Human Chest Wall Function during Epidural Anesthesia

This study was approved by our institution's review board. Six healthy persons who did not smoke were studied after they gave informed consent (Table 1). Results of physical examination, chest roentgenogram, and pulmonary function tests were normal for each person. A pregnancy test was performed on the one female participant. Participants were brought to the laboratory the day before the actual experiment to become familiar with the experimental procedures. They were not allowed to consume oral fluids after midnight the day before the experiment.

A lumbar epidural catheter was placed at the L3-L4 interspace, and a test dose of 3 ml 2% lidocaine with 1:200,000 epinephrine was administered. Each participant was placed in the supine position. Respiratory impedance plethysmography (RIP) belts (Respitrace, Ardsley, NY) were placed around the upper rib cage and mid-abdomen. An intravenous catheter was inserted, and the radial artery was cannulated to obtain samples for blood gas analysis (IL 1302; Instrumentation Laboratories, Lexington, MA) and to monitor blood pressure. Ringer's lactate was infused at approximately 30 ml/hr via the intravenous catheter throughout the experiment.

Bipolar EMG electrodes were inserted into several respiratory muscles. The electrodes were fashioned by removing 1 mm of insulation from the end of 0.002-inch teflon-insulated wires. Two wires were passed through an insulated 30-gauge needle and then bent 1 mm from the end to form hooks. The electrodes were inserted under ultrasonic guidance [30]while electrical activity was monitored in the desired location, the needle was removed, leaving the wires in the desired muscle. Electrodes were placed in the transversus abdominis muscle in anterior axillary line approximately 4 cm inferior to the costal margin (at approximately the T10 dermatome), in the parasternal intercostal muscle (PS) at the third right interspace, approximately 3 cm lateral to the midline, and in the right anterior scalene muscle (SCA) near the base of the neck. An electrode was also placed in the diaphragms of three participants. [31]To insert the diaphragm electrode, ultrasound was first used to define the diaphragm position at full inspiration. The needle was then inserted one interspace inferior to this point in the mid-axillary line and advanced until phasic inspiratory electrical activity was detected. [32]This procedure minimized the possibility of causing pneumothorax as the electrode was placed through the intercostal muscles into the diaphragm close to its insertion on the rib cage. Electromyogram signals were amplified (Grass P511, Quincy, MA), band-pass filtered between 30 and 3000 Hz, and recorded on digital audio tape (TEAC RT100, Montebello, CA) for later processing.

The RIP bands were calibrated using the method of Mankikian and colleagues. [33]Changes in dimensions of the rib cage and abdomen measured with the RIP bands were related to the volumes displaced by the rib cage and abdominal compartments during tidal breathing; these relationships were expressed as volume-motion coefficients of the rib cage and abdomen. These coefficients were calculated from data obtained by asking the participants to alternate predominately abdominal and thoracic breathing. Calibrations were performed at the beginning and end of each experimental condition (intact or EA), and average values were calculated for each condition. Inspiratory and expiratory gas flows were measured using a pneumotachograph (Fleisch 3, Richmond, VA) connected to a differential pressure transducer (Validyne MP-45, Northridge, CA). Gas flows were integrated to obtain changes in lung volume, which were corrected to BTPS conditions. Subjects breathed through a mouthpiece and nose clip during all measurements. Similar to our previous study, [31]we found that this apparatus did not change the pattern of breathing (data not shown).

Heart muscle activity, blood pressure, and arterial oxygen saturation were monitored throughout the study.

Participants were placed in the dynamic spatial reconstructor (DSR), a high-speed X-ray scanner that uses the computed tomography principle to provide three-dimensional images of the thorax. This technique has been described in detail elsewhere. [31,34–37]The DSR has sufficient temporal resolution to visualize thoracic structures during quiet breathing and sufficient volume resolution to determine a known volume to within 2%.

The participants were scanned while breathing quietly through the mouthpiece and nose clip, with their arms brought above their heads with the humeri vertical to allow for thoracic imaging. Scans lasting 300 ms were triggered manually at end expiration and end inspiration. Scans at the same point in the breathing cycle were recorded during three consecutive breaths and gated together during later analysis to provide end-expiratory and end-inspiratory images. Because cephalocaudal height of the imaging field was not sufficient to include the entire thorax, these initial scans included only the superior half of the thorax. The participants were then shifted cephalad, and a similar sequence of scans was obtained to visualize the inferior portion of the thorax. [36]Respiratory impedance plethysmography measurements, tidal volume, and EMGs were recorded simultaneously to ensure stability of the breathing pattern. During later analysis, the superior and inferior images were joined to produce end-inspiratory and end-expiratory images of the entire thorax.

Immediately after scanning, the FRC was measured in duplicate using a nitrogen-dilution technique. [31,38,39]The participants each performed six vital capacity maneuvers into a 4-liter bag initially filled with 100% oxygen after the bag was connected to the mouthpiece at end expiration. Nitrogen concentrations in the bag were determined before and after this maneuver using a mass spectrometer (Perkin-Elmer MGA 1100, Buffalo Grove, IL).

After these measurements were taken during quiet breathing (referred to hereafter as “intact” breathing), local anesthetic solution was injected incrementally through the epidural catheter to produce sensory block to approximately the T6 dermatome. In participants 1 and 2, 2% lidocaine with 1:200,000 epinephrine was used; in the other four participants, 1.5% etidocaine with 1:200,000 epinephrine was used to maximize motor block. At this level of block, subsequently called “mid-thoracic EA,” activity in the transversus abdominis muscle was eliminated (both during quiet breathing, if present, and during voluntary activation), but activity in the PS was intact. After a stable level was achieved, measurements were made during quiet breathing; DSR scans were not performed because the cumulative amount of radiation exposure had to be limited in these volunteers. More local anesthetic was injected through the catheter to produce motor block of the PS at the third interspace, a condition called “high EA.” After a stable level was confirmed, DSR scans and FRC measurements were repeated.

Details of image processing to define chest wall boundaries and to validate the DSR in measuring chest wall motion have been previously described. [31,34–37]Briefly, each scan produced a three-dimensional volume image of the thorax composed of cubic volume elements (voxels) with edge lengths of 1.3 mm. Images were processed to define each voxel in the image as being in the thoracic cavity, the abdominal cavity, or the background. Thoracic volume (V^th) was determined by counting the number of voxels in the thoracic cavity. Changes in thoracic liquid volume during inspiration (Delta V^liq), presumably representing changes in thoracic blood volume, [36]were calculated as the difference between changes in V^thfrom the beginning to the end of inspiration (Delta V^th) and tidal volume (V sub T) measured by the integration of gas flow (Delta V^liq= Delta V^th- V^T). Changes in thoracic volumes between any two scans were partitioned into volumes swept by the diaphragmatic and rib cage surfaces. [34]Changes in the pattern of chest wall motion were quantified as previously described [34](Figure 1).

Changes in the volume of the heart and major vessels were also estimated directly from these images. [31,38]The heart and other mediastinal structures were isolated in each image using a combination of computer thresholding and operator interaction. The pulmonary vessels were truncated at approximately the same location in all scans from each participant. The total volume of these mediastinal structures (V^ms) was measured by counting the number of voxels within these structures and multiplying by the volume of one voxel. Because the images were obtained by gating several scans together at random times in the cardiac cycle, this volume represents an average during this cycle.

Electromyogram signals recorded on tape were processed with a third-order Paynter filter to provide a 100-msec moving time average (MTA). [40]To quantify this activity, the MTA tracings were digitized and the mean MTA activity per breath was calculated as the area under the MTA signal divided by its duration. [41]If EMG activity persisted into expiration, only the portion of the signal before the onset of expiratory flow was used to calculate this mean. [42].

Paired statistical comparisons were made by paired t tests. Multiple comparisons were made using repeated measures analysis of variance, with Dunnett's test used for post hoc comparisons. P < 0.05 was taken as significant. Values are expressed as M +/- SE.

Mid-thoracic EA required a total dose of 16 +/- 2 ml of anesthetic and produced sensory levels ranging from T4 to T8 dermatomes (Table 2). High EA required a total dose of 34 +/- 2 ml anesthetic and produced sensory levels ranging from C6 to T2 in 5 participants. In participant 4, the sensory level did not exceed T4.

Compared with values before block, high EA did not significantly change arterial P^O2(from 93 +/- 3 to 97 +/- 5 mmHg) or arterial P^CO2(from 39 +/- 1 to 36 +/- 2 mmHg). High EA significantly decreased mean arterial pressure (from 101 +/- 5 to 83 +/- 6 mmHg) without changing the heart rate (from 66 +/- 3 to 69 +/- 4 min sup -1). Participant 6 required a 500-ml bolus of intravenous fluid for transient hypotension when establishing high EA.

During intact breathing, consistent phasic EMG activation, predominantly during inspiration, was observed in the PS and the diaphragms in the three participants in whom this activity was measured (Figure 2, Table 2). In these three persons, phasic expiratory activity was also detected by the diaphragm electrode (Figure 2). Phasic inspiratory activity was present in the SCA of every participant except 6. In participant 5, tonic activity was present in the PS and SCA in addition to phasic activity. Tonic activity was noted in the transversus abdominis muscle in two participants.

Mid-thoracic EA abolished any activity in the transversus abdominus muscle and produced tonic activity in the SCA of participant 6. This level of EA did not significantly change phasic inspiratory EMG activity in the SCA or the PS (mean MTA of 1.57 +/- 0.69 and 0.93 +/- 0.18, respectively, normalized to the activity present before block).

High EA abolished activity in the PS of every participant except in the person designated number 4. Additional tonic activity was noted in the SCA of some persons (Figure 3, Table 2). High EA did not significantly change phasic inspiratory EMG activity in the SCA (mean MTA of 0.85 +/- 0.20, normalized to the MTA activity before block).

High EA did not significantly change tidal volume or breathing frequency (Table 3). The change in intrathoracic volume during inspiration, as measured by the DSR, was significantly greater than the tidal volume measured by integrating gas flow, indicating that V^liqconsistently increased during inspiration (Table 3). This increase in V^liqcomprised a substantial (20 +/- 4%) fraction of Delta V^thduring intact breathing and corresponded closely to the change in V^mgduring inspiration, measured directly from DSR images of the mediastinal structures. These findings are consistent with our previous observations during quiet breathing. [31]During high EA, V^liqand V^mgincreased with inspiration in five of the six participants (Table 3).

High EA significantly decreased inspiratory rib cage volume displacement, compared with intact breathing, but did not change inspiratory diaphragm displacement (Table 3). Therefore, high EA decreased the relative contribution of rib cage expansion to inspiratory increases in thoracic volume (from 27 +/- 2 to 10 +/- 11% of Delta V^th).

The pattern of inspiratory rib cage expansion before block varied among participants (Figure 4). The effect of high EA also varied. Paradoxic rib cage motion (i.e., a net inward inspiratory rib cage motion) was noted during inspiration in participant 5. In participants 2, 4, and 6, the expansion of the most cephalad portion of the rib cage was unchanged or augmented by high EA. Consistent with earlier reports, the greatest motion of the diaphragm in the cephalocaudal axis was always in the most dependent regions (Figure 5).

The effects of mid-thoracic EA were assessed by analyzing RIP measurements (Table 4). This level of EA did not significantly affect compartmental volume displacements estimated by RIP. Consistent with the DSR measurements, high EA significantly reduced the inspiratory rib cage volume expansion estimated by RIP (Table 4). Neither level of EA changed any measure of ventilatory timing (Table 4).

High EA significantly increased the FRC, measured by nitrogen dilution (Table 5). In five participants, the change in FRC was greater than the change in V^th, showing that V^liqat end expiration decreased in these participants. Consistent with this observation, V^mgmeasured directly from the thoracic images also decreased in these five participants (Table 5). These results suggest that in these participants, high EA decreased the intrathoracic blood volume at end expiration and thus contributed to the increase in FRC. In contrast, intrathoracic blood volume at end expiration increased in participant 6, who was unique in requiring an additional 500-ml fluid bolus during the establishment of high EA.

In four participants, an expansion of the rib cage contributed to an increase in end-expiratory V^th(Table 5). There was also a net caudad motion of the end-expiratory position of the diaphragm in five of the six participants. In every participant, the anterior portion of the diaphragm moved caudad (Figure 6). In five participants, the posterior portion of the diaphragm also moved caudad. This posterior portion moved cephalad in one person; this was the same participant (#4) in whom a sensory block above T4 could not be achieved.

Major findings of this study include the following:(1) Rib cage expansion contributes to tidal volume during high EA in most participants, even when most of the muscles of the rib cage are paralyzed;(2) the mean phasic electrical activity of intact respiratory muscles such as scalenes does not significantly increase in response to rib cage muscle paralysis produced by EA; and (3) high EA increases the FRC, an increase produced in most participants by a caudad motion of the diaphragm and a decrease in intrathoracic blood volume.

Inspiratory rib cage expansion requires the coordinated activity of several respiratory muscles, including phasic inspiratory activation of PS and scalene muscles, and phasic expiratory activation of internal intercostal muscles in the lower lateral interspaces. [1,2,32]We confirmed the presence of phasic activity in these muscles in most participants during quiet breathing. As in a previous study, [31]we again noted phasic expiratory activity in the diaphragm electrode. The present study confirms our previous speculation that its source was the overlying internal intercostal muscles, because paralysis of this myotome by EA consistently eliminated this activity.

Studies in animals and in quadriplegic humans have demonstrated that isolated diaphragm contraction constricts the upper rib cage and expands the lower rib cage. [1,29,35,43]Constriction of the upper rib cage is caused by an inspiratory decrease in intrathoracic pressure that pulls this portion of the rib cage inward. Expansion of the lower rib cage results from the action of the diaphragm through its insertion on the rib cage, which elevates and expands the ribs (an “insertional” component), and inspiratory increases in abdominal pressure that act via the portion of the diaphragm that is apposed to the inner rib cage surface (an “appositional” component). Paralysis of rib cage muscles by regional anesthetic techniques might be expected to produce this pattern of chest wall motion.

Data regarding chest wall motion during epidural or spinal anesthesia from previous human studies are limited. Eisele and coworkers [44]mentioned passive retraction of the upper rib cage during continuous spinal anesthesia to the T1 level in three persons, but they did not quantify chest wall motion. Pascucci and associates [45]noted paradoxic inward motion of the rib cage during inspiration in four of six infants receiving spinal anesthesia. The deformability of the infant rib cage makes it difficult to extrapolate these results to adults. Kochi and coworkers [22]reported that thoracic EA with 2% lidocaine (with resulting sensory levels from approximately T1 to T6) caused a small but significant decrease in the rib cage contribution to tidal volume during quiet breathing. Yamakage and associates [17]found that spinal anesthesia to a T6 sensory level actually increased the rib cage contribution to tidal volume. This increase disappeared when the patients later fell asleep, suggesting that it represented a voluntary response to the sensations engendered by spinal anesthesia.

We found that although high EA diminished the contribution of the rib cage to tidal volume, it did not cause inspiratory constriction of the upper rib cage in most participants, suggesting that other muscles still active during EA must have expanded the rib cage. The scalene muscles, formerly thought to be “accessory” muscles of respiration, are in fact regularly active during quiet breathing in humans in supine and upright postures. [28,46]There is evidence that SCA activity can significantly affect the motion of the upper rib cage, especially when the intercostal muscles are paralyzed. [47]Estenne and De Troyer [29]found that the upper rib cage moved inward during inspiration in quadriplegic participants when SCA activity was absent. In most participants with phasic inspiratory SCA activity, the upper rib cage expanded during inspiration. De Troyer and Estenne [46]found similar results in healthy participants performing voluntary respiratory maneuvers. These results suggest that the continued SCA activity observed during high block could account for rib cage expansion despite paralysis of most intercostal muscles. We previously found that high EA caused paradoxic inspiratory rib cage motion in pentobarbital-anesthetized dogs. [35]In these dogs, SCA activity was never present in implanted EMG electrodes. Although there are many species differences between the chest wall of dogs and humans, this may represent further evidence for the importance of SCA in preventing paradoxic rib cage motion during high EA in humans.

It is also possible that intercostal muscles cephalad to the T3 myotome (at which parasternal intercostal activity was monitored) were active and contributed to rib cage expansion during high block. We could not eliminate PS activity at this level in participant 4; continued rib cage expansion can be attributed to this activity in this person. However, the complete block at the T3 level achieved in the other participants makes it likely that muscles at more cephalad interspaces were at least partially paralyzed. We used etidocaine in the last four participants we studied to maximize intercostal muscle paralysis, yet we noted paradoxic rib cage motion in participant 5 only.

We are uncertain why this participant could not maintain inspiratory rib cage expansion. Estenne and De Troyer [29]also found that some quadriplegic patients exhibited paradoxic rib cage motion despite phasic inspiratory SCA activity. This participant also experienced the highest sensory level of block and was unique in demonstrating tonic activity in both the PS and SCA before block.

Acute paralysis of selected respiratory muscles can increase activity in other respiratory muscles. [48,49]Possible factors responsible for such recruitment include reflexes, with afferents arising from receptors in the lung or the chest wall, or increases in Pa sub CO², if ventilation is sufficiently depressed by paralysis. Brichant and colleagues [48]found that high EA in pentobarbital-anesthetized dogs increased the neural activation of the diaphragm, probably in response to increases in Pa^CO2. We had hypothesized that SCA EMG activity might increase to compensate for paralysis of the PS during EA. That we could not show a change in this activity during high EA suggests that perturbations of chest wall motion, lung volume, or Pa^CO2were insufficient to trigger compensatory responses, that blockade of chest wall afferents interfered with any reflex responses, or that other factors such as effects of absorbed local anesthetic on respiratory control affected any response. Similarly, mid-thoracic EA elicited no compensatory increase in PS or SCA activation; neither did it significantly affect the pattern of chest wall motion. Measurements of ventilatory timing also provided no evidence for a reflex response to EA, because they were not affected. The development of hypercarbia during high EA in pentobarbital-anesthetized dogs studied by Brichant and colleagues may also suggest a greater dependence of the dog on the actions of extradiaphragmatic respiratory muscles during quiet breathing compared with humans. [34,35]Alternatively, pentobarbital anesthesia may have interfered with compensatory responses in the dogs.

Although we did not measure gas exchange in detail, arterial blood gases were consistent with previous reports that epidural or spinal anesthesia has little effect on Pa^O2. [13–17,30]Apparently the alterations in chest wall motion are insufficient to significantly affect regional ventilation and gas exchange, or sufficient compensatory mechanisms such as hypoxic pulmonary vasoconstriction prevent significant gas exchange abnormalities. Participant 5, who demonstrated the greatest change in rib cage function, showing paradoxic inspiratory rib cage motion, actually experienced an increase in Pa^O2(from 94 to 108 mmHg), accompanied by a decrease in Pa^CO2(from 37 to 31 mmHg). This trend toward a decrease in Pa^CO2, although not statistically significant, is consistent with previous reports and may be related to effects of absorbed local anesthetic on ventilatory control. [11,12]Maintenance of gas exchange may be related to the increase in FRC produced by EA, as discussed in the following section.

Previous studies found that thoracic or lumbar EA has no effect [14–16]or decreases [13]the FRC. Paralysis of rib cage muscles by intercostal nerve blocks decreases the FRC. [51,52].

In contrast to these reports, we found that high EA significantly increased the FRC (by approximately 15%). The mechanism responsible for this increase varied among participants. In five of the six, V^liqdecreased, presumably caused by a shift of blood out of the thorax. This decrease, calculated on the basis of differences between changes in gas volume (measured by nitrogen dilution) and total thoracic volume (measured by the DSR), was confirmed by direct measurements of the volume of images of the heart and other mediastinal structures (V^mg). In these five participants, this change represented a 6.4 +/- 2.1% decrease in the volume of tissue in the thorax. Arndt and coworkers [9]found that EA to a T4 sensory level decreased thoracic blood volume by 8.1 +/- 1.2% as measured by radiolabeled erythrocytes, with this blood redistributed to the capacitance vessels of the legs in most persons studied. [9,10]Although direct comparison of our values with those of Arndt and coworkers requires knowledge of the proportion of thoracic tissue composed of blood, they appear to be consistent. These shifts in blood volume presumably reflect the cardiovascular effects of the sympathectomy produced by EA. As found in previous work, most of these healthy volunteers could maintain an adequate heart rate and mean arterial pressure without requiring substantial amounts of intravenous fluids. [8].

In contrast, high EA increased end-expiratory thoracic liquid volume in participant 6. This person was unique in requiring additional intravenous fluid for transient hypotension during local anesthetic injection. We speculate that this increase in V^liqin participant 6 reflected the administration of this additional intravenous fluid. Consistent with this concept, Stanton-Hicks and associates [10]found that decreases in thoracic blood volume produced by EA could be reversed by increasing cardiac filling by peripheral venoconstriction. In addition, we found that high EA increased end-expiratory thoracic liquid volume in a previous study of pentobarbital-anesthetized animals. [35]Although many other factors may have been responsible, these animals received a similar amount of intravenous fluid per hour as did participant 6 when scaled for weight (approximately 2 ml [dot] kg [dot] hr sup -1). Measures directed toward the cardiovascular system, such as volume infusion or vasoactive drugs, may affect thoracic blood volume during EA and thus may influence the effect of EA on thoracic gas volumes.

Other factors contributed to an increased FRC in some of the persons we studied. In four of the six participants, the end-expiratory position of the rib cage moved outward, and in five of six the end-expiratory position of the diaphragm moved caudad. These increases in thoracic volumes could be caused by a decrease in lung elastic recoil, which was not measured. However, it is difficult to postulate a possible mechanism for such a decrease, directing attention toward direct effects of EA on the chest wall.

The rib cage could be expanded by the tonic activity in the SCA observed in some participants during high EA; however, there was no direct correspondence between the presence of such activity and rib cage expansion (see Table 2and Table 5). Other muscles inserting on the rib cage, such as the sternocleidomastoids, could have also contributed. In a previous study, we found that halothane anesthesia changed the curvature of the thoracic spine, which may have affected rib cage dimensions [31]; however, we saw no such changes in the present study (data not shown). Here, as in previous studies, [31,32]we found evidence for phasic expiratory activity in the lateral intercostal muscles; expiratory activity may also be present in the transversus thoracis muscle in supine persons when breathing quietly. [53]Because this activity constricts the rib cage during expiration, [54]its abolition should increase end-expiratory rib cage dimensions. This mechanism is partly responsible for the increase in FRC noted with high EA in dogs. [38].

If tonic activity is normally present in muscles that expand the rib cage, [4–6]such as the PS muscles, then paralysis of these muscles should decrease rib cage dimensions. A similar mechanism has been advanced to explain decreases in rib cage dimensions produced by general anesthesia. [4]However, many studies have failed to find evidence for significant tonic activity in rib cage muscles during quiet breathing in humans lying supine. [2,31,55,56]We also could find no evidence of tonic activity in the PS or SCA during intact breathing, with the exception of participant 5. The decrease in rib cage dimensions produced by high EA in this person could be explained by the abolition of tonic PS activity. However, the increase in rib cage dimensions produced in our study by high EA in four other participants is not consistent with the presence of such tonic activity.

The caudad displacement of the end-expiratory position of the diaphragm could be explained by an increase in tonic activation of the diaphragm with high EA. No evidence for this possibility was observed in EMG recordings of three participants. However, these recordings were obtained only from the costal portion of the diaphragm, and it is possible that tonic activity could have increased in the crural diaphragm. Decreases in abdominal pressure would also favor caudad diaphragm displacement. Two participants exhibited tonic abdominal muscle activity before EA, and block of this activity could have decreased abdominal pressure. Arndt and coworkers [9]found that EA decreased splanchnic blood volume in most participants, a factor that would also decrease abdominal pressure.

These measurements of chest wall function were obtained using dermatomal levels of block that are at the limit of those desired in clinical practice, using agents that provide substantial motor blockade. Thus they can be viewed as representing the maximal effect that might be expected during the conduct of lumbar EA in the clinical setting. Despite considerable block of the extradiaphragmatic muscles under these maximal conditions, most of these healthy persons maintained some degree of rib cage expansion when breathing quietly, with no significant effects on gas exchange as measured by arterial blood gases. This maintenance of gas exchange may have been aided by a concomitant increase in FRC. Unlike common clinical practice, most of these participants did not receive intravenous fluid boluses during the establishment of block. Data from the one participant who did receive such a bolus suggest that this practice may significantly affect intrathoracic blood volume, and hence the FRC.

Individual patients undergoing surgery may respond in different ways, as shown by the variability within our group of participants. One healthy participant in our study could not maintain inspiratory rib cage expansion, for reasons that are unknown. Patients with abnormal chest wall mechanics, such as the obese or those with chronic obstructive pulmonary disease, may rely more on rib cage expansion to generate tidal volume compared with normal participants, [57]and they may be more susceptible to rib cage muscle paralysis produced by EA. Additional effects of surgical stimulation must also be considered. For example, intraabdominal manipulation produces diaphragm dysfunction, probably due to reflex inhibition of the diaphragm. [58,59]During EA, the ability of the rib cage muscles to compensate for this inhibition may be impaired.

The authors thank Kathy Street and Darrell Loeffler for technical assistance, Janet Beckman for secretarial support, Dr. Brad Narr for performing preanesthetic medical evaluations, Drs. Bill Lichty and Michael Joyner for assistance with EMG techniques, Mike Rhyner and Don Erdman for operating the dynamic spatial reconstructor, and Dr. Jamil Tajik for assistance with ultrasonography.