Excessive accumulation of glutamate and the subsequent increase in intracellular calcium are important causes of ischemic neuronal injury. There have been extensive studies that examined the effects of various drugs, including glutamate receptor antagonists (1–3) and calcium entry blockers (4,5), against ischemic spinal cord injury. However, no drugs have been established as being protective in the clinical setting. Because magnesium blocks the Ca2+ channel in the N-methyl-d-aspartate (NMDA) receptor and acts as a physiologic Ca2+ blocker (6), it might be expected to provide beneficial effects.
IV administration of magnesium has been used for eclampsia for many years and has been effective in protecting the brain (7) and spinal cord (8) against ischemic injuries in animal experiments. However, the problem is its poor passage across the blood-brain (spinal cord) barrier. If a large dose of magnesium is administered IV to increase its concentration in the brain and spinal cord, systemic hypotension and bradycardia will become prominent. Therefore, the intrathecal administration of magnesium may be preferable. Simpson et al. (9) and Follis et al. (10) have demonstrated protective effects of magnesium intrathecally given against ischemic spinal cord injury in dogs and rats, respectively. However, in their studies, only a single dose was tested. To accept intrathecal magnesium as a protective intervention in the clinical setting, it is required to investigate its safety at variable dose ranges and then to determine an optimal dose. Chanimov et al. (11) have systematically investigated the neurotoxicity of intrathecal magnesium in the spinal cord in rats and have concluded that the injection of magnesium sulfate in an iso-osmolar concentration of 6.3% (4.6 mg/kg) is safe (11). However, there is no study that systematically investigated the safety of intrathecal magnesium in other species.
In the present study, we sought to investigate the effects of various doses of magnesium sulfate intrathecally administered on the neurologic function and histopathologic changes of the spinal cord in detail and to determine its optimal dose for protection, if any, against ischemic spinal cord injury in rabbits.
The protocol of this study was approved by the Ethics Committee for Animal Experiments at Yamaguchi University School of Medicine. Fifty-seven New Zealand white rabbits weighing 2.9 ± 0.4 kg (mean ± sd) were used in this study. The methods for surgical preparation, implanting an intrathecal catheter, producing spinal cord ischemia, postischemic management, and neurologic and histopathologic evaluation were similar to those used in our previous studies (12–15).
Anesthesia was induced with 5% sevoflurane in oxygen. An ear-vein catheter was inserted for administration of fluid and drugs. Intubation of the trachea was performed, and lungs were mechanically ventilated with 2% isoflurane in 40% oxygen/60% nitrogen. A PE-10 catheter, for administration of magnesium sulfate or saline, was intrathecally implanted through the L5–6 interlamina space so that the tip of the catheter was located at the level of the cauda equina. Two to 3 days after implantation, rabbits that showed no sign of neurologic deficit were used for the following three sets of the studies.
Rabbits were randomly assigned to one of the following groups: magnesium sulfate 0.3 mg/kg (n = 6), 1 mg/kg (n = 6), 2 mg/kg (n = 6), 3 mg/kg (n = 6), and the saline (control) (n = 6). Magnesium solutions were prepared by diluting 1 M of the magnesium sulfate solution (Otsuka Pharmaceutical Co, Ltd, Tokyo, Japan) with distilled water and saline. The osmolarity range of the magnesium sulfate solutions administered in the present study was 278–295 mOsm/L (Osmostat OM-6040; Arkray, Shiga, Japan). The rabbits intrathecally received 0.3 mL of magnesium sulfate solutions (4 different doses) or saline under general anesthesia maintained by 5% inhaled sevoflurane with a nonsealing face mask device.
The rabbits were neurologically assessed daily until 1 wk after the administration by an observer unaware of the treatment groups. Hindlimb motor function was assessed by a 5-point grading scale: 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; 1 = poor lower extremity motor function and weak antigravity movement only; and 0 = paraplegic with no lower extremity motor function (16). Sensory function was evaluated by seeking an aversive response to pinprick stimulation with a 23-gauge needle, progressing from sacral to thoracic dermatomes. The score of the sensory function was assessed by the following 3-point grading scale: 2 = normal; 1 = a region with diminished response is present; and 0 = a region with no response is present (15).
After completion of neurologic function scoring at 1 wk, the rabbits were reanesthetized, and transcardiac perfusion and fixation of the spinal cord were performed by 10% phosphate-buffered formalin. Coronal sections of the spinal cord (8 μm) at the L5 level were stained with hematoxylin and eosin. Morphological changes of the spinal cord were assessed at a magnification of 100–400× by an observer unaware of the treatment groups. The degree of the vacuolation of the dorsal funiculus was assessed with a 4-point grading scale: 0 = no vacuolation; 1 = <10% area of the dorsal funiculus vacuolated; 2 = 10%–50% area of the dorsal funiculus vacuolated; and 3 = >50% area of the dorsal funiculus vacuolated (15).
We investigated the temporal profile of the neurotoxicity of intrathecal magnesium sulfate. Rabbits were randomly assigned to one of the following groups: 6-h interval between the administration of magnesium and perfusion-fixation of the spinal cord (6-h group; n = 3), 2-day interval (2-day group; n = 3), and 4-day interval (4-day group; n = 3). All 3 groups intrathecally received 3 mg/kg of magnesium sulfate solution (0.3 mL). After completion of neurologic function scoring at the prescheduled time point, the lumbar spinal cord was histopathologically assessed, as described above.
The neuroprotective effects of 0.3 mg/kg and 1 mg/kg of magnesium sulfate were tested, avoiding magnesium sulfate 2 mg/kg and 3 mg/kg because these two larger doses showed apparent neurotoxicity in Experiments 1A and 1B. Rabbits received 0.3 mg/kg of magnesium sulfate, 1 mg/kg of magnesium sulfate, or saline intrathecally 20 min before ischemia (n = 6 per group).
Anesthesia was induced and maintained in the same manner for the intrathecal catheter implantation. Paravertebral muscle temperature measured by a needle-type thermistor (Model PTC-201; Unique Medical, Tokyo, Japan) was controlled to 38.0°C with a heating lamp and warming pad throughout the study. PE-60 catheters were inserted into both femoral arteries to measure arterial blood pressure above and below the aortic occlusion. The right-side catheter was advanced 3 cm into the abdominal aorta, whereas the other was advanced 17 cm.
In the right lateral position, the retroperitoneal abdominal aorta was exposed at the level of 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 (16-French rubber tube) was tunneled to the skin.
After completion of surgery, end-tidal isoflurane concentration was maintained at 2%. Rabbits received either 0.3 mL of magnesium solutions (0.3 mg/kg or 1 mg/kg) or saline 20 min before aortic occlusion. Heparin 400 U was administered IV immediately before aortic occlusion. Ischemia was induced by pulling the PE catheter and clamping an occluder tube for 15 min.
After all incisions were sutured, tracheal extubation was performed. At 24, 48, 72, and 96 h after recirculation, hindlimb motor function was assessed by an observer unaware of the treatment groups, as described in Experiment 1A.
After completion of motor function scoring at 96 h, histopathologic assessment (hematoxylin and eosin staining) of the L5 level of the spinal cord was performed. 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 by the presence of pyknotic homogenous nuclei.
Physiological variables were analyzed by a repeated-measures analysis of variance. Hindlimb motor function, cutaneous sensation, vacuolation of the dorsal funiculus, 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.
All rabbits survived until the prescheduled time period for neurologic assessment and other evaluations.
At first, we planned to administer magnesium sulfate in awake rabbits already implanted with an intrathecal catheter. However, the intrathecal administration of large doses of magnesium sulfate (2–3 mg/kg) gave the rabbit unpleasant stimuli. Therefore, we administered magnesium solutions under a brief interval of general anesthesia.
At 1 h after the intrathecal administration of magnesium solutions, 1 rabbit in the 1-mg/kg group, 2 rabbits in the 2-mg/kg group, and 5 rabbits in the 3-mg/kg group could not hop (motor function score, 0–1; Table 1). Although there was no significant difference in hindlimb motor function among the 5 groups 7 days after the administration, 4 rabbits (2-mg/kg and 3-mg/kg groups) showed persistent motor dysfunction (motor function score, 0–3). For sensory function, there was a significant difference among the five groups (Table 2). The sensory function score in the 3-mg/kg group at 7 days was worse than in the 0.3-mg/kg group, 1-mg/kg group, or saline group (P = 0.019). All rabbits in the saline, 0.3-mg/kg, and 1-mg/kg groups were neurologically normal 7 days after the administration.
There was no apparent change in motor neurons in the ventral horn. In contrast, an area of destruction with inflammatory changes was observed in the intermediate area of the gray matter (laminae V-VII) in 1 rabbit in the 1-mg/kg group (motor function score, 4), 2 rabbits in the 2-mg/kg group (motor function score, 1 and 2), and 1 rabbit in the 3-mg/kg group (motor function score, 0) (Fig. 1). The vacuolation that was confined within the dorsal funiculus was observed in 3 rabbits in the 1-mg/kg group (sensory score, 2, 2, and 2, respectively), 3 rabbits in the 2-mg/kg group (sensory score, 0, 0, and 2, respectively), and 3 rabbits in the 3-mg/kg group (sensory score, 0, 1, and 1, respectively). However, there was no significant difference in the degree of the vacuolation among the groups.
The motor-sensory function in each rabbit at the prescheduled time point was 0-0, 0-0, and 0-0 in the 6-h group, 0-0, 0-0, and 4-1 in the 48-h group, and 1-0, 1-0, and 3-1 in the 96-h group. There was no apparent morphological change in the lumbar spinal cord in the 6-h group. However, the area of destruction with inflammatory changes was observed in the intermediate area of the gray matter in 2 rabbits in the 48-h group (scores, 0-0 and 0-0) and 3 rabbits in the 96-h group (scores, 1-0, 1-0, and 3-1). The vacuolation of the dorsal funiculus was observed in 2 rabbits in the 48-h group (scores, 0-0 and 0-0) and 2 rabbits in the 96-h group (scores, 1-0 and 3-1).
Physiologic values are shown in Table 3. There were no significant differences among the groups. The time course of hindlimb motor function after recirculation is shown in Figure 2. There was no significant difference in hindlimb motor function 7 days after recirculation. The number of normal neurons in the anterior spinal cord 96 h after recirculation is shown in Figure 3. There was no significant difference in the number of normal neurons among the groups. In the rabbits with severe motor dysfunction, few normal neurons in the anterior spinal cord and destruction of the entire gray matter with inflammatory changes were seen.
In the present study, we demonstrated in rabbits that magnesium administered in the lumbar subarachnoid space had neurotoxicity in a dose-dependent manner and that intrathecal magnesium did not have any neuroprotective property in transient spinal cord ischemia.
In ischemia of the central nervous system, one of the important triggers that initiate a cascade of events resulting in cell death is thought to be Ca2+ entry into neurons through the NMDA-operated calcium channel (17). In the resting (i.e., nondepolarized) state, the NMDA-operated calcium channel is blocked by extracellular Mg2+ at a site deep within the channel. When both the binding of glutamate to the NMDA receptor and sufficient depolarization of the neuron occur, the block by Mg2+ is removed, and Ca2+ enters neurons through the NMDA-operated calcium channel (17). Therefore, in ischemia of the central nervous system, the increase in extracellular Mg2+ concentration may inhibit Mg2+ from being removed from the channel and reduce glutamate-induced NMDA receptor responses, leading to neuroprotection.
Because passage of magnesium across the blood-brain (spinal cord) barrier is poor, intrathecal administration seems preferable to obtain sufficient effects. If the intrathecal administration of magnesium is safe and has neuroprotective effects, it would be a useful strategy because cerebrospinal fluid drainage has become popular during thoracoabdominal aneurysmal surgery; thus, it would be easy to intrathecally administer magnesium during surgery. The intrathecal administration is also preferable from a hemodynamic standpoint. Therefore, we sought to investigate the safety of intrathecal magnesium and then to determine an optimal dose of magnesium for spinal cord protection in a rabbit model of spinal cord ischemia.
The intrathecal administration of large doses of magnesium sulfate was neurotoxic in rabbits, as demonstrated in Experiment 1A. Experiment 1B was undertaken to see the time-dependent features of histopathologic changes. The noncompetitive NMDA receptor antagonists, such as phencyclidine and dizocilpine (MK-801), are reported to have neurotoxic side effects consisting of vacuolation of neuronal cytoplasm in the brain (18). Vacuoles become evident 4 to 12 hours after the administration of the NMDA receptor antagonists and subside later (18). However, in the present study, there was no apparent morphological change six hours after the administration. We confirmed the area of destruction in the intermediate area of the gray matter in the rabbit with motor dysfunction at 48 and 96 hours. Therefore, the mechanism for neuronal injury by magnesium seems to be different from that by noncompetitive NMDA receptor antagonists.
The results are contrary to those reported by Chanimov et al. (11) who have shown that repeated intrathecal administrations of 4.6 mg/kg of magnesium sulfate in an iso-osmolar solution in the spinal cord are safe in rats. We have no clear explanation for this discrepancy except the species difference. However, as they have mentioned that the hyper-osmolar solution of 12.6% magnesium sulfate (9.2 mg/kg) may have been responsible for the increased degree of vacuolation of the ganglion cells (11), the larger doses of magnesium sulfate may cause neurotoxicity in rats.
The intrathecal drugs often used are not necessarily free of neurotoxicity. We have reported the neurotoxicity of the intrathecal administration of large doses of local anesthetics in the spinal cord in rabbits (15,19). In those studies, rabbits with severe motor dysfunction showed the central chromatolysis of motor neurons that is thought to be the result of axonal injury (15,19). In the present study, despite the existence of the area of destruction with inflammatory changes in the intermediate area of the gray matter, the morphological changes of motor neurons were not obvious. Therefore, the mechanism for motor dysfunction by magnesium seems to be different from that by local anesthetics. The difference in the mechanism for motor dysfunction might be associated with the difference in the site of action. Karasawa et al. (20) have investigated neurologic function after the intrathecal administration of magnesium sulfate (12.3% [4.1 mg/kg] or 24.6% [8.2 mg/kg]) or lidocaine (4% or 8%) in rats and have observed that motor paralysis caused by magnesium is spastic in character, whereas that caused by lidocaine is paralytic. Although they did not investigate histopathologic damage, they speculated that the inhibitory interneuron that affects motor neurons is inhibited by large concentrations of magnesium, resulting in spastic paralysis (20). The mechanism for the destruction in the intermediate area of the gray matter is uncertain at present. However, the area of destruction in the intermediate area of the gray matter may explain why motor dysfunction occurred without obvious changes of motor neurons in the present study because the damage of the interneurons in laminae III-VII has been reported to cause spastic paraparesis without apparent changes of motor neurons in the transient spinal cord ischemia in rats (21).
The lumbar intrathecal administration of magnesium sulfate (50 mg) has been used to prolong spinal opioid analgesia in laboring parturients, and no sensory or motor complications have been observed (22). However, in the present study, 3 of 6 rabbits showed vacuolation in the dorsal funiculus, and 1 rabbit showed the area of destruction in the intermediate area of the gray matter in the 1-mg/kg group, although no rabbits showed sensory or motor dysfunction seven days after administration. Therefore, further evaluation of the effects of intrathecal magnesium and the mechanisms for its neurotoxicity are warranted.
Regarding neuroprotection with magnesium, our results are contrary to two previous studies. Simpson et al. (9) administered 3 mg/kg of magnesium sulfate into the cisterna cerebellomedullaris 20 minutes before an aortic clamp in dogs. Although they observed the dogs until 24 hours after recirculation, all dogs in the magnesium group (n = 8) were neurologically and histopathologically normal (9). Follis et al. (10) administered 20 μL of 1-M magnesium sulfate (≈9 mg/kg) intrathecally in the lumbar spinal cord more than two hours before aortic occlusion in rats and observed neurologic function for 96 hours (no histopathologic examination). In Follis et al.'s study (10), magnesium did not prevent neurologic deficit one day after reperfusion but improved neurologic recovery four days after reperfusion. The reasons for the different results between these two studies and ours is not known. But the differences in the dose of magnesium sulfate and severity of ischemia might be the cause. The doses of magnesium sulfate (0.3 and 1 mg/kg) used in our study were smaller than those used in the studies by Simpson et al. (9) and Follis et al. (10). However, the dose of 1 mg/kg represents the maximum tolerable dose in rabbits from a neurotoxic standpoint. Simpson et al. (9) and Follis et al. (10) only monitored rectal temperature and maintained it at 36°C and 37°C ± 0.5°C, respectively, whereas we maintained paravertebral muscle temperature at 38°C (esophageal temperature at ≈38.5°C, which is the normal temperature for rabbits). Small differences in the spinal cord temperature might have affected the severity of ischemia. Differences in animal species may have influenced the results. Nevertheless, because we demonstrated that magnesium intrathecally administered had obvious neurotoxicity in a dose-dependent manner, we feel that further evaluation of the safety and efficacy of intrathecal magnesium is required before this drug is used in the clinical setting.
In summary, the present results indicate that intrathecal administration of magnesium has a risk of neurotoxicity and shows no evidence of protective effects against ischemic spinal cord injury in rabbits.
1. Hao J-X, Xu X-J, Aldskogius H, et al. The excitatory amino acid receptor antagonist MK-801 prevents the hypersensitivity induced by spinal cord ischemia in the rat. Exp Neurol 1991;113:182–91.
2. Saganova K, Marsala M. Intrathecal administration of dizocilpine maleate (MK-801) attenuates ischemic damage in the rabbit spinal cord. Exp Neurol 1994;130:337–43.
3. Ehrlich M, Knolle E, Ciovica R, et al. Memantine for prevention of spinal cord injury in a rabbit model. J Thorac Cardiovasc Surg 1999;117:285–91.
4. Johnson SH, Kraimer JM, Graeber GM. Effects of flunarizine on neurological recovery and spinal cord blood flow in experimental spinal cord ischemia in rabbits. Stroke 1993;24:1547–53.
5. Burns LH, Jin Z, Bowersox SS. The neuroprotective effects of intrathecal administration of the selective N-type calcium channel blocker ziconotide in a rat model of spinal ischemia. J Vasc Surg 1999;30:334–43.
6. Iseri LT, French JH. Magnesium: nature's physiologic calcium blocker. Am Heart J 1984;108:188–93.
7. Miles AN, Majda BT, Meloni BP, Knuckey NW. Postischemic intravenous administration of magnesium sulfate inhibits hippocampal CA1 neuronal death after transient global ischemia in rats. Neurosurgery 2001;49:1443–50.
8. Vacanti FX, Ames A III. Mild hypothermia and Mg++ protect against irreversible damage during CNS ischemia. Stroke 1984;15:695–8.
9. Simpson JI, Eide TR, Schiff GA, et al. Intrathecal magnesium sulfate protects the spinal cord from ischemic injury during thoracic aortic cross-clamping. Anesthesiology 1994;81:1493–9.
10. Follis F, Miller K, Scremin OU, et al. NMDA receptor blockade and spinal cord ischemia due to aortic crossclamping in the rat model. Can J Neurol Sci 1994;21:227–32.
11. Chanimov M, Cohen ML, Grinspun Y, et al. Neurotoxicity after spinal anaesthesia induced by serial intrathecal injections of magnesium sulphate: an experimental study in a rat model. Anaesthesia 1997;52:223–8.
12. Matsumoto M, Iida Y, Sakabe T, et al. Mild and moderate hypothermia provide better protection than a burst-suppression dose of thiopental against ischemic spinal cord injury in rabbits. Anesthesiology 1997;86:1120–7.
13. Matsumoto M, Iida Y, Wakamatsu H, et al. The effects of NG
-nitro-L-arginine-methyl ester on neurologic and histopathologic outcome after transient spinal cord ischemia. Anesth Analg 1999;89:696–702.
14. Matsumoto M, Ohtake K, Wakamatsu H, et al. The time course of acquisition of ischemic tolerance and induction of heat shock protein 70 after a brief period of ischemia in the spinal cord in rabbits. Anesth Analg 2001;92:418–23.
15. Yamashita A, Matsumoto M, Matsumoto S, et al. A comparison of the neurotoxic effects on the spinal cord of tetracaine, lidocaine, bupivacaine, and ropivacaine administered intrathecally in rabbits. Anesth Analg 2003;97:512–9.
16. Drummond JC, Moore SS. The influence of dextrose administration on neurologic outcome after temporary spinal cord ischemia in the rabbit. Anesthesiology 1989;70:64–70.
17. Hudspith MJ. Glutamate: a role in normal brain function, anaesthesia, analgesia and CNS injury. Br J Anaesth 1997;78:731–47.
18. Olney JW, Labruyere J, Price MT. Pathological changes induced in cerebrocortical neurons by phencyclidine and related drugs. Science 1989;244:1360–2.
19. Ohtake K, Matsumoto M, Wakamatsu H, et al. Glutamate release and neuronal injury after intrathecal injection of local anesthetics. Neuroreport 2000;11:1105–9.
20. Karasawa S, Ishizaki K, Goto F. The effect of intrathecal administration of magnesium sulphate in rats. Anaesthesia 1998;53:879–86.
21. Taira Y, Marsala M. Effect of proximal arterial perfusion pressure on function, spinal cord blood flow, and histopathologic changes after increasing intervals of aortic occlusion in the rat. Stroke 1996;27:1850–8.
© 2004 International Anesthesia Research Society
22. Buvanendran A, McCarthy RJ, Kroin JS, et al. Intrathecal magnesium prolongs fentanyl analgesia: a prospective, randomized, controlled trial. Anesth Analg 2002;95:661–6.