The use of the vertical expandable prosthetic titanium rib (VEPTR) device has become a common surgical treatment for the correction of spine and chest wall deformities in patients diagnosed with thoracic insufficiency syndrome and severe early-onset scoliosis (EOS). Unlike prior spinal fusion techniques, the VEPTR device is lengthened over multiple surgical procedures to maintain correction of the deformity, while facilitating continued growth and development of the chest, spine, and lungs.1,2
Neurological injury is a risk inherent to VEPTR procedures, secondary to the mechanical and vascular stresses on the spinal cord and brachial plexus.1,3 Intraoperative neurological monitoring (IONM) is a technique utilized to alert surgeons of impending neurological deficits, at which point intervention can reduce the risk of permanent neurological injury. The validity of this technique has been well described in spinal fusion procedures for the treatment of spinal deformity; however, the utility of IONM during VEPTR procedures remains debated.4,5
The purpose of this study is to evaluate the utility of IONM during VEPTR procedures, and to describe circumstances in which IONM may not be indicated. We hypothesize that in the absence of significant intraoperative neurological signal changes during initial VEPTR implantation, IONM will not identify new neurological injuries with routine expansion procedures.
A retrospective evaluation of a prospectively collected multicenter database was conducted after IRB approval. Data were collected from a total of 17 institutions on all consecutive patients surgically treated with primary VEPTR device implantation from 2005 to 2011. All included patients had minimum 2 years of continued clinical follow-up from their index implant procedure. Because of concurrent studies within the multicenter database, 4 patients who met study criteria were included in this study that had previously been reported regarding the rate of neurological injury during VEPTR procedures.6 Patients who had received prior growing rods or other invasive spine-based growth modulation treatments were excluded from this study.
There were a total of 2355 consecutive VEPTR procedures (in 352 patients) consisting of 299 implant, 377 revision, 1587 expansion, and 92 removal procedures. In total, 620 VEPTR procedures had reported the use of IONM in the multicenter database; however, only 539 of those procedures had IONM records available for review (87%). A total 256 patients had IONM performed during the course of their VEPTR treatment, and of those 218 had at least 1 IONM record available for review (85%).
The primary outcome variable was postoperative neurological injury. A neurological injury was defined as a new loss of sensory or motor function identified during postoperative neurological examination. Neurological deficits were recorded prospectively in the multicenter database. A secondary outcome variable of interest was the presence of significant IONM signal change during VEPTR surgery. A significant change in somatosensory-evoked potential (SSEP) monitoring was defined as, a >50% decrease in amplitude and/or a >10% increase in latency of the subcortical and/or cortical waveforms, whereas a significant change in motor-evoked potential (MEP) monitoring was defined as a complete loss of waveforms. All available IONM reports were collected and reviewed for surgeries where it was utilized. All postoperative clinic notes in cases with IONM alerts were also reviewed for documentation of neurological injury.
Demographic, surgical, and postoperative data were obtained from the multicenter database. General baseline data before the index procedure was collected including the patient’s age, sex, classification of early-onset scoliosis (C-EOS), sagittal and coronal deformity measures, and abnormalities within the spinal canal identified by computed tomography or magnetic resonance imaging (MRI).7 Relevant operative data collected included VEPTR procedure type, use of IONM, and surgical responses to changes in IONM signal or neurological deficit. There were 4 types of surgical procedures: (1) implant, defined as the index procedure at which the VEPTR construct was placed, (2) revision, a procedure to replace or compensate for a failed implant or undesirable sequelae of previous surgery, (3) expansion, routine lengthening of the device to maintain correction of the deformity as the patient grows, and (4) removal, the explantation of VEPTR implants when the patient has attained skeletal maturity and is ready to proceed to definitive spinal fusion, or explantation secondary to intolerable sequelae. Postoperative data collected included details of neurological injury, as well as time to full recovery.
Descriptive statistics were used to review instances of neurological injury or IONM alert. Calculations of sensitivity, specificity, positive predictive value, and negative predictive value were used to evaluate IONM as a diagnostic test for neurological injury.
New neurological injury occurred in 3/2355 procedures (0.1%), or 3/352 patients (0.9%) (subsequently referred to as patients a, b, and c). All 3 injuries were related to implant procedures. Only one of these procedures had an IONM alert. All 3 patients were diagnosed with congenital scoliosis and fused ribs, and had transient neurological deficits (Table 1).
Patient (a) was diagnosed with congenital scoliosis with fused ribs secondary to Klippel-Feil Syndrome, and had preoperative coronal and sagittal deformities of 80 and 35 degrees, respectively. MEP were attempted, but not obtainable for this patient; SSEP were obtained and monitored for the surgical procedure. The patient had an IONM alert, which occurred during device distraction—and was interpreted as a brachial plexus compression. The device was removed from its attachment sites at the second and seventh ribs with subsequent recovery of the SSEP waveforms, recontoured, and reapplied with less distraction. At completion of the case the SSEP waveforms were at their baseline values. Throughout the postoperative follow-up the patient did not have any sensory changes. The patient had intact motor function on postoperative evaluation, but subsequently developed partial brachial plexus palsy with upper extremity weakness during the second postoperative week. The treating team presumed this neurological change to be due to surgical edema. The patient underwent physical rehabilitation and had full recovery at 124 days after presentation of the injury.
Patient (b) had congenital scoliosis with fused ribs and thoracic insufficiency syndrome secondary to VATER syndrome, and had preoperative coronal and sagittal deformities of 32 and 33 degrees, respectively. This patient underwent VEPTR device implantation without IONM alert and had a normal initial postoperative evaluation, but on the sixth postoperative day presented with upper extremity sensory and motor deficits. The patient completed physical rehabilitation and had full recovery at 30 days after the reported injury.
Patient (c) had congenital scoliosis with fused ribs, a preoperative coronal deformity of 62 and a sagittal deformity of 50. The patient was without IONM alert during implantation and had upper extremity motor weakness on postoperative examination that was reported to be recovering. One day after the procedure the rate of recovery was determined to be inadequate; the treating surgeon attributed the neurological injury to overdistraction of the VEPTR device leading to elevation of the first rib and compression of the brachial plexus. The patient was subsequently taken back to the operating room to have the device removed. Two weeks later the patient returned to the operating room for the third time, and the device was successfully reimplanted. The patient had an uneventful postoperative period with full recovery 17 days after the injury.
Intraoperative Neuromonitoring Alerts
IONM alerts occurred in 9/539 procedures (1.7%). Four of these alerts consisted of significant changes in SSEP waveforms, 3 consisted of complete loss of MEP waveforms and 2 had both significant changes and loss of SSEP and MEP waveforms. The alerts occurred in 7/352 patients (2%), with 1 patient having alerts in 3 procedures. Three of these patients had diagnoses of congenital scoliosis (2 with rib fusions), 2 patients had syndrome-related scoliosis, 1 patient had neuromuscular scoliosis, and 1 had infantile idiopathic scoliosis. The mean preoperative coronal and sagittal deformities were 77 (range, 32 to 103) degrees and 62 (range, 29 to 85) degrees, respectively. The rate of IONM alerts were 1.6% for implant procedures (3/192; all 3 had changes in SSEP), 5.2% for revision procedures (3/58; 2 with loss of MEP and 1 with SSEP change and MEP loss), and 1.2% for expansion procedures (3/258; 1 with change in SSEP, 1 with MEP loss, and 1 with both SSEP change and MEP loss). Two of the IONM alerts were in the upper extremity (SSEP changes), 4 in the lower extremity (2 with SSEP changes, 2 MEP loss), and 3 were in both the upper and lower extremities (1 with MEP loss, 2 with both change in SSEP and MEP loss) (Table 2).
Two cases of SSEP alerts were determined not to represent any evident complication and no intervention was pursued by the surgeon. A significant reduction in 1 of these patient’s SSEP signals occurred during closure—well after any manipulation or correction of the spine—and resolved spontaneously (patient d), whereas another patient had significantly reduced SSEP signals that remained stable throughout the procedure (patient e).
One patient had SSEP and MEP alerts before incision (patient f). The patient was repositioned; however, a full recovery of SSEP and MEP waveforms did not occur and the VEPTR procedure was canceled. An MRI was ordered and revealed stenosis of the foramen magnum, the patient subsequently underwent suboccipital decompression. An uneventful VEPTR procedure was later performed. However, the patient’s stenosis reoccurred and resulted in SSEP and MEP alerts and cancellation of 2 subsequent VEPTR procedures.
One patient with no prior history of IONM alerts presented for excision and drainage of VEPTR hardware and then underwent VEPTR expansion (patient g). This patient developed decreased SSEP waveforms during tightening of the VEPTR device and closure. The patient’s blood pressures were noted to be low at this time and fluids were administered. The patient’s blood pressure increased and the SSEP waveforms returned to baseline.
Three patients had IONM alerts after surgical maneuvers. Two of these patients had loss of MEP waveforms after placement of the VEPTR device—in both cases the waveforms returned after surgeons decreased the distraction of the rod (patient h) or repositioned the patient (patient i). The third patient had a significant reduction of SSEP waveforms during distraction of the rod and presumed compression of the brachial plexus; the device was subsequently removed until waveforms returned, at which point the rod was recontoured and applied with less distraction. This was the only patient to present with neurological injury following an IONM alert (patient a).
Although the VEPTR device has become an established treatment for the management of severe chest wall and spinal deformities in the growing pediatric patient, there is limited evidence describing the utility of IONM during the course of VEPTR treatment. More specifically, the utility of IONM during expansion procedures in patients with no prior history of neurological injury or IONM alert. This study represents the largest known reported series of VEPTR procedures evaluating the utility of IONM. Our data demonstrate a low incidence of neurological injury during VEPTR surgery. All neurological injuries occurred during primary device implantation. There were a total of 1587 expansion procedures without a single incident of neurological injury.
Only 1 expansion procedure had an IONM alert without a prior history of IONM signal changes. This occurred during closure in the setting of an infection and hypotension. After administering fluids intraoperatively, the patient’s pressures and IONM signals returned to baseline. It is well established that hemodynamic changes, such as hypotension in the setting of infection, can contribute to changes in IONM signals. These signal changes can be readily reversed by an increase in blood pressure.4,8 Although an IONM alert did occur, it is not clear whether it provided value in prompting the intervention of the surgical team, as a drop in blood pressure was also evident.
Skaggs and colleagues first reported a series of 1736 VEPTR procedures in 299 patients describing a 1.5% and 1.3% rate of neurological injury during device implantation and device exchange, respectively. They also described 1185 expansion procedures with no instances of neurological injury.9 Gauthier and colleagues subsequently reviewed a series of 3347 rib-based distraction procedures in 524 patients and reported a 1.8% rate of neurological injury during device implant procedures and 0.3% rate of neurological injury during revision procedures; there were no neurological injuries associated with 2193 routine lengthening procedures.6 In a related study of growing rod procedures, Sankar and colleagues reported 782 procedures in 282 patients with only 1 event of neurological injury during an implant exchange, a neurological injury rate of 0.1%, and in 362 lengthening procedures there were no instances of neurological injury.10 In these 3 studies, neurological injury only occurred during device implantation, revision, or exchange procedures. There were no instances of neurological injury during a total of 3740 routine lengthening procedures. Furthermore, both Skaggs and colleagues and Sankar and colleagues had described that IONM had not demonstrated any alerts in patients during expansion or lengthening procedures without a previous VEPTR or growing rod-related IONM alert.
Neurological injury during rib-based distraction for EOS is most commonly associated with injury to the brachial plexus and typically occurs in patients with congenital abnormalities of the spine and chest wall.6,9 All of the reported neurological injuries in our cohort were upper extremity injuries, presumably insults to the brachial plexus, and all neurological injuries occurred in patients with congenital scoliosis and fused ribs. The mechanisms of brachial plexus injury during distraction-based implants have been previously described and include: injury as a result of the first rib’s superior displacement due to rib-anchored instrumentation, direct injury to the brachial plexus with retraction of the scapula and injury to the brachial plexus during Sprengel deformity reconstruction as the scapula is moved inferiorly.3
Neurological injuries are rare in the treatment of thoracic insufficiency syndrome and EOS with growth sparing surgery. In our study, there were 3 instances of neurological injury, 2 of which, IONM failed to alert the surgeons of impending neurological injury. Although the procedure where the IONM alert occurred may have protected the patient from more serious injury, it did not protect the patient from a neurological injury, itself. When considering IONM as a test for postoperative neurological injury after device implantation, our data indicate a sensitivity of 33%, a specificity of 99%, a positive predictive value of 33%, and a negative predictive value of 99%.
There were 3 cases of IONM alerts during 258 cases of monitored expansion procedures with no subsequent neurological injuries. As discussed above, an IONM alert occurred in the setting of a patient with hypotension and a surgical-site infection who underwent incision, drainage, and subsequent expansion. The other case of IONM alert during an expansion procedure occurred in a patient previously diagnosed with a chromosome 1 deletion and congenital scoliosis. This patient had 2 separate instances of expansion monitoring alerts after a prior history of IONM alert during a revision procedure, the etiology of which had been confirmed by subsequent MRIs as foramen magnum and upper cervical stenosis.
Congenital anomalies of the craniocervical junction and upper cervical spine can have devastating outcomes if not properly recognized and treated. Although rare, Hosalkar et al11 reported 68 patients with such disorders referred to their tertiary care hospital over a 15-year period. Several studies have cited an increased risk of upper cervical spinal anomalies and potential neurological injury in subsets of syndromic patients and patients with genetic disorders. For example, Jeune syndrome patients have been reported to have up to a 60% incidence of upper cervical stenosis.12 Another study reported that 43% of Klippel-Feil syndrome patients had upper cervical pathology that resulted in neurological injury that occurred spontaneously or due to minor trauma.13 Patients found to have an upper cervical spine or craniocervical anomaly based on underlying diagnosis, clinical symptoms, or imaging are recommended to undergo surgical stabilization of the defect before being treated for other spinal deformities.12
Limitations of this study include an incomplete evaluation of all IONM records in this study population. The prospectively collected multicenter database had indicated 620 VEPTR procedures with IONM; however, IONM records could only be retrieved for a total of 539 procedures. Although the multicenter database had not indicated any episodes of IONM alert or neurological injury in the procedures with unavailable IONM reports, the official neurological reports were not available for verification. In addition, major neurological injuries, including spinal cord injury, are a rare event during surgery for spinal deformity and have been reported to range from 0.09% to 0.68%.14,15 Despite the large number of patients and surgeries in our cohort, it may not be of sufficient size to document such a rare event.
In this study, the rate of neurological injury during VEPTR procedures was low at 0.1%. All neurological injuries occurred during primary device implantation, a rate of 1%, justifying the use of IONM in these procedures. However, careful clinical examination of these patients postoperatively is warranted given the poor sensitivity and positive predictive value of IONM during implant procedures. There were a total of 1587 expansion procedures without a single neurological injury, and no child without prior IONM alert or neurological injury developed a new neurological injury during routine expansion procedures.
1. Campbell RM, Smith MD, Mayes TC, et al. The effect of opening wedge thoracostomy on thoracic insufficiency syndrome associated with fused ribs and congenital scoliosis. J Bone Joint Surg Am. 2004;86-A:1659–1674.
2. Hell AK, Campbell RM, Hefti F. The vertical expandable prosthetic titanium rib implant for the treatment of thoracic insufficiency syndrome associated with congenital and neuromuscular scoliosis in young children. J Pediatr Orthop B. 2005;14:287–293.
3. Joiner ER, Andras LM, Skaggs DL. Mechanisms and risk factors of brachial plexus injury in the treatment of early-onset scoliosis with distraction-based growing implants. J Bone Joint Surg Am. 2013;95:e161.
4. Devlin VJ, Schwartz DM. Intraoperative neurophysiologic monitoring during spinal surgery. J Am Acad Orthop Surg. 2007;15:549–560.
5. Schwartz DM, Auerbach JD, Dormans JP, et al. Neurophysiological detection of impending spinal cord injury during scoliosis surgery. J Bone Joint Surg Am. 2007;89:2440–2449.
6. Gauthier LE, Mandourah Y, Soroceanu A, et al. Perioperative neurologic injury associated with rib-based distraction surgery. Spine Deformity. 2014;2:481–488.
7. Williams BA, Matsumoto H, McCalla DJ, et al. Development and initial validation of the classification of early-onset scoliosis (C-EOS). J Bone Joint Surg Am. 2014;96:1359–1367.
8. Seyal M, Mull B. Mechanisms of signal change during intraoperative somatosensory evoked potential monitoring of the spinal cord. J Clin Neurophysiol. 2002;19:409–415.
9. Skaggs DL, Choi PD, Rice C, et al. Efficacy of intraoperative neurologic monitoring in surgery involving a vertical expandable prosthetic titanium rib for early-onset spinal deformity. J Bone Joint Surg Am. 2009;91:1657–1663.
10. Sankar WN, Skaggs DL, Emans JB, et al. Neurologic risk in growing rod spine surgery in early onset scoliosis: is neuromonitoring necessary for all cases? Spine (Phila Pa 1976). 2009;34:1952–1955.
11. Hosalkar HS, Sankar WN, Wills BP, et al. Congenital osseous anomalies of the upper cervical spine. J Bone Joint Surg Am. 2008;90:337–348.
12. Campbell RM. Spine deformities in rare congenital syndromes: clinical issues. Spine (Phila Pa 1976). 2009;34:1815–1827.
13. Nagib MG, Maxwell RE, Chou SN. Identification and management of high-risk patients with Klippel-Feil syndrome. J Neurosurg. 1984;61:523–530.
14. Raynor BL, Bright JD, Lenke LG, et al. Significant change or loss of intraoperative monitoring data: a 25-year experience in 12,375 spinal surgeries. Spine (Phila Pa 1976). 2013;38:E101–E108.
15. Skovrlj B, Cho SK, Caridi JM, et al. Association between surgeon experience and complication rates in adult scoliosis surgery: a review of 5117 cases from the Scoliosis Research Society database 2004-2007. Spine (Phila Pa 1976). 2015;40:1200–1205.