| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the Department of Anesthesiology, Yamaguchi University Graduate School of Medicine, Yamaguchi, Japan.
Address correspondence and reprint requests to Mishiya Matsumoto, MD, Department of Anesthesiology, Yamaguchi University Graduate School of Medicine, 1-1-1 Minami-Kogushi, Ube, Yamaguchi 755-8505, Japan. Address e-mail to mishiya{at}yamaguchi-u.ac.jp.
Abstract
BACKGROUND: It is not well established whether insulin protects against ischemic spinal cord injury. We examined the effects of a single dose of insulin that corrects mild hyperglycemia on the outcome after transient spinal cord ischemia in rabbits.
METHODS: We assigned rabbits to four groups (n = 8 in each); untreated control (C) group, preischemic insulin (Pre-I) group, preischemic insulin with glucose (GI) group (glucose concentrations were maintained at levels similar to the C group by the administration of glucose), and postischemic insulin (Post-I) group. Insulin (0.5 IU/kg) was administered 30 min before ischemia in the Pre-I and GI groups, and just after reperfusion in the Post-I group. Spinal cord ischemia was produced by occluding the abdominal aorta for 13 min. Neurologic and histopathologic evaluations were performed 7 days after ischemia.
RESULTS: The mean blood glucose concentration before ischemia in the Pre-I group (118 mg/dL) was significantly lower than in the other three groups (158–180 mg/dL) and those of 30 min after reperfusion in the Pre-I (92 mg/dL) and Post-I (100 mg/dL) groups were significantly lower than in the C (148 mg/dL) and GI (140 mg/dL) groups. The motor function score and number of normal neurons in the anterior lumbar spinal cord in the Pre-I group were significantly greater than in the other three groups.
CONCLUSIONS: These results suggest that a relatively small dose of preischemic insulin protects against ischemic spinal cord injury, and that the protective effects result from tight glycemic control before ischemia.
One of the most serious complications of thoracoabdominal aortic surgery is paraplegia, which is thought to result from spinal cord ischemia during aortic occlusion. Although animal experiments have shown several protective measures against ischemic spinal cord injury (1), no single strategy has been widely accepted for use in the clinical setting.
A recent clinical trial of intensive insulin therapy in critically ill patients (2,3) has aroused interest in brain protection in acute stroke (4). However, the idea of brain protection by insulin is not new. Indeed, many investigations have demonstrated that hyperglycemia aggravates ischemic brain injury (5), and that ischemic brain injury is generally reduced by insulin (6). Insulin has been reported to reduce ischemic brain injury by two mechanisms. One is that insulin acts indirectly by controlling blood glucose concentrations. The indirect effect seems to be the principal mechanism of protection in focal cerebral ischemia (6,7). The other is that insulin interacts directly with the brain. The direct effect seems to contribute, at least in part, to the protection in global cerebral ischemia (6,8). Interestingly, recent investigations have highlighted the direct protective effect of insulin (4,9,10).
In the spinal cord, as in the brain, hyperglycemia has been reported to aggravate ischemic injury (11,12). Concerning the efficacy of insulin on the ischemic spinal cord, the results are inconsistent. Preischemic insulin was shown to be effective in two studies (13,14) and ineffective in another (15). Studies reporting the protective effect of insulin have not determined whether this effect can be attributed to the direct or indirect (secondary glucose-decreasing) effect (13,14).
Insulin has been widely used clinically to control blood glucose concentrations. If a clinically relevant dose of insulin is proven protective against ischemic spinal cord injury in animals, insulin could provide an important strategy for spinal cord protection in thoracoabdominal aortic aneurysmal surgery. For the possible clinical therapeutic application of insulin, we raised the following three questions: 1) Is a relatively small dose of preischemic insulin (0.5 IU/kg) protective? 2) Does insulin, given together with glucose, affect the outcome? 3) Is postischemic insulin protective? We designed the present study to answer those questions using a rabbit spinal cord ischemia model.
METHODS
The protocol of this study was approved by the Ethics Committee for Animal Experiment at Yamaguchi University Graduate School of Medicine. Thirty-two New Zealand white rabbits weighing 2.3 ± 0.1 kg (mean ± sd) were used in this study.
After an overnight fast with unrestricted access to water, rabbits were anesthetized in a plastic box with 5% sevoflurane in oxygen. A catheter was inserted in an ear vein for administration of fluid (lactated Ringer's solution 4 mL · kg–1 · h–1) and drugs, and pentobarbital (30 mg) was administered to facilitate tracheal intubation. After placing a 3-mm cuffed endotracheal tube, the inspired gas mixture was changed to isoflurane 2%–3% in 40% oxygen/60% nitrogen, and the rabbits' lungs were mechanically ventilated. Temperatures were monitored with a calibrated esophageal thermistor (Model MG-Type 209; Nihon Koden, Tokyo, Japan) and a needle-type thermistor (Model PTC-201; Unique Medical, Tokyo, Japan) inserted into the paravertebral muscle at the level of L4–5. The paravertebral muscle temperature was controlled throughout the study at approximately 38.0°C with a heating lamp and warming pad. PE-60 catheters were inserted into both femoral arteries to measure arterial blood pressure proximal and distal to the level of the aortic occlusion. The right-side catheter was advanced 3 cm into the abdominal aorta, and the left one was advanced 17 cm.
Spinal cord ischemia was produced as previously reported (16,17). In brief, in the right lateral decubitus position, the abdominal aorta was exposed retroperitoneally at the level of the left renal artery. A PE-60 catheter was placed around the aorta immediately distal to the left renal artery for later occlusion of the aorta. Then, an occluder tube (16F rubber tube) was tunneled to the skin. After completion of surgery, end-tidal isoflurane concentration was maintained at 2%.
Rabbits were randomly assigned to one of the following groups (n = 8 in each): an untreated control group (C group), a preischemic insulin group (Pre-I group), a preischemic insulin with glucose (GI) group, or a postischemic insulin group (Post-I group). In the C group, saline (1 mL) was administered IV 30 min before aortic occlusion and saline infusion (25 mL/h) was started simultaneously for 60 min. Saline (1 mL) was again administered just after recirculation. Instead of saline, 1 mL of insulin solution (0.5 IU/kg, dissolved in saline) was administered IV 30 min before aortic occlusion in the Pre-I group and just after reperfusion in the Post-I group. Other regimens in the Pre-I and Post-I groups were the same as those in the C group. In the GI group, 1 mL of insulin solution (0.5 IU/kg, dissolved in saline) was administered IV 30 min before aortic occlusion and glucose infusion (5% solution, approximately 25 mL/h) was started simultaneously for 60 min. Saline (1 mL) was administered just after recirculation. The dose of glucose was adjusted to maintain the blood glucose concentration comparable to that in the C group.
Segmental spinal cord evoked potentials (SSCEPs) were monitored, stimulating the left sciatic nerve with square-wave pulses of 0.1 ms duration and 0.6 mA intensity delivered at 3 Hz, and recording in a bipolar fashion (L5 and L6) with silver needle electrodes. SSCEPs were recorded every 1 min for 7 min after aortic occlusion, and then every 2 min until 15 min after reperfusion. The typical recording of SSCEPs demonstrated two positive waves and four negative waves (N1–N4; N1, N2: presynaptic components, N3, N4: postsynaptic components). We measured the amplitude of N3. Heparin 400 U was administered immediately before aortic occlusion. Ischemia was induced by pulling the PE catheter placed around the aorta and clamping an occluder tube for 13 min. We verified the severity of ischemia by the time of disappearance of N3 wave (comparable time periods in all groups, see Results) and distal arterial blood pressure during clamping. After reperfusion, phenylephrine was administered IV to maintain 60 mm Hg of mean distal arterial blood pressure for preventing hypotension.
After the final recording of SSCEPs, all catheters and electrodes were removed, and then all incisions were sutured. Isoflurane was discontinued and the lungs were mechanically ventilated with 100% oxygen. Extubation of the trachea was performed when vigorous spontaneous ventilation and movement occurred. The animals were allowed to recover in a warmed plastic box that contained supplemental oxygen until the conscious animal appeared alert. IV fluid was provided until the animals began to drink. Antibiotic (cephazolin 30 mg/kg, IM) was administered once daily for 3 days, and bladder contents were expressed manually as required.
In all groups, we measured blood glucose concentrations at five points: before administration of insulin or vehicle (saline), just before ischemia, 30 min, 3 h, and 6 h after reperfusion. To sample the blood for glucose measurement at 3 and 6 h after reperfusion, the rabbits were briefly anesthetized with 5% sevoflurane with a nonsealing facemask. The rabbits were fasted until 6 h after ischemia. The person performing the neurologic and histopathologic assessments was not aware of the treatment group on any given animal.
The rabbits were neurologically assessed at 12 h and then daily for 7 days after reperfusion by an observer unaware of the treatment group using the five-point score system proposed by Drummond and Moore (12): 4 = normal motor function, 3 = ability to draw legs under body and hop, but not normally, 2 = some lower-extremity function with good antigravity strength but inability to draw legs under body and/or hop, 1 = poor lower-extremity function but weak antigravity movement only, 0 = paraplegic with no lower-extremity function.
After the final neurologic assessment (7 days after reperfusion), the rabbits were reanesthetized with 2% isoflurane in oxygen. Transcardiac perfusion and fixation were performed with 10% phosphate-buffered formalin. Coronal sections of the spinal cord at the level of L5 were cut at a thickness of 8 µm, and stained with hematoxylin and eosin. Normal neurons in the anterior spinal cord (anterior to a line drawn through the central canal perpendicular to the vertical axis) were counted in two sections for each rabbit and averaged. Ischemic neurons were identified by cytoplasmic eosinophilia with loss of Nissl substances and the presence of pyknotic homogenous nuclei.
For statistical evaluation, physiologic variables were analyzed by a repeated-measures analysis of variance followed by factorial analysis of variance. Where differences were identified, Scheffé post hoc test for intergroup comparisons was performed. The time for N3 wave of SSCEPs to disappear after aortic occlusion or to appear after reperfusion were analyzed by a factorial analysis of variance. Hindlimb motor function and the number of normal neurons in the anterior spinal cord were analyzed with a nonparametric method (Kruskal–Wallis test followed by the Mann–Whitney U-test). P < 0.05 was considered statistically significant.
RESULTS
Blood glucose concentrations are shown in Table 1. At the baseline, they were mildly to moderately hyperglycemic, probably due to the stress of surgical preparation and anesthesia, but there were no significant differences among the four groups. The blood glucose concentration just before aortic occlusion in the Pre-I group was normal and significantly lower than in the other three groups, including the C group. At 30 min after reperfusion, the blood glucose concentrations in the GI and C groups were significantly higher than in the Pre-I and Post-I groups. At 3 h after reperfusion, the blood glucose concentration in the Post-I group was normal and significantly lower than in the C group.
|
In other physiologic variables, there were no significant differences among the four groups except for slightly increased heart rate in the Pre-I group, which was within physiologic range (Table 2).
|
In all groups, the time required for N3 wave of SSCEPs to disappear was 4–5 min, whereas the time required for N3 wave to appear in each group was 10 ± 3, 7 ± 2, 10 ± 4, and 10 ± 3 min in the C, Pre-I, GI, and Post-I groups, respectively. There were no significant differences in those times among the four groups.
All rabbits survived until the final neurologic assessment (7 days after reperfusion). The time course of changes in motor function score in each group is shown in Figure 1. The C group showed variable degrees of motor dysfunction and seven of eight rabbits showed score 2 to 0 at 7 days after reperfusion. The scores in the Pre-I group were significantly better than in the other three groups from 2 to 7 days after reperfusion. There were no significant differences in motor function score among the C, GI, and Post-I groups at any time point.
|
Figure 2 shows microphotographs of the lumbar spinal cord in each group. In the animals of the C, GI, and Post-I groups, the structure of the spinal cord gray matter was destroyed, most motor neurons disappeared, and the prominent inflammatory cell infiltration was observed. Vacuolation in the white matter was also observed in those three groups. In contrast, the structure of the gray and white matter of the spinal cord was well maintained and motor neurons preserved normal appearance in the Pre-I group (Fig. 2). The number of morphologically normal appearing neurons 7 days after reperfusion is shown in Figure 3. There were significantly more normal appearing neurons in the Pre-I group than in the other three groups.
|
|
DISCUSSION
In the present study, we demonstrated that a relatively small dose of preischemic, but not postischemic, insulin that corrects mild to moderate hyperglycemia protected against ischemic spinal cord injury, and that the protective effect of insulin was easily cancelled by a concomitant glucose infusion that maintained blood glucose concentrations comparable to those of the untreated control group (mild hyperglycemic).
We used a well-characterized animal model of spinal cord ischemia in which the outcome after various durations of ischemia (13, 15, and 20 min) (16,18,19) and the influence of changes in temperature (32°C, 35°C, and 38°C) (16) have been described. To verify the proper ischemic insult, we monitored SSCEPs and arterial blood pressures proximal and distal to the aortic occlusion. We selected 13 min ischemia in the present study, because a longer duration of ischemia has been known to cause severe acute onset of motor dysfunction, which may be impossible to treat.
Our results are consistent with those of Robertson and Grossman (13) and LeMay et al. (14), despite the large differences in the doses of insulin. Robertson and Grossman (13) demonstrated a better recovery in SSCEPs 2 h after reperfusion in the rabbits treated with preischemic insulin (8–13 IU/kg). LeMay et al. (14) showed that preischemic insulin (2–2.5 IU/kg) improved neurologic function 24 h after reperfusion in rats. Although a detailed comparison among the studies by Robertson and Grossman, LeMay et al. and ours is difficult because of their short observation period and no histopathologic examination, the doses of insulin that correct mild to moderate hyperglycemia may have enough power to protect against ischemic spinal cord injury. Contrary results reported by Nakao et al. (15) deserve consideration. They examined the effects of insulin-like growth factor 1 (0.3 mg/kg) and insulin (0.3 IU/kg), at equivalent doses to decrease plasma glucose concentrations (approximately 55 and 50 mg/dL, respectively), on neurologic function and histopathology 48 h after transient spinal cord ischemia in rabbits (15). Insulin-like growth factor 1, but not insulin, exhibited protective effects on neurologic function and histopathology (15). The animal model and the insulin dose used by Nakao et al. (15) were similar to those in our study. However, two factors may have caused different results. First, the glucose concentration induced by insulin was so low (approximately 50 mg/dL) that the protective effect may have been undetected. This is likely because hypoglycemia <3 mM (54 mg/dL) has been shown to cancel the protective effects of insulin in brain ischemia (20). Second, the longer duration of aortic occlusion (15 min) in Nakao et al.'s study (15) than ours (13 min) might have made the protective effect of insulin undetectable.
The important finding in the present study is that the protective effect of insulin given before ischemia was cancelled by a concomitant glucose infusion that maintained blood glucose concentrations (mildly hyperglycemic) comparable to those of the untreated control group. The results indicate that the protection against ischemic spinal cord injury by the relatively small dose of insulin results from tight glycemic control before ischemia. Our results are consistent with the investigation in transient focal cerebral ischemia, where preischemic insulin (2–3 IU/kg), but not preischemic GI, significantly reduced ischemic damage (7). Given that glycemic control is the principal mechanism of protection by insulin, it may be reasonable that postischemic insulin showed no ameliorating effect in the present study. Because postischemic hyperglycemia has been reported not to aggravate ischemic brain injury (21), the glycemic control by insulin in the postischemic period may not be expected to provide any beneficial effect. Nevertheless, in a rat forebrain ischemia model, postischemic insulin (7 IU/kg) was reported to attenuate ischemic brain damage (8). Moreover, the protective effect of postischemic insulin was still observed when blood glucose concentrations were maintained at the level of the control group (mean preischemic glucose concentration, 8.1 mM) by glucose administration (8). From these results, the authors concluded that the attenuation of ischemic brain damage by insulin is independent of its hypoglycemic effect, and that the neuroprotective mechanism involves insulin's direct action on the brain (8). The different results may be related to the regimen of insulin. Indeed, in Voll and Auer's study (8), a large dose of insulin (7 IU/kg) was administered immediately or about 30 min after reperfusion and during the first 3 days of recovery (7 IU/kg every 12 h). In contrast, we administered a relatively small single dose of insulin (0.5 IU/kg) just after reperfusion. Large and repeated doses may be needed if the treatment with insulin starts after an ischemic episode.
Besides decreasing blood glucose concentrations, insulin has several inherent actions including antiinflammatory, antioxidant, and vasodilatory effects (4). Insulin was also reported to up-regulate Bcl-xL and down-regulate Bax in the motor neurons in the spinal cord after transient ischemia, although those effects were weak compared with those of insulin-like growth factor 1 (15). In a model of oxygen-glucose deprivation in cultured cortical neurons, insulin was reported to provide neuroprotective effects by counteracting the decrease in cell surface 
-aminobutyric acid type A receptors (10). However, it is not known how intensely those mechanisms contribute to the protection in the brain and spinal cord. Further studies may be needed to clarify this issue. Nonetheless, because it is difficult to administer large doses of insulin in the clinical setting because of the possibility of severe hypoglycemia, we think that it is practical to start the tight glycemic control with a relatively small dose of insulin before spinal cord ischemia and to maintain it in the postischemic period.
In conclusion, the present study suggests that tight glycemic control by insulin started in the preischemic, but not postischemic, period improves the neurologic and histopathologic outcome after transient spinal cord ischemia. Although the involvement of a neuroprotective mechanism other than the glucose-decreasing effect may need to be further examined, our results suggest that glycemic control is the principal mechanism for the neuroprotective effect of a relatively small dose of insulin in our experimental setting.
Footnotes
Accepted for publication July 5, 2007.
Supported, in part, by the Grant-in-Aid for Scientific Research (Japan Society for the Promotion of Science) No. 17591633.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
