- Email: [email protected]
Postischemic hyperglycemia worsens neurologic outcome after spinal cord ischemia
Postischemic hyperglycemia worsens neurologic outcome after spinal cord ischemia M a r k R. H e m m i l a , MS, G e r a l d B. Zelenock, M D , and Louis G. D'Alecy, D M D , P h D , Ann Arbor, Mich.
Purpose: The neurologic effect of induced hyperglycemia in the postischemic period was investigated with a rat aortic occlusion model. Alethods: Sprague-Dawley rats weighing 200 to 350 gm were anesthetized, intubated, and ventilated with 1% to 1.5% halothane. Tempera~tre was continuously monitored and maintained at 37 ° + 0.5 ° C. The chest was opened, the thymus excised, and the aortic arch exposed. Snares were placed around the aorta distal to the left subclavian artery and the right and left subclavian arteries. The three vessels thus isolated were occluded for 8 rninutes. With snare release and withdrawal, the rats received an intraperitoneal injection of 5% dextrose in water (2 gm/kg) or an equivalent volume of 0.9% saline solution. In a second group of rats the administration of glucose or saline solution was delayed until 30 rninutes after snare release. Blood samples for blood glucose determination were obtained before operation, before occlusion, immediately after occlusion, and 15, 30,45, 60, and 240 minutes after occlusion. A neurologic deficit score was assigned at 1, 4, 18, and 24 hours after occlusion to quantify hindlimb neurologic deficit based on 15-point scale (0 = normal, 15 = severe deficit). Sham-operated rats received the same operation and injection, but the snares were only manipulated and not made occlusive. Results: The rats that were administered glucose immediately after snare release showed a statistically significant exacerbation of lower extremity neurologic deficit at 24 hours after occlusion (p --- 0.05, Mann-Whitney IJ test). The sham-operated rats were normal (0 score) at 24hours. Significant elevation of blood glucose (321 -+ 33 mg]dl) was seen in the glucose-injected rats at 15 minutes and continued for up to 4 hours after occlusion (p -- 0.040 and 0.014, respectively; Student's t test). Conclusion: Postischemic hyperglycemia immediately after a standard spinal cord ischemic slxess worsens neurologic outcome. (J VAsc SURG 1993;17:661-8.)
Paraplegia resulting from spinal cord ischemia is a potential complication o f surgical procedures that involve clamping o f the thoracicoabdominal aorta. In a study o f 605 patients who underwent vascular repair ofl~horacicoabdominal aortic aneurysms, it was noted that the overall incidence o f lower extremity neurologic dysfunction was 11%} For the subset o f patients whose operation involved the complex repair of an anemTsm that extended from just distal to the From the Departments of Physiology(Dr. D'Alecy)and Surgery (Drs. Zelenock and D'Alecy), the University of Michigan Medical School, and the Veteran's Administration Medical Center, Ann Arbor. Supported by a Veteran's Administration merit review grant and the Uniw:rsity of Michigan Medical School Committee on Student BiomedicalResearch. Reprint requests: Louis G. D'Alecy, PhD, the University of Michigan Medical School, 7799 Medical Science II, 1301 Catherine St., Ann Arbor, MI 48109-0622. 2411/39830
left subclavian artery to involve most o f the descending thoracic aorta and most or all o f the abdominal aorta, the occurrence o f postoperative paraplegia was 28%. If the aneurysm was dissecting, this percentage increased even further to 40%. Because paraplegia is a c o m m o n and extremely debilitating complication o f thoracicoabdominal aneurysm repair, there has been an active experimental effort to identify potential perioperative protocols, both pharmacologic and mechanistic, that could reduce the risk o f lower extremity neurologic deficit resulting from spinal cord ischemia. 2-n In addition there are other situations in which the delivery o f blood to the spinal cord is temporarily impaired, such as in trauma or profound shock, that have the potential to place the functional integrity o f the spinal cord at risk from ischemic injury. Therefore it would be advantageous if the spinal cord could be protected from ischemic injury by control o f physiologic 661
662 Hemmila,Zelenock, and D'Alecy
variables identified as potential risk factors for the development of paraplegia. Previous work in our laboratory 13,14and by other investigators is has shown that hyperglycemia before and during an aortic cross-clamp/declamp sequence exacerbates paraplegia in both the rat and rabbit spinal cord ischemia models. It has also been demonstrated that pretreatment with insulin, to produce a moderate degree of hypoglycemia, protects from paraplegia in the rat 16 and rabbit ~7 aortic occlusion models. It must be emphasized that the aforementioned studies examined only the effect of preocclusion induction of hyperglycemia on neurologic deficit in spinal cord ischemia models. The possibility that hyperglycemia induced after an ischemic insult could also be detrimental to lower extremity neurologic function has received very little attention but has obvious therapeutic implications. The issue is not resolved. Pulsinelli et al)8 reported that glucose given after global central nervous system (CNS) ischemia produced no greater damage than did a saline injection. Conversely, our laboratory has found that, when brain temperature is controlled, administration of glucose to produce hyperglycemia immediately after global CNS ischemia does indeed worsen neurologic outcome compared with salineinjected controls (unpublished data). 19 The purpose of this study was to determine the effect of hyperglycemia in the postischemic period on lower extremity neurologic function after a standard spinal cord ischemic stress in the rat aortic occlusion model. A secondary objective was to characterize better the relevant time course to determine if there was any significant difference in postischemic lower extremity neurologic deficit with hyperglycemia induced immediately after occlusion compared with that induced 30 minutes after the end of the occlusion period. MATERIAL A N D M E T H O D S
Thirty-six male Sprague-Dawley rats, aged 2 to 3 months and weighing 200 to 350 gm, were housed individually with free access to food and water. Animal care complied with the Principles ofLaboratory
Animal Care and the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication No. 85-23, revised 1985). On the day of surgery the rats were weighed and anesthesia was induced by placing them in a chamber containing 3.5% halothane. After induction, the rats were transferred from the chamber to a mask apparatus. Tracheal intubation was then performed with a neonatal laryngoscope reduced to a blade width of 7 mm to assist in the placement of a 2.5 mm outside
JOURNALOF VASCULARSURGERY April 1993
diameter tube 8 cm long. The rats were then ventilated with a nonrebreathing pressure ventilator (model 212; Harvard Apparatus Co., Inc., S. Natick, Mass.) at 90 cycles/min with 1% to 1.5% halothane and 5 cm H 2 0 positive end expiratory pressure (PEEP). Temperature was monitored continuously with a probe inserted 4.5 cm into the rectum and maintained at 37 ° -+ 0.5 ° C with a homeothermic blanket control unit (model 513; Harvard Apparatus Co., Inc.). The left chest wall was opened just lateral to the sternum from the apex of the manubrium to thesuperior aspect of the third rib. The thymus was excised and the aorta isolated just distal to the left subclavian artery. At this location a 15 cm segment of polyethylene 10 tubing was placed around the aorta, carefully avoiding the left vagus and recurrent laryngeal nerves. The free ends of the tubing were then passed through a 3 cm segment of polyethylene 160 tubing, creating a snare. Snares were placed around the proximal right and left subclavian arteries in a similar manner. A t the beginning of the ischemic period, the three snares were pulled taut and secured, thus occluding each of the isolated vessels. Vessel occlusion was verified by visual inspection of the snare site and the vessel distal to the snare. During the 8-minute occlusion period, the PEEP was increased to 10 cm H 2 0 and maintained at this level. The snares exited the chest superior to the m a n u b r i ~ and the incision was closed. At the end of the occlusion period, defined as time t = 0, the snares were released and withdrawn. Chest closure was completed in three layers with 3-0 silk, PEEP was discontinued, and halothane anesthesia was terminated. Concomitant with snare withdrawal, the rats received an intraperitoneal injection of 5% dextrose in water (2 gm/kg) or an equivalent volume of 0.9% sodium chloride solution. Extubation was performed after the rat demonstrated whisker activity and maintained voluntary ventilation when disconnected from the ventilator. Blood samples for determinafion of blood glucose concentration (0.5 ml) were obtained immediately after induction of anes-~ thesia, before aortic occlusion, immediately after release of the occluding snares, and 15, 30, 45, 60, and 240 minutes after occlusion. Plasma blood glucose concentrations were determined by the glucose oxidase method with an automated glucose analyzer (model 23A; YSI, Yellow Springs, Ohio). A neurologic deficit score was assigned 1, 4, 18, and 24 hours after occlusion to quantify hindlimb motor deficit, based on a 15-point scale used extensively by our laboratory (Table I). 13'16 The sham-operated group received the same
~OURNAL OF VASCULAR SURGERY Volume 17, Number 4
anesthetic, ,operation, and intraperitoneal injection of 5% dextrose in water (2 gm/kg). The only difference between sham and experimental groups was that the snares were pulled taut only momentarily and then released immediately. Thus the sham animals did not sustain aortic occlusion or spinal cord ischemia. In a separate experiment, rats received delayed intraperitoneal injections of 5% dextrose or 0.9% sodium cl~Loride solution 30 minutes after occlusion instead of immediately after occlusion. The experimental protocol used for these rats was otherwise the ~2rne as the one described above except that one additional blood sample was obtained 90 minutes after occlusion. Twenv£-one rats were used in the first experiment: glucose treated (n = 7), saline control (n = 6), and sham (n = 2). Six rats died either during or after surgery of technical complications associated with the experimental preparation. In the second experiment (glucose-treated 30 minutes after occlusion) 15 rats were used: glucose-treated (n = 7) or saline control (n = 7). One rat died after surgery of technical complications. Neurologic deficit scores were compared nonparametrically with the single-tailed Mann-Whitney U test based on previous demonstration of the adverse effect of glucose. 13 Blood glucose concentrations were compared parametrically with the un7fired single-tailed Student's t test. The p values determined by the Mann-Whitney U test are expressed as less than or equal to table values. Exact p values are given for Student's t test. All average values are expressed as means _+ 1 SEM. Data analysis was performed with a StatView 512 + statistical software package on a Macintosh SE personal computer. RESULTS Neurologic deficit scores. The group of rats administered glucose immediately on reperfusion (time t = 0) showed a statistically significant exacerbation of their hindlimb neurologic deficit 24 hours after occlusion compared with saline-treated rats ~p ___ 0.0!51, Mann-Whitney U test). At earlier testing periods (11, 4, and 18 hours after occlusion), no statistically significant difference in neurologic deficit was detected (Fig. 1, top panel). The sham-operated rats had an initial transient postoperative neurologic deficit atl:ributable to the anesthetic and open-chest procedure but had no deficit 24 hours after occlusion. These sham results are similar to those produced previously by the sham procedure in our laboratory for this model by LeMay et al. ~3 In the rats with delayed administration of glucose (injected at t = 30 minutes after occlusion) no statistically significant
Hemmila, Zelenock, and DMlecy 6 6 3
Table I. Rat spinal cord ischemia model neurologic deficit scoring Deficit score Walking with lower extremities 0 N o evidence o f deficit 1 Toes flat under body when walking but ataxia exists 2 Knuckle walks 3 Movement in lower extremities but unable to knuckle walk 4 N o movement, drags lower extremities Horizontal rope 0 Grasps rope and pulls up with lower extremities 1 Raises lower extremities and grasps rope without pulling 2 Raises lower extremities but cannot grasp rope 3 Does not raise lower extremities Screen 0 Lower extremities grasp screen to I80 degrees for more than 5 sec 1 Lower extremities grasp screen to 180 degrees for less than 5 sec 2 Lower extremities grasp screen past vertical but not to 180 degrees 3 Lower extremities fall from screen past vertical (270- to 180-degree range) Bar at 45 degrees (best o f three trials) 0 Lower extremities grasp bar for more than 10 sec 1 Lower extremities grasp bar for more than 5 sec 2 Lower extremities grasp bar for less than 5 sec 3 Lower extremities slide off bar without grasping Pain sensation 0 Strong withdrawal to toe pinch 1 Weak withdrawal to toe pinch 2 N o reaction to toe pinch Maximum deficit = 15 Minimum deficit = 0 Data from LeMay et al. l VAsc SURG 1987;6:383-90; )~Surg Res 1988;44:352-8. Note: For each test a score o f zero indicates normal performance. Increasing test scores correspond to an increase in lower extremity neurologic deficit.
difference was detected in any of the neurologic deficit scores between the glucose-treated and salinetreated rats. Comparing rats administered glucose at time t = 0 (Fig. 1, toppanel) with rats given glucose at time t = 30 minutes (Fig. 1, bottompand) revealed that the earl), postocclusion administration of glucose significantly worsened lower extremity neurologic deficit scores 18 and 24 hours after occlusion (p -< 0.049 18 hours andp _< 0.027 24 hours after occlusion, Mann-Whitney U test). Blood glucose concentrations. Baseline blood glucose concentrations were slightly elevated in both glucose and saline groups (172 + 7 mg/dl and 177 +-6 mg/dl, respectively). In response to the stress associated with the surgical procedure, both groups had further elevation of blood glucose to postocclusion concentrations of 261 + 23 mg/dl in the glucose-treated group and 234 __ 18 mg/dl in the
IOURNALOF VASCULARSURGERY April 1993
664 Hemmila, Zelenock, and D'Alecy
Injection at t = 0 min Postocclusion 15 kg. 111
10
0 m
¢3 0
III Glucose (n=7)
,_1
"1 Saline
(n=6)
0 n-
~] Sham
(n=2)
uJ Z
1
4
18
24
Injection at t = +30 min Postocclusion 15
_P 0 m
M. IJJ ¢2
10
0 |
(3 0
Glucose (n=7)
"1 Saline
..I
(n=7)
0 rr UJ Z 1
4 HOURS
18
24
POSTOCCLUSlON
Fig. 1. Neurologic deficit score 1, 4, 18, and 24 hours after occlusion for glucose- and saline-treated groups injected at time t = 0 and t = 30 minutes after occlusion. For rats injected at time t = 0, glucose-administered group showed significant exacerbation ofneurologic deficit compared with saline-treated group 24 hours after occlusion. *p - 0.051, Mann-Whitney U test. Top, Neurologic deficit score was not obtained 18 hours after occlusion for one saline-treated rat.
saline-treated group when injected at t = 0 minutes (Fig. 2), Blood glucose concentrations further diverged, reaching a maximum at t = 60 minutes after occlusion o f 391 --- 37 mg/dl for glucose-treated and a maximum o f 249 _+ 13 mg/dl for saline-treated rats 15 minutes after occlusion. Significant elevations o f blood glucose concentration were seen at all time points in the glucose-treated rats compared with the saline-treated rats beginning 15 minutes after occlusion and continuing for 4 hours after occlusion
= 0.040 15 minutes andp -- 0.014 4 hours after occlusion, Student's t test). The rats that received delayed administration o f glucose or saline solution (t = 30 minutes after occlusion) had similar increases in baseline blood glucose concentrations (185 _ 9 mg/dl and 174 _+ 4 mg/dl in the glucose- and saline-treated groups, respectively). Postocclusion but preinjecfion blood glucose maximums o f 225 __ 22 mg/dl and 234 __ 18 mg/dl in the glucose- and saline-treated
}'OURNALOF VASCULARSURGERY Volume 17, Number 4
Hemmila, Zelenock, and D'Alecy 665
Injection at t = 0 min Postocclusion 400'
¢3 r~ O O
1
100 IDeolemp + InJeotlon !
Glucose Injection (n=7) Saline Injection (n=6)
z,
4
.J
o
-1
|
|
|
!
i
o
1
2
3
4
Injection at t = +30 min Postocclusion 400.
300'
r~
o
lOO
0
Injection Declamp~/~
Glucose Injection (n=7) Saline Injection (n=7)
&
..J m
o
|
0
!
I
1 2 TIME (hours)
!
i
3
4
Fig. 2. Blood glucose concentrations versus time for glucose- and saline-injected rats. When treated at time t = 0, glucose-injected rats showed significant elevation in blood glucose concentration compared with saline-injected rats starnng 15 minutes after occlusion and continuing to 4 hours after occlusion (p = 0.040 at 15 minutes and p = 0.014 at 4 hours, Student's t test). For rats treated at t = 30 minutes, glucose-injected rats showed significant elevation in blood glucose concentration compared with saline-injected rats starting 45 minutes after occlusion and continuing to 4 hours after occlusion (p = 0.006 45 minutes and p = 0.008 4 hours after occlusion, Student's t test).
rats were: n o t e d at time t = 15 minutes. After the intraperitoneal injection o f glucose and saline solution at time t = 30 minutes after occlusion, b l o o d glucose levels diverged with the glucose group, reaching a m a x i m u m o f 369 + 59 m g / d l at time t = 4 hours after occlusion. T h e saline-treated rats had a slight decline in b l o o d glucose levels f r o m a m a x i m u m o f 233 + 23 m g / d l at time t = 45 minutes to a :minimum o f 181 +_ 20 m g / d l at time t = 4 hours afi:er occlusion. In animals receiving delayed adminisu:ation o f glucose, statistically significant
elevation o f b l o o d glucose concentrations was seen 45 minutes after occlusion and continuing to 4 hours after occlusion c o m p a r e d the saline-treated g r o u p (p = 0.006 45 minutes a n d p = 0.008 4 hours after occlusion, Student's t test). DISCUSSION
Several prior studies in rat and rabbit spinal cord ischemia models indicate that preischemic hyperglycemia significantly worsens neurologic o u t c o m e after t e m p o r a r y aortic occlusion 131s and insulin
666 Hemmila, Zelenock, and D'Alecy
pretreatment; to produce a moderate degree of hypoglycemia has been demonstrated to protect against paraplegia. 16'17 Despite its obvious therapeutic importance, postischemic hyperglycemia and the issue of glucose administration during the postischemic period has received little attention. It must be emphasized that all prior studies in spinal ischemia examined only the effect of preocclusion physiologic variables on lower extremity neurologic deficit. This study demonstrated for the first time that postocclusion hyperglycemia, caused by treatment with exogenous glucose, can produce a significant exacerbation of hindlimb neurologic deficit, specifically paraplegia, in the rat spinal cord ischemia model. This suggests that at least some of the neurologic damage resulting from spinal cord ischemia must occur during the reperfusion period and postocclusion hyperglycemia worsens its outcome. It also rather precisely defined the period of increased vulnerability; if administration of exogenous glucose was delayed until 30 minutes after occlusion, the resulting hyperglycemia did not significantly increase the degree of paraplegia compared with salineinjected controls. The mechanism by which postischemic hyperglycemia exacerbates spinal cord ischemic injury caused by aortic clamping is not completely clear. Oxygen radical mechanisms of brain injury after ischemia and reperfusion have been reviewed recently, 2° and by inference these mechanisms may be operable in other settings of CNS injury such as spinal cord ischemia. Emerging evidence supports the theory that a variety of pathologic events are generated within spinal cord tissue during the ischemic period that create a biochemical environment whereby additional cellular damage is produced after reperfusion. In experimental models, alterations in antioxidant enzyme systems within ischemic tissues may predispose the cells to the formation of reactive oxygen metabolites on reperfusion. 21 Because hyperglycemia in the early reperfusion period produced an increase in neurologic damage to the spinal cord, it suggests either a mechanistic linkage between abnormally high blood glucose concentration and reperfusion oxygen radical injury or mechanistically discrete but synergistic adverse affects. The importance of the temporal sequence is emphasized by the lack of adverse effect of hyperglycemia when delayed for 30 minutes. Other physical parameters of reperfusion may also influence outcomes. Rapid reperfusion and reoxygenation of ischemic tissue has been noted to have a
JOURNALOF VASCULARSURGERY April 1993
significantly worse effect on cells compared with gradual establishment of reperfusion. An experiment by Danielisova et al.s showed that a decreased partial pressure of oxygen in arterial blood during reperfusion can have a beneficial effect after an ischemic reperfusion sequence in the spinal cord. It appears that a gradual return to aerobic glycolysis after an ishemic stress may be more effective in preventing reperfusion neuronal injury than the approach of maximal reoxygenation coupled with hyperglycemia. It has long been known that ischemia causes shift in cellular metabolism from aerobic to anaerobic glycolysis. Recent studies by Robertson and Grossman 17have shown that after reperfusion there is a lag in the return to aerobic metabolism. This delay can lead to an increase in lactic acidosis during reperfusion as a result of the increased availability of glucose as a substrate for anaerobic glycolysis. From experiments in cardiac resuscitation in dogs it is known that elevated brain lactate accumulation and increased neurologic deficit are associated with modest hyperglycemia in the setting of global CNS ischemia. 22 These results suggest lactate accumulation as a possible mechanism to explain worsened neurologic outcome after spinal cord ischemia in the rat with hyperglycemia. Based on this and our prior studies, 13'16 we Can conclude that there are two critical phases wh¢.~,, hyperglycemia can significantly affect the degree ot' injury produced by an episode of spinal cord ischemia. The most critical phase is before and during the actual ischemic period. Hyperglycemia before the ischemic period produces a maximal exacerbation of paraplegia in the rat spinal cord ischemia model. The second phase during which hyperglycemia worsens neurologic outcome is the immediate reperfusion period after an ischemic episode. Hyperglycemia produced immediately after occlusion significantly exacerbates the incidence and severity of paraplegia in the rat aortic occlusion model. However, the detrimental effect of exogenous glucose administration on neurologic outcome is quite circumscribed. When, glucose is given 30 minutes after reperfusion, there is no incremental effect. Although pretreatment to avoid hyperglycemia is practical and relevant in elective operative procedures involving aortic cross-clamping, not all clinical cases with the potential for ischemic spinal cord injury present as a pretreatment situation. The avoidance of induced hyperglycemia after an ischemic event involving the spinal cord is also readily possible. 23 Even
JOURNAL OF VASCULARSURGERY Volume 17, Number 4
while recognizing the concern that p e r m a n e n t neurologic & a n a g e m a y result f r o m the delayed recognition and t r e a t m e n t o f p r o f o u n d hypoglycemia, it m u s t be t m d e r s t o o d that euglycemia or even mildm o d e r a t e hypoglycemia is n o t associated with injury and m a y even be protective. T h e ready availability and capacity for rapid assessment o f b l o o d glucose concentration in the e m e r g e n c y r o o m , the operating r o o m , and at the bedside m a k e it possible to maintain patients' b l o o d glucose concentrations in the euglycemic range w h e n e v e r there is the risk o f spinal cord .dchemia. A review by Sieber et al.24 reevaluated the use o f intraoperative glucose in neurosurgery and concluded that intraoperative glucose administration appears to have m i n i m a l beneficial effects. T h e only instance in w h i c h hyperglycemia has been associated with a possible decreased a m o u n t o f neuronal injury is in the setting o f focal cerebral ischemia. 2s'26 These t w o studies d e m o n s t r a t e d that hyperglycemia during the focal ischemic period resulted in less cortical damage, by histologic assessment, c o m p a r e d with saline-treated controls. N o functional o u t c o m e s were reported. I t m u s t be emphasized that m o r p h o l o g i c data and timctional data in this setting have yet to be positively correlated. W i t h the one a f o r e m e n t i o n e d exception, hyperglycemia has been d e m o n s t r a t e d to have little if "3Y bene.ficial effect in the setting o f neuronal ischemia. T h e evidence continues to collect that, in spinal cord ischemia, hyperglycemia has adverse consequences w h e n present either before or after the ischemic event. Therefore glucose-containing intravenous fluids should be used only w h e n there is a clinically compelling justification for supplemental glucose. We thank Rakesh Patel for his assistance in the start-up of this experiment.
REFERENCES 1. Crawford ES, Crawford JL, Hazim JS, 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. 2. &gee JM, Flanagan T, Blackbourne LH, Kron IL, Tribble CG. Reducing postischemic paraplegia using conjugated superoxide dismutase. Ann Thorac Surg 1991;51:911-4. 3. Barsan WG, Hedges JR, Syverud SA, Dronen SC, Dimlich RV. Effects of dichloroacetate in spinal stroke in the rabbit. Life Sci 1987;41:1065-9. 4. Breckwoldt WL, Genco CM, Connolly RJ, Cleveland RJ, Diehl JT. Spinal cord protection during aortic occlusion:
Hemmila, Zelenock, and D'Alecy 667 efficacy of intrathecal tetracaine. Ann Thorac Surg 1991;51: 959-63. 5. Danielisova V, Marsala M, Chavko M, Marsala J. Postischemic hypoxia improves metabolic and functional recovery of the spinal cord. Neurology 1990;40:1125-9. 6. Dasmahapatra HK, Coles JG, Wilson GI, et al. Relationship between cerebrospinal fluid dynamics and reversible spinal cord ischemia during experimental thoracic occlusion. J Thorac Cardiovasc Surg 1988;95:920-3. 7. DelRossi AJ, Cernaianu AC, Cilley Jt-I, et al. Preventive effect of Fluosol-DA for paraplegia encountered after surgical treatment of the thoracic aorta: preliminary results in a dog model. J Thorac Cardiovasc Surg 1990;99:665-9. 8. I-Iall ED. Effects of the 21-aminosteroid U74006F on posttraumatic spinal cord ischemia in cats. J Neurosurg 1988;68:462-5. 9. Kochhar A, Zivin JA, Lyden PD, Mazzarella V. Glutamate antagonist therapy reduces neurologic deficits produced by focal central nervous system ischemia. Arch Neuro11988;45: 148-53. 10. LeMay DR, Zetenock GB, D'Alecy LG. Neurological protection by dichloroacetate depending on the severity of injury in the paraplegic rat. J Neurosurg 1990;73:118-22. 11. Wadouh F, Wadouh R, Hartmann M, Crisp-Lindgren N. Prevention of paraplegia during aortic operations. Ann Thorac Surg 1990;50:543-52. 12. Yum SW, Faden AI. Comparison of the ueuroprotective effects of the N-methyl-D-aspartate antagonist MK-801 and the opiate-receptor antagonist nalmefene in experimental spinal cord ischemia. Arch Neurol 1990;47:277-81. 13. LeMay DR, Neal S, Neal S, Zelenock GB, D'Alecy LG. Paraplegia in the rat induced by aortic cross clamping: model characterization and glucose exacerbation of neurological deficit. J VASCSURG 1987;6:383-90. 14. Lundy EF, Ball TD, MandeUMA, Zelenock GB, D'Alec3TLG. Dextrose administration increases sensory/motor impairment and paraplegia following infrarenal aortic occlusion in the rabbit. Arch Surg 1987;102:737-42. 15. Drummond JC, Moore SS. The li~luence of dextrose administration on neurologic outcome after temporary spinal cord ischemia in the rabbit. Anesthesiology 1989;70:64-70. 16. LeMay DR, Lu AC, Zelenock GB, D'Atecy LG. Insulin administration protects from paraplegia in the rat aortic occlusion model. J Surg Res 1988;44:352-8. 17. Robertson CS, Grossman RG. Protection against spinal cord ischemia with insulin-induced hypoglycemia. J Neurosurg 1987;67:739-44. 18. Pulsinelli WA, Waldman S, Rawlinson D, Plum F. Moderate hyperglycemia augments ischemic brain damage: aneuropathologic study in the rat. Neurology 1982;32:1239-46. 19. Lundy EF, Kuhn JE~ Kwon JM, Zelenock GB, D'Alecy LG. Infusion of five percent dextrose increases mortality and morbidity following six minutes of cardiac arrest in resuscitated dogs. J Crit Care 1987;2:4-14. 20. Traystman RJ, Kirsch JR, Koehler RC. Oxygen radical mechanisms of brain injury following ischemia and reperfusion. J AppI Physiol 1991;71:1185-95. 21. Fantone JC. Pathogenesis of ischemia-reperfusion injury: an overview. In: Zelenock GB, ed. Clinical ischemic syndromes: mechanisms and consequences of tissue injury. 1st ed. St Louis: Mosby-Year Book, 1990:141-2. 22. Natale JE, Stante SM, D'Alecy LG. Elevated brain lactate
668
Hemmila, Zelenock, and D'Alecy
accumulation and increased neurologic deficit are associated with modest hyperglycemia in global brain ischemia, lq,esuscitation 1990;19:271-89. 23. Browning RG, Olson DW, Stueven HA, Mateer JR. Fifty percent dextrose: antidote or toxin? Ann Emerg Med 1990; 19:683-7. 24. Sieber FE, Smith DS, Traystman RJ, Wollman H. Glucose: a reevaluation of its intraoperative use. Anesthesiology 1987;67:72-81.
IOURNAL OF VASCULARSURGERY April 1993
25. Kraft SA, Larson P, Shuer LM, et al. Effect of hyperglycemia on neuronal changes in a rabbit model of focal cerebral ischemia. Stroke 1990;21:447-50. 26. Zasslow MA, pearl RG, Shuer LM, et al. Hyperglycemia decreases acute neuronal ischemic changes after middle cerebral artery occlusion in cats. Stroke 1989;20:519-23.
Submitted Jan. 21, 1992; accepted May 28, 1992.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the JOURNAL Or VASCU~R SURGERY for 1993 are available to subscribers only. They may be purchased from the publisher at a cost of $69.00 for domestic, $89.83 for Canadian, and $85.00 for international subscribers for Vol. 17 (January to June) and Vol. 18 (July to December). Price includes shipping charges. Each bound volume contains a subject and author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Subscription Services, Mosby, 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, USA. In the United States call toll free (800)325-4177, ext. 4351. In Missouri or foreign countries call (314)453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.
Purpose: The neurologic effect of induced hyperglycemia in the postischemic period was investigated with a rat aortic occlusion model. Alethods: Sprague-Dawley rats weighing 200 to 350 gm were anesthetized, intubated, and ventilated with 1% to 1.5% halothane. Tempera~tre was continuously monitored and maintained at 37 ° + 0.5 ° C. The chest was opened, the thymus excised, and the aortic arch exposed. Snares were placed around the aorta distal to the left subclavian artery and the right and left subclavian arteries. The three vessels thus isolated were occluded for 8 rninutes. With snare release and withdrawal, the rats received an intraperitoneal injection of 5% dextrose in water (2 gm/kg) or an equivalent volume of 0.9% saline solution. In a second group of rats the administration of glucose or saline solution was delayed until 30 rninutes after snare release. Blood samples for blood glucose determination were obtained before operation, before occlusion, immediately after occlusion, and 15, 30,45, 60, and 240 minutes after occlusion. A neurologic deficit score was assigned at 1, 4, 18, and 24 hours after occlusion to quantify hindlimb neurologic deficit based on 15-point scale (0 = normal, 15 = severe deficit). Sham-operated rats received the same operation and injection, but the snares were only manipulated and not made occlusive. Results: The rats that were administered glucose immediately after snare release showed a statistically significant exacerbation of lower extremity neurologic deficit at 24 hours after occlusion (p --- 0.05, Mann-Whitney IJ test). The sham-operated rats were normal (0 score) at 24hours. Significant elevation of blood glucose (321 -+ 33 mg]dl) was seen in the glucose-injected rats at 15 minutes and continued for up to 4 hours after occlusion (p -- 0.040 and 0.014, respectively; Student's t test). Conclusion: Postischemic hyperglycemia immediately after a standard spinal cord ischemic slxess worsens neurologic outcome. (J VAsc SURG 1993;17:661-8.)
Paraplegia resulting from spinal cord ischemia is a potential complication o f surgical procedures that involve clamping o f the thoracicoabdominal aorta. In a study o f 605 patients who underwent vascular repair ofl~horacicoabdominal aortic aneurysms, it was noted that the overall incidence o f lower extremity neurologic dysfunction was 11%} For the subset o f patients whose operation involved the complex repair of an anemTsm that extended from just distal to the From the Departments of Physiology(Dr. D'Alecy)and Surgery (Drs. Zelenock and D'Alecy), the University of Michigan Medical School, and the Veteran's Administration Medical Center, Ann Arbor. Supported by a Veteran's Administration merit review grant and the Uniw:rsity of Michigan Medical School Committee on Student BiomedicalResearch. Reprint requests: Louis G. D'Alecy, PhD, the University of Michigan Medical School, 7799 Medical Science II, 1301 Catherine St., Ann Arbor, MI 48109-0622. 2411/39830
left subclavian artery to involve most o f the descending thoracic aorta and most or all o f the abdominal aorta, the occurrence o f postoperative paraplegia was 28%. If the aneurysm was dissecting, this percentage increased even further to 40%. Because paraplegia is a c o m m o n and extremely debilitating complication o f thoracicoabdominal aneurysm repair, there has been an active experimental effort to identify potential perioperative protocols, both pharmacologic and mechanistic, that could reduce the risk o f lower extremity neurologic deficit resulting from spinal cord ischemia. 2-n In addition there are other situations in which the delivery o f blood to the spinal cord is temporarily impaired, such as in trauma or profound shock, that have the potential to place the functional integrity o f the spinal cord at risk from ischemic injury. Therefore it would be advantageous if the spinal cord could be protected from ischemic injury by control o f physiologic 661
662 Hemmila,Zelenock, and D'Alecy
variables identified as potential risk factors for the development of paraplegia. Previous work in our laboratory 13,14and by other investigators is has shown that hyperglycemia before and during an aortic cross-clamp/declamp sequence exacerbates paraplegia in both the rat and rabbit spinal cord ischemia models. It has also been demonstrated that pretreatment with insulin, to produce a moderate degree of hypoglycemia, protects from paraplegia in the rat 16 and rabbit ~7 aortic occlusion models. It must be emphasized that the aforementioned studies examined only the effect of preocclusion induction of hyperglycemia on neurologic deficit in spinal cord ischemia models. The possibility that hyperglycemia induced after an ischemic insult could also be detrimental to lower extremity neurologic function has received very little attention but has obvious therapeutic implications. The issue is not resolved. Pulsinelli et al)8 reported that glucose given after global central nervous system (CNS) ischemia produced no greater damage than did a saline injection. Conversely, our laboratory has found that, when brain temperature is controlled, administration of glucose to produce hyperglycemia immediately after global CNS ischemia does indeed worsen neurologic outcome compared with salineinjected controls (unpublished data). 19 The purpose of this study was to determine the effect of hyperglycemia in the postischemic period on lower extremity neurologic function after a standard spinal cord ischemic stress in the rat aortic occlusion model. A secondary objective was to characterize better the relevant time course to determine if there was any significant difference in postischemic lower extremity neurologic deficit with hyperglycemia induced immediately after occlusion compared with that induced 30 minutes after the end of the occlusion period. MATERIAL A N D M E T H O D S
Thirty-six male Sprague-Dawley rats, aged 2 to 3 months and weighing 200 to 350 gm, were housed individually with free access to food and water. Animal care complied with the Principles ofLaboratory
Animal Care and the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication No. 85-23, revised 1985). On the day of surgery the rats were weighed and anesthesia was induced by placing them in a chamber containing 3.5% halothane. After induction, the rats were transferred from the chamber to a mask apparatus. Tracheal intubation was then performed with a neonatal laryngoscope reduced to a blade width of 7 mm to assist in the placement of a 2.5 mm outside
JOURNALOF VASCULARSURGERY April 1993
diameter tube 8 cm long. The rats were then ventilated with a nonrebreathing pressure ventilator (model 212; Harvard Apparatus Co., Inc., S. Natick, Mass.) at 90 cycles/min with 1% to 1.5% halothane and 5 cm H 2 0 positive end expiratory pressure (PEEP). Temperature was monitored continuously with a probe inserted 4.5 cm into the rectum and maintained at 37 ° -+ 0.5 ° C with a homeothermic blanket control unit (model 513; Harvard Apparatus Co., Inc.). The left chest wall was opened just lateral to the sternum from the apex of the manubrium to thesuperior aspect of the third rib. The thymus was excised and the aorta isolated just distal to the left subclavian artery. At this location a 15 cm segment of polyethylene 10 tubing was placed around the aorta, carefully avoiding the left vagus and recurrent laryngeal nerves. The free ends of the tubing were then passed through a 3 cm segment of polyethylene 160 tubing, creating a snare. Snares were placed around the proximal right and left subclavian arteries in a similar manner. A t the beginning of the ischemic period, the three snares were pulled taut and secured, thus occluding each of the isolated vessels. Vessel occlusion was verified by visual inspection of the snare site and the vessel distal to the snare. During the 8-minute occlusion period, the PEEP was increased to 10 cm H 2 0 and maintained at this level. The snares exited the chest superior to the m a n u b r i ~ and the incision was closed. At the end of the occlusion period, defined as time t = 0, the snares were released and withdrawn. Chest closure was completed in three layers with 3-0 silk, PEEP was discontinued, and halothane anesthesia was terminated. Concomitant with snare withdrawal, the rats received an intraperitoneal injection of 5% dextrose in water (2 gm/kg) or an equivalent volume of 0.9% sodium chloride solution. Extubation was performed after the rat demonstrated whisker activity and maintained voluntary ventilation when disconnected from the ventilator. Blood samples for determinafion of blood glucose concentration (0.5 ml) were obtained immediately after induction of anes-~ thesia, before aortic occlusion, immediately after release of the occluding snares, and 15, 30, 45, 60, and 240 minutes after occlusion. Plasma blood glucose concentrations were determined by the glucose oxidase method with an automated glucose analyzer (model 23A; YSI, Yellow Springs, Ohio). A neurologic deficit score was assigned 1, 4, 18, and 24 hours after occlusion to quantify hindlimb motor deficit, based on a 15-point scale used extensively by our laboratory (Table I). 13'16 The sham-operated group received the same
~OURNAL OF VASCULAR SURGERY Volume 17, Number 4
anesthetic, ,operation, and intraperitoneal injection of 5% dextrose in water (2 gm/kg). The only difference between sham and experimental groups was that the snares were pulled taut only momentarily and then released immediately. Thus the sham animals did not sustain aortic occlusion or spinal cord ischemia. In a separate experiment, rats received delayed intraperitoneal injections of 5% dextrose or 0.9% sodium cl~Loride solution 30 minutes after occlusion instead of immediately after occlusion. The experimental protocol used for these rats was otherwise the ~2rne as the one described above except that one additional blood sample was obtained 90 minutes after occlusion. Twenv£-one rats were used in the first experiment: glucose treated (n = 7), saline control (n = 6), and sham (n = 2). Six rats died either during or after surgery of technical complications associated with the experimental preparation. In the second experiment (glucose-treated 30 minutes after occlusion) 15 rats were used: glucose-treated (n = 7) or saline control (n = 7). One rat died after surgery of technical complications. Neurologic deficit scores were compared nonparametrically with the single-tailed Mann-Whitney U test based on previous demonstration of the adverse effect of glucose. 13 Blood glucose concentrations were compared parametrically with the un7fired single-tailed Student's t test. The p values determined by the Mann-Whitney U test are expressed as less than or equal to table values. Exact p values are given for Student's t test. All average values are expressed as means _+ 1 SEM. Data analysis was performed with a StatView 512 + statistical software package on a Macintosh SE personal computer. RESULTS Neurologic deficit scores. The group of rats administered glucose immediately on reperfusion (time t = 0) showed a statistically significant exacerbation of their hindlimb neurologic deficit 24 hours after occlusion compared with saline-treated rats ~p ___ 0.0!51, Mann-Whitney U test). At earlier testing periods (11, 4, and 18 hours after occlusion), no statistically significant difference in neurologic deficit was detected (Fig. 1, top panel). The sham-operated rats had an initial transient postoperative neurologic deficit atl:ributable to the anesthetic and open-chest procedure but had no deficit 24 hours after occlusion. These sham results are similar to those produced previously by the sham procedure in our laboratory for this model by LeMay et al. ~3 In the rats with delayed administration of glucose (injected at t = 30 minutes after occlusion) no statistically significant
Hemmila, Zelenock, and DMlecy 6 6 3
Table I. Rat spinal cord ischemia model neurologic deficit scoring Deficit score Walking with lower extremities 0 N o evidence o f deficit 1 Toes flat under body when walking but ataxia exists 2 Knuckle walks 3 Movement in lower extremities but unable to knuckle walk 4 N o movement, drags lower extremities Horizontal rope 0 Grasps rope and pulls up with lower extremities 1 Raises lower extremities and grasps rope without pulling 2 Raises lower extremities but cannot grasp rope 3 Does not raise lower extremities Screen 0 Lower extremities grasp screen to I80 degrees for more than 5 sec 1 Lower extremities grasp screen to 180 degrees for less than 5 sec 2 Lower extremities grasp screen past vertical but not to 180 degrees 3 Lower extremities fall from screen past vertical (270- to 180-degree range) Bar at 45 degrees (best o f three trials) 0 Lower extremities grasp bar for more than 10 sec 1 Lower extremities grasp bar for more than 5 sec 2 Lower extremities grasp bar for less than 5 sec 3 Lower extremities slide off bar without grasping Pain sensation 0 Strong withdrawal to toe pinch 1 Weak withdrawal to toe pinch 2 N o reaction to toe pinch Maximum deficit = 15 Minimum deficit = 0 Data from LeMay et al. l VAsc SURG 1987;6:383-90; )~Surg Res 1988;44:352-8. Note: For each test a score o f zero indicates normal performance. Increasing test scores correspond to an increase in lower extremity neurologic deficit.
difference was detected in any of the neurologic deficit scores between the glucose-treated and salinetreated rats. Comparing rats administered glucose at time t = 0 (Fig. 1, toppanel) with rats given glucose at time t = 30 minutes (Fig. 1, bottompand) revealed that the earl), postocclusion administration of glucose significantly worsened lower extremity neurologic deficit scores 18 and 24 hours after occlusion (p -< 0.049 18 hours andp _< 0.027 24 hours after occlusion, Mann-Whitney U test). Blood glucose concentrations. Baseline blood glucose concentrations were slightly elevated in both glucose and saline groups (172 + 7 mg/dl and 177 +-6 mg/dl, respectively). In response to the stress associated with the surgical procedure, both groups had further elevation of blood glucose to postocclusion concentrations of 261 + 23 mg/dl in the glucose-treated group and 234 __ 18 mg/dl in the
IOURNALOF VASCULARSURGERY April 1993
664 Hemmila, Zelenock, and D'Alecy
Injection at t = 0 min Postocclusion 15 kg. 111
10
0 m
¢3 0
III Glucose (n=7)
,_1
"1 Saline
(n=6)
0 n-
~] Sham
(n=2)
uJ Z
1
4
18
24
Injection at t = +30 min Postocclusion 15
_P 0 m
M. IJJ ¢2
10
0 |
(3 0
Glucose (n=7)
"1 Saline
..I
(n=7)
0 rr UJ Z 1
4 HOURS
18
24
POSTOCCLUSlON
Fig. 1. Neurologic deficit score 1, 4, 18, and 24 hours after occlusion for glucose- and saline-treated groups injected at time t = 0 and t = 30 minutes after occlusion. For rats injected at time t = 0, glucose-administered group showed significant exacerbation ofneurologic deficit compared with saline-treated group 24 hours after occlusion. *p - 0.051, Mann-Whitney U test. Top, Neurologic deficit score was not obtained 18 hours after occlusion for one saline-treated rat.
saline-treated group when injected at t = 0 minutes (Fig. 2), Blood glucose concentrations further diverged, reaching a maximum at t = 60 minutes after occlusion o f 391 --- 37 mg/dl for glucose-treated and a maximum o f 249 _+ 13 mg/dl for saline-treated rats 15 minutes after occlusion. Significant elevations o f blood glucose concentration were seen at all time points in the glucose-treated rats compared with the saline-treated rats beginning 15 minutes after occlusion and continuing for 4 hours after occlusion
= 0.040 15 minutes andp -- 0.014 4 hours after occlusion, Student's t test). The rats that received delayed administration o f glucose or saline solution (t = 30 minutes after occlusion) had similar increases in baseline blood glucose concentrations (185 _ 9 mg/dl and 174 _+ 4 mg/dl in the glucose- and saline-treated groups, respectively). Postocclusion but preinjecfion blood glucose maximums o f 225 __ 22 mg/dl and 234 __ 18 mg/dl in the glucose- and saline-treated
}'OURNALOF VASCULARSURGERY Volume 17, Number 4
Hemmila, Zelenock, and D'Alecy 665
Injection at t = 0 min Postocclusion 400'
¢3 r~ O O
1
100 IDeolemp + InJeotlon !
Glucose Injection (n=7) Saline Injection (n=6)
z,
4
.J
o
-1
|
|
|
!
i
o
1
2
3
4
Injection at t = +30 min Postocclusion 400.
300'
r~
o
lOO
0
Injection Declamp~/~
Glucose Injection (n=7) Saline Injection (n=7)
&
..J m
o
|
0
!
I
1 2 TIME (hours)
!
i
3
4
Fig. 2. Blood glucose concentrations versus time for glucose- and saline-injected rats. When treated at time t = 0, glucose-injected rats showed significant elevation in blood glucose concentration compared with saline-injected rats starnng 15 minutes after occlusion and continuing to 4 hours after occlusion (p = 0.040 at 15 minutes and p = 0.014 at 4 hours, Student's t test). For rats treated at t = 30 minutes, glucose-injected rats showed significant elevation in blood glucose concentration compared with saline-injected rats starting 45 minutes after occlusion and continuing to 4 hours after occlusion (p = 0.006 45 minutes and p = 0.008 4 hours after occlusion, Student's t test).
rats were: n o t e d at time t = 15 minutes. After the intraperitoneal injection o f glucose and saline solution at time t = 30 minutes after occlusion, b l o o d glucose levels diverged with the glucose group, reaching a m a x i m u m o f 369 + 59 m g / d l at time t = 4 hours after occlusion. T h e saline-treated rats had a slight decline in b l o o d glucose levels f r o m a m a x i m u m o f 233 + 23 m g / d l at time t = 45 minutes to a :minimum o f 181 +_ 20 m g / d l at time t = 4 hours afi:er occlusion. In animals receiving delayed adminisu:ation o f glucose, statistically significant
elevation o f b l o o d glucose concentrations was seen 45 minutes after occlusion and continuing to 4 hours after occlusion c o m p a r e d the saline-treated g r o u p (p = 0.006 45 minutes a n d p = 0.008 4 hours after occlusion, Student's t test). DISCUSSION
Several prior studies in rat and rabbit spinal cord ischemia models indicate that preischemic hyperglycemia significantly worsens neurologic o u t c o m e after t e m p o r a r y aortic occlusion 131s and insulin
666 Hemmila, Zelenock, and D'Alecy
pretreatment; to produce a moderate degree of hypoglycemia has been demonstrated to protect against paraplegia. 16'17 Despite its obvious therapeutic importance, postischemic hyperglycemia and the issue of glucose administration during the postischemic period has received little attention. It must be emphasized that all prior studies in spinal ischemia examined only the effect of preocclusion physiologic variables on lower extremity neurologic deficit. This study demonstrated for the first time that postocclusion hyperglycemia, caused by treatment with exogenous glucose, can produce a significant exacerbation of hindlimb neurologic deficit, specifically paraplegia, in the rat spinal cord ischemia model. This suggests that at least some of the neurologic damage resulting from spinal cord ischemia must occur during the reperfusion period and postocclusion hyperglycemia worsens its outcome. It also rather precisely defined the period of increased vulnerability; if administration of exogenous glucose was delayed until 30 minutes after occlusion, the resulting hyperglycemia did not significantly increase the degree of paraplegia compared with salineinjected controls. The mechanism by which postischemic hyperglycemia exacerbates spinal cord ischemic injury caused by aortic clamping is not completely clear. Oxygen radical mechanisms of brain injury after ischemia and reperfusion have been reviewed recently, 2° and by inference these mechanisms may be operable in other settings of CNS injury such as spinal cord ischemia. Emerging evidence supports the theory that a variety of pathologic events are generated within spinal cord tissue during the ischemic period that create a biochemical environment whereby additional cellular damage is produced after reperfusion. In experimental models, alterations in antioxidant enzyme systems within ischemic tissues may predispose the cells to the formation of reactive oxygen metabolites on reperfusion. 21 Because hyperglycemia in the early reperfusion period produced an increase in neurologic damage to the spinal cord, it suggests either a mechanistic linkage between abnormally high blood glucose concentration and reperfusion oxygen radical injury or mechanistically discrete but synergistic adverse affects. The importance of the temporal sequence is emphasized by the lack of adverse effect of hyperglycemia when delayed for 30 minutes. Other physical parameters of reperfusion may also influence outcomes. Rapid reperfusion and reoxygenation of ischemic tissue has been noted to have a
JOURNALOF VASCULARSURGERY April 1993
significantly worse effect on cells compared with gradual establishment of reperfusion. An experiment by Danielisova et al.s showed that a decreased partial pressure of oxygen in arterial blood during reperfusion can have a beneficial effect after an ischemic reperfusion sequence in the spinal cord. It appears that a gradual return to aerobic glycolysis after an ishemic stress may be more effective in preventing reperfusion neuronal injury than the approach of maximal reoxygenation coupled with hyperglycemia. It has long been known that ischemia causes shift in cellular metabolism from aerobic to anaerobic glycolysis. Recent studies by Robertson and Grossman 17have shown that after reperfusion there is a lag in the return to aerobic metabolism. This delay can lead to an increase in lactic acidosis during reperfusion as a result of the increased availability of glucose as a substrate for anaerobic glycolysis. From experiments in cardiac resuscitation in dogs it is known that elevated brain lactate accumulation and increased neurologic deficit are associated with modest hyperglycemia in the setting of global CNS ischemia. 22 These results suggest lactate accumulation as a possible mechanism to explain worsened neurologic outcome after spinal cord ischemia in the rat with hyperglycemia. Based on this and our prior studies, 13'16 we Can conclude that there are two critical phases wh¢.~,, hyperglycemia can significantly affect the degree ot' injury produced by an episode of spinal cord ischemia. The most critical phase is before and during the actual ischemic period. Hyperglycemia before the ischemic period produces a maximal exacerbation of paraplegia in the rat spinal cord ischemia model. The second phase during which hyperglycemia worsens neurologic outcome is the immediate reperfusion period after an ischemic episode. Hyperglycemia produced immediately after occlusion significantly exacerbates the incidence and severity of paraplegia in the rat aortic occlusion model. However, the detrimental effect of exogenous glucose administration on neurologic outcome is quite circumscribed. When, glucose is given 30 minutes after reperfusion, there is no incremental effect. Although pretreatment to avoid hyperglycemia is practical and relevant in elective operative procedures involving aortic cross-clamping, not all clinical cases with the potential for ischemic spinal cord injury present as a pretreatment situation. The avoidance of induced hyperglycemia after an ischemic event involving the spinal cord is also readily possible. 23 Even
JOURNAL OF VASCULARSURGERY Volume 17, Number 4
while recognizing the concern that p e r m a n e n t neurologic & a n a g e m a y result f r o m the delayed recognition and t r e a t m e n t o f p r o f o u n d hypoglycemia, it m u s t be t m d e r s t o o d that euglycemia or even mildm o d e r a t e hypoglycemia is n o t associated with injury and m a y even be protective. T h e ready availability and capacity for rapid assessment o f b l o o d glucose concentration in the e m e r g e n c y r o o m , the operating r o o m , and at the bedside m a k e it possible to maintain patients' b l o o d glucose concentrations in the euglycemic range w h e n e v e r there is the risk o f spinal cord .dchemia. A review by Sieber et al.24 reevaluated the use o f intraoperative glucose in neurosurgery and concluded that intraoperative glucose administration appears to have m i n i m a l beneficial effects. T h e only instance in w h i c h hyperglycemia has been associated with a possible decreased a m o u n t o f neuronal injury is in the setting o f focal cerebral ischemia. 2s'26 These t w o studies d e m o n s t r a t e d that hyperglycemia during the focal ischemic period resulted in less cortical damage, by histologic assessment, c o m p a r e d with saline-treated controls. N o functional o u t c o m e s were reported. I t m u s t be emphasized that m o r p h o l o g i c data and timctional data in this setting have yet to be positively correlated. W i t h the one a f o r e m e n t i o n e d exception, hyperglycemia has been d e m o n s t r a t e d to have little if "3Y bene.ficial effect in the setting o f neuronal ischemia. T h e evidence continues to collect that, in spinal cord ischemia, hyperglycemia has adverse consequences w h e n present either before or after the ischemic event. Therefore glucose-containing intravenous fluids should be used only w h e n there is a clinically compelling justification for supplemental glucose. We thank Rakesh Patel for his assistance in the start-up of this experiment.
REFERENCES 1. Crawford ES, Crawford JL, Hazim JS, 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. 2. &gee JM, Flanagan T, Blackbourne LH, Kron IL, Tribble CG. Reducing postischemic paraplegia using conjugated superoxide dismutase. Ann Thorac Surg 1991;51:911-4. 3. Barsan WG, Hedges JR, Syverud SA, Dronen SC, Dimlich RV. Effects of dichloroacetate in spinal stroke in the rabbit. Life Sci 1987;41:1065-9. 4. Breckwoldt WL, Genco CM, Connolly RJ, Cleveland RJ, Diehl JT. Spinal cord protection during aortic occlusion:
Hemmila, Zelenock, and D'Alecy 667 efficacy of intrathecal tetracaine. Ann Thorac Surg 1991;51: 959-63. 5. Danielisova V, Marsala M, Chavko M, Marsala J. Postischemic hypoxia improves metabolic and functional recovery of the spinal cord. Neurology 1990;40:1125-9. 6. Dasmahapatra HK, Coles JG, Wilson GI, et al. Relationship between cerebrospinal fluid dynamics and reversible spinal cord ischemia during experimental thoracic occlusion. J Thorac Cardiovasc Surg 1988;95:920-3. 7. DelRossi AJ, Cernaianu AC, Cilley Jt-I, et al. Preventive effect of Fluosol-DA for paraplegia encountered after surgical treatment of the thoracic aorta: preliminary results in a dog model. J Thorac Cardiovasc Surg 1990;99:665-9. 8. I-Iall ED. Effects of the 21-aminosteroid U74006F on posttraumatic spinal cord ischemia in cats. J Neurosurg 1988;68:462-5. 9. Kochhar A, Zivin JA, Lyden PD, Mazzarella V. Glutamate antagonist therapy reduces neurologic deficits produced by focal central nervous system ischemia. Arch Neuro11988;45: 148-53. 10. LeMay DR, Zetenock GB, D'Alecy LG. Neurological protection by dichloroacetate depending on the severity of injury in the paraplegic rat. J Neurosurg 1990;73:118-22. 11. Wadouh F, Wadouh R, Hartmann M, Crisp-Lindgren N. Prevention of paraplegia during aortic operations. Ann Thorac Surg 1990;50:543-52. 12. Yum SW, Faden AI. Comparison of the ueuroprotective effects of the N-methyl-D-aspartate antagonist MK-801 and the opiate-receptor antagonist nalmefene in experimental spinal cord ischemia. Arch Neurol 1990;47:277-81. 13. LeMay DR, Neal S, Neal S, Zelenock GB, D'Alecy LG. Paraplegia in the rat induced by aortic cross clamping: model characterization and glucose exacerbation of neurological deficit. J VASCSURG 1987;6:383-90. 14. Lundy EF, Ball TD, MandeUMA, Zelenock GB, D'Alec3TLG. Dextrose administration increases sensory/motor impairment and paraplegia following infrarenal aortic occlusion in the rabbit. Arch Surg 1987;102:737-42. 15. Drummond JC, Moore SS. The li~luence of dextrose administration on neurologic outcome after temporary spinal cord ischemia in the rabbit. Anesthesiology 1989;70:64-70. 16. LeMay DR, Lu AC, Zelenock GB, D'Atecy LG. Insulin administration protects from paraplegia in the rat aortic occlusion model. J Surg Res 1988;44:352-8. 17. Robertson CS, Grossman RG. Protection against spinal cord ischemia with insulin-induced hypoglycemia. J Neurosurg 1987;67:739-44. 18. Pulsinelli WA, Waldman S, Rawlinson D, Plum F. Moderate hyperglycemia augments ischemic brain damage: aneuropathologic study in the rat. Neurology 1982;32:1239-46. 19. Lundy EF, Kuhn JE~ Kwon JM, Zelenock GB, D'Alecy LG. Infusion of five percent dextrose increases mortality and morbidity following six minutes of cardiac arrest in resuscitated dogs. J Crit Care 1987;2:4-14. 20. Traystman RJ, Kirsch JR, Koehler RC. Oxygen radical mechanisms of brain injury following ischemia and reperfusion. J AppI Physiol 1991;71:1185-95. 21. Fantone JC. Pathogenesis of ischemia-reperfusion injury: an overview. In: Zelenock GB, ed. Clinical ischemic syndromes: mechanisms and consequences of tissue injury. 1st ed. St Louis: Mosby-Year Book, 1990:141-2. 22. Natale JE, Stante SM, D'Alecy LG. Elevated brain lactate
668
Hemmila, Zelenock, and D'Alecy
accumulation and increased neurologic deficit are associated with modest hyperglycemia in global brain ischemia, lq,esuscitation 1990;19:271-89. 23. Browning RG, Olson DW, Stueven HA, Mateer JR. Fifty percent dextrose: antidote or toxin? Ann Emerg Med 1990; 19:683-7. 24. Sieber FE, Smith DS, Traystman RJ, Wollman H. Glucose: a reevaluation of its intraoperative use. Anesthesiology 1987;67:72-81.
IOURNAL OF VASCULARSURGERY April 1993
25. Kraft SA, Larson P, Shuer LM, et al. Effect of hyperglycemia on neuronal changes in a rabbit model of focal cerebral ischemia. Stroke 1990;21:447-50. 26. Zasslow MA, pearl RG, Shuer LM, et al. Hyperglycemia decreases acute neuronal ischemic changes after middle cerebral artery occlusion in cats. Stroke 1989;20:519-23.
Submitted Jan. 21, 1992; accepted May 28, 1992.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the JOURNAL Or VASCU~R SURGERY for 1993 are available to subscribers only. They may be purchased from the publisher at a cost of $69.00 for domestic, $89.83 for Canadian, and $85.00 for international subscribers for Vol. 17 (January to June) and Vol. 18 (July to December). Price includes shipping charges. Each bound volume contains a subject and author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Subscription Services, Mosby, 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, USA. In the United States call toll free (800)325-4177, ext. 4351. In Missouri or foreign countries call (314)453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.