Comparison of Arginine Vasopressin, Terlipressin, or Epinephrine to Correct Hypotension in a Model of Anaphylactic Shock in Anesthetized Brown Norway Rats

This study, including care of the animals involved, was conducted according to the official edict presented by the French Ministry of Agriculture (Paris, France) and the recommendations of the Helsinki Declaration. Therefore, these experiments were conducted in an authorized laboratory under the supervision of an authorized researcher (P.M.M.). Because the ovalbumin-sensitized BN rat is a suitable model for inducing specific immunoglobulin E responses,9we used 10-week-old BN rats weighing 290–340 g (Janvier, Le Genest-St-Isle, France). They were kept under standard conditions (temperature: 21°± 1°C; lights from 6:00 am to 6:00 pm) and given a standardized diet (A04; UAR, Villemoisson-sur-Orge, France) and water (Aqua-clear; Culligan, Northbrook, IL) ad libitum . Rats were sensitized by subcutaneous administrations on days 0, 4, and 14 with grade VI chicken egg albumin (OVA, 1 mg; Sigma-Aldrich, Saint-Quentin Fallavier, France) and aluminum hydroxide (OHA¹, 3.5 mg; Merck Eurolab, Briare Le Canal, France) diluted in 1 ml saline solution (Chlorure de sodium, Cooper 0.9%; Melun, France).

The surgical procedure was performed during general anesthesia on day 21. Anesthesia was induced with 60 mg/kg intraperitoneal thiopentone sodium (Sanofi Santé animale, Libourne, France) and maintained with intravenous additional doses (2 mg/kg) when required. A fluid-filled polyethylene catheter (ID: 0.58 mm; OD: 0.96 mm; Biotrol Diagnostic, Chennevières Les Louvres, France) was inserted in the right common carotid artery for arterial pressure monitoring. Another fluid-filled catheter was inserted in the left external jugular vein for administration of drugs and fluid maintenance (10 ml · kg⁻¹· h⁻¹) with Ringer's solution (Braun Medical SA, Boulogne, France). The trachea was cannulated, and the lungs were mechanically ventilated with 100% oxygen using a Harvard Rodent respirator model 683 (Harvard Apparatus, Cambridge, MA).

Mean arterial pressure was recorded using a strain gauge pressure transducer (DA-100; Biopac Systems, Northborough, MA). The pressure transducer and the electrocardiogram were connected to a desktop computer for continuous data acquisition (Acqknowledge software and MP 100 hardware; Biopac Systems).

A flexible Clark-type polarographic oxygen electrode (diameter and length of the oxygen partial pressure-sensitive area of the probe: 0.5–0.6 mm and 1 mm, respectively) computer-supported Licox system (GMS, Mielkendorf, Germany) was introduced in one of the quadriceps muscles. For correction of tissue oxygen partial pressure (Ptio²) measurements, temperature within the muscle was monitored, and Ptio²values were adjusted to quadriceps temperature by means of the computer software Licox. The electrode was calibrated before and after each experiment with room air. The Licox system was connected to a desktop computer for continuous data acquisition (Acqknowledge). Hemodynamic values were allowed to stabilize for 30 min (stabilization period) after surgery, and pre-anaphylactic shock values were recorded before shock induction. In each group, anaphylactic shock was induced by injecting 1 mg ovalbumin diluted in 250 μl saline solution intravenously in 1 min. Time 0 (T0) corresponds to the beginning of ovalbumin injection.

In a first series of rats, we compared ovalbumin-sensitized BN rats randomly allocated to the anaphylactic shock or to the nicardipine group in which hypotension of amplitude similar to that of the anaphylactic shock group was obtained by injecting 2 μg/100 g (by weight) nicardipine (Novartis Pharma SA, Rueil-Malmaison, France) intravenously in 1 min followed by a continuous intravenous infusion of nicardipine (500 μg · 100 g⁻¹· h⁻¹), as previously done.8At the end of the experiments (1 h after shock induction), 3 ml blood was withdrawn from the arterial catheter, and endogenous plasma AVP measurements were performed with a standard radioimmunologic method.10A polyclonal antivasopressin antibody and a radioactivated vasopressin tracer (RIA; Amersham Biosciences, Orsay, France) were used. The lower limit of detection for this assay in our laboratory is 1.2 pg/ml. The intraassay and interassay variation coefficients were of 9.4% and 19.2%, respectively.

Animals were randomly allocated into three groups (epinephrine, AVP, and terlipressin) according to a computer-generated randomization list. In each animal, post-anaphylactic shock values for MAP, heart rate, and Ptio²were recorded at T0 + 5 min corresponding to the time when MAP decreased at 40% of pre-anaphylactic shock values as previously described8(table 1). The first injection of the different vasoconstrictors was performed at T0 + 5 min (fig. 1). In each animal, a dose-response relation was obtained by injecting increasing doses (intravenous bolus in saline solution, 2 ml/kg) of seven doses of epinephrine (doses 1–7: 0.3–300 μg/kg, Adrénaline; Aguettant, Lyon, France), eight doses of AVP (doses 1–8: 0.03–100 U/kg, (Arginine8)-Vasopressin Grade VI Solution APP; Sigma-Aldrich), or seven doses of terlipressin (doses 1–7: 0.03–30 μg/kg, Glypressine; kindly provided by Ferring, Gentilly, France). For each group, maximal MAP values obtained after each injection were averaged, then plotted (y-axis) against log dose (x-axis). Maximal MAP effect (Emax MAP) and ED⁵⁰were determined and expressed as mean and 95% confidence interval of the mean by using the four-parameter logistic equation according to the Hill model: Y = minimum + (Emax − minimum)/(1 + 10^{(dose-logED50)}), where minimum equals return to the hypotension level (Prism 3 algorithm for statistical analysis; GraphPad Software, San Diego, CA). For each rat, the following measurements were also performed: time interval in seconds between the end of the injection dose and (1) MAP maximal effect (peak) and (2) time required to return to hypotension values (recovery). Durations (in minutes) of the different experiments were recorded. At the end of the experiments, animals were killed by an overdose of thiopentone.

Results are expressed as mean ± SD for convenience, although several statistical tests were nonparametric. Endogenous plasma AVP measurements of anaphylactic shock and nicardipine groups were compared using a Mann-Whitney test. Intragroup and intergroup differences in epinephrine, AVP, or terlipressin groups were tested by one-way and two-way analysis of variance for repeated measures (Statview; SAS Institute Inc., Cary, NC). When a significant interaction was observed with two-way analysis of variance, the mean values were compared with the postanaphylactic shock values by the Fisher exact test. The slopes of the MAP versus  Ptio²value curves were compared using a Wilcoxon test. The criterion of significance was P < 0.05 (two tailed).

Two series of experiments involving 45 ovalbumin-sensitized BN rats were performed. Endogenous plasma AVP concentrations were measured in a first series of rats (290 ± 16 g) randomly allocated to the anaphylactic shock (n = 12) or the nicardipine group (n = 12). Blood samples for AVP measurements were withdrawn 60 min after the induction of the anaphylactic shock or of the pharmacologic hypotension. A second series of 21 rats (317 ± 32 g) were randomly allocated to the epinephrine group (n = 7), the AVP group (n = 7), or the terlipressin group (n = 7). Endogenous plasma AVP concentrations were significantly lower (P < 0.05) in the anaphylactic shock group than in the nicardipine group: 57 ± 26 versus  91 ± 43 pg/ml, respectively.

Pre-anaphylactic shock values (MAP, heart rate, Ptio²) were not significantly different among the three groups (T0 − 5; table 1). Five minutes after induction of anaphylactic shock (post-anaphylactic shock values equivalent to mean values before the first dose of drug injection), a profound and similar reduction in MAP was observed in the three groups (table 1). The duration of the experiments was similar in the epinephrine (38 ± 22 min) and AVP (38 ± 20 min) groups, whereas it was significantly shorter (11 ± 5 min; P < 0.05) in the terlipressin group.

Dose-response curves for epinephrine, AVP, and terlipressin versus  MAP values are presented in figure 2. Injections of epinephrine and AVP induced dose-dependent increases in MAP values. Epinephrine injection resulted in higher Emax MAP values as compared with AVP (P < 0.05). ED⁵⁰was 10.6 μg/kg for epinephrine and 4.1 U/kg for AVP (table 2). Injections of terlipressin did not change MAP values, regardless of the dose used. We had previously performed experiments (n = 5) that revealed that terlipressin increased MAP values in nonshocked, ovalbumin-sensitized, anesthetized BN rats. Increases of 18%, 27%, and 29% as compared with the MAP baseline values were respectively obtained within 2–3 minutes after a terlipressin dose of 1 μg/kg and within 1 min after doses of 3 and 10 μg/kg (data not shown). Heart rate values for the epinephrine, AVP, and terlipressin groups before and after each dose injection are presented in table 3. Heart rate values before and after the highest two doses of epinephrine were significantly increased when compared with the post-anaphylactic shock values (P < 0.05). When high doses of epinephrine were injected, there were 3 out of 7 rats that presented episodes characterized by prolonged ventricular depolarization (QRS) complexes that lasted for approximately 50 s, which were reversible and consistent with episodes of ventricular tachycardia. These episodes of arrhythmia occurred at high doses of epinephrine (> 100 μg/kg) when MAP values were higher than 170 mmHg. In contrast, in rats that received AVP, there were no such episodes.

The time to peak MAP values and the time to return to hypotensive values for the three groups are presented in table 4. There were significant time-dependent increases in both the delay to peak and the delay to return to arterial hypotension values for both the epinephrine and AVP groups. The time to peak effect increased significantly after the seventh dose (300 μg/kg) in the epinephrine group (P < 0.05) and after the fourth dose (1 U/kg) in the AVP group (P < 0.05). The time to recovery increased dose dependently and was significantly longer in the epinephrine group (P < 0.05) after the seventh dose (300 μg/kg) and in the AVP group after the sixth dose (10 U/kg).

The Ptio²profiles for the three groups are presented in figure 3. Ptio²values decreased significantly (P < 0.05) after induction of anaphylactic shock in the three groups (table 1). Ptio²decreased after the first injection in the AVP group and after the second injection in the epinephrine and terlipressin groups (P < 0.05 vs.  Ptio²measured before injection of the first dose of each drug). Ptio²decreased rapidly in the terlipressin group, whereas MAP values did not change. The slopes of the MAP versus  Ptio²values were significantly different (P < 0.05) between the epinephrine and AVP groups and for the epinephrine and AVP groups versus  the terlipressin group.

Ptio²values measured at MAP values (below 120 mmHg) observed after injections of 1–10 μg/kg epinephrine or after injections of 1–10 U/kg AVP were not significantly different between groups, whereas they remained higher (P < 0.05) in the epinephrine group at MAP values obtained after injections of 30 μg/kg epinephrine or 30–100 U/kg AVP.

The main findings of this study were that (1) anaphylactic shock occurring in anesthetized BN rats, as compared with a model of nicardipine-induced arterial hypotension of similar magnitude,8was characterized by significantly lower concentrations of endogenous plasma AVP; (2) the effects of exogenous AVP and epinephrine on MAP, heart rate, and skeletal muscle Ptio²values were comparable at low doses and differed for the highest doses; (3) the ED⁵⁰values for epinephrine and AVP were 10.6 μg/kg and 4.1 U/kg, respectively; and (4) terlipressin, even at very high doses, did not increase MAP.

Epinephrine is the first-line vasopressor to treat cardiovascular collapse in most guidelines on perioperative management of anaphylaxis. ††2,4However, epinephrine is not always successful, and situations where epinephrine does not restore cardiovascular homeostasis are considered as “epinephrine resistant,” although no clear criteria define this term. In other situations of catecholamine resistance (i.e. , in septic shock), exogenous AVP and terlipressin have been proposed and used.5,6,11–14 

Part of the rationale for the use of exogenous AVP in several shock states is based on the observation that under those conditions, endogenous plasma AVP concentrations are inadequately low.5,6In the current study, we demonstrate that during the short period of time of this experimental anaphylactic shock, endogenous plasma AVP concentrations are significantly lower as compared with a nicardipine-induced model of arterial hypotension of similar profile. In response to arterial hypotension during hemorrhage or severe sepsis, endogenous AVP is released from the neurohypophysis, and its concentration in plasma increases (20- to 200-fold).15As shock worsens, the initially high plasma concentrations of endogenous AVP decrease. This was reported both in animal models during hemorrhagic shock16and in patients with septic shock5,7or presenting with arterial hypotension refractory to catecholamines after left ventricular assist device implantation.17,18Different mechanisms have been proposed to explain the late decrease of endogenous plasma AVP concentrations, such as (1) depletion of neurohypophyseal stores of AVP in advanced shock, (2) impaired baroreceptor-mediated AVP secretion,5and (3) increased norepinephrine levels responsible for central inhibitory effect on AVP release.19A causal role for a relative AVP deficiency in several forms of shock is suggested by studies that reported a correlation between the endogenous plasma AVP concentrations and the magnitude of arterial hypotension5,7and between the endogenous plasma AVP concentrations and survival after resuscitation of cardiopulmonary arrest.20 

Because measurement of endogenous plasma AVP concentrations requires 3–4 ml blood per sample withdrawn, whereas the total blood volume in our rat model is approximately 18–20 ml, we could not perform a time-dependent profile of endogenous plasma AVP during anaphylactic shock. Therefore, further studies are needed to investigate whether lower endogenous plasma AVP concentrations in anaphylactic shock animals are preceded by a transient increase. A likely cause of decreased endogenous plasma AVP concentrations could be the increase of endogenous norepinephrine concentrations.19We previously demonstrated that a stronger sympathetic response was observed in the anaphylactic shock group as compared with a nicardipine-induced arterial hypotension of similar amplitude.8This could support the mechanism proposed by Day et al.  19Nevertheless, we cannot exclude the fact that our results could be related to the choice of nicardipine as the hypotensive drug, and therefore, our results require confirmation with different hypotensive drugs before the profile of endogenous AVP during anaphylactic shock can be clearly defined.

The cardiovascular effects of epinephrine, AVP, and terlipressin are complex,21but in the context of resuscitation of anaphylactic shock occurring during anesthesia, the therapeutic target is restoration of flow and function of vital organs. In the clinical context, MAP values are an acceptable surrogate because they can be measured with minimum effort. Therefore, the current study focused on MAP values to construct dose-response curves. The heart rate response was investigated because it is one of the determinants of left ventricular minute work and its increase is associated with decreased diastolic time, possibly impairing left ventricular myocardial perfusion. Because we recently demonstrated that anaphylactic shock in anesthetized BN rats is associated with intense vasoconstriction and rapidly decreased Ptio²values in skeletal muscle,8we investigated this compartment, which rapidly responds to changes in the central and peripheral circulation, to compare the effects of the three drugs on the oxygen delivery-oxygen consumption balance.

Both exogenous AVP and epinephrine injections resulted in increased MAP values, and for both drugs, there was a wide 95% confidence interval. When investigating the time to peak effect on MAP and the delay to return to arterial hypotension after repeated bolus injection, we observed that for both drugs, at higher doses there was a statistically significant increase in time to peak effect. The design of our experiments does not allow us to identify the causes of the increase of the time to peak effect. In addition, for the lowest doses of the two drugs, the delays to return to arterial hypotension were short, thus suggesting the interest for both drugs of a continuous infusion.

At the highest doses, there were differences in the hemodynamic effects of epinephrine and exogenous AVP. These differences are as follows: (1) MAP values were significantly higher with epinephrine as compared with AVP; (2) episodes of ventricular arrhythmia occurred with epinephrine, a proarrhythmogenic effect that was not observed in animals treated with AVP; and (3) AVP injection was associated, at higher doses (30 and 100 U/kg), with lower Ptio²values (P < 0.05) as compared with epinephrine (30 μg/kg) at equivalent MAP values. When MAP values after either epinephrine or AVP was in the range of 80–120 mmHg, Ptio²values were not significantly different between the two groups.

The interpretation of the skeletal muscle Ptio²values in shock states is complex. In a previous study of anaphylactic shock in anesthetized rats, we associated measurements of skeletal muscle Ptio²values with measurements of interstitial lactate and pyruvate concentrations. We demonstrated that the most likely interpretation of low Ptio²values and increased lactate/pyruvate ratios in this model of anaphylactic shock was rapid failure of energy production in the skeletal muscle.8We interpret the lower Ptio²values in rats resuscitated with high doses of AVP as the effect of excessive vasoconstriction, increased oxygen consumption, or both, and we consider the lower Ptio²values as a potential deleterious effect of AVP. Nevertheless, AVP was associated with significantly lower Ptio²values only at the highest doses, which will probably never be used in clinical practice. However, at lower AVP doses, for similar increased on MAP values, AVP and epinephrine resulted in comparable Ptio²values.

Terlipressin is a synthetic derivative of vasopressin. It is a prodrug that is converted to lysine vasopressin in the circulation after the N-triglycyl residue is cleaved by endothelial peptidases.21This results in a rapid and sustained release of the structural analog lysine vasopressin. The reason for which terlipressin was investigated in the current study is that despite differences with AVP for its hemodynamic effects, terlipressin has been used as an alternative to exogenous AVP, especially in France, where AVP is not commercially available.12,22–25Unexpectedly, terlipressin, which has been reported to be effective in catecholamine-resistant septic shock in adults12,24,26,27and in children,28did not increase MAP values during anaphylactic shock, regardless of the dose injected. The lack of increase of MAP in terlipressin-treated animals explains the shorter duration of the experiments as compared with the two other groups. The explanations for this lack of MAP effect are not clear. In experimental studies, terlipressin significantly increased MAP values without decreasing aortic blood flow in fluid-challenged endotoxic rats but not in hypodynamic rats29or in an endotoxemic sheep,30where MAP increase was accompanied by a progressive decrease in cardiac index. The authors concluded that it might therefore be advantageous to infuse terlipressin gradually instead of administering it as a bolus.30In an isolated rabbit heart model, terlipressin induced coronary blood flow reduction and, consequently, myocardial depression, only at supratherapeutic concentrations (≥ 30 nm).31The lack of effect of terlipressin is not strain related because, as stated in the Results section, we could observe an increase in MAP values after terlipressin infusion in the absence of anaphylactic shock. Therefore, we propose two possible explanations for the lack of effect of terlipressin in this experimental model: (1) hypovolemia occurring after onset of anaphylactic shock and (2) a deleterious effect of terlipressin bolus infusion on cardiac output.

The Ptio²profile in rats injected with terlipressin is unexpected. In the absence of an increase of MAP values upon terlipressin injection, one would have expected a Ptio²profile comparable to that observed in rats with unresuscitated anaphylactic shock.8Nevertheless, as compared with a nonresuscitated model, terlipressin injections resulted in a faster decrease of skeletal muscle Ptio². One possible explanation of these results could have been increased skeletal muscle oxygen consumption (Vo²) in the absence of an increased blood flow and oxygen transport. This hypothesis is not supported by several published articles that showed that terlipressin-induced a decrease of global Vo²in an endotoxemic sheep30or pig model32and in patients treated with renin-angiotensin system inhibitors receiving terlipressin to counteract the anesthesia-induced hypotension.25Therefore, the mechanisms of the rapidly decreasing Ptio²values in anaphylactic shock rats injected with terlipressin require further studies.

Our results provide experimental arguments for the possible use of AVP in humans with anaphylactic shock, as already reported by several authors.33–35Another experimental argument favoring the use of exogenous AVP, in addition to the lower endogenous plasma AVP concentrations demonstrated in our study, could be the fact that an in vitro  study showed that whereas epinephrine only partially reversed histamine-induced vasodilatation in human internal mammary arteries, vasopressin leads to a complete reversal of the vasculature relaxation.36 

The ED⁵⁰value calculated for AVP in our study was approximately 4 U/kg. A possible explanation of these ED⁵⁰values calculated in the current experiments may be due to the use of cumulative doses. Hiruta et al.  37reported that the highest increase in MAP values in a rabbit model of anaphylactic shock was observed with 0.8 U/kg AVP, but the highest survival was observed with 0.08 U/kg AVP. The detrimental effect of the higher AVP dose was assumed by the authors to be related to decreased AVP-induced myocardial function. Taken together, the results of Hiruta et al.  37are consistent with our observations in that lower doses of AVP seem to have the most favorable results in the experimental setting.

Arguments against the use of exogenous AVP in anaphylactic shock, but also in other forms of shock, could be due to excessive vasoconstriction in the splanchnic circulation38as well as other territories.39,40Indeed, in the current experiments, high doses of exogenous AVP that will probably never be used in clinical practice did decrease Ptio²values in skeletal muscle. Nevertheless, for lower doses, there were no differences for skeletal muscle Ptio²values between AVP and epinephrine. Whether this observation can be extrapolated to other organs in rats or to patients with comorbidities is not known. Our results should be interpreted by taking into account the species in which they were performed, the model of induction of anaphylaxis (ovalbumin), and also the model that was used to induced the arterial hypotension in the control group, i.e. , the injection of nicardipine.

In conclusion, we provide information on dose-response curves for epinephrine and AVP and demonstrate that in this experimental model, AVP, but not terlipressin, may restore blood pressure during anaphylactic shock, corroborating recent clinical reports.33–35Further work is needed to study the optimal role of AVP in treatment of patients experiencing anaphylactic shock.

The authors thank Pierre Mairose, Viviane Camaeti, and Maryline Schwartz (Technicians, Laboratoire d'Explorations Fonctionnelles Métaboliques et Endocriniennes, Centre Hospitalier Universitaire Hôpital de Brabois, Vandoeuvre-lès-Nancy, France) for their technical assistance.