Early-onset scoliosis (EOS) refers to scoliosis that presents in children aged below 5 years, representing a heterogeneous group with a variety of etiologies.1 In general, the goals of treatment include the control of spinal deformity, preservation of spinal growth, and encouragement of normal development of the lungs and chest wall.2 Accordingly, various growth-friendly spine implants have emerged and evolved into an effective and safe treatment option for EOS. On the basis of repeated distractive forces exerted on the spine, single or dual growing rods (GRs) and vertical expandable prosthetic titanium ribs are classified as distraction-based implants by Skaggs et al.3 Of these, GR is the most commonly reported technique in the literature.
It is well known that GRs only provide limited fusion at the proximal and distal foundation sites, whereas the rods span the curve without fusion. In EOS with increased kyphosis, the middle part of the rod may be under high compressive loading, and intermittent distractions may lead to cyclical fatigue of growth implants, which therefore increases the risk of rod fracture. Schroerlucke et al4 found that EOS patients with hyperkyphosis during the GR treatment presented more challenges with respect to complications, especially implant-related complications. Moreover, stress may be transmitted to the proximal and distal foundation sites, increasing the loading on the junctional area, which leads to junctional problems such as fixation anchor failure and proximal junctional kyphosis (PJK). El-Hawary et al5 study on EOS treated with distraction-based implants (rib based in 24 cases and spine based in 16 cases) demonstrated that patients with preoperative hyperkyphosis (≥50 degrees) were 2.8 times more likely to experience postoperative PJK than those with kyphosis <50 degrees. Eventually, the more often that complications occurred, the more unplanned surgeries would be needed. However, there is a paucity of data regarding the effect of GRs on the sagittal spinal profiles in hyperkyphotic EOS. Therefore, it is essential to investigate how hyperkyphotic EOS responds to GRs and to determine the complication rate during the treatment.
After approval of Institutional Review Board, a retrospective review of a single database of patients with EOS was performed. The inclusion criteria were patients who (1) were treated with GRs, (2) had at least 2 lengthening procedures, and (3) had a follow-up duration of >2 years. Exclusion criteria were (1) a prior history of spine surgery and (2) halogravity traction before index surgery. Finally, 40 patients who met these criteria were enrolled in this study, consisting of 11 boys and 29 girls. None of these patients had preoperative thoracic kyphosis (TK)<20 degrees. With reference to the definition of normal thoracic kyphosis in children,6 patients were categorized into 2 groups on the basis of preoperative T2-12 thoracic kyphosis: 20 to 50 degrees [normal (N) group] and >50 degrees [hyperkyphotic (K) group]. Patient demographic data were collected as follows: diagnosis, age at the index surgery, type of GR, location of proximal and distal foundations, treatment time, and the number and intervals of lengthening procedures. The following sagittal spinal parameters were measured on the lateral radiographs before and after the index surgery and at the last follow-up: major coronal Cobb angle, T1-S1 spinal height, T2-12 thoracic kyphosis, L1-S1 lumbar lordosis, and proximal junctional angle (PJA). PJA was defined as the angle between the caudal endplate of the upper instrumentation vertebra and the cephalad endplate of the second supradjacent vertebrae above the upper instrumentation vertebra.7
Complications were recorded for the patients in each group. Implant-related complications included rod fracture and failure of foundation fixation, such as hook or screw pullout. In terms of alignment-related complication, we particularly focused on PJK, which is defined by the following 2 criteria7: (1) PJA≥10 degrees and (2) PJA at least 10 degrees greater than the preoperative measurement. The presence of both criteria was necessary to be considered as PJK. The complications were also evaluated using the surgical complications grading system proposed by Smith et al,8 where severity refers to the level of care and urgency required to treat the complication and can be classified as follows: (1) grade I does not require unplanned surgery and can be corrected at the next scheduled surgery; (2) grade II requires an unplanned surgery, with IIA requiring a single trip and IIB needing multiple trips for resolution; (3) grade III alters the planned course of treatment; and (4) grade IV is defined as death.
Statistical analyses were performed using SPSS version 20.0 software (SPSS Inc., Chicago, IL). Comparisons between variables of interest were made using the independent samples t test. Fisher exact tests were performed to compare the incidence of complications between the N and K groups. The significance was defined as P<0.05.
There were 19 and 21 patients in the N and K groups, respectively. The diagnosis and instrumentation data are summarized in Table 1. A majority of patients in each group had a diagnosis of congenital scoliosis [N group: n=11 (57.9%); K group: n=12 (57.1%)], all of which were mixed type anomalies with long curves. The overall distribution of diagnoses was not significantly different between groups (P=0.993). The mean age before surgery in the N and K groups was 6.2±1.8 (range, 4.2 to 8.6) and 6.4±2.1 (range, 4.3 to 9.2) years, respectively. The N and K groups had an average number of lengthenings of 4.2±2.0 (range, 2 to 7) and 4.3±2.3 (range, 2 to 7) with an average length of follow-up of 4.7±1.9 (range, 2.3 to 6.8) and 4.6±2.1 (range, 2.4 to 7.2) years, respectively. The mean lengthening intervals were 11.3±2.3 (range, 9 to 12) and 10.9±1.9 (range, 8 to 11) months in the N and K groups, respectively. No statistically significant differences were observed in the above data between these 2 groups. There were 226 procedures in total, including 40 implantations, 170 lengthenings, and 16 revisions.
Correction in the Coronal Plane
The major Cobb angle in both groups decreased remarkably after index surgery and slightly increased over time with subsequent distractions (Table 2). The T1-S1 spinal height in both groups increased significantly. During the treatment, the T1-S1 height gain of 4.0±1.8 cm was achieved after 4.2±2.0 lengthenings in the N group, whereas 4.0±2.1 cm for 4.3±2.3 lengthenings in the K group. The T1-S1 growth gain per lengthening was 0.99±0.17 and 0.97±0.19 cm in the N and K group, respectively.
Restoration in the Sagittal Plane
In the N group, TK changed from 35.7±9.6 degrees preoperatively to 33.2±8.8 degrees at the last follow-up, indicating insignificant changes (P=0.408). In contrast, TK in the K group notably decreased from 66.4±10.4 to 43.7±9.8 degrees (P<0.001) after the index surgery and slightly increased to 47.7±9.2 degrees (P=0.180) over time with subsequent distractions (Fig. 1). Lumbar lordosis (LL) in the N group changed from 49.6±13.9 to 46.2±12.4 degrees after index surgery and to 44.1±13.0 degrees at the last follow-up, whereas in the K group, it significantly decreased from 53.4±12.6 to 43.5±17.9 degrees after index surgery and slightly increased to 44.3±14.0 degrees at the last follow-up, which was consistent with the change of TK.
Nine of 19 (47.4%) patients in the N group and 17 of 21 (81.0%) patients in the K group had at least 1 complication (Table 3), indicating that complications in the K group were significantly more frequent than in the N group (P=0.046). The most common implant-related and alignment-related complication was rod fracture (17.5%) and PJK (22.5%), respectively. The rod fracture rate in the N and K groups was 10.5% and 23.8%, respectively. Repeat fractures were observed in 3 patients, with 1 in the N group and 2 in the K group. The rate of PJK in the N and K groups was 15.8% and 28.6%, respectively. Although there were no statistically significant differences between the N and K groups for these 2 complications, the tendency showed that the patients with hyperkyphosis had higher rate of rod fracture and PJK than those with normal kyphosis. Moreover, for general complications, there was one superficial infection in each group. No neurological deficit, pneumonia or other complications occurred.
Among these complications, PJK, sagittal imbalance and superficial infection did not require unplanned surgery. Thus, 5 and 8 complication events in the N and K groups, respectively, were evaluated with grade I. Rod fracture and implant dislodgment needed a single unplanned surgery, such as replacement or enhancement of instrumentation. Thus, 5 and 11 complication events in the N and K groups, respectively, were classified as grade IIA. However, in both groups, there were no complication events reaching grade III that were too severe to change the planned course of treatment.
Single rod had a relatively higher fracture rate than dual rods (4 of 10, 40% vs. 3 of 30, 10%, P=0.052). Repeat fracture was more common in patients with single rod (2 of 10, 20% vs. 1 of 30, 3.3%, P=0.149). In addition, implant dislodgement was also more common in patients with single rod (5 of 10, 50% vs. 1 of 30, 3.3%, P=0.002). However, dual rods had a relatively higher PJK rate than single rod (8 of 30, 26.7% vs. 1 of 10, 10%, P=0.404).
According to the classification of EOS,9 etiologies are listed in prioritized order from the highest to the lowest: congenital or structural, neuromuscular, syndromic, and idiopathic. In this study, we included a large number of congenital scoliosis patients, all of whom were mixed type and accompanied with long curves. Importantly, brace was ineffective in controlling the deformities in this group of patients. Thus, congenital scoliosis patients with young age, long curve and failure of brace treatment should be candidates for GR techniques. Congenital scoliosis accounted for a large proportion of cases, which did affect the average spinal growth to some extent. Although <1.2 cm per year between the ages of 5 and 10, which was reported by Dimeglio,10 the data in our study were almost close to the normal value. Moreover, our data also followed the “law of diminishing returns” that the T1-S1 gain decreased significantly with repeated lengthenings.11 We are only maintaining the postoperative result with those repeat distractions.
Previous studies of GRs for EOS mainly focused on the coronal curves, more or less referred to the sagittal profiles, such as TK and LL changes.12,13 Akbarnia et al12 evaluated 23 children with EOS who underwent GR instrumentation. The mean TK was 50 degrees preoperatively, 35 degrees after initial lengthening, and 45 degrees at the last follow-up. LL was −45 degrees preoperatively, −42 degrees after initial lengthening, and −48 degrees at the last follow-up. The above studies demonstrated that TK decreased after index surgery and increased between the index surgery and the latest follow-up. However, as the patients in these studies were not stratified by the magnitude of TK, whether differences exist in the evolutionary tendency of TK between patients with preoperative hyper and normal kyphosis under GR treatment has not yet been determined.
Recently, in a multicenter retrospective review performed by Shah et al,14 patients were categorized into 3 groups in terms of preoperative T5-12 kyphosis (hypokyphotic group, <10 degrees; normal kyphotic group, 10 to 40 degrees; and hyperkyphotic group, >40 degrees). It was found that the patients with hypokyphosis improved, and the patients with normal kyphosis maintained their alignment. According to the references for sagittal parameters in the pediatric population,6 normal T2-T12 kyphosis is defined as a range of 20 to 50 degrees, and hyperkyphosis as >50 degrees. On the basis of the cut-off point of 50 degrees, we collected patients with preoperative TK between 20 and 50 degrees as the N group and those with preoperative TK>50 degrees as the K group. The demographic data were similar between 2 groups. In our study, TK in the K group was greater than that in the N group at all time points, but the change from the preoperative time point to the last follow-up was greater in the K group, which indicates that GRs did result in significant improvements to hyperkyphosis.
The present study identified the complication rate for the N and K groups as 47.4% and 81.0%, respectively, demonstrating that total complication events in the K group were significantly more frequent than in the N group (P=0.046). In this cohort, implant-related complications included rod fracture and implant dislodgment, among which rod fracture ranked most frequently with an overall incidence of 17.5%, and repeat fractures were observed in 3 patients. Yang et al15 reviewed 327 patients who underwent GR surgery from 17 centers. They found that 86 rod fractures occurred in 49 patients, with an incidence of ∼15%. A comparison of preoperative kyphosis between patients with and without fracture was not statistically significant (41 vs. 43 degrees, P=0.29), revealing that preoperative kyphosis was not a risk factor for GR fracture. Conversely, Schroerlucke et al4 reported that the hyperkyphosis group had 25 rod fractures in 14 patients compared with 13 in 9 patients in the normal group. Although there was no statistically significant difference, the survival curve analysis revealed that the incidence of rod fracture increased incrementally as TK increased and that the rod survival was 73% at a TK of 40 degrees. In our study, there was also no statistically significant difference (P=0.412) between the N and K groups in terms of rod facture rate, probably due to the limited sample size. Nonetheless, the incidence of rod fracture in the K group was >2 times that of the N group (23.8% vs. 10.5%), demonstrating that patients with hyperkyphosis may experience more rod fractures.
In this study, alignment-related complications included PJK and sagittal imbalance, among which PJK ranked most frequently with an overall incidence of 22.5%. We found that the K group had more PJK than the N group (28.6% vs. 15.8%), which implied that preoperative hyperkyphosis may be a potential risk for PJK. Watanabe et al16 reported the incidence of PJK was 26% (23/88) in patients with EOS who underwent dual-rod GR surgery and identified main thoracic kyphosis of 60 degrees or more as a significant independent risk factor for PJK. El-Hawary et al5 performed the radiographic comparisons between patients with and without PJK and found TK was higher in a PJK group (45.1 vs. 28.7 degrees, P<0.05).
According to the grading system by Smith et al,8 the complications in the K group were 8 with grade I and 11 with grade IIA. None were assigned to grade III, which would have been severe enough to change the planned course of treatment. Although the complication rate was higher in the K group, some complications, such as PJK, are considered to be “expected,” or even unavoidable complications during the course of treatment. As reported by Hart et al,17 PJK is a recurrent deformity with limited clinical impact, and it should be distinguished from proximal junctional failure, which is a significant complication following adult spinal deformity surgery with potential for neurological injury and increased need for surgical revision. In contrast, it is believed that the ultimate goal of GRs is to correct EOS while allowing for spinal growth and pulmonary development until near skeletal maturity. PJK is only a radiographic phenomenon, which is definitely not a cause to terminate the GR treatment. Therefore, the impact of PJK on the overall treatment plan is minimal, and in our center is often corrected by extending the fusion at the final fusion surgery.
In this study, rod fracture and implant dislodgement were more common in patients with single rod, while PJK was more frequently occurred in patients with dual rods. For single rod, the mechanical stress may be concentrated on the implants, easily leading to rod fracture15 or implant dislodgement. However, dual rods attached to solid foundations and dissipated the amount of mechanical stress to junctional region, which was prone to develop PJK.18 Therefore, dual rods are strongly recommended, especially in patients with hyperkyphosis, to reduce the implant complication rate.
The infection rate in our study was 5%. Such a low infection rate may be related to adequate soft tissue coverage of surgical site and submuscular placement of the implant. Also, the infection rate was related to the number of surgical procedures.19 The fewer procedures may lead to a lower infection rate. In addition, the infection rate was related to etiology.20 In our study, neuromuscular scoliosis accounted for the lowest, which may result in a low infection rate.
The main limitation of this study is its relatively smaller sample size. Another limitation is the length of follow-up which does not include full maturity. Thus, in the future, a long-term follow-up study should be conducted. In addition, EOS is a complex and heterogeneous condition of significant diversity with respect to etiology, which creates inherent drawbacks in generalized statements. Despite these limitations, all patients in this single-center study were treated by surgeons working with the same surgical decision making, which minimized the variations in operative strategies and techniques between patients and maximally guaranteed the consistency of surgical results.
In conclusion, the sagittal profile of EOS can be successfully reestablished and maintained within the normal range during the treatment with GRs. Although hyperkyphotic EOS tends to increase complications, such as rod fracture and PJK, the complications were not severe enough to change the planned course of treatment. Nevertheless, greater attention must be paid to patients with hyperkyphosis.
1. Gillingham BL, Fan RA, Akbarnia BA. Early onset idiopathic scoliosis. J Am Acad Orthop Surg. 2006;14:101–112.
2. Dimeglio A, Canavese F. The growing spine: how spinal deformities influence normal spine and thoracic cage growth. Eur Spine J. 2012;21:64–70.
3. Skaggs DL, Akbarnia BA, Flynn JM, et al. A classification of growth friendly spine implants. J Pediatr Orthop. 2014;34:260–274.
4. Schroerlucke SR, Akbarnia BA, Pawelek JB, et al. How does thoracic kyphosis affect patient outcomes in growing rod
surgery? Spine (Phila Pa 1976). 2012;37:1303–1309.
5. El-Hawary R, Sturm P, Cahill P, et al. What is the risk of developing proximal junctional kyphosis during growth friendly treatments for early-onset scoliosis
? J Pediatr Orthop. 2015. [Epub ahead of print].
6. Boseker EH, Moe JH, Winter RB, et al. Determination of “normal” thoracic kyphosis: A roentgenographic study of 121 “normal” children. J Pediatr Orthop. 2000;20:796–798.
7. 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 (Phila Pa 1976). 2005;30:1643–1649.
8. Smith JT, Johnston C, Skaggs D, et al. A new classification system to report complications in growing spine surgery: a multicenter consensus study. J Pediatr Orthop. 2015;35:798–803.
9. 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;96:1359–1367.
10. Dimeglio A. Growth in pediatric orthopaedics. J Pediatr Orthop. 2001;21:549–555.
11. Sankar WN, Skaggs DL, Yazici M, et al. Lengthening of dual growing rods and the law of diminishing returns. Spine (Phila Pa 1976). 2011;36:806–809.
12. 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 (Phila Pa 1976). 2005;30:S46–S57.
13. Thompson GH, Akbarnia BA, Kostial P, et al. Comparison of single and dual growing rod
techniques followed through definitive surgery: a preliminary study. Spine (Phila Pa 1976). 2005;30:2039–2044.
14. 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 (Phila Pa 1976). 2014;39:E1311–E1317.
15. Yang JS, Sponseller PD, Thompson GH, et al. Growing rod
fractures: Risk factors and opportunities for prevention. Spine (Phila Pa 1976). 2011;36:1639–1644.
16. 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
. Clin Spine Surg. 2016;29:E428–E433.
17. Hart RA, Mccarthy I, Ames CP, et al. Proximal junctional kyphosis and proximal junctional failure. Neurosurg Clin N Am. 2013;24:213–218.
18. Cho SK, Kim YJ, Lenke LG. Proximal junctional kyphosis following spinal deformity surgery in the pediatric patient. J Am Acad Orthop Surg. 2015;23:408–414.
19. Kabirian N, Akbarnia BA, Pawelek JB, et al. Deep surgical site infection following 2344 growing-rod procedures for early-onset scoliosis
: Risk factors and clinical consequences. J Bone Joint Surg Am. 2014;96:e128.
20. Cahill PJ, Warnick DE, Lee MJ, et al. Infection after spinal fusion for pediatric spinal deformity: Thirty years of experience at a single institution. Spine (Phila Pa 1976). 2010;35:1211–1217.