Optimal blood glucose levels while using insulin to minimize the size of infarction in focal cerebral ischemia

Transcription

1 J Neurosurg 101: , 2004 Optimal blood glucose levels while using insulin to minimize the size of infarction in focal cerebral ischemia CHANG Z. ZHU, M.D., AND ROLAND N. AUER, M.D., PH.D. Departments of Pathology and Clinical Neuroscience, Faculty of Medicine, University of Calgary, Alberta, Canada Object. Insulin has been shown to ameliorate cerebral necrosis in global and, more recently, in focal cerebral ischemia. The goal of this study was to determine the relationship between this neuroprotective effect and blood sugar levels in a rat model of focal ischemia. Methods. Thirty-four rats were subjected to 80 minutes of transient middle cerebral artery occlusion at a mean arterial blood pressure of 60 mm Hg and a temperature of 37 C. Insulin (3.5 IU/kg) was administered 1 hour before (12 rats) and 20 minutes after (12 rats) ischemia; 10 animals served as controls. A quantitative histopathological study conducted after 1 week of survival showed that insulin was not beneficial in reducing the size of the infarction or selective neuronal necrosis in the penumbra when administered before or after ischemia. In addition to infarction, six animals from the insulintreated groups had bilateral selective neuronal necrosis in the hippocampus or the neocortex. A nonlinear regression analysis in which glucose levels were compared with both cortical necrosis and total infarction yielded a U-shaped curve with a nadir for cerebral necrosis that lay in the 6- to 7-mM blood glucose range. The increased brain damage induced by insulin occurred in animals with very low blood sugar values in the range of 2 to 3 mm. Conclusions. These results in rats indicate that if insulin is used following ischemia, blood glucose levels should be maintained at approximately 6 to 7 mm. From these data one can infer that hypoglycemia of less than 3 mm should be avoided in situations of focal cerebral ischemia in which insulin is used. Additional animal studies and clinical trials in humans are needed to study the effects of insulin on ischemia. KEY WORDS insulin ischemia focal hypoglycemia normoglycemia rat I NSULIN has been demonstrated to be beneficial in both global 14,25,28,31,35 and focal 10,12,20,22 models of cerebral ischemia and even in ischemic paraplegia of the spine. 15 The relationship between insulin-related hypoglycemia and the neuroprotective effect of insulin, however, differs in these two models of ischemia. In animals with global ischemia the neuroprotection remains active even if glucose is coadministered with the insulin to maintain the same blood sugar level present in untreated control animals. 13 The same species of animal subjected to focal ischemia induced by MCA occlusion, however, loses the neuroprotection conferred by insulin when given a coadministration of glucose. 10 These data together indicate that the neuroprotective effect of insulin is dependent on blood sugar levels in focal cerebral ischemia but works directly and centrally in global ischemia. The latter effect was demonstrated in global ischemia by direct intracerebroventricular injection; based on the findings of that study we infer that part of the central effect of insulin in global ischemia is due to the action of a growth factor mediated at insulin-like growth factor I receptors. 18,35 In rats with focal ischemia, Nedergaard and Diemer 20 Abbreviations used in this paper: ECA = external carotid artery; ICA = internal carotid artery; MCA = middle cerebral artery; SD = standard deviation. 664 found that insulin-induced hypoglycemia attenuated brain damage even when blood glucose levels were only 2 mm. Izumi and colleagues, 12 however, argued that the most favorable outcome of permanent focal ischemia following insulin administration would occur without marked and sustained hypoglycemia. At our own laboratory we found that blood glucose levels of 3 to 4 mm allowed neuroprotection in a model of transient focal ischemia in rats. 10 Nevertheless, experiments in cats conducted by de Courten-Myers, et al., 7 in which transient clip occlusion of the MCA was performed, have indicated that the rate of mortality and the size of the infarction are paradoxically increased by blood glucose levels in the range of 3 to 4 mm. In the present experiments we originally intended to determine the effectiveness of an infusion of insulin after focal ischemia, following up on our previous demonstration of insulin s effectiveness when administered before ischemia. 10 Lack of a demonstrable benefit, however, prompted us to examine our data to determine why no benefit occurred. A correlation between blood glucose levels and infarction size in a model of transient focal cerebral ischemia provided the answer. Materials and Methods Fifty-nine male Wistar rats (Charles River Breeding Centre, St. Constant, Quebec), each weighing 250 to 330 g, were used in the ex-

2 Insulin in focal brain ischemia periment. To determine the maximum tolerable dosage of insulin, 7 IU/kg and 5 IU/kg of insulin were given to groups of 25 rats on separate occasions in pilot experiments. Both these doses proved to be lethal for all rats, whereas the dose of 2.5 IU/kg, which we had used in a previous study, 10 produced only mild hypoglycemia. On the basis of these findings we selected a dose of 3.5 IU/kg insulin for the present study in which we used 34 rats divided among three groups. Experimental Groups The insulin that we used consisted of bovine crystalline zinc insulin (CZI; Novo Nordisk Canada Inc., Mississauga, ON, Canada) and longer acting zinc insulin suspension (NPH; Novo Nordisk Canada Inc.), which were administered intravenously and intraperitoneally, respectively. The rats were randomly allocated to one of three groups and were allowed free access to rat chow and water before and after the ischemic insult. The first group did not receive insulin and served as a control group (10 animals). The second group was treated with insulin 60 to 65 minutes before ischemia (12 animals) and was named the insulin pre group. A third group was treated with insulin 20 minutes after ischemia and was named the insulin post group (12 animals). Both insulin-treated groups received 3.5 IU/kg insulin (a ratio of 1:2 CZI/NPH). Afterward, all rats were allowed to experience a spontaneous level of glycemia, with neither insulin nor glucose administered subsequently. The resulting blood glucose levels are shown in Table 1. Ischemia Model After a 4- to 5-minute inhalation of 3 to 4% halothane added to a mixture of 40% O 2 and 60% N 2 O, each rat was intubated and given mechanical ventilation. Anesthesia was maintained using a 1% halothane mixture. The ventral tail artery was cannulated for both blood pressure monitoring and sampling of blood gas and glucose levels (Table 1). Blood pressure was automatically recorded by a computer every 5 seconds. A lateral tail vein was dissected and connected with P-50 polyethylene tubing to an infusion of 0.9% normal saline given at a rate of 4 ml/kg/hr to avoid dehydration and changes in the hematocrit level. The animal s body temperature was controlled using a thermistor-regulated feedback-controlled heating blanket with the core temperature set at 37 C. The temperature of the animal s head was monitored separately and controlled by an overhead heating lamp, which was located symmetrically over the skull vertex. A thermometer was located in the external auditory meatus and the head temperature was maintained at a mean of C. Blood glucose was measured using glucose oxidase reagent strips and a reflectance glucometer (Ames Co., Mishawaka, IN). A ph/blood gas analyzer (model 1304; Instrumentation Laboratories, Milan, Italy) was used to monitor blood gases and ph before, during, and 15 minutes after MCA occlusion. Glucose was monitored at 3, 12, and 24 hours and 7 days postischemia. In a previous study we examined blood pressure and the duration of MCA occlusion in this reperfusion model. 34 Based on the results of that study, we chose 80 minutes of occlusion at a blood pressure of 66 mm Hg to obtain an infarction of the desired size, which was smaller than 100 mm 3. To produce ischemia, the sternomastoid, digastric, and omohyoid muscles were separated by a right ventrolateral incision in the neck. With the aid of an operating microscope, the common carotid artery was isolated and the occipital and superior thyroid artery branches of the ECA were isolated and coagulated. The ECA was cut with microsurgical bipolar forceps and tied loosely with a 6-0 silk suture. A microvascular clip was placed temporarily at the origin of the ECA. The ECA stump was cut with fine surgical scissors, and a 22-mm 3-0 monofilament nylon thread, the tip of which was smoothed using fine-grade sandpaper, was introduced into the ICA through the ECA. The silk suture surrounding both the ECA stump and the intraluminal thread was tightened to prevent bleeding, and the microclip was removed. A 22-mm nylon thread was advanced into the ICA until a slight resistance was encountered and a sharp decrease of blood pressure was displayed on the computer screen, indicating that the tip of the thread had reached the proximal segment of the anterior cerebral artery. The length of the nylon thread introduced into the ICA was approximately 20 mm, TABLE 1 Physiological measurements in rats treated with insulin for focal ischemia Control Insulin pre Insulin post Parameter Group (10 rats) Group (12 rats) Group (12 rats) mean weight (g) mean weight loss (g) glucose (mm) preischemia intraischemia min postischemia hrs postischemia hrs postischemia hrs postischemia days postischemia ph preischemia intraischemia postischemia PCO 2 (mm Hg) preischemia intraischemia postischemia PO 2 (mm Hg) preischemia intraischemia postischemia body temperature (ºC) preischemia intraischemia postischemia head temperature (ºC) preischemia intraischemia postischemia blocking the ICA and the anterior and posterior cerebral arteries to the MCA. The mean arterial blood pressure was kept at 60 mm Hg by regulating the intake of halothane, requiring its concentration to be approximately 2% throughout the entire 80-minute ischemic insult. The nylon suture was pulled out, the ECA was permanently tied, and the blood pressure was allowed to recover by discontinuing the administration of halothane. A neurological assessment was conducted at 3, 12, and 168 hours (1 week) after the ischemic insult by a researcher blinded to the experimental groups to which the animals were assigned. The Bederson neurological grading criteria 2 were used to rank the severity of the neurological deficit, with the resulting grades of 0, 1, 2, and 3, respectively, for normal, mild, moderate, and severe neurological deficit. Quantification of the Size of the Infarction The rats in all groups were allowed to survive for 1 week after the operation. Animals that died before 1 week were excluded from the histopathological data analysis, because it was possible that their infarctions had not completely matured and were only included in the calculation of the mortality rate. After anesthesia had been induced by administration of a 2% halothane mixture, transcardiac perfusion was performed using saline followed by 4% phosphate-buffered formaldehyde, ph of 7.3. The brains were removed and cut into 3-mm coronal slices. After the tissue had been processed, embedded in paraffin, serially sectioned at 8- m thickness, and stained with hematoxylin and eosin, we prepared slides of brain tissue cut at 250- m intervals from 5.2 mm anterior to the bregma to 6.8 mm posterior to that structure. On each of these sections, the areas of infarcted cortex and striatum were traced by an operator blinded to the experimental groups, who used a video analysis system (Jandel Scientific, San Raphael, CA). The total area of each hemisphere was also measured separately because obvious atrophy was seen in the 665

3 C. Z. Zhu and R. N. Auer FIG. 1. Bar graph demonstrating that the differences in cortical necrosis, striatal necrosis, and the total volume of necrosis and atrophy do not differ significantly (means SD) among the three experimental groups, although there is an indication that the damage in Group 2 was most severe. Insulin-pre = insulin given before MCA occlusion; insulin-post = insulin given after MCA occlusion. ischemic hemisphere compared with the contralateral one; atrophy was calculated by subtracting the area of the ischemic hemisphere from the nonischemic hemisphere. Areas of cortical and striatal necrosis plus atrophy were summed to provide the total area of lost cerebral tissue. Volumes of infarction in the cortex and striatum, and the total tissue loss in the hemisphere were then determined using a three-dimensional reconstruction program (PC3D; Jandel Scientific), which aligns the sections along a reference point (the center of the corpus callosum) and integrates the area measurement in three dimensions to calculate the volume accurately. Quantification of Selective Neuronal Necrosis In addition to infarction, selective neuronal necrosis was found within 1 mm of the sharply demarcated rim of the infarction and, hence, was quantitated with the aid of a microscope by counting the absolute number of dead neurons in this histological penumbra. The absolute numbers of dead neurons were counted in the entire section and the sum from three standardized coronal levels was used in the statistical analysis. Neuronal necrosis in the ipsilateral and contralateral hippocampus and any contralateral hemispheric infarction were also noted qualitatively. Statistical Analysis Measures of physiological parameters were subjected to a oneway analysis of variance, with the Scheffé test for multiple comparisons. A two-way analysis of variance was used to determine the effect of hypotension and the duration of MCA occlusion on the volume of the infarction and on selective neuronal necrosis. The rate of mortality and the Bederson neurological scores were analyzed using nonparametric chi-square and Wilcoxon rank-sum tests, respectively. Second-order regression was used to analyze the relationship between blood glucose and infarction size. All data were summarized as means SDs. Differences were considered significant at a probability level lower than Results Physiological variables are presented in Table 1. Blood glucose levels before, during, and 10 minutes after ischemia were lower in the insulin pre group. In the insulin post group, the postischemic glucose level at 3 hours was considerably lower than the control group. The rate of mortality was FIG. 2. Bar graph showing that the differences in quantitated selective neuronal necrosis cells (mean SD) in the ischemic penumbra of the infarction were not significant in the three groups. Pre-insulin = insulin given before MCA occlusion; post-insulin = insulin given after MCA occlusion. greater in both insulin-treated groups eight rats in the insulin pre group, and five in the insulin post group compared with only two rats in the control group but failed to reach statistical significance (p 0.05, chi-square test). The majority of rats died within 12 hours after the initiation of ischemia and displayed the typical hypoglycemic shock pattern, progressing from stupor to coma, convulsion, and respiratory failure. At the time of death the animals blood glucose levels were below 1.5 mm. Neurological assessment was impossible in most insulin-treated rats at 3 hours postischemia because they were in a comatose or stuporous state. Neurological grades did not differ significantly between groups at 12 hours or at 7 days postischemia (p 0.05, data not shown). The histopathological findings showed no apparent benefit of insulin treatment when a groupwise analysis was performed. The area of striatal infarction in the control group was mm 3, compared with mm 3 in the insulin pre group and 26 8 mm 3 in the insulin post group (Fig. 1). The area of cortical infarction was insignificantly larger in the insulin pre group (55 49 mm 3 ), followed by the control (33 24 mm 3 ) and insulin post (26 42 mm 3 ) groups. The total extent of cerebral tissue loss, including atrophy, was mm 3 in the control group, mm 3 in the insulin pre group, and mm 3 in the insulin post group (p 0.05, Fig. 1). Visual counting of selective necrotic neurons in the ischemic penumbra (Fig. 2) showed that the number of necrotic neurons (mean SD) in the control group was , compared with cells in the insulin pre group and cells in the insulin post group (p 0.05, Fig. 2). Six rats from insulin-treated groups had selective neuronal necrosis in the hippocampus (including the CA1 and CA3 sectors and the dentate gyrus). Four of these animals had ipsilateral cell loss and two had bilateral cell loss. One insulin-treated rat had selective neuronal necrosis in the contralateral neocortex. No control animals had any necrosis outside the territory of infarction either in the hippocampus or anywhere in the contralateral hemisphere. 666

4 Insulin in focal brain ischemia Because our earlier findings that insulin reduces the damage of focal ischemia in the rat 10 were apparently contradicted by the present findings, we analyzed the blood glucose level in relation to the infarction volume, in an attempt to determine a possible link between blood glucose levels and quantitative neuropathology. A graph showing a comparison of the blood glucose level and the volume of cortical necrosis and total infarction demonstrated a U-shaped curve, and nonlinear regression analysis yielded a nadir for cerebral necrosis at a blood glucose level of approximately 7 mm (Fig. 3). Discussion It is widely accepted that hyperglycemia accentuates the brain damage produced by cerebral ischemia in animals and humans. 5,6,17,19,32,33 Stroke-related damage is augmented by hyperglycemia regardless of whether diabetes mellitus is present. 4,23 The mechanism likely involves the enhanced accumulation of lactic acid in the tissue and glucose-sensitive acidosis in the ischemic penumbra. 1 The idea behind initial experiments with insulin was to mitigate this hyperglycemic acidotic mechanism of ischemic damage. 28 Another possibility is that corticosteroids play a detrimental role in hyperglycemic ischemia. 21 In cases of global cerebral ischemia, either profound hypoglycemia or hyperglycemia is detrimental. 28 Warner and colleagues 32 found that insulin-induced normoglycemia (7 9 mm) occurring immediately before the onset of forebrain ischemia reduced brain damage. Strong and colleagues demonstrated improved cognitive function 25 and brain mitochondrial function 26 in rats rendered mildly hypoglycemic with insulin (3 mm) before they experienced global ischemia. At the upper end of normal glucose values, a step function for worsening seizures and histological outcomes was found in rats at a hyperglycemic threshold of 10 to 13 mm. 16 The lower limits of the effects of blood glucose, and insulin in focal ischemia have been less well studied. In the situation of focal ischemia, coadministration of glucose to raise the blood glucose level from mm to mm effectively cancelled most of the neuroprotective effect of insulin, when compared with untreated control animals in which the mean blood glucose level was mm. 10 The results of these experiments thus indicated that lowering blood sugar by administering insulin is desirable in cases of focal ischemia, but the lower limit of effectiveness was not studied to determine whether the curve relating blood glucose levels to brain damage is linear or U-shaped. Similar experiments of MCA suture occlusion in which orthovanadate was used to lower blood glucose level showed that a blood glucose level of 76 mg/100 ml (4.2 mm) is neuroprotective. 11 In our present series these levels averaged 2.5 and 2.2 mm in the groups that received insulin, lower than those attained in previous studies from this laboratory. 10 Indeed, some animals had glucose levels in the range of 1 to 2 mm. The production of seizures, an independent exacerbating factor in determining ischemic outcome, 30 was no doubt detrimental in our insulin-treated rodents. A series of experiments in cats by de Courten-Myers, et al., 7 indicated that very low blood sugar levels may be harmful in the presence of focal ischemia. The reduction of blood glucose levels to just above 3 mm in the transient clip FIG. 3. Graph demonstrating that the nonlinear regression analysis between infarction size and mean blood glucose levels in both insulin-treated groups yields a U-shaped curve with a nadir for infarction volume at approximately 7 mm. occlusion model led to early death and larger infarction. 7 Nine groups comprised this study, however, and some were treated earlier than other groups. 7,8 The present series of concurrently performed experiments confirms the findings of de Courten-Myers, et al., in cats, and indicates that too low a blood glucose level, on the order of 2 mm, is harmful in rats. Some experimental results have shown that glucose can be beneficial in some circumstances, 3,9,24 perhaps by fueling glutamate uptake. 27. Only results in humans can delineate any negative effects of clinical hypoglycemia, and the threshold at which such effects occur in humans when insulin is used to treat focal ischemia. In the past we conducted a pilot study on the use of insulin in humans who undergo cardiac bypass surgery and in whom high glucose levels are seen in combination with the possibility of air or other embolism; we found no detrimental effects, even in patients with glucose levels of 1 to 2 mm who were clinically well while their blood glucose levels were this low (Borger and Auer, unpublished observations). Conclusions In the present study conducted in rats, our results indicate that blood glucose levels should be maintained at approxi- 667

5 C. Z. Zhu and R. N. Auer mately 6 to 7 mm when insulin is used following ischemia. From these data one can infer that hypoglycemia of less than 3 mm should be avoided in cases of focal cerebral ischemia in humans in which insulin is used. Our findings underscore the need to study insulin in human ischemia in surgical as well as medical settings. References 1. Anderson RE, Tan WK, Martin HS, et al: Effects of glucose and PaO2 modulation on cortical intracellular acidosis, NADH redox state, and infarction in the ischemic penumbra. Stroke 30: , Bederson JB, Pitts LH, Tsuji M, et al: Rat middle cerebral artery occlusion: evaluation of the model and development of a neurologic examination. Stroke 17: , Callahan DJ, Engle MJ, Volpe JJ: Hypoxic injury to developing glial cells: protective effect of high glucose. Pediatr Res 27: , Candelise L, Landi G, Orazio EN, et al: Prognostic significance of hyperglycemia in acute stroke. Arch Neurol 42: , de Courten-Myers GM, Kleinholz M, Holm P, et al: Hemorrhagic infarct conversion in experimental stroke. Ann Emerg Med 21: , de Courten-Myers GM, Kleinholz M, Wagner KR, et al: Fatal strokes in hyperglycemic cats. Stroke 20: , de Courten-Myers GM, Kleinholz M, Wagner KR, et al: Normoglycemia (not hypoglycemia) optimizes outcome from middle cerebral artery occlusion. J Cereb Blood Flow Metab 14: , de Courten-Myers GM, Wagner KR, Myers RE: Insulin reduction of cerebral infarction. J Neurosurg 84: , Ginsberg MD, Prado R, Dietrich WD, et al: Hyperglycemia reduces the extent of cerebral infarction in rats. Stroke 18: , Hamilton MG, Tranmer BI, Auer RN: Insulin reduction of cerebral infarction due to transient focal ischemia. J Neurosurg 82: , Hasegawa Y, Hamada J, Morioka M, et al: Neuroprotective effect of postischemic administration of sodium orthovanadate in rats with transient middle cerebral artery occlusion. J Cereb Blood Flow Metab 23: , Izumi Y, Pinard E, Roussel S, et al: Insulin protects brain tissue against focal ischemia in rats. Neurosci Lett 144: , Katz LM, Wang Y, Ebmeyer U, et al: Glucose plus insulin infusion improves cerebral outcome after asphyxial cardiac arrest. Neuroreport 9: , LeMay DR, Gehua L, Zelenock GB, et al: Insulin administration protects neurologic function in cerebral ischemia in rats. Stroke 19: , LeMay DR, Lu AC, Zelenock GB, et al: Insulin administration protects from paraplegia in the rat aortic occlusion model. J Surg Res 44: , Li PA, Shamloo M, Smith ML, et al: The influence of plasma glucose concentrations on ischemic brain damage is a threshold function. Neurosci Lett 177:63 65, Longstreth WT Jr, Inui TS: High blood glucose level on hospital admission and poor neurological recovery after cardiac arrest. Ann Neurol 15:59 63, Mackay KB, Loddick SA, Naeve GS, et al: Neuroprotective effects of insulin-like growth factor-binding protein ligand inhibitors in vitro and in vivo. J Cereb Blood Flow Metab 23: , Nedergaard M: Neuronal injury in the infarct border: a neuropathological study in the rat. Acta Neuropathol 73: , Nedergaard M, Diemer NH: Focal ischemia of the rat brain, with special reference to the influence of plasma glucose concentration. Acta Neuropathol 73: , Payne RS, Tseng MT, Schurr A: The glucose paradox of cerebral ischemia: evidence for corticosterone involvement. Brain Res 971:9 17, Prado R, Ginsberg MD, Dietrich WD, et al: Hyperglycemia increases infarct size in collaterally perfused but not end-arterial vascular territories. J Cereb Blood Flow Metab 8: , Pulsinelli WA, Levy DE, Sigsbee B, et al: Increased damage after ischemic stroke in patients with hyperglycemia with or without established diabetes mellitus. Am J Med 74: , Romano C, Price M, Bai HY, et al: Neuroprotectants in Honghua: glucose attenuates retinal ischemic damage. Invest Ophthalmol Vis Sci 34:72 80, Strong AJ, Fairfield JE, Monteiro E, et al: Insulin protects cognitive function in experimental stroke. J Neurol Neurosurg Psychiatry 53: , Strong AJ, Miller SA, West IC: Protection of respiration of a crude mitochondrial preparation in cerebral ischaemia by control of blood glucose. J Neurol Neurosurg Psychiatry 48: , Swanson RA, Chen J, Graham SH: Glucose can fuel glutamate uptake in ischemic brain. J Cereb Blood Flow Metab 14:1 6, Voll CL, Auer RN: The effect of postischemic blood glucose levels on ischemic brain damage in the rat. Ann Neurol 24: , Voll CL, Auer RN: Insulin attenuates ischemic brain damage independent of its hypoglycemic effect. J Cereb Blood Flow Metab 11: , Voll CL, Auer RN: Postischemic seizures and necrotizing ischemic brain damage: neuroprotective effect of postischemic diazepam and insulin. Neurology 41: , Voll CL, Whishaw IQ, Auer RN: Postischemic insulin reduces spatial learning deficit following transient forebrain ischemia in rat. Stroke 20: , Warner DS, Gionet TX, Todd MM, et al: Insulin-induced normoglycemia improves ischemic outcome in hyperglycemic rats. Stroke 23: , Yip PK, He YY, Hsu CY, et al: Effect of plasma glucose on infarct size in focal cerebral ischemia-reperfusion. Neurology 41: , Zhu CZ, Auer RN: Graded hypotension and MCA occlusion duration: effect in transient focal ischemia. J Cereb Blood Flow Metab 15: , Zhu CZ, Auer RN: Intraventricular administration of insulin and IGF-1 in transient forebrain ischemia. J Cereb Blood Flow Metab 14: , 1994 Manuscript received November 7, Accepted in final form June 9, Address reprint requests to: Roland N. Auer, M.D., Ph.D., Departments of Pathology and Clinical Neuroscience, University of Calgary, 3330 Hospital Drive N.W, Calgary, Alberta T2N 4N1, Canada. rauer@ucalgary.ca. 668