INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE DIAPHRAGM IS DESIGNED like a piston powered by two muscles, costal and crural, on either side of a noncontractile central tendon that acts as the piston head (Fig. 1) (30). The two muscles have been described to have different embryological origins, segmental innervations, and mechanical functions (6, 7). Territories of motor innervation within and between the two muscles are sufficiently localized for precise control of local responses to mechanical loading (11). Our research has focused on how these apparently different muscles work together to support high levels of ventilation. At rest, blood flow per gram of muscle, neural activation, and muscle fiber shortening with each breath are not uniformly distributed between or within the two muscles (5, 8, 9). This was presumed to reflect uneven mechanical loading at rest that did not require full effort from all regions of the diaphragm. In aerobic quadrupeds, the diaphragm becomes complexly loaded during exercise not only by high levels of ventilation but also by inertial forces shifting abdominal contents back and forth against the lower rib cage and diaphragm (3, 4). We expected that efficient design would require full recruitment of all parts of the diaphragm during heavy exercise and that muscle mass should be distributed in proportion to energy requirements. We designed an experiment to examine this hypothesis.

View larger version (64K): [in this window] [in a new window]   Fig. 1.   Surface of a dog diaphragm viewed by looking dorsally from inside the rib cage [redrawn after Zietzchmann et al. (30)]. The diaphragm is composed of 2 muscles, costal and crural, separated by a central tendon. The tendon lies in the dome and acts as a piston head during breathing. When the 2 muscles contract, the dome is pulled dorsally and fills the lungs with air.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Six adult male foxhounds (mean body weight 32 kg) were trained to run on a treadmill while wearing a leak-free respiratory mask (1) connected to a one-way valve (model 2700, Hans-Rudolph, Kansas City, MO). Each dog had bilateral carotid loops constructed under general anesthesia for subsequent catheterization studies. During a study, the inspired port of the mask was connected to a screen pneumotachometer (model 3813, Hans-Rudolph) for measuring inspired flow and the expired port was connected to a heated screen pneumotachometer for expired flow and calibrated for volume measurement by the method of Yeh et al. (29). Expired O₂, CO₂, and N₂ were measured continuously distal to a mixing chamber. An esophageal balloon was inserted into the lower third of the esophagus through a port in the mask for measuring transpulmonary pressure by using a differential Validyne transducer. Pressure-volume loops were recorded continuously for calculating work of breathing per minute on both lungs and breathing apparatus (13). Expired ventilation, O₂ uptake, CO₂ output, and work of ventilation were averaged over each 20 breaths and monitored continuously on a computer screen. We measured regional diaphragmatic blood flow by a fluorescent-microsphere technique (10) at rest and three different treadmill workloads up to ~65% of maximal O₂ uptake (12, 14) (Table 1). Briefly, under local anesthesia, a cardiac catheter was threaded percutaneously through the carotid artery into the left ventricle and fixed in place. A venous catheter (Cook, Bloomington, IN) was inserted in the other carotid loop for collecting reference blood after microsphere injections. At each workload, after a steady state was reached, 3 × 10⁶ latex microspheres (15 µm in diameter labeled with a fluorescent marker) were injected into the left ventricle. Reference blood was collected from the other carotid catheter at a fixed sampling rate, beginning just before the microsphere bolus and continuing for 2 min after the injection. Microspheres were distributed and trapped in the microcirculation of different organs in proportion to regional blood flow. At each workload, a different fluorescent marker was used. At postmortem, costal diaphragm was subdivided into eight regions and crural diaphragm into five regions (Fig. 2). Levels of fluorescence trapped within each region were used to estimate regional blood flow. Support for adequate mixing of injected microspheres was provided by blood flow to the right and left kidney at each workload.

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Subdivision of costal and crural diaphragm into regions for blood flow measurement.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Contrary to expectations, specific blood flow (per gram of muscle) remained unevenly distributed in the same fixed pattern from rest to heavy exercise (Fig. 3). At each workload, a similar dorsal-to-ventral gradient of increasing specific blood flow occurred in adjacent regions of costal and crural muscles up to the middiaphragm where the crural muscle ends. Ventral to this point, specific blood flow in the costal muscle progressively declined toward the sternum. Anatomically, costal and crural muscle fibers are aligned along the same directions across the dorsal arms of the central tendon (Fig. 1); hence, the two muscles must operate against a similar load. This is supported by the observation that specific blood flow to costal and crural muscles facing each other across the dorsal arms of the central tendon are approximately matched. To compare gradients across workloads, we normalized blood flow data in each region with respect to that in costal region 7 and superimposed the results (Fig. 4). The dorsal-to-ventral pattern of uneven blood flow per gram of muscle remained the same regardless of workload. Data across the four different workloads were compared by repeated-measures ANOVA. Regional differences in blood flow in the dorsal-to-ventral directions from rest to exercise in costal and crural muscles were highly significant statistically (P < 0.0001). Differences in recruitment across workloads from rest to heavy exercise were not significant (P = 0.974), nor were there significant interactions among workloads (P = 0.460). Hence patterns of blood flow recruitment were the same at all workloads. Microsphere mixing in the left ventricle was adequate as evidenced by the fact that right and left renal blood flow was not significantly different at rest or exercise [290 ± 46 and 268 ± 39 (SE) ml/min, respectively] at peak exercise.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Regional blood flow per gram of muscle (d) in costal and crural diaphragm. Specific d in the costal diaphragm plotted with respect to region increases in a dorsal-to-ventral direction from region 8 to region 5, followed by a gentle decline from region 5 to region 1 at the sternum. At each workload, the dorsal-to-ventral increase in specific d in the crural diaphragm, region E to region A, parallels that in adjacent dorsal costal diaphragm matched with similar fiber directions. The 2 muscles work together to pull the central tendon down and fill the lungs with air.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Comparison of specific d in the costal and crural diaphragm normalized as a ratio to that measured in costal region 7 at each workload. The pattern of regional diaphragmatic recruitment remains the same from rest to heavy exercise in all regions. Solid symbols and solid lines, costal diaphragm. Open symbols and dashed lines, crural diaphragm. Work level is represented by different symbols: , , , and , rest, light, medium, and heavy exercise, respectively. Dorsal-to-ventral differences in regional d recruitment from rest to heavy exercise were highly significant statistically (P < 0.0001), but patterns of d recruitment at different workloads were not significantly different.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

On the basis of blood flow data, there were no differences in the pattern of muscle recruitment between costal and crural diaphragm at rest or exercise; the two muscles behaved as a single unit. Heterogeneity of blood flow only occurred in the dorsal-to-ventral direction in both costal and crural muscles; this dorsal-to-ventral heterogeneity remained fixed at all levels of exercise. Specific blood flow was approximately twofold greater in the midcostal and ventral crural diaphragm than in the dorsal regions of either. Distribution of muscle capillary densities and fiber types is similar throughout the dog diaphragm (20, 21); hence, differences in regional capacity for O₂ uptake and utilization cannot explain these gradients. Similar fixed dorsal-to-ventral patterns of blood flow heterogeneity per gram of muscle have been described in the costal and crural diaphragm of small rodents (19, 22, 23). In these animals, specific blood flow is higher in the dorsal than ventral regions, opposite to what we describe in dogs. Nevertheless the pattern of heterogeneity remains fixed from rest to heavy exercise, even though oxidative capacity is uniformly distributed in the two muscles. There is heterogeneity in blood flow between whole costal and whole crural diaphragms at rest and exercise in ponies similar to that which we have measured in foxhounds, but regional blood flow within each muscle has not been measured in ponies for comparison (17). Blood flow heterogeneity may be a general feature in the mammalian diaphragm with variable patterns among different species.

If it were not for the insight afforded by the studies of Wilson and De Troyer (27, 28) on mechanical advantage of respiratory muscles (µ), we would be compelled to conclude that either muscle mass or blood flow is inefficiently distributed and utilized in the diaphragm. Wilson and De Troyer introduced the important concept that the ability of a respiratory muscle to generate the alveolar pressures required for ventilation is determined not only by the tension developed in the muscle but also by the regional µ of the muscle. Regional µ is a function of regional anatomy and defined as the alveolar pressure (P) that a muscle can generate by an active stress (T) during contraction against closed airways

(1)

where L_m is muscle length, L/L_m is fractional change in muscle length, VL is lung volume, V_m is muscle volume, and µ is in liters¹. The value of µ is solely a function of the muscle anatomy and constitutive properties. It can be measured from the strain induced in regional muscle fibers, L/L_m, by a passive increase in lung volume (VL) under relaxed conditions. The contribution to power generated per gram of muscle during an active volume change, on the basis of Eq. 1, would be as follows

(2)

where sp gr is specific gravity converting muscle volume to mass and dVL/dt is rate of change of lung volume. Theoretically regions with a high µ can deliver more power per gram of muscle at a given tension development. For optimal efficiency, energy loss from distortion of the thorax must be minimized by coordinated action of all interacting respiratory muscles in such a way that their passive undistorted configuration is maintained throughout a breath. Rates of contractile shortening in different regions must remain proportional to regional µ to prevent one region from being distorted by contractile tension developed in another region. Optimal distribution of power should be approached by preferential distribution of muscle mass and neural activation to those regions with a high µ. Given these assumptions, power output, O₂ uptake, and blood flow per gram of muscle should be distributed in direct proportion to regional µ rather than being uniformly distributed throughout the diaphragm.

The value of µ has been measured in the diaphragm. Boriek et al. (2) and Wilson et al. (26) estimated µ in different regions of the diaphragm of dogs during active as well as passive volume changes. They found a similar ventral-to-dorsal distribution of µ as we reported for specific blood flow in costal muscle, but they were unable to show a significant dorsal-to-ventral change in µ in crural muscle. Fiber directions in the crural muscle are not the same on the abdominal side of the diaphragm, where the radiopaque markers were placed, as on the thoracic side, leading to some ambiguity in interpretation. Wait et al. (24) perfusion fixed the left hemidiaphragm in dogs with 6.25% buffered glutaraldehyde either at functional residual capacity (FRC) or at total lung capacity (TLC) in the supine position and measured regional muscle thickness near TLC (T_TLC) and at FRC (T_FRC). Average µ values of the dorsal, middle, and ventral regions of the costal and crural diaphragm were estimated from their data as µ = (1  T_FRC/T_TLC)/VL (Fig. 5). The value of µ calculated from passive thickness changes in costal and crural muscles measured by Wait et al. followed a similar pattern of distribution as that of specific blood flow in the present studies.

View larger version (13K): [in this window] [in a new window]   Fig. 5.   A: dorsal-to-ventral distribution of mechanical advantage in the canine costal and crural diaphragm calculated from the data of Wait et al. (24). B: regional d averaged over all workloads, normalized to costal region 7, and plotted with respect to regional mechanical advantage. d is preferentially distributed to regions of high mechanical advantage.

Distribution of muscle mass among the costal and crural regions is shown in Fig. 6. Most of the muscle mass in both muscles is distributed in the midcostal and ventral crural regions in a similar distribution to regions of highest µ. We did not measure regional thickness of the diaphragm but data of Brancatisano et al. (5) and of Margulies (18) indicate a significant dorsal-to-ventral increase in thickness of both the costal and crural diaphragm. Total costal and crural muscle mass values are 102 ± 15 and 63 ± 8 (SD) g, respectively.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Regional distribution of muscle mass in the costal and crural diaphragm. Muscle mass in the costal diaphragm is preferentially distributed in midcostal regions 3, 4, 5, and 6 where mechanical advantage is greatest. Muscle mass in the crural diaphragm is preferentially distributed in ventral regions A and B where mechanical advantage is also highest. Muscle mass in the crural diaphragm (regions A, B, C, D, and E) equals the muscle mass of regions 4-8 of the costal diaphragm and face each other across the dorsal arms of the central tendon, providing balanced forces to maintain symmetry during contraction. Data were analyzed by repeated-measures ANOVA showing significant differences between regions in both costal and crural diaphragm (P < 0.0001). Pattern of mass distribution was the same in all dogs. * Significantly different from regions 2, 7, and B (P < 0.02).  Region B is significantly different from regions C, D, and E (P < 0.0001), and region A is significantly different from regions B, C, D, and E (P < 0.0001).

Collectively, the results are consistent with the following conclusions. 1) Costal and crural muscles are recruited as a single unit during ventilation. 2) Regional patterns of neural activation and blood flow in both muscles are anatomically fixed from rest to heavy exercise and 3) regional levels of neural activation and blood flow are matched to regional µ and not to regional muscle mass. 4) Muscle mass in the smaller costal diaphragm approximately matches the muscle mass of that part of the costal diaphragm that it faces across the dorsal arm of the central tendon; this should provide a balance of force generation to ensure symmetry during contraction.

There is good theoretical and experimental evidence that similar patterns of neural activation and blood flow distribution apply to all muscles of breathing, although none have been validated in the exercising animal (15, 16, 25). Nevertheless, combined data suggest that the efficient anatomic design of the thoracic pump reduces complexity of neural controls and minimizes power requirements of breathing and muscle mass required. These design features may help aerobic quadrupeds, such as the foxhound, to achieve exceedingly high levels of ventilation during pursuit and/or escape.