Acute normovolemic hemodilution reduces the circulating erythrocyte mass and, thus, the hemoglobin concentration. After extreme acute normovolemic hemodilution to the critical hemoglobin concentration (Hbcrit), oxygen demand of the tissues is no longer met by oxygen supply, and death occurs with increasing oxygen debt. The aim of the current study was to investigate whether ventilation with 100% oxygen (fraction of inspired oxygen [FiO2] = 1.0; hyperoxic ventilation) initiated at Hbcrit could restore adequate tissue oxygenation and prevent death.
Fourteen anesthetized pigs ventilated with room air (FiO2 = 0.21) were hemodiluted by exchange of whole blood for 6% hydroxyethyl starch (200,000:0.5) until the individual Hbcrit was reached. Hbcrit was defined as the onset of oxygen supply dependency of oxygen consumption and was identified with indirect calorimetry. For the next 6 h, animals were either ventilated with an FiO2 of 0.21 (n = 7) or an FiO2 of 1.0 (n = 7).
All animals in the 0.21 FiO2 group died within the first 3 h at Hbcrit (i.e., 6-h mortality 100%). Death was preceded by an increase of serum concentrations of lactate and catecholamines. In contrast to that, six of the seven animals of the 1.0 FiO2 group survived the complete 6-h observation period without lactacidosis and increased serum catecholamines (i.e., 6-h mortality 14%; FiO2 0.21 vs. FiO2 1.0, P < or = 0.05). After 6 h at Hbcrit, the FiO2 was reduced from 1.0 to 0.21, and five of the six animals died within the next 3 h.
In anesthetized pigs submitted to lethal anemia, hyperoxic ventilation enabled survival for 6 h without signs of circulatory failure.
REPLACEMENT of intraoperative blood loss with erythrocyte-free (i.e. , crystalloidal or colloidal) solutions avoids premature unnecessary erythrocyte transfusion. 1Although hemodilution reduces the erythrocyte mass and the hemoglobin concentration, 2it has long been known that adequate tissue oxygenation does not depend on “normal” hemoglobin concentrations. 1,3,4As long as normovolemia is maintained, tissue oxygenation is preserved during acute dilutional anemia until a so-called critical hemoglobin concentration (Hbcrit) is reached. 1
At Hbcrit, the compensatory mechanisms of the body to sustain tissue oxygen demand (i.e. , increase of cardiac output and arterial oxygen extraction) become exhausted. 2Further hemodilution to a hemoglobin concentration lower than Hbcritreduces oxygen supply below the oxygen demand of the tissues, and tissue hypoxia occurs. To avoid serious organ dysfunction and death, arterial oxygen content (Cao2) must be increased instantaneously. This is usually realized by transfusion of erythrocytes (autologous and/or allogenic). However, if transfusion must be started despite ongoing blood loss, part of the erythrocytes transfused are immediately spilled out of the circulation. Therefore, an intervention other than transfusion would be advantageous to bridge the time between the indication and the start of transfusion. As a short-term alternative to erythrocyte transfusion, Cao2can simply and instantaneously be increased by hyperoxic ventilation (HV), i.e. , ventilation with 100% oxygen (fraction of inspired oxygen [Fio2] 1.0) via the increase of oxygen physically dissolved in plasma. 5
Previous studies have shown beneficial effects of HV on tissue oxygenation and organ function at different degrees of acute dilutional anemia. 6–9However, at Hbcrit, HV failed to increase global, myocardial, and intestinal oxygen delivery (Do2) 10: despite the increase of Cao2, global and in particular myocardial Do2remained unchanged during HV because of hyperoxia-induced arteriolar constriction and a decrease in nutritive organ blood flow. 10
However, at Hbcrit(i.e. , at the limit of oxygen transport), the definite impact of vasoconstriction on survival was not yet investigated. 11,12In particular, it was unknown whether the overproportional constriction of myocardial arterioles 10would counteract the expected increase in myocardial oxygen supply and thus lead to an even higher risk of myocardial failure and death.
Therefore, the aim of the current study was to investigate the effects of HV on mortality after acute normovolemic hemodilution to Hbcritbecause mortality is the most critical consequence of inadequate tissue perfusion and oxygenation. We speculated that in case of high biologic availability of physically dissolved plasma oxygen, mortality should be reduced by HV at Hbcritdespite an impaired myocardial blood flow and a lack of increased myocardial Do2.
Materials and Methods
After governmental approval, the study was performed in 14 domestic pigs of either sex (weight, 26.7 ± 5.7 kg). The animals were treated in accordance with the Principles of Laboratory Animal Care. 13
After one night’s fasting with free access to water, intramuscular premedication was performed with midazolam (1–2 mg/kg) and ketamine (10 mg/kg). Anesthesia was induced by intravenous injection of fentanyl (0.01 mg/kg) and propofol (2 mg/kg) and maintained by continuous infusion of the same drugs (0.01–0.02 mg · kg−1· h−1fentanyl and 20 mg · kg−1· h−1propofol). Muscular relaxation was omitted.
Estimated fluid losses were replaced by intravenous infusion of balanced electrolyte solution (3 mg · kg−1· h−1). A warming pad and a warming lamp were used to keep core body temperature constant (baseline temperature ± 0.25°C).
The animals were endotracheally intubated and mechanically ventilated at a rate of 14 min−1and a positive end-expiratory pressure of 5 cm H2O. On the basis of multiple blood gas analyses, tidal volume was individually adjusted before the onset of the protocol to maintain normocapnia and was then maintained throughout the whole protocol.
Instrumentation and Monitoring
All animals were placed in a supine position. A five-lead electrocardiogram (II, V5) was installed for detection of arrhythmias and ST-segment changes. A catheter was inserted into each femoral artery for continuous arterial pressure recording and blood withdrawal during the hemodilution procedure. Another catheter was inserted into the external jugular vein for infusion of the plasma substitute in the course of the hemodilution process and for monitoring of central venous pressure. To avoid any stress related to surgical preparation, all catheters were inserted percutaneously using the technique of Seldinger. Total body oxygen consumption (V̇o2) was measured noninvasively at intervals of 1 min with a Deltatrac metabolic monitor (DeltaTrac II™ MBM-200; Datex, Helsinki, Finland) connected to the respirator.
The study protocol is depicted in figure 1. After catheter insertion, a 60-min period was allowed to elapse for stabilization of V̇o2(see Determination of Hbcrit), and a first set of data were collected (time point: baseline). Subsequently, all animals were hemodiluted with 6% hydroxyethyl starch (200,000:0.5) to their individual Hbcrit(blood exchange rate 1 ml · kg−1· min−1). Hbcritwas defined as a significant decrease of total body V̇o2as compared to the baseline value (see below). When Hbcritwas met, a second set of data were collected (time point: Hbcrit). At this time, the animals were randomized to the two experimental groups: seven animals were further ventilated with room air (0.21 Fio2group), and the other seven animals were ventilated with 100% oxygen (1.0 Fio2group). During the next 6 h, data were collected every 15 min (observation period 1). If a pig randomized to the 1.0 group survived the 6-h observation period, the Fio2was reduced to 0.21, followed by a second 6-h observation period (observation period 2). In both groups, CVP measured at time point Hbcritwas kept constant throughout the observation period by infusion of 0.9% NaCl to maintain intravascular volume.
Determination of Hbcrit
Under resting conditions, V̇o2of the tissues equals their oxygen demand. However, when Do2decreases below a critical value, V̇o2becomes oxygen-supply dependent. The sudden decrease in V̇o2reflects the onset of tissue hypoxia (fig. 2). 14,15
The detection of Hbcritwas automated with computer software particularly designed for this purpose (DeltaCrit System). 16V̇o2values collected during a stable 60-min observation period were included in an on-line regression analysis. Every V̇o2value measured during the subsequent hemodilution process was compared to the mean value predicted by the DeltaCrit system: if the actual value was outside a predefined range (3 × SD of regression line), a significant decrease of V̇o2was assumed, and the computer alerted visually and acoustically. 16
Measurements were performed at baseline, Hbcrit, and every 15 min after randomization (observation periods 1 and 2;fig. 1). A blood gas analyzer (Chiron Diagnostics, Fernwald, Germany) was used to measure partial pressure of oxygen (Po2), partial pressure of carbon dioxide (Pco2), and pH. Hemoglobin concentration and arterial hemoglobin–oxygen saturation (Sao2) were measured by spectrophotometry adjusted to pig hemoglobin (682 CO-oximeter; Instrumentation Laboratory, Lexington, MA). Serum catecholamine concentrations (epinephrine and norepinephrine) were determined using high-performance liquid chromatography. Electrocardiogram printouts were analyzed off-line by two blinded investigators. Electrocardiographic criteria for myocardial ischemia included the occurrence of arrhythmia and ST-segment depression of greater than 0.1 mV. 17
Data are presented as mean ± SD. The 6-h mortality rates observed in the two study groups were compared using the Fisher exact test (SigmaStat 2.0; Jandel Corporation, San Rafael, CA). The time effect on the different variables as well as differences between groups at the different time points were tested by repeated-measures analysis of variance for the time points baseline, Hbcrit, and 15 min after Hbcrit. Post hoc analysis of within-group differences and between-group differences was performed with a Student-Newman-Keuls test (Statistica 5.1; StatSoft, Tulsa, OK). Subsequent time points were not tested because of relevant differences in the sample sizes between groups. Statistical significance was accepted at P < 0.05 for all tests.
At baseline, no differences were detected between the two groups regarding all parameters investigated (table 1).
Hemodilution to Hbcrit
Normovolemic hemodilution to Hbcritrequired the exchange of 63 ± 17 ml/kg (0.21 group: 59 ± 15 ml/kg; 1.0 group: 67 ± 16 ml/kg; P < 0.05) blood for hydroxyethyl starch, and took approximately 1 h. Hbcritwas 3.1 ± 0.4 g/dl in the 0.21 group and 2.9 ± 0.7 g/dl in the 1.0 group (not significant).
At Hbcrit, hemodynamic changes were comparable in both groups (table 1). Mean arterial pressure decreased in both groups (0.21 group:−42%; 1.0 group:−42%), whereas heart rate was not altered (table 1). Central venous pressure increased in both groups (0.21 group:+30%; 1.0 group:+44%). V̇o2decreased in both groups to a similar degree (0.21 group:−11%; 1.0 group:−10%).
There was no evidence for increased sympathetic tone: plasma epinephrine and norepinephrine concentrations remained unchanged. Arterial lactate also remained unchanged at Hbcrit, whereas standard base excess decreased (0.21 group:ΔBE =−4.1 mmol; 1.0 group:ΔBE =−5.5 mmol). Cao2decreased to the same degree in both groups (0.21 group:−61%; 1.0 group:−62%); the percent contribution of physically dissolved plasma oxygen to Cao2increased from 2.4% to 7.0% in the 0.21 group and from 2.3% to 7.4% in the 1.0 group.
In 11 of the 14 animals investigated, hemodilution to Hbcritprovoked significant electrocardiographic changes: 10 animals presented ST-segment depression or elevation (fig. 3); one developed negative T waves at Hbcrit. Ventricular arrhythmia or ectopia was not encountered in any of the animals. Three animals showed no significant electrocardiographic changes at time point Hbcrit(two in the 0.21 group and one in the 1.0 group).
Observation Period 1
All animals in the 0.21 group died within the first 3 h of observation period 1 (minimum: 15 min; maximum: 2 h 54 min; 6-h mortality 100%) as a result of asystole, whereas six of seven animals in the 1.0 group survived the complete 6-h period (6-h mortality 14%; P ≤ 0.05, 0.21 group vs. 1.0 group;fig. 4). No animal had ventricular fibrillation.
In the 1.0 group, the onset of HV (time point 15 min) was associated with a significant increase in mean arterial pressure (+56%). In contrast, in the 0.21 group, mean arterial pressure decreased continuously until the death of the animals (fig. 5). Heart rate did not differ between groups. After an initial peak, the arterial serum lactate concentration normalized after 1 h in the 1.0 group, whereas a continuous increase was observed in the 0.21 group (fig. 6). The concentrations of epinephrine and norepinephrine remained unchanged in the 1.0 group, whereas in the 0.21 group, an extreme increase of catecholamines preceded the death of each animal (fig. 7).
Pathologic electrocardiographic readings reflecting myocardial ischemia remained unchanged or even worsened in the 0.21 group during the first 15 min of observation period 1. In contrast, the electrocardiogram improved in five of the seven animals of the 1.0 group within 15 min of HV (disappearance of ST-segment changes as well as T-wave negativity). In one animal, the electrocardiogram did not change within 15 min (persisting ST-segment depression), and another animal died after 13 min (fig. 3).
Observation Period 2
In the surviving animals of the 1.0 group, the reduction of Fio2to 0.21 after 6 h of HV provoked changes similar to those observed in the 0.21 group immediately after having reached Hbcrit(table 2). Five of the six animals died within 1 h 5 min as a result of asystole. One animal survived the second observation period of 6 h (fig. 8).
Arterial oxygen content (Cao2) decreased by 21% after the reduction of Fio2. Simultaneously, the serum concentrations of epinephrine (3 ± 6 vs. 307 ± 586 pg/ml) and norepinephrine (29 ± 10 vs. 658 ± 1,283 pg/ml) both increased. Heart rate increased by + 41%. Subsequently, pH and base excess decreased, whereas arterial lactate increased (+340%;table 2).
In our experimental pig model, hemodilution to Hbcritcould be shown to be 100% lethal within 6 h if the animals were ventilated with room air. HV initiated at Hbcritreduced this 6-h mortality from 100% to 14%. With termination of HV after 6 h, mortality increased to 83%. Therefore, in animals confronted with lethal anemia, HV allowed “bridging” of a time period of 6 h and protected the animals from death.
Hyperoxic ventilation increases the amount of oxygen physically dissolved in plasma and by that Cao2. Although at physiologic hemoglobin concentrations the percent contribution of plasma oxygen to V̇o2may be considered negligible (approximately 5%), the situation changes after acute normovolemic hemodilution: with the shrinking numeric importance of erythrocytes as oxygen transporters, the percent contribution of physically dissolved oxygen to V̇o2increases to 47% (hemoglobin 7 g/dl) 18and 74% (hemoglobin 3.0 g/dl). 19Moreover, hyperoxemia (i.e. , the increase of Cao2by HV) initiated after moderate hemodilution partially reversed the compensatory mechanisms for anemia (hemoglobin 7 g/dl) 18and created a margin of safety for tissue oxygenation and organ function that allowed further extension of hemodilution (hemoglobin 3.0 g/dl) 19without signs of tissue hypoxia and organ dysfunction. In our study, HV increased Cao2from 4.1 to 5.8 ml/dl. This corresponds to the quantity of oxygen, which can be transported by an additional amount of 1.25 g/dl of 98%-saturated hemoglobin (ΔHb = (5.8 ml/dl − 4.1 ml/dl)/(0.98 × 1.39 ml/g)).
Unexpectedly, despite the increase in Cao2, HV always failed to restore the hemodilution-induced impaired systemic, myocardial, and intestinal Do2. The onset of HV is accompanied by generalized arteriolar constriction mediated by arachidonic acid metabolites and reduced endothelial nitric oxide release. 18,20–24In a previous experimental model in pigs hemodiluted to Hbcrit, we could demonstrate that the increase of coronary vascular resistance exceeded the increase of systemic vascular resistance by far. Therefore, it could not be excluded that HV, initiated at a lethal degree of anemia, could be deleterious for myocardial Do2, myocardial contractility, and ultimately survival.
It is not known to what extent anemic anoxia leads to “stunning” of the myocardium and whether HV should improve stunning and myocardial function immediately. In a previous pig model, HV led to an increase of intestinal Do2but barely improved myocardial function within half an hour. 10Survival is a cumulative parameter for tissue perfusion, oxygen delivery, and oxygenation. Any improvement in survival observed with the onset of HV should be interpreted as an indirect sign of adequate tissue oxygenation.
After acute normovolemic hemodilution to Hbcrit, all animals that were further ventilated with room air died during the 6-h observation period. The 86% survival rate of the animals ventilated with pure oxygen can only be explained with the high biologic availability of plasma oxygen for tissue oxygenation. Even though we did not measure coronary blood flow and regional myocardial perfusion in the current experimental model, we speculate that the impact of HV on these parameters should be comparable to that in previous reports. 10
One could also speculate that the improvement of 6-h survival was partially realized by the better maintenance of mean arterial pressure during HV. Because coronary perfusion pressure is defined as the difference between diastolic aortic pressure and left ventricular end-diastolic pressure, arteriolar constriction might even be considered beneficial for myocardial perfusion, oxygenation, and performance. It cannot be excluded that the animals became slightly hypovolemic during the 6-h observation period despite maintenance of CVP. However, if so, hypovolemia should have occurred in both groups to a similar extent, and the effect on survival should have been comparable.
We have observed no stress reaction or arterial lactacidosis during HV despite the presence of a critical level of anemia. Therefore, the positive effects of HV on oxygen transport seem to outweigh the previously observed, potentially disadvantageous microcirculatory perfusion mismatch.
Recently, van Bommel et al. 6reported that HV initiated during extreme anemia increased cerebral microvascular oxygen partial pressure and improved cerebral oxygen delivery. Weiskopf et al. 25found that HV reversed anemia-induced deficits of cognitive function and memory in awake volunteers having undergone extreme hemodilution to a hemoglobin concentration of 5.7 ± 0.4 g/dl. Therefore, it may be suggested that HV not only sustains myocardial function, but it also sustains cerebral tissue oxygenation and function at the presence of extreme hemodilution.
After the Fio2was returned from 1.0 to 0.21 in the surviving animals of the 1.0 group, all animals but one died within the next 2 h. Therefore, HV initiated at Hbcritallowed bridging of a time interval of 6 h and guaranteed adequate tissue oxygenation, organ function, and survival. The sudden death of the animals after termination of HV indicated that survival was not explained by undetected augmentation of hemoglobin-based Cao2or by genetic differences among the animals.
The clinical application of HV may contribute to the avoidance of premature erythrocyte transfusion in the intraoperative setting. In patients experiencing major intraoperative bleeding during orthopedic surgery, the indication of erythrocyte transfusion—based on the appearance of physiologic trigger parameters—could be reversed in two thirds of patients by the simple switch from ventilation with an Fio2of 0.4 to HV. Therefore, a time interval of 27–60 min (median, 30 min) and an additional blood loss of approximately 850 ml could be bridged by further intraoperative hemodilution under the protection of HV. 26
In summary, in anesthetized pigs hemodiluted to their individual oxygen transport limit, HV significantly reduced 6-h mortality. This finding may be interpreted as an indirect proof of the biologic availability and effective use of physically dissolved plasma oxygen for tissue oxygenation. Despite a potential maldistribution of organ perfusion at the microcirculatory level, lethal myocardial dysfunction was absent during HV. Therefore, HV can be considered a rescue therapy in the presence of acute bleeding and life-threatening acute anemia. It creates a margin of safety for tissue oxygenation, allowing for extreme hemodilution without the risk of tissue hypoxia, and may bridge the time until the availability of erythrocytes.