Thoracolumbar Kyphosis in Patients with Mucopolysaccharidoses: Clinical Outcomes and Predictive

1

Thoracolumbar kyphosis in patients with mucopolysaccharidoses: clinical outcomes and predictive radiographic factors for progression of deformity

1. Simon B. Roberts, MSc, MRCS(Ed);
2. Ross Dryden, Medical Student;
3. Athanasios I. Tsirikos, MD, FRCS, PhD.

Study conducted at the Scottish National Spine Deformity Centre, Royal Hospital for Sick Children, Edinburgh, United Kingdom.

Address all correspondence to:

Athanasios I. Tsirikos, MD, FRCS, PhD

Consultant Orthopaedic and Spine Surgeon,

Honorary Clinical Senior Lecturer University of Edinburgh,

Scottish National Spine Deformity Centre,

Sciennes Road, Edinburgh, EH9 1LF, United Kingdom.

Email: , Tel. number: 0044-131-662-1265, Fax number: 0044-131-536-0924.

Conflict of Interest Statement:No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

Word Count: 4351 (abstract and text)

Abstract

Clinical and radiological data were reviewed for all mucopolysaccharidoses (MPS) patients with thoracolumbar kyphosis managed non-operatively or operatively in our institution. 16 patients were included (8 female-8 male). Nine patients had Hurler, 5 Morquio and 2 Hunter syndrome. Six patients were treated non-operatively (mean age at presentation 6.3 years; kyphosis progression +1.5o/year; follow-up 3.1 years) and 10 patients operatively (mean age at presentation 4.7 years; kyphosis progression +10.8o/year; follow-up 8.2 years) by circumferential arthrodesis with posterior instrumentation in patients with flexible deformities.

In the surgical group (mean age at surgery 6.6 years; mean post-operative follow-up 6.3 years), mean preoperative thoracolumbar kyphosis of +74.3o was corrected to mean +28.6o post-operatively, relating to a mean deformity correction of 66.9%. Surgical complications included a deep wound infection treated by early debridement, apical non-union treated by posterior re-grafting, and stable adjacent segment spondylolisthesis managed non-operatively. Thoracolumbar kyphosis +38o at initial presentation was identified as predicting progressively severe deformity with 90% sensitivity and 83% specificity.

In conclusion, severe thoracolumbar kyphosis in patients with MPS can be effectively treated by circumferential arthrodesis. Severity of kyphosis at initial presentation may predict progression of thoracolumbar deformity. Patients with MPS may be particularly susceptible to post-operative complications due to the underlying connective tissue disorder and inherent immunological compromise.

Introduction

Mucopolysaccharidoses (MPS) are a group of rare genetic disorders characterised by deficiency in enzymes required for glycosaminoglycan catabolism. Accumulation of glycosaminoglycans leads to progressive cellular damage, multiple organ failure, and reduced life expectancy. Musculoskeletal manifestations are a prominent early feature of MPS. In particular, thoracolumbar kyphosis develops in up to 80% of MPS patients.1 The advent of haematopoetic stem cell transplantation and enzyme replacement therapy has significantly improved morbidity and mortality in MPS, with improvement of most antecedent organ damage.2-4 However, the natural history of musculoskeletal manifestations is relatively unaffected by haematopoetic stem cell transplantation or enzyme replacement therapy.5,6 This is a result of incomplete penetrance of the replenished enzyme into musculoskeletal tissue and extracellular matrix disruption occurring early in the aetiopathogenesis of MPS.7,8The underlying problems associated with progressive skeletal deformity remain unresolved by current enzyme replacement therapy.9-12 Following medical treatment, MPS patients are, therefore, increasingly developing progressive and severe thoracolumbar kyphosis requiring surgical intervention.

There is limited evidence guiding the strategy for surgical stabilisation of thoracolumbar kyphosis in MPS. In 2004, Dalvie et al.13 reported the first series of MPS patients with thoracolumbar kyphosis treated by anterior instrumented fusion with good correction and no complications. Yasin et al.6 described good correction of kyphosis following excision of the apical vertebra and adjacent discs with anterior autologous bone grafting and posterior instrumented fusion extending 3 levels proximal and distal to the apical vertebra. Genevois et al.14 recorded successful fusion and correction of thoracolumbar kyphosis in MPS patients managed by anterior corpectomies and discectomies, anterior reconstruction with autologous tibial strut graft, and short posterior instrumented fusion. The optimal surgical strategy to achieve acceptable correction and stabilisation of thoracolumbar deformity in MPS patients remains undetermined.

The principal indications for surgical treatment of thoracolumbar kyphosis in MPS patients include a progressively severe deformity and impending or evident neurological dysfunction. In 2014, Yasin et al.6 suggested that a kyphosis angle greater than +45o at initial presentation could predict subsequent progression of deformity greater than 10o. It is unclear whether other patient factors such as age at diagnosis or commencement of haematopoetic stem cell transplantation or enzyme replacement therapy also affect subsequent progression of thoracolumbar kyphosis.

The aims of our study were: 1) to investigate the efficacy of combined antero-posterior fusion with or without posterior segmental pedicle screw instrumentation in the treatment of progressively severe thoracolumbar kyphosis in MPS; 2) to investigate patient factors predictive of kyphosis progression and the development of a severe thoracolumbar deformity.

Patients and Methods

We retrospectively reviewed prospectively-collected data of all patients with MPS presenting with thoracolumbar kyphosis between 2001 and 2013 in our centre. All patients met the following inclusion criteria: biochemically-confirmed diagnosis of MPS; patient information comprising demographics, clinical manifestations of MPS, and haematopoetic stem cell transplantation or enzyme replacement therapy; radiographic data available from initial assessment to final follow-up; and operative details including pre- and post-operative radiographic parameters, and peri-operative complications.

The study cohort comprised 16 patients (8 female-8 male) including 9 patients with Hurler (type I MPS), 5 with Morquio (type IV MPS), and 2 with Hunter syndrome (type II MPS). Mean follow-up was 6.2 years (range: 1 to 11.8), at which stage 6 patients (38%) were skeletally mature. Our patients were divided into 2 groups: patients managed non-operatively throughout available follow-up (n = 6), and those treated by surgical stabilisation of spinal deformity (n = 10). The demographic characteristics, extra-spinal manifestations, and spinal features of all patients are shown in Tables 1 & 2.

Radiological parameters of thoracolumbar kyphosis and associated scoliosis were measured using the Cobb method15 on an electronic system by 2 authors (AIT, SBR) based on consensus agreement regarding anatomical landmarks as to the extent of the thoracolumbar deformity. The same landmarks were used for measurements on consecutive radiographs. Thoracic kyphosis was measured between T2 and T12 vertebrae. Deformity correction was determined as follows: Correction Rate (%) = (pre-operative Cobb angle– post-operative Cobb angle) / (pre-operative Cobb angle) x 100. Baseline magnetic resonance imaging of the whole neural axis was performed in all patients to assess for cranio-cervical junction or intra-spinal anomalies. No intraspinal anomalies were detected on radiological assessment. Flexion-extension radiographs were also performed in all patients to assess for atlanto-axial instability. If detected patients were referred to the regional neurosurgical unit for consideration of occipito-cervical fusion prior to surgical stabilisation of thoracolumbar kyphosis.

Patients with thoracolumbar kyphosis <+40o and no neurological deficits, or in whom surgery was contraindicated (Tables 2 & 3; patient 3 – surgery contraindicated due to severe cognitive and functional impairment subsequent to perinatal cerebral infarction), were managed with observation. Bracing was not attempted, as this is ineffective to control the deformity and often poorly tolerated in patients with MPS.13,16 Indications for surgical stabilisation in our group of patients included the presence of severe, progressive thoracolumbar kyphosis >+40o in order to prevent neurological complications and back pain.

Surgical technique. Surgery comprised single-stage circumferential arthrodesis performed through a thoraco-abdominal retroperitoneal and a midline posterior spinal approach. The fibro-cartilaginous extents of the abnormal apical vertebrae were excised and complete discectomies were performed across the levels of the deformity. The apical defect following the corpectomy was bridged with autologous rib strut grafts and morsellised rib graft was placed into the intervening disc spaces. If the thoracolumbar kyphosis remained inflexible after the anterior spinal release upon testing intraoperatively (n = 4; 40% of operated patients), a long rib strut graft was placed anteriorly under compression between the proximal and distal end vertebrae to be included in the fusion to achieve fusion with positional correction [non-vascularised rib strut in 3 patients (Figure 1) - vascularised rib strut in one patient].

Posterior arthrodesis was then performed with autologous and allograft bone across the extent of deformity. Instrumentation was used in patients in whom the thoracolumbar kyphosis became flexible and correctable after the extensive anterior release as assessed intra-operatively (ability to manually correct the kyphosis at completion of anterior discectomies and apical vertebral body resection) (Figure 2; n = 6; 60% of operated patients). The posterior instrumented arthrodesis included the cephalad end vertebra of the kyphosis and caudally the first lordotic lumbar disc to prevent the development of distal junctional deformity. The fusion extended distally to L3 or L4 in order to preserve maximum motion segments as all patients retained walking capacity at the time of surgery. In only one patient (patient 9, Table 4) the lowest vertebra was selected at L5 as revision instrumented arthrodesis was required due to proximal and distal junctional kyphosis developing following previous limited anterior apical fusion performed at another centre. Bilateral paediatric segmental pedicle screw/hook instrumentation was used in all patients undergoing instrumented spinal arthrodesis. Intra-operative spinal cord monitoring was performed recording transcranial motor and cortical/cervical somatosensory evoked potentials, which remained stable throughout in all patients. The postoperative protocol included immediate transfer of all patients to the intensive care unit, weaning from invasive to non-invasive ventilatory support as soon as possible and subsequent transfer to regular ward care. All patients were immobilised in a custom-moulded plaster jacket for 4 months post-operatively until fusion was evident on radiographs, and were permitted to ambulate as able.

Statistical analysis. Demographics, chronology of medical therapy, and radiological parameters between non-operated/operated patients and non-instrumented/instrumented fusions were compared using the two-tailed Student’s t-test. Chi-squared test was used to analyse nominal data between groups. Receiver-Operator Characteristics (ROC) analysis was used to evaluate the initial kyphosis severity predictive of subsequentprogression of deformity. Statistical significance was set at p<0.05. All statistical analyses were performed using SPSS v 21.0.0 software (SPSS Inc., Chicago, Illinois).

Results

Our cohort comprised 16 MPS patients with thoracolumbar kyphosis and mean age at clinical presentation 5.3 years (range: 0.4 to 14.4 years) of which 6 patients were managed non-operatively and 10 patients were treated surgically. There was no significant difference in the distribution of MPS sub-type between patients treated non-operatively or operatively (X2=2.110, p=0.35). All patients with Hurler syndrome had received haematopoetic stem cell transplantation at mean age 1.2 years (range: 0.6 to 1.6). Patients with Hunter or Morquio syndrome were all treated by enzyme replacement therapy, commencing at mean age 6.9 years (range: 5 to 9.5), except 3 patients for whom enzyme replacement therapy was not available from the regional health boards. Occipito-cervical fusion for atlanto-axial instability was performed in 4 patients (3 Morquio; 1 Hurler) prior to management of thoracolumbar kyphosis.

The characteristics of spinal deformity in patients managed non-operatively are shown in Table 3. Mean age at presentation was 6.3 years (range: 0.4 to 13). Mean extent of thoracolumbar kyphosis was 6.2 vertebrae (range: 4 to 9). The thoracolumbar kyphosis progressed from mean +30.5o (range: +8 to +60o) to +35.3o (range: +15 to +75o) over a mean follow-up of 3.1 years (range: 1 to 5.1). The mean rate of kyphosis progression was +1.5o per year. A patient (patient 3, Table 3) with severe thoracolumbar kyphosis and untreated Hunter syndrome continued non-operative management (+60o kyphosis at presentation progressing to +75o at follow-up) as he had suffered a perinatal cerebral infarct with resultant hemiplegia and surgical intervention was contra-indicated following discussion with the patient’s family and the multidisciplinary team. Two patients had an associated thoracic and/or thoracolumbar scoliosis, which was also kept under monitoring (Table 3).

The characteristics of spinal deformity and peri-operative complications in patients treated surgically are shown in Table 4. Mean age at initial presentation was 4.7 years (range: 2.4 to 16.8). Mean extent of thoracolumbar kyphosis was 4.3 vertebrae (range: 3 to 6). Mean thoracolumbar kyphosis at presentation was +54.6o (range: +36 to +95o), progressing to mean +74.3o (range: +42 to +110o) prior to surgical stabilisation over a mean 1.9 years pre-operative follow-up (range: 1.1 to 3.2). The mean rate of kyphosis progression was +10.8o per year prior to surgical treatment. The mean age at surgery was 6.6 years (range 2.4 to 16.8) and mean post-operative follow-up was 6.3 years (range: 3.5 to 10.3). The mean post-operative thoracolumbar kyphosis was +28.6o (range: 0 to +65o), corresponding to a mean correction of 66.9% (range: 31 to 100%). Patients treated by un-instrumented arthrodesis (Figure 1) had a greater mean pre-operative kyphosis [+92.3o (range: +55 to +110o) compared to +62.3o (range: +37 to +78o); p=0.04], and less absolute [+39o (range: +27 to +55o) compared to +50.2o (range: +37 to +64o); p=0.09] and percentage correction of kyphosis [42.7% (range: 30.9 to 50%) compared to 83.1% (range: 55.1 to 100%); p=0.002], compared to patients who underwent an instrumented arthrodesis (Figure 2). There was no significant loss of kyphosis correction between immediate post-operative and latest follow-up radiographs for all patients. Thoracic kyphosis improved from mean -0.7o (range: -50 to +14o) pre-operatively to mean +20.7o (range: -15 to +37o) postoperatively. Two patients had a thoracolumbar and 2 a thoracic scoliosis in association with the thoracolumbar kyphosis. The thoracolumbar scoliosis was corrected from mean +35.5o (range: +26 to +45o) pre-operatively to mean +7.5o(range: 0 to +15o) post-operatively; the thoracic scoliosis improved spontaneously from mean +29o (range: +25 to +33o) pre-operatively to mean +17o (range: +14 to +20o) post-operatively following correction of the thoracolumbar kyphosis.

Prior to spinal surgery, 2 patients required establishment of an elective tracheotomy. Mean surgical time was 170 minutes (range: 150 to 240), and mean intra-operative blood loss was 387 mls (range: 200 to 1000). Three patients developed surgical complications in our cohort. One patient developed an early deep wound infection treated with surgical debridement and intravenous antibiotics, permitting retention of instrumentation with no subsequent loss of deformity correction. The same patient also developed proximal junctional kyphosis which remained stable at follow-up and did not require extension of the fusion. Another patient who underwent an un-instrumented fusion developed non-union at the apex of the kyphosis posteriorly treated with posterior re-grafting using allograft bone followed by application of a custom-moulded plaster jacket for 3 months. The only patient with Hunter syndrome treated operatively developed a stable grade II spondylolisthesis at the adjacent segment distal to the instrumentation; as the patient remained asymptomatic, this was managed conservatively to skeletal maturity and the spondylolisthesis showed no evidence of deterioration. Post-operative medical complications included chest infection in 3 patients, prolonged ileus and need for tracheotomy for prolonged mechanical ventilation in one patient each. Mean hospital stay was 31 days (range: 10 to 142). All patients had been ambulant pre-operatively and there was no alteration in ambulatory status post-operatively and at follow-up. Mean post-operative follow-up was 6.3 years (range: 3.5 to 10.3), at the conclusion of which the mean age of operated patients was 14.5 years (range: 7 to 24).

The demographics, surgical techniques, radiographic outcomes and peri-operative complications in our cohort of MPS patients with thoracolumbar deformity are compared with previously published series in Table 5.

Variables at initial presentation that may predict progression to severe thoracolumbar kyphosis and need for surgical stabilisation were investigated. Significant association was found for neither age at presentation (p=0.836) nor age at commencement of haematopoetic stem cell transplantation or enzyme replacement therapy (p=0.513). The size of thoracolumbar kyphosis at initial clinical presentation was associated with progression to severe deformity and need for surgical treatment (p<0.001). ROC analysis demonstrated that an initial thoracolumbar kyphosis +38o predicts progression to severe deformity and need for surgical correction with 90% sensitivity and 83% specificity (AUC = 0.892).

Discussion

We believe this is one of the largest reported series reviewing the outcomes of treatment for thoracolumbar kyphosis in patients with MPS and contains the longest reported follow-up after non-operative management or single-stage circumferential arthrodesis, with 38% of the study cohort reaching skeletal maturity. Consensus treatment recommendations have not yet emerged due to the limited reports of both the natural history and management of thoracolumbar kyphosis in patients with MPS. In our series, there was similar distribution of patients with each MPS subtype (I, II, and IV) in the non-operative and operative groups. Hurler syndrome patients in our cohort under the age of 2 years at presentation had a mean initial thoracolumbar kyphosis of +42.5o, which is similar to the mean +38o described by Yasin et al.6 Despite all patients with Hurler syndrome receiving haematopoetic stem cell transplantation and tending to present at a younger age, a higher proportion of these patients (7 of 9; 78%) developed progressively severe thoracolumbar kyphosis requiring surgical stabilisation than did patients with Hunter (1 of 2; 50%) or Morquio (2 of 5; 40%). The mean age at presentation in this cohort was slightly greater than previously reported cohorts of MPS patients6,14 due to the inclusion of patients with Hunter and Morquio syndromes who may present at a later age. The thoracolumbar deformity in our series does not appear to universally progress in patients of each MPS subtype, as has been recognised previously.6 Thoracolumbar kyphosis appears to follow a variable course within each subtype, but Hurler syndrome patients are more likely than patients with other MPS to develop early a progressively severe thoracolumbar kyphosis. Factors influencing the heterogeneity of phenotypic severity of thoracolumbar deformity remain unclear, but may include genotype, engraftment, or enzyme levels.6,17

The surgical techniques described for correction of thoracolumbar deformity in patients with MPS include anterior instrumented fusion, anterior fusion using vascularised rib graft, and combined anterior/posterior instrumented fusion.6,13,14 Dalvie et al.13 described good correction of kyphosis in MPS patients treated by anterior-only instrumented fusion, though this series included predominantly patients with Morquio syndrome and less severe kyphosis. Although no complications were described, the duration of post-operative follow-up was limited. Anterior fusion using vascularised rib graft has also been described for MPS patients with thoracolumbar kyphosis but is associated with a high rate of post-operative complications and progression of deformity.6 This was also the experience in our study in which a single patient was managed by anterior reconstruction using a vascularised rib graft; the patient required prolonged post-operative ventilatory support and a tracheotomy, underwent re-grafting surgery to address an apical non-union, and developed long-term chest wall disruption across the harvesting site. In contrast, non-vascularised rib grafting permits placement of a strut across the deformity, while the defect at the donor rib site gradually heals as the vascular supply across the rib bed is preserved. This produces, in our experience, much less disruption of the chest wall when compared to a vascularised rib which creates a permanent thoracic defect across the harvesting site. The disadvantage of a non-vascularised rib graft is that it takes longer to consolidate to the vertebral bodies and therefore postoperative support in the form of a custom-moulded spinal jacket is recommended.Posterior-only fusion is also associated with a higher risk of pseudoarthrosis. By comparison with the management of other skeletal dysplasias, surgical correction of developmental kyphosis secondary to failure of anterior vertebral body formation may be optimally achieved by anterior spinal release with strut graft correction and circumferential arthrodesis with the use of posterior instrumentation.18-21 Antero-posterior arthrodesis is also associated with a reduced risk of neurological complications,22,23 and has been successfully described in the management of patients with Hurler syndrome (Table 5).6,14