Effect of diaphragmatic contraction on the action of the canine parasternal intercostals.

The inspiratory intercostal muscles enhance the force generated by the diaphragm during lung expansion. However, whether the diaphragm also alters the force developed by the inspiratory intercostals is unknown. Two experiments were performed in dogs to answer the question. In the first experiment, external, cranially oriented forces were applied to the different rib pairs to assess the effect of diaphragmatic contraction on the coupling between the ribs and the lung. The fall in airway opening pressure (deltaPa(O)) produced by a given force on the ribs was invariably greater during phrenic nerve stimulation than with the diaphragm relaxed. The cranial rib displacement (Xr), however, was 40-50% smaller, thus indicating that the increase in deltaPa(O) was exclusively the result of the increase in diaphragmatic elastance. In the second experiment, the parasternal intercostal muscle in the fourth interspace was selectively activated, and the effects of diaphragmatic contraction on the deltaPa(O) and Xr caused by parasternal activation were compared with those observed during the application of external loads on the ribs. Stimulating the phrenic nerves increased the deltaPa(O) and reduced the Xr produced by the parasternal intercostal, and the magnitudes of the changes were identical to those observed during external rib loading. It is concluded, therefore, that the diaphragm has no significant synergistic or antagonistic effect on the force developed by the parasternal intercostals during breathing. This lack of effect is probably related to the constraint imposed on intercostal muscle length by the ribs and sternum.     

The action of the canine diaphragm on the lower ribs depends on activation

Conventional wisdom maintains that the diaphragm lifts the lower ribs during isolated contraction. Recent studies in dogs have shown, however, that supramaximal, tetanic stimulation of the phrenic nerves displaces the lower ribs caudally and inward. In the present study, the hypothesis was tested that the action of the canine diaphragm on these ribs depends on the magnitude of muscle activation. Two experiments were performed. In the first, the C5 and C6 phrenic nerve roots were selectively stimulated in 6 animals with the airway occluded, and the level of diaphragm activation was altered by adjusting the stimulation frequency. In the second experiment, all the inspiratory intercostal muscles were severed in 7 spontaneously breathing animals, so that the diaphragm was the only muscle active during inspiration, and neural drive was increased by a succession of occluded breaths. The changes in airway opening pressure and the craniocaudal displacements of ribs 5 and 10 were measured in each animal. The data showed that 1) contraction of the diaphragm causes the upper ribs to move caudally; 2) during phrenic nerve stimulation, the lower ribs move cranially when the level of diaphragm activation is low, but they move caudally when the level of muscle activation is high and the entire rib cage is exposed to pleural pressure; and 3) during spontaneous diaphragm contraction, however, the lower ribs always move cranially, even when neural drive is elevated and the change in pleural pressure is large. It is concluded that the action of the diaphragm on the lower ribs depends on both the magnitude and the mode of muscle activation. These findings can reasonably explain the apparent discrepancies between previous studies. They also imply that observations made during phrenic nerve stimulation do not necessarily reflect the physiological action of the diaphragm.     

Role of pleural pressure in the coupling between the intercostal muscles and the ribs.

The inspiratory intercostal muscles elevate the ribs and thereby elicit a fall in pleural pressure (DeltaPpl) when they contract. In the present study, we initially tested the hypothesis that this DeltaPpl does, in turn, oppose the rib elevation. The cranial rib displacement (Xr) produced by selective activation of the parasternal intercostal muscle in the fourth interspace was measured in dogs, first with the rib cage intact and then after DeltaPpl was eliminated by bilateral pneumothorax. For a given parasternal contraction, Xr was greater after pneumothorax; the increase in Xr per unit decrease in DeltaPpl was 0.98+/-0.11 mm/cmH2O. Because this relation was similar to that obtained during isolated diaphragmatic contraction, we subsequently tested the hypothesis that the increase in Xr observed during breathing after diaphragmatic paralysis was, in part, the result of the decrease in DeltaPpl, and the contribution of the difference in DeltaPpl to the difference in Xr was determined by using the relation between Xr and DeltaPpl during passive inflation. With diaphragmatic paralysis, Xr during inspiration increased approximately threefold, and 47+/-8% of this increase was accounted for by the decrease in DeltaPpl. These observations indicate that 1) DeltaPpl is a primary determinant of rib motion during intercostal muscle contraction and 2) the decrease in DeltaPpl and the increase in intercostal muscle activity contribute equally to the increase in inspiratory cranial displacement of the ribs after diaphragm paralysis.     

Pleural pressure increases during inspiration in the zone of apposition of diaphragm to rib cage

The zone of apposition of diaphragm to rib cage provides a theoretical mechanism that may, in part, contribute to rib cage expansion during inspiration. Increases in intra-abdominal pressure (Pab) that are generated by diaphragmatic contraction are indirectly applied to the inner rib cage wall in the zone of apposition. We explored this mechanism, with the expectation that pleural pressure in this zone (Pap) would increase during inspiration and that local transdiaphragmatic pressure in this zone (Pdiap) must be different from conventionally determined transdiaphragmatic pressure (Pdi) during inspiration. Direct measurements of Pap, as well as measurements of pleural pressure (Ppl) cephalad to the zone of apposition, were made during tidal inspiration, during phrenic stimulation, and during inspiratory efforts in anesthetized dogs. Pab and esophageal pressure (Pes) were measured simultaneously. By measuring Ppl's with cannulas placed through ribs, we found that Pap consistently increased during both maneuvers, whereas Ppl and Pes decreased. Whereas changes in Pdi of up to -19 cmH2O were measured, Pdiap never departed from zero by greater than -4.5 cmH2O. We conclude that there can be marked regional differences in Ppl and Pdi between the zone of apposition and regions cephalad to the zone. Our results support the concept of the zone of apposition as an anatomic region where Pab is transmitted to the interior surface of the lower rib cage.     

Mechanism of the lung-deflating action of the canine diaphragm at extreme lung inflation

When lung volume in animals is passively increased beyond total lung capacity (TLC; transrespiratory pressure = +30 cmH(2)O), stimulation of the phrenic nerves causes a rise, rather than a fall, in pleural pressure. It has been suggested that this was the result of inward displacement of the lower ribs, but the mechanism is uncertain. In the present study, radiopaque markers were attached to muscle bundles in the midcostal region of the diaphragm and to the tenth rib pair in five dogs, and computed tomography was used to measure the displacement, length, and configuration of the muscle and the displacement of the lower ribs during relaxation at seven different lung volumes up to +60 cmH(2)O transrespiratory pressure and during phrenic nerve stimulation at the same lung volumes. The data showed that 1) during phrenic nerve stimulation at 60 cmH(2)O, airway opening pressure increased by 1.5 ± 0.7 cmH(2)O; 2) the dome of the diaphragm and the lower ribs were essentially stationary during such stimulation, but the muscle fibers still shortened significantly; 3) with passive inflation beyond TLC, an area with a cranial concavity appeared at the periphery of the costal portion of the diaphragm, forming a groove along the ventral third of the rib cage; and 4) this area decreased markedly in size or disappeared during phrenic stimulation. It is concluded that the lung-deflating action of the isolated diaphragm beyond TLC is primarily related to the invaginations in the muscle caused by the acute margins of the lower lung lobes. These findings also suggest that the inspiratory inward displacement of the lower ribs commonly observed in patients with emphysema (Hoover's sign) requires not only a marked hyperinflation but also a large fall in pleural pressure.     

Determinants of diaphragm motion in unilateral diaphragmatic paralysis.

Cranial displacement of a hemidiaphragm during sniffs is a cardinal sign of unilateral diaphragmatic paralysis in clinical practice. However, we have recently observed that isolated stimulation of one phrenic nerve in dogs causes the contralateral (inactive) hemidiaphragm to move caudally. In the present study, therefore, we tested the idea that, in unilateral diaphragmatic paralysis, the pattern of inspiratory muscle contraction plays a major role in determining the motion of the inactive hemidiaphragm. We induced a hemidiaphragmatic paralysis in six anesthetized dogs and assessed the contour of the diaphragm during isolated unilateral phrenic nerve stimulation and during spontaneous inspiratory efforts. Whereas the inactive hemidiaphragm moved caudally in the first instance, it moved cranially in the second. The parasternal intercostal muscles were then severed to reduce the contribution of the rib cage muscles to inspiratory efforts and to enhance the force generated by the intact hemidiaphragm. Although the change in pleural pressure (DeltaPpl) was unaltered, the cranial displacement of the paralyzed hemidiaphragm was consistently reduced. A pneumothorax was finally induced to eliminate DeltaPpl during unilateral phrenic nerve stimulation, and this enhanced the caudal displacement of the inactive hemidiaphragm. These observations indicate that, in unilateral diaphragmatic paralysis, the motion of the inactive hemidiaphragm is largely determined by the balance between the force related to DeltaPpl and the force generated by the intact hemidiaphragm.     

The effect of lung inflation on the inspiratory action of the canine parasternal intercostals.

Inflation induces a marked decrease in the lung-expanding ability of the diaphragm, but its effect on the parasternal intercostal muscles is uncertain. To assess this effect, the phrenic nerves and the external intercostals were severed in anesthetized, vagotomized dogs, such that the parasternal intercostals were the only muscles active during inspiration, and the endotracheal tube was occluded at different lung volumes. Although the inspiratory electromyographic activity recorded from the muscles was constant, the change in airway opening pressure decreased with inflation from -7.2+/-0.6 cmH2O at functional residual capacity to -2.2+/-0.2 cmH2O at 20-cmH2O transrespiratory pressure (P<0.001). The inspiratory cranial displacement of the ribs remained virtually unchanged, and the inspiratory caudal displacement of the sternum decreased moderately. However, the inspiratory outward rib displacement decreased markedly and continuously; at 20 cmH2O, this displacement was only 23+/-2% of the value at functional residual capacity. Calculations based on this alteration yielded substantial decreases in the change in airway opening pressure. It is concluded that, in the dog, 1) inflation affects adversely the lung-expanding actions of both the parasternal intercostals and the diaphragm; and 2) the adverse effect of inflation on the parasternal intercostals is primarily related to the alteration in the kinematics of the ribs. As a corollary, it is likely that hyperinflation also has a negative impact on the parasternal intercostals in patients with chronic obstructive pulmonary disease.     

Triangularis sterni: a primary muscle of breathing in the dog

The isolated action, pattern of neural activation, and mechanical contribution to eupnea of the triangularis sterni (transversus thoracis) muscle were studied in supine anesthetized dogs. Linear displacement transducers were used to measure the axial displacements of the ribs and sternum. Tetanic stimulation of the triangularis sterni in the apneic animal caused a marked caudal displacement of the ribs, a moderate cranial displacement of the sternum, and a decrease in lung volume. During quiet breathing, there was invariably a rhythmic activation of the muscle in phase with expiration that was independent of the presence or absence of activity in the abdominal and internal interosseous intercostal muscles. This phasic expiratory activity in the triangularis sterni was of large amplitude and caused the ribs to be more caudal and the sternum to be more cranial during the spontaneous expiratory pause than during relaxation. Additional studies on awake animals showed that rhythmic activation of the triangularis sterni occurs in all body positions and is not caused by anesthesia. These findings indicate that expiration in the dog is not a passive process and that the end-expiratory volume of the rib cage is not determined by an equilibrium of static forces alone. Rather, it is actively determined and maintained below its relaxation volume by contraction of the triangularis sterni throughout expiration. The use of this muscle is likely to facilitate inspiration by increasing the length of the parasternal intercostals and taking on a portion of their work.     

Inspiratory elevation of the ribs in the dog: primary role of the parasternals

To assess the relative contributions of the different groups of inspiratory intercostal muscles to the cranial motion of the ribs in the dog, we have measured the axial displacement of the fourth rib and recorded the electromyograms of the parasternal intercostal, external intercostal, and levator costae in the third interspace in 15 anesthetized animals breathing at rest. In eight animals, the parasternal intercostals were denervated in interspaces 1-5. This procedure caused a marked increase in the amount of external intercostal and levator costae inspiratory activity, and yet the inspiratory cranial motion of the rib was reduced by 55%. On the other hand, the external intercostals in interspaces 1-5 were sectioned in seven animals, and the reduction in the cranial rib motion was only 22%; the amount of parasternal and levator costae activity, however, was unchanged. When the parasternals in these animals were subsequently denervated, the levator costae inspiratory activity increased markedly, but the inspiratory cranial motion of the rib was abolished or reversed into an inspiratory caudal motion. These studies thus confirm that, in the dog breathing at rest, the parasternal intercostals have a larger role than the external intercostals and levator costae in causing the cranial motion of the ribs during inspiration. A quantitative analysis suggests that the parasternal contribution is approximately 80%.     

Dysfunction of the canine respiratory muscle pump in ascites.

Ascites, a complicating feature of many diseases of the liver and peritoneum, commonly causes dyspnea. The mechanism of this symptom, however, is uncertain. In the present study, progressively increasing ascites was induced in anesthetized dogs, and the hypothesis was initially tested that ascites increases the impedance on the diaphragm and, so, adversely affects the lung-expanding action of the muscle. Ascites produced a gradual increase in abdominal elastance and an expansion of the lower rib cage. Concomitantly, the caudal displacement of the diaphragm and the fall in airway opening pressure during isolated stimulation of the phrenic nerves decreased markedly; transdiaphragmatic pressure during phrenic stimulation also decreased. To assess the adaptation to ascites of the respiratory system overall, we subsequently measured the changes in lung volume, the arterial blood gases, and the electromyogram of the parasternal intercostal muscles during spontaneous breathing. Tidal volume and minute ventilation decreased progressively as ascites increased, leading to an increase in arterial PCO2 and parasternal intercostal inspiratory activity. It is concluded that 1) ascites, acting through an increase in abdominal elastance and an expansion of the lower rib cage, impairs the lung-expanding action of the diaphragm; 2) this impairment elicits a compensatory increase in neural drive to the inspiratory muscles, but the compensation is not sufficient to maintain ventilation; and 3) dyspnea in this setting results in part from the dissociation between increased neural drive and decreased ventilation.     

Inspiratory function of the levator costae and external intercostal muscles in the dog

The shortening of the canine parasternal intercostals during inspiration may have a passive component, and we have previously speculated that this might result from the actions of the levator costae and external intercostals (J. Appl. Physiol. 66: 1421-1429, 1989). The present studies were designed, therefore, to evaluate the pattern of activation of these muscles in the dog and to define their action on the rib cage during breathing. The results indicate that 1) the levator costae and external intercostals in the cranial part of the rib cage are active during inspiration, both in the supine and in the prone posture; 2) the inspiratory activation of the two muscles is increased after bilateral phrenicotomy; 3) it is increased even more when the parasternal intercostals in the different interspaces are also denervated; and 4) when the levator costae and external intercostals are the only muscles active during inspiration, the ribs continue to move cranially, and the sternum, rather than moving caudally as it does in the intact animal, moves cranially as well. Therefore, we conclude that the levator costae and external intercostals in the dog have a true inspiratory function. When needed, they are capable of causing a significant expansion of the rib cage and the lung during breathing.     

Modeling the kinematics of the canine midcostal diaphragm

The hypotheses that the chest wall insertion (CW) is displaced laterally during inspiration and that this displacement is essential in maintaining muscle curvature of the costal diaphragmatic muscle fibers were tested. With the use of data from three dogs, caudal, lateral, and ventral displacements of CW during both quiet, spontaneous inspiration and during inspiratory efforts against an occluded airway were observed and recorded. We have developed a kinematic model of the diaphragm that incorporates these displacements. This model describes the motions of the muscle fibers and central tendon; the displacements of the midplane, muscle-tendon junction (MTJ), CW, and center of the muscle fiber-central tendon arcs are modeled as functions of muscle fiber length. In the model, the center of the fiber arcs and MTJ both move caudally parallel to the midplane during inspiration, whereas CW moves both caudally and laterally. The observed lateral displacement of CW and the observed caudal displacement of MTJ, as functions of muscle fiber length, both approximate well the theoretical displacements that would be necessary to maintain curvature of the fiber arcs. In confirming our hypotheses, we have found that lateral displacement of CW is a mechanism by which changes in the shape of the costal diaphragm, as described by its curvature, are limited.     

Interaction between the canine diaphragm and intercostal muscles in lung expansion.

Changes in intrathoracic pressure produced by the various inspiratory intercostals are essentially additive, but the interaction between these muscles and the diaphragm remains uncertain. In the present study, this interaction was assessed by measuring the changes in airway opening (DeltaPao) or transpulmonary pressure (DeltaPtp) in vagotomized, phrenicotomized dogs during spontaneous inspiration (isolated intercostal contraction), during isolated rectangular or ramp stimulation of the peripheral ends of the transected C(5) phrenic nerve roots (isolated diaphragm contraction), and during spontaneous inspiration with superimposed phrenic nerve stimulation (combined diaphragm-intercostal contraction). With the endotracheal tube occluded at functional residual capacity, DeltaPao during combined diaphragm-intercostal contraction was nearly equal to the sum of the DeltaPao produced by the two muscle groups contracting individually. However, when the endotracheal tube was kept open, DeltaPtp during combined contraction was 123% of the sum of the individual DeltaPtp (P < 0.001). The increase in lung volume during combined contraction was also 109% of the sum of the individual volume increases (P < 0.02). Abdominal pressure during combined contraction was invariably lower than during isolated diaphragm contraction. It is concluded, therefore, that the canine diaphragm and intercostal muscles act synergistically during lung expansion and that this synergism is primarily due to the fact that the intercostal muscles reduce shortening of the diaphragm. When the lung is maintained at functional residual capacity, however, the synergism is obscured because the greater stiffness of the rib cage during diaphragm contraction enhances the DeltaPao produced by the isolated diaphragm and reduces the DeltaPao produced by the intercostal muscles.     

Mechanics of the parasternal intercostals during occluded breaths in dogs

The electrical activity and the respiratory changes in length of the third parasternal intercostal muscle were measured during single-breath airway occlusion in 12 anesthetized, spontaneously breathing dogs in the supine posture. During occluded breaths in the intact animal, the parasternal intercostal was electrically active and shortened while pleural pressure fell. In contrast, after section of the third intercostal nerve at the chondrocostal junction and abolition of parasternal electrical activity, the muscle always lengthened. This inspiratory muscle lengthening must be related to the fall in pleural pressure; it was, however, approximately 50% less than the amount of muscle lengthening produced, for the same fall in pleural pressure, by isolated stimulation of the phrenic nerves. These results indicate that 1) the parasternal inspiratory shortening that occurs during occluded breaths in the dog results primarily from the muscle inspiratory contraction per se, and 2) other muscles of the rib cage, however, contribute to this parasternal shortening by acting on the ribs or the sternum. The present studies also demonstrate the important fact that the parasternal inspiratory contraction in the dog is really agonistic in nature.     

Partitioning of inspiratory pressure swings between diaphragm and intercostal/accessory muscles

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 (Pic). If correct, Pic 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 during Mueller maneuvers in which deltaPab = 0. Because there was no outward displacement of the rib cage, Pic must have contributed to deltaPpl, as did Pdi. Under these conditions the total pressure developed by the inspiratory muscles across the rib-cage pathway was less than Pdi + Pic. 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.     

Mechanics of the canine diaphragm in pleural effusion

Pleural effusion is a complicating feature of many diseases of the lung and pleura, but its effects on the mechanics of the diaphragm have not been assessed. In the present study, radiopaque markers were attached along muscle bundles in the midcostal region of the diaphragm in anesthetized dogs, and the three-dimensional location of the markers during relaxation before and after the stepwise introduction of liquid into the left or right pleural space and during phrenic nerve stimulation in the same conditions was determined using computed tomography. From these data, accurate measurements of diaphragm muscle length and displacement were obtained, and the changes in pleural and abdominal pressure were analyzed as functions of these parameters. The effect of liquid instillation on the axial position of rib 5 was also measured. The data showed that 1) liquid leaked through the dorsal mediastinal sheet behind the pericardium so that effusion was bilateral; 2) effusion caused a caudal displacement of the relaxed diaphragm; 3) this displacement was, compared with passive lung inflation, much larger than the cranial displacement of the ribs; and 4) the capacity of the diaphragm to generate pressure, in particular pleural pressure, decreased markedly as effusion increased, and this decrease was well explained by the decrease in active muscle length. It is concluded that pleural effusion has a major adverse effect on the pressure-generating capacity of the diaphragm and that this is the result of the action of hydrostatic forces on the muscle.     

Action of neck accessory muscles on rib cage in dogs

To investigate the action of the neck accessory muscles on the rib cage, we stimulated the sternocleidomastoid and the scalenus muscles separately in supine anesthetized dogs. Hooks screwed into the sternum and ribs were used to measure their axial displacements and the changes in anteroposterior (AP) and transverse (T) diameters of the rib cage. We found that the sternocleidomastoid and scalenus muscles, when they contract alone, cause a large axial displacement of the sternum and the ribs in a cephalad direction and expand the rib cage along both its AP and T diameters. Opening the abdomen increased the cephalad displacement of the ribs and the expansion of the lower rib cage, particularly along its T diameter, but reduced the increase in lung volume. These experiments indicate 1) that the action of the sternocleidomastoid and scalenus muscles on the rib cage is essentially the consequence of a rotation of the ribs' neck axes, resulting from the cephalad displacement of the ribs, and 2) that the fall in abdominal pressure, almost certainly by acting through the zone of apposition of the diaphragm to the rib cage, has a deflationary action on the lower rib cage, more markedly so on its lateral than its anterior wall. The experiments also suggest that the fall in abdominal pressure prevents the diaphragm from moving cephalad and aids the neck accessory muscles in inflating the lungs.     

The two mechanisms of intercostal muscle action on the lung.

The mechanisms of respiratory action of the intercostal muscles were studied by measuring the effect of external forces (F) applied to the ribs and by modeling the effect of F exerted by the intercostal muscles. In five dogs, with the airway occluded, cranial F were applied to individual rib pairs, from the 2nd to the 11th rib pair, and the change in airway opening pressure (Pao) was measured. The ratio Pao/F increases with increasing rib number in the upper ribs (2nd to 5th) and decreases in the lower ribs (5th to 11th). These data were incorporated into a model for the geometry of the ribs and intercostal muscles, and Pao/F was calculated from the model. For interspaces 2-8, the calculated values agree reasonably well with previously measured values. From the modeling, two mechanisms of intercostal muscle action are identified. One is the well-known Hamberger mechanism, modified to account for the three-dimensional geometry of the rib cage. This mechanism depends on the slant of an intercostal muscle relative to the ribs and on the resulting difference between the moments applied to the upper and lower ribs that bound each interspace. The second is a new mechanism that depends on the difference between the values of Pao/F for the upper and lower ribs.     

Linkage between parasternals and external intercostals during resting breathing

To assess the mechanical coupling between the parasternal and external intercostals in the cranial portion of the rib cage, we measured the respiratory changes in length and the electromyograms of the two muscles in the same third or fourth intercostal space in 24 spontaneously breathing dogs. We found that 1) the amount of inspiratory shortening of the external intercostal was considerably smaller than the amount of shortening of the parasternal; 2) after selective denervation of the parasternal, the inspiratory shortening of both the parasternal and the external intercostal was almost abolished; 3) on the other hand, after selective denervation of the external intercostal, the inspiratory shortening of the parasternal was unchanged, and the inspiratory shortening of the external intercostal was reduced but not suppressed; and 4) this persistent shortening of the external intercostal was reversed into a clear-cut inspiratory lengthening when the parasternal was subsequently denervated. We conclude that in the dog 1) the inspiratory contraction of the external intercostals in the cranial portion of the rib cage is agonistic in nature as is the contraction of the parasternals; 2) during resting breathing, however, the changes in length of these external intercostals are largely determined by the action of the parasternals. These observations are consistent with the idea that in the dog, the parasternals play a larger role than the external intercostals in elevating the ribs during resting inspiration.     

Gas exchange during separate diaphragm and intercostal muscle breathing.

In patients with diaphragm paralysis, ventilation to the basal lung zones is reduced, whereas in patients with paralysis of the rib cage muscles, ventilation to the upper lung zones in reduced. Inspiration produced by either rib cage muscle or diaphragm contraction alone, therefore, may result in mismatching of ventilation and perfusion and in gas-exchange impairment. To test this hypothesis, we assessed gas exchange in 11 anesthetized dogs during ventilation produced by either diaphragm or intercostal muscle contraction alone. Diaphragm activation was achieved by phrenic nerve stimulation. Intercostal muscle activation was accomplished by electrical stimulation by using electrodes positioned epidurally at the T(2) spinal cord level. Stimulation parameters were adjusted to provide a constant tidal volume and inspiratory flow rate. During diaphragm (D) and intercostal muscle breathing (IC), mean arterial Po(2) was 97.1 +/- 2.1 and 88.1 +/- 2.7 Torr, respectively (P < 0.01). Arterial Pco(2) was lower during D than during IC (32.6 +/- 1.4 and 36.6 +/- 1.8 Torr, respectively; P < 0.05). During IC, oxygen consumption was also higher than that during D (0.13 +/- 0.01 and 0.09 +/- 0.01 l/min, respectively; P < 0.05). The alveolar-arterial oxygen difference was 11.3 +/- 1.9 and 7.7 +/- 1.0 Torr (P < 0.01) during IC and D, respectively. These results indicate that diaphragm breathing is significantly more efficient than intercostal muscle breathing. However, despite marked differences in the pattern of inspiratory muscle contraction, the distribution of ventilation remains well matched to pulmonary perfusion resulting in preservation of normal gas exchange.