The effect of nimodipine and dextran on axonal function and blood flow following experimental spinal cord injury

Full access

✓ There is evidence that posttraumatic ischemia is important in the pathogenesis of acute spinal cord injury (SCI). In the present study spinal cord blood flow (SCBF), measured by the hydrogen clearance technique, and motor and somatosensory evoked potentials (MEP and SSEP) were recorded to evaluate whether the administration of nimodipine and dextran 40, alone or in combination, could increase posttraumatic SCBF and improve axonal function in the cord after acute SCI. Thirty rats received a 53-gm clip compression injury on the cord at T-1 and were then randomly and blindly allocated to one of six treatment groups (five rats in each). Each group was given an intravenous infusion of one of the following over 1 hour, commencing 1 hour after SCI: placebo and saline; placebo and dextran 40; nimodipine 0.02 mg/kg and saline; nimodipine 0.02 mg/kg and dextran 40; nimodipine 0.05 mg/kg and saline; and nimodipine 0.05 mg/kg and dextran 40.

The preinjury physiological parameters, including the SCBF at T-1 (mean ± standard error of the mean: 56.84 ± 4.51 ml/100 gm/min), were not significantly different (p > 0.05) among the treatment groups. Following SCI, there was a significant decrease in the SCBF at T-1 (24.55 ± 2.99 ml/100 gm/min; p < 0.0001) as well as significant changes in the MEP recorded from the spinal cord (MEP-C) (p < 0.0001), the MEP recorded from the sciatic nerve (MEP-N) (p < 0.0001), and the SSEP (p < 0.002). Only the combination of nimodipine 0.02 mg/kg and dextran 40 increased the SCBF at T-1 (43.69 ± 6.09 ml/100 gm/min; p < 0.003) and improved the MEP-C (p < 0.0001), MEP-N (p < 0.04), and SSEP (p < 0.002) following SCI. With this combination, the changes in SCBF were significantly related to improvement in axonal function in the motor tracts (p < 0.0001) and somatosensory tracts (p < 0.0001) of the cord. This study provides quantitative evidence that an increase in posttraumatic SCBF can significantly improve the function of injured spinal cord axons, and strongly implicates posttraumatic ischemia in the pathogenesis of acute SCI.

Abstract

✓ There is evidence that posttraumatic ischemia is important in the pathogenesis of acute spinal cord injury (SCI). In the present study spinal cord blood flow (SCBF), measured by the hydrogen clearance technique, and motor and somatosensory evoked potentials (MEP and SSEP) were recorded to evaluate whether the administration of nimodipine and dextran 40, alone or in combination, could increase posttraumatic SCBF and improve axonal function in the cord after acute SCI. Thirty rats received a 53-gm clip compression injury on the cord at T-1 and were then randomly and blindly allocated to one of six treatment groups (five rats in each). Each group was given an intravenous infusion of one of the following over 1 hour, commencing 1 hour after SCI: placebo and saline; placebo and dextran 40; nimodipine 0.02 mg/kg and saline; nimodipine 0.02 mg/kg and dextran 40; nimodipine 0.05 mg/kg and saline; and nimodipine 0.05 mg/kg and dextran 40.

The preinjury physiological parameters, including the SCBF at T-1 (mean ± standard error of the mean: 56.84 ± 4.51 ml/100 gm/min), were not significantly different (p > 0.05) among the treatment groups. Following SCI, there was a significant decrease in the SCBF at T-1 (24.55 ± 2.99 ml/100 gm/min; p < 0.0001) as well as significant changes in the MEP recorded from the spinal cord (MEP-C) (p < 0.0001), the MEP recorded from the sciatic nerve (MEP-N) (p < 0.0001), and the SSEP (p < 0.002). Only the combination of nimodipine 0.02 mg/kg and dextran 40 increased the SCBF at T-1 (43.69 ± 6.09 ml/100 gm/min; p < 0.003) and improved the MEP-C (p < 0.0001), MEP-N (p < 0.04), and SSEP (p < 0.002) following SCI. With this combination, the changes in SCBF were significantly related to improvement in axonal function in the motor tracts (p < 0.0001) and somatosensory tracts (p < 0.0001) of the cord. This study provides quantitative evidence that an increase in posttraumatic SCBF can significantly improve the function of injured spinal cord axons, and strongly implicates posttraumatic ischemia in the pathogenesis of acute SCI.

There is evidence that spinal cord injury (SCI) in the acute phase is caused by two separate mechanisms: the initial mechanical damage and secondary changes due to vascular or biochemical effects.2,6,36,37 Indeed, measurements of spinal cord blood flow (SCBF)2,9,12,18,19,35,47 as well as microangiographic studies4,14,47 have shown ischemia at and extending away from the injury site. With recording of SCBF and motor and somatosensory evoked potentials (MEP's and SSEP's) from the cord with intraspinal microelectrodes, Fehlings, et al.,9,12 recently characterized the relationships among the severity of SCI, the extent of posttraumatic cord ischemia, and the resultant changes in axonal function. Their data showed that the reduction in posttraumatic SCBF varied as a linear function of injury severity and that axonal dysfunction after SCI was related both to the force of cord injury and to the degree of ischemia. These data suggested that it may be possible to improve axonal function after SCI by increasing the posttraumatic SCBF. To date, this hypothesis has not been validated.

There is considerable evidence that Ca++ ions play a key role in the pathogenesis of posttraumatic ischemia and ischemic cell death.1,5 For example, ischemic depolarization of cell membranes is associated with an intracellular shift of Ca++ which promotes smooth-muscle contraction and vasospasm, impairs mitochondrial function, and enhances the production of vasoactive prostanoids.20,22,39 Nimodipine, a dihydropyridine calcium channel blocker, has recently been shown to increase posttraumatic SCBF,19 although the effect on recovery of posttraumatic axonal function is not known. Guha, et al.,19 reported that the effect of nimodipine on postinjury SCBF required the concomitant administration of the vasopressor adrenaline to correct posttraumatic hypotension and to prevent further lowering of the mean arterial blood pressure (MABP) by peripheral vasodilatation. However, there is evidence that catecholamines may exacerbate neuronal injury after cord trauma,28,31 hence an alternative method of maintaining the MABP would theoretically enhance the possibility of achieving recovery of axonal function. Recently, Wallace and Tator48 reported that hypervolemic hemodilution with dextran 40 significantly increased MABP and SCBF after SCI in rats; therefore, dextran 40 would appear to be a suitable adjunct to nimodipine. In the present study, the combination of MEP and SSEP recordings and SCBF measurement has been used to examine whether the administration of nimodipine and dextran 40, alone or together, could increase posttraumatic SCBF and improve axonal function in the cord after acute experimental SCI.

Materials and Methods
Operative Procedures

Thirty adult Wistar rats* (mean weight ± standard error of the mean (SEM): 341 ± 22 gm) were anesthetized by intraperitoneal injection of 75 mg/kg α-chloralose and 525 mg/kg urethane. After tracheostomy and insertion of bilateral femoral artery and vein catheters, a right parietal craniectomy was performed to expose the hindlimb sensorimotor cortex.8,21 A laminectomy was made from C-6 to T-2, and each rat received a 1-minute extradural clip compression injury34 of the cord at T-1 with a modified aneurysm clip exerting a force of 53 gm, an injury which causes complete paraplegia.27 Neuromuscular blockade was accomplished with intravenous pancuronium bromide (0.6 mg every hour) and the rats were mechanically ventilated and maintained by a 2:1 mixture of oxygen and nitrous oxide. The MABP, electrocardiogram, rectal temperature, hematocrit, and arterial blood gas pressures were monitored. The temperature was maintained at 37° to 38°C with a heating pad, the arterial blood gases were maintained in a physiological range (pO2 > 100 mm Hg; pCO2 > 35 to 45 mm Hg), and the pH was held steady at 7.35 to 7.45.

Measurement of Spinal Cord Blood Flow

The SCBF was measured by the hydrogen clearance technique using a microcomputer-based recording system.32 Platinum-iridium microelectrodes with a 10-µm tip diameter were stereotactically inserted into the dorsal column of the cord, 0.5 mm lateral to the midline, at T-1 (injury site), C-6, and T-10 to a depth of 500 µm. The SCBF was recorded from these sites 1 hour before and at 1, 2, 3, and 4 hours after SCI (Fig. 1). Linear regression was used on-line to compare the presaturation polarographic baseline level with the desaturation baseline level. Rows with a drift in baseline measurement of more than 5% were discarded. The initial slope index method23 was used to calculate the SCBF from the logarithmically transformed desaturation curves.

Fig. 1.
Fig. 1.

A schematic representation of the experimental protocol. For description see text.

Motor and Somatosensory Evoked Potentials

The MEP's and SSEP's were recorded concomitantly with the SCBF (Fig. 1) 1 hour before and at 1, 2, 3, and 4 hours after SCI. The MEP's were elicited by applying anodal stimuli to the SMC via a 0.8-mm platinum ball electrode, referenced to an Ag/AgCl disc electrode under the hard palate. The MEP's from the spinal cord (MEP-C) were recorded from two platinum-iridium microelectrodes stereotactically inserted to a depth of 500 µm in the dorsal columns and 0.5 mm lateral to the dorsal vein; at T-10 the MEP's from the sciatic nerve (MEP-N) were recorded from a bipolar electrode on the left sciatic nerve. The stimulus parameters were: intensity 10 mA, pulse duration 50 µsec, and presentation rate 8.2 Hz for the MEP-C; and intensity 30 mA, pulse duration 300 µsec, and presentation rate 4.2 Hz for the MEP-N. A total of 512 MEP-C and 50 MEP-N responses were recorded at a bandwidth of 30 to 3000 Hz, averaged, and replicated. The spinal evoked potential (SEP) recorded from the microelectrodes at T-10 and the SSEP recorded from the platinum electrode over the sensorimotor cortex were elicited by applying cathodal stimuli (intensity 10 mA, pulse duration 50 µsec, and presentation rate 8.2 Hz) to the left sciatic nerve. A total of 1024 SEP and SSEP responses were recorded at a bandwidth of 30 to 3000 Hz, averaged, and replicated. The latency and amplitude of the evoked potential (EP) peaks were scored by two independent blinded observers using previously established criteria;7,10,11 the scoring methodology used to label peaks is illustrated for the MEP-C and SSEP in Fig. 2. The MEP-N and SEP peaks were labeled as positive (P) and negative (N) deflections in a manner similar to the SSEP.

Fig. 2.
Fig. 2.

Left: Computer-derived grand mean of the motor evoked potentials recorded from the spinal cord (MEP-C) of 20 normal rats at T-10 (average of 20,480 responses). The normal MEP-C consists of a series of positive deflections: an initial d (direct) wave which reflects direct pyramidal cell activation, and a series of subsequent i waves which result from indirect activation of pyramidal cells by cortical inter-neurons as well as activation of nonpyramidal tracts. Right: Computer-derived grand mean of somatosensory evoked potentials (SSEP) recorded from 16 normal rats from the sensorimotor cortex following sciatic nerve stimulation (average of 32,768 responses). The normal rat SSEP consists of four positive (P) and three negative (N) peaks.

Experimental Protocol

The experimental protocol is schematically shown in Fig. 1. Following the preinjury recordings, the rats each received a 53-gm clip compression injury of the cord at T-1. The microelectrodes were removed just prior to injury and reinserted immediately thereafter. Following the postinjury recordings of SCBF and EP's, the rats were randomly and blindly assigned to one of the following groups (five animals in each): placebo and saline; placebo and dextran 40; nimodipine 0.02 mg/kg and saline; nimodipine 0.02 mg/kg and dextran 40; nimodipine 0.05 mg/kg and saline; and nimodipine 0.05 mg/kg and dextran 40. Nimodipine was supplied as a 0.02% stock solution in a polyethylene glycol, ethanol, and water diluent. The stock solution was diluted with placebo (diluent alone) to obtain the appropriate concentrations of nimodipine. The infusion syringes were protected from light with aluminum foil, and a sodium vapor lamp was used during dilution and delivery of the drug due to the photosensitivity of nimodipine to normal light.40 Two Harvard infusion pumps were used to deliver a fixed volume of drug at a constant rate over 1 hour via separate femoral venous access sites: one pump for nimodipine or placebo (3.0 ml) and the second for dextran 40 or saline (7.0 ml). Postinfusion SCBF and EP recordings were made immediately following the infusion and at 1 and 2 hours after the cessation of drug delivery.

Histological Assessment and Quantitation of Cord Hemorrhages

At the conclusion of the experiment, the microelectrode sites were electrolytically marked (10 µA direct current for 15 sec) and the rats were transcardially perfused with 10% formalin. Serial 8-µm cross sections of the injury site and electrode sites were stained with Luxol fast blue. The sections were examined to confirm the position of the microelectrodes and to quantitate the amount of hemorrhage at the injury site. For the latter, 130 consecutive cross sections through the injury site (a 1.04-mm segment of fixed cord) containing the maximum amount of hemorrhage were selected by microscopy, and quantitative evaluation was undertaken on every 10th section with an interactive image analysis system§ in a “blinded” fashion. The sections were displayed at × 25 magnification on a television screen by a camera interfaced to a photomicroscope and digitized to achieve a gray scale range of 0 to 255. A binary discrimination algorithm was created to interactively separate hemorrhagic from nonhemorrhagic tissue on the basis of their different levels of gray. The area of hemorrhage and the total cross-sectional area of each slice were quantitated from the digitized image. The volume of hemorrhage and the total volume of the cord were calculated from the 13 sections by means of an integration algorithm based on Simpson's theorum:41 Vx = ⅓ dx(AO + 4A1 + 2A2 + 4A3 + 2A4 + … + 2An-2 + 4An-1 + An), where Vx = volume of cord or hemorrhage, dx = interval between sections = 80 µm, AO = area of first section analyzed, and n = 12.

Statistical Analysis

The data were evaluated in a blinded fashion and the treatment code was not broken until the conclusion of the statistical analysis. Interobserver differences in the analysis of the EP data were assessed by two-tailed paired t-tests. Each dependent measure was analyzed by multivariate analysis of variance (MANOVA) for repeated measures.38 Post hoc comparisons at each time point between groups were made by univariate one-way analysis of variance (ANOVA), and Tukey's HSD test, whereas post hoc comparisons within each treatment group were made by two-way ANOVA and Tukey's HSD test. The quantitative assessment of cord volume (VCORD), hemorrhage volume (VHEM), and percent hemorrhage by volume (VHEM/VCORD × 100%) was compared between groups by one-way ANOVA. The results have been expressed as the mean ± SEM, and differences were considered significant at p < 0.05.

Results
Physiological Parameters

The mean preinjury physiological parameters including the MABP (129 ± 2.3 mm Hg), hematocrit (46.2% ± 1.6%), heart rate (545 ± 7 beats/min), pH (7.34 ± 0.02), pO2 (168.6 ± 5.8 mm Hg), and pCO2 (38.7 ± 0.6 mm Hg) were not significantly different (p > 0.05) among the groups (Table 1). Furthermore, multivariate analysis showed that the pH, pO2, and pCO2 remained constant throughout the experiment. After SCI, there was a significant decrease in the MABP to 60.2 ± 4.7 mm Hg (F = 127.97, df = 1,23, p < 0.0001) which was similar among all groups (F = 0.58, df = 5,24, p > 0.05). However, the MABP was significantly increased (F = 3.55, df = 5,24, p < 0.02) following infusion of placebo and saline (74.0 ± 8.3 mm Hg), placebo and dextran 40 (85.0 ± 7.7 mm Hg), and nimodipine 0.02 mg/kg and dextran 40 (88.4 ± 8.4 mm Hg). At 1 hour after drug infusion, the MABP was again similar among the six treatment groups (F = 1.25, df = 5,22, p > 0.05). Although the postinjury and postinfusion hematocrit was similar among the groups, at 1 hour postinfusion the mean hematocrit of the three groups that received dextran alone or in combination with nimodipine (24.4% ± 1.5%) was significantly lower (F = 4.14, df = 5,22, p < 0.02) than that of the other three groups (33.8% ± 0.3%). Multivariate analysis revealed that the heart rate was similar among the six treatment groups at all times during the experiment (F = 1.47, df = 20,40.75, p > 0.05); however, it decreased significantly after drug infusion in all groups (F = 19.11, df = 4,25, p < 0.0001), and then remained constant for the remainder of the experiment.

TABLE 1

Physiological parameters*

Time & Experimental GroupMABP (mm Hg)Heart Rate (bts/min)Hematocrit (%)pHpO2 (mm Hg)pCO2 (mm Hg)
preinjury 
 placebo & saline128.4 ± 4.7 528 ± 1547.5 ± 3.17.33 ± 0.01164.1 ± 14.737.7 ± 1.5
 placebo & dextran 40127.0 ± 4.0 552 ± 2040.0 ± 1.67.32 ± 0.01170.8 ± 17.539.5 ± 1.9
 nimodipine 0.02 mg/kg & saline132.8 ± 8.7 546 ± 644.9 ± 1.47.33 ± 0.01186.5 ± 12.239.9 ± 2.2
 nimodipine 0.02 mg/kg & dextran 40132.0 ± 6.7 552 ± 2050.2 ± 1.87.38 ± 0.02183.2 ± 18.037.7 ± 1.6
 nimodipine 0.05 mg/kg & saline129.2 ± 5.6 540 ± 2844.8 ± 2.77.36 ± 0.01154.8 ± 16.637.6 ± 0.6
 nimodipine 0.05 mg/kg & dextran 40128.0 ± 6.0 553 ± 749.8 ± 2.27.33 ± 0.02152.2 ± 7.339.5 ± 1.7
postinjury 
 placebo & saline58.6 ± 6.4 510 ± 939.0 ± 2.67.33 ± 0.01167.3 ± 12.238.6 ± 2.4
 placebo & dextran 4056.2 ± 6.0 540 ± 2738.6 ± 1.27.35 ± 0.01159.7 ± 8.838.6 ± 1.5
 nimodipine 0.02 mg/kg & saline61.2 ± 6.5 522 ± 1238.8 ± 1.27.35 ± 0.01171.9 ± 21.339.4 ± 1.4
 nimodipine 0.02 mg/kg & dextran 4069.0 ± 3.7 564 ± 1144.2 ± 2.27.34 ± 0.01179.4 ± 6.236.4 ± 0.7
 nimodipine 0.05 mg/kg & saline56.6 ± 7.3 546 ± 3138.4 ± 2.87.35 ± 0.02151.9 ± 16.437.9 ± 1.3
 nimodipine 0.05 mg/kg & dextran 4059.8 ± 6.4 536 ± 2243.1 ± 2.37.36 ± 0.02159.1 ± 10.137.9 ± 1.4
postinfusion 
 placebo & saline74.0 ± 8.3 473 ± 2733.5 ± 1.97.40 ± 0.02174.3 ± 8.439.0 ± 2.4
 placebo & dextran 4085.0 ± 7.7 486 ± 3326.8 ± 4.07.38 ± 0.02166.7 ± 19.540.0 ± 1.6
 nimodipine 0.02 mg/kg & saline64.6 ± 6.8 420 ± 1633.4 ± 2.17.37 ± 0.02181.8 ± 31.439.6 ± 1.5
 nimodipine 0.02 mg/kg & dextran 4088.4 ± 8.4 492 ± 2829.5 ± 2.27.36 ± 0.01180.8 ± 14.643.1 ± 0.4
 nimodipine 0.05 mg/kg & saline52.6 ± 6.4 396 ± 2031.3 ± 2.07.36 ± 0.02157.6 ± 21.038.7 ± 1.6
 nimodipine 0.05 mg/kg & dextran 4064.4 ± 5.2 442 ± 2829.8 ± 4.87.35 ± 0.01164.9 ± 10.738.9 ± 1.5
1 hr postinfusion 
 placebo & dextran 4061.8 ± 7.3 480 ± 3122.4 ± 2.37.40 ± 0.01184.8 ± 17.338.5 ± 1.5
 nimodipine 0.02 mg/kg & saline58.2 ± 6.6 450 ± 1933.8 ± 1.77.37 ± 0.03171.7 ± 38.838.6 ± 0.9
 nimodipine 0.02 mg/kg & dextran 4075.6 ± 7.8 504 ± 2623.6 ± 1.47.39 ± 0.02177.0 ± 21.241.3 ± 0.9
 nimodipine 0.05 mg/kg & saline50.8 ± 9.1 414 ± 2033.5 ± 1.57.39 ± 0.02143.9 ± 15.338.2 ± 2.1
 nimodipine 0.05 mg/kg & dextran 4064.4 ± 6.7 467 ± 3527.3 ± 3.17.37 ± 0.03144.4 ± 13.139.2 ± 1.7
2 hrs postinfusion 
 placebo & saline55.3 ± 5.3 480 ± 3528.0 ± 1.07.38 ± 0.03162.4 ± 12.437.1 ± 1.0
 placebo & dextran 4052.0 ± 8.4 468 ± 3122.0 ± 1.87.39 ± 0.02166.7 ± 21.836.5 ± 0.5
 nimodipine 0.02 mg/kg & saline60.0 ± 7.3 450 ± 1931.5 ± 3.57.38 ± 0.02168.9 ± 28.040.4 ± 0.9
 nimodipine 0.02 mg/kg & dextran 4064.0 ± 2.0 510 ± 3023.0 ± 2.47.41 ± 0.01179.9 ± 8.140.4 ± 2.1
 nimodipine 0.05 mg/kg & saline54.0 ± 10.7 420 ± 1232.1 ± 1.67.37 ± 0.03141.9 ± 7.838.3 ± 2.0
 nimodipine 0.05 mg/kg & dextran 4062.3 ± 5.9 530 ± 2626.8 ± 3.17.41 ± 0.02161.4 ± 13.337.5 ± 2.0

Values are means ± standard error of the means. MABP = mean arterial blood pressure.

Spinal Cord Blood Flow

The SCBF data recorded from the injury site at T-1, adjacent to the injury site at C-6, and distal to the injury site at T-10, are summarized in Figs. 3, 4, and 5, respectively. The preinjury SCBF at T-1 was similar among the six treatment groups (mean: 56.83 ± 3.82 ml/100 gm/min; F = 1.20, df = 5,24, p > 0.05). Multivariate analysis showed that the SCBF at T-1 was influenced by SCI and drug infusion (F = 36.48, df = 4,20, p < 0.0001). After SCI, there was a significant decrease in the SCBF at T-1 to 24.55 ± 2.99 ml/100 gm/min (F = 69.99, df = 1,23, p < 0.0001), which was similar for all groups (F = 0.69, df = 5,24, p > 0.05). After drug infusion, there was a significant difference in the SCBF at T-1 among the treatment groups (F = 4.92, df = 5,24, p < 0.004). Post hoc analysis showed that the SCBF at T-1 in the group treated with nimodipine 0.02 mg/kg and dextran 40 (43.69 ± 6.09 ml/100 gm/min) was significantly greater than that of the other five groups (mean 24.63 ± 1.5 ml/100 gm/min). Although the mean SCBF in the group treated with nimodipine 0.05 mg/kg and dextran 40 increased from 22.68 ± 2.78 to 29.88 ± 4.46 ml/100 gm/min, post hoc analysis showed that this increase was not significant. At 1 hour postinfusion, the SCBF at T-1 in the group that received nimodipine 0.02 mg/kg and dextran 40 had declined to 30.54 ± 3.47 ml/100 gm/min, but was still significantly greater than that of the other five groups (mean: 17.3 ± 1.5 ml/100 gm/min; F = 8.36, df = 5,24, p < 0.0001). By 2 hours postinfusion, however, the SCBF at T-1 was again similar in all groups (mean: 12.29 ± 0.91 ml/100 gm/min; F = 1.42, df = 5,24, p > 0.05).

Fig. 3.
Fig. 3.

Spinal cord blood flow (SCBF) recorded from the injury site at T-1 in 30 rats. The preinjury SCBF (56.83 ± 3.82 ml/100 gm/min) did not differ significantly among the groups. Following spinal cord injury, there was a significant decrease in SCBF in all groups (mean ± standard error of the mean: 24.55 ± 2.99 ml/100 gm/min). In the group treated with nimodipine (Nim) 0.02 mg/kg and dextran 40, there was a significant increase in SCBF immediately after infusion (43.69 ± 6.09 ml/100 gm/min) and at 1 hour postinfusion (30.54 ± 3.47 ml/100 gm/min) in comparison with the other groups (mean: 17.30 ± 1.50 ml/100 gm/min). At 2 hours postinfusion, the SCBF was similar for all groups.

Fig. 4.
Fig. 4.

Spinal cord blood flow (SCBF) data recorded adjacent to the injury site from the cord at C-6 in 30 rats. The preinjury SCBF was similar in all groups (mean ± standard error of the mean: 51.88 ± 3.72 ml/100 gm/min). With injury, the SCBF at C-6 declined significantly in each group (30.43 ± 1.99 ml/100 gm/min). Infusion of dextran alone or in combination with nimodipine (Nim) at 0.02 mg/kg or 0.05 mg/kg resulted in a significant increase in SCBF in comparison with the postinjury values. At 1 and 2 hours postinfusion the improvement in SCBF at C-6 in these groups had abated and there was no difference in SCBF among the six treatment groups.

Fig. 5.
Fig. 5.

Spinal cord blood flow (SCBF) data recorded distal to the injury site from the cord at T-10 in 30 rats. The SCBF remained constant in all groups throughout the experiment with the exception of the group treated with nimodipine (Nim) 0.02 mg/kg and dextran 40. In the latter, there was a significant increase in the postinfusion SCBF at T-10 to 81.49 ± 8.45 ml/100 gm/min (mean ± standard error of the mean).

The preinjury SCBF measured from the cord at C-6 was similar among the groups (mean 51.88 ± 3.72 ml/100 gm/min; F = 0.33, df = 5,24, p > 0.05). Based on the MANOVA, the SCI at T-l and drug infusion significantly affected the SCBF at C-6 (F = 26.11, df = 4,20, p < 0.0001). Indeed, after SCI, there was a significant decline in the SCBF at C-6 to 30.43 ± 1.99 ml/100 gm/min (F = 41.79, df = 1,23, p < 0.0001), although the SCBF at C-6 remained similar among the treatment groups (F = 0.80, df = 5,24, p > 0.05). Furthermore, MANOVA showed that drug infusion resulted in a significant increase in the SCBF at C-6 (F = 7.34, df = 1,23, p < 0.03). Two-way ANOVA revealed that the SCBF at C-6 was significantly increased in the groups treated with placebo and dextran 40 (46.66 ± 9.70 ml/100 gm/min; F = 4.05, df = 4,16, p < 0.02), nimodipine 0.02 mg/kg and dextran 40 (53.97 ± 9.52 ml/100 gm/min; F = 6.65, df = 4,16, p < 0.004), and nimodipine 0.05 mg/kg and dextran 40 (36.08 ± 8.93 ml/100 gm/min; F = 4.20, df = 4,16, p < 0.03). Moreover, there was no significant difference in the postinfusion SCBF among these groups (F = 1.82, df = 5,24, p > 0.05). Post hoc analysis with Tukey's HSD test revealed a significant reduction in the SCBF at C-6 at 1 hour (24.09 ± 1.85 ml/100 gm/min) and 2 hours (21.25 ± 2.32 ml/100 gm/min) following termination of the infusion in all groups.

The preinjury SCBF at T-10 (mean 56.96 ± 4.73 ml/100 gm/min) was similar in all treatment groups (F = 0.66, df = 5,24, p > 0.05). Although the SCBF at T-10 tended to be lower after cord injury due to the decline in MABP, this change was not significant (p > 0.05) in any group. Indeed, two-way ANOVA showed that the SCBF at T-10 did not change significantly (p > 0.05) in any group at any point during the course of the experiment, with the notable exception of the group that received nimodipine 0.02 mg/kg and dextran 40. In the latter group, the postinfusion SCBF at T-10 2 hours after injury (81.49 ± 8.45 ml/100 gm/min) was significantly greater than at other times in the experiment (F = 5.00, df = 4,16, p < 0.02).

Electrophysiological Analysis

The interobserver differences in the scoring of the latency (t = 1.22, p > 0.05) and amplitude (t = 1.01, p > 0.05) of the EP peaks were not significant. Representative EP tracings from a control rat treated with placebo and saline and a rat treated with the combination of nimodipine 0.02 mg/kg and dextran 40 are presented in Fig. 6 upper and lower, respectively. Preinjury, the latency of each MEP-C peak was similar (p > 0.05) among the six groups. With the labeling system shown in Fig. 2, MANOVA revealed that the amplitude of the d (F = 17.94, df = 4,20, p < 0.0001), il (F = 25.06, df = 4,20, p < 0.0001), i2 (F = 21.06, df = 4,20, p < 0.0001), i3 (F = 4.17, df = 4,19, p < 0.002), and i4 (F = 9.51, df = 4,10, p < 0.004) peaks changed during the course of the experiment. Two-way ANOVA showed that SCI resulted in a significant reduction in the amplitude of each of these peaks in all treatment groups (p < 0.01). For example (as illustrated in Fig. 7), the d wave amplitude was significantly decreased in the groups receiving placebo and saline (F = 4.28, df = 4,16, p < 0.02), placebo and dextran 40 (F = 3.72, df = 4,16, p < 0.04), nimodipine 0.02 mg/kg and saline (F = 10.05, df = 4,16, p < 0.0004), nimodipine 0.02 mg/kg and dextran 40 (F = 25.73, df = 4,16, p < 0.0001), nimodipine 0.05 mg/kg and saline (F = 7.63, df = 4,16, p < 0.002), and nimodipine 0.05 mg/kg and dextran 40 (F = 4.07, df = 4,16, p < 0.03). Indeed, the reduction in d wave amplitude with injury was similar in all groups (F = 1.33, df 5,24, p > 0.05). With the exception of the group treated with nimodipine 0.02 mg/kg and dextran 40, there was no recovery of the MEP-C (Figs. 7 and 8). The latter treatment resulted in a significant improvement in axonal function as evidenced by an increase in the amplitude of the d (F = 3.96, df = 5,24, p < 0.01), il (F = 3.41, df = 5,24, p < 0.02), and i2 (F = 7.59, df = 5,24, p < 0.0002) peaks (Figs. 7 and 8). Furthermore, the increase in the d wave amplitude persisted at 1 hour (F = 3.37, df = 5,24, p < 0.03) and 2 hours (F = 9.94, df = 5,24, p < 0.0005) following infusion of nimodipine 0.02 mg/kg and dextran 40.

Fig. 6.
Fig. 6.

Upper: Representative evoked potential recordings from a control rat which received placebo and saline. After cord injury, the motor evoked potentials recorded at the spinal cord (MEP-C) and at the sciatic nerve (MEP-N) and the somatosensory evoked potentials (SSEP) were abolished and did not recover postinfusion. The spinal evoked potentials (SEP) recorded from the cord at T-10 (caudal to the injury site) following stimulation of the sciatic nerve were unchanged pre- and postinjury, confirming the physiological integrity of the experimental preparation. Lower: Representative evoked potential recordings from a rat treated with nimodipine 0.02 mg/kg and dextran 40. After cord injury, the MEP-C and MEP-N were abolished and the amplitude of the SSEP was significantly attenuated. Postinfusion, there was a significant recovery of the MEP-C, MEP-N, and SSEP. The SEP recorded from the cord at T-10 (caudal to the injury site) following stimulation of the sciatic nerve were unchanged pre- and postinjury, confirming the physiological integrity of the experimental preparation.

Fig. 7.
Fig. 7.

The amplitude of the d wave of the motor evoked potential recorded at the spinal cord and plotted as a function of time. Preinjury, the d wave amplitude was similar in all groups. After spinal cord injury, there was a significant reduction in d wave amplitude in all groups. With the exception of the groups treated with nimodipine (Nim) 0.02 mg/kg or 0.05 mg/kg in combination with dextran 40, there was a progressive attenuation of d wave amplitude after injury. After treatment with nimodipine 0.02 mg/kg and dextran 40, there was a significant recovery of d wave amplitude which persisted until the conclusion of recording. In the group that received nimodipine 0.05 mg/kg and dextran 40, there was transient delayed recovery of d wave amplitude at 1 hour postinfusion.

Fig. 8.
Fig. 8.

Upper: Superimposed postinfusion motor evoked potential (MEP) traces from the spinal cord of rats treated with placebo and saline (left) or nimodipine 0.02 mg/kg and dextran 40 (right). Lower: The respective grand mean MEP's are shown. These data illustrate a significant return of axonal function in the motor tracts of the cord after infusion of nimodipine 0.02 mg/kg and dextran 40.

Preinjury, the latency of each MEP-N peak did not differ significantly among the groups. It was shown by MANOVA that SCI and drug infusion significantly altered the amplitude of the N, (F = 39.26, df = 4,20, p < 0.0001), P1 (F = 63.67, df = 4,20, p < 0.0001), and N2 (F = 93.81, df = 4,20, p < 0.0001) peaks. Indeed, with SCI there was a significant decrease in the amplitude of each of these peaks (N1: F = 79.56, df = 1,23, p < 0.0001; P1: F = 70.18, df = 1,23, p < 0.0001; and N2: F = 153.74, df = 1,20, p < 0.0001) (Fig. 6). Oneway ANOVA revealed that infusion of nimodipine 0.02 mg/kg and dextran 40 was associated with a significant recovery of the P1 (F = 4.98, df = 5,24, p < 0.01) and N2 (F = 2.72, df = 5,24, p < 0.05) peaks, whereas no improvement was seen in the other groups. Furthermore, the improvement in the amplitude of P1 (F = 5.38, df = 5,24, p < 0.005) and N2 (F = 4.42, df = 5,24, p < 0.02) persisted at 1 hour following infusion but was no longer apparent at 2 hours after the conclusion of drug administration (P1: F = 0.91, df = 5,24, p > 0.05; N2: F = 1.09, df = 5,24, p > 0.05).

Preinjury, the latency and amplitude of each SSEP peak were not significantly different among the six treatment groups. Multivariate analysis showed that the amplitude of these peaks changed significantly over the course of the experiment (P1: F = 14.22, df = 4,7, p < 0.003; N2: F = 6.44, df = 4,8, p < 0.02; P2: F = 54.33, df = 4,20, p < 0.0001; N2: F = 79.25, df = 4,20, p < 0.0001; P3: F = 18.63, df = 4,20, p < 0.0001; N3: F = 8.96, df = 4,18, p < 0.003; P4: F = 37.82, df = 4,18, p < 0.0001) and that SCI resulted in a significant attenuation of the amplitude of each of these peaks (P1: F = 30.97, df = 1,10, p< 0.0003; N1:F= 13.96, df = 1,11, p < 0.005; P2: F = 102.07, df = 1.23, p < 0.0001; N2: F = 161.10, df = 1,23, p < 0.0001; P3: F = 31.05, df = 1,23, p < 0.0001; N3: F = 11.45, df = 1,21, p < 0.0004; and P4: F = 51.49, df = 1,21, p < 0.0001). One-way ANOVA revealed that, following infusion of nimodipine 0.02 mg/kg and dextran 40, there was a significant improvement in the amplitude of the P2 (F = 4.04, df = 5,24, p < 0.01) and N2 (F = 7.43, df = 5,24, p < 0.001) peaks of the SSEP (Fig. 9), whereas no improvement was seen with the other treatment combinations. Two-way ANOVA showed that the P2 amplitude remained significantly increased (F = 12.70, df = 4,16, p < 0.001) at 1 and 2 hours following infusion, whereas the N2 amplitude was significantly increased at 1 hour (F = 24.75, df = 4,16, p < 0.001) but not at 2 hours postinfusion.

Fig. 9.
Fig. 9.

The amplitude of the N2 wave of the somatosensory evoked potentials plotted as a function of time. Preinjury, the N2 amplitude was similar for all groups. After spinal cord injury, there was a significant reduction of N2 wave amplitude in all groups. With the exception of the group treated with nimodipine (Nim) 0.02 mg/kg and dextran 40, there was a progressive attenuation of N2 amplitude after injury. Following treatment with nimodipine 0.02 mg/kg and dextran 40, however, there was a significant recovery of the N2 amplitude which persisted at 1 hour but was no longer apparent at 2 hours postinfusion.

One-way ANOVA revealed that the latency and amplitude of the SEP peaks were similar (p > 0.05) among the six groups. Furthermore, multivariate analysis of the SEP's revealed that neither their latency nor their amplitude changed significantly (p > 0.05) over the course of the experiment, thus confirming that the EP changes described above reflect specific changes in axonal function at the injury site and not changes in axonal physiology due to an unstable experimental preparation.

Multiple regression revealed that the changes in axonal function were significantly correlated with the changes in SCBF. Indeed, the amplitude of the d (F = 90.47, df = 1,149, p < 0.0001) (Fig. 10) and i2 (F = 3.62, df = 1,149, p < 0.05) peaks of the MEP-C and the amplitude of the P2 (F = 102.43, df = 1,149, p < 0.0001) and N2 (F = 117.52, df = 1,149, p < 0.0001) peaks of the SSEP's were significantly related to the changes in SCBF.

Fig. 10.
Fig. 10.

Relationship between the d wave amplitude of the motor evoked potentials recorded at the spine and the spinal cord blood flow (SCBF) in 30 rats (five observations per rat: 150 observations in total).

Histological Assessment and Quantitation of Cord Hemorrhages

Examination of the electrode tracks revealed that the tips of the microelectrodes were located at the base of the dorsal columns, near the gray/white matter interface in all cases. The C-6 electrode tracks were located 6.1 ± 0.4 mm rostral to the tracks in the cord at T-1. The results of the quantitative assessment of the volume of hemorrhage in a 1.04-mm section of cord encompassing the injury site are summarized in Table 2. The total cord volume ranged from 6.59 ± 0.50 to 8.97 +1.13 cu mm and was not significantly different (F = 2.14, df = 5,24, p > 0.05) among the groups. Furthermore, neither the total volume of hemorrhage at the injury site (F = 2.12, df = 5,24, p > 0.05) nor the percent hemorrhage by volume (F = 2.12, df = 5,24, p > 0.05) varied significantly among treatment groups, indicating that pharmacological improvement of posttraumatic SCBF did not exacerbate the degree of hematomyelia.

TABLE 2

Quantitation of cord hemorrhage*

Experimental GroupCord Volume (cu mm)Hemorrhage Volume (cu mm)Hemorrhage By Volume (%)
placebo & saline6.74 ± 0.490.27 ± 0.084.26 ± 1.48
placebo & dextran7.39 ± 0.240.30 ± 0.084.19 ± 1.17
nimodipine 0.02 mg/kg & saline8.97 ± 1.130.34 ± 0.093.76 ± 0.95
nimodipine 0.02 mg/kg & dextran7.48 ± 0.470.13 ± 0.031.70 ± 0.39
nimodipine 0.05 mg/kg & saline7.10 ± 0.410.47 ± 0.066.70 ± 1.14
nimodipine 0.05 mg/kg & dextran6.59 ± 0.410.35 ± 0.105.08 ± 1.32
significanceNSNSNS

Values are means ± standard error of the means. NS = not significantly different by analysis of variance.

Discussion

The present investigation provides quantitative evidence supporting the hypothesis that reversal of posttraumatic ischemia can promote recovery of axonal function after acute SCI. Infusion of nimodipine at 0.02 mg/kg in combination with dextran 40 significantly improved posttraumatic SCBF and axonal function as assessed by MEP's and SSEP's, a technique which accurately reflects axonal function in the descending and ascending tracts of the cord, respectively,7,10,11 and is highly correlated with clinical neurological function.8 Moreover, the improvement in axonal function attained with nimodipine and dextran was highly correlated with the improvement in SCBF (Fig. 10).

Within 1 hour after trauma, there was a marked decline in SCBF both at and adjacent to the injury site, similar to the findings of others.18,19,35,47 For example, Sandier and Tator36,37 demonstrated that the area of ischemia spread rostrally and caudally from the point of injury and resulted in areas of spinal cord infarction up to 2 cm away. Senter and Venes,42 using the hydrogen clearance technique, also reported decreased SCBF at and 1 cm remote from the site of trauma.

The MABP shows a characteristic response pattern to major cervical cord injury with a brief initial hypertensive peak19,48 followed by a sustained period of posttraumatic hypotension (Table 1). Because the injured cord has impaired autoregulation,43 it has been recommended that posttraumatic hypotension in patients should be corrected in order to maximize neuronal recovery,45,46 but the experimental or clinical proof of this practice is lacking. Although the administration of saline, dextran, and nimodipine 0.02 mg/kg with dextran 40 increased the postinjury MABP significantly (Table 1 and Fig. 11), only the latter treatment resulted in a significant increase in SCBF at T-1 (Figs. 3 and 11), indicating that correction of posttraumatic hypotension alone is insufficient to restore SCBF after SCI. It should be emphasized that the beneficial effect of nimodipine on posttraumatic ischemia was critically dependent on the maintenance of an adequate MABP by dextran 40, confirming the results of Guha, et al.19 (Fig. 11). The ability of the normal cord to autoregulate in the face of decreased MABP is evidenced by the lack of a significant change in the SCBF at T-10 after cord injury. Indeed, only the administration of nimodipine 0.02 mg/kg and dextran 40 significantly increased the SCBF at T-10 over preinjury levels.

Fig. 11.
Fig. 11.

Changes in mean arterial blood pressure (MABP), spinal cord blood flow at T-1 (SCBF T1), d wave amplitude of the motor evoked potentials recorded at the spine (MEP D), and the somatosensory evoked potential (SSEP) N2 wave amplitude with drug infusion (postinfusion-preinfusion value). Although the infusion of saline, dextran, and nimodipine (Nim) 0.02 mg/kg with dextran 40 increased the MABP, only the latter increased the SCBF and promoted a recovery of axonal function.

The mechanisms by which nimodipine and dextran affect SCBF may be understood more clearly by referring to the Hagen-Poiseuille equation,30 which may be used to model blood flow in linear conductance vessels to a first approximation. Given a tube of length L and radius r with a pressure gradient of (P1−P2) across the length of the tube, then the laminar flow (Q) of a newtonian fluid of viscosity n is defined by:

mu1

The administration of dextran 40 resulted in hypervolemic hemodilution as evidenced by an increase in the MABP and a decrease in the hematocrit. Since hematocrit is an important determinant of blood viscosity24 and blood flow is critically dependent on MABP in the absence of autoregulation (Equation 1), the increased SCBF at C-6 with dextran 40 is readily explained and suggests that rheological factors contribute significantly to the pathogenesis of ischemia adjacent to the injury site, as has been postuated in cerebral ischemia.51–53 In contrast, the infusion of dextran without nimodipine failed to improve SCBF at the injury site, indicating that nonrheological factors modulate posttraumatic ischemia at this location. Because blood flow varies as the fourth power of vessel radius (Equation 1), even small decreases in vessel caliber, due to a mechanism such as vasospasm, will drastically alter flow. Microangiographic studies have shown that posttraumatic ischemia is in part due to the impairment of structurally intact vessels to conduct blood,4,14,49 a phenomenon which has been termed posttraumatic vasospasm.19 There is evidence that the mechanism of posttraumatic ischemia is critically dependent on the intracelluar influx of Ca++, which promotes smooth-muscle contraction, initiates synthesis of vasoactive prostanoids, and promotes progressive microvascular lipid peroxidation.1,20 Nimodipine, a blocker of voltage-sensitive Ca++ channels,40 has previously been shown to produce vasodilatation in the brain and spinal cord.19,26,44 The ability of nimodipine to increase posttraumatic SCBF when the MABP was maintained at 80 to 90 mm Hg with dextran provides strong evidence that nimodipine counteracted a Ca++-mediated mechanism such as vasospasm.

There is evidence that nimodipine may also have a direct cytoprotective effect. For example, although the increase in SCBF with nimodipine 0.02 mg/kg and dextran 40 had waned by 2 hours postinfusion, the improvement in axonal function persisted (Fig. 7). Furthermore, a transient recovery of the MEP-C occurred after treatment with nimodipine 0.05 mg/kg and dextran 40 despite the absence of a significant restoration of SCBF (Fig. 7). It is therefore possible that nimodipine may have prevented or reversed further axonal injury by inhibiting Ca++-mediated cytotoxic processes such as impairment of mitochondrial function22 or breakdown of membrane phospholipids.29,50 Indeed, nimodipine has been found to promote functional and metabolic recovery after cerebral ischemia in experimental animals17,25,44,54 and improved clinical outcome in patients with acute stroke.15,16 Faden, et al.,3 on the other hand, reported that nimodipine at a dose of 0.06 or 0.12 mg/kg failed to promote histopathological or clinical neurological recovery 24 to 48 hours after an ischemic cord injury in rabbits. However, in the latter study neither SCBF nor MABP was measured and no adjuvant agents were used to maintain the MABP. Similarly, Ford and Malm13 reported that nimodipine failed to reverse SCI in cats, but they did not administer an adjuvant agent to maintain the MABP.

Quantitative assessment of the volume of hemorrhage at the injury site showed that the improvement in SCBF achieved with nimodipine and dextran did not exacerbate the extent of hematomyelia (Table 2). Rawe, et al.,33 examined the effect of hypertension on the degree of posttraumatic intramedullary hemorrhage in cats with the thoracic cord injured by the weight-drop technique: in animals with MABP raised to greater than 150 mm Hg by the administration of metaraminol, there was a significant increase in the extent of intramedullary hemorrhage. In the present experiment, the MABP in the group that received nimodipine 0.02 mg/kg and dextran 40 was 88.4 mm Hg (Table 1), and this may in part account for the lack of a significant change in the volume of hemorrhage at the injury site.

In conclusion, the administration of nimodipine at 0.02 mg/kg in combination with dextran 40 significantly increased posttraumatic SCBF and promoted a significant recovery of axonal function in the motor and sensory tracts of the cord. This study provides quantitative evidence that increasing the SCBF at the site of cord trauma can significantly improve the function of acutely injured spinal axons. Furthermore, these data strongly indicate that posttraumatic ischemia is of major importance in the pathogenesis of axonal dysfunction after acute cord injury and is reversible by pharmacological intervention. Further experiments to assess whether the improvement in function can be maintained for longer periods of time are required before the clinical applicability of these encouraging results can be determined.

Acknowledgments

Nimodipine and placebo were donated by Miles Pharmaceuticals, Inc. The IBAS II image analysis system was kindly loaned to us by Zeiss Canada, Ltd. Excellent technical assistance was provided by J. Loukides, D. Wilken, and R. Wunderlich.

References

  • 1.

    Cheung JYBonventre JVMalis CDet al: Calcium and ischemic injury. N Engl J Med 314:167016761986N Engl J Med 314:

  • 2.

    Collins WF: A review and update of experimental and clinical studies of spinal cord injury. Paraplegia 21:2042191983Collins WF: A review and update of experimental and clinical studies of spinal cord injury. Paraplegia 21:

  • 3.

    Faden AIJacobs TPSmith MT: Evaluation of the calcium antagonist nimodipine in experimental spinal cord ischemia. J Neurosurg 60:7967991984J Neurosurg 60:

  • 4.

    Fairholm DJTurnbull IM: Microangiographic study of experimental spinal cord injuries. J Neurosurg 35:2772861971J Neurosurg 35:

  • 5.

    Farber JL: Minireview. The role of calcium in cell death. Life Sci 29:128912951981Farber JL: Minireview. The role of calcium in cell death. Life Sci 29:

  • 6.

    Fehlings MGTator CH: A review of models of acute experimental spinal cord injuryIllis LS (ed): Spinal Cord Dysfunction: Assessment. Oxford: University Press1988333Spinal Cord Dysfunction: Assessment.

  • 7.

    Fehlings MGTator CHLinden RD: The combined recording of motor and somatosensory evoked potentials to assess spinal cord function in the rat. in Barber CBlum T (eds): Evoked Potentials III. New York: Butterworth1987462468Evoked Potentials III.

  • 8.

    Fehlings MGTator CHLinden RD: The effect of directcurrent field on recovery from experimental spinal cord injury. J Neurosurg 68:7817921988J Neurosurg 68:

  • 9.

    Fehlings MGTator CHLinden RD: The relationship between posttraumatic spinal cord ischemia, motor and somatosensory evoked potentials and injury severity. J Cereb Blood Flow Metab 7 (Suppl):61987 (Abstract)J Cereb Blood Flow Metab 7 (Suppl):

  • 10.

    Fehlings MGTator CHLinden RDet al: Motor and somatosensory evoked potentials recorded from the rat. Electroencephalogr Clin Neurophysiol 69:65781988Electroencephalogr Clin Neurophysiol 69:

  • 11.

    Fehlings MGTator CHLinden RDet al: Motor evoked potentials recorded from normal and spinal cord-injured rats. Neurosurgery 20:1251301987Neurosurgery 20:

  • 12.

    Fehlings MGTator CHLinden RDet al: Relationship between spinal cord blood flow and axonal function in the motor and sensory tracts of the cord after experimental spinal cord injury. Surg Forum 38:5085091987Surg Forum 38:

  • 13.

    Ford RWMalm DN: Failure of nimodipine to reverse acute experimental spinal cord injury. Cent Nerv Syst Trauma 2:9171985Cent Nerv Syst Trauma 2:

  • 14.

    Fried LCGoodkin R: Microangiographic observations of the experimentally traumatized spinal cord. J Neurosurg 35:7097141971J Neurosurg 35:

  • 15.

    Gelmers HJ: The effects of nimodipine on the clinical course of patients with acute ischemic stroke. Acta Neurol Scand 69:2322391988Gelmers HJ: The effects of nimodipine on the clinical course of patients with acute ischemic stroke. Acta Neurol Scand 69:

  • 16.

    Gelmers HJGorter Kde Weerdt CJet al: A controlled trial of nimodipine in acute ischemic stroke. N Engl J Med 318:2032071988N Engl J Med 318:

  • 17.

    Germano IMBartkowski HMCassel MEet al: The therapeutic value of nimodipine in experimental focal cerebral ischemia. J Neurosurg 67:81871987J Neurosurg 67:

  • 18.

    Griffiths IR: Spinal cord blood flow after acute experimental cord injury in dogs. J Neurol Sci 27:2472591976Griffiths IR: Spinal cord blood flow after acute experimental cord injury in dogs. J Neurol Sci 27:

  • 19.

    Guha ATator CHPiper I: Effect of a calcium channel blocker on posttraumatic spinal cord blood flow. J Neurosurg 66:4234301987J Neurosurg 66:

  • 20.

    Hall EDWolf DL: A pharmacological analysis of the pathophysiological mechanisms of posttraumatic spinal cord ischemia. J Neurosurg 64:9519611986J Neurosurg 64:

  • 21.

    Hall RDLindholm EP: Organization of motor and somatosensory neocortex of the albino rat. Brain Res 66:23381974Brain Res 66:

  • 22.

    Hoehner PJWilson RF: Mitochondrial O2 use and ATP synthesis: kinetic effects of Ca++ and HPO modulated by glucocorticoids. Ann Emerg Med 9:39640319802 use and ATP synthesis: kinetic effects of Ca++ and HPO modulated by glucocorticoids. Ann Emerg Med 9:

  • 23.

    Ingvar DHLassen NA: Regional blood flow of the cerebral cortex determined by krypton85. Acta Physiol Scand 54:325338196285. Acta Physiol Scand 54:

  • 24.

    Kee DB JrWood JH: Rheology of the cerebral circulation. Neurosurgery 15:1251311984Neurosurgery 15:

  • 25.

    Mabe HNagai HTakagi Tet al: Effect of nimodipine on cerebral functional and metabolic recovery following ischemia in the rat brain. Stroke 17:5015051986Stroke 17:

  • 26.

    McCalden TANath RGThiele K: The effects of a calcium antagonist (nimodipine) on basal cerebral blood flow and reactivity to various agonists. Stroke 15:5275301984Stroke 15:

  • 27.

    Midha RFehlings MGTator CHet al: Assessment of spinal cord injury by counting corticospinal and rubrospinal neurons. Brain Res 410:2993081987Brain Res 410:

  • 28.

    Naftchi NEDemeny MDeCrescito V: Biogenic amine concentrations in traumatized spinal cords of cats. Effect of drug therapy. J Neurosurg 40:52571974J Neurosurg 40:

  • 29.

    Nemoto EMShiu GKNemmer JPet al: Free fatty acid accumulation in the pathogenesis and therapy of ischemic-anoxic brain injury. Am J Emerg Med 1:1751791983Am J Emerg Med 1:

  • 30.

    Norwich KH: Classical Theoretical Physiology. Toronto: Intercept Press1980Chapt 105Norwich KH: Classical Theoretical Physiology.

  • 31.

    Osterholm JLMathews GJ: Altered norepinephrine metabolism following experimental spinal cord cord injury. Part 1: Relationship to hemorrhagic necrosis and post-wounding neurological deficits. J Neurosurg 36:3863941972J Neurosurg 36:

  • 32.

    Piper IGuha ATator CHet al: A microcomputer system for on-line collection of blood flow and related physiological data. Comput Biol Med 17:2792911987Comput Biol Med 17:

  • 33.

    Rawe SELee WAPerot PL Jr: The histopathology of experimental spinal cord trauma. The effect of systemic blood pressure. J Neurosurg 48:100210071978J Neurosurg 48:

  • 34.

    Rivlin ASTator CH: Effect of duration of acute spinal cord compression in a new acute cord injury model in the rat. Surg Neurol 10:39431978Surg Neurol 10:

  • 35.

    Rivlin ASTator CH: Regional spinal cord blood flow in rats after severe cord trauma. J Neurosurg 49:8448531978J Neurosurg 49:

  • 36.

    Sandier ANTator CH: Effect of acute spinal cord compression injury on regional spinal cord blood flow in primates. J Neurosurg 45:6606761976J Neurosurg 45:

  • 37.

    Sandier ANTator CH: Review of the effect of spinal cord trauma on the vessels and blood flow in the spinal cord. J Neurosurg 45:6386461976J Neurosurg 45:

  • 38.

    SAS Users Guide: Cary, NC: SAS Institute1985SAS Users Guide:

  • 39.

    Schanne FAXKane ABYoung EEet al: Calcium dependence of toxic cell death: a final common pathway. Science 206:7007021979Science 206:

  • 40.

    Scriabine ABattye RHohhmeister F: NimodipineScriabine A (ed): New Drugs Annual: Cardiovascular Drugs. New York: Raven Press1985197218New Drugs Annual: Cardiovascular Drugs.

  • 41.

    Selby S: Standard Mathematical Tables. Cleveland: CRC Press197114Selby S: Standard Mathematical Tables.

  • 42.

    Senter HJVenes JL: Altered blood flow and secondary injury in experimental spinal cord trauma. J Neurosurg 49:5695781978J Neurosurg 49:

  • 43.

    Senter HJVenes JL: Loss of autoregulation and posttraumatic ischemia following experimental spinal cord trauma. J Neurosurg 50:1982061979J Neurosurg 50:

  • 44.

    Steen PANewberg LAMilde JHet al: Nimodipine improves cerebral blood flow and neurologic recovery after complete cerebral ischemia in the dog. J Cereb Blood Flow Metab 3:38431983J Cereb Blood Flow Metab 3:

  • 45.

    Tator CHRowed DW: Current concepts in the immediate management of acute spinal cord injuries. Can Med Assoc J 121:145314641979Can Med Assoc J 121:

  • 46.

    Tator CHRowed DWSchwartz MLet al: Management of acute spinal cord injuries. Can J Surg 27:2892931984Can J Surg 27:

  • 47.

    Wallace MCTator CH: Spinal cord blood flow measured with microspheres following spinal cord injury in the rat. Can J Neurol Sci 13:91961986Can J Neurol Sci 13:

  • 48.

    Wallace MCTator CH: Successful improvement of blood pressure, cardiac output, and spinal cord blood flow after experimental spinal cord injury. Neurosurgery 20:7107151987Neurosurgery 20:

  • 49.

    Wallace MCTator CHFrazee P: Relationship between posttraumatic ischemia and hemorrhage in the injured rat spinal cord as shown by colloidal carbon angiography. Neurosurgery 18:4334391986Neurosurgery 18:

  • 50.

    Wei EPLamb RGKontos HA: Increased phospholipase C activity after experimental brain injury. J Neurosurg 56:6956981982J Neurosurg 56:

  • 51.

    Wood JHSimeone FAFink EAet al: Correlative aspects of hypervolemic hemodilution: low-molecular-weight dextran infusions after experimental cerebral arterial occlusion. Neurology 34:24341984Neurology 34:

  • 52.

    Wood JHSimeone FAFink EAet al: Hypervolemic hemodilution in experimental focal cerebral ischemia. Elevation of cardiac output, regional cortical blood flow, and ICP after intravascular volume expansion with low molecular weight dextran. J Neurosurg 59:5005091983J Neurosurg 59:

  • 53.

    Wood JHSnyder LLSimeone FA: Failure of intravascular volume expansion without hemodilution to elevate cortical blood flow in region of experimental focal ischemia. J Neurosurg 56:80911982J Neurosurg 56:

  • 54.

    Young WLJosovitz KMorales Oet al: The effect of nimodipine on post-ischemic cerebral glucose utilization and blood flow in the rat. Anesthesiology 67:54591987Anesthesiology 67:

Rats obtained from Charles River Canada, Inc., St. Constant, Quebec, Canada.

Modified aneurysm clip manufactured by Walsh Manufacturing Co., Ltd., Oakville, Ontario, Canada.

Signal averager, Model 8400, manufactured by Cadwell Laboratories Inc., Kennewick, Washington.

IBAS II image analysis system provided on loan by Zeiss Canada Ltd., Don Mills, Ontario, Canada.

This research was supported by grants from the Medical Research Council of Canada, the Canadian Paraplegic Association, and Miles Pharmaceuticals, Inc.

Article Information

Address reprint requests to: Michael G. Fehlings, M.D., Lab 12-423, Playfair Neuroscience Unit, Toronto Western Hospital, 399 Bathurst Street, Toronto, Ontario M5T 2S8, Canada.

Dr. Fehlings is a Fellow of the Medical Research Council of Canada.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    A schematic representation of the experimental protocol. For description see text.

  • View in gallery

    Left: Computer-derived grand mean of the motor evoked potentials recorded from the spinal cord (MEP-C) of 20 normal rats at T-10 (average of 20,480 responses). The normal MEP-C consists of a series of positive deflections: an initial d (direct) wave which reflects direct pyramidal cell activation, and a series of subsequent i waves which result from indirect activation of pyramidal cells by cortical inter-neurons as well as activation of nonpyramidal tracts. Right: Computer-derived grand mean of somatosensory evoked potentials (SSEP) recorded from 16 normal rats from the sensorimotor cortex following sciatic nerve stimulation (average of 32,768 responses). The normal rat SSEP consists of four positive (P) and three negative (N) peaks.

  • View in gallery

    Spinal cord blood flow (SCBF) recorded from the injury site at T-1 in 30 rats. The preinjury SCBF (56.83 ± 3.82 ml/100 gm/min) did not differ significantly among the groups. Following spinal cord injury, there was a significant decrease in SCBF in all groups (mean ± standard error of the mean: 24.55 ± 2.99 ml/100 gm/min). In the group treated with nimodipine (Nim) 0.02 mg/kg and dextran 40, there was a significant increase in SCBF immediately after infusion (43.69 ± 6.09 ml/100 gm/min) and at 1 hour postinfusion (30.54 ± 3.47 ml/100 gm/min) in comparison with the other groups (mean: 17.30 ± 1.50 ml/100 gm/min). At 2 hours postinfusion, the SCBF was similar for all groups.

  • View in gallery

    Spinal cord blood flow (SCBF) data recorded adjacent to the injury site from the cord at C-6 in 30 rats. The preinjury SCBF was similar in all groups (mean ± standard error of the mean: 51.88 ± 3.72 ml/100 gm/min). With injury, the SCBF at C-6 declined significantly in each group (30.43 ± 1.99 ml/100 gm/min). Infusion of dextran alone or in combination with nimodipine (Nim) at 0.02 mg/kg or 0.05 mg/kg resulted in a significant increase in SCBF in comparison with the postinjury values. At 1 and 2 hours postinfusion the improvement in SCBF at C-6 in these groups had abated and there was no difference in SCBF among the six treatment groups.

  • View in gallery

    Spinal cord blood flow (SCBF) data recorded distal to the injury site from the cord at T-10 in 30 rats. The SCBF remained constant in all groups throughout the experiment with the exception of the group treated with nimodipine (Nim) 0.02 mg/kg and dextran 40. In the latter, there was a significant increase in the postinfusion SCBF at T-10 to 81.49 ± 8.45 ml/100 gm/min (mean ± standard error of the mean).

  • View in gallery

    Upper: Representative evoked potential recordings from a control rat which received placebo and saline. After cord injury, the motor evoked potentials recorded at the spinal cord (MEP-C) and at the sciatic nerve (MEP-N) and the somatosensory evoked potentials (SSEP) were abolished and did not recover postinfusion. The spinal evoked potentials (SEP) recorded from the cord at T-10 (caudal to the injury site) following stimulation of the sciatic nerve were unchanged pre- and postinjury, confirming the physiological integrity of the experimental preparation. Lower: Representative evoked potential recordings from a rat treated with nimodipine 0.02 mg/kg and dextran 40. After cord injury, the MEP-C and MEP-N were abolished and the amplitude of the SSEP was significantly attenuated. Postinfusion, there was a significant recovery of the MEP-C, MEP-N, and SSEP. The SEP recorded from the cord at T-10 (caudal to the injury site) following stimulation of the sciatic nerve were unchanged pre- and postinjury, confirming the physiological integrity of the experimental preparation.

  • View in gallery

    The amplitude of the d wave of the motor evoked potential recorded at the spinal cord and plotted as a function of time. Preinjury, the d wave amplitude was similar in all groups. After spinal cord injury, there was a significant reduction in d wave amplitude in all groups. With the exception of the groups treated with nimodipine (Nim) 0.02 mg/kg or 0.05 mg/kg in combination with dextran 40, there was a progressive attenuation of d wave amplitude after injury. After treatment with nimodipine 0.02 mg/kg and dextran 40, there was a significant recovery of d wave amplitude which persisted until the conclusion of recording. In the group that received nimodipine 0.05 mg/kg and dextran 40, there was transient delayed recovery of d wave amplitude at 1 hour postinfusion.

  • View in gallery

    Upper: Superimposed postinfusion motor evoked potential (MEP) traces from the spinal cord of rats treated with placebo and saline (left) or nimodipine 0.02 mg/kg and dextran 40 (right). Lower: The respective grand mean MEP's are shown. These data illustrate a significant return of axonal function in the motor tracts of the cord after infusion of nimodipine 0.02 mg/kg and dextran 40.

  • View in gallery

    The amplitude of the N2 wave of the somatosensory evoked potentials plotted as a function of time. Preinjury, the N2 amplitude was similar for all groups. After spinal cord injury, there was a significant reduction of N2 wave amplitude in all groups. With the exception of the group treated with nimodipine (Nim) 0.02 mg/kg and dextran 40, there was a progressive attenuation of N2 amplitude after injury. Following treatment with nimodipine 0.02 mg/kg and dextran 40, however, there was a significant recovery of the N2 amplitude which persisted at 1 hour but was no longer apparent at 2 hours postinfusion.

  • View in gallery

    Relationship between the d wave amplitude of the motor evoked potentials recorded at the spine and the spinal cord blood flow (SCBF) in 30 rats (five observations per rat: 150 observations in total).

  • View in gallery

    Changes in mean arterial blood pressure (MABP), spinal cord blood flow at T-1 (SCBF T1), d wave amplitude of the motor evoked potentials recorded at the spine (MEP D), and the somatosensory evoked potential (SSEP) N2 wave amplitude with drug infusion (postinfusion-preinfusion value). Although the infusion of saline, dextran, and nimodipine (Nim) 0.02 mg/kg with dextran 40 increased the MABP, only the latter increased the SCBF and promoted a recovery of axonal function.

References

1.

Cheung JYBonventre JVMalis CDet al: Calcium and ischemic injury. N Engl J Med 314:167016761986N Engl J Med 314:

2.

Collins WF: A review and update of experimental and clinical studies of spinal cord injury. Paraplegia 21:2042191983Collins WF: A review and update of experimental and clinical studies of spinal cord injury. Paraplegia 21:

3.

Faden AIJacobs TPSmith MT: Evaluation of the calcium antagonist nimodipine in experimental spinal cord ischemia. J Neurosurg 60:7967991984J Neurosurg 60:

4.

Fairholm DJTurnbull IM: Microangiographic study of experimental spinal cord injuries. J Neurosurg 35:2772861971J Neurosurg 35:

5.

Farber JL: Minireview. The role of calcium in cell death. Life Sci 29:128912951981Farber JL: Minireview. The role of calcium in cell death. Life Sci 29:

6.

Fehlings MGTator CH: A review of models of acute experimental spinal cord injuryIllis LS (ed): Spinal Cord Dysfunction: Assessment. Oxford: University Press1988333Spinal Cord Dysfunction: Assessment.

7.

Fehlings MGTator CHLinden RD: The combined recording of motor and somatosensory evoked potentials to assess spinal cord function in the rat. in Barber CBlum T (eds): Evoked Potentials III. New York: Butterworth1987462468Evoked Potentials III.

8.

Fehlings MGTator CHLinden RD: The effect of directcurrent field on recovery from experimental spinal cord injury. J Neurosurg 68:7817921988J Neurosurg 68:

9.

Fehlings MGTator CHLinden RD: The relationship between posttraumatic spinal cord ischemia, motor and somatosensory evoked potentials and injury severity. J Cereb Blood Flow Metab 7 (Suppl):61987 (Abstract)J Cereb Blood Flow Metab 7 (Suppl):

10.

Fehlings MGTator CHLinden RDet al: Motor and somatosensory evoked potentials recorded from the rat. Electroencephalogr Clin Neurophysiol 69:65781988Electroencephalogr Clin Neurophysiol 69:

11.

Fehlings MGTator CHLinden RDet al: Motor evoked potentials recorded from normal and spinal cord-injured rats. Neurosurgery 20:1251301987Neurosurgery 20:

12.

Fehlings MGTator CHLinden RDet al: Relationship between spinal cord blood flow and axonal function in the motor and sensory tracts of the cord after experimental spinal cord injury. Surg Forum 38:5085091987Surg Forum 38:

13.

Ford RWMalm DN: Failure of nimodipine to reverse acute experimental spinal cord injury. Cent Nerv Syst Trauma 2:9171985Cent Nerv Syst Trauma 2:

14.

Fried LCGoodkin R: Microangiographic observations of the experimentally traumatized spinal cord. J Neurosurg 35:7097141971J Neurosurg 35:

15.

Gelmers HJ: The effects of nimodipine on the clinical course of patients with acute ischemic stroke. Acta Neurol Scand 69:2322391988Gelmers HJ: The effects of nimodipine on the clinical course of patients with acute ischemic stroke. Acta Neurol Scand 69:

16.

Gelmers HJGorter Kde Weerdt CJet al: A controlled trial of nimodipine in acute ischemic stroke. N Engl J Med 318:2032071988N Engl J Med 318:

17.

Germano IMBartkowski HMCassel MEet al: The therapeutic value of nimodipine in experimental focal cerebral ischemia. J Neurosurg 67:81871987J Neurosurg 67:

18.

Griffiths IR: Spinal cord blood flow after acute experimental cord injury in dogs. J Neurol Sci 27:2472591976Griffiths IR: Spinal cord blood flow after acute experimental cord injury in dogs. J Neurol Sci 27:

19.

Guha ATator CHPiper I: Effect of a calcium channel blocker on posttraumatic spinal cord blood flow. J Neurosurg 66:4234301987J Neurosurg 66:

20.

Hall EDWolf DL: A pharmacological analysis of the pathophysiological mechanisms of posttraumatic spinal cord ischemia. J Neurosurg 64:9519611986J Neurosurg 64:

21.

Hall RDLindholm EP: Organization of motor and somatosensory neocortex of the albino rat. Brain Res 66:23381974Brain Res 66:

22.

Hoehner PJWilson RF: Mitochondrial O2 use and ATP synthesis: kinetic effects of Ca++ and HPO modulated by glucocorticoids. Ann Emerg Med 9:39640319802 use and ATP synthesis: kinetic effects of Ca++ and HPO modulated by glucocorticoids. Ann Emerg Med 9:

23.

Ingvar DHLassen NA: Regional blood flow of the cerebral cortex determined by krypton85. Acta Physiol Scand 54:325338196285. Acta Physiol Scand 54:

24.

Kee DB JrWood JH: Rheology of the cerebral circulation. Neurosurgery 15:1251311984Neurosurgery 15:

25.

Mabe HNagai HTakagi Tet al: Effect of nimodipine on cerebral functional and metabolic recovery following ischemia in the rat brain. Stroke 17:5015051986Stroke 17:

26.

McCalden TANath RGThiele K: The effects of a calcium antagonist (nimodipine) on basal cerebral blood flow and reactivity to various agonists. Stroke 15:5275301984Stroke 15:

27.

Midha RFehlings MGTator CHet al: Assessment of spinal cord injury by counting corticospinal and rubrospinal neurons. Brain Res 410:2993081987Brain Res 410:

28.

Naftchi NEDemeny MDeCrescito V: Biogenic amine concentrations in traumatized spinal cords of cats. Effect of drug therapy. J Neurosurg 40:52571974J Neurosurg 40:

29.

Nemoto EMShiu GKNemmer JPet al: Free fatty acid accumulation in the pathogenesis and therapy of ischemic-anoxic brain injury. Am J Emerg Med 1:1751791983Am J Emerg Med 1:

30.

Norwich KH: Classical Theoretical Physiology. Toronto: Intercept Press1980Chapt 105Norwich KH: Classical Theoretical Physiology.

31.

Osterholm JLMathews GJ: Altered norepinephrine metabolism following experimental spinal cord cord injury. Part 1: Relationship to hemorrhagic necrosis and post-wounding neurological deficits. J Neurosurg 36:3863941972J Neurosurg 36:

32.

Piper IGuha ATator CHet al: A microcomputer system for on-line collection of blood flow and related physiological data. Comput Biol Med 17:2792911987Comput Biol Med 17:

33.

Rawe SELee WAPerot PL Jr: The histopathology of experimental spinal cord trauma. The effect of systemic blood pressure. J Neurosurg 48:100210071978J Neurosurg 48:

34.

Rivlin ASTator CH: Effect of duration of acute spinal cord compression in a new acute cord injury model in the rat. Surg Neurol 10:39431978Surg Neurol 10:

35.

Rivlin ASTator CH: Regional spinal cord blood flow in rats after severe cord trauma. J Neurosurg 49:8448531978J Neurosurg 49:

36.

Sandier ANTator CH: Effect of acute spinal cord compression injury on regional spinal cord blood flow in primates. J Neurosurg 45:6606761976J Neurosurg 45:

37.

Sandier ANTator CH: Review of the effect of spinal cord trauma on the vessels and blood flow in the spinal cord. J Neurosurg 45:6386461976J Neurosurg 45:

38.

SAS Users Guide: Cary, NC: SAS Institute1985SAS Users Guide:

39.

Schanne FAXKane ABYoung EEet al: Calcium dependence of toxic cell death: a final common pathway. Science 206:7007021979Science 206:

40.

Scriabine ABattye RHohhmeister F: NimodipineScriabine A (ed): New Drugs Annual: Cardiovascular Drugs. New York: Raven Press1985197218New Drugs Annual: Cardiovascular Drugs.

41.

Selby S: Standard Mathematical Tables. Cleveland: CRC Press197114Selby S: Standard Mathematical Tables.

42.

Senter HJVenes JL: Altered blood flow and secondary injury in experimental spinal cord trauma. J Neurosurg 49:5695781978J Neurosurg 49:

43.

Senter HJVenes JL: Loss of autoregulation and posttraumatic ischemia following experimental spinal cord trauma. J Neurosurg 50:1982061979J Neurosurg 50:

44.

Steen PANewberg LAMilde JHet al: Nimodipine improves cerebral blood flow and neurologic recovery after complete cerebral ischemia in the dog. J Cereb Blood Flow Metab 3:38431983J Cereb Blood Flow Metab 3:

45.

Tator CHRowed DW: Current concepts in the immediate management of acute spinal cord injuries. Can Med Assoc J 121:145314641979Can Med Assoc J 121:

46.

Tator CHRowed DWSchwartz MLet al: Management of acute spinal cord injuries. Can J Surg 27:2892931984Can J Surg 27:

47.

Wallace MCTator CH: Spinal cord blood flow measured with microspheres following spinal cord injury in the rat. Can J Neurol Sci 13:91961986Can J Neurol Sci 13:

48.

Wallace MCTator CH: Successful improvement of blood pressure, cardiac output, and spinal cord blood flow after experimental spinal cord injury. Neurosurgery 20:7107151987Neurosurgery 20:

49.

Wallace MCTator CHFrazee P: Relationship between posttraumatic ischemia and hemorrhage in the injured rat spinal cord as shown by colloidal carbon angiography. Neurosurgery 18:4334391986Neurosurgery 18:

50.

Wei EPLamb RGKontos HA: Increased phospholipase C activity after experimental brain injury. J Neurosurg 56:6956981982J Neurosurg 56:

51.

Wood JHSimeone FAFink EAet al: Correlative aspects of hypervolemic hemodilution: low-molecular-weight dextran infusions after experimental cerebral arterial occlusion. Neurology 34:24341984Neurology 34:

52.

Wood JHSimeone FAFink EAet al: Hypervolemic hemodilution in experimental focal cerebral ischemia. Elevation of cardiac output, regional cortical blood flow, and ICP after intravascular volume expansion with low molecular weight dextran. J Neurosurg 59:5005091983J Neurosurg 59:

53.

Wood JHSnyder LLSimeone FA: Failure of intravascular volume expansion without hemodilution to elevate cortical blood flow in region of experimental focal ischemia. J Neurosurg 56:80911982J Neurosurg 56:

54.

Young WLJosovitz KMorales Oet al: The effect of nimodipine on post-ischemic cerebral glucose utilization and blood flow in the rat. Anesthesiology 67:54591987Anesthesiology 67:

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 2 2 2
Full Text Views 63 63 23
PDF Downloads 55 55 43
EPUB Downloads 0 0 0

PubMed

Google Scholar