The influence of heart rate on baroreceptor fibre activity in the carotid sinus and aortic depressor nerves of the rabbit

The arterial baroreceptors and their afferent fibres provide the sensory arm of the reflex that regulates systemic arterial pressure. We have examined whether the relationship between mean baroreceptor discharge and mean arterial pressure is altered when heart rate changes. Experiments were performed on pentobarbitone-anaesthetized rabbits. We recorded the activity of single and multifibre preparations of the carotid sinus (CSN) and aortic depressor nerves (ADN). Data were collected under control conditions and while heart rate was increased by approximately 30-35% by right atrial pacing. Baroreceptor regions were exposed to ramps of pressure (from approximately 25 to 140 mmHg, at approximately 0.5-1 mmHg s(-1)), generated by inflation and deflation of cuffs placed around the inferior vena cava and descending thoracic aorta. Response curves relating baroreceptor discharge to mean pressure were constructed and fitted with third-order polynomial expressions. To provide a measure of an effect of an increase in heart rate on the response curve in the region of the normal operating pressure, we calculated the position of the test response curve relative to the position of the control curve at 90 mmHg (deltaBP(90)). For the ADN, the activity of single fibres (presumptive myelinated fibres) was unaffected by increasing heart rate (deltaBP(90) = +0.1 +/- 1.0 mmHg), while single fibres in the CSN showed a small increase in activity (deltaBP(90) = -1.5 +/- 0.3 mmHg). In multifibre preparations there was a small increase in activity that may be attributable to additional activity in unmyelinated fibres (ADN, deltaBP(90) = -3.4 +/- 1.2 mmHg; CSN, deltaBP(90) = -5.2 +/- 0.9 mmHg). We conclude that the mean discharge of arterial baroreceptors remains a reliable index of mean arterial pressure in the presence of substantial changes in heart rate.     

Interaction of somatic and cardiopulmonary receptors in control of renal circulation.

The purpose of this study was to determine if there is an important interaction between somatic and cardipulmonary receptors in the control of vasomotor outflow to the kidney. This interaction was examined by determining the renal vasoconstrictor responses to afferent electrical stimulation (30 V, 1 ms, 20--40 Hz) of the sectioned sciatic nerve in 8 chloralose-anesthetized dogs with sinoaortic deafferentation. During isovolemia, sciatic stimulation resulted in significant increases in arterial pressure and heart rate, and in renal vasoconstriction. Volume expansion significantly attenuated and vagotomy significantly augmented the renal vasoconstrictor responses to sciatic stimulation. These interventions did not significantly influence the arterial pressure or heart rate responses to sciatic stimulation. In 4 dogs with aortic nerves sectioned and carotid sinuses isolated and perfused at constant pressure (135 mmHg), the renal vasoconstrictor responses to stimulation were attenuated by volume expansion and augmented by vagotomy. These data show that in the absence of the arterial baroreceptors or with intermediate levels of carotid baroreceptor activation, volume expansion and, thus, augementation of discharge of cardiopulmonary receptors (vagal afferents) markedly attenuated the renal vasoconstrictor responses to somatic afferent stimulation.     

Participation of the anterior hypothalamus in the baroreceptor reflex

1. On the basis of discrete electrical stimulation in the pre-optic region and anterior hypothalamus of anaesthetized cats, a depressor area has been defined, stimulation of which elicits a fall of arterial blood pressure of 30-50 mm Hg and a bradycardia of some 25%, caused by inhibition of sympathetic vasomotor tone and by vagal activation respectively. These are accompanied by a reduction in rate and depth of respiration.2. The depressor area, from which this pattern of response is elicited, lies ventral and caudal to the anterior commissure, and extends caudally in the dorsal hypothalamus, dorsal to the fornix.3. The pattern of response elicited from identified points in the depressor area was shown to be indistinguishable from that to baroreceptor afferent stimulation.4. A lesion destroying the hypothalamic depressor area bilaterally reduced the response to baroreceptor afferent stimulation. Lesions in the medullary depressor area which spared a large part of the nucleus of the tractus solitarius also reduced, but did not abolish, the baroreceptor reflex response. The two lesions combined abolished the reflex.5. It is concluded that the whole brain-stem depressor area, from the hypothalamus through the mid-brain to the medulla, constitutes a functional unit which integrates the response to baroreceptor afferent stimulation.     

Central interaction between the baroreceptor reflexes from the carotid sinus and aortic arch.

The cardiovascular response to combined stimulation of the aortic nerve (AN) and carotid sinus nerve (CSN) with trains of electrical stimuli was studied in dogs under morphine-chloralose anesthesia. Combined stimulation of both nerves ipsilaterally resulted in systemic arterial depressor responses and cardiac slowing that were 59% and 77% greater (P less than 0.05) than the respective sum of the responses to separate stimulation of these nerves. The depressor response to combined stimulation of these nerves contralaterally was not different (P greater than 0.05) from the sum of the separate responses. However, cardiac slowing to contralateral stimulation of the nerves remained 43% greater (P less than 0.05) than the sum of the individual responses. The results suggest that an interaction between the aortic and carotid baroreceptor reflexes occurs centrally, which causes a greater than additive effect on the autonomic output. The degree of interaction between the reflexes did not change as the delay between the stimulus trains was varied over an interval of 0 and 100 ms. At intervals greater than 100 ms the reflex responses were additive.     

Contrasting reflex influences of cardiac afferent nerves during coronary occlusion

Afferent neurons contained within cardiac sympathetic nerves may have important influences on the circulation when activated during myocardial ischemia. Although such activation is known to reflexly excite upper thoracic sympathetic efferent neurons, effects on other components of sympathetic outflow are unknown. Therefore, cardiac sympathetic afferent nerves were stimulated by occlusion of coronary arteries to investigate their reflex influences on renal sympathetic nerve activity and systemic arterial blood pressure. Responses were observed in anesthetized cats in which sympathetic and/or vagal cardiac afferent nerves remained intact and arterial baroreceptors remained intact or had been denervated. Stimulating sympathetic afferent neurons caused excitation of renal nerve activity, which was accompanied by variable changes in arterial pressure. Stimulation of vagal afferents by coronary occlusion consistently produced inhibition of renal nerve activity and marked depressor responses. When both components of cardiac innervation remained intact, increases or decreases in renal nerve activity and blood pressure were elicited by coronary artery occlusion in the presence or absence of arterial baroreceptors. These results illustrate that cardiac sympathetic afferent nerves can contribute significantly to cardiovascular control during myocardial ischemia.     

Role of cardiac sympathetics in the tonic circulatory restraint by vagal afferents.

In anesthetized cats with aortic nerves sectioned and carotid arteries occluded, we determined the role of cardiac sympathetic nerves on the tonic inhibitory restraint by cardiac vagal afferents on the cardiovascular system. The effect of afferent vagal blockade on mean arterial pressure and cardiac contractility was determined when sympathetic tone to the heart was altered. Bilateral cardiac sympathectomy produced a significant decrease in left ventricular dP/dt and attenuated the arterial pressure response to afferent vagal cold block to less than 40% of the control. The increase in dP/dt normally observed with vagal blockade was also reduced significantly. Increasing dP/dt by efferent stimulation of cardiac sympathetic nerves restored the arterial pressure response to vagal blockade to near control levels. While the vagal inhibitory activity appeared to be dependent on the resting dP/dt, left ventricular peak pressure did not seem to be contributing to the reflex. Thus, the inhibitory effects of vagally mediated reflexes from the heart which contribute to arterial pressure regulation appear to be influenced by changes in cardiac contractility induced by cardiac sympathetic nerve stimulation.     

The effect of efferent discharges in renal nerves on the activity of arterial mechanoreceptors in the kidney in rabbit

1. The effect of sympathetic efferent flow in renal nerves on afferent discharges from renal arterial mechanoreceptors was observed in dissected filaments of renal nerves in the rabbit.2. An increase in afferent discharge rate was observed after renal nerve stimulation.3. Perfusate accumulated during renal nerve stimulation in an isolated kidney preparation showed a vasoconstrictor effect on blood vessels in rabbit ear.4. Infusion of a small amount of (+/-)-noradrenaline increased afferent discharge rate.5. Modification of blood pressure levels by chemical means modified renal afferent discharge rate.6. It is suggested that there is an efferent control mechanism on the activity of renal arterial mechanoreceptors by a vasomotor reflex.     

Studies on blood-pressure regulation by the use of a baroreceptor analog

Summary In rabbits anesthetized with urethane the aortic and carotid sinus nerves were stimulated by an electric baroreceptor analog. In closed loop experiments the analog buffered blood pressure alterations and alterations of heart rate following the blockade of aortic nerves and clamping the common carotid arteries. The carotid sinus nerves appeared to be not more sensitive to baropacing by the analog than the aortic nerves. Blood pressure and heart rate oscillations induced by vagal stimulation with the analog were damped and finally disappeared when the electric receptor analog was used to stimulate simultaneously the depressor nerves. Analysis of the baroreceptor analog in open loop experiments gave the same results as obtained on natural baroreceptors: pulse pressure sensitivity changed with mean pressure, amplitude, and pulse rate. The transfer function of the analog is the same as transfer functions of carotid sinus nerves are. It is concluded, that the baroreceptor analog may be useful for closed loop baropacing treatment of hypertension.This work was supported by the Deutsche Forschungsgemeinschaft (Ze 67/1).     

The relative roles of the aortic and carotid sinus nerves in the rabbit in the control of respiration and circulation during arterial hypoxia and hypercapnia

1. The respiratory and circulatory effects of graded arterial hypoxia, alone or with superadded hypercapnia, were studied in four groups of unanaesthetized rabbits including normal animals and those with selective section of the aortic nerves, selective section of the carotid sinus nerves and section of both sets of nerves.

2. When measured 2-4 days after selective section of the carotid sinus nerves the resting respiratory minute volume and arterial P_O2 were lower and the P_CO2 higher than normal. These effects were not observed after selective section of the aortic nerves. Selective aortic nerve section, and selective carotid sinus nerve section each produced a similar increase in the resting arterial pressure and heart rate, but were without effect on the resting cardiac output.

3. During arterial hypoxia reflex respiratory and circulatory effects ascribable to arterial chemoreceptor stimulation (hyperventilation, bradycardia, vasoconstriction) were mediated for the most part through the carotid sinus nerve. In animals with only the aortic nerves intact the circulatory response was determined largely by the opposing effects of aortic baroreceptor reflexes and the local peripheral dilator action of hypoxia.

4. The circulatory effects of hyperventilation induced by hypercapnia during arterial hypoxia, in animals with both aortic and carotid sinus nerves cut were small.

5. The results suggest that relatively few chemoreceptor fibres originate from the aortic region in the rabbit, though the carotid sinus and aortic nerves both contain baroreceptor fibres.

     

Baroreceptor sensitive neurones in the anterior hypothalamus of the cat

1. The activity of 372 neurones, located within or close to the hypothalamic depressor area, was recorded on raising the pressure within a vascularly isolated carotid sinus.2. When the intrasinusal pressure was raised abruptly to 200 mm Hg the activity of fifteen neurones was enhanced and of six inhibited.3. The neurones exhibiting excitation on baroreceptor afferent stimulation were broadly divisible into two groups. Nine neurones responded with a marked increase in discharge frequency, the other six responded with a ;burst-on' discharge with a minimum latency of 90 msec. Excitation in both cases was always within 500 msec of the presentation of the stimulus. No functional distinction has been drawn from these responses.4. Evidence has been obtained suggesting that the responses of both types of neurone showing excitation may be potentiated using a pulsatile rather than static stimulus.5. The role of the neurones in the integration of the carotid sinus baroreceptor reflex is discussed. It is suggested that the hypothalamic depressor area represents the rostral extension of the integrative centre for the reflex.     

Reflux mechanisms involved in cardiac arrhythmias induced by hypothalamic stimulation

Electrical stimulation of widespread areas in the CNS has been shown to cause cardiac arrhythmias, which occur most frequently after cessation of stimulation. To determine the reflex and autonomic mechanism responsible for the poststimulation arrhythmias, we anesthetized cats with chloralose, and recorded arterial pressure, ECG, and cardiac vagal nerve activity. Stimulation of the hypothalamus consistently caused increases in blood pressure and heart rate during stimulation and caused arrhythmias, accompanied by vagal hyperactivity, immediately following stimulation. The arrhythmias were mediated solely by the vagus nerves because vagotomy or propantheline administration prevented them, whereas propranolol did not. Administration of either phentolamine or spinal cord transection prevented both the rise in blood pressure during stimulation and the poststimulation arrhythmias, but sectioning the carotid sinus and aortic depressor nerves had no preventative effect. However, when this denervation was combined with sectioning of vagal afferents, bursts of vagal activity (used as an index of cardiac rhythm disturbances) were prevented in three of six animals. Subsequent administration of phentolamine prevented the bursts in the remaining animals. It is concluded that poststimulation arrhythmias are elicited by the rise in blood pressure occurring during stimulation causing a sudden surge in parasympathetic outflow to the heart. The reflexogenic areas involved appear to be stretch receptors innervated by afferent vagal fibers.     

Baroreflex control of the circulation in patients with congestive heart failure

Vagal and glossopharyngeal afferents from cardiopulmonary and arterial baroreceptors exert supraspinal tonic restraint on sympathetic efferent outflow. The baroreceptor inhibitory influence is directly related to physiological changes in cardiac filling and arterial pressures. Increased cardiac pressures and dimensions during CHF may provide chronic stimulation that reduces responsiveness of these receptors and thereby influence the neurohumoral control of the circulation. Patients with chronic and severe CHF of ischemic cause were compared with control subjects whose ischemic heart disease did not affect cardiac performance. Orthostatic pooling of blood with use of upright tilt (45 degree), provided an apparently sufficient stimulus to unload baroreceptors in patients like controls. In contrast to peripheral vasoconstriction in controls, the patients dilated their resistance vessels during upright tilt. This abnormal vasodilation was systemic and uniform in skeletal muscle and subcutaneous tissue of the forearm remaining at heart level. Such an inability to vasoconstrict in the patients, could not be attributed to depression of local vasoconstrictor reflex or autoregulatory responsiveness of forearm vascular beds. Neural blockade carried out separately or in combination with blockades of forearm vascular effector receptors revealed; increased neural efferent activity to the forearm during tilting the patients which mediated beta-adrenergic vasodilation in both vascular beds. The patients had augmented circulating catecholamine levels, those for epinephrine increased in venous effluents but were maintained in brachial arterial inflow, and those for norepinephrine increased in arterial rather than venous plasma in the forearm. Following the patients during a course of therapy with a selective vasodilator calcium antagonist, the beta-adrenergic reflex vasodilation became substantially attenuated but was preserved during a placebo course of therapy. The beta-adrenergic reflex effect evidenced in the studied patients is most probably a manifestation of reduced baroreceptor afferent restraint and it could subsequently relate to the severity of depression of baroreceptor sensitivity during the course of CHF.     

Carotid-sinus baroreflex modulation of core and skin temperatures in rats: an open-loop approach

The neural mechanisms of the thermoregulatory control of core and skin temperatures in response to heat and cold stresses have been well clarified. However, it has been unclear whether baroreceptor reflexes are involved in the control of core and skin temperatures. To investigate how the arterial baroreceptor reflex modulates the body temperatures, we examined the effect of pressure changes of carotid sinus baroreceptors on core and skin temperatures in halothane-anesthetized rats. To open the baroreflex loop and control arterial baroreceptor pressure (BRP), we cut vagal and aortic depressor nerves and isolated carotid sinuses. We sequentially altered BRP in 20-mmHg increments from 60 to 180 mmHg and then in 20-mmHg decrements from 180 to 60 mmHg while measuring systemic arterial pressure (SAP), heart rate (HR), and core blood temperature (T(core)) at the aortic arch and skin temperature (T(skin)) at the tail. In response to the incremental change in BRP by 120 mmHg, SAP, HR, and T(core) fell by 90.3 +/- 5.1 mmHg, 60.3 +/- 10.5 beats min(-1), and 0.18 +/- 0.01 degrees C, respectively. T(skin) rose by 0.84 +/- 0.10 degrees C. The maximum rate of change per unit BRP change was -2.1 +/- 0.2 for SAP, -1.5 +/- 0.4 beats min(-1) mmHg(-1) for HR, -0.003 +/- 0.001 degrees C mmHg(-1) for T(core), and 0.011 +/- 0.002 degrees C mmHg(-1) for T(skin). After the administration of hexamethonium or bretylium, these baroreflexogenic responses were completely abolished. We concluded that T(core) and T(skin) are modulated by the arterial baroreceptor reflex.     

Afferent discharges from arterial mechanoreceptors in the kidney of the rabbit

1. Afferent discharges were observed in dissected filaments or single nerve fibres of renal nerves in the rabbit.2. Increasing the arterial perfusion pressure in excised kidney preparations caused an increase in afferent discharge rate. Increasing the venous pressure was without effect. Receptor discharge patterns were compatible with those of the slowly adapting type.3. In vivo, increasing the systemic blood pressure by means of the sinus reflex caused an increase in afferent discharge rate. Afferent discharge patterns showed no synchronicity with heart beat, irrespective of the blood pressure levels.4. It is suggested that mechanoreceptors are present around the arterial walls in the kidney and that they send information about blood pressure levels in the renal artery to the central nervous system.     

Sympathetic nervous response to induced fall and rise of arterial blood pressure in anesthetized rabbits

Summary The baroreceptor reflex is thought to counteract both a rise and a fall in arterial blood pressure, but the evidence for this is incomplete. In 19 anesthetized rabbits, the relationship between blood pressure/aortic nerve activity and efferent sympathetic activity in renal and splanchnic nerves was therefore investigated during induced rise and fall of arterial pressure. Baroreceptor activity increased and sympathetic activity decreased with raised pressure. In contrast, reduction of baroreceptor activity during acute hemorrhagic hypotension had little or no effect on activity in the two sympathetic nerves. Following complete barodenervation, however, sympathetic activity rose on average to 238% of control values. We therefore conclude that the baroreceptor reflex mainly counteracts a rise in arterial blood pressure. Reduction of baroreceptor activity in the first few minutes of acute hemorrhagic hypotension does not lead to increased sympathetic activity. The small activity remaining in baroreceptor nerves at low pressures had a very potent sympathetic inhibitory effect.     

Depressed responsiveness of the carotid sinus reflex in conscious newborn animals.

The responsiveness of the carotid sinus reflex was evaluated by comparing the effects of bilateral carotid occlusion (BCO) in conscious adult dogs and puppies on measurements of arterial pressure, cardiac output, heart rate, and calculations of total peripheral resistance (TPR). In eight adult dogs, BCO increased mean arterial pressure by 57 +/- 6%, TPR by 48 +/- 5%, and heart rate by 45 +/- 15%. In puppies, BCO induced smaller increases (P less than 0.05) in mean arterial pressure (30 +/- 5%) and TPR (29 +/- 4%), while heart rate did not change. After elimination of opposing vagal and aortic baroreceptor reflexes, the differences in responses to BCO of mean arterial pressure and TPR between adults and newborns were even greater. Thus, the carotid baroreceptor reflex appears to be depressed in the newborn when compared with the fully developed reflex in the normal, conscious adult.     

The effects of cyclopropane, halothane and ether on central baroreceptor pathways

1. The reductions in arterial pressure and preganglionic sympathetic activity evoked by aortic nerve stimulation in the rabbit were studied before and during administration of constant inspired concentrations of the inhalation anaesthetics cyclopropane, halothane, and ether. The background anaesthetic was pentobarbitone, gallamine triethiodide was given, and pulmonary ventilation was with 100% oxygen.2. During light pentobarbitone anaesthesia, aortic nerve stimulation usually induced similar reductions in arterial pressure and preganglionic discharge, expressed as the maximum percentage reduction from prestimulation levels. There were two components in the sympathetic responses, attributable to A and C fibre excitation in the aortic nerve, which was also shown to contain a third fibre group with properties similar to those of B fibres.3. The arterial pressure, heart rate, and preganglionic sympathetic responses to aortic nerve stimulation were rapidly and profoundly inhibited by 50% cyclopropane, which also produced arterial hypertension.4. Halothane (3%) significantly inhibited the depressor responses, but even in the presence of severe hypotension the arterial pressure could usually be further reduced by aortic nerve stimulation. The inhibitory effects of 2% halothane were slow in onset and not pronounced. In the concentrations used, these actions of halothane were significantly less than those of cyclopropane.5. The inhibitory effects of ether on the depressor responses were roughly intermediate between those of cyclopropane and halothane; complete suppression of the responses occurred with high ether concentrations, which were also liable to cause circulatory collapse.6. It is concluded that the three anaesthetics significantly inhibit impulse transmission through central baroreceptor pathways; the implications of the findings are discussed in relation to the different circulatory actions of these anaesthetics.     

Pressure-Independent Inhibition of Sympathetic Activity by Noradrenaline: Role of Baroreceptor C Fibres*

The effect of i.v. infusion of noradrenaline on activity in the renal sympathetic nerve was studied in rabbits anesthetized with chloralose and urethane. Noradrenaline (3–8 μg/kg·min) initially increased mean arterial pressure 20–40 mmHg and consequently reduced renal nerve activity. However, studies over a wide range of pressures—obtained by changing the blood volume, revealed that noradrenaline after a few minutes had induced a pressure-independent reduction of sympathetic discharge. The effect disappeared with baroreceptor denervation. An unchanged relationship between arterial pressure and integrated activity in the whole left aortic nerve (which is largely a measure of activity in A fibres) suggested that the sympathetic depression was due to excitation of aortic nerve C fibres. This conclusion was supported by studies of sympathetic responses to selective stimulation of aortic nerve A and C fibres at equal pressures before and during infusion of noradrenaline. Compared to the reflex activity from A fibres, C fibre stimulation was invariably less effective in suppressing renal nerve activity during the infusion. Our studies indicate that noradrenaline may effect a negative feedback control of sympathetic discharge through activation of baroreceptor C fibres.     

Abnormal baroreflex control in renal hypertension is due to abnormal baroreceptors

We determined if baroreflex control (BC) of lumbar sympathetic nerve activity (LSNA) is preserved despite impaired control of heart rate (HR) in rabbits with 6 wk of renal hypertension (HT). Baroreflex responses were determined during transient or steady-state increases (phenylephrine, PE) or decreases (nitroglycerin or caval occlusion) in arterial pressure. Impaired BC of HR was confirmed in conscious and anesthetized HT rabbits with all baroreflexes intact. In contrast, BC of LSNA was preserved in anesthetized HT rabbits. We further determined whether this selective impairment of BC of HR but not of LSNA could be due to an abnormality in the central nervous system (CNS) or in the afferent limb of the baroreflex. With only the left aortic depressor nerve (ADN) intact (other arterial baroreceptor afferents cut), BC of both HR and LSNA in HT was significantly impaired during infusion of PE. However, responses of HR and LSNA to afferent electrical stimulation of the left ADN (all arterial baroreceptor afferents cut) were similar in HT and normotensive controls. We conclude that 1) BC of LSNA is preserved in renal HT even though control of HR is impaired; 2) selective impairment of BC of HR in HT results from an abnormality in the afferent limb of baroreflex and not in CNS; 3) this abnormality in the afferent limb is not sufficient to impair BC of LSNA when all baroreflexes are intact but is sufficient after partial arterial baroreceptor denervation.     

Enhanced sympathetic activity and cardiac sympathetic afferent reflex in rats with heart failure induced by adriamycin

Our previous studies have shown that the cardiac sympathetic afferent reflex is enhanced in rats with chronic heart failure(CHF) induced by coronary artery ligation and contributes to the over-excitation of sympathetic activity.We sought to determine whether sympathetic activity and cardiac sympathetic afferent reflex were enhanced in adriamycin-induced CHF and whether angiotensin II(Ang II) in the paraventricular nucleus(PVN) was involved in enhancing sympathetic activity and cardiac sympathetic afferent reflex.Heart failure was induced by intraperitoneal injection of adriamycin for six times during 2 weeks(15 mg/kg).Six weeks after the first injection,the rats underwent anesthesia with urethane and α-chloralose.After vagotomy and baroreceptor denervation,cardiac sympathetic afferent reflex was evaluated by renal sympathetic nerve activity and mean arterial pressure(MAP) response to epicardial application of capsaicin(1.0 nmol).The response of MAP to ganglionic blockade with hexamethonium in conscious rats was performed to evaluate sympathetic activity.The renal sympathetic nerve activity and cardiac sympathetic afferent reflex were enhanced in adriamycin rats and the maximum depressor response of MAP induced by hexamethonium was significantly greater in adriamycin rats than that in control rats.Bilateral PVN microinjection of angiotensin II(Ang II) caused larger responses of the cardiac sympathetic afferent reflex,baseline renal sympathetic nerve activity and MAP in adriamycin rats than control rats.These results indicated that both sympathetic activity and cardiac sympathetic afferent reflex were enhanced and Ang II in the PVN was involved in the enhanced sympathetic activity and cardiac sympathetic afferent reflex in rats with adriamycin-induced heart failure.