Effect of Oncotic Pressure of Diaspirin Cross-Linked Hemoglobin (DCLHb Trademark) on Brain Injury After Temporary Focal Cerebral Ischemia in Rats

Abstract

In patients at risk for focal cerebral ischemia, hemodilution has been proposed as both a prophylactic and resuscitative therapy to ameliorate brain injury [1]. Although controversial, the rationale for hemodilution therapy is based on an observed correlation between reductions in blood viscosity and increased cerebral blood flow (CBF) [2-5]. When nonoxygenating fluids are used as therapeutic hemodiluents, the recommended hematocrit (in terms of limiting infarction volume) is approximately 30% [5]. Molecular hemoglobin, by decreasing viscosity while maintaining oxygen-carrying capacity, might in theory improve oxygen delivery to ischemic tissue after hemodilution [3]. Our laboratory has performed several investigations to evaluate the therapeutic effect of alpha-alpha diaspirin cross-linked hemoglobin (DCLHb Trademark; Baxter Healthcare Corporation, Round Lake, IL) for focal cerebral ischemia, and has observed that hypervolemic hemodilution with DCLHb Trademark augmented CBF and decreased infarction volume in a dose-dependent manner [3,6,7]. The maximum effect was observed at a hematocrit of 9% which required large volumes of DCLHb Trademark. Cardiovascular disease is often concomitantly present in patients with cerebrovascular disease, and hemodilution therapy has been associated with increased morbidity in clinical trials [8]. Thus, if attempting to maximize therapy with DCLHb Trademark for an ischemic brain, intentional hypervolemic hemodilution might result in meaningful cardiac morbidity and effectively lessen the overall therapeutic efficacy of molecular hemoglobin for stroke patients. Although a more concentrated preparation of DCLHb Trademark might lessen the required volume for a given dose, it would be expected to increase oncotic pressure, which has been shown to attenuate the development of ischemic cerebral edema [9]. In a rat model of temporary focal cerebral ischemia, we assessed the effect of equal doses of 20% DCLHb Trademark (20.2 g/dL, oncotic pressure 129 mm Hg) and 10% DCLHb Trademark (10.2 g/dL, oncotic pressure 43 mm Hg) on cerebral infarction volume after 90 min of middle cerebral artery occlusion (MCAo) and 72 h of reperfusion. Methods After approval by the institutional research committee of Loma Linda University, male, spontaneously hypertensive rats (350-400 g, 16-20 wk) were anesthetized with isoflurane, orotracheally intubated, and mechanically ventilated (Harvard Co., Boston, MA). The femoral vessels were cannulated for blood pressure monitoring (Full Scale Transducer/TA 2000 Recorder Registered Trademark; Gould Inc., Cerritos, CA), blood sampling, and fluid administration. During MCAo and for the initial 6 h of reperfusion, cranial temperature was servo-controlled at 37 degrees C (Mon-a-Therm Registered Trademark temperature sensor; Mallinckrodt Anesthesia Products, St. Louis, MO). Arterial blood (125 micro Liter) was analyzed (30-min intervals throughout MCAo) for pHa, PaCO2, PaO2, blood glucose, and hematocrit (IL-1306 pH Blood Gas Analyzer Registered Trademark; Instrumentation Laboratory, Lexington, MA; YSI Model 23-A Glucose Analyzer Registered Trademark; Yellow Springs Instruments, Yellow Springs, OH; IEC MB Centrifuge Microhematocrit Registered Trademark; DAMON/IEC Division, Needham Heights, MA, respectively). A midline neck incision was made and ischemia induced by advancing a 0.26-mm filament (Berkley Trimax Registered Trademark; Berkley Outdoor Technologies Group, Spirit Lake, IA) 17 mm, through the left external carotid artery, into the left middle cerebral artery. The incision was closed and the wound infiltrated with 0.25% bupivicaine. The isoflurane was discontinued, and the animals were extubated and allowed to recover. Immediately after MCAo, each rat was randomized to one of the following groups: Control (n = 9), 8.0 mL of fresh donor blood was given (hematocrit not manipulated); 10/Hb (n = 9), blood volume and hematocrit (30%) were manipulated by a 5.0-mL exchange transfusion with 10% DCLHb Trademark (10.2 g/dL, oncotic pressure 43 mm Hg), followed by an additional 8.0-mL bolus (0.5 mL/min); 7.5/Alb (n = 9), blood volume and hematocrit (30%) were manipulated by a 5.0-mL exchange transfusion with 7.5% human serum albumin (Armour Pharmaceutical Company, Kankakee, IL; oncotic pressure 43 mm Hg), followed by an additional 8.0-mL bolus (0.5 mL/min); 20/Hb (n = 9), blood volume and hematocrit were manipulated by a 2.5-mL exchange transfusion with 20% DCLHb Trademark (20.2 g/dL, oncotic pressure 129 mm Hg), followed by an additional 4.0-mL bolus (0.5 mL/min); and 15/Alb (n = 9), blood volume and hematocrit were manipulated by a 2.5-mL exchange transfusion with 15% human serum albumin (Armour Pharmaceutical Company, Kankakee, IL, oncotic pressure 130 mm Hg), followed by an additional 4.0-mL bolus (0.5 mL/min). The initial dose of DCLHb Trademark (see Appendix and Table 1) was given as an exchange transfusion to maintain a stable mean arterial blood pressure (MABP). If DCLHb Trademark is initially given as a hypervolemic bolus, MABP increases by 25-30 mm Hg for approximate equals 120-min [via nitric oxide binding [10]]. However, if given as an exchange transfusion normotension is maintained. Aside from the isovolemic exchange portion of the fluid regimens for the 10/Hb, 7.5/Alb, 20/Hb, and 15/Alb groups, all groups were administered an 8.0-mL hypervolemic fluid bolus.Table 1: Chemical Assay of 10% and 20% alpha-alpha Diaspirin Cross-Linked Hemoglobin (a)The animals recovered in an incubator with auto-temperature control and humidified oxygen. Cranial temperature was servo-controlled at 37 degrees C throughout MCAo and for the first 6 h of reperfusion. Although pilot studies with this model demonstrated that after 60 min of reperfusion the animals maintained normal cranial temperature, cranial temperature was servo-controlled for 6 h as a reasonable margin of insurance that cranial temperature was not different among groups. After 90 min of MCAo the suture was quickly removed (<5 min) during isoflurane anesthesia, and a 72 h period of reperfusion allowed. At the conclusion of the 72 h reperfusion period the animals were reanesthetized and perfused transcardially with 2% triphenyltetrazolium chloride (TTC) solution (37 degrees C) for 20 min. Fixation was then achieved with a 10% buffered formalin/10% gluteraldehyde solution. The brains were removed, embedded in an egg albumin:gelatin medium, and mounted on a vibratome. Ten standardized coronal brain sections (1.0-mm increments), spanning the area of middle cerebral artery distribution, were cut and photographed with color slide film (Ektachrome Registered Trademark, Tungsten 160 ASA; Eastman Kodak, Rochester, NY). The photographs were analyzed using a Drexel/DUMAS Image Analysis System and the volume of the entire hemisphere and of deficient TTC staining (infarct) defined and calculated. [See Cole et al. [7] for a photograph of a TTC stained brain in which normal and injured brain are defined.] The corpus callosum does not routinely stain with TTC in normal tissue, thus the rim of tissue representing the corpus callosum was excluded from analysis. All image analysis was performed by an independent observer who was blinded to study protocol. All between-groups data were evaluated by an analysis of variance, and, as appropriate, mean values compared by t-tests with Sheffe's test for multiple comparisons. P < 0.05 was considered significant. Results All data are presented as mean +/- SD. Except for hematocrit, there were no between-groups differences in the physiologic data for any of the 30-min sampling intervals during MCAo. Accordingly, the data are reported as an average of each group over the entire 90-min period of ischemia Table 2. There were no abnormalities in TTC staining for the hemisphere contralateral to MCAo. The calculated total volume (mm3) for the hemisphere ipsilateral to MCAo was 443 +/- 37 in the Control group, 448 +/- 34 in the 10/Hb group, 451 +/- 39 in the 7.5/Alb group, 457 +/- 30 in the 20/Hb group, and 441 +/- 35 in the 15/Alb group.Table 2: Physiologic Values During Middle Cerebral Artery OcclusionInfarction volume (mm3) was less in the DCLHb Trademark groups (10/Hb, 79 +/- 17; 20/Hb, 51 +/- 14) than the oncotically matched albumin groups (7.5/Alb, 124 +/- 21; 15/Alb, 85 +/- 18); was less in the 10/Hb, 20/Hb, and 15/Alb groups than the Control (135 +/- 17) and 7.5/Alb groups; and was less in the 20/Hb group than the other four groups (P < 0.05). There was no difference between the 7.5/Alb group and the Control group Table 3.Table 3: Infarction Volume (mm3) After 72 h of RecoveryDiscussion The results of this study support the hypothesis that hemodilution decreases ischemic brain injury. In addition, the data demonstrate that, in this model of cerebral ischemia, DCLHb Trademark decreases ischemic brain injury more proficiently than albumin, and that a hyperoncotic hemodiluent is preferable to a relatively normooncotic fluid. Previous studies in our laboratory, using a craniectomy model of MCAo (180 min MCAo and 120 min of reperfusion), demonstrated a dose-dependent favorable effect of DCLHb Trademark on CBF and infarction volume [3,6,7]. The maximum reduction in ischemia and infarction volume was achieved at a hematocrit of 9% which required a near total volume exchange. Volume manipulation of this magnitude may, in theory, entail a meaningful physiologic stress in stroke patients with preexisting cardiovascular disease [8]. Accordingly, the purpose of this study was to assess whether a dose of DCLHb Trademark as a more concentrated solution (20%) would maintain the efficacy of hemodilution therapy observed with 10% DCLHb Trademark. Hemodilution is postulated to favorably affect cerebral ischemia by augmenting CBF via one of two mechanisms. The first mechanism is by decreasing viscosity, of which hematocrit is the major determinant [1], while the second is a direct myogenic vasodilatory response to a reduction in oxygen content [11]. Although controversial, the majority of the data suggests that in ischemic brain a reduction in viscosity is the predominant mechanism for a hemodilution-induced increase in CBF [2-4,12]. Accordingly, we anticipated that an equal molecular dose of 20% DCLHb Trademark (greater viscosity) in one half the volume would not be as effective in reducing ischemic injury as 10% DCLHb Trademark. (It is likely that the direct hemodilution properties of 20% DCLHb Trademark were partially compensated for by Starling forces resulting in a net intravascular flux of fluid from the extravascular space. This would also act to reduce the potential volume advantages of 20% DCLHb Trademark.) Considering these differences, we were surprised that ischemic injury was less in the 20/Hb group than the 10/Hb group. Hemodilution has been associated with an increase in intracranial pressure, edema formation, and myocardial infarction [8,13]. It is plausible that there is a beneficial aspect of hemodilution therapy that increases flow to ischemic areas of the brain which is countered by inherent side effects in certain high risk patients susceptible to intravascular volume overload. If a similar beneficial effect of hemodilution therapy on CBF is attained by one half the volume of DCLHb Trademark (as a 20% preparation) without incurring as great a potential downside as a larger volume of 10% DCLHb Trademark, overall therapeutic efficacy might be improved. Several investigators have evaluated the effect of decreases in oncotic pressure on traumatic brain edema. The current understanding, in the setting of traumatic brain injury, is that oncotic pressure has a negligible effect on edema formation [14]. However, few studies have assessed the effect of oncotic pressure on focal cerebral ischemia. Korosue et al. [15] observed that in the setting of focal cerebral ischemia, brain injury was greater in animals hemodiluted with lactated Ringer's solution than in animals hemodiluted with hetastarch. In a model of forebrain ischemia, Hakamata et al. [9] observed less edema in animals given 25% albumin than in animals given normal saline. Thus, one mechanism that may account for a beneficial effect of hyperoncotic fluids on the ischemic brain is via a decrease of cerebral edema formation. Two types of brain edema are typically described: cytotoxic edema which is caused by an inability of cell membranes to regulate ionic fluxes and intracellular water balance, and vasogenic edema which is caused by increased permeability of the blood-brain barrier and a transvascular leak of macromolecules and ionic solutes into the interstitial space. Normally, the blood-brain barrier is relatively impermeable to macromolecules, and osmotic forces are the primary determinant of water movement. However, in the injured brain, the blood-brain barrier is disrupted to variable degrees, and moderate ischemic insults are postulated to result in a blood-brain barrier that is selectively permeable to ionic solutes with preservation of impermeability to macromolecules (e.g., hemoglobin) [15]. If this speculation were true, in the ischemic brain, an increase in intravascular oncotic pressure after 20% DCLHb Trademark therapy (oncotic pressure 129 mm Hg) may be a more meaningful factor in edema formation and secondary injury [16]. The possibility that portions of an infarction have an intact blood-brain barrier is supported by data obtained during the early period of reperfusion which demonstrated areas of impermeability to albumin [17]. A second explanation for the observed differences in infarction volume between the 10/Hb and 20/Hb groups might have been a difference in cardiac output. Although controversial, it appears that in normal brain CBF is not dependent on cardiac output, while in ischemic brain an increase in cardiac output effects an increase in CBF [18,19]. Cardiac output was not measured in the present study and intuitively one might have expected the group that received the largest fluid challenge (10/Hb group) to have had the greatest cardiac output and hence a greater increase in CBF. However, it is possible that spontaneously hypertensive rats may not tolerate the fluid load given to the 10/Hb group, and consequently suffered cardiac decompensation, a decrease in cardiac output, and decreased CBF relative to the 20/Hb group. The differences in infarction volume between oncotically matched DCLHb Trademark and albumin groups might in part be explained by mechanisms other than the obvious oxygen-carrying capacity of DCLHb Trademark. There are limited data to support the possibility that, in the ischemic brain with a damaged blood-brain barrier, albumin might be neurotoxic [20]. Moreover, molecular hemoglobin solutions are known to bind nitric oxide [10], which has been hypothesized to have a neurotoxic role during cerebral ischemia [21]. Accordingly, if this hypothesis were true, in the present study, the scavenging of nitric oxide by DCLHb Trademark might have favorably affected cerebral infarction volume. Criticisms of this study include that hemodilution was maintained only during MCAo. Although hemodilution therapy during reperfusion is a logical study question, it was our objective to evaluate the efficacy of hemodilution during MCAo only in a paradigm analogous to a monitored intraoperative ischemic event. However, based on an intravascular retention half-life as long as 24 h [22] it is likely that residual hemodilution was in effect during the initial period of reperfusion. A second critique is that intravascular oncotic pressure was not measured. It seems logical that fluids with unique oncotic pressures, administered as a regimen that consisted of a 25% exchange transfusion and a 40% hypervolemic bolus (of the intravascular volume) [23], would differentially affect intravascular oncotic pressure. It was not the purpose of this study to titrate a fluid regimen to a given intravascular oncotic pressure, but to test the effect of hemodiluents with different oncotic pressures on infarction volume. However, it is acknowledged that the exact quantitative effect is unknown. A final critique concerns the intraluminal filament which was not visually confirmed to totally occlude the middle cerebral artery. It is possible that residual blood flow in the surrounding area of the filament was present, and that decreasing intravascular viscosity would improve residual flow in the margins of the occlusive device in a manner analogous to a partial thromoembolic occlusion of a cerebral artery. Although the use of hemodilution has been successful in laboratory studies, clinical trials have been inconsistent [5-8,24,25]. Several explanations may account for this inconsistency. The first is the likelihood of a therapeutic window after the onset of ischemia, after which therapeutic maneuvers that augment CBF are not effective in limiting infarction [8]. If hemodilution is instituted after this window, ischemic injury may have matured to a point at which therapeutic efficacy is absent and only detrimental side effects are manifest [25]. In the present study, hemodilution was instituted immediately after the onset of ischemia. This protocol should maximize therapeutic efficacy but limit model relevance to circumstances in which hemodilution is used immediately after the detection of a monitored ischemic event (e.g., carotid endarterectomy, cerebral aneurysm surgery). The second is the observation that isovolemic hemodilution is often associated with a decrease in blood pressure [24]. In the present study, by utilizing a hypervolemic regimen and a hemodiluting fluid (DCLHb Trademark) that binds nitric oxide and maintains MABP [10], any hypotensive effects of hemodilution therapy were obviated. Finally, there is the issue of decreased oxygen content when hemodiluting with non-oxygen-binding fluids. Such fluids place inherent limits on oxygen transport, and therefore the effectiveness and magnitude of therapy. In theory, hemodilution with oxygen-binding fluids such as DCLHb Trademark may convey a unique advantage in the treatment of temporary focal cerebral ischemia. In summary, the effect of hypervolemic hemodilution on temporary focal cerebral ischemic injury was evaluated. Two preparations of DCLHb Trademark with different oncotic pressures (10% DCLHb Trademark, oncotic pressure 43 mm Hg; 20% DCLHb Trademark, oncotic pressure 129 mm Hg) were compared to each other and to oncotically matched preparations of albumin. The results indicate that DCLHb Trademark decreases ischemic brain injury more proficiently than albumin, and in terms of treating cerebral ischemia, a hyperoncotic preparation of DCLHb Trademark is preferable. The authors gratefully acknowledge the technical assistance of Suzzanne Marcantonio. Appendix The DCLHb Trademark solution was prepared by Baxter Healthcare Corporation (Round Lake, IL) as follows. Outdated red blood cells (human) were lysed by exposure to hypotonic buffer. Red cell stroma was removed by ultrafiltration, after which the diaspirin compound bis(3,5-dibromosalicyl) fumarate was used to cross-link molecular hemoglobin at the alpha chain. Viral contamination was eliminated and protein purification achieved by heat pasteurization. The DCLHb Trademark was formulated to a concentration of 10.2 or 20.2 g/dL by diafiltration against electrolyte and buffer solution (see Table 1 for lot release analysis). The solution was kept at -70 degrees C until needed for the current study, when it was thawed to 5 degrees C, and on the day of the study passively warmed to room temperature. Oxygen transport of DCLHb Trademark is similar to that of whole blood. The alpha-alpha cross-linking with bis(3,5-dibromosalicyl) fumarate stabilizes the hemoglobin and prolongs intravascular retention. The viscosity of 10% DCLHb Trademark (1.1 centipoiseuilles) is comparable to that of serum albumin and considerably less than that of whole blood (>3.5 centipoiseuilles).