Journal of Clinical Neuroscience 22 (2015) 1403–1407
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Clinical Study
Intraoperative monitoring during decompression of the spinal cord and spinal nerves using transcranial motor-evoked potentials: The law of twenty percent Satoshi Tanaka a,⇑, Jun Hirao a, Hidehiro Oka b, Jiro Akimoto c, Junko Takanashi d, Junichi Yamada e a
Department of Neuro-Oncology and Neurosurgery, Tokyo Nishi Tokushukai Hospital, 3-1-1 Matsubara-cho, Akishima-city, Tokyo 196-0003, Japan Department of Neurosurgery, Kitasato University Medical Center, Kitamoto-city, Saitama, Japan Department of Neurosurgery, Tokyo Medical University, Shinjuku-ku, Tokyo, Japan d Department of Clinical Laboratory, Kitasato University Medical Center, Kitamoto-city, Saitama, Japan e Department of Clinical Laboratory, Tokyo Nishi Tokushukai Hospital, Akishima-city, Tokyo, Japan b c
a r t i c l e
i n f o
Article history: Received 10 September 2014 Accepted 3 March 2015
Keywords: Compound muscle action potential Compressive spinal disease Japan Orthopedic Association score Motor-evoked potential Transcranial stimulation
a b s t r a c t Motor-evoked potential (MEP) monitoring was performed during 196 consecutive spinal (79 cervical and 117 lumbar) surgeries for the decompression of compressive spinal and spinal nerve diseases. MEP monitoring in spinal surgery has been considered sensitive to predict postoperative neurological recovery. In this series, transcranial stimulation consisted of trains of five pulses at a constant voltage (200–600 V). For the normalization of MEP, we recorded compound muscle action potentials (CMAP) after peripheral nerve stimulation, usually on the median nerve at the wrist 2 seconds before or after each transcranial stimulation of the motor area, for all operations. The sensitivity and specificity of MEP monitoring was 100% and 97.4%, respectively, or 96.9% with or without CMAP compensation (if the threshold of postoperative motor palsy was defined as 20% relative amplitude rate [RAR]). The mean RAR after CMAP normalization, of the most affected muscle in the patient group with excellent postoperative results (recovery rate of a Japan Orthopedic Association score of more than 50%) was significantly higher than that in the other groups (p = 0.0224). All patients with an amplitude increase rate (AIR) with CMAP normalization of more than 20% achieved neurological recovery postoperatively. Our results suggest that if the RAR is more than 20%, postoperative motor palsy can be avoided in spinal surgery. If the AIR with normalization by CMAP after peripheral nerve stimulation is more than 20%, neurological recovery can be expected in spinal surgery. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Since most spinal surgeries are functional surgeries, neurological worsening after a spinal operation is never acceptable. To prevent such postoperative neurological worsening, intraoperative neurophysiological monitoring is widely applied in spinal surgery [1–5]. Motor symptoms are most important in operations for the treatment of compressive spinal and spinal nerve diseases, and the motor-evoked potential (MEP) is an important target for neurophysiological monitoring in spinal surgery [6,7]. MEP has become popular due to recent rapid advances with propofol anesthesia and the train stimulation method [1,8]. To record MEP, we must stimulate the primary motor cortex in the frontal lobe or ⇑ Corresponding author. Tel.: +81 425004433. E-mail address: [email protected] (S. Tanaka). http://dx.doi.org/10.1016/j.jocn.2015.03.011 0967-5868/Ó 2015 Elsevier Ltd. All rights reserved.
pyramidal tract by one of two methods. The first involves direct stimulation of the motor cortex at 10–20 milliamps using subdural electrodes over the primary motor area by phase reversal of the somatosensory evoked potential (SEP; cortical MEP) [7,9–13]. The other method involves high voltage (several hundred volts) transcranial stimulation using screw electrodes that have been placed in the scalp [11,14–16]. Transcranial stimulation of the motor area with recording of the peripheral electromyogram (EMG) is the most popular method of MEP monitoring during spinal surgery and it has also been shown to be very sensitive [17–19]. Intraoperative MEP monitoring has been performed solely for the prevention of newly developed postoperative motor palsy. Interestingly, recovery from mild motor symptoms such as a decrease in grasp force, impairment of fine motor skills and intermittent claudication, which are commonly seen in patients
1404
S. Tanaka et al. / Journal of Clinical Neuroscience 22 (2015) 1403–1407
undergoing spinal surgery, may be predicted by MEP monitoring [5]. In this study, we examined not only the threshold relative amplitude for postoperative motor palsy but also the usefulness of intraoperative monitoring by MEP to predict postoperative neurological recovery in compressive spinal and spinal nerve diseases.
2. Materials and methods 2.1. Patient population Intraoperative MEP monitoring was performed in 196 consecutive spinal surgeries for the treatment of compressive spinal and spinal nerve diseases from April 2006 to February 2015. The surgeries consisted of 79 cervical spinal operations including 42 cervical laminoplasties (29 at C3–6) and 37 cervical anterior fusions (16 for two levels and 32 included C5/6), 117 lumbar spinal operations including 41 lumbar laminectomies (12 were at L4/5, 10 were at L3–5), 43 lumbar fenestrations (24 were for two levels and 32 included L4/5), 27 lumbar discectomies (12 were at L4/5 and 13 were at L5/S1) and five posterior lumbar-thoracic interbody fusions. All of the patients had motor symptoms to some extent and we did not operate on patients with only sensory symptoms, except for pain. Most of the patients had a history of neurological worsening 3 months before surgery. We obtained written informed consent from all of the patients, including permission for MEP monitoring.
2.2. MEP monitoring With regard to anesthesia, total intravenous anesthesia with propofol (usually at a constant dose of about 0.06 mg/kg/hour) was used in all operations [8]. As a muscle relaxant, vecuronium bromide at 0.1 mg/kg was usually used only for tracheal intubation. A set of screw electrodes (Unique Medical, Tokyo, Japan) with the cathode on the more affected side and the anode on the contralateral side were placed 2 cm anterior to C3 or C4 by the international 10–20 electroencephalogram system. The screw electrodes were inserted into the scalp and contacted the skull surface. They are stronger than conventional needle electrodes and resistant to high voltage electric stimulation. Stimulation consisted of trains of five pulses at a constant voltage by a Multi-Path D185 (Digitimer, Hertfordshire, UK) or Electric Stimulator SEN-4100 (Nihon Kohden, Tokyo, Japan). Stimulation at 200–400 V was most common in our series, except for patients in whom recording was difficult. The duration of each pulse was 0.2 milliseconds and the inter-pulse interval was 2 milliseconds. Surface electrodes for recording EMG responses were placed on the abductor pollicis brevis and abductor hallucis muscles as well as on other affected muscles. EMG were recorded with Neuropack-2, MEB-2208, MEB-9204, MEB-2306, or MEE-1208 (Nihon Kohden). The amplitude of each MEP was measured from the baseline to the first negative peak of the waves. Surface electrodes for applying stimulation for the normalization of MEP by compound muscle action potential (CMAP) after peripheral nerve stimulation were usually placed on the median nerve at the wrist. CMAP by single, bipolar supramaximum stimulation (20–50 milliamps), which had been determined at the beginning of the operation, usually on the median nerve at the wrist 2 seconds before or after each transcranial stimulation of the motor area, was recorded in all operations as previously described [19,20]. The amplitudes of MEP and CMAP after peripheral nerve stimulation were measured. The relative amplitude rate (RAR) and RAR normalized by the amplitude of CMAP were calculated
automatically with Microsoft Excel (Microsoft, Redmond, WA, USA). 2.3. Outcome assessment Postoperative motor palsy was defined as less than 2/5 in the Medical Research Council (MRC) manual muscle test grading system 1 week after the operation. Preoperative and 1 week, 1 month, 3 month, 6 month, and 1 year postoperative Japan Orthopedic Association (JOA) scores were assessed by a third party, and the recovery rate was defined as described by Hirabayashi [21]. Using the best JOA score postoperatively, all patients were divided into four groups according to recovery rate: excellent (E; JOA recovery rate P 50%), good (G; recovery rate 0 < JOA score < 50%), no change (N; JOA recovery rate = 0%), and worsened (W; JOA recovery rate < 0%). In each group, the mean RAR, with or without normalization by CMAP after peripheral nerve stimulation, was calculated and the RAR for functional recovery from compressive spinal and spinal nerve diseases was estimated from these results. The numbers in groups E and G compared to groups N and W, above and below the amplitude increasing rate (AIR) of the most affected muscle, were compared at every 10% increment within the range of 0–100% for convenience, with or without CMAP normalization in cervical operations. 3. Results MEP could be recorded in all 196 consecutive spinal operations. Overall, no adverse events were noted with high voltage transcranial stimulations or EMG recordings. In all patients, sufficient postoperative decompression was proved by postoperative imaging. 3.1. Sensitivity and specificity of MEP monitoring For each spinal operation, sensitivities and specificities were calculated according to the RAR (70–0%, every 10% interval) with or without normalization by CMAP after peripheral nerve stimulation. Receiver operating characteristic (ROC) analyses were employed to calculate the threshold for postoperative motor palsy (less than MRC grade 2/5), as shown in Figure 1 [22]. The threshold of postoperative motor palsy was 20% of RAR, and was also calculated by ROC analysis. In 196 spinal operations, the RAR of 10 patients without CMAP normalization and nine patients with CMAP normalization at the end of the operation was less than 20%. Among these patients, four who received cervical operations experienced new postoperative motor palsy of less than MRC grade 2/5. The other six and five patients, respectively, were considered false positives. If a 20% RAR of MEP was defined as the threshold for postoperative motor palsy, the specificity of MEP monitoring in our cervical operations was (196 6)/196 100 = 96.9% without CMAP normalization, and (196 5)/196 = 97.4% with CMAP normalization. None of the patients except the four mentioned above had new postoperative motor palsy. None of the patients were false negatives; none had postoperative motor palsy without a less than 20% decrease in RAR. Therefore, the sensitivity of MEP monitoring in cervical operations in our consecutive series was 100% at 20% RAR, with or without CMAP normalization. 3.2. Monitoring results and neurological recovery Among all the 196 patients who underwent surgery for compressive spinal and spinal nerve diseases, 110 were in group E, 75 were in group G, eight were in group N, and three were in group W. Box plots of the RAR in the four groups are shown in Figure 2.
S. Tanaka et al. / Journal of Clinical Neuroscience 22 (2015) 1403–1407
1405
Fig. 1. Plots of receiver operating characteristic (ROC) analyses employed to calculate the threshold for postoperative motor palsy of less than Medical Research Council grade 2/5. For each spinal operation, sensitivities and specificities were calculated according to relative amplitude rate (RAR; 70–0%, every 10% interval) with or without (±) normalization by compound muscle action potential (CMAP) after peripheral nerve stimulation. The threshold of postoperative motor palsy was 20% of RAR, calculated by ROC analysis.
Fig. 2. Box plots of the relative amplitude index of transcranial motor-evoked potentials during all 150 spinal surgeries for the treatment of compressive spinal and spinal nerve diseases, with (A; +) or without (B; ) normalization by compound muscle action potential (CMAP) after peripheral nerve stimulation. All patients were divided into four groups according to their recovery rate on the Japan Orthopedic Association (JOA) score at 1 month postoperatively: excellent (E; JOA recovery rate P 50%), good (G; recovery rate 0 < JOA < 50%), no change (N; JOA recovery rate = 0%), and worsened (W; JOA recovery rate < 0%). The mean relative amplitude rate in the excellent group was significantly higher than in the three other groups with CMAP normalization (A; Student’s t-test p = 0.0274). On the other hand, those without CMAP normalization were not significantly different from those in the three other groups (B; Student’s t-test p = 0.1326). Relative amplitude rates are presented as the mean ± standard deviation.
The mean RAR in group E was significantly higher than in the three other groups (Student’s t-test p = 0.0224). On the other hand, those without CMAP normalization were not significantly different from those in the three other groups (Student’s t-test p = 0.1427). The numbers in groups E and G compared to groups N and W, above and below the AIR of the most affected muscle, were compared at every 10% interval within the range of 0% to 100%, with or without CMAP normalization, in all spinal operations (Table 1). In all spinal operations with CMAP normalization, significant differences were observed between the rates in groups E and G above and below the AIR within the range of 10–70% by Fisher’s exact probability test (Table 1). All of the patients with an AIR above 20% were in group E or G. Therefore, 20% seemed to be the
cut off AIR using CMAP normalization for neurological recovery from compressive spinal and spinal nerve diseases. 4. Discussion Prior to the introduction of propofol anesthesia, intraoperative monitoring by SEP had been used in neurosurgery. SEP is easily obtained by placing electrodes on the scalp and is less susceptible to anesthesia, but it cannot detect changes in motor function. Since the introduction of propofol anesthesia, MEP monitoring has become more widespread [8]. In brain surgery, MEP has been reported to be more sensitive than SEP for detecting motor disturbances due to cerebral ischemia [12].
1406
S. Tanaka et al. / Journal of Clinical Neuroscience 22 (2015) 1403–1407
Table 1 Operative results judged by the change in JOA score and the amplitude increasing rate of MEP monitoring during 175 spinal operations Amplitude increasing rate (%)
60 >0 610 >10 620 >20 630 >30 640 >40 650 >50 660 >60 670 >70 680 >80 690 >90 6100 >100
CMAP normalization ( )
CMAP Normalization (+) *
Total
E and G
N and W
p value
Total
E and G
N and W
p value*
137 59 126 70 116 80 103 93 97 99 93 103 85 111 80 116 75 121 70 126 68 128
132 53 122 63 113 72 101 84 95 90 91 94 83 102 78 107 74 111 69 116 67 118
5 6 4 7 3 8 2 9 2 9 2 9 2 9 2 9 1 10 1 10 1 10
0.0906
105 91 93 103 86 110 80 116 76 120 69 127 59 137 55 141 51 145 50 146 44 152
102 83 92 91 86 99 80 105 76 109 69 116 59 126 55 130 51 134 50 135 44 141
3 8 1 12 0 11 0 11 0 11 0 11 0 11 0 11 0 11 0 11 0 11
0.1170
0.0572 0.0320 0.0270 0.0582 0.0618 0.1181 0.2048 0.0539 0.1009 0.1009
0.0029 0.0027 0.0082 0.0074 0.0178 0.0361 0.0362 0.0696 0.0688 0.1281
* p values were calculated by Fisher’s exact probability test. + = with, = without, CMAP = compound muscle action potential, E = excellent result, G = good result, JOA = Japan Orthopedic Association, MEP = motor-evoked potential, N = no change, W = worsened.
The JOA score reflects not only motor symptoms but also sensory symptoms [21]. However, MEP can only evaluate motor function. All of our patients had motor symptoms to some degree, and the recovery of motor function was the most important goal of the surgery. As we did not operate upon patients who only had sensory symptoms, the recovery of sensory symptoms was always accompanied by the recovery of motor symptoms. As they are so sensitive, MEP by transcranial stimulation and EMG recording are very effective for detecting postoperative motor palsy [6]. However, there is little consensus regarding the evaluation of changes in amplitude and the threshold for concern in MEP [7]. In direct cortical MEP recording with cervical epidural electrodes, a 50% reduction in the amplitude of the D-wave generally seems to be significant [13]. An 80% reduction in amplitude has been reported to be the threshold for postoperative motor palsy in direct cortical MEP with EMG recording [23]. Langeloo et al. also showed that an 80% reduction in amplitude was the threshold point in MEP with EMG recording in spinal surgery [3]. We previously reported that in 121 spinal operations for patients without preoperative motor palsy (MRC grading P 3/5), the sensitivity and specificity of MEP by EMG recording with CMAP normalization was 100% and 96.4%, respectively, using a criterion of an 80% reduction in amplitude [19,20]. These results are consistent with those of Langeloo et al. [3]. Thus, we defined a RAR of 20% as a threshold for new postoperative motor palsy. It is well known that transcranial MEP amplitude, even under physiological conditions, exhibits substantial trial to trial variability. CMAP normalization may overcome this variability. Furthermore, this variability is a result of the high sensitivity of transcranial MEP in spinal surgery. We suggest that a decrease in MEP is the result of a transient motor deficit which recovers when the patient awakens from anesthesia. MEP is so sensitive in spinal surgery that a subclinical transient motor deficit can be detected [4]. It is also well known that acute decompression can cause hyperemia in the spinal cord and is a major reason for false positive MEP findings in spinal surgery. Therefore, the specificity of MEP during a spinal operation is low compared to its very high sensitivity.
We used CMAP after peripheral nerve stimulation to normalize the changes in the MEP amplitude due to muscle relaxants, not anesthetics [20]. While a comparison of the CMAP throughout the surgical procedure should provide information about the muscle endplate, the effects of anesthesia, for example, excitation of alpha motor neurons, may result in a reduction of MEP that is not related to the surgical procedure. The development of propofol anesthesia has made it possible to perform MEP monitoring and propofol anesthesia can preserve the motor system in a relatively stable condition [8]. This may only be detected by comparing MEP on the healthy and affected sides. Indeed, while a comparison to MEP on the healthy side can normalize MEP for the effect of anesthetics on the excitation of alpha motor neurons, the control side is not necessarily unaffected by the operative procedure, especially in operations involving the brain and cervical spine. In a thoracic-lumbar spinal operation, comparison of the MEP of the upper and lower extremities appears to be useful. CMAP after peripheral nerve stimulation can constantly normalize for the effect of muscle relaxants in neurosurgical operations. Our CMAP normalization was shown to be effective under the constant administration of anesthetics. Although CMAP normalization provided only a very minor improvement in the sensitivity and specificity of MEP monitoring, as previously reported [19], CMAP normalization is necessary when using MEP to predict functional recovery from compressive spinal and spinal nerve disease. MEP has been used to predict and prevent neurological worsening postoperatively [6,7]. In our lumbar spinal disorder series, no neurological worsening occurred. MEP monitoring was not useful for predicting postoperative neurological worsening. Most of our spinal operation patients had relatively mild motor dysfunction, including a decrease in grasp force, impairment of fine motor skills or intermittent claudication. An increase of RAR may provide the information to decide on the decompression range in lumbar spinal canal fenestration. Also, in cervical anterior discectomies, an increase of RAR may provide the information required to determine whether to resect an ossified posterior longitudinal ligament. It can be difficult to assess the improvement of mild motor symptoms by MEP monitoring. There have been few reports on
S. Tanaka et al. / Journal of Clinical Neuroscience 22 (2015) 1403–1407
the prediction of recovery from motor function by MEP monitoring. Voulgaris et al. performed intraoperative MEP monitoring during 25 lumbar laminectomies for lumbar spinal canal stenosis [5]. Seventeen patients who had a greater than 50% increase in the amplitude of MEP experienced significant pain relief or recovery from motor dysfunction postoperatively. In our series, intraoperative MEP monitoring made it possible to predict, to some degree, motor neurological recovery from compressive spinal and spinal nerve disorders. If the AIR of intraoperative MEP monitoring predicts the postoperative neurological recovery, it would be very helpful to surgeons in deciding the range of decompression in spinal surgery. Our results suggest that if RAR is more than 20%, postoperative motor worsening can be avoided in spinal surgery. If AIR with normalization by CMAP after peripheral nerve stimulation is more than 20%, neurological recovery can be expected from spinal surgery. In cervical spinal surgery, MEP monitoring can be essential to predict and prevent postoperative motor palsy. In lumbar spinal surgery, MEP monitoring can be expected to predict postoperative motor functional recovery. Conflicts of Interest/Disclosures The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication. References [1] Mochida K, Shinomiya K, Komori H, et al. A new method of multisegment motor pathway monitoring using muscle potentials after tRARn spinal stimulation. Spine (Phila Pa 1976) 1995;20:2240–6. [2] Morota N, Deletis V, Constantini S, et al. The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery 1997; 41:1327–36. [3] Langeloo DD, Lelivelt A, Louis Journëe H, et al. Transcranial electrical motorevoked potential monitoring during surgery for spinal deformity: a study of 145 patients. Spine (Phila Pa 1976) 2003;28:1043–50. [4] Quinones-Hinojosa A, Lyon R, Zada G, et al. Changes in transcranial motor evoked potentials during intramedullary spinal cord tumor resection correlate with postoperative motor function. Neurosurgery 2005;56:982–93 [discussion 982–93]. [5] Voulgaris S, Karagiorgiadis D, Alexiou GA, et al. Continuous intraoperative electromyographic and transcranial motor evoked potential recordings in spinal stenosis surgery. J Clin Neurosci 2010;17:274–6. [6] Burke D, Hicks RG. Surgical monitoring of motor pathways. J Clin Neurophysiol 1998;15:194–205.
1407
[7] Kombos T, Kopetsch O, Suess O, et al. Does preoperative paresis influence intraoperative monitoring of the motor cortex? J Clin Neurophysiol 2003;20: 129–34. [8] Biebuyck JF. Propofol. An update on its clinical use. Anesthesiology 1994;81: 1005–43. [9] Levy WJ. Clinical experience with motor and cerebellar evoked potential monitoring. Neurosurgery 1987;20:169–82. [10] Kaneko M, Fukamachi A, Sasaki H, et al. Intraoperative monitoring of the motor function: experimental and clinical study. Acta Neurochir Suppl (Wien) 1988; 42:18–21. [11] Yamamoto T, Katayama Y, Fukaya S, et al. Comparison of the descending spinal cord evoked potentials with direct motor cortex stimulation and with transcranial brain stimulation. Clin Electroencephalogr 1998;40:162–6. [12] Szelenyi A, Bueno de Camargo A, Flamm E, et al. Neurophysiological criteria for intraoperative prediction of pure motor hemiplegia during aneurysm surgery. Case report. J Neurosurg 2003;99:575–8. [13] Yamamoto T, Katayama Y, Nagaoka T, et al. Intraoperative monitoring of the corticospinal motor evoked potential (D-wave): clinical index for postoperative motor function and functional recovery. Neurol Med Chir (Tokyo) 2004;44:170–80 [discussion 181–2]. [14] Rothwell J, Burke D, Hicks RG. Transcranial electrical stimulation of the motor cortex in man: further evidence for the site of activation. J Physiol 1994;481: 243–50. [15] MacDonald DB. Safety of intraoperative transcranial electrical stimulation motor-evoked potential monitoring. J Clin Neurophysiol 2002;19:416–29. [16] Iwasaki M, Kuroda S, Niiya Y, et al. Sensitivity of motor evoked potential (MEP) to intraoperative cerebral ischemia: case report. Jpn J Neurosurg (Tokyo) 2008; 17:622–6. [17] Deletis V, Isgum V, Amassian VE. Neurophysiological mechanisms underlying motor-evoked potentials in anesthetized humans Part 1. Recovery time of corticospinal tract waves elicited by pairs of transcranial electrical stimulation. Clin Neurophysiol 2001;112:438–44. [18] Deletis V, Rodi Z, Amassian VE. Neurophysiological mechanisms underlying motor-evoked potentials in anesthetized humans. Part 2. Relationship between epidurally and muscle recorded MEPs in man. Clin Neurophysiol 2001;112:445–52. [19] Tanaka S, Tashiro T, Gomi A, et al. Sensitivity and specificity in transcranial motor-evoked potential monitoring during neurosurgical operations. Surg Neurol Int 2011;2:111. [20] Tanaka S, Kobayashi I, Sagiuchi T, et al. Compensation of intraoperative transcranial motor-evoked potential monitoring by compound muscle action potential after peripheral nerve stimulation. J Clin Neurophysiol 2005;22: 271–4. [21] Japanese Orthopaedic Association: [Japanese Orthopaedic Association scoring system for cervical myelopathy (17–2 version and 100 version)]. Nippon Seikeigeka Gakkai Zasshi 1994;68:49003 [Japanese with English translation]. [22] Daniels L, Worthingham C. Muscle testing: techniques of manual examination. 5th ed. Philadelphia: WB Saunders Co.; 1986. [23] Kombos T, Suess O, Ciklatekerlio O, et al. Monitoring of intraoperative motor evoked potentials to increase the safety of surgery in and around the motor cortex. J Neurosurg 2001;95:608–14.
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Clinical Study
Intraoperative monitoring during decompression of the spinal cord and spinal nerves using transcranial motor-evoked potentials: The law of twenty percent Satoshi Tanaka a,⇑, Jun Hirao a, Hidehiro Oka b, Jiro Akimoto c, Junko Takanashi d, Junichi Yamada e a
Department of Neuro-Oncology and Neurosurgery, Tokyo Nishi Tokushukai Hospital, 3-1-1 Matsubara-cho, Akishima-city, Tokyo 196-0003, Japan Department of Neurosurgery, Kitasato University Medical Center, Kitamoto-city, Saitama, Japan Department of Neurosurgery, Tokyo Medical University, Shinjuku-ku, Tokyo, Japan d Department of Clinical Laboratory, Kitasato University Medical Center, Kitamoto-city, Saitama, Japan e Department of Clinical Laboratory, Tokyo Nishi Tokushukai Hospital, Akishima-city, Tokyo, Japan b c
a r t i c l e
i n f o
Article history: Received 10 September 2014 Accepted 3 March 2015
Keywords: Compound muscle action potential Compressive spinal disease Japan Orthopedic Association score Motor-evoked potential Transcranial stimulation
a b s t r a c t Motor-evoked potential (MEP) monitoring was performed during 196 consecutive spinal (79 cervical and 117 lumbar) surgeries for the decompression of compressive spinal and spinal nerve diseases. MEP monitoring in spinal surgery has been considered sensitive to predict postoperative neurological recovery. In this series, transcranial stimulation consisted of trains of five pulses at a constant voltage (200–600 V). For the normalization of MEP, we recorded compound muscle action potentials (CMAP) after peripheral nerve stimulation, usually on the median nerve at the wrist 2 seconds before or after each transcranial stimulation of the motor area, for all operations. The sensitivity and specificity of MEP monitoring was 100% and 97.4%, respectively, or 96.9% with or without CMAP compensation (if the threshold of postoperative motor palsy was defined as 20% relative amplitude rate [RAR]). The mean RAR after CMAP normalization, of the most affected muscle in the patient group with excellent postoperative results (recovery rate of a Japan Orthopedic Association score of more than 50%) was significantly higher than that in the other groups (p = 0.0224). All patients with an amplitude increase rate (AIR) with CMAP normalization of more than 20% achieved neurological recovery postoperatively. Our results suggest that if the RAR is more than 20%, postoperative motor palsy can be avoided in spinal surgery. If the AIR with normalization by CMAP after peripheral nerve stimulation is more than 20%, neurological recovery can be expected in spinal surgery. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Since most spinal surgeries are functional surgeries, neurological worsening after a spinal operation is never acceptable. To prevent such postoperative neurological worsening, intraoperative neurophysiological monitoring is widely applied in spinal surgery [1–5]. Motor symptoms are most important in operations for the treatment of compressive spinal and spinal nerve diseases, and the motor-evoked potential (MEP) is an important target for neurophysiological monitoring in spinal surgery [6,7]. MEP has become popular due to recent rapid advances with propofol anesthesia and the train stimulation method [1,8]. To record MEP, we must stimulate the primary motor cortex in the frontal lobe or ⇑ Corresponding author. Tel.: +81 425004433. E-mail address: [email protected] (S. Tanaka). http://dx.doi.org/10.1016/j.jocn.2015.03.011 0967-5868/Ó 2015 Elsevier Ltd. All rights reserved.
pyramidal tract by one of two methods. The first involves direct stimulation of the motor cortex at 10–20 milliamps using subdural electrodes over the primary motor area by phase reversal of the somatosensory evoked potential (SEP; cortical MEP) [7,9–13]. The other method involves high voltage (several hundred volts) transcranial stimulation using screw electrodes that have been placed in the scalp [11,14–16]. Transcranial stimulation of the motor area with recording of the peripheral electromyogram (EMG) is the most popular method of MEP monitoring during spinal surgery and it has also been shown to be very sensitive [17–19]. Intraoperative MEP monitoring has been performed solely for the prevention of newly developed postoperative motor palsy. Interestingly, recovery from mild motor symptoms such as a decrease in grasp force, impairment of fine motor skills and intermittent claudication, which are commonly seen in patients
1404
S. Tanaka et al. / Journal of Clinical Neuroscience 22 (2015) 1403–1407
undergoing spinal surgery, may be predicted by MEP monitoring [5]. In this study, we examined not only the threshold relative amplitude for postoperative motor palsy but also the usefulness of intraoperative monitoring by MEP to predict postoperative neurological recovery in compressive spinal and spinal nerve diseases.
2. Materials and methods 2.1. Patient population Intraoperative MEP monitoring was performed in 196 consecutive spinal surgeries for the treatment of compressive spinal and spinal nerve diseases from April 2006 to February 2015. The surgeries consisted of 79 cervical spinal operations including 42 cervical laminoplasties (29 at C3–6) and 37 cervical anterior fusions (16 for two levels and 32 included C5/6), 117 lumbar spinal operations including 41 lumbar laminectomies (12 were at L4/5, 10 were at L3–5), 43 lumbar fenestrations (24 were for two levels and 32 included L4/5), 27 lumbar discectomies (12 were at L4/5 and 13 were at L5/S1) and five posterior lumbar-thoracic interbody fusions. All of the patients had motor symptoms to some extent and we did not operate on patients with only sensory symptoms, except for pain. Most of the patients had a history of neurological worsening 3 months before surgery. We obtained written informed consent from all of the patients, including permission for MEP monitoring.
2.2. MEP monitoring With regard to anesthesia, total intravenous anesthesia with propofol (usually at a constant dose of about 0.06 mg/kg/hour) was used in all operations [8]. As a muscle relaxant, vecuronium bromide at 0.1 mg/kg was usually used only for tracheal intubation. A set of screw electrodes (Unique Medical, Tokyo, Japan) with the cathode on the more affected side and the anode on the contralateral side were placed 2 cm anterior to C3 or C4 by the international 10–20 electroencephalogram system. The screw electrodes were inserted into the scalp and contacted the skull surface. They are stronger than conventional needle electrodes and resistant to high voltage electric stimulation. Stimulation consisted of trains of five pulses at a constant voltage by a Multi-Path D185 (Digitimer, Hertfordshire, UK) or Electric Stimulator SEN-4100 (Nihon Kohden, Tokyo, Japan). Stimulation at 200–400 V was most common in our series, except for patients in whom recording was difficult. The duration of each pulse was 0.2 milliseconds and the inter-pulse interval was 2 milliseconds. Surface electrodes for recording EMG responses were placed on the abductor pollicis brevis and abductor hallucis muscles as well as on other affected muscles. EMG were recorded with Neuropack-2, MEB-2208, MEB-9204, MEB-2306, or MEE-1208 (Nihon Kohden). The amplitude of each MEP was measured from the baseline to the first negative peak of the waves. Surface electrodes for applying stimulation for the normalization of MEP by compound muscle action potential (CMAP) after peripheral nerve stimulation were usually placed on the median nerve at the wrist. CMAP by single, bipolar supramaximum stimulation (20–50 milliamps), which had been determined at the beginning of the operation, usually on the median nerve at the wrist 2 seconds before or after each transcranial stimulation of the motor area, was recorded in all operations as previously described [19,20]. The amplitudes of MEP and CMAP after peripheral nerve stimulation were measured. The relative amplitude rate (RAR) and RAR normalized by the amplitude of CMAP were calculated
automatically with Microsoft Excel (Microsoft, Redmond, WA, USA). 2.3. Outcome assessment Postoperative motor palsy was defined as less than 2/5 in the Medical Research Council (MRC) manual muscle test grading system 1 week after the operation. Preoperative and 1 week, 1 month, 3 month, 6 month, and 1 year postoperative Japan Orthopedic Association (JOA) scores were assessed by a third party, and the recovery rate was defined as described by Hirabayashi [21]. Using the best JOA score postoperatively, all patients were divided into four groups according to recovery rate: excellent (E; JOA recovery rate P 50%), good (G; recovery rate 0 < JOA score < 50%), no change (N; JOA recovery rate = 0%), and worsened (W; JOA recovery rate < 0%). In each group, the mean RAR, with or without normalization by CMAP after peripheral nerve stimulation, was calculated and the RAR for functional recovery from compressive spinal and spinal nerve diseases was estimated from these results. The numbers in groups E and G compared to groups N and W, above and below the amplitude increasing rate (AIR) of the most affected muscle, were compared at every 10% increment within the range of 0–100% for convenience, with or without CMAP normalization in cervical operations. 3. Results MEP could be recorded in all 196 consecutive spinal operations. Overall, no adverse events were noted with high voltage transcranial stimulations or EMG recordings. In all patients, sufficient postoperative decompression was proved by postoperative imaging. 3.1. Sensitivity and specificity of MEP monitoring For each spinal operation, sensitivities and specificities were calculated according to the RAR (70–0%, every 10% interval) with or without normalization by CMAP after peripheral nerve stimulation. Receiver operating characteristic (ROC) analyses were employed to calculate the threshold for postoperative motor palsy (less than MRC grade 2/5), as shown in Figure 1 [22]. The threshold of postoperative motor palsy was 20% of RAR, and was also calculated by ROC analysis. In 196 spinal operations, the RAR of 10 patients without CMAP normalization and nine patients with CMAP normalization at the end of the operation was less than 20%. Among these patients, four who received cervical operations experienced new postoperative motor palsy of less than MRC grade 2/5. The other six and five patients, respectively, were considered false positives. If a 20% RAR of MEP was defined as the threshold for postoperative motor palsy, the specificity of MEP monitoring in our cervical operations was (196 6)/196 100 = 96.9% without CMAP normalization, and (196 5)/196 = 97.4% with CMAP normalization. None of the patients except the four mentioned above had new postoperative motor palsy. None of the patients were false negatives; none had postoperative motor palsy without a less than 20% decrease in RAR. Therefore, the sensitivity of MEP monitoring in cervical operations in our consecutive series was 100% at 20% RAR, with or without CMAP normalization. 3.2. Monitoring results and neurological recovery Among all the 196 patients who underwent surgery for compressive spinal and spinal nerve diseases, 110 were in group E, 75 were in group G, eight were in group N, and three were in group W. Box plots of the RAR in the four groups are shown in Figure 2.
S. Tanaka et al. / Journal of Clinical Neuroscience 22 (2015) 1403–1407
1405
Fig. 1. Plots of receiver operating characteristic (ROC) analyses employed to calculate the threshold for postoperative motor palsy of less than Medical Research Council grade 2/5. For each spinal operation, sensitivities and specificities were calculated according to relative amplitude rate (RAR; 70–0%, every 10% interval) with or without (±) normalization by compound muscle action potential (CMAP) after peripheral nerve stimulation. The threshold of postoperative motor palsy was 20% of RAR, calculated by ROC analysis.
Fig. 2. Box plots of the relative amplitude index of transcranial motor-evoked potentials during all 150 spinal surgeries for the treatment of compressive spinal and spinal nerve diseases, with (A; +) or without (B; ) normalization by compound muscle action potential (CMAP) after peripheral nerve stimulation. All patients were divided into four groups according to their recovery rate on the Japan Orthopedic Association (JOA) score at 1 month postoperatively: excellent (E; JOA recovery rate P 50%), good (G; recovery rate 0 < JOA < 50%), no change (N; JOA recovery rate = 0%), and worsened (W; JOA recovery rate < 0%). The mean relative amplitude rate in the excellent group was significantly higher than in the three other groups with CMAP normalization (A; Student’s t-test p = 0.0274). On the other hand, those without CMAP normalization were not significantly different from those in the three other groups (B; Student’s t-test p = 0.1326). Relative amplitude rates are presented as the mean ± standard deviation.
The mean RAR in group E was significantly higher than in the three other groups (Student’s t-test p = 0.0224). On the other hand, those without CMAP normalization were not significantly different from those in the three other groups (Student’s t-test p = 0.1427). The numbers in groups E and G compared to groups N and W, above and below the AIR of the most affected muscle, were compared at every 10% interval within the range of 0% to 100%, with or without CMAP normalization, in all spinal operations (Table 1). In all spinal operations with CMAP normalization, significant differences were observed between the rates in groups E and G above and below the AIR within the range of 10–70% by Fisher’s exact probability test (Table 1). All of the patients with an AIR above 20% were in group E or G. Therefore, 20% seemed to be the
cut off AIR using CMAP normalization for neurological recovery from compressive spinal and spinal nerve diseases. 4. Discussion Prior to the introduction of propofol anesthesia, intraoperative monitoring by SEP had been used in neurosurgery. SEP is easily obtained by placing electrodes on the scalp and is less susceptible to anesthesia, but it cannot detect changes in motor function. Since the introduction of propofol anesthesia, MEP monitoring has become more widespread [8]. In brain surgery, MEP has been reported to be more sensitive than SEP for detecting motor disturbances due to cerebral ischemia [12].
1406
S. Tanaka et al. / Journal of Clinical Neuroscience 22 (2015) 1403–1407
Table 1 Operative results judged by the change in JOA score and the amplitude increasing rate of MEP monitoring during 175 spinal operations Amplitude increasing rate (%)
60 >0 610 >10 620 >20 630 >30 640 >40 650 >50 660 >60 670 >70 680 >80 690 >90 6100 >100
CMAP normalization ( )
CMAP Normalization (+) *
Total
E and G
N and W
p value
Total
E and G
N and W
p value*
137 59 126 70 116 80 103 93 97 99 93 103 85 111 80 116 75 121 70 126 68 128
132 53 122 63 113 72 101 84 95 90 91 94 83 102 78 107 74 111 69 116 67 118
5 6 4 7 3 8 2 9 2 9 2 9 2 9 2 9 1 10 1 10 1 10
0.0906
105 91 93 103 86 110 80 116 76 120 69 127 59 137 55 141 51 145 50 146 44 152
102 83 92 91 86 99 80 105 76 109 69 116 59 126 55 130 51 134 50 135 44 141
3 8 1 12 0 11 0 11 0 11 0 11 0 11 0 11 0 11 0 11 0 11
0.1170
0.0572 0.0320 0.0270 0.0582 0.0618 0.1181 0.2048 0.0539 0.1009 0.1009
0.0029 0.0027 0.0082 0.0074 0.0178 0.0361 0.0362 0.0696 0.0688 0.1281
* p values were calculated by Fisher’s exact probability test. + = with, = without, CMAP = compound muscle action potential, E = excellent result, G = good result, JOA = Japan Orthopedic Association, MEP = motor-evoked potential, N = no change, W = worsened.
The JOA score reflects not only motor symptoms but also sensory symptoms [21]. However, MEP can only evaluate motor function. All of our patients had motor symptoms to some degree, and the recovery of motor function was the most important goal of the surgery. As we did not operate upon patients who only had sensory symptoms, the recovery of sensory symptoms was always accompanied by the recovery of motor symptoms. As they are so sensitive, MEP by transcranial stimulation and EMG recording are very effective for detecting postoperative motor palsy [6]. However, there is little consensus regarding the evaluation of changes in amplitude and the threshold for concern in MEP [7]. In direct cortical MEP recording with cervical epidural electrodes, a 50% reduction in the amplitude of the D-wave generally seems to be significant [13]. An 80% reduction in amplitude has been reported to be the threshold for postoperative motor palsy in direct cortical MEP with EMG recording [23]. Langeloo et al. also showed that an 80% reduction in amplitude was the threshold point in MEP with EMG recording in spinal surgery [3]. We previously reported that in 121 spinal operations for patients without preoperative motor palsy (MRC grading P 3/5), the sensitivity and specificity of MEP by EMG recording with CMAP normalization was 100% and 96.4%, respectively, using a criterion of an 80% reduction in amplitude [19,20]. These results are consistent with those of Langeloo et al. [3]. Thus, we defined a RAR of 20% as a threshold for new postoperative motor palsy. It is well known that transcranial MEP amplitude, even under physiological conditions, exhibits substantial trial to trial variability. CMAP normalization may overcome this variability. Furthermore, this variability is a result of the high sensitivity of transcranial MEP in spinal surgery. We suggest that a decrease in MEP is the result of a transient motor deficit which recovers when the patient awakens from anesthesia. MEP is so sensitive in spinal surgery that a subclinical transient motor deficit can be detected [4]. It is also well known that acute decompression can cause hyperemia in the spinal cord and is a major reason for false positive MEP findings in spinal surgery. Therefore, the specificity of MEP during a spinal operation is low compared to its very high sensitivity.
We used CMAP after peripheral nerve stimulation to normalize the changes in the MEP amplitude due to muscle relaxants, not anesthetics [20]. While a comparison of the CMAP throughout the surgical procedure should provide information about the muscle endplate, the effects of anesthesia, for example, excitation of alpha motor neurons, may result in a reduction of MEP that is not related to the surgical procedure. The development of propofol anesthesia has made it possible to perform MEP monitoring and propofol anesthesia can preserve the motor system in a relatively stable condition [8]. This may only be detected by comparing MEP on the healthy and affected sides. Indeed, while a comparison to MEP on the healthy side can normalize MEP for the effect of anesthetics on the excitation of alpha motor neurons, the control side is not necessarily unaffected by the operative procedure, especially in operations involving the brain and cervical spine. In a thoracic-lumbar spinal operation, comparison of the MEP of the upper and lower extremities appears to be useful. CMAP after peripheral nerve stimulation can constantly normalize for the effect of muscle relaxants in neurosurgical operations. Our CMAP normalization was shown to be effective under the constant administration of anesthetics. Although CMAP normalization provided only a very minor improvement in the sensitivity and specificity of MEP monitoring, as previously reported [19], CMAP normalization is necessary when using MEP to predict functional recovery from compressive spinal and spinal nerve disease. MEP has been used to predict and prevent neurological worsening postoperatively [6,7]. In our lumbar spinal disorder series, no neurological worsening occurred. MEP monitoring was not useful for predicting postoperative neurological worsening. Most of our spinal operation patients had relatively mild motor dysfunction, including a decrease in grasp force, impairment of fine motor skills or intermittent claudication. An increase of RAR may provide the information to decide on the decompression range in lumbar spinal canal fenestration. Also, in cervical anterior discectomies, an increase of RAR may provide the information required to determine whether to resect an ossified posterior longitudinal ligament. It can be difficult to assess the improvement of mild motor symptoms by MEP monitoring. There have been few reports on
S. Tanaka et al. / Journal of Clinical Neuroscience 22 (2015) 1403–1407
the prediction of recovery from motor function by MEP monitoring. Voulgaris et al. performed intraoperative MEP monitoring during 25 lumbar laminectomies for lumbar spinal canal stenosis [5]. Seventeen patients who had a greater than 50% increase in the amplitude of MEP experienced significant pain relief or recovery from motor dysfunction postoperatively. In our series, intraoperative MEP monitoring made it possible to predict, to some degree, motor neurological recovery from compressive spinal and spinal nerve disorders. If the AIR of intraoperative MEP monitoring predicts the postoperative neurological recovery, it would be very helpful to surgeons in deciding the range of decompression in spinal surgery. Our results suggest that if RAR is more than 20%, postoperative motor worsening can be avoided in spinal surgery. If AIR with normalization by CMAP after peripheral nerve stimulation is more than 20%, neurological recovery can be expected from spinal surgery. In cervical spinal surgery, MEP monitoring can be essential to predict and prevent postoperative motor palsy. In lumbar spinal surgery, MEP monitoring can be expected to predict postoperative motor functional recovery. Conflicts of Interest/Disclosures The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication. References [1] Mochida K, Shinomiya K, Komori H, et al. A new method of multisegment motor pathway monitoring using muscle potentials after tRARn spinal stimulation. Spine (Phila Pa 1976) 1995;20:2240–6. [2] Morota N, Deletis V, Constantini S, et al. The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery 1997; 41:1327–36. [3] Langeloo DD, Lelivelt A, Louis Journëe H, et al. Transcranial electrical motorevoked potential monitoring during surgery for spinal deformity: a study of 145 patients. Spine (Phila Pa 1976) 2003;28:1043–50. [4] Quinones-Hinojosa A, Lyon R, Zada G, et al. Changes in transcranial motor evoked potentials during intramedullary spinal cord tumor resection correlate with postoperative motor function. Neurosurgery 2005;56:982–93 [discussion 982–93]. [5] Voulgaris S, Karagiorgiadis D, Alexiou GA, et al. Continuous intraoperative electromyographic and transcranial motor evoked potential recordings in spinal stenosis surgery. J Clin Neurosci 2010;17:274–6. [6] Burke D, Hicks RG. Surgical monitoring of motor pathways. J Clin Neurophysiol 1998;15:194–205.
1407
[7] Kombos T, Kopetsch O, Suess O, et al. Does preoperative paresis influence intraoperative monitoring of the motor cortex? J Clin Neurophysiol 2003;20: 129–34. [8] Biebuyck JF. Propofol. An update on its clinical use. Anesthesiology 1994;81: 1005–43. [9] Levy WJ. Clinical experience with motor and cerebellar evoked potential monitoring. Neurosurgery 1987;20:169–82. [10] Kaneko M, Fukamachi A, Sasaki H, et al. Intraoperative monitoring of the motor function: experimental and clinical study. Acta Neurochir Suppl (Wien) 1988; 42:18–21. [11] Yamamoto T, Katayama Y, Fukaya S, et al. Comparison of the descending spinal cord evoked potentials with direct motor cortex stimulation and with transcranial brain stimulation. Clin Electroencephalogr 1998;40:162–6. [12] Szelenyi A, Bueno de Camargo A, Flamm E, et al. Neurophysiological criteria for intraoperative prediction of pure motor hemiplegia during aneurysm surgery. Case report. J Neurosurg 2003;99:575–8. [13] Yamamoto T, Katayama Y, Nagaoka T, et al. Intraoperative monitoring of the corticospinal motor evoked potential (D-wave): clinical index for postoperative motor function and functional recovery. Neurol Med Chir (Tokyo) 2004;44:170–80 [discussion 181–2]. [14] Rothwell J, Burke D, Hicks RG. Transcranial electrical stimulation of the motor cortex in man: further evidence for the site of activation. J Physiol 1994;481: 243–50. [15] MacDonald DB. Safety of intraoperative transcranial electrical stimulation motor-evoked potential monitoring. J Clin Neurophysiol 2002;19:416–29. [16] Iwasaki M, Kuroda S, Niiya Y, et al. Sensitivity of motor evoked potential (MEP) to intraoperative cerebral ischemia: case report. Jpn J Neurosurg (Tokyo) 2008; 17:622–6. [17] Deletis V, Isgum V, Amassian VE. Neurophysiological mechanisms underlying motor-evoked potentials in anesthetized humans Part 1. Recovery time of corticospinal tract waves elicited by pairs of transcranial electrical stimulation. Clin Neurophysiol 2001;112:438–44. [18] Deletis V, Rodi Z, Amassian VE. Neurophysiological mechanisms underlying motor-evoked potentials in anesthetized humans. Part 2. Relationship between epidurally and muscle recorded MEPs in man. Clin Neurophysiol 2001;112:445–52. [19] Tanaka S, Tashiro T, Gomi A, et al. Sensitivity and specificity in transcranial motor-evoked potential monitoring during neurosurgical operations. Surg Neurol Int 2011;2:111. [20] Tanaka S, Kobayashi I, Sagiuchi T, et al. Compensation of intraoperative transcranial motor-evoked potential monitoring by compound muscle action potential after peripheral nerve stimulation. J Clin Neurophysiol 2005;22: 271–4. [21] Japanese Orthopaedic Association: [Japanese Orthopaedic Association scoring system for cervical myelopathy (17–2 version and 100 version)]. Nippon Seikeigeka Gakkai Zasshi 1994;68:49003 [Japanese with English translation]. [22] Daniels L, Worthingham C. Muscle testing: techniques of manual examination. 5th ed. Philadelphia: WB Saunders Co.; 1986. [23] Kombos T, Suess O, Ciklatekerlio O, et al. Monitoring of intraoperative motor evoked potentials to increase the safety of surgery in and around the motor cortex. J Neurosurg 2001;95:608–14.
