INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

WHEN OXYGEN DELIVERY to active skeletal muscle is insufficient for the ongoing metabolic demands, metabolites accumulate and stimulate afferents within the active skeletal muscle that elicit a powerful pressor response known as the muscle metaboreflex. Activation of the muscle metaboreflex during exercise elicits increases in heart rate, cardiac output, systemic arterial pressure, ventricular performance, central blood volume mobilization, and vasoconstriction in the renal and nonischemic active skeletal muscle vasculatures (9, 11, 12, 14, 16, 17, 22, 23, 25, 33). In addition, meta-boreflex activation can also increase the circulating levels of vasoactive hormones (13, 18). These marked cardiovascular responses are buffered by arterial (23) and cardiopulmonary (4) baroreflexes. Studies from our laboratory and from others (11, 12, 16-18, 22, 23, 33) have shown that in dogs during mild exercise a clear threshold exists for metaboreflex activation, i.e., initial reductions in blood flow to active skeletal muscle (i.e., hindlimbs) do not elicit any metaboreflex responses. Only when oxygen delivery is reduced below a threshold level do substantial metaboreflex pressor responses occur (25). In contrast, during moderate workloads no clear threshold exists, indicating that the metaboreflex may be tonically active or that the prevailing level of blood flow approximates the threshold for the reflex, and any reduction in perfusion to the active skeletal muscle will engage the muscle metaboreflex (20, 22, 33).

The cardiovascular responses to dynamic exercise are markedly altered in subjects with ventricular dysfunction. Previous studies have shown in normal dogs that little if any reductions in blood flow occur in inactive areas even during severe exercise (10, 29). However, during exercise at or near maximal levels in dogs with cardiac impairment, substantial vasoconstriction occurs in the renal and mesenteric vascular beds, and even active skeletal muscle becomes relatively vasoconstricted (e.g., the rise in iliac blood flow with exercise is depressed) (10, 29). The mechanisms mediating these altered cardiovascular responses to dynamic exercise in subjects with ventricular dysfunction are unknown. Inasmuch as blood flow to active skeletal muscle is lower in this setting, this relative underperfusion could potentially elicit exaggerated sympathoactivation via the muscle metaboreflex because at even moderate workloads in normal animals no clear threshold is discernible for the reflex.

Little is known regarding the effects of heart failure (HF) on the strength and mechanisms of the muscle metaboreflex. To our knowledge, only three studies have directly addressed this issue. Sterns et al. (28) used the technique of posthandgrip circulatory arrest (PHG-CA) to isolate the effects of the muscle metaboreflex in normal humans and subjects with HF (i.e., inflation of a blood pressure cuff on the upper arm to suprasystolic levels at the cessation of static handgrip, thereby entrapping metabolites within the formerly active muscle). They observed that the normally sustained or even elevated levels of muscle sympathetic nerve activity (MSNA) in this setting were suppressed in the subjects with HF, indicating that muscle metaboreflex responses may be impaired in HF. These authors proposed that the attenuation of the muscle metaboreflex may be due to desensitization of the receptor afferents. In a subsequent study from the same laboratory, Silber et al. (27) showed that during fatiguing rhythmic handgrip MSNA increased sooner in subjects with congestive HF and that during PHG-CA MSNA remained elevated, whereas in normal humans MSNA decreased toward control levels. The authors also observed markedly elevated forearm acidosis as well as greater H₂PO₄ accumulation (metabolites thought to be important in mediating metaboreflex responses) in the HF subjects during PHG-CA. Subsequently, Shoemaker et al. (26) observed similar differences between normal human and HF subjects in response to 5 min of partially ischemic rhythmic handgrip exercise, i.e., higher venous lactate and H⁺ levels and a larger pressor response in the HF group. These data indicate that during fatiguing rhythmic handgrip the muscle metaboreflex may become tonically active in HF subjects (possibly due to lower skeletal muscle blood flow, although this was not measured in either study) and thereby contribute to the increased sympathetic activity. However, with the PHG-CA technique, the muscle metaboreflex responses are observed during the recovery from exercise rather than during exercise per se. This may be an important issue because the mechanisms mediating the pressor response may be quite different in the two settings. For example, many studies have shown that during PHG-CA heart rate declines toward resting levels with a time course similar to that during the normal recovery from static handgrip without circulatory arrest (30). In contrast, in normal dogs muscle ischemia during exercise results in a marked increase in heart rate, and the increase in cardiac output is the primary mechanism of the metaboreflex pressor response (16, 17, 22, 33).

The objective of this study was to investigate the effects of HF on the strength and mechanisms of the muscle metaboreflex in conscious, chronically instrumented dogs during dynamic exercise. We hypothesized that the cardiac component of the reflex would be attenuated in HF, thereby limiting the reflex pressor response. In addition, we hypothesized that in HF skeletal muscle blood flow would be decreased during exercise, which may elicit tonic activation of the muscle metaboreflex. Some of these data have been published previously in abstract form (7, 8).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental Subjects

Surgical Procedures

In the second procedure a retroperitoneal incision was made on the left flank just above the iliac crest, and an 8- or 10-mm ultrasonic or electromagnetic blood flow transducer was placed on the terminal aorta to measure blood flow to the hindlimbs (TAQ). All vessels arising from the terminal aorta between the flow transducer and the iliac arteries were ligated. A hydraulic occluder (InVivo Metric) was also placed around the aorta distal to the flow probe to provide the means to progressively occlude the aorta. A 20-gauge polyvinyl catheter (Tygon, S54-HL, Norton) was introduced into the midthoracic aorta to monitor systemic arterial pressure (SAP). A pulsed Doppler blood flow probe (Crystal Biotech) was implanted on the left renal artery to measure renal blood flow (RBF) in seven animals, and ultrasonic transit-time probes (Transonic Systems) were installed in three dogs. All leads and the catheter were tunneled subcutaneously to the interscapular exit site. In a final procedure, a 20-gauge catheter (Tygon, S54-HL, Norton) was passed into atriocaval junction via the right external jugular vein to measure right atrial pressure (RAP). Catheters were also inserted through small side branches of the femoral artery and the femoral vein for separate studies. All catheters were exteriorized in the interscapular region.

After all surgical procedures, postoperative pain was controlled by parenteral administration of buprenorphrine (0.02-0.04 mg · kg¹ · 4 h¹) and acepromazine maleate (0.1 mg · kg¹ · 4 h¹) as needed. Antibiotic therapy included prophylactic perioperative administration of cefazolin (500 mg iv) and postoperative treatment with cephalexin (30 mg/kg, per os, 3 times/day) for at least 1 wk.

Data Collection

Exercise Protocols

Data Reduction and Statistical Analysis

The response of each dog was characterized by combining the data analyzed for each run under the control or HF conditions, and then the individual averages for the free flow, threshold, and maximum occlusion points were combined to describe the overall averaged response of the group (means ± SE) as plotted in Fig. 3. Utilizing the averaged responses for each subject, further statistical analyses were performed on the data with Systat software (Version 5, SPSS). Statistical significance was defined as P < 0.05 (one-tailed). For comparison of standing rest and free-flow conditions at low workload or moderate workload, paired t-tests were performed. For data collected during exercise, one- and two-way repeated measures analysis of variance was used for comparison with respect to cardiac condition and workload. When the analysis of variance indicated significant interactions, comparison between individual means was performed by a C-matrix test for simple effects. For the neurohumoral data (NE, PRA, and AVP), a Friedman two-way analysis of variance was performed with subsequent comparisons by a Wilcoxon signed rank test. Because of technical difficulties, we were unable to obtain all measurements for all subjects in all conditions. Table 1 shows the number of subjects (n) for each measurement in each condition.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Comparison of Control and HF States at Standing Rest

Comparison of Standing Rest to Exercise at 3.2 km/h at 0% Grade

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Hemodynamic parameters at standing rest versus exercise under free-flow conditions. Left panels, 3.2 km/h at 0% grade; right panels, 6.4 km/h at 10% grade; solid bars, control; open bars, heart failure. * P < 0.05 vs. control,  P < 0.05 vs. rest;  P < 0.05 vs. 3.2 km/h at 0% grade.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Neurohumoral responses at standing rest, free-flow exercise, and muscle metaboreflex activation. Left panels, 3.2 km/h at 0% grade; right panels, 6.4 km/h at 10% grade; solid bars, control; open bars, heart failure. * P < 0.05 vs. control,  P < 0.05 vs. rest; § P < 0.05 vs. free flow.

Comparison of Standing Rest to Exercise at 6.4 km/h at 10% Grade

In HF at moderate workload (Fig. 1), CO, SV, TAQ, RBF, RVC, and SVC were significantly depressed with respect to control; however, SAP was not significantly different from control. HR was elevated with respect to control, and CO and HR both rose significantly from standing rest at the onset of exercise. SV increased significantly beyond standing rest and low workload. RAP was markedly elevated from standing rest and from control. SVC was greater at moderate workload than at standing rest but depressed with respect to control. In HF, PRA, AVP, and NE all rose significantly at moderate workload from standing rest and their respective control values (Fig. 2).

Circulatory Response to Metaboreflex Activation During Dynamic Exercise

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Hemodynamic parameters during free-flow exercise and on muscle metaboreflex activation. Filled circles, control, 3.2 km/h at 0% grade; open circles, heart failure, 3.2 km/h at 0% grade; filled triangles, control, 6.4 km/h at 10% grade; open triangles, heart failure, 6.4 km/h at 10% grade.

In HF at the low workload (Fig. 3, open circles), striking and significant changes occurred in the responses to metaboreflex activation vs. the control experiments. During free flow and subthreshold reductions in TAQ, HR and RAP were elevated significantly with respect to control, whereas SAP, CO, SV, and RBF were depressed. On elicitation of the metaboreflex, SAP, RAP, and HR rose significantly; however, no significant increase in CO occurred. In addition, with metaboreflex activation during mild exercise in HF a significant decrease in SV occurred. SVC was maintained comparable to control values until the metaboreflex was activated, and then it fell significantly. As in the control runs, RBF did not fall significantly with elicitation of the reflex. Under subthreshold conditions, RVC was depressed and comparable to the value observed during metaboreflex activation (i.e., at maximal occlusion) in the control runs but fell significantly on elicitation of the reflex. PRA (Fig. 2) was higher than control but did not increase significantly from the free-flow value on metaboreflex activation. Concentrations of AVP and NE rose significantly from free-flow levels with activation of the reflex.

At moderate workload under control conditions, HR, CO, SAP, and RAP were much higher than at low workload (Fig. 3, filled triangles). Generally, any reduction of TAQ resulted in increases in HR, CO, and SAP, however, RAP and SV did not increase significantly on elicitation of the metaboreflex. RBF dropped significantly at the maximal occlusion. RVC was significantly lower under free-flow conditions or during subthreshold aortic occlusions than at rest or low workload and fell further on reflex activation. Metaboreflex activation did not result in higher PRA but did significantly increase concentrations of AVP and NE (Fig. 2).

In HF during exercise at moderate workload, any reduction in TAQ resulted in increases in HR and SAP. With maximal aortic occlusion, SAP increased significantly but was substantially lower than the control value, and CO did not increase significantly. SV remained ~75% of control. RAP was strikingly elevated beyond the control values under free-flow conditions but rose insignificantly at maximal aortic occlusion. SVC was lower than control but tended to fall insignificantly with reductions of TAQ. Before any aortic occlusion, RBF was similar to that observed in HF at low workload and fell significantly at maximal occlusion. RVC was significantly lower than control values (comparable to the control value observed at maximal occlusion) and, on partial aortic occlusion, RVC dropped significantly more. Plasma NE, AVP, and PRA were greatly elevated under free-flow conditions with respect to control values and became markedly elevated at maximum occlusion.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The major new finding in this study is that, in dogs with moderate HF, activation of the muscle metaboreflex still produces a significant pressor response; however, the mechanisms of this pressor response are markedly altered compared with those observed in control experiments. Specifically, in HF, the CO component of this reflex is virtually abolished and the pressor response occurs via peripheral vasoconstriction by markedly exaggerated sympathoactivation and the substantial release of vasoactive hormones. Finally, during moderate exercise in HF, skeletal muscle blood flow is reduced below the metaboreflex threshold as established in control experiments; therefore the exaggerated sympathoactivation and higher levels of vasoactive hormones in this setting may stem from tonic activation of the metaboreflex.

Effect of HF on Hemodynamic Responses to Metaboreflex Activation

In control experiments under free-flow conditions from standing rest to the moderate workload, substantial increases in HR, SV, CO, and RAP occurred. As described previously (33) the metaboreflex may already be active at this workload under free-flow conditions in normal dogs, so positive inotropic stimulation mediated by the metaboreflex may already be engaged. Furthermore, the rise in RAP from standing rest to free-flow exercise, probably in part due to increased activity of the skeletal muscle pump (24), indicates that the rise in SV may result from both increased contractility as well as the Frank-Starling mechanism. During the stepwise partial aortic occlusions, HR increased progressively yet SV was maintained constant across the measured range from free flow to maximum occlusion despite a progressively increasing afterload and decreasing filling time. This sustained stroke volume coupled with increased heart rate caused substantial increases in CO. As observed previously (17, 22, 33), during both mild and moderate workloads, the pressor response under control conditions was mediated via the increase in CO inasmuch as little net decrease in systemic vascular conductance occurred.

The metaboreflex cardiac responses were markedly altered after induction of HF. CO was depressed during free-flow exercise at both workloads, and the increase in CO with metaboreflex activation was virtually abolished. Despite a sustained tachycardic response, little change in CO occurred due to a concomitant fall in stroke volume. Previously, White et al. (32) demonstrated that tachycardia in and of itself does not augment CO in normal dogs within the range of HR observed in this study, owing to a fall in SV due to decreased filling times. O'Leary et al. (17) concluded that improvements in ventricular contractility must accompany increases in HR to augment CO. Thus it appears that, after induction of HF, metaboreflex-mediated increases in ventricular performance were markedly attenuated. Although still chronotropically competent, the hearts became relatively inotropically insensitive and likely afterload sensitive. With meta-boreflex activation in HF, despite a significant increase in filling pressure, stroke volume could not be maintained with the increase in afterload and decreased ventricular filling time. However, CO was significantly higher during free-flow exercise at the moderate workload versus the low workload in HF. In this setting, stroke volume was higher, probably due to the increased filling pressure and relatively low afterload (compared with that during metaboreflex activation).

Despite the inability to increase CO, activation of the metaboreflex in HF did result in a significant increase in SAP. However, in contrast to the control conditions wherein the rise in SAP was virtually solely due to the increase in CO, after induction of HF this pressor response was almost solely due to peripheral vasoconstriction. Quantification of the strength of the metaboreflex based on the ability to increase SAP can be problematic in this model. As described previously by Wyss et al. (33), in this experimental model of metaboreflex activation, SAP will increase due to reflex activation and also due to the mechanical effects of the partial vascular occlusion, i.e., a substantial mechanical reduction in conductance to a vascular bed that receives a large fraction of the CO can cause increases in SAP (15). Because perfusion pressure is a function of CO and SVC, the effects of the observed changes in CO and SVC on SAP can be calculated. This calculated change in SAP would therefore reflect the pressor response generated via the metaboreflex-mediated cardiac and peripheral vasomotor changes, independent of the passive mechanical effects of the occluder. Based solely on the observed reflex changes in CO and SVC, in control experiments arterial pressure would increase by 35 and 28 mmHg at the low and moderate workloads, respectively, whereas the observed changes were 56 and 47 mmHg. This difference between the changes in SAP calculated from the changes in CO and SVC versus those actually observed can be attributed to the mechanical effects of the vascular occluder (33). In HF, the average predicted changes in SAP during metaboreflex activation based on the observed reflex changes in CO and SVC were only 25 and 14 mmHg at the low and moderate workloads, respectively. Thus, despite significant peripheral vasoconstriction with metaboreflex activation, after induction of HF the inability to increase CO likely limits the ability of the metaboreflex to increase SAP and therefore limits the ability of the reflex to partially restore blood flow to the ischemic active skeletal muscle (20).

In the control experiments, no significant increase in RAP occurred with metaboreflex activation at either workload, although at the higher workload there tended to be a small increase that was not statistically significant. As discussed by Sheriff et al. (22), the lack of a change in RAP with a large increase in CO provides indirect evidence for metaboreflex-mediated central mobilization of blood volume that maintains ventricular filling pressure despite a large increase in CO. These conclusions are supported by the data obtained after induction of HF. In this setting, little or no change in CO occurred with metaboreflex activation, and large increases in RAP were observed at the low workload. At the higher workload RAP tended to increase with metaboreflex activation, but this change was not statistically significant, although it should be noted that RAP was already markedly elevated at this workload and further central volume mobilization via the metaboreflex may be limited.

Systemic vascular responses. Although SVC did not change appreciably with metaboreflex activation, this does not imply that regional circulations were unaffected (9, 11, 22). As shown previously, metaboreflex activation causes significant reductions in RVC (12, 22), and Mittelstat et al. (11) have shown that metaboreflex activation also decreases vascular conductance in nonischemic active skeletal muscle. In HF, although the absolute values of RVC and RBF were much lower with respect to the values observed at standing rest in the control experiments, the patterns of the responses were similar to those observed in the control experiments during both free-flow exercise and partial aortic occlusion. During low workload under free-flow conditions, RVC did not vary significantly from standing rest in HF. With partial aortic occlusion during the low workload a large decrease in RVC occurred. During moderate workload in HF, RBF under free-flow conditions was decreased ~10% from the values observed at standing rest; this approached statistical significance (P = 0.06, one-tailed). Previously, Mittelstadt et al. (12) used a similar model to study the responses to metaboreflex activation of paired normal and denervated kidneys at an intermediate workload of 10 km/h at 0% grade in otherwise normal dogs. Their results were similar to those obtained in the present investigation during control experiments with mild exercise (i.e., no change in RBF but decreases in RVC). They concluded that the renal vascular responses of the both normal and denervated kidneys in dogs likely reflect an autoregulatory response to increased aortic pressure or a humorally mediated response, rather than increased renal sympathetic nerve activity. In contrast, in this study, with metaboreflex activation during moderate exercise after the induction of HF, significant decreases in both RBF and RVC occurred. This is not likely to be due solely to autoregulation. Millard et al. (10) also studied renal vascular responses in paired normal and denervated kidneys during severe exercise in groups of dogs with either intact or impaired hearts and with or without alpha receptor blockade. They demonstrated that the changes in RVC were not dependent on renal innervation or circulating catecholamines in normal dogs. However, in HF, blood flow to the innervated kidneys was depressed at rest and fell drastically during severe exercise. In contrast, blood flow in the denervated kidneys at rest was comparable to the control value, and the mild decline observed during severe exercise was probably due to denervation supersensitivity to circulating catecholamines because this could be prevented by -receptor blockade. They concluded that autoregulatory and nonadrenergic mechanisms control RBF in normal dogs but in HF substantial renal sympathetic activity reduces RBF during severe exercise, and an adrenergic vasoconstrictive response to circulating catecholamines serves as a secondary response mechanism to reduce RBF. Similarly, Vatner et al. (29) observed marked renal vasoconstriction in response to severe exercise in dogs after induction of atrioventricular block, which likely limited the normal increase in CO. These findings suggest that tonic renal sympathetic nerve activity in HF decreases RVC and reduces RBF but it is probably not due to the muscle metaboreflex during mild exercise. However, during moderate workload in HF, the TAQ under free-flow conditions is below the metaboreflex threshold for RVC as established in control experiments, and thus the reduced RVC under these conditions may be partially due to metaboreflex activation. With further reductions in TAQ by stepwise aortic compression, RVC fell dramatically, and RBF fell significantly as well. Indeed, in two animals metaboreflex activation caused RBF and RVC to decrease to virtually zero after induction of HF. Thus the metaboreflex causes significant decreases in RBF only in circumstances where it is strongly activated, such as the combination of a moderate workload with muscle ischemia or possibly severe exercise in HF.

Neuroendocrine Responses

In HF, plasma NE tripled at standing rest, approximating the maximum response observed with metaboreflex activation during low workload. However, NE did not rise further with free-flow exercise at low workload but only upon metaboreflex activation. This, coupled with the demonstrable thresholds for HR and SAP responses, supports the concept that the metaboreflex is not active during mild exercise in HF, and the plasma NE concentrations observed at standing rest and free flow at low workload are due to some other mechanism. During moderate workload under free-flow conditions, the fivefold increase in plasma NE indicates intense sympathetic activation. On partial aortic occlusion, the plasma NE response became extreme, indicating massive sympathetic activation.

PRA. A number of studies have indicated that PRA rises with exercise workload under free-flow conditions in a number of species (5, 31), but few studies have directly addressed the role of the metaboreflex with respect to recruitment of the renin-angiotensin system. In a recent article, Nishiyasu et al. (13) demonstrated that metaboreflex activation stimulates renin and AVP release. Neither mild free-flow exercise nor metaboreflex activation during mild free-flow exercise elicited any increase of PRA in normal dogs (18). Similar results during control experiments at the low workload were observed in this study, but at the moderate workload under free-flow conditions, PRA was significantly elevated compared with standing rest, yet it did not increase further with definitive metaboreflex activation by partial aortic occlusion. Upon the induction of HF, PRA was elevated with respect to control at standing rest and during exercise at both workloads. As in the control experiments, no metaboreflex-induced release of renin occurred at the low workload, although sympathetic activation was evident at standing rest and increased with exercise and reflex activation. Apparently, the increased sympathetic activity mediated by the metaboreflex was insufficient to induce renin secretion at this workload. During moderate free-flow exercise in HF, the PRA increased dramatically (approximately twice that of standing rest) and, on partial occlusion of the terminal aorta, PRA rose further. Thus at this workload the renin response to metaboreflex activation likely contributed to the systemic vasoconstriction under free-flow conditions as well as during experimental metaboreflex activation.

AVP. As in the secretion of renin, previous studies have demonstrated that AVP secretion occurs during exercise, but relatively little is known about the relationship between the metaboreflex and AVP secretion. Yamashita et al. (34) have shown that stimulation of the group III and IV nerve afferents (the same classes of muscle afferents that mediate the muscle metaboreflex) increases the activity of AVP-secretory neurons located within the supraoptic nucleus. As noted above, Nishiyasu et al. (13) concluded that in humans metaboreflex activation elicits significant AVP secretion. There are two previous reports from this laboratory dealing with this subject. In the first, O'Leary et al. (18) demonstrated that the metaboreflex activation can stimulate AVP secretion during mild exercise, but only after ganglionic blockade of the large arterial pressor response. They attributed this effect to an inhibitory influence of the arterial baroreflex. In the second study, examining responses to graded exercise under free-flow conditions (19), heavy exercise induced AVP secretion during -blockade, but the authors could not discriminate whether this response was mediated by the baroreflex (in response to relative hypotension) or by the metaboreflex (in response to decreased hindlimb blood flow).

In this study, we demonstrated in control experiments a very small but statistically significant increase of plasma AVP in response to metaboreflex activation at the low workload and, under free-flow conditions, at moderate workload. On metaboreflex activation at moderate workload, a larger response occurred that may have directly induced vasoconstriction (6). In HF at standing rest and under free-flow conditions at low workload, although the values are significantly higher than respective controls, they probably do not have appreciable vasoconstrictive effects (6). However, in HF, metaboreflex activation induced increases of plasma AVP at the low workload that approach vasoactive concentrations (6). Furthermore, in HF at moderate workload, under free-flow conditions, AVP levels are substantially higher than observed under normal circumstances and rise dramatically on partial aortic occlusion. In this circumstance, plasma AVP probably produces direct vasoconstrictive effects (6). In all, these data further support the concept that the muscle metaboreflex is inactive during mild exercise in HF, is likely active during moderate exercise in HF, and on experimentally induced muscle ischemia (in normal or HF states) is clearly capable of recruiting potent humoral mechanisms to support or increase blood pressure.

Perspectives: Muscle Metaboreflex in HF

We do not know whether similar reductions in hindlimb blood flow elicit the same extent of metaboreceptor activation in HF. Inasmuch as the free-flow levels of TAQ are depressed in this setting and the metaboreflex may indeed be already tonically active at the moderate workload, it is possible that similar reductions in TAQ elicit larger increases in metabolite concentrations and thereby higher levels of metaboreceptor activation, as evidenced by the often markedly higher levels of metaboreflex-mediated increases in circulating vasoactive hormones and occasionally extreme renal vasoconstriction. In the present study the experiments were performed at the same absolute level of exercise before and after the induction of HF, and these workloads likely reflected different relative levels of exercise intensity (e.g., % maximal O₂ consumption). Theoretically, the strength of the muscle metaboreflex in normal dogs may become less as workload increases toward the maximal levels of CO. In this setting, further increases in CO via the muscle metaboreflex would become progressively limited. In this study, the failure of the muscle metaboreflex to elicit any significant increase in CO at the low workload is not likely due to such limitations because, during the moderate workload, the free-flow level of CO increased substantially above that at the lower workload. To what extent the maximal CO is limited in this model has not been established, and this may be reflected in the limited ability to increase output further at the moderate workload. Finally, the metaboreflex is buffered by the arterial and cardiopulmonary baroreflexes (4, 23) and the gains of both of these baroreflexes are reduced in HF (2) such that the same level of metaboreceptor activation may yield larger reflex responses due to reduced baroreflex buffering. The extent to which HF alters the competing relationship between baroreflexes and the metaboreflex is unknown.

In summary, in normal dogs the metaboreflex is inactive at low workloads and possibly somewhat active at moderate workloads. On engagement, it normally operates by increasing cardiac performance, with adjustments to the peripheral vasculature that direct blood flow to the active muscles and support RAP. Humoral mechanisms are recruited only with strong metaboreflex activation, and therefore usually serve as secondary response mechanisms. However, with failure of the cardiac pump, during moderate and heavier exercise the metaboreflex is likely tonically activated, and the intense sympathetic stimulation plus the recruitment of potent vasoactive hormones become the primary mechanisms to support and increase blood pressure by peripheral vasoconstriction.