Motor and somatosensory evoked potentials in a primate model of experimental spinal cord injury.
Motor and somatosensory evoked potentials (MEP and SSEP) were compared after experimental spinal cord injury in Bonnet monkeys (macaca radiata). The MEP and SSEP changes following graded injuries were related to clinical outcome. Eight healthy mature monkeys with a mean weight of 4.2 + 0.9 Kg were chosen for the study. Graded spinal cord injury was caused using 50, 100, 200, 300 gm-cm force by modified Allens' weight drop device. MEP and SSEP recordings were done before injury and at 0, 2, 4 and 6 hours after injury and on the 7th postoperative day. Neurological assessment was done at 24 hours and on the 7th day following injury. 50, 100, 200 gm-cm force caused partial injuries and 300 gm-cm force caused severe spinal cord injury. The predictive value of MEP and SSEP following partial injuries was 80% and 66.67% respectively. Both MEP and SSEP were 100% predictive in severe injury. MEP and SSEP monitoring can therefore be complementary to each other in predicting the neurological outcome in partial injuries to the spinal cord.
Evoked potentials are valuable tools in assessing the functional status of the nervous system. It is vital to monitor the motor system as the blood supply and its anatomical location are different from that of the sensory system in the spinal cord. Somatosensory evoked potentials (SSEP) have been used in the past as an indicator for predicting the outcome following spinal cord injuries.,,,,, However, some studies have also demonstrated that the presence of a normal SSEP may not be associated with normal power,,,, while others have shown that SSEP changes did not predict recovery of motor function., There have been various reviews on spinal cord injuries in nonprimates which showed motor evoked potential (MEP) to be a more sensitive measure of post injury motor performance.,,,,,,, We performed graded spinal MEP and SSEP in Experimental Spinal Cord Injury cord injury in primate model and analysed the signal changes in MEP and SSEP to determine their reliability as predictors of outcome.
Eight young mature and healthy Bonnet monkeys (macaca radiata) of either sex were chosen for the study. Anaesthesia was induced and maintained by Inj. ketamine hydrochloride 15 mg/kg, intramuscular for induction, and 10-15 mg/kg/hr, intravenous as maintenance. Continuous invasive arterial blood pressure using the femoral artery, pulse rate, rectal temperature and SaO2 were monitored.
The animal was placed prone on the operating table. A mid thoracic laminectomy was performed under aseptic conditions to expose the dura over the spinal cord. Bipolar epidural electrodes were placed two segments cranial and caudal to the proposed site of injury and the motor evoked potentials (MEP) and somatosensory evoked potentials (SSEP) prior to injury were recorded. For obtaining MEPs, the stimulating electrode was placed at the vertex (anode) with the reference (cathode) at the forehead. Transcranial electrical stimulation (Digitimer-185) was used at 220 V. The recording was done using bipolar epidural electrode caudal to the proposed site of the cord injury. For obtaining SSEP, the posterior tibial nerve was stimulated percutaneously just above the motor threshold. The recording was done using bipolar epidural electrodes which were placed cranial to the proposed site of injury. The stimulus was a square pulse of 100 microsecond at 7.5 mA at 4.7Hz. 600 responses were averaged using a Nicolect Viking IV.
The cord injury was produced by Allen's weight drop method (1911) modified by us using a caliberated electromagnetic device and galvanometer operated by a 6 volt battery to provide uniformity of 'contact time' during injury. After a midthoracic laminectomy, 25 gm weight which was held by a electromagnet was dropped from a predetermined height on to the cord through a caliberated tube. The galvanometer was used to lift the weight from the cord after a contact time of 50 msec. The weight dropped from a height of 2, 4, 8 and 12 cm, gave a 50, 100, 200 and 300 gm-cm force respectively. Two monkeys were used for each of these above mentioned categories. MEPs and SSEPs were recorded at 0, 2, 4 and 6 hours after injury and on the 7th post operative day. The 7th day recordings were done by reanaesthetising the animal with Inj. ketamine hydrochloride, 15mg/kg intramuscularly, after assessing the neurological status or muscle power in the lower limbs. Motor power in the lower limbs was graded using Tarlov scale modified by Ducker et al in 1978.
Grade 0 : Paraplegia
Grade 1 : Minor movements at the joints.
Grade 2 : Major movements at the joints but inability to stand.
Grade 3 : Animal can stand and possibly walk.
Grade 4 : Animal can run and has a normal motor system with no obvious weakness.
'Predictive value' was calculated using the formula: True positive/ (true positive + false positive) x 100. In partial injuries true positives were those where the waveforms became normal (+) at 6 hours following injury and motor status corresponded to grade 3 or above. False positives were those in which the waveforms were normal (+) at 6 hours after injury, but motor power was grade 2 or less on the 7th post operative day [Table:1]. For uniformity in analysis, the value which indicated less than 5% change over the 'baseline value' in either latency or amplitude in MEPs and SSEPs were taken as 'Normal' (Table II A - D). In severe injuries true positive corresponded to recordings of waveform being negative at 6 hours following injury and the motor power grade 0 or plegic at the end of one week.
50 to 200 gm-cm force (n=6) caused partial injuries (paretic or grade 1 - 4). 300 gm-cm force (n=2) caused severe injury (plegic or grade 0), which remained so at the end of one week. The motor power was grade 3 or more in 4 out of the 6 monkeys, and grade 2 in the remaining two monkeys with partial cord injuries (n=6) on assessment at the end of one week.
(a) Motor evoked potentials : Latency (Table IIA) : Percentage changes in MEP latency was prominent (prolonged latency) at the end of the second hour after injury with 50 and 100 gm-cm force. 200 gm-cm force produced notable changes soon after injury. Signal changes in one primate (case 6) failed to recover at the sixth hour, and remained so at the end of one week following injury with a 200 gm-cm force.
Amplitude (Table IIB) : Marked changes in MEP amplitude (a decrease in amplitude) were seen in the initial hours with partial injuries to the cord (ie. at 'zero' and 2 hours). This indicated that in almost every case of partial injury a distinct change in the amplitude can be expected. The reversal of the MEP amplitude was seen from four to six hours after injury, except for one case (no. 6), which did not recover to the baseline values even at the end of one week.
(b) Somatosensory evoked potentials : Latency (Table IIC) : A uniform change in the SSEP latencies (an increase in latent period) was observed with partial injuries in all cases, irrespective of the force used. Reversal of the latency changes was seen in only 50% of the cases (3/6 primates) towards the end of 6 hours. In the 200 gm-cm force group (cases 5 and 6), there was no recovery of the latency at the sixth hour or even at one week following injury. The SSEP
waveforms seemed to be susceptible to greater magnitude of force within partial injuries, though the animals were only 'paretic' (and not 'plegic') on clinical assessment done at the end of one week [Table:1].
Amplitude (Table IID) : Uniform decrease in SSEP amplitude was noticed irrespective of the force used. Similar to the changes in SSEP latencies with 200 gmcm force (cases 5 and 6), there was no recovery in the amplitude inspite of some motor response seen at 6 hours and one week following injury.
An illustrated example of a partial injury : (50 gm-cm force) : The changes in MEP and SSEP signals with 50 gm-cm force (case 1) have been reviewed [Figure - 1]. The baseline (before injury) latency value obtained in SSEP for the 'PlN1' complex was 7.2 msec. Soon after the injury (0 hour) it increased to 7.7 msec (6.9% change). At 2, 4 and 6 hours following the injury, the latency prolonged to 8.4, 8.9 and 9.2 msec (ie. a 16.7%, 23.6% and 27.8% change) respectively. The amplitude of the 'P1N1' complex decreased to a certain extent. It was 8.9 µV before injury. At 0 hour after injury the amplitude was 8.1µV (9% change or decrease in amplitude); at 2, 4 and 6 hours the amplitude remained low at 7.5 uV (a 15.7% change or decrease), but reversed to normal only at the end of one week.
The baseline (before injury) latency of the 'D' wave in MEP recording was 5.5 msec. At 0 and 2 hours following injury, it prolonged to 5.7 and 6.0 msec (3.6% and 9% change) respectively. At the end of 6 hours, the latency reversed to normal. Nevertheless, the changes in MEP amplitude was well established. Before injury, the amplitude of the 'D' wave was 11.2 µV. The amplitude at 0 and 2 hours decreased to 2.5 µV and 1.9 µV (a 77.7% and 83% change or decrease in amplitude) respectively. At the end of 4 hours, it reversed to 10.9 µV (a 1.8% change or decrease) and became almost normal (11.0 µV) at 6 hour time intervel (0.9% change). A change in percentage to less than 5% of the baseline values (latency or amplitude) were considered normal.
300 gm-cm force weight drop (n=2) caused severe injury. The motor power was grade 0 at 24 hours following injury and remained so even on the 7th post operative day. Both MEPs and SSEPs did not reappear at the end of one week.
In partial injuries, the predictive value of MEP was 80% and SSEP was 66.67%. In severe injury, the predictive value of both the MEP and SSEP were 100%.
Activation of the motor system either by direct stimulation of the brain or spinal cord with recording from either the spinal cord or the peripheral nerve produce reproducible and reliable waveforms (MEP), which indicates the degree of dysfunction of the nervous system., It is known that the 'D' wave in MEP signifies direct stimulation of the pyramidal tract neurons which is much more consistent and recovers faster, both in amplitude and latency than the 'I' wave, which are generated by 'relayed excitation of pyramidal neurons through cortical interneurons'.
In experimental spinal cord injury, it has been proven that MEP obtained from the spinal cord below the lesion was found to be a significant correlate of the 'ambulation recovery', though signals obtained from the peripheral nerve (EMG) were equally sensitive. In motor evoked potentials, changes in amplitude were gross and therefore these were found to correlate well with the degree of neurological dysfunction and histological damage.,,,,
On the other hand, somatosensory evoked potentials are being used for predicting outcome in experimental studies and in humans.,,,,, It is known that somatosensory evoked potentials can be of limited use in predicting spinal cord injury, as they are transmitted primarily by the dorsal columns, and hence do not reflect the integrity of the important ventral motor pathways. The blood supply and the location of the dorsal column are also different from that of the corticospinal tract, and therefore ischaemic changes do vary in these tracts.,,,,,
However, it is interesting to note that there is a set of second order fibers that send collaterals to the dorsal column nuclei. This pathway has been located outside the dorsal column in the dorsolateral fasciculus of the spinal cord. They consist of axons arising from the region of the spinal cord above the lumbar enlargement.,, These dorsolateral fibres transmit more or less the same information as the ascending dorsal column pathway, but end in the reticular region of the dorsal column nuclei.,, The presence of this pathway indicates that the dorsal column nuclei are not totally deafferented by sectioning the dorsal columns or injury to it, as there is an alternate pathway carried by the dorsolateral fasciculus. Therefore, if an injury to the spinal cord is mild or if it occurs in the midline with relatively less force transmitted to the lateral fasciculus, there is a possibility that the cortico spinal tract be damaged relatively more when compared to the ascending dorsal column tracts, which is found in the outer aspect of the dorsolateral fasciculus. This perhaps provides an explanation, why SSEP signals are not completely lost sometimes, though, there is a significant impairment in motor power.
In the present study, we found that partial injuries produce changes in both amplitude and latency of the MEPs. However, the amplitude changes were very marked as compared to changes in latency, especially at zero and 2 hours following injury. Both latency and amplitude of the MEPs had a significant correlation with the recovery in partial injuries, but were not 100% predictive. MEPs reversed almost in all except one (l/6) at the end of 6 hours, thus having a predictive value of 80% at the end of one week. It was also seen that the predictive value of SSEP was about 66.67%.
The latency of the SSEP in partial injuries was prolonged more than MEP, though the latency had reversed at the end of one week following the injury.
Nevertheless, MEP and SSEP monitoring could still be ideal prognostic indicators of post injury neurological recovery, as long as they are used 'complementary' to one another, because the MEP signals by itself were 100% predictive with partial injuries in primates. Not much significance was achieved in motor and sensory evoked potential monitoring in the case of severe cord injuries as both of them were 100% predictive of the clinical outcome.
In a primate model of spinal cord injury, the predictive value of MEP was 80% and SSEP 66.67% (partial injuries); MEP and SSEP signals were cent percent predictive of the outcome in severe injuries. MEP signals, especially the amplitude were found to be highly sensitive to changes in the cord following partial injuries to the spinal cord. Percentage changes of both MEP and SSEP must be precisely monitored as they can be complementary to each other in predicting the final neurological outcome.
The authors would like to thank the FLUID Research Committee of the Christian Medical College, Vellore for financial assistance to carry out this study.