THORACIC SURGERY DIRECTORS ASSOCIATION AWARD

The Thoracic Surgery Directors Association (TSDA) Resident Research Award, sponsored by Medtronic, lnc, was established in 1990 to encourage resident research in cardiothoracic surgery. Abstracts submitted to The Society of Thoracic Surgeons (STS) Program Committee representing research performed by residents were forwarded to the TSDA to be considered for this award. The abstracts were reviewed by the TSDA Executive Committee consisting of: Martin F . McKneally, President; Gordon F . Murray, President-Elect; Mark B. Orringer, SecretarylTreasurer; Stanton P. Nolan, Executive Committeeman; and Sidney Levitsky, Executive Committeeman. The second TSDA Resident Research Award was given to Dr Robert E. Maughan, resident in training at Maimonides Medical Center, Brooklyn, New York, who received a sum of $2,500 and had his expenses paid to the STS meeting. The TSDA, with support by Medtronic, lnc, will make this award annually, using the above selection procedure. The resident author of the selected study will be recognized at the STS meeting.

Intrathecal Perfusion of an Oxygenated Perfluorocarbon Prevents Paraplegia After Aortic Occlusion Robert E. Maughan, MD, Chittur Mohan, MD, Ira M. Nathan, PhD, Enrico Ascer, MD, Peter Damiani, BS, Israel J. Jacobowitz, MD, Joseph N. Cunningham, Jr, MD, and Corrado P. Marini, MD Division of Thoracic and Cardiovascular Surgery, Department of Surgery, Maimonides Medical Center, Brooklyn, New York, and Polyclinic Medical Center, Harrisburg, PennsylGania

A canine model was used to evaluate the effects of continuous intrathecal perfusion of an oxygenated perfluorocarbon emulsion on systemic and cerebral hemodynamics and neurologic outcome after 70 minutes of normothermic aortic occlusion. Twelve mongrel dogs were instrumented to monitor proximal and distal arterial blood pressure, cerebrospinal fluid pressure, spinal cord perfusion pressure, and somatosensory evgked potentials. The intrathecal perfusion apparatus consisted of two perfusing catheters, placed in the intrathecal space through a laminectomy, and a draining catheter percutaneously inserted in the cisterna cerebellomedullaris. The aorta was cross-clamped just distal to the left subclavian artery for 70 minutes. Animals were randomized into two groups: group 1 (n = 6) animals were treated with intrathecal perfusion of saline solution, whereas group 2 (n = 6) animals received oxygenated Fluosol-DA 20%. Data were acquired at baseline, during the cross-clamp period, and after reperfusion. Normothermic Fluosol or saline solution was infused at a rate of 15 mL/min beginning 15 minutes before cross-clamping and continued throughout the ischemic interval. There was no difference in proximal arterial blood pressure (97.2 versus 95.4 mm Hg; p > 0.05) or distal arterial blood pressure (14.6 versus 15.0; p > 0.05) between the two Resented, in part, at the Twenty-eighth Annual Meeting of The Society of Thoracic Surgeons, Orlando, FL, Feb 3-5, 1992. Address reprint requests to Dr Marini, Polyclinic Medical Center, Hamsburg, PA 17110.

0 1992 by The Society of Thoracic Surgeons

groups throughout the cross-clamp interval. Cerebrospinal fluid pressure rose significantly in both groups with the onset of intrathecal perfusion of either saline solution or Fluosol(7 1 versus 24 & 5 and 8 +- 1 versus 40 2 4 mm Hg, respectively; p < 0.05). The rise in cerebrospinal fluid pressure was sustained throughout the perfusion interval in both groups. Negative spinal cord perfusion pressure values were recorded in both groups throughout the cross-clamp interval. All Fluosol-treated animals regained electrophysiologic conduction within 10 minutes of reperfusion and were spared neurologic injury. In contrast, all but 1 animal treated with saline solution suffered spastic paraplegia. Based on the results of this study, we conclude that the intrathecal space can be used as an alternate vascular tree to perfuse the spinal cord with an oxygenated perfluorocarbon emulsion. In this canine model, paraplegia after 70 minutes of normothermic aortic occlusion can be uniformly prevented by the intrathecal perfusion of normothermic Fluosol-DA

*

20%.

(Ann Thorac Surg 1992;54:818-25)

D

espite recent improvements in anesthetic management, operative technique, and spinal cord protection, paraplegia continues to be a devastating complication of procedures requiring cross-clamping of the descending thoracic aorta. Although the etiology of paraplegia is multifactorial, the final common pathway to 0003-4975/92/$5.00

Ann Thorac Surg 1992;54818-25

spinal cord injury is ischemia of the spinal cord from either inadequate collateral blood flow during aortic crossclamping or failure to reimplant critical intercostal vessels [l, 21. The incidence of paraplegia is related to the depth and duration of ischemia; the risk of paraplegia increases from 0.5% for ischemic intervals shorter than 30 minutes to 95% to 100% after 60 minutes of ischemia [3, 41. The incidence of paraplegia after repair of extensive dissecting thoracoabdominal aneurysms can be as high as 40% [5]. Efforts to reduce the incidence of paraplegia have centered on maintenance of distal aortic perfusion with shunts and bypasses [6, 71, rapid identification and reimplantation of critical intercostal vessels [8, 91, increasing spinal cord perfusion pressure with cerebrospinal fluid (CSF) drainage [lo, 111, and mitigation of the effects of ischemia with pharmacological agents such as steroids [121, superoxide dismutase [131, and calcium-channel blockers [14]. Although such efforts have been shown to be of benefit both experimentally and clinically, there is currently no method available to uniformly prevent paraplegia after prolonged aortic occlusion. Because the current modalities have been less than satisfactory in preventing paraplegia after prolonged cross-clamping, we decided to investigate an alternate means of perfusing the spinal cord during aortic occlusion. As the perfusion of the subarachnoid space with oxygenated perfluorocarbon emulsion has been shown to maintain electrical activity during brain ischemia [151, we investigated whether the intrathecal perfusion of Fluosol-DA 20% could prevent paraplegia after 70 minutes of normothermic aortic occlusion.

Material and Methods Experimental Preparation Twelve mongrel dogs weighing 25 to 35 kg were anesthetized with intravenous pentobarbital sodium (20 mg/kg). Animals were intubated and placed on a Harvard ventilator (Harvard Apparatus, Millis, MA) using room air. Respiratory settings included a tidal volume of 12 mL/kg, respiratory rate to maintain the partial pressure of carbon dioxide between 25 and 35 mm Hg, and 5 cm H,O of positive end-expiratory pressure. Pressure monitoring catheters were inserted into the left carotid and femoral arteries and the left external jugular vein to monitor aortic pressures proximal and distal to the cross-clamp and central venous pressure, respectively. All pressures were monitored with Bentley Trantec physiological pressure transducers (Model 60-800; Santa Ana, CA). Lactated Ringer's solution was infused intravenously at a rate of 5 mL/kg/h. Through a left thoracotomy in the fourth intercostal space, the aorta was isolated 1 cm distal to the left subclavian artery. After aortic cross-clamping, partial exsanguination (removal of 40% of circulating blood volume) was used to maintain mean proximal arterial blood pressure between 85 and 95 mm Hg in both groups. With the animal in a prone position, a lumbar laminectomy was performed. The spinous processes of L4 and L5 were removed, and the bone was shaved until the epidural fat was exposed. The dura over the cauda equina

TSDA AWARD MAUGHAN ET AL INTRATHECAL PERFUSION OF FLUOSOL

819

was exposed, and two cruciate incisions were made. Two polyethylene inflow catheters (PE90; Clay Adams, Division of Becton Dickinson & Co, Parsippany, NJ) were inserted into the intrathecal space and advanced for 2 cm along the lateral aspect of the cord, one directed cephalad and the other caudad. Purse string sutures of 6-0 Prolene (Ethicon, Somerville, NJ) were placed to close the dura around the catheters and prevent leakage of CSF. Using this technique, there was loss of less than 10 mL of CSF during the insertion of the catheters and no appreciable loss of Fluosol or saline solution during perfusion. The catheters were secured to the paraspinal muscles, exteriorized through the wound and connected to a Y connector. With the animal again in a right lateral decubitus position, a 20-gauge, 3.8-cm spinal catheter was percutaneously inserted into the cisterna cerebellomedularis to monitor CSF pressure (CSFP) and serve as a drainage catheter. The CSFP transducer was zero referenced to the level of the cisterna cerebellomedularis. Somatosensory evoked potentials (SEPs) were monitored with a Nicolet IV apparatus (Nicolet, Madison, WI) after cortical electrodes and a right posterior tibia1 nerve stimulator were placed.

Preparation and Delivery of Fluorocarbon Emulsion Fluosol DA-20% emulsion (The Green Cross Corporation, Osaka, Japan) consists of three separate parts that must be mixed before use: (1)the Fluosol emulsion; (2) solution 1; and (3) solution 2. The additive solutions serve to adjust pH ionic strength and osmotic pressure in the final 20% emulsion, and must be added separately and sequentially before administration. Solution 1 contains sodium bicarbonate and potassium chloride; solution 2 contains sodium chloride, dextrose, magnesium, and calcium. Before use, the emulsion was thawed and preoxygenated to an oxygen content greater than 5.5 vol% by bubbling a 95% 0, and 5% CO, mixture at 2 L/min for 20 minutes; thereafter, this level of oxygenation was maintained with continuous bubbling of the same gas mixture at 2 L/min throughout the cross-clamp interval. Temperature of the emulsion was maintained between 32" and 37°C by placement in a water bath. The delivery system consisted of Fluosol emulsion suspended 1 m above the animal. Perfusion occurred by gravity at a flow rate of 12 to 15 mL/min through polyethylene tubing connected to the two inflow catheters positioned in the cauda equina. Drainage was assured by the spinal catheter placed in the cisterna cerebellomedularis, and the antigravitary return of the solution to the suspended reservoir was provided by a roller pump. Cerebrospinal fluid pressure was maintained at less than 50 mm Hg throughout the experiment. Oxygen content of the emulsion was measured at the inflow and outflow sites using Lex-0,-Con apparatus (Lexington Instruments Corp, Waltham, MA); oxygen extraction ratio was calculated by dividing the difference between inflow and outflow content by inflow content. To evaluate whether normal saline solution could be used to oxygenate the spinal cord, we oxygenated normal saline solution in the same fashion and measured the

820

TSDA AWARD MAUGHAN ET AL INTRATHECAL PERFUSION OF FLUOSOL

oxygen content of samples obtained from the inflow and outflow catheters. The oxygen content was 0.3 vol% at both sites, indicating inadequate oxygen content and absence of extraction; therefore, we decided to omit oxygenation of the solution in the group receiving normal saline solution. After the acquisition of baseline hemodynamics, animals were randomly assigned to two groups: group 1 animals (n = 6) underwent intrathecal perfusion with normothermic saline solution beginning 15 minutes before aortic cross-clamping and throughout the 70-minute cross-clamp interval. Group 2 animals (n = 6) underwent normothermic intrathecal perfusion with oxygenated Fluosol at the same flow rates throughout the same ischemic interval. Hemodynamic parameters and SEPs were monitored at 5-minute intervals during the 70-minute crossclamp period and for 30 minutes after reperfusion. After removal of the aortic cross-clamp, the thoracotomy was closed over a thoracostomy tube, which was removed upon reversal of anesthesia. The inflow catheters were removed and the dura closed with 6-0 Prolene sutures. A matrix of bone chips and bone wax was placed into the bony defect for added stability, and the lumbar laminectomy was closed. A Penrose drain was placed in the subcutaneous tissue and exteriorized through a separate stab incision to prevent seroma formation. All animals received, preoperatively, 1 g of cefazolin intravenously.

Postoperative Evaluation Animals were evaluated at 24 and 48 hours postoperatively, and neurologic function was graded according to Tarlov’s score by an observer unaware of the experimental protocol. At 48 hours all animals were anesthetized with intravenous pentobarbital sodium (20 mgkg), laminectomy was performed, and the spinal cord was removed and fixed in 10% formalin. Animals were then killed by intravenous injection of a lethal dose of potassium chloride. All animals received 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” (NIH publication No. 85-23, revised 1985).

mmHg 120

-

-

--

100

CONTROL FLUOSOL

‘p < 0.05 v8 Fluosol

_. BL

PM

AXC -on

T20

T30

T40

T50

T60

T70

.---

AXC

-on

AXC -on2

Fig 1. Time course of proximal blood pressure in control (saline) and Fluosol-treated animals. (AXC-off = removal of aortic cross-clamp; AXC-off2 = thirty minutes after repetfusion; AXC-on = application of aortic cross-clamp; BL = baseline; Perf = intrathecal perfusion of saline or Fluosol; T26T70 = 20, 30, 40, 50, 60, and 70 minutes after aortic cross-clamping.)

4 mm Hg to 119 -t 6 mm Hg ( p < 0.05). In contrast, the intrathecal perfusion of Fluosol, before aortic occlusion, did not cause any significant change in systemic blood pressure (85 f 4 mm Hg at baseline versus 83 4 mm Hg during Fluosol infusion; p > 0.05) (Fig 1). There was no significant difference in distal arterial blood pressure between the two groups after aortic cross-clamping (15 2 3 versus 15 2 3 mm Hg; p > 0.05) (Fig 2).

*

Spinal Fluid Dynamics Cerebrospinal fluid pressure increased significantly from a baseline value of 7 -t 1to 24 f 5 mm Hg ( p < 0.05) with the onset of intrathecal perfusion with saline solution. Intrathecal perfusion with Fluosol caused an even greater rise in CSFP (8 1 at baseline versus 40 & 4 mm Hg; p < 0.05). The CSFP during the infusion of Fluosol was significantly higher than that observed during saline infusion (40 2 4 versus 24 5 mm Hg; p < 0.05). The rise in CSFP persisted throughout the cross-clamp interval in both groups (Fig 3). The sustained rise in CSFP, in the presence of low distal arterial blood pressure after aortic cross-clamping caused negative spinal cord perfusion

*

*

Statistical Method All data are presented as means k standard error of the mean, and were analyzed with analysis of variance for repeated measures. When present, differences were localized by the method of Newman and Keuls. Somatosensory evoked potential data and neurologic outcome were analyzed with Fisher’s exact test. Statistical significance was accepted to correspond to a p value of less than 0.05.

mm Hg 140 120

60

p < 0.05 vs Flu0801

40

Results Hemodynamics The intrathecal perfusion of saline solution, before aortic cross-clamping, caused an immediate rise in baseline systemic blood pressure: blood pressure rose from 88 -C

t

2o I

.

BL

Perf

AXC -on

T20

T30

T40

T50

T60

.

T70

-

7

AXC

-on

Fig 2 . Time course of distal blood pressure for control (saline) and Fluosol-treated groups. (Abbreviations are as in Figure 1 .)

AXC

-on

TSDA AWARD MAUGHAN ET AL INTRATHECAL PERFUSION OF FLUOSOL

Ann Thorac Surg 1992;54:81%25

821

rnm Hg

T

1

.

EL

Perfusion

AXC-on

T30

. -

T70

7

AXC-ofl

Fig 3. Time course of cerebrospinal fluid pressure for control (saline) and Fluosol-treated groups. (Abbreviations are as in Figure 1 .)

pressure throughout the cross-clamp interval in both 7 mm Hg; group 2, -22.9 f groups (group 1, -14.5 5 mm Hg) (Fig 4). After reperfusion, systemic hemodynamics and cerebrospinal fluid pressures returned to baseline values.

*

Somatosensory Evoked Potential Return and Neurologic Outcome All Fluosol-treated animals regained electrophysiologic conduction within 10 minutes of reperfusion, whereas only one of the saline-treated animals displayed a similar response ( p < 0.01). The return of SEPs correlated with neurologic outcome. All but 1 animal in group 1 (saline perfusion) suffered spastic paraplegia upon recovery from anesthesia; at reevaluation at 24 and 48 hours postoperatively, they had not improved neurologically. In contrast, all of the Fluosol-treated animals were spared neurologic injury ( p < 0.01).

Oxygen Extraction Mean oxygen content of Fluosol at the inflow catheter was significantly higher than that measured at the outflow

mm Hg

*p < 0.05 vs Fluosol

EL

-20

--

-40

I

Perfusion

Fig 4. Time course of spinal cord perfusion pressure for control (saline) and Fluosol-treated groups. (Abbreviations are as in Figure 1 .)

Fig 5. Comparison of oxygen content of Fluosol samples from the inflow and outflow catheters.

catheter (5.293 -C- 0.144 versus 3.686 & 0.096 ~ 0 1 %p; < 0.01). The oxygen extraction ratio was 29.6% (Fig 5).

Histologic Examination No significant difference in pathology of cord tissue was detected using light microscopy. There was no evidence of arachnoid or meningeal inflammation in either group.

Comment Inherent to repair of descending thoracic and thoracoabdominal aortic aneurysms is temporary occlusion of the descending thoracic aorta with the inevitable occurrence of transient spinal cord ischemia, if adjunctive measures are not implemented to provide distal aortic perfusion. However, paraplegia cannot be uniformly prevented, even when distal aortic perfusion is provided with mechanical devices [16]. Other modalities employed to reduce the unacceptably high risk of paraplegia after procedures on the thoracic and thoracoabdominal aorta include methods to detect intraoperatively spinal cord ischemia with SEPs or motor evoked potentials [17, 181, reimplantation of critical intercostal vessels identified with hydrogen-induced current impulse technique [ 191, and drainage of CSF to increase spinal cord perfusion [ l l , 201. Although all of these techniques have shown some experimental or clinical success, they have not been uniformly successful in eliminating the risk of postoperative paraplegia. The two major causes of postoperative paraplegia after procedures on the thoracic and thoracoabdominal aorta are (1) the degree of ischemia, which is related to the extent of the aneurysmal disease, the degree of collateralization, and the duration of the ischemic interval [21, 221, and (2) failure to reimplant, in the graft, intercostal vessels that supply the spinal cord [23]. Therefore, the approaches used to prevent paraplegia after aortic crossclamping have been directed at balancing the two major risk factors for neurologic injury: provide distal aortic perfusion with mechanical devices and identify and reimplant critical intercostal vessels. The pitfalls of these approaches are the inability to uniformly protect the spinal cord with retroperfusion even when "adequate"

822

TSDA AWARD MAUGHAN ET AL INTRATHECAL PERFUSION OF FLUOSOL

distal aortic blood pressure is achieved and the prolongation of the ischemic interval required to anastomose critical intercostal vessels, as well as the necessity to declamp the proximal aorta to reestablish perfusion through the reattached arteries to provide blood supply to the spinal cord. It is for this reason that we decided to investigate a new approach that involves the use of the intrathecal space to provide oxygen to the spinal cord. This approach, if successful, would allow the surgeon ample time to reimplant all intercostal vessels without time constraints and would not require proximal declamping after reattachment of the intercostal arteries. Because previous studies have demonstrated the ability of oxygen to diffuse from the subarachnoid space across the ependyma and maintain electrical activity in brains subjected to ischemia [15], we decided to use the intrathecal space as an alternate vascular tree to perfuse the spinal cord during a period of prolonged normothermic aortic occlusion. To provide a high diffusion gradient sufficient to allow passive diffusion of oxygen across the ependyma to the external white matter and the central gray matter, oxygenated Fluosol-DA 20%, a perfluorocarbon emulsion, was used. This agent was selected because of its ability to deliver substantial amounts of oxygen without apparent tissue toxicity [24]. The combined surface area of the Fluosol particles is larger than that of blood (1.82 x 10' cm2/L versus 1.86 x lo6 cm2/L, respectively), the time taken for oxygen exchange or release is twice as fast as hemoglobin, and the time of CO, transfer is a few milliseconds [25]. In addition to having excellent oxygen transport properties, Fluosol may prevent vasospasm, reach areas of ischemia through collateral microcirculation, and inhibit the CDll receptors of polymorphonuclear cells, therefore preventing leukocyte aggregation and diapedesis. Additionally, Fluosol has been successfully used in the laboratory setting to minimize reperfusion injury after normothermic ischemia of the myocardium and skeletal muscles [26, 271. In the present study, the intrathecal perfusion of saline solution in control animals caused a significant rise in systemic blood pressure, whereas Fluosol perfusion did not affect systemic blood pressure. This differential response to intrathecal perfusion can be explained by the difference in pH between the two solutions used in this study: the pH of normal saline solution is 5.0, whereas the pH of Fluosol-DA 20% is 7.40. The infusion of normal saline solution in the intrathecal space caused systemic hypertension, probably through a direct stimulation of the vasomotor center. In contrast, Fluosol can be safely infused in the intrathecal space because it has the characteristics of CSF with regard to pH, electrolyte content, and osmotic pressure [24]. After aortic cross-clamping there was no significant difference in systemic pressures between the two groups because the pressures were controlled with partial exsanguination to prevent proximal hypertension. Cerebrospinal fluid pressure rose significantly, in both groups, during the intrathecal perfusion. This response is

Ann Thorac Surg 1992;54:81&?-25

not surprising because CSF was not drained before the onset of intrathecal perfusion; hence, there was an increase in CSF volume which can explain the rise in CSFP. Because in this study, we did not have a group in which CSF was drained before the onset of intrathecal perfusion to prevent the volume expansion of the CSF, we cannot say whether removal of CSF will prevent or minimize the increase in CSFP. The higher pressures seen with Fluosol perfusion may be attributed to the higher viscosity of Fluosol (2.7 centipoise at shear rate of 77 s-l at 37°C) as compared with saline solution. Fluosol is a nonNewtonian fluid like blood, and although it has lower viscosity than blood at lower shear rates, it does have higher viscosity than saline solution. Therefore, we believe that the intrathecal perfusion of Fluosol-DA 20% will cause a rise in CSFP, even if CSF volume is drained before the onset of perfusion. The elevation in CSFP in the presence of decreased distal blood pressure after cross-clamping resulted in negative spinal cord perfusion pressure, defined as the difference between distal arterial blood pressure and CSFP, throughout the cross-clamp interval in both groups. Therefore, as previously shown in our laboratory [ll],in the absence of adjunctive measures, all animals should have suffered spastic paraplegia. Somatosensory evoked potentials were monitored at baseline, during the cross-clamp interval, for 30 minutes after reperfusion, and at 48 hours postoperatively. In all animals, SEPs were lost 5 to 12 minutes after placement of the aortic cross-clamp, and there was no difference between the two groups. We believe that the loss of SEPs during the intrathecal perfusion may be attributed to a small decrease in temperature of the spinal cord caused by the intrathecal perfusion. However, all of the Fluosoltreated animals regained electrophysiologic conduction within 10 minutes of reperfusion, whereas only 1 of the saline-treated animals displayed a similar response. The return of SEPs correlated with neurologic outcome. All of the Fluosol-treated animals were spared neurologic injury; in contrast, all but 1of the saline-treated animals had spastic paraplegia at 48 hours. Obviously, because the temperature of the solutions varied from 32°C to 37"C, one could question whether the protective effect on the spinal cord was secondary to hypothermia or actual oxygenation. Although local hypothermia achieved with intrathecal perfusion of cold saline solution has been shown to be protective for relatively short ischemic intervals [28], we do not believe that the relatively small variations in perfusate temperature played a major role in our experiment. Hypothermia has been shown to reduce oxygen consumption of neural tissue by 5% for each degree of temperature drop between 37" and 22°C [29]; therefore, with a maximum drop of 5"C, the decrease in oxygen consumption would have been at most 25%, not adequate to provide sufficient protection for the ischemic interval chosen in this study. It is well known that spinal cord temperature must be lowered to 18" to 20°C to prevent ischemic injury to the spinal cord [30]. Furthermore, the statistically significant difference in oxygen content between the samples obtained from the inflow and outflow

Ann Thorac Surg 1992;54:81%25

catheters with an extraction ratio of 29.6% indicates, indirectly, that protection to the spinal cord was afforded by the delivery of oxygen to the cord. Of interest, the only animal in the control group that was spared paraplegia regained electrophysiologic conduction within 10 minutes of reperfusion. This animal displayed positive spinal cord perfusion pressure throughout the ischemic interval, with a mean greater than 10 mm Hg, a threshold value for spinal cord integrity in the canine model [ 111. Based on the results of this study, we believe that it may be possible to use the intrathecal space as an alternate vascular tree to perfuse the spinal cord during periods of prolonged aortic occlusion. The intrathecal perfusion of an oxygenated perfluorocarbon emulsion can uniformly prevent paraplegia after extended normothermic aortic cross-clamping in the canine model. From the results of the oxygen extraction, it seems reasonable to conclude that the mechanism of spinal cord protection is by delivery of oxygen to the cord by passive diffusion across the ependyma. Additionally, the intrathecal perfusion of Fluosol may provide a "washout" of toxic metabolites and reduce reperfusion injury. Further investigations should address whether this perfusion can be accomplished through a double-lumen catheter inserted percutaneously in the intrathecal space, and whether the changes in CSF dynamics caused by this methodology can be minimized by changing flow rate and viscosity of the perfusate. Supported by a grant from the Maimonides Research and Development Foundation. Additional support provided by Cardiothoracic Surgical Associates. We gratefully acknowledge Anthony Siconolfi for technical operative assistance and Mary Ann Bottali for manuscript preparation.

References 1. Wadouh F, Lindeman E, Arndt CF, et al. The arteria radicularis magna anterior as a decisive factor influencing spinal cord damage during aortic occlusion. J Thorac Cardiovasc Surg 1984;88:1-10. 2. Williams GM, Perler BA, Burdick JF, et al. Angiographic localization of spinal cord blood supply and its relationship to postoperative paraplegia. J Vasc Surg 1991;13:23-35. 3. Katz NM, Blackstone EH, Kirklin JW, et al. Incremental risk factors for spinal cord injury following operations for acute traumatic transection. J Thorac Cardiovasc Surg 1981;81: 669-74. 4. Jex RK, Schaff HV, Piehler JM, et al. Early and later results following repair of dissections of the descending thoracic aorta. J Vasc Surg 1991;3:226-37. 5. 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;3:389-404. 6. Grossi EA, Kreiger KH, Cunningham JN Jr, et al. Venoarterial bypass: a technique for spinal cord protection. J Thorac Cardiovasc Surg 1985;89:22&34. 7. Svensson LG, Von Ritter CM, Groeneveld HT, et al. Crossclamping of the thoracic aorta. Influence of aortic shunts, laminectomy, papaverine, calcium channel blocker, allopurinol, and superoxide dismutase on spinal cord blood flow and paraplegia in baboons. Ann Surg 1986;204:38-47. 8. Laschinger JC, Cunningham JN Jr, Baumann FG, et al. Monitoring of SEP during surgical procedures on the thora-

TSDA AWARD MAUGHAN ET AL INTRATHECAL PERFUSION OF FLUOSOL

823

coabdominal aorta; 111: Intraoperative identification of vessels critical to spinal cord blood supply. J Thorac Cardiovasc Surg 1987;94:2714. 9. Svensson LG, Patel V, Robinson MF, et al. Influence of preservation or perfusion of intraoperatively identified spinal cord blood supply on motor evoked potentials and paraplegia after aortic surgery. J Vasc Surg 1991;13:35545. 10. McCullough JL, Hollier LH, Nugent M. Paraplegia after thoracic aortic occlusion: influence of cerebrospinal fluid drainage. J Vasc Surg 1988;9:153-60. 11. Grubbs P, Marini CP, Toporoff 8, et al. Somatosensory evoked potentials and spinal cord perfusion pressure are significant predictors of postoperative neurologic dysfunction. Surgery 1988;104:216-23. 12. Laschinger JC, Cunningham JN Jr, Cooper MM, et al. Prevention of ischemic spinal cord injury following aortic crossclamping: use of corticosteroids. Ann Thorac Surg 1984;38: 500-7. 13. Lim KH, Connolly M, Rose D, et al. Prevention of reperfusion injury of the ischemic spinal cord: use of recombinant superoxide dismutase. Ann Thorac Surg 1986;42282-6. 14. Gelbfish JS, Phillips T, Rose DM, et al. Acute spinal cord ischemia: prevention of paraplegia with verapamil. Circulation 1986;74(Suppl 1):5-10. 15. Osterholm JL, Alderman JB, Triolo AJ, et al. Severe cerebral ischemia treatment by ventriculosubarachnoid perfusion with an oxygenated fluorocarbon emulsion. Neurosurgery 1983;13:381-7. 16. Crawford ES, Mizhari EM, Hess KR, et al. The impact of distal aortic perfusion and somatosensory evoked potential monitoring on prevention of paraplegia after aortic aneurysm operation. J Thorac Cardiovasc Surg 1988;95:35747. 17. Cunningham JN Jr, Laschinger JC, Spencer FC, et al. Monitoring of SEP during surgical procedures on the thoracoabdominal aorta: IV. Clinical observations and results. J Thorac Cardiovasc Surg 1987;94:27585. 18. Laschinger JC, Owen J, Rosenbloom M, et al. Direct noninvasive monitoring of spinal cord motor function during thoracic occlusion: use of motor evoked potentials. J Vasc Surg 1988;7161-71. 19. Svensson LG, Patel V, Robinson MF, et al. Influence of preservation or perfusion of intraoperatively identified spinal cord blood supply on spinal motor evoked potentials and paraplegia after aortic surgery. J Vasc Surg 1991;13355-65. 20. Bower GT, Murray MJ, Gloviczki P, et al. Effects of thoracic aortic occlusion and cerebrospinal fluid drainage on regional spinal cord blood flow in dogs: correlation with neurologc outcome. J Vasc Surg 1988;9:13544. 21. 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;3:389404. 22. Crawford ES, Palamara AE, Saleh SA, Roehm JOF. Aortic aneurysms: status of surgical treatment. Surg Clin North Am 1979;59:597-636. 23. Laschinger JC, Izumoto H, Kouchoukos NT. Evolving concepts in prevention of spinal cord injury during operations on the descending thoracic and thoracoabdominal aorta. Ann Thorac Surg 1987;44:667-74. 24. Hansebout RR, van der Jagt RCH, Sohal SS, Little JR. Oxygenated fluorocarbon perfusion as treatment of acute spinal cord compression injury in dogs. J Neurosurg 1981;55: 725-32. 25. Naito R, Yokoyama K. Perfluorochemical blood substitutes Fluosol43, Fluosol-DA-20% and 35%. In: Technical information booklet, Serial No. 5. Osaka, Japan: Green Cross Corporation, 1978. 26. Kolodgie FD, Virmani R, Farb A. Limitation of no reflow injury by blood-f+ee reperfusion with oxygenated perfluorochemical (Fluosol-DA 20%). J Am Coll Cardiol 1991;18: 215-23. 27. Forman MB, Bingham SE, Kopelman HA, et al. Reduction of

824

TSDA AWARD

MAUGHAN ET AL

INTRATHECAL PERFUSION OF FLUOSOL

infarct size with intracoronary perfluorochemical in a canine preparation of reperfusion. Circulation 1985;71:1060-8. 28. Beurger R, Porto J, Fedoronko BS, et al. Selective deep hypothermia of the spinal cord prevents paraplegia after aortic cross-clamping in the dog model. J Vasc Surg 1992;15: 62-72.

Ann Thorac Surg 1992;54:818-25

29. Hagerdal M, Harp J, Nilsson L, et al. The effect of induced hypothermia upon oxygen consumption in the rat brain. J Neurochem 1975;24:311-6. 30. Hansebout RR, Kuchner EF, Romero-Sierra C. Effects of local hypothermia and of steroids upon recovery from experimental spinal cord compression injury. Surg Neurol1975;4:531-5.

DISCUSSION DR AHMAD RAJII-KHORASANI (Neptune, NJ): I would like to thank The Society for this opportunity and congratulate the authors for their award. I see some wrinkles in the literature on thoracic aortic surgery, and in the interest of time I am going to discuss only the issue of transspinal perfusion pressure. There are two parameters in this measurement. One of them is distal aortic pressure, the other one is spinal fluid pressure. Let’s look at them carefully. The confined cranial-spinal space contains brain, meninges, vessels, and CSF. Because it is confined, any change in the volume of this space will cause a change in CSF pressure. Cerebrospinal fluid volume and its dynamics and the volume of brain and meninges are probably not affected by aortic clamping; however, the vascular volume is likely to be affected by clamping of the aorta. We know from own clinical experience that increasing the jugular venous pressure will increase the CSF pressure, cough or increased intrathoracic pressure increases CSF pressure, hypertension increases CSF pressure, and animal experiments have shown that if you increase cardiac output with hypervolemia you will have an increase in CSF pressure. The interesting thing that the same studies have shown is that not only is the CSF pressure increased, but also there is an increase in cortical blood flow. So what we do not know is what CSF pressure means. Is it a protective mechanism for some reason, for brain against, let’s say, increased intravascular pressure, or is it a harmful effect? If it is a harmful effect, what are the ranges, when does it become harmful? The neurosurgical model of severe increased intracranial pressure where we see impedance of blood flow to the brain is perhaps the end of the spectrum and should not be adopted in any other circumstance. So let’s look for a moment at the arterial pressure. In the clinical model, as soon as you cross-clamp the aorta, when you use a pump or a partial bypass or a shunt you have three different areas of pressure, potential pressure for spinal cord in your system, and in this experiment you have two areas. The proximal pressure, which is usually elevated, could be your spinal perfusion pressure. In fact all the cases that the surgeons do with the so-called clamp and go, during their clamp the only circulation to their spinal cord is going through the cephalic source of circulation. So that is the spinal perfusion pressure. On the other hand, if you use a partial bypass or a shunt and you are hoping to take advantage of possible distal position of the spinal cord circulation or the retrograde collaterals, then you add to that the potential of that pressure to be your spinal cord perfusion pressure, or it might be your only spinal cord perfusion. It depends on the anatomy. And in the patients whom you have operated on who have had adequate proximal circulation and adequate distal circulation but became paraplegic, then your excluded segment pressure was your spinal cord perfusion pressure. Now, in relation to the arterial pressure elevation with clamping and the elevated CSF pressure, if the mechanism is vascular engorgement, if suddenly we are exposing the proximal circulation to the same or slightly reduced cardiac output, we will see the cerebral circulation as increased cardiac output, which was

demonstrated to increase your cerebral blood flow. So if the engorgement is the cause of increased pressure, then it is no surprise that when you give patients nitrite you will have your CSF pressure go up, but did you measure the blood flow to show that there is a decrease in blood flow? I do not know if it is really accurate to do this mathematical measurement using only your distal pressure. So the question I have is, what does it mean in physical terms, in mechanical terms, when you have a negative transspinal perfusion pressure? My second question is, in your second group was there a significant drop in your proximal blood pressure? DR MAUGHAN Thank you, Dr Khorasani. Let me answer your last question concerning negative spinal cord perfusion pressures measured as the difference between distal blood pressure and CSF pressure. Negative spinal cord perfusion pressures are uniformly associated with absence of meaningful spinal cord blood flow as measured with radioisotopes and are associated with a paraplegia rate of 100% in animal models. In this experiment, the intrathecal infusion of both saline and Fluosol solutions caused an increase in CSF pressure; this in turn resulted in negative spinal cord perfusion pressures. Therefore, based on previous studies in our laboratory, all animals should have suffered spastic paraplegia. What is remarkable is that although Fluosol-treated animals had lower spinal cord perfusion pressures compared with saline-treated animals, none of them suffered neurologic injury. DR LARS G. SVENSSON (Houston, TX): I would also like to congratulate the authors on a very fine study and potentially a major breakthrough in the prevention of paraplegia. Since our early work in 1985we have believed that the use of the intrathecal space was the way to protect the spinal cord, and in 10 animals no paralysis resulted with intrathecal papaverine. We use intrathecal papaverine in patients, and it reduces the incidence of immediate postoperative paraplegia and paralysis. The problem we have encountered in patients, however, particularly in association with hypotension or respiratory failure, is delayed development of paraplegia. Similarly, I think your technique will be effective in probably two-thirds of the patients in preventing immediate postoperative paraplegia, particularly in the patients with short proximal descending thoracic aortic aneurysms, in whom ischemia to the cord is probably the predominant cause of the paraplegia. In the more extensive thoracoabdominal aortic aneurysm, intercostal artery reimplantation becomes probably more important. I have a few questions. First of all, in the neurology literature it has been said that when one studies brain oxygenation, oxygen will only diffuse from the surface of the brain to about a depth of 1 mm into the cortex. My question is, have you thought about looking at the oxygenation of the spinal cord both on its surface and deeper inside the spinal cord? This could easily be done with a polygraphic technique of measuring oxygen with a fine needle. Excellent probes are available for this. Second, I might have misunderstood you, but did you measure

Ann Thorac Surg 1992;54:8~-25

the temperature of your Fluosol before you perfused the spinal cord? Was it possibly cold? In other words, are you mimicking the study of Ramon Berguer in which he showed that intrathecal cooling of the spinal cord was equally protective? Was the Fluosol inflow and effluent oxygenation significantly different? Because Fluosol has some vasodilatory effects like papaverine but normal saline solution will probably cause spasm, might some of the effect of your high paraplegia rate in the control group been due to spasm in that group, and vasodilatation been the protective effect in the Fluosol group? Nevertheless, 70 minutes of aortic cross-clamping in a dog model will virtually always result in paraplegia. Finally, do you have any thoughts about the use of something lie this in humans? Have you thought about the problem of getting FDA approval or any other way of trying to apply the technique to humans? If this technique proves to be effective in humans, it could be a major breakthrough. Reimplantation of the intercostal vessels, however, continues to be a problem.

DR MAUGHAN Thank you, Dr Svensson. From the neurological literature there are examples of local perfusion of the spinal cord with oxygenated perfluorocarbons, and what they found was that you required a 600 mm Hg gradient to have passive diffusion of the oxygen across the ependyma and into the spinal cord. Although we did not measure specifically whether or not there was oxygenation of the cord, we used the measurement of oxygen content, via the inflow and outflow catheters, as an indirect method of oxygen consumption. Future work, which we have already begun, is to examine more closely the exact mech-

TSDA AWARD MAUGHAN ET AL INTRATHECAL PERFUSION OF FLUOSOL

825

anism by which Fluosol works and the exact mechanism of oxygen uptake. As far as measuring the temperature of the Fluosol, this is a very pertinent question because we are oxygenating the Fluosol during the experiment, and the question arises whether or not the protective effects may be the result of some degree of hypothermia. We measured temperature at the inflow catheter and found that it ranged from an initial temperature of 37°C down to a low of 32°C at 60 minutes of the experiment. Based on previous studies, which show a decrease in metabolic rate of 5% per degree of temperature drop, this would result in a maximum 25% decrease in metabolic rate toward the end of the experiment. This would not be enough to explain the protective effects of the Fluosol. The third question concerned the effects of saline solution being due to vasospasm. That is a very pertinent question. In fact we noticed that the saline infusion did elicit some type of vasomotor response with a resultant hypertension. This could have also resulted in some vasospasm in the spinal canal. But, as you suggested, the 70-minute cross-clamp interval would have resulted in paraplegia whether we had infused saline solution or not, based on previous experiments. As far as clinical application, we are currently working on a more clinical approach to this problem by development of an infusion and drainage apparatus that can be inserted percutaneously into the lumbar area. This may make this approach more clinically applicable. As far as FDA approval of Fluosol, I think we will just have to wait and see how far the animal experimentation can go and if we can document the lack of adverse effects of the Fluosol.