Emergency preservation and resuscitation with profound hypothermia, oxygen, and glucose allows reliable neurological recovery after 3 h of cardiac arrest from rapid exsanguination in dogs

/Home/EPR and Neurological Recovery

Emergency preservation and resuscitation with profound hypothermia, oxygen, and glucose allows reliable neurological recovery after 3 h of cardiac arrest from rapid exsanguination in dogs

This study was supported by United States Army Medical Research and Materiel Command DAMD 17-01-2-0038.

Journal of Cerebral Blood Flow & Metabolism (2008) 28, 302–311; doi:10.1038/sj.jcbfm.9600524; published online 11 July 2007

Xianren Wu1,2, Tomas Drabek1,2, Samuel A Tisherman1,3,4, Jeremy Henchir1, S William Stezoski1, Sherman Culver1, Jason Stezoski1, Edwin K Jackson5, Robert Garman6 and Patrick M Kochanek1,3

1Safar Center for Resuscitation Research, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
2Department of Anesthesiology, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
3Department of Critical Care Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
4Department of Surgery, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
5Department of Pharmacology, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
6Consultants in Veterinary Pathology Inc., Murrysville, Pennsylvania, USA
Correspondence: Dr P Kochanek, Safar Center for Resuscitation Research, Department of Critical Care Medicine, University of Pittsburgh, 3434 Fifth Avenue, Pittsburgh, PA 15260, USA. E-mail: kochanekpm@ccm.upmc.edu

Received 29 March 2007; Revised 11 May 2007; Accepted 19 May 2007; Published online 11 July 2007.

Top of page
Abstract
We have used a rapid induction of profound hypothermia (<10°C) with delayed resuscitation using cardiopulmonary bypass (CPB) as a novel approach for resuscitation from exsanguination cardiac arrest (ExCA). We have defined this approach as emergency preservation and resuscitation (EPR). We observed that 2 h but not 3 h of preservation could be achieved with favorable outcome using ice-cold normal saline flush to induce profound hypothermia. We tested the hypothesis that adding energy substrates to saline during induction of EPR would allow intact recovery after 3 h CA. Dogs underwent rapid ExCA. Two minutes after CA, EPR was induced with arterial ice-cold flush. Four treatments (n=6/group) were defined by a flush solution with or without 2.5% glucose (G+ or G−) and with either oxygen or nitrogen (O+ or O−) rapidly targeting tympanic temperature of 8°C. At 3 h after CA onset, delayed resuscitation was initiated with CPB, followed by intensive care to 72 h. At 72 h, all dogs in the O+G+ group regained consciousness, and the group had better neurological deficit scores and overall performance categories than the O−groups (both P<0.05). In the O+G− group, four of the six dogs regained consciousness. All but one dog in the O−groups remained comatose. Brain histopathology in the O−G+ was worse than the other three groups (P<0.05). We conclude that EPR induced with a flush solution containing oxygen and glucose allowed satisfactory recovery of neurological function after a 3 h of CA, suggesting benefit from substrate delivery during induction or maintenance of a profound hypothermic CA.

Keywords: cardiac arrest, energy metabolism, neuropreservation
Top of page
Introduction
Traumatic exsanguination cardiac arrest (ExCA) remains a condition with a nearly 100% mortality (Rhee et al, 2000). Conventional resuscitation using basic and advanced trauma life support as recommended by the American Heart Association as well as the American College of Surgeons is futile due to profoundly reduced blood volume and ongoing bleeding. In the 1980s, a novel approach to ExCA was proposed by Safar and Bellamy (Bellamy et al, 1996), who conceived the idea that emergency preservation with rapid induction of profound hypothermia and/or administration of pharmacological treatments, allowing time for transport, damage control surgery, and delayed resuscitation using cardiopulmonary bypass (CPB) would eventually allow intact survival. This concept of emergency preservation and resuscitation has been given the acronym EPR (Wu et al, 2006) and its efficacy was demonstrated in dogs (Behringer et al, 2001d, 2003; Nozari et al, 2004b), pigs (Alam et al, 2006), and rats (Drabek et al, 2007). Recently, novel agents have been found to induce a hibernation-like state with carbon monoxide in worms under hypoxia (Nystul and Roth, 2004) or with hydrogen sulfide in mice (Blackstone et al, 2005). However, rapid induction of profound hypothermia remains the only approach effective in large animal models and thus the only currently feasible clinical approach.

For induction of EPR in a series of prior reports, we have used ice-cold normal saline. Although pharmacological adjuncts added to the flush solution could have theoretical advantages in preserving vital tissues, long-term neurological outcome in our models was not improved versus normal saline with a number of mechanism-based pharmacological approaches (Behringer et al, 2001a, 2001b) except for the antioxidant tempol, which had a modest effect (Behringer et al, 2002). Similarly, the use of conventional or novel alternative flush solutions such as albumin, Unisol (Behringer et al, 2001c), or the University of Wisconsin solution (Wu et al, 2005) did not augment the protection afforded by profound hypothermia. Taking a different approach, Taylor et al (1994) demonstrated that continuous perfusion with an asanguineous preservation solution allowed satisfactory recovery over 3 h of ultraprofound hypothermia (<5°C). Alam et al (2006) have had success with a similar approach in pigs with traumatic hemorrhage. However, it is unlikely that in a clinical trauma scenarios continuous perfusion will be an option (Bellamy et al, 1996; Wu et al, 2005). Using an ice-cold saline flush for induction of EPR, preservation efficacy was improved with the use of either lower temperature or with faster cooling rates; however, efficacy reached its plateau at a core temperature between 7 and 10°C (Behringer et al, 2003; Alam et al, 2006) or when cooling rate was maximized (Alam et al, 2004). However, at the maximal cooling rate that could be achieved, it still took between 12 and 15 mins to achieve a brain temperature of ~10°C in dogs (Behringer et al, 2003). This 12–15-min period required to reach target temperatures might represent a key limiting factor in the ultimate success of EPR based on the report that the brain oxygen demand in pigs remains ~50% of baseline at 28°C, ~19% at 18°C, and ~10% at 8°C, respectively (Ehrlich et al, 2002).

Logically, we speculated that providing energy substrates during induction of EPR might either avoid further energy depletion, or even restore energy reserves that would probably be reduced during the 5 mins period of shock, 2 mins normothermic CA, and cooling duration—before achieving profound hypothermia. Perfusion of dissolved oxygen at deep hypothermia (without circulatory arrest) can deliver considerable substrate, particularly in a setting of markedly reduced metabolic demands (Dexter et al, 1997). Robbins et al (1990) reported that intermittent flush of energy substrates into the brain during profound hypothermic CA delayed ATP and creatine phosphate depletion in brain. However, there is no solution that has been convincingly shown to improve neurological outcome in a prolonged CA model, without intermittent perfusion during the arrest.

Using our modified EPR model (Nozari et al, 2004a), the current study was designed to test if profound hypothermia induced by aortic flush with a solution that was enriched with energy substrates, that is oxygen and glucose, could successfully produce intact long-term neurological outcome despite a prolonged (3 h) CA.

Top of page
Materials and methods
The experimental protocol was approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh and the Department of Defense and followed the National Guidelines for Treatment of Animals.

Experimental Design
The model included three phases: (1) exsanguination (5 min) and CA (2 min); (2) EPR (3 h); (3) delayed resuscitation, including CPB (2 h) and intensive care (72 h). At the end of the exsanguination and CA phase, dogs were randomized into four groups based on the specific additives in the ice-cold normal saline flush solution, namely (1) oxygen+glucose (O+G+), (2) oxygen without glucose (O+G−), (3) glucose alone (O−G+), and (4) neither oxygen nor glucose (O−G−).
Anesthesia and Preparation
Custom-bred, male hunting dogs, weighing 19.5 to 24.0 kg, were housed for at least 3 days before the experiment. A total of 24 dogs were used, and no dog was excluded from the protocol once entered. Dogs were fasted with free access to water for 12 h. Ketamine 10 mg/kg and atropine 0.4 mg were administered intramuscularly. A cannula (18 G) was inserted into a peripheral vein and fluid infusion (D5W/0.45 NaCl at 4 mL/kg per h) was started. After anesthesia induction with 4% halothane via face mask, endotracheal intubation (ID 8 to 9 mm) was performed. Continuous anesthesia was provided with halothane titrated during preparation (O2/N2: 50%:50%). Controlled ventilation was initiated with tidal volume of 12 to 15 mL/kg, PEEP 2 cm H2O, and frequency of 20 to 25 per min, titrated to maintain PaCO2 of 35 to 45 mm Hg. Electrocardiogram lead II was continuously monitored. A Foley catheter was placed into the urinary bladder. Temperature probes were inserted for rectal, esophageal, and both tympanic membrane temperatures (Tty). Sterile cut downs were made in both groins and the right side of the neck. A PE 90 catheter was inserted into the left femoral artery for blood pressure monitoring and blood sampling. A pulmonary artery catheter (7.5 F) was inserted via the left femoral vein into the pulmonary artery to monitor pressure, cardiac output, and core temperature (Tpa). A CPB arterial cannula (7 or 9 G) was inserted into the right femoral artery. A multiple hole, CPB venous cannula (19 F) was inserted 25 cm into the right atrium via the right external jugular vein. The CPB system consisted of a hollow-fiber membrane oxygenator (Medtronic, Grand Rapids, MI, USA) and centrifugal pump (Biomedicus, Eden Prairie, MN, USA). For induction of EPR hypothermia, the CPB system was primed with normal saline at 2°C; for delayed resuscitation after EPR, the system was primed with shed blood (30 mL/kg) and Plasma-Lyte A (Baxter, Deerfield, IL, USA).
Baseline measurements (hemodynamics, arterial and venous blood gases, and body temperatures) were determined when hemodynamics was stable for 15 to 30 mins after surgical preparation and Tpa was controlled at 37.5 to 38.5°C.

Exsanguination and Cardiac Arrest Phase
After two baseline measurements, heating, intravenous fluids, and halothane were discontinued, and the dogs were weaned to spontaneous breathing of air via a T-tube. When the canthal reflex returned, rapid exsanguination was initiated via the right external jugular cannula and the blood was collected in bags with sodium citrate anticoagulant for later reinfusion. Exsanguination was conducted stepwise to a mean arterial pressure of 20 mm Hg at 4 mins. At 5 mins, ventricular fibrillation was induced with transthoracic AC at 95 V to ensure zero blood flow. Ventricular fibrillation was confirmed with electrocardiogram and arterial blood pressure.

Emergency Preservation and Resuscitation Phase
Two minutes after the onset of CA, flush solution (80 mL/kg) at 2°C was infused into the aorta at a rate of 80 mL/kg per min using the CPB pump. Close-chest CPB from the right external jugular vein to the right femoral artery was then initiated for induction of hypothermia until Tty on the right side, which was arbitrarily chosen to represent brain temperatures, reached 8°C. Either 100% oxygen or nitrogen was supplied to the oxygenator throughout the flush interval and induction of hypothermia. The gas flow to the CPB oxygenator was adjusted to maintain PaCO2 35 to 45 mm Hg. Once Tty of 8°C was reached, the CPB was stopped. The entire body was covered with ice from the onset of flush to the end of 3 h of CA.

Delayed Resuscitation Phase
Cardiopulmonary Bypass After 3 h of CA, reperfusion was started with CPB that was primed with shed blood plus heparin 1,000 U. Before the start of CPB, sodium bicarbonate (1 mEq/kg) and epinephrine 0.01 mg/kg were injected into the circuit. The temperature of the water bath of the CPB heat exchanger was set to 5°C above Tpa until Tpa reached 34°C. Cardiopulmonary bypass was started with a flow of 50 mL/kg per min when Tpa was <20°C, increased to 75 mL/kg per min when Tpa was 21 to 30°C, and to 100 mL/kg per min when Tpa was >30°C. Repetitive doses of epinephrine (0.01 mg/kg) were given to increase mean arterial pressure to 60 mm Hg at Tpa <20°C, to 80 mm Hg at Tpa 21 to 30°C, and to 100 mm Hg at Tty >30°C. When Tpa reached 32°C, defibrillation was attempted with external DC countershocks of 150 J, increased by 50 J for repeated shocks. Oxygen flow through the oxygenator was adjusted to keep PaCO2 between 35 and 40 mm Hg and PaO2 ≥100 mm Hg. During CPB, controlled ventilation was given with 100% oxygen at a rate of between 8 and 10 breaths/min. Intravenous fluids were restarted at 4 mL/kg per h. A base deficit of >6.0 mEq/L was corrected with sodium bicarbonate. Mean arterial pressure was maintained at 90 to 150 mm Hg. The CPB flow rate for assisted circulation was sequentially reduced to 75, 50 mL/kg per min, and stopped at 120 mins. During CPB, activated clotting times were maintained at >300 secs with additional heparin.

ICU The details of life support, including mechanical ventilation, hemodynamic monitoring and support, and correction of acid–base or electrolyte abnormalities, were published previously (Behringer et al, 2003). Body temperature was kept at 34°C until 36 h of resuscitation, followed by slow controlled rewarming (0.3°C/h) to 36.5°C, as per our previous study that found the benefit of more prolonged postresuscitation mild hypothermia (Wu et al, 2006). Consequently, mechanical ventilation was continued to 48 h with morphine sulfate analgesia, diazepam sedation, and pancuronium. Hemodynamic support included additional fluids and pressors as needed. Acid–base and electrolyte abnormalities were corrected as per typical clinical patient management. At 48 h, neuromuscular blockade was reversed, and sedation and analgesia were discontinued. Dogs were then weaned from mechanical ventilation. After extubation, they were transferred to the step down unit where continuous intravenous fluids and vital sign monitoring were provided until 72 h.

Outcome Evaluation
Functional outcomes were evaluated every 6 h in the step down unit according to the overall performance categories (OPC: 1=normal or slight disability; 2=moderate disability; 3=severe disability; 4=coma; 5=death) and the neurological deficit scores (NDS: 0 to 10%=normal; 100%=brain death), which include level of consciousness, breathing pattern, cranial nerve function, sensory and motor function, and behavior. At 72 h, a final functional assessment was performed and animals were then re-anesthetized with ketamine and halothane. Perfusion fixation was performed with cephalad infusion of 10% neutral-buffered formalin via the thoracic aorta. The entire brain was removed ~2 h after perfusion fixation and retained in 10% neutral-buffered formalin until dissection.
Neuropathology
Whole perfusion-fixed brains were divided into multiple coronal slices. Six coronal brain slices plus three transverse sections of the medulla oblongata and upper cervical cord were selected for microscopic evaluation. These represented entire brain slices taken at the following levels: (1) the optic chiasm; (2) the anterior thalamus; (3) the posterior thalamus; (4) the midbrain; (5) posterior portions of the occipital lobes; (6) middle of the cerebellum and underlying brainstem; (7) medulla oblongata and upper cervical cord. These slices were processed for paraffin embedding, resulting in 20 tissue blocks from each brain. The paraffin blocks were sectioned at 5 μm, and the resulting sections stained with hematoxylin and eosin (H&E) and with Fluoro-Jade B. The examining neuropathologist (RG) was masked as to the treatment groups. Each neuroanatomic region with evidence of damage on microscopic examination received a subjective pathological grade ranging from 1+ (minimal) to 5+ (severe). Each affected region on each side of the brain (right and left) received separate scores for the degrees of neuropathological damage detected in H&E- and Fluoro-Jade B-stained sections. The histological damage scores (HDS) in each region were compared with other groups.
Statistical Analysis
Data are presented as mean±s.d. unless otherwise stated. Repeated measures analysis of variance followed by Bonferroni post hoc tests was performed to identify differences in hemodynamic parameters and temperatures between groups. Overall performance categories, NDS, and HDS scores were analyzed using Kruskal–Wallis and then Mann–Whitney U-test with correction of multiple comparisons. The Spearman rank analysis was used to examine the correlation of histological findings from different histological techniques. P<0.05 was considered statistically significant.
Top of page
Results
Induction of Emergency Preservation and Resuscitation
The total EPR induction time to reach Tty 8°C was similar between groups. During the induction of EPR, glucose levels in the G+ groups were approximately five times higher than in the G− groups (P<0.01), and the PaO2 values in the O+ groups were ~20 times higher than in the O− groups (P<0.01) (Table 1). PaCO2 did not differ among four groups. Compared with G− groups, the G+ groups had lower levels of plasma sodium (P<0.05) and hematocrit (NS) at the end of flush (Table 1). The O+ groups had significantly lower plasma potassium than the other two groups (P<0.01) (Table 1). The brain temperatures did not differ among four groups over 3 h of CA (Figure 1).
Figure 1.
Right side tympanic temperatures during induction of EPR and no-flow cardiac arrest. O+G+: 100% O2 with 2.5% glucose in normal saline; O+G−:100% O2 with normal saline; O−G+: 100% N2 with 2.5% glucose in normal saline; O−G−:100% N2 with normal saline.

Full figure and legend (43K)

Table 1 - Physiological parameters at the end of EPR induction.
Full table

Resuscitation
After 3 h of CA, rewarming and stable mean arterial pressure (>60 mm Hg without need for vasopressors) were achieved in all dogs with CPB. When Tpa reached 32°C (40 to 50 mins after delayed resuscitation with CPB), restoration of spontaneous circulation was achieved in all dogs with 1 to 2 defibrillation attempts at 150 J. Cardiopulmonary bypass was then weaned off at 2 h in all dogs without any need for pharmacological support. During the delayed resuscitation phase, the lactate levels were significantly higher in the two O− groups (P<0.05) (Figure 2). All dogs were maintained with stable vital signs to 72 h.
Figure 2.
Arterial lactate levels over 3 h of ExCA and 48 h delayed resuscitation. O+G+: 100% O2 with 2.5% glucose in normal saline; O+G−:100% O2 with normal saline; O−G+: 100% N2 with 2.5% glucose in normal saline; O−G−:100% N2 with normal saline.

Full figure and legend (42K)

Functional Outcome
At 72 h, all dogs in the O+G+ group regained consciousness with a significantly better OPC (Figure 3) and better NDS (Figure 4) (both P<0.05), compared with the O− groups. In the O+G− group, four of the six dogs regained consciousness (NS versus other groups). In contrast, 11 of the 12 dogs in the O− groups remained comatose.
Figure 3.
Final overall performance category (OPC) at 72 h after 3 h of cardiac arrest. O+G+: 100% O2 with 2.5% glucose in normal saline; O+G−:100% O2 with normal saline; O−G+: 100% N2 with 2.5% glucose in normal saline; O−G−:100% N2 with normal saline.

Full figure and legend (17K)

Figure 4.
Final neurological deficit scores (NDS) at 72 h after 3 h of cardiac arrest. O+G+: 100% O2 with 2.5% glucose in normal saline; O+G−:100% O2 with normal saline; O−G+: 100% N2 with 2.5% glucose in normal saline; O−G−:100% N2 with normal saline.

Full figure and legend (37K)

Neuropathology
Total HDS combining scores in all regions on both sides was significantly higher (worse) in the O−G+ group compared with the other groups (P<0.05) (H&E staining (median and range): O−G+: 145 (120 to 150), O+G+: 51 (10 to 76), O+G−: 57 (27 to 70), and O−G−: 48 (10 to 112); Fluoro-Jade B staining: O−G+: 132 (106 to 174), O+G+: 56 (38 to 74), O+G−: 36 (0 to 86), and O−G−: 57 (2 to 96)). Of 25 brain regions, there were 11 regions that had substantial brain injury in at least one group (Figures 5A and 5B). Histological damage scores of the O−G+ group was consistently worse than all other groups in all involved brain regions (P<0.05) (Figures 5A and 5B). The results with H&E and Fluoro-Jade B staining were significantly correlated in all 25 brain regions (r=0.52 to 1.0, all P<0.01, Spearman rank test).
Figure 5.
(A) Brain histological damage score (HDS) (median, 25–75% range): H&E staining. O+G+: 100% O2 with 2.5% glucose in normal saline; O+G−:100% O2 with normal saline; O−G+: 100% N2 with 2.5% glucose in normal saline; O−G−:100% N2 with normal saline; N: neurons; Cx: cortex; *P<0.05 compared with other three groups. (B) Brain histological deficit score (HDS): Fluoro-Jade B staining; O+G+: 100% O2 with 2.5% glucose in normal saline; O+G−:100% O2 with normal saline; O−G+: 100% N2 with 2.5% glucose in normal saline; O−G−:100% N2 with normal saline; N: neurons; Cx: cortex *P<0.05 compared with other three groups.

Full figure and legend (152K)

Top of page
Discussion
With the induction of profound hypothermia for EPR with ice-cold saline following rapid exsanguination to CA, we have previously been able to achieve good outcomes in dogs after up to 2 h CA. In the clinical situation of the exsanguinated trauma victim, more time may be needed for transport, resuscitative surgery, and initiation of delayed resuscitation. Thus, there is a clinical need to extend the duration of EPR. In the current study, an energy preservation strategy with oxygen and glucose allowed good recovery after 3 h of CA.

The significance of this finding could be demonstrated when it is placed along our many trials to improve EPR efficacy. In our pursuit of pharmacological preservation (Wu et al, 2005) (Behringer et al, 2001a, 2001b), only tempol improved outcome in a 20-min CA EPR model (Behringer et al, 2002); in the pursuit of hypothermic preservation, which was reliable, we came to realize that the maximal effect of profound hypothermia alone could allow consistent intact neurological outcome, but only to 2 h of CA (Behringer et al, 2001d; Nozari et al, 2003). Given this background, our current study, showing consistently good functional outcome after a 3-h ExCA (2.5 h of no flow), represents a significant step of advance in neuropreservation in EPR studies.

Although we have benefited from the earlier explorations in cryobiology and deep hypothermic CA (DHCA) for neuropreservation, EPR studies have distinct challenges and goals. The normothermic hypotension and subsequent normothermic CA before induction of hypothermia, modeling ExCA, probably increased the difficulties in achieving successful preservation. In contrast, as early as 1986, successful recovery of neurological functions was achieved after 3 h profound hypothermic circulatory arrest in some healthy dogs (Haneda et al, 1986). For DHCA, 3 h circulatory arrest is rarely indicated clinically, and thus seldom addressed in labs. While EPR studies target survival with satisfactory neurological functions (OPC: 1 to 2) as an acceptable goal in trauma victims who would otherwise have near 100% mortality after ExCA (Rhee et al, 2000), DHCA studies, however, target the reduction of neurological morbidity after ~60 mins of bloodless surgery (Amir et al, 2005). In fact, the ‘safe’ duration of DHCA appears to be as short as 20 to 30 mins in patients who underwent thoracic cardiac surgery (Immer et al, 2004). Neurological recovery after up to 3 to 3.5 h of profound hypothermic CA has been reported using continuous perfusion with an experimental tissue preservation solution (Taylor et al, 1994). However, using the same solution significantly worsened NDS and OPC were noticed when intermittent no flow (circulatory arrest) was allowed in a 100-min DHCA model (Miura et al, 1996).

The decision to add oxygen and glucose to the perfusate was a logical step in our pursuit of more effective neuropreservation. First, we reported previously (Behringer et al, 2003) that despite using an arterial flush catheter of maximal diameter, maximal flush rate, and concurrent surface cooling, target temperature of 7 to 10°C could not be achieved for 12 to 15 mins in dogs. Second, metabolic demands are much greater during cooling than at profound hypothermia (Ehrlich et al, 2002). Third, in our EPR model, we previously screened 14 pharmacological strategies covering a broad spectrum of mechanism-based and empiric strategies—including agents targeting apoptosis, mitochondrial failure, anticonvulsants, barbiturates, MK-801 and calcium antagonists, among others (Wu et al, 2005). Unfortunately, these drugs targeting secondary injury cascades of neuronal death were disappointing, suggesting the need to consider alternative approaches. Similar disappointing results were observed by Aoki et al (1994) with addition of MK-801 to profound hypothermia in DHCA in piglets. Fourth, 2 h of DHCA (12 to 15°C) in sheep could be achieved with preservation of high-energy phosphate levels in brain (to ~60% of baseline) via intermittent infusion of a crystalloid solution containing dissolved oxygen and 2.5% dextrose (Robbins et al, 1990). That solution was called ‘cerebroplegia’ and also contained lidocaine, sodium bicarbonate, nitroglycerine, and mannitol. Fifth, as temperature decreases below 37°C, the affinity of hemoglobin for oxygen is greatly enhanced, restricting delivery, and increasing the importance of the dissolved oxygen component (Dexter et al, 1997). Sixth, the solubility of oxygen in saline nearly doubles between 37 and 18°C (Pearl et al, 2000). Grist (1996) suggested that the use of hyperoxia before DHCA can take advantage of enhanced oxygen solubility and reduced metabolic demands of hypothermia to prevent tissue injury. Hyperoxic perfusion during induction of hypothermia has been suggested to attenuate tissue acidosis in the clinical use of DHCA (Pearl et al, 2000). Finally, our flush rates of 20 L delivered over ~20 mins suggest that substrate delivery during the flush could be substantial, particularly in the setting of reduced metabolic demands. Thus, it was logical to propose that we could meet better metabolic demands with oxygen and glucose added to the flush solution during the induction of hypothermia. Different from DHCA, clinical application of EPR in management of ExCA would only be feasible after a normothermic ExCA has occurred. Brain energy reserve is depleted ~5 mins after normothermic CA (Shaffner et al, 1999; Eleff et al, 1991). Thus, to postpone energy failure, it may be important for preservation strategies to prevent energy depletion in brain and restore energy levels during induction of hypothermia. Based on our favorable outcomes in the dogs flushed with oxygen (with or without glucose) and on the aforementioned study by Robbins et al (1990) in which the cerebral ATP depletion was attenuated with the addition of dissolved oxygen and glucose in the perfusate, flush with oxygen and glucose solution after ExCA may have prevented the development of critical energy depletion. Additional studies of ATP or energy charge would be needed to prove that hypothesis.

Recently, important work by Vereczki et al (2006) demonstrated deleterious effects of hyperoxic reperfusion after a 10-min ventricular fibrillation cardiac arrest in dogs (Vereczki et al, 2006). In contrast, the powerful favorable effect of oxygen in our study probably relates to the fact that it is used to mitigate energy failure during cooling—before it results in cellular disturbances that set the stage for oxidative reperfusion injury. This intriguing hypothesis also needs to be further evaluated.

The best functional outcome was achieved only with the combination of oxygen and glucose in our model. It is possible that added glucose is important in delaying energy depletion during the prolonged hypothermic CA. However, the effects of glucose in cerebral ischemia are complex. On one the hand, high glucose may enhance energy production via glycolysis during anoxia/hypoxia (Tian and Baker, 2002), and/or provide beneficial osmolar effects as did mannitol in a cat middle cerebral artery occlusion model (Little, 1978). However, it is not clear how much glucose was transferred across the blood–brain barrier during hypothermia induction in our model. The G+ groups did not have increased arterial lactate levels either at the end of flush or during early reperfusion. Instead, lactate levels were consistently higher only in both O− groups during early reperfusion, and there was no difference in lactate levels between two O− groups. Cerebral lactate production was not examined in this study. It is certainly possible that cerebral effects were masked by systemic effects. We recognize that it is possible that the choice of 2.5% dextrose does not represent the optimal concentration and could be excessive.

Alternatively, glucose could be detrimental during cerebral ischemia (Vannucci et al, 1996), with higher tissue lactate levels, acidosis, oxidative stress, glutamate, DNA fragmentation, and other deleterious effects (Li et al, 2001). It is possible that a net benefit of glucose is seen if energy failure is prevented, while injury exacerbation dominates if frank ischemia occurs. Consistent with a potential dichotomous effect of glucose depending on whether energy failure is prevented, in our EPR model, the combination of oxygen and glucose produced the best functional outcomes, whereas the O−G+ group exhibited the worst histological injury. Additional studies of brain glucose utilization and energy charge would be helpful. Further studies would also be important to define the possible benefits of the osmotic effects of glucose in this model (Shin'oka et al, 1998). The findings of hyponatremia and decreased hemoglobin levels in the glucose groups suggest that the systemic osmotic effects were substantial. The effects on the brain, however, are unclear.

Given the dichotomous effects of additional glucose, the ability to successfully achieve 3 h of preservation might be attributed mostly to the addition of oxygen to profound hypothermia. Postresuscitation lactate was lower in both groups with oxygen added to the flush versus those without oxygen, and there was a trend toward improved functional outcome in the O+G− group. Also, the O+G− and O+G+ groups had similar HDS. Based on the observed OPC and HDS, our study was insufficiently powered to test for differences between groups with and without glucose in the presence of oxygen in the flush. The variable effects of glucose on histology in our model also suggest the possibility that alternative fuels such as β-hydroxybutyric acetate may be worthy of investigation as an adjunct to dissolved oxygen in EPR (Suzuki et al, 2001).

Unlike our previous EPR studies where we used one-way flush for induction, we used CPB to induce profound hypothermia in the current study. The hematocrit during induction of EPR was around 5 to 10% in all groups. This deviation from one-way flush with saline could be important. On one hand, higher hematocrit (>10%) during induction of hypothermia before CA was associated with better energy reserve and neurological outcome (Shin'oka et al, 1996). If so in the current study, it again suggests the importance of supporting oxygen delivery during induction of hypothermia. However, after cooling to 10 to 12°C, hemodilution to <5% improved neurological function (Sekaran et al, 2001). It is therefore difficult for us to predict the impact of the residual hemoglobin without further studies.

Among limitations, the lack of biochemical data (i.e., brain ATP, energy charge, glucose, and lactate levels) suggests the need for caution in making conclusions about a clear relationship between energy metabolism and improved outcome. In addition, it is important to recognize that it would be technically difficult to induce profound hypothermia in the field with CPB—as used in this proof of concept investigation. We also recognize that it is impossible to determine in this work whether beneficial effects of oxygen and glucose added to the flush are being produced during the induction of hypothermia, during the period of no flow, during reperfusion, and/or any combination of these time intervals. Biochemical studies of both ATP preservation and osmolar effects during induction and maintenance of hypothermia, and of secondary injury mechanisms during reperfusion are necessary to understand further this intervention and optimize its application. Finally, we recognize that hypothermia can delay the appearance of damage after cerebral ischemia and that assessment of brain histopathology after longer outcome intervals would be valuable.

In summary, EPR using a combination of oxygen and 2.5% glucose plus profound hypothermia allowed satisfactory recovery of neurological function after 3 h of CA. Prevention or reversal of energy failure and other mechanisms may be responsible for the benefit. Adding oxygen and possibly glucose in the cooling solution might augment the efficacy of either resuscitative or elective deep hypothermia.

© 2010 International Society for Cerebral Blood Flow & Metabolism