10 Proximal Junctional Deformity

Proximal junctional kyphosis (PJK) is one of the most common complications associated with spinal fusions in adult spinal deformity. PJK came about with the advent of modern fusion and instrumentation techniques, when spine surgeons began to note a progressive imbalance at the junction between fused and unfused segments. First outlined by Lee et al, a case of PJK was initially defined as a patient having kyphosis of 5 degrees greater than normal from T2 to the proximal level of the instrumented fusion,¹ with the reference for “normal” angulation across segments derived from the work of Bernhardt and Bridwell.² Using this definition, 46% of patients with adolescent idiopathic scoliosis (AIS) were found to have developed PJK at a minimum of 2 years postoperatively. Several years later, Glattes et al made a slight modification to the definition of PJK, characterizing this phenomenon as a proximal junctional Cobb’s angle in the sagittal plane greater than or equal to 10 degrees and 10 degrees greater than the preoperative measurement, measuring from the lower endplate of the uppermost instrumented vertebra (UIV) and the upper endplate of vertebra two levels above ( Fig. 10.1).³ In their retrospective study of 81 adults undergoing long posterior spinal fusion for scoliosis or sagittal imbalance, these authors found that after an average follow-up of 5.3 years, 26% of patients had developed radiographic evidence of PJK.

Since this work, others have put forth a variety of definitions of PJK,^1,3,4,5,6 with no current consensus about which definition most accurately represents this process. For example, Helgeson et al defined PJK using only one segment superior to the UIV, positing that the disruption of posterior soft tissues and facet capsule at the level above are likely to play a major role, and that focusing on this superior level (rather than at two levels above) would better isolate the area of interest.⁶ The reported incidence of PJK is quite varied, with the majority of studies reporting PJK incidence in the range of 20 to 40%, using the definition put forth by Kim et al.⁷ Given the significant prevalence of this complication, much effort has been put forth to better understand—and to prevent the occurrence—of this phenomenon. Although the incidence of radiographic PJK following spinal fusion is significant, its clinical implications and subsequent management remain controversial. This chapter presents our current understanding of PJK, including etiology, risk factors, natural history of disease, clinical outcomes, as well as potential prevention strategies.

10.2 Etiology and Risk Factors

Though much effort has been devoted to studying PJK, its underlying etiology is not fully understood. Prior work, however, has highlighted various risk factors which make patients more susceptible to developing PJK. To better understand those variables predisposing patients to developing PJK, we break them down into demographic, radiographic, and surgical risk factors.

10.2.1 Demographic Risk Factors

Age, body mass index (BMI), and bone density have all been identified as potential risk factors for the development of PJK. Age is perhaps the most well-documented risk factor for PJK. In a study on adult spinal deformity by Kim et al, the authors found that those who developed PJK had a significantly older age at operation (p = 0.007), with 69% of patients older than 55 years in the PJK group (p < 0.0001).⁸ Similarly, in a retrospective study of 364 patients with adult scoliosis, Kim et al found that age greater than 60 years posed a greater risk for PJK (p = 0.021),⁹ while in a different study the same group reported significantly older age in patients demonstrating PJK requiring revision (60.1 years) compared to those without PJK (49.9 years; p = 0.03).¹⁰ Finally, Bridwell et al found that patients with PJK greater than or equal to 20 degrees were older (56 vs. 46 years; p < 0.001).¹¹ Although the mechanism is not yet clear, it has been posited that age-dependent disc and facet joint degeneration and weaker spinal extensors in older patients could contribute to these differences.⁸ Other studies, however, have demonstrated no significant difference in PJK with increasing age when accounting for other factors through multivariate analyses.^12,13 In a systematic review of patients undergoing surgery for scoliosis and/or kyphosis, Kim et al concluded that there was a low strength of evidence to suggest age as a potential risk factor for PJK.⁷

Fig. 10.1 (a) Lateral radiograph demonstrating the proximal junctional angle, defined as the angle between the inferior endplate of the upper instrumented vertebra and the superior endplate two levels above. (b) Lateral radiograph of the same patient at 2-week follow-up demonstrating proximal junctional kyphosis, as defined degrees by Glattes et al (proximal junctional angle ≥ 10 degrees and at least 10 degrees greater than the preoperative measurement).

Whereas age is a well-studied risk factor, the association between PJK and BMI is less clear. As with age, Bridwell et al also found that PJK greater than or equal to 20 degrees was associated with significantly higher BMI (p = 0.015) in adult deformity patients.¹¹ Similarly, in their study of AIS, Helgeson et al found that patients with a greater than 2 standard deviation increase in PJK had significantly increased BMI (26.7) compared to those without such an increase (20.9; p = 0.013).⁶ Other studies, however, have failed to demonstrate this association between BMI and PJK.^{3,8,9,10,13,14}

Similarly, equivocal results have been demonstrated for the association between low bone density and PJK development. Yagi et al found a trend in PJK development in adult scoliosis patients with low bone mineral density (BMD) (osteoporosis and/or osteopenia), though this did not reach significance (p = 0.055).¹⁴ In multiple studies, Kim et al have found differing results, with one study demonstrating a significant difference between the proportion of patients with osteoporosis in a PJK versus non-PJK group (20.4 vs. 9.8%; p = 0.016),⁹ and another study demonstrated no significant association between low bone density in PJK after performing multivariate analyses.¹² While not yet fully proven, the association between PJK and low bone density is logical, with experts suggesting that PJK could occur due to the prevalence of compression fractures and loosening of the pedicle screw at the UIV.^4,15,16,17 Indeed, biomechanical studies support the idea that bone density may play a vital role in PJK.^18,19

10.2.2 Radiographic Risk Factors

Preoperatively, sagittal malalignment has been found to have perhaps the greatest association with PJK.^{4,13,15,20,21} From a global alignment perspective, increased preoperative sagittal vertebral axis (SVA) has been demonstrated to show significant association with the development of postoperative PJK, as reported by Annis et al who found preoperative SVA greater than 5 cm to be a risk factor for acute proximal junctional failure (PJF).²⁰ Similarly, the association between PJK and preoperative thoracic kyphosis (TK) has been shown in several studies. Maruo et al, for example, found that preoperative TK greater than 30 degrees was a significant predictor of postoperative PJK.¹⁵ Similarly, in a study of AIS patients, Kim et al found that increased preoperative TK (T5–T12 > 40 degrees) showed a significant association with PJK development (p = 0.015).²² Finally, in their study of adult deformity, Mendoza-Lattes et al found that the average magnitude of TK was significantly larger than lumbar lordosis (LL) at baseline in the group of patients that developed PJK compared to the group that did not (37.3° ± 19.2° vs. 25.9° ± 12.4°, p = 0.044). Preoperatively, the PJK patients had a significantly smaller difference between LL and TK compared to the non-PJK group (−6.6° ± 14.2° vs. 6.6° ± 23.2°; p = 0.012), and demonstrated pelvic retroversion as well as reduced sacral slope preoperatively.¹³

More locally, proximal junctional angle (PJA) at the level of interest has also been associated with the development of PJK in AIS patients, although results are mixed. The work of Lee et al found that preoperative proximal kyphosis greater than 5 degrees from normal from T2 to the proposed UIV was indicative of the need to extend the fusion to a higher level with high sensitivity (78%) and specificity (84%).¹ Nevertheless, other reports demonstrate conflicting evidence. Hollenbeck et al studied patients with AIS and reported no difference in preoperative proximal junctional flexion in those who developed increased junctional flexion, as defined by the angle formed between the posterior wall of the UIV and that of the vertebra two levels proximal.²³ Similarly, Kim et al did not find preoperative proximal junctional flexion to be associated with increased risk in AIS patients, with the UIV level not affecting the incidence of PJK.²²

10.2.3 Surgical Risk Factors

Surgical risk factors for the development of PJK can be broken down into approach-related risk factors, such as maintenance of soft-tissue integrity, fusion construct rigidity, UIV level selection, and overall magnitude of correction.

There is general consensus that intraoperative disruption of the posterior soft tissues is a significant risk factor for the development of PJK, with the thought that violation of the posterior tension band and other intervertebral elements are major contributors to this phenomenon.^{3,7,8,11,14,24,25,26} This belief stems from cadaveric and biomechanical studies. In their study on cadavers, Anderson et al demonstrated the importance of posterior structures in adjacent segment flexion stiffness. In their study, they found incremental reductions in adjacent segment motion stiffness associated with individual procedures (supratransverse process hook, supralaminar hook, pedicle screw placement, or pedicel screw removal), with supraspinous and interspinous ligament transection adding 6.59% flexion stiffness loss alone, and transection of the remaining posterior structures, including facet joints and soft tissues, contributing 44.72% to stiffness loss. They further found that transection of all posterior structures resulted in flexion stiffness loss of 67.61%.²⁷ Similarly, in a biomechanical simulation of various models of adult scoliosis, Cammarata et al found that complete bilateral facetectomy, resection of the posterior ligaments, and a combination of the two predicted increase in proximal junctional kyphotic angle by 10, 28, and 53%, respectively, in addition to increasing the proximal flexion force and moment.²⁸ These findings suggest that disruption of these posterior stabilizing structures is important and that their preservation may reduce the risk of PJK.

In addition to the risk of posterior violation, the combined anterior–posterior approach has been reported as a significant risk factor in several studies. In a retrospective review, Kim et al reported that a combined anterior–posterior approach was the strongest risk factor for the development of PJK ( Fig. 10.2) with an odds ratio of 3.04 (95% confidence interval [CI] 1.56–5.93)¹²; in another retrospective review, combined anterior–posterior fusion demonstrated significantly greater rates of PJK compared to posterior fusion alone (p = 0.041).⁸ Though less is known about the risk of anterior approach alone, it has been reported to be an independent risk factor for PJF.²⁹ Whether this is due to anterior soft tissues, or confounded by other risk factors such as larger corrections or greater construct stiffness has not been clearly delineated.

In addition to approach-related issues, construct rigidity and stiffness have been demonstrated to play an important role in PJK risk. Higher rates of PJK have been shown with pedicle screw use in long-segment instrumentation in AIS^6,22,25 and in adult deformity,³⁰ contributing to the risk both by increasing construct rigidity and due to the greater likelihood of facet disruption.²² Less stiff constructs have been shown to reduce these effects. In their biomechanical simulation of adult scoliosis, Cammarata et al found that, compared with pedicle screws at the UIV, proximal transverse process hooks reduced all biomechanical indices (proximal junctional kyphotic angle, proximal flexion force and moment) by approximately 26%, while the use of reduced diameter proximal transition rods (from 5.5 to 4 mm) decreased these indices as well.²⁸ Thawrani et al similarly found that transverse process hooks demonstrated less stiffness and provided a more gradual transition to normal motion compared to pedicle screws.³¹ Although these findings suggest instrumentation type at the UIV plays a critical role in the development of PJK, clinical studies have not been as consistent.

Fig. 10.2 Risk of proximal junctional kyphosis (PJK) development with combined anterior–posterior approach—cumulative hazard plot for PJK development for combined anterior–posterior approach versus anterior or posterior approach alone. (Adapted from Kim et al.¹²)

Finally, fusion extending to the sacrum has been associated with an increased risk of PJK, as demonstrated by Yagi et al¹⁴ and Kim et al.⁸ In further support of this idea, Bridwell et al reported that PJK greater than or equal to 20 degrees was associated with fusion to the sacrum with iliac screws (p = 0.029),¹¹ while Kim et al found patients with PJK requiring revision had higher rates of fusion to the pelvis (91%), compared to those with PJK not requiring revision (85%) and those without PJK (74%).¹⁰

In addition to construct rigidity, the choice of UIV level has been demonstrated to contribute to the risk of PJK. Studies have demonstrated a greater risk of both PJK¹¹ and PJF^4,21 with fusion to the lower thoracic (LT) spine, compared to the upper thoracic (UT) spine. Interestingly, Annis et al found that T10 specifically demonstrated a significantly higher risk of PJF than adjacent levels (T9, p = 0.03; T11, p = 0.01).²⁰ On the other hand, Kim et al reported opposite results, noting that patients with UIV to T1 through T3 were almost twice as likely to develop PJK compared to those with UIV T4 to T12 (p = 0.034).¹² Finally, equivocal results were demonstrated in a study by Fujimori et al. In patients undergoing fusion from sacrum to either UT or LT spine, at more than or equal to 2 years postoperatively the authors found no significant differences in PJK (0.9 degrees for UT, 2.8 degrees for LT; p = 0.4), asymptomatic PJK (32% for UT, 41% for LT; p = 0.4), or PJK requiring revision (6.4% for UT, 10% for LT; p = 0.6).³² These authors suggest that, while UT segments are more stabilized due to the surrounding rib cage and scapulae, PJK may occur in this region due to stress being distributed over fewer proximal motion segments but that the LT region is more biomechanically vulnerable.³² Which UIV region is more likely to result in PJK remains unclear.

Finally, the magnitude of deformity correction has been associated with PJK, with greater operative correction translating to higher rates of PJK.^{10,13,14,15,20,26} Several studies have found that greater SVA correction leads to higher rates of PJK, with the thought that this may be due to either overcorrection or pelvic retroversion.^10,13,14 Indeed, adults exhibit a progressive increase in positive sagittal balance over the course of a lifetime,^33,34 so correcting postoperative sagittal balance (e.g., to 0 cm) may in fact represent overcorrection, with subsequent PJK occurring as a compensatory mechanism.¹³ In support of this overcorrection theory, Mendoza-Lattes et al report that SVA is a significant predictor of PJK, with every centimeter increase in the C7 plumbline translating to a decrease in PJK by 30%.¹³ Biomechanically, increased deformity correction results in increased forces in the spine. Cammarata et al report that an increase in sagittal rod curvature from 10 to 20, 30, and 40 degrees increased the proximal junctional kyphotic angle (by 6, 13, and 19%, respectively) as well as the proximal flexion force (3, 7, and 10%, respectively) and moment (9, 18, and 27%, respectively).²⁸ Other studies have found a lack of association between postoperative positive sagittal balance and PJK development, further lending support to this notion.^3,8 In addition to SVA, increased LL is a demonstrated risk factor for PJK. In multivariate analyses, two studies found that greater correction in LL was an independent risk factor for PJK (LL > 30 degrees)¹⁵ and PJF.²⁰ Mendoza-Lattes et al found that the greater the difference between LL and TK (i.e., in cases where LL/TK mismatch was addressed), the lower the risk of developing PJK, with every 10-degree difference decreasing the risk of PJK by 140%.¹³ Similarly, Kim et al noted that patients requiring revision after PJK were more likely to have a LL closer to the pelvic incidence, whereas those who did not develop PJK had a lower LL.¹⁰ These results further support the risk associated with overcorrecting the SVA and the importance of considering global alignment when addressing sagittal imbalance. Further studies are needed to define the appropriate spinopelvic parameters to target in these deformity cases.

There are a number of strategies to address modifiable risk factors to limit the likelihood of PJK. As summarized by Lau et al, these are extending fusion to level with kyphosis greater than 5 degrees; decreasing instrumentation stiffness (including use of composite metals, fewer implants, and transverse process hooks/transition rods); preserve more soft tissues at the UIV; and attempt to achieve optimal spinal balance and alignment, taking into account the potential for overcorrection.¹⁷ Although numerous factors have been identified as being associated with the development of PJK, it is important to remember that the vast majority of studies assessing PJK are retrospective in nature and many do not include multivariate analyses, so it is possible—and likely—that one or more of these risk factors is mistakenly identified due to the effects of confounding, which has been reported previously.¹²

10.3 Classification

As PJK is merely a radiographic finding, its existence encompasses a wide range of pathologies, from those with mild kyphosis due to soft-tissue failure to those with more severe Cobb angles from bony or implant failure. As a result, in order to better stratify these patients based on severity of disease, Yagi et al first proposed a classification scheme in 2011,¹⁴ which has since been updated to include spondylolisthesis.²⁶ In their initial scheme, the authors categorized patients based on grade (amount of kyphosis) and type (underlying etiology of failure) as follows: proximal junctional increase 10 to 14 degrees (grade A), proximal junctional increase 15 to 19 degrees (grade B), proximal junctional increase greater than or equal to 20 degrees; PJK from disc/ligamentous failure (type 1), bone failure (type 2), implant/bone interface failure (type 3). According to this classification, a majority of patients were found to be grade A (56%, 18/32 patients) and type 1 (81%, 26/32).¹⁴ While more detailed, this classification nevertheless still relies only on radiographic findings without consideration of clinical manifestations, which are similarly diverse.

Clinically speaking, many have found that a significant number of patients with PJK are asymptomatic, making a classification scheme considering only radiographic components less useful. In the study by Yagi et al, for example, only 6/32 (19%) of patients who had developed PJK were symptomatic. Perhaps to address this question, these authors further refined their scheme to stratify those with PJF, defined as symptomatic PJK requiring surgery. From their original scheme, they modified the grades (proximal junctional increase 10 to 19 degrees, grade A; proximal junctional increase 20 to 29 degrees, grade B; proximal junctional increase greater than or equal to 30 degrees, grade C) and added a component based on the presence/absence of spondylolisthesis above the UIV ( Table 10.1). They found that the most common form of PJF was type 2N (bony failure, no spondylolisthesis above UIV), whereas the most debilitating form of PJF was observed most commonly in type 2S (bony failure, spondylolisthesis above UIV).²⁶ While this updated classification is straightforward and better predicts which patients are likely to become symptomatic, it is neither prognostic regarding the course/severity of disease nor does it help guide treatment.¹⁷

Table 10.1 Classification of PJK/PJF by Yagi et al^26,35

In an attempt to improve upon these shortcomings, the International Spine Study Group (ISSG) put forth a new classification taking into account PJK etiology, radiographic parameters, symptoms, and disease severity. As reported by Lau et al, these authors proposed a numeric scale including six components: neurological deficit, focal pain, instrumentation problem, change in kyphosis/posterior ligament complex integrity, fracture location, and level of UIV ( Table 10.2). The proposed classification reportedly has good reliability and repeatability and correlates strongly with recommended treatment, with pain, kyphosis, neurological status, and instrumentation failure as the strongest predictors of the need for revision surgery.¹⁷ Further research is needed to validate this classification scheme.

Table 10.2 Proposed PJK classification and severity scale

PJK typically occurs in the acute or subacute postoperative period. In a retrospective study of adult spinal deformity over a minimum of 5 years, Kim et al found that the majority (59%) of those developing PJK demonstrate a significant increase in PJA within the first 8 weeks postoperatively.⁸ Similarly, Yagi et al reported that the majority (75%) of PJK is identified within 3 months postoperatively,³⁵ while Wang et al found that 80% would develop PJK within 18 months postoperatively.³⁶

While evidence of PJK can often be seen in this acute/subacute postoperative period, it tends to progress over years. Kim et al noted that 35% of patients with PJK continued to show significant increases in PJA from 2 years postoperatively to final follow-up.⁸ Similarly, Yagi et al reported that only half (53%) of the total average PJ angle increase occurred by 3 months postoperatively, with the other half occurring by 2 years postoperatively, with average PJ angle increasing from 1.2 to 14.9 (2 years) to 18.5 degrees (final follow-up, minimum 5 years) among patients with PJK³⁵ ( Fig. 10.3). While these patients tended to progress over several years, they found no patients progressing after 5 years postoperatively.

Though PJK is common and patients tend to progress postoperatively, the clinical relevance of these radiographic findings is controversial. Many studies have focused on clinical outcomes and symptoms associated with PJK and have found mixed results. Most agree that PJK does not commonly carry significant clinical consequences. In prior studies of adult deformity, Yagi and colleagues found a PJK incidence of 22%, whereas they report a rate of symptomatic PJK of only 4%,¹⁴ and a rate of symptomatic PJK requiring surgery of only 1.4%.²⁶ Similarly, several studies have demonstrated that PJK has not been shown to correlate with postoperative patient-reported outcomes. Glattes et al, in their study of PJK after long instrumented posterior fusion, reported no difference in Scoliosis Research Society-24 (SRS-24) scores between those with PJK and those without at a minimum of 2 years of follow-up³; Yagi et al studied postoperative SRS-22r and Oswestry Disability Index (ODI) in adult deformity cases and found no significant differences between the PJK and non-PJK groups³⁵; and Kim et al noted similar findings in their study of patients undergoing combined anterior–posterior surgery, with no difference in SRS-22 scores between the PJK and non-PJK groups.¹²

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Clinical Presentation of Thoracic Spinal Compression
2. Neuromuscular Scoliosis
3. Intramedullary Spinal Cord Tumors
4. Intraoperative Neuromonitoring During Thoracic Spine and Spinal Cord Surgery

Stay updated, free articles. Join our Telegram channel

Join