Patients with cerebral palsy (CP) have a high prevalence of neuromuscular scoliosis (NMS), with 1 study reporting 28% of a cohort of 666 patients exhibiting mild (17%) or moderate to severe (11%) scoliosis on clinical examination.1 Many CP patients have severely limited mobility and require care providers for transportation, bathing, and activities of daily living. Surgical correction has been shown to improve the quality of life for both patients and caregivers, reducing pain and facilitating ease of transfers for nonambulatory patients.2 After surgical correction of scoliosis, caretakers have reported significant improvements in patients’ physical appearance, head control, trunk balance, sitting ability, respiratory function, and overall quality of life.2,3
A major complication of surgical treatment, however, is spinal cord injury and subsequent neurological insult. In patients with CP, surgical correction of scoliosis results in higher morbidity rates than in patients with adolescent idiopathic scoliosis (AIS).4,5 Reames et al5 showed that surgical correction of NMS had the highest rate of complications (17.9%) when compared with correction of congenital scoliosis (10.6%) and AIS (6.3%). In regard to neurological deficits in particular, surgical correction of NMS had a complication rate of 1.1% compared with 0.8% for correction of AIS. Intraoperative neuromonitoring (IONM) has been shown to reduce postoperative neurological deficits and is currently the standard of care in spinal fusion procedures.3,6–9 Initially used in the correction of idiopathic scoliosis, IONM has been shown to be efficacious in NMS correction as well.8,10–12 IONM allows intraoperative detection of neurological insults to the spinal cord and lets surgeons adjust and react accordingly. This is particularly important in patients with baseline motor deficits, in whom it is critical to preserve existing mobility and motor function.13
IONM for spine surgery is typically conducted by monitoring posterior tibial nerve somatosensory-evoked potentials (SSEP) and transcranial motor-evoked potentials (TcMEP). SSEP monitoring is specific to the ascending dorsal tracts of the spinal cord but provides no information regarding the descending anterior motor tracts.14 TcMEP monitoring uses direct depolarization of the pyramidal tract axons and conduction down the anterior motor tracts. Although it is possible to use either TcMEP or SSEP separately to assess the integrity of the spinal cord, combined multimodal monitoring of SSEP and TcMEP provides parallel redundancy. This results in concomitant assessment of the dorsal sensory and ventral motor columns with combined sensitivity and specificity approaching 100%.7,10,15,16
It can be difficult to obtain viable baseline signals in patients with preexisting neurological deficits. In patients with severe CP, the rate of IONM failure was 61% in 1 study.10 CP is a syndrome of motor impairment resulting from a static lesion that occurred in the developing brain.17 CP is a static encephalopathy. The underlying cause of CP varies in each patient and is generally perinatal or prenatal in nature. Once established, it does not progress over the course of a lifetime.18 Insults during the first trimester are associated with cerebral maldevelopment, such as schizencephaly; in the second trimester, periventricular white matter damage; and in the third trimester, cortical and deep gray matter damage.19 Other causes include asphyxia, traumatic brain injury, cerebrovascular accidents, and infection. Despite the variety of causes of CP, both known and unknown, there is much overlap in the clinical presentation of patients. CP patients requiring scoliosis correction often have preexisting neuroimaging in the form of computed tomographic (CT) and magnetic resonance imaging (MRI) at the time of preoperative planning. Such imaging is useful in determining the presence of cerebral lesions, including hydrocephalus, periventricular leukomalacia (PVL), encephalomalacia, abnormalities of the corpus callosum, and cerebral atrophy.
Most patients in our study population have limited function and mobility, characterized by spasticity and quadriplegia. Although clinically indistinguishable, some patients had complete interpretable signals (both SSEP and TcMEP), whereas others had “no sensory” (no SSEP regardless of interpretable TcMEP), “no motor” (no TcMEP regardless of interpretable SSEP), or “no signals” (neither TcMEP nor SSEP). This observation is counterintuitive because one may expect the degree of physical disability to correlate with the quality of signals. However, this was not the case in our study. Typically, inability to obtain signals is caused by a deficit in signal conduction, such as along the conductance tract.20 Inability to obtain TcMEP, in particular, usually suggests a lesion within the motor tracts. Patients with CP, however, typically have supratentorial brain lesions rather than lesions within the spinal cord or tract.21,22 A possible explanation for this is that cerebral abnormalities may be the reason for the lack of interpretable signals in these patients because they may interfere with signal generation. Analysis of anatomic findings on preexisting imaging may clarify this.
The purpose of this study was to identify potential anatomic predictors of the absence of interpretable IONM signals during surgical correction of scoliosis in patients with CP. We posit that such predictors, once identified, can improve preoperative planning for patients with scoliosis secondary to CP and, consequently, patient care.
In this single-institution retrospective cohort study, we examined data from patients diagnosed with CP who underwent surgical correction of scoliosis from 2002 through 2013. After institutional review board approval, we obtained and deidentified patient records for all scoliosis correction procedures during this period. Patients were excluded if no neuroimaging data were available and if IONM records did not include both SSEP and TcMEP data. IONM data were initially categorized into 4 groups: (1) the ability to attain both SSEP and TcMEP; (2) the inability to obtain both SSEP and TcMEP; (3) the ability to obtain SSEP irrespective of TcMEP; and (4) the ability to obtain TcMEP irrespective of SSEP. We examined TcMEP and cortical SSEP because of the high efficacy and wide use of multimodal IONM. Neuromonitoring data were reviewed for the presence and quality of subcortical SSEP in situations in which there was an inability to obtain cortical SSEP and TcMEP. In these patients, however, subcortical SSEP were also not present.
Gross Motor Function Classification System (GMFCS) level was assessed for each patient within the cohort. GMFCS level was excluded from analysis because of insufficient sample sizes within the different levels.
Existing neuroimaging records were available in the form of CT and/or MRI. The records were reviewed for the presence and severity of anatomic abnormalities, such as abnormalities of the corpus callosum, generalized cerebral atrophy, hydrocephalus, and encephalomalacia. Descriptions, classifications, and degree of severity of anatomic predictors were derived from radiologic reports. Because of the unique clinical history of each CP patient, imaging was limited. The neuroanatomic abnormalities were equally visible on both CT and MRI. From the total patient cohort, 26 patients had only CT data, 28 had only MRI data, 39 had both MRI and CT data, and 57 had no imaging. Because MRI and CT data were equally informative for determining neurological deficits, we excluded patients without any relevant imaging and included those with either MRI data, CT data, or both.
The primary outcome was the inability to obtain IONM signals. Outcome categories were defined as “no signals” (neither reliable SSEP nor reliable TcMEP), “no sensory” (no reliable SSEP regardless of TcMEP status), and “no motor” (no TcMEP regardless of SSEP status).
Most patients received a narcotic-based anesthetic with hydromorphone, fentanyl, or methadone, in addition to a volatile agent (isoflurane or desflurane) at approximately 0.5 minimum alveolar concentration in 60% oxygen in air. Nitrous oxide was not used. Vecuronium was used for paralysis and titrated to 2 of 4 twitches in the train-of-four test.
Multimodal IONM (SSEP and TcMEP monitoring) was performed on all patients. Somatosensory-evoked responses were monitored via bilateral median nerve stimulation at the wrist and posterior tibial nerve stimulation at the ankle or popliteal fossa. Recordings were made from multiple scalp channels, a C5-cervical to Fpz channel, and in the legs, from the popliteal fossae. TcMEP were elicited after anodal train stimulation (5 to 7 stimuli at 300 to 500 Hz) with C1 (C2 cathode) or C2 (C1 cathode) anodes and an additional Cz site as needed. All scalp electrode locations were based on the International 10/20 System.23 Recordings were made via electrodes placed over the tibialis anterior, abductor hallucis, and abductor pollicis brevis muscles bilaterally with a reference electrode placed nearby in a relatively inactive site.20
Statistical analysis of univariate and multivariate associations was performed using logistic regression. Odds ratios (ORs) were calculated with significance set at P<0.05. Statistical analysis was performed using Stata/SE, version 12.1, software (StataCorp LP, College Station, TX).
Patient Demographic Characteristics
Of 206 patients initially identified, 93 met the study criteria. Patient characteristics are detailed in Table 1. In the complete cohort, 47 (51%) patients were male and 46 (49%) were female. Most patients [79 (85%)] underwent posterior spinal fusion procedures with pedicle screw and rod implantation. Thirteen patients (14%) underwent combined anterior and posterior spinal fusion, and 1 patient (1%) had posterior spinal fusion procedures with growing rod implantation. The mean age was 14.3±3.2 years (range, 5 to 26 y); 83 patients (89%) were less than 18 years old. GMFCS distribution was as follows: 78 (84%) patients were GMFCS level V, 8 (9%) were GMFCS level IV, 2 (2%) were GMFCS level III, 3 (3%) were GMFCS level II, and 2 (2%) were GMFCS level I. Statistical analysis did not show any added value to the GMFCS in these children. In our cohort, 31 patients had no signals, 19 had SSEP only, 9 had TcMEP only, and 34 had both SSEP and TcMEP. When regrouped into previously defined IONM groups, 31 had no signals, 9 had “no sensory,” and 19 had “no motor.” Most patients had suboptimal verbal communication, cognitive deficits, and motor deficits. Motor deficits were documented in 47 patients (51%). Of these patients, 40 were quadriplegic and 7 were diplegic. Spasticity was documented in 72 (77%) patients. Sixty-six patients (71%) had a history of seizures, 2 of whom (2%) had a remote history of seizures.
Neuroanatomic findings derived from preexisting imaging consisted of hydrocephalus (n=43), encephalomalacia (n=33), abnormalities of the corpus callosum (n=27), cerebral atrophy (n=23), PVL (n=21), porencephaly (n=7), and lissencephaly (n=7) (Table 2).
Univariate Analysis Results
Unadjusted significant relationships were found between neuroanatomic findings and the defined outcome groups of no signals, no sensory, and no motor (Table 3). Hydrocephalus, both moderate and marked, was significantly associated with all 3 outcomes. Moderate hydrocephalus was associated with no signals (OR=9.22; P<0.01), no motor (OR=8.17; P=0.01), and no sensory (OR=5.05; P=0.02). Marked hydrocephalus was associated with no signals (OR=4.50; P<0.01), no motor (OR=19.07; P<0.01), and no sensory (OR=2.94; P=0.02). Encephalomalacia was significantly associated with no signals (OR=4.16; P<0.01) and no motor (OR=4.75; P<0.01). Atrophy of the corpus callosum was significantly associated with no signals (OR=5.20; P=0.01) and no sensory (OR=5.04; P=0.02).
Multivariate Analysis Results
Adjusted multivariate analysis by logistic regression showed significant associations of focal PVL, hydrocephalus, and encephalomalacia with lack of interpretable signals (Table 4). Focal PVL (Fig. 1) was associated with no motor (OR=39.95; P=0.04). Marked hydrocephalus (Fig. 2) was associated with no motor (OR=20.46; P<0.01) and no signals (OR=8.83; P=0.01). Moderate hydrocephalus was associated with no signals (OR=32.25; P<0.01), no motor (OR=10.14; P=0.04), and no sensory (OR=8.44; P=0.03). Finally, encephalomalacia (Fig. 3) was associated with no motor (OR=6.99; P=0.01) and no signals (OR=4.26; P=0.03).
The results of our analysis show significant associations between IONM signals and anatomic findings in the brain. The strongest associations were with TcMEP, in particular in patients with hydrocephalus, showing associations with 2 of 3 defined outcome groups (Table 4). Moderate and severe hydrocephalus were significantly associated with decreased likelihood of attaining TcMEP signals and were also correlated with no signals (neither TcMEP nor SSEP).
The quality of interpretable IONM signals can differ substantially in patients with CP despite similar clinical presentations. Because IONM is the standard of care in the surgical correction of scoliosis, it is important to understand factors limiting the acquisition of utilizable signals.14,24–28 Standard TcMEP monitoring involves stimulating pyramidal cells of the motor cortex via electrical currents, resulting in a wave of depolarization that travels down the corticospinal tract.29 In SSEP monitoring, a peripheral nerve is stimulated, and the resulting evoked potential travels across the surgical field and is then recorded from the somatosensory cortex. If the motor or somatosensory cortex is anatomically altered or shifted, TcMEP generation, SSEP recording, or both may be adversely affected. Supratentorial cerebral lesions are a possible explanation for impaired signal quality.30 Alternatively, lesions within the spinal cord itself, either in the anterior or dorsal tract, may interfere with IONM by limiting or inhibiting signal conduction.31,32
The associations of hydrocephalus with no motor or no signals may be explained mechanistically. We hypothesize that with more severe hydrocephalus, enlargement of the ventricles causes displacement of cortical structures and subcortical pathways. This may adversely affect the generation of motor volleys after voltage stimulation for TcMEP. Furthermore, conduction of the electric stimulation impulses through cerebrospinal fluid in the ventricles may be ineffective, and signal transmission may be blocked at this point. In our patients, IONM-stimulating electrodes were placed in accordance with standards that are based on normal cortical architecture. Alternative electrode placement in patients with hydrocephalus may lead to successful TcMEP acquisition. This hypothesis may be verified in future studies.
Similarly, encephalomalacia, although categorized binarily as present or not present, was diffuse and cystic for many of our patients. In patients with severe encephalomalacia, there are greater distances between brain matter and bone. We believe this may cause increased resistance in the electrical circuit and therefore ineffective transcranial stimulation. Of note, focal PVL was found to be the most statistically significant factor associated with unobtainable IONM signals, particularly TcMEP, despite showing no association on univariate analysis. A potential explanation for the significance of PVL is focal lesions in the brain matter inhibiting signal generation. Alternatively, PVL may result in interference with impulse conduction to the spinal cortical tracts.
For the purposes of this study, we examined SSEP and TcMEP as separate outcomes, with an emphasis on TcMEP for 2 reasons. Although a valuable modality when used in isolation, SSEP monitoring has been reported to be at risk for false-negative findings. There are well-documented instances of poor neurological outcomes despite intact SSEP signals intraoperatively and postoperatively, suggesting neurological insult despite intact SSEP.14,33–35 Anatomically, a selective insult to the descending anterior motor tracts can occur while preserving the ascending dorsal tracts of the spinal cord. For all of these reasons, multimodal IONM is optimal.36 Direct monitoring of the integrity of the motor tracts is the most reliable method to detect spinal cord insult to the motor tracts. The second reason we emphasize TcMEP over SSEP stems from the fact that the anesthetic environment is typically optimized for TcMEP (at the expense of SSEP) during multimodal IONM at our institution.
There are several limitations to the current study. The retrospective nature of the study does not allow for the prospective determination of variables or accumulation of data. In addition, availability of variables and patient information were varied within the study cohort. A further consideration is that this study suggests correlations between predictors and outcomes without any means of establishing causality, which is left to further investigation.
The results of this study contribute to the understanding of when and how supratentorial brain lesions may affect IONM signal acquisition in patients with CP. These results could prove useful for preoperative prediction of meaningful IONM signals. They may also allow for the modification of current IONM techniques to enhance signal acquisition. Specifically, IONM personnel could review patient neuroimaging studies before procedures. All procedures should be treated and approached systematically. We are currently not suggesting any change in the preoperative evaluation of patients. Our findings suggest that neuroimaging, when available, would be useful to review preoperatively. We believe it will be possible to use preoperative imaging to identify patients with decreased likelihood of signal acquisition and to alert monitoring personnel and anesthesia staff before the case, as early as imaging is performed and reviewed. Patients with hydrocephalus, for example, might benefit from modified IONM electrode placement on the basis of these preoperative analyses. Wider electrode spacing might allow the signal generated after TcMEP stimulation to reach the displaced pathways. Therefore, on the basis of our findings, we believe it is possible to improve signal acquisition. This should be studied in parallel to current IONM practice. In the interim, conventional IONM practices are the standard of care and should be performed and troubleshot as appropriate.
IONM provides a valuable and tested means of reducing neurological insults. Neuroanatomic findings of hydrocephalus, encephalomalacia, and PVL are significantly correlated with unobtainable IONM signals, in particular TcMEP. It is therefore worth reviewing all available brain imaging studies in CP patients before scoliosis correction. Analysis of anatomic findings in these images is a promising method to predict IONM outcomes. Our findings have implications for how IONM is performed in patients with CP. Modified electrode placement deserves to be tested.
1. Persson-Bunke M, Hagglund G, Lauge-Pedersen H, et al. Scoliosis
in a total population of children with cerebral palsy
. Spine (Phila Pa 1976). 2012;37:E708–E713.
2. Obid P, Bevot A, Goll A, et al. Quality of life after surgery for neuromuscular scoliosis
. Orthop Rev (Pavia). 2013;5:e1.
3. Fehlings MG, Brodke DS, Norvell DC, et al. The evidence for intraoperative neurophysiological monitoring in spine surgery: does it make a difference? Spine (Phila Pa 1976). 2010;35:S37–S46.
4. Murphy NA, Firth S, Jorgensen T, et al. Spinal surgery in children with idiopathic and neuromuscular scoliosis
. What’s the difference? J Pediatr Orthop. 2006;26:216–220.
5. Reames DL, Smith JS, Fu KMG, et al. Complications in the surgical treatment of 19,360 cases of pediatric scoliosis
: a review of the Scoliosis
Research Society Morbidity and Mortality database. Spine (Phila Pa 1976). 2011;36:1484–1491.
6. Eager M, Jahangiri F, Shimer A, et al. Intraoperative neuromonitoring
: lessons learned from 32 case events in 2095 spine cases. Evid Based Spine Care J. 2010;1:58–61.
7. Gonzalez AA, Jeyanandarajan D, Hansen C, et al. Intraoperative neurophysiological monitoring during spine surgery: a review. Neurosurg Focus. 2009;27:E6.
8. Nuwer MR, Emerson RG, Galloway G, et al. Evidence-based guideline update: intraoperative spinal monitoring with somatosensory and transcranial electrical motor evoked potentials. J Clin Neurophysiol. 2012;29:101–108.
9. Stecker MM. A review of intraoperative monitoring for spinal surgery. Surg Neurol Int. 2012;3:S174–S187.
10. DiCindio S, Theroux M, Shah S, et al. Multimodality monitoring of transcranial electric motor and somatosensory-evoked potentials during surgical correction of spinal deformity in patients with cerebral palsy
and other neuromuscular disorders. Spine (Phila Pa 1976). 2003;28:1851–1855. Discussion 1855-1856.
11. Ecker ML, Dormans JP, Schwartz DM, et al. Efficacy of spinal cord monitoring in scoliosis
surgery in patients with cerebral palsy
. J Spinal Disord. 1996;9:159–164.
12. Pastorelli F, Di Silvestre M, Plasmati R, et al. The prevention of neural complications in the surgical treatment of scoliosis
: the role of the neurophysiological intraoperative monitoring. Eur Spine J. 2011;20(suppl 1):S105–S114.
13. Mehta JS, Gibson MJ. The treatment of neuromuscular scoliosis
. Curr Othop. 2003;17:313–321.
14. Kundnani VK, Zhu L, Tak HH, et al. Multimodal intraoperative neuromonitoring
in corrective surgery for adolescent idiopathic scoliosis
: evaluation of 354 consecutive cases. Indian J Orthop. 2010;44:64–72.
15. 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.
16. Thuet ED, Winscher JC, Padberg AM, et al. Validity and reliability of intraoperative monitoring in pediatric spinal deformity surgery: a 23-year experience of 3436 surgical cases. Spine (Phila Pa 1976). 2010;35:1880–1886.
17. Colver A, Fairhurst C, Pharoah PO. Cerebral palsy
. Lancet. 2014;383:1240–1249.
18. Rethlefsen SA, Lening C, Wren TA, et al. Excessive hip flexion during gait in patients with static encephalopathy: an examination of contributing factors. J Pediatr Orthop. 2010;30:562–567.
19. Krageloh-Mann I, Cans C. Cerebral palsy
update. Brain Dev. 2009;31:537–544.
20. Gutierrez-Hernandez S, Ritzl EKHusain AM. Motor evoked potentials. A Practical Approach to Neurophysiologic Intraoperative Monitoring, 2nd ed. New York: Demos Medical Publishing; 2014:33–45.
21. Truwit CL, Barkovich AJ, Koch TK, et al. Cerebral palsy
: MR findings in 40 patients. AJNR Am J Neuroradiol. 1992;13:67–78.
22. Yin R, Reddihough DS, Ditchfield MR, et al. Magnetic resonance imaging findings in cerebral palsy
. J Paediatr Child Health. 2000;36:139–144.
23. American Clinical Neurophysiology Society. Guideline one: minimum technical requirements for performing clinical electroencephalography. Available at: https://www.acns.org/pdf/guidelines/Guideline-1.pdf
. Accessed on July 27; 2015.
24. Ferguson J, Hwang SW, Tataryn Z, et al. Neuromonitoring
changes in pediatric spinal deformity surgery: a single-institution experience. J Neurosurg Pediatr. 2014;13:247–254.
25. Keith RW, Stambough JL, Awender SH. Somatosensory cortical evoked potentials: a review of 100 cases of intraoperative spinal surgery monitoring. J Spinal Disord. 1990;3:220–226.
26. Kelleher MO, Tan G, Sarjeant R, et al. Predictive value of intraoperative neurophysiological monitoring during cervical spine surgery: a prospective analysis of 1055 consecutive patients. J Neurosurg Spine. 2008;8:215–221.
27. Khan MH, Smith PN, Balzer JR, et al. Intraoperative somatosensory evoked potential monitoring during cervical spine corpectomy surgery: experience with 508 cases. Spine (Phila Pa 1976). 2006;31:E105–E113.
28. Padberg AM, Wilson-Holden TJ, Lenke LG, et al. Somatosensory- and motor-evoked potential monitoring without a wake-up test during idiopathic scoliosis
surgery. An accepted standard of care. Spine (Phila Pa 1976). 1998;23:1392–1400.
29. Koht A, Sloan TB, Toleikis JR. Monitoring the Nervous System for Anesthesiologists and Other Health Care Professionals. New York: Springer; 2012. eds.
30. Neuloh G, Schramm J. Motor evoked potential monitoring for the surgery of brain tumours and vascular malformations. Adv Tech Stand Neurosurg. 2004;29:171–228.
31. Chen X, Sterio D, Ming X, et al. Success rate of motor evoked potentials for intraoperative neurophysiologic monitoring: effects of age, lesion location, and preoperative neurologic deficits. J Clin Neurophysiol. 2007;24:281–285.
32. Weinzierl MR, Reinacher P, Gilsbach JM, et al. Combined motor and somatosensory evoked potentials for intraoperative monitoring: intra- and postoperative data in a series of 69 operations. Neurosurg Rev. 2007;30:109–116. Discussion 116.
33. Ben-David B, Haller G, Taylor P. Anterior spinal fusion complicated by paraplegia. A case report of a false-negative somatosensory-evoked potential. Spine (Phila Pa 1976). 1987;12:536–539.
34. Ginsburg HH, Shetter AG, Raudzens PA. Postoperative paraplegia with preserved intraoperative somatosensory evoked potentials. Case report. J Neurosurg. 1985;63:296–300.
35. Lesser RP, Raudzens P, Luders H, et al. Postoperative neurological deficits may occur despite unchanged intraoperative somatosensory evoked potentials. Ann Neurol. 1986;19:22–25.
36. Hilibrand AS, Schwartz DM, Sethuraman V, et al. Comparison of transcranial electric motor and somatosensory evoked potential monitoring during cervical spine surgery. J Bone Joint Surg Am. 2004;86:1248–1253.