REFLEX VASCULAR RESPONSES TO CHANGES IN LEFT VENTRICULAR PRESSURES, HEART RATE AND INOTROPIC STATE IN DOGS

Transcription

1 Experimental Physiology (1992), 77, Printed in Great Britain REFLEX VASCULAR RESPONSES TO CHANGES IN LEFT VENTRICULAR PRESSURES, HEART RATE AND INOTROPIC STATE IN DOGS J. K. A. AL-TIMMAN AND R. HAINSWORTH School of Medicine, University of Leeds, Leeds LS2 9JT (MANUSCRIPT RECEIVED 7 OCTOBER 1991, ACCEPTED 9 DECEMBER 1991) SUMMARY Dogs were anaesthetized with chloralose, artificially ventilated and the chests widely opened. Left ventricular mechanoreceptors, including those in or near the coronary arteries, were stimulated by changing the pressure in the aortic root. The pressures distending the left atrium and the aortic and carotid baroreceptors were controlled. Reflex vascular responses were assessed from changes in perfusion pressures to a hind limb and to the rest of the systemic circulation, which were perfused independently at constant flows. Physiological increases in peak left ventricular and coronary arterial pressures resulted in vasodilatation in both regions. These responses were not influenced by changes in the heart rate. Stimulation of the left cardiac sympathetic nerves resulted in increases in peak ventricular pressure and in the maximal rate of change of pressure (dp/dtnax). This also resulted in increases in perfusion pressures (vasoconstriction) at all levels of peak ventricular pressure although there was little effect on the responses to changes in ventricular pressure. Sympathetic stimulation had little effect on the relationship between perfusion pressures and aortic root pressure. Increases in ventricular filling also resulted in vasoconstriction at all levels of peak ventricular pressure. Increases in filling, however, did not affect the relationship between either perfusion pressure and aortic root pressure. Conversely, decreases in left ventricular filling, by bypassing some of the left atrial blood, resulted in vasodilatation at all levels of peak ventricular pressures but had no effect on the perfusion pressures at any aortic root pressure. The combination of sympathetic stimulation with decreased ventricular filling resulted in little effect on perfusion pressures or on their responses to changes in either aortic root or ventricular systolic pressures. We conclude that the vascular responses to stimulation of left ventricular mechanoreceptors are not enhanced by sympathetic stimulation, decreases in ventricular filling or the combination of the two. The apparent effects of each of these interventions alone on the relationships between perfusion pressures and ventricular, but not aortic root, pressure, could be explained if the receptors responsible were sensitive more to changes in aortic root and coronary arterial pressures than to pressure changes in the ventricle itself. INTRODUCTION We have previously shown that stimulation of left ventricular mechanoreceptors results in reflex vasodilatation (Challenger, McGregor & Hainsworth, 1987; Tutt, McGregor & Hainsworth, 1988 a, Vukasovic, Tutt, Crisp & Hainsworth, 1989). The technique used was to induce changes in pressure in the aortic root, which affects both ventricular systolic and coronary arterial pressures but not the pressures distending the left atrium or the carotid or aortic baroreceptors. It has been suggested that ventricular receptors are particularly responsive to increases in cardiac inotropic state (see Hainsworth, 1991) although when we induced moderate changes in inotropic state by administration of the inotropic agent, dobutamine, or the,-adrenoceptor antagonist, propranolol, we did not observe any effects on the reflex responses to ventricular pressure changes (Tutt, Al-Timman & Hainsworth,

2 456 J. K. A. AL-TIMMAN AND R. HAINSWORTH 1988 b). Another factor which may enhance the activity of left ventricular receptors is the degree of filling of the ventricle, and it has been claimed that the combination of reduced ventricular filling, together with an enhanced activity in the efferent cardiac sympathetic nerves, provides a particularly potent stimulus to ventricular receptors (Oberg & Thoren, 1972). Indeed, this mechanism has become widely accepted as the likely means whereby vasovagal syncope is initiated in humans following reductions in cardiac filling (e.g. Abboud, 1989). We considered that the evidence that the stimulus to ventricular receptors during states of low cardiac filling and high sympathetic activity was increased sufficiently to cause intense vasodilatation, was not conclusive. In particular, no reflex studies had been carried out to test this hypothesis. Furthermore, even in the one study which showed an increase in activity in some ventricular afferent nerves in response to decreased ventricular filling (Oberg & Thoren, 1972), this effect was seen only in a minority of the fibres studied. Studies of reflex responses to increases in cardiac inotropic state (Fox, Gerasch & Leonard, 1977; Emery, Estrin, Wahler & Fox, 1983) are difficult to interpret due to concomitant increases in ventricular peak pressures. We therefore considered it to be important to investigate the independent effects of changes in heart rate, inotropic state, and ventricular filling on the reflex responses to stimulation of left ventricular mechanoreceptors. METHODS All experiments were conducted on dogs, kg, anaesthetized with chloralose, which was infused through a catheter passed under local anaesthesia (2 % amethocaine) through a saphenous vein into the inferior vena cava. The initial dose of chloralose was 100 mg/kg, dissolved in polyethylene glycol (100 mg/ml). Anaesthesia was subsequently maintained by further doses of chloralose (about 10 mg/kg every 30 min, 10 mg/ml in saline). The trachea was cannulated through a mid-line cervical incision and the lungs were ventilated at 18 strokes/min and initially 17 ml/kg, using a Starling Ideal pump. When the chest was opened an end-expiratory resistance equivalent to 0 3 kpa was applied. Arterial blood partial pressure of 02 and CO2 (PO, Pcj ), and ph were determined frequently during the experiments and P0 was maintained over 212 kila by addition of oxygen, as necessary, to the inspired gas. P,0 and ph were maintained at kpa and units, respectively, by adjustments of the stroke of the respiratory pump and infusion as required of molar sodium bicarbonate solution. An incision was made in the first right intercostal space. The stellate ganglion, containing the cardiac sympathetic nerves, was crushed at the neck of the first rib and the chest incision then closed. The left side of the chest was then widely opened by removing the third to sixth ribs. The descending aorta was mobilized by tying and dividing the upper six pairs of intercostal arteries. The left subclavian artery was dissected free and the left cardiac sympathetic nerves (ansae subclaviae) were mobilized and a thread placed round them for subsequent crushing and stimulation. The pericardium was opened and a cord carefully placed between the ascending aorta and the pulmonary trunk, cm from the aortic root. In some experiments silver pacing electrodes were sewn on to the right atrial appendage. The left hindlimb was prepared for perfusion by placing nylon cords at its proximal end, round the main muscle groups but avoiding the sciatic nerve and the femoral nerve and vessels (Hainsworth, Karim & Stoker, 1975). The animal was then given heparin (500 IU/kg I.v.) and the perfusion circuit (Fig. 1). which was partly filled with a mixture of equal parts of mammalian Ringer solution and dextran in dextrose solution, was connected to the animal. Cannulae leading from a small constant pressure chamber were inserted into both ends of the cut left subclavian artery. This linked to a temporary bypass cannula inserted into the central end of a femoral artery to provide for some flow to the lower part of the animal while the aorta was being cannulated. Next, a 7 mm i.d. polyethylene cannula was inserted into the left atrium through its appendage. This was accompanied by a catheter for recording left atrial pressure. This cannula led

3 LEFT VENTRICULAR REFLEX CP Aortic res 457 LA res and LV LA res SG ~ ~ ~ ~ ~ Lc aortic res SG >~~~~~~~~~S HE0;t LF aorta just distal to the aortic valves. Blood flows from this cannula to a pressurized reservoir. Blood from this reservoir was pumped at a constant pressure to the cephalic circulation, including aortic and carotid baroreceptors, through two cannulae inserted through both ends of the cut left subclavian artery; the aortic end ofthe left subclavian artery perfused a pouch ofaorta outside the cannula and then into the brachiocephalic artery. The vascularly isolated hindlimb was perfused at a constant flow via the femoral artery. The rest of the caudal circulation was perfused at a constant flow via the descending aorta. A cannula inserted in the left atrium was connected to an open reservoir which allowed the control of left atrial and left ventricular end-diastolic pressures. In experiments in which effects of increased left ventricular filling were examined, a cannula was inserted through the left ventricular apex which allowed the increase of left ventricular filling by transferring blood from the aortic reservoir into the left ventricle. Abbreviations: CP, constant pressure; Aortic res, aortic reservoir; LScA, left subclavian artery; BCA, brachiocephalic artery; LFA, left femoral artery; LA res, left atrial reservoir; HE, heat exchanger; P, Pump; D, damping chamber; SG, strain gauge transducer; S, nylon snare.

4 458 J. K. A. AL-TIMMAN AND R. HAINSWORTH to an open reservoir, and a reversible pump (603UR, Watson-Marlow Ltd, Falmouth) maintained a constant level in the reservoir by transferring blood to or from the aortic reservoir. A stainless-steel cannula with a diameter and radius of curvature similar to the aortic arch was passed into the arch through the descending aorta. This led to a reservoir from which blood was pumped at constant flow into the left femoral artery (Watson-Marlow MHRE) and into the descending aorta (Watson-Marlow 603UR). Blood was also pumped (Watson-Marlow 603UR) into the reservoir connected to the left subclavian artery at a rate controlled electronically to maintain a constant level in that reservoir. In some experiments (as shown in Fig. 1) a 7 mm i.d. polyethylene cannula was inserted into the cavity of the left ventricle through a stab incision in the apex and secured by a purse-string suture. This cannula connected via a damping chamber and a reversible pump (type 603UR) to the aortic reservoir. Blood pressures were recorded, using nylon catheters attached to strain gauges (P23Gb, Gould-Statham, Oxnard, CA, USA), from the right brachial artery (cephalic perfusion pressure), the aortic cannula (aortic root pressure), the left atrium, the right femoral artery (systemic arterial perfusion pressure), the left femoral arterial perfusion cannula, and the cavity of the left ventricle. Pressures were recorded on a Gould electrostatic recorder (ES1000). Temperatures of the animal and the perfusing blood were recorded by thermister probes (Yellow Springs Instruments, OH, USA) in the oesophagus and in the perfusion blood. These were maintained at C by heat exchangers in the circuit (see Fig. 1) and by heaters under the animal table. After connection of the circuit, the cords round the ascending aorta and the limb were tightened and the rates of the pumps and the pressures applied to the reservoirs were set to adjust the various arterial pressures to about 12 kpa. During each of the experimental procedures described below, left atrial and cephalic perfusion pressures, and blood flows to the hindlimb and the remainder of the systemic circulation were held constant. Experimental procedures Changes in aortic root and ventricular systolic pressures at different heart rates. In all experiments, we changed the pressure in the aortic reservoir in steps to give changes in mean aortic root pressure of about 4 kpa between 8 and 20 kpa. Two series of experiments were carried out. In one, heart rate was initially controlled by pacing the atrium at a little over the spontaneous rate (Grass stimulator, model S4), and was subsequently increased by about 60 beats/min. It was then decreased to its previous level. At each heart rate, aortic root and ventricular systolic pressures were raised in steps. In the other series, the peripheral end of the crushed right cervical vagosympathetic trunk was stimulated to decrease heart rate to about 80 beats/min. The heart was then paced slightly faster than this, then at about 50 beats/min faster, and again at the slower rate. Tests of changes in ventricular pressures were carried out at each rate. The complete sequence was carried out twice in each dog. Changes in aortic root and ventricular systolic pressures during stimulation of efferent cardiac sympathetic nerves. In these experiments, responses were first obtained to graded changes in aortic root and ventricular pressures as described above. Then the cardiac end of the left ansa subclavia (distal end crushed) was stimulated at 10 V, 2 ms, 2-4 Hz. Ten to fifteen minutes after the onset of stimulation, further series of pressures steps were imposed. Finally, the stimulator was switched off and, after a further min, another test of pressure steps was carried out. This sequence was undertaken twice in each dog. Changes in aortic root and ventricular systolic pressures during increases in ventricular.filling. In these experiments, a cannula was inserted directly through the apex of the ventricle (Fig. 1), to allow ventricular filling to be increased without increases in atrial pressure. The level of the atrial reservoir was initially set so that, at low ventricular peak pressure and without blood being pumped directly into the ventricle, there was no net flow into or out of the reservoir. The reservoir was subsequently adjusted as required to control atrial pressure. Tests of graded increases in aortic root and ventricular peak pressures were undertaken before, during and after increasing ventricular filling through the apical cannula. This was carried out twice in each dog. Changes in aortic root and ventricular systolic pressures during decreases in ventricular filling, alone or during stimulation of efferent sympathetic nerves. The left cardiac sympathetic nerves were crushed and a length on the cardiac side of the crushed region was placed on silver wire electrodes for stimulation at 10 V, 2 ms, 2-4 Hz. Ventricular filling was held either at a level which caused no net

5 LEFT VENTRICULAR REFLEX 459 LVP LO V~1 BCPJ\jO\J - LAP -30 AP J LO -30 LVP SPP -30 SPP LPP LAPI -30 AP LVP -0BCP[ C AP LAP LPP LVP I s Fig. 2. Responses to step changes in aortic root and left ventricular peak pressures. Traces obtained during continuous stimulation of efferent end of cut right vagus nerve. Heart was paced at 90 beats/min (top traces) and 125 beats/min (lower traces). Abbreviations: SPP, systemic perfusion pressure; LPP, hind limb perfusion pressure; LAP, left atrial pressure; AP, aortic root pressure; BCP, brachiocephalic arterial pressure; LVP, left ventricular pressure. All pressures in kpa. Note that similar changes in left ventricular peak and aortic root pressures resulted in decreases in systemic and limb perfusion pressures and that these responses were little different at the different heart rates. flow to the atrial reservoir or was reduced by partially bypassing the ventricle by lowering the reservoir level. Aortic root and ventricular peak pressures were changed in steps (a) with ventricular filling at initial level and no sympathetic stimulation; (b) with ventricular filling decreased, no stimulation; (c) normal filling during sympathetic stimulation and (d) reduced filling during stimulation. Then (c), (b) and (a) were repeated in that order. This sequence was then repeated. RESULTS All values reported were obtained in steady states at least 2 min after a change in aortic root pressure, 5 min after a change in ventricular filling, and 10 min after the onset or cessation of sympathetic stimulation. The applied changes in aortic root pressure resulted in changes in left ventricular peak pressure in steps of about 3 kpa between 16 and 25 kpa. Because

6 460 J. K. A. AL-TIMMAN AND R. HAINSWORTH A B C ~ I '20 ~20- >/ a/ lult 18 ~' _ Left ventricular peak pressure (kpa) Left ventricular peak pressure (kpa) Fig. 3. Responses of systemic perfusion pressure to changes in left ventricular peak pressure at different heart rates. A, heart rate reduced by right vagal stimulation at constant rate and controlled by electrical pacing (O) at beats/min and (-) at beats/min. B, heart rate controlled by pacing above spontaneous rate (U) at beats/min and (C1) at beats/min. Results show means and S.E.M. from five dogs in each series. Heart rate had no significant effect on the relationships between perfusion pressures and left ventricular peak pressures. the pressures actually measured varied in the various tests and experiments, in the analysis of the results we plotted the perfusion pressures against both the mean aortic root pressures and the peak left ventricular pressures. From these plots we estimated by interpolation the values ofperfusion pressures at peak ventricular pressures of 16, 19, 22 and 25 kpa and aortic pressures of 8, 12, 16 and 20 kpa, which were the integral values closest to the actual pressures applied. The interpolated data so obtained were used in all further analyses of the results. Data were calculated from the averages of all tests carried out in each dog and are given as the means of these averages+ one standard error of the mean. Significance was assessed by Student's paired t test. Responses to changes in aortic root and ventricular systolic pressures at diferent heart rates In five dogs, cephalic perfusion pressure was held at kpa, left atrial and ventricular end-diastolic pressures were held at kpa and kpa. When the heart was paced at beats/min, the overall response to an increase in left ventricular peak pressure from 16 to 25 kpa was a decrease in systemic perfusion pressure from to kpa ( %; P < 0 001). An increase in heart rate to beats/min resulted in no significant change in perfusion pressures at any

7 LEFT VENTRICULAR REFLEX o28.-. *T a 26 0~, 24,', I Left ventricular peak pressure (kpa) Fig. 4. Responses of systemic perfusion pressure to step changes in left ventricular peak pressure in the absence of (-) and during stimulation (El) of the efferent left ansa subclavia. Results show means and S.E.M. calculated from averages of all tests in five dogs. Note the parallel shift in the plot during sympathetic stimulation, denoting vasoconstriction. * P < 0 05; ** P < 0-02 compared to corresponding control point. ventricular peak pressure. The overall response ( %) was also not significantly changed (P > 005). In four experiments, increases in ventricular pressure also resulted in decreases in perfusion pressure in the hindlimb and these responses were also not influenced by changes in heart rate. In a further five dogs, in which cephalic, mean atrial and ventricular end-diastolic pressures were controlled at , and kpa, heart rate was reduced by vagal stimulation and subsequently changed by atrial pacing. Figure 2 shows an example of the responses to changes in aortic root and ventricular peak pressures at two controlled heart rates. Overall, an increase in ventricular peak pressure from 16 to 25 kpa, at paced heart rates of and beats/min, decreased systemic perfusion pressure from to kpa and to kpa, respectively. These responses were not significantly different. The values of systemic perfusion pressure at the various left ventricular peak pressures at the different heart rates are plotted in Fig. 3. There was no significant effect of heart rate, in either series of experiments, at any value of ventricular peak pressure. Similar results were seen in the plots of limb perfusion pressure. Furthermore, when the values of the perfusion pressures were plotted against mean aortic root pressure these were also seen to be unaffected by changes in heart rate.

8 462 J. K. A. AL-TIMMAN AND R. HAINSWORTH X27 * _ Aortic root pressure (kpa) Fig. 5. Responses of systemic perfusion pressure plotted against aortic root pressure in the absence of (-) and during (OI) stimulation of the efferent left ansa subclavia. Data from same experiments as in Fig. 4 (n = 5). Note that, unlike plots of data against ventricular pressure, except for data at lowest aortic root pressure, the plots are not significantly different. * P < 0-05 compared with corresponding control value. Effects of stimulation of efferent cardiac sympathetic nerves on responses to changes in aortic root and ventricular peak pressures In five dogs, cephalic perfusion pressure was held at kpa and left atrial and ventricular end-diastolic pressures at and kpa respectively. In the absence of sympathetic stimulation an increase in ventricular peak pressure from 16 to 25 kpa decreased systemic perfusion pressure from to kpa ( %, P < 002). Limb perfusion pressure decreased from P53 to kpa ( %, P < 0-01). Stimulation of the sympathetic nerves resulted in large increases in the maximum rate of change of ventricular pressure (dp/dtmax) as determined using an analog differentiator (Neal, Halpen & Reeves, 1960) at any level of peak ventricular pressure. For example, at 16 kpa, dp/dtmax increased from to kpa/s. During sympathetic stimulation the values of systemic and limb perfusion pressures, at any level of ventricular pressure, were consistently higher than in the absence of stimulation. The responses of the perfusion pressures to changes in ventricular pressure, however, were not changed during sympathetic stimulation; the resulting effect was an upward displacement of the plots relating perfusion pressures to ventricular pressure (Fig. 4). During sympathetic stimulation the difference between left ventricular peak pressure and mean aortic root pressure was consistently increased. When mean aortic root pressure was 8 kpa, ventricular peak pressures, without and during sympathetic stimulation, were

9 LEFT VENTRICULAR REFLEX _ a CL r- Ct *3 20 K 18 _ 16 _ - 1** 1 _* 14 _ 12 _ IC I, L //i Left ventricular peak pressure (kpa) Fig. 6. Responses of limb perfusion pressure to step changes in left ventricular peak pressure during normal ventricular filling (-) and during increased filling (El) achieved by pumping blood directly into the ventricular cavity. Results show means and S.E.M. calculated from the averages of all tests in five dogs. Note the upward shift of the plot during increased ventricular filling denoting vasoconstriction. * P < 0 05; ** P < 0-02; *** P < 0-01 compared to control. j and kpa. At 20 kpa aortic pressure, ventricular pressures were and kpa. The results were, therefore, replotted to show perfusion pressures against mean aortic root pressure (Fig. 5). This analysis shows that, apart from a small although significant (P < 0-05) effect at the lowest mean aortic root pressure, the two lines nearly overlapped. In two dogs, we investigated the effects of cooling both vagosympathetic trunks to 2 C on the vasoconstrictor responses to sympathetic stimulation with ventricular peak pressure held at 16 kpa. In both dogs, the responses were almost abolished. The average changes in systemic perfusion pressures, with the vagi warm and with the vagi cooled were from 27-2 to 32-5 kpa and, from 27-1 to 27-3 kpa respectively. In two dogs, propranolol (0-5 mg/kg, i.v.) was given and this completely prevented the change in dp/dtmax which previously occurred during sympathetic stimulation. This reduced the vasoconstrictor responses to sympathetic stimulation. The average change in systemic perfusion pressures to sympathetic stimulation with ventricular peak pressure at 16 kpa, before propranolol was from 24-6 to 28-8 kpa and, after propranolol, from 25-0 to 26-3 kpa.

10 464 J. K. A. AL-TIMMAN AND R. HAINSWORTH % II Aortic root pressure (kpa) Fig. 7. Responses of limb perfusion pressure plotted against aortic root pressure during normal (-) and increased (l) ventricular filling. Data from same experiments as in Fig. 6 (n = 5). Note that now the plots have closely converged. Effects of increases in ventricular filling on the responses to changes in aortic root and ventricular peak pressures In five dogs, cephalic perfusion pressure and left atrial pressure were controlled at kpa and kpa. Left ventricular end-diastolic pressure was controlled at 034+O 06 and kpa. Increases in ventricular peak pressure from 16 to 25 kpa, at the lower end-diastolic pressure resulted in a decrease in systemic perfusion pressure from to kpa ( %, P < 001) and, at the high end-diastolic pressure, from to kpa ( %, P < 005). During the increased filling the values of perfusion pressure were higher at all levels of ventricular pressure, but the percentage responses to changes in ventricular pressure were not significantly different. Similar responses were also seen in the hindlimb circulation (Fig. 6). When the results were re-plotted to show systemic and limb perfusion pressures against mean aortic root pressure, changes in ventricular filling were seen to have no effect on the relationship (e.g. Fig. 7). Effects of reduction in left ventricular filling alone, and during stimulation of cardiac sympathetic nerves, on responses to changes in aortic root and ventricular peak pressures Experiments were carried out on six dogs in which cephalic perfusion pressure was held at kpa. In the absence of left ventricular bypass, left ventricular end-diastolic pressure was kpa and an increase in ventricular peak pressure from 16 to 25 kpa decreased systemic perfusion pressure by % (P < 0-0 1). Decreasing ventricular filling

11 LEFT VENTRICULAR REFLEX M~~~~~~~~~ I-** Left ventricular peak pressure (kpa) Fig. 8. Responses of limb perfusion pressure to step changes in left ventricular peak pressure during normal filling with (V) and without (-) sympathetic stimulation and during reduced filling with (A) and without (0) sympathetic stimulation. Results show means and S.E.M. calculated from the averages of all tests in six dogs. Note the upward displacement of the plot (vasoconstriction) during sympathetic stimulation and the downward displacement (vasodilatation) during the reduced filling. The combination of the two resulted in perfusion pressures very close to the control values. * P < 0-05; ** P < 0-02, ***P < 0001 compared to control. reduced end-diastolic pressure to and significantly (P < 0 05) reduced the values of perfusion pressures at all levels of peak ventricular pressure. When sympathetic nerves were stimulated during the partial ventricular bypass, the values of systemic perfusion pressure increased to become not significantly different from the values obtained in the control state, with neither ventricular bypass nor sympathetic stimulation. Thus, sympathetic stimulation counteracted the effects resulting from decreasing ventricular filling. These effects of changes in ventricular filling and sympathetic stimulation were also seen in the perfused hindlimb and are plotted in Fig. 8. Reducing ventricular filling resulted in lower perfusion pressures at all levels of ventricular peak pressure and sympathetic stimulation resulted in an upward displacement of the plot. However, the combination of partial ventricular bypass and sympathetic stimulation restored the position of the curve close to its initial position. Analysis of the data relating perfusion pressures to the mean aortic root pressure revealed that partial ventricular bypass either alone or during sympathetic stimulation had no significant effects on the plots. DISCUSSION The technique used in the present investigation has allowed us to apply pressure changes to the left ventricular while controlling pressures to other major reflexogenic regions. This technique has been discussed in detail in previous papers (Crisp, Tutt, McGregor & Hainsworth, 1989; Vukasovic et al. 1989). Briefly, the partial bypass between the left atrium and the aortic reservoir enabled left atrial and left ventricular end-diastolic pressures

12 466 J. K. A. AL-TIMMAN AND R. HAINSWORTH to be held constant during changes in peak left ventricuiar pressure. This is important because it also prevents pressure changes in the pulmonary circulation which otherwise might have induced reflex responses (Marshall & Widdicombe, 1958; Paintal, 1972; Ledsome, 1977). Also the presence of this bypass together with the ventricular apical cannula (in some experiments) allowed left ventricular filling to be changed independently of atrial and pulmonary vascular pressures. However, because the changes in left ventricular peak pressure were effected by applying changes in pressure to the aortic root, the resulting stimulus would have been applied not only to the left ventricle, but also to the coronary circulation and the first cm of the aortic root. The reasons why any contribution from receptors in the first part of the aortic root to the observed reflex response is likely to have been very small have been discussed earlier (Crisp et al. 1989). These reasons can be summarized as follows: the first cm of the ascending aorta contains less than 10 % of vagal afferent fibres projecting from the aortic arch (Coleridge, Coleridge, Dangel, Kidd, Luck & Sleight, 1973). Also the dissection of the aorta from the pulmonary artery and the application of a ligature around the aorta, just distal to the coronary ostea, is likely to have destroyed most of these fibres as well as limiting distension of that part of the aorta. Thus, it is likely that the responses arose from stimulation of afferent nerve fibres in the left ventricle including those in or near the coronary arteries. In experiments of this study, each step increase in aortic root and left ventricular systolic pressure resulted in decreases in perfusion pressures. The decreases in perfusion pressures denote vasodilatation since both the vascularly isolated hindlimb and the remainder of the systemic circulation were perfused at constant flows. In the first part of this study we examined whether the reflex responses to stimulation of ventricular mechanoreceptors were influenced by heart rate. Heart rate was changed by electrically pacing the right atrium and the normal autonomic influences on heart rate were largely prevented by section of the right vagus nerve and both cardiac sympathetic nerves. Over the range of heart rates we studied, there would have been little effect on cardiac output (Berglund, Borst, Duff & Schreiner, 1958, Bristow, Ferguson, Mintz & Rapaport, 1963). The effect of changing the rate would have been predominantly to change the frequency at which the stimulus was applied to the ventricular mechanoreceptors. We showed that the heart rate was without significant effect on the reflex activity from ventricular receptors. In this context, it is of interest to note that not only is there no effect from changes in heart rate on the reflex responses to stimulation 6f ventricular mechanoreceptors but, unlike the chemosensitive cardiac afferents and the arterial baroreceptors, it has been shown that the ventricular mechanoreceptors have no significant effect on heart rate (Tutt et al. 1988a). In the second part of the study, we examined the effects of sympathetic stimulation on the responses of perfusion pressures to increases in ventricular pressure. Tutt et al. (1988 b) showed that moderate changes in cardiac inotropic state, induced by administration of either propranolol or dobutamine, did not significantly change the vascular perfusion pressures at any level of ventricular peak pressure. In the present study, however, in which we induced changes in inotropic state by direct stimulation of efferent cardiac sympathetic nerves, we did see a significant increase in perfusion pressures, denoting vasoconstriction. The difference could be related to the changes in inotropic state in the present series being rather larger. It is also possible that vasoconstriction was not seen in the earlier series because it may have been masked by vasodilatation induced directly by the dobutamine. The vasoconstriction obtained in the present series is clearly a reflex effect because it was

13 LEFT VENTRICULAR REFLEX prevented by cold blockade of the vagus nerves. This also indicates that it is unlikely that the vasoconstriction could have been caused by the release of catecholamines into the systemic circulation from sympathetic nerve terminals. Another finding which excludes the involvement of catecholamine release is the observation that vasoconstriction started immediately the nerve was stimulated and before blood could have passed through the perfusion circuit. The finding that the vasoconstriction in response to stimulation of the cardiac sympathetic nerves was much reduced after administration of propranolol suggests that it was mainly related to the inotropic effect. However, the vasoconstriction at the lowest ventricular pressure was not totally abolished after propranolol, despite the absence of any observed effect on left ventricular dp/dtmax. One possible explanation for this residual change is that it could have resulted from stimulation of the aortic chemoreceptors supplied from the coronary arteries (Coleridge, Coleridge & Howe, 1970) at low perfusion pressures and during reduction of their blood supply as a result of an a-mediated vasoconstriction. The reflex vasoconstriction in response to sympathetic stimulation could only be demonstrated when the aortic root pressure was decreased sufficiently to prevent changes in ventricular peak pressure. It was noted that, when the vascular perfusion pressures were related to aortic root pressure rather than to peak ventricular pressure, except for the lowest aortic pressure studied, neither sympathetic stimulation nor f-receptor blockade had any effect. One hypothesis to explain this is that the receptors responsible for the reflex responses were sensitive more to aortic root (or coronary arterial) pressure than to the peak pressure in the ventricle. We were unable to test this hypothesis because the preparation did not permit independent control of aortic and peak ventricular pressures. In the third part of the study, we determined the effects of increases in ventricular filling on the reflex activity of ventricular receptors. The results showed that when ventricular filling was increased, there were increases in perfusion pressures at all levels of left ventricular peak pressures. This vasopressor effect was similar to that which was seen during sympathetic stimulation and, as suggested in relation to the effects of inotropic changes, it may have been related to the fact that, although the effects of increasing ventricular filling were compared at identical left ventricular peak pressures, the pressures in the aortic root were not the same. It was seen that an increase in left ventricular enddiastolic pressure was associated with increases in ventricular peak pressure at all levels of aortic root pressure. When ventricular peak pressure was adjusted to the same level as that occurring at the lower filling, the aortic root pressure was then lower. This adds further support to the possibility that the resulting vasoconstriction might be a consequence of the lower aortic root or coronary pressures at given ventricular peak pressures. The final and most important part of this study was to test the claim made by others, that the combination of a high level of sympathetic efferent nervous activity with a low level of ventricular filling provided a particularly potent reflexogenic stimulus. The rationale for this was that Oberg & Thoren (1972) had noted that some non-myelinated ventricular afferent nerves in cats became more active when the cats had been bled by about 25 % of estimated blood volume and this increase in activity sometimes preceded the onset of bradycardia. Sleight & Widdicombe (1965) and Muers & Sleight (1972) had also reported that the activity of ventricular mechanoreceptors was enhanced by sympathetic stimulation or by administration of adrenaline either intravenously or directly into a coronary artery. These two observations have led to the widely accepted view that intense stimulation of ventricular receptors as the result of low ventricular volumes and high sympathetic efferent activity is likely to be the cause of vasovagal syncope following haemorrhage or severe 467

14 468 J. K. A. AL-TIMMAN AND R. HAINSWORTH orthostatic stress (Abboud, 1989). Against this view is the fact that in the study of Oberg & Thoren (1972) only about 20 % of the fibres which responded to mechanical stimulation showed any increase during haemorrhage; most of the others decreased their activity. Also there have been other studies which have shown that vasodilatation in response to haemorrhage is not necessarily dependent on afferent vagal activity (see Hainsworth, 1991). In the present study we did show that the reflex responses to changes in ventricular systolic pressure were influenced by ventricular filling. In particular, when the ventricle was contracting when nearly empty, as the result of bypassing blood from the left atrium to the aortic reservoir, there was a pronounced vasodilatation. The effects of sympathetic stimulation, however, were in the 'wrong' direction; during sympathetic stimulation, perfusion pressures were consistently higher at all levels of ventricular pressure. The interpretation of these results is complicated by the change in the relationship between peak ventricular pressure and aortic root pressure. When ventricular ejection was high, during increased ventricular filling or sympathetic nervous stimulation, this gradient was also increased. We noted that when aortic root pressure was considered as the independent variable, instead of peak ventricular pressure, neither changes in ventricular filling nor sympathetic stimulation influenced the position of the plots. The results of the crucial experiments, in which we simultaneously decreased ventricular filling and stimulated sympathetic nerves, fortunately, are quite clear cut. The tendency for sympathetic stimulation to increase the gradient between peak ventricular and mean aortic pressures was almost exactly offset by the tendency of the reduced filling to decrease it. Thus, there were no effects of this combined intervention on the relationship between either the peak ventricular pressure or the mean aortic root pressure and the vascular perfusion pressures. The main conclusion that we would draw from these experiments is that, at least under the controlled conditions of our preparations, there is no evidence that the reflexogenic activity of the ventricular receptors responsible for reflex vasodilatation in response to physiological pressure changes is enhanced by the combination of reduced ventricular filling and increased sympathetic activity. The individual changes brought about by changes in either of these variables alone suggest that the receptors responsible for initiating the reflex dilatation may be excited more by changes in pressure in the coronary arteries than by changes in pressure in the cavity of the ventricle itself. ABBOUD, F. M. (1989). Ventricular syncope: is REFERENCES the heart a sensory organ? New England Journal of Medicine 320, BERGLUND, E., BORST, H. G., DUFF, F. & SCHREINER, G. L. (1958). Effect of heart rate on cardiac work, myocardial oxygen consumption and coronary blood flow in the dog. Acta Physiologica Scandinavica 42, BRISTOW, J. D., FERGUSON, R. E., MINTZ, F. & RAPAPORT, E. (1963). The influence of heart rate on left ventricular volumes in dogs. Journal of Clinical Investigation 4, CHALLENGER, S., MCGREGOR, K. H. & HAINSWORTH, R. (1987). Peripheral vascular responses to changes in left ventricular pressure in anaesthetized dogs. Quarterly Journal of Experimental Physiology 72, COLERIDGE, H. M., COLERIDGE, J. C. G., DANGEL, A., KIDD, C., LUCK, J. C. & SLEIGHT, P. (1973). Impulses in slow conducting vagal fibres from afferent endings in the veins, atria and arteries of dogs and cats. Circulation Research 33, COLERIDGE, H. M., COLERIDGE, J. C. G. & HOWE, A. (1970). Thoracic chemoreceptors in the dog: A histological and electrophysiological study of the location, innervation and blood supply of the aortic bodies. Circulation Research 26,

15 LEFT VENTRICULAR REFLEX 469 CRISP, A. J., TUTT, S. M., McGREGOR, K. H. & HAINSWORTH, R. (1989). The effects of changes in left ventricular pressure on respiratory activity in anaesthetized dogs. Quarterly Journal ofexperimental Physiology 74, EMERY, R. W., ESTRIN, J. A., WAHLER, G. M. & Fox, I. J. (1983). The left ventricular mechanoreceptor reflex: characterisation of the efferent pathway. Cardiovascular Research 17, Fox, I. J., GERASCH, D. A. & LEONARD, J. J. (1977). Left ventricular mechanoreceptors: a haemodynamic study. Journal of Physiology 273, HAINSWORTH, R. (1991). Reflexes from the heart. Physiological Reviews 71, HAINSWORTH, R., KARIM, F. & STOKER, J. B. (1975). The influence of aortic baroreceptors on venous tone in the perfused hindlimb of the dog. Journal of Physiology 244, LEDSOME, J. R. (1977). The reflex role of pulmonary artery baroreceptors. American Review of Respiratory Diseases 115, MARSHALL, R. & WIDDICOMBE, J. (1958). The activity of pulmonary stretch receptors during congestion of the lungs. Quarterly Journal of Experimental Physiology 43, MUERS, M. F. & SLEIGHT, P. (1972). Action potentials from ventricular mechanoreceptors stimulated by occlusion of the coronary sinus in the dog. Journal of Physiology 221, NEAL, T. J., HALPEN, W. & REEVES, T. J. (1960). Velocity and acceleration of pressure changes in heart and circulation. Journal of Applied Physiology 15, OBERG, B. & THOREN, P. (1972). Increased activity in left ventricular receptors during haemorrhage or occlusion of caval veins in the cat. Acta Physiologica Scandinavica 85, PAINTAL, A. S. (1972). Cardiovascular receptors. In Handbook of Sensory Physiology, vol. III, Enteroreceptors, ed. NEIL, E., pp Springer-Verlag, Berlin. SLEIGHT, P. & WIDDICOMBE, J. G. (1965). Action potentials in fibres in the epicardium and myocardium of the dogs left ventricle. Journal of Physiology 181, TuTT, S. M., AL-TIMMAN, J. K. A. & HAINSWORTH, R. (1988b). Reflex responses of vascular resistance in anaesthetized dogs to independent changes in ventricular systolic pressure and cardiac inotropic state. Quarterly Journal of Experimental Physiology 73, TUTT, S. M., MCGREGOR, K. H. & HAINSWORTH, R. (1988a). Reflex vascular responses to changes in left ventricular pressure in anaesthetized dogs. Quarterly Journal ofexperimental Physiology 73, VUKASOVIC, J. L., TuTT, S. M., CRISP, A. J. & HAINSWORTH, R. (1989). The influence of left ventricular pressure on the vascular responses to changes in carotid sinus pressure in anaesthetized dogs. Quarterly Journal of Experimental Physiology 74, EPH 77