Epidural Versus Subdural Spinal Cord Cooling: Cerebrospinal Fluid Temperature and Pressure Changes

Transcription

1 Epidural Versus Subdural Spinal Cord Cooling: Cerebrospinal Fluid Temperature and Pressure Changes Sven A. Meylaerts, MD, Cor J. Kalkman, MD, PhD, Peter de Haan, MD, Marjolein Porsius, and Michael J. H. M. Jacobs, MD, PhD Departments of Vascular Surgery and Anesthesiology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Background. Regional spinal cord cooling can increase the tolerable duration for spinal cord ischemia resulting from aortic clamping. We compared the efficacy of epidural and subdural cooling and the effect of the resulting cerebrospinal fluid-pressure (CSF) increases on spinal cord motor neuron function. Methods. In 8 pigs, CSF temperature and pressure were assessed in the subdural space at L4, T15, and T7. Saline was infused at 333, 666, and 999 ml/h at four consecutive locations: L4 subdural, L4 epidural, T15 subdural, and T15 epidural. First, the influence of CSF-pressure increases during normothermic infusion on transcranial motor evoked potentials (tc-meps) was assessed. Then, hypothermic infusion (4 C) was performed to assess CSF-temperature changes. Results. During normothermic infusion, baseline CSF pressures increased uniformly from 6 4mmHgto34 18, 42 17, and mm Hg with increasing infusion rates (p < 0.001), and did not differ between epidural or subdural infusion. Tc-MEPs indicated spinal cord ischemia in 6 animals when CSF pressures reached mm Hg. During hypothermic infusion, CSF temperatures decreased from 37 to , , and C, but increasing CSF-temperature gradients were observed between the infusion location and distant segments. Subdural cooling resulted in lower CSF temperatures (p < 0.001), but caused larger CSF-pressure increases (p < 0.001). Conclusions. Subdural and epidural infusion cooling produce localized spinal cord hypothermia in pigs. The concurrent pressure increases, however, are uniformly distributed and can result in tc-mep evidence of ischemia. (Ann Thorac Surg 2000;70:222 8) 2000 by The Society of Thoracic Surgeons Aortic cross-clamping during thoracoabdominal aortic aneurysm (TAAA) repair can cause spinal cord ischemia. A prolonged aortic cross-clamp time is associated with an increased incidence of irreversible spinal cord damage [1]. Hypothermia has consistently been shown to increase the time that spinal cord ischemia can be sustained without irreversible neurologic damage [2 4]. Systemic hypothermia has been used as a protective measure in TAAA repair and the results concerning neurologic outcome are promising [5]. However, systemic hypothermia has several disadvantages such as increased cardiac excitability, sometimes resulting in dysrythmias [6], coagulation defects [7], and an increased risk for postoperative wound infections [8]. Regional hypothermia avoids these systemic complications by confining cooling to the spinal cord region. In several animal studies, this adjunct prevented paraplegia following transient aortic clamping [2 4, 9]. One clinical study suggested a benefit of epidural infusion cooling in 70 patients undergoing TAAA repair [10]. Accepted for publication Dec 30, Address reprint requests to Dr Jacobs, Department of Vascular Surgery, Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ, Amsterdam, The Netherlands; m.jacobs@amc.uva.nl. Nonetheless, it remains unclear whether regional spinal cord cooling produces generalized spinal cord hypothermia or the effect is confined to several spinal cord segments only. The latter would imply that the protective effect is only localized. In contrast, regional infusion cooling can cause considerable CSF-pressure increases, which are thought to be uniformly distributed along the subdural space. These pressure increases can result in spinal cord perfusion pressure decreases [3, 11]. The efficacy of epidural and subdural infusion for the production of regional spinal cord hypothermia has never been compared. In this experimental study, we compared the influence of epidural and subdural cooling on CSF temperatures and CSF pressures along the spinal cord. Furthermore, we assessed with transcranial motor evoked potentials (tc-meps) whether CSF-pressure increases were sufficient to impair spinal cord motor function. Material and Methods Animal care and all procedures were performed in compliance with The National Guidelines for Care of Laboratory Animals in The Netherlands. The study protocol 2000 by The Society of Thoracic Surgeons /00/$20.00 Published by Elsevier Science Inc PII S (00)

2 Ann Thorac Surg MEYLAERTS ET AL 2000;70:222 8 EPIDURAL AND SUBDURAL SPINAL CORD HYPOTHERMIA 223 was approved by the Animal Research Committee of the Academic Hospital at the University of Amsterdam, Amsterdam, The Netherlands. Eight female domestic pigs were studied. Weight of the animals varied between 40 and 50 kg. The anesthetic agents used in this experiment have no major effect on tc-mep responses and are similar to those used in TAAA patients in our clinic [12 15]. Ketamine (15 mg/kg, IM) was used as premedication. Anesthesia was induced with 2.0% isoflurane by mask in a mixture of 50% O 2 in N 2 O. After induction, one intravenous line (18 Gauge) was introduced in an ear vein. The animals received sufentanil 15 g/kg. Isoflurane was discontinued and anesthesia was maintained with infusion of ketamine 15 mg kg 1 h 1, sufentanil 5 g kg 1 h 1 and N 2 O (60%). Animals were intubated and ventilated using intermittent positive pressure ventilation. Ventilation was adjusted to maintain an end-tidal CO 2 (mainstream capnograph [Hewlett-Packard, Boebingen, Germany]) within 4.8 to 5.3 kpa (36 to 40 mm Hg) throughout the experiment. Adequacy of ventilation was confirmed with blood gas analysis at 37 C. Arterial blood pressure was measured with a catheter placed into the axillary artery. Electrocardiogram (ECG), mean arterial pressure, end-tidal CO 2, and nasopharyngeal temperature were monitored continuously. Fluids were substituted by Ringers lactate. As demonstrated in previous reports, fluid infusion into the perispinal space can cause a compensatory MAP increase [4, 11]. We aimed to maintain mean arterial pressure between 70 to 90 mm Hg with fluid administration or sodium-nitroprusside titration as required. Because increases in CSF pressure, resulting from regional liquid infusion, can impair spinal cord blood flow (spinal cord perfusion pressure A MAP CSF pressure), this technique might result in spinal cord ischemia. The detection of spinal cord ischemia during regional infusion was performed with tc-meps, a technique that was described in detail previously [14, 15]. Tc-MEPs were evoked using a multipulse transcranial electrical stimulator (Digitimer D 185 cortical stimulator, Welwyn Garden City, UK). The stimuli were applied to the scalp with four needle electrodes. The anode was placed at the occiput and the cathode consisted of three interconnected cathodes placed behind the ears, in the mastoid bone, and in the soft palate. The stimulus consisted of a train of five pulses with a interstimulus interval of 2.0 ms. Compound muscle action potentials were recorded from the skin over the quadriceps muscles and foreleg muscles. Data acquisition, processing, and analysis were performed on computer with an ADconverter and software written in the LabVIEW programming environment (National Instruments, Austin, TX). The supramaximal stimulus was assessed (typically 400 to 500 V, 1.0 to 1.2 A) and tc-meps were recorded at a stimulus intensity of 10% above the level that produced maximal tc-mep amplitude. A 25% intraanimal variation of tc-mep amplitude was accepted as normal. Baseline tc-mep amplitudes were assessed by averaging the five consecutive responses before the first infusion episode. Fig 1. Schematic view of the experimental setup. Four infusion catheters are introduced into the epidural and subdural space at the L4 and T15 level, resulting in four infusion locations. Three CSF-temperature probes (CSFT) and three CSF-pressure catheters (CSFP) are introduced in the subdural space at the L4, T15, and T7 level. (T thoracic; L lumbar; SD subdural; ED epidural.) Ischemic spinal cord dysfunction was defined as an amplitude decrease below 25% of baseline values. Because of interanimal variation, amplitudes are given as percentages of baseline values. Tc-MEP amplitudes of the left leg were used for data analysis. The animals were placed on their right flank. Three laminectomies were performed at the L4, T15, and T7 levels. At these levels, the ligamentum flavum and dura were minimally incised. Four infusion catheters (4F) were introduced into the epidural and subdural space at both the L4 and T15 levels, and advanced in a cranial direction for approximately 3 cm. Three temperature probes (Subcutaneous Temperature Sensor, Monatherm Inc, St. Louis, MO), connected to a Mon-a-therm, model 6510 (Mallinkrodt Medical, Inc, St. Louis, MO) and three 3F catheters for CSF-pressure measurements were positioned in the subdural space at the L4, Th15, and Th7 levels (Fig 1). The catheters and probes were secured by closing the dura and ligamentum flavum with pursestring sutures, and dorsal muscles were approximated to prevent CSF leakage. The inflow catheters were connected to an infusion pump (IVAC 591, IVAC, San Diego, CA), permitting an infusion rate between 1 and 999 ml/h. From reports in literature, it was established that an infusion rate of 500 to 1200 could result in profound CSF-temperature decreases [2, 4, 9]. We equally divided the maximum pump rate into three successive infusion rates; 333, 666, and 999 ml/h. Normal saline solution was used for infusion. Cooling of the infusate was performed with a heat exchanger (Hyp 10, Gambro, Sweden), placed between the infusion pump and the subdural and epidural inflow catheters. The infusate temperature could be varied between 4 C and 40 C. Because evoked potential responses can be altered by spinal cord temperature decreases [16], the influence of subdural and epidural infusion on CSF pressure and the resulting effect on motor neuron function was first assessed during normothermic infusion. This prevented tc-mep changes by CSF-temperature decreases. Thereafter, hypothermic infusion was performed in the same animals to assess the influence of regional infusion cooling on CSF temperatures throughout the subdural space.

3 224 MEYLAERTS ET AL Ann Thorac Surg EPIDURAL AND SUBDURAL SPINAL CORD HYPOTHERMIA 2000;70:222 8 Normothermic Infusion At each of the four infusion locations (L4 subdural, L4 epidural, T15 subdural, and T15 epidural), infusion was performed at three successive rates; 333, 666, and 999 ml/h. Each rate was continued for 10 minutes. CSF pressures and tc-meps were assessed at the end of each 10-minute infusion episode. Every time infusion at the maximum rate was concluded at each of the four infusion locations, spontaneous recovery of CSF pressure was awaited before infusion was recommenced at one of the other infusion locations. In animals 1, 3, 5, and 7, L4 and T15 infusion was started in the epidural space and followed by subdural infusion. The remaining animals were studied in a reverse infusion sequence. Hypothermic Infusion The infusion speed, location, and sequence were identical as during normothermic infusion. However, the infusate temperature was decreased to 4 C. Every time infusion at the maximum rate was concluded at each of the four infusion locations, spontaneous recovery of the CSF temperature was awaited before infusion was recommenced at one of the other infusion locations. At the end of the experiments, animals were euthanized with pentobarbital infusion. Statistical Analysis All data are expressed as mean standard deviation, except for raw tc-mep amplitudes, which are presented as medians ( range). Repeated measures of analysis of variance were performed with a nested model design to analyze differences between the variables. The SPSS package (SPSS 7.5 for Windows; SPSS Inc, Chicago, IL) was used for statistical analysis. Results Mean arterial pressure was maintained between 70 and 90 mm Hg in 5 animals. In the remaining 3 animals, pressures increased to values between 91 and 111 mm Hg, in spite of high doses of Na-nitroprusside. These MAP increases concurred with CSF-pressure increases to values between 70 and 88 mm Hg. Nasopharyngeal temperatures decreased from to C during the experiment (p 0.004). Reproducible tc-meps could be recorded in all animals. Response amplitude was 321 V (180 to 4900 V) at baseline. Before the onset of infusion, CSF pressures at level L4, T15, and T7 were 5 3, 6 5, and 7 5mmHg, respectively and CSF temperatures were , , and C. Normothermic Infusion The influence of epidural and subdural normothermic infusion on CSF pressures at the three measurement locations is shown in Figure 2. CSF pressures increased significantly with increasing infusion rates in all animals during both epidural and subdural infusion (p 0.001). During epidural infusion at the L4 level, CSF pressures at the infusion level were consistently higher than at the T15 and T7 level, but this difference did not reach statistical significance. During infusion at the three other infusion locations, the resulting CSF-pressure increases were uniformly distributed along the subdural space. Overall, epidural infusion resulted in slightly larger CSFpressure increases than did subdural infusion, but this difference did not reach significance (43 20 versus 41 18, respectively, p 0.5). The CSF-pressure increases resulted in tc-meps decreases below 25% of baseline in 6 animals. This occurred when the CSF pressure increased to an average of mm Hg, which resulted in an average spinal cord perfusion pressure reduction to 16 9 mm Hg. In 1 animal, tc-mep loss occurred during infusion at 333 ml/h and in 3 animals tc-mep loss occurred during infusion at 666 ml/h. In the 2 remaining animals, infusion rates of 999 ml/h were necessary to cause motor neuron dysfunction. Figure 3 demonstrates the effect of CSF-pressure variations on tc-mep amplitudes in one of the animals. One animal demonstrated a tc-mep decrease below 25%, immediately following the introduction of the infusion catheters. Tc-MEP amplitudes remained below 25% for the remaining experiment in this animal. Hypothermic Experiments During hypothermic infusion, average CSF temperatures decreased significantly with increasing infusion rate, irrespective of the infusion location (p 0.001). The average CSF temperatures decreased to , , and C at the successive infusion rates. CSFtemperature decreases, however, were not uniformly distributed along the subdural space, but a temperature gradient was consistently present. As demonstrated in Figure 4, the CSF-temperature decreases were most pronounced at the infusion location, during infusion at both the L4 and the T15 level. For example, infusion at a rate of 999 ml/h resulted in an average CSF temperature of 21 2 C at the infusion locations, but remained as high as 34 1 C at the location where the effect was least pronounced. In order to decrease CSF temperatures below 30 C at two different measurement locations, an infusion rate of 999 ml/h was necessary. The maximum CSF-temperature gradient was analyzed. During infusion at L4, the maximum gradient was observed between L4 and T7. During infusion at T15, the maximum gradient was observed between T15 and L4. The CSF-temperature gradient significantly increased with increasing infusion rates (p 0.001). At 333 ml/h, the maximum gradient was 6 3 and increased to 8 5 C and 13 6 C at rates of 666 and 999 ml/h, respectively. Subdural infusion resulted in an average CSF-temperature gradient of 12 5 C while epidural infusion resulted in a gradient of only 9 5 C (p 0.001). Overall, subdural infusion resulted in lower CSF temperatures than did epidural infusion (p ). During hypothermic infusion, the average CSFpressure increases were larger than during normothermic infusion (47 20 versus mm Hg, p 0.02). Furthermore, no CSF-pressure gradient was observed

4 Ann Thorac Surg MEYLAERTS ET AL 2000;70:222 8 EPIDURAL AND SUBDURAL SPINAL CORD HYPOTHERMIA 225 Fig 2. CSF pressures during normothermic infusion. (A) Infusion at the L4 level; (B) infusion at the T15 level. The histogram shows the CSF pressures, measured at the L4, T15, and T7 levels during infusion at the L4 level (A) and the T15 level (B) at the three subsequent infusion rates. At every infusion speed, CSF pressures at the L4 (solid bars), T15 (open bars), and T7 level (hatched bars) are shown. during hypothermic infusion at any location. CSF pressures were consistently higher during subdural hypothermic infusion as compared to epidural hypothermic infusion (p 0.002). Every time regional infusion at the maximum infusion rate was terminated, CSF temperatures spontaneously returned to 37 C within 12 3 minutes. CSF pressures returned to values below 15 mm Hg within 2 minutes in all animals. Comment The results of this porcine experiment demonstrated that infusion of cold saline into the epidural or subdural space, in order to produce regional spinal cord hypothermia, resulted in localized CSF-temperature decreases only. Subdural cooling was more effective in decreasing local CSF temperature, but it did not extend the cooling over a larger number of spinal cord segments. The CSF-temperature decreases were obtained at the expense of significant CSF-pressure increases, distributed along the entire spinal cord. The resulting reduction in spinal cord blood flow was sufficient to produce electrophysiologic evidence of spinal cord ischemia. The rationale of spinal cord hypothermia as a protective measure for spinal cord ischemia during TAAA surgery is the fact that it increases the tolerable duration of spinal cord ischemia to the extent that reattachment of segmental arteries and reestablishment of spinal cord blood flow can be completed before irreversible damage has occurred. While systemic hypothermia is accompanied by coagulation abnormalities, cardiac dysrythmias, and an increased risk for postoperative infection [6 8], regional spinal cord hypothermia avoids these complications at comparable levels of protection. It was demonstrated in animal studies that moderate regional hypothermia (25 C 27 C) offers sufficient protection during transient aortic clamping [3, 4]. In a clinical series, comparable CSF temperatures were obtained with epidural cooling to protect the spinal cord in 70 patients undergo-

5 226 MEYLAERTS ET AL Ann Thorac Surg EPIDURAL AND SUBDURAL SPINAL CORD HYPOTHERMIA 2000;70:222 8 Fig 3. Tc-MEP amplitude and CSF-pressure measurements in an experimental animal. X-axis: infusion rates, Y-axis: tc-mep amplitude (solid circles) and CSF pressure (plain lines) during infusion at different infusion rates and locations. Epidural infusion at the L4 level rapidly resulted in CSF pressures exceeding 60 mm Hg causing tc-mep amplitude to decrease below 25%. Tc-MEPs recovered when CSF pressures remained below 60 mm Hg during subdural infusion at the L4 level and epidural infusion at the T15 level. When CSF pressure again increased above 60 mm Hg during subdural infusion at the T15 level, tc-meps again decreased below 25%. ing TAAA repair, resulting in a low rate of neurologic deficit [10]. Nevertheless, experimental as well as clinical reports emphasize the risk of unavoidable but potentially hazardous CSF-pressure increases. In rabbits, lethal intracranial hypertension was described [3]. Cambria and coworkers described a patient who developed a cervical spinal cord lesion during epidural infusion cooling, which they ascribed to the high CSF pressure [10]. During the normothermic infusion experiments in our series, CSF pressures also increased significantly. Pressure increases were mostly uniformly distributed throughout the subdural space. Only during epidural normothermic infusion at the L4 level was a pressure gradient observed. This can possibly be explained by the fact that infusion at the L4 level only allows fluid to move in a cranial direction. The contents of the epidural space (fat, veins) possibly act as a resistance that induces a pressure gradient. The CSF-pressure increases reduced the spinal cord perfusion pressure to values that impaired motor neuron function, as evidenced by tc-meps, in all but 2 animals. These data imply that regional infusion could impair spinal cord perfusion in spinal cord segments at the infusion site, but also at distant spinal cord segments. The hypothermic infusion experiments demonstrated that spinal cord hypothermia is limited to several spinal cord segments only, since CSF temperatures decreased principally at the infusion location. In order to extend moderate hypothermia over a larger spinal cord segment, infusion rates had to be increased to 999 ml/h. Nonetheless, CSF temperatures at distant levels remained relatively unaffected. The CSF-temperature gradient, which increased with increasing infusion rates, stresses the limitations of this technique. Spinal cord segments, not at risk for ischemia resulting from aortic cross-clamping or protected by regional hypothermia, now become at risk for spinal cord ischemia because of infusion cooling. Fig 4. CSF temperatures during hypothermic infusion. (A) infusion at the L4 and (B) infusion at the T15 level. The histogram shows the CSF temperatures, measured at the L4, T15, and T7 level during infusion at the L4 level (A) and the T15 level (B) at the three subsequent infusion rates. At every infusion speed, CSF temperatures at the L4 (solid bars), T15 (open bars), and T7 level (hatched bars) are shown.

6 Ann Thorac Surg MEYLAERTS ET AL 2000;70:222 8 EPIDURAL AND SUBDURAL SPINAL CORD HYPOTHERMIA 227 Lower CSF temperatures at distant levels could possibly be accomplished when the infusion episodes would have been prolonged. Berguer and associates reached CSF temperatures of 15 C at the cisterna magna after 37 minutes of subdural infusion at the L6 level in dogs, but an inflow rate of 1500 ml/h was used and outflow was obtained by a durotomy [9]. In this series, the CSFtemperature decrease reached a plateau within the 10- minute infusion period. Furthermore, durotomies are not clinically feasible, and our model may therefore represent a better approximation of the clinical situation. Interestingly, CSF temperatures at the L4 level hardly decreased when infusion cooling was performed at the T15 level in this experiment. One explanation for this observation is that the tips of the infusion catheters were directed cranially, and saline inflow was thereby directed to the T7 measurement location. We observed higher CSF pressures during hypothermic infusion compared to normothermic infusion. This could possibly be explained by the fact that perispinal vasoconstriction, caused by the CSF-temperature decrease, decreased the resorption of infused saline, resulting in higher CSF pressures during hypothermic infusion. In the present study, we were unable to find a distinct advantage of either infusion method for the purpose of regional spinal cord cooling. Subdural cooling seemed to be more efficient in decreasing local CSF temperature, but larger CSF-temperature gradients annulled this effect at distant spinal cord segments. Furthermore, subdural catheter introduction was a possible cause for the persistent motor neuron function decrease in one of the animals caused by local spinal cord compression. However, clinical application of this method would not imply multicatheter introduction, so the risk for mechanical damage would be smaller. Introduction of infusion catheters into the epidural space might be safer and therefore turn the scale towards epidural infusion. The difference in the recovery of CSF pressures and temperatures after the termination of infusion suggests a possible solution for the problem of CSF-pressure increases during regional infusion cooling. If spontaneous CSF-pressure normalization occurs within minutes while CSF temperatures remain within protective values for a longer period, efficient spinal cord cooling might be obtained with intermittent infusion. This requires confirmation in further investigations. The experimental setup of this study has a possible limitation. Actual spinal cord tissue temperatures were not assessed. As Berguer demonstrated during regional cooling in dogs, a difference of several degrees Celsius occurs between the infusate and the spinal cord temperature [9]. However, introduction of a temperature needle probe into the spinal cord would have prevented motor neuron function assessment with tc-meps, because of mechanical damage to the spinal cord. In conclusion, both epidural and subdural infusion cooling produce relatively localized areas of spinal cord hypothermia, dependent on the speed of infusion. The concurrent CSF-pressure increases, however, are uniformly distributed along the subdural space and result in reductions of spinal cord perfusion pressure sufficient to produce spinal cord ischemia. References 1. Crawford ES, Crawford JL, Safi HJ, et al. Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining immediate and long-term results of operations in 605 patients. J Vasc Surg 1986;3: Wisselink W, Becker MO, Nguyen JH, Money SR, Hollier LH. Protecting the ischemic spinal cord during aortic clamping: the influence of selective hypothermia and spinal cord perfusion pressure. J Vasc Surg 1994;19: Gonzalez-Fajardo J, Beatriz A, Perez-Burkhardt JL, et al. Epidural regional hypothermia for prevention of paraplegia after aortic occlusion: experimental evaluation in a rabbit model. J Vasc Surg 1996;23: Marsala M, Vanicky I, Galik J, Radonak J, Kundrat I, Marsala J. Panmyelic epidural cooling protects against ischemic spinal cord damage. J Surg Res 1993;55: Kouchoukos NT, Rokkas CK. Hypothermic cardiopulmonary bypass for spinal cord protection: rationale and clinical results. Ann Thorac Surg 1999;67: Okada M. The cardiac rhythm in accidental hypothermia. Electrocardiology 1984;17: Rohrer MJ, Natale AM. Effect of hypothermia on the coagulation cascade. Crit Care Med 1992;20: Kurz A, Sessler DI, Lenhardt R. Perioperative normothermia to reduce the incidence of surgical-wound infection and shorten hospitalization. Study of Wound Infection and Temperature Group. N Engl J Med 1996;334: Berguer R, Porto J, Fedoronko B, Dragovic L. Selective deep hypothermia of the spinal cord prevents paraplegia after aortic cross-clamping in the dog model. J Vasc Surg 1992;15: Cambria RP, Davison JK, Zannetti S, et al. Clinical experience with epidural cooling for spinal cord protection during thoracic and thoracoabdominal aneurysm repair. J Vasc Surg 1997;25: De Haan P, Kalkman CJ, Meylaerts SA, Lips J, Jacobs MJ. Development of spinal cord ischemia after clamping of non-critical segmental arteries in the pig. Ann Thorac Surg 1999;68: Ubags LH, Kalkman CJ, Been HD, Porsius M, Drummond JC. The use of ketamine or etomidate to supplement sufentanil/n2o anesthesia does not disrupt monitoring of myogenic transcranial motor evoked responses. J Neurosurg Anesthesiol 1997;9: De Haan P, Kalkman CJ, de Mol BA, Ubags LH, Veldman DJ, Jacobs MJ. Efficacy of transcranial motor-evoked myogenic potentials to detect spinal cord ischemia during operations for thoracoabdominal aneurysms. J Thorac Cardiovasc Surg 1997;113: Jacobs MJ, Meylaerts SA, de Haan P, de Mol BA, Kalkman CJ. Strategies to prevent neurologic deficit based on motor evoked potentials in type I and II thoracoabdominal aortic aneurysm repair. J Vasc Surg 1999;29: Meylaerts SA, Jacobs MJ, de Haan P, Kalkman CJ. Comparison of transcranial motor evoked potentials and somatosensory evoked potentials during thoracoabdominal aortic aneurysm repair. Ann Surg 1999;230: Browning JL, Heizer ML, Baskin DS. Variations in corticomotor and somatosensory evoked potentials: effects of temperature, halothane anesthesia, and arterial partial pressure of CO2. Anesth Analg 1992;74:643 8.