- Email: [email protected]
Pharmacologic interventions for prevention of spinal cord injury caused by aortic crossclamping
Cardiopulmonary Bypass, Myocardial Management, and Support Techniques
Pharmacologic interventions for prevention of spinal cord injury caused by aortic crossclamping The efficacy of pharmacologic agents for prevention and control of oxygen-derived free radical damage in ischemia-reperfusion injury of the spinal cord was assessed in a swine model of thoracic and tboracoabdominal aortic crossclamping. Animals were exposed to 30 minutes of ischemia that induced lethal, irreversible injury and paraplegia. The experimental groups were as foUows: group A (n = 7), control group, receiving no pharmacologic intenention; group B (n = 7), deferoxamine 50 mg/kgjday administered intravenously over 3 to 4 hours before ischemia; group C (n = 7), aUopurinol pretreatment 50 mgjkg/day for 3 days; and group D (n = 7), superoxide dismutase 60,000 units administered with 50,000 units before removal of the aortic crossclamp and 10,000 units over 10 minutes of reperfusion. Proximal hypertension was controUed with sodium nitroprusside and volume depletion, The methods of assessment were neurologic by a modified Tarlov criteria and blood flow by radiolabeled microspheres. Results of blood flow assessment confirmed a true ischemic episode of 30 minutes for aU animals in aU groups. The blood flow fell significantly during ischemia (p < 0.01) and a hyperemic response was evident in the early reperfusion period. AU animals in control group A were paraplegic. The group B (deferoxamine) results were superior; 85 % had grade In function on a modified Tarlov scale, with animals in the group standing and even walking with difficulty. Only one animal in this group had good movements of hind limbs but was unable to stand or walk. Neurologic recovery was limited in the allopurinol group (group C), with 85 % showing s6gbt neurologic recovery with limited movement of the hind limbs. The animals in the superoxide dismutase group (group D) aU had good recovery, with strong motor response of hind limbs, but were not able to stand. In summary, the results of this experimental protocol confirmed the possible role of oxygen-derived Cree radicals in the pathophysiology of spinal cord injury, induced by aortic crossclamping. Moreover, it proved that ischemiareperfusion injury could be altered by pharmacologic interventions. (J THORAC CARDIOVASC SURG 1992;104:256-61)
A. Karim Qayumi, MD, PhD, Michael T. Janusz, MD, W. R. Eric Jamieson, MD, and Donald M. Lyster, PhD, Vancouver, B.C., Canada
From the Division of Cardiovascular and Thoracic Surgery, Depart-
mentofSurgery,Facultyof Medicine, Faculty of Pharmacological Sciences, University of British Columbia, Vancouver, B.C.,Canada. Received for publication Jan. 3, 1991. Accepted for publication Jan. 20, 1992.
256
Supported by the Rick Hansen Man In Motion Legacy Fund. Addressfor reprints: A. K. Qayumi, MD, PhD, Assistant Professor of Surgery, Department of Surgery, Room 3100-910 West 10th Ave., Vancouver, Be, V5Z 4E3, Canada.
12/1/36840
Volume 104 Number 2 August 1992
Spinal cord injury
Spinal cord injury remains the most dreaded complication of operations on the aorta for treatment of descending thoracic and thoracoabdominal aneurysms. I, 2 The necessity for aortic crossclamping carries a risk of distal organ ischemia as a result of diminished perfusion. The organs at risk of ischemia include kidneys, liver, spinal cord, and intestines. The spinal cord is the organ most sensitive to ischemia. The resulting ischemic injury can
60
produce paraparesis or paraplegia. The risk of paraplegia
20
increases with the length of aorta replaced and with the length of time during which the aorta is crossclamped; the risk of paraplegia is about 3% for resections of the descending thoracic aorta, about 15% for operations involving the rnajority of the descending thoracic aorta plus the upper abdominal aorta, and about 25% for the entire descendingthoracic and abdominal aorta.v 4 Paraplegiais uncommon in operations requiring a crossclamp time less than 25 minutes, and it increases progressively with extended time intervals.l Attempts have been made to reduce spinal cord injury by improving operative techniqueand anesthesia management and by using shunts or bypass techniques to maintain distal perfusion.v'' The complex methods that have been used clinically and experimentally to eliminate or reduce the frequency of this complication have been ineffective. The mechanisms of tissue injury and possibleinterventional therapies have received limited attention in regard to the spinal cord. There is now acceptance that reactive oxidant speciesplaya primary role in the development of ischemia-reperfusion injury of various organs.v!" The oxygen-derived free radicals, namely, superoxide anion, hydrogen peroxide, and hydroxyl ion,are important in the mechanisms of the ischemia-reperfusion injury.!" The experimental protocol using a swine model of spinal cord ischemia,examined in this study, was designed to evaluate pharmacologic interventions in respect to control of oxygen-derived free radical injury to extend the limits of safety of surgically induced spinal cord ischemia. Materials and methods The efficacy of pharmacologic agents for prevention and control of oxygen-derived free radical damage in ischemiareperfusion injury of the spinal cord was assessed in a swine model of thoracic and thoracoabdominal aortic crossclamping. Experiments were performed on 28 swine weighing from 20 to 22 kg. The animals were divided into four groups of seven each. In group A (n = 7), the control group, no pharmacologic intervention was introduced. In group B (n = 7) deferoxamine in a dose of 50 mg/kg was administered intravenously over a 3~ to 4~hour period before ischemia. The group C (n = 7) animals were pretreated with allopurinol, 50 rug/kg/day, for 3 days, and
257
Flow 80
o'-------------.:.:..tPL--------L..--T1
T2
T3
Time
Fig. I. Spinal cord blood flow (mean ± standard deviation) assessed by nuclear imaging technique using specific microspheres at each time period. A Control group A; • group B;. group C; 0 group D; TI, before ischemia; T2, during ischemia; T3, 30 minutes after ischemia. *p < 0.01. in group D (n = 7) superoxide dismutase (SOD) 60,000 units was administered, 50,000 units before removal of the aortic crossclamp and 10,000 units over a lO-minute period after reperfusion. All animals were subjected to 30 minutes of ischemia and observed for 4 hours after recovery from anesthesia for neurologic status. Anesthesia was induced with ketamine (20 mg/kg body weight) and maintained with isoflurane (0.50/0 to 2.00/0). Ventilation was maintained with a volume ventilator and inflatable cuffed endotracheal tube. Hemodynamics were controlled by monitoring the systemic blood pressure (systolic, diastolic, mean) and heart rate. Electrocardiographic monitoring was also conducted throughout the procedure. Proximal hypertension was controlled with 20 to 40 mg sodium nitroprusside and volume depletion (300 to 500 ml). Nitroprusside was administered at a dose of 100 }lgjml and a flow rate of I ml/kg/min during the first 15minutes of ischemia. There were indications for volume depletion when the systemic pressure and heart rate were more than double the original value. The cardiovascular system in swine is sensitive. Therefore the use of both methods (volume depletion and .vasodilator infusion) is important to control the proximal hypertension. Under sterile conditions a limited thoracotomy was performed in the fourth left intercostal space to expose the proximal and midportions of the descending thoracic aorta. The aorta distal to the left subclavian artery was exposed for clamping. Proximal and distal arterial monitoring lines for pressure were inscribed via the left internal mammary artery and left femoral artery. The pericardium was opened for exposure of the left atrial appendage and placement of a polyethylene catheter for microsphere infusion. The external jugular vein was cannulated for intravenous access for the administration of nitroprusside to control proximal hypertension during aortic clamping and volume replacement during the unclamping phase. After aortic unclamping the thoracotomy was closed and the animal was permitted to regain consciousness for neurologic assessment
The Journal of Thoracic and Cardiovascular Surgery
2 5 8 Qayumi et al.
Table I. Effect of various pharmacologic interventions assessed by blood flow and Tarlov criteria Animal
p Value
0
3.0 ± 1.5 52.8 ± 6.2
0.01
1000/0
27.8 ± 4.6
3.4 ± 0.8 49.5 ± 5.8
0.01
7
25.0 ± 3.9
2.6 ± 0.8 45.6 ± 6.8
0.01
7
41.0 ± 10.5 6.1 ± 2.0 66.8 ± 9.4
0.01
No.
I
7
27.6 ± 4.2
7
C: Allopurinol (pretreatment for 3 days, 50 mgjkgjday) D: Superoxide dismutase
A: Control (no pharmacologic intervention) B: Deferoxamine (50 mg/kg)
Tarlov criteria at 4-hour reperfusion
Bloodflow duringischemia II
III
(60,000U)
I
II
III
14% 860/0
860/0
14%
100%
IV
P Value
0.05 compared with groups A and C
NS
0.05 compared with group A
NS, Not significant.
over 2 to 3 hours. The animals received appropriate sedation during the recovery period. Acid-base status was monitored throughout the procedure and metabolic acidosis was corrected with sodium bicarbonate, if necessary. After recovery from anesthesia and neurologic assessment, the animals were again anesthetized for laminectomy. The
laminectomy was performed at the levelof the tenth thoracic
vertebra to obtain spinal cord- specimens for quantitation of microspheres. The animals were killed at completion of the experiments. The experimental protocol was designed and performed in accordance with the principles of the Canadian Council on Animal Care and the University of British Columbia Animal Care Committee regulations.
The end points of evaluation were assessment of neurologic
status and spinal cord blood flow. Neurologic assessment included physical findings of hind limb neurologic function 4 hours after reperfusion. The neurologic status was determined when the animals were fully conscious. The assessment comprisedclinical criteria of hind limb neurologic function, according to the modified Tarlov criteria, as follows: Grade 0: No voluntary function (complete paralysis) Grade I: Movement of joints perceptible Grade II: Active movement of joints (inability to stand) Grade III: Able to stand (unable to walk) Grade IV: Complete recovery (able to stand, walk, run) Blood flow was assessed by the radiolabeled microsphere techniques. The blood flow assessment was used to determine the residual spinal cord blood flow, includingcollateral flow, for each of the study groups. Blood flow wasassessed at three time periods: time one (TI, before ischemia),gadolinium 153; time two (T2, during ischemia, 5 minutes before reperfusion), tin 113; and time three (T3, after ischemia, within 30 minutes after reperfusion, at the time of reestablishment of hemodynamic parameters), ruthenium 103. The radiopharmaceuticals were injected into the left atrium through the left atrial cannula and blood samples were collected simultaneously from the femoral artery. Blood flow was quantified by a method described by Heymann and colleagues. 1 I Blood flow assessments were analyzed for the effectof time and for comparison among the groups by means of repeated measures analysis of variance. The difference in Tarlov criteria among the groups was determined by nonparametric statistical analysis of Mann-Whitney pairwise comparison and Kruskal-
Wallis x2 test. Asterisks on Fig. 1 represent the significant differences among the groups. The significance was assessed at 950/0 and 98% confidence levels as indicated in the figures.
Results Assessment of spinal cord blood flow is shown in Table I and Fig. 1. There is a significant (p < 0.01) fall in blood flow in all groups of animals during ischemia (T2). At the time of reperfusion (T3) a hyperemic response is evident for all experimental groups. Equal spinal cord blood flow for all groups during ischemia (T2) indicates a true ischemicepisode and also indicates that there was essentially no collateral circulation for the spinal cord throughout the ischemic interval in all groups. The neurologic assessment (Table I, Fig. 2) demonstrates complete paraplegia with a modified Tarlov criteria score of zero for all seven animals in the control group (group A, 1000/0 paraplegia). In group B, in which deferoxamine was used, the neurologic status was superior to that of all other groups. Six of seven animals (860/0) were in grade III by modified Tarlov criteria, which indicates that the majority of animals in this group were standing and even walking with difficulty. Only one animal ( 14%) in this group was in grade II, with good movements of the hind limbs, but the animal was unable to stand and walk. The allopurinol-pretreated animals (group C) showed limited neurologic recovery; six animals (860/0) had a modified Tarlov score of I, slight neurologic recovery with limited movement of the hind limbs. Only one animal ( 140/0) had a better neurologic recovery, which was grade II by the modified Tarlov scale. In group D, where SOD was used, all seven animals (100%) had a strong motor response of the hind limbs and good neurologic recovery. All the animals in this group were in grade III of the modified Tarlov criteria. Consequently the animals in group B, where deferoxamine was used, had significantly (p < 0.05) better neurologic recovery than the control
Volume 104
Spinal cord injury
Number 2 August 1992
(group A) and allopurinol-pretreated (group C) animals. Animals in group D, where SOD was used, had significantly (p < 0.05) better recovery than the control group (Table I).
Control
259
Deferoxamine
Discussion The role of oxygen-derived free radicals in the pathogenesis of ischemia-reperfusion injury has been extensively investigated in the past decade.v!" Although a great deal of information is available about the adverse effectof oxygenspeciesin heart, lung, liver,kidney,intestine, and other tissues," little attention has been paid to spinal cord ischemia. Lipid peroxidation of the cellular membranes caused by oxygen species is a concept that generally applies to all tissues; therefore results of this experimental endeavor are mostly discussedin an overall lipid peroxidation concept. Allopurinol, a xanthine oxidase inhibitor, was used to prevent the formation of oxygen-derived free radicals. The protectiveeffect of allopurinolas a xanthine oxidase inhibitor is controversial because of the absence of detectable levels of xanthine oxidase in rabbit, pig, and human rnyocardium.P: 13 This brings into question the mechanism of the protectiveeffect of allopurinolin ischemia-reperfusion injury. Results of our experiments in a swinemodel of heart-lung transplantation, as wellas the experience of other investigators, 14-16 clearlydemonstrate the effectiveness of allopurinol in ischemia-reperfusion injury in species with no apparent xanthine oxidaseactivity whenthe drug is administered repeatedly. In search of other mechanisms of action, Das and colleagues'? proposed a direct scavenging property for allopurinol. A number of other possible mechanisms of allopurinol protection have been proposed, including an increased efficiency of adenosinetriphosphate salvage,17 facilitation of mitochondrial electron transfer, 18 and direct activationof endogenously formed reative species, such as hydroxyl radicalsor myeloperoxidase-derived hypochlorous acid. 19 Presenceof xanthine oxidasehas not been investigated in the swine spinal cord tissue. Therefore results of this experimental endeavor cannot confirm the fact that allopurinol acted as a xanthine oxidase inhibitor or expressed a mechanism of action other than xanthine oxidase inhibition. Although allopurinol does not afford the mostprotectionin this model, its availabilitymakes it clinically relevant. Another substance used in this experimental protocol was SOD. Despite the strong theoretical basis, there are controversies about its protective effect. SOD is thought to be exclusively an intracellular enzyme. Therefore administration of this substance extracellularly may be
100%
Allopurinol 1 2 3 4 5
SOD
• No movement @ Slight movement 0 Good movement without resistance @ Good movement against resistance • Normal
Fig. 2. Neurologic assessment by modified Tarlov criteria assessed 4 hours after reperfusion. Deferoxamine group B had a significantly higher score (p < 0.05) than the control group, with 100% paraplegia. ineffective. Shlafer and coworkers.l" using globally ischemic and reperfused isolated rabbit and cat hearts, showed reduction in cardiac injury with SOD and catalase. Myers and associates" demonstrated improved recoveryof function obtained with SOD and catalase in a canine model of regional ischemia. These investigators occluded the left anterior descendingcoronary artery for 15minutes and then reperfusedthe area for two hours. In contrast, Myers and colleagues" did not find any evidenceof added protection with SOD. These investigators used catalase and SOD in a separate series of experiments, and only the group of animals in which catalase was used in combination with SOD demonstrated a significant improvement of functional parameters . Myers's group concludedthat the protectiveeffectachievedby the former investigatorswasprobablybecauseof catalase and not SOD. There is a strong theoretical support for the combined use of SOD and catalase.20• 21 SOD is able to dismutate superoxideanions to hydrogen peroxide,which
260
The Journal of Thoracic and Cardiovascular Surgery
Qayumi et al.
is a harmful oxygen species in itself. Moreover, in the presence of iron, hydrogen peroxide can be converted into a more potent oxygen-derived free radical, which is the hydroxyl radical. This reaction needs 'catalase or glutathione peroxidase to convert the hydrogen peroxide into water and carbon dioxide. However, results of this experimental endeavor clearly demonstrated that SOD provides some degree of protection by itself without catalase or glutathione peroxidase. Animals in the SOD group had significantly better function in their hind limbs after 30 minutes of ischemia and 2 hours of reperfusion than did the control animals. The protective effect of SOD in this experimental model confirms the importance of the time and method of administration, as it is described by oth-
ers.!'' Svensson and colleaguesf investigated the effect of allopurinol, SOD, calcium channel blocker, and papaverine on spinal cord blood flow and paraplegia caused by 60 minutes' crossclamping of the thoracic aorta. These investigators emphasized the effect of papaverine as a vasodilator and did not find a protective effect using allopurinol and SOD. The time-response relationship of recombinant SOD in the ischemic spinal cord has also been investigated by Lim and associates.P The protective effect of SOD was significant within 30 minutes of spinal cord ischemia. SOD appeared to be ineffective in longer than 3D-minute ischemic intervals. Results of these latter investigators agree with our findings and also explain the failure of Svensson and colleagues to obtain a protective effect with SOD after 60 minutes of spinal cord ischemia. Chelation of iron with agents such as deferoxamine is thought to protect ischemic organs from reperfusion injury. Although Myers and associates-" did not find much benefit from deferoxamine administration, the majority of investigators found deferoxamine highly protective in the event of ischemia-reperfusion injury. 25, 26 Menasche and coworkers/" demonstrated that deferoxamine-containing cardioplegic solutions alone afford the best myocardial protection. Bonser, Fragomen, and Edwards/" also demonstrated in a model of heart-lung transplantation that preliminary reperfusion with deferoxamine improves the quality of ischemic lung protection as manifested by improved gas exchange and pulmonary hemodynamics after 4 hours of hypothermic preservation. OUf experience in a model of heart-lung transplantation (A.K.Q., unpublished data) demonstrated a protective effect using high-molecular-weight deferoxamine. Protective effect of deferoxamine is mainly due to chelation of ferric ions that catalyze the Fantom-Haber-Weiss reaction; consequently, in prevention of the most potent oxygen-derived free radical, the hydroxyl radicals, the deleterious effect of this oxygen species on lipid peroxida-
tion is substantial. 10 Results of this experimental protocol demonstrated a significantly better protection for animals given deferoxamine than for the placebo group. Deferoxamine is a clinically available substance, and its use can be promising for the prevention of spinal cord ischemia resulting from surgery on the thoracic and thoracoabdominal aorta. In summary, results of this experimental endeavor confirm the possible role of oxygen-derived free radicals in the pathophysiology of spinal cord injury induced by aortic crossclamping for surgery of thoracic and thoracoabdominal aneurysms. Prevention of hydroxyl formation by iron-chelating agents such as deferoxamine appears to be one of the most important mechanisms in protecting the spinal cord from ischemia reperfusion injury. Deferoxamine and allopurinol are clinically available agents tha t can be used for this category of patients. Although allopurinol, SOD, and deferoxamine are providing some degree of protection, the search WIll continue for an agent or combination of agents to maximize protection for the duration of ischemia to effectively complete the required operation. We extend our appreciation to Eva Germann for computer programming and statistical assistance and Terry Rihela for technical support. REFERENCES 1. Costello TG, Fisher A. Neurological complications following aortic surgery: case reports and review of the literature. Anaesthesia 1983;38:230-6. 2. Watson N. Paraplegia following cardiovascular surgery.
Paraplegia 1979;17:294-7. 3. Connolly lE. Prevention of paraplegia secondary to operation on the aorta. J Cardiovasc Surg (Torino) 1986;27: 410-7. 4. Wakabayashi A, Connolly lE. Prevention of paraplegia associated with resection of extensive thoracic aneurysms. Arch Surg 1976;11:1186-9. 5. Crawford ES, Schnessler JS. Thoracoabdominal and abdominal aortic aneurysm involving celiac, superior mesenteric, and renal arteries. World J Surg 1980;4:64351. 6. Verdant A, Page A 1 Cossette R, Dontigny L, Page P, Baillot R. Surgery of the descending thoracic aorta: spinal cord protection with the Gott shunt. Ann Thorac Surg 1988;42: 147-54. 7. Galbut DL, Bolooki H. Surgery of descending aorta: a method of autotransfusion and intercostal artery preservation. Chest 1982;5:590-2. 8. Oldfield EH, Plunkett RJ, Nylander WA Jr, Meacham WF. Barbiturate protection in acute experimental spinal cord ischemia. J Neurosurg 1982;56:511-6. 9. McCord JM. Oxygen-derived free radicals in post-ischemic tissue injury. N Eng! J 1983;312: 159-63.
Moo
Volume 104 Number 2 August 1992
10. Menasche P, Piwnica A. Free radicals and myocardial protection: a surgicalviewpoint. Ann Thorac Surg 1989;47: 939-45.
11. Heymann MA, Peyne BD, Hoffman JIE, Rudolph AM. Bloodflow measurement with radionuclide-labelled particles. Prog Cardiovasc Dis 1977;20:55-79. 12. Das DK, Engelman RM, Clement R, Otani H, Prasad MR, Rao PS. Role of xanthine oxidase inhibitors as free radical scavenger: a novelmechanism of action of allopurinol and oxypurinol in myocardial salvage. Biochem Biophys Res Commun 1987;148:314-9. 13. Eddy LJ, Stewart JR, Jones HP, Engerson TD, McCord JM, Downey 1M. Free radical-producing enzyme, xanthineoxidase,is undetectable in human hearts. Am J Physiol 1987;235:H709-11. 14. Godin DV; Bhimji S, McNeill JH. Effects of allopurinol pretreatment on myocardial ultrastructure and arrhythmias following coronary artery occlusion and reperfusion. Virchows Arch (Cell Path) 1986;52:327-41. 15. Godin DV, Ko DM, Garnett ME. Altered antioxidant status in the ischemicjreperfused rabbit myocardium: effects of allopurinol. Can J Cardiol 1989;5:365-71. 16. Qayumi AK, Jamieson WRE, Godin DV, et at. Response to allopurinolpretreatment in a swine model of heart-lung transplantation. J Invest Surg 1990;3:331-40. 17. Simpson PJ~ Miekelson JK, Lucchesi BR. Free radical scavengers in myocardial ischemia. Fed Proc 1987;46: 2413-21. 18. Peterson DA, Kelly B, Gerrard JM. Allopurinolcan act as an electron transfer agent. Is this relevant during reperfusion injury? Biochem Biophys Res Commun 1986;137: 76-9.
Spinal cord injury
261
19. Moorhouse PC, Grootveld M, Halliwell B, Quinlan JG,
Gutteridge JME. Allopurinol and oxypurinol and hydroxyl radical scavengers. FEBS Lett 1987;213:23-8. 20. Shlafer M, Kane PF; Kirsh MM, Mich AA. Superoxide dismutase plus catalase enhances the efficacyof hypothermic cardioplegia to protect the globally ischemic, reperfused heart. J THORAC CARDIOVASC SURG 1982;83:830-9. 21. Myers ML, Bolli R, Lekich RF; Hartley CJ, Roberts R. Enhancement of recoveryof myocardial function by oxygen free radical scavengers after reversible regional ischemia. Circulation 1985;72:915-21. 22. Svensson LG, Von Ritter eM, Groeneveld HT, et al. Cross-clamping of the thoracic aorta: influence of aortic shunts, laminectomy, papaverine,calcium channel blocker, allopurinol and superoxidedismutase on spinal cord blood flow and paraplegia in baboons. Ann Surg 1986;204:38-47. 23. Lim KH~ Conolly M, Rose D, Jacobowitz I, Cunningham IN. Time-responserelationship of recombinant superoxide dismutase in the ischemic spinal cord syndrome. Surg Forum 1986;37:443-5. 24. Myers CL, WeissSJ, Kirsh MM~ Shepard BM, Shlafer M, Mich AA. Effects of supplementing hypothermic crystalloid cardioplegicsolution with catalase, superoxide dismutase, allopurinol,or deferoxamine on functional recoveryof globally ischemicand reperfused isolatedhearts. J THORAC CARDIOVASC SURG
1986;91:281-9.
25. Menasche P, Grousset C, Gouduely M, Piwnica A. Prevention of hydroxylradical formation: a critical concept for improvingcardioplegia.Circulation 1987;76{Pt2):V2-180. 26. Bonser RS, Fragomen LS, Edwards BJ. Allopurinol and deferoxamine improve canine lung preservation. Transplant Proc 1990;22:557-8.
Pharmacologic interventions for prevention of spinal cord injury caused by aortic crossclamping The efficacy of pharmacologic agents for prevention and control of oxygen-derived free radical damage in ischemia-reperfusion injury of the spinal cord was assessed in a swine model of thoracic and tboracoabdominal aortic crossclamping. Animals were exposed to 30 minutes of ischemia that induced lethal, irreversible injury and paraplegia. The experimental groups were as foUows: group A (n = 7), control group, receiving no pharmacologic intenention; group B (n = 7), deferoxamine 50 mg/kgjday administered intravenously over 3 to 4 hours before ischemia; group C (n = 7), aUopurinol pretreatment 50 mgjkg/day for 3 days; and group D (n = 7), superoxide dismutase 60,000 units administered with 50,000 units before removal of the aortic crossclamp and 10,000 units over 10 minutes of reperfusion. Proximal hypertension was controUed with sodium nitroprusside and volume depletion, The methods of assessment were neurologic by a modified Tarlov criteria and blood flow by radiolabeled microspheres. Results of blood flow assessment confirmed a true ischemic episode of 30 minutes for aU animals in aU groups. The blood flow fell significantly during ischemia (p < 0.01) and a hyperemic response was evident in the early reperfusion period. AU animals in control group A were paraplegic. The group B (deferoxamine) results were superior; 85 % had grade In function on a modified Tarlov scale, with animals in the group standing and even walking with difficulty. Only one animal in this group had good movements of hind limbs but was unable to stand or walk. Neurologic recovery was limited in the allopurinol group (group C), with 85 % showing s6gbt neurologic recovery with limited movement of the hind limbs. The animals in the superoxide dismutase group (group D) aU had good recovery, with strong motor response of hind limbs, but were not able to stand. In summary, the results of this experimental protocol confirmed the possible role of oxygen-derived Cree radicals in the pathophysiology of spinal cord injury, induced by aortic crossclamping. Moreover, it proved that ischemiareperfusion injury could be altered by pharmacologic interventions. (J THORAC CARDIOVASC SURG 1992;104:256-61)
A. Karim Qayumi, MD, PhD, Michael T. Janusz, MD, W. R. Eric Jamieson, MD, and Donald M. Lyster, PhD, Vancouver, B.C., Canada
From the Division of Cardiovascular and Thoracic Surgery, Depart-
mentofSurgery,Facultyof Medicine, Faculty of Pharmacological Sciences, University of British Columbia, Vancouver, B.C.,Canada. Received for publication Jan. 3, 1991. Accepted for publication Jan. 20, 1992.
256
Supported by the Rick Hansen Man In Motion Legacy Fund. Addressfor reprints: A. K. Qayumi, MD, PhD, Assistant Professor of Surgery, Department of Surgery, Room 3100-910 West 10th Ave., Vancouver, Be, V5Z 4E3, Canada.
12/1/36840
Volume 104 Number 2 August 1992
Spinal cord injury
Spinal cord injury remains the most dreaded complication of operations on the aorta for treatment of descending thoracic and thoracoabdominal aneurysms. I, 2 The necessity for aortic crossclamping carries a risk of distal organ ischemia as a result of diminished perfusion. The organs at risk of ischemia include kidneys, liver, spinal cord, and intestines. The spinal cord is the organ most sensitive to ischemia. The resulting ischemic injury can
60
produce paraparesis or paraplegia. The risk of paraplegia
20
increases with the length of aorta replaced and with the length of time during which the aorta is crossclamped; the risk of paraplegia is about 3% for resections of the descending thoracic aorta, about 15% for operations involving the rnajority of the descending thoracic aorta plus the upper abdominal aorta, and about 25% for the entire descendingthoracic and abdominal aorta.v 4 Paraplegiais uncommon in operations requiring a crossclamp time less than 25 minutes, and it increases progressively with extended time intervals.l Attempts have been made to reduce spinal cord injury by improving operative techniqueand anesthesia management and by using shunts or bypass techniques to maintain distal perfusion.v'' The complex methods that have been used clinically and experimentally to eliminate or reduce the frequency of this complication have been ineffective. The mechanisms of tissue injury and possibleinterventional therapies have received limited attention in regard to the spinal cord. There is now acceptance that reactive oxidant speciesplaya primary role in the development of ischemia-reperfusion injury of various organs.v!" The oxygen-derived free radicals, namely, superoxide anion, hydrogen peroxide, and hydroxyl ion,are important in the mechanisms of the ischemia-reperfusion injury.!" The experimental protocol using a swine model of spinal cord ischemia,examined in this study, was designed to evaluate pharmacologic interventions in respect to control of oxygen-derived free radical injury to extend the limits of safety of surgically induced spinal cord ischemia. Materials and methods The efficacy of pharmacologic agents for prevention and control of oxygen-derived free radical damage in ischemiareperfusion injury of the spinal cord was assessed in a swine model of thoracic and thoracoabdominal aortic crossclamping. Experiments were performed on 28 swine weighing from 20 to 22 kg. The animals were divided into four groups of seven each. In group A (n = 7), the control group, no pharmacologic intervention was introduced. In group B (n = 7) deferoxamine in a dose of 50 mg/kg was administered intravenously over a 3~ to 4~hour period before ischemia. The group C (n = 7) animals were pretreated with allopurinol, 50 rug/kg/day, for 3 days, and
257
Flow 80
o'-------------.:.:..tPL--------L..--T1
T2
T3
Time
Fig. I. Spinal cord blood flow (mean ± standard deviation) assessed by nuclear imaging technique using specific microspheres at each time period. A Control group A; • group B;. group C; 0 group D; TI, before ischemia; T2, during ischemia; T3, 30 minutes after ischemia. *p < 0.01. in group D (n = 7) superoxide dismutase (SOD) 60,000 units was administered, 50,000 units before removal of the aortic crossclamp and 10,000 units over a lO-minute period after reperfusion. All animals were subjected to 30 minutes of ischemia and observed for 4 hours after recovery from anesthesia for neurologic status. Anesthesia was induced with ketamine (20 mg/kg body weight) and maintained with isoflurane (0.50/0 to 2.00/0). Ventilation was maintained with a volume ventilator and inflatable cuffed endotracheal tube. Hemodynamics were controlled by monitoring the systemic blood pressure (systolic, diastolic, mean) and heart rate. Electrocardiographic monitoring was also conducted throughout the procedure. Proximal hypertension was controlled with 20 to 40 mg sodium nitroprusside and volume depletion (300 to 500 ml). Nitroprusside was administered at a dose of 100 }lgjml and a flow rate of I ml/kg/min during the first 15minutes of ischemia. There were indications for volume depletion when the systemic pressure and heart rate were more than double the original value. The cardiovascular system in swine is sensitive. Therefore the use of both methods (volume depletion and .vasodilator infusion) is important to control the proximal hypertension. Under sterile conditions a limited thoracotomy was performed in the fourth left intercostal space to expose the proximal and midportions of the descending thoracic aorta. The aorta distal to the left subclavian artery was exposed for clamping. Proximal and distal arterial monitoring lines for pressure were inscribed via the left internal mammary artery and left femoral artery. The pericardium was opened for exposure of the left atrial appendage and placement of a polyethylene catheter for microsphere infusion. The external jugular vein was cannulated for intravenous access for the administration of nitroprusside to control proximal hypertension during aortic clamping and volume replacement during the unclamping phase. After aortic unclamping the thoracotomy was closed and the animal was permitted to regain consciousness for neurologic assessment
The Journal of Thoracic and Cardiovascular Surgery
2 5 8 Qayumi et al.
Table I. Effect of various pharmacologic interventions assessed by blood flow and Tarlov criteria Animal
p Value
0
3.0 ± 1.5 52.8 ± 6.2
0.01
1000/0
27.8 ± 4.6
3.4 ± 0.8 49.5 ± 5.8
0.01
7
25.0 ± 3.9
2.6 ± 0.8 45.6 ± 6.8
0.01
7
41.0 ± 10.5 6.1 ± 2.0 66.8 ± 9.4
0.01
No.
I
7
27.6 ± 4.2
7
C: Allopurinol (pretreatment for 3 days, 50 mgjkgjday) D: Superoxide dismutase
A: Control (no pharmacologic intervention) B: Deferoxamine (50 mg/kg)
Tarlov criteria at 4-hour reperfusion
Bloodflow duringischemia II
III
(60,000U)
I
II
III
14% 860/0
860/0
14%
100%
IV
P Value
0.05 compared with groups A and C
NS
0.05 compared with group A
NS, Not significant.
over 2 to 3 hours. The animals received appropriate sedation during the recovery period. Acid-base status was monitored throughout the procedure and metabolic acidosis was corrected with sodium bicarbonate, if necessary. After recovery from anesthesia and neurologic assessment, the animals were again anesthetized for laminectomy. The
laminectomy was performed at the levelof the tenth thoracic
vertebra to obtain spinal cord- specimens for quantitation of microspheres. The animals were killed at completion of the experiments. The experimental protocol was designed and performed in accordance with the principles of the Canadian Council on Animal Care and the University of British Columbia Animal Care Committee regulations.
The end points of evaluation were assessment of neurologic
status and spinal cord blood flow. Neurologic assessment included physical findings of hind limb neurologic function 4 hours after reperfusion. The neurologic status was determined when the animals were fully conscious. The assessment comprisedclinical criteria of hind limb neurologic function, according to the modified Tarlov criteria, as follows: Grade 0: No voluntary function (complete paralysis) Grade I: Movement of joints perceptible Grade II: Active movement of joints (inability to stand) Grade III: Able to stand (unable to walk) Grade IV: Complete recovery (able to stand, walk, run) Blood flow was assessed by the radiolabeled microsphere techniques. The blood flow assessment was used to determine the residual spinal cord blood flow, includingcollateral flow, for each of the study groups. Blood flow wasassessed at three time periods: time one (TI, before ischemia),gadolinium 153; time two (T2, during ischemia, 5 minutes before reperfusion), tin 113; and time three (T3, after ischemia, within 30 minutes after reperfusion, at the time of reestablishment of hemodynamic parameters), ruthenium 103. The radiopharmaceuticals were injected into the left atrium through the left atrial cannula and blood samples were collected simultaneously from the femoral artery. Blood flow was quantified by a method described by Heymann and colleagues. 1 I Blood flow assessments were analyzed for the effectof time and for comparison among the groups by means of repeated measures analysis of variance. The difference in Tarlov criteria among the groups was determined by nonparametric statistical analysis of Mann-Whitney pairwise comparison and Kruskal-
Wallis x2 test. Asterisks on Fig. 1 represent the significant differences among the groups. The significance was assessed at 950/0 and 98% confidence levels as indicated in the figures.
Results Assessment of spinal cord blood flow is shown in Table I and Fig. 1. There is a significant (p < 0.01) fall in blood flow in all groups of animals during ischemia (T2). At the time of reperfusion (T3) a hyperemic response is evident for all experimental groups. Equal spinal cord blood flow for all groups during ischemia (T2) indicates a true ischemicepisode and also indicates that there was essentially no collateral circulation for the spinal cord throughout the ischemic interval in all groups. The neurologic assessment (Table I, Fig. 2) demonstrates complete paraplegia with a modified Tarlov criteria score of zero for all seven animals in the control group (group A, 1000/0 paraplegia). In group B, in which deferoxamine was used, the neurologic status was superior to that of all other groups. Six of seven animals (860/0) were in grade III by modified Tarlov criteria, which indicates that the majority of animals in this group were standing and even walking with difficulty. Only one animal ( 14%) in this group was in grade II, with good movements of the hind limbs, but the animal was unable to stand and walk. The allopurinol-pretreated animals (group C) showed limited neurologic recovery; six animals (860/0) had a modified Tarlov score of I, slight neurologic recovery with limited movement of the hind limbs. Only one animal ( 140/0) had a better neurologic recovery, which was grade II by the modified Tarlov scale. In group D, where SOD was used, all seven animals (100%) had a strong motor response of the hind limbs and good neurologic recovery. All the animals in this group were in grade III of the modified Tarlov criteria. Consequently the animals in group B, where deferoxamine was used, had significantly (p < 0.05) better neurologic recovery than the control
Volume 104
Spinal cord injury
Number 2 August 1992
(group A) and allopurinol-pretreated (group C) animals. Animals in group D, where SOD was used, had significantly (p < 0.05) better recovery than the control group (Table I).
Control
259
Deferoxamine
Discussion The role of oxygen-derived free radicals in the pathogenesis of ischemia-reperfusion injury has been extensively investigated in the past decade.v!" Although a great deal of information is available about the adverse effectof oxygenspeciesin heart, lung, liver,kidney,intestine, and other tissues," little attention has been paid to spinal cord ischemia. Lipid peroxidation of the cellular membranes caused by oxygen species is a concept that generally applies to all tissues; therefore results of this experimental endeavor are mostly discussedin an overall lipid peroxidation concept. Allopurinol, a xanthine oxidase inhibitor, was used to prevent the formation of oxygen-derived free radicals. The protectiveeffect of allopurinolas a xanthine oxidase inhibitor is controversial because of the absence of detectable levels of xanthine oxidase in rabbit, pig, and human rnyocardium.P: 13 This brings into question the mechanism of the protectiveeffect of allopurinolin ischemia-reperfusion injury. Results of our experiments in a swinemodel of heart-lung transplantation, as wellas the experience of other investigators, 14-16 clearlydemonstrate the effectiveness of allopurinol in ischemia-reperfusion injury in species with no apparent xanthine oxidaseactivity whenthe drug is administered repeatedly. In search of other mechanisms of action, Das and colleagues'? proposed a direct scavenging property for allopurinol. A number of other possible mechanisms of allopurinol protection have been proposed, including an increased efficiency of adenosinetriphosphate salvage,17 facilitation of mitochondrial electron transfer, 18 and direct activationof endogenously formed reative species, such as hydroxyl radicalsor myeloperoxidase-derived hypochlorous acid. 19 Presenceof xanthine oxidasehas not been investigated in the swine spinal cord tissue. Therefore results of this experimental endeavor cannot confirm the fact that allopurinol acted as a xanthine oxidase inhibitor or expressed a mechanism of action other than xanthine oxidase inhibition. Although allopurinol does not afford the mostprotectionin this model, its availabilitymakes it clinically relevant. Another substance used in this experimental protocol was SOD. Despite the strong theoretical basis, there are controversies about its protective effect. SOD is thought to be exclusively an intracellular enzyme. Therefore administration of this substance extracellularly may be
100%
Allopurinol 1 2 3 4 5
SOD
• No movement @ Slight movement 0 Good movement without resistance @ Good movement against resistance • Normal
Fig. 2. Neurologic assessment by modified Tarlov criteria assessed 4 hours after reperfusion. Deferoxamine group B had a significantly higher score (p < 0.05) than the control group, with 100% paraplegia. ineffective. Shlafer and coworkers.l" using globally ischemic and reperfused isolated rabbit and cat hearts, showed reduction in cardiac injury with SOD and catalase. Myers and associates" demonstrated improved recoveryof function obtained with SOD and catalase in a canine model of regional ischemia. These investigators occluded the left anterior descendingcoronary artery for 15minutes and then reperfusedthe area for two hours. In contrast, Myers and colleagues" did not find any evidenceof added protection with SOD. These investigators used catalase and SOD in a separate series of experiments, and only the group of animals in which catalase was used in combination with SOD demonstrated a significant improvement of functional parameters . Myers's group concludedthat the protectiveeffectachievedby the former investigatorswasprobablybecauseof catalase and not SOD. There is a strong theoretical support for the combined use of SOD and catalase.20• 21 SOD is able to dismutate superoxideanions to hydrogen peroxide,which
260
The Journal of Thoracic and Cardiovascular Surgery
Qayumi et al.
is a harmful oxygen species in itself. Moreover, in the presence of iron, hydrogen peroxide can be converted into a more potent oxygen-derived free radical, which is the hydroxyl radical. This reaction needs 'catalase or glutathione peroxidase to convert the hydrogen peroxide into water and carbon dioxide. However, results of this experimental endeavor clearly demonstrated that SOD provides some degree of protection by itself without catalase or glutathione peroxidase. Animals in the SOD group had significantly better function in their hind limbs after 30 minutes of ischemia and 2 hours of reperfusion than did the control animals. The protective effect of SOD in this experimental model confirms the importance of the time and method of administration, as it is described by oth-
ers.!'' Svensson and colleaguesf investigated the effect of allopurinol, SOD, calcium channel blocker, and papaverine on spinal cord blood flow and paraplegia caused by 60 minutes' crossclamping of the thoracic aorta. These investigators emphasized the effect of papaverine as a vasodilator and did not find a protective effect using allopurinol and SOD. The time-response relationship of recombinant SOD in the ischemic spinal cord has also been investigated by Lim and associates.P The protective effect of SOD was significant within 30 minutes of spinal cord ischemia. SOD appeared to be ineffective in longer than 3D-minute ischemic intervals. Results of these latter investigators agree with our findings and also explain the failure of Svensson and colleagues to obtain a protective effect with SOD after 60 minutes of spinal cord ischemia. Chelation of iron with agents such as deferoxamine is thought to protect ischemic organs from reperfusion injury. Although Myers and associates-" did not find much benefit from deferoxamine administration, the majority of investigators found deferoxamine highly protective in the event of ischemia-reperfusion injury. 25, 26 Menasche and coworkers/" demonstrated that deferoxamine-containing cardioplegic solutions alone afford the best myocardial protection. Bonser, Fragomen, and Edwards/" also demonstrated in a model of heart-lung transplantation that preliminary reperfusion with deferoxamine improves the quality of ischemic lung protection as manifested by improved gas exchange and pulmonary hemodynamics after 4 hours of hypothermic preservation. OUf experience in a model of heart-lung transplantation (A.K.Q., unpublished data) demonstrated a protective effect using high-molecular-weight deferoxamine. Protective effect of deferoxamine is mainly due to chelation of ferric ions that catalyze the Fantom-Haber-Weiss reaction; consequently, in prevention of the most potent oxygen-derived free radical, the hydroxyl radicals, the deleterious effect of this oxygen species on lipid peroxida-
tion is substantial. 10 Results of this experimental protocol demonstrated a significantly better protection for animals given deferoxamine than for the placebo group. Deferoxamine is a clinically available substance, and its use can be promising for the prevention of spinal cord ischemia resulting from surgery on the thoracic and thoracoabdominal aorta. In summary, results of this experimental endeavor confirm the possible role of oxygen-derived free radicals in the pathophysiology of spinal cord injury induced by aortic crossclamping for surgery of thoracic and thoracoabdominal aneurysms. Prevention of hydroxyl formation by iron-chelating agents such as deferoxamine appears to be one of the most important mechanisms in protecting the spinal cord from ischemia reperfusion injury. Deferoxamine and allopurinol are clinically available agents tha t can be used for this category of patients. Although allopurinol, SOD, and deferoxamine are providing some degree of protection, the search WIll continue for an agent or combination of agents to maximize protection for the duration of ischemia to effectively complete the required operation. We extend our appreciation to Eva Germann for computer programming and statistical assistance and Terry Rihela for technical support. REFERENCES 1. Costello TG, Fisher A. Neurological complications following aortic surgery: case reports and review of the literature. Anaesthesia 1983;38:230-6. 2. Watson N. Paraplegia following cardiovascular surgery.
Paraplegia 1979;17:294-7. 3. Connolly lE. Prevention of paraplegia secondary to operation on the aorta. J Cardiovasc Surg (Torino) 1986;27: 410-7. 4. Wakabayashi A, Connolly lE. Prevention of paraplegia associated with resection of extensive thoracic aneurysms. Arch Surg 1976;11:1186-9. 5. Crawford ES, Schnessler JS. Thoracoabdominal and abdominal aortic aneurysm involving celiac, superior mesenteric, and renal arteries. World J Surg 1980;4:64351. 6. Verdant A, Page A 1 Cossette R, Dontigny L, Page P, Baillot R. Surgery of the descending thoracic aorta: spinal cord protection with the Gott shunt. Ann Thorac Surg 1988;42: 147-54. 7. Galbut DL, Bolooki H. Surgery of descending aorta: a method of autotransfusion and intercostal artery preservation. Chest 1982;5:590-2. 8. Oldfield EH, Plunkett RJ, Nylander WA Jr, Meacham WF. Barbiturate protection in acute experimental spinal cord ischemia. J Neurosurg 1982;56:511-6. 9. McCord JM. Oxygen-derived free radicals in post-ischemic tissue injury. N Eng! J 1983;312: 159-63.
Moo
Volume 104 Number 2 August 1992
10. Menasche P, Piwnica A. Free radicals and myocardial protection: a surgicalviewpoint. Ann Thorac Surg 1989;47: 939-45.
11. Heymann MA, Peyne BD, Hoffman JIE, Rudolph AM. Bloodflow measurement with radionuclide-labelled particles. Prog Cardiovasc Dis 1977;20:55-79. 12. Das DK, Engelman RM, Clement R, Otani H, Prasad MR, Rao PS. Role of xanthine oxidase inhibitors as free radical scavenger: a novelmechanism of action of allopurinol and oxypurinol in myocardial salvage. Biochem Biophys Res Commun 1987;148:314-9. 13. Eddy LJ, Stewart JR, Jones HP, Engerson TD, McCord JM, Downey 1M. Free radical-producing enzyme, xanthineoxidase,is undetectable in human hearts. Am J Physiol 1987;235:H709-11. 14. Godin DV; Bhimji S, McNeill JH. Effects of allopurinol pretreatment on myocardial ultrastructure and arrhythmias following coronary artery occlusion and reperfusion. Virchows Arch (Cell Path) 1986;52:327-41. 15. Godin DV, Ko DM, Garnett ME. Altered antioxidant status in the ischemicjreperfused rabbit myocardium: effects of allopurinol. Can J Cardiol 1989;5:365-71. 16. Qayumi AK, Jamieson WRE, Godin DV, et at. Response to allopurinolpretreatment in a swine model of heart-lung transplantation. J Invest Surg 1990;3:331-40. 17. Simpson PJ~ Miekelson JK, Lucchesi BR. Free radical scavengers in myocardial ischemia. Fed Proc 1987;46: 2413-21. 18. Peterson DA, Kelly B, Gerrard JM. Allopurinolcan act as an electron transfer agent. Is this relevant during reperfusion injury? Biochem Biophys Res Commun 1986;137: 76-9.
Spinal cord injury
261
19. Moorhouse PC, Grootveld M, Halliwell B, Quinlan JG,
Gutteridge JME. Allopurinol and oxypurinol and hydroxyl radical scavengers. FEBS Lett 1987;213:23-8. 20. Shlafer M, Kane PF; Kirsh MM, Mich AA. Superoxide dismutase plus catalase enhances the efficacyof hypothermic cardioplegia to protect the globally ischemic, reperfused heart. J THORAC CARDIOVASC SURG 1982;83:830-9. 21. Myers ML, Bolli R, Lekich RF; Hartley CJ, Roberts R. Enhancement of recoveryof myocardial function by oxygen free radical scavengers after reversible regional ischemia. Circulation 1985;72:915-21. 22. Svensson LG, Von Ritter eM, Groeneveld HT, et al. Cross-clamping of the thoracic aorta: influence of aortic shunts, laminectomy, papaverine,calcium channel blocker, allopurinol and superoxidedismutase on spinal cord blood flow and paraplegia in baboons. Ann Surg 1986;204:38-47. 23. Lim KH~ Conolly M, Rose D, Jacobowitz I, Cunningham IN. Time-responserelationship of recombinant superoxide dismutase in the ischemic spinal cord syndrome. Surg Forum 1986;37:443-5. 24. Myers CL, WeissSJ, Kirsh MM~ Shepard BM, Shlafer M, Mich AA. Effects of supplementing hypothermic crystalloid cardioplegicsolution with catalase, superoxide dismutase, allopurinol,or deferoxamine on functional recoveryof globally ischemicand reperfused isolatedhearts. J THORAC CARDIOVASC SURG
1986;91:281-9.
25. Menasche P, Grousset C, Gouduely M, Piwnica A. Prevention of hydroxylradical formation: a critical concept for improvingcardioplegia.Circulation 1987;76{Pt2):V2-180. 26. Bonser RS, Fragomen LS, Edwards BJ. Allopurinol and deferoxamine improve canine lung preservation. Transplant Proc 1990;22:557-8.