While the traditional focus of adolescent idiopathic scoliosis (AIS) treatment has been in the coronal plane, surgeons now recognize AIS as a more complex 3-dimensional deformity with important sagittal and axial plane deviations as well. Furthermore, recent advances in instrumentation techniques have enabled surgeons to be more aggressive in correcting this complex deformity.1–3 Popular now in the treatment of AIS is the use of a posterior-only approach to gain 3 column fixation using pedicle screw instrumentation segmentally. A relatively common phenomenon with long segment constructs is the development of proximal junctional kyphosis (PJK), defined as a kyphotic deformity in the spinal motion segments immediately cephalad to the instrumented segments. Several recent studies have reviewed the incidence of PJK in AIS. Kim et al4 noted an increased rate of PJK in patients treated with hybrid constructs (distal pedicle screws and proximal hooks) compared with all hook constructs. They hypothesized that increased PJK was secondary to the increased rigidity of the hybrid instrumentation. Lee et al5 found a 46% prevalence of proximal kyphosis at 2-year follow-up after hook and rod instrumentation in AIS. Additionally, in a comparison of anterior and posterior instrumentation techniques in AIS, Rhee et al6 noted an increase of 7° PJK with posterior instrumentation, and proposed that the difference was related to posterior tension band disruption, posterior compression, and decreased thoracic kyphosis. While the above authors have reviewed PJK in AIS, none have done so with all pedicle screw constructs. In 2 studies by Kim et al, no statistically significant difference was found in proximal junctional change between pedicle screws and hooks or pedicle screws and hybrid constructs.7–9 While they did not find a significant difference, a trend was noted in the change in kyphosis from preoperative to 2 years postoperative when comparing pedicle screws (9°) and hook (6°) instrumentation (P = 0.19).7 To date, the incidence of and risk factors for PJK in AIS have not been extensively studied.
While proximal junctional changes following long segmental instrumentation have been recognized recently from a radiographic standpoint, the clinical significance of this is unclear. Lee et al5 initially selected an increase of 5° as being abnormal kyphosis to account for measurement error between radiographs. Subsequent studies have selected 10° presumably based on the study by Carman et al10 on the reliability of radiographic measurements. They found that to be 95% confident, a measured difference had to be 10° for scoliosis and 11° for kyphosis. Vertebral body endplate landmarks for the measurement of kyphosis can be very obscure on the lateral film, especially in the proximal thoracic spine. Therefore, because of the lack of early clinical symptoms associated with proximal kyphosis and the arbitrary selection of a radiographic measurement, what remains “too kyphotic” or diagnosed as PJK is uncertain.
Materials and Methods
Following Institutional Review Board approval, a multicenter database was reviewed for all patients who underwent posterior spinal fusion for AIS and had at least 2 years follow-up. The patients included in this study were from 8 different centers, represented by 12 surgeons. Patients who had additional procedures (i.e., an anterior approach) or in whom the upper instrumented vertebra (UIV) was below T9 were excluded from this study. Furthermore, if we were unable to adequately visualize the vertebral body of the appropriate levels, the patient was excluded. Patients were divided into 4 groups based on method of instrumentation: group 1–hook only constructs (Figure 1), group 2–hybrid constructs, group 3–pedicle screw only constructs (Figure 2), and group 4–all pedicle screw constructs except hooks at the cephalad level (UIV) (Figure 3). Group 4 was created from 2 surgeons (P.O.N. and D.H.C.), who because of the recent interest in proximal kyphosis, attempted to dampen the effect by placing hooks at the upper instrumented vertebrae.
Proximal kyphosis was evaluated on lateral upright radiographs of the spine before surgery and at 2 years after surgery. The proximal junctional angle was measured from the caudal endplate of the UIV to the cephalad endplate of the vertebra adjacent to the UIV (Figure 4). Measurement of one motion segment was selected to isolate the area of interest. One potential cause of increased junctional kyphosis is the disruption of the posterior soft tissues to include the facet capsule at the adjacent level. Furthermore, one would expect less change in a motion segment which was not involved in the surgical dissection and would see less force from the large moment arm created by the instrumented fusion. Therefore, to increase the likelihood that we isolated the area of interest and to reduce the dampening effect of including additional segments, one motion segment was selected.
For evaluation of risk factors for PJK, all patients were divided into 2 groups, patients greater than 2 standard deviations above the mean and patients within 2SD.
Statistical analysis was performed using a statistical software program, SPSS 13.0 for Windows (SPSS Inc., Chicago, IL). For group analysis, we used repeated measures analysis of variance (ANOVA) with between subjects factor (instrumentation group) to compare change over time. For comparisons across the 4 instrumentation groups at one time point, ANOVA with Bonferonni post hoc comparisons was used. All data were checked for normality and equality of variances before ANOVA testing.
Two hundred eighty-three patients underwent posterior spinal fusion with segmental instrumentation for the treatment of AIS and had follow-up of greater than 2 years. There were 229 females and 54 males. The mean age at surgery was 14.6 ± 2 years. The Lenke types were as follows: 156 Lenke 1 curves, 61 Lenke 2 curves, 12 Lenke 3 curves, 14 Lenke 4 curves, 21 Lenke 5 curves, and 19 Lenke 6 curves. The mean number of fused vertebra was 12 ± 2. Risser stage was recorded as zero in 61 patients, 1 in 19 patients, 2 in 43 patients, 3 in 47, 4 in 80, and 5 in 33 patients. The 283 patients were divided into the groups as follows: group 1–hook only constructs, n = 51; group 2–Hybrid constructs, n = 177; group 3–screw only constructs, n = 37; group 4–screws with proximal hook constructs, n = 18. The average T5–T12 kyphosis before surgery was 24.5 ± 12.3°, 26.3 ± 13.3°, 23.9 ± 12.7°, and 29.4 ± 16.7°, respectively, without any statistically significant difference between groups. The average T5–T12 kyphosis after surgery was 24.0 ± 9.2°, 21.8 ± 8.8°, 18.2 ± 8.6°, and 20.9 ± 10.9°. When comparing group 1 to group 3, there was a significant difference in the 2 year T5–T12 kyphosis, 24.0 versus 18.2°, respectively.
Proximal Junctional Kyphosis
The adjacent level change in proximal kyphosis was significantly increased with pedicle screw constructs compared with hook (P = 0.014) and hybrid (P = 0.02) constructs. The preoperative adjacent level kyphosis was 3.6° ± 3.6° in group 1, 3.4° ± 4.1° in group 2, 2.6° ± 2.7° in group 3, and 2.0° ± 3.9° in group 4. Subsequently, at 2-year follow-up the adjacent level kyphosis measured 5.0° ± 3.8°, 5.7° ± 5.6°, 8.2° ± 9.2°, and 6.4° ± 6.4°, respectively. The change in kyphosis between groups was 1.4°, 2.2°, 5.6°, and 4.4°, respectively.
If “Too Kyphotic” Definition Changed
To further distinguish a normal from an abnormal change in kyphosis, the change in PJK (postoperative measurement minus preoperative measurement) was calculated. This measurement for the entire study group was plotted and found to have a normal distribution with a mean change of 3° and a standard deviation of 6° (Figure 5). Therefore, if we choose 2 standard deviations above the mean as being “abnormal” we are redefining abnormal PJK as any increased postoperative kyphosis of 15° or more. Using 15° as our definition for PJK, the incidence of PJK in our study was 0% (hooks only), 2.5% (hybrids), 8.1% (all screws), and 5.6% (screws with proximal hooks).
Additional Risk Factors
Eight patients were found to have a change in kyphosis greater than 2 standard deviations above normal. There was no statistically significant difference between patients who developed an increased kyphosis of 15° and those who did not comparing the SRS-22 scores, age of surgery, percent correction, or T5–T12 kyphosis (pre or after surgery) (P > 0.05). However, there was a difference between the 2 groups in body mass index. Patients with greater than 2 SD increase in PJK had a significantly increased BMI of 26.7 ± 15.2 compared with 20.9 ± 4.5 in those patients whose change in PJK was within 2 SD (P = 0.013).
The surgical treatment of AIS has evolved over the past 10 years and the current trend is for a posterior only approach with segmental instrumentation with pedicle screws. Recently, PJK has been recognized in adolescents and adults following posterior spinal fusion.4–8 While the incidence of PJK has ranged from 0% to 46%, the clinical implication of this is unclear.4–8,11 We would caution assigning the diagnosis of PJK to a postoperative patient in whom the clinical significance may be minimal. This is also corroborated by the patient based outcomes measure, the SRS-22 which showed no statistical difference in outcomes scores between patients with PJK and without PJK in this study. This was true for the total score as well as the individual domains of pain, function, self-image, activity, and satisfaction. Furthermore, the arbitrary selection of 5° or 10° based on the ability to measure radiographs in previous studies does not correlate with clinical symptoms or adjacent level disease. Therefore, we have proposed a new definition of abnormal being greater than 15° based on the standard deviation of all postoperative patients in our study group, and the assumption that normal is within 2 standard deviations. Until long-term follow-up is available and PJK is correlated with clinical symptoms, we recommend diagnosing those patients outside of 2 SD as being abnormal and having PJK.
The change in junctional kyphosis in our study was significantly greater with all pedicle screw constructs compared with hooks. Initially, although not statistically significant, Kim et al7 reported a trend towards an increase in kyphosis with pedicle screws compared with hooks, 9° versus 6°, respectively (P = 0.19). They also reported a trend towards an increased incidence of PJK (defined as greater than 10° change and an absolute value greater than 10°) with pedicle screws (38%) compared with hooks (19%).7 In the most recent study by Kim et al,9 they found a significant increase in PJK (using a definition of PJK which requires the angle across 2 segments to be greater than 10° and increased at least 10° compared with before surgery) when comparing pedicle screws to hooks (P = 0.029) and hybrids to hooks (P = 0.026). While they noted these significant differences, they did not find any difference between pedicle screw constructs and hybrid constructs; however, they used a different measurement of PJK compared to our study. The incidence of PJK in our study was 8.1% with pedicle screws, less than previously reported based on our new definition of PJK. The cause of increased kyphosis with pedicle screws and the clinical correlation of this are unclear.
Several hypotheses exist in an attempt to describe the cause of PJK and it remains likely multifactorial. Kim et al4,9 suggested the increased overall rigidity of pedicle screw and hybrid constructs compared with hooks was related to the increased kyphosis seen in their study. To further extrapolate, one could expect that pedicle screw constructs with their increased rigidity compared with hybrid constructs could have a further detrimental effect on proximal kyphosis. Additionally, Rhee et al6 in comparing anterior to posterior instrumentation described disruption of the posterior ligamentous structures, posterior compression correction techniques and decreased thoracic kyphosis within the construct as potential factors in developing PJK. In our study, all patients were instrumented posteriorly, but in addition to ligamentous disruption, pedicle screw placement in the upper thoracic spine often violates or affects the supra-adjacent facet capsule. One important potential explanation for the increased PJK seen in the pedicle screw only constructs was the significant difference in T5–T12 kyphosis. There was significantly less kyphosis across T5–T12 in the pedicle screw group and this could lead to an increase in PJK when the spine attempts to compensate for the overall changes in sagittal balance. Therefore, sagittal balance and thoracic kyphosis remains an essential consideration during curve correction. PJK seen with pedicle screw constructs was likely the combination of increased posterior soft tissue disruption (capsular, ligamentous, and muscular), increased curve correction, decreased kyphosis, and increased construct rigidity.
One potential method of decreasing PJK is to place hooks at the top of an all pedicle screw construct. Although we admit our study did not find any statistical significance with doing so, we did find a trend towards decreased kyphosis which would correlate with the previously mentioned causes of PJK. Placement of transverse process hooks at the top could decrease the structural rigidity of the construct providing a transition to the noninstrumented spine and allow for less facet capsule and muscle disruption. While this is a potential solution to PJK, the results are not significant, and further investigation is required.
Our study contains several limitations, the first of which is its retrospective nature. Furthermore, this is a multicenter database, which incorporates several different surgeons and just as many different techniques. While we admit this is a limitation, it is the best method of gathering enough data to evaluate for changes between groups, and some may argue that this is a better representation of a cohort of patients than a single surgeon's series over time. Another limitation in our study is that all of our treatment groups are not equal. Matched cohort groups were not created as this would have significantly limited the sample size within each group, reducing the study's power.
Pedicle screws significantly increased the supra-adjacent level kyphosis with BMI serving as a possible risk factor, but the clinical significance of this is unclear. Because of the clinical uncertainty, we have defined the diagnosis of PJK as an increased kyphosis of 15° across one segment based on the standard deviation of all patients in our study. A potential solution to PJK is the substitution of hooks at the upper-instrumented vertebrae, but further investigation is required.
- Adjacent level kyphosis was significantly increased with pedicle screw-only constructs for the surgical treatment of AIS.
- The clinical significance of proximal junctional kyphosis (PJK) is uncertain, and we recommend a change in the definition of PJK as 15° across the adjacent motion segment as diagnostic of PJK.
- Increased body-mass index (BMI) may be a risk factor for PJK.
- There was a trend towards a decrease in proximal kyphosis with placement of hooks at the upper instrumented vertebra compared with screws.
1.Cotrel Y, Dubousset J, Guillaumat M. New universal instrumentation
in spinal surgery. Clin Orthop
2.Kim YJ, Lenke LG, Bridwell KH, et al. Free hand pedicle screw placement in the thoracic spine: is it safe? Spine
3.Webb JK, Burwell RG, Cole AA, et al. Posterior instrumentation
in scoliosis. Eur Spine J
4.Kim YJ, Βridwell KH, Lenke LG, et al. Proximal junctional kyphosis in adolescent idiopathic scoliosis
following segmental posterior spinal instrumentation
and fusion: minimum 5-year follow-up. Spine
5.Lee GΑ, Βetz RR, Clements DH III, et al. Proximal kyphosis
after posterior spinal fusion in patients with idiopathic scoliosis
6.Rhee JM, Bridwell KH, Clements DH III, et al. Sagittal plane analysis of adolescent idiopathic scoliosis
: the effect of anterior versus posterior instrumentation
7.Kim YJ, Lenke LG, Cho SK, et al. Comparative analysis of pedicle screw versus hook instrumentation
in posterior spinal fusion of adolescent idiopathic scoliosis
8.Kim YJ, Lenke LG, Kim J, et al. Comparative analysis of pedicle screw versus hybrid instrumentation
in posterior spinal fusion of adolescent idiopathic scoliosis
9.Kim YJ, Lenke LG, Bridwell KH, et al. Proximal junctional kyphosis in adolescent idiopathic scoliosis
after 3 different types of posterior segmental spinal instrumentation
and fusions. Spine
10.Carman DL, Browne RH, Birch JG. Measurement of scoliosis and kyphosis radiographs. Intraobserver and interobserver variation. J Bone Joint Surg Am
11.McCance SE, Denis F, Lonstein JE, et al. Coronal and sagittal balance
in surgically treated adolescent idiopathic scoliosis
with the King II curve pattern. A review of 67 consecutive cases having selective thoracic arthrodesis. Spine