Large Difference between Proximal Junctional Angle and Rod Contouring Angle is a Risk Factor for Proximal Junctional Kyphosis

Background: There are several risk factors for proximal junctional kyphosis (PJK) in adolescent idiopathic scoliosis (AIS) surgery. Decreased rod contouring angle (RCA) has been proposed as a risk factor for PJK, but the role of difference between proximal junctional angle (PJA) and RCA (PJA-RCA) has not been fully investigated. The aim of this study was to assess the role of PJA-RCA for the development of postoperative PJK in AIS. Methods: We performed a retrospective analysis of 96 AIS patients who underwent posterior segmental spinal instrumentation and fusion between the years 2012 and 2018 (minimum 1.5-year and average 2-year follow-up) at a single institution. Each patient was measured on preoperative, postoperative and final follow-up long-cassette standing radiographs. The PJA-RCA was regarded as a new definition that reflects the match degree between proximal rod contouring and vertebra curvature, and radiographic parameters were compared between PJK and non-PJK group. Results : Among the 96 patients with a mean age of 14.00 years (± 0.82), the overall incidence of PJK was 22%. PJK group showed a significantly greater preoperative SVA (P = 0.032) and larger correction of SVA (P = 0.007) than non-PJK group. At the last follow-up, PJK patients had significantly greater LL (P = 0.046). Patients in the PJK group had significantly greater preoperative PJA-RCA than the non-PJK group (4.07±3.30 vs. 1.42±4.28, P = 0.024). However, RCA was not significantly different between two groups (3.88 ± 4.34 vs. 2.86 ± 3.36, P = 0.405). In addition, Pearson correlation coefficient showed a significant correlation between the change of SVA and the last follow-up PJA (r = -0.208, P = 0.042). Preoperative PJA-RCA and postoperative PJA-RCA demonstrated similar results which showed a strong correlation with the last follow-up PJA (r = 0.528 PJK in AIS, and PJK might be a compensation mechanism rather than complication when spine is shifted and overcorrected.

PJK in AIS, and PJK might be a compensation mechanism rather than complication when spine is shifted and overcorrected.

Background
Patients with all-pedicle screw instrumentation had the greatest curve correction percentage, maintenance of this correction in the coronal and sagittal planes, less blood loss, shorter operation time, higher patient satisfaction and better self-assessment of appearance in operative treatment of adolescent idiopathic scoliosis (AIS). [1][2][3] Based on these advantages, all-pedicle screw construct has been gradually adopted as the current trend of adolescent idiopathic scoliosis surgeon. However, Kim et al [4] reported that allpedicle screw construct demonstrated a significantly higher two-year postoperative prevalence of proximal junctional kyphosis (PJK) (35%) compared with the hook-only (22%) and hybrid (29%) groups. Helgeson et al [5] had similar conclusion and found adjacent level proximal kyphosis was significantly increased with pedicle screws. As one of the most common adjacent pathology, PJK is typically defined as the angle between the inferior endplate of the upper instrumented vertebra (UIV) and the superior endplate of the UIV 2 is 10° or greater, and at least 10° greater than the preoperative value.
[6] The incidence of PJK in AIS patients following posterior instrumented arthrodesis ranges from 14% to 35%.[4, [7][8][9] While patients with PJK may be initially asymptomatic with worsening degrees of PJK, they can develop the structural failures that characterize proximal junctional failure (PJF), which may be accompanied by subsequent pain, neurologic deficit, gait difficulties, sagittal imbalance, social isolation and higher need for revision surgery.
[10] Based on previous studies, multiple risk factors which correlate significantly with PJK have been extensively evaluated and reported. These include a larger preoperative thoracic kyphosis angle, greater immediate postoperative thoracic kyphosis angle decrease, thoracoplasty, the use of a pedicle screw at the top vertebra, autogenous bone graft, distal fusion below L2, and male sex. [4,7] Ferrero et al [8] found that patients with high pelvic incidence and consequently large lumbar lordosis (LL) and thoracic kyphosis (TK) were more at risk of PJK. Furthermore, optimal rod contouring has also been proposed as a means to prevent PJK. Lonner et al [11] reported that decreased rod contour angle (RCA: the Cobb angle between the UIV and one vertebra caudal to the UIV) were one of the major risk factors for PJK in AIS. (Fig. 1) However, as a new definition that reflects the relative match degree between proximal rod contouring and vertebra curvature, the impact of PJA-RCA has not been fully investigated and warrants further study in AIS patients with posterior spinal fusion.

Subjects
We performed a retrospective analysis of 96 AIS patients who underwent posterior segmental spinal instrumentation and fusion between the years 2012 and 2018 at a single institution. Minimum 1.5-year and average 2-year follow-up was required.
Inclusion criteria of this study were as follows: (1) diagnosis of AIS; (2) single-stage posterior spinal fusion with an all-pedicle screw construct; (3) instrumented levels more than 5; (4) complete radiographic follow-up with distinct radio graphic landmarks.
Exclusion criteria included the following: (1) patients diagnosed with early onset scoliosis and neuromuscular scoliosis; (2) hybrid instrumentation constituted of hooks and wires; (3) history of Smith-Petersen osteotomy, pedicle subtraction osteotomy, vertebral column resection, and (5) reoperations related to the pedicle screw system.

Data Collection and Radiographic Parameters
Each patient was measured on preoperative and postoperative long-cassette standing radiographs. Basic demographic and surgical data collected included age at surgery, gender, Risser, and levels instrumented. The radiological parameters were as follows: (1) Pelvic parameters: pelvic incidence (PI) and pelvic tilt (PT). Based on the criteria that PJK is defined as a PJA more than 10°, and at least 10° greater than the preoperative value at the last follow-up, patients were categorized into two groups: PJK and non-PJK, with 21 and 95 patients respectively.

Statistical Analysis
Statistical analysis was performed with IBM SPSS Statistics v.21.0 (IBM Corp., Armonk, New York, USA). For changes in radiographic parameters, a Paired t tests was used and for differences between categorical variables, an independent t test was used. Pearson correlation coefficient (r) was used to establish relationships between potential risk factors and developing PJK. Statistical significance was defined as P < 0.05. and SVA reveal significant differences (P < 0.05, Table 2),, while the PI showed no significant change after surgery (P = 0.079). No statistical difference were demonstrated between PJK and non-PJK group at the preoperative, postoperative, and final follow-up in the sagittal parameters of PI, PT, and TK (P > 0.05, Table 3). PJK group showed a significantly greater preoperative SVA than non-PJK group (P = 0.032). At the last followup, PJK patients had significantly greater LL (P = 0.046). Meanwhile, PJK experienced larger correction of SVA (P = 0.007, Table 3)..
Patients in the PJK group had significantly greater preoperative PJA-RCA than the non-PJK group (4.07±3.30 vs. 1.42±4.28, P = 0.024, Table 1).. However, RCA was not significantly  indicate that PJK mainly occurred within the year of 1.5 or 2 after surgery. We defined PJK as the Cobb angle between the UIV and the UIV 2 is 10° or greater, and at least 10°g reater than the preoperative value.
[6] However, due to the inconsistent definitions, the incidence of PJK could vary hugely. In a 1999 study by Lee et al [12], the authors used 5° as the cutoff value and the incidence of PJK was 46%. Helgeson et al[5] measured the Cobb angle between the UIV and the UIV 1, and redefined PJK as any postoperative kyphosis increase ≥15°. The incidence of PJK was 8.1% in their pedicle screws group.
Thus, a consensual and precise definition of PJK is needed in the further study.
As shown by the current data, we found greater preoperative PJA-RCA in PJK group, a strong correlation between PJA-RCA and the last follow-up PJA, but no significant difference in RCA between PJK and non-PJK group which showed the contrary result with Lonner et al [11]. The reason might be that decreased RCA reflects an absolute straight rod curve while greater PJA-RCA means a relative straight rod curve. Moreover, they found RCA was a risk factor for PJK in Lenke 2 and 4 curves, but our object included other curve types. Theoretically, improper rod contouring would reciprocally change the biomechanical forces in proximal junctional part. A biomechanical study reported that the increase of the sagittal rod curvature from 10° to 20°, 30°, and 40° increased the proximal junctional angle (PJA) (by 6%, 13%, and 19%) and the proximal flexion force (3%, 7%, and 10%) and moment (9%, 18%, and 27%), [13] indicating that absolute increased RCA may be also associated with PJK. Therefore, the risk factor of "relative straight rod curve" would be more convincing. According to Dubousset's concept of ''cone of economy'', the spine alignment tends to balance itself in all three dimensions with minimal energy. Before the operation, AIS patient's body has experienced self-compensation to adapt to spine misalignment. When the spine alignment is corrected and changed, there will be a rebalance. In case of maintaining a similar compensation in junctional area, mismatched rod curve would be easier to create a vertebral angulation, (Fig. 3, A ≈ B) which could result in PJK further. Consequently, proper rod contouring should be seriously taken into consideration during the AIS spinal surgery.
The etiologies of PJK are multifactorial. In particular, residual sagittal imbalance and greater curvature correction were previously revealed as potential risk factors for PJK. [4,8,9,11] In this study, PJK group showed a significantly greater preoperative SVA (P = 0.032) and larger correction of SVA (P = 0.007, Table 3 There were some limitations in our study. First limitation of this study was its retrospective nature. Second, the number of cases was small and we did not categorize the type of AIS curves. Third, there was a lack of evaluation to muscle condition and   Figure 1 The measurement of proximal junctional kyphosis (PJK) and rod contouring angles (RCA).