Congenital scoliosis results from abnormal vertebral development with subsequent asymmetric growth of the spine. While the vertebral anomalies are present at birth, the deformity may not be diagnosed until later in life. These anomalies can affect both the sagittal and coronal alignment of the spine with scoliosis and kyphosis often occurring together in the same patient. Of the possible vertebral anomalies that can cause spinal deformities, hemivertebra is one of the most common.1
The severity of spinal deformity created by a hemivertebra depends on the location, type, and a patient's remaining growth.2–4 Surgical treatment is frequently necessary in the progressive curves. The goal is to achieve a balanced spine at the end of growth with coronal and sagittal contours that are anatomically and cosmetically acceptable.
Controversy remains over the “best” surgical approach for the treatment of these challenging congenital deformities. Numerous techniques, including in situ fusion, hemiepiphysiodesis, instrumented fusion, and hemivertebra excision, have been advocated. The purpose of this retrospective study was to compare the clinical and radiographic outcomes of three surgical treatments for congenital spinal deformity due to a hemivertebra.
MATERIALS AND METHODS
After approval at each of the institutional review boards, a multicenter retrospective study involving 12 centers was performed to evaluate patients with congenital spinal deformity who underwent spinal surgery. All patients were younger than 18 years at surgery and all procedures were performed between 1991 and 2004. Decisions regarding surgical approach and implant type were made by individual surgeons and related to the location of the hemivertebra, the magnitude of the main and compensatory curves, the size of the patient as well as the patient's clinical appearance. Patients were included if they met the following inclusion criteria: diagnosis of scoliosis or kyphosis due to one or two hemivertebrae, underwent surgery for the scoliotic spinal deformity, age less than or equal to 21 years at the time of surgery, and a minimum of 2-year postoperative follow-up.
Patients were divided for analysis into three surgical treatment groups. Group 1 had a fusion without correction with either a hemiepiphysiodesis or an in situ fusion. Group 2 underwent an instrumented correction without a hemivertebra resection. Group 3 had an instrumentation correction with either an anterior and/or posterior hemivertebra excision.
Clinical data retrieved from the medical records included age at surgery, sex, hemivertebra location and type, surgical time, estimated blood loss, blood products given, surgical approach, use of anterior or posterior releases, and complications. Radiographic data evaluated included preoperative and 2-year postoperative coronal and sagittal Cobb angles across congenital and secondary curves as well as percent deformity correction.
Continuous data were checked for normality of distribution and homogeneity of variance prior to utilization of analysis of variance (ANOVA) with Bonferroni post hoc comparisons to examine differences between the three treatment groups. Because of the heterogeneous nature of the patients preoperatively, repeated-measures ANOVA was used to measure the change from preoperative to 2 years postoperative for sagittal profile between the three groups. Statistical significance was set at α = 0.05.
Seventy-six patients treated between 1991 and 2004 were evaluated. The mean age was 8 years (range: 1–18). The hemivertebra were classified as fully segmented, nonincarcerated (n = 51; 67%), incarcerated (n = 1; 1%), and semisegmented (n = 24; 32%). There were 65 patients with single and 11 patients with double hemivertebra. There were 14 (18.4%) group 1 patients (fusion without correction; Figure 1), 20 (26.3%) group 2 patients (correction without hemivertebra resection; Figure 2), and 42 (55.3%) group 3 patients (correction with hemivertebra resection; Figure 3).
In group 1, there were 12 single and 2 double hemivertebrae. All but two patients had improvement in their Cobb angle. One patient's curve remained unchanged while another patient required extension of their fusion for progressive deformity. Within group 2, there were 17 single and 3 double hemivertebrae. Eleven patients had a combined anterior/posterior fusion while nine had a posterior-only procedure. Various instrumentation types, including hooks, wires, pedicle screws, and hybrid constructs, were used. In group 3, there were 36 single and 6 double hemivertebrae. The majority had a posterior hemivertebra excision (33 of 42 patients). While various instrumentation types were utilized, a greater percentage of patients had all pedicle screw constructs compared with group 2.
Group 1 (37 ± 14°) and group 3 (35 ± 26°) had smaller preoperative curves than group 2 (55 ± 26°) (P < 0.01) (Table 1). At 2 years postoperation, group 3 had better percent correction (73%, mean Cobb angle: 10°) than group 1 (27%, mean Cobb angle: 24°) and group 2 (42%, mean Cobb angle: 34°) (P < 0.001). The sagittal deformity was greatest preoperatively in group 2 (52 ± 22°) compared with group 1 (17 ± 30°) and group 3 (18 ± 21°) (P < 0.02) (Table 1). Postoperatively, there was no difference in the change in the sagittal Cobb angle between the groups (P = 0.183).
Perioperative data demonstrated that group 3 had shorter fusion (P = 0.001), less estimated blood loss (EBL, P = 0.03), and a trend toward shorter operative times compared with group 2 (P = 0.10) (Table 2). Groups 1 and 3 had similar fusion length. At the time of surgery, Group 3 patients were significantly younger (5 years) compared with group 2 (10.3 years) (P < 0.001) and group 1 (9.7 years) (P = 0.023).
Eleven patients had their hemivertebrae at either L4 or L5. Three were treated with an in situ fusion/hemiepiphysiodesis (group 1) and had a mean preoperative curve of 25 ± 17°. At 2 years postoperation, there was minimal correction with the mean curve size at 20 ± 15°. The other eight patients had an average preoperative curve of 36 ± 11° and underwent excision. Similar to the entire group 3, there was a 66% correction (mean Cobb angle 12 ± 9°) in the coronal deformity.
The overall complication rate for the entire group was 30%: group 1 (23%), group 2 (17%), and group 3 (44%) (P = 0.09) (Table 3). There were a total of four postoperative infections (three superficial and one deep). All infections resolved with one superficial and the deep infection requiring incision and drainage. One group 1 and one group 3 patient developed a pseudoarthrosis that required revision surgery. Similarly, there were two additional patients each in groups 1 and 3 that had progression of their deformity and required extension of their fusion. There were a total of six instrumentation failures (one in group 2 and five in group 3). Four of these required revision surgery.
There were also a total of six neurologic complications (one in group 2 and five in group 3). One group 3 patient had a postoperative seizure. One group 2 patient had an intraoperative loss of SSEP and MEP that resolved after decreasing the correction and had no postoperative findings. Of the group 3 patients, one patient had no intraoperative events but awoke with some burning sensation in the feet. Computed tomography myelogram was negative and burning resolved on postoperative day 2. Another patient presented at 2 weeks with bilateral lower extremity spasms that were felt to be secondary to stretching of the neural elements. Computed tomography myelogram was negative and symptoms resolved after course of steroids. The third patient underwent an L5 excision and had a significant decrease in spinal evoked potentials. She awoke with ipsilateral lower extremity weakness and pain. At 2 weeks, the patient had full motor recovery and was being treated with neurontin for pain. The fourth patient underwent an L3 hemivertebrae excision and postoperatively had ipsilateral quadriceps weakness. Intraoperative neurologic monitoring was considered normal as well as postoperative computed tomography and magnetic resonance imagery. The patient was found to have complete recovery by 10 months.
Analysis of individual clinical centers demonstrated improved radiographic results at the site with the greater experience with hemivertebra resections (Table 4). This site contributed 17 of the 42 (41%) reported hemivertebra excisions. The subsequent two sites with the largest numbers of group 3 patients each had six patients (14%). The most-experienced site (G3) had a greater percent coronal correction in group 3 than other sites (84 ± 19% vs. 50 ± 25%, P < 0.001). In addition, this was associated with a lower complication rate and a trend to less EBL and operative time.
Numerous techniques for the treatment of congenital scoliosis secondary to a hemivertebra have been published. The results can be variable depending on the location and type of hemivertebra as well as the remaining growth of the patient. Improved neurologic monitoring and spinal instrumentation have also allowed for more aggressive correction techniques.
The aim of this study was to retrospectively compare the clinical and radiographic outcomes of three surgical treatments for congenital spinal deformity due to a hemivertebra. Patients were categorized on the basis of techniques that provided different degrees of acute correction at the index procedure. Patients who underwent in situ fusion or hemiepiphysiodesis (group 1) had little to no acute correction. The use of instrumentation in group 2 allowed for some correction to be attained while hemivertebra excision (group 3) provided the greatest potential for deformity correction at the time of surgery.
Patients in group 1 had the smallest percent of correction (27%) at 2 years follow-up. In all the 12 centers, it was the least frequently used technique, reserved for patients with smaller preoperative curves. Interestingly, the average age at the time of surgery was 9.7 years. In situ fusion is typically performed in younger patients with small deformities at high risk of progression that are otherwise well-balanced.5 Hemiepiphysiodesis is similarly recommended in patients younger than 5 years with less than 40° to 50° curves.5 The goal is to arrest the growth at the convexity and allow gradual correction through concave growth.
The anticipated success of this gradual correction is debatable. Roaf6 and Thompson7 both reported improvement in the Cobb angle in 60% and 76% of patients, respectively. On the contrary, the studies by Winter et al8,9 and Keller et al10 demonstrated that the majority of patients had cessation of curve progression similar to an in situ fusion. The average published correction for a successful hemiepiphysiodesis is 10°.6,9 This is similar to the correction demonstrated in our patients. The final curve correction may be slightly greater considering that not all the patients attained skeletal maturity at the time of analysis.
Group 2 patients had partial correction of their congenital deformity with the use of instrumentation in addition to the fusion. Instrumentation has been shown to be safe and effective in congenital scoliosis.11,12 The goal of an instrumented fusion is to achieve correction through the flexible and segmented portions of the spine.13 Hedequist et al14 demonstrated an improved correction using Harrington instrumentation compared with a noninstrumented fusion in patients with congenital scoliosis (22° vs. 5°, respectively).
The patients in this series were older at the time of their intervention and had the largest preoperative coronal and sagittal deformity. They also had larger secondary curves as well. As a result, they underwent more extensive fusions compared with groups 1 and 3 (7 vertebra vs. 3 vertebra). Uninstrumented fusions or acute correction through a single-level hemivertebrectomy would not have reliably achieved a well-balanced spine. In determining the fusion lengths, bending and/or traction films were used to determine flexibility of the congenital curve as well as the response of the secondary curves. Consideration for an anterior fusion in younger patients was also taken to minimize the risk of crankshaft.15
Hemivertebrectomy has become increasingly popular in the management of congenital scoliosis secondary to a hemivertebra. It removes the source of the deformity, thereby preventing the worsening of the primary curves. Hemivertebra excision (group 3) was the most commonly performed procedure among the 12 sites. It provided the greatest correction, 73%, across the shortest segment of spine. This is similar to the rates of correction found in other series.16–22
The average age in group 3 was significantly younger than that in the other two groups. Hemivertebra excision has been shown to be safe and effective in the very young patient. Klemme et al21 demonstrated a 67% correction at a mean age of 19 months. Callahan et al19 found the greatest curve correction in those patients younger than 4 years. Ruf and Harms18 recommended early excision to avoid the development of severe deformities and secondary structural curves.
The surgical approach to a hemivertebra excision can be done through a combined anterior/posterior or a posterior-only procedure. Among the 12 sites, the majority were performed through a posterior approach (79%). Traditionally, hemivertebrectomy was performed through a combined procedure. Leatherman and Dickson23 reported hemivertebra excisions, utilizing a two-stage anterior and posterior procedure. This was followed by successful excision through a single-stage sequential combined approach.20,21 Simultaneous anterior/posterior hemivertebra excision has also been demonstrated.17 More recent studies have published on hemivertebrectomy through a posterior-only approach. Ruf and Harms18 reported on a 69% correction after posterior excision in 28 patients at 3.5 years follow-up. Shono et al described a similar experience in their 12 patients.22 Nakamura et al have the longest follow-up, demonstrating maintenance of correction at an average of 12.8 years after a posterior-only procedure.24
Hemivertebra located near the lumbosacral junction (L4 or L5) were further analyzed. They can frequently lead to significant deformity, resulting in either major trunk shift or large compensatory curves.4 Excision has generally been recommended for hemivertebra at the lumbosacral junction.5 Eleven patients in this series had a hemivertebra located at either L4 or L5. Three patients (group 1) were felt to have acceptable truncal balance preoperatively and, therefore, underwent in situ fusion or hemiepiphysiodesis. At follow-up, their spinal curvature remained unchanged with no further worsening of their clinical deformity. The other eight patients (group 3) had larger preoperative curves and underwent the recommended treatment of excision. This resulted in an average correction of 66%, which nearly restored the normal lumbosacral take-off compared with the group 1 patients.
When determining the most appropriate method for managing a hemivertebra, it is important to balance the risks with the potential correction that can be attained. Group 1 had the smallest correction in Cobb angle but also had the lowest complication rate and EBL between the groups. Group 2 had the longest operative time and greatest EBL, which was secondary to the longer fusions. Group 3 had the most-aggressive deformity correction but also had a greater complication rate including a higher risk of neurologic injury. Of the six instrumented failures, four (one in group 2 and three in group 3) were considered serious radiographic adverse events or major complications and required revision surgery secondary to coronal or sagittal decompensation.25,26
This risk of neurologic injury after hemivertebra excision is well-documented and may involve either a nerve-root or spinal-cord lesion.16,27–29 Except for one case in group 2 where there was only an intraoperative loss of neurologic signals, all other neurologic injuries were in the excision group. In all cases, the patients had complete recovery. The injuries appeared to be secondary to a stretch of neural elements. Motor findings were specific to the nerve roots around the lumbar hemivertebra being excised.
This study would also suggest that there is a learning curve associated with hemivertebra excisions. The site with the greatest experience demonstrated a greater deformity correction across a smaller spinal segment. In addition, this was associated with a lower complication rate. Aydogan et al30 had a similar conclusion recommending extreme care and experience in performing this technically demanding procedure. There may be other factors beyond experience that were responsible for the differences between sites that were not discernable from data obtained from a retrospective review. In addition, the lack of specific outcome measures prevented any determination of clinical significance.
Another limitation of this retrospective study is the heterogeneous nature of the patients especially in the magnitude of spinal deformity. Procedures for each patient were selected by the individual surgeon at each center and were related to the location of the hemivertebra, the magnitude of the main and compensatory curves, the size of the patient as well as the patient's clinical appearance. Direct comparisons of the three procedures in patients with similar spinal deformities, therefore, could not be done because of the heterogeneity of the patient population preoperatively. In addition, considering these differences, it was not anticipated that each procedure could be equally applied to every patient. When deciding between these three procedures, the treating physician must consider both the magnitude of the initial deformity as well as the desired radiographic and clinical outcome.
Hemivertebra resection produces better deformity correction than either hemiepiphysiodesis/in situ fusion or instrumented fusion without a resection. It is recommended for patients before the development of significant compensatory curves where a significant correction is desired across a small segment. With this improved correction, there is an increased risk of complications including neurologic injury. Our study would suggest that increased experience with hemivertebrectomy may minimize these risks while optimally correcting the congenital scoliosis.
- Hemivertebra resection with posterior instrumentation resulted in greater correction with shorter fusions and less EBL compared with hemiepiphysiodesis/in situ fusion or instrumented fusion without resection.
- Hemivertebra resection with posterior instrumentation had a higher complication rate than hemiepiphysiodesis/in situ fusion or instrumented fusion without resection.
- This investigation suggests that experience with the technique of hemivertebra resection with posterior instrumentation may be important for an optimal result.
1. McMaster. Congenital scoliosis
. In: Weinstein SL, ed. The Pediatric Spine: Principles and Practices. New York: Raven Press; 1994:227–44.
2. Nasca RJ, Stilling FH III, Stell HH. Progression of congenital scoliosis
due to hemivertebrae and hemivertebrae with bars. J Bone Joint Surg Am 1975;57:456–66.
3. McMaster MJ, Ohtsuka K. The natural history of congenital scoliosis
. A study of two hundred and fifty-one patients. J Bone Joint Surg Am 1982;64:1128–47.
4. McMaster MJ, David CV. Hemivertebra as a cause of scoliosis. A study of 104 patients. J Bone Joint Surg Br 1986;68:588–95.
5. Hedden D. Management themes in congenital scoliosis
. J Bone Joint Surg Am 2007;89(Suppl 1):72–8.
6. Roaf R. The treatment of progressive scoliosis by unilateral growth-arrest. J Bone Joint Surg Br 1963;45:637–51.
7. Thompson AG, Marks DS, Sayampanathan SR, et al. Long-term results of combined anterior and posterior convex epiphysiodesis for congenital scoliosis
due to hemivertebrae. Spine (Phila Pa 1976) 1995;20:1380–5.
8. Winter RB. Convex anterior and posterior hemiarthrodesis and hemiepiphyseodesis in young children with progressive congenital scoliosis
. J Pediatr Orthop 1981;1:361–6.
9. Winter RB, Lonstein JE, Denis F, et al. Convex growth arrest for progressive congenital scoliosis
due to hemivertebrae. J Pediatr Orthop 1988;8:633–8.
10. Keller PM, Lindseth RE, DeRosa GP. Progressive congenital scoliosis
treatment using a transpedicular anterior and posterior convex hemiepiphysiodesis and hemiarthrodesis. A preliminary report. Spine (Phila Pa 1976) 1994;19:1933–9.
11. Hedequist DJ, Hall JE, Emans JB. The safety and efficacy of spinal instrumentation
in children with congenital spine deformities. Spine (Phila Pa 1976) 2004;29:2081–6; discussion 7.
12. Ruf M, Harms J. Hemivertebra resection
by a posterior approach: innovative operative technique and first results. Spine (Phila Pa 1976) 2002;27:1116–23.
13. Hedequist D, Emans J. Congenital scoliosis
: a review and update. J Pediatr Orthop 2007;27:106–16.
14. Hall JE, Herndon WA, Levine CR. Surgical treatment of congenital scoliosis
with or without Harrington instrumentation. J Bone Joint Surg Am 1981;63:608–19.
15. Kesling KL, Lonstein JE, Denis F, et al. The crankshaft phenomenon after posterior spinal arthrodesis for congenital scoliosis
: a review of 54 patients. Spine (Phila Pa 1976) 2003;28:267–71.
16. Bollini G, Docquier PL, Viehweger E, et al. Lumbar hemivertebra resection
. J Bone Joint Surg Am 2006;88:1043–52.
17. Hedequist DJ, Hall JE, Emans JB. Hemivertebra excision in children via
simultaneous anterior and posterior exposures. J Pediatr Orthop 2005;25:60–3.
18. Ruf M, Harms J. Posterior hemivertebra resection
with transpedicular instrumentation: early correction in children aged 1 to 6 years. Spine (Phila Pa 1976) 2003;28:2132–8.
19. Callahan BC, Georgopoulos G, Eilert RE. Hemivertebral excision for congenital scoliosis
. J Pediatr Orthop 1997;17:96–9.
20. Bradford DS, Boachie-Adjei O. One-stage anterior and posterior hemivertebral resection and arthrodesis for congenital scoliosis
. J Bone Joint Surg Am 1990;72:536–40.
21. Klemme WR, Polly DW Jr, Orchowski JR. Hemivertebral excision for congenital scoliosis
in very young children. J Pediatr Orthop 2001;21:761–4.
22. Shono Y, Abumi K, Kaneda K. One-stage posterior hemivertebra resection
and correction using segmental posterior instrumentation. Spine (Phila Pa 1976) 2001;26:752–7.
23. Leatherman KD, Dickson RA. Two-stage corrective surgery for congenital deformities of the spine. J Bone Joint Surg Br 1979;61-B:324–8.
24. Nakamura H, Matsuda H, Konishi S, et al. Single-stage excision of hemivertebrae via
the posterior approach alone for congenital spine deformity: follow-up period longer than ten years. Spine (Phila Pa 1976) 2002;27:110–5.
25. Flynn JM, Betz RR, O'Brien MF, et al. Radiographic classification of complications of instrumentation in adolescent idiopathic scoliosis. Clin Orthop Relat Res 2009:665–9.
26. Weis JC, Betz RR, Clements DH III, et al. Prevalence of perioperative complications after anterior spinal fusion for patients with idiopathic scoliosis. J Spinal Disord 1997;10:371–5.
27. Holte DC, Winter RB, Lonstein JE, et al. Excision of hemivertebrae and wedge resection in the treatment of congenital scoliosis
. J Bone Joint Surg Am 1995;77:159–71.
28. Winter RB. Congenital kyphoscoliosis with paralysis following hemivertebra excision. Clin Orthop Relat Res 1976:116–25.
29. Lazar RD, Hall JE. Simultaneous anterior and posterior hemivertebra excision. Clin Orthop Relat Res 1999:76–84.
30. Aydogan M, Ozturk C, Tezer M, et al. Posterior vertebrectomy in kyphosis, scoliosis and kyphoscoliosis due to hemivertebra. J Pediatr Orthop B 2008;17:33–7.