Proximal Junctional Kyphosis Following Spinal Deformity Surgery in the Pediatric Patient

Abstract

Proper understanding and restoration of sagittal balance is critical in spinal deformity surgery, including conditions such as adolescent idiopathic scoliosis and Scheuermann kyphosis. One potential complication following spinal reconstruction is proximal junctional kyphosis. The prevalence of proximal junctional kyphosis varies in the literature, and several patient- and surgery-related risk factors have been identified. To date, the development of proximal junctional kyphosis has not been shown to lead to a negative clinical outcome following spinal fusion for adolescent idiopathic scoliosis or Scheuermann kyphosis. Treatment options range from simple observation in asymptomatic cases to revision surgery with extension of the fusion proximally. Several techniques and technologies are emerging that seek to address and prevent proximal junctional kyphosis.

Proximal junctional kyphosis (PJK) is a postoperative radiographic phenomenon observed in the sagittal plane following instrumentation and fusion for spinal deformity. In a broader sense, PJK is a form of adjacent segment pathology associated with spinal fusion procedures that is most often seen after correction of scoliosis or kyphosis. In pediatric patients, PJK often manifests as a kyphotic change in the disk space above the fusion (Figure 1). In addition to the disk change, PJK in adults also includes fractures, subluxations, and long sweeping kyphosis above the fusion.^1-7 Surgeons need to be alert for this common postoperative finding.

Figure 1: Proximal junctional kyphosis in Scheuermann kyphosis. In pediatric patients, most of the junctional change occurs as a result of kyphosis through the intervertebral disk above the fusion. PJA = proximal junctional angle

Radiographic Definition

In a 1999 study by Lee et al,⁸ the authors systematically studied PJK in 69 patients with adolescent idiopathic scoliosis (AIS) who had undergone posterior spinal fusion (PSF) with a minimum 2-year follow-up. In their study, abnormal PJK was defined as kyphosis from T2 to the proximal level of the instrumented fusion ≥5° above the summed normal angular segments, as reported by Bernhardt and Bridwell.⁹ In 1989, Bernhardt and Bridwell⁹ had published a study analyzing the sagittal plane alignment of thoracic and lumbar spines in normal persons (Figure 2). In addition to reporting on regional Cobb measurements of thoracic kyphosis, thoracolumbar junction, and lumbar lordosis, the authors had determined segmental angulation at each level from T1-T2 to L5-S1. Lee et al⁸ used these segmental angulation values in their analysis and defined abnormal as anything measuring ≥5° from the predicted value. The authors used 5° as the cutoff value because they sought to determine whether the development of PJK following PSF could be predicted on the basis of original preoperative radiographs. The authors also demonstrated that preoperative junctional kyphosis of ≥5° above the proposed proximal instrumented vertebra showed the highest sensitivity (78%) and specificity (84%).

Figure 2: Adjusted mean segmental sagittal angulations of normal thoracic and lumbar curves from T1-T2 to L5-S1. (Reproduced with permission from Bernhardt M, Bridwell KH: Segmental analysis of the sagittal plane alignment of the normal thoracic and lumbar spines and thoracolumbar junction. Spine 1989;14:717-721.)

Subsequently, Glattes et al¹⁰ redefined PJK as a proximal junctional sagittal Cobb angle between the lower end plate of the upper instrumented vertebra (UIV) and the upper end plate of the two supra-adjacent vertebrae of ≥10° and at least 10° greater than the preoperative measurement (Figure 3). The change in definition may be based on the fact that the difference in radiographic measurements had to be 10° for scoliosis and 11° for kyphosis to be 95% confident.¹¹ To date, this is the most frequently used definition of PJK in the literature.

Figure 3: A, Preoperative proximal junctional sagittal Cobb measurement between the lower end plate of the upper instrumented vertebra and the upper end plate of the two supra-adjacent vertebrae measuring +15°. B, Final follow-up (7 years + 10 months) proximal junctional sagittal Cobb measurement demonstrating proximal junctional kyphosis measuring +27°. (Reproduced with permission from Glattes RC, Bridwell KH, Lenke LG, Kim YJ, Rinella A, Edwards C: Proximal junctional kyphosis in adult spinal deformity following long instrumented posterior spinal fusion: Incidence, outcomes, and risk factor analysis. Spine 2005;30:1643-1649.)

Helgeson et al¹² proposed another method to define PJK. In their study, the proximal junctional angle was measured from the caudal end plate of the UIV to the cephalad end plate of the one, instead of two, vertebra above the UIV (Figure 4). The authors noted that the previous definition of PJK is somewhat arbitrary and does not have any physiologic basis. One potential cause of PJK is disruption of the soft tissues, including the facet capsule and the interspinous ligament above the instrumented level; thus, the authors’ method better isolates this area for the purposes of their analysis. They chose two standard deviations above the mean (ie, calculated to be 15°) as being abnormal and redefined PJK as any postoperative kyphosis increase ≥15°.

Figure 4: Proximal junctional kyphosis Cobb measurement from the caudal end plate of the upper instrumented vertebra to the cephalad end plate of the one, instead of two, vertebra above the upper instrumented vertebra. (Reproduced with permission from Helgeson MD, Shah SA, Newton PO, et al: Evaluation of proximal junctional kyphosis in adolescent idiopathic scoliosis following pedicle screw, hook, or hybrid instrumentation. Spine 2010;35:177-181.)

In a study of 19 patients with scoliosis by Sacramento-Dominguez et al,¹³ the authors evaluated the reproducibility of measuring the Cobb angle of PJK using either the first or the second vertebra above the UIV. Intra-surgeon concordance correlation coefficients ranged from 0.78 to 0.92 (high to very high reproducibility), and comparative inter-surgeon concordance correlation coefficients ranged from 0.55 to 0.80 (moderate to high reproducibility). The authors stated that no conclusions could be drawn as to which of the two vertebrae is the better option.

There is currently no consensus on the precise definition of PJK. Although several studies have provided reasons as to why certain values were used in each analysis, none of these values has been validated. Furthermore, there appears to be no clear correlation between the radiographic development of PJK and the clinical outcome following surgery in children, at least for milder radiographic forms of PJK.

Adolescent Idiopathic Scoliosis

Prevalence

Lee et al⁸ first reported a 46% prevalence of PJK in their analysis of 69 patients with AIS who underwent PSF with hook/rod instrumentation using 5° as the cutoff. Only one patient underwent revision surgery for cosmetic reasons; the result was a satisfactory outcome. Rhee et al¹⁴ retrospectively analyzed 110 consecutive patients with AIS who had undergone either anterior or posterior spinal fusion with a minimum 2-year follow-up. For posterior instrumentation, a 5.5-mm rod-based system was used with a combination of screws, hooks, and wires. Hooks were used most commonly at the proximal end of the construct. Of the patients treated with anterior and posterior fusion with instrumentation, 17% and 35% had radiographic PJK, respectively, although no patient required revision surgery. The authors attributed the greater incidence of PJK following posterior instrumentation to dissection of the soft tissue that functions as a tension band during surgery, compression during instrumentation (eg, proximal claw constructs) or correction of proximal thoracic scoliosis, and reciprocal kyphosis in response to a reduction in thoracic kyphosis associated with posterior instrumentation.

Kim et al¹⁵ evaluated 193 consecutive patients with AIS with a minimum 5-year follow-up. Hooks were used in 137 patients, whereas hybrid constructs (ie, proximal hooks and distal pedicle screws) were implanted in 56 patients. In their series, PJK developed in 26% of patients at final follow-up (mean 7.3 years). Furthermore, most patients (82%) had evidence of PJK by the second year following surgery. For patients with PJK at final follow-up, the average increase in the junctional sagittal Cobb angle was 23.9°. In another report, Kim et al¹⁶ analyzed PJK in 410 consecutive patients with AIS with a minimum 2-year follow-up. The patients underwent surgeries that used three different types of posterior segmental instrumentation (ie, hooks-only, hybrid with proximal hooks and distal pedicle screws, and all-pedicle screws). In this study, prevalence was 27%, and patients with PJK had a mean increase of 22° in the proximal junctional angle. In terms of timing, 30% of patients with PJK had this pathology at 8-week follow-up, whereas most patients (70%) had this pathology at the 2-year follow-up.

Hollenbeck et al¹⁷ reported on 174 consecutive patients with AIS with a mean 4.9-year follow-up; PJK was found in 9.2% of the study cohort. With a preoperative Cobb angle of 3.7°, the final Cobb angle was 17.8°, with a net increase of 14.1°. Wang et al¹⁸ retrospectively reviewed 123 patients with AIS who were treated with PSF. Thirty-five patients (28%) experienced PJK. PJK occurred in 11 patients (31%) within 6 months after surgery, in 17 patients (46%) between 6 and 18 months after surgery, and in 7 patients (23%) at ≥18 months after surgery. Using a multi-institutional database, Helgeson et al¹² evaluated PJK in 283 patients with AIS following pedicle screw, hook, or hybrid instrumented fusion with a minimum 2-year follow-up. Using 15° as a cutoff value for PJK, the authors found the prevalence of PJK to be zero (hooks-only), 2.5% (hybrid), 8.1% (all screws), and 5.6% (screws with proximal hooks).

Clinical Outcome

Radiographic development of PJK is not uncommon but correlates poorly with clinical outcome, potentially because of the mild degree of kyphosis inherent in the current working definition. In the original work of Lee et al,⁸ the authors reported a 46% prevalence of PJK but only one patient required extension of fusion proximally. The authors did not report validated health-related quality of life (HRQOL) measures in their report. Rhee et al¹⁴ also did not include HRQOL data in their analysis, but the authors reported that no patient needed revision for PJK despite a 35% and 17% incidence of PJK following posterior and anterior fusions, respectively.

Kim et al¹⁵ were the first authors to report Scoliosis Research Society (SRS) outcomes data for patients with AIS. The PJK group had a total score of 98 (of 120), whereas the non-PJK group had a score of 93 (P = 0.14). There were no differences in the pain and self-image SRS subscores. In a study analyzing three different types of instrumentation, Kim et al¹⁶ reported similar SRS scores between PJK and non-PJK patients (totals 97 and 95 of 120, respectively). Pain (P = 0.16), self-image (P = 0.54), function (P = 0.39), and satisfaction (P = 0.20) subdomains in the SRS questionnaire were also similar between the two groups. None of the patients with PJK required revision surgery. Two potential explanations for these findings were provided: (1) young patients usually do not have significant disk degeneration that can cause pain in the spine, or (2) the SRS questionnaire was not sensitive enough to evaluate PJK.¹⁵ Hollenbeck et al¹⁷ also reported no differences in SRS scores between the normal and the PJK groups. In their series, no patient required reoperation because of PJK.

Risk Factors

Although not considered a formal risk analysis, Lee et al⁸ found that a preoperative one-level junctional kyphosis of ≥5° above the proposed UIV was both sensitive and specific for the development of PJK. Rhee et al¹⁴ compared anterior instrumentation with posterior instrumentation and noted a significant association between posterior instrumentation and greater increases in PJK.

In a 5-year follow-up study, Kim et al¹⁵ identified three significant risk factors: thoracoplasty, preoperative thoracic hyperkyphosis, and hybrid instrumentation (ie, proximal hooks and distal pedicle screws). The authors noted that a hybrid construct, compared with an all-hook construct, may be more rigid and cause junctional stress. Thoracoplasty disrupts the spine-chest cage unit and may place additional stress on the instrumentation; in turn, the authors theorized that this additional stress on the instrumentation had a negative impact at the proximal junction. Preoperative thoracic kyphosis ≥40° was also identified as a significant risk factor. The authors observed that patients with thoracic hyperkyphosis undergo substantial reduction of thoracic kyphosis during PSF with instrumentation, and such an acute change in the sagittal alignment may lead to a postoperative reciprocal kyphosis at the proximal junction.

Kim et al¹⁶ found other risk factors associated with PJK when they analyzed three different types of posterior instrumentation used in patients with AIS. The authors reported that the all-pedicle screw group had the highest prevalence of PJK (35%) compared with the hook-only (22%) and hybrid (29%) groups, with a trend toward significant difference (P = 0.058). This finding was later corroborated by Helgeson et al;¹² the authors reported a significantly higher prevalence of PJK in the all-screw group compared with the hybrid (P = 0.02) and the all-hook (P = 0.014) groups. In addition to findings of preoperative thoracic hyperkyphosis (ie, ≥40°) as a risk factor, the authors discovered that simply reducing thoracic kyphosis by ≥5° can lead to PJK. Male gender was identified as a risk factor without a clear explanation.

Wang et al¹⁸ found several risk factors, some of which were similar to those identified by Kim and colleagues.^15,16 These risk factors include ≥10° intraoperative decrease in thoracic kyphosis, thoracoplasty, the use of pedicle screws at the UIV, autogenous bone graft, and distal fusion below L2. The authors reported that the use of autogenous bone graft may have resulted in earlier fusion, resulting in stronger centralization of stress and subsequent induction of PJK. They also noted that performing the fusion below the L2 vertebra significantly reduced the number of motion segments below the fusion for self-adjustment of the trunk following surgery to achieve overall balance; this action may have resulted in transferring the stress proximally to cause PJK. In contrast to the stated studies, Hollenbeck et al¹⁷ analyzed 15 risk factors but found none of them to be significant.

Scheuermann Kyphosis

Prevalence

The reported prevalence of PJK after surgical correction of Scheuermann kyphosis (SK) varies widely. Lowe¹⁹ used Luque sublaminar wiring to treat SK and reported that this technique may increase the risk of PJK. Sturm et al²⁰ analyzed 30 patients who received Harrington rods; PJK was found in 7%, although what constitutes PJK was not clearly defined in the study. Lowe and Kasten²¹ reviewed 32 patients treated with Cotrel-Dubousset instrumentation and reported a 31% prevalence of PJK; however, parameters for PJK were not clearly defined. Papagelopoulos et al²² noted PJK in 1 of 21 patients (4.8%). Hosman et al²³ retrospectively evaluated 33 patients with SK who were treated with segmental instrumentation with a mean follow-up of 4.5 years; PJK developed in one patient (3%).

Lim et al²⁴ conducted a retrospective review on 23 consecutive patients with SK who underwent surgery with posterior segmental instrumentation. PJK was not clearly defined in the study. However, the authors noted that PJK ≥10° was seen in 13% of patients. At the final follow-up, none of the patients had pain attributable to PJK.

Lonner et al²⁵ performed a retrospective multicenter review consisting of 78 patients with SK who underwent surgical treatment. PJK ≥10° occurred in 32.1% of patients. The preoperative proximal junctional Cobb angle was 1.0°, increasing to 21.5° at final follow-up. Denis et al²⁶ performed a single-center retrospective case review analyzing 67 patients who underwent PSF for SK. In their series, PJK occurred in 30% of patients, with a mean proximal junctional angle of 20°.

Clinical Outcome

Currently, no studies report HRQOL measures for SK patients treated with surgical correction. Lonner et al²⁵ reported a 32.1% prevalence of PJK in their retrospective review of 78 patients. Only 1 of 25 PJK patients developed symptoms and underwent reoperation. Denis et al²⁶ reported a 30% prevalence in a study of 67 patients with SK. Similarly, 4 of 20 patients with PJK required revision surgery. Koller et al²⁷ reviewed a consecutive series of 111 patients with 2-year follow-up and found that an increased junctional kyphosis angle led to a higher revision rate (12%), predominantly the result of implant failure and pseudarthrosis rather than junctional kyphosis itself.

Risk Factors

Lonner et al²⁵ did not offer specific risk factors other than noting that PJK was more likely to occur postoperatively in patients with larger kyphosis. Conversely, Denis et al²⁶ identified several risk factors for PJK following surgery for SK. The authors noted that the most common cause of PJK was fusion constructs that did not include the proximal end vertebra. Disruption of the junctional ligamentum flavum also caused PJK. However, PJK did not develop in any of the patients who underwent PSF that incorporated the proximal end vertebra and included preservation of the junctional ligamentum flavum. The most common cause of inappropriate upper end vertebra selection was poor radiographic visualization of the proximal thoracic spine, highlighting the importance of obtaining high-quality radiographs for preoperative planning and paying particular attention to the ends of the fusion construct. Koller et al²⁷ noted that PJK bore a direct relationship with global spinopelvic morphology (ie, pelvic incidence) and that overcorrection of thoracic kyphosis can lead to junctional issues if unfused proximal thoracic and lumbar spinal segments cannot adjust for the global change brought on by surgical treatment. Interestingly, most of the fusion constructs stopped at T4 (36%) or T5 (32.4%) as the UIV, but this action did not have an impact on PJK postoperatively.²⁷

Suggestions to Minimize Proximal Junctional Kyphosis

Conducting detailed preoperative planning and understanding individual patient characteristics are paramount in minimizing the risk of PJK postoperatively. Careful analysis of a patient’s sagittal alignment and achieving proper global balance with surgical correction have been shown to be critical factors. For example, overcorrection of regional thoracic kyphosis in SK needs to be avoided in the setting of high pelvic incidence because the unfused proximal thoracic and lumbar segments may not be able to adjust for postoperative changes to achieve appropriate overall alignment.²⁷ Maintenance or induction of thoracic kyphosis in AIS appears to reduce the risk of PJK.²⁸

Proper selection of the UIV is important in SK. A consensus appears to exist in recommending inclusion of the proximal end vertebra to the fusion construct.^22,26 In contrast, choosing the UIV is more a function of achieving shoulder balance in AIS, rather than preventing PJK. Several guidelines in the literature discuss this topic, which is beyond the scope of this article.²⁹

Surgical approach plays a role in the development of PJK, and the posterior approach has been identified as a risk factor in several studies.¹⁴ Currently, PSF is the more common and preferred technique, and as such, inherently carries a higher risk of PJK compared with anterior spinal fusion. During surgery, careful dissection and minimizing damage to the soft tissue (eg, facet joint capsule, interspinous ligament) that behaves as a tension band are crucial. In a recent biomechanical study, sequential dissection of facets and posterior ligaments led to an increase in the proximal kyphotic angle by 10% to 53%.³⁰ Some surgeons minimize applying cantilever forces at the top of the construct to preserve the natural kyphosis present in the proximal thoracic spine. Other clinicians use hooks instead of pedicle screws as a “soft landing.” This technique appears to reduce rates of PJK in patients with AIS, unlike the situation in adults.^12,16 The use of proximal transverse process hooks instead of pedicle screws can reduce junctional stress by 26%.³⁰ Although currently not in widespread clinical use, transition rods may potentially distribute stress more gradually to the unfused proximal spine and reduce rates of PJK.^30,31 Minimally invasive surgery has gained immense popularity. Literature supporting its use in spinal deformity surgery is scant.³² However, minimally invasive surgery disrupts the posterior soft tissue significantly less than the traditional open procedures and may play a significant role in the future.

Summary

PJK is a common form of adjacent-level pathology following surgical fusion for spinal deformity. To date, there is no consensus on the radiographic definition of PJK among surgeons and researchers in the spinal deformity community. The most commonly used definition is a proximal junctional sagittal Cobb angle between the lower end plate of the UIV and the upper end plate of the two supra-adjacent vertebrae of ≥10° and at least 10° greater than the preoperative measurement.¹⁰ Despite its prevalence, radiographic development of PJK does not seem to correlate with poor clinical outcome, possibly because most radiographic PJK seen is mild and asymptomatic. Several surgery-related, radiographic, and patient risk factors have been identified in the literature. Careful and detailed preoperative planning and surgical execution may reduce PJK rates in patients with pediatric deformities. Long-term consequences of radiographic development of PJK following spinal fusion for deformity in children remain to be seen and need further investigation in the future.