Abstract
Background: The vertical expandable prosthetic titanium rib (VEPTR) device is used in the treatment of thoracic insufficiency syndrome and certain types of early-onset spinal deformity. The purpose of this study was to evaluate the risk of neurologic injury during surgical procedures involving use of the VEPTR and to determine the efficacy of intraoperative spinal cord neuromonitoring.
Methods: Data were collected prospectively during a multicenter study. Surgical procedures were divided into three categories: primary device implantation, device exchange, and device lengthening. Further retrospective evaluation was undertaken in cases of neurologic injury or changes detected with neuromonitoring.
Results: There were 1736 consecutive VEPTR procedures at six centers: 327 (in 299 patients) consisted of a primary device implantation, 224 were a device exchange, and 1185 were a device lengthening. Perioperative clinical neurologic injury was noted in eight (0.5%) of the 1736 cases: these injuries were identified after five (1.5%) of the 327 procedures for primary device implantation, three (1.3%) of the 224 device exchanges, and none of the 1185 device-lengthening procedures. Of the eight cases of neurologic injury, six involved the upper extremity and two involved the lower extremity. The neurologic deficit was temporary in seven patients and permanent in one patient, who had persistent neurogenic arm and hand pain. Intraoperative neuromonitoring demonstrated changes during six (0.3%) of the 1736 procedures: five (1.5%) of the 327 procedures for primary device implantation and one (0.08%) of the 1185 device-lengthening procedures. The surgery was altered in all six cases, with resolution of the monitoring changes in five cases and persistent signal changes and a neurologic deficit (upper-extremity brachial plexopathy) in one. Two patients had false-negative results of monitoring of somatosensory evoked potentials, and one had false-negative results of monitoring of somatosensory evoked potentials and motor evoked potentials during implant surgery; two had a brachial plexopathy and one had monoplegia postoperatively, with all three recovering.
Conclusions: Neurologic injury during VEPTR surgery occurs much more frequently in the upper extremities than in the lower extremities. The rates of potential neurologic injuries (neurologic injuries plus instances of changes detected by monitoring) during primary implantation of the VEPTR (2.8%) and during exchange of the VEPTR (1.3%) justify the use of intraoperative neuromonitoring of the upper and lower extremities during those procedures. As neuromonitoring did not demonstrate any changes in children without a previous VEPTR-related monitoring change and there were no neurologic injuries during more than 1000 VEPTR-lengthening procedures, intraoperative neuromonitoring may not be necessary during those procedures in children without a history of a neurologic deficit during VEPTR surgery.
Level of Evidence: Therapeutic Level IV. See Instructions to Authors for a complete description of levels of evidence.
Implantation of a vertical expandable prosthetic titanium rib (VEPTR) has become increasingly popular for the treatment of thoracic insufficiency syndrome and certain types of early-onset spinal deformity. Traditional methods for the treatment of early-onset spinal deformity were directed primarily at the spinal deformity, not the associated thoracic insufficiency syndrome, and commonly involved early spinal arthrodesis. Early spinal arthrodesis, however, may contribute to incomplete development and growth of the thorax and further loss of lung volume and pulmonary function. VEPTR surgery was developed to treat deformities of both the spine and the chest wall while allowing growth of both. Implantation of the vertical rib implant expands the thorax and indirectly corrects the spinal deformity, leading to growth of the spine, thorax, and lungs1,2. To accommodate the growth of the chest and spine, patients with this device typically undergo multiple surgical lengthening procedures and, often, device exchanges up to skeletal maturity.
Clinical neurologic injury has been associated with VEPTR surgery for fused ribs and congenital scoliosis. In previous series, we observed three neurologic injuries (two transient brachial plexus injuries and one direct spinal cord injury occurring during dissection) in twenty-seven patients1 and two brachial plexus injuries in thirty-one patients3.
The purpose of the present study was to define the risk of neurologic injury and the role of intraoperative neuromonitoring during all types of VEPTR surgical procedures.
A prospective study of all VEPTR surgical procedures during the U.S. Food and Drug Administration (FDA) investigational stage at six centers was performed. Participating centers included Christus Santa Rosa Children's Hospital, San Antonio, Texas; Children's Hospital of Pittsburgh, Pennsylvania; Children's Hospital of Boston, Massachusetts; Childrens Hospital Los Angeles, California; Children's Hospital and Regional Medical Center, Seattle, Washington; and Primary Children's Medical Center, Salt Lake City, Utah. The prospective study was conducted while the VEPTR device was under an FDA Investigational Device Exemption, with San Antonio serving as the sole site for an FDA feasibility study with the addition of the other sites after 1996 to form a multicenter study. The total number of VEPTR procedures performed at the six centers was 1736, with 327 (in 299 patients) done for primary device implantation, 224 for device exchange, and 1185 for device lengthening. Intraoperative monitoring was performed during the primary implantation procedures at all centers, but the use of such monitoring during VEPTR lengthening varied among the centers. Monitoring of motor evoked potentials was used with increasing frequency as the study progressed and it became more generally available.
When the treating surgeon believed that a real, substantial change had been revealed by neuromonitoring intraoperatively, a "monitoring change" was recorded for that case, and when a patient had clinical neurologic sequelae either postoperatively or during a wake-up test, a "neurologic injury" was recorded. When a change detected by neuromonitoring was thought to be due to artifact, such as anesthetic issues, it was not included in the analysis. Additional retrospective evaluation of all cases with monitoring changes or neurologic injury was performed.
The data collected prospectively included the patient's age at the operation, type of surgery, diagnosis, and positive findings on preoperative magnetic resonance imaging. There were three types of surgery: primary implantation, exchange, and lengthening of the VEPTR. In most of the primary implantation procedures, expansion thoracostomy was performed with separating osteotomies between fused ribs or expansion thoracostomy between adjacent ribs, maximal expansion of the constricted thorax, and implantation of a single or multiple VEPTR devices longitudinally from rib to rib, or from rib to spine or pelvis. The exchange procedures consisted of replacement of a component (the rib sleeve and inferior rib cradle or lumbar hybrid extension) of a completely expanded implant, generally through less extensive incisions than had been used for the original procedure. The lengthening procedures involved lengthening of an implanted device with use of a mild-to-moderate expansion force applied through limited skin incisions.
The maximum scoliosis and kyphosis were measured on radiographs with use of the Cobb method. If abnormal kyphosis (of >50°) was noted, the level of the apex was also recorded. The type of neuromonitoring—i.e., of somatosensory evoked potentials or motor evoked potentials, and of the lower and/or upper extremities—was also noted. We also collected data with regard to the changes detected by neuromonitoring, the nature of neurologic injuries demonstrated by clinical examination, how the surgery was modified as a result of the neuromonitoring event, and the time until recovery if an actual clinical neurologic injury occurred. A neurologic deficit was defined as temporary if there was eventually full recovery and as permanent if the deficit was still present at the latest follow-up evaluation.
Source of Funding
This study was performed through the VEPTR study group, which was funded by Synthes Spine.
Clinical signs of neurologic injury developed in the perioperative period in eight (2.7%) of the 299 patients in this study (Table I). The average age of these eight patients at the time of the procedure was forty-nine months (range, seven to 120 months). Seven had congenital scoliosis and fused ribs, and one had infantile idiopathic scoliosis. The rate of neurologic injury was 2.7% per patient (eight of 299 patients) and 0.5% per VEPTR procedure (eight of 1736 procedures). The rate of temporary neurologic injury was 2.3% (seven of 299 patients), and the rate of permanent neurologic injury was 0.3% (one of 299 patients). A neurologic injury developed after five (1.5%) of the 327 procedures for primary device implantation, three (1.3%) of the 224 device exchanges, and none of the 1185 device-lengthening procedures. The neurologic injury involved the upper extremity in six patients and the lower extremity in two. There were five brachial plexopathies (with four involving both motor and sensory deficits). One case resolved by six months; one, by ten months; and two more, by one year. The remaining patient had resolution of all motor deficits but had persistent neurogenic arm and hand pain four years after the surgery. One child had hypersensitivity in both hands for six months. One child had dense motor deficits in both lower extremities noted during the wake-up test. Distraction was released and the lower implants were removed, resulting in a complete return of motor function postoperatively. One child had monoplegia of the left lower extremity noted during the wake-up test. In this patient, the dura and spinal canal had been inadvertently violated during the thoracostomies, resulting in incomplete injury of the spinal cord. Recovery was almost full by three years. Of the eight cases of neurologic injury in the perioperative period, five occurred in patients with a kyphosis of >50°. One of these patients had lower-extremity involvement. All eight cases of clinical neurologic injury occurred in the latter phase (after 1997) of the FDA investigational stage (1989 through 2004).
Intraoperative neuromonitoring relevant to the site of injury had been utilized in four of the eight patients who experienced a perioperative clinical neurologic injury. In one patient with brachial plexopathy, only the lower extremities had been monitored intraoperatively; in another patient, with dense postoperative lower-extremity neurologic deficits, only the upper extremities had been monitored. Of the four patients who had had monitoring of both the upper and the lower extremities, one had a monitoring change noted intraoperatively, during a primary device implantation. In this case, upper-extremity somatosensory evoked potentials were markedly decreased intraoperatively. In response, soft tissues were released, the scapula was mobilized, fused ribs were partially resected, and distraction was partially released in order to minimize tension on the brachial plexus. The somatosensory evoked potentials remained diminished at the end of the operation, and a motor and sensory brachial plexopathy developed postoperatively; it resolved after ten months. In three cases, intraoperative monitoring had failed to detect a clinical neurologic injury that was evident postoperatively. In one of these patients, who had congenital scoliosis and fused ribs, upper and lower-extremity somatosensory evoked potentials were normal but brachial plexopathy occurred. In another patient, who had congenital scoliosis and fused ribs, upper and lower-extremity somatosensory evoked potentials and motor evoked potentials were normal but brachial plexopathy was noted postoperatively. In the third case, also with congenital scoliosis and fused ribs, upper and lower-extremity somatosensory evoked potentials were normal but unilateral lower-extremity monoplegia occurred.
Monitoring changes developed intraoperatively in five (1.7%) of the 299 patients in this study, with one patient having monitoring changes during two operations (Table II). The average age of these patients at the time of the procedure was 37.5 months (range, twenty-four to forty-eight months). Two patients had infantile idiopathic scoliosis, one had kyphoscoliosis and sacral agenesis, one had myelomeningocele and fused ribs, and one had congenital scoliosis with fused ribs and VATER syndrome. The rate of neuromonitoring changes was 1.7% per patient (five of 299 patients) and 0.3% per VEPTR procedure (six of 1736 procedures). The monitoring changes were noted during five (1.5%) of the 327 procedures for primary device implantation, none of the 224 device exchanges, and one (0.08%) of the device-lengthening procedures. Three of the six monitoring changes were in the upper extremity, and the other three were in the lower extremity. Three of the six monitoring changes occurred in patients with abnormal kyphosis, and two of the three had these changes in the lower extremity. All six cases of monitoring changes occurred in the latter phase (after 1997) of the FDA investigational stage (1989 through 2004).
In five of the six occurrences of monitoring changes, the patient had a full return to normal monitoring results intraoperatively with no postoperative neurologic deficits. The one monitoring change that was noted during a lengthening was a transient loss of somatosensory evoked potential signals in one lower extremity following a device lengthening of 1 cm; the signals returned to normal when the amount of lengthening was decreased to 5 mm, and there were no clinical sequelae. This child had had previous changes in the somatosensory evoked potentials during the primary implantation. In four of the patients with monitoring changes, the changes resolved with reduction of the distraction force on the VEPTR while the VEPTR was left in place. In one child, after the monitoring changes were observed, the VEPTR implant was placed on a different rib and the signal changes resolved. One child with intraoperative monitoring changes had persistently abnormal somatosensory evoked potentials despite corrective maneuvers (detailed above), and a motor and sensory brachial plexopathy subsequently developed.
Implantation of a VEPTR device with or without an expansion thoracostomy can control both spine and chest wall deformities during growth and allow continued growth of the thorax, spine, and lungs1-3. One of us (R.M.C. Jr.) and colleagues previously reported the results in twenty-seven patients who had undergone VEPTR surgery for the treatment of congenital scoliosis associated with fused ribs1. The mean age at the time of surgery was 3.2 years (range, 0.6 to 12.5 years), and the mean duration of follow-up was 5.7 years. The scoliosis decreased from a mean of 74° preoperatively to a mean of 49° at the time of the last follow-up. The height of the thoracic spine increased by a mean of 0.71 cm/yr. At the time of follow-up, the vital capacity was best for the patients who had undergone surgery at the age of two years or younger and worst for those older than two years at the time of surgery and those with a history of spine surgery. Overall, fifty-two complications were reported in twenty-two patients. The most common complication was asymptomatic proximal migration of the device through ribs, which occurred in seven patients. Three neurologic complications were reported. Upper-extremity brachial plexopathy developed postoperatively in two patients. The plexopathy was thought to have resulted from placement of the VEPTR device too far proximally, causing the rib to impinge on the brachial plexus. The plexopathy resolved in time after repositioning of the implant. There was one incomplete spinal cord injury secondary to inadvertent violation of the dura and spinal canal during the most medial extent of an osteotomy between ribs. The patient had almost fully recovered at 2.8 years postoperatively.
In Europe, Hell et al. reported the results in fifteen patients who had undergone VEPTR surgery for treatment of thoracic insufficiency syndrome (nine patients) or severe neuromuscular scoliosis (six patients)2. Cosmetic, functional, and radiographic improvements were noted, with a reduction in the Cobb angle from an average of 76° (range, 40° to 110°) to an average of 55° (range, 30° to 67°). There were three complications (skin breakdown, hook displacement, and rib fracture). No neurologic complications were reported. One of us (J.E.) and colleagues prospectively evaluated thirty-one patients who had undergone VEPTR surgery for the treatment of thoracic insufficiency syndrome associated with congenital scoliosis and fused ribs3. Spinal deformity was controlled, while allowing for continued growth of the thoracic spine and increased lung volume, in thirty of the thirty-one patients. Complications included device migration; infection; and two cases of brachial plexus palsy, one of which was profound. The profound brachial plexus palsy occurred early in the experience with use of the VEPTR, before the routine use of upper-extremity monitoring. Immediate postoperative paralysis and decreased perfusion of the ipsilateral upper extremity were attributed to acute thoracic outlet syndrome and resolved after device expansion was decreased.
The risk of neurologic injury during posterior spinal fusion and instrumentation for the treatment of adolescent idiopathic scoliosis was reported by the Scoliosis Research Society Morbidity and Mortality Committee to be 0.32% (fourteen of 4369 cases)4. The risk of neurologic injury during correction of congenital scoliosis is assumed to be higher, but to our knowledge there are no statistical data. The importance of intraoperative neuromonitoring for the detection and possible prevention of neurologic injury during spinal surgical procedures is well recognized5-8. To our knowledge, no neurologic complications have been reported in association with the growing-rod technique9.
The prevalence of changes detected by spinal cord monitoring or of subsequent clinical neurologic injury following VEPTR surgery has not been previously reported in a large series, to our knowledge. In the current series, the rate of spinal cord monitoring changes without subsequent clinical neurologic sequelae was 1.2% (four of 327) during primary VEPTR implantation, 0% (of 224) during device exchange, and 0.08% (one of 1185) during device lengthening. The rate of clinical neurologic injury, mostly transient brachial plexopathies, was 1.5% (five of 327) after primary VEPTR implantation, 1.3% (three of 224) after device exchange, and 0% (of 1185) after device lengthening.
A review of the literature suggests that neurologic injuries during surgery to address spine deformities primarily involve the lower extremities. However, upper-extremity neurologic injury, which developed in six patients in our series, appears to be more common than lower-extremity neurologic injury, which occurred in two patients, after VEPTR surgery. Three of the six cases of intraoperative neuromonitoring changes in this series involved the upper extremity. The risk of brachial plexopathy is rare but present in traditional spinal surgery10. Schwartz et al. demonstrated the importance of intraoperative monitoring of the upper extremities in patients undergoing posterior spinal fusion in the prone position for the treatment of scoliosis11. On the basis of somatosensory evoked potential signal characteristics, brachial plexopathy was identified intraoperatively in 3.6% (eighteen) of 500 limbs. The signal characteristics, specifically amplitude, improved immediately in response to repositioning, and no patient demonstrated signs of brachial plexopathy after the surgery. Our results with the VEPTR are quite different from those associated with traditional spinal surgery in that a majority of our cases of postoperative clinical neurologic sequelae involved the upper extremity. Why this is the case is worthy of consideration.
We hypothesize several possible mechanisms for acute or delayed brachial plexopathy: (1) direct impingement by the device in the region of the brachial plexus, (2) acute thoracic outlet syndrome from elevation of the upper hemithorax and distal traction on the clavicle, (3) traction on the scapula and the upper extremity, and (4) local postoperative edema. We believe that the brachial plexus draping over the first rib may be subject to injury when a VEPTR implant is placed nearby or when the upper ribs are displaced cephalad by expansion thoracostomy. Brachial plexopathy may occur as a result of the first rib and the upper part of the thorax being elevated by the VEPTR device, resulting in direct pressure on the brachial plexus or entrapment of the brachial plexus between the upper part of the thorax and the clavicle. After expansion thoracoplasty, the scapula and the associated musculocutaneous flap are typically drawn caudad to cover the newly expanded hemithorax. Caudad displacement of the scapula and the attached shoulder and clavicle may contribute to entrapment of the brachial plexus, essentially creating an acute thoracic outlet syndrome. To avoid this problem, we recommend that the superior cradle of the VEPTR not be placed on the first rib or laterally on the second rib. In cases in which the proximal few ribs are fused and less mobile, placement of the rib cradle on these fused ribs seems to be better tolerated, and often necessary as individual second and third ribs may not be available for rib cradle placement. Most importantly, caution is necessary to avoid brachial plexus compression when the proximal ribs are moved cephalad during VEPTR distraction.
Neuromonitoring changes have occurred in the upper extremity while the scapula was brought inferiorly to close the wound of the acutely expanded thorax. The added tension on the wound closure may lead to constriction of the tissue over the first rib and subsequent compression of the brachial plexus. In this case, a decrease in the amount of distraction created by the VEPTR may be necessary to allow safe wound closure.
We observed that simple retraction of the scapula prior to VEPTR implantation resulted in a transient loss of somatosensory evoked potentials and motor evoked potentials to the arm in two patients who were not included in the series reported here. Children with fused ribs and a hypoplastic hemithorax often have marked fibrosis around the scapulae and ribs. We hypothesize that retraction of the scapula from the rib cage by lifting its caudad and medial border may also cause the superior part of the scapula to pinch the brachial plexus. Avoiding overzealous scapular retraction and immediate cessation of scapular retraction at the first signs of neurologic changes may prevent a clinically relevant injury.
It is possible to estimate the efficacy of neuromonitoring during primary implantation and device-exchange procedures with use of these data. On the basis of 327 primary implantations and 224 device exchanges with elimination of four cases of neurologic injury in which there had been no intraoperative monitoring at all (Cases 3 and 4; Table I) or in which the affected extremities had not been monitored intraoperatively (Cases 5 and 7; Table I), there were four false-positive, one true-positive, three false-negative, and 539 true-negative results of neuromonitoring. If neuromonitoring is considered to be a test for postoperative neurologic injury after device implantation, the positive predictive value is 20%, the negative predictive value is 99%, the sensitivity is 25%, and the specificity is 99%. The poor positive predictive value of intraoperative neuromonitoring may be explained in part by the common use of somatosensory evoked potentials alone, without monitoring of motor evoked potentials, in this series, which was performed before intraoperative monitoring of motor evoked potentials was generally available. However, one patient (Case 6) who had had monitoring of both somatosensory and motor evoked potentials experienced a brachial plexopathy. On the basis of these data, we recommend postoperative assessment of the neurologic status of both the upper and the lower extremities. An intraoperative wake-up test may also be justified. In the many cases in which the postoperative course following primary implantation includes mechanical ventilation and sedation, continued postoperative clinical assessment of the neurologic function of the extremities, although inconvenient, may detect neurologic deficits not noted on the basis of intraoperative neuromonitoring. It must also be cautioned that, even with negative findings on monitoring of somatosensory evoked potentials and on performance of a wake-up test, a lower-extremity neurologic deficit may still occur. After this study was completed, an opening-wedge thoracostomy stabilized with a VEPTR in a patient with angular congenital kyphosis as well as moderate scoliosis was complicated postoperatively by paraplegia, and there was no recovery despite implant removal. Both somatosensory evoked potential monitoring and the wake-up test revealed negative findings during the surgery, and the deficit was noted days later, after respirator sedation had been decreased.
In summary, the rates of potential neurologic injury (instances of neurologic injury and instances of neuromonitoring changes) during primary device implantation (2.8%; nine of 327) and device exchange (1.3%; three of 224) justify the use of intraoperative monitoring during these surgical procedures. Because the risk of neurologic injury during VEPTR surgery appears to be higher in the upper extremities than in the lower extremities, both the upper and the lower extremities should be monitored. The poor positive predictive value of intraoperative neuromonitoring also suggests the need for a careful postoperative clinical neurologic assessment following primary implantation and device-exchange procedures. The risk of neurologic injury during device-lengthening procedures seems to be very low on the basis of our series, in which it was 0% of 1185, and the single monitoring change during device lengthening occurred in a child with a history of a monitoring change during the primary implantation. Therefore, neuromonitoring during all subsequent VEPTR procedures is recommended for patients who had had previous monitoring changes during the primary implantation of the VEPTR. 
Campbell RM Jr, Smith MD, Mayes TC, Mangos JA, Willey-Courand DB, Kose N, Pinero RF, Alder ME, Duong HL, Surber JL. The effect of opening wedge thoracostomy on thoracic insufficiency syndrome associated with fused ribs and congenital scoliosis. J Bone Joint Surg Am.2004;86:1659-74.861659
2004
[PubMed]
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;14:287-93.14287
2005
[CrossRef]
Emans JB, Caubet JF, Ordonez CL, Lee EY, Ciarlo M. The treatment of spine and chest wall deformities with fused ribs by expansion thoracostomy and insertion of vertical expandable prosthetic titanium rib: growth of thoracic spine and improvement of lung volumes. Spine.2005;30(17 Suppl):S58-68.30S58
2005
[CrossRef]
Coe JD, Arlet V, Donaldson W, Berven S, Hanson DS, Mudiyam R, Perra JH, Shaffrey CI. Complications in spinal fusion for adolescent idiopathic scoliosis in the new millennium. A report of the Scoliosis Research Society Morbidity and Mortality Committee. Spine.2006;31:345-9.31345
2006
[CrossRef]
Forbes HJ, Allen PW, Waller CS, Jones SJ, Edgar MA, Webb PJ, Ransford AO. Spinal cord monitoring in scoliosis surgery. Experience with 1168 cases. J Bone Joint Surg Br.1991;73:487-91.73487
1991
Noordeen MH, Lee J, Gibbons CE, Taylor BA, Bentley G. Spinal cord monitoring in operations for neuromuscular scoliosis. J Bone Joint Surg Br.1997;79:53-7.7953
1997
[CrossRef]
Nuwer MR, Dawson EG, Carlson LG, Kanim LE, Sherman JE. Somatosensory evoked potential spinal cord monitoring reduces neurologic deficits after scoliosis surgery: results of a large multicenter survey. Electroencephalogr Clin Neurophysiol.1995;96:6-11.966
1995
[CrossRef]
Pelosi L, Lamb J, Grevitt M, Mehdian SM, Webb JK, Blumhardt LD. Combined monitoring of motor and somatosensory evoked potentials in orthopaedic spinal surgery. Clin Neurophysiol.2002;113:1082-91.1131082
2002
[CrossRef]
Akbarnia BA, Marks DS, Boachie-Adjei O, Thompson AG, Asher MA. Dual growing rod technique for the treatment of progressive early-onset scoliosis: a multicenter study. Spine.2005;30(17 Suppl):S46-57.30S46
2005
[CrossRef]
Labrom RD, Hoskins M, Reilly CW, Tredwell SJ, Wong PK. Clinical usefulness of somatosensory evoked potentials for detection of brachial plexopathy secondary to malpositioning in scoliosis surgery. Spine.2005;30:2089-93.302089
2005
[CrossRef]
Schwartz DM, Drummond DS, Hahn M, Ecker ML, Dormans JP. Prevention of positional brachial plexopathy during surgical correction of scoliosis. J Spinal Disord.2000;13:178-82.13178
2000
[CrossRef]