Contemporary spine fusion techniques have facilitated correction of spine deformities. However, with the advent of modern spine deformity correction techniques, the incidence of proximal junctional kyphosis (PJK) has risen steeply. Reported postoperative PJK rates following adult spinal deformity (ASD) surgery range as high as 20% to 40%.1–8 While there is some debate regarding the clinical relevance of PJK (ranging from benign kyphosis to severe deformity associated with pain and/or neurological deficit), the term proximal junctional failure (PJF) has been used to denote the more severe form of PJK. Reported postoperative ASD rates for PJF range from 1.4% to 5.6%, and even more concerning, up to 47% of ASD patients with PJF require revision spine surgery.7,9,10
Initial reports that described PJK attributed surgical overcorrection of sagittal deformities in Scheuermann kyphosis and adolescent idiopathic scoliosis as a risk factor for PJK, however, much of the subsequent research on techniques to prevent PJK have focused on the use of surgical implants rather than methodologies to identify and prevent overcorrection of the sagittal plane.11,12 Several reports have described different prophylactic surgical implants to prevent PJK and/or PJF including: (1) use of transverse process hooks rather than pedicle screws at the upper instrumented vertebra (UIV); (2) use of transition rods at the cranial aspect of long instrumented pedicle screw constructs; (3) insertion of polyethylene tethers at the UIV, UIV+1, and/or UIV+2; and (4) performance of vertebroplasty at the UIV and UIV+1.6,13–19 However, the use of specific surgical implants targeted to prevent PJK has not eradicated the problem, and the occurrence of PJK remains troublesome.
Sagittal malalignment is an established cause of pain and poor quality of life in ASD.20 Consequently, a large emphasis has been placed upon correcting sagittal malalignment. The enthusiasm to restore physiological sagittal alignment, has in part however, led to overcorrection of regional and global sagittal deformities. Recent studies have demonstrated a deleterious effect of overcorrection of sagittal deformities.5,21–27 Lafage et al28 reported that with increasing age, there is an increase in positive sagittal alignment. The authors reported that older patients often do not need the same sagittal alignment, or magnitude of surgical correction, as younger patients, and provided a series of age-adjusted sagittal spinopelvic alignment parameters to serve as guidelines for patient assessment and surgical correction. Further work by Lafage et al29 went on to demonstrate that overcorrection of age-adjusted sagittal alignment is a risk factor for PJK.
PJK and PJF are likely multi-factorial phenomena, with no single solution. However, there is no existing research that has compared the efficacy of prophylactic surgical implants versus the avoidance of sagittal overcorrection to prevent PJF. Therefore, the purpose of this study is to evaluate the ability of surgical implants alone to reduce the incidence of PJF compared with the ability of avoidance of sagittal overcorrection of age-adjusted sagittal spinopelvic alignment parameters to reduce the incidence of PJF. This analysis will be made utilizing a propensity score matched analysis of surgically treated ASD patients from a large, multi-center database.
MATERIALS AND METHODS
ASD patients evaluated at one of 11 participating research sites in the United States were enrolled into a prospective, multi-center ASD database (the database). Database inclusion criteria are more than 18 years of age and a minimum one of the following, scoliosis more than 20°, sagittal vertical axis more than 5 cm (SVA), pelvic tilt more than 25° (PT), and/or thoracic kyphosis more than 60° (TK). All participating sites received institutional review board (IRB) approval prior to enrolling patients. Patients were then provided with the option of surgical versus nonoperative care at the discretion and counseling process of the treating surgeon. All surgery performed was at the discretion and consent process between the treating surgeon and the patient. The database is a prospective, observational database, therefore there was no randomization or study control of any type of treatment provided for any patient. In addition to the database inclusion criteria, additional inclusion criteria for the current study were: (1) patients surgically treated for ASD with more than or equal to five spine levels fused posteriorly; and (2) minimum 1-year follow-up (the 1-year follow-up window for the database was 9–23 months). One year was used as follow-up criteria for PJF assessment in this study because previous literature indicates the majority of PJF cases occur within 1 year postoperatively.30 PJF was defined in this study using a mathematical methodology described previously31:
- Upper thoracic PJF = Proximal junctional angle more than or equal to 28.0° and Δ Proximal junctional angle more than or equal to 21.6° or Proximal junctional anterolisthesis more than or equal to 8 mm and Δ Proximal junctional anterolisthesis more than or equal to 8 mm.
- Lower thoracic PJF = Proximal junctional angle more than or equal to 28.0° and Δ Proximal junctional angle more than or equal to 21.6° or Proximal junctional anterolisthesis more than or equal to 3 mm and Δ Proximal junctional anterolisthesis more than or equal to 3 mm
Demographic, radiographic, and surgical data were evaluated, and postoperative radiographs evaluated for the occurrence of PJF, as defined above. Risk factors for PJF were then identified from this cohort, which included patient age, modified adult spinal deformity frailty index (frailty, as previously defined), preoperative SVA, preoperative T1 pelvic angle (TPA), preoperative pelvic incidence minus lumbar lordosis (PI–LL), preoperative lumbar lordosis (LL), and fusion to pelvis.32,33 A propensity score matched analysis (PSM) was then performed controlling for the identified PJF risk factors. All patients were assessed for use of any surgical implants to prevent PJF (IMPLANT) or no use of surgical implants to prevent PJF (NONE). The specific types of surgical implants used to prevent PJF were identified including: (1) injection of polymethylmethacrylate cement at the UIV and UIV+1 (CEMENT); (2) transverse process hook placed at the UIV (HOOK); or (3) insertion of polyethelyne tether at the spinolaminar junction of the UIV+1 and/or UIV+2 (TETHER) as previously described (Figure 1A–E).6,13,18,34 The use of CEMENT, HOOK, TETHER, or NONE was performed at the discretion of the treating surgeon. There was no study randomization or control for any surgical treatment. Postoperative sagittal alignment for all patients was evaluated and patients were assessed for overcorrection of age-adjusted sagittal alignment using previously described age-adjusted sagittal alignment parameters.28,29 Patients that had a PI–LL less than normative values for the patient's respective age group were categorized as overcorrected (OVER), whereas patients that had a PI–LL more than or equal to normative values for the patient's respective age group were categorized as aligned (ALIGN; Figure 2A, B). As previously recommended, PI–LL was used as the assessment criteria for overcorrection because SVA and PT can increase following the occurrence of PJF, whereas PI–LL typically remains constant and is more consistently reflective of the surgical induced sagittal alignment, especially when the entirety of the lumbar lordosis is within the surgical construct.27–29 The incidence of PJF, as defined above, was then evaluated within the propensity matched cohorts, including NONE, IMPLANT, HOOK, CEMENT, TETHER, OVER, and ALIGN.
Standard analyses were performed using Student t test, analysis of variance with Tukey HSD and chi-squared analyses.2 Propensity score matching (PSM) was performed by matching the NONE and IMPLANT cohorts according to the identified risk factors previously mentioned for PJF. All comparisons were performed using jmp version 14.0.0 (SAS Institute Inc., Cary, NC). PSM was performed using genetic matching with the rgenoud algorithm (Version 5.8–10, http://sekhon.berkeley.edu/rgenoud) within the R (R version 3.4.3, The R Foundation for Statistical Computing Platform) package MatchIt (version 3.0.2, http://gking.harvard.edu/matchit) using RStudio (RStudio, Inc., Boston, MA).35,36
Between 2008 and 2017, 834 patients enrolled into the database met study inclusion criteria of which 625 were evaluated for this study (74.9% follow-up). Mean follow-up was 2.6 years (0.7–6.9), mean age 58.6 years (18.3–86.2), and sex distribution was 76.4% female (Table 1). Mean preoperative and postoperative SVA was 66.4 mm (–83.7 to 326.5) and 25.1 mm (–103.3 to 323.4), respectively (P < 0.05). Mean preoperative and postoperative PI–LL was 16.2° (–140.3° to 82.2°) and 1.1° (–163.2 to 61.7), respectively (P < 0.05). Using previously reported age-adjusted alignment parameters for PI–LL, 60.6% of patients met criteria for postoperative age adjusted sagittal overcorrection.28,29 Mean total spine levels fused was 12.0 (5–19), and 80.5% patients were fused to the pelvis. The total incidence of PJF was 13.9% and 7.4% of total patients were treated surgically for PJF with proximal extension of the posterior spinal fusion. Of the patients that developed PJF (n = 87), 53% were treated surgically (n = 46; Table 1).
Impact of Surgical Implants on PJF; NONE versus IMPLANT
PSM of NONE (n = 390) versus IMPLANT (n = 235) demonstrated both groups had similar mean patient age, sex ratio, frailty, preoperative and postoperative SVA, preoperative and postoperative PI–LL, total spine levels fused, percentage of fusion to the pelvis, and percentage of patients overcorrected for age-adjusted alignment (P > 0.05; Table 2). The incidence of PJF was greater for NONE (20.3%) than for IMPLANT (10.6%), however surgical revision rates for PJF were similar between NONE and IMPLANT (8.4% vs. 6.4%, respectively; P > 0.05; Figure 3).
Impact of Specific Surgical Implants on PJF; NONE versus CEMENT, HOOK, and TETHER
PSM of specific types of implants used to prevent PJF, including CEMENT (N = 58), HOOK (N = 115), TETHER (n = 62), and NONE (n = 390) demonstrated HOOK had lower mean age than the other cohorts, and smaller percentage of patients fused to the pelvis than all other cohorts, and HOOK had smaller preoperative SVA than TETHER (P < 0.05; Table 3). Additionally, CEMENT had fewer total spine levels fused than the other cohorts (P < 0.05; Table 3). All cohorts had a similar percentage of patients overcorrected according to age-adjusted PI–LL (P > 0.05; Table 3). The incidence of PJF was similar for CEMENT (12.1%), HOOK (7.0%), and TETHER (16.1%; P > 0.05; Figure 4). HOOK had a lower incidence of PJF (7.0%) than NONE (20.3%; P < 0.05, Figure 4). Surgical revision rates for PJF were similar for CEMENT, HOOK, TETHER, and NONE (P > 0.05; Figure 4).
Impact of Overcorrection of Age-Adjusted Sagittal Alignment on PJF; OVER versus ALIGN
PSM analysis of patients overcorrected (OVER, n = 379) versus not overcorrected for age-adjusted PI–LL (ALIGN, n = 246), demonstrated similar sex ratio, frailty, preoperative SVA, preoperative LL, and total spine levels fused for both groups (P > 0.05; Table 4). OVER had greater percentage of patients fused to the pelvis than ALIGN (95.1% vs. 74.4%, respectively; P < 0.05). Evaluation of patients that did not receive any surgical implant prophylaxis demonstrated that the PJF incidence for OVER-NONE (n = 225; 24.2%) was greater than ALIGN-NONE (n = 165; 13.2%, P < 0.05), and the incidence of surgical revision for PJF was greater for OVER-NONE (11.7%) compared with ALIGN-NONE (2.6%, P < 0.05; Figure 5). Evaluation of patients that received PJF implant prophylaxis demonstrated that the PJF incidence for OVER-IMPLANT (n = 154; 11.0%) was similar to ALIGN-IMPLANT (n = 81; 9.9%, P > 0.05), and the incidence of surgical revision for PJF was also similar for OVER-IMPLANT (7.1%) compared with ALIGN-IMPLANT (4.9%, P > 0.05). OVER-NONE had the greatest PJF incidence of all cohorts (24.2%; P < 0.05; Figure 5). Kaplan–Meier analysis of treatment cohorts showed OVER-NONE had earlier PJF and worse PJF survival-free duration than all other cohorts (P < 0.05; Figure 6).
Despite substantial research efforts PJK remains an incompletely understood and poorly prevented complication following ASD surgery. However, PJK is most commonly defined as a radiographic phenomenon, consequently PJK severity ranges from asymptomatic kyphosis to catastrophic failure.4 The goals of this study were to: (1) evaluate the postoperative incidence and risk factors of a more severe form of PJK, defined as PJF, in a large ASD population. Then, performing a propensity score matched analysis to control for identified PJF risk factors, (2) evaluate the efficacy of three different surgical techniques commonly described in the literature to prevent PJF; (3) evaluate the impact that overcorrection of age-adjusted sagittal spinopelvic alignment has upon PJF; and (4) evaluate the effects that combining surgical implants with avoidance of overcorrection of age-adjusted sagittal alignment has upon the incidence of PJF. The incidence of PJF in this entire study population was approximately 13.9%, which is consistent with previous reports.6,14–18,37 The identified risk factors for PJF in this study population, including age, large preoperative sagittal deformity, patient frailty, and fusion to the pelvis, are also consistent with previous reports.2,3,7,9,30,33,37–39 However, this study demonstrated that if no PJF prophylaxis is used and no attention is paid to avoid overcorrection of the sagittal plane, the PJF incidence was 24.2% and the surgical revision rate for PJF was 11.7%, both of which are among the highest rates reported in the literature. This study demonstrated that the use of surgical implants alone has a positive effect by reducing the incidence of PJF to 10.6% and surgical revision rates for PJF to 6.6%. Importantly, this study also showed that the use of age-adjusted alignment parameters can also reduce PJF, as patients receiving no prophylactic surgical implants and were not over corrected had a lower PJF incidence (13.2%) than patients that did not receive prophylactic implants but were over corrected (24.2%; P < 0.05). Additionally, if PJF prophylactic implants were combined with avoidance of overcorrection in the sagittal plane, the PJF incidence further reduced to 9.9% and surgical revision rates reduced to 4.9%. These findings further support the theory that attention to sagittal alignment and avoidance of overcorrection is an important component of a surgical strategy to avoid PJF.
Lowe and Kasten11 provided one of the earliest reports on PJK, describing proximal kyphosis after Cotrel-Dubousset instrumentation for kyphosis secondary to Scheuermann disease. The authors indicated that proximal junctional kyphosis was associated with overcorrection of the kyphotic deformity, defined as more than 50% correction of preoperative thoracic kyphosis, and recommended avoidance of overcorrection of more than 50% of kyphotic deformities in Scheuermann disease. However, much of the subsequent research surrounding PJK prevention has focused upon the use of surgical implants to prevent vertebral fracture by (1) hardening the UIV and/or UIV+1 by injecting cement in the vertebrae, or (2) dampen the forces at the cranial aspect of long constructs to prevent abrupt range of motion transitions between the instrumented and noninstrumented spine segments using hooks at the UIV or proximal tethers.6,13–19 Lafage et al28,29 recently refocused the attention of PJK prevention onto spinal alignment, by providing age-adjusted sagittal alignment parameters for adults. Lafage et al28,29 also reported that with increased overcorrection beyond age-adjusted sagittal alignment there is a corresponding increase in PJK severity. Our current findings that patients that received no surgical implant prophylaxis but were overcorrected in the sagittal plane had a higher incidence of PJF compared with patients that received no surgical implant prophylaxis but were not overcorrected further corroborate the findings by Lafage et al by demonstrating the importance of not overcorrecting sagittal alignment to prevent PJF.
There are several important limitations to this study. The data from this study was obtained from a prospective, observational database, therefore, there was no control or randomization of the patients that received surgery and, importantly, there was no control or randomization of patients that received surgical implants for PJK prophylaxis. Additionally, there was no randomization or control of the patients that were overcorrected in the sagittal plane. We attempted to mitigate the limitations of this observational database by (1) performing an initial analysis to identify PJF risk factors for the studied cohort, then (2) performing a propensity score matched analysis that controlled for the identified PJF risk factors. Consequently, many of the demographic, radiographic and surgical risk factors for PJF were similar between the compared cohorts, however, this analysis cannot control for all the confounding variables. Therefore, we recommend caution when interpreting the performance results of the specific implants to reduce PJF, and we do not advocate for use of, or superiority of, one specific implant over another. Instead, we recommend further controlled studies that are adequately powered to determine the superiority of a specific implant type. Additionally, we did not include patient reported outcome measures in this analysis. Therefore, we cannot state if patients with PJF in this study had worse clinical outcomes. Another limitation is that we focused our study efforts upon patients that had severe PJK (defined as PJF) by using a previously reported definition for PJF that is based upon radiographic thresholds for which patients commonly received surgery for severe PJK (also defined as PJF in the referenced study). However, there were no guidelines implemented in this study mandating that patients with PJF receive surgery if they met criteria for PJF.31 Additional research is needed that further refines the definition of PJF and controls for the associated treatments.
In conclusion, this study demonstrates that if no efforts are made to prevent PJF in multilevel fusion procedures for ASD by (1) not using surgical prophylactic techniques, and/or by (2) overcorrecting patients beyond age-adjusted sagittal alignment parameters, the incidence of PJF is unacceptably high, at 24.2%. The use of prophylactic surgical implants can reduce the incidence of PJF, however the ability to prevent PJF is further enhanced by paying strict attention to the sagittal plane and avoiding overcorrecting the patient's sagittal profile appropriate for the patient's age. Based upon these findings the authors have integrated attempts to routinely measure and adhere to age-adjusted sagittal plane alignment parameters, as well as routinely using surgical implants including hooks and spinous process tethers when surgically treating ASD. We are optimistic that the combination of prophylactic implants and use of age-adjusted alignment parameters to help personalize patient spinal alignment can help reduce the potentially catastrophic occurrence of PJF. Additional research efforts are needed to identify techniques that can help categorize and further reduce the incidence of PJF.
- Ideal PJF prevention following ASD surgery requires use of prophylactic surgical implants and avoidance of age-adjusted sagittal overcorrection.
- Adherence to age-adjusted sagittal alignment goals reduced PJF as effectively as use of surgical implant prophylaxis.
- ASD patients that did not receive surgical implant PJF prophylaxis and had overcorrection of the sagittal plane beyond recommended age-adjusted sagittal alignment parameters had the highest rate of PJF.
1. 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)
2. Yagi M, Akilah KB, Boachie-Adjei O. Incidence, risk factors and classification of proximal junctional kyphosis
: surgical outcomes review of adult idiopathic scoliosis. Spine (Phila Pa 1976)
3. Yagi M, King AB, Boachie-Adjei O. Incidence, risk factors, and natural course of proximal junctional kyphosis
: surgical outcomes review of adult idiopathic scoliosis. Minimum 5 years of follow-up. Spine (Phila Pa 1976)
4. Kim YJ, Bridwell KH, Lenke LG, et al. Proximal junctional kyphosis
in adult spinal deformity
after segmental posterior spinal instrumentation and fusion: minimum five-year follow-up. Spine (Phila Pa 1976)
5. 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 (Phila Pa 1976)
6. Hassanzadeh H, Gupta S, Jain A, et al. Type of anchor at the proximal fusion level has a significant effect on the incidence of proximal junctional kyphosis
and outcome in adults after long posterior spinal fusion. Spine Deform
7. Hart R, McCarthy I, O’Brien M, et al. Identification of decision criteria for revision surgery among patients with proximal junctional failure
after surgical treatment of spinal deformity. Spine (Phila Pa 1976)
8. Yagi M, Rahm M, Gaines R, et al. Characterization and surgical outcomes of proximal junctional failure
in surgically treated patients with adult spinal deformity
. Spine (Phila Pa 1976)
9. Nguyen NL, Kong CY, Hart RA. Proximal junctional kyphosis
and failure-diagnosis, prevention, and treatment. Curr Rev Musculoskelet Med
10. Hart RA, McCarthy I, Ames CP, et al. Proximal junctional kyphosis
and proximal junctional failure
. Neurosurg Clin N Am
11. Lowe TG, Kasten MD. An analysis of sagittal curves and balance after Cotrel-Dubousset instrumentation for kyphosis secondary to Scheuermann's disease. A review of 32 patients. Spine (Phila Pa 1976)
12. Lee GA, Betz RR, Clements DH, et al. Proximal kyphosis after posterior spinal fusion in patients with idiopathic scoliosis. Spine (Phila Pa 1976)
13. Hart RA, Prendergast MA, Roberts WG, et al. Proximal junctional acute collapse cranial to multi-level lumbar fusion: a cost analysis of prophylactic vertebral augmentation. Spine J
14. Theologis AA, Burch S. Prevention of acute proximal junctional fractures after long thoracolumbar posterior fusions for adult spinal deformity
using 2-level cement augmentation at the upper instrumented vertebra and the vertebra 1 level proximal to the upper instrumented vertebra. Spine (Phila Pa 1976)
15. Ghobrial GM, Eichberg DG, Kolcun JPG, et al. Prophylactic vertebral cement augmentation at the uppermost instrumented vertebra and rostral adjacent vertebra for the prevention of proximal junctional kyphosis
and failure following long-segment fusion for adult spinal deformity
. Spine J
16. Raman T, Miller E, Martin CT, et al. The effect of prophylactic vertebroplasty on the incidence of proximal junctional kyphosis
and proximal junctional failure
following posterior spinal fusion in adult spinal deformity
: a 5-year follow-up study. Spine J
17. Viswanathan VK, Kukreja S, Minnema AJ, et al. Prospective assessment of the safety and early outcomes of sublaminar band placement for the prevention of proximal junctional kyphosis
. J Neurosurg Spine
18. Buell TJ, Buchholz AL, Quinn JC, et al. A pilot study on posterior polyethylene tethers to prevent proximal junctional kyphosis
after multilevel spinal instrumentation for adult spinal deformity
. Oper Neurosurg (Hagerstown)
19. Bess S, Harris JE, Turner AW, et al. The effect of posterior polyester tethers on the biomechanics of proximal junctional kyphosis
: a finite element analysis. J Neurosurg Spine
20. Lonner BS, Newton P, Betz R, et al. Operative management of Scheuermann's kyphosis in 78 patients: radiographic outcomes, complications, and technique. Spine (Phila Pa 1976)
21. Barrey C, Roussouly P, Le Huec JC, et al. Compensatory mechanisms contributing to keep the sagittal balance of the spine. Eur Spine J
2013; 22: (suppl): S834–S841.
22. Roussouly P, Pinheiro-Franco JL. Biomechanical analysis of the spino-pelvic organization and adaptation in pathology. Eur Spine J
2011; 20: (suppl): 609–618.
23. Laouissat F, Sebaaly A, Gehrchen M, et al. Classification of normal sagittal spine alignment: refounding the Roussouly classification. Eur Spine J
24. Kim HJ, Bridwell KH, Lenke LG, et al. Patients with proximal junctional kyphosis
requiring revision surgery have higher postoperative lumbar lordosis and larger sagittal balance corrections. Spine (Phila Pa 1976)
25. Reames DL, Kasliwal MK, Smith JS, et al. Time to development, clinical and radiographic characteristics, and management of proximal junctional kyphosis
following adult thoracolumbar instrumented fusion for spinal deformity. J Spinal Disord Tech
26. Scheer JK, Lafage R, Schwab FJ, et al. Under correction of sagittal deformities based on age-adjusted alignment thresholds leads to worse health-related quality of life whereas over correction provides no additional benefit. Spine (Phila Pa 1976)
27. Lafage R, Bess S, Glassman S, et al. Virtual modeling of postoperative alignment after adult spinal deformity
surgery helps predict associations between compensatory spinopelvic alignment changes, overcorrection, and proximal junctional kyphosis
. Spine (Phila Pa 1976)
28. Lafage R, Schwab F, Challier V, et al. Defining spino-pelvic alignment thresholds: should operative goals in adult spinal deformity
surgery account for age? Spine (Phila Pa 1976)
29. Lafage R, Schwab F, Glassman S, et al. Age-adjusted alignment goals have the potential to reduce PJK. Spine (Phila Pa 1976)
30. Hostin R, McCarthy I, O’Brien M, et al. Incidence, mode, and location of acute proximal junctional failures after surgical treatment of adult spinal deformity
. Spine (Phila Pa 1976)
31. International Spine Study G, Lafage R, Schwab FJ, et al. Redefining radiographic thresholds for junctional kyphosis pathologies. Spine J
32. Miller EK, Vila-Casademunt A, Neuman BJ, et al. External validation of the adult spinal deformity
(ASD) frailty index (ASD-FI). Eur Spine J
33. Miller EK, Neuman BJ, Jain A, et al. An assessment of frailty as a tool for risk stratification in adult spinal deformity
surgery. Neurosurg Focus
34. Kebaish KM, Martin CT, O’Brien JR, et al. Use of vertebroplasty to prevent proximal junctional fractures in adult deformity surgery: a biomechanical cadaveric study. Spine J
35. Mebane WR Jr, Sekhon JS. Genetic optimization using derivatives: the rgenoud package for R. J Stat Softw
36. Diamond A, Sekhon JS. Genetic matching for estimating causal effects: a general multivariate matching method for achieving balance in observational studies. Rev Econ Stat
37. Lau D, Clark AJ, Scheer JK, et al. Proximal junctional kyphosis
and failure after spinal deformity surgery: a systematic review of the literature as a background to classification development. Spine (Phila Pa 1976)
38. Yagi M, Fujita N, Tsuji O, et al. Low bone-mineral density is a significant risk for proximal junctional failure
after surgical correction of adult spinal deformity
: a propensity score-matched analysis. Spine (Phila Pa 1976)
39. Lafage R, Line BG, Gupta S, et al. Orientation of the upper-most instrumented segment influences proximal junctional disease following adult spinal deformity
surgery. Spine (Phila Pa 1976)