Histopathologic Consequences of Hyperglycemic Cerebral Ischemia During Hypothermic Cardiopulmonary Bypass in Pigs

Transcription

1 Histopathologic Consequences of Hyperglycemic Cerebral Ischemia During Hypothermic Cardiopulmonary Bypass in Pigs Brendan P. Conroy, MD, FFARCSI, Marjorie R. Grafe, MD, PhD, Larry W. Jenkins, PhD, Alejandro H. Vela, MD, Cheng Y. Lin, PhD, Douglas S. DeWitt, PhD, and William E. Johnston, MD Departments of Anesthesiology (Division of Cardiothoracic Anesthesiology), and Pathology, The University of Texas Medical Branch, Galveston, Texas, and Department of Neurosurgery, University of Pittsburgh, Pittsburgh, Pennsylvania Background. This study examined whether 34 C or 31 C hypothermia during global cerebral ischemia with hyperglycemic cardiopulmonary bypass (CPB) in surviving pigs improves electroencephalographic (EEG) recovery and histopathologic scores when compared with normothermic animals. Methods. Anesthetized pigs were placed on CPB and randomly assigned to 37 C (n 9), 34 C (n 10), or 31 C (n 8) management. After increasing serum glucose to 300 mg/dl, animals underwent 15 minutes of global cerebral ischemia by temporarily occluding the innominate and left subclavian arteries. Following reperfusion, rewarming, and termination of CPB, animals were recovered for 24 (37 C animals) or 72 hours (34 C and 31 C animals). Daily EEG signals were recorded, and brain histopathology from cortical, hippocampal, and cerebellar regions was graded by an independent observer. Results. Before ischemia, serum glucose concentrations were similar in the 37 C (307 9 mg/dl), 34 C ( mg/dl), and 31 C (310 15) groups. By the first postoperative day, EEG scores in 31 C animals ( ) had returned to baseline and were greater than those in the 34 C ( ) and 37 C ( ) groups (p < 0.05, respectively, between groups). Cooling to 34 C showed selective improvement over 37 C in hippocampal, temporal cortical, and cerebellar regions, but the greatest improvement in all regions occurred with 31 C. Cumulative neuropathology scores in 31 C animals ( ) exceeded 34 C ( ) and 37 C ( ) animals (p < 0.05, respectively, between groups). Conclusions. Hypothermia during CPB significantly reduced the morphologic consequences of severe, temporary cerebral ischemia under hyperglycemic conditions, with the greatest protection at 31 C. (Ann Thorac Surg 2001;71: ) 2001 by The Society of Thoracic Surgeons Controversy continues over the importance of temperature management during cardiopulmonary bypass (CPB) in patients undergoing cardiac operations. Particularly in patients with intraoperative ischemic strokes, whether cooling during CPB can alter neurologic outcome is not known. In one recent clinical trial of temperature management ranging between 20 C and 37 C, no relationship was found between the size of a cerebral ischemic infarct and the perfusate temperature [1]. In contrast, Gaudino and colleagues [2] found that hypothermic perfusion (nasopharyngeal temperature 27.4 C 0.9 C) compared with normothermic perfusion (nasopharyngeal temperature 36.3 C 0.5 C) reduced perioperative ischemic brain injury as assessed by computed tomographic scans and neurologic examination. Because cooling during CPB requires longer perfusion duration, produces detrimental effects on enzyme kinetics and coagulation, and necessitates rewarming with the possibility of cerebral hyperthermia [3], defining the Accepted for publication Dec 13, Address reprint requests to Dr Johnston, Department of Anesthesiology, The University of Texas Medical Branch, 301 University Blvd, Galveston, TX ; william.johnston@utmb.edu. optimal temperature for maximal neurologic recovery is important. Another complicating factor during extracorporeal circulation is hyperglycemia that is frequently encountered secondary to the endogenous stress response to nonpulsatile perfusion and contact activation, from enhanced renal tubular reabsorption of filtered glucose, and from cardioplegic and pump prime solutions [3 7]. Commonly, serum glucose concentrations may be elevated by 200% to 350% during CPB. In normothermic experimental animals, preischemic hyperglycemia can exacerbate the extent of neurologic injury and potentiate regional lactic acidosis in brain tissue [8, 9]. Although attempts are made to attenuate surges in serum glucose levels during CPB, maintaining normoglycemia is difficult and therapy to combat hyperglycemia is not without substantial independent risk [6]. As hypothermia represents one of the few treatments shown to attenuate enhanced ischemic damage in hyperglycemic animals [10, 11], reason dictates that it may be beneficial under hyperglycemic CPB conditions as well, in which secondary ischemic risk may occur from hemodilution, nonpulsatile flow, and cerebral micro- and 2001 by The Society of Thoracic Surgeons /01/$20.00 Published by Elsevier Science Inc PII S (01)

2 1326 CONROY ET AL Ann Thorac Surg HYPERGLYCEMIC CEREBRAL ISCHEMIA DURING CPB 2001;71: macroemboli. However, the optimal temperature is not known. Several studies in rodents without extracorporeal circulation have shown that preischemic cooling to 33 C to 35 C markedly or completely attenuates excitatory amino acids release and improves cerebral histopathology [12 14]. In contrast, in an earlier study using a porcine CPB model of global cerebral ischemia [15], 31 C hypothermia but not 34 C proved beneficial in terms of reducing excitotoxicity and S-100 protein release and improving electroencephalographic (EEG) and cerebral metabolic recovery. The clinical relevance of the latter study was limited by the lack of hyperglycemic response in pigs during CPB and the lack of postoperative survival data necessary to substantiate any neuroprotective benefit from hypothermia. Accordingly, the present study was designed to examine whether 34 C or 31 C hypothermia during cerebral ischemia with hyperglycemic CPB in surviving pigs would reduce histologic injury and improve EEG recovery when compared with normothermic animals. Material and Methods Animal Profiles and Surgical Specimens Animals were handled according to the guidelines approved by the American Physiological Society and the Guide for the Care and Use of Laboratory Animals, published by the National Institutes of Health (publication 85-23, revised 1985). The Institutional Animals Care and Use Committee approved this study. General anesthesia was induced using ketamine (20 mg/kg intramuscularly) in 27 mature, female Yorkshire pigs weighing kg (range 25 to 36 kg). After endotracheal intubation, each animal was ventilated with a tidal volume of 15 ml/kg using an air and oxygen mixture. The respiratory rate was adjusted to maintain normocapnia as measured by capnography (Datex Instrumentation, Helsinki, Finland). Core temperature was recorded using a precalibrated thermistor probes (Yellow Springs Instruments, Yellow Springs, OH) placed in the distal esophagus 40 cm from the snout, in the rectum 3 cm from the anal opening, and in the left ear abutting the tympanic membrane. Anesthesia was maintained with a bolus dose of fentanyl (15 g/kg) and diazepam (0.2 mg/kg) followed by a continuous infusion of fentanyl (10 g kg 1 h 1 ) and diazepam (0.3 mg kg 1 h 1 ) through a catheter in a lateral auricular vein. Isoflurane (0.5% to 1.5% inspired concentration) was administered as necessary to maintain a sufficient depth of anesthesia during the surgical preparation [15 17]. Isoflurane (0.25% inspired concentration) was continued throughout CPB in all animals. Muscle paralysis was provided by 0.1 mg/kg pancuronium administered intravenously as necessary to prevent shivering. The bladder was catheterized using a 10F Foley catheter under direct vision. All animals were tracheostomized for postoperative ventilation. Via a 4-cm epigastric incision, a 14F Foley catheter was inserted into the jejunum to allow enteral feeding throughout the postoperative period. Through a median sternotomy incision, fluidfilled catheters were tunneled through the abdominal wall and inserted into the aorta through the right and left internal mammary arteries and into the superior vena cava through the right internal mammary vein. Our experimental model for producing cerebral ischemia during CPB has been described in detail previously [15, 16]. Briefly, the innominate and left subclavian arteries, just distal to their origin from the aorta, were isolated and loosely encircled with suture ligatures. The pericardium was incised and tented. Heparin (400 U/kg) was administered before inserting a 20F infusion cannula (model 76020, Medtronic DLP, Grand Rapids, MI) in the ascending aortic arch and a 39F two-stage venous drainage cannula (model TF2937-0, Research Medical Inc, Salt Lake City, UT) through the right atrium to harvest venous return. Heparin doses were repeated during CPB as necessary to maintain the activated clotting time above 400 seconds. A 20-gauge catheter was inserted directly into the transverse aortic arch to measure mean aortic pressure during the cerebral ischemic interval. The oxygenator was primed with 1,000 ml crystalloid solution (Plasmalyte; Baxter Edwards Critical Care, Deerfield, IL) and 500 ml 6% hetastarch (Hespan; Dupont Pharmaceuticals, Wilmington, DE). A 40- m arterial blood filter and bubble trap (model SP3840; Pall Biomedical Products, East Hills, NY) was inserted in the arterial infusion line, and a membrane oxygenator with reservoir unit (model VPCML Plus; Cobe Cardiovascular, Arvada, CO) was used. Nonpulsatile perfusion was provided by a roller pump (model 5000; Sarns, Ann Arbor, MI). During hypothermic CPB, the arterial carbon dioxide tension was regulated using an -stat technique. Protocol Before bypass, animals were randomly assigned to one of three temperature groups as measured by the esophageal thermistor: 37 C (n 9), 34 C (n 10), and 31 C (n 8). Animals were placed on total CPB with pump flow adjusted between 80 and 120 ml kg 1 to maintain mean aortic pressure at 80 mm Hg. After stabilizing on CPB for 15 minutes, hemodynamic measurements were acquired. Next, the animals esophageal temperatures were maintained at normothermia or reduced using a thermostatistically controlled water bath (Sarns) while blood glucose levels were increased to 300 mg/dl by the addition of 50% dextrose solution to the venous reservoir. After stabilizing each animal at the assigned levels of temperature and blood glucose for 15 minutes, repeat measurements were obtained. No attempt was made to alter or regulate serum glucose after this point. Next, all animals were subjected to 15 minutes of global cerebral ischemia by temporarily ligating both innominate and left subclavian arteries. During ischemia, pump flow was adjusted to maintain mean aortic pressure (directly measured in the ascending aorta) at 50 mm Hg, and repeat hemodynamic measurements were acquired. The ligatures were then released and all animals were reperfused for 10 minutes and rewarmed over 20 minutes before repeat hemodynamic measurements were obtained. During

3 Ann Thorac Surg CONROY ET AL 2001;71: HYPERGLYCEMIC CEREBRAL ISCHEMIA DURING CPB 1327 both cooling and rewarming phases, the temperature gradient between the water bath and core (esophageal) sites was maintained less than 10 C. The maximal water bath temperature during rewarming was 39 C. Ventilation was resumed and the animals were weaned from CPB with the use of a dilute epinephrine solution (1 mg mixed in 500 ml saline) to prevent right ventricular dilation during subsequent volume loading. The epinephrine infusion was short-lived (30 to 60 seconds) and immediately discontinued in all animals. After decannulation, protamine (2 mg/kg) was administered, the sternum was wired closed, and the wound sutured. Animals were placed in padded slings, transferred to an intensive care unit, and ventilated mechanically (Servo ventilator 9000, Siemens-Elena, Solna, Sweden) with a 12-mL/kg tidal volume. The frequency of ventilation was set by the respiratory rate of the animal and the respired oxygen tension was adjusted to maintain the arterial oxygen tension at more than 100 mm Hg. During the early postoperative course, the basal infusion of fentanyl and diazepam was continued for 4 to 5 hours and discontinued. Cephazolin 30 mg/kg was administered intravenously every 12 hours. Buprenorphine 0.15 mg was administered intravenously every 12 hours for pain, and diazepam (2.5-mg boluses) was administered as necessary for sedation. Pancuronium in 2-mg boluses was titrated to prevent gross movement. A dilute norepinephrine infusion (4 mg mixed in 500 ml saline) was titrated as necessary to maintain mean aortic pressure above 60 mm Hg. Rectal and tympanic temperatures were measured continuously by thermistor probes (Yellow Springs) and recorded. Tympanic temperature was maintained between 37.5 C and 38.5 C using a servocontrolled temperature-sensing device that regulated an overhead warmer or an oscillating fan for heating or cooling the animal as necessary. All animals were closely monitored postoperatively for 24 hours (37 C animals) or 72 hours (34 C and 31 C animals). During this period, enteral feedings were continued through the jejunostomy catheter using an infusion of full strength isocal (Mead-Johnson Laboratories, Princeton, NJ) at 1 ml kg 1 h 1. Measurements Arterial blood gases and electrolytes (Model 1306, Instrumentation Laboratory, Lexington, MA) and serum hemoglobin (CO-Oximeter Model 482, Instrumentation Laboratory) were measured repeatedly throughout the experiment. Plasma glucose level was determined by a glucometer (Lifescan Inc, Milpitas, CA). Postoperatively, hemodynamic and temperature measurements were recorded every 4 hours. Electroencephalographic Analysis A parietal EEG was recorded from each hemisphere after induction of anesthesia (baseline), on arrival in the intensive care facility, and thereafter every 24 hours until termination of the experiment. Electroencephalographic signals were analyzed using a 5-point scale previously validated during porcine cerebral ischemia [17]. Each recording was reviewed independently by 2 investigators blinded to the animal s treatment group and assigned a numerical score for which 5 normal EEG, indistinguishable from baseline EEG; 4 mildly (25%) depressed amplitude and frequency from baseline; 3 moderately (50%) depressed amplitude and frequency from baseline; 2 markedly (75%) depressed amplitude and frequency from baseline; and 1 isoelectric EEG. Interobserver reliability of the EEG analysis score was tested using a weighted kappa statistic and found to be 0.62, indicating satisfactory agreement between the 2 observers. In addition, the EEG signal was continuously monitored during the first 2 hours after arrival in the intensive care unit to evaluate the presence of seizure activity. S-100 Protein Assay Upon arrival in the intensive care unit, an arterial blood sample was obtained to determine S-100 protein concentration. Repeat samples were obtained on each subsequent day until the termination of the experiment. All samples were immediately centrifuged and the supernatant analyzed for S-100 protein content using a monoclonal two-site immunoradiometric assay (Sangtec 100, Sangtec Medical AB, Bromma, Sweden). Samples were analyzed in duplicate and rejected if more than 10% variability resulted. Using this technique, values as low as 0.1 g/l are quantifiable. Cerebral Histopathology After 1-day (37 C group) or 3-day (34 C and 31 C group) survival, animals were reanesthetized with isoflurane (1.5% to 2.0% inspired concentration). Histologic appearance of necrotic injury in the brain is typically seen 3 days after ischemia [18]; this process can be accelerated to 24 hours under hyperglycemic conditions [19, 20]. The sternotomy incision was reopened and the ascending aorta cannulated. Loose snare tapes were placed around the descending aorta and the inferior vena cava. Cardiac asystole was produced by intravenous potassium chloride and the snares were tightened. After occluding the descending aorta and incising the superior vena cava, the ascending aortic arch was perfused by roller pump with 0.9% saline at 80 mm Hg until the venous effluent was clear. Next, 10 L of cold (4 C) 10% neutral buffered formalin were infused at 80 mm Hg; the brain was removed and refrigerated for at least 5 days in formalin solution before histopathologic examination. Sections for histopathology were taken from (1) the right frontal cortex at the level of the anterior basal ganglia, (2) the right mid-hippocampus, (3) the right occipital cortex, and (4) a midsagittal slice of the cerebellar vermis with a horizontal slice of the right cerebellar hemisphere. Brain tissue was dehydrated in a graded series of alcohols, cleared in xylene, and embedded in paraffin; 6- m thick sections were stained with hematoxylin and eosin. Histopathologic evaluation was performed by a neuropathologist who was blinded to the treatment group. Ischemic neurons were defined by cytoplasmic eosinophilia, shrinkage of the cytoplasm

4 1328 CONROY ET AL Ann Thorac Surg HYPERGLYCEMIC CEREBRAL ISCHEMIA DURING CPB 2001;71: Table 1. Definitions of Ischemic Injury Scores by Brain Region Brain Region Ischemic Injury Grade Frontal, occipital, and temporal cortex Normal Rare ischemic neurons ( 10%) Hippocampus Normal Rare ischemic pyramidal or granule cells Cerebellum Normal Rare ischemic Purkinje cells Frequent ischemic neurons (10% to 50%) Focal ischemic damage 10% to 50% ischemic Purkinje cells Mostly ischemic neurons ( 50%) Severe, diffuse ischemic damage 50% ischemic Purkinje cells with loss of definition of Nissl substance, and homogenization or karyorrhexis of the nucleus. Other changes seen in areas of severe damage included inflammatory infiltrates (polymorphonuclear cells around vessels, microglia in parenchyma, macrophages), vascular proliferation, or cavitation. The extent of ischemic injury in these regions was graded according to the criteria in Table 1. All hippocampal regions were examined, including the dentate gyrus, the CA1 through CA4 subfields, and the subiculum. Cumulative neuropathology score for each animal was obtained by summing the score for each individual region (best possible score 15; worst possible score 0). This histopathologic scoring system had been developed and validated previously in experimental animals [21, 22]. As large areas of necrosis were seen in some animals, the sections containing frontal, temporal, and occipital cortex were further evaluated to assess the degree of necrosis. Morphometry was performed using a Nikon Optiphot microscope with a Hitachi HV-C10 CCD color video camera, and Bioscan OPTIMAS 4.10 software on a computer system equipped with an Imaging Technology Vision Plus-AT CFG digitizing card. On each of these sections, the total cortical area was measured, the regions of necrosis Table 2. Descriptive Data of the Hemodynamic and Oxygenation Variables Throughout the Experimental Bypass Procedure Cardiopulmonary Bypass Variable Group Baseline Temperature and Glucose Management Ischemia Rewarming Pump flow (ml/min/kg) 31 C C C Mean aortic pressure (mm Hg) 31 C C C Hemoglobin (g/dl) 31 C C C Arterial ph 31 C C C Arterial carbon dioxide tension (mm Hg) 31 C C C Arterial oxygen tension (mm Hg) 31 C C C Esophageal temperature ( C) 31 C C C Tympanic temperature ( C) 31 C C C Serum glucose (mg/dl) 31 C C C Data are presented as mean SD with 8 animals per group.

5 Ann Thorac Surg CONROY ET AL 2001;71: HYPERGLYCEMIC CEREBRAL ISCHEMIA DURING CPB 1329 Table 3. Cumulative Anesthetic Drug Dosages During the Experimental Procedure and Postoperative Period Group Intraoperative Drug Dosages Fentanyl ( g/kg) Diazepam (mg/kg) Postoperative Drug Dosage Diazepam (mg/kg/day) 31 C C C Data are presented as mean SD. were outlined, and the percent necrotic cortex was calculated. The percentages from the three cortical regions were averaged to give an estimate of the extent of necrosis. Statistical Analysis Primary outcome variables for repeated measures such as EEG scores and serum S-100 protein concentrations were analyzed using analysis of variance for a two-factor experiment (temperature and time) with repeated measures over time. Other primary variables without repeated measures (ie, cerebral histopathology) were analyzed using the Kruskal Wallis test. Fisher s least significant difference procedure was used for multiple comparisons with a Bonferroni adjustment. Secondary variables that were closely controlled throughout bypass (mean aortic blood pressure, pump flow, oxygenation variables, temperature, serum glucose) and postoperatively (oxygenation variables, mean aortic blood pressure) were not analyzed but are presented in tables below. An a priori decision was made to limit the number of comparisons to reduce the possibility for type I errors. Data are expressed as mean SD with p less than 0.05 to indicate significance where applicable. Results The study was successfully completed in 24 of 27 animals. Two animals (n 1 at 37 C; n 1 at 34 C) died intraoperatively secondary to surgical complications (torn cavoatrial junction; torn left subclavian artery). Another animal (n 1 at 37 C) died the first postoperative night because of spontaneous extubation. Table 4. Descriptive Data of Postoperative Hemodynamic and Oxygenation Variables Expressed as Daily Averages Variable Group Postoperative Day Mean aortic pressure (mm Hg) 31 C C C Heart rate (beats/min) 31 C C C Central venous pressure (mm Hg) 31 C C C Rectal temperature ( C) 31 C C C Tympanic temperature ( C) 31 C C C Arterial ph 31 C C C Arterial carbon dioxide tension (mm Hg) 31 C C C Arterial oxygen tension (mm Hg) 31 C C C Serum glucose (mg %) 31 C C C Serum hemoglobin (g/dl) 31 C C C Data are presented as mean SD.

6 1330 CONROY ET AL Ann Thorac Surg HYPERGLYCEMIC CEREBRAL ISCHEMIA DURING CPB 2001;71: Table 5. Electroencephalographic Scores and S-100 Protein Concentrations During the Postoperative Period Variable Group Baseline Postoperative Day EEG score 31 C a a C a,c a,b,c a,c a,c 37 C a,c a,c,d Serum S-100 protein ( g/l) 31 C C C a p 0.05 versus baseline value. b p 0.05 versus previous value. c p 0.05 versus 31 C value. d p 0.03 versus 34 C value. EEG electroencephalogram. Data are presented as mean SD with 8 animals per group. As seen in Table 2, hemodynamic and oxygenation variables were tightly controlled throughout the experimental bypass procedure, particularly the desired endpoints of esophageal temperature and serum glucose concentrations. At all steady state measurement intervals, tympanic temperatures tracked closely with esophageal temperature. All animals received comparable doses of fentanyl, diazepam, and isoflurane during the experimental protocol (Table 3). The amount of 50% dextrose added to the venous reservoir to achieve a serum glucose concentration approximating 300 mg/dl was 38 10, 50 20, and ml in the 31 C, 34 C, and 37 C groups, respectively. Averaged daily oxygenation and hemodynamic variables are presented in Table 4. All animals maintained acceptable acid base status and only 2 animals in the 37 C group required temporary vasopressor support with norepinephrine infusion to support mean aortic blood pressure above 60 mm Hg during the first postoperative day. Although no attempt was made to regulate serum glucose after initial loading in CPB, levels appeared comparable among groups postoperatively (Table 4). Cumulative diazepam doses in the postoperative period are shown in Table 3. Postoperatively, overt seizure activity was observed during the initial 2 hours of continuous EEG monitoring in 6 of 8 normothermic animals. Seizure activity was repetitive but short-lived, lasting only 3 to 5 seconds, and was self-terminating in all cases. In contrast, none of the animals managed at 31 C or 34 C during bypass had evidence of seizure activity during a similar observational period. In the immediate postoperative period, all groups had significantly lower EEG scores than baseline (Table 5). By postoperative day 2, the 31 C animals had EEG scores that were not different from baseline. In contrast, EEG scores in the 37 C animals were significantly lower without improvement on postoperative day 1. The 34 C animals showed significant improvement on day 1, although their EEG scores remained consistently lower than the 31 C animals. Serum S-100 protein levels were similar in the two hypothermic groups immediately after bypass and during the postoperative period. Although mean S-100 values were 130% greater in 37 C animals than 31 C animals on postoperative day 1, there were no significant intergroup differences (Table 5). Brain histology scores are shown cumulatively in Table 6 and individually in Figure 1. By design, 37 C animals were sacrificed on day 1, and 31 C and 34 C animals on day 3. In each brain region, 37 C and 34 C groups had significantly lower scores than the 31 C animals. In the hippocampus, temporal cortex, and cerebellum, 34 C animals had improved scores over 37 C animals, whereas histology scores were not different in the frontal and occipital cortices. The total histology scores were the worst for the 37 C group, and the 34 C group was worse than the 31 C group. Overall, there was a strong correlation between postoperative EEG scores taken on day 1 (37 C) and day 3 (31 C; 34 C) and total neurohistopathologic scores for all groups (r ; p 0.001). Representative sections from the frontal cortex are shown in Figure 2. With respect to the extent of pannecrosis, although the three treatment groups did not differ significantly, 37 C animals tended to have a higher Table 6. Brain Histology Scores 1 Day (37 C) or 3 Days (34 C; 31 C) After Hyperglycemic Cerebral Ischemia During Cardiopulmonary Bypass Brain Region Group Frontal Cortex Hippocampus Temporal Cortex Occipital Cortex Cerebellum Total 31 C C a a a a a a 37 C a a,b a,b a a,b a,b Data are presented as mean SD. a p 0.05 versus 31 C. b p 0.05 versus 34 C.

7 Ann Thorac Surg CONROY ET AL 2001;71: HYPERGLYCEMIC CEREBRAL ISCHEMIA DURING CPB 1331 Fig 1. Individual neurohistopathology scores for all animals per brain region are depicted. The 31 C animals had consistently higher scores than 34 C or 37 C groups, indicating less ischemic injury with 31 C temperature management. (Circles indicate 31 C; squares indicate 34 C; and triangles indicate 37 C during the period of cerebral ischemia. For the frontal (A), temporal cortex (C), and occipital (D), 3 normal, 2 rare ( 10%) ischemic neurons, 1 frequent (10 50%) ischemic neurons, 0 majority ( 50%) ischemic neurons. For the hippocampus (B), 3 normal, 2 rare ischemic pyramidal or granule cells, 1 focal ischemic damage, 0 severe, diffuse ischemic damage. For the cerebellum (E), 3 normal, 2 rare ischemic Purkinje cells, 1 10 to 50% Purkinje cells, 0 50% ischemic Purkinje cells). percent necrosis in the frontal cortex (34.6% 43.9%) than 31 C (1.4% 3.9%; p 0.07) and 34 C (2.9% 4.3%; p 0.36) animals. Similarly, in the occipital and temporal cortices, there was a trend for greater necrotic injury in 37 C animals (37.5% 51.8% and 19.3% 32.4%, respectively) than in 31 C animals (2.8% 7.4%; p 0.37, and 0% 0%; p 0.06, respectively) and 34 C animals (0.4% 0.7%; p 0.58, and 0% 0%; p 0.06, respectively). Comment Data from this study indicate that 31 C hypothermic management prior to transient but severe cerebral ischemia during hyperglycemic CPB provides superior neuroprotection than 34 C or 37 C. With 31 C, EEG scores returned to near baseline values 3 days later, and histologic recovery was significantly improved. Although some degree of neuroprotection was evidenced with 34 C management when compared with normothermia, 31 C had superior results to either group. This study is consistent with our previous observation that preischemic hypothermia to 31 C during CPB reduces the excitotoxic response to cerebral ischemia by attenuating release of glutamate and aspartate more so than 34 C or 37 C [15]; this attenuation may represent a primary mechanism of hypothermic neuroprotection. Furthermore, even though hyperglycemia alone can increase brain lactate during

8 1332 CONROY ET AL Ann Thorac Surg HYPERGLYCEMIC CEREBRAL ISCHEMIA DURING CPB 2001;71: Fig 2. The spectrum of injury seen in the frontal cortex; similar changes were seen in other cortical regions. (A) 31 C animal: the cortex is normal. Neurons have a large, round nucleus with a single prominent nucleolus. (B) 31 C animal with scattered clusters of angular ischemic neurons with pyknotic nuclei (examples at arrows). (C) 34 C animal with a laminar zone of ischemic neurons and vacuolization (edema) in the neuropil. At the upper edge of the photograph a few intact neurons are present (curved arrows). (D) 37 C animal: all cells are dead in this region. The animal had a 24-hour survival time and this represents an area of early infarct. (Hematoxylin & eosin stains. Magnification for A D, 250.) CPB [23], hypothermia s neuroprotection persisted despite elevation of the serum glucose to 300 mg/dl. Neurologic complications continue to be a major factor after cardiac surgery [24]. Stroke after CPB occurs more frequently than after other major surgical procedures and can increase patient mortality by five- to tenfold. Neurocognitive dysfunction correlates with increased number of microemboli during bypass. Patients at particular risk for ischemic neurologic injury can be predicted [24] so that their course of intraoperative management may be altered to influence outcome [25]. However, the controversy over the optimal temperature management of cardiac surgical patients continues with both proponents and opponents of systemic cooling. Engleman and associates [1, 26] reported no relationship between perfusion temperature and neurologic outcome. Neurologic function deteriorated in 68% of patients immediately after operation regardless of temperature; strokes occurred in 4.5% of patients with equal distribution in cold, tepid, and warm groups [26]. In a more recent report, ischemic strokes were equally distributed in hypothermic and tepid/normothermic groups with similar infarct volumes [1]. In contrast, Gaudino and colleagues [2] observed that the severity of the neurologic insult defined by computed tomographic scan and neurologic outcome was significantly lessened by hypothermia despite a similar incidence of strokes in normothermic and hypothermic patients. Serum glucose data were not presented in the latter study. In terms of neuropsychological outcome, cooling to 28 C to 32 C reduced the degree of deterioration when compared with 37 C management [27], although hyperglycemia may attenuate this benefit [28]. In experimental animals, preischemic hyperglycemia under normothermic conditions can markedly potentiate the severity of cerebral ischemic injury, as demonstrated by brain lactate accumulation [8, 9] causing intracellular and extracellular acidosis [9, 29], edema formation [30], reduced cerebral reperfusion [31], a higher incidence of seizure activity [19, 32, 33], and accelerated morphologic injury demonstrated by histopathology [20] and infarct size [34]. The histopathologic manifestations of cerebral ischemic injury usually require 3 to 4 days to progress and summate. Hyperglycemia can markedly accelerate this process to less than 24 hours after ischemia [20], which is why we sacrificed 37 C animals at the end of the

9 Ann Thorac Surg CONROY ET AL 2001;71: HYPERGLYCEMIC CEREBRAL ISCHEMIA DURING CPB 1333 first postoperative day [10]. Sacrificing the 37 C animals at a later time (eg, 72 hours) would likely have worsened the degree of histopathologic injury seen in our experiments. Hence, the assessment of injury at 24 hours in 37 C animals probably underestimated the damage present at 72 hours. Although numerous studies have documented the detrimental effect of hyperglycemia during normothermic cerebral ischemia, few have examined whether hypothermia can attenuate this damage. Lundgren and colleagues [10] found that cooling hyperglycemic rats to 32 C to 33 C before 15 minutes of cerebral ischemia reduced the incidence of seizures and improved histopathologic scores in some (parietal cortex, hippocampus, caudoputamen) but not all (cingulate cortex, medial and lateral ventero-posterior thalamic nuclei, substantia nigra) brain regions. In a canine model of hypothermic circulatory arrest for 105 minutes, Vannucci and colleagues [35] found that hyperglycemia caused greater ischemic injury, particularly in the amygdaloid nucleus, caudate nucleus, and brainstem than normoglycemia. Our data support the concept of regional vulnerability to ischemia in the brain in which intermediate (34 C) hypothermia showed greatest benefit in the hippocampus, temporal cortex, and cerebellum with less improvement in the frontal and occipital cortices. More profound hypothermia to 31 C provided significant but equal protection in all brain regions. These data suggest that the frontal and occipital cortices are more resistant to therapy and may require lower temperature for neuroprotection. Although potential mechanisms were not examined in this study, hypothermic neuroprotection may have resulted from reduced metabolic demands and excitotoxicity. Hyperglycemia in normothermic animals subjected to 15 minutes of forebrain ischemia increases extracellular glutamate in neocortical regions, which parallels neuronal damage [36]. Hypothermia can attenuate release of excitotoxic amines during ischemia [15, 37] even under hyperglycemic conditions [11]. Hypothermia does not just delay the time when histologic damage is seen [10]. Similar amounts and patterns of injury were found 15 minutes after ischemia as those occurring 1 week later. In addition, cooling to 30 C reduces blood brain barrier permeability changes caused by hyperglycemia and cerebral ischemia [38]. In conclusion, hypothermia during CPB can significantly reduce the morphologic consequences of severe, temporary cerebral ischemia occurring under hyperglycemic conditions during extracorporeal circulation. Greater neuroprotection was found with 31 C than 34 C under these circumstances. We thank Tatsuo Uchida, MS, for statistical review, Christy Perry for manuscript preparation, and Jordan Kicklighter for editorial review. This study was supported by a grant from the John Sealy Memorial Endowment Fund for Biomedical Research, University of Texas Medical Branch, Galveston, Texas (W.E.J.). References 1. Engelman RM, Pleet AB, Hicks R, et al. Is there a relationship between systemic perfusion temperature during coronary artery bypass grafting and extent of intraoperative ischemic central nervous system injury? J Thorac Cardiovasc Surg 2000;119: Gaudino M, Martinelli L, Di Lella G, et al. Superior extension of intraoperative brain damage in case of normothermic systemic perfusion during coronary artery bypass operations. J Thorac Cardiovasc Surg 1999;118: Guyton RA, Gott JP, Brown WM, Craver JM. Cold and warm myocardial protection techniques. Adv Card Surg 1996;7: Lehot JJ, Piriz H, Villard J, Cohen R, Guidollet J. Glucose homeostasis. Comparison between hypothermic and normothermic cardiopulmonary bypass. Chest 1992;102: Braden H, Cheema-Dhadli S, Mazer CD, McKnight DJ, Singer W, Halperin ML. Hyperglycemia during normothermic cardiopulmonary bypass: the role of the kidney. Ann Thorac Surg 1998;65: Chaney MA, Nikolov MP, Blakeman BP, Bakhos M. Attempting to maintain normoglycemia during cardiopulmonary bypass with insulin may initiate postoperative hypoglycemia. Anesth Analg 1999;89: Kuntschen FR, Galletti PM, Hahn C, Arnulf JJ, Isetta C, Dor V. Alterations of insulin and glucose metabolism during cardiopulmonary bypass under normothermia. J Thorac Cardiovasc Surg 1985;89: Chopp M, Welch KM, Tidwell CD, Helpern JA. Global cerebral ischemia and intracellular ph during hyperglycemia and hypoglycemia in cats. Stroke 1988;19: Combs DJ, Dempsey RJ, Maley M, Donaldson D, Smith C. Relationship between plasma glucose, brain lactate, and intracellular ph during cerebral ischemia in gerbils. Stroke 1990;21: Lundgren J, Smith ML, Siesjo BK. Influence of moderate hypothermia on ischemic brain damage incurred under hyperglycemic conditions. Exp Brain Res 1991;84: Li PA, He QP, Miyashita H, Howllet W, Siesjo BK, Shuaib A. Hypothermia ameliorates ischemic brain damage and suppresses the release of extracellular amino acids in both normothermic and hyperglycemic subjects. Exp Neurol 1999; 158: Busto R, Dietrich WD, Globus MY, Ginsberg MD. Postischemic moderate hypothermia inhibits CA1 hippocampal ischemic neuronal injury. Neurosci Lett 1989;101: Busto R, Dietrich WD, Globus MY, Valdes I, Scheinberg P, Ginsberg MD. Small differences in intraischemic brain temperature critically determine the extent of ischemic neuronal injury. J Cereb Blood Flow Metab 1987;7: Minamisawa H, Nordström C-H, Smith M-J, Siesjö BK. The influence of mild body and brain hypothermia on ischemic brain damage. J Cereb Blood Flow Metab 1990;10: Conroy BP, Lin CY, Jenkins LW, et al. Hypothermic modulation of cerebral ischemic injury during cardiopulmonary bypass in pigs. Anesthesiology 1998;88: Conroy BP, Black D, Lin CY, et al. Lamotrigine attenuates cortical glutamate release during global cerebral ischemia in pigs on cardiopulmonary bypass. Anesthesiology 1999;90: Johnston WE, Jenkins LW, Lin CY, et al. Cerebral metabolic consequences of hypotensive challenges in hemodiluted pigs with and without cardiopulmonary bypass. Anesth Analg 1995;81: Pulsinelli WA, Brierley JB, Plum F. Temporal profile of neuronal damage in a model of transient forebrain ischemia. Ann Neurol 1982;11: Li PA, Shamloo M, Smith ML, Katsura K, Siesjo BK. The influence of plasma glucose concentrations on ischemic brain damage is a threshold function. Neurosci Lett 1994;177: Smith ML, Kalimo H, Warner DS, Siesjo BK. Morphological

10 1334 CONROY ET AL Ann Thorac Surg HYPERGLYCEMIC CEREBRAL ISCHEMIA DURING CPB 2001;71: lesions in the brain preceding the development of postischemic seizures. Acta Neuropathol (Berl) 1988;76: Fleischer JE, Nakakimura K, Drummond JC, et al. Effects of levemopamil on neurologic and histologic outcome after cardiac arrest in cats. Crit Care Med 1992;20: Fleischer JE, Tateishi A, Drummond JC, et al. MK-801, an excitatory amino acid antagonist, does not improve neurologic outcome following cardiac arrest in cats. J Cereb Blood Flow Metab 1989;9: Feerick AE, Johnston WE, Jenkins LW, Lin CY, Mackay JH, Prough DS. Hyperglycemia during hypothermic canine cardiopulmonary bypass increases cerebral lactate. Anesthesiology 1995;82: Selnes OA, Goldsborough MA, Borowicz LM, McKhann GM. Neurobehavioural sequelae of cardiopulmonary bypass. Lancet 1999;353: Gaudino M, Glieca F, Alessandrini F, et al. Individualized surgical strategy for the reduction of stroke risk in patients undergoing coronary artery bypass grafting. Ann Thorac Surg 1999;67: Engelman RM, Pleet AB, Rousou JA, et al. What is the best perfusion temperature for coronary revascularization? J Thorac Cardiovasc Surg 1996;112: Regragui I, Birdi I, Izzat MB, et al. The effects of cardiopulmonary bypass temperature on neuropsychologic outcome after coronary artery operations: a prospective randomized trial. J Thorac Cardiovasc Surg 1996;112: Griffin S, Klinger L, Newman S, et al. The effect of substrate load and blood glucose management on cerebral dysfunction following cardiopulmonary bypass. Vasc Surg 1992;26: Katsura K, Folbergrova J, Siesjö BK. Changes in labile energy metabolites, redox state and intracellular ph in postischemic brain of normo- and hyperglycemic rats. Brain Res 1996;726: Huang NC, Wei J, Quast MJ. A comparison of the early development of ischemic brain damage in normoglycemic and hyperglycemic rats using magnetic resonance imaging. Exp Brain Res 1996;109: Palmon SC, Sieber FE, Brown PR, Koehler RC, Eleff SM, Traystman RJ. Poor hemodynamic and metabolic recovery after global incomplete cerebral ischemia associated with short-term diabetes in dogs. J Cereb Blood Flow Metab 1995; 15: Warner DS, Todd MM, Dexter F, Ludwig P, McAllister AM. Temporal thresholds for hyperglycemia-augmented ischemic brain damage in rats. Stroke 1995;26: Uchino H, Smith ML, Bengzon J, Lundgren J, Siesjo BK. Characteristics of postischemic seizures in hyperglycemic rats. J Neurol Sci 1996;139: Prado R, Ginsberg MD, Dietrich WD, Watson BD, Busto R. Hyperglycemia increases infarct size in collaterally perfused but not end-arterial vascular territories. J Cereb Blood Flow Metab 1988;8: Vannucci RC, Rossini A, Towfighi J. Effect of hyperglycemia on ischemic brain damage during hypothermic circulatory arrest in newborn dogs. Pediatr Res 1996;40: Li PA, Shuaib A, Miyashita H, He QP, Siesjo BK, Warner DS. Hyperglycemia enhances extracellular glutamate accumulation in rats subjected to forebrain ischemia. Stroke 2000;31: Busto R, Globus MY, Dietrich WD, Martinez E, Valdes I, Ginsberg MD. Effect of mild hypothermia on ischemiainduced release of neurotransmitters and free fatty acids in rat brain. Stroke 1989;20: Dietrich WD, Alonso O, Busto R. Moderate hyperglycemia worsens acute blood brain barrier injury after forebrain ischemia in rats. Stroke 1993;24:111 6.