Modified Closing-Opening Wedge Osteotomy to Correct Kyphosis in Ankylosing Spondylitis.

BACKGROUND The aim of this study was to analyze the clinical and radiological outcomes of modified closing-opening wedge osteotomy (mCOWO) for correcting kyphosis in ankylosing spondylitis (AS) patients. MATERIAL AND METHODS From April 2012 to April 2017, records of consecutive patients who underwent mCOWO were reviewed. The clinical and radiological outcomes were analyzed preoperatively, postoperatively, and at the most recent follow-up. RESULTS Eleven AS patients underwent mCOWO, with a mean follow-up of 19.4 months (range, 12-45 months). The average sagittal vertical axis (SVA) was corrected from 191.9 mm preoperatively to 75.9 mm postoperatively (P<0.05) and 78.9 mm at the most recent follow-up (P<0.05). The average correction angles at the osteotomy site were 44.5° postoperatively and 45.0° at the most recent follow-up (P>0.05). Sagittal translation (ST) occurred in 2 patients, and 5 mm was the maximum. There was no neurologic damage. Solid fusion was observed at the most recent follow-up in all patients. CONCLUSIONS Modified closing-opening wedge osteotomy (mCOWO) is an effective technique for correcting kyphosis in patients with AS.

Background

Ankylosing spondylitis (AS) is a form of chronic, progressive inflammatory rheumatism that mainly affects young people ages 20–30 years. Its general prevalence is 0.1–1.4%, and men are more often affected than women, with a male-to-female sex ratio of roughly 2 to 1 [1]. The main clinical features are back pain caused by sacroiliitis and inflammation at other locations in the axial skeleton, as well as spinal stiffness caused by osteoproliferation. As the disease progresses, changes in the vertebrae cause hyperkyphosis and sagittal imbalance, which can lead to restriction of looking straight ahead and a decrease in quality of life. Although there are some conservative treatments available [2,3], the progression of AS is irreversible. Therefore, corrective spinal osteotomy is considered to restore the horizontal visual axis and sagittal balance of the spine in late-stage AS patients.

Traditionally, pedicle subtraction osteotomy (PSO) has been commonly used to correct kyphosis in AS patients, and a single-level PSO can achieve a correction of 27° on average [4,5]. In recent years, some researchers have performed closing-opening wedge osteotomy (COWO) to correct kyphosis, obtaining corrections that are 10° larger on average than those obtained with PSO [6-8]. In PSO and COWO, the posterior elements and part of the vertebral body of the target vertebra are removed, causing instability and sagittal translation (ST) of the spine. Such instability and translation have the potential to result in neurologic damage [9].

In this paper we describe a new method, modified COWO (mCOWO), to correct kyphosis in patients with AS. The radiographic outcomes, complications, and risks of this technique were evaluated in a series of 11 patients.

Material and Methods

Patients

Records of 15 consecutive AS patients with kyphotic deformity who underwent the mCOWO procedure to correct kyphosis in our institution from April 2012 to April 2017 were retrospectively reviewed. The inclusion criteria were: diagnosis of AS with thoracolumbar or lumbar kyphosis, single-level mCOWO, and a minimum 12 months of follow-up. The exclusion criteria were: cervical kyphosis, revision surgery, and 2-level osteotomy. One patient with spinal osteotomy history, 1 patient with two-level osteotomy and 2 patients with less than 12 months of follow-up were excluded. The mean age of the remaining patients (9 men, 2 women) was 38.3 years (range, 27–50 years), with a mean follow-up period of 19.4 months (12–45 months). The study was approved by the Institutional Review Board of Xi’an Jiaotong University, and signed informed consents were obtained from all involved individuals before the study.

Surgical techniques

According to the severity of the patients’ kyphosis, the operation bed was positioned into a suitable reverse “V” shape before surgery. After general anesthesia, the patient was placed in prone position. Following a standard posterior median incision, the spinous process, lamina, and facet joints were completely exposed and pedicle screws were inserted into at least 2 vertebrae above and below the osteotomy site. Then, the spinous process, the cephalic lamina, and the superior facets of the osteotomized vertebra were removed, and the caudal lamina and the inferior facets were resected at the proximal adjacent vertebra. Subsequently, a unilateral temporary rod was installed to maintain spinal stability and to avoid any uncontrolled ST during the surgery. The mCOWO was then performed at the level of the osteotomized vertebra (Figure 1). The upper parts of the pedicles, the postero-superior vertebral body, and the upper intervertebral disc adjacent to the osteotomized vertebra were removed in a wedge fashion using the osteotome, rongeur, and curette, and the osteotomy gap was carefully formed into a “V” shape with the anterior vertebral edge being broken, which is crucial to preserve the integrity of the inferior facet joints of the osteotomized vertebra and the intervertebral foramen between the osteotomized vertebra and the one below it. The dural and nerve roots were carefully protected during the osteotomy. After this, another temporary rod was added on the other side and a cage packed with autologous bone particles was inserted into the anterior part of the osteotomy gap. The gap was closed in a stepwise manner by gradually flattening the operation table and providing repeated posterior compression. Then, the temporary rods were replaced with 2 precontoured rods on both sides. All the instrumentation was fastened after the correction was confirmed by intraoperative radiography. A bone autograft was paved on the posterior column and the intervertebral space was filled in to facilitate spinal fusion.

Figure 1

Schematic illustration of modified closing-opening wedge osteotomy (mCOWO): (A) The lateral view outlines the tissue to be resected including the upper parts of the pedicles, the postero-superior vertebral body and the upper intervertebral disc adjacent to the osteotomized vertebra, and the osteotomy gap is carefully formed into a “V” shape with the anterior vertebra edge being broken. (B) The postoperative lateral view shows that correction is achieved hinging on the cage, closing the posterior osteotomy and opening the anterior column of the spine.

All procedures were performed by the senior author (H.T.) under both somatosensory-evoked potential (SEP) and motor-evoked potential (MEP) monitoring throughout the surgery. Autologous blood transfusion was used in each operation.

Postoperatively, the patients were allowed to ambulate with a customized thoracolumbosacral orthosis 3 days after surgery. The orthosis was typically maintained for 3–6 months.

Radiological assessment

Standing anteroposterior and lateral radiographs of the whole spine were taken before surgery, postoperatively (within 1 week after surgery), and at the most recent follow-up (Figure 2). The following 6 radiographic parameters were measured using Surgimap (Nemaris, Inc., New York, NY, USA) by an experienced spine surgeon who was not involved in the treatment: (1) mCOWO angle (MA), defined as the correction angle from preoperation to postoperation and the most recent follow-up, was measured between the lower endplate of the osteotomized vertebra and the upper endplate of the adjacent cephalad vertebra (Figure 2I, 2J); (2) thoracic kyphosis (TK), measured as the Cobb angle between T4 and T12; (3) lumbar lordosis (LL), measured as the Cobb angle between T12 and S1; (4) pelvic incidence (PI), defined as the angle between the line perpendicular to the sacral plate at its midpoint and the line connecting this point to the midpoint of the axis of 2 femoral heads; (5) sacral slope (SS), defined as the angle between the superior plate of S1 and a horizontal line; and (6) sagittal vertical axis (SVA), measured as the horizontal distance from the posterosuperior corner of the S1 body to the sagittal C7 plumb line.

Figure 2

A 32-year-old woman in AS with a follow-up of 45 months. (A, B) Preoperative X-ray films showing thoracolumbar kyphosis and sagittal imbalance. (C, D) The patient underwent L2 mCOWO with sagittal balance being restored. (E, F) At the most recent follow-up, X-ray films showed that the sagittal balance was well maintained. (G, H) Clinical appearance before surgery and at the most recent follow-up. (I, J) Radiographs showed that the MA was 43.3° after the surgery. (K) Radiograph at the most recent follow-up showing solid fusion at the L1–L2 level.

Sagittal translation (ST) was defined as displacement >2 mm, which was measured between the posterior superior edge of the osteotomized vertebra and the posterior inferior edge of the adjacent cephalad vertebra.

The estimated blood loss and operation time were obtained from medical records. Clinical outcomes were evaluated using the Scoliosis Research Society-22 (SRS-22) outcome questionnaire (including 5 domains: pain, appearance, function, mental health, and satisfaction) before surgery and at the most recent follow-up.

Statistical analysis

All data are expressed as means ± standard deviations and were statistically analyzed by SPSS 18.0 (SPSS, Inc., Chicago, IL, USA). A paired t test was performed to compare the radiological parameters preoperation, postoperation, and from the most recent follow-up. The SRS-22 scores before surgery and at the most recent follow-up were also compared using the paired t test. A difference of P<0.05 was considered statistically significant.

Results

Surgical and clinical results

mCOWO was performed on L1 in 1 patient, L2 in 7 patients, and L3 in 3 patients. The average operative time was 345.5 min (range, 260–390 min). The mean intraoperative estimated blood loss was 2045 ml (range, 1000–3000 ml). The mean follow-up was 19.4 months (12–45 months). Regarding the clinical outcomes, the SRS-22 scores for pain, appearance, function, and mental health were significantly improved. Most of the patients (10/11) were satisfied with the surgery (Table 1).

Table 1

SRS-22 scores.

Domain	Preoperative	The most recent follow-up
Pain	2.24±0.67	4.22±0.53*
Appearance	1.69±0.57	4.44±0.66*
Function	2.27±0.68	4.16±0.71*
Mental health	1.91±0.52	4.42±0.66*
Satisfaction	–	4.59±0.50

Statistically significant (P<0.05).

Radiological results

The average MAs were 44.5° postoperatively and 45.0° at the most recent follow-up (P>0.05). The average LL changed from 0.1° preoperatively to -40.9° postoperatively (P<0.05) and −38.4° at the most recent follow-up (P<0.05). The average SS increased from 4.6° preoperatively to 19.3° postoperatively (P<0.05) and 18.6° at the most recent follow-up. The average SVA was corrected from 191.9 mm preoperatively to 75.9 mm postoperatively (P<0.05) and 78.9 mm at the most recent follow-up (P<0.05). However, TK and PI did not significantly differ between preoperation and postoperation (P>0.05). At the most recent follow-up, GK, TK, LL, PI, SS, and SVA did not significantly change compared to the postoperative parameters (P>0.05) (Table 2).

Table 2

Radiological outcomes of patients with AS underwent mCOWO (x±s).

Measurements	Pre-operation	Post-operation	The most recent follow-up
TK/°	52.8±12.5	52.8±14	53.2±13.8
LL/°	0.1±22.0	−40.9±18.5*	−38.4±18.8*
PI/°	46.9±9.8	49.1±9.7	48.8±8.8
SS/°	4.6±11.8	19.3±10.3*	18.6±11.3*
MA/°	–	44.5±8.6	45.0±8.8
SVA/mm	191.9±44.7	75.9±37.0*	78.9±42.1*

Statistically significant (P<0.05).

Complications

No neurological or vascular injury or dural tear occurred intraoperatively. No pseudarthrosis or implant breakage or loosening were found. All patients achieved fusion at the most recent follow-up (Figure 2K). ST occurred in 2 patients (5 mm in one patient and 4 mm in the other one), none of whom experienced neurological deficits. One patient complained of abdominal distension with intermittent postoperative abdominal pain and felt relief within 24 hours.

Discussion

PSO is highly effective for restoring spinal sagittal balance in patients with AS [10]. However, a single-level PSO can only obtain an average correction of 27°, and correction rarely exceeds 40° [4,5,7,10]. For AS patients with severe kyphosis, single-level PSO usually cannot achieve satisfactory correction. Although vertebral column resection (VCR) can obtain greater correction at a single level, it is reserved for the most severe spinal deformity due to its higher risk of neurologic deficits and massive blood loss [11,12]. In an attempt to improve the correction at a single level without VCR, in 2001 Kawahara et al. [6] created the COWO technique to treat patients with kyphosis and obtained corrections between 31° and 55°. Subsequently, other investigators [7,8,13] performed COWO to treat patients with sagittal imbalance of various etiologies, including AS, and achieved an average regional correction of 41.8° at 1 level. Similar to COWO, this study showed that single-level mCOWO achieved an average correction of 44.5°, which was also greater than that obtained with PSO (27°). The correction obtained with these osteotomies is highly associated with the corrective mechanism involved. PSO, COWO, and mCOWO all create a “V”-shaped osteotomy gap that penetrates all 3 columns of the vertebra. PSO corrects the kyphosis with a hinge on the leading edge of the vertebral body, and shortens the middle and posterior columns. In contrast, mCOWO involves placing a cage, which serves as a hinge, in the osteotomy gap before closing the posterior column. The first step of correction in mCOWO is similar to that of PSO, in which the posterior osteotomy gap is closed with the hinge at the leading edge of the vertebral body. In the second step, the position of the hinge is moved posteriorly to the cage, and greater correction is achieved by continuing to close the posterior column while simultaneously opening the anterior cortex. The corrective mechanism of COWO is the same as that of mCOWO. Therefore, COWO and mCOWO can achieve an extra 10° correction at the osteotomy site compared with PSO because of the shift of the hinge point [7].

Postoperative fusion in patients who have undergone osteotomy is important for surgical outcomes because failure of fusion can lead to pseudarthrosis, which causes local pain and implant damage [14]. Almost all patients with pseudarthrosis required revision surgery with another anterior, posterior, or circumferential arthrodesis [14-16]. Previously, researchers found that the prevalence of pseudarthrosis was 6.3–15% after lumbar PSO [14,16], and Chang et al. reported that the rate of pseudarthrosis was 7.3% after lumbar COWO [13]. However, in our cohort, solid fusion was observed on all the patients’ radiographs. We speculated that these differences were associated with the different elements removed by the osteotomies: the inferior articular processes and their continuity with the vertebral body are preserved in mCOWO, while the superior and inferior facets and the pedicle are all resected in PSO and COWO. Due to the bone continuity from the vertebral body to the lamina, there is a larger graft bed and better blood supply for the bone grafts at the osteotomy site in mCOWO. Theoretically, posterior fusion can be more effective in mCOWO than in PSO or COWO. Furthermore, Yang et al. [16] reported that pseudarthrosis invariably occurred in patients whose intervertebral discs above and below the osteotomy site were preserved. In this situation, posterior failure of fusion with intact anterior discs functionally creates a circumferential nonunion. However, in mCOWO, the upper adjacent disc was resected, and interbody fusion can occur between the osteotomy vertebra and the proximal cephalic vertebra. Therefore, patients who undergo mCOWO can obtain better fusion with fewer pseudarthroses and implant failures than those who undergo PSO or COWO.

Sagittal translation (ST) is a common complication that occurs in 3-column osteotomy as a result of the instability of the spine during the operation [7,17-20]. Chang et al. [17] confirmed that patients with ST during surgery had a significantly higher risk of developing neurological complications than those without ST. Moreover, if sagittal subluxation, a severe form of ST, occurs, it can result in nonunion at the osteotomy site, which may require a revision by anterior approach [15]. In our cohort, the maximum ST was 5 mm, far less than the 15 mm reported with PSO [20]. Although the literature has not reported a specific ST value in COWO for comparison, we believe that mCOWO can result in a smaller ST and a lower incidence of severe ST than PSO or COWO. During PSO and COWO, elements including the spinous process, lamina, upper and lower articular processes, bilateral pedicles, and partial vertebral bodies of the target vertebral body are resected; however, mCOWO maintains the integrity of the lower facet joint and inferior pedicle and removes the upper endplate and the upper disc. Theoretically, the intact lower articular processes can enhance the stability of the posterior column, protect the lower nerve roots, and reduce the incidence of ST. In addition, some researchers have demonstrated [21] that PSO with a cage placed in the osteotomy gap can significantly reduce the incidence of ST because the cage would fall into the osteotomized vertebra and act as a stable link between the upper and the lower sides of the osteotomy site during the closure of the gap. In contrast, only the caudal part of the cage sinks into the osteotomized vertebra in mCOWO, while the cranial part does not. Nevertheless, when preforming posterior compression in mCOWO, great friction can occur between the cranial side of the cage and the lower endplate of the proximal vertebra, which could reduce the magnitude of ST.

mCOWO is safer than PSO when the osteotomy site is higher than L1. The position of the conus medullaris usually corresponds to the lower third of L1 [22], and the cauda equina is below it. When PSO, COWO, or mCOWO are performed on a segment below L1, the shortening of the posterior column seems not to injure the redundant cauda equina [23]. However, when the osteotomy site is located at L1 or a higher level of the spine, the increased degree of correction leads to more obvious shortening of the posterior column, increasing the risk of wrinkles, compression, and ischemia in the spinal cord, which cause damage to the central nervous system. In animal experiments on adult dogs, Kawahara et al. [24] demonstrated that the dural sac was shrunken and buckled during spinal shortening between one- and two-thirds of the vertebra, and kinking of the spinal cord was observed after shortening over two-third of the vertebra, which resulted in neurological complications. When the correction angles are the same for the 3 methods, the spinal column shows less shortening in COWO and mCOWO; therefore, the risk of spinal cord injury is lower.

Some limitations of this study should be considered. The foremost limitation is the small sample size resulting from the patients’ choices of medical institutions. Furthermore, the lack of long-term follow-up restricts assessments of the correction angle and the condition of the posterior fusion. In addition, a randomized controlled trial is necessary to refine the relevant results.

Conclusions

mCOWO provides an alternative technique for correcting kyphosis associated with AS. The correction angle of mCOWO is similar to that of COWO and significantly greater than that of PSO. However, mCOWO can obtain better fusion postoperatively, and the risk of the incidence of ST in mCOWO is theoretically lower than that in PSO and COWO.