Conducted somatosensory evoked potentials during spinal surgery

J Neurosurg57:349-353, 1982

Conducted somatosensory evoked potentials during spinal surgery Part 1: Control conduction velocity measurements JAMES B. MACON, M.D., AND CHARLES E. POLETTI, M.D.

Neurosurgical Service, Massachusetts General Hospital, Boston, Massachusetts

~/ Intraoperative recordings of conducted bipolar epidural somatosensory evoked potentials (SEP's) generated by unilateral common peroneal nerve stimulation have been obtained in 27 patients. The SEP's were multiphasic, 0.3 to 1.5/~V in amplitude, and recorded in 100% of patients with normal cords or in patients with spinal lesions, at a site caudal to the lesions. Control spinal conduction velocities (CV's), measured in the midthoracic to lower cervical regions, were in the range of 65 to 85 m/sec. Control lumbar and lower thoracic CV's were in the range of 30 to 45 m/sec. The CV values were obtained periodically throughout the course of surgery and were plotted as a function of time. In control patients with extradural lesions and neuroleptic anesthesia, the CV's remained constant (_ 3%). The consistency, sensitivity, and safety of SEP recordings obtained by this technique make precise monitoring readily available during spinal operations. KEY WORDS spinal evoked potentials 9 epidural electrodes somatosensory evoked potentials 9 spinal cord ~

C

ORTICAL somatosensory evoked potentials (CSEP's) have been proposed as a means of monitoring spinal cord function intraoperativelyJ ~ Measuring CSEP's, especially in response to peroneal stimulation, is limited by a number of factors, including: l) marked variability of both latency and amplitude with changing levels of general anesthesia; 2) insensitivity to incomplete spinal cord lesions because of supraspinal polysynaptic components; 3) excessively long conduction distance masking the effects of spinal lesions; and 4) unreliability of responses from unilateral lower extremity peripheral nerve stimulation, even in normal individuals. 7 Conducted spinal somatosensory evoked potential (SEP) recording has been accomplished in the laboratory setting with bipolar surface electrodes. 2 This method cannot be easily employed in the operating room because the small-amplitude potentials elicited in the upper thoracic and cervical regions are unobtainable under these suboptimal recording conditions. J. Neurosurg. / Volume 57 / September, 1982

9

Direct recording from the subarachnoid space with needle electrodes has allowed consistent measurement of higher amplitude conducted SEP's. 4,5 This method involves considerable risk unless performed under direct observation by the surgeon and is not satisfactory for continuous intraoperative monitoring. A safer method of epidural recording with percutaneously placed electrodes has been proposed. This method is satisfactory for segmental SEP's, but records conducted SEP's in only 20% to 40% of patients on peripheral nerve stimulation of the lower extremities. 4,14,15Conducted SEP's from cauda equina stimulation may be recorded with this technique and used for intraoperative monitoring.16 However, this method does not involve selective activation of somatosensory spinal pathways, may produce undesirable motor responses, provides no lateralizing information, and requires insertion of additional epidural electrodes for stimulation of the cauda equina. We are now reporting a method for intraoperative 349

J. B. Macon and C. E. Poletti

T8

~,oo ....

.

.

.

C.D..

,.ooo~

..........

T4

monitoring of conducted SEP's with bipolar electrodes placed rostral and caudal to the laminectomy site. This method provides a safe and accurate measurement of spinal somatosensory conduction velocity (CV) and avoids most of the problems associated with the other recording techniques. A preliminary report of these results has been presented elsewhere?

1-~.:::::_......

.

%

Clinical Material and Methods

+

-t-

FIG. 1. Bipolar epidural electrode arrangement and polarity used to record somatosensory evoked potentials at caudal (Vc) and rostral (V0 positions are shown separated by the conduction distance (C.D.) measured by x-ray film. 1"8 and T4 indicate vertebral levels. A

Tb

mA

SEP

O.S u V

B

~

/~

T,

c.v. 74.6 m/s

. . . . . .

15.T 0.5 u~

~t O,S vV

Fzo. 2. Conducted spinal somatosensory evoked potentials (SEP's) in response to unilateral common peroneal (CP) nerve stimulation recorded with bipolar epidural electrodes in patients following thoracic or lower cervical laminectomy. A: Epidural bipolar SEP recording at the T-6 level demonstrating the effects on waveform and latency of increasing stimulus current applied to the right CP nerve. Stimulus parameters: 6.3 Hz, 0.2 msec. B: The latency difference between the onset of the first negative peak at T8 and T-5 (a distance of 97 mm) is 1.3 msec. Accordingly, the conduction velocity (C.V.) was 74.6 m/sec. Stimulus parameters at the CP nerve: 12 mA, 6.3 Hz, 0.2 msec. C: A comparison of spinal monopolar epidural recording at T- 11 with paraspinal muscle reference (T- 11 to G) and bipolar (T-I 1 to T-9) epidural recordings (in the same patient). These two records demonstrate the improved resolution using the latter technique. Stimulus parameters at the CP nerve: 12 mA, 6.3 Hz, 0.2 msec. 350

Intraoperative spinal epidural bipolar recordings have been obtained in a total of 27 patients. Thoracic or cervical laminectomy was performed with standard techniques under neuroleptic anesthesia with fentanyl, nitrous oxide, and muscle relaxation with pancuronium bromide (Pavulon). Platinum wire electrode pairs,* insulated with polyurethane except for the tips (I x 3 ram), were inserted in the epidural space rostral and caudal to the laminectomy site (Fig. 1). Electrode pair tips were separated by 30 mm and sutured in place. Leads were secured in position with sutures in the subcutaneous tissue to prevent movement during the spinal surgery. Electrode positions were confirmed on x-ray films before and after the recording sessions to assure constant conduction distances. Leads were connected to the recording system with shielded cables. There were no complications due to placement of the epidural electrodes. The distance between the rostral and caudal pairs of recording electrodes was measured on x-ray films both before and after completion of the recordings. Square-wave electrical pulse stimuli were applied transcutaneously with silver disc electrodes to either common peroneal nerve with the following parameters: 6.3 Hz, 0.2 msec, and 6 to 14 mA. Stimulus intensity was adjusted to be at least three times threshold for production of minimal motor responses prior to muscle relaxation. The SEP's were recorded with a Nicolet 1174B signal-averaging system with electronic artifact rejection.? A total of 256 to 512 stimulation trials of 40-msec duration were averaged. Electrocardiographic (EKG) artifacts were rejected electronically with a 3-msec delay to accommodate the stimulus artifact. The total analysis time was 40.6 to 81.3 seconds. The recording bandwidth was 5 to 3000 Hz. The recording electrodes were arranged so that the negative electrode of each pair was the most caudal (Fig. 1). All recordings were displayed with the negative potential upward. For the fastest conducting somatosensory fibers, CV was calculated from the latency difference of the onset of the initial negative wave (estimated at the peak of the initial positive deflection) of the rostral and caudal pairs, and the xray film confirmed conduction distance. * Avery E 355 wire electrode pairs manufactured by Avery, Inc., 145 Rome Street, Farmingdale, New York. ? Nicolet 1174B signal-averaging system manufactured by Nicolet Instrument Corp., 5225 Verona Road, Madison, Wisconsin. J. Neurosurg. / Volume 57 / September, 1982

Somatosensory evoked potentials in spinal surgery A

SEP 1.9msec

C.V. 76 m/s

~.:~ 70-

T~ T7

,

144 mm

!. --,,A. ,~^

Ouv.25~ ~ ' ~

~. ~ ~

"

9

v- vv -v

SEP

T,-T,

G0-

C.V. 76.6 + 2.5

z

4Omsec c~,c:='

50

CONTROL

0

I 1

l 2

I 3

TI~E (hrs.) FtG. 3. Intraoperative spinal somatosensory conduction velocity (CV) measurement with bipolar epidural somatosensory evoked potentials (SEP's) from a control patient without myelopathy, during a left extradural thoracic (T3-5) ganglionectomy for treatment of pain. A: Intraoperative epidural SEP's were recorded from the T-1 and T-7 spinal levels at 3.25 hrs from the onset of monitoring. Inter-electrode latency was 1.9 msec. Conduction distance was 144 mm, CV 75.8 m/sec. Left common peroneal nerve stimulus parameters: 12 mA, 6.3 Hz, 0.2 msec. B: The CV was plotted as a function of monitoring time intraoperatively. Variability of CV over the 4-hour period was _ 3.3% (76.6 + 2.5 m/sec) despite widely ranging levels of anesthesia. The patient was awake for sensory testing at the time of the final latency recording (3.75 hrs).

Results

Of the 27 patients with intraoperative spinal epidural monitoring, 15 had a variety of spinal lesions. The remaining 12 patients served as a control group, without myelopathy. Using the bipolar epidural technique, conducted SEP's were recorded from the thoracic or lower cervical region in 100% of the patients with normal cords, or at a site caudal to the spinal lesions in the group with myelopathy. The SEP waveform was dependent on the intensity of common peroneal nerve stimulation (Fig. 2A). The SEP response threshold using transcutaneous common peroneal stimulation was typically 3 to 4 mA. Just above the stimulus threshold (6 mA), the waveform at T-6 was triphasic with an initial positive deflection (Fig. 2A). With increasing intensity (8 mA), a polyphasic waveform appeared which at highest intensity (12 mA) consistently exhibited at least three negative peaks and often more (Figs. 2A and 3A). In addition, the onset latency of SEP's decreased progressively with increasing stimulus intensity until it became constant at higher stimulation (Fig. 2A). Accordingly, in order to reduce SEP waveform and latency variability to a minimum, suprathreshold stimulus intensities (10 to 14 mA) were routinely used for monitoring operative procedures. The SEP waveform also depended on the spinal level of recording. Below T-10, segmental evoked potentials were recorded from the upper lumbar cord, and above C-4, the SEP's became increasingly polyphasic and lower in amplitude. Stimulus frequency changes in the range of 1 to 10 Hz did not alter SEP latency or amplitude. A stimulus frequency of 6.3 Hz J. Neurosurg. / Volume 57 / September, 1982

I 4

was routinely used for monitoring intraoperatively. Monopolar recording with a reference electrode in the paraspinal muscles markedly reduced the recorded SEP amplitude and signal resolution (Fig. 2C). This appeared to result from an increased contribution of EKG and paraspinal electromyogram (EMG) artifacts with this electrode arrangement. Amplitudes of SEP's ranged from 0.3 to 1.5/zV at maximal stimulus intensity in the thoracic and cervical regions of control patients or in patients with myelopathy at a site caudal to their spinal lesion. The SEP amplitude was larger with ipsilateral common peroneal nerve stimulation. Amplitudes of SEP's often varied substantially, even during control operations (• 40%), with no apparent relation to the level of anesthesia or cord manipulation during surgery. In different patients, the onset of latency of the initial negative SEP wave ranged from 8.0 to 22.9 msec depending on the patients' height and the level of the lesion and recording. During extradural operations in patients without myelopathy, however, the onset latency of SEP's remained satisfactorily constant, with a standard deviation range of only 0.04 to 0.11 msec. This SEP onset latency did not vary with differing levels of anesthesia throughout an operation. Conduction velocity measurements across the operative field for the controls, derived from the onset latency between caudal and rostral SEP's, ranged from 65 to 85 m/sec in the cervical and upper thoracic region (Figs. 2B and 3B). In the lumbar and lower thoracic regions, conduction velocity was lower, ranging from 30 to 45 m/sec. In operations in the control group in which the dura was not opened, the conduction velocity remained relatively constant (+ 2.5 m/sec in Fig. 3B). 351

J. B. Macon and C. E. Poletti Discussion The bipolar epidural technique reported here was successful under operating room conditions in recording conducted cervical and thoracic SEP's in 100% of patients. This contrasts with only a 20% to 40% success rate using the monopolar epidural techniques reported previously. 4,1~The reason for this improvement probably relates to the reduction of extraneous paraspinal E M G and E K G artifacts with the closer electrode spacing and proximity to the cord when using bipolar leads. Other factors that may have been responsible for the improved signal resolution include the use of electrodes with a larger surface area, 6 muscle relaxation, and electronic artifact rejection. The latter obviated the need to trigger off the E K G artifact TM and allowed a wider range of stimulus frequencies. The principal advantage of the current technique is that it provides a consistent latency measurement of SEP's from unilateral lower extremity peripheral nerve stimulation that does not change significantly with conditions varying during prolonged spinal operations, including different levels of anesthesia. This constant latency of the SEP's, combined with rostralcaudal recording, then permits calculation and monitoring of the CV across any operative spinal cord segment, excluding peripheral and supraspinal multisynaptic conduction pathways. It is this segmental CV across the operative field that proves to be an extremely sensitive monitor of spinal cord function? Additional advantages of the present technique include: 1) the rapid analysis time (less than 2 minutes); 2) stable epidural electrodes that do not interfere with the operative field; 3) a relatively low risk of cord injury or other complications from the epidural electrodes; and 4) lateralizing information with regard to spinal cord lesions due to unilateral cord activation from ipsilateral peripheral nerve stimulation. As noted, the SEP waveform was increasingly polyphasic and greater in amplitude at higher intensity stimulation (Fig. 2A) and showed more polyphasic potentials at more rostral levels (Fig. 3A). These changes probably reflect activation of populations of larger and smaller sensory fibers with faster and slower conduction velocities separating during transit over longer conduction distances. In contrast to the constant SEP latencies, the SEP amplitudes were often variable during the course of a given operation. Accordingly, the SEP amplitude and waveform are of less value for physiological monitoring than the SEP latency. Studies using cortical evoked potentials have calculated a spinal CV of 55 ___9.9 m/sec for the entire cord. 2,3 Similarly, in other studies, direct recordings of "tractus potentials" from the spinal subarachnoid space have yielded overall CV's at different levels of the spinal cord. 4 In the upper thoracic and cervical regions, CV's ranged from 65 to 85 m/sec, whereas in the lower thoracic and lumbar regions CV's ranged from 30 to 45 m/sec. Studies of transcutaneous re352

cordings in normal humans in the laboratory have also found different CV's at different vertebral levels: in the cervical and upper thoracic region, CV's are 72 to 100 m/sec, and in the lower thoracic region, CV's range from 41 to 56 m/sec? Similarly, a slower CV in the lumbar cord and lower thoracic region has been noted in experimental animals, n,13 These results correlate closely with the direct intraoperative measurements under neuroleptanesthesia in the present study. In conclusion, the technique reported here, by measuring spinal cord conduction velocity across the operative field, provides a convenient and very sensitive physiological means of monitoring spinal cord function during spinal operations. Acknowledgments The authors would like to express their appreciation to Drs. W. H. Sweet, N. T. Zervas, R. G. Ojemann, P. H. Chapman, P. M. Black, and H. T. Ballantine, whose patients participated in these studies. We are also indebted to Lori Newman-Sandrew for preparation of this manuscript. References 1. Cracco JB, Cracco RQ, Stolove R: Spinal evoked potential in man: a maturational study. Electroencephalogr Clln Neurophysio146:58-64, 1979 2. Dorfman LJ: Indirect estimation of spinal cord conduction velocity in man. Electroencephalogr Clin Neurophysloi 42:26-34, 1977 3. Dorfman LJ, Perkash I, Bosley TM, et al: Use of cerebral evoked potentials to evaluate spinal somatosensory function in patients with traumatic and surgical myelopathies. J Neurosurg 52:654-660, 1980 4. Ertekin C: Comparison of the human evoked electrospinogram recorded from the intrathecal, epidural and cutaneous levels. Electroeneephalogr Clin Neurophysiol 44:683-690, 1978 5. Ertekin C: Studies on the human evoked electrospinogram. II. Conduction velocity along the dorsal funiculus. Aeta Neurol Stand 53:21-38, 1976 6. Happel LT, LeBlanc HJ, Kline DG: Spinal cord potentials evoked by peripheral nerve stimulation. Electroencephalogr Clin Neurophysio138:349-354, 1975 7. Low MD, Purves SJ, Purves GB: A critical assessment of the use of evoked potentials in diagnosis of peripheral nerve, spinal cord, and cerebral disease, in Morley TP (ed): Current Controversies in Neurosurgery. Philadelphia: WB Saunders, 1976, pp 169-179 8. Macon JB, Poletti CE, Sweet WH, et al: Conducted somatosensory evoked potentials during spinal surgery. Part 2: Clinical applications. J Neurosurg 57:354-359, 1982 9. Macon JB, Poletti CE, Sweet WH, et al: Spinal conduction velocity measurement during laminectomy. Surg Forum 31:453--455, 1980 10. McCallum JE, Bennett MH: Electrophysiologic monitoring of spinal cord function during intraspinal surgery. Surg Forum 26:469-471, 1975 11. McDonald WI, Sears TA: The effects of experimental demyelination on conduction in the central nervous system. Brain 93:583-598, 1970 J. Neurosurg. / Volume57 / September, 1982

Somatosensory evoked potentials in spinal surgery 12. Owen MP, Brown RH, Spetzler RF, et al: Excision of intramedullary arteriovenous malformation using intraoperative spinal cord monitoring. Surg Neurol 12: 271-276, 1979 13. Sarnowski RJ, Cracco RQ, Vogel HB, et al: Spinal evoked response in the cat. J Neurosurg 43:329-336, 1975 14. Shimoji K, Higashi H, Kano T: Epidural recording of spinal electrogram in man. Eleetroeneephalogr Clin Neurophysio130:236-239, 1971 15. Shimoji K, Kano T, Higashi H, et al: Evoked spinal electrograms recorded from epidural space in man. J Appl Physiol 33:468-471, 1972

J. Neurosurg. / Volume 57 / September, 1982

16. Shimoji K, Shimizu H, Maruyama Y: Origin of somatosensory evoked responses recorded from the cervical skin surface. J Neurosurg 48:980-984, 1978

Manuscript received February 12, 1982. Address reprint requests to: Charles E. Poletti, M.D., Neurosurgical Service, Massachusetts General Hospital, Boston, Massachusetts 02114.

353