Preoperatively, vital capacity and peak flow were similar in the two groups. At the time of the first postoperative examination, both parameters decreased significantly more (p = 0.004 and p = 0.014, respectively) in the thoracoscopic surgery group compared with the posterior surgery group. However, at the time of the final follow-up, with the numbers available, there were no significant differences in peak flow (p = 0.18) or vital capacity (p = 0.67) between the groups. Of note, no patient in either group had severe pulmonary disease, although four patients in the thoracoscopic group and five patients in the posterior group had mild asthma.
The mean total SRS-22 scores improved in the thoracoscopic surgery group and remained the same in the posterior surgery group when compared with preoperative values. Although the mean total scores were similar between the two groups preoperatively, the difference in mean total scores at the time of follow-up was significant (p < 0.0001), favoring the thoracoscopic group. The thoracoscopic surgery group had better scores in all domains of the questionnaire when compared with the posterior surgery group at the time of the final follow-up (Fig. 6).
Thoracoscopic spinal fusion compared favorably with posterior spinal fusion for the treatment of thoracic adolescent idiopathic scoliosis in the present series. Coronal plane correction was similar for both groups, and curve correction in the posterior surgery group mirrored that in numerous studies in which hook or combination hook-screw constructs were utilized1-4,10,30. Coronal balance was improved in both groups and was within normal limits in both groups at the time of the final follow-up. Thus, the goal of partial curve correction and coronal balance was achieved similarly in both groups. The tilt angle of the most caudad instrumented vertebra was corrected to a greater extent in the thoracoscopic surgery group (from 23.1° to 8.4°) than in the posterior surgery group (from 17.2° to 8.6°), although the final tilt angles were equivalent. The long-term effects of residual tilt angle are not known, although there may be a greater potential for adjacent segment disc degeneration in association with larger residual tilt angles15.
It has been reported that, in patients managed with Harrington rod instrumentation and fusion, the more caudad a fusion is extended below the third lumbar vertebra, the more likely there will be late caudad disc degeneration and back pain31. In the present study, the most caudad extent of the fusion was in the upper lumbar spine in the majority of patients in the posterior surgery group whereas it was in the lower thoracic spine in the majority of patients in the thoracoscopic surgery group. Whether or not differences in the rate of caudad segment disc degeneration and back pain or thoracolumbar flexibility will be found in the future is unclear and will require longer follow-up.
Curve correction with thoracoscopic instrumentation may increase as the surgeon's experience grows. In one study of fifty patients who were managed with thoracoscopic spinal instrumentation, curve correction averaged 50% in the first forty patients and improved to 69% in the last ten patients22. Recently, Wong et al. reported a mean thoracic curve correction of 62% with use of a thoracoscopic approach32. The radiographic results in the present series are similar to those reported in a number of studies of anterior thoracic instrumentation performed by means of thoracotomy. Thoracic curve correction associated with the use of threaded flexible rods, single solid rods, or dual rod constructs has ranged from 45% to 71%10,14,33-37. Newer posterior techniques involving the use of thoracic pedicle screws may duplicate the larger amounts of correction seen in some studies of anterior surgery. Furthermore, the new posterior techniques appear to increase coronal plane correction more than hook-rod constructs do7,8.
One weakness of the present study is that it was retrospective in nature and therefore was not randomized. However, aside from the slightly greater amount of kyphosis in the posterior group, the two groups were similar in terms of age, preoperative curve size and range, and gender.
The mean preoperative kyphosis (T2-T12) was greater in the posterior surgery group (among the patients in whom it was measured) than in the thoracoscopic group (34° compared with 26°), perhaps reflecting a selection bias in that patients with kyphosis of >40° were not offered the thoracoscopic procedure. Hyperkyphosis following anterior spinal instrumentation with a flexible rod system in patients with preoperative kyphosis of >20° has been reported10,16. Hyperkyphosis was not seen in our patients who were managed with thoracoscopic surgery involving the use of a 4.5-mm titanium rod. Furthermore, junctional kyphosis cephalad or caudad to the construct was not seen in either surgical group.
The majority of patients in both groups had a Lenke type-1A or 1B curve (i.e., a lumbar curve not crossing the midline); thus, a meaningful comparison between the groups for selective thoracic fusion cannot be made. Two patients in the thoracoscopic surgery group who had a Lenke type-1C curve (i.e., a lumbar curve crossing the midline) and four such patients in the posterior surgery group had balanced corrections. Despite early reports of coronal decompensation and junctional kyphosis in patients undergoing posterior segmental fixation, we did not encounter the problem, which is minimized with proper fusion level selection4. Lenke et al. reported better major curve correction and spontaneous lumbar curve correction in patients undergoing selective thoracic fusion by means of the anterior approach as compared with the posterior approach37. It remains to be seen whether this finding will hold true for the thoracoscopic approach.
One advantage of the thoracoscopic approach in the present series was the ability to save caudad fusion levels. In twenty-six of the twenty-eight patients in the thoracoscopic surgery group, the caudad extent of the fusion was at or cephalad to T12. The remaining two patients had fusions to L1. In comparison, ten of the twenty-three patients in the posterior surgery group had fusions into the upper lumbar spine, two of which extended to L3. Betz et al. compared anterior surgery with posterior surgery for the treatment of thoracic scoliosis and found that an average of 2.5 caudad fusion levels were saved with the anterior approach10. It has long been recognized that anterior surgery is expected to save one or more fusion levels when compared with posterior surgery for the treatment of thoracolumbar or lumbar scoliosis14,16,38,39. Fewer levels were fused anteriorly in our patients, with a mean of 5.8 levels fused in the thoracoscopic surgery group and 9.3 levels fused in the posterior surgery group. Although the function domain score on the SRS outcome questionnaire was better for the thoracoscopic surgery group than for the posterior surgery group at the time of follow-up, it is unclear if this finding was a reflection of the smaller number of levels fused and better spinal flexibility.
Complications requiring revision surgery occurred in one patient in each group. In the posterior surgery group, a hook broke through the twelfth thoracic lamina and was successfully revised with pedicle screw instrumentation to L1. In the thoracoscopic surgery group, revision was required in a patient in whom a pseudarthrosis developed in the cephalad two levels of the fusion, with associated implant displacement. On the basis of the intraoperative findings at the time of revision and assessment of screw placement, the pseudarthrosis resulted from residual disc material noted in the disc spaces. Animal studies comparing thoracoscopic techniques with open anterior techniques or evaluating thoracoscopic techniques alone have demonstrated equivalent results in terms of the degree of release, the amount of disc removed, and the rates of radiographic, biomechanical, and histological fusion40-44, but care is required to ensure satisfactory disc excision to allow fusion to occur21.
The fusion rate in the thoracoscopic surgery group in the present series compares favorably with that in the study by Picetti et al., who reported pseudarthrosis in ten (20%) of fifty patients undergoing thoracoscopic fusion22. Nine of the ten pseudarthroses occurred in patients in whom demineralized bone matrix had been utilized as a substitute for autograft. The use of allograft alone for fusion after disc excision has been shown, in an animal model, to result in a lower fusion rate than was the case when autograft was used42. In the present study, the pseudarthrosis rate in the thoracoscopic surgery group was 3.6% (one of twenty-eight). Wong et al. reported no cases of pseudarthrosis in a series of twelve patients who had been managed with thoracoscopic instrumentation32.
Rod breakage occurred in one patient in the posterior surgery group and two patients in the thoracoscopic surgery group. Although rod breakage may be a sign of nonunion, none of these patients had loss of correction after the three-month postoperative evaluation, and none of them had pain at the time of follow-up. In addition, a solid arthrodesis was documented with computed tomography scanning in both of the patients in the thoracoscopic surgery group. While it has been reported that rod breakage may allow controlled settling and gradual arthrodesis, avoiding breakage is preferred17. In both of the patients in the thoracoscopic surgery group in whom rod breakage occurred, 4.5-mm titanium rods had been used and the fracture may have occurred as a result of fatigue due to stressing and notching of the titanium rod during cantilever reduction of the curvature. Currently, we use 4.5-mm stainless steel rods because less rod breakage is predicted45. Betz et al. noted that if an anterior rod breaks, it will do so by two years postoperatively10.
In the thoracoscopic surgery group, one patient had a mucous plug that was thought to be the result of a prolonged operative time with the patient in the lateral decubitus position, in which the dependent lung had become hyperemic and congested21,22. We have not noted this problem recently as our operative times have decreased. Aggressive fiberoptic suctioning through the double-lumen endotracheal tube is performed after each procedure prior to extubation, and we believe that postoperative bronchoscopy should be considered in cases in which the operative time exceeds five hours.
Major complications in the form of vascular or neurological injury did not occur in either group. We are not aware of any reported cases of vascular or neurological injury associated with thoracoscopic instrumentation, although this concern has been raised. Sucato et al.46 documented the proximity of the aorta to the tips of the vertebral screws in the anterolateral aspect of the thoracic spine. The spinal rotation that occurs in patients with right thoracic scoliosis changes the relationship of the vertebrae to the thoracic aorta and may make the aorta more vulnerable to iatrogenic injury from the screws46. Bicortical screw purchase must be done in a manner to minimize screw tip protrusion beyond the vertebral body23. A unicortical screw design is now available and may allow the surgeon to minimize this concern. The guidewire technique that we initially used was abandoned because of the potential for migration into the contralateral side of the chest, risking injury to the aorta or lung47.
The thoracoscopic procedure compared favorably with the posterior procedure in terms of intraoperative blood loss and the requirement for blood transfusions. The mean blood loss was 361 mL in the thoracoscopic surgery group, compared with 545 mL in the posterior surgery group. In the thoracoscopic surgery group, four patients received a one-unit transfusion of autologous blood. In the posterior surgery group, eight patients received a one-unit transfusion and two patients received a two-unit transfusion. Although the estimated blood loss was roughly 200 mL greater in the posterior surgery group, the transfusion rate may have been affected by the anesthesiologists' tendency to return autologous blood to patients after there had been 500 mL of blood loss. The mean operative time in the thoracoscopic surgery group (6.0 hours; range, 4.0 to 9.8 hours) was significantly greater than that in the posterior surgery group (3.3 hours; range, 2.2 to 4.6 hours) (p < 0.0001). The first three thoracoscopic procedures exceeded nine hours in length, with the operative time decreasing to as low as four hours later in the series. The learning curve associated with thoracoscopic surgery has been previously documented for anterior disc releases without instrumentation22,24.
The length of the hospital stay was approximately one and one-half days shorter in the thoracoscopic surgery group, suggesting less immediate postoperative pain and more rapid early recovery from surgery. Data regarding the length of stay in the intensive care unit were not collected. Typically, patients undergoing thoracoscopic surgery were monitored in an intermediate care unit for one to two days before being transferred to a standard nursing care unit. Patients undergoing posterior surgery usually stayed in the standard nursing unit during the entire hospital stay. While a cost-analysis comparison of thoracoscopic and thoracotomy disc releases has been reported48, we did not compare costs in the present study.
Pulmonary function is only temporarily diminished following thoracoscopic surgery. In the present series, vital capacity decreased by 28% at three weeks postoperatively and returned to baseline by one year. This temporary decline is consistent with the findings of other studies assessing the effect of open thoracotomy on pulmonary function19,49-51. Thoracoscopic surgery causes less chest-cage disruption and pulmonary compromise than open thoracotomy does19. In the posterior surgery group, vital capacity initially declined by 11% and then returned to baseline by the time of the final follow-up. This initial decline was likely related to a thoracoplasty being performed in thirteen of the twenty-three patients51,52. Thoracoplasty is less likely to be performed today in association with the use of thoracic pedicle screws but was commonly performed in patients managed with all-hook or hybrid constructs. Peak flow also diminished more in the thoracoscopic surgery group than in the posterior surgery group immediately postoperatively. In summary, our data demonstrated that the thoracoscopic approach had no significant final deleterious effect on pulmonary function and was not significantly different from the posterior approach in this regard.
Patient-based outcomes as assessed with the SRS-22 questionnaire revealed improvement in the total score and in the self-image domain in both groups. Patients in the thoracoscopic surgery group had higher scores in all domains than those in the posterior surgery group did, despite similar scores preoperatively. Overall, both groups of patients fared well and our findings were in agreement with the published results from a multicenter outcomes assessment in patients with adolescent idiopathic scoliosis who were managed with either anterior or posterior surgery5.
To our knowledge, the present report describes the largest study in which thoracoscopic spinal fusion and instrumentation has been compared with posterior spinal fusion and instrumentation for the treatment of thoracic adolescent idiopathic scoliosis and is the only published comparison to date that includes both pulmonary function and patient-based outcomes. Thoracoscopic spinal instrumentation compared favorably with the posterior procedure in terms of coronal plane curve correction and balance, sagittal contour, the complication rate, pulmonary function, and patient-based outcomes. The procedure offers the advantages of fewer levels of spinal fusion, less operative blood loss, lower transfusion requirements, and improved cosmesis as a result of small, well-hidden incisions. Concerns about the procedure include a steep learning curve and the potential for pseudarthrosis and/or rod breakage. Over the course of the present series, the operative time for the thoracoscopic surgery group was nearly double that for the posterior surgery group. Additional time and experience are required in order to determine the precise role of thoracoscopic spinal instrumentation in the treatment of thoracic adolescent idiopathic scoliosis. ▪
The authors did not receive grants or outside funding in support of their research for or preparation of this manuscript. They did not receive payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated.
1. , Hanson DS, Rhee JM, Lenke LG, Baldus C, Blanke K. Correction of thoracic adolescent idiopathic scoliosis with segmental hooks, rods, and Wisconsin wires posteriorly: it's bad and obsolete, correct? Spine. 2002;27: 2059-66.
2. , Birch JG, Herring JA, Johnston CE, Roach JW. Frontal plane and sagittal plane balance following Cotrel-Dubousset instrumentation for idiopathic scoliosis. Spine. 1989;14: 733-7.
3. , Denis F, Lonstein JE, Winter RB. Coronal and sagittal balance in surgically treated adolescent idiopathic scoliosis with the King II curve pattern. A review of 67 consecutive cases having selective thoracic arthrodesis. Spine. 1998;23: 2063-73.
4. , Lonstein JE. A meta-analysis of the literature on the issue of selective thoracic fusion for the King-Moe type II curve pattern in adolescent idiopathic scoliosis. Spine. 2003;28: 948-52.
5. , Haher TR, Brkaric M, Panagopoulos G, Mathur S, Kohani O, Lowe TG, Lenke LG, Wenger DR, Newton PO, Clements DH 3rd, Betz RR. A multicenter study of the outcomes of the surgical treatment of adolescent idiopathic scoliosis using the Scoliosis Research Society (SRS) outcome instrument. Spine. 2002;27: 2046-51.
6. , Gorup JM, Shin TM, Homel P, Merola AA, Grogan DP, Pugh L, Lowe TG, Murray M. Results of the Scoliosis Research Society instrument for evaluation of surgical outcome in adolescent idiopathic scoliosis. A multicenter study of 244 patients. Spine. 1999;24: 1435-40.
7. , Lenke LG, Bridwell KH. Comparative analysis of pedicle screw versus hook instrumentation in posterior spinal fusion of adolescent idiopathic scoliosis: a matched cohort analysis. Read at the Annual Meeting of the Scoliosis Research Society; 2003 Sept 10-13; Quebec City, Canada.
8. , Lee CK, Kim WJ, Chung YJ, Park YB. Segmental pedicle screw fixation in the treatment of thoracic idiopathic scoliosis. Spine. 1995;20: 1399-405.
9. , Bridwell KH, Won DS, Lenke LG, Chotigavanichaya C, Hanson DS. Sagittal plane analysis of adolescent idiopathic scoliosis: the effect of anterior versus posterior instrumentation. Spine. 2002;27: 2350-6.
10. , Harms J, Clements DH 3rd, Lenke LG, Lowe TG, Shufflebarger HL, Jeszenszky D, Beele B. Comparison of anterior and posterior instrumentation for correction of adolescent thoracic idiopathic scoliosis. Spine. 1999;24: 225-39.
11. , Shufflebarger HL. Late-developing infection in instrumented idiopathic scoliosis. Spine. 1999;24: 1909-12.
12. , Newton NC, Sherwood AA. An anterior approach to scoliosis. A preliminary report. Clin Orthop Relat Res. 1969;62: 192-202.
13. , Pellin B. [New instruments and implants for supplementation of the Harrington system]. Z Orthop Ihre Grenzgeb. 1976;114: 5347. German.
14. , Mun Yue W, Moller HB, Hendricks KJ, Burd TA, Gaines RW. Short segment bone-on-bone instrumentation for single curve idiopathic scoliosis. Spine. 2003;28: S224-33.
15. , Hall JE. Solid rod short segment anterior fusion in thoracolumbar scoliosis. J Pediatr Orthop B. 1998;7: 124-31.
16. , Shufflebarger H. Anterior versus posterior instrumentation for the correction of thoracic idiopathic scoliosis. Spine. 2001;26: 1095-100.
17. , Lenke LG, Harms J. Anterior instrumentation. Spine: State of the Art Reviews. 2000;14: 115-26.
18. , Hazelrigg SR, Mack MJ, Dowling RD, Burke D, Gavlick J, Perrino MK, Ritter PS, Bowers CM, DeFino J. Postoperative pain-related morbidity: video-assisted thoracoscopy versus thoracotomy. Ann Thorac Surg. 1993;56: 1285-9.
19. , Marks M, Faro F, Betz R, Clements D, Haher T, Lenke L, Lowe T, Merola A, Wenger D. Use of video-assisted thoracoscopic surgery to reduce perioperative morbidity in scoliosis surgery. Spine. 2003;28: S249-54.
20. , Betz R, Newton P, Lenke L, Clements D, Harms J, Haher T, Lowe T, Merola A, Crawford A, Sucato D. Effects of continued posterior spinal growth on sagittal contour after anterior spinal fusion with instrumentation of different sizes for adolescent idiopathic scoliosis. Read at the Annual Meeting of the Scoliosis Research Society; Quebec City. 2003 Sept 10-13; Quebec City, Canada.
21. . Thoracoscopic anterior instrumentation and fusion for idiopathic scoliosis. J Am Acad Orthop Surg. 2003;11: 221-7.
22. , Ertl JP, Bueff U. Endoscopic instrumentation, correction, and fusion of idiopathic scoliosis. Spine J. 2001;1: 190-7.
23. , Duchene C. The position of the aorta relative to the spine: a comparison of patients with and without idiopathic scoliosis. J Bone Joint Surg Am. 2003;85: 1461-9.
24. , Shea KG, Granlund KF. Defining the pediatric spinal thoracoscopy learning curve: sixty-five consecutive cases. Spine. 2000;25: 1028-35.
25. , Betz RR, Harms J, Bridwell KH, Clements DH, Lowe TG, Blanke K. Adolescent idiopathic scoliosis: a new classification to determine extent of spinal arthrodesis. J Bone Joint Surg Am. 2001;83: 1169-81.
26. . Manual of pulmonary function testing. 8th ed. St. Louis: Mosby; 2003.
27. , Min Lai S, Burton DC. Further development and validation of the Scoliosis Research Society (SRS) outcomes instrument. Spine. 2000;25: 2381-6.
28. , Herring JA, Johnston CE, Birch JG, Roach JW. Treatment of adolescent idiopathic scoliosis using Texas Scottish Rite Hospital instrumentation. Spine. 1994;19: 1598-605.
29. . Theory and mechanisms of posterior multiple hook/screw/rod devices. In: Weinstein SL, editor. Pediatric spine surgery. 2nd ed. Philadelphia: Lippincott Williams and Wilkins; 2001. p 351-66.
30. , Schuette A, Emans JB. Lumbar pedicle screws versus hooks. Results in double major curves in adolescent idiopathic scoliosis. Spine. 1997;22: 1369-79.
31. , Irstam L, Nachemson A. Long-term anatomic and functional changes in patients with adolescent idiopathic scoliosis treated by Harrington rod fusion. Spine. 1983;8: 576-84.
32. , Hee HT, Yu Z, Wong D. Results of thoracoscopic instrumented fusion versus conventional posterior instrumented fusion in adolescent idiopathic scoliosis undergoing selective thoracic fusion. Spine. 2004;29: 2031-9.
33. , Volper J, Heinrich E, Zielke K. Correction of the sagittal plane in idiopathic scoliosis using the Zielke procedure (VDS). SICOT. 1992;16: 213-8.
34. , Zielke K. VDS instrumentation for idiopathic thoracic curvatures. Read at the Annual Meeting of the American Academy of Orthopaedic Surgeons, Las Vegas; 1985 Jan 19-23; Las Vegas, NV.
35. , Lenke LG, Bridwell KH, Blanke KM, Whorton J. Prospective radiographic and clinical outcomes and complications of single solid rod instrumented anterior spinal fusion in adolescent idiopathic scoliosis. Spine. 2001;26: 1956-65.
36. , Shono Y, Satoh S, Abumi K. Anterior correction of thoracic scoliosis with Kaneda anterior spinal system. A preliminary report. Spine. 1997;22: 1358-68.
37. , Betz RR, Bridwell KH, Harms J, Clements DH, Lowe TG. Spontaneous lumbar curve coronal correction after selective anterior or posterior thoracic fusion in adolescent idiopathic scoliosis. Spine. 1999;24: 1663-72.
38. , Asher MA, Lai SM. Patient-based outcomes analysis of patients with single torsion thoracolumbar-lumbar scoliosis treated with anterior or posterior instrumentation: an average 5- to 9-year follow-up study. Spine. 2002;27: 2363-7.
39. , Lee CK, Chung SS. Comparison of Zielke ventral derotation system and Cotrel-Dubousset instrumentation in the treatment of idiopathic lumbar and thoracolumbar scoliosis. Spine. 1994;19: 419-29.
40. , Bylski-Austrow DI, Shelton FS, Crawford AH, Kolata RJ, Baum DS. Endoscopic discectomy increases thoracic spine flexibility as effectively as open discectomy. A mechanical study in a porcine model. Spine. 1998;23: 9-16.
41. , Kotani Y, McNulty PS, Cappuccino A, Kanayama M, Fedder IL, McAfee PC. Video-assisted thoracoscopic surgery versus open thoracotomy for anterior thoracic spinal fusion. A comparative radiographic, biomechanical, and histological analysis in a sheep model. Spine. 1998;23: 1333-40.
42. , Lee SS, Mahar AT, Farnsworth CL, Weinstein CH. Thoracoscopic multilevel anterior instrumented fusion in a goat model. Spine. 2003;28: 1614-20.
43. , Cardelia JM, Farnsworth CL, Baker KJ, Bronson DG. A biomechanical comparison of open and thoracoscopic anterior spinal release in a goat model. Spine. 1998;23: 530-5.
44. , Kassab F, Dempsey M. Thoracoscopic anterior spinal instrumentation and fusion for idiopathic scoliosis: A CT analysis of screw placement and completeness of discectomy [abstract]. Scoliosis Research Society 36th Annual Meeting Book. Rosemont, IL: Scoliosis Research Society; 2001 Sept 19-22. p 90.
45. , Bourgeault CA. Notch sensitivity of titanium alloy, commercially pure titanium, and stainless steel spinal implants. Spine. 2001;26: 1668-72.
46. , Kassab F, Dempsey M. Analysis of screw placement relative to the aorta and spinal canal following anterior instrumentation for thoracic idiopathic scoliosis. Spine. 2004;29: 554-9.
47. , Crawford AH, Berlin RE, Wolf RK. Tension pneumothorax as a complication of video-assisted thoracoscopic surgery for anterior correction of idiopathic scoliosis in an adolescent female. Spine. 2001;26: 448-50.
48. , Wenger DR, Mubarak SJ, Meyer RS. Anterior release and fusion in pediatric spinal deformity. A comparison of early outcome and cost of thoracoscopic and open thoracotomy approaches. Spine. 1997;22: 1398-406.
49. , Lenke LG, Lowe TG, Betz RR, Bridwell KH, Kong Y, Blanke K. Prospective pulmonary function evaluation following open thoracotomy for anterior spinal fusion in adolescent idiopathic scoliosis. Spine. 2000;25: 2319-25.
50. , White DK, Kemp JS, Bridwell KH, Blanke KM, Engsberg JR. Evaluation of ventilatory efficiency during exercise in patients with idiopathic scoliosis undergoing spinal fusion. Spine. 2002;27: 2041-5.
51. , Lenke LG, Bridwell KH, Haas J, Linville DA. A prospective evaluation of pulmonary function in patients with adolescent idiopathic scoliosis relative to the surgical approach used for spinal arthrodesis. Spine. 2000;25: 82-90.
52. , Bridwell KH, Blanke K, Baldus C. Analysis of pulmonary function and chest cage dimension changes following thoracoplasty in idiopathic scoliosis. Spine. 1995;20: 1343-50.