Abstract
Solutions of modified cell-free hemoglobin, prepared from outdated red blood cells, have been developed during the past decade to circumvent the increasing need for allogeneic blood. Despite improvements in the safety and efficacy of these solutions, undesirable effects such as an increase in vascular tone leading to hypertension have not been fully resolved, which might hinder their clinical usefulness. To discriminate between the pharmacological and rheological effects of cell-free hemoglobin, we compared the effects of blood/cell-free hemoglobin mixtures of high versus low viscosity on hemodynamics and vascular hindrance, an index of vascular tone, which was normalized for blood viscosity. Anesthetized rats were subjected to 50% exchange transfusion with (1) high-viscosity solutions: whole blood (n=5) or red blood cells mixed with cell-free hemoglobin (Hb-Hv group, n=5); (2) low-viscosity solutions: cell-free hemoglobin (Hb-Lv group, n=5) or human albumin (n=5). Two hours after hemodilution, vascular hindrance remained unchanged in animals transfused with whole blood and albumin. Hb-Lv induced an immediate and sustained increase in vascular hindrance (208%). Conversely, in Hb-Hv animals, the vascular hindrance increase was delayed and smaller (27% to 147%), whereas peripheral resistance increased gradually (94% after 2 hours). Our results demonstrate the beneficial effects of cell-free hemoglobin in the presence of the animals’ own red blood cells in maintaining physiological viscosity and limiting vasoconstriction because of the pharmacological properties of cell-free hemoglobin.
Compensation of blood loss to restore cardiac output and tissue perfusion is required when patients have reached the transfusion trigger (classic transfusion threshold of 10 g/dL). Current approaches used to overcome the problems associated with the short supply of donor blood and the risks from contaminated blood include the use of plasma expanders (such as colloids, crystalloids, or a combination). Another emerging alternative to blood donation is the use of hemoglobin-based oxygen-carrying solutions (HBOCs), which can restore blood volume and carry and release oxygen to tissues. However, after major compensation from hemorrhage, blood is subjected to a reduction in viscosity that leads to an alteration in shear stress that therefore, affects vessel regulation processes under the control of mechanotransduction (physical mechanisms).1 Furthermore, it is also well documented that a hemoglobin solution can alter cardiovascular homeostasis by a number of mechanisms: (1) pharmacological effects leading to vasoconstriction, such as molecular interactions of the hemoglobin moiety with nitric oxide and/or endothelin-1 and (2) autoregulation mediated by the oxygen supply.2,3 Thus, the beneficial (oxygenation) and deleterious (vasoconstriction) acute effects of HBOCs could result from a combination of pharmacological and mechanical properties.
In a recent experiment with a well known vasoactive hemoglobin solution, diaspirin cross-linked hemoglobin (αα-Hb) in hemodiluted-hemorrhaged rats produced deleterious effects in vascular resistance compared with animals infused with another hemoglobin, a polyethylene glycol–modified hemoglobin. This suggests that the formulation and physicochemical characteristics of the infused solutions are important parameters that need to be considered in the design of second-generation hemoglobin solutions.4 This phenomenon had been observed previously by Tsai et al,5 who reported that increasing the blood oxygen-carrying capacity and lowering blood viscosity had deleterious effects due to microvascular autoregulation processes, leading to vasoconstriction and an impaired oxygen supply to tissues. Recently, Tsai and Intaglietta6 suggested that the “ideal” HBOC should combine high viscosity, low hemoglobin concentration, and a low P50 (O2 half-saturation pressure). However, viscosity plays a major role in the blood circulation through vessels and capillaries, but this parameter has been poorly studied in a dynamic situation, and few studies address this issue. The mechanism underlying vasoconstriction caused by hemoglobin solutions is still controversial. Hypertension, a manifestation of vasoconstriction, appears to be the main unwanted side effect of hemoglobin solutions. This deleterious intrinsic property of hemoglobin limits its use in the urgent clinical situation. The physiological effects of these solutions must therefore be understood to specify their optimal clinical use.
Vascular hindrance (VH; resistance/blood viscosity) reflects the contribution of vascular geometry to flow for a normalized value of viscosity.7 This parameter is therefore more relevant than calculated peripheral resistance (PR), because the former is not affected by changes in blood viscosity.
To discriminate between the effects of viscosity and pharmacological properties of cell free-hemoglobin, we have compared 4 experimental situations of hemodilution: (1) blood with low viscosity (hematocrit [Hct]=0.20 L/L), in which plasma contained albumin or free hemoglobin and (2) blood with high viscosity (Hct=0.40 L/L), in which plasma remained unchanged or contained free hemoglobin. Before and for 2 hours after an isovolemic 50% exchange transfusion, blood samples were withdrawn from anesthetized rats to measure the blood viscosity at different values of shear stress and to calculate VH in each of the 4 hemodilution conditions.
Methods
Test Solutions
Human serum albumin (Alb) was prepared as a 5.0 g/dL solution in Ringer’s medium (Pasteur-Mérieux sérums & vaccins). Solution of αα-Hb (Walter Reed Army Institute of Research) which contains 8.2 g/dL of heat-treated human hemoglobin, was stabilized by use of a cross-linker [bis(3,5-dibromosalicyl)fumarate] between the 2 α-subunits.8 The values of kinematic viscosity of plasma, Alb, and αα-Hb measured in centi-stoke (cSt) with a capillary viscosimeter (module V, Amtec) at 37°C were similar (1.20±0.06, 1.10±0.07, and 1.22±0.05 cSt, respectively).
Preparation of Animals
The experimental protocol was approved by the Animal Protection Bureau of the French Ministry for Fishing, Agriculture and Food and was conducted in accordance with the Guiding Principles for Research Involving Animals and with the national instructions for laboratory animal use. The animals were fed ad libitum before the experiments. At the end of the experiments, the animals were euthanized by an overdose of pentobarbital.
On the day of experiments, 20 male rats (Iffa Credo, France) weighing 270±15 g were anesthetized with pentobarbital (50 mg/kg IP, Sanofi). Anesthesia was maintained by perfusion of a saline solution containing 6 mg/kg pentobarbital through the left jugular vein (delivered at 0.4 mL/h) except during the steps of blood withdrawal.
The right femoral artery was cannulated with a heparin-filled catheter advanced into the abdominal aorta for pulsatile arterial pressure measurements. After midline laparotomy, a hard epoxy Doppler blood-flow transducer (internal diameter, 1.0 mm; HDP-20, Crystal Biotech) was placed on the abdominal aorta above the emergence of the arterial catheter.9 The right carotid artery and jugular vein were cannulated to infuse test solutions and withdraw blood, respectively. After instrumentation, the animals were allowed to stabilize for a 30-minute period, and then the hemodilution protocol was initiated.
Hemodilution Protocol
Hemodilution was performed by exchange transfusion in which 50% of total blood volume (calculated as blood volume [mL]=0.06×body weight [g]+0.77)10 was exchanged gradually with one of the test solutions. The aim of this exchange transfusion was to obtain 2 levels of Hct. Low-viscosity conditions (target Hct=0.20 L/L) were obtained by exchanging blood with either αα-Hb (Hb-Lv) or albumin (Alb). High-viscosity conditions (target Hct=0.40 L/L) were obtained by exchanging blood with either autologous blood diluted with exogenous blood or with a mixture of αα-Hb (Hb-Hv) and exogenous red blood cells. The exogenous blood was obtained form anesthetized donor rats by heart puncture.
Blood Gas and Hematologic Values
Blood samples were withdrawn before (baseline) and 20, 60, and 120 minutes after the completion of exchange transfusion (Figure 1). To avoid changes in blood volume as a result of blood sampling, blood withdrawn was replaced immediately by an equal volume of the test solution mixed with autologous whole blood at the same hematocrit (0.20 L/L or 0.40 L/L) and containing the same concentration of free hemoglobin or Alb as the blood withdrawn. Values of Hct, arterial pH, Pao2, Paco2, plasmatic bicarbonate and acid/base excess, total hemoglobin concentration, blood oxyhemoglobin, methemoglobin, O2 capacity (possible O2 carrying), and O2 content (real O2 carrying) were measured as previously described.9 The concentration of various hemoglobin derivatives is reported as a function of the total hemoglobin concentration in the animal’s blood.
Figure 1. Study design. Measurement parameters were determined before (baseline b1) and at 20, 60, and 120 minutes after exchange transfusion (ET).
Hemodynamic Monitoring and Calculation
The aortic catheter was connected to a pressure transducer (Viggo-Spectramed) to measure the pulsatile arterial pressure. The aortic blood-flow velocity (cm/s) was measured by a transducer connected to a 20-MHz module (PD-20, Crystal Biotech) and a dedicated amplifier (CBI-8000, Crystal Biotech). Values of mean arterial pressure (MAP) and aortic blood flow (ABF) were calculated with the use of acqknowledge software. PR was calculated as MAP/ABF.9
Blood Viscosity and VH
Arterial blood was collected in sterile tubes containing 5% EDTA (wt/vol) before (baseline) and 20, 60, and 120 minutes after the completion of exchange transfusion. The viscosity was measured ex vivo at Hct values of 0.40 and 0.20 L/L. The viscosity was determined at 37°C with a Couette viscometer (Low Shear 30, Contraves) for shear rates ranging from 0.2 to 128 s−1 and expressed in mPa s11 VH was calculated according to Simchon et al12 before and after hemodilution for each group. Calculations of VH were made for a relevant shear rate value, namely, 128 s−1, representing the shear rate in arteries.
Statistical Analysis
The animals were randomly allocated to 1 of the 4 following experimental groups (n=5 each): whole blood, Hb-Hv, Hb-Lv, and Alb. All data are expressed as mean±SE. Statistical comparisons were made before hemodilution and at various posthemodilution time points (baseline and at 20, 60, and 120 minutes) for each group with a 1-factor ANOVA for repeated measures. Comparisons between groups were made for each time point with a 1-way ANOVA for repeated measures with the Tukey-Kramer post test or ANOVA plus the Dunnett post test when appropriate. A value of P<0.05 was considered significant.
Results
Blood Gases and Hematologic Values
The hemoglobin concentration in blood was dependent on both the post–exchange transfusion Hct and the concentration of hemoglobin injected (Table). Accordingly, the following order was observed: Hb-Hv>whole blood>Hb-Lv>Alb. The hemoglobin concentration was unchanged for 2 hours after exchange transfusion (Figure 2). Blood Pao2 was not significantly different between groups, whereas the oxygen content (O2ct) was dependent on the hemoglobin concentration. Methemoglobin slightly increased when the hemoglobin solution was applied (due to autoxidation of hemoglobin). No changes were observed in the other parameters after exchange transfusion by these solutions and at any given time point.
Hematological Parameters Measured in Animal Blood Samples as Function of Time
Figure 2. Total hemoglobin concentration measured in the bloodstream of rats as a function of time. High-viscosity (Hct=0.40 L/L): WB indicates whole blood; Hb-Hv, blood+cell-free hemoglobin. Low-viscosity (Hct=0.20 L/L): Alb indicates albumin-diluted blood; Hb-Lv, hemoglobin-diluted blood. Mean±SE, n=5, with £P<0.05, Alb vs WB; §P<0.05, Hb-Lv vs WB; ^PP<0.05, Hb-Lv vs Hb-Hv; #P<0.05, Alb vs Hb-Hv; ΔP<0.05, Alb vs Hb-Lv; *P<0.05, WB vs Hb-Hv.
Viscosity Data
Blood viscosity remained unchanged in animal transfused with whole blood for the 2 hours after exchange transfusion (Figure 3). Viscosity decreased slightly after injection of Hb-Hv but remained stable until the end of the experiment. Compared whole blood and Hb-Hv at the shear rate of 128 s−1, a significant difference was only observed for the last time point (P<0.01). In contrast, viscosity was dramatically decreased at low hematocrit (Alb and Hb-Lv groups; P<0.01). However, at the shear rate of 128 s−1, the viscosities were similar in the low-Hct groups.
Figure 3. Measurements of blood viscosity of acute hemodiluted situations according to shear rate. Abbreviations are the same as in the legend to Figure 1. Mean±SE, n=5.
Hemodynamics and VH
Heart rate was unchanged whatever the condition (data not shown). At each time point, MAP was unmodified (Figure 4) with whole blood, Hb-Lv, and Hb-Hv, but a decrease in MAP was observed with Alb (P<0.01). As expected, ABF slightly decreased except for Alb, wherein we noted a significant increase compared with baseline values and the other situations (58%, P<0.001; Figure 5A). PR and VH remained unchanged in animals transfused with whole blood. PR decreased immediately after infusion of Alb as the result of increased ABF and decreased MAP, whereas VH remained unchanged. Hb-Lv induced an immediate and sustained increase in VH (>200%) despite unchanged PR (Figure 5B). Conversely, in the Hb-Hv condition, the VH increase (Figure 6) was delayed by 1 hour and was less compared with Hb-Lv (40% to 148%), whereas PR increased gradually (maximum increase after 2 hours, 94%).
Figure 4. MAP of acute hemodiluted situations as a function of time. Abbreviations are the same as in the legend to Figure 1. Mean±SE, n=5, †P<0.01.
Figure 5. Blood flow (A) and vascular resistance (B) of acute hemodiluted situations as a function of time. Abbreviations are the same as in the legend to Figure 1. Mean±SE, n=5, with £P<0.05, Alb vs WB; §P<0.05, Hb-Lv vs WB; ^PP<0.05, Hb-Lv vs Hb-Hv; ΔP<0.05, Alb vs Hb-Lv; *P<0.05, WB vs Hb-Hv.
Figure 6. Variation of VH of acute hemodiluted situations as a function of time. Abbreviations are the same as in the legend to Figure 1. Mean±SE, n=5, *P<0.05, †P<0.01, ‡P<0.001.
Discussion
The aim of this work was to discriminate between pharmacological (hemoglobin moiety–related) and mechanical (viscosity-related) effects of a stoma cell-free hemoglobin solution infused under conditions of hemodilution with either unchanged or decreased hematocrit (0.40 and 0.20 L/L, respectively). We evaluated the effects of an αα-Hb solution with high or low viscosity by measuring VH in rats subjected to a 50% exchange transfusion. The major finding of this study is that at similar hemoglobin concentration, a cell-free hemoglobin solution with high viscosity (Hb-Hv) was able to delay and attenuate vascular constriction compared with the same solution with low viscosity (Hb-Lv).
This was demonstrated by an immediate and sustained increase in VH after infusion of Hb-Lv, whereas Hb-Hv had no effect on this parameter in the acute phase after the injection. This result indicates that hemoglobin solutions with similar pharmacological characteristics (hemoglobin concentration, O2 affinity, nitric oxide affinity, methemoglobin content, P50, etc) might elicit different effects on vascular tone according to their mechanical properties (mainly viscosity and rheologic properties). In other words, in the acute phase of hemodilution, the intrinsic physical properties of hemoglobin solutions have an impact on the mechanisms regulating vascular homeostasis that blunt the pharmacological effects related to the heme moiety. Nevertheless, the lack of change in VH after injection of either solution devoid of stoma-free hemoglobin and, despite their opposite physical characteristics (low viscosity for albumin vs high viscosity for whole blood), makes it possible to rule out the only mechanical properties of the fluids in this phenomenon. Indeed, in addition to the immediate vascular impact of hemoglobin solutions due to their mechanical properties, their pharmacological properties exert a delayed effect, as exemplified by the progressive increase in VH observed in the Hb-Hv group at the latest time point.
A possible mechanism that might explain the limited vasoconstriction action of cell-free hemoglobin when viscosity was maintained in the physiological range is related to shear stress. Shear stress induced by hemodiluted blood flow on endothelial cells of the vessel wall is indeed responsible for the vasomotion. As previously described, one of the major factors of this regulation is the production of nitric oxide released from endothelial cells.13 Two mechanisms related to nitric oxide–dependant shear stress could be proposed to explain the beneficial action of Hb-Hv: first, a direct action due to viscosity itself, which remained unaltered (as shown in Figure 2) and therefore mimicked physiological conditions; and second, a possible extracellular diffusion barrier for hemoglobin of the plasmatic layer along the vessels (also known as the erythrocyte-free zone), responsible for lesser entrapping of nitric oxide by erythrocyte hemoglobin.14,15 The erythrocyte-free zone can influence the consumption of nitric oxide by erythrocytes and free hemoglobin16 Moreover, the vasoconstriction was observed as soon as free hemoglobin was injected into the bloodstream. This phenomenon had been widely described in the literature and could be due to endothelin-1,17,18 microcirculation,19 vasoconstriction,20 vascular reflex induced by the tissue O2 supply excess,21,22 or from a more discrete reaction, such as ATP-sensitive K+ channels or as a specific scavenger hemoglobin receptor.23,24 In fact, the basic common point between these possible actions on vasomotion was the presence of heme and that was consequently, a pure pharmacological effect. Finally, the observed vasoconstriction could be limited by high viscosity (that was demonstrated in this study), but it remained present and will appear only when viscosity decreased.
The low shear stress, partly reflecting erythrocyte aggregation and indicating potential rheologic modifications, is strongly dependent on the Hct level and hence, on hemodilution. The values of low and high viscosity were similar in matched experimental groups, suggesting that the presence of cell-free hemoglobin did not interfere with the in vitro erythrocyte behavior, whereas the blood dilution played a fundamental role. Despite the fact that it had no effects on blood viscosity measured in vitro, cell-free hemoglobin might impact vascular tone through other mechanisms; one of these mechanisms has been shown in our model by measuring VH in vivo.
Hindrance is an absolute indicator of vascular tone able to translate the action on blood circulation. It depends on mechanical and/or pharmacological effects of injected solutions into bloodstream on the vessel wall. However, it could be available mechanisms in the vasomotion induced by free hemoglobin in bloodstream, which are dependent on blood viscosity. If blood viscosity is maintained, the involvement of free hemoglobin is unclear, because vasoconstriction is delayed. These data demonstrate the beneficial effects of physiological viscosity, which limits vasoconstriction due to the pharmacological properties of hemoglobin.
The hemoglobin solution used in this study exhibits low intrinsic viscosity. However, future studies might include other hemoglobin solutions, particularly hemoglobin of high molecular weight, chemically modified to reduce the pharmacological effect (polymerized or cross-linked with macromolecules like dextran25,26 or polyethylene glycol27). Additionally, these high-molecular-weight solutions with a longer half-life than αα-Hb might reside in the circulation for longer than 60 minutes. Such fluids might limit the progressive fall in blood flow and increase in vascular resistance that were observed in the current study, ie, 2 hours after infusion of Hb-Hv. Indeed, such effects might lead to inadequate tissue perfusion and might become detrimental to the subject in extreme hemodilution conditions within a few hours, as observed in a phase III trials with DCLHb.28 As mentioned earlier, the acute hemodilution (50% blood volume) used in this study mimics an extreme clinical situation; it will be interesting to investigate lesser hemodilution levels leading to increased blood viscosity and limited cell-free hemoglobin pharmacological effects.
Perspectives
In conclusion, measurements of VH suggested that the viscosity-related mechanical effects of blood temporarily predominate the pharmacological effects elicited by cell-free hemoglobin in extreme dilution. Thus, maintaining physiological viscosity during hemodilution conditions with cell-free hemoglobin is critical to prevent hemoglobin-mediated vasoconstriction. These observations provide new insights for the use of hemoglobin solutions as substitute fluids in extreme hemodilution.
- Received December 23, 2003.
- Revision received January 16, 2004.
- Accepted February 5, 2004.
References
This Issue
Article Tools
Share this Article
- Hemodilution With Stoma-Free Hemoglobin at Physiologically Maintained Viscosity Delays the Onset of Vasoconstriction
