Secondary Logo

Journal Logo

Basic Science Aspects


Cabrales, Pedro*; Tsai, Amy G.*†; Intaglietta, Marcos

Author Information
doi: 10.1097/
  • Free



Severe hemorrhage is a major cause of morbidity and mortality after trauma (1, 2). The persisting depression of microvascular blood flow despite successful restoration of macrohemodynamics has been suggested to promote multiple organ failure after hemorrhagic shock (3, 4). The proper fluid to correct blood losses has been a long-standing subject of debate, particularly with regard to its colligative properties. Controversy surrounds the timing, type, extent, and pace of intravascular volume expansion without conclusive resolution. The key issues affecting the different hemorrhage models are the oxygen transport and the rheological properties needed to ensure adequate vital organ homeostasis. The hepatosplanchnic region is of particular importance because hypoperfusion of this vascular bed has been implicated in the development of multisystem organ failure (3, 5-7). The liver modulates the host defense response after shock and trauma, and hepatic failure predisposes multiple organ failure (8). Therefore, the instantaneous restoration of hepatic microvascular blood flow and oxygen supply should be a primary goal of shock treatment.

Organ blood flow is mostly determined by perfusion pressure and vascular resistance, the latter being also a function of blood viscosity (9). Furthermore, in vivo changes in viscosity may be accompanied by changes in vascular geometry due to autoregulatory processes driven by changes in the production of vasoactive mediators by the endothelium.

Volume restitution with plasma expanders and autotransfusion as a response to hemorrhage dilute the blood components. In particular, the dilution of red blood cells (RBCs) decreases the blood viscosity and, therefore, the viscosity-dependent component of peripheral vascular resistance (VR). Although it is generally perceived that the decrease in peripheral VR (within limits) is beneficial, the range of this effect is not well defined. Recent studies on the physiology of hemodilution have advanced the hypothesis that the functional lower limit in the decrease of RBC concentration is mainly determined by the decrease in blood viscosity rather than by the decrease in oxygen-carrying capacity (9). The rationale for this hypothesis originates from experimental studies in hemorrhagic shock, showing that a threshold of blood/plasma viscosity is required to maintain microvascular perfusion and, in particular, functional capillary density (FCD) (10-13). According to these studies, the maintenance of FCD in conditions of prolonged hemorrhagic shock differentiates between survival and nonsurvival, regardless of tissue Po2 (10). Clearly, hemoglobin (Hb) level and oxygen-carrying capacity become limited when oxygen delivery and metabolic oxygen needs are no longer maintained. Current studies suggest that this limit of intrinsic blood oxygen-carrying capacity is lower than that at the point when a blood transfusion is deemed necessary if the viscosity of blood remains sufficiently elevated (12, 13).

The present study tests the hypothesis that the restoration of systemic and microvascular conditions after hemorrhage followed by hypovolemic shock depends mostly on the blood rheological properties rather than in maintaining the oxygen-carrying capacity. To test this hypothesis, we used the hamster window chamber model subjected to a 50% blood volume loss for 1 h and provided resuscitation by transfusing 25% of blood volume of RBCs suspended in plasma (hematocrit [Hct] level, 0.30) in which Hb was converted to methemoglobin and, therefore, did not increase the oxygen-carrying capacity. Resuscitation in a group using identical volume transfusions with fully functional RBCs and in a second group using the same volume infusion of plasma was used as control. We measured the blood flow distribution to the different organs to determine if there were changes related to the different resuscitation modalities. These measurements also allowed us to test the hypothesis that the microvascular events measured in the window chamber model correspond to the hemodynamic changes in the major organs.


Animal preparation

Investigations were performed in 55- to 65-g male golden Syrian hamsters (Charles River Laboratories, Boston, Mass) fitted with a dorsal skinfold chamber window. Animal handling and care followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The experimental protocol was approved by the local animal care committee. The hamster chamber window model is widely used for microvascular studies in the unanesthetized state, and the complete surgical technique is described in detail in other references (14, 15). The window chamber preparation was given at least 2 days for recovery before the preparation was assessed under the microscope for any signs of edema, bleeding, or unusual neovascularization. The animals were anesthetized again, and arterial and venous catheters filled with a heparinized saline solution (dose, 30 IU/mL) were implanted. Catheters were tunneled under the skin, were exteriorized at the dorsal side of the neck, and were securely attached to the window frame. The microvasculature was examined 3 to 4 days after the initial surgery, and only animals passing an established systemic and microcirculatory inclusion criteria, including having tissue void of low perfusion, inflammation, and edema, were enrolled in the study (16).

Inclusion criteria

Animals were suitable for the experiments if (1) the systemic parameters were within the reference range (i.e., heart rate greater than 340 beats per minute, mean arterial blood pressure [MAP] greater than 80 mmHg, systemic Hct level greater than 0.45, and arterial oxygen partial pressure [PaO2] greater than 50 mmHg); and (2) microscopic examination of the tissue in the chamber (magnification, ×650) did not reveal signs of edema or bleeding. Hamsters are a fossorial species with a lower arterial PO2 than do other rodents because of their adaptation to the subterranean environment. However, the microvascular PO2 distribution in the chamber window model is the same as in other rodents, such as mice (17).

Acute hemorrhage and volume replacement protocol

Acute hemorrhage was induced by means of withdrawal of 50% of estimated total blood volume via the carotid artery catheter within 5 min. Total blood volume was estimated as 7% of body weight. One hour after hemorrhage induction, the animals received a single-volume infusion of 50% of the shed blood volume (25% of blood volume) of the resuscitation fluid (Experimental Groups section) within 10 min via the jugular vein catheter. Therefore, the restoration of 25% of the blood volume does not cause hypervolemia, reinstating normovolemia in the hamster window model. The animals did not receive any additional fluid during the experiment. Fifty percent of shed blood volume was used as the resuscitation volume because autotransfusion restores about half of the shed volume during the shock period. Therefore, the restoration of 25% of the blood volume does not cause hypervolemia and reinstates normovolemia in the hamster window model. The parameters were analyzed before hemorrhage (baseline), after hemorrhage (shock), and up to 90 min after volume replacement (resuscitation).

Experimental groups

Before the experiment, the animals were randomly divided into the following three experimental groups:

  • Plasma: resuscitation was performed with fresh plasma;
  • Methemoglobin-loaded fresh RBC (MetRBC): resuscitation was performed with fresh RBCs, where Hb has been converted to MetHb by sodium nitrite exposition and suspended in fresh plasma at Hct level of 0.30; and
  • RBC: resuscitation was performed with fresh RBCs suspended in fresh plasma at Hct level of 0.30. Table 1 lists the physical characteristics of the three study groups.
Table 1
Table 1:
Properties of resuscitation materials

Systemic parameters

The MAP and the heart rate were recorded continuously (MP150; Biopac Systems, Santa Barbara, Calif). The level of Hct was measured from centrifuged arterial blood samples taken in heparinized capillary tubes (Hct level, 25 μL [~50% of the heparinized glass capillary tube is filled]). The Hb content was determined spectrophotometrically from a single drop of blood (B-Hemoglobin; Hemocue, Stockholm, Sweden).

Blood chemistry and biophysical properties

Arterial blood was collected in heparinized glass capillaries (capacity, 50 μL) and immediately analyzed for PaO2, PaCO2, base excess, and pH (Blood Chemistry Analyzer 248; Bayer, Norwood, Mass). The comparatively low Pao2 and high Paco2 levels of these animals are a consequence of their adaptation to a fossorial environment. Blood samples for viscosity and colloid osmotic pressure measurements were quickly withdrawn from the animals into a heparinized 5-mL-capacity syringe at the end of the experiment for immediate analysis, or were refrigerated for next-day analysis. Viscosity was measured in a DV-II+ (Brookfield Engineering Laboratories, Middleboro, Mass) cone/plate viscometer with a CPE-40 cone spindle at a shear rate of 160 s−1. Colloid osmotic pressure was measured using a model 4420 colloid osmometer (Wescor, Logan, Utah).

Measurement of cardiac output

Cardiac output (CO) was measured using a modified thermodilution technique (18) in a different group of animals not used for microvascular studies because of the complexity of the setup and the difficulty of positioning the instrumented animal on the microscope. These animals were characterized in terms of systemic parameters to guarantee that they presented the same characteristics as those used in the microcirculatory studies. The animals assessed for CO were randomly assigned to different experimental groups.

Vascular resistance and hindrance calculations

Vascular resistance was calculated as the ratio between MAP and CO. The effect of vascular geometry on flow is termed vascular hindrance (VH) and is calculated as the ratio between VR and viscosity. Cardiac output was measured at baseline, at 50 min after the start of shock, and at 60 and 90 min after the start of resuscitation.

Measurement of blood flow distribution

Fluorescent microspheres (Molecular Probes, Eugene, Ore), 15 μm in diameter and of four different colors (green, yellow, red, and scarlet), were suspended in saline with 0.02% Tween-80. Microspheres of each color (n = 1 × 105) were injected 15 to 20 min after each exchange. Before injection, all fluorescent microspheres were vortexed and mixed in the syringe to obtain a homogenous concentration. The microspheres were injected for 30 s (volume, 100 μL). The animals were killed with a lethal dose of sodium pentobarbital; then, 8 vital organs (brain, heart, lung, liver, kidney, spleen, stomach, and intestine) and the window chamber were removed.

To estimate the number of fluorescent microspheres in each organ, the quantitative relationship between the numbers of fluorescent microspheres was determined from the reference blood samples filtered using Perkin-Elmer centrifugal devices (19). Dyes from the retained fluorescent microspheres were eluted with 2 mL of Cellosolve (Aldrich Chemical, Milwaukee, Wis), and the fluorescent signals were determined by an automated fluorescent system spectrofluorometer, FluoroMax-2 (HORIBA, Ltd, Japan) (19). Traditionally, the number of microspheres lodging in organ pieces (fluorescent dye recovered) estimates the blood flow distribution of each organ (20, 21).

Microvascular experimental setup

The unanesthetized animals were placed in a restraining tube with a longitudinal slit from which the window chamber protruded, and then were fixed to the microscopic stage of a transillumination intravital microscope (BX51WI; Olympus, New Hyde Park, NY). The animals were given 20 min to adjust to the change in the tube environment before the measurements were conducted. The tissue image was projected onto a charge-coupled device camera (COHU 4815), connected to a videocassette recorder, and viewed on a monitor. Measurements were performed using a ×40 (LUMPFL-WIR, numerical aperture 0.8; Olympus) water immersion objective. The same sites of study were followed throughout the experiment so that comparisons could be made directly to baseline levels.

Functional capillary density

Functional capillaries, defined as those capillary segments that have RBC transit of at least a single RBC in a 30- to 45-s period in 10 successive microscopic fields, were assessed, totaling to a region of 0.46 mm2. Each field had between two and five capillary segments with RBC flow. The FCD per centimeter (i.e., total length of RBC perfused capillaries divided by the area of the microscopic field of view) was evaluated by measuring and adding the length of the capillaries that had RBC transit in the field of view. The relative change in FCD from baseline levels after each intervention is indicative of the extent of capillary perfusion (22, 23).


Arteriolar and venular blood flow velocities were measured online by using the photodiode cross-correlation method (24) (Photo Diode/Velocity Tracker model 102B; Vista Electronics, San Diego, Calif). The measured centerline velocity (V) was corrected according to vessel size to obtain the mean RBC velocity (25). A video image-shearing method was used to measure the vessel diameter (D) (26). Blood flow (Q) was calculated from the measured values as Q = π × V (D/2)2. Changes in arteriolar and venular diameter from baseline were used as indicators of a change in vascular tone. This calculation assumes a parabolic velocity profile and has been found applicable to tubes of 15- to 80-μm internal diameters and to Hct levels in the range of 0.06 to 0.60 (25). Wall shear stress was defined by the equation WSS = WSR × η, where WSS is the wall shear stress, WSR is the wall shear rate given by 8VD−1, and η is the microvascular blood viscosity or plasma viscosity.

Microvascular PO2 distribution

High-resolution, noninvasive microvascular PO2 measurements were made using phosphorescence-quenching microscopy (PQM) (16, 27), which is based on the oxygen-dependent quenching of phosphorescence emitted by albumin-bound metalloporphyrin complex after pulsed light excitation. Phosphorescence-quenching microscopy is independent of the dye concentration within the tissue and is well suited for detecting hypoxia because its decay time is inversely proportional to the PO2 level, causing the method to be more precise at low PO2 levels. This technique is used to measure both intravascular and extravascular PO2 levels because the albumin-dye complex continuously extravasates the circulation into the interstitial tissue (16, 27). Tissue PO2 level was measured in tissue regions in-between functional capillaries. The PQM allows for precise localization of the PO2 measurements without subjecting the tissue to injury. The PO2 measurements were performed only at 90 min after resuscitation, at the last observation time point. These measurements provide a detailed understanding of microvascular oxygen distribution and indicate whether oxygen is delivered to the interstitial areas.

Oxygen delivery and extraction

The microvascular methodology used in our studies allows for a detailed analysis of oxygen supply in the tissue. Calculations are made using equations 1 and 2 (22):

where RBCHb is the hemoglobin in RBCs (in grams of Hb per deciliter of blood); γ is the oxygen-carrying capacity of saturated hemoglobin (1.34 mL O2/gHb); SA% is the arteriolar oxygen saturation; 1 − Hct is the fractional plasma volume (in deciliter of plasma per deciliter of blood); α is the solubility of oxygen in plasma (3.14 × 10−3 mL O2/dLplasma mmHg); PO2A is the arteriolar partial pressure of oxygen; AV indicates the arteriolar/venular differences; and Q is the microvascular flow. The oxygen saturation for the hamster RBCs have been published previously (28).

Organ Oxygen Supply

Calculation of oxygen supply is performed using the equation

where Hbblood is the Hb in blood (in grams of Hb per deciliter of blood); SAO2% is the arteriolar oxygen saturation; Qorgan is the flow distribution fraction for each organ; and CO is cardiac output.

Preparation of RBCs containing methemoglobin

Fresh RBCs and plasma were collected from the donor animal before the start of the experiment. Blood was centrifuged, supernatant plasma was removed and stored, and RBC buffy coat was discarded. The RBCs were transferred into tubes, resuspended in an equivalent amount of isotonic sodium chloride solution, and mixed gently for 2 min with sodium nitrite (100 μL of 1 mol/L sodium nitrite per 5 milliliters of RBCs). The cells were then centrifuged at 2,100g, washed three times with 5 mL of heparinized saline, and stored as packed cells at 4°C; blood was used within 4 h of collection. Aliquots of these cells were tested, and only those cells with 0.95 to 1.0 metHb were used. The RBCs containing metHb were resuspended in fresh plasma to produce a 0.30 Hct level (concentration, 9.6 gHb/dL). The RBC manipulation (pipette, reaction, and transfer) was performed in a laminar flow hood for sterility.

Data analysis

Results are presented as mean ± SD. Data within each group were analyzed using analysis of variance for repeated measurements (Kruskal-Wallis test). When appropriate, post hoc analyses were performed using the Dunn test. Microhemodynamic measurements were compared with baseline levels obtained before the experimental procedure. The box-whisker plot separates the data into quartiles, with the top of the box defining the 75th percentile, the line within the box displaying the median, and the bottom of the box showing the 25th percentile. The upper whisker defines the 95th percentile; the lower whisker, the 5th percentile. Microhemodynamic data are presented as absolute values and ratios relative to baseline values. A ratio of 1.0 signifies no change from baseline, whereas lower and higher ratios are indicative of changes proportionally lower and higher than baseline (i.e., 1.5 would mean a 50% increase from the baseline level). The same vessels and functional capillary fields were followed so that direct comparisons with their baseline levels could be performed, allowing for more robust statistics for small sample populations. Pearson product moment correlation was used to test the linear correlation between microvascular flow and organ blood flow. All statistics were calculated using GraphPad Prism 4.01 (GraphPad Software, Inc, San Diego, Calif). Changes were considered statistically significant if P < 0.05.


Thirty animals were entered into this study, and all tolerated the entire acute hemorrhage shock resuscitation protocol without visible signs of discomfort.

Blood typing and crossmatching tests are not necessary with the hamster on the basis of previous experience with the species; no changes in body temperature (as immune responses against inappropriate transfusion) were detected during the protocol. The animals were assigned randomly to the experimental groups: Plasma (n = 6 for microcirculation and n = 4 for organ flow distribution and CO); MetRBC (n = 6 for microcirculation and n = 4 for organ flow distribution and CO); and RBC (n = 6 for microcirculation and n = 4 for organ flow distribution and CO). Systemic and microhemodynamic data sets for baseline and shock were obtained by combining the data from all experimental groups.

Systemic parameters

Table 2 presents systemic and gas parameters measured during the hemorrhagic shock resuscitation protocol. All groups showed a significant reduction in Hct level after hemorrhagic shock (0.30 ± 0.01, 50 min after hemorrhage) from baseline (0.49 ± 0.01). Resuscitation further decreased the Hct level to only 0.24 ± 0.01 for the plasma group (P < 0.05 to baseline and shock). The level of Hct in the MetRBC and RBC after restoration of blood volume was maintained at the same level as during shock (0.29 ± 0.01). Hemoglobin showed the same trend as Hct, decreasing from baseline (Hb level, 15 ± 1 gHb/dL) to 10 ± 1 gHb/dL after hemorrhage and shock. Resuscitation with plasma decreased the total Hb levels (7 ± 1 gHb/dL), and resuscitation with MetRBC or RBC maintained the total Hb levels reached during hemorrhage and shock (Hb level, 10 ± 1 gHb/dL). The MetRBC resuscitation introduced 2.2 ± 0.3 gHb/dL of MetHb, which was 24% ± 7% of the circulating RBCs, consequently reducing the oxygen-carrying capacity during resuscitation by 25% when compared with normal RBC resuscitation.

Table 2
Table 2:
Laboratory parameters

Systemic and microvascular hemodynamic results are presented in detail on Table 2 and Table 3. In brief, hemorrhage and shock decreased MAP from 109 ± 6 mmHg to 40 ± 6 mmHg. Resuscitation with plasma partially restored MAP at 60 and 90 min after resuscitation to 66 ± 8 and 63 ± 6 mmHg, respectively, both statistically significant below that of the transfused groups. For resuscitation with RBCs mixed with plasma, MAP was statistically higher than resuscitation with plasma at 60 and 90 min after transfusion. The MetRBC resuscitation after 60 and 90 min restored MAP to 87 ± 7 and 79 ± 7 mmHg, respectively. Similar results were obtained with oxygen functional RBC, which restored MAP to 85 ± 8 and 78± 6 mmHg after 60 and 90 min, respectively. In all groups, MAP was statistically lower than baseline. Heart rate was not different from baseline at any time point, mostly because of the large variability in this model.

Table 3
Table 3:
Cardiac output and vascular resistance

The gas laboratory parameter and the calculated acid-base balance were parallel to the restoration of MAP. Resuscitation with plasma partially recovered the systemic parameters. Transfusion with functional or oxygen-inactivated RBCs consistently provided better restoration of systemic parameters than did plasma (Table 2). The systemic hemodynamic parameters (Table 3) show that vascular resistance and CO decreased after hemorrhage and shock. Resuscitation with plasma did not restore CO, whereas, in the transfused groups, CO was restored to baseline levels. Vascular resistance was reduced after hemorrhage and was below baseline after resuscitation. Vascular hindrance was only increased after resuscitation with plasma. In the transfused groups, the changes in viscosity and flow were counteracted by changes in vascular geometry due to autoregulatory processes.

Physical properties of blood

Table 2) compares the blood rheological properties and the COP 90 min after resuscitation. Blood viscosities were statistically reduced from baseline in all experimental groups. In contrast, plasma viscosities were preserved. The nondiluted blood values of rheological properties and colloid osmotic pressures were also obtained from hamsters that did not undergo the shock protocol.

Microvascular measurements are presented in Figure 1. Arteriolar and venular diameters statistically decreased when compared with the transfused groups. Arteriolar and venular flows were also statistically lesser in the plasma group compared with those in the transfused groups because of vascular constriction. The differences between volume restitution (plasma) and transfusion with fully functional or oxygen-inactivated RBCs were more pronounced at 90 min after resuscitation. Unexpectedly, the differences in the oxygen-carrying capacity in the transfused groups did not produce differences in vascular tone. Changes in capillary perfusion during the protocol are presented in Figure 2. The FCD was dramatically reduced after hemorrhage and shock. Resuscitation partially restored FCD in all groups. The differences in FCD between groups were observed at 60 min, when the FCDs of plasma-resuscitated animals were statistically lower than those in the transfused groups. At 90 min, the FCD differences among groups evidenced a more significant recovery for the transfused animals (P < 0.01).

Fig. 1
Fig. 1:
Relative changes to baseline in arteriolar and venular hemodynamics for plasma, MetRBC, and RBC. Broken line represents baseline level. The mean diameters ± SDs (in micrometer) of arteriolar (A) and venular (B) hemodynamics at baseline for each animal group were as follows (n = number of vessels studied): (1) plasma (arterioles, 56.8 ± 8.2 [n = 42]; venules, 57.7 ± 9.3 [n = 45]); (2) MetRBC (arterioles, 57.0 ± 8.8 [n = 40]; venules, 59.3 ± 10.4 [n = 48]); (3) RBC (arterioles, 58.2 ± 9.1 [n = 44]; venules, 59.7 ± 10.3 [n = 47]). The mean RBC velocities ± SDs (in millimeter per second) of arteriolar and venular hemodynamics at baseline for each animal group were as follows: plasma (arterioles, 4.6 ± 0.9; venules, 2.7 ± 0.7); MetRBC (arterioles, 4.5 ± 0.8; venules, 2.5 ± 1.0); RBC (arterioles, 4.4 ± 1.0; venules, 2.3 ± 0.8). The mean calculated flows ± SDs (in nanoliter per second) of arteriolar (C) and venular (D) hemodynamics at baseline for each animal group were as follows: plasma (arterioles, 11.9 ± 3.8; venules, 7.1 ± 2.8); MetRBC (arterioles, 12.4 ± 3.6; venules, 7.0 ± 2.3); RBC (arterioles, 10.6 ± 3.2; venules, 6.8 ± 2.2). P < 0.05 relative to baseline; *P < 0.05 among groups.
Fig. 2
Fig. 2:
Effects of plasma viscosity on capillary perfusion during hemodilution. The FCD was unchanged after level 1 exchange and was lower after resuscitation with plasma compared with volume restitution with RBCs. The FCD per centimeter ± SD at baseline was as follows: plasma, 123 ± 12; MetRBC, 116 ± 14; and RBC, 115 ± 17. * P < 0.05 and **P < 0.01 among groups.

Microvascular oxygen tensions are presented in Figure 3, showing that the differences in oxygen-carrying capacity were reflected in the oxygen distribution. Improved perfusion in the group transfused with oxygen-inactivated RBCs, when compared with plasma-only resuscitation, was evident in the measurement of tissue venular Po2, which represents the oxygen reserve (excess) in the tissue. Calculated oxygen delivery and extraction levels are presented in Figure 4. Resuscitation with plasma and with oxygen-inactivated RBCs had lower oxygen delivery and extraction levels compared with transfusion with oxygen functional RBCs.

Fig. 3
Fig. 3:
Microvascular oxygen partial pressure (arterioles, venules, and tissue) 90 min after resuscitation from hemorrhagic shock. * P < 0.05.
Fig. 4
Fig. 4:
Arteriolar oxygen delivery and extraction 90 min after resuscitation. Calculations of global oxygen transport are not directly measurable in our model. However, the changes relative to baseline can be calculated using the measured parameters. The extraction was calculated as the difference of averaged arterioles and venules for each animal. The difference in oxygen delivery and extraction between plasma and MetRBC are not statistically significant. * P < 0.05.

Organ blood flow is presented in Figure 5. Major differences among groups were observed for the liver and the spleen. Organ perfusion was statistically improved for transfused animals compared with those resuscitated with plasma as early as 60 min after volume restoration. Microvascular flow in the window chamber tissue was statistically significantly correlated with the brain (R2 = 0.78; Pearson r = 0.88; P < 0.005), heart (R2 = 0.39; Pearson r = 0.62; P = 0.06), and kidney (R2 = 0.59; Pearson r = 0.78; P < 0.05) blood flows.

Fig. 5
Fig. 5:
Organ blood flow. Distribution of the CO was measured by estimating the number of recovered fluorescent microspheres from the vital organs and from the window chamber.

Oxygen supply to the different organs presented in Figure 6 was calculated using the data on organ blood flow, Hb, arterial blood PO2, and CO. Organ oxygenation was driven mostly by perfusion and by the oxygen-carrying capacity. Transfusion with oxygen functional RBCs during resuscitation restored oxygenation to levels close to baseline in most organs.

Fig. 6
Fig. 6:
Organ oxygen delivery. Regional differences in the circulatory responses during hemodilution may not be adequately reflected by measurements of global oxygen delivery and oxygenation status in individual tissues and organs. * P < 0.05 and ⊗P < 0.10 among groups; P < 0.05 compared with baseline. The oxygen delivery measurements for baseline (in milliliters of oxygen per minute) were as follows: brain, 3.9 ± 0.2; heart, 2.6 ± 0.1; kidney, 7.1 ± 0.5; liver, 3.1 ± 0.2; lung, 2.4 ± 0.2; spleen, 1.6 ± 0.2; and window chamber, 0.9 ± 0.1.


The principal finding of this study is that hemorrhagic shock resuscitation by transfusion of RBCs provides improved restoration of both systemic and microhemodynamic parameters compared with volume restoration using plasma, independent of whether RBCs carry oxygen or not. Inasmuch as there are no physical differences between oxygen-carrying and non-oxygen-carrying RBCs-the difference between oxygen and non-oxygen-carrying RBCs resuscitation-the uncompleted resuscitation with plasma must be due to alterations in blood rheological properties and not due to changes in oxygen-carrying capacity. Our results suggest that the principal factor in ensuring hemodynamic restoration by RBCs, besides volume restitution, is the restoration of blood viscosity, for the levels of oxygen carrying capacity loss tested in this study (decrease in intrinsic oxygen carrying capacity by 40% of baseline).

It should be noted that MetHb remained elevated for a long period, although there are multiple mechanisms available within RBCs to reverse the state of methemoglobinemia. Hemoglobin oxidation to MetHb by sodium nitrite occurs at a rate that is characteristically different for each animal species (29), and the reduction of MetHb presents considerable species variation (30). Many species have a strong reductase system for Hb that is able to maintain Hb in a nonoxidized state or reduce it when oxidation occurs. Hamsters, however, do not reduce MetHb in a very efficient manner, as was found in this study and during preliminary experiments. In other situations when MetHb increased significantly in the hamster, the reduction of MetHb to OxyHb was obtained using methylene blue (9).

Previous work on microcirculatory changes during hemorrhagic shock and resuscitation assumed that the hemodynamic and metabolic events observed in the microcirculation reflected responses in vital organs (12, 13, 31, 32). This assumption was confirmed by the present findings showing a strong correlation between the window chamber (microcirculation) and the flow distribution presented in studied vital organs. Flow restoration was at the lowest level after resuscitation with plasma, and at the highest after transfusion with fresh RBCs; transfusion with non-oxygen-functional RBCs was at the intermediate level. Hemodynamic changes related to surgical preparation and repetitive assessment of CO and organ flow distribution using fluorescent microspheres were reported before and were not statistically different from baseline measurements of the actual study (33). In general, the conclusions drawn for the baseline measurements can also be applied to actual results, when compared to measurements previously reported (33).

The flow in the window chamber was significantly decreased during shock and, after resuscitation, did not return to baseline levels regardless of the fluid used, whether assessed in terms of microvascular measurements or of the microspheres flow distribution technique. Therefore, the hemodynamic analysis of the window chamber seems to present a worst-case scenario, whereas the macro events in other organs tend to show improved restoration versus findings in the microcirculation. The present study shows that the microvascular outcome evidenced in the window chamber could be indicative of related and similar effects in the brain, heart, and kidney because these main organs showed parallel changes in perfusion when the hamster was subjected to hemorrhage resuscitation with and without restoring the oxygen-carrying capacity. The finding that the correlation coefficients were statistically significant (correlating the changes in perfusion in the window tissue and in these organs) supports the hypothesis of similar autoregulatory responses to changes in blood properties. Notably, liver and spleen seemed to be underperfused, independent of the resuscitation, which correlated with the window chamber findings.

The role of blood viscosity in maintaining systemic blood pressure and blood gases are highlighted by the results obtained because transfusion of RBCs, regardless of oxygen-carrying capacity, provided consistent normalization of systemic parameters when compared with resuscitation with plasma. The acid-base balance after volume restitution with plasma showed a weak trend of recovery when compared with RBC transfusion. The beneficial effects of maintaining blood viscosity after resuscitation from hemorrhage were similar to results found using high-viscosity plasma expanders to enhance the blood rheological properties without adding Hct in a similar protocol (12, 13).

Transfusion of RBCs provided a 25% increase in whole blood viscosity relative to resuscitation with plasma. This difference may account for the increase in MAP, organ perfusion, microvascular perfusion, and FCD, and should be because of the factors related to the normalization of shear stress. The shear stress exerted by blood moving near the endothelial surface modulates the release of dilatory autocoids (prostacyclin, nitric oxide) affecting vessel diameter. A previous study by Tsai et al. (34) showed that increased shear stress was associated with an increase in the measured concentration of perivascular nitric oxide, with concomitant vasodilator effects. Another possibility is that the viscous drag exerted is sensed via the glycocalyx, triggering the activation of endothelial mechanisms (35). Even in vital organs such as the heart, where ion channels are operative, the nonselective channels could be activated by stretching the cell membrane. Inwardly rectifying K+ channels activated by fluid shear stress have been described in cell culture (36, 37). Current studies show that if blood viscosity is severely decreased by hemodilution, microvascular function is impaired, and tissue survival is jeopardized because of the local microscopic maldistribution of blood flow and the resulting deficit in oxygen delivery (9, 16). In hemorrhagic shock, a limit is reached when hemodilution is no longer able to maintain the metabolic needs of the tissue. However, microvascular studies in extreme hemodilution show that this limit could be significantly lower than that of the current transfusion triggers if blood viscosity is partially maintained during hemodilution by increasing the plasma or the whole blood viscosity (9, 16, 22, 38).

Macrohemodynamic effects that reflect the cellular responses to shear preservation are evident from the calculation of vascular hindrance (Table 3), which is unchanged from baseline for both RBC transfusion groups. Conversely, plasma resuscitation showed a 25% increase in vascular hindrance, indicating that vascular resistance independent of the effect of blood viscosity is greater for plasma resuscitation, suggesting the presence of vascular constriction. In accordance with previous findings in the microcirculation, major decreases in FCD resulted from a decrease in capillary pressure as a consequence of vasoconstriction (22, 39). Notably, this effect is independent of oxygen-carrying capacity because oxygen tension and oxygen delivery were significantly improved by the use of fresh RBCs.

The restoration of oxygen-carrying capacity with fresh RBCs led to the near normalization of oxygen delivery to the brain, heart, kidney, and lung. However, a significant oxygen deficit remained in the spleen, liver, and window chamber after 90 min after resuscitation. This effect may be due, in part, to the incomplete restoration of the oxygen-carrying capacity. In view of current findings in studies of extreme hemodilution with high-viscosity plasma expanders (9, 33), it is likely that a significant additional improvement may be obtained by increasing the plasma viscosity by suspending RBCs in a viscogenic fluid.

Blood conserved by conventional means for transfusion carries a limited amount of oxygen upon introduction into the circulation. Oxygen transport by transfused RBCs begins several (2-5) hours later (40). Consequently, a conventional blood transfusion (using stored blood) may not fully restore the oxygen-carrying capacity in acute conditions. However, it restores blood volume and blood viscosity. Blood transfusions have immediate subjective, physiological, and clinical beneficial effects, which are not fully explained by the restoration of oxygen-carrying capacity because this occurs as much as several hours later, depending on the storage period of the transfused blood. Recent studies showed that an increase in Hct level in a healthy organism led to a rapid increase in nitric oxide production via increased shear stress (41). Similar effects were obtained when the Hct level decreased via hemodilution and when the plasma viscosity increased, raising shear stress and consequently augmenting the levels of microvascular perivascular nitric oxide, producing a stable and homogeneously perfused microcirculation (34). Therefore, the beneficial effect of blood transfusion may be linked, in part, to the increase or restoration of shear stress and mechanotransduction by blood viscosity.

Although improved heart function may be one of the causes for the difference between RBC and plasma reperfusion, the improvement of cerebral perfusion via sympathetic outflow-mediated processes may be involved in the regulation of heart function. Central NO bioavailability restrains sympathetic outflow. Several of our studies suggest that increased blood and plasma viscosity lead to increased NO production via the increase in shear stress and the corresponding increased production of NO. Therefore, a difference between RBC- and plasma-reperfused animals could be due to the central effects in the brain that reduce the NO bioavailability, an effect that would lead to vasoconstriction via decreased inhibition to overall sympathetic outflow (34, 41, 42).

In conclusion, this study shows that the restoration of homeostatic conditions after hemorrhagic shock resuscitation requires the restoration of blood rheological properties and oxygen-carrying capacity. Furthermore, in the moderate hemorrhagic shock conditions analyzed in this study, the restoration of only the rheological properties seems to be as effective as restoring the rheological conditions and the oxygen-carrying capacity. The restoration of volume using a fluid with the rheological properties of blood that does not carry oxygen yields the same results as resuscitation by transfusion of oxygen-functional RBCs. This result explains why the transfusion of stored RBCs, which does not necessarily raise the effective capacity of blood to transport oxygen upon transfusion, still provides beneficial effects: the transfusion is effective in restoring microvascular perfusion, which is a crucial factor for oxygen delivery and for flushing out the metabolites produced during hypovolemic conditions. Furthermore, it seems that anemia caused by a 50% hemorrhage can be effectively corrected by the restoration of the viscogenic properties of the circulating blood rather than the restoration of the oxygen-carrying capacity.


The authors thank Froilan P. Barra and Cynthia Walser for the surgical preparation of the animals.


1. Chapler CK, Cain SM, Stainsby WN: The effect of hyperoxia on oxygen uptake during acute anemia. Can J Physiol Pharmacol 62:809-814, 1984.
2. Sauaia A, Moore FA, Moore EE, Moser KS, Brennan R, Read RA, Pons PT: Epidemiology of trauma deaths: a reassessment. J Trauma 38:185-193, 1995.
3. Wang P, Ba ZF, Burkhardt J, Chaudry IH: Measurement of hepatic blood flow after severe hemorrhage: lack of restoration despite adequate resuscitation. Am J Physiol 262:G92-G98, 1992.
4. Messmer K, Kreimeier U: Microcirculatory therapy in shock. Resuscitation 18(Suppl):S51-S61, 1989.
5. Razani B, Engelman JA, Wang XB, Schubert W, Zhang XL, Marks CB, Macaluso F, Russell RG, Li M, Pestell RG, et al.: Caveolin-1 null mice are viable but show evidence of hyperproliferative and vascular abnormalities. J Biol Chem 276:38121-38138, 2001.
6. Remmers DE, Wang P, Cioffi WG, Bland KI, Chaudry IH: Testosterone receptor blockade after trauma-hemorrhage improves cardiac and hepatic functions in males. Am J Physiol 273:H2919-H2925, 1997.
7. Ayala A, Tu Y, Flye MW, Chaudry IH: Depressed splenic function after hemorrhage results from gastrointestinal tract stimulation of hepatic-mediator release. Correction with portacaval shunt. Arch Surg 131:1209-1214, 1996. Discussion 1215.
8. Rivkind AI, Siegel JH, Dunham CM: Patterns of organ injury in blunt hepatic trauma and their significance for management and outcome. J Trauma 29:1398-1415, 1989.
9. Cabrales P, Martini J, Intaglietta M, Tsai AG: Blood viscosity maintains microvascular conditions during normovolemic anemia independent of blood oxygen-carrying capacity. Am J Physiol Heart Circ Physiol 291:H581-H590, 2006. Epub 2006 Mar 3.
10. Kerger H, Saltzman DJ, Menger MD, Messmer K, Intaglietta M: Systemic and subcutaneous microvascular Po2 dissociation during 4-h hemorrhagic shock in conscious hamsters. Am J Physiol 270:H827-H836, 1996.
11. Matheson B, Kwansa HE, Bucci E, Rebel A, Koehler RC: Vascular response to infusions of a nonextravasating hemoglobin polymer. J Appl Physiol 93:1479-1486, 2002.
12. Cabrales P, Intaglietta M, Tsai AG: Increased plasma viscosity sustains microcirculation after resuscitation from hemorrhagic shock and continuous bleeding. Shock 23:549-555, 2005.
13. Cabrales P, Tsai AG, Intaglietta M: Hyperosmotic-hyperoncotic versus hyperosmotic-hyperviscous small volume resuscitation in hemorrhagic shock. Shock 22:431-437, 2004.
14. Colantuoni A, Bertuglia S, Intaglietta M: Quantitation of rhythmic diameter changes in arterial microcirculation. Am J Physiol 246:H508-H517, 1984.
15. Endrich B, Asaishi K, Goötz A, Messmer K: Technical report: a new chamber technique for microvascular studies in unanaesthetized hamsters. Res Exp Med 177:125-134, 1980.
16. Tsai AG, Friesenecker B, McCarthy M, Sakai H, Intaglietta M: Plasma viscosity regulates capillary perfusion during extreme hemodilution in hamster skin fold model. Am J Physiol 275:H2170-H2180, 1998.
17. Cabrales P, Tsai AG, Frangos JA, Intaglietta M: Role of endothelial nitric oxide in microvascular oxygen delivery and consumption. Free Radic Biol Med 39:1229-1237, 2005.
18. Cabrales P, Acero C, Intaglietta M, Tsai AG: Measurement of the cardiac output in small animals by thermodilution. Microvasc Res 66:77-82, 2003.
19. Raab S, Thein E, Harris AG, Messmer K: A new sample-processing unit for the fluorescent microsphere method. Am J Physiol 276:H1801-H1806, 1999.
20. Schimmel C, Frazer D, Huckins SR, Glenny RW: Validation of automated spectrofluroimetry for measurement of regional organ perfusion using fluorescent microspheres. Comput Methods Programs Biomed 62:115-125, 2000.
21. Heymann MA, Payne BD, Hoffman JI, Rudolph AM: Blood flow measurements with radionuclide-labeled particles. Prog Cardiovasc Dis 20:55-79, 1977.
22. Cabrales P, Tsai AG, Intaglietta M: Microvascular pressure and functional capillary density in extreme hemodilution with low and high plasma viscosity expanders. Am J Physiol Heart Circ Physiol 287:H363-H373, 2004.
23. Tsai AG, Cabrales P, Winslow RM, Intaglietta M: Microvascular oxygen distribution in awake hamster window chamber model during hyperoxia. Am J Physiol Heart Circ Physiol 285:H1537-H1545, 2003.
24. Intaglietta M, Silverman NR, Tompkins WR: Capillary flow velocity measurements in vivo and in situ by television methods. Microvasc Res 10:165-179, 1975.
25. Lipowsky HH, Zweifach BW: Application of the "two-slit" photometric technique to the measurement of microvascular volumetric flow rates. Microvasc Res 15:93-101, 1978.
26. Intaglietta M, Tompkins WR: Microvascular measurements by video image shearing and splitting. Microvasc Res 5:309-312, 1973.
27. Kerger H, Groth G, Kalenka A, Vajkoczy P, Tsai AG, Intaglietta M: Po2 measurements by phosphorescence quenching: characteristics and applications of an automated system. Microvasc Res 65:32-38, 2003.
28. Cabrales P, Nacharaju P, Manjula BN, Tsai AG, Acharya SA, Intaglietta M: Early difference in tissue pH and microvascular hemodynamics in hemorrhagic shock resuscitation using polyethylene glycol-albumin- and hydroxyethyl starch-based plasma expanders. Shock 24:66-73, 2005.
29. Betke K, Greinacher I, Tietze O: Oxidation of human and animal oxyhemoglobins by sodium nitrite. Naunyn Schmiedebergs Arch Exp Pathol Pharmakol 229:220-232, 1956.
30. Matthies H: Biochem Z 329:341-345, 1957. Article in German.
31. Cheung AT, Jahr JS, Driessen B, Duong PL, Chan MS, Lurie F, Golkaryeh MS, Kullar RK, Gunther RA: The effects of hemoglobin glutamer-200 (bovine) on the microcirculation in a canine hypovolemia model: a noninvasive computer-assisted intravital microscopy study. Anesth Analg 93:832-838, 2001.
32. Sakai H, Hara H, Tsai AG, Tsuchida E, Johnson PC, Intaglietta M: Changes in resistance vessels during hemorrhagic shock and resuscitation in conscious hamster model. Am J Physiol 276:H563-H571, 1999.
33. Cabrales P, Tsai AG: Plasma viscosity regulates systemic and microvascular perfusion during acute extreme anemic conditions. Am J Physiol Heart Circ Physiol. 291:H2445-H2452, 2006. Epub ahead of print.
34. Tsai AG, Acero C, Nance PR, Cabrales P, Frangos JA, Buerk DG, Intaglietta M: Elevated plasma viscosity in extreme hemodilution increases perivascular nitric oxide concentration and microvascular perfusion. Am J Physiol Heart Circ Physiol 288:H1730-H1739, 2005.
35. Cooke JP, Rossitch E Jr, Andon NA, Loscalzo J, Dzau VJ: Flow activates an endothelial potassium channel to release an endogenous nitrovasodilator. J Clin Invest 88:1663-1671, 1991.
36. Krieter H, Hagen G, Waschke KF, Kohler A, Wenneis B, Bruckner UB, van Ackern K: Isovolemic hemodilution with a bovine hemoglobin-based oxygen carrier: effects on hemodynamics and oxygen transport in comparison with a nonoxygen-carrying volume substitute. J Cardiothorac Vasc Anesth 11:3-9, 1997.
37. Lansman JB, Hallam TJ, Rink TJ: Single stretch-activated ion channels in vascular endothelial cells as mechanotransducers? Nature 325:811-813, 1987.
38. Cabrales P, Tsai AG, Intaglietta M, Frangos JA: Shear stress mechanotransduction and the flow properties of blood. In Winslow RM (ed.): Blood Substitutes. London, England: Elsevier Science Ltd, pp 93-101, 2005.
39. Cabrales P, Tsai AG, Winslow RM, Intaglietta M: Effects of extreme hemodilution with hemoglobin-based O2 carriers on microvascular pressure. Am J Physiol Heart Circ Physiol 288:H2146-H2153, 2005.
40. Tsai AG, Cabrales P, Intaglietta M: Microvascular perfusion upon exchange transfusion with stored RBCs in normovolemic anemic conditions. Transfusion 44:1626-1634, 2004.
41. Martini J, Carpentier B, Chavez Negrete A, Frangos JA, Intaglietta M: Paradoxical hypotension following increased hematocrit and blood viscosity. Am J Physiol Heart Circ Physiol 289:H2136-H2143, 2005.
42. Patel KP, Li YF, Hirooka Y: Role of nitric oxide in central sympathetic outflow. Exp Biol Med (Maywood) 226:814-824, 2001.

Microcirculation; extreme hemodilution; plasma expander; intravascular oxygen; methemoglobin; methylene blue; functional capillary density

©2007The Shock Society