Influence of higher brain centres and vasopressin on the haemodynamic response to acute central hypovolaemia in rabbits

Journal of the Autonomic Nertous System, 35 (1991 ) 1- 14 .*'.~ 1991 Elsevier Science Publishers B.V. 0165-1838/91/$03.50 JANS 01175

Influence of higher brain centres and vasopressin on the haemodynamic response to acute central hypovolaemia in rabbits Roger G. Evans 1, John L u d b r o o k ~, Robyn L. Woods : and David Casley 3 I Unicersity of" Melbourne Department q/" Surgeo'. Royal Melbourne Hospital. Parkcilh': 2 Baker Medical Research Institute, Prahran, and Unicersity qf Melbourne Department (~["Medicine, Austin Hospital, Heidelberg. Victoria, Australia (Received 19 December 1990) (Revision received 21 Februa~' 1991 ) (Accepted 4 March 1991)

Key words': A r g i n i n e v a s o p r e s s i n ; B a r o r e f l e x ; F o r e b r a i n ; H e m o r r h a g e ;

Midbrain; Rabbit

Abstract Wc tested whether suprapontine brain centres contribute to the sudden failure of vasoconstriction that occurs in unanacsthetized rabbits during acute reduction in central blood volume, ttaemorrhage was simulated by gradually inflating a culf around the thoracic inferior vena cava so that cardiac output fell by about 8% per min. In intact rabbits, and in rabbits that had undergone craniectomy but not decerebration, the haemodynamic response to simulated haemorrhage was always biphasic. During the first, compensator~ phase, systemic vascular conductance fell almost in proportion to the fall in cardiac output so that arterial pressure fell by only about 10 mmHg. When cardiac output had fallen by about 50c/~, a decompensatory phase supervened in which systemic vascular conductance rose abruptly, arterial pressure fell steeply to less than 40 mmHg, and the plasma arginine vasopressin (AVP) level rose. High mesencephalic decerebration did not affect the compensatory phase, but it abolished the decompensatory phasc and there was no rise in the plasma AVP level. The decompensatory, phase was not restored by intravenous administration of AVP. We came to two conclusions as a result of this study. Suprapontine brain centres do not influence the arterial barorcflex-mediated vasoconstriction that occurs during the first phase of acute central hypovolaemia. However, the sudden failure of vasoconstriction that occurs during the second phase of acute central hypow~laemia, attributable to a signal from the heart and mediated by a a-opioid receptor mechanism in the brainstem, does depend on the integrity of suprapontine brain centres, though not on neurohypophysial AVP release.

Introduction In u n a n a e s t h e t i z e d r a b b i t s , p r o g r e s s i v e r e d u c tion of central blood volume evokes a haemodyn a m i c r e s p o n s e t h a t is b i p h a s i c [11,12,34,41,42]. A t first, as c a r d i a c o u t p u t falls t h e r e is a p a r a l l e l

Correspondence: J. Ludbrook, Cardiovascular Research Laboratory, University of Melbourne Department of Surgery, Royal Melbourne Hospital. Parkville. Victoria 3050, Australia.

fall o f s y s t e m i c v a s c u l a r c o n d u c t a n c e so t h a t b l o o d p r e s s u r e is s u s t a i n e d at a n e a r - n o r m a l level. T h e r e is a l s o a p r o g r e s s i v e rise o f r e n a l s y m p a t h e t i c n e r v e a c t i v i t y {5,37], so t h a t t h i s c o m p e n s a t o r y p h a s e c a n a l s o b e d e s c r i b e d as s y m p a t h o e x c i t a tory. B u t w h e n c a r d i a c o u t p u t h a s f a l l e n by a b o u t 50%, systemic vascular conductance rises abruptly and blood pressure plummets even though there is c o n c o m i t a n t r e l e a s e o f a r g i n i n e v a s o p r e s s i n ( A V P ) i n t o t h e b l o o d s t r e a m {33,38]. R e n a l symp a t h e t i c n e r v e a c t i v i t y d e c l i n e s s t e e p l y [5,37], so

this second, decompensatory phase can also be described as sympathoinhibitory. In rabbits, it appears to be precipitated by a signal from the heart [5,11], and to depend on a 6-opioid receptor mechanism which is probably located in the hindbrain [12]. A similar biphasic response to acute central hypovolaemia has been observed in other laboratory mammals and in conscious human volunteers subjected to venesection or lower body negative pressure (see ref. 41). There is circumstantial evidence to suggest that higher brain centres might also be involved in the decompensatory phase of acute hypovolaemia. In humans, the haemodynamic events of this phase resemble closely those that occur during vasovagal syncope induced by emotional stress [18,19]. In rabbits, we have found that a surgical plane of general anaesthesia [46], including that induced by choralose and urethane (unpublished observations), disguises or abolishes the decompensatory phase. We have tested the effects of high mesencephalic decerebration in rabbits on the haemodynamic response to graded inflation of a cuff on the inferior vena eava so as to simulate haemorrhage. We have also tested whether the decompensatory phase depends on the release of AVP into the bloodstream, since it has been reported that AVP sensitizes cardiac sensory receptors [1] and potentiates cardiac receptor reflexes [4,20].

Materials and Methods

All experiments were conducted according to the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes 1990 [3], and were approved in advance by the Animal Ethics Committee of the Royal Melbourne Hospital.

Preliminary surgical procedures These were carried out with full aseptic precautions under general anaesthesia with 2.5% halothane in O2, after induction with thiopentone sodium (25 mg kg 1, i.v.) (Abbott Australasia Pty. Ltd., Sydney, Australia) and endotracheal intubation. The same anaesthetic regimen

was used for decerebration (see later). Eight New Zealand White rabbits were used, weighing 2.472.84 (mean 2.60) kg at the beginning of the first study. Inflatable cuffs were placed round the thoracic inferior vena cava (caval cuff) and the descending thoracic aorta (aortic cuff), anti an electromagnetic flow probe (Biotronex 5050) was placed on the ascending aorta. The surgical techniques have been described in detail previously [11,12]. The first experiment was carried oul not less than 2 weeks later, when all rabbits were well and gaining weight.

Craniectomy and high mesencephalic decerebration The technique followed was similar to that described by Korner et al. [28]. The operations were carried out with full aseptic precautions. General anaesthesia was induced and maintained as described above, except that a tracheostomy tube was inserted after the induction of anaesthesia. The rabbit's head was placed in a stereotaxic frame. The first step was to perform a bilateral fronto-parietal craniectomy. The bone of the calvarium on either side of the midline was removed subperiosteally to expose the dura mater, using a high-speed drill and bone-nibbling forceps. Bleeding from the bone edges was arrested by applying bone wax. Blood loss was < 1 ml. This step took approximately 30 min. In the second step the dura mater was excised and the occipital cortex was removed by suction through a blunt 19 gauge needle to expose the superior colliculi. The midbrain was then divided with a spatula from just rostral to the superior colliculi to just caudal to the mammillary bodies. All brain tissue rostral to the line of section was removed by suction, including the contents of the pituitary fossa. Bleeding was arrested by packing the base of the skull with gelatin sponge (Gelfoam Size 12, Upjohn, Kalamazoo, MI, U.S.A.), and applying paper tissue to the exposed brain. The dead space within the skull was filled with isotonic saline, and the skin was closed with a running suture. This second step also took approximately 30 rain to perform. Operative blood loss was estimated by measuring the volume of fluid in a suction trap, and subtracting the volume of brain removed. The net blood loss was 13 _+ 2 ml.

It was replaced by i.v. infusion of an equal volume of dextran-saline, consisting of low molecular weight dextran (Rheomacrodex, Pharmacia (South Seas) Pty. Ltd., Melbourne, Australia) diluted in saline so that it was isosmotic with plasma [13]. In the first experiment the first and second steps were carried out in separate operations, 3 - 5 days apart. In the second experiment they were carried out at the same operation.

Preparations for studies These were done under local anaesthesia with 2% procaine HCI. Catheters were inserted into the left central ear artery to measure arterial pressure, and into the right central ear artery for blood sampling. Catheters were inserted into one or both marginal ear veins for i.v. infusions. The lead for the flow probe and the tubes leading to the inflatable cuffs were retrieved from their subcutaneous positions. At the end of each study except the final one the intravascular catheters were removed, and the connecting plug and tubes returned to their subcutaneous positions, under local anaesthesia.

Haemodynamic measurements Ascending aortic flow (cardiac output) was measured by connecting the flow probe to a meter (Biotronex BL-613). The probes had been previously calibrated in vitro with whole blood [13]. H e a r t rate was measured by a tachometer actuated by the flow pulse. Arterial pressure was measured by connecting a strain gauge (Statham P23d) to a catheter inserted into the left central ear artery. The signals were amplified and recorded on a Grass Model 7 Polygraph, and sent to a microcomputer (Olivetti M24) for A-D conversion. Mean values of arterial pressure (MAP, mmHg), heart rate ( M H R , bpm) and cardiac output (MCO, m l / m i n ) were recorded over 2 s or 10 s intervals. Mean cardiac index (MCI, m l / k g / rain), mean systemic vascular conductance index (MSVCI, 102 M C I / M A P ) and mean stroke volume index (MSVI, m l / k g ) were calculated.

Simulated haemorrhage The caval cuff was inflated by a micrometerdriven syringe so that MCI fell at a constant rate

of ~ 8% of its resting level per min [11,12,33]. The caval cuff was deflated when MAP had fallen to 40 m m H g or MCI to 40 + 2% of its resting level, depending on which occurred first.

Baroreceptor-heart rate reflex This was elicited by inflating the caval and aortic cuffs to cause a slow ramp fall or rise of MAP at 1-2 m m H g / s until M H R had reached upper or lower limiting values [31]. The reflex was used to test that the effects of halothane anaesthesia had worn off [46].

Arterial haematocrit and blood gases Haematocrit (Hct) was measured in duplicate on 0.2 ml blood samples (Hawksley). p~,O2 and p~CO 2 were measured on 0.6 ml blood samples (Radiometer A B L 4 acid-base analyser).

Arterial plasma uasopressin concentration Plasma arginine vasopressin (AVP) concentration was measured from 5 ml blood samples. The plasma was immediately separated by centrifugation and stored at - 2 0 ° C for later analysis by radioimmunoassay [47]. The assays for the two experiments (see later) were done in different laboratories, using the same antibody. The packed red cells were diluted with saline and reinfused into the rabbit.

Artificial centilation This was done through the tracheostomy tube, at a tidal volume of 15 ml at 64 cycles/rain (minute volume 960 ml). The rabbit was paralysed with intravenous vecuronium bromide (Norcuron, Organon Teknika Australasia Pty. Ltd., Melbourne, Australia), given as a bolus dose of 100 / ~ g / k g followed by infusion at 5 ~g/kg/min.

Autopsy At the conclusion of the last study the rabbit was killed with i.v. thiopentone sodium (75 m g / k g ) . It was decapitated, and the head immersed in 10% formalin in saline. Two weeks later the brain was removed and inspected. In all rabbits it was confirmed that the brain section had been carried out as described.

Experiment 1: effects of decerebration The aim of this experiment was to study the effects of high mesencephalic decerebration on the haemodynamic response to simulated haemorrhage, while controlling out the potential confounding effects of prior anaesthesia and craniectomy, and of artificial ventilation with neuromuscular blockade. Each of the five rabbits was studied on three occasions, at intervals of 3-5 days. The first study served as a time control The rabbit was prepared for the study as described above, then Hct and blood gases were measured. Three h later Hct and blood gases were measured, the baroreceptor-heart rate reflex was elicited and simulated haemorrhage was performed. These observations were repeated 1 h later. Blood samples for AVP assay were taken immediately before and after the first simulated haemorrhage. The second study served to control out the effects of prior anaesthesia and craniectomy. Hct and blood gases were measured before anaesthesia was induced. Following craniectomy the periosteum and skin were sutured and the rabbit was allowed to recover from the anaesthetic. Ten ml of blood was removed by venesection, heparinized, and stored at 4 °C. It was replaced by i.v. infusion of 10 ml of the dextran-saline mixture. Two h after the operation the rabbit was placed in its study box. Hct and blood gases were again measured, the baroreceptor-heart rate reflex was tested and a simulated haemorrhage was performed. One h later the observations were repeated. Blood samples for AVP assay were taken immediately before and after the first simulated haemorrhage. At the end of the study the stored blood was reinfused. The third study was designed to test the effects of decerebration, while controlling out any effects of artificial ventilation and neuromuscular blockade. Hct and blood gases were measured before anaesthesia was induced. High mesencephalic decerebration was performed through the exposure provided by the previous craniectomy. The rabbit was allowed to recover from the anaesthetic while supported by plastic sponge in a squatting position and insulated from auditory stimuli. Two h later Hct and blood gases were again measured,

the baroreceptor-heart rate reflex was tested, and a simulated haemorrhage was performed. The rabbit was then paralysed and artificaIly ventilated. One h later the observations were repeated. Blood samples for AVP assay were taken immediately before and after the first simulated haemorrhage.

Experiment 2: effects of exogenous A VP The aim of this experiment was to examine the effects of infusing saline or AVP (Sigma Chemical Company, St. Louis, MO, U.S.A.) in intact and decerebrate rabbits on the haemodynamic response to simulated haemorrhage. Each of the three rabbits was studied on two occasions, 7 days apart. The first study was designed to test whether an increase in plasma AVP concentration would alter the response to simulated haemorrhage when the brain was intact. Simulated haemorrhage was performed three times at 90-min intervals, during three different intravenous treatments. The seq u e n c e was saline, high dose A V P (10 n g / k g / m i n ) and low dose AVP (1.0 n g / k g / m i n ) . The i.v. infusions were given at a rate of 100 ~ I / k g / m i n , commencing 10 rain before and continuing throughout simulated haemorrhage. Blood samples for Hct determination were taken before each treatment. Blood samples for plasma AVP assay were taken before commencing each treatment and immediately after each simulated haemorrhage. The second study was designed to test whether exogenous AVP would restore the decompensatory phase of simulated haemorrhage after decerebration. Following decerebration, the rabbit was immediately paralysed and artificially ventilated. After allowing 2 h for recovery from the anaesthetic three simulated haemorrhages were performed at 90-min intervals, under the same treatment conditions as in the first study. Blood samples for plasma AVP assay were taken betore each treatment and immediately after each simulated haemorrhage.

Analysis of reszdts We analysed the data in a similar fashion to that described previously [11,12]. Two- or three-

way analyses of variance t A N O V A ) were used to evaluate differences within and between studies in the levels of the haemodynamic variables, Hct and blood gases. If multiple contrasts were made within the A N O V A the critical value of F was adjusted by the Dunn-Sidfi.k procedure in order to control experimentwise Type I error [35]. The regression analysis confirmed that MCI fell linearly with time during the 6 simulated haemorrhages performed in each rabbit ( r = 0.979 _+ 0.004). As we have found previously [11], there was a near-linear relationship of MAP, M H R and MSVCI to MCI during simulated haemorrhage until the end of the compensatory phase. Family regression analysis was used to compare the slopes of the regressions of MSVCI and M H R on MCI during this phase within and between studies and to calculate the common regression equation for each experiment [14]. We use a graphical method for analysing the effects of treatments on the responses of MSVCI, MAP and M H R to simulated haemorrhage [11,12]. The relationships of MAP, M H R and MSVCI to MCI as a percentage of baseline level were characterized by three sets of co-ordinates: (a) Post-treatment levels, immediately before the onset of simulated haemorrhage. (b) The point at which MSVCI reached a minimum at the end of the compensatory phase before rising abruptly. (c) The final observation before the caval cuff was deflated. These co-ordinates were averaged between rabbits, and were used to distinguish the effects of the different treatments. Unless otherwise indicated, the levels of the

haemodynamic variables are expressed as between rabbit means +_ 1 SEM. Plasma AVP concentrations are expressed as the geometric mean followed by the range in parentheses. The sample sizes were too small to permit further statistical analysis.

Results

All studies were completed in all 8 rabbits. The only difficulties encountered were that after decerebration the rabbits occasionally made spontaneous running movements, and that after they were paralysed and ventilated they exhibited occasional episodes of transient hypertension and bradycardia.

Experiment h effects of decerebration Baseline le~'els of Hct, blood gases and haemodynamic l'ariables Hct remained constant within studies, but it fell from 37% to 29% over the three studies (Table I). There was no significant variation of p~,O2 within or between studies (FH~ = 0.8, F2,1~ =1.1; P > 0 . 5 9 ) . The grand mean was 85_+1 mmHg. There was no significant variation of p~,CO, between studies (F2,1s=0.26; P = 0.95), the grand mean being 35 +_ 2 mmHg. However artificial ventilation in the decerebration study caused p a C O 2 t o fall from 39 _+ 4 to 29 _+ 2 m m H g (Fl,ls = 10.1; P = 0.01). The baseline levels of the haemodynamic vari-

TABLE I

Baseline calues o# haemodynamic variables and haematocrit in Experiment 1 according to study Mean values ± SEM for 5 rabbits. MCI, mean cardiac index ( m l / k g / m i n ) : MAP, mean arterial pressure (mmHg); MHR, mean heart rate (bpm); MSVCI, mean systemic vascular conductance index (102 m l / k g / m i n / m m H g ) ; MSV1, mean stroke volume index ( m l / k g ) ; Hot, haematocrit (% packed red cell volume). P from A N O V A at d.f. 2,20 for overall differences among studies. Pairwise contrasts were made between studies at d.f. 1,20. Time control versus post-craniectomy: ** P < 0.01. Post-craniectomy vs. post-decerebration: t p < 0.05, ++ P < 0.01. Time-control vs. post-decerebration: * P < 0.05. Study

MCI

MAP

MttR

MSVCI

MSVI

lict

Timc control Post-craniectomy Post-decerebration P

152 ± 15 165 _+ 16 131 + 9 ** < 0.001

82 + 2 81 ± 2 84 + 6 0.73

226 + 14 251 ± 10 224 + 17 * 0.009

186 ± 18 204 + 19 156 ± 4 ** < (I.001

0.67 + 0.05 0.65 ± 0.06 0.59 + 0.02 * 0.029

37 + 1 31 ± 1 ** 29 + 1 ~ < 0.001

1.1 m l / k g / m i n per min (8.1 _+ 0.1% of its baseline level per min). There were no significant differences in this rate between studies (F2.2, = 2.33; P = 0.12). There was a small but consistent difference in the rate between the first and second steps of the studies (8.3 _+ 0.2 vs. 7.8 _+ ().2) (F,.2o = 9.63; P = 0.006). However, these was no significant interaction between studies and steps (F2,2o = 0.30; P = 0.74), indicating that the difference was a constant one from study to study. Within and between the time control and post-craniectomy studies the patterns of the haemodynamic responses to simulated haemorrhage were indistinguishable (Figs. 1 and 2). At first, MSVCI fell steadily almost in proportion to the fall in MCI, so that MAP fell only slightly. M H R rose steadily. When MCI had fallen to 84 + 7 m l / k g / m i n (54 _+ 7% of its baseline level), MSVCI rose abruptly and M A P fell steeply to 37 + 1 mmHg. M H R either ceased to rise or fell

ables in the time control study were within the usual ranges for normal rabbits in our laboratory (Table I). Overall, there were no significant differences in the baseline levels between steps within studies (F1,,8 ~< 0.67; P >~ 0.42). In particular, there were no changes with time in the first study (Fl.20 ~< 0.27; P >~ 0.99), no evidence of carryover effects of anaesthesia and craniectomy in the second study (F,,20 ~< 0.41;P >/0.99), and no effects of artificial ventilation in the third study (FL2o <~0.74; P >/0.95). However, there were overall differences between studies in the baseline levels of all the haemodynamic variables except MAP (Table I). In particular, MCI, MSVCI and M H R were lower after decerebration than after craniectomy.

Responses to simulated haemorrhage Across all studies, the rate of fall of MCI during simulated h a e m o r r h a g e averaged 12.2_+

260 -

200 E E ._= E

E

130

I00

0

I

I

I

I

I

I

0

150

400

75

200

I

I

I

I

E

o -2

0

I

I

I

{

I

2

4

6

8

10

o -2

0

I

n

I

I

n

2

4

6

8

I0

Fig. 1. Levels of haemodynamic variables in one rabbit before and during simulated haemorrhage (Expt. 1). MCI, mean cardiac index; MAP, mean arterial pressure; MSVCI, mean systemic vascular conductance index; MHR, mean heart rate. Symbols are 10 s means, o , Intact rabbit, first simulated haemorrhage, z~ ; after craniectomy, first simulated haemorrhage; m, after decerebration, first simulated haemorrha ge (spontaneous respiration).

slightly. When the caval cuff was deflated MCI was 67 + 5 m l / k g / m i n (43 + 2% of its baseline level) and MSVI was 0.23 + 0.02 m l / k g . After decerebration, the pattern of the haemodynamic response was entirely different, whether or not the rabbit was artificially ventilated (Figs. 1 and 2). MCI and MSVCI fell steadily throughout the simulated haemorrhage, and M A P fell only slightly. M H R rose monotonically. When the caval cuff was deflated MCI was 51 + 3 m l / k g / m i n (39 ± 2% of its baseline level), MSVI was 0.20 ± 0.02 m l / k g and M A P was 71 _+ 5 mmHg. Neither the rate of fall of MSVCI on MCI nor INTACT

the rate of rise of M H R on MCI during the first phase of simulated haemorrhage differed significantly within or between the three studies (respectively, FI,46 ~< 0.59, F2.46 ~ 0.20; P >/0.45 and /> 0.82) (Fig. 2). The common regression equations for the experiment were MSVCI = 38.86 + 0.98 MCI and M H R = 2 8 0 . 0 - 0.45 MCI.

Plasma A VP concentrations before and after simulated haemorrhage These are displayed in Fig. 3. In intact rabbits, the plasma AVP levels before and after simulated haemorrhage were 7.7 (6.9-8.8) and 16.6 (7.7-76) (;R,,\NIEC'FOM5

0 71

240

I)ECEREBRATION

0 11

0 62

E

__E

-....

E 120 %

0 001

012

5 0

i

i

i

i

i

023

400

i

i

i

i

097

0 63

200

041 0

i

i

i

i

i

0.85

150

010

i

i

i

086

i

094

-r., E

75

0 00l

0 86 0 200

i 150

i 100

i 50

200

i 150 MCI

i 100

i 50

200

i 150

i 100

i 50

(ml/kg/min)

Fig. 2. Haemodynamic responses to simulated haemorrhage in intact, craniectomized and decerebrate rabbits (Expt. 1). Symbols are the means from five rabbits. Vertical bars represent 1 SEM. The symbols El and zx represent successive simulated haemorrhages. In the decerebration study, the simulated haemorrhage denoted by A was performed after the rabbits were paralysed and artificially ventilated. In the top right hand corner of each panel is the P value from A N O V A at d.f. 1,20 for the within-study difference in the level of the haemodynamic variable at the time of cuff deflation. In the bottom right-hand corner of each panel is the P value from A N O V A at d.f. 1,20 for between-studies differences in the level of haemodynamic w~riable at the time of cuff deflation (craniectomy vs. intact, decerebration vs. intact).

INTACT

CRANIE('TOMY

I)ECEREBR,,\TION

2

E

1

•

•

i

q

0

i

J

B

A

|L

B

A

--i

i

I

B

A

Fig. 3. Arterial plasma levels of immunoreactive arginine vasopressin (log 10 AVP) before (B) anti after (A) simulated haemorrhage in five rabbits. Each rabbit was studied intact, after craniectomy and after high mesencephalie decerebration (Expt. 1).

p g / m l . After craniectomy, the corresponding levels were 9.9 (5.1-12.9) and 22.3 (9.2-79.3) p g / m l . In decerebrate rabbits the plasma A V P level was not affected by simulated haemorrhage, the levels before and after being 5.1 (2.7-8.5) and 5.9 (3.58.9) p g / m l .

Baroreceptor-heart rate reflex We assessed the effects of craniectomy, decerebration and decerebration ptus artificial respira-

tion on the range, upper limit and lower limit of M H R in the baroreceptor-heart rate reflex (Table II). There were no significant differences within studies in any of these variables (FI,:~ ~ <~ 0.23; P >~ 0.64). We conclude from this that neither time nor paralysis and artificial ventilation affected baroreflex control of heart rate. However, there were some differences between studies. In particular, after craniectomy compared with the time control the lower limit of M H R

TABLE 1I Characteristics

of the baroreceptor-HR reJTex in Experiment 1 according to study and step within stm,(v

Heart rate (bpm) under resting conditions, and for the range, upper limit and lower limit of the baroreceptor-heart rate reflex. Mean values _+ 1 SEM for 5 rabbits. P from A N O V A at d.f. 2,20 for overall differences among studies. Pairwise contrasts were madc between studies at d.f. 1,20. Time control vs. post-craniectomy: ** P < 0.01. Post-craniectomy vs. post-decerebration: ~ P < 0.05, ~* P < 0.01. Time control vs. post-decerebration: *~ P < 0.01. Study

Resting

Range

Upper limit

Lower limit

Time control Step I Step 2

234 _+ 9 "~ 223 + 10 /

152 _+ 7 136+5

327 +

175 +

310~_

174+ 6 f

Post-craniectomy Step l Step 2

251 + 10} 248 +

109 -+ 7} ** 114+2

317+ ~} 310 +

2(18+ 2 } * * 196 +

Post-decerebration Step 1 Step 2 P

204_+ 9 ~* 219 _+ 23 J 0.0//9

108 +

27'1_+ 5 } +" 282+ 19 ** 0.001

167+ 8 } ~* 174+ 14 0.001

Z

,°3+Z) < 0.001

l::~

rate was elevated and the range was narrower, and decerebration consistently depressed the upper limit of M H R (Table II).

Experiment 2: effect of exogenous A VP Baseline levels of Hct and haemodynamic variables As in the previous experiment, Hct remained constant within each study but was lower after decerebration (Table III). The effects of decerebrat±on on the haemodynamic variables were similar to those described for the first experiment (Table Ill). That is, it caused significant falls in MCI and MSVCI (F~,24 > 35; P > 0.001), but had no significant effect on MAP or M H R (F~.2~ ~< 4.88; P >~ 0.07). Intravenous administration of AVP lowered MCI, MSVCI and M H R in a dose-dependent manner in both intact and decerebrate rabbits (FL,4>~31; P~<0.001) (TabLe III). MAP was raised by AVP treatment in intact (FL24 = 11.38; P = 0.005)but not decerebrate rabbits.

Responses to simulated haemorrhage and effects" of, AVP The rate of fall of MCI during simulated haemorrhage averaged 11.7 _+ 0.7 m l / k g / m i n per rain (8.5 + 0 . 3 % of its resting level per min). There was no significant difference in this rate within or between studies (respectively, F2.1o = 0.63, FLI,, = 3.90; P = 0.55 and 0.08).

In intact rabbits during saline infusion the haemodynamic response to simulated haemorrhage followed the same course as the first experiment (Fig. 5). At first, MSVCI fell steadily almost in proportion to the fall in MCI, so that MAP fell only slightly. M H R rose steadily. When MCI had fallen to 80 ± 14 m l / k g / m i n (53 ± 7% of its baseline level), MSVCI rose abruptly and MAP fell steeply to 39_+ 0 mmHg. M H R either ceased to rise or fell slightly. When the caval cuff was deflated MCI was 66 _+ 9 m l / k g / m i n (44_+ 4% of its baseline level). After decerebration the rabbits were artifically ventilated. During saline infusion the haemodynamic response to simulated haemorrhagc followed the same pattern as observed in the first experiment (Fig. 4). MCI and MSVCI fell steadily throughout the simulated haemorrhage, and MAP fell only slightly. M H R rose monotonically. When the caval cuff was deflated MCI was 43 ± 1 m l / k g / m i n (38 ± 2 ~ of its baseline level), and MAP was 57 ± 7 mmHg. Intravenous infusion of AVP, even though it affected greatly the resting haemodynamic variables in both intact and decerebrate rabbits (see above), had little or no effect on the patterns of haemodynamic response to simulated haemorrhage (Fig. 4). In intact rabbits, the levels of MCI, MAP, M H R and MSVCI at the onset of the decompensatory phase were unaffected (F2. ~ ~< 2.8; P >~ 0.35).

TABLE 11I

Baseline values of haemodynarnic cariables and haematocrit in Experiment 2 according to study and steps within study, beJore and ~{¢?er intracenous m~usion of saline or arginine vasopressin (AVP) Variables as in Table I. Mean values + SEM for 3 rabbits. Low dose AVP: 1.0 n g / k g / m i n . High dose AVP: 10.0 n g / k g / m i n . Study

MCI

MAP

MHR

MSVCI

Hct

Before

After

Before

After

Before

After

Before

After

Beh)re

Brain intac! Saline LowAVP HighAVP

149 + 10 154± 5 147± 4

15(1 + 6 147±6 102+2

85 +_ 2 79+3 88+1

82 + 1 83_+3 101±5

214 ± 15 259± 8 237± 5

232 + 9 229_+ 7 171+10

176 ± 14 196+ 11 167+5

149 ± 7 178± 111 1(12+ 7

37 + 1 37+ I 37+1

Oecerebrate Saline LowAVP HighAVP

113+_ 6 113+ 2 106+ 3

114+6 107+2 79+3

83±4 76_+1 80±3

84_+3 79+1 82+2

235_+ 9 230± 6 228_+12

230± 11) 22{1+ 6 192+ 8

138+ 12 149+ 3 133+ 7

137± 12 135± 2 96± 6

32* 1 32+ I 31+1

10 DECEREBRATION

INTACT

028

0.13

200 E E 100 %

OO01 > i

0

200 ~P

i

i

~

I

014

0.07

400

E

i

B

089 0

I

080

150

E E m. <

i

095

Plasma A VP concentration beJbre and after simulated haemorrhage The results are displayed in Fig. 5. In rabbits with intact brains, the baseline plasma AVP level was 10.5 (6.5-14.9) p g / m l . After treatment uith saline, low-dose AVP (1.0 n g / k g / m i n ) and highdose AVP (10 n g / k g / m i n ) , the plasma AVP levels at the end of simulated haemorrhage were respectively 65 (23-468), 114 (98-131), and 11197 (931-2125) p g / m l . In decerebrate rabbits, the baseline plasma AVP level was 12.6 (7.8-29.0) p g / m l . After treatment with saline, the plasma AVP level did not rise during simulated haemorrhage and was only 9.8 (7.8-11.5) p g / m l when the caval cuff was deflated. After the low and high doses of AVP, the plasma AVP levels at the end of simulated haemorrhage were respectively 181 ( 159-199) :rod 1148 (675-2300) p g / m l .

o\ 75

Discussion O 001 0 200

i

i

i

150

IO 0

50

i 0

200

i 150

i

i

i

I O0

5O

0

MCI (ml/kg/min) Fig. 4. Haemodynamic responses to simulated hacmorrhage in intact and decerebrate rabbits and the effects of intravenous infusion of saline or arginine vasopressin (AVP) (Expt. 2). Symbols are the m e a n s of three rabbits. Vertical bars represent 1 SEM. (>, saline (100 / z l / k g / m i n ) ; a , A V P (1.0 n g / k g / m i n ) ; D, A V P (10 n g / k g / m i n ) . In the top right-hand corner of each panel is the P value from A N O V A at d.f. 2,10 for within-studies differences in the level of the haemodynamic variable at the time of cuff deflation. In the bottom right-hand corner of each panel is the P value from A N O V A at d.f. 1,10 for the between-studies difference in the level of the haemodynamic variable at the time of cuff deflation (decerebration vs. intact).

The rate of fall of MSVCI on MCI during the first phase of stimulated haemorrhage did not differ significantly within or between the studies (respectively, F2,24 = 2.34, F1,24 = 0 . 3 0 ; P = 0.12 and 0.59). The common regression equation for the experiment was MSVCI = 30.26 + 0.93 MCI. This was indistinguishable from that of the first experiment.

These experiments appear to be the first in which the role of suprapontine brain centres in the haemodynamic response to acute central hypovolaemia has been examined in an unanaesthetized mammal. Our results show that the first, compensatory phase of acute hypovolaemia is independent of suprapontine brain centres whereas their integrity is essential to the occurrence of the second, decompensatory phase. We used simulated, rather than actual, haemorrhage in these experiments for two reasons. The first is that in conscious rabbits the haemodynamic and humoral responses to progressive inflation of a caval cuff are indistinguishable from those of actual haemorrhage when the rates of fall of cardiac output are identical [33]. The second reason is that in normal rabbits the haemodynamic responses evoked by simulated haemorrhage are highly reproducible within and between days [11,12,33] (Fig. 2). After craniectomy the lower limit of heart rate in the baroreceptor-heart rate reflex was elevated (Table II), suggesting that there may have been a small carryover effect of halothane anaesthesia [46]. However, this was not enough to alter signif-

to unloading the arterial baroreceptors [32,40], and this is presumably also true in the case of simulated haemorrhage. There have been several studies in anaesthetized cats of the effect of decerebration on the pressor response to acute reduction of carotid sinus pressure. Some have shown the response to be unaffected [8,24] whereas others have found it to be attenuated [25,39]. The importance of confounding variables such as uneven depth of anaesthesia and artifical ventilation have been pointed out [24]. We believe that our experiments were free from these, and that is it reasonable to infer that arterial baroreflex-mediated vasoconstriction is independent of centres higher than the midbrain in unanaesthetized rabbits. Decerebration prevented the occurrence of the second, decompensatory phase of simulated

icantly the baseline levels of the haemodynamic variables or their responses to simulated haemorrhage (Figs. 1 and 2, Table I). After high mesencephalic decerebration the baseline levels of cardiac output and systemic vascular conductance were low (Tables I and IID and the upper limit of heart rate in the baroreceptor-heart reflex was depressed (Table II). Similar effects of infracollicular decerebration have been described in unanaesthetized rabbits [27,28]. However decerebration did not affect the first, vasoconstrictor phase of simulated haemorrhage. That is, the slopes of the relationship between systemic vascular conductance or heart rate and cardiac output were unaltered (Figs. 2 and 4). In conscious rabbits, the progressive systemic vasoconstriction that is the hallmark of the compens a t o ~ phase of haemorrhage is a reflex response

AVP (1.0 ng/kg/min)

SALINE

AVP (10.0 ng/kg/min)

E 9_.

>

~

2

0

0

I

I

I

I

B

A

B

A

SALINE

AVP (1.0 ng/kg/min)

!

I

B

A

AVP (10.0 ng/kg/min)

4 E e~ ¢X.

>

.<

2

e-

I

|

I

|

I

B

A

B

A

I

B

I

A

Fig. 5. Arterial plasma levels of immunoreactive arginine vasopressin (log L(~ AVP) before (B) and after (A) simulated haemorrhagc in 3 rabbits (Expt. 2). Each rabbit was studied intact (top panel), and after dccercbration with artificial ventilation (bottom panel). during treatment with intravenous saline, low-dose AVP and high-dose AVP.

12

haemorrhage. Systemic vascular conductance continued to fall throughout the whole period of simulated haemorrhage, even though cardiac output was reduced to lower levels than in the time control and post-eraniectomy studies (Figs. 1, 2, 4) and stroke volume fell to the same level. Since the decompensatory phase of haemorrhage and simulated haemorrhage in conscious rabbits is triggered by a signal which is conveyed by cardiac afferent nerves [5,11], it is reasonable to infer that this signal travels to suprapontine brain centrcs or that it interacts in the brainstem with tonic descending influences. In the course of performing mesencephalic decerebration we removed the hypothalamus and the contents of the pituitary fossa. AVP was still detectable in the plasma after decerebration (Figs. 3 and 5), possibly because the surgical procedure caused release of large amounts of AVP from the pituitary and this had not been cleared from the plasma by the time the blood samples were taken. However, simulated haemorrhage no longer caused the level to rise (Figs. 3 and 5). This did not affect the first, vasoconstrictor phase of simulated haemorrhage (Figs. 2 and 4), in conformity with reports that AVP does not contribute to this phase of haemorrhage (see ref. 41). Left ventricular mechanoreceptors are sensitized by artifical elevation of the plasma level of AVP [1]. This also potentiates depressor reflexes from left ventricular mechanoreceptors [20] and chemosensitive receptors [4], whether from an action of AVP at the area postrema {2,21,451 or because of the sensitization of cardiac receptors referred to above. In conscious rabbits the plasma concentration of AVP rises steeply during the decompensatory phase of haemorrhage [38] or simulated haemorrhage [33]. It occurred to us that the cardiac receptor reflex-mediated decompensatory, phase of simulated haemorrhage might depend on a concomitant rise in the plasma concentration of AVP. However, AVP infusion after dccerebration, at rates that caused the same or greater elevations in plasma concentration as occurred in the intact rabbit (Fig. 5), did not restore the decompensatory phase of simulated haemorrhage (Fig. 4). It seems likely, therefore, that the decompen-

satory phase depends on higher brain centres which provide an inhibitory signal to the sympathoexcitatory neurons in the rostral ventrolateral medulla or intermediolateral cell column [6]. Electrical or chemical stimulation of a number of suprapontine regions can evoke sympathoinhibitory or vasodepressor responses. Among these are the fronto-occipital cortex (see ref. 26), the cingulate gyrus [30], the anterior [15,16,22], lateral [43] and paraventricular {23,48] regions of the hypothalamus, the septal region [17], and the amygdaloid nucleus [10]. Which, if any, of these regions are involved in the vasodepressor phase of acute hypovolaemia is purely speculative, though there are anatomical and functional interconnections between many of these regions and cardiovascular centres in the brainstem and spinal cord [?,9,29,30,36,44]. Anterior hypothalamic [15], amygdaloid [7,% 10] and septal neurones [36] appear to be involved in vasodepressor responses to affectivc stimuli. In humans, the haemodynamic events of the decompensatory phase of acute central hypovolaemia (see ref. 41) resemble closely those that occur in emotional fainting [18,19]. It is tempting to suggest that the same higher brain centers are involved in the vasovagal syncope associated with unpleasant emotional stimuli as appear to be involved, at any rate in the conscious rabbit, with the sympathoinhibitory response to acute reduction in central blood volume.

Acknowledgements We thank Gianinna Legudi and Louis Gallagher for their technical assistance with these experiments. This work was supported by the National Health and Medical Research Council of Australia.

References 1 Abboud, F.M., Aylward, P.E., Floras, J.S. and Gupta, B.N., Sensitization of aortic and cardiac baroreceptors by arginine vasopressin in mammals, J. Physiol. (Lond.), 377 (1986) 251-265.

13 2 Applegate, R.J., Hasser, E.M. and Bishop, V.S.. Vagal cold block in area postrema-lesioned dogs: interaction of vasopressin and sympathetic nervous system, Am. J. Physiol., 252 (1987) H135-141. 3 Australian Code of Practice for the Care and Use of Animals for Scientific Purposes, Australian Government Publishing Service, Canberra, 1991L 4 Barazanji, M.W. and Cornish, K.G., Vasopressin potentiates ventricular and arterial reflexes in the conscious nonh u m a n primate, Am. J. Physiol., 256 (1989) H1546-1552. 5 Burke, S.L. and Dorward, P.K., Influence of endogenous opiates and cardiac afferents on renal nerve activity during haemorrhage in conscious rabbits, J. Physiol. (Lond.), 402 (1988) 9-27. 6 Calaresu. F.R. and Yardley. C.P., Medullary basal sympathetic tone. Annu. Rev. Physiol., 50 (1988)511-524. 7 Cechetto, D.F. and Calaresu, F.R., Units in the amygdala responding to activation of carotid baro- and chemoreceptors, Am. J. Physiol., 246 (1984) R832-836. 8 Chai, C.Y., Share, N.N. and Wang, S.C.. Central control of sympathetic cardiac augmentation in lower brain stem of the cat, Am. J. Physiol., 205 (1963) 749 753. 9 Cox, G.E., Jordan, D.. Moruzzi, P.. Schwaber, J.S., Spyer. K.M. and Turner. S.A., Amygdaloid influences on brainstem neurones in the rabbit. J. Physiol. Lond.),381 (10861 135 148. 10 Cox, G.E., Jordan, D.. Paton, J.F.R., Spyer, K.M. and Wood. L.M., Cardiowtscular and phrenic nerve responses to stimulation of the amygdala central nucleus in the anacsthetized rabbit, J. Physiol. (Lond.), 389 (1987) 541556. 11 Evans, R.G., Ludbrook, J. and Potocnik, S.J., Intracisternal naloxone and cardiac nm~'e blockade prevent vasodilatation during simulated haemorrhage in awake rabbits, J, Physiol. {Lond.). 4(19 ( 19891 1- 14. 12 Evans, R.G., Ludbrook. J. and Van keeuwen, A.F.. Role of central opiatc receptor sub-types in the circulatory responses of awake rabbits to graded cawd occlusions, J. Physiol. (Lond.), 419 {19891 15-31. 13 Faris, lB.. lannos, J., Jamieson, G.G. and Ludbrook, J., The circulatou effects of acute hypervolemia and hemodilution in conscious rabbits, Circ. Res., 48 {1981) 825-834. 14 Feldman, It.A., Families of lines: random effects in linear regression, J. Appl. Physiol.. 64 (1988) 1721-1732. 15 Folkow, B., Johansson, B. and (15berg, B., A hypothalamic structure with a marked inhibitory effect on tonic sympathetic actixity, Acta Physiol. Scand., 47 (1959) 262-2711. 16 Francis, J,S., Sampson, L.D., Gerace, T. and Schneiderman, N., Cardkwascular responses of rabbits to ESB: effccls of anesthetization, stimuhls frequency and pulsetrain duration, Physiol. Behav.,ll (19731 195 2(13. 17 Gelsema. A.J. and Calaresu, F.R., Chemical microstimulation of thc septal area lowers arterial pressure in the rat, Am. J. Physiol., 252(1987) R760 767. 18 Glick, G. and Yu, P.N., Hemodynamic changes during spontaneous vasowigal reactions, Am. J. Med., 34 (19631 42 51.

19 Greenfield. A.D.M., An emotional faint. Lancet. i (19511 13(12-1303. 20 Gupta, B.N., Abboud, A.L., Floras. J.S.. Aylward. P.E. and Abboud, F.M.. Vasopressin lacilitates inhibition of renal nerve activity mediated through vagal afferents, Am. J. Physiol., 253 (19871 H1-7. 21 Hasser, E.M., Undesser, K.P. and Bishop, V.S., Interaction of vasopressin with urea postrema during volume expansion, Am. J. Physiol., 253 {1987) R6115 010. 22 Hilton, S.M. and Spyer, K.M.. Participation ol the anterior hypothalamus in the baroreceptor reflex, J. Physiol. (Lond.I. 218 (19711 271-293. 23 Katafuchi, T., Oomura. Y. and Kurosax~a, M , Effects of chemical stimulation of paraventricular nucleus on adrenal and renal ncrvc activity in rats, Neurosci, kct1.. ,S6 (1988) 195-200. 24 Katz, R.L., Kahn. N. and Wang, S.(., Brain stem mechanisms subserving barorcccptor reflexes. Factors affecting the carotid occlusion response, In P. Kezdi (Ed.), Baroreceptors and Hypertension, Pergamon Press, ()xford, 1%7, pp. 169 178. 25 Kent, B.B.. Drane, J.W. and Manning. J.W., Supiapollthle contributions to the carotid sinus reflex in the cat, Circ. Res., 29 (19711 534 541. 26 Korner, P.l,. Integrative ueura] cardiovascular control, Physiol. Rex'., 51 (It;71) 312-367. 27 Korner, P.I., Shaw, J., West. M.J. and Oliver, J . R . ('cntral nc~'ous system control of haroreceptor reflexes in the rabbit. Circ. Res..31 (19721 637 052. 28 Korner, P.l., Uther. J.B. and While, S.W., ('entral nervous integration of the circulatory and respiratory responses to arterial hypoxemia in the rabbit. Circ. Res., 24 (19691 757-776, 29 koewx, A.D. and McKcllar, S., The neuroanatomical hasis of central cardiovascular control, Fed. Proc.. 3~t (198{1) 2495-25{/3. 30 L61\'ing, FI,. Cardiowtscular adjustments induccd from the rostral cingulate gyrus, with special rcfcrcnce to sympathodnhibitory inechanisms, Acta Physiol. Stand., 53 {Suppl. 1841 (1%11 5 82. 31 Ludbrook, J.. Comparison of the reflex effects of arterial barorcceptors and cardiac receptors on the heart rate of conscious rabKts, Clin. Exp. Pharmacol. Physiol.. 11 (19841 245 261L 32 Ludbrook, J, and Graham. W.F,. Thc role of cardiac receptor and arterial baroreceptor reflexes m control of the circulation during aCLlte change of blood x,'ohlnle in thc conscious rabbit. Circ. Res., 54 (19841 424 435. 33 gudbrook, J., Potocnik, S.J. and Woods, R.L., Simulation of acute haemorrhage in ummaesthctiscd rabbits. Clin. Exp. Pharmacol. Physiol., 15 (1988) 575 584. 34 kudbrook, J. and Ruttcr, P.('., Effect of naloxonc on haemodynamic rcsponses to acute blood loss in unanacsthctized rabbits, ,I. Physiol. (Loud.). 400 11988) 1-14. 35 Miller. R.G., Simultaneous Statistical lnlerence. 2nd. edn.. Springer-Verlag. New York, 1981. 36 Miyazawa. T.M.. Gelsema, A.J., and Calarcsu. [:.R.. Sep-

14

37

38

39

40

4l

42

lal neurons respond to activation of baro- and chemoreceptors in the rat, Am. J. Physiol., 254 (1988) R331-337. Morita, H., Nishida, Y,, Motochigawa, H., Uemura, N., 14osomi, H. and Vatner, S.F., Opiate receptor-mediated decrease in renal nerve activity during hypotensive hemorrhage in conscious rabbits, Circ. Res., 63 (1988) 165-172. Quail, A.W., Woods, R.L. and Korner, P.I., Cardiac and arterial baroreceptor influences in release of vasopressin and renin during hemorrhage, Am. J, Physiol., 252 (1987) 111120 1126. Reis, D.J. and Cu~nod, M., Central neural regulation of carotid baroreceptor reflexes in the cat, Am. J. Physiol., 209 (1965) 1267-1277. Schadt, J.C. and Gaddis, R.R., Cardiovascular responses to hemorrhage and naloxone in conscious barodenervated rabbits, Am. J. Physiol., 251 (1986) R909-915. Schadt, J.C. and Ludbrook, J., Hemodynamic and neurohumoral responses to acute hypovolemia in conscious mammals, Am. J. Physiol., 260 (1991) 305-318. Schadt, J.C., McKown, M.D., McKown, D.P. and Franklin, D., Hemodynamic effects of hemorrhage and subsequent naloxone treatment in conscious rabbits, Am, J. Physiol., 247 (1984) R497-5(/8.

43 Spencer, S.E., Sawyer, W.B. and Loewy, A.D., Cardiovascular effects produced by L-glutamate stimulation of the lateral hypothalamic area, Am. J. Physiol., 257 11989) H540-552, 44 Strack, A.M., Sawyer. W.B., Hughes, J.H., Plait, K.B. and Eoewy, A.D., A general pattern of CNS innervation of the sympathetic outflow demonstrated by transneuronal pseudorabies viral infections, Brain Res., 491 (1989) 15~-162, 45 Undesser, K.P., Hasser, E.M., Haywood, J.R., Johnson, A.K. and Bishop, V.S., Interactions of vasopressin with the area postrema in arterial baroreflex function in conscious rabbits, Circ. Res., 56 (1985) 41(I-417. 46 Van Leeuwen, A.F., Evans, R.G. and Ludbrook, J., Effects of halothane, ketaminc, propofol and alfentanil anaesthesia on circulatory control in rabbits, Clin Exp. Pharmacol. Physiol., 17 (1990) 781 798, 47 Woods, R.L. and Johnston, C.l.. Contribution of vasopressin to the maintenance of blood pressure during dehydration, Am. J. Physiol., 245 (19821 F615-621. 48 Yamashita, tt., Kannan, H., Kasai, M. and Osaka, T., Decrease in blood pressure by stimulation of the rat hypothalamic paraventricular nucleus with i.-glutamate or weak current. J. Auton, Nerv. Syst., 19 (1987) 229-234.