Prevention of Proximal Junctional Kyphosis Using Proximal Fixation Techniques

Background Adult spinal deformity (ASD) is a disorder characterized by abnormal curvature of the spine resulting from progressive degeneration of spinal elements. Although operative intervention for ASD is commonplace, it is associated with several complications, including proximal junctional kyphosis (PJK) and proximal junctional failure (PJF). The objective of this review is to outline the role of proximal fixation in preventing PJK and PJF. Methods We conducted a literature search using the Embase, Scopus, Web of Science, CINHAL, Cochrane Library, and PubMed MEDLINE databases. We considered only studies focusing on adult patients and selected clinical studies investigating proximal fixation techniques. Results There was mixed evidence of the efficacy of hooks and other instrumentation methods in preventing PJK, although most studies supported the use of hooks. Selection of lower thoracic vertebrae was associated with higher rates of PJK and PJF in several studies, although the relationship was inconsistent, and many studies reported no significant difference in rates of PJK or PJF between different upper instrumented vertebra (UIV) levels. Other techniques that are not related to specific instrumentation or vertebral selection, such as adjusting UIV screw trajectory, were also referenced. However, the evidence supporting these techniques was limited. Discussion Despite the presence of numerous studies in the literature discussing proximal fixation strategies to reduce the incidence of PJK/PJF, the lack of prospective studies and high variability in study methods make direct comparison challenging. We could not draw strong conclusions regarding the superiority of any one technique, despite promising clinical results with a strong biomechanical basis in several studies. Clinical Relevance This systematic literature review showed that a variety of proximal fixation techniques have been used to prevent PJK/PJF without clear evidence in favor of any particular technique. Level of Evidence 3.


INTRODUCTION
Over the next several decades, as the US population continues to age, the prevalence of adult spinal deformity (ASD) is expected to rise considerably.While operative intervention for ASD has been shown to be beneficial, it has been associated with relatively high rates of complications, including proximal junctional kyphosis (PJK).The prevalence of PJK among patients who have had ASD correction is generally thought to be between 17% and 39%, with some studies reporting incidence as high as 69% at 3-year follow-up. 1,2Many patients ultimately require revision surgery for proximal junctional failure (PJF), which can incur significant morbidity and cost for patients.Furthermore, there remains a risk for recurrence of PJK even after revision surgery, with one study reporting a recurrence rate of 44% among 70 patients. 3Because of the numerous challenges PJK presents, there has been considerable research on prevention.Several patient-specific, radiographic, and surgical risk factors for PJK have been identified. 4Various prevention strategies have been adopted to address these risk factors, but a widely accepted algorithm has yet to emerge.One area under investigation is surgical prevention through proximal fixation.The underlying theory of this strategy is that a more gradual transition of forces to the noninstrumented spine, with less disruption of natural spine biomechanics, can prevent PJK.The objective of this review is to assess the current literature to understand the role of proximal fixation techniques in preventing PJK.We consider the options that have been studied to date, the strength of their supporting evidence, and their efficacy.
Cobb angle of ≥10° between the upper instrumented vertebrae (UIVs) and the superior 2 levels (UIV+2) and an angle ≥10° greater than the preoperative baseline.PJF is commonly recognized as any form of PJK requiring revision surgery.This can be due to UIV or UIV+1 fracture, posterior osseo-ligamentous disruption, or failure of instrumentation at the UIV. 6Many risk factors have been identified as possible contributors to PJK.Patient-specific risk factors such as age, body mass index, bone density, smoking, and the presence of other medical comorbidities have all been proposed.][9][10][11] Several surgical risk factors are also thought to contribute to PJK.These are generally related to disruption of surrounding supportive soft tissues, 12 excessive construct rigidity, 13 choice of UIV, 14 or magnitude of deformity correction. 9Finally, radiographic risk factors include a high preoperative sagittal vertical axis 15,16 and a high degree of thoracic kyphosis. 17,18JK generally becomes apparent within the first 1 to 2 years after surgery.Kim et al 10 reported that 59% of proximal junctional angle (PJA) progression occurs within the first 8 weeks postoperatively.However, in the same study, 35% of the total PJA progression occurred more than 2 years after surgery, suggesting PJK has a progressive component as well.In contrast, PJF is often a relatively early complication.Yagi et al 19 reported that 87% of patients with PJF had undergone revision surgery within 2 years after surgery, with a mean time to revision of 10 months.Other studies have reported similar findings, with average times to revision of approximately 6 months. 6,20There remains controversy over the clinical significance of PJK; although a considerable proportion of patients are asymptomatic, many studies report poorer functional scores and worse patient-reported pain in patients with PJK. 8,21Patients who require revision surgery are subject to perioperative complications and considerable costs.Hart et al 22 estimated an average cost of $77,432 for revision of PJF.

PREVENTION THROUGH PROXIMAL FIXATION
The Table summarizes selected studies.These studies were reviewed independently by 2 of the authors, and their subject matter was deemed relevant to the current review.Conflicts were resolved via discussion.Studies were classified according to the author, study type, quality assessment based on the Newcastle-Ottawa Score, number of patients included, type of operative treatment utilized, mean age of included patients, minimum follow-up (if specified), prophylactic technique or implant investigated, and relationship to PJK incidence.

Hooks
The use of various hooks at the UIV has been investigated as a preventive measure.The underlying theory of this technique is that, compared with pedicle screws (PSs), hooks at the UIV provide a more gradual transition of construct stiffness to the upper, noninstrumented vertebrae and less mechanical stress.1][42] Photographs of PS and laminar hook constructs in a porcine model are shown in Figure 1 for reference. 43Several clinical studies support the use of hooks at the UIV.Cazzulino et al 25 describe a "soft-landing technique" in which they used unilateral preservation of the soft-tissue sleeve combined with one or more hooks on the contralateral side in 39 patients.They observed an incidence of radiographic PJK (which they defined as PJA greater than 10°, differing from the traditional definitions) of 41% over an average of 2.2 years of follow-up.Four patients (10%) required an extension of the construct and met the criteria for PJF.They did not have a control group for comparison.Hassanzadeh et al 21 used transverse process hooks (TPHs) in 47 patients undergoing long spinal fusion, with 20 receiving a TPH and 27 receiving a PS alone as a control cohort.The minimum follow-up was set at 2 years.None of the patients in the TPH group developed PJK, while 8 patients in the PS group developed PJK (P = 0.01).The TPH group also had significantly higher functional scores compared with the PS group at the final follow-up (P < 0.05).In a multicenter study of 625 patients, Line et al 32 analyzed the effect on PJF rates with no supplementary fixation vs various implant options.Augmentation methods included TPHs, cement vertebroplasty, and tethering.PJF occurred at a rate of 20.3% in the no-augmentation group and 10.3% in the augmentation group.TPHs had the lowest rate of PJF at 7% (n = 115).
Despite several promising clinical studies, there is also evidence showing no association and even a negative association between hooks and PJK incidence.Matsumura et al 33 compared TPH with PS in 39 patients; 17 patients received a TPH and 22 received a PS at the UIV.The incidence of PJK (defined solely as a change in the

Table.
Continued.PJA of >20°) was 17.6% in the TPH group and 27.3% in the PS group, but this difference was not found to be statistically significant.They did find a larger change in PJA in the PS group compared with the TPH group, 19° vs 5° (P = 0.04).Only a single study found that TPHs were associated with higher rates of PJK.Tsutsui et al 37 performed a retrospective review of 28 patients with a TPH and 25 with a PS at the UIV at either T-9 or T-10.They found a significantly higher incidence of PJK in the TPH group compared with the PS group (25.7% vs 8.0%) at a 1-year follow-up.Ultimately, there is a lack of high-quality evidence supporting the use of hooks as a definitive PJK prevention measure, albeit with several promising clinical studies and sound underlying biomechanical evidence.Larger trials are needed to investigate this method further.
Another prophylactic option for PJK prevention is the use of vertebral tethers around the UIV.Operating on a similar principle as TPHs, vertebral tethers are hypothesized to provide a more gradual transition of forces from the UIV to the noninstrumented vertebrae above, therefore reducing the incidence of PJK.5][46] There is promising recent clinical evidence that suggests that various tethering techniques are effective in reducing the incidence of PJK. 47,48However, despite this evidence, a variety of tethering techniques are under investigation, and some studies have shown that certain methods are not associated with significantly different PJK incidence. 49Other studies have shown that tethering, while effective at reducing PJK incidence, is not demonstrably superior to other methods such as the use of hooks, 32 with an optimal instrumentation configuration remaining unclear.

Screw Techniques
Alternative screw techniques at the UIV have been implemented to achieve a similar reduction in construct stiffness.One such technique involves cranially directed transvertebral screws, which are inserted at the inferolateral aspect of the most superior pedicles.The screw is then directed obliquely and superiorly across one or more vertebral levels, depending on the technique.Sandquist et al 35 reported on PJF in a sample of 15 patients who received screws placed with this technique, termed multilevel stabilization screws (MLSSs), with at least 1-year follow-up.A 3-dimensional reconstruction of this technique is shown in Figure 2. 35 None of the 15 patients experienced PJF, and there was an average change in their sagittal Cobb angle of 4°, ranging from −0.92° to 9.13°.The same researchers published a later study of 76 patients, including 26 controls receiving standard instrumentation. 29The MLSS group had a significantly lower incidence of PJK when compared with the control group, 10.0% vs 30.8% (P = 0.023), as well as a significantly lower PJA (1.3° ± 5.3° vs 5.2° ± 6.3°, P = 0.014).Another study investigated the role of polyaxial vs monoaxial screws as another means of reducing construct stiffness.Wang et al 38 used monoaxial screws (MAS) instead of the more frequently used polyaxial screws (PAS) in a sample of 242 patients, observing no difference in PJK rates when compared with the polyaxial screw technique (22.2% and 24.0%, respectively).

Implant Orientation and Trajectory
It is well known that PS angle modulates mechanical stress and screw pullout strength in biomechanical models, 50 and this has led some to postulate that screw trajectory may also influence PJK.Harris et al 51 studied PS trajectory as a risk factor for PJK and found that screws angled greater than 3° cranially were associated with significantly increased incidence of PJK and PJF.The authors hypothesized that this was a result of the creation of a larger bone channel when using a caudally directed trajectory, decreasing stress and increasing pullout strength.This benefit was balanced against the risk of facet joint violation with additional caudal angulation.
In addition to screw trajectory, the mismatch between proximal rod contour and PJA has been analyzed in several studies.Yan et al 52 found that greater angle mismatch between the PJA and proximal rod contour was associated with higher rates of postoperative PJK compared with constructs with rod contours that more closely matched the PJA.Numerous additional studies support this finding.For example, in a retrospective case series, Yang et al 53 found that the cohort of patients with PJK included a significantly higher proportion of individuals with poorly matched (greater than 5°) proximal rod contouring compared with the cohort without PJK (69% vs 25%). 53However, many of these studies were conducted with populations that included adolescent patients, making their generalizability to entirely adult populations questionable.

Soft Tissue Preservation
Violation of and excessive damage to soft tissues around the UIV are believed to contribute to the development of PJK. 54,55Various methodologies to preserve local soft-tissue envelopes and ligamentous support have been described.These techniques are typically combined with additional instrumentation prophylaxis as opposed to being used as standalone methods.The increased use of minimally invasive surgical (MIS) techniques has led several groups to investigate MIS and hybrid fixation options for preserving soft tissue integrity.Mummaneni et al 56 compared PJK rates in circumferential MIS (cMIS) with a hybrid approach for ASD and found lower rates of PJK in the cMIS group but no difference when controlling for a number of levels fused.Burks et al 24 studied percutaneous fixation at the upper 2 instrumented levels in hybrid-MIS ASD correction in a case series of 36 patients. 24They observed rates of PJK that were similar to those in the generally published literature but did not have a control group for comparison.Thus far, there has not been a consistently demonstrated advantage of MIS techniques compared with traditional open surgery for the prevention of PJK.

UIV Selection
The influence of UIV level on the incidence of PJK has long been debated.Generally, it is important to select a stable, neutral vertebra and a level that allows for the inclusion of spinal segments with significant baseline kyphosis.However, these decisions are not always straightforward; extension of the construct to the upper thoracic (UT) levels may be associated with its own perioperative complications and may increase overall construct stiffness.Furthermore, the involvement of higher vertebral levels may increase the risk for catastrophic PJK and devastating neurological injuries compared with shorter fusions.The literature has not yet provided clear guidance on an optimal strategy, and there is mixed evidence about a consistent relationship between the UIV level and the development of PJK.
In several studies, the selection of a more caudal UIV was related to higher rates of PJK, PJF, and the need for surgical revision.Daniels et al 27 retrospectively analyzed 303 patients and observed significantly lower rates of PJK (odds ratio [OR] 0.49, 95% confidence interval [CI] 0.24-0.99)and lower rates of PJF (OR 0.54, 95% CI 0.24-1.22)for a UT UIV compared with a lower thoracic (LT) UIV at 2-year follow-up.Cho et al 26 observed 51 patients with 2 years of follow-up and found that the rates of PJK were greater when the UIV was at or below the upper-end vertebra of the deformity, with all 5 cases (5 of 51, 10%) of PJK occurring in this group.A larger cohort study of 165 patients treated with pedicle subtraction osteotomy found a greater number of patients requiring surgical revision for PJK based on UIV location, with 2 of 11 in the UT group and 9 of 11 in the thoracolumbar (TL) group (P = 0.0274).However, overall rates of PJK were not significantly different between the groups. 36In a case-control study of 252 ASD patients who underwent posterior instrumentation and fusion, Lafage et al 31 found that 63.5% of patients in the TL group developed PJK, compared with 49.2% in the UT group (P = 0.02).Another aspect to consider in UIV selection is the position in which patients are evaluated during preoperative imaging.Typically, patients are evaluated in a standing position for radiographs, yet many patients spend significantly larger portions of their lives sitting.Yoshida et al 39 performed a study in which both sitting and standing preoperative radiographs were used to determine the relationship of PJK to various parameters, including coronal distance to C2 and C7 plumb lines.They hypothesized that a sitting radiograph would be a better predictor of PJK as it better represents resting biomechanics.They found that a UIV that was greater than 115 mm from the C2 plumb line on sitting radiographs was significantly associated with a higher incidence of PJK on logistic regression analysis.This finding was not replicated with standing radiographs.The distance to the C7 plumb line was also investigated in this study but was not found to be significantly associated with PJK.
There have been several studies showing no relationship between UIV level and PJK incidence.In a retrospective cohort study of 58 patients, O'Shaughnessy et al 34 found no significant difference in either PJK incidence or surgical revision rates between the UT and LT groups with a minimum of 2 years follow-up.Similarly, Kim et al 9 followed 198 patients undergoing long instrumentation (>5 levels) for ASD for 2 years.When the patients were divided into UT and LT groups, their measured PJK angles and surgical revision rates were not significantly different at follow-up.Buell et al 23 conducted a retrospective analysis of 560 patients who underwent surgical correction for ASD with 2 years of follow-up.They found similar rates of reoperation for PJK in the UT and LT groups, 9.8% vs 8.6%, respectively (P = 0.810).Ha et al 28 compared the incidence of PJK in patients with upper fixation at either proximal thoracic (PT) or distal thoracic (DT) levels.Among the 89 patients, there were no significant differences between the PT and DT cohorts in rates of PJK, PJF, or other measured clinical outcomes.They did find, however, that vertebral compression fracture was more prevalent in the DT group, while subluxation was more prevalent in the PT group.Kim et al, 30 in a study of patients undergoing instrumentation and fusion for ASD, established 3 retrospective cohorts based on UIV location (T9-T10, T11-T12, and L1-L2) with 2 years of follow-up.There was no significant difference in overall rates of PJK (51%, 55%, and 36% for T9-T10, T11-T12, and L1-L2, respectively, P = 0.2).There were also no significant differences in the total change in the PJA from preoperative measurement to final follow-up (P = 0.46).

SUMMARY
PJK and PJF are potentially severe complications of corrective surgery for ASD, with a prevalence that is expected to increase in the United States as the population ages.Prevention of PJK and PJF will be of great importance in providing effective, high-value care and will require the development of well-defined prevention strategies.Proximal fixation techniques and surgical planning represent one subset of these strategies, although effective PJK prophylaxis will undoubtedly require a mix of surgical and patient-specific strategies.While there is compelling biomechanical and clinical data to support more in-depth clinical investigation into proximal fixation techniques such as PJK prophylaxis, current data are insufficient to draw firm conclusions about the superiority of any one method.

Figure 1 .
Figure 1.Posterior (A) and superior (B) views of porcine spinal constructs instrumented with traditional pedicle screws, left, and laminar hooks, right.Source: Reprinted from Figure 4 in Tai et al.Biomechanical comparison of different combinations of hook and screw in one spine motion unit-an experiment in porcine model.BMC Musculoskelet Disord.2014;15:197 43 under Creative Commons CC BY license.

Figure 2 .
Figure 2. Three-dimensional reconstruction demonstrating a multilevel stabilization screw construct.Source: Reprinted from Figure 3 in Sandquist et al.Preventing proximal junctional failure in long segmental instrumented cases of adult degenerative scoliosis using a multilevel stabilization screw technique.Surg Neurol Int.2015;6:112 35 under Creative Commons license CC BY.

Table .
Selected clinical studies of proximal fixation techniques in adult patients.