Adrenergic control of the cardiovascular system in the turtle Trachemys scripta

The circulatory system of vertebrates is primarily regulated by the autonomic nervous system, as well as by hormonal and local factors. The autonomic nervous system consists of a parasympathetic vagal inhibitory component and a sympathetic adrenergic excitatory component. The autonomic nervous system is of paramount importance for regulation of blood pressures,cardiac contractility and distribution of blood flows between various vascular beds (Nilsson, 1983; Morris and Nilsson, 1994). In turtles, the ventricle is undivided, and blood-flow distribution between the systemic and pulmonary circulations and the intracardiac shunt patterns are largely determined by the vascular resistances of these two circulations (e.g. Shelton and Burggren, 1976; Hicks et al., 1996). It is well established that cholinergic control of pulmonary vascular resistance exerts an important role in the distribution of blood flow (e.g. Johansen and Burggren, 1980; Hicks, 1998), but an adrenergic component also seems to be involved in the regulation of vascular resistances of the pulmonary and systemic circulations(Burggren, 1977; Hicks, 1994). Increased sympathetic tone is often associated with exercise and stressful conditions,where oxygen delivery needs to be augmented through increased systemic blood flow through increased heart rate (fH) and/or stroke volume. A sympathetically mediated reduction in the re-circulation of systemic venous blood within the systemic circulation (i.e. a reduced right-to-left shunt)will improve systemic oxygen transport through increased oxygen saturation of systemic arterial blood (Wang and Hicks,1996b). Despite this benefit, a clear picture of the role of adrenergic stimulation in the regulation of pulmonary vascular resistance(R_pul) in turtles has yet to emerge because contradictory results appear in the literature (Hicks,1994); some studies show no effects of catecholamine injections,while others report reductions in R_pul(Luckhardt and Carlson, 1921; Berger, 1972; Milsom et al., 1977; Burggren, 1977; Comeau and Hicks, 1994).

While the overall haemodynamic effects of adrenergic stimulation have been characterised to some degree in turtles, it remains unknown as to what extent the changes in cardiac shunt pattern are caused by increased systemic resistance (R_sys) or decreased R_pul. Furthermore, the specific role of α-adrenergic and β-adrenergic receptors underlying the regulation in the pulmonary circulation is largely unknown. Non-anaesthetised turtles primarily regulate the ratio of the vascular resistances (R_pul/R_sys), and thereby cardiac shunt pattern, through a cholinergically mediated constriction of the pulmonary circulation. However, owing to the lack of vagal tone in anaesthetised turtles, the present study allowed us to examine how the ratios of vascular resistances affects shunt patterns when the resistances are manipulated through adrenergic stimulus or blockade. We measured haemodynamic responses following bolus injections of norepinephrine and the specificα- and β-adrenergic agonists phenylephrine and isoproteronol. Furthermore, we investigated the effects of pharmacological blockade using theα-adrenergic antagonist phentolamine and the β-adrenergic antagonist propranolol. Here, we aim to characterise the adrenergic regulation of the cardiovascular system in turtles and its possible functional implications.

Freshwater turtles Trachemys scripta (Gray) with a body mass ranging between 0.6 kg and 1.5 kg (mean ± S.D., 0.85±0.07 kg)were obtained from Lemberger Inc. (Oshkosh, WI, USA) and air-freighted to Aarhus University, where they were housed in a large fibreglass tank containing fresh water heated to 25±2°C and had free access to dry platforms, allowing for behavioural thermoregulation. All procedures were in accordance with the laws of animal care and experimentation in Denmark.

Turtles were anaesthetised by an intramuscular injection of sodium pentobarbital (Nembumal, Sygehusapotekerne, Denmark; 50 mg kg^-1)and normally ceased to exhibit responses to pinching of the legs or tail within 60 min. The trachea was then exposed by a ventral incision in the neck,and the turtle was tracheotomised for artificial ventilation. During surgery and the entire experimental protocol, turtles were maintained ventral side up and received continuous artificial ventilation with a tidal volume of 20-30 ml kg^-1 at a frequency of 20-30 breaths min^-1 using an HI 665 Harvard Apparatus Respirator (Cambridge, MA, USA). Because of the high ventilation rate, we used a gas mixture of 97% room air and 3% CO₂delivered by a Wösthoff gas-mixing pump (Bochum, Germany) to mimic the arterial blood PCO₂in vivo (e.g. Glass et al., 1983; see Crossley et al., 1999 for arterial blood gas composition of anaesthetised turtles using a similar experimental design).

To access the central vascular blood vessels, a 5 cm × 5 cm portion of the plastron was removed using a bone saw, and the pectoral muscles were gently loosened from the excised piece. A polyethylene catheter, containing heparinised saline, was occlusively inserted into the left carotid artery and advanced into the right aortic arch. The common pulmonary artery was non-occlusively cannulated using an intravenous catheter (Surflo, Terumo Medical Cooporation, Elkton, MD, USA) inserted upstream into the common pulmonary artery to within 0.5 cm of the heart. This required that the pericardium was opened by a 1 cm incision. After insertion, the intravenous catheter was connected to PE-60 tubing, which, in turn, was connected to a Baxter Edward disposable pressure transducer (model PX600, Irvine, CA, USA). The signals from the pressure transducers were amplified using an in-house-built preamplifier and were calibrated daily against a static column of water. For measurements of systemic and pulmonary blood flows, a 1-1.5 cm section of the left aortic arch (LAo) and the left pulmonary artery (LPA) were freed from connective tissue for placements of transit-time ultrasonic blood-flow probes (2R or 2S probes; Transonic System, Inc., NY, USA). To improve the signal, acoustical gel was infused between the blood vessel and flow probe. The flow probes were connected to a Transonic dual-channel blood-flow meter (T206) for measurements of instantaneous blood-flow rates. Signals from the pressure transducer and the blood-flow meter were recorded with a Biopac MP100 data acquisition system (Biopac Systems, Inc., Goleta, CA,USA) at a sampling frequency of 50 Hz.

This study did not measure blood flow in all systemic and pulmonary arteries. Several studies on anaesthetised and non-anaesthetised freshwater turtles have shown that systemic blood flow(Q̇_sys) can be adequately estimated as 2.85Q̇LA_o(Shelton and Burggren, 1976; Comeau and Hicks, 1994; Wang and Hicks, 1996a; this relationship also persists after injection of adrenergic antagonists and agonists in fully recovered turtles; J. A. W. Stecyk, J. Overgaard, A. P. Farrell and T. Wang, unpublished data). Likewise, pulmonary blood flow(Q̇_pul) can be calculated as 2Q̇_LPA under the assumption that blood flow in the right pulmonary artery equals that in the left. fH was calculated from the instantaneous blood-flow trace from the LAo. Total cardiac output(Q̇_tot) was calculated as Q̇_sys+Q̇_pul,and total stroke volume (VS_tot; pulmonary + systemic) was calculated as Q̇_tot divided by fH. Pulmonary and systemic resistances (R_puland R_sys, respectively) were calculated from mean blood pressure (P) and mean blood flow(R_pul=P_pul/Q̇_puland R_sys=P_sys/Q̇_sys)using the assumption that central venous blood pressures are zero. This assumption may lead to a small overestimation of vascular resistances, as pointed out by Badeer and Hicks(1994).

After ensuring steady-state conditions (stable pressures and flows for a minimum of 30 min), the turtles were sequentially injected with adrenergic receptor antagonists and agonists using the protocol described below. The injections were administered through the systemic catheter, which was subsequently flushed with saline. Although not studied in detail, injections of adrenergic agonists in the pulmonary artery yielded similar responses to those following systemic injections. The total volume injected never surpassed 2 ml, and control injections of 2 ml saline did not cause haemodynamic changes. Following injection of each drug, we recorded haemodynamic parameters at a maximal response. This usually occurred after 1-2 min with the agonists and after 30 min with the antagonists. After injections of agonists, the haemodynamic variables were allowed to return to baseline values before continuing the protocol. Following injection of antagonists, we waited for at least 30 min until a new steady-state was established.

The animals were randomly divided into two groups (body mass of 0.78±0.03 kg and 0.92±0.12 kg for group 1 and 2, respectively)and presented with the following pharmacological protocols (all chemicals were purchased from Sigma-Aldrich, Vallensbæk Strand, Denmark).

Group 1: norepinephrine 5 μg kg^-1, phenylephrine 5 μg kg^-1, propranolol 3 mg kg^-1, norepinephrine 5 μg kg^-1, phenylephrine 5 μg kg^-1, phentolamine 3 mg kg^-1. Group 2: these animals were exposed to the same protocol as group 1 with the exception that the injections of phentolamine and propranolol were reversed. In addition, eight of the twelve animals (four from each group)were injected with a lower dose of norepinephrine (1 μg kg^-1),and four animals (two from each group) were injected with isoproteronol (1μg kg^-1). These injections were performed before administration of the first antagonist. After completion of the experimental protocol, all turtles were euthanized by a lethal injection of saturated KCl.

Recordings of blood flows and pressures were analysed using AcqKnowledge data analysis software (Biopac Systems, Inc., Goleta, CA, USA, version 3.7.0)where the mean values for fH, Q̇_LAo, Q̇_LPA, P_sys, P_pul, R_sysand R_pul were determined over a 1-2 min period following each injection. Differences in control values between groups 1 and 2 were tested using a t-test. In each group, the results obtained following injections of agonists and antagonists were tested for significant differences from untreated baseline values using a one-way analysis of variance (ANOVA)for repeated measures. Significant differences among mean values were identified using a post hoc Student—Newman—Keuls (SNK)test. Effects of agonists (phenylephrine and norepinephrine) after injection of antagonists (propranolol or phentolamine) were compared with the baseline value that prevailed 30 min after injection of the antagonist. Similarly, the eight turtles that received two doses of norepinephrine were tested using a one-way ANOVA for repeated measures and a post hoc SNK test. A paired t-test was used to assess for significant differences from control values in the four animals injected with isoproteronol. In all tests, a limit of P<0.05 was applied to identify significant differences. Owing to low N values for some comparisons, type 1 errors are possible, and lack of significant differences should be interpreted cautiously. The data are presented as means ± S.E.M. and normalised to a 1 kg turtle.

The cardiovascular status for each of the two experimental groups during the control condition is shown in Table 1. fH, P_pul and P_sys were similar between the two groups. Although blood flows seemed higher in group 1 compared with group 2, this difference was not significant. Similarly, there were no significant differences between the two groups in any of the other reported parameters, and the ratios of Q̇_pul/Q̇_sysand R_pul/R_sys were very similar in both groups. In both groups, Q̇_pul greatly exceeded Q̇_sys, P_sys was considerably higher than P_pul, and R_sys was approximately threefold greater than R_pul.

The temporal changes in heart rate, blood flows and pressures following injection of norepinephrine are illustrated in Fig. 1A, 2A for two turtles from each experimental group. The group data are presented in Table 1. Data from the eight turtles (four from each group) that received two doses of norepinephrine were pooled and are presented in Fig. 3. Both concentrations of norepinephrine had significant chronotropic and inotropic effects. Q̇_tot, Q̇_pul, fH and V_Stot all increased significantly following injection of 1 μg kg^-1 norepinephrine, and fH also increased further following injection of 5 μg kg^-1norepinephrine (Fig. 3A). Q̇_sys was unchanged following injection of 1 μg kg^-1 norepinephrine but decreased significantly after injection of 5 μg kg^-1 norepinephrine. Norepinephrine caused significant dose-dependent increases in P_pul and P_sys but did not affect R_pul, while R_sys increased significantly from baseline values following injection of 5 μg kg^-1 norepinephrine (Fig. 3D).

Injection of the α-agonist phenylephrine led to an increase in Q̇_tot, although the effect was only significant in group 2 (Table 1). The increase in Q̇_pul and Q̇_tot following injection of phenylephrine was slightly lower than that following norepinephrine, and this correlated with the lack of a chronotropic effect of phenylephrine, but there were similar increases in V_Stot after injection of either phenylephrine or norepinephrine(Table 1). As with norepinephrine, phenylephrine elicited a significant increase in Q̇_pul, while R_pul was not affected. The change in Q̇_sys following phenylephrine was not statistically significant, although R_sys almost doubled (Table 1; Figs 1A, 2A).

Blockade of the α-receptors by phentolamine induced significant reductions in R_sys, P_sys, P_pul and Q̇_pul without altering fH, Q̇_sys and R_pul (Table 1). The pronounced α-adrenergic systemic vasoconstriction was abolished by phentolamine. Thus, when phenylephrine or norepinephrine were injected after phentolamine, there were no significant cardiovascular changes except for an increase in P_sys and Q̇_tot following phenylephrine.

The haemodynamic changes caused by β-adrenergic receptor stimulation using isoproteronol are presented in Fig. 4. Isoproteronol produced significant positive chronotropic effects and significantly reduced Vs_tot, while there were no significant changes in Q̇_pul or Q̇_sys. Although there were no significant changes in R_sys and R_pul,isoproteronol caused significant hypotension in both the pulmonary and systemic circuits.

Abolishment of β-adrenergic tone, using propranolol, significantly reduced fH and Q̇_totbut did not affect Vs_tot. A significant increase in R_pul resulted in a decrease in Q̇_pul and no change in P_pul. The changes in R_sys and Q̇_sys were not statistically significant. Similarly, when propranolol was applied after phentolamine (group 2), fH decreased significantly, without any further change in Q̇_tot and Q̇_pul (which had already been reduced significantly by phentolamine). The increases in R_sys and R_pul following propranolol injection did not reach statistical significance. Compared with norepinephrine injection before β-adrenergic blockade, norepinephrine injection following propranolol (group 1) caused an even larger increase in R_sys, while the systemic hypertension was of similar magnitude. Phentolamine injection following β-adrenergic blockade (group 1) caused cardiac collapse, with Q̇_tot decreasing to 25% of the initial control values, and systemic and pulmonary blood pressures decreasing to 50% of their initial control values(Table 1).

The haemodynamic variables reported during the untreated control condition of our study are in general agreement with previous studies on anaesthetised Trachemys scripta (see table 3 in Crossley et al., 1998). Nevertheless, Q̇_pul in group 1 of our study was somewhat higher, and fH was slightly lower, than previous reports (Hicks et al.,1996; Hicks and Comeau,1994; Comeau and Hicks,1994; Crossley et al., 1998, 2000). As in previous studies on anaesthetised turtles, Q̇_pul always exceeded Q̇_sys, resulting in a net left-to-right cardiac shunt. Furthermore, our results clearly show that an increase in R_sys increased this left-to-right shunt by increasing Q̇_pul, often without compromising Q̇_sys.

We studied anaesthetised turtles to isolate the adrenergic control of the cardiovascular system without the confounding effects of the often-profound cardiovascular changes associated with ventilation in non-anaesthetised animals. The haemodynamic variables of anaesthetised turtles differ markedly from that of non-anaesthetised and recovered animals. For example, during non-ventilatory periods, recovered animals are characterised by a low fH and a large right-to-left shunt, whereas lung ventilation produces increases in fH, Q̇_pul and Q̇_pul/Q̇_sys(White and Ross, 1966; Shelton and Burggren, 1976; West et al., 1992; White et al., 1989; Wang and Hicks, 1996a). The low fH and Q̇_pulduring apnoea appear to stem from a large vagal tone, as injection of atropine increases both fH and Q̇_pul and abolishes their reduction during breath holding (Berger,1972; Burggren,1975; Hicks and Wang,1998; Hicks, 1998; Hicks and Farrell, 2000). Injection of a beta-blocker did not affect the changes in fH associated with intermittent ventilation(Burggren, 1975), although nadolol reduced mid-apnoeic fH(Hicks and Farrell, 2000). It is unknown whether the decreased fH results from blocking the effects of circulating catecholamines or the adrenergic innervation of the heart. Hicks (1994) suggested that part of the increased Q̇_puland fH during ventilation can be attributed to increased adrenergic tone. Indeed, in anaesthetised turtles, afferent stimulation of the vagus causes tachycardia and increased Q̇_pul, which is abolished by inhibiting adrenaline release from nerve endings with bretylium(Comeau and Hicks, 1994). Hence, it seems that the reciprocal changes in fH and blood flows associated with breathing are predominantly caused by alterations in vagal tone, but that changes in adrenergic tone contribute to cardiac rhythm in non-anaesthetised turtles.

Blood flow and fH of anaesthetised turtles are higher than in non-anaesthetised, apnoeic turtles because vagal tone on the heart and pulmonary artery is lost in anaesthetised and ventilated animals(Crossley et al., 1998). As a result, blood flow and fH in anaesthetised turtles are quantitatively more similar to those measured in non-anaesthetised turtles during ventilation, when vagal tone is reduced. Nevertheless, anaesthetised turtles maintain an adrenergic tone similar to recovered animals, as the 35% reduction in fH occurring with propranolol injection in the present study(Table 1) agrees well with the 16-35% reduction in fH after nadolol injection in non-anaesthetised turtles (Hicks and Farrell,2000).

This is the first study to simultaneously examine α- andβ-adrenergic regulation of the heart as well as the pulmonary and systemic circulations in turtles. Previous studies have shown that systemic injection of catecholamines leads to an increased heart rate that can be blocked or greatly attenuated by injection of β-receptor antagonists(Burggren, 1975; Hicks and Farrell, 2000). In our study, propranolol caused a large reduction in fH but did not completely abolish the chronotropic effect of norepinephrine(Fig. 1; Table 1). This is likely to result from competitive binding between propranolol and norepinephrine on the cardiac ß-adrenergic receptors, as phentolamine and phenylephrine did not affect fH, whereas specific stimulation of β-adrenergic receptors using isoproteronol elicited marked tachycardia(Fig. 4). These observations are consistent with studies on isolated ventricular strips from Trachemys, where no chronotropic effects could be demonstrated following α-adrenergic stimulation(Van Harn et al., 1973).

In this study, we used bolus injections of adrenergic agonists and antagonists to evaluate the adrenergic regulation of the heart and vascular tones. This experimental design does not allow for a differentiation between the effects of circulating catecholamines relative to the effects of catecholamine release from nerve endings.

In our study, high doses of norepinephrine caused a marked increase in R_sys (Table 1). This response was mimicked by phenylephrine and could be blocked by phentolamine, demonstrating that the systemic vascular constriction following norepinephrine injection is caused by α-adrenergic receptors(Table 1; Fig. 2). Phentolamine halved R_sys, pointing to a substantial α-adrenergic tone on the systemic circulation in anaesthetised turtles(Table 1). This is also the case in fully recovered turtles, where phentolamine reduced R_sys from 0.08 kPa ml^-1 min kg to 0.06 kPa ml^-1 min kg (A. W. Stecyk, J. Overgaard, T. Wang and A. P. Farrell,unpublished data). Thus, as in virtually all other vertebrates examined so far, stimulation of the α-adrenergic receptors in the systemic vascular beds is associated with constriction(Nilsson, 1983). Conversely,stimulation of β-adrenergic receptors with isoproteronol was associated with relaxation of the systemic vascular beds(Fig. 4), and blockade ofβ-adrenergic receptors increased R_sys by a factor of 1.4 (Table 1; Fig. 1). Similar changes have been observed in fully recovered turtles, where β-adrenergic blockade using nadolol led to a threefold increase in R_sys from 0.05 kPa ml^-1 min kg to 0.15 kPa ml^-1 min kg(Hicks and Farrell, 2000).

Comeau and Hicks (1994)reported a small decrease in R_pul of anaesthetised turtles following catecholamine injection, whereas other studies were unable to detect reductions in R_pul(Luckhardt and Carlson, 1921; Berger, 1972; Milsom et al., 1977). In our study, both norepinephrine and phenylephrine elicited a small, statistically non-significant reduction in R_pul(Table 1; Fig. 1). A potentially minor dilatory role for α-adrenergic receptors was further suggested by the increase in R_pul (albeit not statistically significant)following phentolamine injection (Table 1). Thus, the pulmonary circulation is much less responsive toα-adrenergic stimulation than is the systemic circulation, where R_sys doubled after injection of phenylephrine. Burggren(1977) earlier reported a 15-20% reduction in the resistance of the isolated distal pulmonary artery upon injection of epinephrine. In the lizard Trachydosaurus rugosus,stimulation of β-receptors dilates the pulmonary vasculature, while stimulation of α-receptors appears to cause a constriction(Berger, 1973). In the snake Elaphe obsoleta, there is also evidence for β-adrenergic relaxation of the pulmonary artery (Donald et al., 1990). The role of β-receptors in the pulmonary circulation is less clear in turtles. Thus, while propranolol caused a significant increase in R_pul, suggesting a tonically active β-adrenergic mediated relaxation(Table 1; Fig. 1), injection of isoproteronol did not reduce R_pul(Fig. 4), perhaps because this vasoactive mechanism was already fully activated.

Our study demonstrates marked differences in the vascular response to adrenergic stimulation between the pulmonary and systemic circulations. The pulmonary circulation is much less responsive to catecholamines, andα-adrenergic receptors could even mediate opposite changes in vascular tone. Other studies also show that vasoactivity in the systemic and pulmonary circulations of turtles differs in other respects. Thus, hypoxia causes vasodilation in the systemic circulation as opposed to vasoconstriction in the pulmonary circulation (Crossley et al.,1998). NO exerts a substantial role in maintaining systemic vascular tone but does not seem important in the pulmonary circulation(Crossley et al., 2000). Conversely, pulmonary resistance is under strong cholinergic control(Berger, 1972; Burggren, 1977; Milsom et al., 1977; Hicks and Comeau, 1994),whereas the cholinergic tone on the systemic circulation is either low or absent (Kirby and Burnstock,1969; Berger and Burnstock,1979; Comeau and Hicks,1994). Indeed, more studies are needed to better define the basis of these differences in vasoactivity.

The present study is revealing in terms of a potential role of venous venoconstriction on cardiac function. For example, despite the slight decrease in Q̇_sys and the marked increase in R_sys with both phenylephrine and norepinephrine, there was an increase in Q̇_tot and Vs_tot. The increase in Vs_tot could be a result of an α-adrenergically mediated venoconstriction. Conversely,the apparent negative inotropic effect of isoproteronol on the heart (i.e. the decrease in Vs_tot) might be best explained by aβ-adrenergically mediated venodilation reducing venous return to the heart. These venous effects of adrenergic stimulation need not be manifest in changes in overall vascular resistance but warrant further attention because of their potential effects on Vs_tot and Q̇_tot. Thus, although it seems that arterial vascular resistances determine the relationship between pulmonary and systemic blood flows, it is likely that the magnitude of changes in total blood flow is determined, in part, by the degree of venous vasoactivity.

Our study is the first to comprehensively evaluate effects on adrenergic stimulation on cardiac shunt patterns. Fig. 5 depicts the relationship between blood-flow distribution,expressed as Q̇_pul/Q̇_sys,and the ratio of the vascular resistances in the pulmonary and systemic circulations (R_pul/R_sys). A similar relationship illustrated that the distribution of blood flows between the systemic and pulmonary circuits in turtles correlates closely with the ratio of pulmonary and systemic vascular resistances(Hicks et al., 1996; Crossley et al., 1998). We were able to show that an even more profound range of cardiac shunting can be produced with adrenergic drug injections that act primarily on the systemic circulation. Consequently, anaesthetised turtles exhibit large net left-to-right shunts(Q̇_pul/Q̇_sysof >1) when R_pul is lower than R_sys(R_pul/R_sys of <1), while a net right-to-left shunt(Q̇_pul/Q̇_sysof <1) exists when R_pul is higher than R_sys (R_pul/R_sys of>1). This relationship implies that systemic outflow resistance, regulated through α-adrenergic systemic vasoconstriction at the arterial level (as shown by the fourfold changes in R_sys betweenα-adrenergic stimulation and blockade), is the primary determinant of distribution of Q̇_tot to the blood flows between the two circuits in these anaesthetised turtles. In non-anaesthetised turtles, changes in the shunt pattern are effected primarily by cholinergically mediated constrictions of the pulmonary circulation. Thus,by focusing on adrenergic control, our work has provided evidence that modulation of the systemic vascular resistance produces the same types of cardiac shunts in turtles as does modulation of pulmonary resistance. Consequently, for turtles, it appears that up to 97% of the observed variability in the cardiac shunting (Fig. 5) can be explained by the ratio of the resistances in the two circulations.

It remains unanswered whether the profound changes in cardiac shunting produced here by adrenergic mechanisms play a role in non-anaesthetised turtles. Increased adrenergic tone is often associated with exercise and stressful conditions such as hypoxia, where sympathetic stimulation safeguards systemic oxygen delivery through increased fH and cardiac output. In addition, reduced cardiac right-to-left shunting through increased adrenergic tone and decreased vagal tone increases oxygen delivery (Wang and Hicks, 1996b, 2002). Our study is consistent with this view. However, because anaesthetised turtles have no vagal tone,adrenergic stimulation serves to increase the net left-to-right shunt rather than abolishing the net right-to-left shunt that would occur in vivo. Furthermore, the changes in shunt pattern are primarily accomplished through increased R_sys and are associated with reductions in Q̇_sys. Therefore, adrenergic stimulation of anaesthetised animals is unlikely to improve systemic oxygen delivery. It is probable, however, that oxygen delivery in fully recovered animals would benefit from increased sympathetic tone, as resting and undisturbed animals are characterised by net right-to-left shunts and low arterial oxygen levels (see tables 1, 2 in Wang and Hicks, 1996a).

This study was supported by the Danish Research Council and NSERC Canada.