Submuscular growing rods (GR) have been demonstrated

SPINE Volume 41, Number 9, pp 792 797 ß 2016 Wolters Kluwer Health, Inc. All rights reserved DEFORMITY Low Pelvic Incidence Is Associated With Proximal Junctional Kyphosis in Patients Treated With Growing Rods Christopher N. Carender, BS, William Z. Morris, MD, Connie Poe-Kochert, RN, CNP, George H. Thompson, MD, Jochen P. Son-Hing, MD, and Raymond W. Liu, MD Study Design. Retrospective review of a prospectively collected pediatric orthopedic spine database. Objective. To investigate whether pelvic incidence (PI) changes during growing rod treatment and to report the effects of PI, if any, on complications during treatment. Summary of Background Data. Growing rods have been demonstrated to correct spinal deformity in early onset scoliosis while allowing for spinal growth. There has been little investigation into the potential effects, if any, of abnormal PI on complications, especially proximal junctional kyphosis (PJK). Methods. We retrospectively reviewed clinical and surgical data from our prospectively collected pediatric orthopedic spine database. Our final cohort of 48 patients had at least one lateral radiograph throughout the course of treatment containing the femoral heads and sacral endplate, and a minimum follow-up of 2 years. Defined failures were identified prospectively. Radiographs were measured for PI and development of PJK. Results. Mean age at initial treatment was 6.9 years (range 2.8 10.8 yr), with 35 females and 13 males. The mean length of follow-up was 8.1 years (range 2.0 22.1 yr). No statistical change in PI was observed throughout this study (P ¼ 0.655). Development of any failure as well as total number of failures was associated with younger age at initial treatment (P < 0.0005 for both). Development of PJK was associated with younger age From the Division of Pediatric Orthopaedic Surgery, Case Western Reserve University, Rainbow Babies and Children s Hospital, Cleveland, OH. Acknowledgment date: September 10, 2015. First revision date: October 15, 2015. Acceptance date: October 27, 2015. The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. Relevant financial activities outside the submitted work: board membership, consultancy, royalties, employment, travel/accommodations/meeting expenses. Address correspondence and reprint requests to Raymond W. Liu, MD, Division of Pediatric Orthopaedic Surgery, Case Western Reserve University, Rainbow Babies and Children s Hospital, 11100 Euclid Avenue, RBC 6081, Cleveland, OH 44106; E-mail: raymond.liu@uhhospitals.org DOI: 10.1097/BRS.0000000000001352 at initial treatment (P ¼ 0.030), female sex (P ¼ 0.002), and lower mean PI (P ¼ 0.042). Conclusion. PI remains constant throughout growth and the course of treatment with growing rods. Low PI was associated with increased PJK. When using growing rods in early onset scoliosis patients with decreased PI, increased attention should be paid to sagittal plane balance in an attempt to avoid PJK. Key words: complication, early onset scoliosis, failure, growing rods, pelvic incidence, proximal junctional kyphosis. Level of Evidence: 4 Spine 2016;41:792 797 Submuscular growing rods (GR) have been demonstrated to correct spinal deformity in early onset scoliosis (EOS) while allowing for spinal growth. 1 5 Treatment with GR has a relatively high rate of complication 1 5 that has been attributed to a variety of proposed risk factors, including younger age at initial treatment, 6 increased number of lengthening procedures 6,7 and increased thoracic kyphosis. 7,8 Pelvic incidence (PI) was defined by Legaye et al 9 as the angle between the perpendicular to the sacral plate at its midpoint and the line connecting this point to the middle axis of the femoral heads. Additionally, Legaye et al 9 defined PI as the sum of sacral slope (SS) and pelvic tilt (PT), such that PI ¼ SS þ PT. PI is thought to increase in value until roughly 10 years of age, and remain constant thereafter. 10 13 PI is also described as position independent, and has been shown to play an important role in the maintenance of sagittal alignment. 9,14,15 The presence of an abnormal PI is a risk factor for sagittal imbalance and failure following adult scoliosis fusion surgery, 16 18 which may lead to increased severity of symptoms 19 and decreased quality of life. 20 To our knowledge, there has been minimal investigation into the potential effects of abnormal PI on GR failure. A recent study by Shah et al 21 found that pelvic parameters (PI, PT, SS) remained unchanged in an ambulatory cohort of patients with EOS treated with GR. These findings are 792 www.spinejournal.com May 2016

consistent with current opinion that PI is a constant parameter. 10 13 However, the study by Shah et al 21 did not examine spinopelvic parameters in relation to GR failure. We retrospectively studied ambulatory and nonambulatory patients to further investigate whether PI can vary during GR treatment, and if it is associated with complications in children with GR. MATERIALS AND METHODS We reviewed 75 consecutive patients with EOS who underwent GR treatment and had at least 2 years of follow-up from an institutional review board-approved prospective database. The main inclusion criteria were patients who had at least one lateral radiograph throughout the course of treatment containing the femoral heads and sacral plate. Twenty-seven patients who had no measurable lateral radiographs were excluded from the study, leaving a final cohort of 48 patients. Clinical data collected included age, sex, type of scoliosis (C-EOS classification: idiopathic, neuromuscular, congenital, syndromic), 22 and duration of follow-up. Surgical data collected included the date of initial implant placement, the total number of surgeries, the date of initial implant failure, the number of implant failures, and the number of implant lengthenings until failure. Failures were defined as rod fracture, pedicle screw pullout, hook dislodgement, and hook erosion. Patients were also evaluated for the development of proximal junctional kyphosis (PJK). Radiographic Parameters PI and proximal junctional angle were measured on 25 preoperative, 40 first upright, and 48 final postoperative radiographs which included visible hip joints. PI was measured according to methods outlined by Duval-Beaupère et al 15 and Vaz et al. 23 The proximal junctional angle was measured from the caudal endplate of the uppermost instrumented vertebrae to the cephalic endplate two vertebrae above in a method described by Glattes et al. 24 PJK was defined as a 108 or greater increase in the proximal junctional angle on most recent postoperative images compared to preoperative images. 25 Measurements were made by three separate observers. standardized residual were inspected for normality, scatterplots of the standardized residuals were inspected for homoscedasticity, and the lack of any undue influence from outliers was confirmed with a Cook distance <1. RESULTS The mean age at initial treatment was 6.9 years (range 2.8 10.8 yr), with 35 females and 13 males. Diagnoses were idiopathic scoliosis in 11 patients, neuromuscular scoliosis in 13 patients, congenital scoliosis in four patients, and syndromic scoliosis in 20 patients. The mean length of follow-up was 8.1 years (range 2.0 22.1 yr). Twenty-one patients had measurable preoperative, first upright, and most recent postoperative lateral radiographs. Mean PI was 498 188 on preoperative radiographs, 478 148 on first upright postoperative radiographs, and 508 188 on most recent postoperative radiographs (Figure 1). No statistical change in PI was observed through the study duration using repeat measures analysis of variance analysis (P ¼ 0.655). Based on this, we chose to average all available PI values of each patient for the remainder of the analysis, with an overall PI of 518 158 for the total cohort of 48 patients. Over the course of treatment, 26 patients experienced implant failure, defined as rod fracture, pedicle screw pullout, hook dislodgement, and hook erosion. Failure was associated only with younger age at initial treatment (P < 0.0005). Female sex, etiology, ambulation, mean follow-up time, and mean PI were not found to have any association (Table 1). The total number of failures was 52, with a mean of 2.0 1.5 failures (range 1 8 failures) in 26 patients. Eleven of 26 patients experienced only one failure. An increased number of failures for each patient was also associated with younger age at initial treatment (P < 0.0005). Female sex, neuromuscular etiology, congenital etiology, syndromic etiology, ambulation, mean follow-up time, and mean PI were not found to have any association with total number of failures (Table 2). Preoperative radiographs of 21 patients with measurable preoperative and postoperative lateral radiographs had a Statistical Analysis All statistics were performed using SPSS Statistics Version 22 (IBM, Armonk, New York). Intraclass correlation coefficients (ICCs) were used to determine interobserver reliability for measurements. Repeated measures analysis of variance was used to compare preoperative, first upright, and most recent postoperative PI. For our multiple regression analysis, independent variables were age, sex, etiology, duration of follow-up, and average PI. Etiology was analyzed using standard dummy coding. Analysis was repeated using presence of any failure, total number of failures, and PJK as the dependent variables. In the multiple regression analysis, multicollinearity was assessed as negative based on variance inflation factor <10 and coefficient tolerance >0.1, normal probability plots of the regression Figure 1. Box plot of PI in preoperative, first upright, and most recent postoperative radiographs. Outliers, 25th percentile, median, and 75th percentiles are shown. PI indicates pelvic incidence. Spine www.spinejournal.com 793

TABLE 1. Factors Associated With Any GR Failure in Patients With EOS 95% CI for b Variable Standardized b Unstandardized b Lower Upper P Age 0.549 0.112 0.171 0.053 <0.0005 Female sex 0.058 0.065 0.366 0.236 0.663 Neuromuscular etiology 0.038 0.042 0.380 0.464 0.840 Congenital etiology 0.001 0.002 0.547 0.551 0.994 Syndromic etiology 0.109 0.110 0.241 0.462 0.530 Ambulatory 0.056 0.059 0.405 0.286 0.730 Follow-up time 0.30 0.000 0.003 0.002 0.844 Mean pelvic incidence 0.113 0.004 0.013 0.006 0.445 Idiopathic etiology was the default in the dummy coding for etiology. Standardized A values represent the contribution of each independent variable in predicting the variance in implant failure, while accounting for the contributions of the remaining variables. Value statistically significant. CI indicates confidence interval; EOS, early onset scoliosis; GR, growing rods. mean proximal junctional angle of 18 68, which increased to 48 78 on first upright postoperative radiographs after index surgery and 58 68 on most recent postoperative radiographs. Eight of the 21 patients fulfilled the criteria for PJK (Figure 2). Development of PJK was associated with younger age at initial treatment (P ¼ 0.030), female sex (P ¼ 0.002), and lower mean PI (P ¼ 0.042). Neuromuscular etiology, congenital etiology, syndromic etiology, ambulation, and mean follow-up time were not found to have any association with the development of PJK (Table 3). ICCs were recorded during the measurement of PI (n ¼ 20, ICC ¼ 0.91) and proximal junctional angle (n ¼ 19, ICC ¼ 0.82) with excellent agreement for both. DISCUSSION Complications occur in a fairly high proportion of children treated with GR. 1 5 Overall, the theory that younger age at initial treatment may be associated with the development of GR failure is inconsistently supported in the current literature. 2,7,26 Bess et al 7 identified younger patients as having an increased risk for the development of rod failures, including hook and screw pullout, and rod breakage. However, Akbarnia et al 2 and Yang et al 26 found no such association. This study found strong associations between younger age at initial treatment with both the occurrence of any GR failure and with an increased total number of failures over the course of treatment. At the time of its proposal, PI was described as an anatomical parameter whose value in adults is both unchanging and unaffected by changes in the values of other spinopelvic parameters. 9 Mangione et al 13 was the first to observe an increase in PI from infancy to adulthood via a comparison of 30 infant PI measurements (mean age 26.7 weeks), 30 different child PI measurements (mean age 3.2 yr), and 30 different adult PI measurements (mean age 28.6 yr). They demonstrated an increase in PI, with a mean difference of 24.88 between the infant and adult PI measurements, and a mean difference of 15.98 between the child and adult PI measurements, with the majority of the change occurring by 10 years TABLE 2. Factors Associated With the Total Number of GR Failures in Patients With EOS 95% CI for b Variable Standardized b Unstandardized b Lower Upper P Age 0.610 0.368 0.538 0.198 <0.0005 Female sex 0.057 0.189 0.679 1.058 0.662 Neuromuscular etiology 0.141 0.466 1.684 0.752 0.443 Congenital etiology 0.264 1.404 2.988 0.180 0.081 Syndromic etiology 0.175 0.522 1.537 0.492 0.304 Ambulatory 0.044 0.137 1.134 0.859 0.782 Follow-up time 0.007 0.000 0.008 0.007 0.959 Mean pelvic incidence 0.051 0.005 0.033 0.023 0.723 Idiopathic etiology was the default in the dummy coding for etiology. Standardized A values represent the contribution of each independent variable in predicting the variance in total number of implant failures while accounting for the contributions of the remaining variables. Value statistically significant. CI indicates confidence interval; EOS, early onset scoliosis; GR, growing rods. 794 www.spinejournal.com May 2016

Figure 2. Development of PJK and hook dislodgement in a patient diagnosed with syndromic EOS. A, Preoperative radiograph, diagnosis, syndromic scoliosis; sex, female; age, 2 years; PI ¼ 368; PJA ¼ 98. B, First upright postoperative radiograph; PI ¼ 428; PJA ¼ 18. C, Most recent postoperative radiograph, follow-up time: 28 months; PI ¼ 348; PJA ¼ 98. EOS indicates early onset scoliosis; PI, pelvic incidence; PJA, proximal junctional angle; PJK, proximal junctional kyphosis. TABLE 3. Factors Associated With PJK in Patients Treated With GR for EOS 95% CI for b Variable Standardized b Unstandardized b Lower Upper P Age 0.438 0.087 0.165 0.10 0.030 Female sex 0.635 0.697 0.302 1.091 0.002 Neuromuscular etiology 0.051 0.056 0.609 0.498 0.832 Congenital etiology 0.087 0.153 0.873 0.567 0.653 Syndromic etiology 0.210 0.208 0.669 0.253 0.348 Ambulatory 0.064 0.067 0.520 0.386 0.756 Follow-up time 0.206 0.002 0.005 0.002 0.292 Mean pelvic incidence 0.414 0.013 0.026 0.001 0.042 Idiopathic etiology was the default in the dummy coding for etiology. Standardized A values represent the contribution of each independent variable in predicting the variance in PJK, while accounting for the contributions of the remaining variables. Value statistically significant. CI indicates confidence interval; EOS, early onset scoliosis; GR, growing rods. Spine www.spinejournal.com 795

of age. In contrast, Marty et al 27 observed a smaller mean difference of 7.98 between 32 infant PI measurements and 44 different healthy adult PI measurements (mean age 24.0 yr), suggesting mild changes with development. Shah et al 21 recently demonstrated PI to be relatively unchanging in a cohort of ambulatory patients with a mean preoperative age of 5.6 years who were treated with GR. Our study examined both ambulatory and nonambulatory patients, and also found no statistical change in PI throughout the course of treatment (Figure 3). The mean age of 6.9 years at initial treatment in our cohort suggests that our patients had likely passed through most of the early developmental changes proposed by Mangione et al. 13 Overall, this suggests that PI is not affected by the presumed large amount of stress through the spine and pelvis which can occur with GR. Pelvic parameters and their relationship to PJK have also been studied in the context of adult spinal deformity 28 and Scheuermann kyphosis. 29 Maruo et al 28 identified PI greater than 558 as being associated with development of PJK in a cohort of adult spinal deformity patients treated with spinal fusion surgery. Similarly, Lonner et al 29 identified a greater degree of PI as being associated with an increased degree of PJK in a cohort of adolescent and adult patients with Scheuermann kyphosis treated with spinal implant systems. Higher PI has also been shown to be a risk factor for the development of sagittal imbalance 17 and failure of lumbopelvic fixation 16 following spinal fusion surgery in adults. To our knowledge, this study is the first to identify low PI as being associated with development of postoperative spinal pathology. However, there have been studies demonstrating an association between low PI and hip pathology, such as slipped capital femoral epiphysis 30 and femoroacetabular impingement (cam and pincer lesions). 31 Yoshimoto et al 32 demonstrated that higher PI may predispose patients to the development of osteoarthritis of the hip. As such, the paradigm of both high and low PI being associated with development of hip pathology may be extrapolated to that of the spine. GR has been shown to be an effective means of improving sagittal balance in patients with EOS. 5,21 However, PJK is a known complication of GR surgery in EOS patients. 33 A child with a lower PI would be expected to have decreased lumbar lordosis 9,12,23 and decreased thoracic kyphosis. 23 It is possible that our standard surgical GR technique did not factor in this sagittal relationship, leading to increased PJK in these patients. The delicate balance of the sagittal plane in scoliosis surgery in general has been well described. 16 19 In the circumstance of low PI, if the surgical construct introduces more lumbar lordosis than the patient s PI can accommodate, then the spine would theoretically need to compensate with additional thoracic kyphosis, achieved through PJK. When we re-evaluated our cohort, 16 patients had preoperative and most recent postoperative radiographs with a measurable proximal junctional angle and lumbar lordosis (L1-L5). Five of six patients with PJK had increased lumbar lordosis, versus six of 10 patients without PJK. Although the numbers were too low for statistical analysis, both the tendency to increase lumbar lordosis and its potential impact on PJK merit attention in future studies. This study was limited in that it was a retrospective review of a prospectively collected database. The relatively small number of patients within the study is a consequence of the rarity of the conditions necessitating treatment with Figure 3. Unchanging PI in a patient diagnosed with idiopathic EOS over the course of treatment with GR. A, Preoperative radiograph, diagnosis, idiopathic scoliosis; sex, female; age, 3 years; PI ¼ 438. B, First upright postoperative radiograph; PI ¼ 458. C, Most recent postoperative radiograph, follow-up time: 149 months; PI ¼ 468. EOS indicates early onset scoliosis; GR, growing rods; PI, pelvic incidence. 796 www.spinejournal.com May 2016

GR. All patients were treated by two surgeons working within the same practice, which minimized variation in operative technique between patients. PI can be difficult to measure, particularly in sitting patients where the lateral view of the pelvis can be maloriented. We used the methodology described by Vaz et al 23 to account for offset between the two femoral heads on the lateral view, and our interobserver reliability values for all measurements fell in the excellent category. In summary, this study demonstrates that PI remains constant in ambulatory and nonambulatory patients with EOS during growth and throughout the course of treatment with GR. These findings are consistent with previous studies of PI in ambulatory patients; to our knowledge, this study is the first to demonstrate PI as a constant parameter in nonambulatory patients. Younger age, female sex, and lower PI were associated with increased PJK. When using GR in EOS patients with low PI, increased attention should be paid to sagittal plane balance to avoid PJK. Key Points PI remained constant throughout growth and the course of treatment with GR in a cohort of patients with EOS. Low PI was associated with increased PJK. Attention should be paid to sagittal plane balance when using GR in EOS patients with low PI. References 1. Thompson GH, Akbarnia BA, Kostial P, et al. Comparison of single and dual growing rod techniques followed through definitive surgery: a preliminary study. Spine 2005;30:2039 44. 2. Akbarnia BA, Marks DS, Boachie-adjei O, et al. Dual growing rod technique for the treatment of progressive early-onset scoliosis: a multicenter study. Spine 2005;30 (17 suppl):s46 57. 3. Wang S, Zhang J, Qiu G, et al. Dual growing rods technique for congenital scoliosis: more than 2 years outcomes: preliminary results of a single center. Spine 2012;37:E1639 44. 4. Li QY, Zhang JG, Qiu GX, et al. Primary effect of dual growing rod technique for the treatment of severe scoliosis in young children. Chin Med J 2010;123:151 5. 5. Akbarnia BA, Breakwell LM, Marks DS, et al. Dual growing rod technique followed for three to eleven years until final fusion: the effect of frequency of lengthening. Spine 2008;33:984 90. 6. Bess S, Akbarnia BA, Thompson GH, et al. Complications of growing-rod treatment for early-onset scoliosis: analysis of one hundred and forty patients. J Bone Joint Surg Am 2010;92:2533 43. 7. Watanabe K, Uno K, Suzuki T, et al. Risk factors for complications associated with growing-rod surgery for early-onset scoliosis. Spine 2013;38:E464 8. 8. Schroerlucke SR, Akbarnia BA, Pawelek JB, et al. How does thoracic kyphosis affect patient outcomes in growing rod surgery?. Spine 2012;37:1303 9. 9. Legaye J, Duval-Beaupère G, Hecquet J, et al. Pelvic incidence: a fundamental pelvic parameter for three-dimensional regulation of spinal sagittal curves. Eur Spine J 1998;7:99 103. 10. Mehta VA, Amin A, Omeis I, et al. Implications of spinopelvic alignment for the spine surgeon. Neurosurgery 2012;70:707 21. 11. Vrtovec T, Janssen MM, Likar B, et al. A review of methods for evaluating the quantitative parameters of sagittal pelvic alignment. Spine J 2012;12:433 46. 12. Schwab F, Lafage V, Patel A, et al. Sagittal plane considerations and the pelvis in the adult patient. Spine 2009;34:1828 33. 13. Mangione P, Gomez D, Senegas J. Study of the course of the incidence angle during growth. Eur Spine J 1997;6:163 7. 14. Boulay C, Tardieu C, Hecquet J, et al. Sagittal alignment of spine and pelvis regulated by pelvic incidence: standard values and prediction of lordosis. Eur Spine J 2006;15:415 22. 15. Duval-Beaupère G, Schmidt C, Cosson P. A Barycentremetric study of the sagittal shape of spine and pelvis: the conditions required for an economic standing position. Ann Biomed Eng 1992;20:451 62. 16. Cho W, Mason JR, Smith JS, et al. Failure of lumbopelvic fixation after long construct fusions in patients with adult spinal deformity: clinical and radiographic risk factors: clinical article. J Neurosurg Spine 2013;19:445 53. 17. Cho KJ, Suk SI, Park SR, et al. Risk factors of sagittal decompensation after long posterior instrumentation and fusion for degenerative lumbar scoliosis. Spine 2010;35:1595 601. 18. Le huec JC, Faundez A, Dominguez D, et al. Evidence showing the relationship between sagittal balance and clinical outcomes in surgical treatment of degenerative spinal diseases: a literature review. Int Orthop 2015;39:87 95. 19. Glassman SD, Bridwell K, Dimar JR, et al. The impact of positive sagittal balance in adult spinal deformity. Spine 2005;30:2024 9. 20. Theis J, Gerdhem P, Abbott A. Quality of life outcomes in surgically treated adult scoliosis patients: a systematic review. Eur Spine J 2015;24:1343 55. 21. Shah SA, Karatas AF, Dhawale AA, et al. The effect of serial growing rod lengthening on the sagittal profile and pelvic parameters in early-onset scoliosis. Spine 2014;39:E1311 7. 22. Williams BA, Matsumoto H, Mccalla DJ, et al. Development and initial validation of the Classification of Early-Onset Scoliosis (C-EOS). J Bone Joint Surg Am 2014;96:1359 67. 23. Vaz G, Roussouly P, Berthonnaud E, et al. Sagittal morphology and equilibrium of pelvis and spine. Eur Spine J 2002;11:80 7. 24. Glattes RC, Bridwell KH, Lenke LG, et al. Proximal junctional kyphosis in adult spinal deformity following long instrumented posterior spinal fusion: incidence, outcomes, and risk factor analysis. Spine 2005;30:1643 9. 25. Denis F, Sun EC, Winter RB. Incidence and risk factors for proximal and distal junctional kyphosis following surgical treatment for Scheuermann kyphosis: minimum five-year follow-up. Spine 2009;34:E729 34. 26. Yang JS, Sponseller PD, Thompson GH, et al. Growing rod fractures: risk factors and opportunities for prevention. Spine 2011;36:1639 44. 27. 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;11:119 25. 28. Maruo K, Ha Y, Inoue S, et al. Predictive factors for proximal junctional kyphosis in long fusions to the sacrum in adult spinal deformity. Spine 2013;38:E1469 76. 29. Lonner BS, Newton P, Betz R, et al. Operative management of Scheuermann s kyphosis in 78 patients: radiographic outcomes, complications, and technique. Spine 2007;32:2644 52. 30. Gebhart JJ, Bohl MS, Weinberg DS, et al. Pelvic incidence and acetabular version in slipped capital femoral epiphysis. J Pediatr Orthop 2014;35:565 70. 31. Gebhart JJ, Streit JJ, Bedi A, et al. Correlation of pelvic incidence with cam and pincer lesions. Am J Sports Med 2014;42:2649 53. 32. Yoshimoto H, Sato S, Masuda T, et al. Spinopelvic alignment in patients with osteoarthrosis of the hip: a radiographic comparison to patients with low back pain. Spine 2005;30:1650 7. 33. Watanabe K, Uno K, Suzuki T, et al. Risk factors for proximal junctional kyphosis associated with dual-rod growing-rod surgery for early-onset scoliosis. J Spinal Disord Tech 2014; Epub ahead of print. Spine www.spinejournal.com 797