Complete prevention of postischemic spinal cord injury by means of regional infusion with hypothermic saline and adenosine

Transcription

1 Complete prevention of postischemic spinal cord injury by means of regional infusion with hypothermic saline and adenosine Spinal cord injury after operations on the descending thoracic and thoracoabdominal aorta remains a persistent clinical problem. Previous attempts to decrease the risk of this devastating complication by lowering the rate of metabolism of the spinal cord have met with varying success. We hypothesized that the tolerance of the spinal cord to an ischemic insult could be improved by means of adenosine. Twenty New Zealand white rabbits underwent 40 minutes of isolated infrarenal aortic occlusion after heparin anticoagulation. Oamps were placed both below the left renal vein and above the aortic bifurcation. In 10 rabbits (group A), a bolus of adenosine (100 mg) was infused into the isolated aortic segment immediately after crossclamping and this bolus was fouowed by a flush of hypothermic saline (8 0 C, 30 mljkg) over the first 10 minutes of ischemia. In one control group of five animals (group B), the descending infrarenal aorta was crossclamped without infusion of adenosine or saline. In another control group of five animals (group C), the aortic segment was flushed with normothermic saline (37 0 C) in a fashion identical to that of the study group. The aortic clamps were removed after 40 minutes, the abdomen was closed, and the animals were auowed to recover from anesthesia. Spinal cord function was assessed 12, 24, 48, 72, and 96 hours after operation by the Tarlov scale. AU animals were put to death at 96 hours after operation and spinal cords were harvested for histologic analysis. The spinal cord function of au group A animals was fudy intact with Tarlov scores of 5; group Band group C animals were au paraplegic with Tarlov scores of 0 (p < 0.001, general linear models analysis of variance). Histologic examination of spinal cords from group A rabbits revealed no evidence of cord injury, whereas spinal cords from groups B and C had evidence of extensive cord injury with central gray necrosis, axonal sweuing, dissolution of Nissl substance, and astrocyte and macrophage infiltration. Regional infusion of the crossclamped infrarenal rabbit aorta with hypothermic saline and adenosine completely prevented paraplegia in our model despite a 40-minute ischemic insult. (J THoRAc CARDIOVASC SURG 1994;107:536-42) Jeffrey A. Herold, MD (by invitation), Roanoke, Va., Irving L. Kron, MD, Scott E. Langenburg, MD (by invitation), Lorne H. Blackbourne, MD (by invitation), and Curtis G. Tribble, MD (by invitation), Charlottesville, Va. Qerations on the aorta that require aortic crossclamping result in ischemia to the distal organs. The most sen- From the Department of Surgery, University of Virginia Health Sciences Center, Charlottesville, Va., and Roanoke Memorial Hospital,a Roanoke, Va. Read at the Seventy-third Annual Meeting of The American Association for Thoracic Surgery, Chicago, IlL, April 25-28, Address for reprints: Curtis G. Tribble, MD, Department of Surgery, Box 181, University of Virginia Health Science Center, Charlottesville, V A Copyright 1994 by Mosby-Year Book, Inc /94 $ /6/51651 sitive of these organs is the spinal cord, and paraplegia resulting from an ischemic event is a devastating consequence. The reported prevalence of neurologic injury for operations on the descending thoracic aorta ranges from 0% to 12%.1 Injury is even more prevalent among patients requiring emergency repairs or among those having dissections, with paraplegia rates ranging from 10% to 40%.1 To date there have been numerous clinical and laboratory studies reporting various adjuncts for preserving spinal cord function. Some of these techniques have been beneficial and are presently used in several clinical settings, such as reimplantation of major intercostal or lumbar 536

2 The Journal of Thoracic and Volume 107, Number 2 Herold et at. 537 Fig. 1. Central gray coagulative necrosis and Nissl substance dissolution of the lumbar spinal cord (original magnification X40). vessels, cerebrospinal fluid drainage, shunts, somatosensory evoked potentials, and adjunctive medications Despite their use, paraplegia remains a persistent complication. We used a rabbit model for spinal cord ischemia because of the unique segmental arterial blood supply to the spinal cord from the infrarenal aorta in this animal. We decided to use lessons learned in organ preservation and apply them to the problem of protecting the spinal cord from ischemic injury.20,21 Recognizing that periods of spinal cord ischemia are often inevitable during operations on the aorta, we hypothesized that infusing agents with protective properties into the isolated aortic segment that has blood supply to the cord would protect the spinal cord. This technique would allow delivery of high concentrations of protective agents to the cord that'would not be feasible if the same agents were given systemically. We chose protective agents such as hypothermic saline solution containing adenosine. Hypothermia is a fundamental part of most approaches to organ protection, and adenosine has been used as a protective agent in numerous other situations. 2o - 22 Previous work in our laboratory has demonstrated that regional infusion of hypothermic saline alone into an isolated infrarenal aortic segment has some benefit in preventing neurologic injury when the spinal cord is subjected to a 30-minute ischemic insult. 23 Other studies in our laboratory have shown similar degrees of neurologic protection with regional infusion of adenosine. However, systemic infusion of adenosine had no significant benefit. 24 We, therefore, decided to combine hypothermic saline solution with adenosine given regionally to see if any synergistically Table I. Neurologic status at 96 hours Group a GroupB Group C Tarlov score (n = 10) (n = 5) (n = 5) beneficial effects in protecting the spinal cord from a 40-minute ischemic insult were apparent. Materials and methods Preoperative care and assessment. Twenty-four New Zealand albino rabbits weighing 3.4 to 4.9 kg (mean 4.01 kg) were anesthetized with intramuscular ketamine (50 mg/kg) and xylazine (5 mg/kg). Four rabbits were eliminated from the study because of technical errors or deaths from anesthesia. No animals received blood pressure or ventilatory support. The animals were then placed in a nose cone so that they were breathing a mixture of halothane 1.5% and oxygen at a rate of 5 L/min. A rectal probe and warming blanket (50 0 C) were used to record and support core temperature. An intravenous catheter (24 gauge) was placed in an ear vein, and preoperatively cefazolin 10 mg/kg was given as a single dose. Maintenance fluid of 0.9% saline solution was infused at a rate of 25 ml/hr through the intravenous catheter during anesthesia and was discontinued immediately after the procedure was completed. An arterial line was placed in an ear artery through a catheter (24 gauge), and this catheter was connected to a blood pressure/heart rate transducer and monitor.

3 538 Herold et al. The Journal of Thoracic and February 1994 Fig. 2. Vacuolization ofaxons in the anterior and lateral columns of the lumbar spinal cord (original magnification X40). 300 r---r , ill I :J Z :::;; II: ill Il... (/) l e:( ill Ol o z e:( (!) I :::;; :::;; r\ r~ / 50 ~ ADENOSINE AND HYPOTHERMIC SALINE INFUSION V HEART RATE \ / _--"'""' '"./ SYSTOLIC BLOOD PRESSURE o ~~~~~~~~~~~~~~~~~~~~~~~~~ o TIME (MINUTES) Fig. 3. The heart rate and systolic blood pressure response in a rabbit after the adenosine bolus and flush of hypothermic saline. After sterile preparation, a 10 cm midline incision was performed. The abdominal aorta was mobilized for a length of 1 cm inferior to the left renal vein down to the aortic bifurcation. Each rabbit was anticoagulated with heparin 150 units/kg before aortic crossclamping. The effects of heparin were not reversed at the end of the procedure. Aortic occlusion was obtained by placement of a Schwartz temporary I-inch vascular clamp (catalog No , Biomedical Research of Development Labs, Gaithersburg, Md.) above the aortic bifurcation followed by clamping of the aorta just below the left renal vein with a similar clamp. All experimental animals were subjected to 40 minutes of crossclamp time. In 10 rabbits (group A) a %-inch 24-gauge catheter was inserted into the aorta just distal to the proximal clamp immediately after aortic occlusion. A 10 ml bolus of saline solution containing 100 mg of adenosine was infused through this catheter into the isolated aortic segment, and this bolus was followed by a flush of hypothermic 0.9% saline solution at 8 C and at a rate of 30 ml/kg over the first 10 minutes of ischemia with an infusion pump. After the flush, the aortic catheter was removed and the arteriotomy in the aor-

4 The Journal of Thoracic and Volume 107, Number 2 Herold et al. 539 ta was closed with a 7-0 polypropylene stitch. In one control group of five animals (group B), crossclamping of the descending infrarenal aorta was performed without infusion of adenosine or saline solution. In another control group of five animals (group C), the aortic segment was cannulated and flushed with normothermic 0.9% saline solution (37 C) in a fashion identical to that used in the study group. The aortic clamps were removed after 40 minutes, and the abdomen was closed in two layers. The arterial line was then removed, and the venous line heparin was locked. Blood pressure, heart rate, and core body temperature were recorded with a Hewlett-Packard monitor (model 78353B, Hewlett-Packard Company, Andover, Mass.) before incision, after bowel mobilization, before crossclamping, after crossclamping, at 10, 20, 30, and 40 minutes after clamping, and after removal of the crossclamps. Spinal cord function was assessed by an independent observer at 12, 24, 48, 72, and 96 hours after operation by the Tarlov scale (0 = no movement, 1 = slight movement, 2 = sits with assistance, 3 = sits alone, 4 = weak hop, 5 = normalhop).25 All animals were putto death 96 hours after the operation with a lethal injection of pentobarbital Spinal cords were harvested for histologic examination immediately after lethal injection, and the cords were fixed in 10% formalin solution for 24 hours before being set in paraffin blocks for sectioning. Sections of the lower thoracic and lumbar cord were stained with hematoxylin and eosin, Luxol fast blue, Biels, and glial fibrillary acid protein stains. Because monitoring of the temperature fluctuations in the spinal cord itself was associated in previous studies with significant attrition of the animals, we did not attempt to measure the spinal cord temperature in the animals in this study. In our previous study, cerebrospinal fluid temperature fell from an average of 37 C to an average of 24.6 C without a significant change in the core temperature of the animals. 24 All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the Institute of Laboratory Animal Resources and published by the N ational Institutes of Health (NIH Publication No , revised 1985). Statistical analysis. All values are expressed as mean ± standard error of the mean. Comparisons among experimental groups were performed by means of general linear models analysis of variance and Fisher's post hoc testing. Frequency of data was compared by X 2 analysis. Significance was considered achieved when p was less than All analyses were performed using the PC version ofthe Number Crunching Statistical System (NCSS version 5.03) run on a Dell System 200 personal computer (Dell Computer Corp., Austin, Tex.). Results All animals in group A (n = 10) had intact neurologic function with Tarlov scores of 5 at 12,24,48,72, and 96 hours. Group B animals (n = 5) were all paraplegic with Tarlov scores of O. Group C (n = 5) animals were also all paraplegic with Tarlov scores of 0 at all times. Neurologic status of all rabbits at 96 hours is depicted in Table I. Histologic examination of the lower thoracic and lumbar cords from animals in group A, with the four stains, revealed a normal cord. Cords from animals in groups B and C on hematoxylin and eosin stains revealed marked central gray coagulative necrosis with macrophage infiltration and dissolution of Nissl substance (Fig. 1). The Luxol fast blue stain confirmed anterior and lateral column vacuolization and no evidence of myelin injury (Fig. 2). Axonal swelling within the anterior and lateral columns was demonstrated with the Biels stairi, and the glial fibrillary acid protein stain displayed migration of astrocytes from the subpial region into the anterior, lateral, and dorsal columns. Of note, both heart rate and systolic blood pressure decreased substantially in group A animals immediately after the adenosine bolus. The bradycardia and hypotensive episode was transient and completely resolved during the first 10 minutes after the infusion of adenosine. This response is diagrammed for a typical rabbit that received the adenosine infusion (Fig. 3). Overall, the beginning and ending heart rate and systolic blood pressure were not statistically different among the three groups (p > 0.05). There was a statistically significant difference in heart rate and systolic blood pressure between group A animals and groups Band C after crossc1amping, which corresponded to adenosine infusion, and at the lo-minute time period (p < 0.05). Also, differences in the core body temperature of all groups were slight, and the differences were not statistically significant (p> 0.05). Discussion We believe that two factors are important in the neuroprotective mechanism in our study. First, hypothermia has been used extensively as an organ protection strategy and has been beneficial especially as a neuroprotectant. 26 Second, adenosine has been used as a cardioprotective agent. 22,27-29 It is still unknown whether the beneficial effect of adenosine is mediated via a substrate salvage or enhancement mechanism or through some other receptor-mediated mechanisms. Adenosine has been shown to reduce adenosine triphosphate degradation during ischemia and to increase its resynthesis during reperfusion Adenosine may prevent microvascular injury by preventing white cell and platelet aggregation and microthrombosis 33 and the release of superoxide anions from the activated neutrophils. 34, 35 Adenosine is known to rise in concentration in the cerebrospinal fluid during ischemia. It has long been believed that adenosine is part of the metabolic regulatory system that matches blood flow to energy needs, both in the brain and in neural tissue. 36 Thus the addition of pharmacologic concentrations of adenosine during ischemia may prevent neutrophil accumulation and free radical release from white cells and,

5 540 Herold et al. The Journal of Thoracic and February 1994 thereby, counter neutrophil-mediated injury of neuronal tissue. Adenosine is also one of the most potent vasodilators known and may overcome the endothelial dysfunction that causes vasospasm after ischemia, even in situations in which blood is not part of the reperfusate. 37 Although there was a significant drop in blood pressure and heart rate after adenosine administration, the drops were transient and well tolerated by the animals. These hemodynamic fluctuations obviously did not affect spinal cord protection and might be further decreased by determining the appropriate dose-response curve in future studies. Selective adenosine agonists may well provide the spinal cord protection seen in this study without systemic hemodynamic effects. 38 We believe that the technique of infusing the isolated aortic segment with hypothermic saline solution containing adenosine for a lo-minute period before repair of aortic disease is clinically feasible and that this technique might reduce the prevalence of paraplegia. Further studies are needed to determine the correct dose necessary for maximal benefit as well as the total amount of ischemic time possible before the effects of such an infusion are no longer protective. In conclusion, we have demonstrated complete prevention of neurologic injury in a rabbit model of spinal cord ischemia. Our technique includes three key features: regional administration of neuroprotective agents, use of adenosine as a neuroprotective agent, and use of hypothermia. We believe this technique is simple and has potential for clinical use. Further studies are warranted. REFERENCES 1. Crawford ES, Crawford JL, Safi. HJ, et al. Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining intermediate and long-term results in 605 patients. J Vasc Surg 1986;3: Crawford ES, Rubino P. Reappraisal of adjuncts to avoid ischemia in the treatment of aneurysms of the descending thoracic aorta. J THORAC CARDIOV ASC SURG 1973;66: McCullough JL, Hollier LH, Nugent M. Paraplegia after thoracic aortic occlusion: influence of cerebrospinal fluid drainage. J Vasc Surg 1988;7: Robertson CS, Foltz R, Grossman RG, Goodman Je. Protection against experimental ischemic spinal cord injury. J N eurosurg 1986;46: DeBakey ME, McCollum CH, Graham JM. Surgical treatment of aneurysms of the descending thoracic aorta: long term results in 500 patients. J Cardiovasc Surg 1978; 19: Crawford ES, Walker HS III, Saleh SA, Normann NA. Graft replacement of aneurysms in descending thoracic aorta: results without bypass or shunting. Surgery 1981; 89: Livesay JJ, Cooley DA, Ventemiglia RA, et al. Surgical experience in descending thoracic aneurysmectomy with and without adjuncts to avoid ischemia. Ann Thorac Surg 1985;39: Cunningham IN, Laschinger JC, Merkin HA, et al. Measurement of spinal cord ischemia during operations upon the thoracic aorta: initial clinical experience. Ann Surg 1982;196: Drenger B, Parker SD, McPherson RW, et al. Spinal cord stimulation evoked potentials during thoracoabdominal aortic aneurysm surgery. Anesthesiology 1992;76: Laschinger JC, Cunningham IN, Baumann FG, et al. Monitoring of somatosensory evoked potentials during surgical procedures on the thoracoabdominal aorta. J THORAC CARDIOV ASC SURG 1987;94: II. Cunningham IN, Laschinger JC, Spencer Fe. Monitoring of somatosensory evoked potentials during surgical procedures on the thoracoabdominal aorta. J THORAC CARDIO VASC SURG 1987;94: Crawford ES, Svensson LG, Hess KA, et al. A prospective randomized study of cerebrospinal fluid drainage to prevent paraplegia after high-risk surgery on the thoracoabdominal aorta. J Vasc Surg 1990;13: Verdant A, Page A, Cossette R, et al. Surgery of the descending thoracic aorta: spinal cord protection with the Gott shunt. Ann Thorac Surg 1988;46: Kouchoukos NT, Wareing TH, Izumoto H, et al. Elective hypothermic cardiopulmonary bypass and circulatory arrest for spinal cord protection during operations on the thoracoabdominal aorta. J THORAC CARDIOV ASC SURG 1990;99: Pakrovsky AV, Yermoljuk RS, Sultanalijev TA, et al. General moderate hypothermia in the surgical treatment of descending thoracic aortic aneurysms. J Cardiovasc Surg 1991;32: Breckwoldt WL, Genco CM, Connelly RJ, et al. Spinal cord protection during aortic occlusion: efficacy of intrathecal tetracaine. Ann Thorac Surg 1991;51: Svensson LG, Stewart RW, Cosgrove DM, et al. Intrathecal papaverine for prevention of paraplegia after operation on the thoracic or thoracoabdominal aorta. J THORAC CARDIOVASC SURG 1988;96: Kirshner DL, Kirshner RL, Heggeness LM, et al. Spinal cord ischemia: an evaluation of pharmacologic agents in minimizing paraplegia after aortic occlusion. J Vasc Surg 1989;9: Archer CW, Wynn MM, Archibald J. Naloxone and spinal fluid drainage as adjuncts in the surgical treatment of thoracoabdominal and thoracic aneurysms. Surgery 1990; 108: Belzer FO, Kalayoglu M. The University of Wisconsin solution as an advance in liver transplant preservation. Trans ImmunoI1989;2:4-13.

6 The Journal of Thoracic and Volume 107, Number 2 Herold et al Southard JH, VanGulik TM, Ametani MS, et al. Important components of the UW solution. Transplantation 1990;49: Ely SW, Mentzer RM, Lasley RD, Lee BK, Berne RM. Functional and metabolic evidence of enhanced myocardial tolerance to ischemia and reperfusion with adenosine. J THORAC CARDIOVASC SURG 1985;90: Locher JP, Siebel PS, Kron IL, Tribble CG. Protection against post-ischemic spinal cord injury using hypothermic regional perfusion. FASEB J 1993;7:A Siebel PS, Kron IL, Tribble CG. Regional adenosine attenuates post-ischemic spinal cord injury. J Vasc Surg 1993;18: Tarlov 1M. Spinal cord compression: mechanisms of paralysis and treatment. Springfield: Charles C. Thomas, 1957: Agee JM, Flanagan T, Blackbourne LH, et al. Reducing post-ischemic paraplegia using conjugated superoxide dismutase. Ann Thorac Surg 1991;51: Pasque MK, Wechsler AS. Metabolic intervention to affect myocardial recovery following ischemia. Ann Surg 1984; 200: Reibel DK, Rovetto MJ. Myocardial ATP synthesis and mechanical function following oxygen deficiency. Am J PhysioI1978;234:H Reibel DK, Rovetto MJ. Myocardial adenosine salvage rates and restoration of A TP content following ischemia. Am J PhysioI1979;237:H Silverman NA, Kohler J, Feinberg H, Levitsky S. Beneficial metabolic effect of nucleoside augmentation on reperfusion injury following cardioplegic arrest. Chest 1983; 83: Wyatt DA, Ely SW, Lasley RD, et al. Purine-enriched asanguineous cardioplegia retards adenosine triphosphate degradation during ischemia and improves post-ischemic ventricular function. J THORAC CARDIOV ASC SURG 1989; 97: Wyatt DA, Ely SW, Lasley RD, et al. Purine enriched asanguineous cardioplegia provides superior myocardial protection. Surg Forum 1987;38: Kitakaze M, Hori M, Sato H, et al. Endogenous adenosine inhibits formation of microthromboembolism in ischemic myocardium [Abstract]. Circulation 1990;82(Suppi):III Cronstein BN, Daguma L, Nichols D, etal. The adenosine/ neutrophil paradox resolved: human neutrophils possess both alpha 1 and alpha 2 receptors that promote chemotaxis and inhibit oxygen generation, respectively. J Clin Invest 1990;85: Schrier DJ, Imre KM. The effects of adenosine agonists on human neutrophil function.j ImmunoI1986;137: Rudolphi KA, Schubert P, Parkinson FE, Fredholm BB. Adenosine and brain ischemia. Cerebrovasc Brain Metab Rev 1992;4: Gao WS, Hijioka T, Lindert KA, Caldwell-Kenkel JC, Lemasters JJ, Thurman RG. Evidence that adenosine is a key component in Carolina rinse responsible for reducing graft failure after orthotopic liver transplantation in the rat. Transplantation 1991 ;52: Laing BT. Adenosine receptors and cardiovascular function. Trends Cardiovasc Med 1992;2: Discussion Dr. Lars G. Svensson (Burlington, Mass.). You have not addressed the problem of the intercostal vessels, but I think this is a very important study. We also have tried to evaluate a type of cold spinoplegia solution-lidocaine and Ringer's lactatebut we have not been quite as successful as you have. Your technique seems to be a major breakthrough. I have two questions. First, you used a very large dosage of adenosine, one that would appear to be toxic in the human being. Have you any thoughts about how to overcome this problem in human beings? My second question concerns your previous study reported in The Journal a/vascular Surgery, in which adenosine was used in a similar animal model. Can you explain the differences in histologic findings in this model this time and those of your previous study? Dr. Edward D. Verrier (Seattle, Wash.). We have had quite a bit of experience with this particular rabbit model and it is a superb model. If the infrarenal aorta is occluded for 13 minutes, all evoked potentials are lost. With 20 minutes of occlusion, all animals become paraplegic. We have used monoclonal antibodies to block white cells and oxygen radical scavengers and have made no impact at all on the ability to prevent paraplegia. I have two questinos. First, can you discuss the historical perspective from your other studies related to hypothermia versus adenosine to put into context the improvement of combining the two? Obviously, other investigators have used isolated hypothermic perfusions and have prevented paraplegia in other models and not this one. Second, I think it would be important to allude to the mechanisms. Adenosine is a complex molecule that has endothelial as well as, at least in the heart, specific myocyte functions mediated by different receptors. Do you have any insight into the potential mechanism of protection of adenosine in neurologic tissue? Dr. Frank C. Spencer (New York, N.Y.). I have a comment and two questions. My two questions arise because the data are almost too good to believe but speak for themselves: You mentioned briefly that neither hypothermia alone nor adenosine alone was protective. Why is the combination so protective? Second, how much hypothermia did you induce by this infusion? Do you have any measurement of spinal cord temperature or any regional measurements? The information is very exciting and very clear and it had to be duplicated. I hope it is correct. Dr. Herold. Dr. Svensson, we used the maximum amount of adenosine that is tolerable in human beings. This dose was a starting point, and we need to work out dose-response curves to calculate the appropriate dose that will limit side effects but maintain neurologic protection. In our previous studies, we gave systemic adenosine at the same concentration and there was an unacceptably high mortality rate. When we administered that concentration of adenosine as a regional infusion, our mortality

7 542 Herold et al. The Journal of Thoracic and February 1994 was zero and the rabbits were able to tolerate the adenosine infusion with only a transient drop in systolic pressure, which we think is related to systemic washout of the adenosine through the aorta. Regarding Dr. Verrier's question concerning hypothermia versus adenosine alone: Our previous studies have shown that hypothermia alone was about 60% protective, which was similar to adenosine alone in preventing neurologic injury. By combining these two modalities, we found that we could completely prevent neurologic injury in this model. There are a number of theories concerning the mechanisms of adenosine protection. Mainly, adenosine increases vasodilation, which could increase collateral blood supply to the cord. The adenosine also decreases white blood cell adherence and degranulation, and it prevents platelet aggregation. Thus a number of effects may be contributing to neurologic protection, which we did not evaluate specifically in this study. I believe future studies in which adenosine agonists to type I and type II receptors are used will help us determine what receptor is actually causing the beneficial effect that we are seeing. To answer his question concerning the differences in the histologic features of the spinal cord in the two different studies: We used four different stains to look at the spinal cord in this study; in the previous study we evaluated electron micrographs of the spinal cord. We were able to determine that there was a difference in the histologic injury, mainly in the anterior and lateral columns but also in the central gray area. Now we are able to find a difference, whereas we did not see these changes with the electron microscope. The temperature of the spinal cord during infusion of hypothermic saline was evaluated previously in our laboratory. The cord temperature was 21 0 C with the same rabbit model and techniques.