Comparative Cardiac Effects of Terlipressin, Vasopressin, and Norepinephrine on an Isolated Perfused Rabbit Heart

A total of 55 hearts from New Zealand albino female rabbits weighing 2.75–3.25 kg were used in this study. Animals were fasted for 12 h before the experiment with free access to water. Care of the animals conformed to the recommendations of the Helsinki Declaration, and the study was performed in accordance with the regulations of the official edict of the French Ministry of Agriculture.

The perfusion medium was reconstituted by mixing human erythrocytes (Etablissement Français du Sang, Paris, France) and a modified Krebs-Henseleit bicarbonate buffer containing 118 mm NaCl, 5.9 mm K⁺, 2.5 mm free Ca²⁺, 0.5 mm MgSO⁴, 1.17 mm NaH²PO⁴, 28 mm NaHCO³, 11 mm glucose, 0.9 mm lactate, and 0.5% bovine serum albumin (Sigma-Aldrich; Saint-Quentin Fallavier, France). The human erythrocytes were stored at 4°C in our laboratory for no longer than 1 week. They were centrifuged and washed with saline (Cell-Saver 4; Hemonetics, Braintree, MA). The mixture allowed us to obtain a hemoglobin value of around 8 g/dl. The reconstituted blood was filtered, then continuously oxygenated with a gas mixture comprising 20% oxygen, 5% carbon dioxide, and 75% nitrogen using a membrane oxygenator (Optima; Cobe Cardiovascular, Arvada, CO). After rewarming to 37°C, electrolyte concentrations were adjusted to achieve physiologic concentrations and sodium bicarbonate was added to obtain a pH between 7.35 and 7.45.

Each of New Zealand rabbit was anesthetized with pentobarbital sodium (25–35 mg/kg intravenously). After thoracotomy, the heart and aorta arch were rapidly excised and placed in cold (4°C) isotonic saline solution. Under immersion, the pericardium was quickly removed and the aorta was cannulated. Then, a retrograde perfusion was begun according to the Langendorff’s technique with a hydrostatic perfusion pressure of 80 mmHg. The coronary driving pressure was computed from the signal pressure obtained from a pressure transducer that was positioned above the aortic valves. The apparatus was modified to enable the continuous recording of speed of peristaltic pump, which reflected coronary blood flow19,20and was electronically regulated by the signal of coronary driving pressure to maintain it at 80 mmHg throughout the experiment. Whenever possible, the heart rate was maintained constant during each experiment by atrial pacing. Nevertheless, when the heart developed arrhythmia during pacing or when the spontaneous heart rate was faster than 150 bpm, the heart was not paced. The coronary sinus was drained by a small catheter inserted into the pulmonary artery. The drained venous blood was not recycled. A cannulated fluid-filled balloon connected to a pressure transducer by a rigid catheter was inserted into the left ventricle through a left atrial incision. A 2-ml graduated syringe was connected to this pressure transducer and allowed to increase the intraventricular volume. The latter was then adjusted to obtain a left ventricular end-diastolic pressure of approximately 10 mmHg.21Left ventricular end-systolic and end-diastolic pressures and heart rate were recorded, and the maximal positive and negative left ventricular pressure derivatives were electronically derived from the left ventricular pressure signal. Because intraventricular volume and heart rate were held constant, the maximal positive and negative left ventricular pressure derivatives reflected inotropic and lusitropic properties, respectively. The whole apparatus was enclosed in a thermostatic chamber at 37.5°C.

Arterial and venous coronary oxygen tension, arterial and venous coronary carbon dioxide tension, and pH were measured with standard electrodes at 37°C (GEM 3000; Instrumentation Laboratory, Saint-Mandé, France) and the arterial hemoglobin concentration and arterial and venous coronary saturations were measured with a hemoximeter (Model IL 682, Instrumentation Laboratory). Arterial and venous coronary oxygen content and myocardial oxygen consumption were derived from standard formulae. At the onset of each experiment, a sample of the reconstituted blood was withdrawn to determine the concentration of main electrolytes (Na⁺, K⁺, Cl⁻, NaHCO³, and Ca²⁺).

A 10-min period was allowed for stabilization of ventricular performances and coronary blood flow. After baseline measurements, terlipressin (Glypressine, Ferring, Limhamn, Sweden) was added to the perfusion medium just above the aortic cannula in 10 hearts. Intracoronary infusions were used to produce target concentrations of 1, 3, 10, 30, and 100 nm. The infusion was maintained until a new steady state was obtained (approximately 5 min). After the end of each infusion of terlipressin, a recovery period was allowed to return to baseline values. Before and after each terlipressin infusion, arterial and coronary sinus samples were collected for blood gas analysis. Intrinsic cardiac effects of increased concentrations of [Arg⁸]vasopressin (Sigma-Aldrich) from 10 to 1000 pm and norepinephrine (Aguettant, Lyon, France) from 1 to 100 nm were studied under similar conditions in additional hearts (n = 10 in each group). All these concentrations were chosen to encompass the therapeutic range. It was found that an intravenous bolus of 7.5 μg/kg of terlipressin induced a maximum plasma terlipressin concentration of approximately 17 nm.13In critically ill patients, Tsuneyoshi et al.  7reported that a continuous infusion of vasopressin at 0.04 U/min induced a plasma vasopressin concentration of approximately 300 nm. Finally, a norepinephrine infusion of 0.6 μg · kg⁻¹· min⁻¹in anesthetized and ventilated dogs induced a mean plasma concentration of free norepinephrine of 44 nm.22 

To eliminate changes in myocardial performances induced by ischemic coronary vasopressor effect, we also assessed terlipressin in five hearts in which coronary blood flow was maintained constant. The coronary effect of the drug was reflected by changes of coronary perfusion pressure.

Because erythrocytes may be involved in metabolism of drugs, intrinsic cardiac effects of terlipressin were assessed in eight additional hearts perfused with Krebs-Henseleit bicarbonate buffer.

The involvement of the V^1areceptor was assessed in six hearts after pretreatment with a potent and specific V^1areceptor antagonist [d(CH²)⁵Tyr(Me)]AVP (Sigma-Aldrich).23,24Therefore, after an initial 10-min recovery period and baseline measurements, an intracoronary infusion of [d(CH²)⁵Tyr(Me)]AVP (0.1 μm) was initiated and maintained throughout the whole experiment. Ten minutes later, intracoronary infusion of 100 nm of terlipressin was started. To ensure that a large number of vasopressin receptors were efficiently blocked by this pretreatment, we also assessed the intrinsic cardiac effects of 1,000 pm of AVP in six additional hearts pretreated under similar conditions.

All drugs were freshly prepared and dissolved in distilled water excepted for the V^1areceptor antagonist [d(CH²)⁵Tyr(Me)]AVP, which was dissolved in dimethyl sulfoxide and distilled water as a stock solution. The final intracoronary concentration of dimethyl sulfoxide was 0.001%. Dimethyl sulfoxide concentrations of up to 0.01% have been shown to be devoid of significant effects on our experimental model.25Dilutions of drugs used were adapted so that the volume of intracoronary infusion never exceeded 5% of the resting coronary flow rate. Finally, all syringes were protected from light throughout the study.

Data are expressed as mean ± SD. Comparison of several means were performed using repeated-measures analysis of variance and the Newman-Keuls test. Comparison of two means was performed using a nonpaired Student t  test. A P  value less than 0.05 was required to reject the null hypothesis. All statistical analysis were performed using NCSS 6.0 software (Statistical Solutions Ltd., Cork, Ireland).

There were no significant differences in pH and electrolyte composition of the reconstituted perfusate between terlipressin, vasopressin, and norepinephrine groups (pH, 7.42 ± 0.06; Na⁺, 136 ± 5 mm, Cl⁻, 98 ± 4 mm; K⁺, 5.2 ± 1.2 mm, NaHCO³, 23 ± 3 mm; and Ca²⁺, 2.1 ± 0.3 mm). Similarly, the left ventricular weight was comparable between all groups (4.91 ± 0.68 g)

Baseline values of coronary blood flow were comparable between groups (3.9 ± 1.0 versus  3.1 ± 0.7 versus  3.5 ± 1.1 ml · min⁻¹· g⁻¹, NS). Return to baseline values was obtained before each perfusion of drugs (data not shown). Norepinephrine induced a biphasic coronary effect with a maximal and significant decrease at 3 nm and then increased above basal at 30 and 100 nm concentrations (fig. 1A). AVP and terlipressin induced a concentration-dependent decrease in coronary blood flow starting from 30 pm AVP concentration and 30 nm concentrations of terlipressin (fig. 1A).

Baseline values of myocardial performances were comparable among the groups (table 1). Similarly, heart rate was similar among the three groups (116 ± 13 versus  120 ± 10 versus  128 ± 13 bpm, not significant) and remained stable throughout all experiments (data not shown). The intracoronary infusion of norepinephrine induced a concentration-dependant increase in myocardial performance that was significant from 30 nm (table 1and fig. 1B). Intracoronary infusions of AVP (≥30 pm) and terlipressin (≥30 nm) were associated with significant decreases in myocardial performances of hearts (table 1and fig. 1B). Only an intracoronary infusion of 100 pm of vasopressin provoked a significant increase in left ventricular end-diastolic pressure (table 1).

Baseline values of myocardial oxygen consumption were comparable between groups (table 2). Return to baseline value of oxygenation variables was obtained before each new concentration in three groups (data not shown). Norepinephrine induced a significant increase in myocardial oxygen consumption from a concentration of 30 nm. AVP produced concentration-dependent decreases in myocardial oxygen consumption that reached significance as early as 100 pm concentrations (table 2). Although there was a trend towards a decrease in myocardial oxygen consumption with terlipressin, it did not reach statistical significance.

Intracoronary infusion of terlipressin induced a significant increase in coronary perfusion pressure from a concentration of 30 nm. However, this coronary vasopressor effect was associated with no significant changes in myocardial performances up to 100 nm where maximal positive and negative left ventricular pressure derivatives were 95 ± 2% and 95 ± 6% of baseline values, respectively (not significant).

As expected, the baseline value of coronary blood flow was significantly higher in buffer-perfused hearts (table 3). Before each concentration, return to baseline values was obtained (data not shown). The effects of terlipressin on coronary blood flow and lusitropic properties were not significantly affected by the type of perfusate. However, terlipressin impaired the inotropic parameters in buffer-perfused hearts significantly more (table 3).

The intracoronary infusion of 0.1 μm of [d(CH²)⁵Tyr(Me)]AVP affected neither resting coronary blood flow nor myocardial performances of erythrocyte-perfused hearts (data not shown). The cardiac effects of AVP (1000 pm) and terlipressin (100 nm) were nearly completely abolished in hearts pretreated with [d(CH²)⁵Tyr(Me)]AVP 0.1 μm (fig. 2).

The principle findings of the current study are that terlipressin, a synthetic analog of AVP, causes a reduction in coronary blood flow and, consequently, depression of myocardial performance only at supra-therapeutic concentration (30 nm), terlipressin’s intrinsic cardiac effects are significantly less pronounced that those observed by AVP, and, like AVP, its cardiac effects are essentially mediated via  V^1avasopressin receptors.

AVP, an endogenous nonapeptide of the neurohypophysis, and its long-acting synthetic analog, terlipressin, have been proposed as efficient alternative therapies to restore arterial blood pressure in refractory shock from various origins.1–9,26The physiologic rationale of this treatment is based on low catecholamine effectiveness and inappropriate low vasopressin concentrations in these clinical settings.1–3,7This deficiency in AVP seems to be related to an impaired synthesis because exogenous administration allows restoration of a high plasma concentration of AVP in these patients.1,7Although devoid of vasopressor effect in healthy patients, a vasopressin pressor hypersensitivity has been reported in patients suffering from vasodilatory septic shock.6This discrepancy could be explained by an increase in vascular receptor affinity to norepinephrine by AVP or a septic-induced autonomic insufficiency.6Like AVP, terlipressin vasopressor effect is principally mediated via  V^1avasopressin receptors.10However, their potent vasopressor effect could compromise regional blood flow and, thus, tissue oxygenation.15,16,27,28In the coronary circulation, some in vitro  studies reported that AVP produces a significant coronary vasoconstriction.16,17In the current study, we assessed intrinsic cardiac effects of vasopressinergic system agonists by using an isolated and erythrocyte-perfused rabbit heart. The clear advantage of this model is the presence of erythrocytes in perfusate medium, which avoids using a high degree of arterial oxygen tension, which, in turn, induces coronary vasoconstriction via  ATP-sensitive K⁺channels.25Because experimental studies suggest that these ATP-sensitive K⁺channels are involved in the vasoconstrictor effect of AVP,29,30we thought that previous closure of these channels by high degrees of oxygen tension might interfere with the vascular effects of a vasopressinergic system agonist. In erythrocyte-perfused rabbit hearts, we showed that AVP induces a significant coronary vasoconstriction from a concentration of 30 pm. This concentration is lower than those measured by Tsuneyoshi et al.  7in plasma (up to 500 pm) of patients with vasodilatory septic shock and receiving a continuous infusion of low dose of exogenous AVP. Although myocardial ischemia could complicate these adverse coronary vasopressive effects, no clinical studies have reported ischemic myocardial side effects using these two drugs to correct a refractory hypotensive state in critically ill patients. However, in most of these studies, patients with coronary disease were excluded. In the current study, coronary vasoconstrictor effects of vasopressin were associated with depression of myocardial performances. Our results are consistent with a previous in vitro  study that demonstrated that AVP induced significant depressive myocardial effects.17However, the fact that myocardial depression was the only consequence of ischemic coronary vasoconstrictor effect of AVP cannot be completely ruled out in the current study. Our findings suggest that a negative inotropic effect could also be involved in the decrease in cardiac output that is usually observed when sepsis-induced hypotension is corrected by using AVP.1,27However, although the assessment of intrinsic properties of myocardium remains difficult in vivo , some authors have reported that AVP had no negative effect on myocardial performances in patients suffering from catecholamine-resistant septic shock.31Because terlipressin has more attractive pharmacokinetic properties, some authors have used it preferentially to correct refractory hypotension.8–10,32Classically, terlipressin is described as inactive itself, meaning that it should be metabolized by peptidase into lysine-vasopressin, which is significantly detectable only after 40 min.13However, some authors previously reported a short onset of vasopressor effect of terlipressin, suggesting that it could exert intrinsic vasoactive effects.12,33Our findings suggest that terlipressin is not acting simply as a prodrug but rather, in an isolated rabbit heart, induces an intrinsic coronary vasoconstrictor effect from a concentration of 30 nm. In addition, the fact that these coronary intrinsic effects persisted in buffer-perfused hearts eliminated a potential metabolism of tricyl-lysine vasopressin into lysine-vasopressin by the erythrocytes present in the medium perforate. On the other hand, as previously reported by our group,34we observed that the inotropic effects of a drug can be significantly modified in buffer-perfused hearts. However, a metabolism by tissular peptidase cannot be completely ruled out. Our study is the first to report an intrinsic coronary effect of terlipressin. This coronary vasoconstrictor effect was significantly lower than that of AVP. The fact that terlipressin provokes a lower coronary vasoconstriction should be taken into account for its use in patients with coronary disease. No previous study has compared coronary vasopressor effects between terlipressin and AVP. Although terlipressin-induced coronary vasopressor effects occurred only in supra-therapeutic concentrations, they were powerful enough to provoke ischemic side effects. We showed that terlipressin-induced myocardial dysfunction disappeared when the myocardial oxygen delivery was maintained constant, suggesting the lack of direct negative inotropic effect. These findings are consistent with those obtained by Eyraud et al .,35who observed that terlipressin is devoid of significant negative inotropic effects in patients suffering from anesthesia-induced hypotension. On erythrocyte-perfused and isolated heart, norepinephrine induced a biphasic effect on coronary circulation characterized by an initial vasoconstriction up to 3 nm followed by a concentration-dependant increase. This latter effect is likely to be related to the increase in myocardial performances from a concentration of 3 nm. This well-known positive cardiac effect of norepinephrine36explains that, unlike AVP or terlipressin, its clinical use preserves cardiac output in patients.27These beneficial cardiac effects could be a serious advantage in maintaining the global oxygen transport, which can be significantly altered during clinical use of AVP.28The precise mechanism of the vasopressor effect induced by AVP remains unclear. AVP induces a contraction of the vascular smooth cells via  V^1avasopressin receptor.17However, the ATP-sensitive K⁺channels and nitric oxide both play a role in AVP-induced vasoconstriction.30,37In the current study, we found that intrinsic cardiac effects of AVP, like terlipressin, are principally mediated by the V^1avasopressin receptor. Nevertheless, we observed a residual effect induced by both drugs after V^1areceptor blockade. Our results are consistent with a previous in vitro  study17which reported residual negative cardiac side effects induced by AVP despite a pretreatment of 0.5 10⁻⁷M [d(CH²)⁵Tyr(Me)]AVP, a concentration similar to that used in the current study. These results confirm that several mechanisms are involved in AVP-mediated coronary vasoconstriction. Although the involvement of the vasopressin receptors in AVP-induced cardiac effects has been previously reported,17,37no previous study has reported the mechanisms involved in terlipressin-induced cardiac effects.

The following points must be considered in the assessment of the clinical relevance of our study. First, our findings have been obtained from an isolated rabbit heart preparation. Consequently, any clinical extrapolation must be made with caution. Second, these experiments were performed with constant coronary perfusion pressure. However, when vasopressor agents are used, systemic blood pressure and thus coronary perfusion pressure are increased. Third, although the perfusion medium did not affect the intrinsic cardiac effect of terlipressin, a metabolism by tissular peptidase could be completely abolished. Fourth, we observed a residual cardiac effect induced by AVP and terlipressin after the pretreatment by [d(CH²)⁵Tyr(Me)]AVP 0.1 μm. Although the involvement of other mechanisms could probably explain this residual effect, we did not perform a curve response to eliminate that a larger concentration of V^1aantagonist could completely abolish cardiac effects induced by vasopressinergic system agonists. Finally, these experiments were performed in a normal beating heart model with intact coronary circulation. Therefore, further investigations will be necessary to confirm our results during different pathologic states, such as septic shock, in which the studied vasopressors are usually employed.

In conclusion, terlipressin is not only a precursor drug but also exerts intrinsic cardiac effects on an isolated heart model only at supra-therapeutic concentrations (≥30nm). These effects included coronary vasoconstriction and, in turn, myocardial depression. As with AVP, the cardiac effects induced by terlipressin are principally mediated by V^1areceptors. However, in concentrations at the lower end of the clinically therapeutic range, terlipressin could have a better cardiac hemodynamic profile.

The authors thank David J. Baker, M.D., F.R.C.A. (Staff Anesthesiologist, Department of Anesthesiology, Centre Hospitalier Universitaire Necker-Enfants Malades, Paris, France), for reviewing the manuscript.