We have treated 101 patients with scoliosis

Transcription

1 The treatment of scoliosis in muscular dystrophy using modified Luque and Harrington-Luque instrumentation G. Bentley, F. Haddad, T. M. Bull, D. Seingry From the Royal National Orthopaedic Hospital Trust, Stanmore, England We have treated 101 patients with scoliosis secondary to muscular dystrophy over a 13-year period; 64 had Duchenne s muscular dystrophy, 33 spinal muscular atrophy and four congenital muscular dystrophy. The patients underwent a modified Luque (87) or Harrington-Luque instrumentation (14) combined with a limited Moe fusion in all except 27 cases. A mean of 13 levels was instrumented. The mean preoperative sitting Cobb angle was 84 (10 to 150) and the mean postoperative angle 40 (52% correction). Most patients (96%) were able to discard their braces and there was a high level of patient satisfaction (89.6%). Less correction was seen for severe curves, and there was a greater recurrence of postoperative pelvic tilt in those patients not instrumented to the sacrum. Although the incidence of minor or temporary complications was high, these occurred chiefly in the early high-risk patients with very severe curves and considerable pre-existing immobility. J Bone Joint Surg [Br] 2001;83-B:22-8. Received 18 March 1999; Accepted after revision 7 January 2000 In neuromuscular scoliosis the major disability is attributable to weakness of the spinal musculature which cannot be supported by bracing. This often produces costopelvic impingement, painful pressure areas and further restriction of an already compromised respiratory system (Fig. 1). The main advantage of instrumentation of the spine is to allow the patients to sit, stand or walk more comfortably without a brace, thus improving their ability to function and communicate. Arresting the progression of a curve may reduce the incidence of infection of the respiratory tract and G. Bentley, ChM, FRCS, FMedSci, Professor of Orthopaedic Surgery F. Haddad, BSc, MCh Orth, FRCS Orth, Senior Registrar in Orthopaedics T. M. Bull, FRCS Orth, Senior Registrar in Orthopaedics D. Seingry, FRCA, Consultant Anaesthetist Institute of Orthopaedics, Royal National Orthopaedic Hospital Trust, Brockley Hill, Stanmore, Middlesex HA7 4LP, UK. Correspondence should be sent to Professor G. Bentley British Editorial Society of Bone and Joint Surgery X/00/ $2.00 perhaps increase longevity. 1 An additional benefit is improved cosmesis, although this is not in itself an indication for surgery. Full correction can be achieved in early cases. The patients are often at risk of cardiac and respiratory complications, particularly those with Duchenne s muscular dystrophy (DMD) in which there is involvement of both diaphragmatic and cardiac muscles. There is also an abnormality of blood vessels as part of the generalised mesenchymal defect, which may result in increased blood loss during surgery. 2 All these factors make it imperative that the operation planned should be modest, while still providing spinal stability. For this reason since 1983 a modified Luque segmental spinal instrumentation with limited Moe fusion has been undertaken on patients confined to wheelchairs, and Harrington-Luque instrumentation with Moe fusion on those who were mobile. 3,4 Our aim was to assess the outcome of a single-stage posterior spinal procedure in this high-risk group of patients with neuromuscular scoliosis, and to quantify the morbidity associated with this form of limited surgery. We also studied blood loss, the necessity for Moe interarticular fusion and the relationship between the site of chromosomal deletion and progression of the curve. Patients and Methods Between April 1983 and January 1996, we treated surgically 101 patients with neuromuscular scoliosis secondary to muscular dystrophy under the care of the first author (GB). DMD was present in 64 patients, 33 had spinal muscular atrophy (SMA), and four had congenital muscular dystrophy (CMD). There were 24 girls and 77 boys with a mean age of 12.7 years (5 to 18), although for DMD patients (all boys) the mean age was 13.4 years (5 to 16). The mean time from the development of symptoms to diagnosis of the underlying condition was 9.8 years (3 to 16) and that from diagnosis of the scoliosis to surgery was 3.8 years (1 to 11). All the case notes and radiographs were reviewed, and all surviving patients were reassessed clinically with new spinal radiographs at a mean of 6.8 years (2 to 12.5) after operation. The indications for surgery changed slightly over the years as our experience expanded, but the main indications 22 THE JOURNAL OF BONE AND JOINT SURGERY

2 THE TREATMENT OF SCOLIOSIS IN MUSCULAR DYSTROPHY USING MODIFIED LUQUE INSTRUMENTATION 23 Fig. 1 Photograph of a patient with scoliosis secondary to DMD. The pressure areas from the brace can be seen over the ribs on the right, and the costopelvic impingement on the left. The patient is unable to support himself. were a Cobb angle greater than 50 in an ambulant patient and more than 30 in the wheelchair-bound. Initially, patients presenting for surgery had severe curves and experienced more complications, but earlier operation became the rule in the latter part of this series. Dubowitz 5 has shown that patients develop rapidly progressive curves when they become wheelchair-dependent. Therefore, any progressive curve not controlled by bracing, and in DMD patients any curve progressing more than 20 in six months, underwent stabilisation. It should be noted that all the patients with DMD in our series were already wheelchair-dependent at the time of surgery, for a variety of reasons. Relative indications for stabilisation included increasing pelvic obliquity, difficulty with bracing or noncompliance, skin pressure sores and increased dependence, especially if the upper limbs were required for control in the sitting position. Preoperative assessment. As part of their preoperative assessment, all patients were seen by the occupational therapists and social workers to identify any special needs in their home environment. Both patients and their parents were counselled as to the potential risks and benefits of surgery. Respiratory function tests and a cardiac assessment, including echocardiography and electrocardiography, were carried out before surgery to ensure that the patients were fit for anaesthesia. Standardised erect and suspended radiographs of the entire spine were taken. Cobb angles and pelvic obliquity, defined as the angle between the iliac crests and the horizontal plane on a sitting anteroposterior (AP) radiograph, were measured by the second and third authors (FH and TMB). Mild cardiac impairment was discovered in three patients, but it was not thought to be a significant risk to surgery. The ASA score for each patient was determined by the consultant anaesthetist; 31 were deemed to be ASA II, 44 ASA III, and 26 ASA IV. The forced vital capacity, expressed as a percentage of the expected normal, gave a mean value of 35% (18 to 68). In the DMD group this was 33% (18 to 63) and in the SMA group, 37% (20 to 68). There was no significant difference between the groups. The mean preoperative Cobb angle in the erect position was 84 (10 to 150) and suspended, 53 (3 to 108). Pelvic obliquity measured 21 (0 to 49). Other preoperative data are summarised in Table I. Genetic studies revealed that 48% of DMD patients had a deletion of one or more exons of the dystrophin gene. Most (75%) had a deletion in the central region of dystrophin and only a few at the 5' end. All patients had out-offrame deletions. Operative technique. All the operations were carried out by the first author (GB), under general anaesthesia, with the patient prone. All patients had somatosensory recording of spinal evoked potentials (SSEP) during the procedure. 6 None of the patients underwent preliminary or concurrent anterior surgery. A straight, longitudinal incision was made and the dorsal elements of the spine were exposed as far laterally as the facet joints. For the 85 non-ambulant patients, the sacroiliac joints were exposed and a cross-rod technique used to introduce the Luque rods through the sacroiliac joints into the ilium for pelvic stabilisation (see Fig. 4b). The spine was corrected by manual pressure and the two rods were contoured to produce a slight overcorrection of the curvature. Sublaminar, 18-gauge wires were passed at each level, one for each side. The rods were then inserted, convex side first, and the wires were tightened sequentially from below upwards. Instrumentation was taken high, at least to T4, because of the considerable risk of further deformity developing above the fused level if lower. 7,8 Harrington rods were used instead of Luque s in the 14 ambulant patients. In these, doubled laminar wires were used at alternate levels and secured to the Harrington rod. Most patients (74) had a modified Moe fusion. 4 This involved elevation and overlapping of part of the contiguous margin of the neural arches and curettage of the posterior spinal joints without added bone graft. In the nonambulant patients, which included 24 with DMD, we did not carry out the modified Moe fusion. This was to reduce VOL. 83-B, NO. 1, JANUARY 2001

3 24 G. BENTLEY, F. HADDAD, T. M. BULL, D. SEINGRY the blood loss in patients who were considered to be potentially at risk under general anaesthesia. The mean duration of surgery for the whole group was 197 minutes (115 to 330) and 213 minutes (155 to 330) for the DMD group which was not significantly different. The mean number of levels fused was 13 for all groups. There was a significant increase in operating time and blood loss in those with a modified Moe fusion (p < 0.01), although it should be noted that the patients who did not have this procedure had a degree of cardiorespiratory instability during the operation. For this reason the surgery was conducted at greater speed. The mean blood loss of the DMD group was 3034 ml (500 to 8700) compared with 2315 ml (350 to 8700) for the rest of the series (p < ). Postoperatively, all patients were transferred to the highdependency ward for monitoring and intensive physiotherapy. Most required assisted ventilation for 12 to 48 hours. Prophylactic flucloxacillin (500 mg four times daily) and amoxicillin (500 mg three times daily) were given for five days and, thereafter, if there was any suggestion of wound infection. Urinary output was monitored with an indwelling catheter. A nasogastric tube, passed after intubation, was maintained until bowel sounds returned. This was frequently as long as five days in patients with DMD. Gastric histamine-2 receptor blockade using intravenous and then oral rantidine offset the risk of stress ulceration. The patients were turned frequently and began sitting as soon as they were fully conscious, achieving a full upright position two weeks after surgery. Feeding by mouth was delayed until they were alert, sitting up and had bowel sounds. Mobile patients wore an underarm brace for three to six months after operation. Follow-up was undertaken in the orthopaedic clinic during the early postoperative period and subsequently in a combined neuromuscular/orthopaedic specialist clinic where a multidisciplinary team was available. Table I. The mean preoperative Cobb angle in degrees (range) and pelvic obliquity in the 101 patients with scoliosis secondary to muscular dystrophy Cobb angle Erect Suspended Pelvic obliquity Overall 84 (10 to 150) 53 (3 to 108) 21 (0 to 50) DMD 70 (18 to 125) 48 (3 to 108) 20 (0 to 50) SMA 92 (50 to 150) 59 (24 to 91) 22 (0 to 45) Table II. The mean postoperative Cobb angle and pelvic obliquity, measured erect (with corrections) in the 101 patients with scoliosis secondary to muscular dystrophy Cobb angle Pelvic obliquity Mean Correction (%) Mean Correction (%) Overall DMD SMA Results The mean length of hospital stay was 21 days (16 to 90). The postoperative data are summarised in Table II. The mean postoperative Cobb angle was 40 (0 to 99), a correction of 52% (Table II). There was no significant difference between the DMD and SMA groups. Less correction was seen in patients with severe curves greater than 70, although these did not lose a greater percentage of the correction obtained after surgery (Figs 2 to 4). Pelvic obliquity was corrected by 11 (48%) to 10, but with significantly better corrections (p < 0.01) in the SMA (55%) compared with the DMD patients (35%). All patients lost some correction during the first year, after which they stabilised. The mean loss of correction of the Cobb angle in the whole group at final review was 7. Pelvic correction appeared to be more stable with a mean loss of correction at final follow-up of 3. There was a greater recurrence of pelvic tilt in those patients not instrumented to the sacrum. Again, there was no significant difference between the subgroups. Stabilisation of the spine was seen within the first two years with no difference between the modified Moe fusion group and the rest. Genetic studies revealed no correlation between the location of the intragenic deletion and severity of the curve, blood loss or surgical outcome. Most patients (96%) were free of bracing at the time of final review, which was the major achievement of surgery. Satisfaction was reported as excellent by 89.6% of the patients, who referred in particular to the correction of deformity, freedom from bracing, improved function, sitting balance and independence. There was a marked improvement in seating tolerance and length of time seated. Many parents and carers reported less buttock pain and pressure problems. Complications. The complications are summarised in Table III. Most patients had a degree of respiratory compromise and required ventilation for 12 to 48 hours. The chest infections which occurred were brief. The three pneumothoraces occurred early in the series and included the one patient who suffered a cardiac arrest. Temporary paralytic ileus was common and we saw one case of perforation and one of ischaemia of the bowel; neither was explained but both recovered. Trunk hyperaesthesiae, minor weaknesses of the distal limb or bladder problems occurred commonly in the early severe cases of DMD in whom very large corrections of severe curves were achieved. Late complications were few but included one case of breakage of both rods, resulting in some loss of correction, and four in which instrumentation had to be extended to the pelvis after three years because of increasing curvature below the fusion. There were no perioperative deaths and no deep infections; 14 of the patients have died since surgery because of their underlying neuromuscular condition. None of the deaths has been attributed to the surgery. THE JOURNAL OF BONE AND JOINT SURGERY

4 THE TREATMENT OF SCOLIOSIS IN MUSCULAR DYSTROPHY USING MODIFIED LUQUE INSTRUMENTATION 25 Radiographs showing a) the preoperative AP view of the spine of a 13-year-old boy with DMD with a Cobb angle of 125 and rib impingement on the pelvis, and b) three years after Luque instrumentation and Moe fusion. The curve measures 64 and sitting balance was restored. Fig. 2a Fig. 2b Radiographs showing a) the preoperative AP view of the spine of a ten-year-old boy with spinal muscular atrophy, measuring 107, who was still walking and b) two years after Harrington-Luque instrumentation and Moe fusion with correction of spinal stability and a curve of 67. Fig. 3a Fig. 3b Discussion The use of segmental spinal instrumentation was described by Luque and Cardoso. 9 There have been several reports of the use of this method of spinal stabilisation in neuromuscular scoliosis, but these have been mainly for cerebral palsy, poliomyelitis and spina bifida. 8,10-12 Our unit has a considerable experience in the management of scoliosis in patients with muscular dystrophy. These patients have many special needs and present considerable difficulties, above and beyond those seen in other types of scoliosis. 13,14 Our study confirms the early work of Rideau et al, 1 who described the high risk for anaesthesia because of poor respiratory function and general physical condition, as revealed by the ASA scores. Nevertheless surgery is indicated and possible for those with a vital capacity of less VOL. 83-B, NO. 1, JANUARY 2001

5 26 G. BENTLEY, F. HADDAD, T. M. BULL, D. SEINGRY Radiographs showing a) the preoperative AP spinal view of a 12-year-old boy with DMD late in the series and b) one year after Luque instrumentation and Moe fusion. The 25 curve is clinically negligible. Fig. 4a Fig. 4b Table III. Complications after surgery for scoliosis in 101 patients with muscular dystrophy Number of Mean duration Complication patients (range) Comments Perioperative Cardiac arrest 1 Resuscitated without sequelae Pneumothoraces 3 2 required chest drains Ventilation 1 4 weeks Weaned from ventilator with no long-term ill-effects Early (mild) Paralytic ileus 28 7 days (2 to 21) Trunk hyperaesthesia days (2 to 60) Distal numbness 5 15 days (7 to 21) Distal weakness 3 5 days (4 to 6) Bladder dysfunction 8 10 days (4 to 21) Chest infection 12 4 days (2 to 14) Superficial wound infection 6 7 days (1 to 13) Early (severe) Colonic perforation 1 Laparotomy recovered Distal numbness 1 4 months Recovered Distal numbness/weakness 1 6 months Recovered Bladder dysfunction 4 12 weeks Recovered Permanent bladder dysfunction 1 Borderline function preoperatively Late Protruding rods 2 Trimmed Fusion extended to pelvis 4 Satisfactory stability Broken rods 2 No intervention necessary Prominent wires 1 Trimmed Decreased mobility 2 Some hip pain or seating problems 13 Mainly solved by suitable seating than 35%, but a specialist team of anaesthetists, nurses and physiotherapists is essential in order to deal with respiratory complications. 15,16 The measured blood loss in patients with DMD was significantly higher than that in patients with SMA. 7 Blood loss is directly related to the duration of surgery. As children with DMD lack dystrophin in all muscle types, including smooth muscle, we postulate that the excessive blood loss is due to poor vasoconstrictive response in smooth muscles caused by a lack of dystrophin. We also observed no correlation between the measured blood loss and the type of gene mutation causing DMD. This, and the fact that we could not predict progression of the curve or surgical outcome from the patient s chromosomal abnormality, highlights the fact that most of the lesions totally disrupt the expression of dystrophin and therefore lead to very similar musculoskeletal manifestations. Monitoring of the spinal cord is now accepted as an THE JOURNAL OF BONE AND JOINT SURGERY

6 THE TREATMENT OF SCOLIOSIS IN MUSCULAR DYSTROPHY USING MODIFIED LUQUE INSTRUMENTATION 27 essential part of surgery for scoliosis, and was used in our series. 6 The pre-existing motor compromise necessitated careful preoperative assessment of muscle function so that subtle changes could be noted subsequently. Most temporary neurological deficits after operation were seen early in the study when the most severe curves were treated. The goal of surgery was a stable maximally-corrected spine above a level pelvis, providing freedom from bracing with improved function and comfort. Luque segmental instrumentation was chosen as the treatment of choice at the beginning of the study since it was an established technique providing secure fixation with minimum morbidity. The quality of bone in patients with muscular dystrophy is often poor, but the laminae are sufficiently robust to support sublaminar wires. Our study shows that the technique is effective for neuromuscular scoliosis, even although it is now being used more rarely in other areas. Blood loss is another important consideration in DMD. The duration of surgery should be kept to a minimum; pedicle screw fixation at multiple levels requires fluoroscopic control and would prolong the operating time considerably. Technique of fusion. Decortication was not carried out in this series and in 27 patients Moe fusion was not undertaken in case of excessive blood loss and the wish to shorten the duration of operation. It became apparent that this did not affect the outcome, as has previously been noted by Bellen et al. 19 These patients lost some slight correction during the first year. It has been observed that improved stabilisation encourages fusion, and it is therefore apparent that the near-rigid Luque technique suffices. Sufficient bone is produced by periosteal stripping to fuse the spine. In three patients with discomfort from prominent wires the spine was explored at a later stage. All were found to have a sound fusion, despite the absence of decortication and bone grafting. This suggests that bone grafting and extensive formal fusion are not necessary; it is our policy to carry out only a modified Moe fusion as described. Eberle 11 reported poor results when fusion was not carried out, but that series comprised younger patients with poliomyelitis, and less stiff rods of 4.8 mm diameter were used in most. Pelvic stabilisation. Spinal instrumentation should not be extended to include the whole of the lumbar spine and pelvis in ambulant patients who walk by tilting the pelvis from side to side. Pelvic stabilisation interferes with mobility as demonstrated by Furumasu et al. 20 The cross-wire technique is, however, very useful for correcting pelvic tilt. This has proved important for patients with DMD in which pelvic tilt is often severe and leads to sitting difficulties when they become wheelchair-dependent, which was the case in all patients with DMD in our series. There has been considerable debate in the literature over the need for lumbosacral stabilisation. 8,21-26 The protagonists describe better maintenance of pelvic correction and improved sitting. Lumbosacral fusion also allows greater correction of a curve and better control of sagittal alignment. 26 Secondary surgery was required to extend the fusion to the pelvis early in the series in four patients, whose pelvic obliquity increased rapidly after instrumentation to L3 or L4 vertebrae. The loss of pelvic correction has been reduced by the use of the cross-wire technique. Instrumentation to the pelvis did not increase the incidence of early complications in our series or in that of Ramirez et al. 13 We therefore recommend stabilisation of the pelvis in all cases of DMD. Although instrumentation as high as T2 has been advocated, 16,27 we went no higher than T4 in most cases because this seemed to allow slightly more freedom of movement of the cervicodorsal region, facilitating function of the upper limb and feeding. Some mild curves developed above the level of the fusion, but none that required further surgical intervention. Difficulties with control of the head frequently improved after stabilisation of the spine, but were managed by orthotics and alterations in seating when neccessary. The patients early in our series had severe curves and were, generally, in a worse condition than the later ones. This is explained by an initial reluctance to operate on such high-risk patients. As our unit developed its links with the Muscular Dystrophy Clinic at the Hammersmith Hospital, the patients were referred much earlier. This has led to treatment of less severe curves, obtaining better corrections with a lower morbidity. Our overall results, however, are still affected by the initial cohort. Many attempts have been made to identify factors which may predict progression of the curve and outcome in DMD. 28 Ramirez et al 13 noted that those who died in the first two years after surgery had shown more preoperative respiratory compromise, which probably reflected the severity of the disease. 13 We attempted to link chromosomal lesions, severity of the curve and preoperative respiratory status with outcome, but were unable to establish a clear relationship. Our study shows that most patients with muscular dystrophy undergoing surgical correction of scoliosis, had good objective results with correction of deformity, abandonment of a brace, improved sitting balance and mobility. Eventually, 96% were very satisfied with the outcome, even although some had a difficult postoperative recovery. It is now common for new patients to request surgery because a friend at the Muscular Dystrophy Clinic has recommended it. The freedom from bracing and improvement in seating posture are the most valued benefits. Patients with DMD, particularly, undergo a rapid decline in respiratory function when the scoliosis progresses. 29 There was a general impression that they had fewer chest infections and their respiratory function was slower to deteriorate after surgery. If further studies were to confirm this, its effect would be to prolong life. 29 Complications in our series were related to the underlying condition, and were usually mild and transient. VOL. 83-B, NO. 1, JANUARY 2001

7 28 G. BENTLEY, F. HADDAD, T. M. BULL, D. SEINGRY We consider that muscular dystrophy, even when advanced, should not exclude a patient from early surgical intervention when scoliosis develops. Our youngest patient was aged five years. This type of surgery should be carried out in specialist centres with experienced surgical, anaesthetic and support staff. Luque segmental spinal stabilisation has proved of value specifically for the mobile collapsing spine in patients with neuromuscular disease, especially with respiratory compromise, and will continue to be the method of choice in our unit for patients with mobile curves of 90 or less. All of the patients in this series were managed jointly in the Neuromuscular Clinic at the Hammersmith Hospital headed by Professor V. Dubowitz, MD, FRCP to whom we are most grateful for much advice over the course of the study. Also thanks to Mr A. Manktelow who gave valuable assistance in the preparation of the paper. 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. References 1. Rideau Y, Glorion B, Delaubier A, Tarle O, Bach J. The treatment of scoliosis in Duchenne muscular dystrophy. Muscle Nerve 1984;7: Bhatnaghar RS, Noordeen MHH, Muntoni F, Bentley G. Smooth muscle involvement in Duchenne muscular dystrophy. J Bone Joint Surg [Br] 1995;77-B:Suppl 1, Allen BL Jr, Ferguson RL. The Galveston technique for L rod instrumentation of the scoliotic spine. Spine 1982;7: Moe JH. Method of correction and surgical techniques in scoliosis. Orthop Clin North Am 1972;3: Dubowitz V. Some clinical observations on childhood muscular dystrophy. Br J Clin Pract 1963;17: Noordeen MHH, Lee J, Gibbons CER, Taylor BA, Bentley G. Spinal cord monitoring in operations for neuromuscular scoliosis. J Bone Joint Surg [Br] 1997;79-B: Swank SM, Brown JC, Perry RE. Spinal fusion in Duchenne s muscular dystrophy. Spine 1982;7: Broom MJ, Banta JV, Renshaw TS. Spinal fusion augmented by Luque-rod segmental instrumentation for neuromuscular scoliosis. J Bone Joint Surg [Am] 1989;71-A: Luque ER, Cardoso A. Segmental correction of scoliosis with rigid internal fixation: preliminary report. Orthop Trans 1977;1: Herndon WA, Sullivan JA, Yngve DA, Gross RH, Dreher G. Segmental spinal instrumentation with sublaminar wires: a critical appraisal. J Bone Joint Surg [Am] 1987;69-A: Eberle CF. Failure of fixation after segmental spinal instrumentation without arthrodesis in the management of paralytic scoliosis. J Bone Joint Surg [Am] 1988;70-A: Gau YL, Lonstein JE, Winter RB, Koop S, Denis F. Luque- Galveston procedure for correction and stabilisation of neuromuscular scoliosis and pelvic obliquity: a review of 68 patients. J Spinal Disord 1991;4: Ramirez N, Richards BS, Warren PD, Williams GR. Complications after posterior spinal fusion in Duchenne muscular dystrophy. J Pediatr Orthop 1997;17: Granata C, Merlini L, Cervellati S, et al. Long term results of spine surgery in Duchenne muscular dystrophy. Neuromuscul Disord 1996;6: Galasko CSB, Delaney C, Morris P. Spinal stabilisation in Duchenne muscular dystrophy. J Bone Joint Surg [Br] 1992;74-B: Shapiro F, Sethna N, Colan S, Wohl ME, Specht L. Spinal fusion in Duchenne muscular dystrophy: a multidisciplinary approach. Muscle Nerve 1992;15: Forbes HJ, Allen PW, Waller CS, et al. Spinal cord monitoring in scoliosis surgery: experience with 1168 cases. J Bone Joint Surg [Br] 1991;73-B: Williamson JB, Galasko CSB. Spinal cord monitoring during operative correction of neuromuscular scoliosis. J Bone Joint Surg [Br] 1992;74-B: Bellen P, Hody JL, Clairbois J, Denis N, Soudon PH. Surgical treatment of spinal deformities in Duchenne muscular dystrophy. Rev Chir Orthop Reparatrice Appar Mot 1992;78: Furumasu J, Swank SM, Brown JC, et al. Functional activities in spinal muscular atrophy patients after spinal fusion. Spine 1989;14: Brook PD, Kennedy JD, Stern LM, Sutherland AD, Foster BK. Spinal fusion in Duchenne muscular dystrophy. J Pediatr Orthop 1996;16: LaPrade RF, Rowe DE. The operative treatment of scoliosis in Duchenne muscular dystrophy. Orthop Rev 1992;21: Mubarak SJ, Morin WD, Leach J. Spinal fusion in Duchenne muscular dystrophy: fixation and fusion to the sacropelvis? J Pediatr Orthop 1993;13: Sussman MD. The operative treatment of scoliosis in Duchenne muscular dystrophy. Orthop Rev 1992;21: Sussman MD, Little D, Alley RM, McCoig JA. Posterior instrumentation and fusion of the thoracolumbar spine for treatment of neuromuscular scoliosis. J Pediatr Orthop 1996;16: Miller F, Moseley CF, Koreska J. Spinal fusion in Duchenne muscular dystrophy. Dev Med Child Neurol 1992;34: Smith AD, Koreska J, Moseley CF. Progression of scoliosis in Duchenne muscular dystrophy. J Bone Joint Surg [Am] 1989;71-A: Miller F, Moseley CF, Koreska J, Levison H. Pulmonary function and scoliosis in Duchenne dystrophy. J Pediatr Orthop 1988;8: Kennedy JD, Staples AJ, Brook PD, et al. Effect of spinal surgery on lung function in Duchenne muscular dystrophy. Thorax 1995;50: THE JOURNAL OF BONE AND JOINT SURGERY