Hemorrhage during Isoflurane–Nitrous Oxide Anesthesia: Effects of Endothelin-A or Angiotensin II Receptor Blockade or Both

Permission to perform the experiments was obtained from the Governmental Animal Protection Committee (AZ G0424/99, Berlin, Germany) and conforms with the Guidelines on the Care and Use of Laboratory Animals . 7 

A total of 24 experiments were performed in six purebred female Beagle dogs (body weight, 13 ± 0.5 kg), with four experiments in each dog. The dogs were kept under highly standardized conditions and received a standardized dietary regimen as described previously. 8 

Eight days before an experiment, 100 ml of each dog’s own blood was collected via  puncture of a foreleg vein and stored in a blood bag at 4°C (Biopack®; Biotrans, Dreieich, Germany). The blood served to replace the blood withdrawn for analysis during the experiments. The interval between experiments in the same dog was at least 14 days.

Each of the six dogs underwent four protocols in randomized order:

1. control experiments (controls)

2. ET^Areceptor blockade (ET^AB)

3. angiotensin II receptor blockade (AT¹B)

4. angiotensin II plus ET^Areceptor blockade (AT¹B + ET^AB)

After a baseline period of 30 min in the awake dogs (baseline awake), anesthesia was induced, and the dogs were observed for 60 min during isoflurane-nitrous oxide anesthesia (baseline anesthesia; for details, see Procedures during the Experiments). Thereafter, the dogs had 20 ml blood/kg body weight withdrawn (equal to 25% of the total blood volume) within 5 min and were studied for another 60 min (hemorrhage), after which time the shed blood—which had meanwhile been stored in a blood bag—was rapidly retransfused within 5 min. After retransfusion, the dogs were observed for a final 60-min period (retransfusion).

General procedures were the same as in protocol 1. In addition, a selective ET^Areceptor blocker (ABT-627; Abbott Laboratories, North Chicago, IL) was administered (intravenous bolus of 1 mg/kg body weight 30 min before the start of the experiment, followed by a continuous infusion of 0.01 μg · kg body weight⁻¹· min⁻¹). This ABT-627 infusion regimen has also been established in our laboratory and has been used in other experiments. 6 

General procedures were the same as in protocol 1. In addition, a selective AT¹receptor blocker (losartan; Merck Research Laboratories, Rahway, NJ) was continuously infused (0.1 μg · kg body weight⁻¹· min⁻¹). The infusion was started 30 min before the baseline awake period. The losartan infusion regimen has been established in our laboratory and has been applied in other experiments. 8Adequate AT¹receptor inhibition was tested at the end of the experiment by finding no blood pressure increase after an intravenous bolus injection of 1 μg angiotensin II.

General procedures were the same as in protocol 1, but this time, AT¹plus ET^Areceptor blockers were coinfused in the amounts and concentrations as described in protocols 2 and 3.

Preparation of the dogs started at 7:30 am. Body weight and temperature were recorded. The urinary bladder was catheterized with a self-retaining Foley catheter. A foreleg vein was punctured, and an infusion of creatinine was started (priming dose, 1.4 g for 30 min; maintenance infusion, 4.7 mg/min) for assessment of glomerular filtration rate (GFR; exogenous creatinine clearance). 9After local anesthesia (1% lidocaine; Braun, Melsungen, Germany), an arterial line (20 gauge, No. 4235-8; Ohmeda, Erlangen, Germany) was advanced into the abdominal aorta via  the femoral artery. Also after local anesthesia, a percutaneous sheath (Arrow-Flex, No. SI-09600; Arrow, Erding, Germany) was inserted into the right and left external jugular veins: one for the hemorrhage procedure and one for the insertion of a pulmonary artery catheter (5 French, No. 132F5; Baxter, Unterschleissheim, Germany). Thereafter, the conscious dogs were given approximately 30 min to adjust to the experimental situation, followed by a 30-min observation period (baseline awake). Thereafter, anesthesia was induced with 10 mg/kg intravenous propofol (Disoprivane® 1% lipid emulsion; AstraZeneca GmbH, Wedel, Germany). After induction, the dogs were intubated with a cuffed endotracheal tube (36 gauge) and mechanically ventilated (Servo Ventilator 900D; Siemens Company, Lund, Sweden). The respiratory rate was set to 12 breaths/min, the positive end-expiratory pressure was set to 5 mmHg, and the tidal volume was adjusted to keep the arterial carbon dioxide tension between 35 and 40 mmHg. Inhalation anesthesia was adjusted to maintain an end-tidal isoflurane concentration of 1.2 vol% in a 30% oxygen-70% nitrous oxide mixture. This is equal to a minimum anesthetic concentration (MAC) of 1.3 in dogs. This results because 1 MAC is 1.28 vol% for isoflurane and 188 vol% for nitrous oxide in dogs. The two MAC values are additive. 10 

Heart rate, MAP, and central venous pressure were measured continuously, and data were stored on a computer. CO was determined by using the thermodilution technique (5-ml injection volume at 5–10°C, Vigilance; Baxter Edwards Critical Care, Unterschleissheim, Germany). Five consecutive measurements were performed. The highest and lowest values were rejected. The mean CO was calculated from the three remaining determinations and used for calculation of systemic vascular resistance by the standard formula.

Blood samples were taken to determine plasma hormones (see Measurement of Plasma Values), creatinine, arterial and mixed venous blood gases, actual bicarbonate, and base excess at the following time points: the end of the awake period, 60 min after anesthesia, 5 and 60 min after hemorrhage, and 60 min after retransfusion. The blood withdrawn for analysis was immediately replaced with an equal amount of each dog’s own blood precollected 1 week before the experiment.

At the end of each period, urine flow and creatinine excretion were measured after complete evacuation of the urinary bladder. Exogenous creatinine clearance was calculated by the standard formula to assess GFR.

Plasma creatinine concentration was determined with a creatinine analyzer (modified Jaffé reaction; Beckman Instruments, Brea, CA), blood gas analyses were performed with the ABL 505 analyzer (Radiometer, Copenhagen, Denmark).

Blood samples for hormone measurements were placed into precooled Na-EDTA vials and centrifuged at 4°C immediately. The separated plasma was stored at −20°C until analysis. The hormones were determined by commercially available kits: plasma renin activity (New England Nuclear, North Billerica, MA), angiotensin II concentration (Eurodiagnostika, Arnhem, The Netherlands), plasma aldosterone concentration (Biomedica, ALDO CTK-2R; Sorin Company, Saluggia, Italy), vasopressin (Vasopressin-RIA; Biermann GmbH, Bad Naunheim, Germany), plasma endothelin (Endo; Biomedica, Vienna, Austria). Epinephrine and norepinephrine were determined with high-performance liquid chromatography (Chromosystems, München, Germany).

All values are presented as mean ± SEM (n = 6). For intragroup comparison (time course), a general linear model of analysis of variance for repeated measures was applied (SPSS 10, Chicago, IL). Intergroup comparison was performed using the Student t  test. The level of significance for the error of first order was adjusted according to the Holm procedure. 11Statistical significance was assumed at P < 0.05.

In controls, MAP decreased from 86 ± 3 in the awake state to 65 ± 2 mmHg during anesthesia. Five minutes after hemorrhage, MAP decreased to 45 ± 3 mmHg (P < 0.05), and it increased spontaneously to 56 ± 2 mmHg within the next 60 min. Retransfusion of the shed blood increased MAP to 74 ± 3 mmHg. In the ET^AB protocol, the decrease in MAP during anesthesia was comparable with controls, but it was more severe 60 min after hemorrhage (49 ± 3 mmHg; P < 0.05). After retransfusion, the increase in MAP was less than in controls (59 ± 3 mmHg; P < 0.05;fig. 1). In the AT¹B protocol, MAP decreased to 54 ± 3 mmHg during anesthesia and decreased to 41 ± 2 mmHg 60 min after hemorrhage. Sixty minutes after retransfusion, MAP reached only 51 ± 4 mmHg (P < 0.05;fig. 1). In the AT¹B + ET^AB protocol, the decrease in MAP was most severe, being 49 ± 5 mmHg during anesthesia, 36 ± 2 mmHg 60 min after hemorrhage, and only 43 ± 2 mmHg 60 min after retransfusion (P < 0.05;fig. 1).

Cardiac output in the awake state was 2.1–2.2 l/min in all protocols and decreased to 1.5–1.6 l/min during anesthesia (P < 0.05;fig. 2). After hemorrhage, CO decreased less in controls (1.3 ± 0.1 l/min) than in the AT¹B, ET^AB, and AT¹B + ET^AB protocols (1.0–1.1 l/min; P < 0.05;fig. 2). Sixty minutes after retransfusion, CO increased, less in the ET^AB than in the other protocols (fig. 2).

Heart rate in the awake state was similar in all protocols (88–93 beats/min) and did not change during anesthesia (table 1). Heart rate increased by approximately 10 beats/min 60 min after hemorrhage in all protocols and increased further after retransfusion (P < 0.05;table 1).

Central venous pressure was similar in all protocols. Only 5 min after hemorrhage, the central venous pressure was decreased (P < 0.05), whereas it had returned to prehemorrhage values 60 min after hemorrhage (table 1).

Systemic vascular resistance in controls and in the ET^AB protocol was always greater during anesthesia and after hemorrhage and retransfusion than in the AT¹B and AT¹B + ET^AB protocols (P < 0.05;table 1).

Plasma norepinephrine concentrations were similar in all protocols during the awake state and during baseline isoflurane-nitrous oxide anesthesia. Norepinephrine increased after hemorrhage and decreased after retransfusion in all protocols (P < 0.05;table 2).

Plasma epinephrine concentrations were comparable in all protocols during the awake state (70–157 pg/ml) and did not change during the baseline anesthesia period. After hemorrhage, epinephrine concentrations increased to 827–1,296 pg/ml (table 2) and decreased in all protocols after retransfusion (P < 0.05), remaining above the levels observed during anesthesia alone in all protocols but controls.

Plasma renin activity in the awake state was higher in AT¹B and AT¹B + ET^AB protocols than in controls (P < 0.05;table 2). During anesthesia, plasma renin activity increased in all protocols (P < 0.05). After hemorrhage, plasma renin activity increased further only in controls and in the ET^AB protocol (P < 0.05;table 2).

Plasma angiotensin II concentrations were higher in AT¹B and AT¹B + ET^AB protocols than in controls and the ET^AB protocol in the awake state (table 2). Angiotensin II increased during anesthesia, especially in the AT¹B and AT¹B + ET^AB protocols (P < 0.05;table 2). Angiotensin II increased after hemorrhage in controls and the ET^AB protocol (P < 0.05;table 2) but not in the AT¹B and AT¹B + ET^AB protocols (table 2). After retransfusion, angiotensin II decreased slightly in all protocols.

Plasma aldosterone concentrations were lower in AT¹B and AT¹B + ET^AB protocols during the awake state. Aldosterone increased during anesthesia and 60 min after hemorrhage in all protocols (P < 0.05;table 2) and decreased slightly after blood retransfusion (table 2).

Vasopressin plasma concentrations were low during the awake state (0.5–0.6 pg/ml) and increased during anesthesia (14–27 pg/ml) and after hemorrhage (P < 0.05), being lowest in controls (162 ± 29 pg/ml) and highest in ET^AB (280 ± 62 pg/ml) (table 2). After retransfusion, vasopressin returned almost to baseline levels (P < 0.05;table 2).

Plasma endothelin 1 concentrations were higher in ET^AB and AT¹B + ET^AB protocols compared with control and AT¹B protocols (P < 0.05;table 2). After hemorrhage, the plasma endothelin 1 concentration increased in all protocols (P < 0.05;table 2). Even 60 min after blood retransfusion, the endothelin 1 concentration remained increased in the range of values found after hemorrhage.

Arterial oxygen tension increased after induction of anesthesia (fraction of inspired oxygen = 0.3; P < 0.05) but did not change throughout the experiments in all protocols. Arterial carbon dioxide tensions were similar in all protocols and did not change during the experiments. Mixed venous oxygen tension decreased during hemorrhage in all protocols and increased to anesthesia baseline levels after retransfusion (P < 0.05;table 3). Mixed venous carbon dioxide tension increased after hemorrhage in all protocols (P < 0.05) and decreased to, or even below, anesthesia baseline levels after retransfusion.

Plasma osmolarity was not different between awake (299–300 mOsm) and anesthetized dogs (298–300 mOsm). After hemorrhage, however, plasma osmolarity decreased in all protocols (294–296 mOsm; P < 0.05) and remained in this range also after retransfusion.

Urine flow ranged between 25 and 98 μl · min⁻¹· kg⁻¹in the awake state, being greatest in the protocols in which losartan had been used. Urine flow decreased during anesthesia (P < 0.05;table 4) and was almost zero during hemorrhage. After retransfusion, urine flow increased above baseline anesthesia levels in control and ET^AB dogs (P < 0.05;table 4), whereas in AT¹B and AT¹B + ET^AB dogs, it returned to baseline anesthesia levels.

Glomerular filtration rates were similar in all protocols during the awake state. During anesthesia, the decrease in GFR was most pronounced in the AT¹B and AT¹B + ET^AB protocols (P < 0.05;table 4). After hemorrhage, GFR was close to zero in the AT¹B and AT¹B + ET^AB protocols (P < 0.05;table 4). After retransfusion, GFR increased above baseline anesthesia values in ET^AB and controls, whereas it remained low in AT¹B and AT¹B + ET^AB dogs (table 4).

The current study investigated whether ET^Areceptor antagonism impairs the integrated cardiovascular, hormonal, and renal responses to acute hemorrhage in dogs anesthetized with isoflurane-nitrous oxide, compared with controls and with selective AT¹antagonism. The anesthetic gas mixture consisted of isoflurane (1.2 vol% end-tidal concentration) in oxygen-nitrous oxide (30/70 vol%), which is equal to approximately 1.3 MAC in dogs. 10Thus, the following results were obtained during moderately deep anesthesia.

In the following discussion, the effects of isoflurane-nitrous oxide anesthesia on hemodynamics and hormones in controls, ET^A, and AT¹receptor-inhibited dogs will be addressed first, followed by the discussion of the effects of hemorrhage and, finally, of the blood retransfusion.

In our control experiments, isoflurane-nitrous oxide anesthesia decreased CO by 28% and mean arterial blood pressure by 25% (figs. 1 and 2), leaving systemic vascular resistance almost unchanged (table 1). This is in line with previously published data in dogs. 12Heart rate increased (table 1) because of the vagolytic effect of isoflurane. 13In ET^A-inhibited dogs, MAP and CO were not different from controls, during the awake state as well as during isoflurane-nitrous oxide anesthesia alone (figs. 1 and 2). This supports the notion that ET¹is not substantially involved in the short-term blood pressure and CO regulation in normotensive dogs. 6,14In contrast, AT¹receptor inhibition caused a greater decrease in MAP (−41%) during isoflurane-nitrous oxide anesthesia than in controls (−25%). This emphasizes the importance of angiotensin II in short-term blood pressure regulation.

After induction of isoflurane-nitrous oxide anesthesia, activation of the renin-angiotensin-aldosterone system and an increase in vasopressin concentrations were unanimous findings in all our protocols (table 2). The activation of the renin-angiotensin-aldosterone system is most likely due to the decrease in MAP and thus renal perfusion pressure during isoflurane-nitrous oxide anesthesia, activating the renal baroreceptor mechanism. 15Stimulation of renal sympathetic nerves may also be involved. 16The increase in vasopressin secretion during isoflurane-nitrous oxide anesthesia is probably also due to the decrease in MAP, 17because other factors that might stimulate vasopressin release, such as plasma osmolarity 18and acid-base status 19(table 3), remained unaffected by isoflurane-nitrous oxide anesthesia per se  in all protocols.

Acute withdrawal of approximately 25% of blood volume decreased MAP and CO (figs. 1 and 2) and increased systemic vascular resistance in both controls and ET^A-inhibited dogs (5 min after hemorrhage;table 1). However, the increase in endogenous vasopressors (vasopressin, angiotensin II, epinephrine) was more pronounced during ET^Areceptor inhibition compared with controls (table 2). A direct, stimulating effect of endothelins on vasopressin release is possible 20but seems unlikely in our experiments because the ET¹plasma concentration increased only slightly after hemorrhage. It is possible that ET^Ainhibition induced an additive vasodilative effect via  ET^Breceptors 21—in addition to that caused by isoflurane anesthesia per se , which was counteracted by the more pronounced increase of vasopressors after hemorrhage. During AT¹receptor inhibition, the decrease in MAP after hemorrhage was greater than in controls or ET^areceptor-inhibited dogs, although there was no additive effect of simultaneous ET^Areceptor inhibition (fig. 1), whereas the decrease in CO, the decrease in central venous pressure, and the increase in heart rate (table 1) after hemorrhage were comparable with controls. All endogenous compensatory mechanisms, such as the increases in plasma epinephrine, norepinephrine, vasopressin, aldosterone, sympathetic traffic were insufficient to return MAP into the range of controls. Therefore, other than in ET^Areceptor-antagonized individuals, the anesthesiologist must be prepared to quickly and continuously replace blood loss or to exogenously support the endogenous compensatory vasoregulatory mechanisms or both in AT¹receptor-inhibited individuals.

Mean arterial pressure and CO partly recovered within 60 min after blood loss in controls. This recovery was paralleled by slow increases of plasma norepinephrine, endothelin, and aldosterone concentrations (“late” hormonal regulation;table 2), and was accompanied by an increase in heart rate. Unlike controls, spontaneous recovery of MAP and CO 60 min after blood loss was markedly blunted in ET^A-antagonized dogs (figs. 1 and 2), suggesting that ET^Areceptor inhibition impairs long-term rather than short-term hemodynamic regulation.

After retransfusion of the shed blood, the concentrations of all vasoconstrictive hormones decreased and MAP returned to baseline anesthesia levels in controls but not with AT¹and ET^Areceptor inhibition (fig. 1). The relative increase in CO was even greater than for MAP, returning CO almost to awake levels in all protocols. The increase in CO can mainly be attributed to the further increase in heart rate after blood retransfusion (fig. 2and table 1). This increase in heart rate probably ensued from the swift (5-min) blood retransfusion into a constricted vascular system. The phenomenon of hypervolemia-induced tachycardia was first described by Bainbridge 22(Bainbridge reflex) in anesthetized dogs. The Bainbridge reflex increases CO in response to an increased venous return to the heart, thereby reducing the elevation of cardiac preload. After retransfusion, the increase in CO may also serve to overcome the slight metabolic acidosis that developed during the hemorrhage period (table 3). In ET^A-inhibited dogs, blood retransfusion did restore CO but not MAP to prehemorrhage values (figs. 1 and 2). This indicates that endothelins are essential in the long-term regulation of MAP and for MAP restoration after blood retransfusion in anesthetized dogs (fig. 2). However, compared with ET^A-inhibited dogs, blood pressure regulation was much more compromised during both AT¹inhibition protocols, suggesting the importance of angiotensin II in both short- and long-term responses to hypovolemic shock. The smaller decrease in epinephrine concentrations after blood retransfusion in both AT¹inhibition protocols (table 2) may indicate continued endogenous sympathetic support.

Anesthesia reduced urine flow markedly in all protocols (table 4), probably because of a decrease in MAP combined with the stimulation of water- and sodium-retaining hormones (angiotensin II, vasopressin, aldosterone). GFR decreased in all protocols during anesthesia, but the decrease was most severe with AT¹receptor inhibition (table 4). This can probably be explained by the decrease of MAP and renal perfusion pressure below the threshold pressure for renin release of approximately 70–80 mmHg in AT¹-antagonized dogs. Below this pressure, the kidney starts to release renin that induces angiotensin II production, which normally constricts the efferent arteriole of the glomerulus to prevent a decrease in GFR. AT¹receptor inhibition impairs this response because it impairs the angiotensin II-mediated constriction of the efferent glomerular arteriole, 23leading to GFR values close to zero after hemorrhage (table 4). Urinary excretion after hemorrhage almost ceased in all protocols as a result of the renal vasoconstriction during hypovolemic hypotension mediated by reflex sympathetic excitation, the increase in circulating catecholamines, 24the increase in sodium- and water-retaining hormones (table 2), and the decrease in GFR (table 4).

After retransfusion, urinary excretion rate and GFR increased to awake levels in controls and ET^A-antagonized dogs, whereas in both AT¹inhibition protocols, renal function remained compromised, possibly because MAP, and thus renal perfusion pressure, only increased to 40–50 mmHg even after blood retransfusion in these protocols.

In conclusion, endothelin-A receptor inhibition impairs long-term blood pressure regulation after hemorrhage and retransfusion during isoflurane-nitrous oxide anesthesia despite an impressive activation of vasoconstrictive hormones such as vasopressin, angiotensin II, and epinephrine and activation of the sympathetic nervous system during hemorrhage. This indicates that endothelins seem to have a place especially in long-term (hours range) cardiovascular regulation. In contrast, angiotensin II (AT¹) receptor inhibition impairs both short and long-term blood pressure regulation after hypovolemic hemorrhage to a greater extent than during ET^Ainhibition. Moreover, renal function after hemorrhage and after retransfusion was by far more compromised by AT¹than by ET^Ainhibition.

The authors thank Rainer Mohnhaupt (Engineer, Clinic of Anesthesiology and Intensive Care Medicine, Charité, Campus Virchow, Berlin, Germany) for statistical advice, and Daniela Bayerl and Birgit Brandt (Medical Laboratory Assistants, Experimental Anesthesia, Charité, Campus Virchow) for expert technical assistance during the experiments.