Journal of Applied Physiology Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J Appl Physiol 81: 1455-1468, 1996;
8750-7587/96 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hussain, S. N. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hussain, S. N. A.

Journal of Applied Physiology
Vol. 81, No. 4, pp. 1455-1468, October 1996

INVITED REVIEW

Regulation of ventilatory muscle blood flow

Sabah N. A. Hussain

Critical Care and Respiratory Divisions, Department of Medicine, Royal Victoria Hospital, McGill University, Montreal, Quebec H3A 1A1, Canada

ABSTRACT
INTRODUCTION
FOOTNOTES
REFERENCES

ABSTRACT

Hussain, Sabah N. A. Regulation of ventilatory muscle blood flow. J. Appl. Physiol. 81(4): 1455-1468, 1996.---The ventilatory muscles perform various functions such as ventilation of the lungs, postural stabilization, and expulsive maneuvers (e.g., coughing). They are classified in functional terms as inspiratory muscles, which include the diaphragm, parasternal intercostal, external intercostal, scalene, and sternocleidomastoid muscles; and expiratory muscles, which include the abdominal muscles, internal intercostal, and triangularis sterni. The ventilatory muscles require high-energy phosphate compounds such as ATP to fuel the biochemical and physical processes of contraction and relaxation. Maintaining adequate intracellular concentrations of these compounds depends on adequate intracellular substrate levels and delivery of these substrates by arterial blood flow. In addition to the delivery of substrates, blood flow influences muscle function through the removal of metabolic by-products, which, if accumulated, could exert negative effects on several excitatory and contractile processes. Skeletal muscle substrate utilization is also dependent on the ability to extract substrates from arterial blood, which, in turn, is accomplished by increasing the total number of perfused capillaries. It follows that matching perfusion to metabolic demands is critical for the maintenance of normal muscle contractile function. In this article, I review the factors that influence ventilatory muscle blood flow. Major emphasis is placed on the diaphragm because a large number of published reports deal with diaphragmatic blood flow. The second reason for focusing on the diaphragm is because it is the largest and most important inspiratory muscle.

inspiratory muscles; expiratory muscles; diaphragm; substrates

INTRODUCTION

VENTILATORY MUSCLE BLOOD FLOW is a function of two main variables: the driving pressure and the caliber of resistance vessels. Driving pressure, in turn, is determined by the difference between upstream or inflow pressure and downstream or outflow pressure. The caliber of resistance vessels is determined by two groups of factors: central factors, which include sympathetic adrenergic drive, and local factors, which mainly consist of two opposing mechanisms. The myogenic mechanism is pressure sensitive and couples pressure with vessel diameter. The metabolic vasoactive mechanism matches blood flow with local metabolic demands. The final state of muscle blood flow depends on complex interactions between all these factors.

Numerous investigators have measured ventilatory muscle blood flow in response to a variety of ventilatory, metabolic, and mechanical perturbations. This review is dedicated to a discussion of how these perturbations (namely, alterations in the pattern and degree of muscle activation, changes in arterial blood gas composition, augmentation of abdominal and pleural pressure swings, and modifications in muscle length and geometry) affect ventilatory muscle perfusion. It also addresses the role of adrenergic neural drive, blood-borne catecholamines, and locally released vasoactive substances in the regulation of ventilatory muscle perfusion.

Anatomy of the Diaphragmatic Circulation

Few investigators have assessed the differences and similarities in microcirculatory configurations between the diaphragm and other skeletal muscles. An early study indicated that the branching patterns of capillaries and arterioles in the diaphragm are similar to those of the intercostal and triceps muscles (106). In addition, no differences were detected in capillary distribution and configuration between the costal and crural diaphragmatic portions in dogs and rats (16, 97). However, Brancatisano et al. (19) reported that costal diaphragmatic blood flow in dogs exhibited a dorsoventral gradient that was independent of gravity, whereas no such gradient could be detected in the crural diaphragm. Because capillary density and configuration are similar in the two diaphragmatic segments (97), the dorsoventral gradient in costal blood flow can be attributed to a difference in muscle thickness between the dorsal and ventral portions of the costal diaphragm.

While microcirculatory configuration in the diaphragm may be similar to that in other skeletal muscles, the diaphragmatic vascular bed has a unique arrangement for arterial supply and venous drainage. The diaphragm is perfused by multiple arterial sources including the inferior phrenic, intercostal, and internal mammary arteries (26). In dogs and cats, the crural and sternal segments and most of the costal areas of the diaphragm are perfused by the inferior phrenic arteries (15, 105). In comparison, peripheral areas of the costal diaphragm adjacent to the ribs in dogs and humans are supplied by the intercostal arteries, whereas the anterior sternal segment of the costal diaphragm is perfused by the internal mammary arteries (26, 46). These arterial sources also form extensive intramuscular and extramuscular anastomoses in canine and human diaphragms (26, 28). On the other hand, there appears to be no arterial shunting between the two sides of the hemidiaphragms (28, 55). These intra- and extramuscular arterial anastomoses are designed to provide collateral sources to maintain adequate blood flow to diaphragmatic muscle fibers and to protect these fibers from ischemia. The resistance of the diaphragm to ischemia has been confirmed by several investigators (27, 75, 86). Most of the venous drainage of the diaphragmatic vascular bed in various species is routed through the inferior phrenic veins that join the inferior vena cava just distal to the hepatic veins (47). Local drainage of the costal insertion and the sternal portion of the diaphragm is channeled through the intercostal and internal mammary veins, respectively (39, 46, 83). In specific conditions such as the presence of a strong cast around the abdominal compartment and open thoracic cage, the bulk of diaphragmatic venous drainage goes through the intercostal veins, whereas only 25% of venous drainage is sent through the inferior phrenic veins (55).

Influence of Muscle Activity

Under normal conditions, metabolic demands are the most important determinant of ventilatory muscle blood flow. Numerous investigators have assessed the relationship between diaphragmatic metabolic demands and blood flow. In early experiments, changes in minute ventilation served as an index of ventilatory muscle metabolic demands. Using the radiolabeled microsphere technique to measure local blood flow, several investigators reported that diaphragmatic blood flow during mechanical ventilation in dogs averaged ~12 ml · 100 g-1 · min-1 (ranges between 4 and 18 ml · 100 g-1 · min-1) (100, 101, 121). During quiet breathing, this value increases to a mean value of ~16 ml · 100 g-1 · min-1 (range 8-22 ml · 100 g-1 · min-1). When minute ventilation is augmented during unobstructed CO2 rebreathing in anesthetized animals, a significant but variable rise in blood flow of the diaphragm and other inspiratory muscles has been reported (6, 84, 100, 101).

Ventilatory muscle blood flow has also been measured during physical exercise in different species under various exercise workloads. In exercising dogs performing moderate to maximal exercise on a treadmill, several investigators reported a 2- to 11-fold increase in diaphragmatic blood flow compared with resting values (40, 53, 89, 94, 104). A significant augmentation of diaphragmatic blood flow has even been reported in exercising pigs and rats (88, 104). Increased blood flow to the ventilatory muscle during exercise appears to be directly related to heightened muscle activation, as indicated by a linear relationship between ventilatory muscle blood flow and ventilation or systemic O2 uptake. Interestingly, Hsia et al. (53) found that the relationship between diaphragmatic blood flow and ventilation was similar during exercise and CO2 rebreathing. In contrast, blood flow to the parasternal intercostal muscle at a given ventilation was higher during exercise than during CO2 rebreathing, suggesting more vigorous recruitment of this muscle during exercise. Despite the substantial augmentation of metabolic demands, it is still unclear whether diaphragmatic blood flow reaches maximum values during physical exercise. In dogs exercising at maximum workloads, blood flow to the diaphragm remains lower than predicted maximum values based on the regression equation of diaphragmatic blood flow and mean arterial pressure (see below). This appears to be true even in pneumonectomized dogs (53). On the other hand, diaphragmatic blood flow in maximally exercising ponies reaches peak values ranging between 245 and 265 ml · 100 g-1 · min-1 (79-82). These values were similar to those reached by the main muscles of propulsion such as the biceps femoris and gluteus medius. In these experiments, systemic infusion of adenosine during exercise did not result in a further increase in diaphragmatic blood flow, suggesting that maximal dilating capacity is attained in the diaphragm during severe exercise. These results also indicate that the absence of any vasodilator reserve in the diaphragm of maximally exercising ponies may constitute a limiting factor to prevent further exertion in this species.

Augmentation of ventilatory muscle metabolic demands during loaded breathing has also been shown to elicit a moderate rise in blood flow to the diaphragm and other inspiratory muscles (74, 99, 101). The highest rise in ventilatory muscle blood flow during resistive loading was reported in conscious sheep (93). Peak parasternal intercostal and diaphragmatic blood flows reached approx 340 and 280 ml · 100 g-1 · min-1, respectively, in response to severe inspiratory resistive loading. Figure 1 illustrates the main findings of Pang et al. (93). Notice the substantial rise in parasternal and diaphragmatic blood flow during resistive loading. These results suggest that in conscious animals parasternal intercostal perfusion may exceed that of the diaphragm. Unlike unobstructed hyperventilation, the relationship between ventilatory muscle blood flow and minute ventilation is poor during resistive loading, but a better relationship was found between diaphragmatic blood flow and pleural pressure tension-time index (TTdi; integral of peak inspiratory pleural pressure swing with respect to time) (93, 101). The highest ventilatory muscle blood flow values were reported during artificial electrical stimulation. For instance, Buchler et al. (23) reported peak diaphragmatic blood flow values of 273 ± 13 ml · 100 g-1 · min-1 during maximal intermittent stimulation of the phrenic nerves in dogs.

Fig. 1. Changes in blood flow to inspiratory muscles during severe inspiratory flow resistive-loaded breathing (n = 7). There is a significant increase in blood flow at all points during loaded breathing compared with baseline for muscles shown. * P < 0.05 compared with point B. Ext. IC, external intercostal muscles. [From Pang et al. (93).]
[View Larger Version of this Image (25K GIF file)]

Effects of Force and Pattern of Contraction

It has been more than 100 years since the discovery that blood flow to skeletal muscles may be hindered during contraction by mechanical compression of the vessels. Limitation of ventilatory muscle blood flow during muscle contraction was first reported by Anrep and co-workers (2, 3). It was reported later that absolute diaphragmatic blood flow peaked at sustained isometric tensions as high as 80% of maximum, with a progressive decline thereafter (35). The first comprehensive report quantifying the relationship between diaphragmatic blood flow and the pattern of muscle activation was published by Bellemare et al. (14), who introduced the TTdi concept, which is the product of normalized force or transdiaphragmatic pressure (Pdi) [as %maximum force (Pdimax)] and the duty cycle (ratio of contraction time to total duration of the contraction-relaxation cycle). These authors described a parabolic relationship between TTdi and diaphragmatic blood flow with the latter increasing as TTdi rises until a TTdi of 0.20 is reached. With further increase in TTdi, diaphragmatic blood flow begins to decline and postcontraction reactive hyperemia begins to increase, suggesting that significant flow obstruction occurs when the TTdi reaches 0.20. During sustained diaphragmatic contractions (duty cycle of 1), diaphragmatic blood flow becomes limited at tensions equivalent to 20% of maximum. Figure 2 illustrates the relationship between diaphragmatic blood flow and TTdi described by Bellemare et al. (14). It was reported later that the decline in diaphragmatic blood flow at high TTdi values is not a unique function of TTdi but the pattern of contraction (13). Indeed, while blood flow was curvilinearly and inversely related to the increase in tension, it was linearly and inversely related to the rise in duty cycle. Multiple relationships between diaphragmatic blood flow and the TTdi have also been confirmed by others (11, 54). Using several combinations of duty cycles and tensions, our group reported that phrenic arterial blood flow is related to the TTdi by a quadratic function
<A><AC>Q</AC><AC>˙</AC></A>di = 0.25 + 2.22 TTdi − 4.02 TTdi<SUP>2</SUP>
where Qdi is diaphragmatic blood flow expressed in ml · min-1 · kg body wt-1 (62).
Fig. 2. Relationship between final diaphragmatic blood flow (Qdi) during a contraction period (broken line), debt measured during recovery (x...x), and diaphragmatic tension-time index (TTdi) in 4 animals during periodic contractions with duty cycles of 0.25, 0.5, 0.75 and 1.0. Results obtained at 20 and 50 Hz are shown. Bars indicate ± SE for the contraction: during recovery SE is 5.85 at 20 Hz and 17.4 at 50 Hz. Pdi, transdiaphragmatic pressure. Best fitted lines are drawn by eye. [From Bellemare et al. (14).]
[View Larger Version of this Image (19K GIF file)]

In addition to the duty cycle and tension or pressure, diaphragmatic blood flow during intermittent stimulation is also modulated by contraction frequency (number of intermittent trains of contractions per minute). During maximum diaphragmatic stimulation at two different duty cycles (0.25 and 0.75), Buchler et al. (23) showed that diaphragmatic blood flow increased with a rise in contraction frequencies up to 80 contractions/min, beyond which diaphragmatic blood flow eventually declined, especially at a long duty cycle (0.75). The main data of Buchler et al., depicting the relationship between diaphragmatic blood flow and contraction frequency in dogs, is shown in Fig. 3. The exact mechanism through which breathing frequency leads to an increase in muscle blood flow is not clear yet. It was proposed, however, that metabolic demands at a given TTdi rise with elevated breathing frequencies as a result of an increase in heat of activation (23).

Fig. 3. Effect of contraction frequency on diaphragmatic blood flow at duty cycles of 1/4 (bullet ) and 3/4 (open circle ). Points represent mean results at contraction frequency intervals of 10-49 contractions/min (low frequency), 50-100 contractions/min (medium frequency), and 101-160 contractions/min (high frequency). Bars indicate ±SE. Flow increased significantly (P < 0.02) from low to medium contraction frequencies at both duty cycles and was significantly higher (P < 0.01) at a duty cycle of 1/4 than at 3/4 at medium and high contraction frequencies. [From Buchler et al. (23).]
[View Larger Version of this Image (15K GIF file)]

In summary, these studies indicate that, when diaphragmatic metabolic demands are increased during phasic contractions, blood flow to the diaphragm is influenced by the force, pattern, and frequency of contractions. At low levels of tension, blood flow increases during both the contraction and relaxation phases, but with rising contractile force blood flow during the contraction phase becomes impeded, and metabolic requirements must be met through increasing blood flow during the relaxation phase. Although flow during the relaxation phase may be adequate to meet metabolic demands at low duty cycles, increasing duty cycles exert a linear negative influence on relaxation phase flow and, hence, on total diaphragmatic blood flow. This is because very little time is allowed for flow to reach its peak during the relaxation phase. These studies also indicate that the relationship between diaphragmatic blood flow and the TTdi varies depending on the animal model and the pattern and frequency of phasic contractions.

Effects of Mechanical Factors

It has long been postulated that the negative influence of strong contractions on muscle blood flow is mediated by a significant rise in intramuscular pressure, which leads to compression of blood vessels (2, 3). More recent studies have confirmed that intramuscular pressure inside the ventilatory muscle increases significantly during muscle contraction (33, 112); however, at a given muscle force, intramuscular pressure inside the diaphragm is significantly less than that generated by rounded limb muscles such as the vastus medialis (112). Simultaneous measurements of phrenic vascular tone and intramuscular pressure indicate that, even though diaphragmatic intramuscular pressure swings are smaller than those generated in thick muscles, the magnitude of these swings is sufficient to account for the mechanical impedance of flow during muscle contraction. We measured intramuscular pressure inside the canine diaphragm during phrenic nerve stimulation and reported impedance of blood flow during muscle contraction when intramuscular pressure reached 50 mmHg (59). We concluded that intramuscular pressure swings are important in limiting diaphragmatic blood flow during muscle contraction.

Another important mechanical factor influencing diaphragmatic blood flow is the change in intrathoracic pressure. Buchler et al. (22) conducted elegant experiments in which diaphragmatic blood flow during sustained contractions was measured when the abdomen was bound and the chest was open on the one hand, and when the chest was closed and the abdomen was open on the other hand. Their results indicate that diaphragmatic blood flow was higher during strong negative pleural pressure swings, compared with contractions during which strong positive abdominal pressure was generated (bound abdomen and open chest). It was hypothesized that strong positive abdominal pressure may be transmitted to the diaphragm, leading to the compression of diaphragmatic blood vessels and flow limitation, whereas negative pleural pressure swings may aid in lowering diaphragmatic vascular resistance. Hyperinflation also affects diaphragmatic blood flow through changes in intrathoracic pressure. In a recent study, Kawagoe et al. (68) measured diaphragmatic blood flow during hyperinflation with intrinsic positive end-expiratory pressure (PEEP) and compared it to inspiratory resistive loading. Their results suggest that, despite the generation of similar diaphragmatic TTdi, intrinsic PEEP produced much lower diaphragmatic blood flow values than inspiratory resistive loading. In addition to the generation of strong positive pleural and abdominal pressure swings, which are likely to compress diaphragmatic blood vessels, hyperinflation may influence diaphragmatic vascular tone directly by increasing downstream vascular pressure, leading to a reduction of perfusion pressure.

Diaphragmatic vascular tone is also influenced by muscle fiber length. An increase in muscle fiber length beyond the optimum length results in a significant decline in diaphragmatic blood flow measured at rest and during intermittent contractions (111). This effect is mediated through a rise in passive tension that exerts strong stress along the longitudinal axis of intramuscular blood vessels, leading to their progressive compression with increasing muscle fiber length.

Effects of Arterial Pressure

There is strong evidence indicating that ventilatory muscle blood flow is strongly influenced by arterial pressure. An early study (105) suggested that phrenic arterial flow is linearly related to arterial pressure over a mean arterial range of 60-100 mmHg during quiet breathing and resistive loading. However, when phrenic arterial flow was measured over a wider range of arterial pressure changes, a small degree of autoregulation was evident in spontaneously breathing dogs, especially between mean arterial pressure values of 90-120 mmHg (61). The effectiveness of autoregulation of phrenic arterial flow improved significantly when diaphragmatic metabolic demands were raised by resistive loading but the range of arterial pressure over which autoregulation occurred narrowed, compared with that of quiet breathing.

When the diaphragmatic vasculature is maximally dilated, arterial pressure becomes the sole determinant of diaphragmatic blood flow. However, the exact mathematical function describing the relationship between maximum diaphragmatic blood flow and arterial pressure varies between different studies. When pharmacological dilators were infused into dogs spontaneously breathing 6% inspiratory O2 concentration and exposed to severe resistive loading, a parabolic function best described the relationship between mean arterial perfusion pressure (Pa) and diaphragmatic blood flow
<A><AC>Q</AC><AC>˙</AC></A>di<SUB>max</SUB> = (1.32 Pa<SUP>2</SUP> + 29.9 Pa) × 10<SUP>−4</SUP>
where Qdimax is maximum diaphragmatic blood flow, expressed in ml · min-1 · g-1, and Pa is in mmHg (96). On the other hand, Magder (76) combined phrenic nerve stimulation with pharmacological dilators to achieve maximum diaphragmatic dilation. He reported the following linear relationship between maximum diaphragmatic blood flow and arterial pressure
<A><AC>Q</AC><AC>˙</AC></A>di<SUB>max</SUB> = 3.13 Pa × 10<SUP>−2</SUP> − 0.52
where Qdimax is expressed in ml · min-1 · g-1 and Pa is in mmHg. The higher flow values in Magder's study could be attributed to the use of artificial phrenic nerve stimulation, which allows greater levels of metabolic demands than during spontaneous breathing.

Effects of Sympathetic Tone and Adrenergic Receptors

Skeletal muscle vascular tone, like the splanchnic and renal vasculature, is reflexly controlled by the sympathetic noradrenergic nerves. In the resting limb muscles, complete abolition of sympathetic activity results in the doubling of blood flow, whereas maximum stimulation of noradrenergic nerves reduces resting muscle blood flow by 75%. This significant modulation of muscle blood flow by sympathetic nerves is smaller in magnitude than that elicited by local mediators but allows for greater overall changes in systemic vascular resistance because of the large proportion of body mass being contributed by the skeletal muscles.

Little is known about the degree to which ventilatory muscle vascular tone is influenced by sympathetic nerves. DeHaan and Kendrick (34) stimulated depressor reflexes in the aortic nerve and measured the vascular resistance of ventilatory muscles at different levels of metabolic demands. Stimulation of aortic nerve depressor reflexes at 100 Hz, which inhibits sympathetic vasoconstrictor activity, reduced vascular resistance of the paralyzed diaphragm by 79%. In comparison, diaphragmatic vascular resistance during quiet breathing and moderate resistive loading declined by 72 and 66%, respectively, in response to aortic nerve stimulation. Similar findings were reported in the intercostal muscle vascular bed. These results suggest that sympathetic vasoconstrictor drive exerts a significant influence on ventilatory muscle vascular tone at various levels of metabolic demands.

Like sympathetic neural drive, exogenous adrenergic agonists are known to influence ventilatory muscle vascular tone and contractile function. In endotoxemic dogs, systemic norepinephrine infusion elicits a small but significant decline in diaphragmatic blood flow with no effect on muscle contractile performance (63). Supinski et al. (114), on the other hand, showed that norepinephrine infusion increased arterial blood pressure and phrenic arterial flow and improved the force-frequency curve of the fatigued diaphragm. However, when the diaphragmatic vascular bed was pump perfused at a fixed flow rate, norepinephrine infusion elicited a significant rise in phrenic vascular resistance, with no improvement in contractile function. Thus norepinephrine appears to exert a direct vasoconstrictor effect on the phrenic vascular bed, presumably through the activation of alpha -adrenoceptors on resistance vessels, whereas the improvement in contractile muscle function is mediated indirectly through augmentation of arterial blood pressure.

More recently, we reported that infusion of phenylephrine (selective alpha -adrenoreceptor agonist) into the phrenic artery of pump-perfused diaphragms increased phrenic vascular resistance (128). Phenylephrine also augmented diaphragmatic muscle tension and O2 uptake during 3-Hz stimulation with no effect on maximum O2 extraction, whereas no changes in muscle O2 uptake were observed in the resting diaphragm. These data provide clear evidence that alpha -adrenoceptor stimulation not only mediates an increase in vascular tone but also causes significant augmentation of muscle force and O2 consumption.

In addition to the alpha -subtype, ventilatory muscles express beta -adrenoceptors that modulate muscle function and vascular tone. This modulation was postulated by Aubier et al. (4), who measured phrenic venous flow in chronic obstructive pulmonary disease patients by inserting a catheter into the inferior phrenic vein of these patients and reported that systemic dopamine infusion (10 µg · kg body wt-1/min) elicited an ~30% rise in phrenic venous flow. Along with the increase in blood flow, diaphragmatic tension generation was also augmented by dopamine infusion. Because dopamine is a selective beta -adrenoreceptor stimulant, it was concluded that beta -adrenoceptor exerts an inhibitory effect on diaphragmatic vascular tone. However, Aubier et al. did not exclude the possibility that dopamine might have influenced diaphragmatic blood flow and contractility through alterations in systemic arterial pressure.

Effects of Hypoxia

The effects of hypoxia on ventilatory muscle vascular tone have been examined in several studies. Lowering inspiratory O2 fraction (FIO2) to 10% in dogs was associated with a twofold increase in diaphragmatic blood flow and only 30% rise in intercostal muscle blood flow (1). When FIO2 was lowered to 5%, diaphragmatic and intercostal blood flow increased by 450 and 129%, respectively. This augmentation of ventilatory muscle blood flow appears to be due to the activation of chemoreceptors, resulting in heightened ventilatory drive and increased ventilatory muscle activity. A similar conclusion was reached by Kendrick et al. (69), who showed that blood flow to the diaphragm and intercostal muscles rose significantly in spontaneously breathing hypoxemic rabbit (PaO2 = 25 Torr). However, when ventilatory muscle activity was reduced by paralysis, systemic hypoxia had no effect on ventilatory muscle blood flow. These results were interpreted to mean that hypoxia exerts a negligible direct dilatory influence on ventilatory muscle vascular tone. This conclusion was challenged by two other groups of investigators. Using an in situ isolated diaphragmatic preparation, Bark et al. (12) found that severe arterial hypoxemia (phrenic venous PO2 <15 Torr) elicited a sharp decline in diaphragmatic vascular resistance. Figure 4 shows the influence of hypoxemia on the relationship between phrenic venous PO2 and diaphragmatic vascular conductance at rest and during rhythmic contractions at two levels of tension-time index. Notice the sharp increase in diaphragmatic conductance when phrenic venous PO2 declined below 15 Torr. Similarly, Reid and Johnson (96) observed that diaphragmatic vascular conductance in hypoxemic dogs increased slowly as phrenic venous PO2 declined below 30 Torr, with a sharp rise in conductance when phrenic venous PO2 fell below 10-12 Torr. On the basis of these studies, it is difficult to assess the direct effect of hypoxia or hypoxemia on local regulatory mechanisms of ventilatory muscle vascular tone because the induction of systemic arterial hypoxemia is likely to perturb systemic acid base balance, blood pressure and central respiratory motor output. All of these systemic changes may obscure local regulatory mechanisms.
Fig. 4. Relationship between phrenic venous oxygen tension (PvO2) and diaphragmatic vascular conductance. bullet , Resting muscles during rhythmic contraction at a tension time index of 0.05 (black-triangle) and 0.15 (black-square). [From Bark et al. (12).]
[View Larger Version of this Image (16K GIF file)]

More recently, Ward (124) quantified the effect of hypoxia on diaphragmatic vascular tone by perfusing in situ isolated diaphragms at a constant flow rate with blood containing various levels of PO2. Moderate and severe hypoxia (phrenic venous PO2 of 25 and 13 Torr, respectively) resulted in significant phrenic vascular dilation, but the magnitude of the dilatory response varied according to vessel size. Small arterioles (baseline diameter of 40 µm and less) dilated significantly more than relatively larger arterioles. These results confirm that hypoxia exerts a selective dilatory influence on the tone of diaphragmatic arterioles.

Effects of Hypercapnia

The effects of alterations in PCO2 on vascular tone have been studied by numerous investigators using in vivo and in vitro blood vessel preparations. The results, however, are somewhat contradictory and appear to depend on other experimental variables such as the method of altering PCO2 (systemic vs. local), the presence or absence of intact autonomic innervation, hypoxia or adrenergic agonists. For instance, it has been demonstrated that hypercapnia elicits either constriction (48), dilation (48), or no effect (36) on arteriolar diameter measured in vivo. Modulation of vascular tone by PCO2 appears to be influenced by the level of tissue oxygenation (20, 85). Indeed, a synergistic interaction between PO2 and PCO2 in determining vascular tone of gastrocnemius (85) and coronary vascular bed (20) has been described. It is clear from these results that one should always take into account the level of tissue oxygenation when assessing the influence of PCO2 on vascular tone.

In the respiratory system, the influence of hypercapnia on ventilatory muscle tone has attracted little attention. Initial reports have not supported a major effect of hypercapnia on ventilatory muscle blood flow (69). Reid and Johnson (96) used regression analysis to identify the effects of hypercapnia on diaphragmatic blood flow during resistive breathing. Their results indicated that the inclusion of PCO2 did not alter the regression equation model predicting maximum diaphragmatic flow from arterial pressure during combined resistive breathing, hypoxia, and hypotension. The direct effect of PCO2 on the diaphragmatic vascular bed was recently investigated in in situ isolated diaphragmatic preparations in which arterial PCO2 was altered without changing systemic arterial PCO2 (124). High arterial PCO2 elicited significant phrenic vasodilation only when phrenic venous PCO2 values exceeded 80 Torr. Moreover, when hypercapnia and hypoxia were presented combined, the increase in diaphragmatic blood flow was greater than the mathematical sum of individual hypercapnic and hypoxic responses. These results confirm that only relatively high levels of PCO2 have any effect on diaphragmatic vascular tone. These data also indicate that hypercapnia-induced phrenic vasodilation is synergistically augmented by hypoxia.

Effects of Anesthesia

The vast majority of experiments on the regulation of ventilatory muscle blood flow have been conducted in anesthetized animals. Yet little is known about the influence of general anesthesia on ventilatory muscle vascular tone. Systemically, general anesthetics have a substantial inhibitory effect on both cardiac function and peripheral circulatory control, resulting primarily in a significant decline in cardiac output and, to a lesser extent, of peripheral vasodilation. In the individual vascular bed or isolated blood vessel, general anesthetics exert variable and complex actions on vascular tone. For instance, halothane may increase resistance in a few vascular beds and reduce resistance in others, leading to small changes in total peripheral vascular resistance. Vascular responses to anesthetic administration also vary among vessels and species. For example, 2% halothane is reported to elicit significant dilation of isolated canine carotid arteries and rabbit aortas (87), whereas a rise in tension occurs in isolated canine femoral arteries. Seyde and Longnecker (107) assessed the effects of ketamine, halothane, isoflurane, and enflurane on blood flow distribution in spontaneously breathing rats. While blood flow to the gastrointestinal tract and spleen remained unchanged, blood flow to the diaphragm and rectus abdominous muscles declined significantly in response to all these anesthetics when compared with conscious rats. Because anesthetic administration was associated with secondary changes in arterial blood pressure, blood gases and lactate concentration, it is difficult to infer whether anesthetics have a direct and selective vasoconstrictor influence on ventilatory muscle vascular tone. More recently, Leon et al. (74) evaluated the additional effects of halothane and isoflurane on the diaphragmatic microcirculation in pentobarbital sodium-anesthetized rats. They reported that neither isoflurane nor halothane had any influence on the diameter of second- and third-order arterioles, but, unlike isoflurane, halothane induced dose-dependent vasoconstriction of fourth-order arterioles and significantly depressed functional capillary density. Because the state of capillary perfusion is an important determinant of substrate delivery to muscle fibers, reduction of capillary density in response to halothane is likely to exert a negative influence on muscle mechanical performance. Indeed, diaphragmatic force generation has been shown to be depressed in response to halothane anesthesia. Clearly, more attention needs to be paid to the effects of various anesthetics on the regulation of ventilatory muscle vascular tone and muscle function.

Local Mediators of Metabolic Vasodilation

A number of factors have been proposed as potential mediators of metabolic vasodilation. They include O2, CO2, potassium, adenosine, lactate, phosphate, and osmolarity. However, none of these mediators has fully accounted for the increase in muscle blood flow during exercise or stimulation (for reviews see Refs. 92 and 129). A discussion of the importance of these factors is beyond the scope of this review. Instead, the focus will be on progress made recently in elaborating the role of nitric oxide (NO), prostaglandins, potassium channels, and endothelins.

Role of NO

NO is a highly reactive species that is synthesized by a group of hemeproteins known as NO synthases (NOS). Three NOS isoforms have been so far characterized, two of which are constitutively expressed and were first discovered in the endothelial cells (ecNOS) and brain cells (bNOS). A third isoform, namely inducible NOS (iNOS), is not usually seen under normal conditions but is induced in response to cytokine exposure and is expressed in a wide variety of cells such as macrophages, hepatocytes, and smooth muscle cells. The role of endothelium-derived NO in the regulation of vascular tone has been studied extensively in isolated arteries, perfused organs, and anesthetized animals, and it is now widely accepted that NO is synthesized by the endothelium of virtually all resistance vessels. In the skeletal muscle, there is accumulating evidence that under normal conditions endothelial NO synthesis and release play an important role in the in vivo regulation of basal vascular tone. Indeed, infusion of NOS inhibitors is known to induce significant vasoconstriction in rat hindquarter (42) and cremaster muscle (67), rabbit tenuissimus muscle (95), human brachal artery (118), and dog hindlimbs (130). In the respiratory system, our group was the first to explore the importance of NO in the regulation of ventilatory muscle blood flow (126). In our early studies, we found that infusion of various NOS inhibitors into the phrenic artery of the in situ isolated resting canine diaphragm resulted in a dose-dependent increase in phrenic vascular resistance. Figure 5 shows an example of the rise in phrenic perfusion pressure in response to intraphrenic infusion of NG-nitro-L-arginine methyl ester, an inhibitor of NOS, in a pump-perfused diaphragm. We also found that infusion of increasing L-arginine concentrations had no effect on phrenic vascular resistance, suggesting that basal endothelial-derived NO release acts as a tonic vasodilatory element opposing myogenic vasoconstriction in resistance vessels of various skeletal muscles. In addition to basal vascular resistance, enhanced NO release plays a significant role in the reactive hyperemia of the canine diaphragm (123).
Fig. 5. Representative tracing of changes in phrenic perfusion pressure (Pphr) in response to intraphrenic infusion of 6 × 10-4 M NG-nitro-L-arginine methyl ester (L-NAME) in pump-perfused diaphragm. Notice progressive rise in Pphr. Qphr, phrenic blood flow. [From Ward and Hussain (126).]
[View Larger Version of this Image (8K GIF file)]

Unlike the reactive hyperemic response, NO's contribution to the active vasodilation of skeletal muscles is still being debated. Early reports (103) suggest that functional hyperemia of the isometrically contracted canine gastrocnemius muscle is significantly attenuated after the infusion of methylene blue (an inhibitor of guanosine 3',5'-cyclic monophosphate formation) or gossipol (an endothelium-derived relaxing factor inhibitor). In exercising dogs, systemic infusion of NOS inhibitors results in a significant rise in arterial pressure and vascular resistance in active muscles including the diaphragm (108). NOS inhibitors also attenuate the rise in limb muscle blood flow in exercising rats (52). This effect appeared to be more severe in muscles with a high percentage of type I fibers. In normal humans, intra-arterial infusion of NOS inhibitors attenuated exercise-induced forearm hyperemia by ~11% (43). In the respiratory system, we found that increased NO release accounted for 25-41% of the active hyperemic response of the canine diaphragm (65). This was calculated by comparing the degree to which NOS inhibitors reverse phrenic vascular dilation induced by low-frequency stimulation of the phrenic nerve, endothelium-dependent dilator (acetylcholine) and endothelium-independent dilator such as sodium nitroprusside (Fig. 6). These results imply that exercise-induced vasodilation in limb and ventilatory muscles is mediated in part by NO synthesis and that the magnitude of NO's contribution varies, depending on the species, muscle fiber type, method, and duration of exercise.

Fig. 6. Influence of nitric oxide (NO) synthase inhibitor NG nitro-L-arginine (L-NNA) on dilator response to intraphrenic infusion of acetylcholine (ACh), sodium nitroprusside (SNP), and low-frequency stimulation of phrenic nerve (2 Hz). When infused alone, ACh or SNP dilated phrenic vasculature by 80% of maximum. When L-NNA was infused along with either ACh, SNP, or during 2-Hz stimulation of phrenic nerve, it completely reversed phrenic vascular dilation, suggesting that this inhibitor was able to block enhanced NO release. By comparison, L-NNA reversed SNP-induced dilation by ~28%. We attributed this reversal to inhibition of baseline NO release. When infused during phrenic nerve stimulation, L-NNA reversed active vasodilation by ~53%. By subtracting the contribution of baseline NO release, we estimated that enhanced NO release accounted for ~25% of active phrenic dilation during 2-Hz stimulation. ** P < 0.01, * P < 0.05 compared with ACh. + P < 0.05 compared with SNP. [From Hussain et al. (65).]
[View Larger Version of this Image (20K GIF file)]

Little information is available about the influence of NO release on skeletal muscle O2 consumption. Our data in dogs indicate that inhibition of NO synthesis and release in the auotperfused, actively contracting diaphragm results in a significant reduction of diaphragmatic O2 consumption. This effect appears to be due to diminished blood flow, since NOS inhibition has no effect on the ability of the diaphragm to extract O2 (24). The lack of effect of NOS inhibitors on maximum diaphragmatic O2 extraction indicates that capillary exchange capacity is not influenced by NO release. This was confirmed by our subsequent study (127). These observations can be explained on the basis of the metabolic theory of vascular tone regulation (44). According to this theory, at low levels of skeletal muscle metabolic demands (venous PO2 >= 40 Torr), tissue oxygenation is maintained by increasing capillary density as a result of terminal arteriolar relaxation with little contribution from resistance vessels. Thus inhibition of NO release from the endothelium of resistance arterioles will not have a major effect on skeletal muscle O2 uptake. This hypothesis is supported by the findings of Persson et al. (95), who reported that NOS inhibition in the rabbit tenuissimus had no effect on the response of terminal arterioles to locally released metabolites. The metabolic theory states, however, that with increasing metabolic demands, the contribution of capillary density is diminished because most of the capillaries are already recruited and, therefore, tissue oxygenation will be increasingly dependent on resistance arterioles to dilate in response to locally released metabolites. Accordingly, the importance of endothelial NO release in the regulation of tissue oxygenation rises with heightened increasing levels of metabolic demands. In addition to the regulation of tissue oxygenation, NO release also appears to play a role in the autoregulation of diaphragmatic blood flow in response to alterations in arterial pressure. Our group reported, however, that NO plays a role in flow autoregulation mainly by counterbalancing myogenic vasoconstriction at relatively high perfusion pressure values (125). In this respect, the diaphragmatic vascular bed behaves in a similar fashion to the heart and kidney.

In all of these studies dealing with the role of NO release in skeletal muscle perfusion, it was assumed that the effect of NOS inhibitors on muscle blood flow is indicative of endothelial NO release. With the recent discovery of NOS expression in striated muscle fibers, this assumption may not hold. The expression of NOS in muscle fibers was described by Nakane et al. (90), who noticed abundant bNOS mRNA in human skeletal muscle samples. Kobzik et al. (71) later confirmed the existence of bNOS in striated muscle fibers and added that bNOS expression is localized to the sarcolemma. In another report, in vitro incubated rat limb muscles were found to release NO, and it was found that NO release rises with increased muscle activity (7). More recently, Kobzik et al. (72) reported the presence of ecNOS isoform in oxidative muscle fibers. By using immunohistochemistry, these authors reported that, unlike bNOS, ecNOS staining was more diffuse across the sarcoplasm. The existence of ecNOS and bNOS isoforms in ventilatory muscle fibers has recently been confirmed by our group (S. N. A. Hussain, M. Abdul-Hussain, and D. Sakkal, unpublished observations). We also found significant intra- and interspecies differences in muscle NOS activity, with canine ventilatory muscle showing significantly higher NOS activities than those of rabbit, mouse, and rat ventilatory muscles. The exact functions of ecNOS and bNOS isoforms in ventilatory muscle function remain unclear. It has been proposed, however, that muscle NOS expression is involved in the regulation of muscle contractility (71), glucose uptake (7), and mitochondrial respiration (21, 72). These newly described roles of NO in muscle function suggest that nonendothelial NO release may participate in the regulation of muscle blood flow. This is particularly true in the case of NOS expression in muscle mitochondria. The exact role of NO in muscle function and blood flow is being investigated by a number of investigators, and significant progress in this field is likely to occur in the coming years.

Endothelium-Derived Prostaglandins

Prostaglandins are a family of compounds produced by the action of cyclooxygenase on arachidonic acid. Both vasoconstrictor and vasodilatory prostaglandins have been identified, with prostacyclin, which is produced by endothelial cells, being the most important member of the latter group. A number of investigators have suggested that prostaglandin release contributes to the vascular dilation of skeletal muscle arterioles, especially during physical exercise (active hyperemia) (51, 70). However, the exact estimate of this contribution varies between different studies (30, 132).

The only available evidence of prostaglandin involvement in the regulation of ventilatory muscle blood flow was provided by Boczkowski et al. (18). These authors studied the effects of selective inhibitors of prostaglandin synthesis (mafenamic acid) on the diameter of second-order arterioles in the rat diaphragm. Inhibition of prostaglandin synthesis resulted in a small but significant reduction in baseline arteriolar diameter ranging from 0.3 to 3.9%. In contrast, inhibition of NO synthesis elicited a 3 to 15% decrease in baseline arteriolar diameter. These results suggest that vascular tone of the resting diaphragm in rats is more dependent on NO release than on prostaglandin synthesis. Interestingly, when inhibitors of NO and prostaglandin synthesis were applied simultaneously, the reduction in diaphragmatic arteriolar diameter was more significant than the additive effects of the individual inhibitors, indicating an interaction between the two pathways of vascular dilation in the diaphragm. The nature of this interaction remains to be determined.

Role of K+ Channels

There is increasing evidence that smooth muscle contractility is regulated in part by changes in transmembrane potential. Alterations in transmembrane potential influence muscle contractility by interfering with Ca2+ influx through voltage-dependent Ca2+ channels. ATP-sensitive potassium (KATP) channels activated by a reduction of intracellular ATP concentration are one group of membrane ionic channels that modulates transmembrane potential. KATP channels exist in a variety of cells, including rabbit mesenteric and pulmonary vascular smooth muscles and cardiac and skeletal muscle fibers (25, 32, 109). Recent studies suggest that smooth muscle KATP channels participate in the regulation of coronary vascular tone under normal conditions (66) and in response to hypoxia and ischemia (31). In addition to KATP channels, the sarcolemma of smooth muscle cells contains a group of K+ channels that are activated by intracellular Ca2+ (KCa). When intracellular Ca2+ increases, the probability of KCa opening rises, leading to membrane hyperpolarization as a result of K+ outflux. On the basis of this mechanism of activation, smooth muscle KCa channels have been postulated to serve as a negative-feedback mechanism controlling the degree of membrane polarization and vasoconstriction.

The contribution of various K+ channels to the regulation of basal tone in the ventilatory muscle vasculature has recently been investigated by our group (119, 120). Our data indicate that K+ channels, particularly those sensitive to intracellular ATP, participate in the regulation of basal vascular tone of the canine diaphragm and mediate a significant portion of the dilator response to brief occlusions of the phrenic artery (120). KATP channels also participate in active diaphragmatic dilation during low-frequency (0.5-4 Hz) stimulation (119). Interestingly, inhibition of diaphragmatic KATP channels had no effect on the ability of the diaphragm to extract O2. The involvement of KATP channels in the regulation of vascular tone of the intact diaphragm was recently confirmed in spontaneously breathing dogs (29). These data suggest that under normal conditions the tone of resistance vessels of the diaphragm is regulated in part by the activity of sarcolemmal K+ channels, particularly those that are sensitive to intracellular ATP concentration. The exact mechanism(s) leading to activation of these channels, however, remains unclear. Local tissue hypoxia, reduced intracellular pH, and adenosine release from skeletal muscle myocytes have all been proposed to heighten the probability of opening smooth muscle KATP channels.

Role of Endothelins

The endothelins are a family of 21-amino-acid peptides with potent sustained vasoconstrictor and vasodilator actions. Endothelin-1 is the main isopeptide generated by the endothelial cells, whereas endothelin-2 and endothelin-3 have more widespread tissue distribution. Endothelins were reported to arise through proteolytic processing of preproendothelin to produce proendothelin, which is then cleaved by endothelin-converting enzyme to endothelins. Two specific receptors for the endothelins have been isolated so far. ETA receptor has a high affinity for endothelin-1 and is mainly expressed in vascular smooth muscle cells. This receptor is thought to mediate the potent vasoconstrictor influence of endothelin-1. ETB receptor, on the other hand, has equal affinity for all three endothelins and was previously thought to be endothelium associated and to mediate vasorelaxation. ETB receptor was proposed initially to mediate the release of endothelium-derived vasorelaxants causing vasodilation in response to endothelin exposure. Recent studies indicate, however, that ETB receptor is also expressed in vascular smooth muscles of various organs and may mediate vasoconstriction as well (110).

The role of endothelins in the regulation of ventilatory muscle blood flow is unknown. Recent studies on limb muscles indicate that endothelins exert quantitatively different vascular responses. For instance, intra-arterial infusion of endothelin-1 in cat gastrocnemius muscle elicits, after initial transient dilation, strong vasoconstrictor effect that is more pronounced in small arterioles and veins (38). A similar infusion of endothelin-2 or endothelin-3 evokes smaller increase in total gastrocnemius vascular resistance compared with endothelin-1 (37). In both experiments, selective inhibition of ETA receptor by FR-139317 abolishes the vasoconstrictor responses to the three endothelins, suggesting that smooth muscles cells of cat gastrocnemius vessels express ETA receptors. Infusion of endothelin-1 into the brachial artery of humans elicits a slow-onset dose-dependent forearm vasoconstriction (49). Endothelin-3 evokes a similar but less pronounced increase in forearm vascular resistance in humans (49). These results indicate that, in human peripheral vessels, both ETA and ETB receptors have a vasoconstrictor effects. Despite the significant progress in endothelin research, the contribution of endogenous endothelins to the regulation of skeletal muscle vascular tone under normal conditions remains debatable. Recently, Terelink et al. (116) reported that infusion of bosentan, a blocker of both ETA and ETB receptors, had no hemodynamic effects in anesthetized dogs. Similarly, neither selective ETA blockade nor administration of phosphoramidon, inhibitor of endothelin-converting enzyme, had any influence on baseline vascular resistance of cat gastrocnemius muscle, indicating no significant endogenous endothelin-1 release in this muscle (38). In contrast to these findings, Haynes and Webb (50) reported that infusion of phosphoramidon or BQ-123 (selective inhibitor of ETA receptor) increased human brachial blood flow by 37 and 64%, respectively. It was concluded that endogenous production of endothelin-1 contributes to the maintenance of peripheral vascular tone in humans.

Ventilatory Muscle Blood Flow and Fatigue

The cause-and-effect relationship between muscle blood flow and fatigue has been debated for decades. The difficulties faced in solving this debate are attributed in part to the fact that the cellular processes involved in muscle fatigue are complex and vary, depending on the workload, pattern of muscle activation, muscle fiber composition, and species. In limb muscles, some investigators have assessed whether blood flow limitation precipitates contractile failure by measuring muscle vasodilator reserve during the course of fatiguing contractions. The results so far are contradictory. For instance, it has been demonstrated in maximally exercising limb muscles that vascular tone is independent of central adrenergic vasoconstrictor drive, indicating that complete vasodilation has been achieved (98). On the other hand, others (91) have illustrated that activation of sympathetic vasoconstrictor drive by baroreceptor reflexes results in a significant vasoconstriction in maximally exercising limb muscles. These findings suggest that "functional sympatholysis" does not develop during fatiguing tasks in these muscles, and a significant vasodilator reserve is still available. A similar conclusion was reached in exercising swine (73).

In another series of experiments, the association between muscle blood flow and rate of fatigue was tested by artificially altering blood flow during muscle contractions. Barclay and Stainsby (10) reported that in highly oxidative limb muscles artificial augmentation of blood flow was associated with a significant rise in twitch tension. Subsequently, they showed that maintaining blood flow at the steady-state level achieved during active hyperemia delayed the rate of fatigue during repetitive brief tetanic contractions (9). This effect of blood flow on the rate of muscle fatigue appears to be independent of the delivery of O2 and other blood-borne substrates (8).

In the respiratory system, there is evidence suggesting that maximum vasodilation may be reached in the diaphragm during fatiguing contractions. As mentioned above, Manohar (79) showed that systemic infusion of adenosine in exercising ponies did not result in a further increase in diaphragmatic blood flow, indicating the absence of vasodilator reserve in the diaphragm during maximum exercise. Complete diaphragmatic vasodilation has also been reported during fatiguing contractions elicited by artificial phrenic nerve pacing in dogs (115). When sodium nitroprusside, a potent vasodilator, was infused systemically, no further increase in diaphragmatic blood flow was seen. Pang et al. (93) reported that peak diaphragmatic blood flow approaches 300 ml · 100 g-1 · min-1 in conscious sheep during severe inspiratory resistive loading, which leads to diaphragm fatigue and hypercapnic respiratory failure. This value falls in the range of maximal diaphragmatic blood flow predicted by Magder (76). Thus diaphragmatic vasodilator reserve may be completely exhausted during fatiguing contractions. The influence of hyperperfusion on muscle contractility previously shown in limb muscles has been confirmed by several investigators (113, 122).

One of the most important cellular mechanisms through which blood flow regulates muscle contractile performance is the delivery of O2 and other substrates necessary to maintain oxidative phosphorylation and glycolysis. The latter are major processes through which ATP is directly resynthesized from ADP and phosphate ion (Pi). Limitation of these two processes leads to a greater reliance on creatine kinase and myokinase reactions to resynthesize ATP from ADP only, eventually resulting in Pi accumulation. Several authors have reported that muscle fatigue correlates closely with Pi (131, 133). Adequate delivery of O2 by arterial flow influences muscle contractility by reducing the reliance on anaerobic metabolism, which results in an accumulation of H+ ions. These ions exert a negative effect on muscle contractile mechanisms (78). Another mechanism through which blood flow may influence the rate of muscle fatigue is by washing out by-products of muscle metabolism (8, 113, 122).

Ventilatory Muscle Blood Flow in Shock

Interest in assessing ventilatory muscle perfusion in various forms of shock has been initiated by the original observations that ventilatory muscle contractile failure develops in animals with cardiac tamponade and leads to hypercapnic respiratory failure and death (5). Measurement of ventilatory muscle perfusion indicated a significant rise in diaphragmatic blood flow within 60 min cardiac tamponade induction (121). Perfusion of other ventilatory muscles also increases but to a lesser extent than that of the diaphragm. The rise in ventilatory muscle blood flow reflects increased metabolic demands as a result of increased minute ventilation. Compared with maximum predicted values based on the equation of Magder, peak diaphragmatic blood flow during cardiogenic shock was submaximum, indicating that the diaphragmatic vasodilatory reserve was not exhausted. Despite the substantial augmentation of diaphragmatic perfusion, during the course of cardiogenic shock, lactate production and glycogen utilization by the diaphragm increased significantly, suggesting that blood flow was inadequate in meeting rising metabolic demands. Muscle fatigue ensues in response to this deficit of oxidative metabolism and leads to the development of hypercapnenic ventilatory failure. Significant augmentation of diaphragmatic blood flow has also been reported when cardiogenic shock is combined with either elastic loading (77) or oleic acid-induced pulmonary edema (102). In both these studies, peak diaphragmatic blood flow was significantly lower than the predicted maximum values but correlated well with such indexes of diaphragmatic metabolic demands as the TTdi. Interestingly, the linear relationship between diaphragmatic blood flow and the TTdi found during combined elastic loading and cardiogenic shock is similar to that described during inspiratory resistive loading.

The relationship between diaphragmatic function and energetics has also been assessed during septic shock. Earlier reports by our group (64) indicated that, similarly to cardiogenic shock, ventilatory muscle contractile failure develops during the course of hypodynamic septic shock in spontaneously breathing dogs and that this failure occurs despite a significant rise in diaphragmatic blood flow, particularly at 60 min after the induction of shock (58, 60). Unlike cardiogenic or hemorrhagic shock, increased muscle activity as a result of hyperpnea is not the only factor involved in alterations of diaphragmatic blood flow and vascular tone during septic shock. An additional factor that may cause a rise in diaphragmatic blood flow at a given level of metabolic demand is the apparent impairment of the diaphragm to extract O2 during the course of septic shock, rendering diaphragmatic O2 consumption more dependent on blood flow and O2 delivery. In addition, septic shock has been shown to be associated with a significant increase in plugging of diaphragmatic capillaries, the result of granulocyte adherence to endothelial cells. (17). More direct evidence indicating that increased muscle activity is not the sole cause of vascular dilation in septic shock was provided by a recent study in which diaphragmatic vascular resistance of the pump-perfused resting diaphragm declined significantly at 90 min after endotoxin infusion in dogs (56). These data suggest that septic shock elicits a significant reduction of arterial pressure associated with a paradoxical decline in diaphragmatic vascular resistance. This fall in peripheral vascular tone has been termed "vascular decompensation." Several mechanisms have been implicated in the vascular decompensation of skeletal muscles during septic shock. These include direct inhibition of the sympathetic system by circulating bacterial endotoxin, direct suppression of vascular smooth muscle metabolism, increased metabolic demands with a release of local vasodilator substances, as well as augmented NO synthesis and release in smooth muscle cells as a result of induction of iNOS. Although the first three mechanisms are being challenged, there is ample evidence that iNOS plays a major role in the depression of vascular reactivity in in vivo and in vitro preparations (41, 45, 117). Whether iNOS is expressed in the ventilatory muscle vasculature during septic shock remains to be investigated. Preliminary results in our laboratory indicate that endotoxin exposure is associated with iNOS expression in ventilatory muscle blood vessels as well as in striated muscle fiber. The degree to which iNOS induction in ventilatory muscle fiber contributes to vascular decompensation and depressed contractility during septic shock needs to be assessed in further detail.

In summary, ventilatory muscle blood flow is determined by complex interactions between physical, neurological, anatomic, and chemical factors acting on various sites along the vascular tree. A few of these factors are investigated in more detail than others. In the past few years, significant progress has been made in the area of vascular biology of NO, endothelins, heme oxygenase, O2 radicals, and other endothelium- and smooth muscle-derived factors. More details about how these factors interact and influence ventilatory muscle blood flow, O2 consumption, and contractile function under normal and pathological conditions are anticipated in the near future.

FOOTNOTES

Address for reprint requests: S. Hussain, Rm. L3.05, Royal Victoria Hospital, 687 Pine Av. West, Montreal, Quebec H3A 1A1, Canada (E-mail: SHUSSAIN{at}RVHMED.LAN.McGill.Ca).

REFERENCES

1. Adachi, H., H. W. Strauss, H. Ochi, and H. N. J. Wagner. The effect of hypoxia on the regional distribution of cardiac output in the dog. Circ. Res. 39: 314-319, 1976.
2. Anrep, G. V., S. Cerqua, and A. Samaan. The effect of muscular contraction upon the blood flow in the skeletal muscle, in the diaphragm and in the small intestine. Proc. R. Soc. Lond. B Biol. Sci. 114: 245-357, 1933.
3. Anrep, G. V., and E. V. Saalfeld. The blood flow through the skeletal muscle in relation to its contraction. J. Physiol. Lond. 85: 375-399, 1935.
4. Aubier, M., D. Murciano, Y. Menu, J. Boczkowski, H. Mal, and R. Pariente. Dopamine effects on diaphragmatic strength during acute respiratory failure in chronic obstructive pulmonary disease. Ann. Int. Med. 110: 17-23, 1989.
5. Aubier, M., T. Trippenbach, and C. Roussos. Respiratory muscle fatigue during cardiogenic shock. J. Appl. Physiol. 51: 499-508, 1981.
6. Axen, C., and P. E. Janson. Diaphragmatic blood flow at various levels of ventilation in the rabbit. Acta Physiol. Scand. 106: 1-3, 1979.
7. Balon, T. W., and J. L. Nadler. Nitric oxide release is present from incubated skeletal muscle preparations. J. Appl. Physiol. 77: 2519-2521, 1994.
8. Barclay, J. K. A delivery-independent blood flow effect on skeletal muscle fatigue. J. Appl. Physiol. 61: 1084-1090, 1986.
9. Barclay, J. K., C. M. Boulianne, B. A. Wilson, and S. J. Tiffin. Interaction of hyperoxia and blood flow during fatigue of canine skeletal muscle in situ. J. Appl. Physiol. 47: 1018-1024, 1979.
10. Barclay, J. K., and W. N. Stainsby. The role of blood flow in limiting maximal metabolic rate in muscle. Med. Sci. Sports Exercise 7: 116-119, 1976.
11. Bark, H., and S. M. Scharf. Diaphragmatic blood flow in the dog. J. Appl. Physiol. 60: 554-561, 1986.
12. Bark, H., G. Supinski, R. Bundy, and S. Kelsen. Effect of hypoxia on diaphragm blood flow, oxygen uptake and contractility. Am. Rev. Respir. Dis. 138: 1535-1541, 1988.
13. Bark, H., G. S. Supinski, J. C. Lamanna, and S. G. Kelsen. Relationship of changes in diaphragmatic muscle blood flow to muscle contractile activity. J. Appl. Physiol. 62: 291-299, 1987.
14. Bellemare, F., D. Wight, C. M. Lavigne, and A. Grassino. Effect of tension and timing of contraction on the blood flow of the diaphragm. J. Appl. Physiol. 54: 1597-1606, 1983.
15. Biscoe, T. J., and A. Bucknell. The arterial supply of the cat diaphragm with a note on the venous drainage. Q. J. Exp. Physiol. 48: 27-33, 1963.
16. Boczkowski, J., E. Vicaut, and M. Aubier. A preparation for in vivo study of the diaphramatic microcirculation in the rat. Microvasc. Res. 40: 157-167, 1990.
17. Boczkowski, J., E. Vicaut, and M. Aubier. In vivo effects of Escherichia coli endotoxemia on diaphragmatic microcirculation in rats. J. Appl. Physiol. 72: 2219-2224, 1992.
18. Boczkowski, J., E. Vicaut, G. Danialou, and M. Aubier. Role of nitric oxide and prostaglandins in the regulation of diaphragmatic arteriolar tone in the rat. J. Appl. Physiol. 77: 590-596, 1994.
19. Brancatisano, A., T. C. Amis, A. Tully, W. T. Kelley, and L. A. Engel. Regional distribution of blood flow within the diaphragm. J. Appl. Physiol. 71: 583-589, 1991.
20. Broten, T. P., J. L. Romson, D. M. Fullerton, D. M. Van Winkle, and E. O. Feigl. Synergistic action of myocardial oxygen and carbon dioxide in controlling coronary blood flow. Clin. Res. 68: 531-542, 1991.
21. Brown, G. C. Nitric oxide regulates mitochondrial respiration and cell functions by inhibiting cytochrome oxidase. FEBS Lett. 369: 136-139, 1995.
22. Buchler, B., S. Magder, H. Katsardis, Y. Jammes, and C. Roussos. Effects of pleural pressure and abdominal pressure on diaphragmatic blood flow. J. Appl. Physiol. 58: 691-697, 1985.
23. Buchler, B., S. Magder, and C. Roussos. Effect of contraction frequency and duty cycle on diaphragmatic blood flow. J. Appl. Physiol. 58: 265-273, 1985.
24. Chang, H. Y., M. E. Ward, and S. N. A. Hussain. Regulation of diaphragmatic oxygen uptake by endothelium-derived relaxing factor (EDRF). Am. J. Physiol. 265 (Heart Circ. Physiol. 34): H123-H130, 1993.
25. Clapp, L. H., and A. M. Gurney. ATP-sensitive K+ channels regulate resting potential of pulmonary arterial smooth muscle cells. Am. J. Physiol. 262: H916-H920, 1992.
26. Comtois, A., W. Gorczyca, and A. Grassino. Anatomy of diaphragmatic circulation (Abstract). J. Appl. Physiol. 62: 238-244, 1987.
27. Comtois, A., and A. Grassino. Restriction of regional blood flow and diaphragmatic contractility. J. Appl. Physiol. 70: 2439-2447, 1991.
28. Comtois, A., F. Hu, and A. Grassino. Anatomy of human diaphragmatic circulation (Abstract). Am. Rev. Respir. Dis. 137: 384, 1988.
29. Comtois, A., C. Sinderby, N. Comtois, A. Grassino, and J. Renaud. An ATP-sensitive potassium channel blocker decreases diaphragmatic circulation in anesthetized dogs. J. Appl. Physiol. 77: 127-134, 1994.
30. Cowley, A. J., K. Stainer, J. M. Rowley, and R. G. Wilcox. Effect of aspirin and indomethacin on exercise-induced changes in blood pressure and limb blood flow in normal volunteers. Cardiovasc. Res. 19: 177-180, 1985.
31. Daut, J., W. Maier-Rudolph, N. von Beckerath, G. Mehrke, K. Gunther, and L. Goedel-Meinen. Hypoxic dilation of coronary arteries is mediated by ATP-sensitive potassium channels. Science Wash. DC 247: 1341-1344, 1990.
32. Davies, N. W., N. B. Standen, and P. R. Stanfield. The effect of intracellular pH on ATP-dependent potassium channels of frog skeletal muscle. J. Physiol. Lond. 445: 549-568, 1992.
33. Decramer, M., T. X. Jiang, and M. B. Reid. Respiratory changes in diaphragmatic intramuscular pressure. J. Appl. Physiol. 68: 35-43, 1990.
34. DeHaan, S. J., and J. E. Kendrick. Neural and metabolic control of blood flow to respiratory muscles of rabbits. Proc. Soc. Exp. Biol. Med. 167: 485-492, 1981.
35. Donovan, E., T. Trippenbach, A. Grassino, P. T. Macklem, and C. Roussos. Effect of isometric diaphragmatic contraction in cardiac output and muscle blood flow. Federation Proc. 38: 1382, 1979.
36. Duling, B. R. Changes in microvascular diameter and oxygen tension induced by carbon dioxide. Circ. Res. 32: 370-376, 1973.
37. Ekelund, U. In vivo effects of endothelin-2, endothelin-3 and ETA receptor blockade on arterial, venous and capillary functions in cat skeletal muscle. Acta Physiol. Scand. 150: 47-56, 1994.
38. Ekelund, U., U. Albert, L. Edvinson, and S. Mellander. In vivo effects of endothelin-1 and ETA receptor blockade on arterial, venous and capillary functions in skeletal muscle. Acta Physiol. Scand. 148: 273-283, 1993.
39. Evans, H. E., and G. C. Christensen. Millers's Anatomy of the Dog. Philadelphia, PA: Saunders, 1979, p. 22-323.
40. Fixler, D. E., J. M. Atkins, J. H. Mitchell, and L. D. Horwitz. Blood flow to respiratory, cardiac, and limb muscles in dogs during graded exercise. Am. J. Physiol. 231: 1515-1519, 1976.
41. Fleming, I., G. A. Gary, G. Julou-Schaeffer, J. Parratt, and J. C. Stoclet. Incubation with endotoxin activates the L-arginine pathway in vascular tissue. Biochem. Biophys. Res. Commun. 171: 562-568, 1990.
42. Gardiner, S. M., A. M. Compton, T. Bennet, R. M. J. Palmer, and S. Moncada. Control of regional blood flow by endothelium-derived nitric oxide. Hypertension Dallas 15: 486-492, 1990.
43. Gilligan, D. M., J. A. Panza, C. M. Kilcoyne, M. A. Waclawiw, P. R. Casino, and A. A. Quyyumi. Contribution of endothelium-derived nitric oxide to exercise-induced vasodilation. Circulation 90: 2853-2858, 1994.
44. Granger, H. J., A. H. Goodman, and D. N. Granger. Role of resistance and exchange vessels in local microvascular control of skeletal muscle oxygenation. Circ. Res. 38: 379-385, 1976.
45. Gray, G. A., C. Schott, G. Julou-Schaeffer, I. Fleming, J. R. Parratt, and J. C. Stoclet. The effect of inhibitors of the L-arginine/nitric oxide pathway on endotoxin-induced loss of vascular responsiveness in anesthetized rats. Br. J. Pharmacol. 103: 1218-1224, 1991.
46. Gray, H. Anatomy, Descriptive, and Surgical. Philadelphia, PA: Running Press, 1980.
47. Greig, H. W., B. J. Anson, and S. S. Coleman. Inferior phrenic artery: types of origin in 850 body-halves and diaphragmatic relationship. Q. Bull. Northwest. Univ. Med. Sch. 25: 345-350, 1951.
48. Harris, P. D., D. E. Longnecker, F. N. Miller, and D. L. Wiegman. Sensitivity of small subcutaneous vessels to altered respiratory gases and local pH. Am. J. Physiol. 231: 244-252, 1976.
49. Haynes, W. G., F. E. Strachan, and D. J. Webb. Endothelin ETa and ETb receptors cause vasoconstriction of human resistance and capacitance vessels in vivo. Circulation 92: 357-363, 1995.
50. Haynes, W. G., and D. J. Webb. Contribution of endogenous generation of endothelin-1 to basal vascular tone. Lancet 344: 852-854, 1994.
51. Herbaczynaska-Cedro, K., J. Staszewska-Barczak, and H. Janczewska. Muscular work and the release of prostaglandin-like substances. Cardiovasc. Res. 10: 413-420, 1976.
52. Hirai, T., M. D. Visneski, K. J. Kearns, R. Zelis, and T. I. Musch. Effects of NO synthase inhibition on the muscular blood flow response to treadmill exercise in rats. J. Appl. Physiol. 77: 1288-1293, 1994.
53. Hsia, C. C. W., M. Ramanathan, J. L. Pean, and R. L. Johnson, Jr. Respiratory muscle blood flow in exercising dogs after pneumonectomy. J. Appl. Physiol. 73: 240-247, 1992.
54. Hu, F., A. Comtois, and A. Grassino. Optimal diaphragmatic blood perfusion. J. Appl. Physiol. 72: 149-157, 1992.
55. Hu, F., A. Comtois, E. Shadram, and A. Grassino. Effect of separate hemidiaphragm contraction on left phrenic artery flow and O2 consumption. J. Appl. Physiol. 69: 86-90, 1990.
56. Hussain, S. N. A. Role of nitric oxide in endotoxin-induced metabolic and vascular dysregulation of the canine diaphragm. Am. Rev. Respir. Dis. 152: 683-689, 1995.
58. Hussain, S. N. A., R. Graham, F. Rutledge, and C. Roussos. Respiratory muscle energetics during endotoxic shock in dogs. J. Appl. Physiol. 60: 486-493, 1986.
59. Hussain, S. N. A., and S. Magder. Diaphragmatic intramuscular pressure in relation to tension, shortening and blood flow. J. Appl. Physiol. 71: 159-167, 1991.
60. Hussain, S. N. A., and C. Roussos. Distribution of respiratory muscle and organ blood flow during endotoxic shock in dogs. J. Appl. Physiol. 59: 1802-1808, 1985.
61. Hussain, S. N. A., C. Roussos, and S. Magder. Autoregulation of diaphragmatic blood flow in dogs. J. Appl. Physiol. 64: 329-336, 1988.
62. Hussain, S. N. A., C. Roussos, and S. Magder. Effect of tension, duty cycle, and arterial pressure on diaphragmatic blood flow. J. Appl. Physiol. 66: 968-976, 1989.
63. Hussain, S. N. A., F. Rutledge, R. Graham, S. Magder, and C. Roussos. Effects of norepinephrine and fluid administration on diaphragmatic oxygen consumption in septic shock. J. Appl. Physiol. 62: 1368-1376, 1987.
64. Hussain, S. N. A., G. Simkus, and C. Roussos. Respiratory muscle fatigue: a cause of ventilatory failure in septic shock. J. Appl. Physiol. 58: 2033-2040, 1985.
65. Hussain, S. N. A., D. J. Stewart, J. P. Ludemann, and S. Magder. Role of endothelium-derived relaxing factor in active hyperemia of the canine diaphra