Inferences about Respiratory Muscle Use after Cardiac Surgery from Compartmental Volume and Pressure Measurements

Eight male American Society of Anesthesiologists physical status 2 patients scheduled for elective cardiac surgery were included in this study. Their age (mean plus/minus SD) was 43 plus/minus 11 yr (range 27-57 yr). Their weight was 77 plus/minus 9 kg (range 68-97 kg). None of them had constant thoracic pain in the immediate preoperative period, and none was taking opioids before the operation. Three patients smoked, but none had clinical evidence of chronic obstructive pulmonary disease, and none was obese. The characteristics of the patients are shown in Table 1.

Institutional ethical committee approval was obtained, and all subjects volunteered to take part in the study after the nature and purpose of the experiment were explained to them.

Tidal changes in rib cage and in abdominal dimensions were measured simultaneously using a respiratory inductance plethysmograph (RIP) (Respitrace, White Plains, NY). A calibration of RIP measurements to spirometry was performed to calculate volume-motion coefficients for rib cage (VMC^RC) and abdomen (VMC^AB). Patients were asked to alternate between predominantly abdominal and thoracic breathing, while changes in the cross section of the two compartments were matched to simultaneously measured spirometric volumes (V^T), From these measurements, a regression line was found by the least squares method between abdominal cross section and V^Ton the x axis and between rib cage cross section and V^Ton the y axis. The intercepts of the regression line with the x axis and y produced 1/VMC^ABand 1/VMC sub RC, respectively. Details concerning the calibration procedure and the calculation of VMC have been reported. [12]Volume was calculated from the instantaneous sum of the RIP values for abdominal and rib cage volumes. Because cardiac surgery may change the geometric characteristics of the chest wall and result in changes in the VMCs of the two compartments, calibration of the RIP was repeated immediately before the measurements made on each day of observation. Spirometric measurements were made with a nose clip and a mouthpiece, connected to a heated number 2 Fleisch pneumotachograph, connected to a differential pressure transducer and integrator (17212, Godart, Bilthoven, The Netherlands).

The results of each set of measurements are the means of all cycles during a 5-min period of quiet breathing. Cross-section changes and VMC calculation of the abdominal and rib cage tidal volumes (V^ABand V^RC, respectively) were used to calculate abdominal and rib cage contributions to tidal breathing (V^AB/V^Tand V^RC/V sub T). Minute ventilation was obtained by the product of the RIP sum tidal volume (V^SUM) and respiratory rate.

P^gaand P^eswere used as indexes of pleural and abdominal pressure changes. They were measured using two thin-walled 10-cm-long balloons connected to polyethylene tubing (120 cm long and 1.7 mm ID) which were positioned in the middle third of the esophagus and in the stomach and filled with 0.5 and 1 ml air, respectively. [13]Each balloon was connected to a pressure transducer (DP 15, Validyne, Northridge CA). Transdiaphragmatic pressure (P^di) was calculated later by subtraction (P^di= P^ga- P^es) using instantaneous values of P^gaand P^es.

Changes in P^gaand P^eswere measured during inspiration by subtracting the pressure measured during the last 100 ms of the expiratory phase from the peak positive or negative pressure developed during inspiration determined on the rib cage tracing of the RIP, as previously reported by Duggan and Drummond. [2]All values presented for P^esand P^garepresent the means of all respiratory cycles recorded during a 5-min period of quiet breathing. During the postoperative period, the stomach was emptied before each set of measurements using a gastric suction tube inserted and removed before the ventilatory measurements. Many patients had "grunting" respiration (described in Results and Discussion) on the 1st postoperative day. Sections of the record where this occurred were avoided in the quantitative analysis.

Changes in cross-section dimensions, in P^es, P^ga, and P^diwere recorded on a strip chart recorder (ES 1000 Gould, Valley View, OH). All values were corrected to body temperature, pressure, saturated.

All of the patients were studied while in the supine 30 degrees head-up position. Preoperative baseline ventilatory measurements were performed the day before surgery. Anesthetic agents used during surgery included diazepam 25-50 mg, fentanyl 1.6-4 mg, and pancuronium 14-25 mg. All patients had a median sternotomy. The surgical procedure consisted of a coronary artery bypass graft in five patients, an atrial septal defect closure in one patient, a ventricular septal defect closure in one patient, and an aortic valve replacement in one patient. During cardiopulmonary bypass, myocardial preservation was achieved in all patients by systemic cooling, cold cardioplegia, and topical cooling of the heart with iced saline slush.

At each set of measurements, after the calibration procedure of the RIP, patients were left in a quiet environment without nose clip or mouthpiece for 10-15 min to obtain a steady-state breathing pattern. They were fully conscious throughout the measurements and were able to breathe quietly comfortably without pain, although many felt pain on maximal breathing maneuvers. No opioid analgesics were used in the postoperative period. Pain was controlled by nonsteroidal antiinflammatory drugs.

All values are reported as means plus/minus SD. Comparisons between the periods of measurements were performed by a two-way analysis of variance for repeated measurements and a Newman-Keuls test for multiple comparisons, when appropriate. A P value less than 0.05 was considered significant.

Group data (Table 2) show changes in RIP sum tidal volume, respiratory rate, minute ventilation, rib cage and abdominal expansion, and changes in P^gaand pleural pressure rather similar to those reported in series of patients after upper abdominal surgery. [3,4]There was a shift to rapid shallow breathing on the 1st postoperative day with no significant change in minute ventilation The V^AB/V^Tdecreased, without a change in V^RC. Inspiratory P^gachanges were smaller, actually becoming negative in five of the eight patients, and the ratio of the maximum change in P^gafrom the end-expiratory point during inspiration and the maximum change in P^es(delta P^ga/delta P^es) decreased. When the pattern of breathing was assessed in more detail by drawing plots of P^esagainst P^gaand rib cage against abdominal expansion through the respiratory cycle, it was found that the patients could be categorized into two groups. Six patients shifted their breathing pattern to one of mainly rib cage motion with positive changes in P^diin inspiration (Figure 1and, for interpretation of it, Figure 2). The other two (Figure 3and Figure 4) shifted to breathing mainly with the abdomen, with negative changes in P^diin inspiration. The two groups are considered separately, and the data displayed in Figure 1and Figure 3are explained in detail below (Discussion).

Average data for the first group show changes in V^T, respiratory rate, minute ventilation, V^RC, and P^ga/P^essimilar to those for the whole set of eight patients. For these six, the change in P^di, averaged 5.6 cmH²O preoperatively and 4.8 cmH sub 2 O on the 1st day after surgery. Rate of increase in P^di(change in P^dibetween end-expiration and end-inspiration per change in inspiratory time taken from rib cage excursion) was 4.2 cmH sub 2 O/s preoperatively and 5.0 cmH²O/s postoperatively. The ratio of Delta P^esto RIP sum tidal volume, a rough estimate of pulmonary impedance, was 6.3 cmH²O/ml preoperatively and 13.3 cmH sub 2 O/ml postoperatively. V^RC/V^Tchanged from 0.54 preoperatively to 0.78 postoperatively.

The other two patients showed a very unusual pattern of breathing on the 1st postoperative day, corresponding to the rib cage-abdomen and Delta P^es/Delta P^gaplots in Figure 3and the polygraph tracings in Figure 4. P^di, decreased from the beginning of inspiration, reaching its lowest level just beyond the beginning of expiration, and then increased through expiration and leveled of at end-expiration. In subject 8, there was a small superimposed positive deflection in P^diat end-inspiration. Both of these patients expanded their rib cage less than preoperatively. Subject 8 reduced his rib expansion to nearly zero. Subject 7 did not move his rib cage for the first half of inspiration but let it expand in the second half.

Several patients (subjects 3, 6, 7, and 8) in many breaths on the 1st postoperative day showed a peculiar pattern of P^esin expiration shown in the loops of Figure 1and Figure 3and the example tracing of Figure 5. After an initial negative deflection through inspiration, P^esswung positive in early expiration, to as great as 30 cmH²O in some patients. The positive change in P^eswas often accompanied by a movement on the rib cage-abdomen plot in an isovolume direction: rib cage out and abdomen in. The positive waves in P^eswere variable from breath to breath but occurred in most breaths postoperatively in three patients, not at all in three patients, and intermittently in the others.

The patterns of breathing in all eight patients had reverted toward normal by the 5th postoperative day although respiratory frequency remained high, and most of the loops had not returned completely to the preoperative configuration.

Chest radiographs showed evidence of lower lobe atelectasis in all patients on the 1st postoperative day. This had cleared in most cases by the 5th day, but there were still definite loss of volume and bilateral silhouette signs of the diaphragm in subject 8. No patient had clinical or radiographic suggestions of unilateral diaphragm paralysis, and none had a prolonged syndrome of atelectasis or elevated diaphragm typical of severe cold block diaphragm dysfunction. Transient diaphragm paresis related to the cold block could not be excluded, however.

Results very similar to those shown in Table 2have been reported in several studies of patients after upper abdominal surgery and by Maeda et al. in a group of 21 patients after thoracotomy for pneumonectomy or lobectomy. [9]In the earlier studies, [3,4]the data were interpreted to indicate a substantial decrease in diaphragm activity on the 1st postoperative day but that interpretation has been challenged. Estimates of diaphragm contribution to inspiration were based on the relation between changes in P^gaand pleural pressure according to the theory of Macklem et al. [14]for inspirations from relaxed functional residual capacity. More recently a study of chest wall mechanics has modified these concepts, [15]and the recognition that abdominal muscle contraction may occur in patients after surgery has required a more complicated analysis of the pressure-volume data. [16,17].

A decrease in the ratio Delta P^ga/Delta P^eswas originally considered to indicate a decrease in diaphragm activity, [3,4]but other factors in patients who have undergone surgery can affect this ratio. [17]First, by relaxing as the diaphragm contracts, abdominal muscles tend to decrease the changes in P^gaand can even shift the maximum change in P^gafrom the end-expiratory point during inspiration from positive to negative. Second, after surgery, patients probably have an increase in the overall drive to breathe, suggested by an increase in occlusion pressure, [18]an increase in the work of breathing, [9]and normal ventilation despite an increase in lung impedance. In our patients, for example the ratio of the maximum change in P^esto V^T, which gives a rough indication of impedance, increased from 6.3 preoperatively to 13.3 ml/cmH²O postoperatively. When overall drive to breathe increases, intercostal and accessory muscle activity increases proportionately more than diaphragm activity, [19]which tends to decrease the ratio Delta P sub ga /Delta P^es. Third, the increase in mechanical load by itself tends to decrease Delta P^ga/Delta P^es. [17].

Likewise, the increase in the ratio of rib cage to abdominal expansion does not necessarily imply an abnormal decrease in the contribution of the diaphragm to breathing. This increased ratio may be attributable in part to increased intercostal inspiratory activity associated with increased overall drive to breathe. In addition, after surgery some patients show tonic activity of abdominal muscles in inspiration,* which makes the abdominal wall less extensible and enhances the ability of the diaphragm to expand the rib cage. [20]Finally, unusual use of abdominal and intercostal muscles by patients after surgery may distort the chest wall and thus decrease the accuracy of calibrated RIP signals for measuring compartmental volumes. Potential problems with the method after thoracic surgery have been described. [11].

We have tried to resolve the question about the use of respiratory muscles by analysis in our patients of the time course of P sub ga, P^es, V^RC, and V^ABthrough the respiratory cycle. Although quantitative estimates of the contributions of the various muscles groups cannot be made, some useful qualitative conclusions can be drawn. Figure 2gives a scheme for interpretation of the P^es/P^gaplots in Figure 1and Figure 3. Negative P^esis plotted on the ordinate, positive P^gaon the abscissa. Lines of iso-P^di, (iso-P^ga- P^es) are downward-sloping 45 degrees diagonals, indicated by dashed lines. Movement on the plot upward and to the right crosses iso-P^dilines in the direction of increasing P sub di. Movement along a dashed line implies no change in P^di. [15]Examination of the matched P^es- P^gaand V^RC- V^AB(Konno-Mead) plots for the six patients of the first group (Figure 1) shows that all the preoperative loops are normal for supine quiet breathing, with smooth increases and decreases in rib cage and abdominal diameters in phase with each other and with P^esbecoming more negative, P^diincreasing and P^gaincreasing from the beginning of inspiration. On the 1st day after surgery, by contrast, 4 of the 6 (subjects 2, 3, 5, 6) begin inspiration with a drop in P^gaaccompanied by some expansion of the abdomen. This can only be explained by supposing that abdominal muscles (or possibly, expiratory intercostal muscles) are active at the end of inspiration and relax in the early part of inspiration, allowing the abdominal wall to expand outward and negative pressure to appear in the abdomen. Intercostal inspiratory muscle activity can produce negative abdominal pressure but drags the diaphragm upward and produces a decrease in abdominal diameter. [21]Diaphragm contraction can expand the abdomen but only by making positive abdominal pressure. Even if these patients use all three sets of muscles, relaxation of expiratory muscles must be the predominant action in early inspiration.

All six of these patients increase P^diin inspiration, implied by a pathway on the P^es- P^gaplot that crosses iso-P sub di lines in an upward or rightward direction (Figure 2) although two of them (subjects 3 and 5) begin inspiration by moving upward to the left, close to the iso-P^diline, suggesting that the first part of inspiration is accomplished without significant contribution by the diaphragm.

We considered whether an estimate of diaphragm activity in isolation could be made from P^di. In the six patients, changes in P sub di, averaged 5.6 cmH²O before the operation and 4.8 cmH²O on the 1st day after surgery. Part of this difference is explained by the change in respiratory frequency. The shorter inspiratory time means that even with the same inspiratory flow rate and the same rate of increase of inspiratory electromyographic activity, the value of P^direached by the end of the shortened inspiration must be less than before the operation. An approximate correction can be made for this by using the mean rate of increase in P^diin inspiration (change in P sub di between end-expiration and end-inspiration per change in inspiratory time taken from rib cage excursion), which was 4.2 cmH²O/s before and 5.0 cmH²O/s after the operation (not significantly different). The rate of increase in P^didoes not change significantly because peak inspiratory P^didepends not only on the level of activation of the diaphragm but also on the length of the diaphragm at end-inspiration. Because of abdominal muscle contraction, the end-inspiratory configuration of the chest wall postoperatively has a relatively smaller abdominal volume and a relatively longer diaphragm, implying that P^dioverestimates the degree of diaphragm activation postoperatively compared with preoperatively. Correction for this force-length effect cannot be estimated accurately. In normal humans, however, it is possible to double the P^diobtained for a given level of electric activation of the diaphragm by contracting abdominal muscles and lengthening the diaphragm. [22].

As illustrated by subjects 7 and 8, described below, end-expiratory abdominal muscle activity confounds interpretation of P sub di by a second mechanism. Abdominal muscle contraction can place a passive stretch and a passive positive P^dion the diaphragm. As the abdominal muscles relax, P^dithen moves in the negative direction. To make positive inspiratory P^dithe diaphragm must first contract actively enough to make up for the loss of tension with release of passive stretch. Only active contraction beyond this compensatory one will show up as a positive wave in P^di. The change in P^dimay then underestimate the degree of diaphragm active contraction.

Overall, it is hard to be certain that there is any important change in activation of the diaphragm in any of the subjects except 7, and 8, in whom diaphragm activation is probably reduced considerably (see below).

Consideration of the balance of forces acting on the rib cage can be used to estimate activity of the inspiratory intercostal and accessory muscles postoperatively. Volume of the rib cage at end-inspiration depends on the balance between elastic recoil of the rib cage, pleural pressure, the expanding force exerted by the diaphragm and the expanding force exerted by inspiratory intercostal and accessory muscles. In the postoperative period, pleural pressure tended to be more negative (end-inspiratory P^eswas -4.8 cmH²O compared with -3.5 cmH²O preoperatively), a condition that should reduce rib cage volume. Inspiratory increase in diaphragm tension, estimated by P sub di, was not statistically changed (the change in P^di, was 4.8 cmH²O compared with 5.6 cmH²O preoperatively), but the effectiveness of the diaphragm at expanding the rib cage will have been greatly impaired because the normal positive inspiratory wave in P^gahad disappeared (the maximum change in P^gafrom the end-expiratory point during inspiration was -2.0 cmH²O postoperatively compared with 2.6 cmH²O preoperatively). Positive P^gapushes directly outward on the rib cage in the area of apposition and also provides resistance to diaphragm descent so that tension in diaphragm fibers inserted into the rib cage margin tends to elevate the ribs. Despite changes in measured pressure that should reduce its movement, the rib cage expands normally in the patients after surgery. This can best be explained by an increase in inspiratory force exerted by intercostal and accessory muscles (although another possibility is that surgery somehow caused a substantial increase in rib cage compliance).

Altogether, the mechanics data in these six patients do not provide convincing evidence of a change in diaphragm activity, but do show the appearance of phasic expiratory abdominal muscle activity and suggest an increase in intercostal and accessory muscle activity. In other circumstances where overall drive to breathe is increased, such as exercise [23]and CO²-stimulated breathing, [24]abdominal muscles are recruited and intercostal and accessory muscles increase their activity, but this is normally associated with a considerable increase in activity of the diaphragm. Abdominal expiratory muscles do not show significant phasic expiratory activity with CO²-stimulated breathing until minute ventilation is about five times the resting level. [24]By contrast, in our patients and others in whom abdominal muscle activity has been studied after surgery, [16]the abdominal muscles have been found to be very active when the overall drive to breathe is only mildly increased. There is thus a shift in emphasis of respiratory neuromuscular output from the diaphragm to other major muscles of respiration.

The postoperative pattern of breathing of the remaining two patients was dramatically different from the others. As shown in the Konno-Mead diagrams of Figure 3, they shifted from a normal preoperative pattern to one in which abdominal motion was exaggerated and rib cage motion was greatly reduced. The pressure data, shown both as loops in Figure 3and pressure-time traces in Figure 4, indicate that P^diactually decreased in inspiration. The changes in P^gaincreased slightly in magnitude and reversed in sign.

Because it is unreasonable to suppose that diaphragm muscle fibers are active through expiration and are deactivated in inspiration, the explanation of these data must be that there is a positive deflection in P^diat end-expiration as a result of passive stretching caused by contraction of abdominal muscles at a low lung volume. At the beginning of inspiration, abdominal muscles relax, releasing the passive stretch as the diaphragm descends and abdominal volume increases. In both subjects, P^direaches its lowest value just after the end of inspiration. In subject 7, there is an increase in P^diat end-inspiration that must be attributed to active contraction of the diaphragm. Throughout inspiration, P^diin subject 7 thus has a decreasing positive component attributable to passive stretch, superimposed on a transient positive component attributable to active contraction. It is not possible to separate these components with precision but the amount of active P^diin this patient may be best estimated from the sudden drop in P^diwhen the diaphragm relaxes at a time when its length is changing little according to the rib cage abdomen plot. In subject 8, who showed no positive bump in P^diat end-inspiration, it is plausible to suppose that the diaphragm does not contract at all, being just passively stretched and released by abdominal muscles but it is impossible to exclude some superimposed active component to P^diin inspiration. The pattern of breathing is similar to the one described for certain patients with bilateral diaphragm paralysis. [25].

A curious pattern of pressure and motion during expiration was seen in the majority of patients on the 1st postoperative day. These patients apparently constricted their larynx or pharynx at end-inspiration to give themselves a partial upper airway obstruction and then relaxed their inspiratory muscles. They thus chose to retard expiratory flow by adding an expiratory resistance instead of by prolonging inspiratory muscle activity into expiration. An example is shown in Figure 5. After a normal inspiration, a plateau in volume begins at point A. During the plateau from A to B, pleural pressure shifts to a positive value and the rib cage volume increases while abdominal volume decreases. The shift in the rib cage-abdomen relation during the plateau probably signifies the return of the chest wall from a configuration distorted by inspiratory muscle tension to a relaxed configuration. There is a positive deflection in P^gaat the same time as the one in pleural pressure but the amplitude of the gastric deflection was smaller because of rapid relaxation of the diaphragm occurring at the same time as the appearance of upper airway resistance.

This pattern of breathing corresponds to grunting respiration because when the airway is eventually opened, the positive end-expiratory pressure can cause a sudden, noisy expulsion of air. It is seen in newborn infants, [26]in some patients with weak respiratory muscles, [27]and in other patients with respiratory difficulties. It may help to keep mean lung volume increased in expiration at a comparatively low cost in terms of respiratory muscle work.

We found two types of breathing patterns among the eight patients observed after thoracic surgery. Six of the eight breathed in much the same way as patients described after abdominal surgery with a reduction in abdominal excursion relative to rib cage excursion and an increase in the ratio of changes in P^gato changes in P^es. The detailed analysis of this pattern suggests it is caused by an increase in expiratory activity of abdominal or intercostal muscles and in inspiratory activity of intercostal muscles, with no increase in diaphragm activity. This pattern tends to reduce abdominal excursion and therefore probably reduce diaphragm excursion. The other two patients showed an increase in abdominal excursion relative to rib cage excursion, and appeared to be generating much of their tidal volume by contraction and relaxation of abdominal muscles. In these cases diaphragm activity was probably reduced below the control level.

The latter two patients used intercostal muscles enough to stabilize the rib cage from collapse but not enough to expand it a great deal, and thus reduced motion of the rib cage. The other six patients breathed in such a way as to limit movement and pressure changes in the abdomen, and consequently to limit diaphragm motion.

Our observations do not give any information about the mechanism or consequences of the altered pattern of breathing in these patients. We speculate that the purpose of it is simply to immobilize the part of the chest wall that gives rise to the most activity from nociceptive receptors. The diaphragm may be the most sensitive region, accounting for the predominance after thoracic surgery of a pattern like that seen after abdominal surgery, but in some cases perhaps the rib cage compartment dominates in its requirement to be still. We could find no clinical correlates for the different patterns.

The use of abdominal muscles in expiration in most of these patients implies that their end-expiratory volume is less than functional residual capacity, particularly in the second group. The magnitude of the decrease in functional residual capacity was not measured but in extreme cases may constitute most of the tidal volume. This change may have a detrimental effect on gas exchange. The grunting maneuver seen in several patients may be an advantage in this regard because it tends to prolong the time spent at end-inspiratory volume.

*Pansard JL, Clergue F, Lesage R, Viars P: Effects de la bupivacaine a 0.5% par voie epidurale thoracique sur le travail inspiratoire apres chirurgie abdominale haute (abstract). Ann Fr Anesth Reanim 10:R59, 1991.