Tolerance to Acute Isovolemic Hemodilution: Effect of Anesthetic Depth

All experimental procedures were performed in accordance with the Belgian guidelines for care of laboratory animals and were approved by the animal research committee of the School of Medicine of the Free University of Brussels (Brussels, Belgium). Twenty-eight mongrel dogs were studied. In the first study, animals were anesthetized with halothane (n = 14; weight 23.8 ± 4.8 kg). After administration of intravenous thiopental 20 mg/kg, tracheal intubation was performed, and mechanical ventilation was started (Elema 900B; Siemens, Solna, Sweden) with air. Ventilatory frequency was set at 12 min⁻¹, and tidal volume was adapted to maintain an arterial partial pressure of carbon dioxide (Paco²) of approximately 35 mmHg. Respiratory conditions remained unchanged until the end of the study. Anesthesia was maintained with halothane, anesthetic depth being adapted to the surgical stimulation of the experimental protocol (end-tidal minimum alveolar concentration [MAC] between 1.0 and 2.0). After completion of the surgical procedure, the dogs were randomly allocated to receive either 1.0 MAC halothane (0.96%; n = 7) or 1.5 MAC halothane (1.4%; n = 7). 22 

In a subsequent study, animals were anesthetized with ketamine (n = 14; 24.3 ± 4.7 kg). In this group, dogs were randomly allocated to receive ketamine either at low dose (an induction bolus of 5 mg/kg followed by a continuous infusion of 0.2 mg · kg⁻¹· min⁻¹; n = 7) or at high dose (an induction bolus of 10 mg/kg followed by a continuous infusion of 0.4 mg · kg⁻¹· min⁻¹; n = 7). After tracheal intubation was performed, controlled mechanical ventilation was started with respiratory conditions as described previously.

In both studies, animals received pancuronium bromide to facilitate controlled ventilation. This agent was given as an initial bolus of 0.1 mg/kg followed by hourly boluses of 1–2 mg.

Electrodes were attached to the four limbs for monitoring heart rate (HR). Two catheters (16-gauge; Becton Dickinson and Co., Rutherford, NJ) were inserted: one into a peripheral vein for fluids and drug infusion and one into the distal aorta through a femoral artery for monitoring of blood pressure and sampling of arterial blood. A pulmonary arterial catheter (Swan-Ganz catheter, model 93A-131-7F; Baxter Healthcare, Irvine, CA) was inserted through a jugular vein. A gas analyzer (Capnomac AGM 103; Datex, Helsinki, Finland) was inserted in the respiratory breathing system for monitoring of end-tidal carbon dioxide tension (Petco²) and halothane concentration. Exhaled gases were directed through a mixing chamber for sampling to measure mixed expired oxygen fraction (Feo²) using a paramagnetic method (semirapid oxygen analyzer, model 500D PK; Morgan Co., Chatam, United Kingdom). The oxygen analyzer and the capnometer were calibrated before each experiment. Expired minute ventilation was measured with a spirometer (model Magtrak II, London, United Kingdom) located proximal to the mixing chamber. Respiratory circuit was carefully checked for air leakage before each experiment.

A splenectomy was performed through a midline laparotomy to prevent autotransfusion. To compensate for insensible fluid losses, the dogs received an intravenous infusion of 0.9% sodium chloride at a rate of 10 ml/kg during the splenectomy and 1 ml · kg⁻¹· h⁻¹thereafter, until the end of the experiment. Body temperature was maintained at 36–38°C throughout the study using warming blankets and humidified heated gases.

Heart rate, arterial pressure, and pulmonary arterial pressure were monitored continuously (Sirecust; Siemens AG, Erlangen, Germany) and recorded together with Petco²and Feo²(model 8000S; Gould Electronique, Ballainvilliers, France). Zero pressure was set at the midchest level of the animal.

Heart rate and intravascular pressures, including mean arterial pressure (MAP), mean pulmonary arterial pressure (MPAP), pulmonary arterial occlusion pressure, and right atrial pressure, were measured at end-expiration from the paper trace. Feo²and Petco²were measured from the paper trace over 1 min. Expired volume was measured over the same period. CO was measured with the thermodilution technique (module E 2261/A; Siemens AG) by repeated injections of 10 ml cold (<5°C) dextrose, 5%, in water. Each injection was started at the end of expiration. 23Three to five measurements, with a variability of less than 10%, were averaged for each CO measurement. Immediately thereafter, arterial and mixed venous blood samples were withdrawn for measurement of blood gas tensions (ABL 2; Radiometer, Copenhagen, Denmark). Hemoglobin and oxygen saturation were measured using a CO oximeter calibrated for dog's blood (OSM 3; Radiometer). Each sample was analyzed twice, and the two values were averaged. Arterial blood lactate concentration (lactate) was measured enzymatically using an automated analyzer (Kontron, Basel, Switzerland).

Cardiac index (CI), stroke index (SI), systemic vascular resistance (SVR), left ventricular stroke work index, and arterial and mixed venous oxygen content were calculated using standard formulae. Oxygen delivery (Do²) was calculated from CO measurements and arterial oxygen content and indexed for weight. Oxygen consumption (V̇o²) was calculated from measured Feo²and minute ventilation, using the appropriate mass balance equation and accounting for the difference between inspired and expired nitrogen fraction, indexed for weight. 24Oxygen extraction (O²ER) was obtained by dividing V̇o²by Do².

Baseline measurements were performed 30 min after completion of splenectomy. Each animal was then slowly hemodiluted by repeated withdrawals of 5-ml/kg aliquots of arterial blood simultaneously replaced by the same volume of 6% hydroxyethyl starch, 200/0.5 (Haes Steril 6%; Fresenius AG, Bad Hombourg, Germany). After each exchange procedure, a 10-min period was allowed to achieve a new steady state defined by a stable MAP, Petco², and Feo². The exchange procedure was continued until the animal could no longer maintain a stable arterial pressure. Death followed shortly thereafter, with the animal still anesthetized.

In each animal, critical Do²was determined from a plot of V̇o²versus  Do²using the dual lines regression method. 25Critical Do²was defined as the point of intersection of the two best-fit regression lines as determined by a least sum of squares technique. Critical O²ER was calculated as the ratio of V̇o²/Do²at critical point. Critical Do²was also determined in each animal using the same dual regression lines analysis from a plot of lactate versus  Do². The critical value of hemoglobin was determined using the same mathematical analysis from a plot of V̇o²versus  hemoglobin (hemoglobin from V̇o²) and from a plot of lactate versus  hemoglobin (hemoglobin from lactate). Figures 1 and 2illustrate the V̇o²–Do², lactate–Do², V̇o²–hemoglobin, and lactate–hemoglobin relationships that were obtained in each dog in the halothane and ketamine studies.

The hemoglobin–Do²relationship was also analyzed in each animal by determining the polynomial that best fit the data. Equations of these polynomials were used to determine Do²values corresponding to fixed hemoglobin values. This allowed for comparison of the Do²response to progressive hemodilution between the two anesthetic dosages in the halothane and the ketamine study.

For both halothane and ketamine, sample size was calculated to detect a difference in critical hemoglobin of 40% between the two anesthetic dosages with a power of 0.8 and an α of 0.05.

Hemodynamic and respiratory variables obtained at baseline and at the experimental point closest to critical Do²with the two anesthetic dosages were compared using a two-way analysis of variance, followed by pairwise comparisons using a Tukey honest significance difference post hoc  test. Data obtained at critical point in the two anesthetic levels were compared using the Mann–Whitney U test. In each group, the hemoglobin–Do²relationship was compared between the two anesthetic dosages using a two-way analysis of variance for repeated measurements. Because anesthetic depth between both agents could not be compared, no statistical analysis was performed between data obtained with the two anesthetic agents.

All values are expressed as mean ± SD. A P  value less than 0.05 was considered statistically significant.

In all animals, acute isovolemic hemodilution resulted in a significant decrease in MAP and in SVR (table 1). In the 1.0 MAC halothane group, CI increased, mainly through an increase in SI. In the 1.5 MAC halothane group, neither CI nor SI increased. As compared to 1.0 MAC, 1.5 MAC halothane anesthesia resulted in a lower HR and MAP, a higher pulmonary arterial occlusion pressure, and a lower CI, SI, and left ventricular stroke work index. Between the two anesthetic dosages, there was a significantly different response to acute hemodilution for CI and SI. In all dogs, acute isovolemic hemodilution resulted in a decrease in mixed venous oxygen saturation. Do²decreased but V̇o²remained constant as O²ER increased (table 1). Just before the critical point, hemoglobin tended to be higher, and V̇o²to be lower, in the 1.5 MAC halothane group.

Critical Do²was significantly lower in the 1.5 MAC halothane group (table 2). This was due to a lower critical V̇o²as critical O²ER was not significantly different between the two groups. Critical Do²obtained from lactate was also significantly lower in the 1.5 MAC halothane group. Lactate at the critical point was not different in the 1.5 MAC halothane and 1.0 MAC halothane groups. Critical hemoglobin obtained through either V̇o²or lactate was significantly higher in the 1.5 MAC halothane group than in the 1.0 MAC halothane group. V̇o²at critical hemoglobin was lower in the 1.5 MAC halothane group than in the 1.0 MAC halothane group (3.6 ± 0.3 and 4.5 ± 1.2 ml · min⁻¹· kg⁻¹, respectively;P < 0.05). Lactate at critical hemoglobin was not different between the two groups (1.5 MAC halothane: 2.8 ± 1.1 mm/l; 1.0 MAC halothane: 2.0 ± 1.0 mm/l.

With the two anesthetic concentrations, isovolemic hemodilution resulted in a progressive decrease in Do²(fig. 3, top ). Do²changes in response to the stepwise decrease in hemoglobin concentration were significantly different among the two halothane concentrations.

In all animals, acute isovolemic hemodilution resulted in a significant decrease in MAP (table 3). In the low-dose ketamine group, CI increased, mainly through an increase in SI. SVR decreased. In the high-dose ketamine group, neither CI nor SI increased. As compared to low-dose ketamine, high-dose ketamine resulted in a lower HR and MPAP, a higher SVR, and a lower CI, SI, and left ventricular stroke work index (table 3). Between the two anesthetic dosages, there was a significantly different response to acute hemodilution for HR, CI, SI, and SVR. In all dogs, acute isovolemic hemodilution resulted in a decrease in arterial pH and mixed venous oxygen saturation. Do²decreased but V̇o²remained constant as O²ER increased (table 3). Just before the critical point, hemoglobin concentration tended to be higher and V̇o²was significantly lower in the high-dose ketamine group.

Critical Do²was significantly lower in the high-dose ketamine group (table 4). This was due to a lower critical V̇o²as critical O²ER was not significantly different between the two groups. Critical Do²obtained from lactate was also significantly lower in the high-dose ketamine group. Critical hemoglobin obtained from V̇o²was significantly higher in the high-dose ketamine group. V̇o²at critical hemoglobin was significantly lower in the high-dose ketamine group than in the low-dose ketamine group (4.4 ± 1.2 and 6.0 ± 0.9 ml · min⁻¹· kg⁻¹, respectively;P < 0.05). Critical hemoglobin obtained from lactate tended to be higher in the high-dose ketamine group, although the difference did not reach statistical significance. Lactate at critical hemoglobin was not different between the two groups (high-dose ketamine: 1.7 ± 0.7 mm/l; low-dose ketamine: 1.8 ± 0.8 mm/l.

With the two anesthetic concentrations, isovolemic hemodilution resulted in a progressive decrease in Do²(fig. 3, bottom ). Do²changes in response to the stepwise decrease in hemoglobin concentration were significantly different among the two ketamine dosages.

Increased anesthetic depth obtained with either halothane or ketamine resulted in a decreased tolerance to acute isovolemic anemia, as reflected by an increase in critical hemoglobin, the value of hemoglobin below which V̇o²becomes supply dependent. With both agents, the increase in critical hemoglobin at the higher anesthetic level was related to a complete blunting of the CO response usually observed during isovolemic hemodilution. 1–3Indeed, CI and SI increased only with the lower dose of each agent but not with the higher one.

These results were observed with two anesthetic agents having quite different cardiovascular properties. 17–21The blunting of the CO response with the higher halothane concentration occurred in the presence of higher filling pressures and, despite a lower SVR, reflecting the negative inotropic properties of this agent. A depressant effect of halothane on the sympathetic system could also be implicated. 26,27In the ketamine study, the blunting of the CO response with the higher anesthetic dosage also seemed to be related to a direct depression of the myocardial function, as reflected by the lower left ventricular stroke work index. This phenomenon occurred despite the effects of this agent on the sympathetic system, as evident from the better maintenance of MAP and SVR during the progressive hemodilution. Although ketamine possesses cardiostimulatory effects through an increase in sympathetic outflow, it also presents direct myocardial depressant properties. 28,29In conditions associated with increased sympathetic stimulation, the direct effects of ketamine on the myocardium could exceed its sympathomimetic properties, resulting in global depressant hemodynamic effects. 30,31 

Our results may have been affected by the concomitant use of other agents such as thiopental and pancuronium. Thiopental, however, was used only in the halothane study and was administered at least 2 h before the experimental protocol. Therefore, it seems unlikely that this drug would have influenced our observations. Pancuronium was used to facilitate mechanical ventilation. This agent, which decreases vagolytic tone and increases sympathetic tone, 32could have different effects on animals anesthetized with halothane or with ketamine. However, within each anesthetic study, it probably had little influence on our results, as the dose of muscle relaxant used was similar in animals receiving the low and the high anesthetic dosages.

Medium-molecular-weight hydroxyethyl starch was used for the hemodilution protocol as this kind of solution has been shown to produce stable plasma volume expansion. 33Because of the high chloride concentration in the diluent of the solution, its uses may be associated with the development of metabolic acidosis, as reflected by the decrease in arterial pH observed in both the halothane and the ketamine studies. This effect, however, was similar in animals receiving the low and the high anesthetic dosages.

In this protocol of progressive hemodilution, with both agents, critical Do²was significantly lower in the animals receiving the higher anesthetic dosage. This resulted from lower oxygen requirements as indicated by a lower critical V̇o². This could be in part related to the dose-dependent negative inotropic properties of halothane and ketamine, resulting in a decreased myocardial oxygen demand. Central body temperature tended to be lower in the animal receiving the higher anesthetic dosage, despite the use of warming blankets and humidified heated gases, A difference in less than 1° in central temperature, however, could hardly explain the marked difference in oxygen consumption between the two anesthetic dosages in the halothane and ketamine studies. Indeed, hypothermia decreases whole body metabolic rate by approximately 8% by Celsius degrees. 34The lower central body temperature observed in the animals receiving the higher anesthetic dosage could be a result of the decreased perfusion but could also reflect the effects of anesthesia on the body's thermoregulation system in a cold operating room environment. 35 

During the hemodilution protocol, Do²decreased progressively in both the halothane and the ketamine studies as a result of the blunting of the CO response in our anesthetized animals (fig. 3). This is in contrast with previous works reporting that when hematocrit is reduced, systemic oxygen transport capacity first increases slightly, reaches a peak at approximately 30% hematocrit, and does not fall below control levels as long as the hematocrit remains above 20%. 36,37The Do²response to hemodilution could have been predicted with halothane but was more surprising for ketamine. This agent has a dual effect on the cardiovascular system, the balance of its indirect stimulant properties and the direct myocardial depressant properties being altered by concentration and time as shown in figure 3.

We took a number of precautions to ensure the validity of our observations. V̇o²was determined from expired gas analysis, whereas Do²was calculated from thermodilution CO and blood gas analysis, so that there was no mathematical coupling of the data. A similar protocol has been used by Cain et al.  9in evaluating the effects of isovolemic anemia on critical Do²in pentobarbital-anesthetized dogs. Determining critical Do²“by eye,” this author evaluated that the hematocrit value for which V̇o²became Do²-dependent was around 10%. In the current study, critical hemoglobin was obtained from both V̇o²and lactate determinations. Both approaches demonstrated similar hemoglobin values that were consistently larger in dogs anesthetized with the higher anesthetic dosage. Interestingly, critical Do²and critical hemoglobin values obtained from lactate measurements were in agreement with those derived from V̇o²measurements, confirming the usefulness of this parameter to detect the development of tissue hypoxia during profound anemia. 10,11,38 

To test the hypothesis that anesthesia can decrease tolerance to acute isovolemic hemodilution, we used two anesthetic agents with well-defined cardiovascular and metabolic effects. In the current study, when the depth of anesthesia was increased, the depressant effects of halothane and ketamine on the cardiovascular system outweighed their possible beneficial effects on tissue metabolism, whatever their effects on the sympathetic system. Our observations may not be directly applicable to other inhaled anesthetics, especially desflurane. However, newer halogenated anesthetics also have negative inotropic properties 39,40that can decrease the CO response associated with acute isovolemic hemodilution. Although isoflurane has been shown to maintain CO better than halothane, Schou et al.  41demonstrated in the pig that isoflurane induced a dose-dependent decrease in CO response during profound normovolemic hemodilution. Several studies using fentanyl–isoflurane–N²O, 42fentanyl–droperidol–N²O, 43or fentanyl–midazolam 44anesthetic-based protocols also reported a blunting of the CO response during isovolemic hemodilution in man. A parasympathetic stimulation related to the central vagal stimulation induced by some opioids such as fentanyl 45could be responsible for a lack of increase in HR, thereby contributing to a decrease in the CO response. Therefore, a decreased tolerance to acute isovolemic anemia when anesthesia is deepened could be relevant for other anesthetic agents with cardiovascular depressant properties. Our observations emphasize the importance of carefully titrating the level of anesthesia and monitoring the cardiovascular response in patients with moderate to severe anemia.