
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
202
MACKLEM,
and Vab because the decrease in Ppl with diaphragmatic contraction would deflate the rib cage, whereas the increase in Pab would inflate the abdomen. Thus AVab = APab Cab = - AVrc = - APpl Crc l
they did not recruit the maneuvers.
GROSS,
GRASSINO,
other expiratory
AND
ROUSSOS
muscles during
RESULTS
l
where Cab and Crc are the compliance of abdomen and rib cage respectively. To prevent the inward displacement of the rib cage it would be necessary to recruit the intercostal/accessory muscles. If this were accomplished just to the extent that AVrc = 0, then AVab = 0 = APab. In this instance, the Mueller maneuver performed at constant Pab would result in intercostal/ accessory muscle recruitment whereas if Goldman and Mead’s approach is correct such recruitment would not occur. Furthermore, because AVrc = 0 the pressure change across the rib cage must be zero and any intercostal/accessory muscle contraction that occurred must have lowered Ppl without lowering Pab. We studied three normal subjects who were highly trained to perform respiratory maneuvers. We measured inspiratory intercostal and diaphragmatic EMG with surface electrodes placed in the 2nd intercostal space parasternally and in the 6th and 7th intercostal space in the anterior axillary line, respectively. Although the latter electrodes presumably detect electrical activity originating from muscles other than the diaphragm, the activity detected during the Mueller maneuver with Pdi = 0 was very small (Fig. 3) but much larger when Pdi >O (Figs. 2 and 4). Gastric and esophageal pressures were used as indices of Pab and Ppl, respectively, and were measured with balloon-catheter systems in the stomach and midesophagus (balloon length 5 cm connected to polyethylene tubing 100 cm long). The gastric balloon contained 1 ml air whereas the esophageal balloon contained 0.5 ml. Abdominal and rib cage dimensions were measured with two pairs of magnetometers placed on the midline anteriorly and posteriorly, one at a level half way between the angle of Louis and the xiphoid, the other 2 cm above the umbilicus. HP267B transducers were used to measure mouth pressure and Pab relative to atmospheric pressure, PL (mouth pressure relative to Ppl) and Pdi (Pab - Ppl). Mouth pressure was measured with a catheter between the mouthpiece and one of the transducers. All signals were recorded on an eight-channel, directwriting Hewlett Packard oscillograph. The subjects were instructed to perform inspiratory efforts against a closed airway, producing a gradual fall in mouth pressure, while keeping Pab constant. To assist in this, Pab was displayed to them on an oscilloscope. The signal from the EMG electrode in the 2nd intercostal space was used to detect inspiratory intercostal muscle recruitment. The subjects were then asked to perform the Mueller maneuver solely with their intercostal/accessory muscles producing an inspiratory effort accompanied by no change in Pdi. Finally, they were asked to perform the Mueller maneuver solely with the diaphragm (as judged by the absence of an inspiratory intercostal EMG signal). In all maneuvers the subjects were asked to keep their abdominal muscles relaxed and it was assumed that
During the Mueller maneuver in which the subjects attempted to keep Pab constant, inspiratory intercostal electrical activity was always detected. Its amplitude increased as mouth pressure became more negative as shown in Fig. 1. In Fig. 1 Pab remained constant until mouth pressure fell to -8 cmH,O after which it increased. In spite of the increase in Pab intercostal electrical activity continued to increase. A similar maneuver performed by another subject is shown in Fig. 2. There was no abdominal motion indicating that abdominal muscle tone did not change and that shape did not change. Pdi and the amplitude of diaphragmatic electrical activity increased as mouth pressure decreased. In this example EMG activity from the electrodes in the 2nd intercostal space was not detected until mouth pressure fell by about 5-8 cmH&)O. AIthough intercostal electrical activity was detected during all Mueller maneuvers in which APab = 0 and there was no abdominal motion (at least three such tracings were obtained in each subject), the threshold at which intercostal electrical activity was first detected was quite variable. That this is due to insensitivity of surface electrodes is suggested by additional experiments (C. ROUSSOS, J* P. Derenne, L. Delhez, and P. T. Macklem, unpublished observations) in which needle electrodes placed in the external intercostal muscle in the 2nd intercostal space detected strong electrical activity immediately upon initiating a Mueller maneuver with constant abdominal pressure. The rise in transpulmonary pressure during the Mueller maneuver in Fig. 2 is presumably an artifact possibly explicable by an increasing thoracic blood volume causing the heart to impinge upon the esophagus. Fig. 3 shows that during the Mueller maneuver in which Pdi remained close to zero, Pab fell, the abdomen was displaced inward, the rib cage outward and there was electrical activity of the inspiratory intercostals. Some electrical activity was also recorded from the electrode in the 7th intercostal space in the anterior axillary line probably because Pdi increased slightly. During the Mueller maneuver in which the subjects attempted to contract only the diaphragm, (Fig. 4) there was very little electrical activity recorded from the 2nd intercostal space, marked activity recorded from the 7th intercostal space, an increase in Pab, an outward displacement of the abdomen, and a reciprocal inward displacement of the rib cage. At least three similar tracings were obtained in all three subjects. It is unlikely that shape changes of abdomen and rib cage materially influenced these results, because during similar isovolume maneuvers, Konno and Mead found that each compartment behaved with essentially a single degree of freedom (8). DISCUSSION
Interpretation
of results.
We interpret
these results
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
PARTITIONING
OF
RESPIRATORY
PRESSURES
EMG OF INTERCOSTALS I
,
8/ s
I
J’
:
EMG OF DIAPHRAGM
Pab cm Ii*0
FIG. 1. Mueller maneuver performed while trying to maintain abdominal pressure (Pab) constant. Pdi = transdiaphragmatic pressure; Ppl = pleural pressure; Pm = mouth pressure. Second marker between Ppl and Pm tracings. For further explanation see text.
Pdi cm H,O
EMG
IC
EMG :,
_‘(
,,,
,’ :
:,
,, .’ ‘\ .,, ‘(
I ( \
r
” !‘:)I
1 ..’
,i)
i,
:, _I
.-
\
.._- li-~
30
Pab cm H,O
z
m\
oj -10
-20 pdi cmH,O
PL cm ii,0
Pm cm H,O
‘i
’ “_____y__----
25 0 -25I 10
\
. -
,: 1;; -40 1 -, .,,” _ -20 -30 -40 /
FIG. 2. Another example of Mueller maneuver performed at constant Pab. Abdominal (A-P AB) dimensions are also shown. Paper speed 25 mm/s. For further explanation see text.
4
d-k/
i
yv-*L-”
r r-
- CII-
- ._1___^“1_
,
I
A-P Abdominal
as indicating that the Mueller maneuver performed purely with intercostal/accessory muscles produced changes in chest wall configuration that were the opposite to those in which the subjects attempted to produce the maneuver solely with the diaphragm. In the former instance, Pab fell, the abdomen was displaced inward and the rib cage outward. In the latter instance Pab increased, the abdomen was displaced outward, and the rib cage inward. This is in keeping with the notion
--L.
that Ppl drives the relaxed rib cage during the Mueller maneuver and contrary to the hypothesis that Pab drives the relaxed rib cage. If the latter pertained, the Mueller maneuver performed solely with the diaphragm would not have resulted in any change in configuration of the chest wall and Pab would not have changed. We were concerned that the electrical signals we picked up in the second intercostal space might have
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
204
MACKLEM,
GROSS,
GRASSINO,
AND
ROUSSOS
Pm cm Hz0
Pab cm H,O
Pdi cm H,O
PL cm YO
A-P
AB I
A-P
RC I
EMG
IC
EMG Diaphragm FIG. 3. Mueller maneuver performed primarily with intercostal/ accessory muscles. PL = transpulmonary pressure; A-P RC indicates rib cage dimensions; arrows indicate direction of increase in dimensions. Otherwise, symbols are same as in Fig. 2. Second marker between A-P RC and EMG IC tracings.
resulted from contraction of either the platysma or pectoralis major muscles. However, voluntary contraction of the platysma produced no signals in the second intercostal space. Voluntary contraction of the pectoralis major did, but we were unable to detect contraction of this muscle by electrodes placed over its main body during Mueller maneuvers. Thus, the electrical activity recorded from the 2nd intercostal space as shown in Fig. 2 indicates that in order to keep Pab constant during the Mueller maneuver, it was necessary to recruit the inspiratory intercostals. Because there was no displacement of the rib cage, the pressure developed by the inspiratory intercostals did not act on the rib cage. It must therefore have contributed to the fall in Ppl. This occurred even though Pab did not become negative (in fact it became slightly more positive when mouth pressure fell lower than -8 cmH,O) indicating that the primary effects of inspiratory intercostal contraction are not on Pab but on the pleural surface. Conceivably Pab could remain constant with contraction of the abdominal muscles simultaneously with the inspiratory intercostals but this would have resulted in an inward displacement of the abdomen which was not observed. We conclude that Pab does not have to become negative in order for intercostal/accessory muscle con-
FIG. 4. resulting of AB and as in Fig.
Mueller maneuver performed primarily with diaphragm, in an immediate increase in Pab an outward displacement reciprocal inward displacement of RC. Symbols are same 3. Second marker between A-P RC and EMG IC tracings.
traction to lower Ppl. Furthermore, the tracings in Figs. 1 and 2 indicate that both the diaphragm and intercostal/accessory muscles are responsible for the changes in Ppl and that under conditions of quasiisometric diaphragmatic contraction the pressure produced across the lung by the two sets of muscles are developed in parallel. Each muscle group develops the total pleural surface pressure change that is generated so that the total pressure across the rib-cage pathway is not the sum of the pressures developed by the diaphragm and the other inspiratory muscles. Clearly, the analysis proposed by Goldman and Mead (3-5, 7) does not account for our results. This does not mean that the hypothesis is wrong that Pab drives the relaxed rib cage in normal upright man when the diaphragm is the only contracting muscle and the respiratory system is not constrained by a Mueller maneuver to have equal and opposite changes in rib cage and abdominal volume. Indeed, it is established that normal upright subjects breathe along the relaxation line for rib-cage and abdominal motion (8) under circumstances when the respiratory system is by definition, not relaxed. Furthermore, as pointed out earlier, the relationship between Pab and Vab and Vrc respectively, remains identical during quiet breathing in upright man when the system is not relaxed to what it is during relaxation (5). Goldman and Meads hypothesis that the diaphragm displaces the relaxed rib cage and abdomen by changing abdominal pressure (5) provides an eminently reasonable and satisfying explanation of these apparently paradoxical observations. Apart from the particular constraints of the Mueller
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
PARTITIONING
OF
RESPIRATORY
205
PRESSURES
maneuver, our data does not invalidate their hypothesis and indeed we know of no other data that does. What we do not accept is the explicit extension of this hypothesis to the claim that the total pressures developed across the rib cage pathway are the sum of Pdi and Pit. Nor do we accept the implicit concept that when the abdominal muscles remain relaxed, the change in abdominal pressure must be negative in order for intercostal accessory muscle contraction to influence transpulmonary pressure. We suggest that whenever the intercostal/accessory muscles and the diaphragm contract simultaneously, the diaphragm must, first of all develop the same changes in Ppl as the intercostal/accessory muscles do, in order to prevent its fibers from lengthening. It is only when APdi > -APpl so that APab is positive that the diaphragm produces positive external work. If these ideas are correct, the sum of the pressures developed across the rib cage pathway are always less than Pdi + Pit whenever the two muscles contract simultaneously. Because this is so, intercostal/accessory muscle contraction can contribute to transpulmonary pressure swings when APab 3 0. The Pub, Pel diagram. The Goldman and Mead hypothesis that, Pab drives the relaxed rib cage (5) is particularly important in relation to the present work, because it allows measurement of the separate contributions of the diaphragm and intercostal/accessory muscles to the respiratory pressure swings. When the diaphragm is the only muscle contracting, the increase in Pab displaces both the abdomen and the rib cage and the fall in Ppl inflates the lung. If this is performed sufficiently slowly so that flow-resistive pressure losses can be neglected, it follows that the relationship between Pab and Ppl is unique when the diaphragm is the only muscle contracting. When flow-resistive losses can be neglected, lung elastic recoil pressure (Pel) equals - Ppl, so that the relationship between Pel and Pab is also unique. The unique relationship between Pel and Pab when the diaphragm is the only muscle contracting is shown in the four-quadrant diagram of Fig. 5, The upper right quadrant is the unique relationship between Vrc and Pab during relaxation. According to Goldman and Mead’s hypothesis, when the diaphragm is the only muscle contracting, the relationship between Vrc and Pab is identical to relaxation because Pab drives the relaxed rib cage (5). The upper left quadrant is the relationship between Vrc and lung volume, VL. This is also the same during relaxation and when the diaphragm is the only muscle contracting and is similarly unique. The lower left quadrant displays another unique relationship, the relationship betwen VL and Pel. The lower right quadrant completes the diagram and is the Pab, Pel relationship. Because all three other curves are unique, the Pab, Pel curve is also unique. Because all three other curves pertain both during relaxation and during inspirations when the diaphragm is the only muscle contracting, the Pab, Pel curve is also identical under these circumstances. This property of the curve allows its measurement with little difficulty. One only needs to plot Pab vs. Pel
during relaxation at different lung volumes. An example obtained in a normal subject is shown in Fig. 6 in which Pel is plotted above the origin rather than below. During quiet breathing when this subject was seated the relationship between Pel and Pab remained on this line. This suggests that the diaphragm was the only muscle contracting and confirms the earlier findings of Goldman and Mead that the relationship between Vrc and Pab is identical during relaxation to what it is in subjects seated and breathing quietly (5). If when the diaphragm is the only muscle contracting, Pel and Pab move along the relaxation line, then during an inspiration when other muscles are recruited, the relationship will be displaced off the relaxation line. Thus such displacements denote recruitment of other muscles and can be used to quantify the pressures produced by these muscles.
FIG. 5. Four-quadrant diagram illustrating unique relationship between Pab and lung elastic recoil pressure (Pel) during relaxation and during inspirations in which diaphragm is the only respiratory muscle contracting. Upper right quadrant is relationship between rib cage volume (Vrc) and Pab during relaxation. Upper left quadrant is relationship between Vrc and lung volume (VL). Lower left quadrant is static pressure volume curve of lung (PeI = elastic recoil pressure) and lower right quadrant is relationship between Pel and Pab. For further explanation see text.
zo-
Pd cm H,O Pdi =
Pab
cm H,O
FIG. 6. Pab, Pel diagram. Solid line pertains during relaxation and duri .ng inspirations when diaphragm is the only muscle contracting. Dashed line is the locus of points where Pdi = 0. For further explanation see text.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
206 In Fig. 6 the dashed line gives the locus of points where Pdi = 0. FRC is given by point A. An inspiration that proceeds along the relaxation line from A to B indicates that the diaphragm is the only muscle contracting. It produces an increase in Pel given by CB which acts to inflate the lung. It produces an increase in Pab given by AC which acts to displace both the abdomen and the rib cage. The Pdi during such a breath is given by AC + CB. When the intercostal/accessory muscles are the only ones contracting, Pdi remains zero and the relationship between Pab and Pel follows the dashed line to point D (provided that no passive tension is produced in the diaphragm). The change in Pel is given by ED. The pressures required to displace the chest wall are more complex. When the rib cage and abdomen are displaced along their reIaxation characteristics a pressure equal to AC is required to increase the volume of the chest wall by an amount equal to the increase in VL produced by an increase in Pel given by CB. However, when Pdi remains zero, Pab decreases, and the abdomen is sucked inward. Thus the rib cage volume must be increased to a substantially greater extent than when the diaphragm alone contracts. The pressures required to do this are given very approximately by EC. EC is not ?he exact. pressure, because by definition the rib cage and abdomen have departed from their relaxation characteristics. As a result the pressures required to displace the chest wall are somewhat greater than EC (3, 4). A third type of inspiration is illustrated by a breath from A to F. During such a breath Pab remains constant and for reasons stated above, the diaphragm has contracted quasi-isometrically, When this occurs Pdi = -Ppl = AF. Because the diaphragm has performed no external work the work of inflating the lung must have been accomplished by the intercostal/accessory muscles. Thus, they have produced an elastic recoil pressure change given by AF. In order to accomplish all the work they have also increased the volume of the chest wall by the same amount as the lungs. However, as in the previous example the chest wall has departed from its relaxation characteristic. Therefore, the pressures required to displace the chest wall are solnewhat greater than FB. However, they are substantially less than those required when the intercostal/accessory muscles were the only ones contracting. Thus when it contracts quasi-isometrically, the diaphragm prevents the transmission of Ppl to the abdomen. As a result the abdomen is not displaced inward and the increase in Vrc and Pit necessary to accommodate the increase in VL is proportionately reduced. It is apparent that under these circumstances the diaphragm is acting as a fixator and preventing the intercostal/ accessory muscles from performing inefficient work to decrease Vab during inspiration. A fourth type of breath is illustrated in Fig. 7 in which Pel and Pab have followed a pathway described by AG. A breath such as this can be thought -of as occurring in two stages, the first from A to H and the second from H to G. The first stage is along the relaxation line and therefore is accomplished by the diaphragm alone. The diaphragm is acting as an ago-
MACKLEM,
GROSS,
GRASSINO,
AND
ROW&OS
Pel cm H,O
-15
-IO
-5
0 Pab
7. Pab, Pel diagram. 6. For further explanation
FIG.
Fig.
+5
t IO
cmH20
Solid and dashed see text.
lines
are same
as in
nist. During the second stage Pab remains constant and thus, for reasons stated earlier the diaphragm remains at constant length and performs no further work. In this instance it is acting as a fixator. In reality, of course, both stages occur simultaneously. It is apparent that the diaphragm can both fixate and agonize at the same time, but it only performs positive external work to the extent that it contracts agonistitally. In this breath Pab has increased by AJ. This change is sufficient to increase lung volume and Pel by JH. Thus to the extent that Pab increases, the diaphragm contracts agonistically and produces external work. The pressure change AJ has produced external vwork on the chest wall whereas JH has produced external work on the lung. The total Pdi during the breath is given by GJ + AJ. The pressure HG developed by the diaphragm has, however, nut produced any external work and the work of inflating the lung from a Pel of H to a Pel of G must have been accomplished. by the intercostal/accessory muscles. The chest wall Imust have increased by the same volume as the lung and the pressure developed by the intercostal/accessory Imuscles to accomplish this is given approximately by GK. Again the pressure is somewhat greater than this because of departures from the relaxation characteristic. In acting simultaneously as a fixator and as an agonist the diaphragm has produced a fixating pressure given by HG, an agonizing pressure on the lung by HJ, and an agonizing pressure on the chest wall by AJ. It is apparent that an infinite variety of inspirations are possible in which the diaphragm and intercostal/ accessory muscles contract simultaneously. In all conditions except the particular case when A Pab = 0, the diaphragm performs external work. When A Pab > 0 the diaphragm shortens and to the extent that Pab increases it produces positive external work. When A Pab < 0 the diaphragm lengthens and to the extent that Pab decreases it produces negative external work. Furthermore, during all possible inspirations in which the diaphragm and intercostal/accessory muscles contract simultaneously the diaphragm acts in part, as a fixator and assists the intercostal/accessory muscles by
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
PARTITIONING
OF
RESPIRATORY
PRESSURES
207
reducing or eliminating the tendency to suck the abdomen into the chest. Limitations of the Pub, PeL diugram. The equation of motion of the lung states that A PL = ~/CL. V, + RL* v whereas that of the chest wall states that A Pw = l/ Cw *VT + Rw .v where Pw = pressure driving the chest wall, CI, and Cw are the compliance of lung and wall respectively; RL and Rw are the resistance of lung and wall respectively. Thus
ence between esophageal and mouth pressure) against Pab. This will quantify the contribution of each muscle to the lung volume changes but it will not reveal directly how the two muscle groups contribute to the pressure which overcomes the load. For a load placed at the mouth, this pressure is given by the difference between mouth pressure and atmosphere. In the case of an elastic load, the solution is simple. Each muscle contributes work to overcoming an elastic load at any instant in time that is in direct proportion to its agonisAPL VT/CL Jr RL. v ----tic contribution to lung elastic recoil pressure at that (1 ~ APw = VrlCw + Rwvv instant. Again it is possible for the diaphragm to pressure equal to the Under static conditions, and if Pab is the pressure generate a transdiaphragmatic sum of lung elastic recoil pressure and the pressure driving the relaxed wall, Eq. 1 simplifies to across the elastic load at the mouth, yet if Pab dues not A P&l Cw increase, one cannot attribute the work of overcoming -=(2) the load to the diaphragm. A Pab CL In the case of flow-resistive loads the problem is Thus, the slope of the Palo, Pel diagram is the ratio of more complex. Each muscle contributes work to overchest wall compliance to lung compliance. Under dy- coming the resistive load which is in direct proportion namic conditions, the slope of the Pab, PL relationship to the rate of change of its agonistic contribution to is given by Eq- 1, If it is the same as under static lung elastic recoil pressure at that instant. In order to conditions, then, solve for this, one needs to determine the agonistic contributions of each group to lung elastic recoil presVT/CL + RL-T~ ~cw - ------" sure as a function of time, which may be accomplished CL vT/cW + Rw* v if one has tracings of the pressure swings on a strip and chart as well as an LTI(*-Y plot. The rate of change of the muscle contribution to elastic recoil pressure can then RL-CL = RwXw be determined graphically. Thus, the Pab, PL relationship when the diaphragm is The influence of abdominal muscle contraction. To the only muscle contracting, is unique under dynamic this point our analysis of the diaphragmatic and interconditions, if and only if the time constant of the lung costal/accessory miscie contributions to the respiratory equals that of the chest wall. Strictly speaking the pressure swings has assumed that the abdominal musPab, PL diagram can only be used to partition the cles remain relaxed. Grimby, Goldman, and Mead have respiratory pressure swings under static conditions. shown that under particular circumstances the abdomHowever, in normal subjects the time constants of lung inal muscles have an inspiratory action and that they and chest wall may be sufficiently similar so that perform a useful function by displacing the diaphragm during quiet breathing and moderate increases in ven- into the chest so that it becomes more efficient as a tilation, the diagram may still prove useful. pressure generator (7). Clearly, if our approach is to A crucial feature of our analysis is the assumption have much practical value, abdominal muscle. contracthat an isometrically contracting diaphragm can per- tion will have to be taken into account. form no external work. This may not necessarily be the When the abdominal muscles contract, Vab decreases case. If the pressure difference across the diaphragm and the relationship between Vrc and VL is no longer can be approximated by the LaPlace Law so that Pdi = unique. Because it is no longer unique, the relationship T/R where T is the tension in the diaphragm and R, its between Pab and Pel is also no longer unique. Theoretradius of curvature, an increase in Pdi when A Pab = 0 ically the solution to this problem is straightforward. might be accompanied by an increase R with a propor- For any given value of Vab there is still a unique tionately greater increase in T. Thus, it is conceivable relationship between Vrc and VL and therefore between that diaphragmatic flattening during an isometric con- Pab and Pel. By measuring the relationships between traction might perform positive external work. How- Vrc and Vab at different lung volumes as described by ever, to the extent that the diaphragm produces no net Konno and Mead (S), a series of isopleths relating Vrc force to displace the chest wall when A Pab = 0, it to VL at different constant values of Vab may be cannot have performed external work on the chest wall. constructed. Using the four-quadrant diagram in Fig. Under these circumstances it is difficult to see how it 5, these yield a series of Pab, Pel isopleths fur constant can perform external work on the lung. Nevertheless, values of Vab. Only one needs to monitor Vab (which one of the tasks of future research in this field will be can be accomplished by measuring its AP dimensions to define the exact changes in Pel and Pab when with magnetometers) to determine which isopleth perdiaphragmatic contraction performs no external work. tains at that instant and by comparing the measured Loading of the respiratory system. Under circumPab and Pel with that predicted from the appropriate stances such as resistive or elastic loading of the respi- isopleth, the diaphragmatic and intercostal/accessory ratory system at the mouth one can still-partition t-he muscle contributions can be determined as described respiratory pressure swings by plotting PT, (the differabove. Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
208
MACKLEM,
Of course this would be exceedingly tedious to do manually. We are currently attempting to do it electritally by describing the isopleths by form fitting equations, monitoring Vab as well as Pab and PL continuously, and determining digitally, on-line the instantaneous displacements from the appropriate isopleth and thus the diaphragmatic and intercostal/accessory muscle contributions to the respiratory pressure swings for
GROSS,
any degree of abdominal successful, the approach be more practical. This search Council. Received
research Council
GRASSING,
ROUSSOS
muscle contraction. If we are described in this paper should
was supported by a grant of Canada. P. T. Macklem
for publication
AND
10 February
from the Medical Reis an Associate of the
1977.
REFERENCES E., G. SANT’AMBROGIO, AND E. AGOSTONL Effect of diaphragm activity or paralysis on distribution of pleural pressure. J. Appl. Physiol. 37: 311-315, 1974. DOSMAN, J., A. GRASSINO, P. T. MACKLEM, AND L. ENGEL. Factors influencing the esophageal pressure gradient in upright man (Abstract), Physiologist 18: 194, 1975. GOLDMAN, M. D. Mechanical coupling of the diaphragm and rib cage. In: Loaded Breathing, edited by L. D. Pengelly, A. S. Rebuck, and E. J. M. Campbell. Edinburgh and London: Churchill Livingstone, 1974, p. 50-63. GOLDMAN, M. D., G. GRIMBY, AND 3. MEAD. Mechanical work of breathing derived from rib cage and abdominal V-P partitioning J. Appl. Physiol. 4: 752-563, 1976. GOLDMAN, M. D., AND J. MEAD. Mechanical interaction between the diaphragm and rib cage. J. AppZ. PhysioZ. 35: 197-204, 1973.
1. D’ANGELO,
2.
3.
4.
5.
6. GRASSINO, A. Influence of chest wall configuration on the static and dynamic characteristics of the contracting diaphragm. In: Loaded Breathing, edited by L. D. Pengelly, A. S. Rebuck, and E. J. M. Campbell. Edinburgh and London: Churchill Livingstone, 1974, p. 64-72. 7. GRIMBY, G., M. GOLDMAN, AND J. MEAD. Respiratory muscle action inferred from rib cage and abdominal V-P partitioning. J. Appl. Physiol. 41: 739-751, 1976. 8. KONNO, K., AND J. MEAD. Measurement of the separate volume changes of the rib cage and abdomen during breathing. J. Appl. Physiol. 22: 407-422, 1967. 9. MEAD, J., N. PETERSON, G. GRIMBY, AND J. MEAD. Pulmonary ventilation measured from body surface movements. Science 156: 1383-1384, 1967.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 16, 2019.
