The Comparison of Posterior Intervertebral Release Combined with Posterior Column Osteotomy and Posterior Column Osteotomy Alone for the Treatment of Moderate‐to‐Severe Rigid Scoliosis: A Prospective Controlled Study

Abstract

Objectives

There is no consensus on the treatment of moderate‐to‐severe rigid scoliosis. Anterior release and three‐column osteotomy are excessively traumatic, whereas posterior column osteotomy (PCO) alone results in poor outcomes. An emerging surgical technique, posterior intervertebral release (PR), can release the rigid spine from the posterior approach. This study was performed to compare the multi‐segment apical convex PR combined with PCO and PCO alone in patients with moderate‐to‐severe rigid scoliosis.

Methods

From June 2021 to June 2022, this prospective study of moderate‐to‐severe (Cobb: 70–90°) rigid scoliosis (flexibility of main curve <25%) involved two groups defined by surgical procedure: the PR group, the patients undergoing PR combined with PCO; and the PCO group, the patients undergoing PCO alone. Follow‐up was at least 12 months. Radiographic results mainly included main curve Cobb, correction of per PR/PCO segment, apical vertebra rotation (AVR) and apical vertebra translation (AVT). Demographics, surgical data, complications were also recorded. Student's independent samples t test and Pearson's chi‐square test were used to compare the differences between groups.

Results

Forty patients with an average age of 16.65 years were included (PR group, n = 20; PCO group, n = 20). The main curves averaged 77.56° ± 5.86° versus 78.02° ± 5.72° preoperatively and 20.07° ± 6.73° versus 33.58° ± 5.76° (p < 0.001) at the last follow‐up in the PR and PCO groups, respectively. The mean correction rates were 74.30% and 56.84%, respectively (p < 0.001). The average coronal curve correction was 13.49° per release segment, which was significantly higher than the PCO correction of 6.20° (p < 0.001). The correction of apical vertebra rotation and translation in the main thoracic curve was significantly better in the PR group than in the PCO group (p < 0.05). Several minor complications in the two groups improved after conservative treatment.

Conclusion

The multi‐segment apical convex PR combined with PCO offers more advantages than PCO alone in the treatment of patients with moderate‐to‐severe rigid scoliosis. Owing to its excellent corrective effect and few complications, this is a high benefit–risk ratio surgical strategy for rigid scoliosis.

The study compared the multi‐segment apical convex intervertebral release (PR) combined with posterior column osteotomy (PCO) with PCO alone in patients with moderate‐to‐severe rigid scoliosis. The multi‐segment apical convex PR combined with PCO offers better corrective outcomes than PCO alone in the treatment of patients with moderate‐to‐severe rigid scoliosis. Owing to its excellent corrective effect and few complications, this is a high benefit–risk ratio surgical strategy for rigid scoliosis.

Introduction

Surgical correction of rigid scoliosis has been a considerable challenge for spine surgeons. With the development of spinal osteotomy and pedicle screws, posterior‐only approaches have received increasing attention in the area of corrective surgery for the treatment of rigid scoliosis. ¹ , ² , ³ Three‐column osteotomy, pedicle subtraction osteotomy (PSO), and vertebral column resection (VCR) are commonly used for severe (cobb >90°) and rigid scoliosis with excellent corrective results. ⁴ , ⁵ Although the three‐column osteotomy can provide satisfactory correction in the coronal and sagittal planes, the technique exposes the patients to a great risk of neurological injury, long operative time, massive hemorrhage, and potential morbidity. ⁶ Three‐column osteotomy in patients with moderate‐to‐severe (Cobb: 70–90°) rigid curves is not an ideal procedure due to the low risk–benefit ratio. Posterior column osteotomy (PCO) is a relatively safe and versatile technique for scoliosis. In a biomechanical experiment modeled on a cadaver, PCO increased the range of motion in flexion, extension, and axial rotation but not in lateral bending. ⁷ Therefore, the correction outcomes of posterior column osteotomy alone could be poor for rigid scoliosis, and surgical procedures with high benefit and low risk for patients with moderate‐to‐severe rigid curves deserve further discussion.

Intervertebral discectomy in the apical region is essential for adequate release of the rigid spine and achieves three‐dimensional release in axial rotation, anterior–posterior, and coronal translation more notably than posterior osteotomy. ⁸ Considering the complications of anterior discectomy, posterior intervertebral discectomy has been reported in the previous literature. ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ Three of the studies involved rigid scoliosis and suggested that posterior intervertebral discectomy combined with PCO can achieve a good correction rate for rigid scoliosis. ¹⁰ , ¹¹ , ¹⁵ However, these reports were mainly presented as technical notes, with a small sample size, and included mild and severe scoliosis. At the same time, no comparison was performed with PCO, and the benefit of intervertebral release compared to PCO remains unclear. Moreover, a biomechanical study showed that releasing the costovertebral joint can reduce the force required for axial rotation and lateral bending while increasing the displacement of vertebral bodies. ¹⁶ This procedure of posterior discectomy may make it possible to release the costovertebral joint concomitantly, and it may further facilitate segmental correction in rigid scoliosis.

From this, we applied the posterior intervertebral release (PR), including intervertebral discectomy and costovertebral joint release, combined with PCO to the treatment of moderate‐to‐severe rigid scoliosis. Our hypothesis is that the surgical approach is a high‐benefit and low‐risk technique for patients with moderate‐to‐severe rigid curves. And we speculate that PR combined with PCO can release the rigid spine and achieve better corrective outcomes than PCO alone. Therefore, the aim of this study was to investigate the following: (i) the safety and efficacy of PR combined with PCO for treatment of moderate‐to‐severe rigid scoliosis; (ii) the benefit of PR combined with PCO compared to PCO alone for treatment of moderate‐to‐severe rigid scoliosis; (iii) review the surgical strategies of moderate‐to‐severe rigid scoliosis and summarize the advantages and technical experience of PR.

Materials and Methods

This prospective study was registered as a clinical trial (ChiCTR2100017565) and approved by the Clinical Trial Ethics Committee of our hospital (No. 2021206). On the basis of previous studies, ¹⁰ , ¹¹ , ¹⁵ , ¹⁷ , ¹⁸ , ¹⁹ the corrective rate was estimated to be 70% ± 10% in the PR group and 60% ± 10% in the PCO group. A power analysis was conducted to estimate the minimum sample size needed to assess the significant difference between the two groups. The type I error was set at 0.05 (α < 0.05) and type II error at 0.2 (80% power). The minimum sample size of 17 subjects in each group would be required. The inclusion criteria were as follows: (1) main thoracic/thoracolumbar curve between 70° and 90° (moderate‐to‐severe scoliosis); (2) flexibility of main curve <25%; (3) age >12 years and Risser sign ≥1; The exclusion criteria were as follows: (1) Cobb <70° or Cobb >90° curves; and (2) congenital scoliosis and the formation of bone bridge on anterior spinal column (which may involve different osteotomy). All patients provided written informed consent prior to enrolment (Figure 1). Patients were assigned according to the choice of surgical technique into the PR group (those who underwent PR combined with PCO) and the PCO group (those who underwent PCO alone).

FIGURE 1.

The flow diagram of patients through the study.

Surgical Procedure

All surgeries were carried out by the same surgical team, and the main correction maneuverer was performed by the senior surgeon (Y. S.). After general anesthesia and positioning, exposure of the posterior elements of the target segments was reached via subperiosteal dissection. All patients underwent Smith–Petersen osteotomy (SPO) or Ponte osteotomy in which the inferior facet process was resected in all segment and both superior and inferior facet processes were resected to expose the interpedicular foramen around the apical vertebrae. For the PR group, the intervertebral discs at the convex side of the apical vertebrae were carefully exposed (Figure 2). Then a sharp knife was used to incise the annulus of the disc and a reamer was inserted in the intervertebral space to remove the disc material. The endplate of the adjacent vertebral was removed. To guarantee a thorough release of the disc space, the lateral annulus of the intervertebral disc was cut off and the rib head bridging the adjacent vertebral bodies should be removed. The procedure was similar to that of the PR group except that intervertebral release was not performed. Subsequently, routine correction maneuverer was performed to correct the deformity, included distraction of the concave side, compression of the convex side, and de‐rotation of the vertebra by twisting the monopolar screws. The intraoperative photos of the PR group were shown in Figure 3. The incision was closed layer by layer after placing a drainage tube alongside the wound. The transcranial electric motor and sensory evoked potential monitoring were performed in each surgery.

FIGURE 2.

Illustration of the surgical approach. The facet processes and part of the lamina were resected to expose the intervertebral disc. Removal of the disc and rib head to complete intervertebral release. (A) The dotted line shows the extent of the osteotomy. (B) The white arrow shows surgical access for discectomy.

FIGURE 3.

(A) After Ponte osteotomy, the convex intervertebral disc was exposed, and posterior intervertebral release (PR) (white arrows) was performed (T6–7, 7–8, 8–9). (B) After completing the release, placing the screws, rods, and compression, we closed the convex gaps adequately, and the scoliosis was satisfactorily corrected (white ellipse).

Demographic and Surgical Data

An extensive record and review of demographic, surgical, and complication data was performed using the Electronic Medical Records System. Information regarding age at surgery, sex, diagnosis, follow‐up time, body mass index (BMI), implant density, osteotomy segment, operative time, estimated blood loss (EBL), perioperative red blood cell transfusion, length of hospitalization, and complications were recorded.

Radiographic Measurements

Radiographs of the whole spine were taken preoperatively, 1‐week postoperatively, and after at least 12 months of follow‐up. Measurements were obtained from the picture archiving and communication system (PACS). According to the Cobb angle protocol, the degrees of main curve (MC) and thoracic kyphosis (TK) were assessed. Preoperative side‐bending X‐ray films were used to assess the flexibility of the MC. The horizontal distance between the C7 plumb line and the central sacral vertical line (C7PL‐CSVL) and sagittal vertical axis (SVA) represented the coronal and sagittal balance, respectively. Apical vertebra translation (AVT) was measured as the distance from the centre of the apical vertebra to the C7PL (if the apical vertebra was in the lumbar segment, CSVL was used as the reference). Three‐dimensional computed tomography (CT) reconstruction of the thoracic and lumbar vertebrae was performed before and after surgery. Apical vertebra rotation (AVR) was measured on CT according to the method described by previous report. ²⁰ To further clarify the corrective effect of the posterior release and PCO, we measured the change in Cobb angle of the release and PCO segments on the coronal CT reconstruction images preoperatively and postoperatively and calculated the correction of each segment (the change in Cobb angle/release and PCO segments). The intra‐ and interobserver reliability and ICCs were analyzed for all main variable measurements, including curve Cobb angle and AVR. The ICC of intra‐observer reliability was 0.982–0.993, and the ICC of interobserver reliability was 0.965–0.987.

Statistical Analysis

SPSS 23.0 (SPSS Inc., Chicago, Illinois) was used for statistical analysis. Numerical variables were expressed as the mean ± standard deviation and analyzed using unpaired Student's t test and Mann–Whitney U test, depending on the distribution of the variables. Multiple subgroups were compared using One‐way ANOVA and Kruskal–Wallis tests. Categorical variables were compared using the χ ² test and Fisher's exact test. p values <0.05 were considered significant differences.

Results

Demographic and Surgical Variables

The present study included 40 patients (19 males, 21 females) with an average age of 16.65 years (range 13–27 years). The mean duration of follow‐up was 18.3 months (range 12–30 months). There were 20 patients in each PR group and PCO group. The demographic and surgical data are shown in Table 1. The PR group had 12 idiopathic scoliosis (IS) and eight neuromuscular scoliosis (NMS), including 19 thoracic curves and one thoracolumbar curve. The flexibility of the main curve averaged 14.6% ± 6.6% (range 0–22.9). SPO and Ponte osteotomy were performed at 11.2 ± 1.1 (range, 9–13) and 3.5 ± 0.5 (range, 3–4) segments, respectively. In addition, PR was performed at 3.4 ± 0.6 (range, 2–4) segments in the apical region of the main curve. The PCO group had 13 idiopathic curves and seven neuromuscular curves, including 19 thoracic curves and one thoracolumbar curve. The flexibility of the main curve averaged 16.1% ± 4.4% (range 8.8–22.8). SPO osteotomy was performed at 11.5 ± 1.1 (range, 8–13) segments. Ponte osteotomy was performed at 5.1 ± 0.7 (range, 4–6) segments in the apical region of the main curve without PR. Apart from the differences in Ponte osteotomy and PR segments, there were no significant differences between the two groups in terms of Cobb angle of the main curve, flexibility of the main curve, age, sex, BMI, implant density, operation time, estimated blood loss, transfusion rate, length of stay, or follow‐up time(Table 1).

TABLE 1.

Demographics and surgical data.

	PR group (n = 20)		PCO group (n = 20)		t/χ 2	p
	$\overline{X}$ ± SD	Range	$\overline{X}$ ± SD	Range
Age at surgery (years)	16.9 ± 3.4	13–27	16.4 ± 3.2	13–27	0.475	0.637
Male/female	9/11		10/10		0.100	1
BMI (kg/m2)	18.8 ± 3.1	16.7–30.5	19.5 ± 2.7	15.6–26.4	−0.737	0.465
Diagnosis (IS/NMS)	12/8		13/7		0.107	1
Thoracic/Thoracolumbar curves	19/1		19/1		‐	1
Flexibility of MC (%)	14.6 ± 6.6	0–22.9	16.1 ± 4.4	8.8–22.8	−0.836	0.409
Implant density	1.4 ± 0.5	1.2–1.9	1.4 ± 0.2	1.2–1.9	0.597	0.554
SP osteotomy segments	11.2 ± 1.1	9–13	11.5 ± 1.1	8–13	−0.900	0.374
Ponte osteotomy segments	3.5 ± 0.5	3–4	6.1 ± 0.7	4–6	−13.594	<0.001
PR segments	3.4 ± 0.6	2–4	None
Operation time (min)	337.2 ± 66.2	246–471	310.3 ± 63.7	210–430	1.308	0.199
Estimated blood loss (mL)	790.0 ± 255.3	500–1500	765.0 ± 249.8	500–1400	0.313	0.756
Transfusion rate	11/20		9/20		0.400	0.752
Follow‐up duration (months)	17.8 ± 4.5	12–28	18.5 ± 5.3	12–30	−0.484	0.631
Length of stay (days)	12.3 ± 4.6	8–27	13.4 ± 4.4	8–22	−0.704	0.486

Note: The bold values represents a statistically significant difference.

Abbreviations: IS, idiopathic scoliosis; MC, main curve; NMS, Neuromuscular Scoliosis; PCO, posterior column osteotomy; PR, posterior intervertebral releases.

Comparison of Radiological Outcomes between PR and PCO

In the PR group, the preoperative main curve of 77.6° ± 5.9° (range 70°–90°) was corrected to 19.4° ± 6.6° (range 6°–30.9°) at the immediate postoperative evaluation, showing that the correction rate was 75.2% ± 7.7% (range 60.6%–91.6%). At the last follow‐up, the main curve was 20.1° ± 6.7° (range 6.5°–30.9°), showing a 74.3% ± 7.9% (range 59.9%–91.6%) scoliosis correction compared to the preoperative evaluation and a 0.9% loss of correction compared to the immediate postoperative evaluation. In the PCO group, the preoperative main curve of 78.1° ± 5.7° (range 70°–90°) was corrected to 32.7° ± 5.7° (range 27.3°–51.6°) at the immediate postoperative evaluation, showing that the correction rate was 58.0% ± 7.1% (range 37.8%–66.3%). At the last follow‐up, the main curve was 33.6° ± 5.8° (range 28.2°–52.5°), showing a 56.8% ± 7.3% (range 36.8%–64.8%) scoliosis correction compared to the preoperative evaluation and a 1.15% loss of correction compared to the immediate postoperative evaluation. Compared to patients in the PCO group, the patients in the PR group had a significantly smaller postoperative and final follow‐up main curve and a significantly higher correction rate immediately postoperative and at follow‐up. A Cobb correction of 13.5° ± 2.3° (range 8.8°–16.9°) could be obtained per segment PR, and there was a significantly higher correction than that of per segment PCO alone (6.2° ± 1.4°, range 3.6°–8.8°) (Table 2).

TABLE 2.

Radiographic results.

	PR group (n = 20)		PCO group (n = 20)		t	p
	$\overline{X}$ ± SD	Range	$\overline{X}$ ± SD	Range
Main Curve Cobb (°)
Preoperative	77.6 ± 5.9	70–90	78.1 ± 5.7	70–90	−0.250	0.804
Postoperatively	19.4 ± 6.6	6–30.9	32.7 ± 5.7	27.3–51.6	−6.827	<0.001
Post‐correction rate (%)	75.2 ± 7.7	60.6–91.6	58.0 ± 7.1	37.8–66.3	7.382	<0.001
Last follow‐up	20.1 ± 6.7	6.5–30.9	33.6 ± 5.8	28.2–52.5	−6.841	<0.001
Last‐Correction rate (%)	74.3 ± 7.9	59.9–91.6	56.8 ± 7.3	36.8–64.8	7.312	<0.001
Loss of correction rate	0.9 ± 0.9	0–3.1	1.2 ± 1.0	0–3.5	−1.115	0.348
Correction of per PR/PCO	13.5 ± 2.3	8.8–16.9	6.2 ± 1.4	3.6–8.8	12.031	<0.001
AVT (mm)
Preoperative	63.9 ± 12.6	39.7–89.7	67.2 ± 14.8	41.9–103.2	−0.757	0.454
Postoperatively	13.6 ± 10.9	0–36.9	21.1 ± 8.5	10.90–45.00	−2.448	0.019
Last follow‐up	14.0 ± 8.0	0–26.8	22.1 ± 8.1	8.9–45.7	−3.201	0.003
AVR (°)
Preoperative	23.6 ± 7.8	12.8–42.6	22.9 ± 7.5	9.7–38.1	0.302	0.764
Postoperatively	10.0 ± 5.4	3.1–25.4	14.0 ± 5.9	4.7–33.7	−2.215	0.033
Correction rate (%)	57.6 ± 16.2	28.6–82.6	39.2 ± 11.9	10.6–57.2	4.253	<0.001
C7PL‐CSVL (mm)
Preoperative	11.4 ± 7.5	0–24.4	9.4 ± 8.6	0–29.8	0.760	0.452
Postoperatively	16.3 ± 9.7	0–42.0	16.0 ± 9.4	0–34.6	0.106	0.916
Last follow‐up	13.5 ± 7.5	0–28.1	12.1 ± 7.1	0–25.0	0.592	0.557
TK (°)
Preoperative	33.0 ± 17.1	1.5–62.2	31.6 ± 15.1	1.2–49.8	0.169	0.867
Postoperatively	21.5 ± 7.3	9.8–34.7	23.2 ± 8.1	10.5–34.3	−0.688	0.495
Last follow‐up	21.4 ± 6.4	10.1–33.4	23.8 ± 7.6	10.3–37.8	−1.053	0.299
SVA (mm)
Preoperative	27.8 ± 17.6	4.1–67.9	25.5 ± 14.1	10–57.8	0.456	0.651
Postoperatively	17.7 ± 9.5	3.5–42.0	14.3 ± 13.7	0–39.9	0.902	0.373
Last follow‐up	12.2 ± 6.5	0–25.5	11.8 ± 10.3	0–34.4	0.133	0.895

Note: The bold values represents a statistically significant difference.

Abbreviations: AVR, apical vertebra rotation; AVT, apical vertebra translation; C7PL‐CSVL, the distance from the C7 plumb line to the midline of the sacrum; SVA, sagittal vertical axis; TK, thoracic kyphosis.

The preoperative AVT in the PR group and PCO group was 63.9 ± 12.6 and 67.2 ± 14.8 mm, and the preoperative AVR in the PR group and PCO group was 23.6 ± 7.8° and 22.9 ± 7.5°, with the values showing no significant difference between the two groups. The AVT and AVR significantly improved after surgery in both groups. After surgery, the AVT in the PR group was smaller than that in the PCO group (13.6 ± 10.9 mm vs. 21.1 ± 8.5 mm, p = 0.019, postoperatively and 14.0 ± 8.0 vs. 22.1 ± 8.1 mm, p = 0.003 at the last follow‐up). The postoperative AVR in the PR group and the PCO group were 10.0 ± 5.4° and 14.0 ± 5.9°, respectively (p = 0.033), and patients in the PR group achieved better de‐rotation than patients in the PCO group (correction rate: 57.6% ± 16.2% vs. 39.2% ± 11.9%, respectively, p < 0.001, Table 2). Preoperatively, immediately postoperatively and at the final follow‐up, no significant difference was found in thoracic kyphosis, C7PL‐CSVL, or the sagittal vertical axis between the patients of the two groups. Typical cases in the PR group and PCO group were shown in Figure 4 and Figure 5, respectively.

FIGURE 4.

A 17‐year‐old boy with neuromuscular scoliosis underwent posterior multisegment apical convex discectomy plus costovertebral joint release (PR) combined with Ponte osteotomies (PCO), three segments of PCO and three segments of PR. (A–C) Preoperative anteroposterior and lateral radiographs showed a 76.6° main thoracic curve, and the bending radiograph suggested a flexibility of 6.7% of the main curve. Thoracic kyphosis was 28.2°. (D, E) Postoperative anteroposterior and lateral radiographs. The main thoracic curve was corrected to 13.3°, and the thoracic kyphosis was corrected to 14.9°. The correction rate of the main curves was 82.6%. (F, G) The coronal preoperative and postoperative CT reconstruction images showed a Cobb change of 45.5° in the released segments, with a mean correction of 16.2° per segment. (H, I) The preoperative thoracic apical vertebral (T6) rotation of 27.4° was improved to 12.2°. The correction rate of axial vertebral rotation was 55.5%. (J, K) Anteroposterior and lateral radiographs taken at the 18‐month follow‐up. The coronal and sagittal alignments were well maintained.

FIGURE 5.

A 13‐year‐old girl with idiopathic scoliosis underwent Ponte osteotomies alone (PCO) and four segments of PCO. (A–C) Preoperative anteroposterior and lateral radiographs showed a 74.5° main thoracic curve, and the bending radiograph suggested a flexibility of 13.7% of the main curve. Thoracic kyphosis was 8.2°. (D, E) Postoperative anteroposterior and lateral radiographs. The main thoracic curve was corrected to 32.4°, and the thoracic kyphosis was corrected to 16.3°. The correction rate of the main curves was 56.5%. (F, G) The coronal preoperative and postoperative CT reconstruction images showed a Cobb change of 27° in the released segments, with a mean correction of 6.8° per segment. (H, I) The preoperative thoracic apical vertebral (T9) rotation of 28.5° was improved to 18.3°. The correction rate of axial vertebral rotation was 35.8%. (J, K) Anteroposterior and lateral radiographs taken at the 24‐month follow‐up. The coronal and sagittal alignments were well maintained.

Complications

In the PR group, one patient suffered pneumonia as a complication and improved after 5‐day anti‐infection therapy. One patient underwent closed chest drainage for pneumothorax due to pleural rupture, and was removed on postoperative day 4. One patient had delayed removal of a plasma drain for 5 days due to cerebrospinal fluid leakage. One patient had intercostal neuralgia on the release side, which resolved after 1 month of conservative treatment. In the PCO group, one patient had pleural effusions and complained of chest discomfort that improved with 3‐day conservative treatment. Two patients experienced pneumonia and improved within a week of anti‐infection therapy. One patient was readmitted 1 week after discharge with a superficial incisional infection and recovered after multiple dressing changes and antibiotic treatment for 2 weeks. No neurological or implant‐related complications were found in the two groups (Table 3).

TABLE 3.

Complications.

	PR group (n = 20)	PCO group (n = 20)
Pulmonary complications
Effusion	0	1
Pneumonia	1	2
Intercostal neuralgia	1	0
Pleural rupture	1	0
Pleural drainage	1	0
Surgical site infection	0	1
Cerebrospinal fluid leak	1	0
Neurological injury	0	0
Fixation failure	0	0

Subgroup Analysis of Different Etiologies

Subgroup analysis was performed in PR group and PCO group according to different etiologies (Table 4). For the IS patients and the NMS patients, the postoperative main curve Cobb of the PR group was significantly lower than that of the PCO group. And, PR group achieved higher the correction of per segment and the AVR correction rate than that of the PCO group. Compared with the PCO group, the postoperative AVT in the PR group were smaller, but the difference was not statistically significant.

TABLE 4.

Subgroup analysis of different etiologies.

	IS (n = 25)		NMS (n = 15)		F/H	p a
	PR group (n = 12)	PCO group (n = 13)	PR group (n = 8)	PCO group (n = 7)
	$\overline{X}$ ± SD	$\overline{X}$ ± SD	$\overline{X}$ ± SD	$\overline{X}$ ± SD
Flexibility of MC (%)	16.6 ± 5.6	16.5 ± 4.4	11.6 ± 7.1	15.3 ± 4.5	1.673	0.190
Implant density	1.4 ± 0.2	1.3 ± 0.2	1.3 ± 0.1	1.3 ± 0.1	0.897	0.452
SP osteotomy segments	11.1 ± 1.2	11.1 ± 1.1	11.1 ± 0.7	12.1 ± 0.7	1.851	0.155
Ponte osteotomy segments	3.5 ± 0.5	6.0 ± 0.7 b	3.4 ± 0.5	6.1 ± 0.7 c	59.220	<0.001
PR segments	3.4 ± 0.7	None	3.3 ± 0.5	None
Main curve Cobb (°)
Preoperative	77.3 ± 7.0	76.0 ± 5.5	78.0 ± 4.0	81.8 ± 4.2	1.699	0.185
Postoperatively	18.4 ± 6.7	31.3 ± 3.2 b	20.9 ± 6.6	35.4 ± 8.4 c	16.838	<0.001
Post‐correction rate (%)	76.5 ± 7.4	58.6 ± 5.7 b	73.3 ± 8.2	56.9 ± 9.5 c	18.129	<0.001
Last follow‐up	18.9 ± 6.9	32.3 ± 3.6 b	21.8 ± 6.5	36.1 ± 8.3 c	16.776	<0.001
Last correction rate (%)	75.7 ± 7.6	57.2 ± 6.3 b	72.1 ± 8.2	56.0 ± 9.3 c	17.839	<0.001
Loss of correction rate	0.8 ± 0.8	1.4 ± 0.9	1.1 ± 0.9	0.8 ± 0.6	1.364	0.269
Correction of per PR / PCO	13.1 ± 2.6	6.0 ± 1.2 b	14.1 ± 1.8	6.5 ± 1.8 c	47.950	<0.001
AVT (mm)
Preoperative	60.8 ± 13.1	64.8 ± 12.3	68.5 ± 11.0	71.6 ± 18.8	1.079	0.370
Postoperatively	12.0 ± 12.0	19.5 ± 5.6	15.9 ± 9.2	24.1 ± 12.2	2.574	0.069
Last follow‐up	12.6 ± 9.1	19.9 ± 6.7	16.0 ± 5.9	26.2 ± 9.2	4.884	0.006
AVR (°)
Preoperative	26.1 ± 7.9	24.7 ± 8.0	20.0 ± 6.4	19.7 ± 5.6	1.842	0.157
Postoperatively	11.6 ± 5.9	15.7 ± 6.3	7.6 ± 3.8	10.8 ± 3.5 c	3.941	0.016
Correction rate (%)	54.7 ± 15.9	35.8 ± 12.9 b	61.9 ± 13.8	45.6 ± 6.8 c	7.609	<0.001

Note: The bold values represents a statistically significant difference.

Abbreviations: AVR, apical vertebra rotation; AVT, apical vertebra translation; IS, idiopathic scoliosis; MC, main curve; NMS, neuromuscular scoliosis; PCO, posterior column osteotomy; PR, posterior intervertebral releases.

^a The figures were calculated from the comparison of the postoperative values among four subgroups.

^b Statistically significant difference in the value compared with the value of IS in the PR group (p < 0.05).

^c Statistically significant difference in the value compared with the value of NMS in the PR group (p < 0.05).

Discussion

There is no consensus on a surgical treatment strategy for moderate‐to‐severe rigid scoliosis. To achieve ideal corrective outcomes, the release of the rigid spine seems to be inevitable. Anterior release for rigid scoliosis is a very classical surgical technique, but surgeons must be wary of the major trauma and complications associated with it. To our knowledge, this is the first report of a prospective study of posterior release techniques, including posterior discectomy and costovertebral joint release, combined with posterior column osteotomy for the treatment of moderate‐to‐severe rigid scoliosis in comparison to Ponte osteotomy applied alone. The most important finding of our study was that application of the posterior release contributed to significant reductions in curves, apical vertebra translation and rotation without any serious complications. Furthermore, the amount of correction obtained by each release segment was also further analyzed. The results of this study are also consistent with the results of previous studies on the biomechanics of intervertebral discectomy release and Ponte osteotomy. ⁷ , ⁸ Therefore, posterior release could be a highly effective and low‐risk procedure for the treatment of moderate‐to‐severe rigid scoliosis.

Surgical Strategies of Moderate‐to‐Severe Rigid Scoliosis

Surgical treatment strategies of moderate‐to‐severe rigid curves mainly include anterior release and posterior fusion, posterior alone approach, and improved surgical technique. Traditionally, anterior–posterior surgery is performed by lateral‐positioned open thoracotomy or video‐assisted thoracoscopic removal of the disc to increase spinal flexibility, followed by prone posterior fixation and fusion. In order to reduce the operating time and the amount of bleeding, Berry et al. reported on a new technique by performing prone anterior video‐assisted thoracoscopic surgical release and posterior spinal fusion simultaneously. ²¹ The mean preoperative Cobb angle of 81.47 ± 12.37° with mean percent flexibility of 26.38% ± 10.75% was corrected to 32.11 ± 8.52°, and the mean correction rate was 60.24% ± 10.73%. With the popularity of pedicle screws, which passes through the three columns and has three‐dimensional correction function, posterior alone approach for moderate to severe scoliosis has gradually become the mainstream. ¹⁷ , ¹⁸ , ¹⁹ According to these studies, the correction rates of 41.4%–58.5% were achieved, similar to the results of 56.8% of the PCO group in the study. Shen et al. introduced an in situ rod‐contouring technique combined with extensive release of the posterior tissues and the average Cobb angle with the flexibility less than 30% was corrected from 85.7° (range 77°–94°) preoperatively to 33.1° (range 21°–52°) postoperatively, resulting in the correction ratio of 61.3% (43.7%–72.4%). ²² Axia traction is a low‐morbidity adjunct to facilitate scoliosis surgery and is widely used in severe scoliosis. Recently, Hu et al. performed a trial to explore the efficacy of intraoperative halo‐femoral traction in patients with adolescent idiopathic scoliosis, Cobb angles between 70° and 100°, and flexibility <35%. The mean preoperative Cobb angle of 83.20 ± 7.42° with the flexibility of 20.63% ± 6.80%, was corrected to 18.27 ± 6.52°, and the mean correction rate was 77.74% ± 8.81%. ²³ The study had a correction rate a little more than the PR group in our study, which could be because they recruited patients with only adolescent idiopathic scoliosis whose curves were more flexible and easily to correct.

PR for Treatment of Scoliosis

The transpedicular approach allows simultaneous discectomy and anterior column release during posterior osteotomy correction for the treatment of scoliosis. There have been seven studies on posterior discectomy and release for scoliosis, ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ three of which were on rigid scoliosis ¹⁰ , ¹¹ , ¹⁵ ; the mean correction rates for these three studies were 72.4%, 72.6%, and 75.3%, respectively. One of these reports had a 75.3% correction rate for apical convex plus concave intervertebral release, with a mean preoperative Cobb angle of 75.2° (range, 58.7°–110.2°) corrected to 19.0° (range, 8.2°–36.3°). ¹⁵ The other two reports performed only convex intervertebral release. ¹⁰ , ¹¹ In this study, the convex side was released, and an average correction rate of 74.3% was achieved. Although more complete discectomy could generate greater spinal mobility in the axial, sagittal, and coronal directions, we believe that convex release may be a better and more acceptable option for moderate‐to‐severe rigid scoliosis. The narrow operative space on the concave side of the apical region and the drift of the spinal cord towards the concave side make intervertebral release on the concave side difficult and dangerous as opposed to the relatively safe and easy release on the convex side. Furthermore, the release of the costovertebral joints can avoid obstruction of the rib head during convex compression and facilitates the manipulation of the corrective surgery.

Comparison of Corrective Effect between PR and PCO

We found that the correction rate of 74.3% in the PR group was significantly higher than 56.8% in the PCO group, and further analyzed the average correction of 13.5° per release segment, which was significantly higher than that of the 6.2° of PCO, indicating that posterior release had a significant advantage. Similar to the coronal Cobb correction, the coronal AVT and the axial VR also achieved better correction in the PR group than in the PCO group. This is due to the increased mobility in all directions resulting from removal of the intervertebral discs. ⁸ Ponte osteotomy has been the technique for correcting sagittal alignment. The Ponte osteotomy segments were more common in the PCO group than in the PR group, whereas the variation in TK was slightly higher in the PR group, possibly as a result of the easier matching of the pre‐bent rods to the release of the posterior column of the spine.

The current study included idiopathic scoliosis and neuromuscular scoliosis. Different etiologies could bring bias. To minimize bias, we included patients with NMS whose curvature was similar to IS. At the same time, only mild cerebellar tonsillar hernia and/or syringomyelia were found on MRI in this cohort of NMS patients. A subgroup comparison between idiopathic and neuromuscular was performed, and PR group can achieve better orthosis in both etiologies. In the subgroup analyses, the difference between AVT and AVR in the PCO and PR groups did not reach a statistical difference, which may be due to the small sample size. Larger sample sizes for different etiologies are worthy of further research.

Complications and Technical Experience

Pulmonary complications were slightly higher in the PR group than in the PCO group due to the possibility of pleural damage during the release of the vertebral space. All patients improved with conservative treatment, except for one patient who required invasive treatment. One patient in the PR group had a postoperative cerebrospinal fluid leak due to a tear on the nerve root cuff caused during intraoperative piezosurgery osteotomy of the superior articular process. Theoretically, anterior discectomy via a posterior approach could lead to anterior macrovascular injury, whereas the barrier of the anterior longitudinal ligament and manipulation by a senior surgeon could have prevented catastrophic consequences. In addition, attention should be given to hemostasis of the intraspinal venous plexus to avoid possible epidural haematoma. Overall, there were no neurovascular complications in the PR group, and we found it easier to perform posterior discectomy than three‐column osteotomy. In our experience, it took approximately 10 min to release each segment, and the PR group did not have significantly increased operative time or intraoperative bleeding compared to the PCO group.

Strengths and Limitations

The present study first compared the safety and efficacy of posterior PR with PCO alone in patients with moderate‐to‐severe rigid scoliosis through a prospective comparative study. Furthermore, the amount of correction obtained by each release segment was also further analyzed, which can provide a reference for preoperative planning. However, there were several limitations in this study. First, our study was a single‐center and small‐sample study, which may bias the results. Second, the follow‐up period was relatively short, and long‐term follow‐up would be reasonable to assess the efficacy. Considering the radiation exposure, three‐dimensional computed tomography reconstruction of the spine was not performed routinely at follow‐up, and we failed to analyze the condition of the PR region. However, spontaneous interbody fusion has been reported in segments of PR. ⁹

Conclusion

PR combined with PCO has more advantages than PCO alone in the treatment of patients with moderate‐to‐severe rigid scoliosis. Owing to its excellent corrective effect and few complications, this is a high benefit–risk ratio surgical strategy for rigid thoracic/thoracolumbar scoliosis.

Conflict of Interest Statement

The authors have no conflict of interest to disclose.

Ethical Statement

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the West China Hospital of Sichuan University Clinical Trial Ethics Committee, No.2021206. Written informed consent was obtained from the parents.

Author Contributions

All authors had full access to the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Xi Yang and Yueming Song. Acquisition of data: Zhipeng Deng, Peng Xiu, Lei Wang, and Chunguang Zhou. Analysis and interpretation of the data: Zhipeng Deng, Peng Xiu, Chunguang Zhou, and Limin Liu. Drafting of the manuscript: Chunguang Zhou and Peng Xiu. Critical revision of the manuscript for important intellectual content: Lei Wang, Limin Liu, Yueming Song, and Xi Yang. Statistical analysis: Zhipeng Deng, Peng Xiu, and Lei Wang. Obtained funding: Xi Yang and Yueming Song. Study supervision: Xi Yang.

Funding Information

This study was supported by the 1‐3‐5 project for disciplines of excellence, West China Hospital, Sichuan University (ZYGD21001); the National Natural Science Foundation of China (82072386).

Authorship Declaration

All authors listed meet the authorship criteria according to the latest guidelines of the International Committee of Medical Journal Editors, and all authors are in agreement with the manuscript.