Surgical intervention for an adult spinal deformity (ASD) is mainly aimed at correcting the spinopelvic malalignment and subsequently restoring the global sagittal balance [1]. To achieve spinal balance, surgical long-segment fixation to the pelvis is often necessary [2,3]. However, long-segment spinal fixation, particularly to the pelvis, is associated with the potential risks of postoperative complications including pseudoarthrosis [4], implant failure [5,6], and adjacent problems [7,8]. Proximal junctional kyphosis (PJK) is an important concern with ASD surgery [7,8]. PJK following ASD surgery is reported in approximately 30% of cases [8] and may result in poor clinical outcomes [9] or neurological problems [10,11]. Several reports found that possible risk factors were long and rigid fusion constructs and pedicle screw (PS) instrumentation techniques [11,12], upper instrumented vertebrae (UIV) [13,14], integrity of the posterior soft-tissue tension band [15,16], magnitude of the correction in the sagittal plane [17,18], combined anteroposterior (AP) spinal surgery [19], and osteoporosis [20].

Some previous reports that targeted adolescent idiopathic scoliosis described the possibility of the implant type at UIV in spinal deformity surgery as a factor in the development of PJK [21,22]. To avoid PJK, it is thought that the type of UIV implant may be important; however, few comparative clinical studies have evaluated the incidence of PJK according to the type of UIV implant in ASD surgery. The aim of this study is to compare the incidence of PJK and changes in the proximal junctional angle (PJA) following ASD surgery using two types of UIV implants: transverse process hooks (TPHs) and PSs.

The medical records of 39 consecutive patients (two men and 37 women) with ASD who underwent corrective surgery by a single surgeon at a single institution and were followed up for at least 2 years were retrospectively reviewed. The mean age at the time of surgery was 67.4 years (range, 42–83 years), and the average follow-up period was 42.9 months (range, 24–91 months). All patients complained of multiple years of lower back pain that was not alleviated with conservative treatment. T9 UIV was selected in 13 patients and T10 UIV in 26 patients. Spinopelvic fusion was performed in all patients using iliac or sacral-alar-iliac screws. PSs as UIV implants were used in the first 22 patients until 2013, and TPHs were used in the next 17 patients from 2013 onward. Sublaminar taping with two polyethylene tapes at UIV was also used in all patients. Augmentation by sublaminar taping at UIV was an effective method of preventing the pull-out of spinal instruments, including TPH and PS. It also added semirigid stabilization by holding the entire lamina. Among the 39 patients, 10 patients underwent three-column osteotomy (pedicle subtraction osteotomy or vertebral column resection).

The following radiographic parameters were assessed with standing AP and lateral plain radiography immediately prior to surgery (PreO), 2 weeks after surgery (PO), and 2 years after surgery (PO2Y): degree of thoracic kyphosis (TK, T5–T12), degree of thoracolumbar kyphosis (TLK, T10–L2), degree of lumbar lordosis (LL, T12– S1), pelvic tilt (PT), pelvic incidence (PI), sagittal vertical axis (SVA), Cobb angle of the lumbar curve (L-curve), and coronal vertical axis. Each patient’s curve flexibility was also evaluated with side bending (L-curve flexibility) films in the coronal plane and with fulcrum backward bending (FBB) films in the sagittal plane [23]. PJA was measured between the caudal endplate of UIV to the cephalad endplate located two vertebrae proximal, as previously described by Glattes et al. [24].

PJK was defined by two criteria: (1) UIV, UIV+1, and UIV−1 vertebral fracture, and (2) change in PJA (dPJA) of at least 20° compared with the preoperative measurement. Using all data, two comparative analyses were performed: (1) between PJK and non-PJK groups and (2) between TPH and PS groups.

Clinical outcomes were evaluated using the Japanese editions of the Scoliosis Research Society-22 Outcomes Questionnaire (SRS-22) and the Medical Outcomes Study Questionnaire Short-Form 36 (SF-36) administered at PreO and PO2Y. All data were statistically analyzed using analysis of variance, Mann–Whitney U-test, or paired or unpaired t-test as appropriate (JMP ver. 11.0; SAS, Cary, NC, USA). For all tests, p<0.05 was considered statistically significant.

This study was approved by the ethics committee of Osaka City General Hospital (approval no., 1305011).

PJK is a recognized complication following ASD surgery and has recently attracted increased attention [7-11,16-18]. PJK had been defined as a change in PJA of >10° [24]. However, some studies have indicated that these changes do not affect clinical outcomes. Bridwell et al. [25] evaluated PJK following ASD surgery using PJA >20° and found that the clinical outcomes were worse in patients with PJK. Yagi et al. [10,18] classified PJK into three types: ligamentous failure, bone failure, and implant/bone interface failure. Therefore, in the present study, changes in PJA of >20° and the presence of UIV or UIV±1 fracture were used to define PJK. The incidence of PJK in the current study was 24% (9/37), and eight of the nine PJK patients showed bone failure, including UIV fracture and UIV−1 fracture. Six UIV fractures occurred within 4 weeks after surgery. Several previous reports have cited PJK incidence ranging from 20% to 40% [8]. Yagi et al. [10] reported that 76% of PJK occurred within 3 months after surgery [10], and Maruo et al. [17] reported that 62% of PJK occurred within 8 weeks after surgery [17]. They also stated that UIV fracture was the most common mechanism of PJK. Our results approximately matched their results. Therefore, we believe that UIV fracture is the most common etiology of PJK and that it occurs relatively early after ASD surgery.

Several risk factors of PJK, including older age [11,17,18], spinopelvic fusion [6,18,26], magnitude of sagittal imbalance [10], magnitude of TK [17], and a greater PI–LL change [17] have been previously reported. Recently, Scheer et al. [26] of the Invasive Species Specialist Group reported a computer-based preoperative predictive model for proximal junctional failure (PJF). The seven strongest predictors of PJF are age, lowest instrumented vertebra, preoperative SVA, UIV implant type, UIV, preoperative PT, and preoperative PI–LL. Our comparative analysis indicates that older age (p=0.01) and fracture risk defined by FRAX (p=0.04) are possible risk factors of PJK. Preoperative SVA (p=0.11) shows some tendency in that regard.

According to previous reports, the impact of PJK on quality of life is variable [9,18]. However, the severity of PJK may vary, and PJF may result in compromised structural integrity and neurological deficit, with severe cases potentially needing revision surgery [10,11]. Unlike PJK, patients with PJF have a worse clinical course. The current comparative analysis of health-related quality of life between PJK and non-PJK shows that PJK has no significant impact on surgical outcomes. It may be because none of the patients had a neurological deficit and radiographic measurements improved in all patients with PJK.

Preventing PJK in patients who undergo ASD surgery is one of the most important issues for successful correction surgery and improvement in the postoperative quality of life [26]. As described, PJK is a multifactorial phenomenon that may occur after long-segment fixation surgery. Therefore, there is no definitive solution for preventing PJK, although there are some prophylactic procedures. Some reported strategies are the treatment of osteoporosis with teriparatide [20]; surgical intervention including cement augmentation [13], UIV anchors [21,22], and percutaneous screws [27]; transitional rods [28]; UIV selection [13-15]; and preservation of the posterior ligamentous complex [16].

Among these strategies, we were particularly interested in the types of UIV anchors. The comparative analysis of TPH and PS as UIV anchors indicates no significant difference in the incidence of PJK. However, changes in PJA in patients with PJK were significantly more common in the PS group (Fig. 1) than in the TPH group (Fig. 2). The magnitude of vertebral body collapse following vertebral fracture was more severe in the PS group. Watanabe at al. [11] reported proximal junctional vertebral fracture following ASD surgery that used PS constructs. Additionally, a comparative study of adolescent idiopathic scoliosis surgery indicated significant advantages for TPH over PS fixation to prevent PJK [21,22]. In a biomechanical study, TPH fixation produced a significant reduction of flexion movement compared to PS fixation [15]. Our results were consistent with these reports. We speculate that the mechanism of PJK using PS as the UIV anchor may be as follows. In the patient with osteoporosis, PS as the UIV anchor may directly damage the vertebral body, allowing subsequent vertebral fracture. If a fracture occurs, PS may break the upper endplate and penetrate the disc space, leading to severe collapse of UIV. Conversely, TPH as the UIV anchor may be a better choice because it provides a soft landing at the transitional segment of the long-segment fixation construct [29]. There are some limitations to our study. Our study included a small number of cases and had a short-term follow-up period, and this was not a biomechanical study. Future work should involve a longer follow-up period and a larger sample size.

Our results show that using TPH as a UIV implant may not prevent PJK; however, using TPH as the UIV anchor may prevent vertebral collapse if cases of UIV fracture. In this study, the increased risk of collapse with UIV fracture in the PS group may have been the result of a higher mechanical load on UIV when using PS fixation. Further biomechanical studies should be performed to clarify our results.