Spinal cord protection with intravenous nimodipine A functional and morphologic evaluation The purpose of this study is to determine the effects of ischemia in the spinal cord when a calcium channel blocker, nimodipine, is administered intravenously before, during, and after crossclamping of the thoracic aorta. In this series of experiments, 18 pigs underwent thoracotomies and had 17.5 to 18.0 em of the thoracic aorta clamped for 30 minutes. By random selection, nine animals received intravenous nimodipine (2 Ilg/kg per minute) and nine control animals received only a carrier solution. Of the nine animals that received nimodipine, eight walked after the operation. In contrast, in the control group only two of nine animals walked. The ninth nimodipine-treated animal walked but had a severe delayed deterioration response. All animals, except one control animal, had a negative central spinal perfusion pressure. Morphologic examination of serial sections of spinal cords from control animals showed diffuse neuronal necrosis. In comparison, cords from the nimodipine group had swollen neurons accompanied by an inflammatory infiltrate and only occasional necrotic neurons. With this data, we conclude that certain calcium channel blockers, when administered in sufficient doses, can lend a protective effect to the spinal cord during ischemic events even when cord perfusion pressure has dropped to dangerously low levels. (J THoRAe CARDIOVASC SURG 1992;104:1100-5)
Anton Schittek, MD,a Ger B. W. E. Bennink, MD,a Denton A. Cooley, MD,a and Lauren A. Langford, MD, DrMed,b Houston, Tex.
NurologiC deterioration occurring with ischemic spinal cord damage is considered to be directly related to a decrease in central spinal perfusion pressure; consequently, many investigators have concentrated on modulating determinants of central spinal perfusion pressure. 1, 2 How the decrease in central spinal perfusion pressure results in death of the neuron is a complicated problem. When the central nervous system (CNS) perfusion pressure drops, systemic vasomotor and respiratory responses attempt to maintain the perfusion pressure by increasing the systemic blood pressure. These systemic responses in turn produce increases in spinal cord sympathetic neuronal discharge to flood the CNS with neurotransmitters. However, in a situation such as aortic surgery, the cord still may not receive additional blood supply and hypoxia continues. With sustained hypoxic conditions, the preFrom the Texas Heart Institute, Texas Medical Center,' and the Section of Neuropathology, St. Luke's Episcopal Hospital," Houston, Tex. Received for publication Feb. 25,1991. Accepted for publication Sept. 16, 1991. Addressfor reprints:Denton Cooley, MD, Texas Heart Institute,P.O. Box 20345, Houston, TX 77225-0345.
synaptic uptake of peptide-bound and free amino acids is impaired and leads to massive overexposure of neurons to the potentially toxic substances. The mechanism by which large quantities of neurotransmitters, amino acids, and other substances kill neurons is not understood, but most data suggest that elevated levels of intracellular calcium are responsible.':" Logically, if calcium metabolism can be altered in a manner to protect, or prime, the CNS microenvironment, more cells might remain viable and capable of recovery after a critical reduction of blood supply. Thus the aim of the present study is to test this hypothesis. Materials and methods Eighteen adult pigs (20 to 30 kg) were used in this experiment. Nine animals were randomly assigned to the control group and nine were randomly assigned to the experimental group. All were anesthesized by an initial intramuscular dose of ketamine hydrochloride (25 rug/kg) and acepromazine maleate (1.0 mg/kg) followed by inhalation of isoflurane (I -chloro2,2,2-trifluoroethyl difluoromethyl ether) (0.5% to 3.0%) with nitrous oxide (30% to 80%). The animal was positioned for a left thoracotomy, shaved, draped, and cleansed with povidoneiodine (Betadine) surgical scrub. An 18-gauge catheter attached to a manometer was then placed in the lumbosacral spinal canal. The right femoral artery, vein, left carotid artery,
Volume 104 Number 4 October 1992
Spinal cord protection
Table I Time Factor
Mean proximal BP (mm Hg)
Nimodipine Control Nimodipine Control Nimodipine Control Nimodipine Control Nimodipine Control
Mean distal BP (rnrn Hg) Heart rate (beats/min) CSF pressure (mm Hg) CSPP (mm Hg)
59 ± 55 ± 53 ± 53 ± 109 ± 103 ± 11.1 ± 9.9 ± 41 ± 49 ±
14 20 12 9 14 21 2.oct I.<)d
87 ± 115 ± 85 ± 72 ± 157 ± 162 ± 11.9 ± 10.6 ± -3.6 ± -2.3 ±
25" 25" 2.1 2.1 36 49 2.4e 3.0< 1.7 2.6
Postclamp 48 ± 8b 56 ± lOb 36 ± 7c 48 ± 7c 133 ± 31 130 ± 21 9.1 ± 2.3 8.7 ± 3.1 26 ± 8f 39 ± 9f
Significant statistical differences between groups are designated by superscripts a through! These differences occurred in mean proximal blood pressure during clamping (a) and after clamping (b); mean distal blood pressure in postclamping (e), CSF pressure in preclamping (d), and CSF pressure in clamping (e) intervals; and central spinal perfusion pressure during postclamping If) interval. 'p < 0.0001. bp = 0.0010. 'p = 0.0028. '» = 0.0012. "p = 0.0029. fp = 0.0106.
and left jugular vein were cannulated. Monitoring of blood pressure, heart rate, cerebrospinal fluid (CSF) pressure, and central venous pressure (CYP) were begun. Under aseptic conditions,the left side of the chest was opened at intercostal spaces four and nine. The thoracic aorta was exposed and encircled with umbilicaltape at twosites:proximal,just downstream from the left subclavian artery, and distal, above the diaphragm, for a total clamping distance of 17.5 to 18.0 em including eight to nine intercostal vessels. Forty-fiveminutes before the aorta was crossclamped at the unbilical tape sites, an infusion of an unknownsolution (either carrier solution, i.e., polyethylene glycol400, or nimodipinesolution, 2 p.g/kg per minute) was begun intravenouslyand continued throughout the 30-minute period of occlusion and an additional 30 minutes after the clamp release. In addition, blood samples and CSF samples were taken at 3D-minute intervals and stored in darkness for later measurement of drug concentration. Systemic heparinization (2 mg/kg intravenously) was done 3 minutes before the application of the vascular clamps on the aorta. Protamine sulfate (I mg/kg intravenously) was administered immediately after clamp release.The thoracotomy was closedin a routine manner. The pig was placed in a panepinto sling for recovery. All animals were evaluated neurologically (without knowledge of whether the pig was a recipient of carrier solution or nimodipine) for bowel and bladder dysfunction, forelimb and hind limb movement, and general discomfort three to four times a day for 48 hours after the operation. Gait status was graded according to Tarlov's criteria'': Grade 0: Spastic paraplegia and no movement of the lower limbs Grade I: Spastic paraplegia and slight movement of the lower limb joint Grade 2: Good movement of the lower limbs but unable to stand Grade 3: Able to stand but not able to walk normally Grade 4: Complete recovery
The animals were killed by intravenous injection of sodium barbital (l mg/kg) 48 hours after the operation. After the descending aorta was removed, the length between the umbilical clamps and the number of intercostal branches were determined. The entire spinal cord was removed and fixed in 10% buffered formalin for 3 days. Serial cross sections of the cord were taken at I em intervals, rostral to caudal, processed for routine light microscopy, and examined in a compound light microscope. The data are reported as the mean ± the standard deviation and were analyzed with Student's t test for equal variances. Statistical significance was accepted at a p value of less than 0.05. All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH Publication No. 86-23, revised 1985).9
Results Heart rate, blood pressure, and blood flow. Heart rate and blood pressures in the control and nimodipine groups were divided into preclamping values, values during clamping, and postclamping values (Table I). Blood pressure was further classified into mean proximal blood pressure and mean distal blood pressure. Differences in the mean proximal blood pressure were significant in the clamping period (p < 0.0001) and in the postclamping period (p = 0.0010) but not during the preclamping time. During clamping intervals, the nimodipine group had a mean proximal blood pressure of 87 ± 25 mm Hg, whereas the control group had an increased mean proximal blood pressure calculated at l l S ± 25 mm Hg. The
1 1 02
The Journal of Thoracic and Cardiovascular Surgery
Schittek et al.
Fig. 1. Statistical analysis of Tarlov grades for each animal showed significant difference (p = 0.0(03) between nimodipine and control groups.
postclamping mean proximal blood pressure was also higher in the control group: 56 ± 10 mm Hg compared with 48 ± 8 mm Hg in the nimodipine group. Measurements of mean distal blood pressures did not differ significantly between the two groups in the preclamping and clamping times, but there were significant differences (p = 0.0028) in mean distal blood pressures between groups in the postclamping period, with the nimodipine group having a mean distal blood pressure of 36 ± 7 mm Hg and the control group having a value of 48 ± 7 mm Hg. Heart rate did not significantly differ at any time between the two groups. CSF pressure. CSF pressures as demonstrated in Table I were significantly different in preclamping and clamping periods (p = 0.0012; P = 0.0029) but not after clamp release. The mean CSF pressure for nimodipinetreated animals just before clamping was 11.1 ± 2.0 mm Hg and during clamping it was 11.9 ± 2.4 mm Hg. The control group had a mean preclamping CSF pressure of 9.9 ± 1.9 mm Hg; during clamping, the CSF pressure measured 10.6 ± 3.0 mm Hg. Central spinal perfusion pressure. Central spinal perfusion pressure was calculated as the mean distal aortic blood pressure minus CSF pressure (Table I). Control animals tended to have higher central spinal perfusion pressures than nimodipine-treated animals during all
times, with a significant difference (p = 0.0106) noted only in the time after clamp release. The control group had a mean central spinal perfusion pressure of 39 ± 9 mm Hg during the postclamping period, whereas the nimodipine group had a lower mean central spinal perfusion pressure of 26 ± 8 mm Hg. Only one animal in the entire study had a consistently positive central spinal perfusion pressure. This animal was in the control group, with the central spinal perfusion pressure measuring +3.5 mmHg. Neurologic assessment. The Tarlov neurologic scores are shown in Fig. I. The majority of the control animals (6) had complete hind limb paralysis and were unable to stand or walk (Tarlov 0). One control animal retained slight movement of one hind limb, but this animal was unable to stand or walk (Tarlov 1). Five animals received a Tarlov score of 3. Ofthese, four had received nimodipine and one was in the control group. One nimodipine-treated animal (Tarlov 3) had morphologic evidence of spinal cord puncture by the CSF catheter. The puncture tract was ipsilateral to the side of hind limb weakness. Another nimodipine-treated animal scored Tarlov 3, but 36 hours after the operation its condition began to deteriorate. Deterioration was manifested by refusal of food, withdrawn behavior, and fever. Six animals received a Tarlov score of 4. Of these, five had received nimodipine
Volume 104 Number 4 October 1992
Spinal cord protection I I 0 3
Fig. 2. Complete cross sections of lumbar spinal cord from control animal No. 10 (A), nimodipine-treatedanimal No. II (B); and nimodipine-treated animal No.7 (0. (Original magnifications X25.) A, This cord is from a control animal that could not stand or walk. The majority of the neurons in the gray matter are pyknotic and eosinophilic (inset X450). Note that the inflammatory response is minimal. B, Spinal cord from a nimodipine-treated animal. Manyofthe neuronsare normalwithintact Nisslsubstance (N); however, someare swollen (8) yetviable(inset X450). Note the increased quantity of mononuclear inflammation (arrowheads). C, The condition of the animal with the lowest nimodipine bloodlevels deteriorated. Sectionsof spinalcord showedstrikingmononuclearcellinfiltration,loss of neurons, and swollen endothelial cells. (Holes in the dorsal and lateral columns are sectioningartifacts.) and only one was in the control group. The mean score for control animals was 0.889 ± 1.5 and the mean for the nimodipine-treated animals was 3.556 ± 0.527, with an overall significance of p = 0.0003. Morphologic evaluation of spinal cords. Spinal cord sections were reviewed without knowledge of whether the animal received nimodipine or carrier solution. Control animals had eosinophilic neuronal necrosis in the gray matter at all spinal cord levels although lumbosacral cord regions were more extensively involved than either thoracic or cervical levels (Fig. 2, A). Most notable, however, was the absence of an inflammatory response in spinal cord sections from control animals. The animals that received nimodipine had fewer eosinophilic neurons yet more swollen neurons within the ventral horns (Fig. 2, B).
Thus these neurons were damaged but still viable cells. In addition, there was a mononuclear cell infiltrate consisting of lymphocytes and macrophages as documented by immunohistochemical staining to antibodies against these cell types. These inflammatory cells appeared to be arriving through branches of the circumflex vessels coursing through the white matter to end in the central gray matter (Fig. 2, B). Lymphocytes and macrophages clustered about normal and pyknotic neurons, but not all neurons. Mitotic figures and migratory inflammatory cells were mostly perineuronal. The spinal cord from the nimodipine-treated animal whose condition deteriorated showed extensive inflammation, neuronophagia, dropout of neurons, and swollen endothelial cells (Fig. 2, C). Nimodipine serum measurements. Serum from the
The Journal of Thoracic and Cardiovascular Surgery
1 1 4 Schittek et al.
... Pg7 ...... Pg 9 ... Pigl' . . Pig16 -0- Pig 17 ... Pig18
Fig. 3. Nimodipine serum concentrations in six animals were measured at 30, 60,90, and 120 minutes. Animals (Nos. 9, 11, and 16)hadthedrugdosage increased to 41Lg/kg perminute before clamping. Thecondition oftheanimal with the lowest serum concentration (No.7) deteriorated after 36 hours.
animals was drawn 30, 60,90, and 120 minutes after the infusion was begun. Serum assays to determine if the animal had received nimodipine and the serum concentrations were performed by Miles, Inc., Elkhart, Indiana. Serum concentrations for six of the nimodipine-treated animals are shown in Fig. 3. Some of the animals (Nos. 9, 11, and 16) had the nimodipine dosage doubled from 2 to 4 J.lgjkg per minute just before clamping. This increase is reflected in the higher serum concentration measurements at 1 hour. Discussion
Nimodipine has been used in numerous experimentsto prevent or alleviate CNS injury, but the reported extent of neuroprotection is inconsistent. 10 This might be explained in part by the fact that the drug must be of sufficient concentration at the cell membrane before the hypoxiceventoccurs;consequently,the importance of the timing and dosage of drug administration are critical variables.I I The present study reemphasizes the importance of sufficientserum concentration, because the animal with the lowestbloodlevelof nimodipinebecame sick after 36 hours. This animal's spinal cord showed active myelitis similar to lesions reported in localized forms of experimental allergic encephalomyelitis produced with anoxic injury.F
The disparityof bloodpressurecontrolbetweenthe two groups is also a variable that must be addressed. Nimodipine-treated animals may have benefitedfrom the lower blood pressures during clamping and postclamping intervals. Lower blood pressure should diminish sympathetic stimulatory responses and increaseflow to the cord. Loweringof blood pressure and vasodilatation, functions directly related to the inhibition of calcium ion transfer into vascular smooth muscle, are additional effects of nimodipine that make this drug even more attractive in the present situation.I 3 The use of sodiumnitroprusside to control blood pressure in either group was purposely avoided because of questionable detrimental effects on spinal cord perfusion.lv 15 An important result of this experiment is that nimodipine protects CNS function despite negative central spinal perfusion pressures. Multiple prior studies have emphasized the effectsof positive central spinalperfusion pressuresand many studies have concentrated on manipulation of the central spinal perfusion pressures.v l'' Indeed, the control animal that had a central spinal perfusion pressuresof +3.5 mm Hg scoreda TarIov4; yet all of the nimodipine group had a negative central spinal perfusion pressure and walked. The present seriesof experimentsconcludesthat intravenous nimodipine, when given in sufficient concentra-
Volume 104 Number 4 October 1992
tions, can play an important neuroprotective role in controlled situations known to be linked to episodes of eNS hypoxia. Nevertheless, until the dynamic changes of the inflammatory response and the local immunologic events are clear, clinical trials should be discouraged. Further animal studies with longer survival times are now being conducted to determine the outcome of cord function with particular attention directed to the development and resolution of a localized myelitis syndrome. We thank Dr. H. Noda (Cardiovascular Research Fellow) for surgical assistance; William Hare for technical assistance; Dr. Marilyn Duke-Woodside for neurologic advice; Dr. Stephen Adams for pharmaceutical help; Dr. George Krol (Miles, Inc.) for the plasma and CSF nimodipine concentration assays, and Dr. Alexander Scriabine (Miles, Inc.) for making this series of experiments possible. Our consultant statistician was Peggy Lynn, MBA, Department of Biostatistics, Texas Heart Institute.
Spinal cord protection
8. 9. 10.
REFERENCES I. Blaisdell FW, Cooley DA. The mechanism of paraplegia after temporary thoracic aorta occlusion and its relationship to spinal fluid pressure. Surgery 1962;51:351-5. 2. Grubbs PE, Marini C, Toporoff B, et a!' Somatosensory evoked potentials and spinal cord perfusion pressure are significant predictors of postoperative neurologic dysfunction. Surgery 1988;104:216-23. 3. Berger L, Hakim AM. Calcium channel blockers correct the acidosis in ischemic brain without altering cerebral blood flow. Stroke 1988;19:1257-61. 4. Deshpande JK, SiesjoBK, Wieloch T. Calcium accumulation and neuronal damage in the rat hippocampus following cerebral ischemia. J Cereb Blood Flow Metab 1987; 7:89. 5. Farber JL, Chien KR, Mittnacht S. The pathogenesis of
I 10 5
irreversible cell injury in ischemia. Am J Pathol 1981; 102:271-81. Hakim AM, Gjedde A, Berger L, Popow D. Calcium channel activation in models of cerebral ischemia. Can J Neurol Sci 1988;15:190. Martiniak J, Saganova K, Chavko M. Free and peptidebound amino acids as indicators of ischemic damage of the rabbit spinal cord. J Neuropathol Exp NeuroI1991;50:7381. Tarlov 1M. Spinal cord compression: mechanism of paralysis and treatment. Springfield, Ill.: Thomas, 1957;147. 1986 Report of the AVMA Panel on Euthanasia. J Am Vet Moo Assoc 1986;188:252-68. Faden AL, Jacobs TP, Smith MT. Evaluation of calcium channel antagonist nimodipine in experimental spinal cord ischemia. J Neurosurg 1984;60:796-9. Levine S. Hyperacute, neutrophilic, and localized forms of experimental allergic encephalomyelitis: a review. Acta Neuropathol 1974;28:179-89. Uematsu D, Greenberg JH, Hickey WF, Reivich M. Nirnodipine attenuates both increase in cytosolic free calcium and histologic damage following focal cerebral ischemia and reperfusion in cats. Stroke 1989;20:1531-7. Scriabine A, Schurman T, Traber J. Pharmacological basis for the use of nimodipine in central nervous system disorders. FASEB 1989;3:1799-1806. Laschinger JC, Owen J, Rosenbloom M, Cox J, Kouchoukos NT. Detrimental effects of sodium nitroprusside on spinal cord motor tract perfusion during thoracic aortic cross-clamping. Surg Forum 1987;38:195-6. Marini CP, Grubbs PE, Toporoff B, et a!' Effect of sodium nitroprusside on spinal cord perfusion and paraplegia during aortic cross-clamping. Ann Thorac Surg 1989;47:37983. Woloszyn TT, Marini CP, Coons MS, eta!' Cerebrospinal fluid drainage and steroids provide better spinal cord protection during aortic crossc1amping than does either treatment alone. Ann Thorac Surg 1990;49:78-83.