The management of progressive scoliosis in very young children is a demanding undertaking fraught with complications and failure.1–8 There is a risk of spinal cord injury between 0.5% and 0.72% during scoliosis surgery.9,10 These injuries occur by 2 primary mechanisms during surgical manipulation: mechanical damage from stretching and compression of nerve fibers or ischemia due to impaired blood supply.11 In 1973, the Stagnara test was first reported as a means to assess the patient’s neurological status after deformity correction before completion of the operation.12 Although this test decreased the incidence of postoperative neurological deficits it does not provide continuous monitoring and cannot be performed on young children, patients with primary muscle disease or those with significant cognitive impairment.13,14 Throughout the 1970s investigation of the use of electrophysiological techniques to monitor spinal cord integrity translated into the early intraoperative modality of somatosensory-evoked potentials (SSEPs) which provides monitoring of the dorsal column-medial lemniscus pathway.15
In 1980, Merton and Morton showed that a high-voltage pulse applied transcranially produced contralateral motor activity. This technique known as transcranial electric motor-evoked potentials (tceMEPs) was popularized in the early 1990s and became the gold standard for intraoperative neuromonitoring of the motor tracts during spinal correction surgery.16 Eccher17 reports that the combined use of SSEPs and tceMEPs produces sensitivities and specificities usually >90%, though variability across studies is recognized.
A previous report suggested that neuromonitoring of routine growing rod surgery lengthening may not be necessary due to the low incidence of neuromonitoring alerts.18 At our institution we became aware of a much higher rate of intraoperative monitoring alarms and took a different view of the subject. This review is an account of 9 years of spinal surgery for early-onset scoliosis (EOS) with its associated spinal cord monitoring alerts at a single institution involving 3 pediatric spine surgeons and a single neuromonitoring team led by a PhD neurophysiologist.
An Institutional Review Board-approved retrospective review of all surgical cases in EOS was performed from July 2003 to July 2012. All monitoring alerts were confirmed by the collaborative judgment of the attending orthopaedic and attending neurophysiologist-led neuromonitoring team. Two sources of this information were used: both the surgeons’ dictated operative note and the dictated neuromonitoring note. The 2 sources were cross referenced to assure that no monitoring events were missed. Both SSEPs and tceMEPs were studied. Triggered electromyography potentials at pedicle screw placement were routinely performed19 but not studied for this report.
After induction of anesthesia in the supine position, both stimulating and recording electrodes were appropriately placed for elicitation of posterior tibial SSEPs, tceMEPS for the upper and lower extremities, and free-run and triggered electroymyography from the instrumented levels (see Appendix I, Supplemental Digital Content 1, http://links.lww.com/BPO/A65). Total intravenous anesthesia was utilized given the significant hindering effects of halogenated agents on the elicitation of tceMEPs.11 Anesthetic management involved a total intravenous protocol of propofol (180 to 100 mcg/kg/min) and remifentanil (0.4 to 0.2 mcg/kg/min) with neuromuscular blockade (dose) administered only for intubation and on several occasions not at all. Bite blocks were appropriately placed for elicitation of tceMEPs. In addition to the other forms of neuromonitoring, bispectral index monitoring (Covidien; Plainfield, IN) was used and correlated with the anesthetic agents, clinical depth of anesthesia, blood pressure, pulse rate, and oxygen saturation. Preoperative hemoglobin assessment was used but often no intraoperative measurement of hematocrit was needed because the surgical time was so short.
All possible monitoring changes were assumed to be significant and reported.20 Transient changes associated with blood pressure fluctuation were included as were total loss of and unrecoverable signal changes due to merely positioning the patient. No attempt was made to distinguish what other reports deem to be “significant” changes from changes which might not lead to permanent neurological loss except that a 50% decrease in SSEPs or doubling of latency was utilized (see the Discussion section).
Monitoring of residual inhaled agents used at induction allowed assessment of the robustness of the monitoring data in the initial several minutes of the procedures. Baseline recordings of the SSEPs and tceMEPs were made before the patient being turned prone. If it was suspected that the SSEPs and/or tceMEPS were being degraded by any residual inhalational anesthesia used for induction before a steady-state, maintenance total intravenous anesthesia, the incision was postponed until improved baselines were obtained. Blood pressure and pulse rate were maintained at levels that would not compromise spinal cord blood flow.
The surgical procedures were performed with standard technique minimally disrupting soft tissue at anchor sites. Continuous monitoring of SSEPs and selective or intermittent monitoring of tceMEPs before and after routine and critical events during the surgical procedure was the standard monitoring protocol. Any changes in the neuromonitoring waveforms from postinduction baselines that met the traditional alarm criteria cited above for SSEPs and tceMEPs, or those that did not meet those criteria but were persistent were reported to the operating surgeon and a rescue protocol was followed.21 From an anesthesia standpoint this protocol included: elevating a depressed blood pressure either with intravenous fluid or pressors, maintaining adequate O2 saturation levels, and checking hematocrit if blood loss had been significant. From the surgical standpoint the protocol included: assessing the position of spinal or rib anchors both clinically and fluoroscopically, lessening distraction of the lengthening mechanism and, in cases where simple prone positioning of the patient led to loss of potentials, assessing and correcting the position of the patient on the table. Special attention was paid to the position of the neck in flexion or extension. If all the above measures failed to produce normalization of the potentials within several minutes then the procedure was aborted, a wake-up test was performed for clinical evaluation. The protocol was developed overtime at our institution and similar practices have been reported elsewhere (Fig. 1).22,23
Thirty patients underwent 180 cases. Thirty cases (17%) were not monitored. These were implant removal, incision and drainage of infection or cases where preexisting neurological impairment prevented monitoring of signals. Of the 150 neuromonitored cases, there were 18 (12%) neuromonitoring alerts, motor, or sensory (Table 1). This represented 40% of the patient cohort over the 9-year study period. Some alerts were transient and allowed surgery to proceed once rescue strategies had been applied. Others necessitated abandonment of the surgery and further study of spinal cord axis with magnetic resonance imaging. In 2 of these aborted cases patients needed corrective cervical spine surgery and went on to uneventful further lengthening and, in 1 case, later definitive fusion.
In 1 patient at time of definitive but staged fusion, there was a delayed L5 nerve root injury. Intraoperative neuromonitoring was normal, the nerve root was decompressed but despite this there was a delayed foot drop present 1 hour after surgery in the recovery room. This recovered by 90% over the next 12 months. No patient had a permanent neurological injury apart from this. No motor or sensory disturbance emerged clinically. Figure 2 shows the baseline and immediate postprone positioning of a patient with duplication of chromosome 1 and EOS previously successfully controlled with vertical expanding prosthetic titanium rib. At this subsequent surgery, merely positioning the patient prone resulted in virtually a complete loss of tceMEPs from the upper and lower extremities. The neck was flexed immediately with only partial recovery of signals so the case was aborted. The maneuver of neck flexion was chosen because of junctional kyphosis at the cervicothoracic level. Subsequent magnetic resonance imaging of the cervical spine under the same anesthetic (Fig. 3) showed a marked narrowing of the foramen magnum posteriorly, a phenomenon that was not present 2 years before and which had changed dramatically. The stenosis was decompressed at a later surgery and she underwent uncomplicated lengthening a few weeks later.
Routine neuromonitoring of spinal cord function potentials has gradually become standard of practice among spinal surgeons though not all procedures warrant the added expense of this technique. The lifetime cost of living with complete paraplegia after injury at 27 years old was estimated to be $1.6 million after adjusting for inflation.11,24 Although the cost of neuromonitoring at our institution for EOS implant lengthening ranges from $800 to 1000, it may warrant the added cost of monitoring even for lower risk procedures to prevent an injury from a cost analysis standpoint.
The questioning of this practice by Sankar et al18 led to this study revealing an alarming discrepancy between their data and the present findings. Explaining this discrepancy is not easy.
Although standards for spinal cord monitoring have been suggested, for instance by the American Society of Neurophysiological Monitoring, the implementation of these standards is variable. Recommendations for implementation of recovery strategies have also been made.
Possible scenarios that would explain the much higher incidence of monitoring alerts in our series include the possibility that our monitoring is more sensitive, that there are more neurophysiologically fragile patients in our population or that the methodological assessment of our data is more accurate. This last explanation is thought to be likely.
In Sankar and colleagues’ study the sentinel events were identified by querying a database. Further information was obtained by sending out questionnaires to the surgeons whose cases were complicated by monitoring alerts. This presupposes that every single monitoring alert, however small, made its way into the database. Allowing for the multicenter nature of this retrospective study and the lack of standardization of monitoring protocols it is unlikely that every alert was identified. Such methodologies may exclude significant data points that support the counterpoint to Sankar's argument: “absence of evidence does not constitute evidence of absence.”25
In the present study there was a set monitoring protocol for every patient, no patient data was lost, the surgeon’s report was cross referenced with the neurophysiologist’s monitoring report, and thus, every monitoring alert no matter how small was identified. According to Stecker and Robertshaw26 study, the neurophysiologist’s level of education and years of experience are positive predictors of neurological-deficit reduction.
Our neuromonitoring team is led by a PhD physiologist with more than 30 years’ experience. There were no particular identifiable intraoperative maneuvers that resulted in alerts. Several patients had alerts with merely prone positioning whereas others had alerts late into procedures. Baselines are routinely done in our practice before the patient is placed prone. There is universal agreement as to what constitutes a significant change in SSEPs namely a 50% reduction in amplitude or a doubling of latency. However, the quoted 60% decrease in TcMEPs is arbitrary and has no agreed methodology of measurement. TcMEPs are highly polyphasic and not averaged over many minutes as are SSEPs. The waveforms vary greatly within patients and over time. Attempts have been made to measure the “area under the curve” but agreement is lacking. The only consensus is on “all or nothing.” In other words, if any motor response is obtained to central stimulation then this is assumed to indicate an intact motor pathway. The problem with this theory is that TcMEPs are estimated to measure activity in at best 15% of the anatomic motor pathways so missing a large percentage of pathways which may be important. In addition there is evidence that the phenomenon of increased required numbers of stimuli centrally and increases voltage requirements to obtain any peripheral response are important. This may indicate a gradual fatigue of the monitored pathways and not a threshold number that defines significance. Until more objective evidence of the validity of a single number is produced, the validity of the arbitrary 60% number remains in question.
This study reports a 12% per surgery and 40% per patient incidence of neuromonitoring alerts in EOS surgery. This may actually underestimate the incidence of alerts; 1 patient did not incur any alerts during the study period, but had 3 monitoring alerts in subsequent surgeries. Index versus routine lengthening rate of alerts showed no difference. However, several patients whose primary implantation surgeries were uneventful had monitoring alerts later in their treatment course. We believe that the incidence of alerts is underreported because of methodological errors in previous reports. The evidence presented in this study supports spinal cord monitoring throughout all stages of surgical management of EOS.
1. Owen J. The application of intraoperative monitoring during surgery for spinal deformity. Spine. 1999;24:2649–2662.
2. Phillips J, Knapp D, Herrera-Soto J. Mortality and morbidity in early-onset scoliosis
surgery. Spine. 2013;38:324–327.
3. Pelosi L, Lamb J, Grevitt M, et al. Combined monitoring of motor and somatosensory evoked potentials in orthopaedic spinal surgery. Clin Neurophysiol. 2002;113:1082–1091.
4. MacDonald D. Safety of intraoperative transcranial electrical stimulation motor evoked potential monitoring. J Clin Neurophysiol. 2002;19:416–429.
5. Frei F, Ryhult S, Duitmann E, et al. Intraoperative monitoring of motor-evoked potentials in children undergoing spinal surgery. Spine. 2007;32:911–917.
6. Nuwer M, Dawson E, Carlson L, et al. Somatosensory evoked potential spinal cord monitoring reduces neurologic deficits after scoliosis surgery: results of a large multicenter survey. Electroencephalogr Clin Neurophysiol. 1995;96:6–11.
7. Qiu Y, Wang S, Wang B, et al. Incidence and risk factors of neurological deficits
of surgical correction for scoliosis: analysis of 1373 cases at one Chinese institution. Spine. 2008;33:519–526.
8. Fehlings M, Brodke D, Norvell D, et al. The evidence for intraoperative neurophysiological monitoring in spine surgery: does it make a difference? Spine. 2010;35(suppl):S37–S46.
9. Scoliosis Research Society. Neuromonitoring information statement; 2009. Available at: http://www.srs.org/professionals/
. Accessed January 2, 2014.
10. Anthony LRM, Peter RL, Abhay SR, et al. A survey of current controversies in scoliosis surgery in the United Kingdom. Spine. 2012;37:1573–1578.
11. Møller AR, Ansari S, Cohen-Gadol AA. Techniques of intraoperative monitoring for spinal cord function: their past, present, and future directions. Neurol Res. 2011;33:363–370.
12. Vauzelle C, Stagnara P, Jouvinroux P. Functional monitoring of spinal cord activity during spinal surgery. Clin Orthop Relat Res. 1973;93:173–178.
13. Vitale MG, Moore DW, Matsumoto H, et al. Risk factors for spinal cord injury during surgery for spinal deformity. J Bone Joint Surg Am. 2010;92:64–71.
14. Sutter M, Eggspuehler A, Grob D, et al. The diagnostic value of multimodal intraoperative monitoring (MIOM) during spine surgery: a prospective study of 1,017 patients. Eur Spine J. 2007;16(suppl):S162–S170.
15. Gonzalez A, Jeyanandarajan D, Hansen C. Intraoperative neurophysiological monitoring during spine surgery: a review. Neurosurg Focus. 2009;27:E6.
16. Modi H, Suh S-W, Yang J-H, et al. False-negative transcranial motor-evoked potentials during scoliosis surgery causing paralysis: a case report with literature review. Spine. 2009;34:E896–E900.
17. Eccher M. Intraoperative neurophysiologic monitoring: are we really that bad? J Clin Neurophysiol. 2012;29:157–159.
18. Sankar W, Skaggs D, Emans J, et al. Neurologic risk in growing rod spine surgery in early onset scoliosis: is neuromonitoring necessary for all cases? Spine. 2009;34:1952–1955.
19. Blas GD, Barrios C, Regidor I, et al. Safe pedicle screw placement in thoracic scoliotic curves using t-EMG. Spine. 2012;37:E387–E395.
20. Calancie B, Harris W, Broton J, et al. “Threshold-level” multipulse transcranial electrical stimulation of motor cortex for intraoperative monitoring of spinal motor tracts: description of method and comparison to somatosensory evoked potential monitoring. J Neurosurg. 1998;88:457–470.
21. Kamerlink J, Errico T, Xavier S, et al. Major intraoperative neurologic monitoring deficits in consecutive pediatric and adult spinal deformity patients at one institution. Spine. 2010;35:240–245.
22. Ziewacz J, Berven S, Mummaneni V, et al. The design, development, and implementation of a checklist for intraoperative neuromonitoring
changes. Neurosurg Focus. 2012;33:E11.
23. Langeloo D-D, Journée H-L, de Kleuver M, et al. Criteria for transcranial electrical motor evoked potential monitoring during spinal deformity surgery. A review and discussion of the literature. Clin Neurophysiol. 2007;37:431–439.
24. United States Department of Labor. CPI inflation calculator; 2013. Available at: http://www.bls.gov/data/inflation_calculator.htm
. Accessed January 2, 2014.
25. Altman D, Bland J. Statistics notes: absence of evidence is not evidence of absence. Br Med J. 1995;311:485.
26. Stecker MM, Robertshaw J. Factors affecting reliability of interpretations of intra-operative evoked potentials. J Clin Monit Comput. 2006;20:47–55.