Posterior spinal arthrodesis for adolescent idiopathic scoliosis using pedicle screw instrumentation

Transcription

1 SPINE Posterior spinal arthrodesis for adolescent idiopathic scoliosis using pedicle screw instrumentation DOES A BILATERAL OR UNILATERAL SCREW TECHNIQUE AFFECT SURGICAL OUTCOME? A. I. Tsirikos, A. S. Subramanian From Scottish National Spine Deformity Centre, Royal Hospital for Sick Children, Edinburgh, United Kingdom We reviewed 212 consecutive patients with adolescent idiopathic scoliosis who underwent posterior spinal arthrodesis using all pedicle screw instrumentation in terms of clinical, radiological and Scoliosis Research Society (SRS)-22 outcomes. In Group 1 (51 patients), the correction was performed over two rods using bilateral segmental pedicle screws. In Group 2 (161 patients), the correction was performed over one rod using unilateral segmental pedicle screws with the second rod providing stability of the construct through two-level screw fixation at proximal and distal ends. The mean age at surgery was 14.8 years in both groups. Comparison between groups showed no significant differences with regard to age and Risser grade at surgery, pre- and post-operative scoliosis angle, coronal Cobb correction, length of hospital stay and SRS scores. Correction of upper thoracic curves was significantly better in Group 1 (p = 0.02). Increased surgical time and intra-operative blood loss was recorded in Group 1 (p < and p = 0.04, respectively). The implant cost was reduced by mean 35% in Group 2 due to the lesser number of pedicle screws. Unilateral and bilateral pedicle screw techniques have both achieved excellent deformity correction in adolescent patients with idiopathic scoliosis, which was maintained at twoyear follow-up. This has been associated with high patient satisfaction and low complication rates. A. I. Tsirikos, MD, FRCS, PhD, Consultant Orthopaedic and Spine Surgeon, Honorary Clinical Senior Lecturer A. S. Subramanian, FRCS(Tr & Orth), Senior Trainee in Orthopaedics, Spine Fellow Scottish National Spine Deformity Centre, Royal Hospital For Sick Children, Sciennes Road, Edinburgh EH9 1LF, UK. Correspondence should be sent to Mr A. I. Tsirikos; atsirikos@hotmail.com 2012 British Editorial Society of Bone and Joint Surgery doi: / x.94b $2.00 J Bone Joint Surg Br 2012;94-B: Received 20 February 2012; Accepted after revision 8 August 2012 Adolescent idiopathic scoliosis (AIS) is the most common type of spinal deformity affecting the coronal, sagittal and axial planes, often producing a severe curvature. 1 Surgical treatment evolved from the initial concave Harrington rods that applied distraction at the proximal and distal ends of the curve, through to the Luque rods and Cotrel-Dubousset instrumentation that increased the fixation anchors and exerted forces bilaterally across multiple vertebrae in three-dimensional planes. The use of sublaminar wire and hook fixation as well as bilateral rods allowed for improved correction of the deformity and increased stability of the construct. Pedicle screws were initially introduced in the lumbar spine combined with mainly thoracic hooks to form hybrid constructs. Pedicle screws provide three-column rigid fixation and apply force vectors in all three planes to allow for correction of the coronal and sagittal imbalance, as well as part of the associated vertebral rotation. Pedicle screw instrumentation systems have been increasingly used because of their superior biomechanical properties, their relative safety of insertion, and effectiveness in correcting complex deformities. 2-5 Other potential benefits of pedicle screw constructs include lower rates of pseudarthrosis, reduced loss of correction of the deformity, greater pull-out strength, lesser risk of neurological complications and preservation of more mobile spinal segments. 6-8 It has been shown previously that pedicle screw instrumentation is capable of correcting scoliotic curves of 100 without the need for anterior surgery, thus avoiding the associated morbidity. 3,9,10 The superiority of pedicle screws to correct scoliosis has resulted in a tendency to increase the anchor points and use bilateral segmental pedicle screw fixation. It has been suggested that increasing the number of pedicle screws within the construct could achieve better correction of deformity, 11 but as the number of pedicle screws increase, so does the risk of screw malposition impinging on neurological or vascular structures and causing permanent injury. The purpose of this study was to review retrospectively a single surgeon s experience of 1670 THE JOURNAL OF BONE AND JOINT SURGERY

2 POSTERIOR SPINAL ARTHRODESIS FOR ADOLESCENT IDIOPATHIC SCOLIOSIS USING PEDICLE SCREW INSTRUMENTATION 1671 treating patients with AIS using pedicle screw constructs and to compare the efficacy of unilateral and bilateral correction techniques as assessed by clinical, radiological and functional results. To our knowledge, this is the largest consecutive single surgeon s series of AIS patients treated with pedicle screw instrumentation and determining whether implant density and technique can affect surgical outcome. Patients and Methods The medical records and radiographs of 212 consecutive patients with AIS treated by posterior spinal arthrodesis between 2006 and 2010, using pedicle screw/rod instrumentation, were reviewed. Inclusion criteria were: a) patients with AIS (all Lenke types 12 ) who were treated with isolated posterior spinal arthrodesis; b) all pedicle screw constructs using extended tab polyaxial screws with the ability to lock into monoaxial screws and attached to 5.5 mm titanium rods (DePuy Expedium 5.5 mm dual innie favoured angle reduction screws; DePuy Spine, Raynham, Massachusetts); c) absence of a thoracoplasty; d) use of autologous bone graft supplemented by allograft bone when required; e) minimum two-year post-operative follow-up with clinical, radiological and functional outcomes using the Scoliosis Research Society (SRS)-22 questionnaire. 13 Local institutional review board approval was obtained for the study. There were 188 females and 24 males, who were divided into two groups. Group 1 (51 patients; 42 female, nine male) in which the deformity was corrected over two rods, using bilateral segmental pedicle screw fixation; and Group 2 (161 patients; 143 female, 18 male) in which the scoliosis was corrected over one rod ( corrective rod ) using unilateral segmental pedicle screws, with the second rod ( supportive rod ) providing stability to the construct through screw fixation at the proximal and distal ends. There were no selection criteria for patients included in Group 1 and Group 2, as Group 1 was the historical series of the senior author. The trend in the literature 14,15 for bilateral segmental screw fixation at the time of initial patient enrolment to the study influenced the author s practice towards the use of bilateral segmental screws with further refinement of the correction technique and reduction in the number of screw anchors in Group 2. The patients gender, age, Risser grade, 16 scoliosis angle, thoracic kyphosis (T5 T12), and lumbar lordosis (L1 L5) at the time of surgery, post-operatively and at latest follow-up were recorded. All spinal radiographs were measured using the Cobb method 17 on an electronic system by both authors based on a consensus decision on anatomical landmarks. Spinal radiographs were taken in the standing position and in the supine maximum lateral bending position. The same anatomical landmarks were used for measurement of the curve. The following measurements were calculated: Flexibility index (FI, %) = 100 ([pre-operative Cobb angle supine maximum lateral bending Cobb angle] / preoperative Cobb angle) Correction rate (%) = 100 ([pre-operative Cobb angle post-operative Cobb angle] / pre-operative Cobb angle) All patients had pre-operative MRI scans that excluded intraspinal anomalies. Duration of the operation (skin incision to skin closure), intra-operative blood loss (both in ml and in blood volumes, defined as body weight in kilograms 70 ml for patients older than ten years) and length of hospital stay were recorded (Table I). Complications during scoliosis surgery, as well as in the post-operative period were documented. Early post-operative complications occurred within three months and late beyond three months after surgery. SRS-22 outcomes were obtained before surgery, as well as at six months and two years post-surgery. The SRS questionnaires were completed by the patients via a telephone interview conducted by a specialist nurse. Surgical technique. An anatomical exposure of the spine was performed using subperiosteal dissection of the paraspinal muscles from the midline out to the tips of the transverse processes. All patients in this series underwent complete inferior facet resections at every level across the thoracic spine, with associated decortication of the superior facets before bone graft was placed. The capsule and opposing cartilage was removed from the lumbar facets, which were preserved to allow for interfacetal bone grafting. The spinous processes were excised across the levels of the fusion to further mobilise the spine and provide bone in addition to harvested autologous iliac crest graft. The ligamentum flavum was preserved at every level and the spinal canal was not violated. In Group 1, pedicle screws were inserted bilaterally at every level to be included in the arthrodesis. In Group 2, pedicle screws were used segmentally on one side, whereas only two screws were placed at the cephalad and caudal ends of the construct on the contralateral side in order to allow for a wider area across the levels in between for bone grafting. In this group, for single thoracic or thoracolumbar/lumbar scoliosis the corrective rod with segmental fixation was always placed on the convexity of the curve, with the supportive rod being on the concavity. For double thoracic and lumbar scoliosis the corrective rod was placed on the convexity of the thoracic curve and the concavity of the lumbar curve. For double thoracic scoliosis the corrective rod was placed on the convexity of the main thoracic curve and the concavity of the upper thoracic curve. All pedicle screws were inserted using a free hand technique and confirmed with intra-operative fluoroscopic imaging before rod placement. The pedicle entry point from T2 to T5 was at the junction of the bisected transverse process and the lateral margin of the facet joint. From T6 to T10 it was at the junction of the upper to the middle third of the transverse process and the lateral margin of the facet joint. At T11 and T12 the entry point was at the base of the superior facet and in the lumbar spine it was at the junction of the transverse process, pars interarticularis and superior facet. The long tab polyaxial VOL. 94-B, No. 12, DECEMBER 2012

3 1672 A. I. TSIRIKOS, A. S. SUBRAMANIAN Table I. The characteristics and surgical results in Groups 1 and 2. Data are shown as mean and range, and p-values are by t-test Group 1 Group 2 p-value Female:male 42:9 143:18 - Age at surgery (yrs) 14.8 (10.9 to 19.4) 14.8 (11 to 23) 0.92 Mean Risser grade 2.2 (0 to 5) 2.3 (0 to 5) 0.88 Upper thoracic Pre-operative ( ) 37.5 (26 to 57) 42 (30 to 61) 0.24 Flexibility index (%) 32 (20 to 60) 28 (10 to 52) 0.65 Post-operative ( ) 11 (0 to 22) 19 (11 to 30) 0.03 Correction (%) 71 (47 to 100) 55 (31 to 74) 0.02 Main thoracic Pre-operative ( ) 68 (40 to 98) 65 (38 to 95) 0.57 Flexibility index (%) 43 (25 to 64) 39 (19 to 60) 0.64 Post-operative ( ) 20 (0 to 43) 19 (5 to 45) 0.86 Correction (%) 71 (37 to 100) 70 (39 to 91) 0.85 Thoracolumbar/lumbar Pre-operative ( ) 61 (38 to 84) 59 (32 to 105) 0.92 Flexibility index (%) 53 (42 to 70) 57 (23 to 80) 0.71 Post-operative ( ) 16 (0 to 35) 17 (0 to 40) 0.63 Correction (%) 74 (58 to 100) 71 (62 to 100) 0.29 Thoracic kyphosis ( ) Pre-operative 26 (4 to 60) 30 (5 to 63) 0.26 Post-operative 22 (8 to 38) 27 (10 to 45) 0.15 Loss of TH kyphosis 4 (2 to 8) 3 (2 to 6) 0.89 Lumbar lordosis ( ) Pre-operative 52 (35 to 65) 47 (25 to 69) 0.16 Post-operative 50 (37 to 62) 44 (30 to 62) 0.22 Loss of lumbar lordosis 2 (1 to 4) 3 (2 to 4) 0.78 Operating time (mins) 320 (240 to 420) 240 (150 to 350) < Blood loss (ml) 1679 (700 to 4500) 1343 (400 to 3000) 0.04 Blood volume loss 0.5 (0.2 to 1.1) 0.3 (0.1 to 0.8) 0.04 Length of hospital stay (days) 8.3 (7 to 12) 8.4 (6 to 22) 0.55 screws allowed for easier engagement of the rods, which were pre-contoured to accommodate thoracic kyphosis and lumbar lordosis. In Group 1, the scoliosis was corrected using a translocation manoeuvre combined with segmental vertebral derotation using the screws converted into monoaxial against the two rods (Fig. 1). In Group 2, sequential correction was performed using the same technique on the side with the segmental screw anchors while the second supportive rod was placed after correction to augment the construct (Fig. 2). In both groups distraction and compression was performed before finalising the construct at the proximal and distal ends as needed in order to level the shoulders and lumbar spine. Autologous iliac crest bone was harvested through the same midline incision and placed across the excised facet joints, transverse processes and laminae to achieve a fusion. Iliac crest bone was harvested in 202 patients. In ten patients with a single or double thoracic scoliosis in whom the fusion involved only the thoracic spine, locally harvested bone was used and then supplemented by allograft bone. The laminae and transverse processes were decorticated and bone graft was used to achieve a solid fusion. Cross connectors between the rods were not used and bone substitutes were not added in any patient. Intraoperative spinal cord monitoring (IOM) was performed recording transcranial motor and somatosensory (cortical and cervical) evoked potentials. Hypotensive anaesthesia and a cell saver were used for all patients. Selection of the levels of fusion was based on clinical examination in conjunction with the flexibility of the curve as determined by the maximum side bending radiographs. Structural curves producing a clinical deformity particularly when the patient was bending forward to 90 with prominent thoracic or waistline line asymmetry were generally included in the fusion. For double structural thoracic and lumbar scoliosis the fusion extended distally to the stable vertebra or the one immediately above the stable vertebra (stable minus one). Selective fusions were performed in patients with primary thoracolumbar/lumbar and compensatory thoracic curves (thoracolumbar fusion) and in those with primary thoracic and compensatory lumbar scoliosis with lumbar modifier A 12 (thoracic fusion). Selective fusions were not performed in this cohort when the patients were skeletally immature with open tri-radiate cartilage, Risser grade 0 and pre-menarchal. Statistical analysis. This was performed using the twotailed t-test between pre-operative and post-operative radiological measurements within each group. The two- THE JOURNAL OF BONE AND JOINT SURGERY

4 POSTERIOR SPINAL ARTHRODESIS FOR ADOLESCENT IDIOPATHIC SCOLIOSIS USING PEDICLE SCREW INSTRUMENTATION 1673 tailed unpaired t-test was used to compare the data between Groups 1 and 2. A p-value < 0.05 was considered statistically significant. The Pearson correlation coefficient was calculated to assess the strength of linear dependence between intragroup parameters such as scoliosis angle, surgical time, intra-operative blood loss, age at surgery and length of hospital stay, as well as FI and deformity correction. An r- value of > 0.2 was considered statistically significant. Results The characteristics of the two groups of patients and their surgical results are presented in Table I. Mean postoperative follow-up was 3.5 years (2.5 to 6) with all patients being skeletally mature at latest review (Risser grade 5). The Lenke curve types for both groups are given in Table II. There were 89 patients with major thoracic (single or double), 56 patients with major thoracolumbar/ lumbar, and 67 patients with double major thoracic and lumbar curvatures. There were a greater number of Lenke 5 curves in Group 2, which is an incidental finding and reflects the senior author s surgical activity during the recorded period. During the study period all patients with Lenke 5 curves in Groups 1 and 2 were routinely treated by a posterior rather than an anterior spinal arthrodesis using a pedicle screw technique. The lumbar modifier for double major thoracic and lumbar scoliosis was 8B, 10C in Group 1 and 21B, 28C in Group 2. In all, ten patients (4.7%) had isthmic spondylolysis and associated lumbosacral spondylolisthesis (two in Group 1, eight in Group 2). This was symptomatic Meyerding Grade 2 to 3 18 in five patients, with a thoracic scoliosis, who underwent a posterolateral lumbosacral arthrodesis using iliac crest bone, without instrumentation, through a Wiltse approach 19 at the same time as the scoliosis procedure. The presence of the spondylolisthesis did not affect the levels of scoliosis surgery or the correction achieved. In Group 1 a higher degree of main thoracic scoliosis correlated with a lower FI (r = -0.4) (Table I). The degree of scoliosis correction correlated with the FI for main thoracic (r = 0.9) and thoracolumbar/lumbar curves (r = 0.8). There was no correlation between the amount of scoliosis correction (% or degrees) and the post-operative loss of thoracic kyphosis in patients with primary thoracic curves (r = -0.1 and -0.09, respectively). Pre-operative scoliosis size of upper thoracic and thoracolumbar/lumbar curves correlated with increased surgical time (r = 0.6 and 0.2, respectively). Pre-operative scoliosis size of main thoracic and thoracolumbar/lumbar curves weakly correlated with increased intra-operative blood loss (r = 0.2 and 0.2, respectively). Pre-operative upper thoracic scoliosis size strongly correlated with age at surgery (r = 0.8) and length of hospital stay (0.9). In Group 2 a higher degree of main thoracic scoliosis correlated with a lower FI (r = -0.4) (Table I). The degree of scoliosis correction correlated weakly with the FI for Table II. Types of curves in Groups 1 and 2 according to the Lenke classification 12 Lenke type of scoliosis Group 1 Group main thoracic (r = 0.2) and thoracolumbar/lumbar curves (r = 0.2). There was no correlation between the amount of scoliosis correction (% or degrees) and the post-operative loss of thoracic kyphosis in patients with primary thoracic curves (r = and -0.12, respectively). Pre-operative scoliosis size for all types of curves correlated with increased surgical time (r = 0.6, 0.4 and 0.4, respectively). Pre-operative scoliosis size of main thoracic and thoracolumbar/lumbar curves weakly correlated with increased intra-operative blood loss (r = 0.3 and 0.3, respectively). Complications are shown in Table III. In Group 1, one patient with a 95 thoracic scoliosis had transient true positive loss of motor traces on IOM during preparation of the screw canal using the pedicle probe for insertion of an apical concave T8 screw. The screw was not inserted, the motor traces returned to normal and the patient had no neurological deficit. One patient developed a late deep wound infection that presented with an abscess at 11 months post-operatively. She underwent surgical debridement and instrumentation removal. There was evidence of nonunion, which was treated with revision surgery after six months, resulting in a good outcome. One patient required trimming of a prominent proximal rod end at two years due to muscular pain. In Group 2, one patient each developed an early deep and an early superficial wound infection treated with surgical debridement and three months of antibiotics with no evidence of recurrence and no loss of scoliosis correction as the instrumentation was retained. One patient had transient brachial plexus neuropraxia due to intra-operative positioning. One patient developed superior mesenteric artery syndrome that resolved with conservative treatment. The data regarding the assessment of quality of life are shown in Figures 3 and 4. In Group 1, the mean pre-operative SRS-22 score was 3.9 (3.5 to 4.1), which improved significantly to 4.5 (4.3 to 4.8) at follow-up (p < 0.001). Pain and self-image demonstrated significant improvements (p = and p < 0.001, respectively) with a mean satisfaction score of 4.9 (4.8 to 5). In Group 2, the mean pre-operative SRS-22 score was 3.7 (3.3 to 4.1), improving significantly to 4.5 (4.2 to 4.8) at follow-up (p < 0.001). Pain, function, self-image and mental health demonstrated significant improvements (all p < 0.001), with mean satisfaction score of 4.8 (4.7 to 5). Comparison between Groups 1 and 2 showed no significant difference in regard to age (p = 0.92; Table I) and VOL. 94-B, No. 12, DECEMBER 2012

5 1674 A. I. TSIRIKOS, A. S. SUBRAMANIAN Table III. Complications Complications Group 1 Group 2 Transient loss of intra-operative spinal cord 1 - monitoring traces/no neurological deficit Deep wound infection 1 (late: 11 months post-operatively, 1 (early: surgical debridement/antibiotics) nonunion requiring revision surgery) Superficial wound infection - 1 (early: surgical debridement/antibiotics) Prominent instrumentation 1 (trimming of upper end of rod) - Transient brachial plexus neuropraxia - 1 Superior mesenteric artery syndrome - 1 (resolved with conservative treatment) Fig. 1a Fig. 1b Fig. 1c Fig. 1d Pre-operative a) posteroanterior and b) lateral radiographs of the spine in a patient aged 12.4 years, showing a severe thoracic and lumbar scoliosis, as well as thoracic hypokyphosis producing a negative sagittal balance of the spine. Post-operative c) posteroanterior and d) lateral radiographs at three years after posterior spinal arthrodesis using a bilateral segmental pedicle screw technique, showing satisfactory correction of both curves. There was some degree of negative sagittal imbalance, but the patient has no complaints regarding her back and took part in normal activities including sports. Risser grade at surgery (p = 0.88; Table I), pre- and postoperative main thoracic and thoracolumbar/lumbar Cobb angle, length of hospital stay (Table I) and SRS scores. Correction of upper thoracic curves was significantly better in Group 1 (p = 0.02). Increased surgical time and intra-operative blood loss was found in Group 1 (p<0.001 and p=0.04, respectively). The implant cost was reduced by a mean of 35% (18% to 41%) depending on the type of curve and length of arthrodesis in Group 2, due to the lesser number of pedicle screws (implant density of 2 and 1.38 (1.2 to 1.8) in groups 1 and 2, respectively). Discussion Pedicle screw instrumentation has been used increasingly in the treatment of AIS with better correction of the deformity when compared with hook or hybrid constructs. 11,15,20 Pedicle screw placement can be performed using a free-hand technique, radiological imaging or navigation guidance. We routinely place all pedicle screws through a free-hand technique using anatomical landmarks as described earlier. In our experience, there is a steep learning curve for insertion of thoracic pedicle screws. Even though we did not check the accuracy of screw positioning with CT scans, we had no post-operative pedicle screw-related neurological or visceral complications with 3202 pedicle screws inserted in this cohort. The construct selected for Group 2 used a convex rod with segmental screw fixation in order to avoid higher-risk concave pedicle screws at the apex of the thoracic curve. Scoliosis correction over a convex rod is a novel idea and has two distinct advantages: 1) it requires placement of convex pedicle screws, which carry lesser risk of pedicle wall penetration, as the convex apical pedicles are wider than those on the concave side and the spinal cord is positioned adjacent to the medial wall of the apical concave pedicles, leaving it susceptible to injury; and 2) the concave supportive rod is sitting higher than the spine, especially across the apex of the scoliosis as this is attached to the vertebrae only proximally and distally, allowing for a large THE JOURNAL OF BONE AND JOINT SURGERY

6 POSTERIOR SPINAL ARTHRODESIS FOR ADOLESCENT IDIOPATHIC SCOLIOSIS USING PEDICLE SCREW INSTRUMENTATION 1675 Fig. 2a Fig. 2b Fig. 2c Fig. 2d Pre-operative a) posteroanterior and b) lateral radiographs of the spine in a patient aged 13.4 years, showing a severe thoracic and lumbar scoliosis with associated thoracic hypokyphosis. Post-operative c) posteroanterior and d) lateral radiographs at 3.5 years after posterior spinal arthrodesis using a unilateral segmental pedicle screw technique, showing satisfactory correction of the deformity and a good sagittal balance of the spine. SRS score SRS 22 Function Pain Self image Mental health Pre-operative 6 months 2 years SRS score SRS 22 Function Pain Self image Mental health Pre-operative 6 months 2 years Fig. 3 Fig. 4 Bar chart showing the Scoliosis Research Society (SRS)-22 outcomes for Group 1 (pre-operatively, and at six months and two years post-operatively). The satisfaction rate at two years was 4.9. Bar chart showing the Scoliosis Research Society (SRS)-22 outcomes for Group 2 (pre-operatively, and at six months and two years post-operatively). The satisfaction rate at two years was 4.8. area for bone grafting and reducing the risk of nonunion. We performed the same deformity reduction technique using a concave corrective and convex supportive rod in patients not included in this cohort applying identical surgical principles and did not record any difference in terms of ability to correct the scoliosis or restore thoracic kyphosis when compared with the present series. The use of extended tab reduction screws allowed for easier capturing of the rods within the screw heads and we used the polyaxiality of the screws with the favoured angle towards the concavity of the curve in order to perform segmental translational correction of the apical deformity. By converting the pedicle screws into monoaxial it was then possible to apply derotational forces and partly correct the rotatory component of the scoliosis. This would have been much less effective if standard polyaxial screws had been used. Scoliosis correction in both Groups 1 and 2 was achieved mainly through segmental vertebral translation of the apex of the curve with a small complementary correction of vertebral rotation, accomplished once the pedicle screws were locked into monoaxial. Previous studies have described scoliosis correction using the direct vertebral derotational technique, 15 bilateral apical vertebral derotation using an outrigger quadrilateral frame, 14 or the vertebral coplanar alignment technique. 21 The mean scoliosis correction for main thoracic and thoracolumbar/lumbar curves of over 70% in our series compares favourably with these previous studies. 14,15,21 VOL. 94-B, No. 12, DECEMBER 2012

7 1676 A. I. TSIRIKOS, A. S. SUBRAMANIAN Complications included three deep and superficial wound infections (1.4%), one transient IOM trace loss with no neurological consequences and one re-operation for technical reasons (0.5%). Our data suggests no difference in the rate of wound infection between the two groups despite the fact that scoliosis correction in Group 1 required a higher density of implants and prolonged surgical times. In contemporary scoliosis correction techniques, increasing numbers of screw anchors have been used over bilateral rods, in order to gain maximal stability and allow for optimum deformity correction. However this can increase the neurological risk due to screw misplacement, prolong surgical time and blood loss, and result in higher instrumentation cost. It has also been suggested that deformity correction in the coronal plane can be achieved at the expense of the sagittal contour of the spine with segmental pedicle screw constructs having a hypokyphotic effect. 4,9-11,22,23 Clements et al 11 have reported that implant density has a positive correlation with the degree of curve correction. In contrast, Quan and Gibson 22 found no advantage of bilateral segmental pedicle screws over unilateral or alternate screw fixation in terms of scoliosis correction. The only factor that correlated with increased coronal correction rate was the flexibility of the curve when assessed by push-prone radiographs (radiographs taken with the patient in the supine position with longitudinal traction and apical corrective forces applied). Decrease in thoracic kyphosis was associated with greater scoliosis correction. In our study, we calculated the FI based on standardised supine maximum lateral bending radiographs. Increased scoliosis size for thoracic curves correlated with a lower FI in both groups. The only predictor of greater scoliosis correction was the FI for both main thoracic and thoracolumbar/lumbar curves in Groups 1 and 2. Increased degree of pre-operative scoliosis correlated with prolonged surgical time and greater blood loss in both groups. The presence of severe upper thoracic scoliosis in Group 1 resulted in longer hospital stay, probably due to the longer extent of the spinal exposure and the arthrodesis. Contrary to the study by Quan and Gibson, 22 the degree of scoliosis correction in our series did not correlate with decrease in thoracic kyphosis. The use of all pedicle screw instrumentation in our cohort resulted in a mean loss of thoracic kyphosis of 3 to 4, which was similar in Groups 1 and 2 and does not appear to affect the clinical outcome as demonstrated by the SRS scores. Groups 1 and 2 were similar in terms of demographics and there was no significant difference on any of the deformity parameters either pre-operatively or at followup, aside from the degree of upper thoracic scoliosis correction, which was greater in Group 1. However, the size of the upper thoracic curves was higher with a lesser FI in Group 2, accounting possibly for the smaller degree of surgical correction. The use of the unilateral pedicle screw correction technique reduced significantly the surgical time and intra-operative blood loss but had no impact on the length of hospital stay, which reflects the overall postoperative morbidity of the procedure. Other potential benefits of the unilateral technique include the ability to preserve a larger area without implants for placement of bone grafts, which could reduce the rate of nonunion and wound infection, as this may be related to a higher density of implants and could both manifest at a longer follow-up. The unilateral technique with the reduced number of pedicle screws also reduced by a mean of 35% the cost of the instrumentation without compromising the clinical results. The limitations of our study include the retrospective nature of the analysis, the lack of randomisation between the two groups, as well as the relatively small numbers of patients in individual Lenke subtypes, to allow comparison of results across the two techniques when using the Lenke classification. Data on rotatory deformity correction were not included, as in the presence of segmental pedicle screws it would have been inaccurate to measure vertebral rotation on plain radiographs. Measurements of rib and chest wall deformity with the use of the scoliometer are equally inaccurate, as these depend on the selected level that is not standardised, as well as on the patient s position in space and during forward bending that is uncontrolled; therefore, such data have not been included in our study. However, we did not observe inferior rib or rotatory deformity correction and maintenance resulting from the absence of more screw anchors over the second supportive rod in Group 2. In addition, this was a large prospectively collected consecutive series of patients treated by a single surgeon in the same institution, comparing two clearly described techniques of pedicle screw correction. We assessed our clinical outcomes in terms of degree of deformity correction controlling for preoperative FI, rate of complications, and parameters determining surgical morbidity such as surgical time, blood loss and hospital stay. The amount of scoliosis correction reflects a measure of technical competence but does not necessarily relate to patient satisfaction, which is the primary goal of surgery. We, therefore, included in our analysis the SRS-22 scores that provide a more accurate representation of clinical outcomes, as perceived by the patients. In conclusion, both groups achieved excellent deformity correction with low rates of complication and high patient satisfaction. Higher pedicle screw implant density and a bilateral segmental correction technique have not achieved greater scoliosis correction apart for upper thoracic curves in our series. Scoliosis surgery has entered an era when spinal surgeons have to demonstrate good clinical practice based on outcomes validated through patient quality of life assessments. Cost implications have to be taken into account as health economics play increasingly a determining role in provision of care within modern health systems. This study supports the use of the unilateral segmental pedicle screw technique that resulted in reduced surgical time, and blood loss, while maintaining comparable deformity correction and excellent clinical outcomes. THE JOURNAL OF BONE AND JOINT SURGERY

8 POSTERIOR SPINAL ARTHRODESIS FOR ADOLESCENT IDIOPATHIC SCOLIOSIS USING PEDICLE SCREW INSTRUMENTATION 1677 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. Riseborough EJ, Wynne-Davies R. A genetic survey of idiopathic scoliosis in Boston, Massachusetts. J Bone Joint Surg [Am] 1973;55-A: Cheng I, Kim Y, Gupta MC, et al. Apical sublaminar wires versus pedicle screws: which provides better results for surgical correction of adolescent idiopathic scoliosis? Spine (Phila Pa 1976) 2005;30: Kuklo TR, Potter BK, Polly DW Jr, et al. Monoaxial versus multiaxial thoracic pedicle screws in the correction of adolescent idiopathic scoliosis. Spine (Phila Pa 1976) 2005;30: Lowenstein JE, Matsumoto H, Vitale MG, et al. Coronal and saggital place correction in adolescent idiopathic scoliosis: a comparison between all pedicle screw versus hybrid thoracic hook lumbar screw constructs. Spine (Phila Pa 1976) 2007;32: Vora V, Crawford A, Babekhir N, et al. A pedicle screw construct gives an enhanced posterior correction of adolescent idiopathic scoliosis when compared with other constructs: myth or reality. Spine (Phila Pa 1976) 2007;32: Kim YJ, Lenke LG, Cho SK, et al. Comparative analysis of pedicle screw versus hook instrumentation in posterior spinal fusion for adolescent idiopathic scoliosis. Spine (Phila Pa 1976) 2004;29: Liljenqvist U, Hackenberg L, Link T, Halm H. Pullout strength of pedicle screws versus pedicle and laminar hooks in the thoracic spine. Acta Orthop Belg 2001;67: Suk SI, Lee CK, Kim WJ, et al. Segmental pedicle screw fixation in the treatment of adolescent idiopathic scoliosis. Spine (Phila Pa 1976) 1995;20: Luhmann SJ, Lenke LG, Kim YJ, Bridwell KH, Schootman M. Thoracic adolescent idiopathic scoliosis curves between 70 and 100 degrees: is anterior release necessary? Spine (Phila Pa 1976) 2005;30: Dobbs MB, Lenke LG, Kim YJ, Luhmann SJ, Bridwell KH. Anterior/posterior spinal instrumentation versus posterior instrumentation alone for the treatment of adolescent idiopathic scoliotic curves more than 90 degrees. Spine (Phila Pa 1976) 2006;31: Clements DH, Betz RR, Newton PO, et al. Correlation of scoliosis curve correction with the number and type of fixation anchors. Spine (Phila Pa 1976) 2009;34: Lenke LG, Betz RR, Harms J, et al. Adolescent idiopathic scoliosis: a new classification to determine extent of spinal arthrodesis. J Bone Joint Surg [Am] 2001;83- A: Asher MA, Lai SM, Burton D, Manna B. The reliability and concurrent validity of the Scoliosis Research Society-22 patient questionnaire for idiopathic scoliosis. Spine (Phila Pa 1976) 2003;28: Potter BK, Kuklo TR, Lenke LG. Radiographic outcomes of anterior spinal fusion versus posterior spinal fusion with thoracic pedicle screws for treatment of Lenke Type I adolescent idiopathic scoliosis curves. Spine (Phila Pa 1976) 2005;30: Lee SM, Suk SI, Chung ER. Direct vertebral rotation: a new technique of threedimensional deformity correction with segmental pedicle screw fixation in adolescent idiopathic scoliosis. Spine (Phila Pa 1976) 2004;29: Risser JC. The iliac apophysis: an invaluable sign in the management of scoliosis. Clin Orthop Relat Res 1958;11: Cobb JR. Outline for the study of scoliosis: Instructional course lectures. Vol. 5. American Academy of Orthopaedic Surgeons. Ann Arbor: JW Edwards, Meyerding HW. Low backache and sciatic pain associated with spondylolisthesis and protruded intervertebral disc: incidence, significance, and treatment. J Bone Joint Surg 1941;23: Wiltse LL, Jackson DW. Treatment of spondylolisthesis and spondylolysis in children. Clin Orthop Relat Res 1976;117: Mulpuri K, Perdios A, Reilly CW. Evidence-based medicine analysis of all pedicle screw constructs in adolescent idiopathic scoliosis. Spine (Phila Pa 1976) 2007;32: Vallespir GP, Flores JB, Trigueros IS, et al. Vertebral coplanar alignment: a standardized technique for three dimensional correction in scoliosis surgery: technical description and preliminary results in Lenke type 1 curves. Spine (Phila Pa 1976) 2008;33: Quan GM, Gibson MJ. Correction of main thoracic adolescent idiopathic scoliosis using pedicle screw instrumentation: does higher implant density improve correction? Spine (Phila Pa 1976) 2010;35: Lehman RA Jr, Lenke LG, Keeler KA, et al. Operative treatment of adolescent scoliosis with posterior pedicle screw-only constructs: minimum three-year follow-up of one hundred fourteen cases. Spine (Phila Pa 1976) 2008;33: VOL. 94-B, No. 12, DECEMBER 2012