Analysis of the length of xation during hemivertebra resection. A monocentric cohort

Background: We aimed at reviewing results of surgical correction of spinal deformity due to hemivertebra with regard to length of instrumentation. Methods: Study comprised 117 patients with congenital spinal deformity treated between 2010 and 2018. Patients’ aged 1 to 18 years. Mean follow-up was 3 years (1 to 8). Posterior approach was used in all cases. Patients were subdivided into 4 groups. Result: Unilateral monosegmental xation (Group I) for 15 children with mean age of 48.3 months. Cobb angle for scoliosis and kyphosis were 31.1 0 + 6.4 0 and 29.3° + 11.9 0 corrected to 7.8 0 + 6.8 0 and 4.7 0 + 4.0 0 respectively. Blood loss was 213.6 mL. Operating time was 165 minutes. Bilateral monosegmental xation (Group II) for 24 patients with mean age of 53.1 months. Cobb angle for scoliosis and kyphosis were 32.5° + 8.6 0 and 30.9° + 4.3 0 corrected to 5.3 0 + 2.8 0 and 0.2 0 + 11.6 0 respectively. Blood loss was 215.4 mL. Operating time was 160.5 minutes. Bilateral three-segmental xation (Group III) for 29 patients with mean age of 78.2 months. Cobb angle for scoliosis and kyphosis were 36.6° + 10.6 0 and 37.6° + 14.6 0 corrected to 6.2 0 + 6.1 0 and 5.0 0 + 5.2 0 respectively. Blood loss was 342.7 mL. Operating time was 197.0 minutes. Bilateral polysegmental xation (Group IV) for 49 children with mean age of 112.7 months. Cobb angle for scoliosis and kyphosis were 40.2° + 14.8 0 and 58.2° + 35.6 0 corrected to 10.7 0 + 9.6 0 and 10.7 0 + 10.3 0 respectively. Blood loss was 549.3 mL. Operating time was 288.8 minutes. Conclusion: Choice of spinal xation technique and the length of xation were determined on the basis of patient’s age, magnitude of the deformity and concomitant


Background
The rst report of the removal of a portion of the spine for a xed deformity was by N. Royle (1928). [1] Initially hemivertebra excision was produced via an anterior-only approach without metal construct. [2] Thereafter, the procedure was performed from two approaches without usage of implants and provided a good correction of scoliosis but later presented with local kyphotic deformity and progressive neurological de cit leading to paraplegia. [3,4] The complications could be avoided with application of metal constructs. M. Bergoni (1981) offered a one-stage hemivertebra excision via a combined anterior-posterior approach with the use of metal xators. [5,6,7] Over time, with steady advances in instrumentation and metal constructs, the procedure could be produced through a posterior approach. [8,9,10,11] Screw xation systems could be used for younger children. Follow-up dispelled worries of several researchers and proved to be of value demonstrating pedicle growth potential and no spinal stenosis. [12,13,14,15] Choice of techniques and options of surgical treatment of congenital scoliosis resulting from hemivertebra remains a challenging issue. This is con rmed by stability in incidence rate and evolution of surgical correction techniques. There is paucity of data regarding optimal length of xation in the literature and this moved the authors to retrospectively review own results.

Objective
We aimed at reviewing results of surgical correction of spinal deformity due to hemivertebra with regard to the type and length of instrumentation and show the correlation of the type and length of xation with the age of children.

Design of the study
Retrospective cohort study with intracohort intergroup comparison; level of evidence: 4 (Case series and poor quality cohort and case-control studies, UK Oxford, v.2009) [16] .
Inclusion criteria: Patients with congenital spinal deformity with a leading hemivertebra, not previously operated on up to the age of 18.
Exclusion criteria: Age over 18 years old, previously performed surgery to correct spinal deformity. Multiple congenital spinal deformities for more than 6 segments.

Methods
The study was based on the results of examination and treatment of 117 patients aged from 14 months to 18 years (mean age 83.7 ± 57.5 months) having different variants of hemivertebra and asymmetric types of vertebral fusion.
Surgical intervention was achieved from posterior approach in all cases and comprised hemivertebra resection from pedicular approach and instrumented xation. Patients were subdivided into 4 groups depending on type of xation ( Fig. 1): Group I -unilateral monosegmental xation; Group II -bilateral mono-segmental xation; Group III -bilateral three-segmental xation; Group IV -bilateral polysegmental xation.
Types of vertebral malformations, defects of vertebra formation, and variants of segmentation are presented for each of the groups (Table 1) Long-term follow-ups were available in all the cases from 12 months to 8 years.

Statistical analysis
Statistical analysis was performed using Statistical Package for the Social Sciences (SPSS), version 22.0 (SPSS Inc., Chicago, IL, USA). All evaluation parameters had normal distribution. Independent t test was used for contrasting means in two study groups: scoliotic and kyphotic deformity before and after surgery, operation time; and blood loss. A p value < 0.05 was considered statistically signi cant.
Neurologic status was preoperatively classi ed as Frankel E level in all the patients. No intra-and postoperative neurologic complications were reported. Complications detected in 5 patients of the group (30%) included progression of the curve that required re-operation in three cases at 3-year follow-up with removal of residual hemivertebra (n=1), the crankshaft phenomenon (n=1), adding-on in lumbar spine (n=1) and progression of deformation outside the zone of instrumental xation in the remaining patients (n=2).
Speci c feature with the group included progressive curve extending outside instrumentation segments due to the second hemivertebra in case of multiple vertebral malformations (n=2).
The second surgical intervention comprised either hemivertebra excision or increased length of xation in these patients.
Neurologic status was preoperatively classi ed as Frankel E level in all the patients. No intra-and postoperative neurologic complications were reported.
Complications were detected in 7 patients of the group (29%) and included necrosis of the wound edges (n=1), unstable screws (n=2, one with pedicle fracture) followed by revision surgery; broken screw at 2-year follow-up (n=1 with radicular syndrome) that entailed reassembly of metal construct; progressive curve at the site of instrumentation xation that resulted in PJK (n=3) ; they had intervention which included bracing in one and re-operation in two.
Speci c features with the group included progressive curve above instrumented xation level due to the second hemivertebra in cases of multiple vertebral malformations (n=3) that was treated at the second stage by hemivertebra excision or increased length of xation at one-to-three-year follow-up. A screw appeared to cut through the bone intra-operatively being unstable (n=1) due to evidently dysplastic pedicle with xation increased to one level and was thereafter stable at the xation site (n=1).
Two patients developed neurological de cit postoperatively: 1. , 4-year-and-5-month-old. Distal low paraparesis without progression in postoperative period, Frankel type D. Speci c features with the group included intra-operative instability with a screw cutting through C6 vertebra and being replaced with a hook. Progressive curve resulted from a poor quality brace in one patient.
Speci c features. One patient died 5 years after surgery, and the cause of death was unknown.
Design of the table is intended as an illustrative tool for comparison between the groups showing the difference in absolute values and statistical signi cance of the variables.

Discussion
Surgical correction is indicated in patients with congenital spinal deformities early before deformity progression and development of compensatory curves prior to the rst growth spurt at the age of 1 to 5 years. [17,18,19] (The curve in Group IV was signi cantly greater than that in Groups I and II with Р=0.003 and Р=0.010). Volume of operative intervention is considered with regard to patient's age, pattern of underlying disorder, 'maturation' of bone structures, anatomy of malformation, abnormality of adjacent vertebrae, spread of major arch and structural properties of compensatory curves.
Unilateral monosegmental xation is no longer used for children because it has not demonstrated signi cant advantages in the volume of deformity correction (no possibility of distractional maneuver on the contralateral side of the curve), operating time and blood loss compared to those provided by other xation techniques. The method has shown worse outcome at long-term follow-up as compared to those secured with other techniques (loss of correction was 4.6° on average). [20,21] Radical pedicular hemivertebra excision within the range of endplates of adjacent vertebrae (PVCR) and bilateral monosegmental transpedicular xation with acute deformity correction was shown to be effective among correction techniques employed for preschool children aged from 1 to 3 years with single, two or more hemivertebrae located between two segments in addition to accompanying spine pathology that requires no surgical correction, with local major curve without structural compensatory curves. [22] The practice provides comparable results of deformity correction of 27.2° of baseline curve, signi cant reduction in blood loss (215.4 mL) in comparison to group III (342.7 mL Р=0.018) and group IV (549.3 mL Р 0.000), minimal operating time of 160.5 minutes in comparison to group III (200.6 min. Р=0.077) and group IV (288.8 mL Р 0.000) with good long-term follow-up (loss of correction 2.2°). A staged surgical procedure can be performed in two or more hemivertebrae with an interval of 3 to 6 months between the stages.
Congenital kyphosis or kyphoscoliosis in early and middle childhood (6-7 to 10-12 years old) is characterized by structural changes in the apical vertebrae with decreased anterior-posterior size, increased wedging at the apex of the curve and agenesis or hypokinetic vertebral arches. [23,24] Surgical treatment is indicated for the patients including radical hemivertebra excision within the range of endplates of adjacent vertebral arches combined with posterior bilateral pedicular xation two segments cranial and caudal to the apex of the curve and acute correction of the physiological curves (three segmental xation). [22,25,26,27] The method provided excellent results of deformity correction with comparable blood loss and operating time, good long-term follow-up (loss of correction for scoliosis 0.6° and for kyphosis -2.1°). [22,25,26,27] More aggressive technique is indicated for severe deformity of adjacent segments employing vertebrectomy and resection of adjacent segments PSO or VCR types (Schwab grade 4 and 5) and repair of defect of anterior column with mesh implant that constitutes the supporting point for reclination maneuver during deformity correction. [28] Radical excision of the leading anomaly or multiple anomalies and polysegmental transpedicular screw xation is practical for middle childhood (7 to 11 years old) and adolescence (12 to 17-18 years old) with extensive deformity and structural compensatory curves but the procedure is associated with greater blood loss and operating time. [28] The choice of the xation method and length is determined by the patient's age, the magnitude of the deformity, and accompanying anomalies in the development of the vertebrae.
Monosegmental bilateral xation with extirpation of an abnormal vertebra can be used in children under 3 years of age. Monosegmental unilateral xation is not justi ed because of the comparable blood loss and the timing of the operation; the scoliotic arch clearly progresses in the long-term period.
At the age of 3-7 years, with a kyphotic component and an aggressive course of scoliotic deformity, the use of volumetric vertebrotomy and bilateral threesegment xation for deformity correction is justi ed, but it increases blood loss and the duration of the operation.
The use of bilateral polysegmental xation is effective in children over 7 years of age with extensive deformities. However, this increases blood loss and surgery time. The data are not nal and require further monitoring until the end of the growth period.
Compliance with ethical standards/Con ict of interest: The authors declare that they have no con ict of interest.

Declarations
Ethics approval and Consent to participate -The study was approved by the local ethics committee of the Ilizarov Scienti c Center, Kurgan, Russian Federation and consent to participate obtained from the parent/guardian of patients less than 16 years. The ethics committee approval attached as le in the submission.
Consent for publication -A copy of the consent is attached as le.
Availability of data and materials -The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Compelling interest -The authors declare that they have no competing interestsF. Funding -No funding received from any body/organization.
Authors' contribution -RSO:-conception and design; data acquisition; analysis of data; drafting of manuscript; administrative support; supervision.