Efficacy and Surgical Outcomes of Posterior Apical Spinal Osteotomy in Severe Thoracic/Thoracolumbar/Lumbar Kyphoscoliosis in Dystrophic Curves of Neurofibromatosis Type 1

Article information

J Minim Invasive Spine Surg Tech. 2024;9(Suppl 2):S172-S184

Publication date (electronic) : 2024 July 31

doi : https://doi.org/10.21182/jmisst.2024.01235

Ashwinkumar Vasant Khandge^,1, Jenil Patel², Vishnu Vikraman Nair², Jwalant Patel³, Vishal Kundnani⁴, Sunil Chodavadiya²

¹Department of Orthopaedics, Dr D Y PATIL Medical College, Hospital and Research Centre, Pune, India

²Department of Orthopaedics, Bombay Hospital & Medical Research Centre, Marine Lines, Mumbai, India

³GCS Medical College and Hospital, Ahmedabad, India

⁴Mumbai Institute of Spine Surgery, Bombay Hospital & Medical Research Centre, Marine Lines, Mumbai, India

Corresponding Author: Ashwinkumar Vasant Khandge Department of Orthopaedics, Dr D Y PATIL Medical College, Hospital and Research Centre, Pune 411018, India Email: ashwin2087@gmail.com

Received 2024 March 3; Revised 2024 April 2; Accepted 2024 April 10.

Abstract

INTRODUCTION

Neurofibromatosis (NF) is one of the commonest autosomal dominant single gene disorders occurring with frequency of 1:3000 [1]. NF-1 (type 1) commonly presents with orthopedic manifestations—kyphoscoliosis being the most frequent of them with an incidence of 10%–60% [1,2]. Two types of spinal deformity curves are encountered in NF-1: nondystrophic and dystrophic. The nondystrophic curves behave like idiopathic scoliosis and are amenable to casting, bracing and surgical procedures akin to idiopathic curves. Dystrophic curves have a characteristic of severe, acute, short and rigid curves with the dystrophic signs like vertebral scalloping, penciling of ribs, spindled transverse process, foraminal widening and wedging of the apical vertebra, often requiring major corrective surgical procedures including osteotomies of spine [3-5]. Due to the high complexity of the dystrophic curves, the patients are at a risk of developing cardiopulmonary insufficiency, neurodeficits and have a comparatively worse prognosis [5]. The decision regarding operative management of these curves was based on magnitude of the curve, location of the apex, curve flexibility and the coronal and sagittal Cobb angle. Anterior release and posterior corrective surgery, posterior corrective surgery with in situ fusion and posterior corrective surgery with osteotomy are the options that have been evaluated and studied with no obvious consensus [6]. These techniques use combined one stage or staged anterior and posterior procedures, but are often associated with increased morbidity, blood loss, surgical time and a very steep learning curve accompanied by increased hospital stay and costs. Neurological deficits, delayed implant failure with loss of correction and pseudoarthrosis are frightful complications in dystrophic curve surgery by all the options described in literature. All posterior single stage pedicle screw fixation with corrective osteotomy in dystrophic NF-1 curves is a favorable alternative reported by few authors in the literature [2,6–9]. However, the literature evaluating the clinicalradiological efficacy and surgical outcome of single stage posterior corrective surgery is scanty.

The objective of this study was to evaluate the clinicalradiological outcomes of an apical spinal osteotomy (ASO) and to assess the surgical efficacy in thoracic/thoracolumbar/lumbar deformities with dystrophic curves in NF-1.

MATERIALS AND METHODS

This study was a prospective analysis of cases with dystrophic NF-1 curves operated at single center from 2009 to 2017. After approval from the hospital institutional ethical and review board (BH-EC-0779), the study conducted on 26 consecutive patients and data collected over a period of 24 months. Written informed consent and approval wear obtained from the parents/guardians A diagnosis of NF-1 made according to the criteria of Consensus Development Conference of the National Institute of Health on NF-1. X-rays, computed tomography (CT) scans and magnetic resonance imaging were assessed for dystrophic features and patients with 3 or more of the dystrophic features were included in the study [1,3].

3. Surgical Technique

Under general anesthesia the patient was placed in prone position on vertical bolsters, the apex of the deformity placed over a hinge of the operation table. All cases were performed under intra operative somatosensory evoked potential, motor evoked potential, and electromyography neuromonitoring as per standard protocols-taking baseline values and noting any signal changes intraoperatively and taking last signal after turning the patient postoperatively before extubation. A posterior midline incision and subperiosteal dissection performed from the upper and lower planned instrumental levels. Facets included in the fusion levels were decorticated to promote arthrodesis. Titanium pediatric pedicle screws (3.5–5.5 mm) and rods were inserted using free hand technique under fluoroscopy. Minimal 4 points of fixation were secured and connected with a contoured rod on one side before starting the osteotomy. Resection was carried out at the apex of the deformity utilizing the lever arm to the maximum effect. The transverse process and rib heads were removed in order to complete dissection of the lateral wall of the vertebral body. Segmental vessels in the line of dissection were ligated and cut on only one side. Nerve roots in the thoracic region were sacrificed, if deemed necessary, whereas nerve roots in the lumbar region were preserved throughout the procedure. Vertebral body resection was done using pituitary rongeurs, 4-mm cutting burr and curettes. Dissection proceeded until the anterior wall of the vertebral body was removed. The anterior longitudinal ligament, medial wall of the pedicle and posterior wall of the vertebral body was kept intact. A temporary rod was attached on the side of this osteotomy and same steps were repeated on the other side. At the end of osteotomy, a thin bony shell remained around the dural sac circumferentially. The posterior vertebral wall, anterior to the dural sac, the lamina on the dorsal aspect of the dura and the medial pedicle wall were removed. This step is performed once appropriate amount of vertebral body is excised. This helps in reducing blood loss from epidural vessels and protects the dural sac from injury. The posterior cortex of the vertebral body was initially thinned out with a diamond burr and was then removed (pushed down) with the help of reverse angle curettes. Deformity correction was attempted by closure of the osteotomy with sequential rod contouring. Gradual correction of the kyphosis angle was done at every sequence of rod bending, with extension of the operating table if needed. While correcting scoliosis, compression and shortening was done more on the convex side and was asymmetrical. The end point of column shortening was visualization of buckling of the dura or any neuromonitoring changes or more than 15 mm of closing of the wedge on the convex side of the osteotomy. The number of levels to be fused was decided intraoperatively depending upon the extent of correction achieved and the gap present anteriorly after osteotomy. The anterior vertebral gaps after osteotomy if less than 10 mm with sufficient bone on bone contact, only 2 levels above and below were fused. If the residual bony defect after closing the osteotomy was less than 10 mm, auto graft was placed. If the osteotomy gap was more than 10 mm, a titanium mesh cage with bone graft was placed. Additional compression was applied on the cage to lock it into place and maintain optimum correction. Wound closure performed in layers over a negative suction drain. Patients were given a customized thoracolumbar sacral orthosis brace for 6 weeks in the postoperative period and were allowed to sit in the bed or allowed to mobilize on the second postoperative day. Figures 1 and 2 demonstrate the case examples from this series including their preoperative clinicalradiological photos, intraoperative images and postoperative correction clinicalradiological images.

Figure 1.

A 13-year-old boy with neurofibromatosis type 1 and dystrophic kyphoscoliosis with right-sided plexiform neurofibroma. (A) The back. (B) Bending forward demonstrating the curve rotation. (C, D) The right and left side, respectively. (E) An axial computed tomography scan at the T8 level demonstrating rib penetration and thin and atrophic pedicles. (F) Postoperative x-ray demonstrating fixation from T4 to L2 with correction of the curve. (G) A final clinical picture from the back and side demonstrating well-balanced shoulders and sagittal profile.

Figure 2.

A 14-year-old boy with thoracolumbar dystrophic neurofibromatosis type 1. (A) A preoperative anteroposterior radiograph showing scoliosis (Cobb angle of 49.4°). (B) A preoperative lateral radiograph showing a kyphotic Cobb angle of 82.7°. (C–E) a preoperative computed tomography (CT) scan showing the curve morphology in coronal and sagittal cuts and dystrophic features seen on axial cuts. (F, G) Sagittal and axial magnetic resonance imaging cuts showing neurofibroma at the curve. (H) Three-dimensional reconstruction of the curve through CT images. (I, J) clinical images of the patient from the back and side. (K) Intraoperative images of apical spinal osteotomy (ASO) performed at the T11 level showing the cord free from all the sides. (L–N) intraoperative C-arm images taken after pedicle crew placement, after ASO and applying a corrective rod. (O) immediate postoperative lateral image of the patient showing visible correction of kyphosis. (P) postoperative anteroposterior and lateral radiographs showing a corrected coronal Cobb angle of 6.4° and a sagittal Cobb angle of 39.8°.

A total 21 cases satisfied the inclusion criteria and the data collected from over a period of 24 months. After an informed/written consent and approval from the parents/guardians, the clinical parameters were recorded preoperatively, postoperatively, and then subsequently at 3 months, 6 months, 1 year, 2 years, and at latest follow-up. The patients were evaluated by the operating surgeon and a spine fellow. Demographics parameters like age, sex, height, weight, pain-visual analogue scale (VAS), Oswestry Disability Index (ODI), neurology grading (Frankel grade) and ambulatory status were collected. Full-length standing spine x-ray anterior-posterior and lateral were evaluated for the level of apical vertebra, coronal, and sagittal Cobb angle and sagittal vertical axis (SVA) preoperatively, postoperatively, and at subsequent follow-ups. Intraoperative surgical parameters like surgical time, blood loss, dural leak, hospital stay, fusion levels, number of levels instrumented were documented. Postoperative clinicalradiological outcome and complications like neurological worsening, cardiopulmonary issues and implant failure were noted at every follow-up intervals. The fusion was assessed with the postoperative CT scan at one near follow-up and then every 6 months interval. The statistical analysis was conducted using IBM SPSS Statistics ver. 20.0 (IBM Co., Armonk, NY, USA). VAS, ODI, Frankel grade, and deformity angles were statistically analyzed by paired t-test pre- and postoperatively and at final follow-up. A p-value < 0.0001 was considered significant.

RESULTS

Twenty-one patients with dystrophic NF-1 curves were included in this study. The mean age at surgery was 9.3±2.3 years (Table 1). There were no intraspinal anomalies but 2 patients had plexiform neurofibroma near the curve. The mean intraoperative blood loss was 364.5±89.0 mL and mean operative time was 256.6±67.64 minutes (Table 1). Osteotomies were performed from levels T3 to L1 depending on the apex of kyphosis, most commonly at the T3 levels (in 4 of 21 patients). Osteotomy was done at T10 or above in 17 (80.95%) cases and at T11 or below in 4 of 21 patients (19.04%). The mean number of levels fused was 5.23±1.17 (range, 4–8 levels). The mean preoperative kyphotic Cobb angle improved significantly from 98.33°±10.87° to 36.52°±8.05° (p<0.005) (Table 2). The mean scoliotic Cobb angle also improved significantly from preoperative angle of 54.5°±9.61° to 24.0°±5.39° at final follow-up (p<0.005) (Table 2). The mean loss of correction for kyphotic Cobb angle was 7.0°±2.04° and for scoliotic Cobb angle was 4.0°±1.30° at final follow-up. The mean percentage of final correction for kyphotic Cobb angle was 62.85°±3.71° and for scoliotic Cobb angle was 55.59°±6.84°. The mean SVA also improved significantly from 7.40±1.34 cm to 4.20±1.23 cm (p<0.005) (Table 2). The mean VAS score of patients improved significantly from preoperative value of 6.19±1.80 to 3.62±1.49 in postoperative period and was 1.09±1.20 at final follow-up (p<0.005) (Table 3). The mean ODI score was 54.85±13.67 preoperatively and had decreased significantly to 22.1±4.69 in postoperative period and was 18.35±3.21 at final follow-up (p<0.005) (Table 3). In this study, the overall complication rate was 30.77% that included 2 cases of dural tear and 1 case each of implant failure and superficial infection (Table 4). The neurological complication rate was 14.2%. One patient had a new-onset neurological deficit involving the left L4 nerve root (quadriceps power from preoperative 5/5 to 3/5). The patient was managed by screw revision following a postoperative CT scan evaluation, which showed inferior screw breach of the L4 pedicle. The patient improved to power 4/5 at 3 months follow-up and 5/5 at the final follow-up. In 2 other patients, the neurological status worsened from preoperative Frankel grade C to A in the immediate postoperative period. In these 2 patients, neuromonitoring signals were lost without any specific maneuver. They were intraoperatively managed by using reversal mechanisms like raising the blood pressure, stopping the paralytic anesthetic agent and warm saline irrigation. One of the 2 cases was managed conservatively and regained Frankel grade C power after 6 months follow-up. Whereas 1 patient developed complete paraplegia (ASO at the L1 level) that resolved only partially by the final follow-up evaluation (Frankel grade B). The patient was managed with aggressive physiotherapy and bracing. One patient had pseudoarthrosis with rod breakage and deterioration of neurological status was managed by reoperation and rod change with full neurology in postoperative period.

Table 1.

Demographic and clinical data

Table 2.

Clinical and radiological parameters

Table 3.

Clinical and neurological evaluation

Table 4.

Fusion rate (Bridwell criteria) and complications

DISCUSSION

Thoracic deformity being the most common orthopedic manifestation of NF-1 which may pose significant challenges including osteoporotic bones, narrow pedicles, rigidity of the curves and postoperative complications like pseudoarthrosis and rod breakage [2,3,10-16]. Surgical management of dystrophic NF-1 curves involves anterior with posterior surgery or only posterior surgery [9,12]. The 2-staged surgeries are associated with increased morbidities like increased intraoperative blood loss and risk of postoperative pulmonary dysfunction in cases where thoracotomy is performed [17-20]. On the other hand, posterior approaches could achieve acceptable correction through osteotomies like posterior vertebral column resection (PVCR) and ASO with relatively lesser postoperative morbidities [15,21,22]. Relatively short posterior fixations with minimum levels spanned after the procedure that aids in retaining the postoperative longitudinal spinal growth and chest cavity growth [9,15,21]. ASO has been previously described for rigid kyphoscoliosis and its use in dystrophic NF-1 curves is hardly reported in literature [15,21]. In this study of 21 patients of dystrophic NF-1 curves, the curve correction improved from 98.33° to 36.52° for kyphosis and from 54.51° to 24.0° for scoliosis at final follow-up. The excellent fusion rates and correction achieved indicates that ASO procedure can be an answer to this enigmatic question of dystrophic curves that has haunted spine surgeons for generations. For large stiff curves, a combined anterior release and fusion followed by posterior spinal fusion and instrumentation has often been performed [20]. Both open and endoscopic approaches have shown a negative impact on pulmonary function than the posterior-only approach [19,20]. With the advent of modern pedicle screw instrumentation system, a single posterior approach is preferable to reduce morbidity to the patients. Closing wedge osteotomies such as Smith-Peterson osteotomy (SPO), pedicle subtraction osteotomy (PSO), and vertebral column resection (VCR) can provide acceptable clinicalradiological results [15,21-24]. Suk et al. [21] reported on a posterior-only approach with a VCR for fixed lumbar spinal deformities [15] as well as for severe rigid scoliosis. These authors reported excellent surgical correction with minimal long-term complications. In this study, surgical technique has allowed correction of severe rigid deformities with the help of single-level apical osteotomy for all patients. The benefits of performing an apical osteotomy are efficient correction at the very site of rigid deformity, the potential to correct deformity in both sagittal and coronal planes, and direct bone-to-bone contact instead of spanning reconstruction, giving better fusion rates and fewer implant-related complications. Deng et al. [25] in their series of 31 patients with dystrophic NF-1 using one stage posterior multiple anchor point method reported correction of scoliosis from mean Cobb angle of 69.1° to 30.2° and that of kyphosis from a mean Cobb angle of 58.3° to 24.1°. Wang and Lenke [26] in their paper discussed 16 patients (8 males, 8 females) with dystrophic NF-1 curves with a mean age of 13 years and all operated through a single stage all posterior pedicle screw fixation with PVCR. They reported a comparable surgical time (mean: 309.7 minutes vs. 256.66 minutes in this study). The blood loss in their series was 1,393.8 mL (range, 300–4,400 mL) as compared to this study (364 mL). This difference may be due to the procedure of ASO where we keep the medial wall of the pedicles and posterior wall of the vertebral body intact before removing the vertebral body in piecemeal fashion. The intact medial pedicle wall may help in keeping the integrity of the epidural vessels intact and minimizing the intraoperative blood loss, hence this difference. Li et al. [2] in their series describe 41 patients (mean age, 13 years) with dystrophic NF-1 curves operated via a 1-stage all posterior surgery using pedicle screw fixation. The mean follow-up of 28.8 months and the deformity improved 50% at the final follow-up. Their mean levels fused were 11.2 as compared to this study of 5.23. The less number of levels fused in this study contributed to less morbidity and earlier surgical recovery and blood loss. The ASO done in this study contributed to more final deformity correction at the final follow-up as compared to their series. Chen et al. [27], Shimode et al. [28], Bakaloudis et al. [29], Lenke et al. [30], and Wang et al. [31] studied rigid kyphoscoliosis of varied etiologies and have documented the efficacy of thoracic PSO, VCR, and apical resection procedures with final corrections in the range of 54%–69%. The mean age of the patients in these studies was in the range of 12–34 years. Similarly, Lenke et al. [30] reported the results of VCR in 35 pediatric patients with a mean follow-up of 2 years. The mean age of the patients in their series was 11 years (range, 2–18 years). The patients in that case series underwent either single-level or multilevel osteotomies, and the authors documented overall correction rates of 51%–60%, with an average blood loss of 691 mL. In our series, the patients’ mean age was lower (9 years; range, 6–14 years), all of the patients underwent single-level ASO for dystrophic NF-1 curves and correction rates of 62.85% and 55.59% were achieved for kyphotic and scoliotic curves, respectively. Various studies on the treatment of rigid kyphoscoliosis with apical osteotomy, Smith Peterson wedge osteotomy (SWO), multilevel modified VCR, and Closing opening wedge osteotomy (COWO) have documented blood loss in the range from 717–3,340 mL [31-35]. In correction of high-magnitude rigid curves, similar correction rates have been reported for both conventional (SPO, PSO, and VCR) and newer unconventional osteotomies (apical resection osteotomy, SWO, and COWO). However, conventional osteotomies may require correction at multiple levels, which increases the morbidity of surgery. Bakaloudis et al. [29] reported an overall correction of 65% with PSO in pediatric patients. Chen et al. [27] documented correction of 69.87% with apical resection osteotomy. In this series, the mean kyphotic curve was 98.34° preoperatively and 36.52° at final follow-up, indicating an average final correction of 62.85%, which remained essentially consistent at the end of 2 years (the change of 7° between postoperative and final measurements was not clinically significant). Some authors [7,15,17,36], believe that dystrophic scoliosis progresses despite spinal arthrodesis. These conclusions were mainly based on the hook-rod based techniques. Present study revealed no progression of deformity with the 3-column pedicle screw system. This suggests that the progression of deformity should mainly be attributed to limitations of the fixation, correction and fusion techniques and theories rather than inadequate fusion. After performing PSO in 27 patients with fixed sagittal imbalance, Lenke et al. [30] noted that the mean ODI score improved from 51.21 preoperatively to 35.75 at the last follow-up evaluation, and 92.3% of the patients were found to be satisfied with the treatment according to the overall satisfaction score. Corresponding to the aforementioned results, this study did demonstrate a significant decrease in the mean ODI score after surgery (from, 54.85 to 22.1) and good functional improvement at the final follow-up evaluation. Bakaloudis et al. [29] reported 5 medical complications and 1 neurological complication in their series of 12 patients for a complication rate of 50%. Chen et al. [28] documented complications in 10 of 23 patients. Lenke et al. [30] reported complications in only 4 of 35 patients in their series of pediatric deformity patients operated on with VCR. However, an alarming rate of complications associated with VCR in pediatric patients was noted in a multicenter trial that included 147 patients [34]. The authors observed complications in 59% of these patients. Sixty-eight patients had a complication during surgery (the most common being changes in spinal cord monitoring data and blood loss >2 L), and 43 had a postoperative complication (most often respiratory related). No patient in his series had a complete permanent neurological deficit. In this series, the overall complication rate was 30.77%. There was a general complication rate of 15%, which included 2 cases of dural tear and 1 case each of implant failure and superficial infection [37–43]. The neurological complication rate was 7.6% (Table 4).

The major limitation of this study is its retrospective design and short follow-up; however, the database for this study was constructed without any hypothesis, and the data were thus collected free from any bias related to the study. The larger group of cases with similar etiology and longer follow-up would allow for more observation of the pitfalls of long-segment fusion in the pediatric population, and the effect of residual deformity resulting in mechanical back pain, implant loosening, and loss of correction can be evaluated better. Another limitation of the study was the use of the ODI as an outcome measure, although it has not been validated in the pediatric age group. The ODI was used because there were no validated pediatric outcome measures available in the local language at the time of data collection. Moreover, the lack of comparison with anterior approach or combined approach in this study could be one of the limiting factors.

CONCLUSION

In this study, the authors described a technique of posterior approach single-level ASO for the treatment of severe dystrophic NF-1 curves. The technique can be performed in children with mild to moderate curves with good clinicoradiological outcomes and satisfactory correction rates. Although the surgeon’s familiarity with the dystrophic curves and technique is the key parameter. Its versatility and effectiveness for correction of even rigid deformities in children could be further explored for other etiologies and through multicenter studies.

Article information Continued