Sternal Split Approach to the Cervicothoracic Junctionin Children
Kishore Mulpuri, MBBS, MS(Ortho),* Jacques G. LeBlanc, MD, FRCSC,‡Christopher W. Reilly, MD, FRCSC,§ Kenneth J. Poskitt, MD, FRCPC,† Rachel L. Choit, BSc,*Vic Sahajpal, MD, FRCSC,§ and Stephen J. Tredwell, MD, FRCSC§
Study Design. We present a descriptive case seriesoutlining the surgical technique and outcome in six pa-tients managed with a combined anterior neck and ster-nal splitting approach.
Objectives. To describe a surgical approach used inthe management of severe cervicothoracic kyphosisand/or scoliosis in pediatric patients.
Summary of Background Data. There are few reportsin the literature that address the problem of accessingmultileveled spinal deformities around the cervicotho-racic junction requiring stabilization in the pediatric pop-ulation.
Methods. A detailed chart and radiographic reviewwas completed of six consecutive patients managed atour center with a combined anterior neck and sternalsplitting approach. The indications, surgical technique,and outcome are reviewed for each case. This techniquewas employed in 6 pediatric patients, aged 3–15 years, atthe authors’ institution. Diagnoses included Klippel-FeilSyndrome (2 patients), Proteus Syndrome, Larsen Syn-drome, and neurofibromatosis type I (2 patients). All pa-tients had severe cervicothoracic kyphosis requiring sur-gical instrumentation. This technique allowed surgicalaccess from C5-T6.
Results. This approach was invaluable in gaining ac-cess to the cervicothoracic junction to address complexspinal deformities in pediatric patients. In one patient, aseparate thoracotomy was performed to access the lowerthoracic spine. The only significant complication relatedto the approach was recurrent laryngeal nerve palsy ex-perienced by one patient. This approach allowed stabili-zation of severe scoliotic and/or kyphotic deformities toimpede curve progression.
Conclusions. This approach was invaluable in gainingmultileveled access to the cervicothoracic junction to ad-dress complex spinal deformities in pediatric patients.
Addressing complex spinal deformities around the cervi-cothoracic junction is technically challenging because thecervicothoracic vertebral bodies are relatively inaccessi-ble due to the anatomic structures that impede the ante-rior approach.1–3 Posterior approaches alone are insuf-ficient in severe cervicothoracic kyphosis. There havebeen several different approaches described in the litera-ture to resolve this difficulty.4–9 Most reports that ad-dress the anterior approach to the cervicothoracic junc-tion describe gaining access to short segments of thespine. This may be adequate to manage primary or met-astatic tumors affecting the upper thoracic vertebrae, thesequelae of Pott’s disease, or thoracic ossification of theposterior longitudinal ligamentin in adults.1–3, 10–13 Butthese reports do not address the problem of accessingmultileveled spinal deformities requiring stabilization inthe pediatric population. There are few reports in theliterature concerning the anterior approach to the cervi-cothoracic junction for instrumentation and fusion ofpediatric scoliosis.5
Cauchoix and Binet4 were the first in the literature todescribe a trans-sternal approach to the upper thoracicspine. They advocated that access to the cervical spinewas best gained via an incision along the anterior marginof the sternomastoid muscle, and thoracic access isgained by an incision in the midline of the sternum that iscarried down to the xiphoid process. The two-part inci-sion is then merged into one operation area by dividingthe subhyoid muscles. Cauchoix used this anterior cervi-cothoracic approach on 2 patients: a 4-year-old childwith quadriplegia secondary to Pott’s disease involvingthe first and second dorsal vertebrae with a sharp kypho-sis and bilateral abscess requiring cord decompression,and a 19-year-old man with complete spastic paraplegiawhose plain films demonstrated an osteolytic tumor de-stroying the first and second dorsal vertebrae and theupper part of the third vertebra.4 The authors achievedsatisfactory results with their two patients, but otherswho later used this approach demonstrated a high op-erative mortality and morbidity, and it was subse-quently advocated that direct anterior exposures shouldbe abandoned.1
Hall5 outlined the specific indications for different an-terior approaches to spinal deformities. In this review, hesuggests that if access to the upper four thoracic verte-brae is needed for correction of a rigid congenital scoli-osis then a staged approach be used whereby the apical
From the *Department of Orthopaedics, British Columbia’s Children’sHospital; and the Departments of †Radiology, ‡Surgery, Division ofCardiovascular Surgery, and §Orthopaedics, Division of PaediatricOrthopaedics, University of British Columbia, British Columbia’sChildren’s Hospital, Vancouver, British Columbia, Canada.Acknowledgment date: April 15, 2004. First revision date: September10, 2004. Second revision date: September 20, 2004. Acceptance date:September 23, 2004.The manuscript submitted does not contain information about medicaldevice(s)/drug(s).No funds were received in support of this work. No benefits in anyform have been or will be received from a commercial party relateddirectly or indirectly to the subject of this manuscript.Address correspondence and requests for reprints to Kishore Mulpuri,Department of Orthopaedics, British Columbia’s Children’s Hospital,A200-4480 Oak Street, Vancouver, British Columbia, V6H 3V4. E-mail: [email protected]
E305
vertebral bodies are resected by splitting the sternum,and posterior correction is then done with Harringtoninstrumentation and spinal fusion. Hall noted that ante-
rior access is needed in children with severe kyphosisbecause posterior fusion alone will not resist all theforces on it nor will it resist curve progression as the childgrows. However, he does note that a posterior fusiondone on those with slight deformities may be sufficient toarrest progression, and should further stabilization beneeded, the posterior fusion mass affords some stabilityto simplify later attempts at anterior stabilization.5
Maciejczak et al8 recommended an anterior transtho-racic, transpleural approach as the best route to gainaccess to the vertebral bodies of the thoracic spine as itallows wide exposure of the anterior spinal column andmakes surgical intervention possible. However, the au-thors acknowledged that this approach does not allowadequate exposure of the upper three thoracic vertebrae,which can be best exposed by a direct anterior approach.
Knoller and Brethner6 reviewed sternotomy ap-proaches for surgical treatment of the spine at the cervi-cothoracic junction. They recommended the approachfirst described by Cauchoix and Binet,4 noting that itallowed adequate cervical spine exposure and had fewercomplications than the other anterior exposure tech-niques such as the high anterior transthoracic approachused by Hodgson14 or the modified anterior approachdescribed by Kurtz et al.7 We present six pediatric caseswith severe cervicothoracic kyphosis and/or scoliosis re-
Figure 1. The patient is placed on the operating table in a supineposition and the neck is hyperextended and turned to the left,allowing a right sided approach.
Figure 2. The sternum isopened, the thymus gland re-sected and the brachiocephalictrunk is mobilized to allow con-tiguous access to the anteriorcervical spine and upper tho-racic spine.
E306 Spine • Volume 30 • Number 11 • 2005
quiring extensive access. In all cases, the lower cervicaland upper thoracic spine was accessed using a neck dis-section and sternal splitting technique with the aid of acardiothoracic surgeon.
Materials and Methods
A detailed chart and radiographic review was completed of sixconsecutive patients managed at our center with a combinedanterior neck and sternal splitting approach. The indications,surgical technique, and outcome are reviewed for each case.Follow-up radiographs and clinical status were reviewed ineach case to determine the degree of correction and complica-tions, if any.
Anterior Neck and Sternal Splitting Surgical TechniqueThe surgical approach is completed with the assistance of acardiothoracic surgeon. A standard extensile anterior cervicalspine approach is used, incorporating an anterior sternal ex-tension (Figure 1). The neck dissection is completed first in astandard fashion. An incision is made along the medial borderof the sternomastoid muscle, extending down to the sternalnotch. The sternomastoid muscle is retracted laterally with theneurovascular sheath, including the carotid artery, the jugularvein, and the vagus nerve. Division of the omohyoid, sternohy-oid and sternothyroid muscles facilitates extensile exposure.The incision is extended as a midline sternotomy approach.Blunt digital dissection is used to mobilize the retrosternal softtissues. The stenum is then split using a sternal saw in a stan-dard fashion. After opening the sternum, the thymus gland isresected to provide exposure and the brachiocephalic trunk is
mobilized and controlled with a vessel loop. At this point in theprocedure the anterior cervical spine and upper thoracic spineare able to be accessed contiguously (Figure 2). The pericar-dium was opened in two patients in this series to increase mo-bility of the brachiocephalic trunk. However, the dissection ofthe brachiocephalic trunk can be done down to the pericardialreflection without having to open the pericardium.
The trachea and esophagus are retracted slightly away fromthe middle line with a right angle retractor, and a deep rightangle retractor is placed under the innominate artery andpulled forward and downward as necessary to provide accessto the lower cervical and upper thoracic spine (Figure 3). Thedistal extent of the exposure at this point will depend on thepatient’s anatomy and deformity; in most cases, the surgeonwill be able to access T4. Disc removal and instrumentation cannow be safely completed. Aggressive distal exposure will placethe recurrent laryngeal nerve under traction and must be donecarefully. Typically, left sided anterior cervical approaches arepreferred because of the distal course of the recurrent laryngealnerve on that side. We have elected to perform a right-sidedapproach with mobilization of the brachiocephalic trunk. Me-dial displacement of the trunk facilitates exposure of more dis-tal segments of the thoracic spine on the right side, down to T6in one patient in this series (Figure 4).
After completion of the orthopaedic procedure, hemostasisis established. The sternum is reapproximated with wires orsuture according to the age. The sterno-thyroid and omohyoidmuscles are reattached. The neck incision is closed in usualfashion. A small silastic drain may be required under thesterno-thyroid muscle if hemostasis is a problem in the cervicalportion of the approach. A mediastinal tube is placed as incardiac surgical procedures.
Results
The clinical features of these patients are summarized inTables 1 and 2.
Figure 3. Retraction of the trachea, esophagus and innominateartery provides access to the lower cervical and upper thoracicspine.
Figure 4. Medial displacement of the brachiocephalic trunkallows more distal access to the thoracic spine.
Table 1. Clinical Summary of Six Cases
Patient Diagnosis Sex Age (yrs) Prior Treatment Access Gained Follow-up
1 Klippel-Feil M 11 None C5-T5 9 years2 Proteus Syndrome M 15 None C5-T4 3 years, 10 months3 Neurofibromatosis I F 15 2 posterior fusions C6-T5 2 years, 8 months4 Neurofibromatosis I M 4 None C5-T4 (thoracotomy performed for T5–T9 access) 2 years, 4 months5 Larsen’s Syndrome F 3 3 posterior fusions C6-T6 22 months6 Klippel-Feil F 13 2 C6-T5 6 months
E307Approach to the Cervicothoracic Junction in Children • Mulpuri et al
Representative Cases
Case 4. A four-year-old boy with severe neurofibroma-tosis type I presented with a rapidly evolving scoliosis ofthe cervical thoracic spine that progressed from 40° to80° over the course of 1 year. There were no neurologicalsequelae from his curve. Preoperative computed tomog-raphy (CT) showed extensive involvement of the spinewith dystrophic changes at the apex of the curve. Mag-netic resonance imaging revealed large plexiform neu-roma in both sides of the chest with a huge neurofibromain the upper left chest. It was elected to attempt to stabi-lize the curve both anteriorly and posteriorly because ofthe rapid collapse and dystrophic nature of the defor-mity. Surgical correction was done in a two step proce-dure; the first stage was anterior cervical thoracic fusionfrom C5 to T4 via a neck dissection and sternal splitapproach. A right lateral thoracotomy was required foranterior fusion from T5–T9. Five days later, the patientwas brought back to the operating room for the secondstage of correction, posterior instrumentation, and fu-sion from C5 to T10. At 18 months after surgery, thepatient had returned to full activities and no complica-tions were observed with regard to his spinal fusion (Fig-ure 5).
Case 5. A 3-year-old girl with Larsen’s Syndrome pre-sented with a collapsing high thoracic kyphosis thatthreatened the viability of her spinal cord. Two previousposterior spinal fusions had been attempted. The initialsurgery resulted in pseudarthrosis and ongoing angula-tion. Halo traction was instituted and posterior fusionwas repeated in the corrected position. Following revi-sion and prolonged immobilization, the patient couldnot be removed from her halo because of curve progres-sion and clinically obvious spinal instability. No neuro-logical signs or symptoms were present. Magnetic reso-nance imaging demonstrated that the apex of the curvewas at T3–T4 and that the subarachnoid space was ex-tremely compromised though that region. A focus of in-creased signal intensity was seen within the spinal cordjust distal to the apex of the kyphosis. A CT demon-strated pseudoarthroses at two levels above the apex ofkyphosis. Incidentally it was found that she did not havevertebral arteries in her vertebral foramens. Instead, thevertebral arteries lay on the posterior aspect of the lam-inae of the cervical spine (Figure 6). She had a staged
procedure with a neck dissection and sternal split ap-proach with fusion from C6 to T6 and then went intohanging traction. One month later, she underwent a pos-terior instrumentation and fusion from C6 to T7 with acombination of pedicle and lateral mass screws. She wasconverted from hanging traction to a halo thoracic vestfor 8 weeks, and her postoperative course was unevent-ful. At 1 year after surgery imaging revealed a stablefusion mass (Figure 7).
Discussion
This small case series demonstrates the usefulness of thedirect anterior approach in gaining access to the cervico-thoracic junction to address complex spinal deformitiesin the pediatric population. With curves of this magni-tude, the goal of surgical intervention is to stabilize seg-ments and impede curve progression. This extensile neckdissection and sternal split approach gave good exposure
Table 2. Pre- and Postoperative Scoliosis and Kyphosis and Summary of Complications
Figure 5. A and B, Postoperative radiographs of a 4-year-old boywith neurofibromatosis type I (case 4) demonstrating extent ofanterior exposure. Additional thoracotomy was done to achieveaccess to distal 3 levels.
E308 Spine • Volume 30 • Number 11 • 2005
from C5 to T6 to allow instrumentation and fusion. Inone patient an adjunctive thoracotomy was performed toaccess the bodies of T5 through T9, an alternative to thetrap door approach to the lower thoracic spine. Fewreports in the literature address the extensive access re-quired for instrumentation and stabilization of severescoliotic and kyphotic deformities. Although it appearsthat little scoliotic and/or kyphotic correction wasachieved (Table 2), the goal of stabilizing the spine andpreventing curve progression was realized.
The potential complications related to this approachare iatrogenic damage to the recurrent laryngeal nerve,the pericardium, and the esophagus. It is recommendedthat the aid of a cardiothoracic surgeon be used becausewe feel it is not an approach to be attempted by a spinalsurgeon alone due to the complexity of the exposure. Inthis series the most significant complication related tothis approach was recurrent laryngeal nerve palsy in onepatient, a complication that had also been reported byothers using an anterior approach.2 The central neuro-logical impairment experienced by one patient fully re-solved and was not related to the sternal splitting tech-nique. This patient’s preoperative CT scan haddemonstrated an almost spondyloptosis of C7 on T1with the body of T1 impinging into the spinal cord canalwith clinical evidence of cord compression and progres-sive myelopathy. Postoperative imaging showed that thecord lay in the channel, and there was no focal anomalythat would suggest a total cord lesion. It was felt that thepatient had suffered a vascular injury to his cord that wasunrelated to the sternal spitting approach.
This surgical approach is complex and invasive. It isrecommended in rare cases where the treating surgeonmust achieve contiguous anterior cervicothoracic fusionto stabilize the spine and protect the patients’ neurology.The cases presented in this paper include severe dystro-phic deformities, complex congenital kyphotic defor-mities, and cases in which previous isolated posteriorfusion attempts had failed. In our hands, with the as-sistance of a cardiothoracic surgeon, we have foundthe anterior sternal split approach to be a safe and
Figure 6. A-C, Preoperative CT(case 5). Note that the vertebralarteries are not in the vertebralforamens, they lie on posterioraspect of the laminae of the cer-vical spine.
Figure 7. Pre- (A) and postoperative (B) radiographs of a 3-year-old girl with Larsen’s Syndrome and collapsing high thoracickyphosis (case 5).
E309Approach to the Cervicothoracic Junction in Children • Mulpuri et al
effective approach to gain access to the cervicothoracicjunction in children.
Key Points
● The sternal splitting approach was useful in gain-ing multileveled access in pediatric patients withsevere cervicothoracic kyphosis and/or scoliosis.● The sternal split and neck dissection techniqueare described● The aid of a cardiothoracic surgeon to provideanterior exposure is recommended
References1. Calliauw J, Dallenga A, Caemaert J. Trans-sternal approach to intraspinal
tumours in the upper thoracic region. Acta Neurochir 1994;127:227–31.2. Hanakita J, Suwa H. Sternal splitting approach to upper thoracic lesions
located anterior to the spinal cord. Neurol Med Chir 1999;39:428–33.3. Sundaresan N, Shah J, Foley KM, et al. An anterior surgical approach to the
upper thoracic vertebrae. J Neurosurg 1984;61:686–90.4. Cauchoix J, Binet JP. Anterior surgical approaches to the spine. Ann R Coll
Surg Engl 1957;21:237–43.
5. Hall JE. The anterior approach to spinal deformities. Orthop Clin North Am1972;3:81–98.
6. Knoller SM, Brethner L. Surgical treatment of the spine at the cervicothoracicjunction: an illustrated review of a modified sternotomy approach with thedescription of tricks and pitfalls. Arch Orthop Trauma Surg 2002;122:365–8.
7. Kurtz LT, Pursel SE, Herkowitz HN. Modified anterior approach to thecervicothoracic junction. Spine 1991;16:S542–7.
8. Maciejczak A, Radek A, Kowalewski, J et al. Anterior transsternal approachto the upper thoracic spine. Acta Chir Hung 1999;38:83–6.
9. Micheli LJ, Hood RW. Anterior exposure of the cervicothoracic spine usinga combined cervical and thoracic approach. JBJS 1983;65A:992–7.
10. Kojima T, Waga S, Kubo Y, et al. Surgical treatment of ossification of theposterior longitudinal ligament in the thoracic spine. Neurosurg 1994;34:854–8.
11. Fujimura Y, Nishi Y, Toyama Y, et al. Anterior decompression and fusionfor ossification of the posterior longitudinal ligament of the upper thoracicspine causing myelopathy using the manubrium splitting approach. SpinalCord 1996;34:387–93.
12. McElvein RB, Nasca RJ, Dunham WK, et al. Transthoracic exposure foranterior spinal surgery. Ann Thorac Surg 1988;45:278–83.
13. Smith TK, Stallone RJ, Yee JM. The thoracic surgeon and anterior spinalsurgery. J Thorac Cardiovasc Surg 1979;77:925–8.
14. Hodgson AR, Stock FE, Fang HSY et al. Anterior spinal fusion. The opera-tive approach and pathologic findings in 412 patients with Pott’s disease ofthe spine. Br J Surg 1960;48:172–8.
