Does Thoracic Hypokyphosis Matter in Lenke Type 1 Adolescent Idiopathic Scoliosis?

Transcription

1 Spine Deformity 1 (2013) 40e45 Does Thoracic Hypokyphosis Matter in Lenke Type 1 Adolescent Idiopathic Scoliosis? Steven D. Glassman, MD a, Daniel J. Sucato, MD, MSc b, Leah Y. Carreon, MD, MSc a, *, James O. Sanders, MD c, Michael G. Vitale, MD, MPH d, Lawrence G. Lenke, MD e a Norton Leatherman Spine Center, 210 East Gray Street, Suite 900, Louisville, KY 40202, USA b Department of Pediatric Orthopedic Surgery, Texas Scottish Rite Hospital for Children, 2222 Welborn Street, Dallas, TX 75219, USA c Department of Orthopaedics, University of Rochester School of Medicine and Dentistry, 601 Elmwood Avenue, Box 665, Rochester, NY 14642, USA d Columbia University and New York Presbyterian Hospital, 600 W. 168th Street, 7th Floor, New York, NY 10032, USA e Department of Orthopaedic Surgery, Washington University School of Medicine, 660 S. Euclid Avenue, Campus Box 8233, Saint Louis, MO 63110, USA Received 20 February 2012; revised 31 August 2012; accepted 2 September 2012 Abstract Study Design: We analyzed a prospective cohort of Lenke 1ABC adolescent idiopathic scoliosis (AIS) patients based on differences in T5eT12 sagittal alignment. Objective: Our objective was to determine whether patients with hypokyphotic scoliosis demonstrate unique characteristics in terms of baseline health status and response to surgery. Background Summary: Right thoracic AIS often presents as a hypokyphotic scoliosis, with rotatory deformity resulting in a diminution of the normal thoracic kyphosis. The perceived importance is indicated by the inclusion of a sagittal plane modifier within Lenke s classification system for AIS and studies examining reduction strategies to restore thoracic kyphosis. Methods: We grouped patients based on thoracic kyphosis, measured from T5 to T12, as either less than 10 (hypokyphotic) or greater than or equal to 10 (normal/kyphotic). We used Student t test for independent samples to compare continuous variables between the hypokyphotic and normal/kyphotic groups. Results: There was a significant difference in age between groups (mean age, 14.9 years in the hypokyphotic group versus 13.4 years in the normal/kyphotic group; p 5.007). Differences in baseline health status measures were statistically significant but small. Hypokyphotic patients reported less pain than normal/kyphotic patients, with a mean Scoliosis Research Society (SRS) Pain score of 4.15 versus 4.03 (p 5.044), better SRS Mental Domain scores (4.06 vs. 3.92; p 5.026), and better SRS Total scores (3.92 vs. 3.83, p 5.027). The hypokyphotic group also had better SRS Appearance (3.36 vs. 3.30) and Activity scores (4.20 vs. 4.13), but these differences did not reach statistical significance. Conclusions: There were no differences in baseline or 2-year postoperative outcome scores in Lenke 1 AIS patients with hypokyphosis compared with patients with normal sagittal alignment. Maintenance of or restoration to normal kyphosis in hypokyphotic patients with Lenke 1 AIS may not improve clinical outcome compared with patients who remain lordotic after surgical correction. Ó 2013 Scoliosis Research Society. Keywords: Adolescent idiopathic scoliosis; Sagittal alignment; Thoracic lordosis; Health status; SRS-22; Thoracic hypokyphosis Author disclosures: SDG (grant from Norton Healthcare; patent with Medtronic; royalties from Medtronic; travel accommodations from Orthopedic Research and Education Fund; Vice President of Scoliosis Research Society, unpaid position); DJS (grants from OREF; lectures for Medtronic royalties from Medtronic); LYC (employment by Norton Healthcare; travel accommodations from Orthopedic Educational Research Fund, National Institutes of Health, University of Louisville IRB); JOS (grants from POS- NA and CWSDSG); MGV (consulting for CWSDSG, Stryker, Biomet; grants from OREF, CWSDRF, SRS, AOSpine, POSNA; royalties from Biomet; travel accommodations from CWSDSG and FoxPSDSG); LGL (grants from Axial Biotech and DePuy; royalties from Medtronic and Quality Medical Publishing; travel accommodations from Medtronic; Immediate past President of the Scoliosis Research Society, unpaid poaition). *Corresponding author. Norton Leatherman Spine Center, 210 East Gray Street, Suite 900, Louisville, KY 40202, USA. Tel.: (502) ; fax: (502) address: leah.carreon@nortonhealthcare.org (L.Y. Carreon) X/$ - see front matter Ó 2013 Scoliosis Research Society.

2 S.D. Glassman et al. / Spine Deformity 1 (2013) 40e45 41 Introduction Right thoracic adolescent idiopathic scoliosis (AIS) is a relatively common spinal deformity; yet, the etiology remains poorly defined. Furthermore, the pattern of curve progression is variable, even among curves of similar magnitude and location. Some authors have suggested that right thoracic AIS most frequently presents as a lordoscoliosis, with rotatory deformity resulting in a diminution of the normal thoracic kyphosis [1]. Although this is certainly one of the observed patterns of progression in AIS, the relative frequency of thoracic lordoscoliosis has not been defined. Also, this description presumes an established baseline for normal thoracic kyphosis, whereas in reality normal thoracic kyphosis encompasses a substantial range. Although there is obviously individual variation, sagittal Cobb angle between 10 and 40 has been cited as a normal range [2-5]. The perceived importance of sagittal plane alignment is emphasized by the inclusion of a sagittal plane modifier within Lenke s commonly used classification system for AIS [6,7]. The Lenke classification denotes normal sagittal alignment (10 to 40 ), hypokyphosis (less than 10 ), and kyphosis (greater than 40 ). Significant attention has also been devoted to the restoration of thoracic kyphosis as a component of surgical correction. Multiple studies have examined the capacity of various instrumentation constructs and reduction strategies to restore thoracic kyphosis [8,9]. Despite the emphasis on sagittal plane alignment in both diagnostic classification and treatment planning, the clinical importance of thoracic hypokyphosis in right thoracic AIS is not well understood. In this study, we sought to determine whether patients with hypokyphotic scoliosis demonstrate unique characteristics in terms of baseline health status. As a subanalysis, we examined whether there is a difference in response to treatment in patients with hypokyphotic scoliosis, compared with patients with normal thoracic kyphosis or mild hyperkyphosis. Methods We queried a prospectively collected multicenter database of pediatric spinal deformity (Spinal Deformity Study Group) for patients with right-sided main thoracic AIS (Lenke 1A, B, and C curves). Inclusion criteria were complete preoperative clinical and radiographic data, age 10 to 18 years at surgery, and absence of associated conditions that might suggest a nonidiopathic etiology. We excluded patients with curve types other than Lenke 1A, B, or C. We excluded patients with curves greater than 70 based on the concern that large curve magnitude might decrease the accuracy of sagittal plane measures on plain radiographs. Patients who were undergoing a revision spine surgery or had a staged or concurrent anterior procedure were also excluded. Preoperative data included standard demographics, history of bracing, baseline pulmonary function, and baseline health status, as determined by the Scoliosis Research Society-22 (SRS) instrument [10]. Preoperative radiographic parameters included standard coronal and sagittal measures for the primary as well as secondary curves, and measures of coronal and sagittal balance. We also recorded scoliometer readings. We determined sagittal alignment based on the T5eT12 Cobb angle as defined by the Lenke Classification Sagittal Modifier [6]. We also examined a subgroup with complete preoperative and 2-year postoperative data. Surgical and perioperative data included operative time, estimated blood loss, use of osteotomies, number of levels fused, as well as complications and length of hospital stay. At 2-year followup, we analyzed health status and pulmonary function, as well as radiographic parameters, in terms of both raw score and change from baseline. We grouped patients based on thoracic lordosis, measured from T5 to T12, as either less than 10 (hypokyphotic) or greater than or equal to 10 (normal/kyphotic). We performed a sensitivity analysis, excluding patients identified as kyphotic in Lenke s classification (greater than 40 ), and thus compared only the normal and hypokyphotic groups. We used Student t test for independent samples to compare continuous variables between the hypokyphotic and normal/kyphotic groups. We used Fisher exact test to compare categorical or dichotomous variables, and Mann-Whitney test to compare ordinal noncontinuous, rank-ordered variables such as the number of levels fused. A p value of.05 was considered statistically significant. We performed all analyses using PASW version 17 (Chicago, IL). Results A total of 793 patients met initial enrollment criteria and 780 patients (98%) had complete preoperative data. The 780 Lenke 1A, B, or C patients analyzed in this study included 615 patients with thoracic sagittal Cobb (T5eT12) greater than or equal to 10 and 165 patients with thoracic sagittal Cobb less than 10 (Table 1). There was a significant difference in age between groups: a mean age of 14.9 years in the hypokyphotic group, compared with 13.4 years in the normal/kyphotic group (p 5.007). There was Table 1 Preoperative clinical measures. Variable Sagittal Cobb angle T5eT12 p value N (total5780) Age (y) Females Prior bracing Clinical measures Shoulder difference Proximal thoracic scoliometer reading Main thoracic scoliometer reading Thoracolumbar scoliometer reading Coronal decompensation Clinical trunk shift Limb length discrepancy

3 42 S.D. Glassman et al. / Spine Deformity 1 (2013) 40e45 Table 2 Preoperative radiographic measures. Radiographic measure Sagittal Cobb angle T5eT12 p value Proximal thoracic Cobb angle Main thoracic Cobb angle Thoracolumbar Cobb angle Proximal thoracic AVT Main thoracic AVT Thoracolumbar AVT Radiographic coronal balance T1 tilt Clavicle angle Radiographic shoulder height Floman shift Pelvic obliquity Sacral obliquity Radiographic leg length discrepancy T2eT12 sagittal Cobb angle T2eT5 sagittal Cobb angle T5eT12 sagittal Cobb angle T10eL2 sagittal Cobb angle T12eS1 sagittal Cobb angle Sagittal balance Abbreviation: AVT, apical vertebral translation. a trend toward difference in gender distribution: 21% males in the hypokyphotic group and 15% males in the normal/ kyphotic groups (p 5.058). There was no difference in history of brace wear between groups. Comparison of baseline radiographic parameters (Table 2) revealed the defined difference in sagittal Cobb measurements: a mean 2.9 T5eT12 Cobb angle in the hypokyphotic group and a mean 24.6 T5eT12 Cobb angle in the normal/kyphotic group (p 5.000). We observed no differences in T2eT5 sagittal Cobb angle. We saw a greater lumbar lordosis (sagittal Cobb T12eS1) in the normal/ kyphotic group (60.3 ), compared with the hypokyphotic group (53.8 ). The normal/kyphotic group also had a more negative global sagittal balance (e23.2 mm vs. e14.1 mm; p 5.002). Coronal plane measures were similar in the two groups, with a mean main thoracic curve magnitude of 54 in both groups. The extent of clinical coronal decompensation, radiographic coronal balance, and Floman shift [11] was similar between groups. We noted significant differences between groups for the compensatory thoracolumbar/ lumbar Cobb angle (32.2 vs ;p5.012) and thoracolumbar/lumbar apical vertebral translation (9.9 mm vs mm; p 5.003) for the hypokyphotic and normal/ kyphotic groups, respectively. The hypokyphotic group had a 4.8-mm coronal imbalance to the right compared with a 0.8-mm list to the left in the normal/kyphotic group (p 5.037). This statistical difference was not duplicated on the Floman shift analysis. We observed small but statistically significant differences in baseline health status measures (Table 3). Hypokyphotic patients reported less pain than normal/kyphotic Table 3 Preoperative SRS-22 scores. Domain Sagittal Cobb angle T5eT12 p value Pain Appearance Activity Mental Total patients with, a mean SRS Pain score of 4.15 versus 4.03 (p 5.044). Hypokyphotic patients also recorded better SRS Mental Health scores (4.06 vs. 3.92; p 5.026) and better SRS Total scores (3.92 vs. 3.83; p 5.027). The hypokyphotic group also had better SRS Appearance (3.36 vs. 3.30) and Activity scores (4.20 vs. 4.13), but these differences did not reach statistical significance. None of these differences approached minimally important clinical difference for the SRS instrument [12]. Surgical and perioperative parameters were equivalent in the two groups (Table 4). There was no difference in operative time (261 vs. 266 minutes), frequency of osteotomies (6% vs. 7%), or levels fused (10.0 vs. 9.9; p 5.680). However, the estimated blood loss in the hypokyphotic patients was statistically greater than in the normal/ kyphotic group (918 vs. 715 ml; p 5.004). There was no difference in hospital length of stay or incidence of perioperative complications. Complete 2-year postoperative data were available in 401 patients, including 91 hypokyphotic patients and 310 normal/kyphotic patients. We analyzed cases without 2-year follow-up and saw no differences in baseline demographics or radiographic parameters in the missing patients compared with the ultimate study cohort. There was no difference in the proportion between groups of patients who received stainless-steel or titanium rods. There was also no difference in the methods used to Table 4 Surgical data. Surgical parameter Sagittal Cobb angle T5eT12 p value Operative time Estimated blood loss Number of levels fused Implant material.186 Stainless steel Titanium Other 4 0 Correction technique.286 Cantilever 2 28 In situ translation 4 38 Compression Distraction Rod rotation 6 61 Direct apical vertebral rotation Smith-Petersen osteotomy 36 (6%) 12 (7%).466

4 S.D. Glassman et al. / Spine Deformity 1 (2013) 40e45 43 Table 5 Two-year postoperative SRS-22 scores. Domain Sagittal Cobb angle T5eT12 p value N Pain Appearance Activity Mental Satisfaction Total achieve correction between groups. The SRS-22 scores improved in both groups from baseline to 2 years postoperatively. The magnitude of change in health status was not different between groups. All of the SRS instrument component scores and SRS Total scores were equivalent in the 2 groups at 2 years postoperatively (Table 5). We performed a subanalysis of the 91 hypokyphotic patients, evaluating the SRS scores based on whether they remained hypokyphotic after surgery. The subanalysis showed no difference in any of the SRS domains (Table 6). Observed changes in radiographic parameters at 2 years postoperatively included a decrease in main thoracic Cobb angle of 32.7 in the hypokyphotic group and 34.5 in the normal/kyphotic group (Table 7). With this correction, the hypokyphotic group had a 13.9 increase in T5eT12 sagittal Cobb angle, whereas the normal/kyphotic group had a 1.2 decrease (p 5.000). There was also a difference with regard to lumbar lordosis: The hypokyphotic thoracic curve group gained 5.7 of lumbar lordosis and the normal/kyphotic group lost 1.2 of lumbar lordosis (p 5.000). Magnitude of change was not significantly different between groups for any of the other radiographic parameters. A sensitivity analysis, excluding the kyphotic patients, and thus comparing only the patients with normal versus hypokyphotic sagittal alignment, did not substantially alter any of the primary observations. We excluded 66 kyphotic patients; correspondingly, the size of the control group decreased to 549 patients. The preoperative mean SRS Pain scores were unchanged (4.03 vs. 4.15), but the p value increased above the.05 level (.054 vs..044). The significant differences in SRS Mental Health and SRS Total scores were maintained. Preoperative T5eT12 mean Table 6 Two-year postoperative SRS-22 scores in Hypokyphotic patients. Domain Postoperative sagittal Cobb angle T5eT12 p value N Pain Appearance Activity Mental Satisfaction Total Table 7 Two-year change in radiographic parameters. Radiographic parameter Sagittal Cobb angle T5eT12 p value Proximal thoracic Cobb angle Main thoracic Cobb angle Thoracolumbar Cobb angle Proximal thoracic AVT Main thoracic AVT Thoracolumbar AVT Radiographic coronal balance T1 tilt Clavicle angle Floman shift Pelvic obliquity Sacral obliquity Radiographic leg length discrepancy T2eT12 sagittal Cobb angle T2eT5 sagittal Cobb angle T5eT12 sagittal Cobb angle T10eL2 sagittal Cobb angle T12eS1 sagittal Cobb angle Sagittal balance Abbreviation as in Table 2. sagittal Cobb angle decreased slightly in the control group (22.2 vs ), but the significant differences between hypokyphotic and normal patients were unchanged. Similarly, there were no changes in either clinical or radiographic parameters at 2 years postoperatively, based on the exclusion of the kyphotic patients. Discussion Right thoracic AIS has been described as a lordoscoliosis, but that is not always the case. The marked predominance of main thoracic Lenke type 1A, B, or C curves analyzed in this study were associated with either a normal (10 to 40 )or kyphotic (greater than 40 ) sagittal thoracic profile. Our investigation focused on whether the hypokyphotic curves differ from the normal/kyphotic curves (either at baseline or with surgical treatment). We segregated the groups as T5eT12 sagittal Cobb angle less than 10 or greater than or equal to 10 based on the sagittal plane modifier within the Lenke classification system. We excluded right thoracic curves greater than 70 based on the concern that the rotation observed in large curves can be misinterpreted as increased kyphosis on biplanar radiography. Analysis of the preoperative data suggests that patients with hypokyphotic thoracic curves differ only slightly at baseline from patients with normal sagittal or kyphotic right thoracic curves. Patients with hypokyphotic curves had minimally better health status in all domains of the SRS instrument. Although differences in pain, mental health, and the SRS total score reached statistical significance, the differences did not approach a minimally important clinical difference threshold [12]. We performed a sensitivity analysis to determine whether differences

5 44 S.D. Glassman et al. / Spine Deformity 1 (2013) 40e45 between study groups might be attributed to poorer health status in the most kyphotic patients; however exclusion of patients with greater than 40 thoracic kyphosis changed none of the primary observations. In addition to the differences in thoracic sagittal profile, which defined the groups, patients with hypokyphotic curves also had slightly less lumbar lordosis and less negative sagittal balance. There was no difference between groups with regard to pulmonary function. Some of the measured differences seem to be intuitively linked to the observed sagittal alignment. In particular, the increase in lumbar lordosis in the normal/kyphotic group is likely to be a direct compensatory mechanism aimed at maintaining overall sagittal balance. Whereas the normal/ kyphotic group had slightly more negative sagittal balance, both groups fell within an acceptable range. A more detailed analysis would probably require an assessment of pelvic parameters, which were not available in this dataset. Perhaps more relevant, the hypokyphotic curves were more frequently listing to the right, which is consistent with the classic clinical description that primary right thoracic curves list to the right. With regard to baseline health status, we observed statistically significant differences between groups, but they were small. One plausible explanation might be that whereas kyphoscoliosis in theory may lead to inferior SRS-22 scores, the SRS instrument may not be sensitive enough to clearly detect a difference in these otherwise young, healthy patients with well-compensated spinal deformity. However, several observations conflict with this premise. First, the normal/ kyphotic group had a mean sagittal profile well within the normal range at baseline, and stratification of that group revealed no deterioration of health status as the T5eT12 Cobb angle increased. Second, health status improved postoperatively in the normal/kyphotic group with no measurable change in sagittal alignment. Another observation was that the hypokyphotic group was older at the time of study enrollment, which may suggest that hypokyphotic curves progress at a different rate than nonhypokyphotic ones. With surgical treatment, the hypokyphotic group had a 13.9 increase in thoracic sagittal Cobb magnitude, whereas the normal/kyphotic group had essentially no change (þ1.2 ). It is not clear whether this represents a concerted effort to restore kyphosis through specific reduction techniques, or a byproduct of patient positioning and standard curve correction. The 2 groups had a similar magnitude of coronal correction and equivalent coronal and sagittal standing balance postoperatively. Finally, the small differences in health status observed preoperatively equalized, such that the groups were similar for all domains of the SRS instrument at 2 years postoperatively. A subanalysis within the hypokyphotic group comparing those who remained hypokyphotic after surgery with those who had a normal/kyphotic sagittal curve after surgery showed no difference in any domain of the SRS instrument at 2 years. We could not determine a specific explanation for the slightly more favorable preoperative SRS-22 scores in the hypokyphotic right thoracic AIS patients. Global sagittal balance has been correlated with symptoms in adult scoliosis patients, but positive sagittal balance is unusual in children, and we did not observe it in either of our study groups. Another potential concern, decreased pulmonary function in the hypokyphotic curves, was similarly absent. An important limitation of the study was the use of standard posteroanterior and lateral 36-inch radiographs, which are suboptimal as a measure of rotational deformity. It is possible that a 3-dimensional assessment of lordosis at the apex would have altered our categorization of some curves. In particular, concern for inaccurate assessment of rotational deformity in larger curves led to the exclusion of curves greater than 70 in the study protocol. In addition, we examined multiple variables such that, despite a large study population, a Bonferroni effect might generate a spurious statistical observation. Measurable differences may be found in compensatory changes of the mobile axial spine in the cervical and pelvic regions, information that is unavailable in this database. Finally, we had 2-year follow-up data for only half the study population, which diminished the power and relevance of the observations regarding surgical treatment. Despite these limitations, the dataset challenges the concept that patients with thoracic hypokyphotic scoliosis differ from those with a normal or kyphotic sagittal alignment. Our results show no difference in baseline outcome scores in Lenke 1 AIS patients with hypokyphosis compared with patients with normal sagittal alignment. This raises the question as to whether restoration of thoracic lordosis in mildly hypokyphotic curves will alter outcome. Our limited 2-year postoperative data revealed no differences in clinical outcome, we observed although some restoration of thoracic kyphosis. References [1] Dickson RA, Lawton JO, Archer IA, Butt WP. The pathogenesis of idiopathic scoliosis. J Bone Joint Surg 1984;66B:8e15. [2] Ascani E, Bartolozzi P, Logroscino CA, et al. Natural history of untreated idiopathic scoliosis after skeletal maturity. Spine (Phila Pa 1976) 1986;11:784e9. [3] 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;14:717e21. [4] Betz RR. Kyphosis of the thoracic and thoracolumbar spine in the pediatric patient: normal sagittal parameters and scope of the problem. AOS Instruct Course Lect 2004;53:479e84. [5] Mac-Thiong JM, Roussouly P, Berthonnaud E, Guigui P. Sagittal parameters of global spinal balance: normative values from a prospective cohort of seven hundred nine Caucasian asymptomatic adults. Spine (Phila Pa 1976) 2010;35:E1193e8. [6] Lenke LG, Edwards CC, Bridwell KH. The Lenke classification of adolescent idiopathic scoliosis: how it organizes curve patterns as a template to perform selective fusions of the spine. Spine 2003;28: S199e207. [7] Lenke LG, Betz RR, Harms J, et al. Adolescent idiopathic scoliosis. A new classification to determine extent of spinal arthrodesis. J Bone Joint Surg 2001;83A:1169e81. [8] de Jonge T, Dubousset JF, Illes T. Sagittal plane correction in idiopathic scoliosis. Spine 2002;27:754e61.

6 S.D. Glassman et al. / Spine Deformity 1 (2013) 40e45 45 [9] Suk SI, Kim WJ, Kim JH, Lee SM. 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;12:489e95. [10] Asher MA, Lai SM, Glattes RC, et al. Refinement of the SRS-22 Health-Related Quality of Life questionnaire Function domain. Spine (Phila Pa 1976) 2006;31:593e7. [11] Floman Y, Penny JN, Micheli LJ, et al. Osteotomy of the fusion mass in scoliosis. J Bone Joint Surg 1982;64A:1307e16. [12] Carreon LY, Sanders JO, Diab M, et al. Spinal Deformity Study Group. The minimum clinically important difference in Scoliosis Research Society-22 Appearance, Activity, and Pain domains after surgical correction of adolescent idiopathic scoliosis. Spine (Phila Pa 1976) 2010;35:2079e83.