The surgical treatment of adolescent idiopathic scoliosis (AIS) has evolved during the past several decades. Improved coronal and axial plane correction has come at the expense of maintaining or improving sagittal contour.1–3 In 1989, Bernhardt and Bridwell4 defined the normal parameters for sagittal contour of the thoracic spine (20–50° of kyphosis from T3 to T12), lumbar spine (20–60° of lordosis from L1 to L5), and thoracolumbar junction (from T10 to L2, where there is a transition from kyphosis to lordosis) on the basis of radiographs obtained from a series of normal adolescent lateral spine radiographs.5 In addition to the traditional goals of AIS surgery, such as cessation of progression and correction to obtain coronal balance, recent attention has focused on maintenance or restoration of normal sagittal contour.6
The purpose of this study is to evaluate the radiographical and surgical factors in AIS surgery that significantly affect kyphosis maintenance. Suk et al6 have identified that thoracic pedicle screw (TPS) fixation is effective in maintaining kyphosis. Clements et al7 identified that increased anchor density has been shown to decrease kyphosis. Clement et al8 identified rod diameter and reduction techniques that have implications for sagittal plane correction, and Sucato et al9 evaluated surgical approaches; however, there are to date no comprehensive studies that perform a multivariate analysis of numerous factors that are involved in kyphosis maintenance.
Improper sagittal spine balance after long spinal fusion has been identified as a possible cause of early lumbar and cervical spine pathology.1,10,11 Low back pain has been attributed to fatigue in the extensor muscles of the hip and low back, which are overactive when the patient with sagittal imbalance strives to maintain a balanced upright posture.1 Thoracic hypokyphosis may also be associated with proximal junctional kyphosis.12 Thoracic kyphosis of less than 20° (as measured by the Cobb angle at the end vertebrae) has been correlated with decreased pulmonary function.13 In another study, severe thoracic hypokyphosis (T5–T12) of less than 10° was associated with diminished pulmonary function.14
Decreased kyphosis in the thoracic spine has been long recognized as a component of the 3-dimensional deformity in AIS, and by some, it has been determined as the “essential lesion.”15 Cotrel and Dubousset introduced the concept of 3-dimensional deformity correction to include the sagittal and the coronal and transverse planes.16 The principal maneuver in Cotrel and Dubousset instrumentation consisted of rotating a pre–bent rod on the concave side by 90º and then distracting intersegmentally to correct the coronal deformity and increase kyphosis.17 Although this maneuver works very well for correcting scoliosis, it was often inadequate in the sagittal plane, leaving many patients with residual hypokyphosis.6 “Flatback syndrome” is a well-described phenomenon in adult patients in whom long posterior fusions into the lumbar spine were performed, with inadequate maintenance of lumbar lordosis after distraction-type posterior spinal instrumentation, resulting in improper sagittal spine balance. Analogous problems may be seen in the cervical thoracic spine, with inadequate kyphosis restoration in the thoracic region.18
TPS facilitates 3-column spinal fixation and the ability to perform more powerful posterior-only correction of coronal plane deformities. However, posterior-only correction may not be the best way to improve sagittal contour, and some have advocated anterior-only approaches.9 Although traditional all-hook and hybrid constructs have produced excellent results, a greater appreciation for the nuances of TPS instrumentation and insertion techniques in the treatment of AIS are developing.19
We hypothesize that anterior approaches, all-screw constructs, lower anchor density, and fewer number of levels fused will have significant clinical and radiographical ramifications on thoracic kyphosis maintenance after AIS surgery.
MATERIALS AND METHODS
After study approval by an institutional review board at each investigating site a retrospective analysis of prospectively collected data from a multicenter database was performed. Data were collected from 1995 to 2009 by multiple surgeons, all of whom routinely perform AIS surgery. Patients were included for this study if they had a diagnosis of AIS with a Lenke curve type 1 (main thoracic) or type 2 (double thoracic) and a minimum 2-year follow-up. A complete radiographical series consisted of preoperative standing posteroanterior, lateral, and bending radiographs, and postoperative posteroanterior and lateral radiographs. Operative data, including number of levels fused, surgical approach, and type of instrumentation, were recorded. We studied kyphosis from T5 to T12 because this corresponds to the region of the typical main thoracic curve and is most reliably assessed on lateral radiographs. This was a standard lateral view, not necessarily in the plane of the apex of the deformity. Patients stood upright, with hands supported with modified “ski poles” and bend in elbows to accommodate shoulder flexion to 30° (Figure 1).
A total of 526 patients met the inclusion criteria. The mean preoperative thoracic kyphosis (T5–T12) was 22º. In total, 269 patients had kyphosis of less than 22° and comprised the group to be included in this analysis. This cutoff value was chosen so that the study group comprised only those patients in whom the desired surgical goal was to increase thoracic kyphosis. In patients with normal kyphosis, there was presumably no desire to change that with surgical correction. The multivariate analysis included assessment of surgical approach, number of levels fused, anchor type/density, rod diameter, metal types, preoperative coronal curve magnitude, and preoperative thoracic kyphosis magnitude. Categorical variables were coded into a response of 0 or 1. The change in kyphosis was calculated as the 2-year value minus the preoperative value. Correlation and multivariate linear stepwise regression were performed. The data were checked for normality and equal variances, and the level of significance was set at P < 0.05. Data were analyzed using SPSS 12.0 (SPSS, Inc., Chicago, IL).
The average age of the patients was 14.2 ± 1.9 years. A total of 79% (213 of 269) of the curves were Lenke type 1 and 21% (56 of 269) were Lenke type 2. A total of 44% (118 of 269) of the cases were anterior-only approaches, 52% (140 of 269) were posterior-only approaches, and 4% (11 of 269) underwent anterior release with posterior spinal fusion.
The average preoperative kyphosis from T5 to T12 for this patient population was 12° ± 7.0°. The average kyphosis at 2 years postoperative was 19.9 ± 9.9. The average change in kyphosis was 7.8 ± 10.2. The preoperative coronal main thoracic curve was 52.0 ± 10.0 and the 2-year postoperative curve was 19.7 ± 8.4 for a 62% ± 16% coronal curve correction (Table 1).
Eight variables were found to be significantly correlated with change in kyphosis—surgical approach, number of levels fused, preoperative kyphosis, percentage of hooks in the construct, percentage of screws in the construct, use of standard stainless steel rods, main coronal thoracic curve magnitude, and percent change in thoracic curve (Table 2). Anterior approaches were correlated with increasing kyphosis compared with posterior approaches (r = 0.37, P < 0.001). Those who had anterior releases with posterior fusions (11 of 269 patients) trended toward increased kyphosis; however, these data were not statistically significant, possibly due to the small sample size (r = 0.119, P < 0.16). Increasing the number of levels fused correlated with decreasing kyphosis (r = −0.33, P < 0.001). A greater preoperative T5–T12 kyphosis was associated with decreasing kyphosis (r = −0.39, P < 0.001) because these patients had more kyphosis to be lost and the curves that had preoperative hypokyphosis trended toward greater increases in kyphosis. A greater 2-year postoperative thoracic curve magnitude was related to increasing kyphosis (r = 0.20, P = 0.001) and a greater 2-year percentage of correction of thoracic curves was related to decreasing kyphosis (r = −0.23, P < 0.001).
Of the patients who had a posterior approach, a greater percentage of hooks in the construct was associated with increasing kyphosis (r = 0.18, P = 0.034), and conversely, a greater percentage of screws in the construct was related to decreasing kyphosis (r = −0.18, P = 0.03). Increasing anchor density had a trend toward decreasing kyphosis (P = 0.053). Although there was no difference between stainless steel and titanium, standard-strength stainless steel (125 KSI, 5.5-mm diameter) was related to a decrease in kyphosis compared with high- and ultra-high-strength stainless steel (r = −0.47, P = 0.011). Therefore, metals with higher strength and stiffness were better at maintaining kyphosis postoperatively.
Factors associated with increasing kyphosis were lower preoperative kyphosis (R2 = 0.15, P < 0.001) and performance of an anterior approach (R2 = 0.16, P < 0.001). An anterior approach was associated with an approximately 7.7° increase in kyphosis compared with posterior approaches. A greater percentage of correction of the main thoracic coronal curves resulted in decreased kyphosis (R2 = 0.04, P < 0.001), with an approximately 1° loss of kyphosis for every 10% of thoracic coronal curve correction. Therefore, less coronal curve correction may imply improved kyphosis maintenance. These 3 variables predicted a total of 34% of the variation in change in kyphosis from preoperative to 2 years postoperative.
In examining only posterior approaches, a multivariate analysis was done and the only significant independent predictor of change in kyphosis was the use of standard-strength stainless steel rods (R2 = 0.22, P = 0.011). This was associated with an approximately 8° decrease in kyphosis compared with the use of high- or ultra-high-strength stainless steel. However, this factor explains only 22% of the variance in change in kyphosis for posterior approaches.
The 3-dimensional deformity associated with AIS is postulated to be the result of transverse plane vertebral asymmetry. This is thought to result in coronal curvature, axial plane tension, and relative overgrowth of the anterior column of the spine, resulting in hypokyphosis or even lordosis.6,15 One of the goals of corrective surgery for AIS is to restore or maintain thoracic kyphosis. This may be important to reduce junctional problems and to maintain pulmonary function, as outlined in the introductory paragraphs of the article.13,14,20
Historically, treatment of scoliosis has focused on correcting the coronal plane deformity, with little focus on the sagittal plane. However, several studies have attributed improper sagittal restoration to long-term disability.10,11,21 Proximal junctional kyphosis is also prevalent after posterior spinal fusion. Kim et al20 reported a 26% (50 of 193) incidence in a series with 5-year follow-up. Risk factors for proximal junctional kyphosis were preoperative thoracic kyphosis of greater than 40°, hybrid instrumentation (hooks and pedicle screws in the construct), and more than 11 levels included in the fusion.20 Thoracic hypokyphosis may also be responsible for diminution of lung function in patients with scoliosis.13,14 As such, there has been an effort to better delineate those factors involved in correcting and maintaining sagittal balance.6,9
Our multivariate analysis has demonstrated that the use of an anterior approach improves thoracic kyphosis to within expected population norms more reliably than posterior approaches. This has been previously reported in comparisons of anterior and hook-and-pedicle screw hybrid posterior constructs.9,22–24 Greater number of levels fused and greater coronal thoracic curve correction are associated with less kyphosis restoration and hypokyphosis.
Despite powerful corrective ability in the coronal and axial planes with pedicle screw constructs, maintenance or restoration of kyphosis with this anchor type has been unreliable.25–28 Anchor density and type (polyaxial vs. monoaxial screws vs. hybrid constructs), rod diameter, metal type, and use of posterior facet release have all been implicated as important factors in achieving kyphosis.8,23
There have been conflicting data with regard to the efficacy of pedicle screws in maintaining appropriate thoracic kyphosis. Lowenstein et al27 compared their series of all-pedicle screw constructs with hybrid constructs and found that the all-screw group demonstrated a significant decrease in kyphosis (P = 0.012), which was not seen in their hybrid group (P = 0.248). Vora et al28 also compared the ability of wires, hooks, and pedicle screw constructs at restoring thoracic kyphosis. Patients in the sublaminar wire group and all-pedicle screw group lost much of their preoperative kyphosis compared with the hook-and-screw hybrid group, which had significant kyphosis correction at 2 years postoperatively as opposed to a loss of kyphosis correction in the other 2 groups (P < 0.05).28 Kim et al26 compared an all-screw group with an all-hook group at 2 years postoperatively and found thoracic kyphosis to be significantly higher in the all-hook group than in the all-screw group (P = 0.0003). Lonner et al29 compared polyaxial pedicle screw constructs with monoaxial screw constructs and hybrid (hooks, wires, and screws) constructs and found that in the sagittal plane, the polyaxial screw constructs maintained thoracic kyphosis at 2 years, whereas there was a trend toward loss of thoracic kyphosis at 2 years in both the hybrid and the monoaxial groups (P < 0.07). In our series, we found that a greater percentage of hooks in the constructs was associated with increased kyphosis (r = 0.18, P = 0.034). Increased percentage of screws in the construct was associated with decreasing kyphosis (r = −0.18, P = 0.03). There is less coronal plane correction with hook constructs, which may account for why kyphosis was better maintained.
Corrective technique, rod diameter, and metal type may all play a role in kyphosis restoration. Clement et al7 evaluated the reduction techniques used and compared cantilever reduction, which used 5.5-mm stainless steel rods, and simultaneous translation technique on two 6.0-mm titanium rods Patients with preoperative hypokyphosis had significantly better restoration of kyphosis when the 2-rod technique was performed, which, they hypothesize, may be related to the diameter and metal type of the rods, given the anchor density and curve magnitudes were equivalent.8 Our data suggest that standard-strength stainless steel may be related to decreased kyphosis as opposed to high- and ultra-high-strength stainless steel (r = −0.47, P = 0.011), suggesting that stronger rods are better at maintaining kyphosis. We were unable to report any significant differences between titanium and stainless steel constructs.
de Jonge et al1 analyzed preoperative hypokyphosis and how alignment affected postoperative kyphosis correction. In their series of 306 patients with AIS using an all-hook construct, 159 (52%) patients had preoperative hypokyphosis, which is consistent with the percentage in our series (269 of 526 patients who were analyzed). Their results demonstrate that the more severe the preoperative hypokyphosis, the better the correction achieved in the restoration of kyphosis to normal; however, in cases of normal preoperative kyphosis, the average value decreased slightly and some patients became hypokyphotic.1 Our study demonstrated similar results in that patients with greater preoperative kyphosis were associated with decreasing kyphosis (r = −0.39, P < 0.001) postoperatively. This may reflect the goals of surgery. Techniques such as differential rod contouring (concave more kyphotic), concave intersegmental distraction, metals with greater stiffness, and aggressive posterior releases may be employed to improve kyphosis.
In our series, the use of an anterior approach was associated with increasing kyphosis (R2 = 0.16, P < 0.001), by an approximately 7.7° increase in kyphosis, as compared with posterior approaches. Our data are consistent with those of Betz et al22 and Sucato et al9, who compared thoracic kyphosis maintenance by surgical approach used. In the study of Betz et al, all-hook posterior constructs were compared with anterior fusion. In patients with preoperative thoracic hypokyphosis (<20°), 81% of the patients in the anterior group were in a normal range postoperatively (mean: 33° ± 6.2°) and 60% of the patients in the posterior group remained hypokyphotic after surgery (mean: 14° ± 3.8°).22 Sucato et al9 compared anterior instrumentation with posterior instrumentation with all-hook and hybrid constructs. The patients in the anterior all-hooks and hybrid group had significantly greater restoration of thoracic kyphosis at 2 years postoperative (29.9° vs. 23.8° vs. 19.7°, P < 0.05) than patients in the posterior all-hooks and hybrid group, respectively. Based on our data and the above studies, it would seem that the anterior approach is more successful in maintaining thoracic kyphosis.
Lonner et al30 have demonstrated restoration and maintenance in thoracic kyphosis using anterior video-assisted thoracoscopic spinal fusion and instrumentation (VATS) without having junctional problems. When Lonner et al31 compared a matched group of patients who underwent VATS with patients who underwent posterior-only spinal fusion, although fewer levels were fused in the VATS group, there was no increased tendency to produce thoracic kyphosis. In our study, there was no significant difference detected in kyphosis restoration in the subset of patients who underwent anterior release in addition to posterior fusion because the sample set was likely too small (11 of 269 patients). Our data, however, do demonstrate a significant correlation of kyphosis with the number of levels fused. Increasing the number of levels fused was correlated with decreasing kyphosis (r = −0.33, P < 0.001). There is then potential that combining VATS with posterior fusion may help save levels and thus have a kyphosis-restoring effect.
A greater 2-year postoperative thoracic curve magnitude was associated with increasing kyphosis (r = 0.20, P = 0.001) and a greater 2-year percentage of correction of thoracic curves was associated with decreasing kyphosis (r = −0.23, P < 0.001). Our regression analysis also demonstrated that a greater percentage of correction of the main thoracic coronal curve resulted in a hypokyphosing effect, with an approximately 1° loss of kyphosis for every 10% of thoracic coronal curve correction. This may be related to powerful derotation associated with coronal curve correction. This would lead to a relatively lengthened anterior column being aligned in the midsagittal plane. Whether anterior releases, causing a relative shortening of the anterior column, will result in improved kyphosis in larger curves will require a larger series.
In conclusion, our multivariate analysis demonstrates that in patients with AIS who have thoracic hypokyphosis as part of their deformity, anterior approach, fusing as few levels as possible, and using a hybrid construct are associated with improved thoracic sagittal contour.
- Maintenance or restoration of thoracic kyphosis is one of the goals of AIS surgery.
- Anterior approach, increased percentage of hooks in the construct, and increased coronal curve are associated with increased kyphosis maintenance.
- Number of levels fused, greater degree of kyphosis preoperatively, increasing percentage of screws in construct, using standard stainless steel rods, and greater percentage decrease in thoracic coronal curve correction are associated with decreased kyphosis maintenance.
- These variables should be considered in preoperative surgical planning for AIS surgery.
1. de Jonge T, Dubousset JF, Illes T. Sagittal plane correction in idiopathic scoliosis. Spine (Phila Pa 1976) 2002;27:754–60.
2. Petcharaporn M, Pawelek J, Bastrom T, et al. The relationship between thoracic hyperkyphosis and the Scoliosis Research Society outcomes
instrument. Spine (Phila Pa 1976) 2007;32:2226–31.
3. Upasani VV, Tis J, Bastrom T, et al. Analysis of sagittal alignment in thoracic and thoracolumbar curves in adolescent idiopathic scoliosis: how do these two curve types differ? Spine (Phila Pa 1976) 2007;32:1355–9.
4. Bernhardt M, Bridwell KH. Segmental analysis of the sagittal plane alignment of the normal thoracic and lumbar spines and thoracolumbar junction. Spine (Phila Pa 1976) 1989;14:717–21.
5. Stagnara P, De Mauroy JC, Dran G, et al. Reciprocal angulation of vertebral bodies in a sagittal plane: approach to references for the evaluation of kyphosis and lordosis. Spine (Phila Pa 1976) 1982;7:335–42.
6. Suk SI, Kim WJ, Kim JH, et al. Restoration of thoracic kyphosis in the hypokyphotic spine: a comparison between multiple-hook and segmental pedicle screw fixation in adolescent idiopathic scoliosis. J Spinal Disord 1999;12:489–95.
7. 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;34:2147–50.
8. Clement JL, Chau E, Kimkpe C, et al. Restoration of thoracic kyphosis by posterior instrumentation in adolescent idiopathic scoliosis: comparative radiographic analysis of two methods of reduction. Spine (Phila Pa 1976) 2008;33:1579–87.
9. Sucato DJ, Agrawal S, O'Brien MF, et al. Restoration of thoracic kyphosis after operative treatment of adolescent idiopathic scoliosis: a multicenter comparison of three surgical approaches. Spine (Phila Pa 1976) 2008;33:2630–6.
10. Hilibrand AS, Tannenbaum DA, Graziano GP, et al. The sagittal alignment of the cervical spine in adolescent idiopathic scoliosis. J Pediatr Orthop 1995;15:627–32.
11. Moskowitz A, Moe JH, Winter RB, et al. Long-term follow-up of scoliosis fusion. J Bone Joint Surg Am 1980;62:364–76.
12. Yagi M, Akilah KB, Boachie-Adjei O. Incidence, risk factors and classification of proximal junctional kyphosis: surgical outcomes
review of adult idiopathic scoliosis. Spine (Phila Pa 1976) 2011;36:E60–8.
13. Winter RB, Lovell WW, Moe JH. Excessive thoracic lordosis and loss of pulmonary function in patients with idiopathic scoliosis. J Bone Joint Surg Am 1975;57:972–7.
14. Johnston CE II, Richards S, Sucato D, et al. Correlation of Preop curve severity with pulmonary function tests (PFTs) in AIS Surgery. Paper presented at: 44th Annual Meeting of the Scoliosis Research Society; , 2009; San Antonio, TX.
15. Dickson RA, Lawton JO, Archer IA, et al. The pathogenesis of idiopathic scoliosis: biplanar spinal asymmetry. J Bone Joint Surg Br 1984;66:8–15.
16. Labelle H, Dansereau J, Bellefleur C, et al. Comparison between preoperative and postoperative three-dimensional reconstructions of idiopathic scoliosis with the Cotrel-Dubousset procedure. Spine (Phila Pa 1976) 1995;20:2487–92.
17. Gray JM, Smith BW, Ashley RK, et al. Derotational analysis of Cotrel-Dubousset instrumentation in idiopathic scoliosis. Spine (Phila Pa 1976) 1991;16:S391–3.
18. Potter BK, Lenke LG, Kuklo TR. Prevention and management of iatrogenic flatback deformity. J Bone Joint Surg Am 2004;86-A:1793–808.
19. Lonner BS, Auerbach JD, Estreicher MB, et al. Thoracic pedicle screw instrumentation: the learning curve and evolution in technique in the treatment of adolescent idiopathic scoliosis. Spine (Phila Pa 1976) 2009;34:2158–64.
20. Kim YJ, Bridwell KH, Lenke LG, et al. Proximal junctional kyphosis in adolescent idiopathic scoliosis following segmental posterior spinal instrumentation and fusion: minimum 5-year follow-up. Spine (Phila Pa 1976) 2005;30:2045–50.
21. Connolly PJ, Von Schroeder HP, Johnson GE, et al. Adolescent idiopathic scoliosis: long-term effect of instrumentation extending to the lumbar spine. J Bone Joint Surg Am 1995;77:1210–6.
22. Betz RR, Harms J, Clements DH III, et al. Comparison of anterior and posterior instrumentation for correction of adolescent thoracic idiopathic scoliosis. Spine (Phila Pa 1976) 1999;24:225–39.
23. Anakwenze OA, Auerbach JD, Milby AH, et al. Sagittal cervical alignment after cervical disc arthroplasty and anterior cervical discectomy and fusion: results of a prospective, randomized, controlled trial. Spine (Phila Pa 1976) 2009;34:2001–7.
24. Levin R, Matusz D, Hasharoni A, et al. Mini-open thoracoscopically assisted thoracotomy versus video-assisted thoracoscopic surgery for anterior release in thoracic scoliosis and kyphosis: a comparison of operative and radiographic results. Spine J 2005;5:632–8.
25. Kim YJ, Lenke LG, Kim J, et al. Comparative analysis of pedicle screw versus hybrid instrumentation in posterior spinal fusion of adolescent idiopathic scoliosis. Spine (Phila Pa 1976) 2006;31:291–8.
26. Kim YJ, Lenke LG, Cho SK, et al. Comparative analysis of pedicle screw versus hook instrumentation in posterior spinal fusion of adolescent idiopathic scoliosis. Spine (Phila Pa 1976) 2004;29:2040–8.
27. Lowenstein JE, Matsumoto H, Vitale MG, et al. Coronal and sagittal plane correction in adolescent idiopathic scoliosis: a comparison between all pedicle screw versus hybrid thoracic hook lumbar screw constructs. Spine (Phila Pa 1976) 2007;32:448–52.
28. 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;32:1869–74.
29. Lonner BS, Auerbach JD, Boachie-Adjei O, et al. Treatment of thoracic scoliosis: are monoaxial thoracic pedicle screws the best form of fixation for correction? Spine (Phila Pa 1976) 2009;34:845–51.
30. Lonner BS, Auerbach JD, Levin R, et al. Thoracoscopic anterior instrumented fusion for adolescent idiopathic scoliosis with emphasis on the sagittal plane. Spine J 2009;9:523–9.
31. Lonner BS, Auerbach JD, Estreicher M, et al. Video-assisted thoracoscopic spinal fusion compared with posterior spinal fusion with thoracic pedicle screws for thoracic adolescent idiopathic scoliosis. J Bone Joint Surg Am 2009;91:398–408.