- Email: [email protected]
Intraoperative spinal cord monitoring during surgery for aortic aneurysm: application of spinal cord evoked potential
Electroencephalography and clinical Neurophysiology, 84 (1992) 315-320 © 1992 Elsevier Scientific Publishers Ireland, Ltd. 0168-5597/92/$05.00
315
EVOPOT 90202
Intraoperative spinal cord monitoring during surgery for aortic aneurysm: application of spinal cord evoked potential Yuzuru Okamoto, Masazumi Murakami, Takeo Nakagawa, Atsushi Murata and Hideshige Moriya Department of Orthopaedic Surgery, School of Medicine, Chiba Uniuersity, Chiba (Japan) (Accepted for publication: 24 February 1992)
Summary Spinal cord evoked potentials elicited by direct stimulation of the spinal cord were monitored in 21 patients during thoracic or thoraco-abdominal aortic aneurysm surgery. Flexible catheter-type electrodes were used for both stimulating and recording. The basic pattern of the spinal cord evoked potential consisted of an initial spike and a subsequent polyphasiccomponent. The earliest and most frequent alterations after cross-clampingof the aorta were changes in the configuration or amplitude of the polyphasic component. In 13 patients who exhibited no change except minor alterations of the polyphasic component during the initial test clamping for 15 or 20 min, subsequent graft replacements were safely performed without reimplantation of intercostal vessels. In 2 patients who had sudden cardiac arrests, the evoked potential completely disappeared. The polyphasic component disappeared first, followed by the initial spike. Another patient developed acute loss of the potential after the aneurysm was incised, presumably due to distal aortic hypoperfusion. In this case, prolonged distal hypotension resulted in flaccid paraplegia. Intraoperative monitoring of the spinal cord evoked potential is a useful method for the early detection of spinal cord ischemia during surgery requiring aortic occlusion. Key words: Spinal cord monitoring; Spinal cord evoked potential; Aortic aneurysm; Spinal cord ischemia; Paraplegia
Neurological complications due to spinal cord ischemia are among the major problems encountered in cardiovascular surgery requiring aortic occlusion. The most disastrous complication is permanent paraplegia. Various surgical techniques have been employed to prevent ischemic spinal cord injury (Wakabayashi and Connolly 1976). So far, however, these techniques have not obviated postoperative paraplegia. One reason is thought to be the extreme complexity and diversity of the blood supply to the spinal cord (Dommisse 1974). Spinal cord evoked potentials (SpEPs) elicited by direct stimulation of the spinal cord have been used clinically to monitor intraoperative spinal cord function in a wide variety of spine and spinal cord surgeries (Imai et al. 1984; Tamaki et al. 1984b, 1985b). We recently began using this technique for early detection of spinal cord ischemia during operations on the descending aorta. This report presents the authors' initial experiences with 21 consecutive patients undergoing surgical repair of thoracic or thoraco-abdominal aortic aneurysms.
Correspondence to: Dr. Y. Okamoto, Department of Orthopaedic Surgery, School of Medicine, Chiba University, 1o8-1 Inohana, Chiba-City, Chiba 280 (Japan).
Materials and methods
Patient population From October 1985 to March 1990, 21 patients underwent intraoperative monitoring of spinal cord evoked potentials (SpEPs) during surgical procedures for aortic lesions. There were 15 men and 6 women, with ages ranging between 40 and 74 years (mean 59.6 years). The clinical diagnoses were thoracic aortic aneurysms in 12 patients, thoraco-abdominal aortic aneurysms in 5 and dissecting aortic aneurysms (DeBakey III) in 4. In case 2, a Dacron patch was employed after excision of a localized saccular aneurysm. Graft replacement of the aneurysm with a Dacron tube was designed for the remaining 20 patients. To maintain distal aortic perfusion, an extra-anatomical temporary bypass was produced in 11 patients and femorofemoral bypass with a pump-oxygenator was used in the remaining 10. All patients were anesthetized with high doses of morphine. Partial pressures of oxygen and body temperature were well controlled except in two cases in whom there was some fluctuation (described later). A detailed profile of each patient is illustrated in Table I. Recording o f S p E P According to the method described by Tamaki et al.
(yrs), sex
58, 60, 59, 61,
54, 63, 65, 49,
no.
1 2 3 4
5 6 7 8
69, M 65, M
49, 58, 67, 40, 70, 69, 74, 59,
60, M
11 12
13 14 15 16 17 18 19 20
21
DB III b
TAAA DB III b TAA TAA TAA TAAA TAA TAA
TAA TAA
TAA TAAA
TAA DB III a DB III b TAA
TAA TAA TAAA TAAA
Diagnosis
D5
distal to LSA D3 D5 D5 . . D6 D3 D6
distal to LSA ascending aorta
D5 distal to LSA
distal to LSA distal to LSA distal to LSA D5
Ds localized in distal to LSA D8 D10
Proximal
135 unidentified
145 260
unidentified 120 160 120
60 75 110 305
Duration (rain)
D12
142
proximal to SMA 143 D9 105 D10 80 Dn 132 . . proximal to CA 85 Dl0 67 Dn 74
D5 D6
D8 proximal to CIA
D12 D10 proximal to CA Dio
proximal to CA proximal to CIA
D10
Distal
Aortic cross-clamping placement
60-100
4 0 - 80 6 0 - 90 8 0 - 90
60-100 5 0 - 70 4 0 - 80 70-110
5 0 - 90 40-110
5 0 - 90 3 0 - 50
70-120 unknown 7 0 - 90 80-120
unknown 8 0 - 90 70-110 70- 80
pressure during occlusion (mm Hg)
Distal aortic
none none none none none 26% decrease none latency shortening (0.3 msec) 47% decrease
none 85% decrease
none total disappearance
none none 56% decrease latency prolongation (0.9 msec) total disappearance none total disappearance none
Spike
SpEP changes
39% decrease
33% increase none none none none 25% decrease none 30% increase
33% decrease 77% decrease
pattern alteration none
none
32% decrease none 45% decrease marked increase (scaled out)
Polyphasic
yes
yes yes no
yes yes
no
yes yes no
yes yes no
Recovery
none
none none none none none none none none
none (died)
none paraplegia
(died) none none none
none none none none
complication
Neurological
massive bleeding from intercostal arteries c
distal hypoperfusion ~
total o c c l u s i o n / distal hypoperfusion b,c
cardiac arrest c
cardiac arrest c
total occlusion b.c b
Comments
a T A A = thoracic aortic aneurysm; T A A A = thoraco-abdominal aortic aneurysm; DB l i i a and DB III b = dissecting aortic aneurysm; DeBakey's type III a and IIl b, respectively; LSA = left subclavian artery; CA = celiac artery; C IA = common lilac artery. b Patients undergoing reimplantation of intercostal vessels. c Presumed causes of SpEP changes.
M F F M M F F M
60, M 42, M
9 10
M F M M
F M M M
Age
Case
Profile of 21 patients with spinal cord monitoring during aortic occlusion a
TABLE I
r-
0 rn
0
©
SPINAL CORD MONITORING FOR AORTIC ANEURYSM
(1984b), specially designed flexible catheter-type electrodes (Unique Medical Co. Ltd., Tokyo) were inserted by means of a Tuohy's needle into the subarachnoid space at the level of the conus medullaris and into the epidural space at the cervical level. One (mainly the rostral electrode) was used for stimulating, and the other for recording. Square-wave pulses of 0.2 msec duration and supramaximal intensity (10-20 mA) were applied at a rate of 10-30/sec. Records were obtained using one of the following: DISA N-2000 or Medelec MS-6 or MS-92B. Low and high frequency filters were at 20 and 2000 Hz, respectively, and the evoked potentials were averaged 20-100 times to eliminate electrical noise. The electrodes were removed more than 24 h after the operations in view of the hemorrhagic tendency due to the intraoperative use of heparin.
Results
Operative results The only patient receiving a Dacron patch (case 2) had an uneventful perioperative course. Of the remaining 20 patients who were scheduled to undergo a Dracon graft replacement, one (case 17) developed systemic hypotension (mean aortic pressure 30 mm Hg) due to acute heart failure prior to aortic occlusion, and the subsequent graft replacement was aborted. Two patients (cases 5 and 12) developed acute heart failure during aortic occlusion and died intraoperatively. One patient (case 10) with an extensive thoraco-abdominal aneurysm exhibited permanent paraplegia postoperatively. The remaining 16 patients successfully underwent graft replacements and were awakened without neurological complications.
Basic pattern of SpEP If both the electrodes are placed properly, SpEP can be recorded with either the rostral or caudal electrode (Fig. 1). In this series, SpEP was principally elicited by rostral stimulation and recorded caudally (descending SpEP). In some cases in which the descending SpEP could not be appropriately recorded, presumably due to incorrect placement of the caudal electrode, the ascending SpEP, using the caudal electrode to stimulate and the rostral to record, was used for monitoring. The basic SpEP pattern consisted of an initial spike and a subsequent polyphasic component (Fig. 1). As the wave configurations varied from patient to patient, the evoked potential initially recorded prior to the surgical procedure on the aorta was considered the control pattern for that patient. The peak-to-peak amplitudes of both the initial spike and the largest component of the polyphasic response were used as the main
317
initialspike
polyphasiccomponent
descendingSpEP
ascendingSpEP
+25JUV 2ms Fig. 1. Basic pattern of SpEP. Upper trace: descending SpEP elicited by rostral stimulation. Lower trace: ascending SpEP elicited by caudal stimulation. Typical wave form of SpEP consists of an initial spike and a polyphasic component. Upward deflections are negative in this and all other figures.
indicators, and pared with the cording to the peak latency of
an alteration of more than 20% comcontrol was considered significant accriteria of Tamaki et al. (1984b). The the spike was also used as an indicator.
SpEP changes during aortic occlusion An initial test clamping was performed in most of the cases before the aneurysm was incised. Both proximal and distal clamping was done, and the evoked potential was observed for 15 or 20 min to estimate collateral circulation to the spinal cord. The earliest and most frequent SpEP alterations after cross-clamping of the aorta were changes in the configuration or amplitude of the polyphasic component. Six patients (cases 1, 4, 8, 11, 13 and 20) demonstrated alterations of the polyphasic component without any change in the amplitude of the spike. In 13 patients who showed no change in the spike amplitude during the test clamping, subsequent graft replacements were successfully performed without reimplantation of intercostal vessels. Significant changes in the amplitude of the initial spike were observed in 7 patients (cases 3, 5, 7, 10, 12, 18 and 21), including 3 patients with total disappearance of SpEP. Among them, 2 patients (cases 5 and 7) developed sudden cardiac arrest during aortic occlusion and the evoked potential gradually faded out, first the polyphasic component and then the initial spike (Fig. 2). In case 10, SpEP amplitude decreased immediately after aortic cross-clamping and disappeared after the aneurysm was incised. The evoked potential did not return during the remainder of the surgery and the patient was found to have flaccid paraplegia postoperatively (Fig. 3). Changes in peak latency of the initial spike were observed in 2 patients. One was prolongation by 0.9 msec (case 4) and the other was shortening by 0.3 msec
318
Y. O K A M O T O E T AL.
(case 20). Since these alterations in latency were not accompanied by any change in the spike amplitude, body temperature fluctuation (a fall of 2.0°C and a rise of 1.1°C, respectively) was thought to be the main cause.
conlro] -.-,-- 14:44
aortic clamp
-4-- Itt: tt 7
aneurysm incised
l,uV
i; 2ms
Discussion
Paraplegia is one of the most devastating complications of cardiovascular surgery requiring temporary occlusion of the aorta. The frequency of paraplegia following surgical repair of a thoracic or thoraco-abdominal aortic aneurysm is considerable and estimated to be approximately 5% (Wakabayashi et al. 1976; Crawford et al. 1981). The exact causative mechanism of the ischemic spinal cord injury remains unclear because the circulation of the spinal cord is extremely complex and the patterns vary with the individual
2ms
~,
~",~,,~".b~ J.
17.51
t
cardiac arrest 18.17
,82,
1:
18.33
18.45 18.52 --
F
^
.,~ x ~ . _ . ~ . _ ~
CPB on 19.07
~
19.37
lOpV
L
zo.3o
14;53
,L
Fig. 3. Intraoperative SpEP in case 10. The initial spike (arrow heads) decreased in amplitude after aortic cross-clamping, but the aneurysm was unfortunately incised before an alarm was given. The distal aortic pressure fell to approximately 40 m m Hg and the latency of the spike was markedly prolonged. SpEP then rapidly disappeared. No recovery of the potentials was observed and the patient was found to have flaccid paraplegia postoperatively.
control
<---
'''''~''J~'~
I; 2ms
Fig. 2. Intraoperative SpEP in case 5. Note the gradual change after cardiac arrest. The evoked potential gradually faded out, initially the polyphasic component and then the initial spike. Reappearance of SpEP was observed in the reverse order after initiation of the pump-oxygenator. This patient, however, could not be weaned from the pump-oxygenator and died intraoperatively.
(Dommisse 1974). However, 4 major risk factors have been proposed as follows: (1) prolonged aortic clamping (Laschinger et al. 1984), (2) elevated spinal fluid pressure due to proximal hypertension (Blaisdell and Cooley 1962), (3) distal aortic hypotension (Krieger and Spencer 1985), and (4) interruption of critical radicular arteries (Wadouh et al. 1984). Several investigators have recently reported that the somatosensory evoked potential (SEP) is a useful method for early detection of spinal cord ischemia and for prevention of paraplegia during surgery on the thoracic or thoraco-abdominal aorta (Coles et al. 1982; Cunningham et al. 1982; Laschinger et al. 1983; Kaplan et al. 1986; Krieger and Spencer 1985). Although this technique has the great merit that it is non-invasive, it also has several disadvantages. SEPs appear to be too sensitive to changes of body temperature, hemodilution, blood pressure, anesthetic drugs and operative manipulations (McWilliam et al. 1985; Thurner et al. 1985; Danto et al. 1988). Moreover, there are some reports of false negative results clinically (Mizrahi and Crawford 1984) or experimentally (Laschinger et al. 1984). Since conduction of the SEP is principally ascribed to the dorsal columns, patients with anterior spinal artery syndromes may demonstrate a normal SEP.
SPINAL CORD MONITORING FOR AORTIC ANEURYSM
Spinal cord evoked potentials (SpEPs) elicited by direct stimulation of the spinal cord have been developed and used clinically since 1972 in Japan. The effectiveness of SpEP monitoring in spine or spinal cord surgery has already been established (Imai et al. 1984; Tamaki et ai. 1984b, 1985b). If the stimulating electrode is placed in the cervical epidural space and the recording electrode in the subarachnoid space at the level of the conus medullaris, SpEPs (termed descending conductive SpEP) from normal subjects consist of an initial spike and a subsequent polyphasic component. There have been several studies to determine the pathways through which both components of SpEP are transmitted. Although the initial spike has been thought to be transferred through the dorsolateral column, Harada et al. (1984) and Toyoda and Kanda (1984) demonstrated that this component originates from many fibers of large diameter in all quadrants of the spinal cord. The polyphasic component is considered to be conducted mainly through the dorsal column. Some experimental results, however, seem to indicate that postsynaptic potentials also contribute to it (Tamaki et al. 1985a). In the present study, the most frequent alternations of the SpEP after cross-clamping of the aorta were those in the configuration or amplitude of the polyphasic component. There was no case in which change in the spike amplitude was observed prior to that of the polyphasic component. Our clinical results agree with the experimental data of Tamaki et al. (1984a, 1985a). The authors feel that this is chiefly due to the differential vulnerability of gray matter and white matter to ischemia (Gelfan and Tarlov 1955). Since long-tract conduction in the spinal cord is thought to be rather resistant to ischemia (Kobrine et al. 1979), the polyphasic component which contains synaptic activities is more sensitive than the initial spike. Several investigators have emphasized the importance of re-implantation of intercostal arteries (Cunningham et al. 1982; Wadouh et al. 1984). In our experience, however, all of the 13 patients who demonstrated no change in the amplitude of the initial spike during the test clamping underwent graft replacement without reimplantation of the intercostal arteries, and none had neurological changes postoperatively. Since a rapid decrease in the spike amplitude was observed during test clamping in case 3, re-implantation of intercostal arteries was performed. The authors believe that initial test clamping for 15 or 20 min is valuable for estimating collateral circulation to the spinal cord. In case 10, the only patient with postoperative paraplegia, SpEP acutely disappeared after an extreme decline in distal aortic pressure. Thus, distal hypoperfusion is though to be the major cause of spinal cord injury (Fig. 3). Moreover, SpEP varied proportionally to the distal aortic pressure in case 12 (Fig. 4). As
319
~
control aorll,celamp DAoPImmHgl
~-~,~
,,.S~,,..,,~.,~, 18:30
55
18:55
50
19:00
42
,.J',-~,-19:18
52
19:51
78
~---~
. . . .
~
^
~ t~' r~.
• ----=
~ , '~,,~.a~'~'-,--'----'--'
1 2uv 2ms Fig. 4. Intraoperative SpEP in case 12. The wave pattern is atypical. There are two spikes, presumably due to incorrect placement of the recording electrode. During aortic cross-clamping, distal aortic pressure (DAoP) fluctuated between 40 and 110 mm Hg. The amplitudes of both the initial spike and the polyphasic component were proportional to the distal aortic pressure (DAoP).
Kobrine et al. (1976) indicated, from experimental data in the rhesus monkey, an autoregulation mechanism of spinal cord blood flow (SCBF) is postulated to operate when mean arterial blood pressure (MAP) is above 50 mm Hg, whereas SCBF fails passively with MAP decrease below 50 mm Hg. Laschinger et al. (1983) reported that distal aortic pressure of less than 40 mm Hg resulted in spinal cord ischemia experimentally and clinically. Thus, maintenance of distal aortic perfusion seems to be extremely important for prevention of ischemic spinal cord injury. In summary, these results demonstrate that monitoring of SpEP is a useful method for real-time evaluation of spinal cord function and for early detection of spinal cord ischemia during cardiovascular surgery requiring aortic occlusion. The authors gratefully acknowledge Yasutsugu Nakagawa, M.D., First Department of Surgery, School of Medicine, Chiba University, and Shigeaki Uemura, M.D., Division of Cardiovascular Surgery,
320 Chiba Kaihin Hospital, for giving us the opportunity to monitor their patients. We also acknowledge Ahmmed Ally, M.D., for kindly checking this manuscript.
References Blaisdell, F.W. and Cooley, D.A. The mechanism of paraplegia after temporary thoracic aortic occlusion and its relationship to spinal fluid pressure. Surgery, 1962, 51: 351-355. Coles, J.G., Wilson, G.J., Sima, A.F., Klement, P. and Tait, G.A. Intraoperative detection of spinal cord ischemia using somatosensory cortical evoked potentials during thoracic aortic occlusion. Ann. Thorac. Surg., 1982, 34: 299-306. Crawford, E.S., Walker, H.S.J., Saleh, S.A. and Normann, N.A. Graft replacement of aneurysm in descending thoracic aorta: results without bypass or shunting. Surgery, 1981, 89: 73-85. Cunningham, Jr., J.N., Laschinger, J.C., Merkin, H.A., Nathan, I.M., Colvin, S., Ransohoff, J. and Spencer, F.C. Measurement of spinal cord ischemia during operations upon the thoracic aorta. Initial clinical experience. Ann. Surg., 1982, 196: 285-296. Danto, J., Cataletto, M. and Wolpin, M. Evoked potential monitoring of anesthetic and operative manipulation. In: T.B. Ducker and R.H. Brown (Eds.), Neurophysiology and Standards of Spinal Cord Monitoring. Springer, New York, 1988: 157-162. Dommisse, G.F. The blood supply of the spinal cord. A critical vascular zone in spinal surgery. J. Bone Jt Surg., 1974, 56B: 225-235. Gelfan, S. and Tarlov, I.M. Differential vulnerability of spinal cord structures to anoxia. J. Neurophysiol., 1955, 18: 170-188. Harada, Y., Takemitsu, Y., Atsuta, Y. and Imai, M. Determination of the pathways of ascending and descending conductive spinal cord evoked potentials (SCEP). In: S. Homma and T. Tamaki (Eds.), Fundamentals and Clinical Application of Spinal Cord Monitoring. Saikon Press, Tokyo, 1984: 33-43. Imai, K., Kobayashi, H., Nakagawa, T., Inoue, S. and Tamaki, T. Experiences and analysis of spinal cord monitoring during surgery. In: S. Homma and T. Tamaki (Eds.), Fundamentals and Clinical Application of Spinal Cord Monitoring. Saikon Press, Tokyo, 1984:211-221. Kaplan, B.J., Friedman, W.A., Alexander, J.A. and Hampson, S.R. Somatosensory evoked potential monitoring of spinal cord ischemia during aortic operations. Neurosurgery, 1986, 19: 82-90. Kobrine, A.I., Doyle, T.F. and Rizzoli, H.V. Spinal cord blood flow as affected by changes in systemic arterial blood pressure. J. Neurosurg., 1976, 44: 12-15. Kobrine, A.I., Evans, D.E. and Rizzoli, H.V. The effects of ischemia on long-tract neural conduction in the spinal cord. J. Neurosurg., 1979, 50: 639-644.
Y. OKAMOTO ET AL. Krieger, K.H. and Spencer, F.C. Is paraplegia after repair of coarctation of the aorta due principally to distal hypotension during aortic cross-clamping?. Surgery, 1985, 97: 2-7. Laschinger, J.C., Cunningham, Jr., J.N., Nathan, I.M., Knopp, E.A., Cooper, M.M. and Spencer, F.C. Experimental and clinical assessment of the adequacy of partial bypass in maintenance of spinal cord blood flow during operations on the thoracic aorta. Ann. Thorac. Surg., 1983, 36: 417-426. Laschinger, J.C., Cunningham, Jr., J.N., Cooper, M.M., Krieger, K., Nathan, I.M. and Spencer, F.C. Prevention of ischemic spinal cord injury following aortic cross-clamping: use of corticosteroids. Ann. Thorac. Surg., 1984, 38: 500-507. McWilliam, R.C., Conner, A.N. and Pollock, J.C.S. Cortical somatosensory evoked potentials during surgery for scoliosis and coarctation of the aorta. In: J. Schramm and S.J. Jones (Eds.), Spinal Cord Monitoring. Springer, Berlin, 1985: 167-172. Mizrahi, E.M. and Crawford, E.S. Somatosensory evoked potentials during reversible spinal cord ischemia in man. Electroenceph. clin. Neurophysiol., 1984, 58: 120-126. Tamaki, T., Noguchi, T., Takano, H., Tsuji, H. and Dincer, M.D. The effects of hypovolemic hypotension and hypoxia on the jeopardized spinal cord. In: S. Homma and T. Tamaki (Eds.), Fundamentals and Clinical Application of Spinal Cord Monitoring. Saikon Press, Tokyo, 1984a: 145-154. Tamaki, T., Noguchi, T., Takano, H., Tsuji, H., Nakagawa, T., Imai, K. and Inoue, S. Spinal cord monitoring as a clinical utilization of the spinal evoked potential. Clin. Orthop., 1984b, 184: 58-64. Tamaki, T., Takano, H. and Takakuwa, K. Spinal cord monitoring: basic principles and experimental aspects. CNS Trauma, 1985a, 2: 137-149. Tamaki, T., Takano, H., Takakuwa, K., Tsuji, H., Nakagawa, T., Imai, K. and Inoue, S. An assessment of the use of the spinal cord evoked potentials in prognosis estimation of injured spinal cord. In: J. Schramm and S.J. Jones (Eds.), Spinal Cord Monitoring. Springer, Berlin, 1985b: 221-226. Thurner, F., Schramm, J. and Romstock, J. Effects of fentanyl and enflurane on cortical and subcortical SEP during general anesthesia in man. In: J. Schramm and S.J. Jones (Eds.), Spinal Cord Monitoring. Springer, Berlin, 1985: 82-89. Toyoda, A. and Kanda, K. Origins of spinal cord potentials evoked by stimulation of the cat spinal cord. In: S. Homma and T. Tamaki (Eds.), Fundamentals and Clinical Application of Spinal Cord Monitoring. Saikon Press, Tokyo, 1984: 99-111. Wadouh, F., Lindemann, E., Arndt, C.F., Hetzer, R. and Borst, H.G. The arteria raducularis magna anterior as a decisive factor influencing spinal cord damage during aortic occlusion. J. Thorac. Cardiovasc. Surg., 1984, 88: 1-10. Wakabayashi, A. and Connolly, J.E. Prevention of paraplegia associated with resection of extensive thoracic aneurysms. Arch. Surg., 1976, 111: 1186-1189.
315
EVOPOT 90202
Intraoperative spinal cord monitoring during surgery for aortic aneurysm: application of spinal cord evoked potential Yuzuru Okamoto, Masazumi Murakami, Takeo Nakagawa, Atsushi Murata and Hideshige Moriya Department of Orthopaedic Surgery, School of Medicine, Chiba Uniuersity, Chiba (Japan) (Accepted for publication: 24 February 1992)
Summary Spinal cord evoked potentials elicited by direct stimulation of the spinal cord were monitored in 21 patients during thoracic or thoraco-abdominal aortic aneurysm surgery. Flexible catheter-type electrodes were used for both stimulating and recording. The basic pattern of the spinal cord evoked potential consisted of an initial spike and a subsequent polyphasiccomponent. The earliest and most frequent alterations after cross-clampingof the aorta were changes in the configuration or amplitude of the polyphasic component. In 13 patients who exhibited no change except minor alterations of the polyphasic component during the initial test clamping for 15 or 20 min, subsequent graft replacements were safely performed without reimplantation of intercostal vessels. In 2 patients who had sudden cardiac arrests, the evoked potential completely disappeared. The polyphasic component disappeared first, followed by the initial spike. Another patient developed acute loss of the potential after the aneurysm was incised, presumably due to distal aortic hypoperfusion. In this case, prolonged distal hypotension resulted in flaccid paraplegia. Intraoperative monitoring of the spinal cord evoked potential is a useful method for the early detection of spinal cord ischemia during surgery requiring aortic occlusion. Key words: Spinal cord monitoring; Spinal cord evoked potential; Aortic aneurysm; Spinal cord ischemia; Paraplegia
Neurological complications due to spinal cord ischemia are among the major problems encountered in cardiovascular surgery requiring aortic occlusion. The most disastrous complication is permanent paraplegia. Various surgical techniques have been employed to prevent ischemic spinal cord injury (Wakabayashi and Connolly 1976). So far, however, these techniques have not obviated postoperative paraplegia. One reason is thought to be the extreme complexity and diversity of the blood supply to the spinal cord (Dommisse 1974). Spinal cord evoked potentials (SpEPs) elicited by direct stimulation of the spinal cord have been used clinically to monitor intraoperative spinal cord function in a wide variety of spine and spinal cord surgeries (Imai et al. 1984; Tamaki et al. 1984b, 1985b). We recently began using this technique for early detection of spinal cord ischemia during operations on the descending aorta. This report presents the authors' initial experiences with 21 consecutive patients undergoing surgical repair of thoracic or thoraco-abdominal aortic aneurysms.
Correspondence to: Dr. Y. Okamoto, Department of Orthopaedic Surgery, School of Medicine, Chiba University, 1o8-1 Inohana, Chiba-City, Chiba 280 (Japan).
Materials and methods
Patient population From October 1985 to March 1990, 21 patients underwent intraoperative monitoring of spinal cord evoked potentials (SpEPs) during surgical procedures for aortic lesions. There were 15 men and 6 women, with ages ranging between 40 and 74 years (mean 59.6 years). The clinical diagnoses were thoracic aortic aneurysms in 12 patients, thoraco-abdominal aortic aneurysms in 5 and dissecting aortic aneurysms (DeBakey III) in 4. In case 2, a Dacron patch was employed after excision of a localized saccular aneurysm. Graft replacement of the aneurysm with a Dacron tube was designed for the remaining 20 patients. To maintain distal aortic perfusion, an extra-anatomical temporary bypass was produced in 11 patients and femorofemoral bypass with a pump-oxygenator was used in the remaining 10. All patients were anesthetized with high doses of morphine. Partial pressures of oxygen and body temperature were well controlled except in two cases in whom there was some fluctuation (described later). A detailed profile of each patient is illustrated in Table I. Recording o f S p E P According to the method described by Tamaki et al.
(yrs), sex
58, 60, 59, 61,
54, 63, 65, 49,
no.
1 2 3 4
5 6 7 8
69, M 65, M
49, 58, 67, 40, 70, 69, 74, 59,
60, M
11 12
13 14 15 16 17 18 19 20
21
DB III b
TAAA DB III b TAA TAA TAA TAAA TAA TAA
TAA TAA
TAA TAAA
TAA DB III a DB III b TAA
TAA TAA TAAA TAAA
Diagnosis
D5
distal to LSA D3 D5 D5 . . D6 D3 D6
distal to LSA ascending aorta
D5 distal to LSA
distal to LSA distal to LSA distal to LSA D5
Ds localized in distal to LSA D8 D10
Proximal
135 unidentified
145 260
unidentified 120 160 120
60 75 110 305
Duration (rain)
D12
142
proximal to SMA 143 D9 105 D10 80 Dn 132 . . proximal to CA 85 Dl0 67 Dn 74
D5 D6
D8 proximal to CIA
D12 D10 proximal to CA Dio
proximal to CA proximal to CIA
D10
Distal
Aortic cross-clamping placement
60-100
4 0 - 80 6 0 - 90 8 0 - 90
60-100 5 0 - 70 4 0 - 80 70-110
5 0 - 90 40-110
5 0 - 90 3 0 - 50
70-120 unknown 7 0 - 90 80-120
unknown 8 0 - 90 70-110 70- 80
pressure during occlusion (mm Hg)
Distal aortic
none none none none none 26% decrease none latency shortening (0.3 msec) 47% decrease
none 85% decrease
none total disappearance
none none 56% decrease latency prolongation (0.9 msec) total disappearance none total disappearance none
Spike
SpEP changes
39% decrease
33% increase none none none none 25% decrease none 30% increase
33% decrease 77% decrease
pattern alteration none
none
32% decrease none 45% decrease marked increase (scaled out)
Polyphasic
yes
yes yes no
yes yes
no
yes yes no
yes yes no
Recovery
none
none none none none none none none none
none (died)
none paraplegia
(died) none none none
none none none none
complication
Neurological
massive bleeding from intercostal arteries c
distal hypoperfusion ~
total o c c l u s i o n / distal hypoperfusion b,c
cardiac arrest c
cardiac arrest c
total occlusion b.c b
Comments
a T A A = thoracic aortic aneurysm; T A A A = thoraco-abdominal aortic aneurysm; DB l i i a and DB III b = dissecting aortic aneurysm; DeBakey's type III a and IIl b, respectively; LSA = left subclavian artery; CA = celiac artery; C IA = common lilac artery. b Patients undergoing reimplantation of intercostal vessels. c Presumed causes of SpEP changes.
M F F M M F F M
60, M 42, M
9 10
M F M M
F M M M
Age
Case
Profile of 21 patients with spinal cord monitoring during aortic occlusion a
TABLE I
r-
0 rn
0
©
SPINAL CORD MONITORING FOR AORTIC ANEURYSM
(1984b), specially designed flexible catheter-type electrodes (Unique Medical Co. Ltd., Tokyo) were inserted by means of a Tuohy's needle into the subarachnoid space at the level of the conus medullaris and into the epidural space at the cervical level. One (mainly the rostral electrode) was used for stimulating, and the other for recording. Square-wave pulses of 0.2 msec duration and supramaximal intensity (10-20 mA) were applied at a rate of 10-30/sec. Records were obtained using one of the following: DISA N-2000 or Medelec MS-6 or MS-92B. Low and high frequency filters were at 20 and 2000 Hz, respectively, and the evoked potentials were averaged 20-100 times to eliminate electrical noise. The electrodes were removed more than 24 h after the operations in view of the hemorrhagic tendency due to the intraoperative use of heparin.
Results
Operative results The only patient receiving a Dacron patch (case 2) had an uneventful perioperative course. Of the remaining 20 patients who were scheduled to undergo a Dracon graft replacement, one (case 17) developed systemic hypotension (mean aortic pressure 30 mm Hg) due to acute heart failure prior to aortic occlusion, and the subsequent graft replacement was aborted. Two patients (cases 5 and 12) developed acute heart failure during aortic occlusion and died intraoperatively. One patient (case 10) with an extensive thoraco-abdominal aneurysm exhibited permanent paraplegia postoperatively. The remaining 16 patients successfully underwent graft replacements and were awakened without neurological complications.
Basic pattern of SpEP If both the electrodes are placed properly, SpEP can be recorded with either the rostral or caudal electrode (Fig. 1). In this series, SpEP was principally elicited by rostral stimulation and recorded caudally (descending SpEP). In some cases in which the descending SpEP could not be appropriately recorded, presumably due to incorrect placement of the caudal electrode, the ascending SpEP, using the caudal electrode to stimulate and the rostral to record, was used for monitoring. The basic SpEP pattern consisted of an initial spike and a subsequent polyphasic component (Fig. 1). As the wave configurations varied from patient to patient, the evoked potential initially recorded prior to the surgical procedure on the aorta was considered the control pattern for that patient. The peak-to-peak amplitudes of both the initial spike and the largest component of the polyphasic response were used as the main
317
initialspike
polyphasiccomponent
descendingSpEP
ascendingSpEP
+25JUV 2ms Fig. 1. Basic pattern of SpEP. Upper trace: descending SpEP elicited by rostral stimulation. Lower trace: ascending SpEP elicited by caudal stimulation. Typical wave form of SpEP consists of an initial spike and a polyphasic component. Upward deflections are negative in this and all other figures.
indicators, and pared with the cording to the peak latency of
an alteration of more than 20% comcontrol was considered significant accriteria of Tamaki et al. (1984b). The the spike was also used as an indicator.
SpEP changes during aortic occlusion An initial test clamping was performed in most of the cases before the aneurysm was incised. Both proximal and distal clamping was done, and the evoked potential was observed for 15 or 20 min to estimate collateral circulation to the spinal cord. The earliest and most frequent SpEP alterations after cross-clamping of the aorta were changes in the configuration or amplitude of the polyphasic component. Six patients (cases 1, 4, 8, 11, 13 and 20) demonstrated alterations of the polyphasic component without any change in the amplitude of the spike. In 13 patients who showed no change in the spike amplitude during the test clamping, subsequent graft replacements were successfully performed without reimplantation of intercostal vessels. Significant changes in the amplitude of the initial spike were observed in 7 patients (cases 3, 5, 7, 10, 12, 18 and 21), including 3 patients with total disappearance of SpEP. Among them, 2 patients (cases 5 and 7) developed sudden cardiac arrest during aortic occlusion and the evoked potential gradually faded out, first the polyphasic component and then the initial spike (Fig. 2). In case 10, SpEP amplitude decreased immediately after aortic cross-clamping and disappeared after the aneurysm was incised. The evoked potential did not return during the remainder of the surgery and the patient was found to have flaccid paraplegia postoperatively (Fig. 3). Changes in peak latency of the initial spike were observed in 2 patients. One was prolongation by 0.9 msec (case 4) and the other was shortening by 0.3 msec
318
Y. O K A M O T O E T AL.
(case 20). Since these alterations in latency were not accompanied by any change in the spike amplitude, body temperature fluctuation (a fall of 2.0°C and a rise of 1.1°C, respectively) was thought to be the main cause.
conlro] -.-,-- 14:44
aortic clamp
-4-- Itt: tt 7
aneurysm incised
l,uV
i; 2ms
Discussion
Paraplegia is one of the most devastating complications of cardiovascular surgery requiring temporary occlusion of the aorta. The frequency of paraplegia following surgical repair of a thoracic or thoraco-abdominal aortic aneurysm is considerable and estimated to be approximately 5% (Wakabayashi et al. 1976; Crawford et al. 1981). The exact causative mechanism of the ischemic spinal cord injury remains unclear because the circulation of the spinal cord is extremely complex and the patterns vary with the individual
2ms
~,
~",~,,~".b~ J.
17.51
t
cardiac arrest 18.17
,82,
1:
18.33
18.45 18.52 --
F
^
.,~ x ~ . _ . ~ . _ ~
CPB on 19.07
~
19.37
lOpV
L
zo.3o
14;53
,L
Fig. 3. Intraoperative SpEP in case 10. The initial spike (arrow heads) decreased in amplitude after aortic cross-clamping, but the aneurysm was unfortunately incised before an alarm was given. The distal aortic pressure fell to approximately 40 m m Hg and the latency of the spike was markedly prolonged. SpEP then rapidly disappeared. No recovery of the potentials was observed and the patient was found to have flaccid paraplegia postoperatively.
control
<---
'''''~''J~'~
I; 2ms
Fig. 2. Intraoperative SpEP in case 5. Note the gradual change after cardiac arrest. The evoked potential gradually faded out, initially the polyphasic component and then the initial spike. Reappearance of SpEP was observed in the reverse order after initiation of the pump-oxygenator. This patient, however, could not be weaned from the pump-oxygenator and died intraoperatively.
(Dommisse 1974). However, 4 major risk factors have been proposed as follows: (1) prolonged aortic clamping (Laschinger et al. 1984), (2) elevated spinal fluid pressure due to proximal hypertension (Blaisdell and Cooley 1962), (3) distal aortic hypotension (Krieger and Spencer 1985), and (4) interruption of critical radicular arteries (Wadouh et al. 1984). Several investigators have recently reported that the somatosensory evoked potential (SEP) is a useful method for early detection of spinal cord ischemia and for prevention of paraplegia during surgery on the thoracic or thoraco-abdominal aorta (Coles et al. 1982; Cunningham et al. 1982; Laschinger et al. 1983; Kaplan et al. 1986; Krieger and Spencer 1985). Although this technique has the great merit that it is non-invasive, it also has several disadvantages. SEPs appear to be too sensitive to changes of body temperature, hemodilution, blood pressure, anesthetic drugs and operative manipulations (McWilliam et al. 1985; Thurner et al. 1985; Danto et al. 1988). Moreover, there are some reports of false negative results clinically (Mizrahi and Crawford 1984) or experimentally (Laschinger et al. 1984). Since conduction of the SEP is principally ascribed to the dorsal columns, patients with anterior spinal artery syndromes may demonstrate a normal SEP.
SPINAL CORD MONITORING FOR AORTIC ANEURYSM
Spinal cord evoked potentials (SpEPs) elicited by direct stimulation of the spinal cord have been developed and used clinically since 1972 in Japan. The effectiveness of SpEP monitoring in spine or spinal cord surgery has already been established (Imai et al. 1984; Tamaki et ai. 1984b, 1985b). If the stimulating electrode is placed in the cervical epidural space and the recording electrode in the subarachnoid space at the level of the conus medullaris, SpEPs (termed descending conductive SpEP) from normal subjects consist of an initial spike and a subsequent polyphasic component. There have been several studies to determine the pathways through which both components of SpEP are transmitted. Although the initial spike has been thought to be transferred through the dorsolateral column, Harada et al. (1984) and Toyoda and Kanda (1984) demonstrated that this component originates from many fibers of large diameter in all quadrants of the spinal cord. The polyphasic component is considered to be conducted mainly through the dorsal column. Some experimental results, however, seem to indicate that postsynaptic potentials also contribute to it (Tamaki et al. 1985a). In the present study, the most frequent alternations of the SpEP after cross-clamping of the aorta were those in the configuration or amplitude of the polyphasic component. There was no case in which change in the spike amplitude was observed prior to that of the polyphasic component. Our clinical results agree with the experimental data of Tamaki et al. (1984a, 1985a). The authors feel that this is chiefly due to the differential vulnerability of gray matter and white matter to ischemia (Gelfan and Tarlov 1955). Since long-tract conduction in the spinal cord is thought to be rather resistant to ischemia (Kobrine et al. 1979), the polyphasic component which contains synaptic activities is more sensitive than the initial spike. Several investigators have emphasized the importance of re-implantation of intercostal arteries (Cunningham et al. 1982; Wadouh et al. 1984). In our experience, however, all of the 13 patients who demonstrated no change in the amplitude of the initial spike during the test clamping underwent graft replacement without reimplantation of the intercostal arteries, and none had neurological changes postoperatively. Since a rapid decrease in the spike amplitude was observed during test clamping in case 3, re-implantation of intercostal arteries was performed. The authors believe that initial test clamping for 15 or 20 min is valuable for estimating collateral circulation to the spinal cord. In case 10, the only patient with postoperative paraplegia, SpEP acutely disappeared after an extreme decline in distal aortic pressure. Thus, distal hypoperfusion is though to be the major cause of spinal cord injury (Fig. 3). Moreover, SpEP varied proportionally to the distal aortic pressure in case 12 (Fig. 4). As
319
~
control aorll,celamp DAoPImmHgl
~-~,~
,,.S~,,..,,~.,~, 18:30
55
18:55
50
19:00
42
,.J',-~,-19:18
52
19:51
78
~---~
. . . .
~
^
~ t~' r~.
• ----=
~ , '~,,~.a~'~'-,--'----'--'
1 2uv 2ms Fig. 4. Intraoperative SpEP in case 12. The wave pattern is atypical. There are two spikes, presumably due to incorrect placement of the recording electrode. During aortic cross-clamping, distal aortic pressure (DAoP) fluctuated between 40 and 110 mm Hg. The amplitudes of both the initial spike and the polyphasic component were proportional to the distal aortic pressure (DAoP).
Kobrine et al. (1976) indicated, from experimental data in the rhesus monkey, an autoregulation mechanism of spinal cord blood flow (SCBF) is postulated to operate when mean arterial blood pressure (MAP) is above 50 mm Hg, whereas SCBF fails passively with MAP decrease below 50 mm Hg. Laschinger et al. (1983) reported that distal aortic pressure of less than 40 mm Hg resulted in spinal cord ischemia experimentally and clinically. Thus, maintenance of distal aortic perfusion seems to be extremely important for prevention of ischemic spinal cord injury. In summary, these results demonstrate that monitoring of SpEP is a useful method for real-time evaluation of spinal cord function and for early detection of spinal cord ischemia during cardiovascular surgery requiring aortic occlusion. The authors gratefully acknowledge Yasutsugu Nakagawa, M.D., First Department of Surgery, School of Medicine, Chiba University, and Shigeaki Uemura, M.D., Division of Cardiovascular Surgery,
320 Chiba Kaihin Hospital, for giving us the opportunity to monitor their patients. We also acknowledge Ahmmed Ally, M.D., for kindly checking this manuscript.
References Blaisdell, F.W. and Cooley, D.A. The mechanism of paraplegia after temporary thoracic aortic occlusion and its relationship to spinal fluid pressure. Surgery, 1962, 51: 351-355. Coles, J.G., Wilson, G.J., Sima, A.F., Klement, P. and Tait, G.A. Intraoperative detection of spinal cord ischemia using somatosensory cortical evoked potentials during thoracic aortic occlusion. Ann. Thorac. Surg., 1982, 34: 299-306. Crawford, E.S., Walker, H.S.J., Saleh, S.A. and Normann, N.A. Graft replacement of aneurysm in descending thoracic aorta: results without bypass or shunting. Surgery, 1981, 89: 73-85. Cunningham, Jr., J.N., Laschinger, J.C., Merkin, H.A., Nathan, I.M., Colvin, S., Ransohoff, J. and Spencer, F.C. Measurement of spinal cord ischemia during operations upon the thoracic aorta. Initial clinical experience. Ann. Surg., 1982, 196: 285-296. Danto, J., Cataletto, M. and Wolpin, M. Evoked potential monitoring of anesthetic and operative manipulation. In: T.B. Ducker and R.H. Brown (Eds.), Neurophysiology and Standards of Spinal Cord Monitoring. Springer, New York, 1988: 157-162. Dommisse, G.F. The blood supply of the spinal cord. A critical vascular zone in spinal surgery. J. Bone Jt Surg., 1974, 56B: 225-235. Gelfan, S. and Tarlov, I.M. Differential vulnerability of spinal cord structures to anoxia. J. Neurophysiol., 1955, 18: 170-188. Harada, Y., Takemitsu, Y., Atsuta, Y. and Imai, M. Determination of the pathways of ascending and descending conductive spinal cord evoked potentials (SCEP). In: S. Homma and T. Tamaki (Eds.), Fundamentals and Clinical Application of Spinal Cord Monitoring. Saikon Press, Tokyo, 1984: 33-43. Imai, K., Kobayashi, H., Nakagawa, T., Inoue, S. and Tamaki, T. Experiences and analysis of spinal cord monitoring during surgery. In: S. Homma and T. Tamaki (Eds.), Fundamentals and Clinical Application of Spinal Cord Monitoring. Saikon Press, Tokyo, 1984:211-221. Kaplan, B.J., Friedman, W.A., Alexander, J.A. and Hampson, S.R. Somatosensory evoked potential monitoring of spinal cord ischemia during aortic operations. Neurosurgery, 1986, 19: 82-90. Kobrine, A.I., Doyle, T.F. and Rizzoli, H.V. Spinal cord blood flow as affected by changes in systemic arterial blood pressure. J. Neurosurg., 1976, 44: 12-15. Kobrine, A.I., Evans, D.E. and Rizzoli, H.V. The effects of ischemia on long-tract neural conduction in the spinal cord. J. Neurosurg., 1979, 50: 639-644.
Y. OKAMOTO ET AL. Krieger, K.H. and Spencer, F.C. Is paraplegia after repair of coarctation of the aorta due principally to distal hypotension during aortic cross-clamping?. Surgery, 1985, 97: 2-7. Laschinger, J.C., Cunningham, Jr., J.N., Nathan, I.M., Knopp, E.A., Cooper, M.M. and Spencer, F.C. Experimental and clinical assessment of the adequacy of partial bypass in maintenance of spinal cord blood flow during operations on the thoracic aorta. Ann. Thorac. Surg., 1983, 36: 417-426. Laschinger, J.C., Cunningham, Jr., J.N., Cooper, M.M., Krieger, K., Nathan, I.M. and Spencer, F.C. Prevention of ischemic spinal cord injury following aortic cross-clamping: use of corticosteroids. Ann. Thorac. Surg., 1984, 38: 500-507. McWilliam, R.C., Conner, A.N. and Pollock, J.C.S. Cortical somatosensory evoked potentials during surgery for scoliosis and coarctation of the aorta. In: J. Schramm and S.J. Jones (Eds.), Spinal Cord Monitoring. Springer, Berlin, 1985: 167-172. Mizrahi, E.M. and Crawford, E.S. Somatosensory evoked potentials during reversible spinal cord ischemia in man. Electroenceph. clin. Neurophysiol., 1984, 58: 120-126. Tamaki, T., Noguchi, T., Takano, H., Tsuji, H. and Dincer, M.D. The effects of hypovolemic hypotension and hypoxia on the jeopardized spinal cord. In: S. Homma and T. Tamaki (Eds.), Fundamentals and Clinical Application of Spinal Cord Monitoring. Saikon Press, Tokyo, 1984a: 145-154. Tamaki, T., Noguchi, T., Takano, H., Tsuji, H., Nakagawa, T., Imai, K. and Inoue, S. Spinal cord monitoring as a clinical utilization of the spinal evoked potential. Clin. Orthop., 1984b, 184: 58-64. Tamaki, T., Takano, H. and Takakuwa, K. Spinal cord monitoring: basic principles and experimental aspects. CNS Trauma, 1985a, 2: 137-149. Tamaki, T., Takano, H., Takakuwa, K., Tsuji, H., Nakagawa, T., Imai, K. and Inoue, S. An assessment of the use of the spinal cord evoked potentials in prognosis estimation of injured spinal cord. In: J. Schramm and S.J. Jones (Eds.), Spinal Cord Monitoring. Springer, Berlin, 1985b: 221-226. Thurner, F., Schramm, J. and Romstock, J. Effects of fentanyl and enflurane on cortical and subcortical SEP during general anesthesia in man. In: J. Schramm and S.J. Jones (Eds.), Spinal Cord Monitoring. Springer, Berlin, 1985: 82-89. Toyoda, A. and Kanda, K. Origins of spinal cord potentials evoked by stimulation of the cat spinal cord. In: S. Homma and T. Tamaki (Eds.), Fundamentals and Clinical Application of Spinal Cord Monitoring. Saikon Press, Tokyo, 1984: 99-111. Wadouh, F., Lindemann, E., Arndt, C.F., Hetzer, R. and Borst, H.G. The arteria raducularis magna anterior as a decisive factor influencing spinal cord damage during aortic occlusion. J. Thorac. Cardiovasc. Surg., 1984, 88: 1-10. Wakabayashi, A. and Connolly, J.E. Prevention of paraplegia associated with resection of extensive thoracic aneurysms. Arch. Surg., 1976, 111: 1186-1189.