- Email: [email protected]
Sagittal Spinopelvic Parameters of Young Children With Scoliosis
Spine Deformity 1 (2013) 343e347 www.spine-deformity.org
Sagittal Spinopelvic Parameters of Young Children With Scoliosis Ron El-Hawary, MD, MSc, FRCS(C)a,*, Peter F. Sturm, MDb, Patrick J. Cahill, MDc, Amer F. Samdani, MDc, Michael G. Vitale, MDd, Peter G. Gabos, MDe, Nathan D. Bodin, MDf, Charles R. d’Amato, MDg, Colin Harris, MDh, Jason J. Howard, MDa, Susan H. Morris, PhDi, John T. Smith, MDj a
IWK Health Centre, 5850 University Avenue, P.O. Box 9700, Halifax, Nova Scotia B3K-6R8, Canada Crawford Spine Center, Pediatric Orthopaedic Surgery, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave. MLC 2017, Cincinnati, OH 45239, USA c Orthopedic Surgery, Shriners Hospital Philadelphia, Shriners Hospital Outpatnt Clinic, 3551 N Broad St Suite 4, Philadelphia, PA, 19140, USA d Clinical Orthopaedic Surgery, Columbia University, 116th Street and Broadway New York, NY, 10027 USA e Orthopaedic Surgery, A. I. duPont Institute, 1600 Rockland Rd, Wilmington, DE, 19803, USA f Orthopedic Surgery, Orthopedic and Neurosurgical Specialists, LLC807 N Haddon Ave Suite 1, Haddonfield, NJ, 08033, USA g Shriners Hospital Portland, Children Orthpaedics, 3101 SW Sam Jackson Park Rd, Portland, OR, 97239, USA h Orthopaedic Surgery, University of Medicine and Dentistry of New Jersey, United States, 2005-2009 i IWK Health Centre & Dalhousie University Division of Neurosurgery, Halifax, Nova Scotia, B3K 6R8, Canada j Primary Children’s Hospital, Pediatric Orthopaedics, 100 N Mario Capecchi Drive, Suite 4550, Salt Lake City, UT, 84113, USA Received 3 September 2012; revised 24 June 2013; accepted 4 July 2013 b
Abstract Study Design: Retrospective, multicenter review of the spinopelvic parameters in young children with scoliosis. Objectives: To describe sagittal alignment of the spine and pelvis in young children with scoliosis. Summary of Background Data: The natural history of spinopelvic parameters has been defined for the first 10 years of life in normal children; however, they have not been described for children with scoliosis. Such information is important because these values can be used as a baseline for the assessment of radiographic outcomes after surgical intervention. Methods: Seven measures of sagittal alignment were taken from standing lateral radiographs of 80 children with scoliosis (coronal Cobb angle greater than 50 ) and compared with age-matched normal children described in the literature. Statistical analysis was performed using 2-tailed Student t tests (level of significance 5 .05) and Pearson correlation coefficient. Results: Patients had a mean age of 4.8 years (range, 1e10 years) and a mean Cobb angle of 72.0 16 . Mean sagittal spine parameters were sagittal balance (2.2 4 cm), thoracic kyphosis (38.0 20.8 ), and lumbar lordosis (49.0 16.6 ). These values were similar to those of children without scoliosis. Mean sagittal pelvic parameters were: pelvic incidence (46.5 15.8 ), pelvic tilt (10.7 13.6 ), sacral slope (35.5 12.1 ), and pelvic radius (55.7 21.3 ). Pelvic incidence was not significantly different from that of age-matched
Author disclosures: RE (support for travel to meetings for the study or other purposes from the Chest Wall and Spine Deformity Study Group; consultancy for DePuy Spine; grants from Synthes Spine, DePuy Spine, Medtronic Canada, Atlantic Innovation Fund/Atlantic Canada Opportunities Agency, Canadian Institute of Health Research, Pediatric Orthopaedic Society of North America, Tecterra); PFS (consultancy for DePuy Spine; grants from DePuy Spine; stock/stock options from Pioneer Surgical); PJC (consultancy for DePuy Synthes Spine; payment for manuscript preparation from DePuy Synthes Spine; editorial board member for Orthopedics [unpaid]); AFS (consultancy for DePuy Synthes Spine, Zimmer Spine, SpineGuard, Stryker; payment for manuscript preparation for DePuy Synthes Spine); MGV (board membership with Chest Wall and Spine Deformity Study Group, AAP Section on Orthopedics, POSNA; consultancy for Stryker, Biomet, Chest Wall and Spine Deformity Study Group; grants from Scoliosis Research Society; grants from Chest Wall and Spine Deformity Research Foundation, POSNA, OREF, OMeGA, 2212-134X/$ - see front matter Ó 2013 Scoliosis Research Society. http://dx.doi.org/10.1016/j.jspd.2013.07.001
AOSpine, Medtronic; royalties from Biomet; travel/accommodations/meeting expenses from Chest Wall and Spine Deformity Study Group, Fox PSDSG, Broadwater [funded by Biomet, Synthes, Stryker, Medtronic, K2]); PGG (consultancy for DePuy Spine; payment for lectures including service on speakers bureaus from DePuy Spine); NDB (grants from Synthes to author’s institution); CRD (support for travel to meetings for the study or other purposes from Chest Wall and Spine Deformity Study Group); CH (none); JJH (grants from Atlantic Innovation Fund and Tecterra); SHM (grants from Globus Medical Quality, Safety, Value Spinal Deformity Research Grant); JTS (board membership with Chest Wall and Spine Deformity Study Group; consultancy for DePuy-Synthes Spine; royalties from VEPTR 2 Device). *Corresponding author. IWK Health Centre, 5850 University Avenue, P.O. Box 9700, Halifax, Nova Scotia B3K-6R8, Canada. Tel.: (902) 4707245; fax: (902) 470-7341 E-mail address: [email protected] (R. El-Hawary).
344
R. El-Hawary et al. / Spine Deformity 1 (2013) 343e347
normal children; however, pelvic tilt was significantly higher and sacral slope was significantly lower in children with scoliosis. Conclusions: Sagittal plane spine parameters and some pelvic parameters were similar for young children with scoliosis versus age-matched normal children; however, children with scoliosis showed signs of increased pelvic tilt and decreased sacral slope. These values can be used as a baseline for both the natural history and the assessment of radiographic outcomes after surgical intervention. Ó 2013 Scoliosis Research Society. Keywords: Early-onset scoliosis; Sagittal plane spinopelvic parameters; Baseline measures; Prognosis; Surgical treatment
Introduction Spinopelvic parameters derived from standing lateral radiographs are used to describe the orientation, shape, and morphology of the spine and pelvis. These measures of sagittal alignment include sagittal balance, thoracic kyphosis, lumbar lordosis, pelvic incidence, pelvic tilt, and sacral slope (Fig. 1). These measures are unique to individuals; taken together, they provide the basis for assessing postural alignment and standing balance [1-5]. Spinopelvic parameters are not strictly of radiographic interest, but have been shown in the adult literature to be correlated to health-related quality of life [6]. Spinopelvic parameters and their natural history during the first 10 years of life have been defined for normal children [7]; however, like health-related quality of life, they have not been described for young children with scoliosis. Such information may be important for patients undergoing surgical treatment, especially those being treated with growing systems that require periodic lengthening through to skeletal maturity. If
the progression and stabilization of sagittal alignment and standing balance in this population are different from those of normal children, the timing and aggressiveness of these lengthenings may need to be re-examined. Furthermore, prediction of postoperative complications such as proximal junctional kyphosis and implant failure will be better understood if the natural history of spinopelvic parameters in children with scoliosis is defined. The purpose of this study was twofold: 1) to define sagittal plane spinopelvic parameters in a group of young children with scoliosis; and 2) to define the variability associated with the measurement of these parameters. Spinopelvic parameters defined in a group of young children with scoliosis will act as a baseline for the prognosis and treatment of this population. These new baselines will improve surgical planning and prediction of postoperative complications, such as proximal junctional kyphosis and implant failure, after implantation of growth-friendly systems. Materials and Methods A total of 80 young children with scoliosis from 7 different institutions were reviewed. Inclusion criteria included age 1e10 years, ambulatory status, Cobb angle of greater than 50 , and no previous history of spine surgery. Standing lateral radiographs were taken with each child assuming a comfortable position with the hips and knees fully extended (Fig. 2). Radiographs were evaluated for the following sagittal plane spine and pelvic measurements: sagittal balance, thoracic kyphosis, lumbar lordosis, pelvic incidence, pelvic tilt, sacral slope, and pelvic radius angle. These parameters were then compared with previously established data from age-matched children without spinal deformity [7] using 2-tailed Student t tests, employing the OpenEpi software program (www.openepi.com). Two independent observers then assessed the variability of the sagittal-plane pelvic parameters. In a subgroup of 11 subjects, they performed and repeated these measurements after 15 days. Interobserver and intra-observer variability were tested using the 2-tailed Student t test and Pearson correlation coefficient, respectively (SPSS 15.0), ibm, Armonk, NY). Results
Fig. 1. Radiograph of a skeletally immature patient with idiopathic scoliosis demonstrating sagittal pelvic parameters. PI, pelvic incidence; PT, pelvic tilt; SS, sacral slope.
A total of 80 subjects, with a mean age of 4.8 years (range, 1 e10 years) and a mean Cobb angle of 72.0 16 , were included in this study. Using the Classification for Early Onset Scoliosis, their diagnoses included
R. El-Hawary et al. / Spine Deformity 1 (2013) 343e347
345
Fig. 2. Standing posteroanterior and lateral radiographs of a 7-year-old boy with left thoracic infantile idiopathic scoliosis. The sagittal profile includes 51 thoracic kyphosis, 60 lumbar lordosis, 66 pelvic incidence, 28 pelvic tilt, 39 sacral slope, and 58 pelvic radius angle.
congenital (n 5 38), neuromuscular (n 5 5), syndromic (n 5 21), and idiopathic (n 5 16). The most common subclassifcations were C3N (n 5 15), S3N (n 5 8), and I3N (n 5 7). Mean sagittal balance was 2.2 cm 4 cm for this group of subjects with scoliosis. A comparison between the remaining sagittal-plane spinopelvic parameters for this group of subjects with scoliosis and the published age-matched normal population [7] demonstrated that thoracic kyphosis and lumbar lordosis were not significantly different. Of the mean sagittal-plane pelvic parameters, pelvic incidence was not statistically different from the normal group; however, pelvic tilt and sacral slope were significantly different. The mean pelvic tilt for the scoliosis group was 10.7 13.6 versus 4.3 8.1 for the normal group, and the mean sacral slope for the scoliosis group was 35.5 12.1 versus 40.3 8.7 for the normal group (p ! .05) (Table 1). Compared with age-matched controls, these findings of increased pelvic tilt and decreased sacral slope for the entire study group were also observed for the congenital (p ! .05), idiopathic, and neuromuscular diagnoses (Table 2). For the syndromic group, pelvic incidence and pelvic tilt were higher than those published for the control group (p ! .05) (Table 2). A comparison was performed between 2 age groups of the scoliosis population: 1e5 years (n 5 53; mean, 3.4 years) versus 6e10 years (n 5 27; mean, 7.7 years). The
older age group had significantly greater pelvic incidence, greater modified pelvic radius angle, and greater sacral slope than the younger age group (p ! .05) (Table 3). The subgroup of subjects who were included in the variablity analysis had a mean age of 5.7 years and a mean scoliosis of 80.8 . Repeated measurements by a single observer (intraobserver variability) demonstrated no significant differences: pelvic incidence (56.4 vs. 52.5 ), pelvic tilt (15.4 vs. 12.6 ), sacral slope (41.5 vs. 39.9 ), and pelvic radius angle (70.5 vs. 69.2 ). Evaluation of these data with paired-sample correlations demonstrated a moderate correlation between measurements of pelvic incidence (.564), whereas stronger correlations were shown
Table 1 Comparison of mean sagittal spinopelvic parameters in scoliosis patients aged 1e10 years versus asymptomatic children aged 0e10 years [7]. Parameter
Scoliosis patients (n580)
Mac-Thiong et al. [7] (n535)
p
Sagittal balance, cm Thoracic kyphosis ( ) Lumbar lordosis ( ) Pelvic incidence ( ) Pelvic tilt ( ) Sacral slope ( ) Modified PRA ( )
þ2.2 38 49 46.5 10.7 35.5 55.7
e 38.3 45.6 44.6 4.3 40.3 e
e .92 .22 .46 .002* .02* e
4 20.8 16.6 15.8 13.6 12.1 21.3
PRA, proximal radius angle. Indicates statistical significance.
*
9.8 12.1 10.6 8.1 8.7
346
R. El-Hawary et al. / Spine Deformity 1 (2013) 343e347
Table 2 Comparison of mean sagittal spinopelvic parameters in scoliosis patients aged 1e10 years, separated by type of scoliosis, versus asymptomatic children aged 0e10 years [7]. Scoliosis type and parameter
Scoliosis patients
Mac-Thiong et al. [7]
p
Congenital (n538) Mean age 5 4.4 years Sagittal balance, cm Thoracic kyphosis Lumbar lordosis Pelvic incidence Pelvic tilt Sacral slope Modified PRA Idiopathic (n516) Mean age 5 4.5 years Sagittal balance, cm Thoracic kyphosis Lumbar lordosis Pelvic incidence Pelvic tilt Sacral slope Modified PRA Neuromuscular (n55) Mean age 5 3.7 years Sagittal balance, cm Thoracic kyphosis Lumbar lordosis Pelvic incidence Pelvic tilt Sacral slope Modified PRA Syndromic (n521) Mean age 5 6.07 years Sagittal balance, cm Thoracic kyphosis Lumbar lordosis Pelvic incidence Pelvic tilt Sacral slope Modified PRA
þ2.5 4.5
e
e
38.8 45 42.7 11.1 32.1 52.3
38.3 45.6 44.6 4.3 40.3 e
*
22 17.4 15.2 16.9 11.1 19.9
þ1.5 3
e
41.1 52.2 45.4 8.3 36.8 55.2
38.3 45.6 44.6 4.3 40.3 e
18 12.7 11.6 11.8 4.8 21.9
e
44.4 50 45.8 12 30 52.4
38.3 45.6 44.6 4.3 40.3 e
15 11 16.8 12.2 16.9 32.3
e
32.8 53.7 54.4 12.2 41.7 63
38.3 45.6 44.6 4.3 40.3 e
21.7 18.1 17.4 8.6 14.4 20
.90 .86 .55 .04* .0008* e
9.8 12.1 10.6 8.1 8.7
.57 .09 .81 .23 .07 e e
þ1.7 3.6
9.8 12.1 10.6 8.1 8.7
e
þ3.5 4.7
9.8 12.1 10.6 8.1 8.7
.42 .45 .88 .23 .25 e e
9.8 12.1 10.6 8.1 8.7
.28 .08 .03* .002* .69 e
Indicates statistical significance.
for measurements of sacral slope (.947), pelvic tilt (.816), and pelvic radius angle (.789). Average measurements between observers for the interobserver data were: pelvic Table 3 Mean sagittal spinopelvic parameters in scoliosis patients aged 1e5 years versus scoliosis patients aged 6e10 years. Parameter
Scoliosis 1e5 years (mean age, 3.4 years) (n553)
Scoliosis 6e10 years (mean age, 7.7 years) (n527)
p
Sagittal balance, cm Thoracic kyphosis ( ) Lumbar lordosis ( ) Pelvic incidence ( ) Pelvic tilt ( ) Sacral slope ( ) Modified PRA ( )
þ2.3 36.9 46.5 43.6 8.4 32.8 50.6
þ1.9 40.2 54.1 52.4 15.5 40.6 66.1
.66 .51 .053 .03* .07 .008* .0029*
*
4.3 20.8 16.4 14.1 10.7 11.2 20
Indicates statistical significance.
3.5 20.9 16.2 17.5 17.6 12.4 20.3
incidence (56.4 vs. 50.0 ), pelvic tilt (15.4 vs. 15.1 ), sacral slope (41.3 vs. 33.9 ), and pelvic radius angle (70.5 vs. 71.5 ). There were no significant differences between observers in spinopelvic measures except for sacral slope, which was highly significantly different (p 5 .003). Discussion The major goals of this study were to define sagittalplane spinopelvic parameters for a group of young children with scoliosis and to compare them with a previously published group of children without spinal deformity. A weakness of this study is that it was a retrospective, multicenter study in which radiographic techniques were not strictly standardized. In an effort to provide stronger conclusions for future research, this study group is in the process of standardizing radiographic techniques between institutions. In the meantime, the authors believe that these results are still valid and provide valuable information on the sagittal plane alignment of young children with scoliosis. With the exception of increased pelvic tilt and decreased sacral slope, sagittal-plane spinopelvic parameters were similar between children with scoliosis and those reported for age-matched children without spinal deformity. These findings are important because they establish a baseline for determining the natural history and prognosis for children with scoliosis undergoing treatment with growing systems. Historically, prognostication for this population was based on the natural history of children without spinal deformity. This turned out to be a valid proposition with the exception of pelvic tilt and sacral slope, which were significantly different in children with spinal deformity. The role of pelvic tilt is to keep the sacral plate posterior to the axis of the hips, to establish the center of gravity over the lower extremities and maintain standing balance [7,8]. This study found that the pelvic tilt of children with spinal deformity was significantly greater than for normal children. Because the scoliosis group had a mean positive sagittal balance of 2.2 cm, pelvic tilt may have increased secondarily in an effort to maintain the center of gravity from displacing too far anterior. The pelvic incidence was not significantly different between the scoliosis (mean, 46.5 15.8 ) and normal groups (mean, 44.6 10.6 ) (p 5 .46). Because pelvic incidence is the summation of pelvic tilt and sacral slope, it would be expected that because there was no intergroup difference in pelvic incidence, the decreased sacral slope (35.5 12.1 vs. 40.3 ; p ! .05) was a compensatory mechanism to counterbalance the increased pelvic tilt in the scoliosis group. These alterations did not significantly affect lumbar lordosis or thoracic kyphosis. Compared with age-matched controls, the findings of increased pelvic tilt and decreased sacral slope for the entire study group (p ! .05) were also found for the congenital, idiopathic, and neuromuscular diagnoses;
R. El-Hawary et al. / Spine Deformity 1 (2013) 343e347
however, they were significant only for the congenital diagnosis. This is likely related to the small sample sizes for the other diagnoses. This finding was not observed for the syndromic diagnosis and was likely related to more heterogeneity within the syndromic etiology. Because it was documented for children without spinal deformity that sagittal plane alignment changes with age, the authors compared 2 age groups within the study population (1e5 vs. 6e10 years) [4,7,9,10]. These 2 age groups were chosen based on the age-related changes in spinal growth velocity. The findings of higher pelvic incidence, higher pelvic radius angle, and higher sacral slope in the older age group were similar to those found in the literature for children without scoliosis [4,7,10]. In addition, the current study’s trends toward higher thoracic kyphosis, higher lumbar lordosis, and less positive sagittal balance in the 6- to 10-year-old group agree with what Cil et al. [10] found for patients without scoliosis. The long-term effects of positive sagittal balance and increased pelvic retroversion were defined by Lafage et al. [6] for adult patients with spinal deformity. That group determined that increased pelvic retroversion resulted in poor health-related quality of life outcomes, including an increased Oswestry Disability Index and a decreased Short Forme12 Physical Component score. Thus, the long-term natural history of increased retroversion in children with scoliosis may be increased pain and disability. As a result, one of the goals of treatment for young children with scoliosis should be to improve their sagittal alignment, specifically pelvic tilt. Awareness of the nuances and compensatory mechanisms that scoliosis patients make to maintain standing balance will invariably affect the end point expectations of treatment with growing systems. It may also predict implant failure and other complications such as junctional kyphosis. Pelvic retroversion (ie, increased pelvic tilt) has been linked to spondylolisthesis in adolescents and young adults [9]. Given that the scoliosis group in this study demonstrated significant baseline pelvic retroversion, this factor could help predict construct failure and junctional kyphosis in a subset treated with growing systems. For distraction-based surgeries, it is possible that further positive sagittal imbalance may result from periodic posterior lengthening procedures. The second goal of this study was to determine the intrarater and inter-rater reliability of the different spinopelvic measures. Interestingly, sacral slope was the most reliable measure when a single observer performed repeated measures, but the least reliable between observers. Intraobserver variability was lowest for the measurements of pelvic incidence, because only a moderate correlation was noted with paired-sample correlations. These differences
347
may be related to difficulties in identifying the landmarks on the immature sacrum that are needed to measure pelvic incidence (Fig. 2). At the authors’ institution, a study is currently being performed to compare the variability of pelvic incidence measured on radiographs versus pelvic incidence measured on magnetic resonance imaging, where the cartilaginous sacral end plate can be better visualized. The pelvic radius angle, which is used by some investigators as a surrogate for pelvic incidence, performed well with both intra-observer and interobserver variability. This measure is not as reliant on the sacral end plate anatomy and is potentially more useful than pelvic incidence for this immature population. In conclusion, baselines values for spinopelvic parameters have been established for young children with scoliosis. These parameters were compared with those of children without spinal deformity and were the same except for a significantly greater pelvic tilt and a significantly lower sacral slope in children with scoliosis. The intra-rater and inter-rater reliability of these measures was fairly strong, with the exception of poor inter-rater reliability for the measurement of sacral slope. These values can be used as a baseline for both the natural history and for assessment of radiographic outcomes after surgical intervention. References [1] Jackson RP, Hales C. Congruent spinopelvic alignment on standing lateral radiographs of adult volunteers. Spine (Phila Pa 1976) 2000;25:2808e15. [2] Legaye J, Hecquet J, Marty C, et al. Sagittal equilibration of the spine: relationship between pelvis and sagittal spinal curves in the standing position [in French]. Rachis 1993;5:215e26. [3] Legaye J, Duval-Beaupere G, Hecquet J, et al. Pelvic incidence: a fundamental pelvic parameter for three-dimensional regulation of spinal sagittal curves. Eur Spine J 1998;7:99e103. [4] Marty C, Boisaubert B, Descamps H, et al. The sagittal anatomy of the sacrum among young adults, infants, and spondylolisthesis patients. Eur Spine J 2002;11:119e25. [5] Rajnics P, Pomero V, Templier A, et al. Computer-assisted assessment of spinal sagittal plane radiographs. J Spinal Disord 2001;14: 135e42. [6] Lafage V, Schwab F, Patel A, et al. Pelvic tilt and truncal inclination: two key radiographic parameters in the setting of adults with spinal deformity. Spine (Phila Pa 1976) 2009;34:E599e606. [7] Mac-Thiong JM, Berthonnaud IE, Dimar JR, et al. Sagittal alignment of the spine and pelvis during growth. Spine (Phila Pa 1976) 2004;29:1642e7. [8] Vaz G, Roussouly P, Berthonnaud E, Dimnet J. Sagittal morphology and equilibrium of pelvis and spine. Eur Spine J 2002;11:80e7. [9] Labelle H, Roussouly P, Berthonnaud E, et al. Spondylolisthesis, pelvic incidence, and spinopelvic balance: a correlational study. Spine (Phila Pa 1976) 2004;18:2049e54. [10] Cil A, Yazici M, Uzumcugil A, et al. The evolution of sagittal segmental alignment of the spine during childhood. Spine (Phila Pa 1976) 2005;30:93e100.
Sagittal Spinopelvic Parameters of Young Children With Scoliosis Ron El-Hawary, MD, MSc, FRCS(C)a,*, Peter F. Sturm, MDb, Patrick J. Cahill, MDc, Amer F. Samdani, MDc, Michael G. Vitale, MDd, Peter G. Gabos, MDe, Nathan D. Bodin, MDf, Charles R. d’Amato, MDg, Colin Harris, MDh, Jason J. Howard, MDa, Susan H. Morris, PhDi, John T. Smith, MDj a
IWK Health Centre, 5850 University Avenue, P.O. Box 9700, Halifax, Nova Scotia B3K-6R8, Canada Crawford Spine Center, Pediatric Orthopaedic Surgery, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave. MLC 2017, Cincinnati, OH 45239, USA c Orthopedic Surgery, Shriners Hospital Philadelphia, Shriners Hospital Outpatnt Clinic, 3551 N Broad St Suite 4, Philadelphia, PA, 19140, USA d Clinical Orthopaedic Surgery, Columbia University, 116th Street and Broadway New York, NY, 10027 USA e Orthopaedic Surgery, A. I. duPont Institute, 1600 Rockland Rd, Wilmington, DE, 19803, USA f Orthopedic Surgery, Orthopedic and Neurosurgical Specialists, LLC807 N Haddon Ave Suite 1, Haddonfield, NJ, 08033, USA g Shriners Hospital Portland, Children Orthpaedics, 3101 SW Sam Jackson Park Rd, Portland, OR, 97239, USA h Orthopaedic Surgery, University of Medicine and Dentistry of New Jersey, United States, 2005-2009 i IWK Health Centre & Dalhousie University Division of Neurosurgery, Halifax, Nova Scotia, B3K 6R8, Canada j Primary Children’s Hospital, Pediatric Orthopaedics, 100 N Mario Capecchi Drive, Suite 4550, Salt Lake City, UT, 84113, USA Received 3 September 2012; revised 24 June 2013; accepted 4 July 2013 b
Abstract Study Design: Retrospective, multicenter review of the spinopelvic parameters in young children with scoliosis. Objectives: To describe sagittal alignment of the spine and pelvis in young children with scoliosis. Summary of Background Data: The natural history of spinopelvic parameters has been defined for the first 10 years of life in normal children; however, they have not been described for children with scoliosis. Such information is important because these values can be used as a baseline for the assessment of radiographic outcomes after surgical intervention. Methods: Seven measures of sagittal alignment were taken from standing lateral radiographs of 80 children with scoliosis (coronal Cobb angle greater than 50 ) and compared with age-matched normal children described in the literature. Statistical analysis was performed using 2-tailed Student t tests (level of significance 5 .05) and Pearson correlation coefficient. Results: Patients had a mean age of 4.8 years (range, 1e10 years) and a mean Cobb angle of 72.0 16 . Mean sagittal spine parameters were sagittal balance (2.2 4 cm), thoracic kyphosis (38.0 20.8 ), and lumbar lordosis (49.0 16.6 ). These values were similar to those of children without scoliosis. Mean sagittal pelvic parameters were: pelvic incidence (46.5 15.8 ), pelvic tilt (10.7 13.6 ), sacral slope (35.5 12.1 ), and pelvic radius (55.7 21.3 ). Pelvic incidence was not significantly different from that of age-matched
Author disclosures: RE (support for travel to meetings for the study or other purposes from the Chest Wall and Spine Deformity Study Group; consultancy for DePuy Spine; grants from Synthes Spine, DePuy Spine, Medtronic Canada, Atlantic Innovation Fund/Atlantic Canada Opportunities Agency, Canadian Institute of Health Research, Pediatric Orthopaedic Society of North America, Tecterra); PFS (consultancy for DePuy Spine; grants from DePuy Spine; stock/stock options from Pioneer Surgical); PJC (consultancy for DePuy Synthes Spine; payment for manuscript preparation from DePuy Synthes Spine; editorial board member for Orthopedics [unpaid]); AFS (consultancy for DePuy Synthes Spine, Zimmer Spine, SpineGuard, Stryker; payment for manuscript preparation for DePuy Synthes Spine); MGV (board membership with Chest Wall and Spine Deformity Study Group, AAP Section on Orthopedics, POSNA; consultancy for Stryker, Biomet, Chest Wall and Spine Deformity Study Group; grants from Scoliosis Research Society; grants from Chest Wall and Spine Deformity Research Foundation, POSNA, OREF, OMeGA, 2212-134X/$ - see front matter Ó 2013 Scoliosis Research Society. http://dx.doi.org/10.1016/j.jspd.2013.07.001
AOSpine, Medtronic; royalties from Biomet; travel/accommodations/meeting expenses from Chest Wall and Spine Deformity Study Group, Fox PSDSG, Broadwater [funded by Biomet, Synthes, Stryker, Medtronic, K2]); PGG (consultancy for DePuy Spine; payment for lectures including service on speakers bureaus from DePuy Spine); NDB (grants from Synthes to author’s institution); CRD (support for travel to meetings for the study or other purposes from Chest Wall and Spine Deformity Study Group); CH (none); JJH (grants from Atlantic Innovation Fund and Tecterra); SHM (grants from Globus Medical Quality, Safety, Value Spinal Deformity Research Grant); JTS (board membership with Chest Wall and Spine Deformity Study Group; consultancy for DePuy-Synthes Spine; royalties from VEPTR 2 Device). *Corresponding author. IWK Health Centre, 5850 University Avenue, P.O. Box 9700, Halifax, Nova Scotia B3K-6R8, Canada. Tel.: (902) 4707245; fax: (902) 470-7341 E-mail address: [email protected] (R. El-Hawary).
344
R. El-Hawary et al. / Spine Deformity 1 (2013) 343e347
normal children; however, pelvic tilt was significantly higher and sacral slope was significantly lower in children with scoliosis. Conclusions: Sagittal plane spine parameters and some pelvic parameters were similar for young children with scoliosis versus age-matched normal children; however, children with scoliosis showed signs of increased pelvic tilt and decreased sacral slope. These values can be used as a baseline for both the natural history and the assessment of radiographic outcomes after surgical intervention. Ó 2013 Scoliosis Research Society. Keywords: Early-onset scoliosis; Sagittal plane spinopelvic parameters; Baseline measures; Prognosis; Surgical treatment
Introduction Spinopelvic parameters derived from standing lateral radiographs are used to describe the orientation, shape, and morphology of the spine and pelvis. These measures of sagittal alignment include sagittal balance, thoracic kyphosis, lumbar lordosis, pelvic incidence, pelvic tilt, and sacral slope (Fig. 1). These measures are unique to individuals; taken together, they provide the basis for assessing postural alignment and standing balance [1-5]. Spinopelvic parameters are not strictly of radiographic interest, but have been shown in the adult literature to be correlated to health-related quality of life [6]. Spinopelvic parameters and their natural history during the first 10 years of life have been defined for normal children [7]; however, like health-related quality of life, they have not been described for young children with scoliosis. Such information may be important for patients undergoing surgical treatment, especially those being treated with growing systems that require periodic lengthening through to skeletal maturity. If
the progression and stabilization of sagittal alignment and standing balance in this population are different from those of normal children, the timing and aggressiveness of these lengthenings may need to be re-examined. Furthermore, prediction of postoperative complications such as proximal junctional kyphosis and implant failure will be better understood if the natural history of spinopelvic parameters in children with scoliosis is defined. The purpose of this study was twofold: 1) to define sagittal plane spinopelvic parameters in a group of young children with scoliosis; and 2) to define the variability associated with the measurement of these parameters. Spinopelvic parameters defined in a group of young children with scoliosis will act as a baseline for the prognosis and treatment of this population. These new baselines will improve surgical planning and prediction of postoperative complications, such as proximal junctional kyphosis and implant failure, after implantation of growth-friendly systems. Materials and Methods A total of 80 young children with scoliosis from 7 different institutions were reviewed. Inclusion criteria included age 1e10 years, ambulatory status, Cobb angle of greater than 50 , and no previous history of spine surgery. Standing lateral radiographs were taken with each child assuming a comfortable position with the hips and knees fully extended (Fig. 2). Radiographs were evaluated for the following sagittal plane spine and pelvic measurements: sagittal balance, thoracic kyphosis, lumbar lordosis, pelvic incidence, pelvic tilt, sacral slope, and pelvic radius angle. These parameters were then compared with previously established data from age-matched children without spinal deformity [7] using 2-tailed Student t tests, employing the OpenEpi software program (www.openepi.com). Two independent observers then assessed the variability of the sagittal-plane pelvic parameters. In a subgroup of 11 subjects, they performed and repeated these measurements after 15 days. Interobserver and intra-observer variability were tested using the 2-tailed Student t test and Pearson correlation coefficient, respectively (SPSS 15.0), ibm, Armonk, NY). Results
Fig. 1. Radiograph of a skeletally immature patient with idiopathic scoliosis demonstrating sagittal pelvic parameters. PI, pelvic incidence; PT, pelvic tilt; SS, sacral slope.
A total of 80 subjects, with a mean age of 4.8 years (range, 1 e10 years) and a mean Cobb angle of 72.0 16 , were included in this study. Using the Classification for Early Onset Scoliosis, their diagnoses included
R. El-Hawary et al. / Spine Deformity 1 (2013) 343e347
345
Fig. 2. Standing posteroanterior and lateral radiographs of a 7-year-old boy with left thoracic infantile idiopathic scoliosis. The sagittal profile includes 51 thoracic kyphosis, 60 lumbar lordosis, 66 pelvic incidence, 28 pelvic tilt, 39 sacral slope, and 58 pelvic radius angle.
congenital (n 5 38), neuromuscular (n 5 5), syndromic (n 5 21), and idiopathic (n 5 16). The most common subclassifcations were C3N (n 5 15), S3N (n 5 8), and I3N (n 5 7). Mean sagittal balance was 2.2 cm 4 cm for this group of subjects with scoliosis. A comparison between the remaining sagittal-plane spinopelvic parameters for this group of subjects with scoliosis and the published age-matched normal population [7] demonstrated that thoracic kyphosis and lumbar lordosis were not significantly different. Of the mean sagittal-plane pelvic parameters, pelvic incidence was not statistically different from the normal group; however, pelvic tilt and sacral slope were significantly different. The mean pelvic tilt for the scoliosis group was 10.7 13.6 versus 4.3 8.1 for the normal group, and the mean sacral slope for the scoliosis group was 35.5 12.1 versus 40.3 8.7 for the normal group (p ! .05) (Table 1). Compared with age-matched controls, these findings of increased pelvic tilt and decreased sacral slope for the entire study group were also observed for the congenital (p ! .05), idiopathic, and neuromuscular diagnoses (Table 2). For the syndromic group, pelvic incidence and pelvic tilt were higher than those published for the control group (p ! .05) (Table 2). A comparison was performed between 2 age groups of the scoliosis population: 1e5 years (n 5 53; mean, 3.4 years) versus 6e10 years (n 5 27; mean, 7.7 years). The
older age group had significantly greater pelvic incidence, greater modified pelvic radius angle, and greater sacral slope than the younger age group (p ! .05) (Table 3). The subgroup of subjects who were included in the variablity analysis had a mean age of 5.7 years and a mean scoliosis of 80.8 . Repeated measurements by a single observer (intraobserver variability) demonstrated no significant differences: pelvic incidence (56.4 vs. 52.5 ), pelvic tilt (15.4 vs. 12.6 ), sacral slope (41.5 vs. 39.9 ), and pelvic radius angle (70.5 vs. 69.2 ). Evaluation of these data with paired-sample correlations demonstrated a moderate correlation between measurements of pelvic incidence (.564), whereas stronger correlations were shown
Table 1 Comparison of mean sagittal spinopelvic parameters in scoliosis patients aged 1e10 years versus asymptomatic children aged 0e10 years [7]. Parameter
Scoliosis patients (n580)
Mac-Thiong et al. [7] (n535)
p
Sagittal balance, cm Thoracic kyphosis ( ) Lumbar lordosis ( ) Pelvic incidence ( ) Pelvic tilt ( ) Sacral slope ( ) Modified PRA ( )
þ2.2 38 49 46.5 10.7 35.5 55.7
e 38.3 45.6 44.6 4.3 40.3 e
e .92 .22 .46 .002* .02* e
4 20.8 16.6 15.8 13.6 12.1 21.3
PRA, proximal radius angle. Indicates statistical significance.
*
9.8 12.1 10.6 8.1 8.7
346
R. El-Hawary et al. / Spine Deformity 1 (2013) 343e347
Table 2 Comparison of mean sagittal spinopelvic parameters in scoliosis patients aged 1e10 years, separated by type of scoliosis, versus asymptomatic children aged 0e10 years [7]. Scoliosis type and parameter
Scoliosis patients
Mac-Thiong et al. [7]
p
Congenital (n538) Mean age 5 4.4 years Sagittal balance, cm Thoracic kyphosis Lumbar lordosis Pelvic incidence Pelvic tilt Sacral slope Modified PRA Idiopathic (n516) Mean age 5 4.5 years Sagittal balance, cm Thoracic kyphosis Lumbar lordosis Pelvic incidence Pelvic tilt Sacral slope Modified PRA Neuromuscular (n55) Mean age 5 3.7 years Sagittal balance, cm Thoracic kyphosis Lumbar lordosis Pelvic incidence Pelvic tilt Sacral slope Modified PRA Syndromic (n521) Mean age 5 6.07 years Sagittal balance, cm Thoracic kyphosis Lumbar lordosis Pelvic incidence Pelvic tilt Sacral slope Modified PRA
þ2.5 4.5
e
e
38.8 45 42.7 11.1 32.1 52.3
38.3 45.6 44.6 4.3 40.3 e
*
22 17.4 15.2 16.9 11.1 19.9
þ1.5 3
e
41.1 52.2 45.4 8.3 36.8 55.2
38.3 45.6 44.6 4.3 40.3 e
18 12.7 11.6 11.8 4.8 21.9
e
44.4 50 45.8 12 30 52.4
38.3 45.6 44.6 4.3 40.3 e
15 11 16.8 12.2 16.9 32.3
e
32.8 53.7 54.4 12.2 41.7 63
38.3 45.6 44.6 4.3 40.3 e
21.7 18.1 17.4 8.6 14.4 20
.90 .86 .55 .04* .0008* e
9.8 12.1 10.6 8.1 8.7
.57 .09 .81 .23 .07 e e
þ1.7 3.6
9.8 12.1 10.6 8.1 8.7
e
þ3.5 4.7
9.8 12.1 10.6 8.1 8.7
.42 .45 .88 .23 .25 e e
9.8 12.1 10.6 8.1 8.7
.28 .08 .03* .002* .69 e
Indicates statistical significance.
for measurements of sacral slope (.947), pelvic tilt (.816), and pelvic radius angle (.789). Average measurements between observers for the interobserver data were: pelvic Table 3 Mean sagittal spinopelvic parameters in scoliosis patients aged 1e5 years versus scoliosis patients aged 6e10 years. Parameter
Scoliosis 1e5 years (mean age, 3.4 years) (n553)
Scoliosis 6e10 years (mean age, 7.7 years) (n527)
p
Sagittal balance, cm Thoracic kyphosis ( ) Lumbar lordosis ( ) Pelvic incidence ( ) Pelvic tilt ( ) Sacral slope ( ) Modified PRA ( )
þ2.3 36.9 46.5 43.6 8.4 32.8 50.6
þ1.9 40.2 54.1 52.4 15.5 40.6 66.1
.66 .51 .053 .03* .07 .008* .0029*
*
4.3 20.8 16.4 14.1 10.7 11.2 20
Indicates statistical significance.
3.5 20.9 16.2 17.5 17.6 12.4 20.3
incidence (56.4 vs. 50.0 ), pelvic tilt (15.4 vs. 15.1 ), sacral slope (41.3 vs. 33.9 ), and pelvic radius angle (70.5 vs. 71.5 ). There were no significant differences between observers in spinopelvic measures except for sacral slope, which was highly significantly different (p 5 .003). Discussion The major goals of this study were to define sagittalplane spinopelvic parameters for a group of young children with scoliosis and to compare them with a previously published group of children without spinal deformity. A weakness of this study is that it was a retrospective, multicenter study in which radiographic techniques were not strictly standardized. In an effort to provide stronger conclusions for future research, this study group is in the process of standardizing radiographic techniques between institutions. In the meantime, the authors believe that these results are still valid and provide valuable information on the sagittal plane alignment of young children with scoliosis. With the exception of increased pelvic tilt and decreased sacral slope, sagittal-plane spinopelvic parameters were similar between children with scoliosis and those reported for age-matched children without spinal deformity. These findings are important because they establish a baseline for determining the natural history and prognosis for children with scoliosis undergoing treatment with growing systems. Historically, prognostication for this population was based on the natural history of children without spinal deformity. This turned out to be a valid proposition with the exception of pelvic tilt and sacral slope, which were significantly different in children with spinal deformity. The role of pelvic tilt is to keep the sacral plate posterior to the axis of the hips, to establish the center of gravity over the lower extremities and maintain standing balance [7,8]. This study found that the pelvic tilt of children with spinal deformity was significantly greater than for normal children. Because the scoliosis group had a mean positive sagittal balance of 2.2 cm, pelvic tilt may have increased secondarily in an effort to maintain the center of gravity from displacing too far anterior. The pelvic incidence was not significantly different between the scoliosis (mean, 46.5 15.8 ) and normal groups (mean, 44.6 10.6 ) (p 5 .46). Because pelvic incidence is the summation of pelvic tilt and sacral slope, it would be expected that because there was no intergroup difference in pelvic incidence, the decreased sacral slope (35.5 12.1 vs. 40.3 ; p ! .05) was a compensatory mechanism to counterbalance the increased pelvic tilt in the scoliosis group. These alterations did not significantly affect lumbar lordosis or thoracic kyphosis. Compared with age-matched controls, the findings of increased pelvic tilt and decreased sacral slope for the entire study group (p ! .05) were also found for the congenital, idiopathic, and neuromuscular diagnoses;
R. El-Hawary et al. / Spine Deformity 1 (2013) 343e347
however, they were significant only for the congenital diagnosis. This is likely related to the small sample sizes for the other diagnoses. This finding was not observed for the syndromic diagnosis and was likely related to more heterogeneity within the syndromic etiology. Because it was documented for children without spinal deformity that sagittal plane alignment changes with age, the authors compared 2 age groups within the study population (1e5 vs. 6e10 years) [4,7,9,10]. These 2 age groups were chosen based on the age-related changes in spinal growth velocity. The findings of higher pelvic incidence, higher pelvic radius angle, and higher sacral slope in the older age group were similar to those found in the literature for children without scoliosis [4,7,10]. In addition, the current study’s trends toward higher thoracic kyphosis, higher lumbar lordosis, and less positive sagittal balance in the 6- to 10-year-old group agree with what Cil et al. [10] found for patients without scoliosis. The long-term effects of positive sagittal balance and increased pelvic retroversion were defined by Lafage et al. [6] for adult patients with spinal deformity. That group determined that increased pelvic retroversion resulted in poor health-related quality of life outcomes, including an increased Oswestry Disability Index and a decreased Short Forme12 Physical Component score. Thus, the long-term natural history of increased retroversion in children with scoliosis may be increased pain and disability. As a result, one of the goals of treatment for young children with scoliosis should be to improve their sagittal alignment, specifically pelvic tilt. Awareness of the nuances and compensatory mechanisms that scoliosis patients make to maintain standing balance will invariably affect the end point expectations of treatment with growing systems. It may also predict implant failure and other complications such as junctional kyphosis. Pelvic retroversion (ie, increased pelvic tilt) has been linked to spondylolisthesis in adolescents and young adults [9]. Given that the scoliosis group in this study demonstrated significant baseline pelvic retroversion, this factor could help predict construct failure and junctional kyphosis in a subset treated with growing systems. For distraction-based surgeries, it is possible that further positive sagittal imbalance may result from periodic posterior lengthening procedures. The second goal of this study was to determine the intrarater and inter-rater reliability of the different spinopelvic measures. Interestingly, sacral slope was the most reliable measure when a single observer performed repeated measures, but the least reliable between observers. Intraobserver variability was lowest for the measurements of pelvic incidence, because only a moderate correlation was noted with paired-sample correlations. These differences
347
may be related to difficulties in identifying the landmarks on the immature sacrum that are needed to measure pelvic incidence (Fig. 2). At the authors’ institution, a study is currently being performed to compare the variability of pelvic incidence measured on radiographs versus pelvic incidence measured on magnetic resonance imaging, where the cartilaginous sacral end plate can be better visualized. The pelvic radius angle, which is used by some investigators as a surrogate for pelvic incidence, performed well with both intra-observer and interobserver variability. This measure is not as reliant on the sacral end plate anatomy and is potentially more useful than pelvic incidence for this immature population. In conclusion, baselines values for spinopelvic parameters have been established for young children with scoliosis. These parameters were compared with those of children without spinal deformity and were the same except for a significantly greater pelvic tilt and a significantly lower sacral slope in children with scoliosis. The intra-rater and inter-rater reliability of these measures was fairly strong, with the exception of poor inter-rater reliability for the measurement of sacral slope. These values can be used as a baseline for both the natural history and for assessment of radiographic outcomes after surgical intervention. References [1] Jackson RP, Hales C. Congruent spinopelvic alignment on standing lateral radiographs of adult volunteers. Spine (Phila Pa 1976) 2000;25:2808e15. [2] Legaye J, Hecquet J, Marty C, et al. Sagittal equilibration of the spine: relationship between pelvis and sagittal spinal curves in the standing position [in French]. Rachis 1993;5:215e26. [3] Legaye J, Duval-Beaupere G, Hecquet J, et al. Pelvic incidence: a fundamental pelvic parameter for three-dimensional regulation of spinal sagittal curves. Eur Spine J 1998;7:99e103. [4] Marty C, Boisaubert B, Descamps H, et al. The sagittal anatomy of the sacrum among young adults, infants, and spondylolisthesis patients. Eur Spine J 2002;11:119e25. [5] Rajnics P, Pomero V, Templier A, et al. Computer-assisted assessment of spinal sagittal plane radiographs. J Spinal Disord 2001;14: 135e42. [6] Lafage V, Schwab F, Patel A, et al. Pelvic tilt and truncal inclination: two key radiographic parameters in the setting of adults with spinal deformity. Spine (Phila Pa 1976) 2009;34:E599e606. [7] Mac-Thiong JM, Berthonnaud IE, Dimar JR, et al. Sagittal alignment of the spine and pelvis during growth. Spine (Phila Pa 1976) 2004;29:1642e7. [8] Vaz G, Roussouly P, Berthonnaud E, Dimnet J. Sagittal morphology and equilibrium of pelvis and spine. Eur Spine J 2002;11:80e7. [9] Labelle H, Roussouly P, Berthonnaud E, et al. Spondylolisthesis, pelvic incidence, and spinopelvic balance: a correlational study. Spine (Phila Pa 1976) 2004;18:2049e54. [10] Cil A, Yazici M, Uzumcugil A, et al. The evolution of sagittal segmental alignment of the spine during childhood. Spine (Phila Pa 1976) 2005;30:93e100.