Neuronophagia by leukocytes in experimental spinal cord injury.

Experimental spinal cord injury was produced in cats by compression trauma. The major histologic features in the first 24 hours following injury included hemorrhagic necrosis, edema, and acute inflammation of gray and white matter. Neutrophilic leukocytes (NL) were apparent in the walls of and adjacent to veins and venules within four hours of injury, but large numbers of NL were not observed in tissue until eight to 24 hours. Acute inflammation was especially prominent in gray matter, and NL frequently surrounded and phagocytized neuronal somata. Large numbers of NL and neuronophagia by leukocytes were evident only in areas of hemorrhage. The role of blood in producing the acute inflammatory response and in generating chemotactic factors responsible for neuronophagia is incompletely understood. Neutrophilic leukocytes, in addition to their response as phagocytes, release histolytic enzymes, reactive species of oxygen, and proinflammatory factors which lead to further tissue necrosis and inflammation.

Susceptibility of feline spinal cord energy metabolism to severe incomplete ischemia.

Feline spinal cords were subjected to 10 to 30 minutes of severe incomplete ischemia (average reduction in blood flow of 92%) with and without 90 minutes of recirculation, and the L-2 segment was analyzed for high-energy phosphates and certain glycolytic metabolites. Spinal cord tissue lactic acid levels were stepwise elevated, and adenosine triphosphate (ATP), phosphocreatine (P-creatine), and glucose were progressively consumed by increasing durations of ischemia. However, upon restoration of blood flow, there was extensive recovery of energy metabolites and normalized lactic acid, demonstrating resumption of mitochondrial oxidative metabolism. These data indicate that the spinal cord can tolerate at least 30 minutes of severely reduced blood flow before recovery of energy metabolism is significantly impaired upon restitution of blood flow.

Bromocriptine treatment of neuroleptic malignant syndrome.

Five patients with neuroleptic malignant syndrome were treated with bromocriptine mesylate 7.5-45 mg/day in three divided doses for at least 10 days. Response to therapy was assessed by monitoring vital signs and serum creatine kinase (CK) levels. In addition, a disability score was determined using a scale based on variables assessed on neurologic examinations. In all five patients, significant improvement was noted 24-72 hours after initiation of bromocriptine treatment and was accompanied in four patients by a rapid drop in serum CK levels. Resolution of confusion and mutism was noted within 24-48 hours after treatment. Normalization of vital signs occurred within 48 hours to 4 days, and resolution of extrapyramidal rigidity occurred within 1 week. In two patients, early discontinuation of bromocriptine resulted in relapse of neuroleptic malignant syndrome, which responded to reinstitution of the drug. The results suggest a therapeutic role for bromocriptine in the treatment of neuroleptic malignant syndrome.

Pretreatment with alpha tocopherol enhances neurologic recovery after experimental spinal cord compression injury.

Lipid hydrolysis with subsequent production of eicosanoids and lipid peroxidation are two of the earliest potentially pathochemical events induced in spinal cord tissue by mechanical trauma. Although these membrane lipid disturbances are thought to contribute to the paralysis that occur subsequent to spinal cord injury, such a correlation has not been demonstrated directly. Consequently, the purpose of this study was to test the capacity of alpha tocopherol, the major lipid antioxidant in cellular membranes and a compound that limits the injury-induced lipid hydrolysis and peroxidation in spinal cord tissue, to promote functional recovery in a static loading model of spinal cord injury. After laminectomy, the L2 spinal cord of cats was compressed with 180 g for 5 min. For 5 days before injury and for 5 days postinjury, treated cats received orally 1000 IUD-alpha tocopherol acetate daily. Control cats were similarly injured but untreated. All cats were blindly evaluated weekly for 4 weeks for their neurologic recovery based on an 11 point behavioral scale that assessed walking, running, and stair climbing. By the second postinjury week, alpha tocopherol-pretreated cats demonstrated significantly better recovery than untreated controls. By 4 weeks, treated cats had recovered 72% of their preinjury function as compared with 20% for untreated controls, i.e., a 3.5-fold difference. These results strongly suggest that lipid peroxidation and/or hydrolysis is primarily involved in the genesis of posttraumatic paralysis and that alpha tocopherol exerts its protection of injured spinal cord tissue, at least in part, by its antioxidant and/or antilipolytic activity.

Effect of delayed administration of U74006F (tirilazad mesylate) on recovery of locomotor function after experimental spinal cord injury.

Beginning at either 30 minutes, 2 hours, 4 hours, or 8 hours after 180 g compression of the cat L2 spinal cord for 5 minutes, infusion of U74006F was initiated. In this series, the cats received a total U74006F dose of 5 mg/kg/48 hours. An additional group of injured cats was treated at 8 hours postinjury with a three-fold higher dose of U74006F (i.e., a total 48-hour dose of 15 mg/kg). Controls received an equal volume of vehicle (citrate-buffered saline) delivered over 48 hours. The cats were evaluated weekly for 4 weeks for recovery of overground locomotion based on an 11-point scale by an investigator blinded to the time and type (i.e., vehicle or drug) of material administered. By 4 weeks postinjury, there was no significant difference in the locomotor recovery of cats that received U74006F at either 30 minutes, 2 hours, 4 hours, or 8 hours after injury. However, only recovery in the groups treated at 30 minutes, 2 hours, or 4 hours after injury was significantly greater than vehicle-treated controls. Locomotor function in cats receiving either 5 mg/kg/48 hours or 15 mg/kg/48 hours of U74006F at 8 hours postinjury was not significantly different from that of the vehicle-treated animals. Mean (+/- SEM) 4-week recovery scores were 6.8 +/- 0.9, 5.9 +/- 1.0, 7.2 +/- 1.1, and 4.7 +/- 2.9 out of 11 for cats treated at 30 minutes, 2 hours, 4 hours, or 8 hours postinjury, respectively, with the 5 mg/kg/48 hour dose. The mean recovery score for cats treated at 8 hours after injury with the 15 mg/kg/48 hour dose was 3.4 +/- 1.8. The average score for the vehicle-treated controls was 1.8 +/- 0.8. These findings demonstrate that U74006F can significantly protect locomotor function in our model of compression spinal cord injury if administered as late as 4 hours postinjury. Delaying administration of the compound to 8 hours after injury results in considerable loss of its protective capabilities even if the dose is increased threefold.

Thrombin interactions with central nervous system tissue and implications of these interactions.

A monospecific antibody was developed to human alpha-thrombin. This antibody stained neurons but not astrocytes in murine spinal cord cultures incubated with 1-10 nM alpha-thrombin using the avidin-biotin-peroxidase technique. Staining did not occur when the primary or linking antibodies were eliminated, and staining was blocked with hirudin. Preliminary studies showed release of arachidonic acid from the cultures when exposed to thrombin. It was proposed that arachidonate release from the membranes of neurons upon exposure to thrombin was similar to that observed in platelets, for example, by activation of phospholipases. Moreover, prostanoids were formed that could have a deleterious effect on cellular elements in the central nervous system. The potential role of thrombin receptors on neurons was discussed.

The kappa opioid-related anticonvulsants U-50488H and U-54494A attenuate N-methyl-D-aspartate induced brain injury in the neonatal rat.

The neuroprotective effects of the kappa opioid-related anticonvulsants U-50488H and U-54494A were tested in a model of N-methyl-D-aspartate (NMDA)-induced brain injury in the neonatal rat. Seven-day-old rat pups were injected intracerebrally with 7.5 nmol NMDA. Five days later, the ensuing unilateral hemisphere weight reduction was measured and used to assess the severity of insult. Control animals (n = 85) exhibited a 21.7 +/- 0.5% hemisphere weight reduction. Animals treated with U-54494A in split doses before and after NMDA administration showed significant neuroprotection at 10, 15, and 20 mg/kg, with the maximum effect observed at 15 mg/kg (33.8% protection). Animals treated with U-50488H on a similar dosing schedule showed significant neuroprotection at all doses tested, with peak protection observed at 30 mg/kg (51.8% protection). Both compounds exhibited a neuroprotective effect when hemisphere cross-sectional area and hippocampal histology were assessed. Treatment with U-54494A after NMDA administration also afforded neuroprotection at various doses, as measured by hemisphere weight disparity, with peak protection occurring at a dose of 20 mg/kg (32.4% protection). These data show that both U-50488H and U-54494A afford neuroprotection against NMDA-induced neuronal injury in the neonatal rat brain.

Methylprednisolone and membrane properties of primary cultures of mouse spinal cord.

The present study attempts to define the capacity of methylprednisolone sodium succinate (MP) to protect neuronal membranes against a free radical challenge in primary cultures of fetal mouse spinal cord. Incubation of these cultures with MP significantly increased the Na+,K(+)-ATPase activity, an effect that was blocked by the RNA synthesis inhibitor, actinomysin D and the protein synthesis inhibitor, cycloheximide, suggesting an induction of protein synthesis by MP. In contrast, incubation with FeCl2 for 1 or 2 h significantly inhibited Na+,K(+)-ATPase activity and elevated the levels of thiobarbituric acid-reactive substances (TBARS). Pretreatment with MP prevented the rise in TBARS and partially prevented the decrease in Na+,K(+)-ATPase activity for the first hour of FeCl2 incubation, an effect that was lost during the second hour. A second dose of MP after the first hour of incubation with FeCl2 partially restored Na+,K(+)-ATPase activity and reduced TBARS levels after the second hour of exposure to FeCl2. Co-incubation of MP with cycloheximide completely prevented the decrease in Na+,K(+)-ATPase activity seen after a 2-h incubation with FeCl2 and eliminated the need for a second dose of MP after the first hour of incubation with FeCl2. These findings suggest a capacity for rapid protein induction and antioxidant activity for MP in vitro.

Effect of methylprednisolone in compression trauma to the feline spinal cord.

The purpose of this study was to determine the effect of methylprednisolone sodium succinate on clincal recovery and tissue preservation following compression trauma of feline spinal cord. Cats were anesthetized with pentobarbital and injured by placing a 170-gm weight on the spinal cord for 5 minutes. One hour after injury, the animals were given intravenous steroid (15 mg/kg/day) for 2 days in three devided doses, 15 mg/kg/day for 1 day intramuscularly, 7.5 mg/kg/day intramuscularly for 3 days, and 3.75 mg/kg/day intramuscularly for 3 days, for a total of 9 days. In a control group, the animals were injured but untreated. At 60 days after injury, the animals were sacrifieced by perfusion fixation with 10% formalin. The spinal cord was removed and evaluated for a number of morphometric parameters, including percentage of spinal cord cross-sectional area containing the cavity (%area) and percentage of spinal cord volume occupied by the cavity (%volume). A clinical recovery score (recovery index) was devised to evaluate neurological recovery. Steroid-treated cats showed significantly greater recovery than the untreated controls (p less than 0.001). Moreover, the spinal cord of treated cats displayed greater tissue preservation as measured by %area (p leass than 0.005) and %volume (p less than 0.004). Correlation coefficients comparing the recovery index with morphometric parameters revealed a negative correlation between cavity size and recovery. These data provide evidence for a beneficial effect of methylprednisolone in promoting recovery and preserving spinal cord tissue following blunt injury to the feline spinal cord.

Spinal cord energy metabolism in normal and postlaminectomy cats.

The purpose of this study was to determine 1) normal concentrations of adenosine triphosphate (ATP), adenosine diphosphate (ADP), adenosine monophosphate (AMP), phosphocreatine (P-creatine), glucose, lactate, and pyruvate in upper (L-2) and lower (L-5) feline lumbar spinal cord, and 2) whether previously reported laminectomy-induced reduction in spinal cord blood flow (SCBF) resulted in disturbance of spinal cord energy metabolism. Concentrations of ATP, P-creatine, pyruvate, and glucose were significantly higher at L-5 than at L-2, probably as the result of larger amounts of gray matter at L-5 than L-2. Significant increases in ADP and AMP levels were the only metabolite changes noted 15 minutes following laminectomy. The authors speculate that the increase in ADP and AMP is due to a laminectomy-induced stimulation of ATP utilization. However, lack of change in other metabolites implies an efficient energy homeostasis. These results indicate that although laminectomy can reduce SCBF, the degree of this reduction is insufficient to adversely affect spinal cord energy metabolism. Thus, tissue from beneath or near the laminectomy site is viable and essentially normal.

Evaluation of an intensive methylprednisolone sodium succinate dosing regimen in experimental spinal cord injury.

Beginning 30 minutes after compression trauma of the upper lumbar (L-2) spinal cord, cats were treated with either a high-dose regimen of methylprednisolone (MP) administered as the sodium salt of the 21-succinate ester (Solu-Medrol sterile powder) or the MP vehicle. Animals were randomly assigned to either treatment group (10 cats per group), and all personnel were blind as to which animals received vehicle or drug. The intensive 48-hour dosing regimen was designed to maintain therapeutic tissue levels of MP and consisted of an initial 30 mg/kg intravenous bolus of MP; 2 and 6 hours later additional 15 mg/kg MP doses were administered by intravenous bolus. Immediately following the bolus given at 6 hours, a continuous MP infusion of 2.5 mg/kg/hr was started. The infusion was stopped abruptly at 48 hours with no dose tapering. Animals in the vehicle group received an equivalent volume of MP vehicle. The total MP dose administered over 48 hours was 165 mg/kg. Animals were evaluated weekly for neurological recovery based upon a 12-point functional scale which assessed general mobility, running, and stair-climbing. Mean recovery scores at 1 month after injury (+/- standard error of the mean) were: vehicle group (seven cats) 3.7 +/- 0.9, and MP group (10 cats) 8.7 +/- 0.2; (p less than 0.001). Histological evaluation of the spinal cords revealed a strong negative correlation between neurological recovery and size of the spinal cord cavity at 1 month (r = -0.88). Three of 10 animals in the vehicle group became ill and had to be dropped from the study, whereas all of the 10 MP-treated animals survived in excellent health. The results demonstrate the therapeutic effectiveness and low incidence of side effects associated with an intensive MP dose regimen for treatment of experimental spinal cord injury.

Cerebrospinal fluid lactate and electrolyte levels following experimental spinal cord injury.

Cerebrospinal fluid (CSF) lactate, sodium (Na+), potassium (K+), calcium (Ca++), magnesium (Mg++), and chloride (C1-) levels were determined for 17 to 21 days following experimental spinal cord compression in cats. Laminectomies were performed at L-2 under general anesthesia with aseptic techniques. Paraplegia was produced by applying a 170-gm weight transdurally for 5 minutes. Significant increases in CSF lactate levels were observed on the first through ninth days post injury with peak levels (50% above normal) occurring at Day 5. The only significant postinjury CSF electrolyte changes were elevation in Ca++ concentration on Days 3, 9, 11, 13, and 15, elevation in K+ concentration on Days 9 and 11 and decline in C1- levels on the first day. The CSF K+ increase probably reflected cellular loss of K+ from damaged tissue whereas the Ca++ rise may have resulted from increased CSF protein levels. The prolonged elevation of CSF lactate indicates that tissue hypoxia plays a role in spinal cord compression paralysis, and that there is a continuing hypoxia of metabolically active spinal cord tissue for several days post injury.

Spinal cord energy metabolism following compression trauma to the feline spinal cord.

The purpose of this study was to determine the spinal cord metabolic state for 24 hours after compression trauma to the feline spinal cord. Cats were anesthetized with pentobarbital and injured by placing a 190-gm weight on the spinal cord for 5 minutes. Biochemical analysis of the injured segment revealed a significant depletion in the levels of adenosine triphosphate (ATP), phosphocreatine (P-creatine), and total adenylates for the entire 24-hour recovery period. Glucose levels initially declined, but by 1 hour had normalized, and at 8 and 24 hours were significantly supranormal. The lactate/pyruvate ratio and tissue lactate concentrations increased four and five and half times, respectively, for the first 4 hours after injury. Between 8 and 24 hours, lactate levels remained elevated, whereas the lactate/pyruvate ratio declined to contol levels as the result of a significant rise in the tissue pyruvate concentration. This sequence of metabolic changes suggested that metabolism was probably not homogeneous throughout the injured segment, and that tissue metabolic rate was depressed for the initial 4 hours after trauma then increased in metabolically active tissue for the remainder of the 24-hour recovery period. This model of spinal cord trauma results in a severe, prolonged ischemia and metabolic injury to the affected tissue. Whether these metabolic changes results from or cause the tissue damage and irreversible paraplegia associated with this type of spinal cord injury remains to be determined.

Effects of laminectomy on spinal cord blood flow.

The effect of a one-segment (L-2) laminectomy on spinal cord blood flow (SCBF) was determined by the reference sample method using isotope-labeled microspheres. The SCBF was measured before laminectomy (control) and at 15 minutes postlaminectomy with the dura exposed (Series 1), 1 hour postlaminectomy with the laminectomy site closed (Series 2), 24 hours postlaminectomy with the laminectomy site closed (Series 3), and 24 hours postlaminectomy with the dura exposed (Series 4). With the laminectomy site open, SCBF was significantly depressed (22% to 45%) along the entire length of the spinal cord at 15 minutes postlaminectomy. At 1 hour postlaminectomy (with the laminectomy site closed), SCBF approached control values, although areas with significantly lowered flow were still observed in all portions of the spinal cord. By 24 hours postlaminectomy, SCBF had returned to prelaminectomy levels. However, if within 1 hour preceding the 24-hour SCBF measurement, the laminectomy site was reopened, SCBF tended to fall at and caudad to the laminectomy site. These data indicate that laminectomy can cause a significant decline in SCBF. At the present time, the mechanism(s) for this laminectomy-induced depression of SCBF are unknown, although a temperature-induced vasoconstriction is suspected.

Microvascular perfusion and metabolism in injured spinal cord after methylprednisolone treatment.

The purpose of this study was to determine the effect of treatment with the synthetic glucocorticoid, methylprednisolone, on the microvasulature and metabolism of the traumatized spinal cord. Spinal cords of cats were compressed with a 170-gm weight for 5 minutes and were treated with either high-dose methylprednisolone (HDMP, 15 mg/kg/24 hrs) or megadose methylprednisolone (MDMP,60 mg/kg/24 hrs). Animals were sacrificed at 2, 8, or 24 hours following injury. Treatment with HDMP resulted in substantial preservation of injured spinal cord microvascular perfusion at 8 hours as compared with injured untreated cats. Compression trauma caused a partial derangement of energy metabolism and a shift toward anaerobic glycolysis in both treated and untreated groups for the entire 24-hour postinjury period. Tissue levels of adenosine triphosphate, phosphocreatine, and total adenylates in the HDMP-treated cats sacrificed at 8 hours after trauma were significantly elevated over untreated controls, but those in the 2- and 24-hour groups were not. Concentration of energy intermediates in MDMP-treated cat were either equal to or below those of injured untreated animals al all three postinjury time period. The postinjury metabolite pattern and concentrations seen in this study possibly result from differing levels of blood flow and neuronal activity in the injured untreated, HDMP-, and MDMP-treated spinal cords. Better tissue perfusion in the HDMP-treated cats might be expected to result in an improved tissue energy state in these animals. However, intensive high-dose glucocorticoid treatment has been demonstrated to augment spinal cord monosynaptic and polysynaptic reflex transmission and primary afferent excitability. Furthermore, acute single intravenous dose studies have shown this direct neuronal action to be dose-related. Thus, additional high-energy phosphate molecules that may be reformed as a result of HDMP treatment were perhaps used as the energy source for any increased neuronal activity caused by steroid administration. The beneficial effects of glucocorticoid treatment in experimental spinal cord trauma might derive from preserved cellular structural integrity. This could result in increased levels of neuronal activity, energy utilization, and production in treated as compared with untreated tissue.

Effects of treatment with U-74006F on neurological outcome following experimental spinal cord injury.

The compound U-74006F is one of a series of 21-aminosteroids that lack glucocorticoid or mineralocorticoid activity. These potent inhibitors of lipid peroxidation have been specifically developed for the acute treatment of central nervous system trauma and ischemia. This study evaluated the dose-response characteristics and capability of U-74006F to promote functional recovery in cats subjected to compression trauma of the upper lumbar (L-2) spinal cord. Thirty minutes following injury, randomized and investigator-blinded treatment was initiated with the intravenous administration of either vehicle (citrate-buffered saline) or one of eight doses of U-74006F. Initial doses of U-74006F ranged from 0.01 to 30 mg/kg. Subsequent doses consisted of intravenous bolus injections followed by a continuous 42-hour intravenous infusion. Over the 48-hour treatment period, cats received total U-74006F doses ranging from 0.048 to 160 mg/kg. The animals were evaluated weekly for neurological recovery based upon an 11-point behavioral scale. With the exception of two cats in one group, the animals receiving accumulated doses of U-74006F (ranging from 1.6 to 160.0 mg/kg/48 hrs) exhibited nearly 75% of normal neurological function by 4 weeks after injury. Lower total doses of 0.16 and 0.48 mg/kg/48 hrs were associated with approximately 50% return of normal function, which was not significantly better than the recovery in the vehicle-treated control group. The lowest total dose tested (0.048 mg/kg/48 hrs) gave results indistinguishable from those in vehicle-treated cats, which had recovered only 20% of their preinjury neurological function by 4 weeks. These findings demonstrate that over a 100-fold range of doses, U-74006F has a remarkable capacity to promote functional recovery in spinal cord-injured cats.

A multicenter trial on the efficacy of using tirilazad mesylate in cases of head injury.

OBJECT: The authors prospectively studied the efficacy of tirilazad mesylate, a novel aminosteroid, in humans with head injuries. METHODS: A cohort of 1120 head-injured patients received at least one dose of study medication (tirilazad or placebo). Eighty-five percent (957) of the patients had suffered a severe head injury (Glasgow Coma Scale [GCS] score 4-8) and 15% (163) had sustained a moderate head injury (GCS score 9-12). Six-month outcomes for the tirilazad- and placebo-treated groups for the Glasgow Outcome Scale categories of both good recovery and death showed no significant difference (good recovery in the tirilazad-treated group was 39% compared with the placebo group in which it was 42% [p=0.461]; death in the tirilazad-treated group occurred in 26% of patients compared with the placebo group, in which it occurred in 25% [p=0.750]). Subgroup analysis suggested that tirilazad mesylate may be effective in reducing mortality rates in males suffering from severe head injury with accompanying traumatic subarachnoid hemorrhage (death in the tirilazad-treated group occurred in 34% of patients; in the placebo group it occurred in 43% [p=0.026]). No significant differences in frequency or types of serious adverse events were shown between the treatment and placebo groups. CONCLUSIONS: Striking problems with imbalance concerning basic prognostic variables were observed in spite of the large population studied. These imbalances concerned pretreatment hypotension, pretreatment hypoxia, and the incidence of epidural hematomas. In future trials of pharmacological therapy for severe head injury, serious consideration must be given to alternative randomization strategies. Given the heterogeneous nature of head injury and the identification of populations that do relatively well with standard therapy, target populations with a higher risk for mortality and morbidity may be more suitable for clinical trials of such agents.

Pathology of lacquer thinner induced neuropathy.

Seven men aged 17 to 22 years developed severe distal symmetrical predominately motor polyneuropathy after repeated inhalation of a commercially available brand of lacquer thinner. Motor nerve conduction velocities were markedly slowed. Fascicular biopsy specimens of sural nerve showed a striking loss of myelinated nerve fibers. Prominent neurofilamentous masses resulted insegmental paranodal distention of axons with secondary thinning and retraction of myelin from the node of Ranvier. Autopsy material from one case revealed central chromatolysis of anterior horn cells in the lumbosacral enlargement and axonal swellings in the fasciculus gracilis. One or more volatile hydrocarbons contained in the lacquer thinner involved may be neurotoxic if inhaled to excess any many cause a neuropathy with characteristic pathological features.

Acute spontaneous subdural hematoma.

A review of the literature reveals 20 reported cases of acute spontaneous subdural hematoma. We present an additional case with a tabulation of pertinent data. The prototype case is described, emphasizing the benign and insidious onset followed, invariably, by a deteriorating neurological status leading, almost always, to surgical evacuation. A high index of suspicion for this entity is of great importance, because a delay in diagnosis may lead to an increase in the morbidity and the mortality of this condition.

Microvascular perfusion experimental spinal cord injury.

Microperfusion of the spinal cords in cats was studied using a colloidal carbon perfusion technique following compression injury at 1/2 hour, 2 hours, 4 hours, 8 hours, 24 hours, and 51 days. Quantitative estimates of vascular filling were determined at these post-compression intervals. Microperfusion diminished in both the gray and white matter at 1/2 hour following injury and severe lack of perfusion was evident at 8 and 24 hours. Diminished filling of the vessels of gray and white matter seemed to parallel the degree of hemorrhagic necrosis of the gray matter. An increased number of vessels were evident in the spinal cords of long term survivals. The observation that microperfusion in the white matter of the spinal cord was diminished at 24 hours is at variance with some previous investigations. The hypoperfusion of the white matter found in this study suggests that ischemia plays a role in paraplegia resulting from experimental compression injury of the spinal cord.