INTRODUCTION
Adult spinal deformity (ASD) refers to a spectrum of pathologies that primarily affect thoracolumbar spinal alignment as patients age [1]. ASD encompasses various conditions, including scoliosis, sagittal malalignment, kyphosis, spondylolisthesis, rotatory subluxation, and axial plane deformity, with degenerative scoliosis being the most prevalent form [2]. The prevalence of scoliosis increases with age, showing an almost linear rise from the fifth to eighth decade of life [3]. Notably, Schwab et al. reported an ASD prevalence of 68% in individuals older than 60 years [4].
With the global trend of population aging, the prevalence of ASD has increased and has become an area of growing interest [5]. Although nonoperative management is often the first-line approach, evidence supporting its effectiveness remains relatively limited [6]. Surgical correction of ASD typically requires long-segment fusion combined with complex procedures such as pedicle subtraction osteotomy or vertebral column resection [1,2,5]. Despite the complexity of these procedures, prior studies have shown that surgical treatment provides superior pain relief and functional improvement compared with nonoperative care [6-9]. In recent years, the frequency of surgical interventions has increased, accompanied by a rise in complex procedures and hospital admissions [10,11].
Although surgical correction offers advantages over nonoperative management, it is associated with high complication and morbidity rates, particularly in older adults [12]. Among these complications, proximal junctional kyphosis (PJK) and proximal junctional failure (PJF) are well-recognized, especially after long-segment spinal fusion. Glattes et al. [13] defined PJK as a proximal junctional angle of at least 10° or more from the preoperative measurement. Hostin et al. [14] defined PJF as a structural complication at or near the upper instrumented vertebra (UIV), including vertebral fracture, posterior ligamentous disruption, instrumentation failure, or the need for proximal extension of the fusion within 6 months postoperatively. However, numerous alternative definitions have been proposed by different authors [15-17].
Despite variability in definitions, PJK and PJF are generally considered part of a spectrum of proximal junctional complications following ASD correction surgery [18]. PJK occurs in approximately 20%–40% of surgically treated patients [13,19,20], while PJF occurs in 2%–18% [14,20-22]. PJF is often regarded as a severe, symptomatic form of PJK, commonly presenting with pain, neurological deficits, and reduced quality of life [18,19,21-23]. These complications also substantially increase the likelihood of revision surgery [21,24], which can be particularly burdensome for older patients with multiple comorbidities [12,25].
Given these risks, prevention of PJK and PJF is often more critical than treatment, especially in older or medically complex patients. Various surgical techniques, including hooks, tethers, and prophylactic vertebroplasty, have been introduced to reduce the incidence of proximal junctional complications [26-28]. However, these techniques remain in the early stages of clinical adoption, and robust evidence is lacking, as most available studies are retrospective. Additionally, there is substantial variability in surgical strategies, with no established consensus on optimal preventive methods.
In this meta-analysis, we focused on surgical techniques proposed to prevent PJK and PJF in ASD correction or long-segment fusion surgery. Specifically, we evaluated the effects of tethers, hooks, prophylactic vertebroplasty, rod characteristics, and UIV level to support more evidence-based surgical strategies in clinical practice.
MATERIALS AND METHODS
1. Search Strategy
This systematic review and meta-analysis was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) 2020 guidelines [29]. The study protocol was registered in PROSPERO (CRD420251107159). An electronic search was performed in PubMed, the Cochrane Library, and Embase to identify studies published through March 2025 that reported and compared the effectiveness of surgical techniques for preventing PJK and PJF, using the keywords “proximal junctional kyphosis,” “proximal junctional failure,” and “spine disease.” The full search strategy is provided in the Supplementary Methods 1-3. Only eligible full-text articles published in nonpredatory journals and written in English were included, with no restrictions on study design.
2. Study Selection
Two reviewers independently screened studies using predefined inclusion and exclusion criteria. Studies were included if they involved adult patients (>18 years) who underwent surgery for ASD or long-segment posterior surgery (≥4 levels) and compared the incidence of PJK or PJF based on preventive surgical techniques. Comparative studies of any design (randomized controlled trials, prospective, or retrospective) were eligible. Preventive techniques of interest included tethering, hook fixation, prophylactic vertebral augmentation, rod characteristics, and UIV level. Studies were excluded if they did not evaluate a preventive surgical strategy. Specifically, we excluded nonsurgical management studies (e.g., osteoporosis or sarcopenia treatment), radiographic-only analyses (e.g., alignment predictors), and observational comparison studies lacking a defined preventive intervention (e.g., demographic or clinical risk-factor analyses comparing PJK vs. non-PJK). Studies involving patients who did not undergo surgery were excluded, as were studies focusing on Scheuermann kyphosis, early-onset scoliosis, adolescent idiopathic scoliosis, or congenital scoliosis. Non-English publications and studies in predatory journals listed in Beall’s list [30] were also excluded. Disagreements between reviewers were resolved through discussion and consensus.
3. Data Extraction
Two reviewers independently extracted data from each included study. Extracted variables included the first author, publication year, study design, patient diagnosis, type of surgery, definition of PJK or PJF used, type of intervention, follow-up duration, and incidence of PJK or PJF. Revision surgery owing to PJK or PJF was collected as a secondary outcome. During extraction, studies were grouped into 5 intervention categories: tethering, hook fixation, prophylactic vertebral augmentation, rod characteristics, and UIV level. When multiple studies appeared to share overlapping patient populations within an intervention category, the study with the largest sample size or most complete dataset was selected to avoid duplication bias. Potential overlaps were identified by cross-checking institutions, author groups, study periods, and inclusion criteria. Two reviewers independently assessed possible duplicates, resolving discrepancies through discussion.
4. Assessing the Quality of Studies
Two reviewers independently assessed the risk of bias using the Newcastle-Ottawa Scale (NOS). Discrepancies were resolved by discussion and, when needed, consultation with a senior author. As this meta-analysis included only observational studies, the NOS—designed for evaluating nonrandomized studies—was used [31]. Risk of bias was assessed across 3 domains: selection, comparability, and outcome assessment. Scores of 7–9 were considered high quality, 4–6 moderate quality, and 0–3 low quality.
Funnel plots were not used to assess publication bias, as their reliability is limited in meta-analyses with fewer than 10 studies, according to Egger et al. [32]. To further evaluate the robustness of our findings and the impact of individual studies or potential bias, a leave-one-out (LOO) sensitivity analysis was performed [33].
5. Statistical Analysis
A random-effects model with inverse variance weighting was applied to account for expected heterogeneity arising from variations in patient populations, surgical techniques, and study methodologies. Because PJK and PJF represent a pathological continuum affecting the proximal junction, the meta-analysis used the reported definitions from the original studies despite definitional variability. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated to estimate effect sizes for PJK, PJF, and revision surgery due to PJK or PJF across the different surgical interventions. Revision surgery owing to PJK was initially grouped with PJF but was analyzed separately when direct PJF assessment was not feasible. Heterogeneity was assessed using the I2 statistics, with values of 25%, 50%, and 75% considered low, moderate, and high heterogeneity, respectively, according to Higgins et al. [34]. All analyses were performed using R v4.4.2 (R Core Team 2024), primarily utilizing the meta (v8.0.1) [35] package.
RESULTS
1. Study Selection
The study selection process is illustrated in Fig. 1. Forty-two studies [20,23,26-28,36-72] were included in the data extraction process. Among them, 4 studies [41-44] were excluded because of overlapping patient populations, resulting in 38 studies [20,23,26-28,36-40,45-72] included in the final systematic review. Although studies of all designs were eligible for inclusion, all screened studies were retrospective. The included studies were categorized into 5 groups: tethering, hook fixation, prophylactic vertebral augmentation, rod characteristics, and level of UIV. Of these, 33 studies [20,23,26-28,45-72] were included in the meta-analysis, based on methodological and clinical considerations described in subsequent sections.
2. Quality Assessment of Studies
The quality assessment of the 38 studies included in the systematic review using the NOS is presented in Supplementary Table 1. Across all studies, the selection domain ranged from 3–4, the comparability domain from 1–2, and the outcome domain from 2–3, resulting in total scores between 7 and 9, indicating high methodological quality. Although a formal GRADE assessment was not performed, the overall certainty of the evidence was judged as low to moderate because all included studies were retrospective.
The inclusion criteria and radiologic outcome definitions were clearly stated in most studies, and measurement methods were adequately described. In the comparability domain, studies that performed multivariate analyses or matched comparisons for key confounders (e.g., age, bone quality, alignment) received higher scores, whereas those without adjustment were downgraded. In the outcome domain, studies with unclear or subjective definitions of PJK/PJF were also downgraded by one point. Information on assessor blinding was generally absent owing to the retrospective design, but this was considered acceptable given the objectivity of radiologic outcomes.
3. Tethering
Six studies [26,37,40,41,59,65] were initially included in the data extraction process; however, the study by Buell et al. [41] was excluded because of overlapping patient populations. The characteristics of the remaining 5 studies [26,37,40,59,65] included in the systematic review that evaluated the efficacy of tethering are summarized in Table 1. Among these, 3 studies [26,59,65] were included in the meta-analysis of spinous process tethering. Mikula et al. [37] was excluded for reporting combined PJK and PJF rates, and Yagi et al. [40] was excluded for reporting sublaminar tethering.
4. Hook Fixation
Fourteen studies [23,27,42,43,47,53,54,56,57,59,60,66,67,72] were initially included in the data extraction process; however, 3 studies [42,43,54] were excluded because of overlapping patient populations. The characteristics of the remaining 11 studies [23,27,47,53,56,57,59,60,66,67,72] included in the systematic review that evaluated the efficacy of hook fixation are summarized in Table 2. All 11 studies were included in the meta-analysis.
5. Prophylactic Vertebral Augmentation
Eight studies [28,45,51,54,57,59,60,66] were included in the data extraction, systematic review, and meta-analysis, with no overlapping patient populations and no exclusions. The characteristics of the 8 studies that evaluated the efficacy of prophylactic vertebral augmentation at UIV or UIV+1 are summarized in Table 3.
6. Rod Characteristics
Seven studies [20,36,49,52,57,62,64] were included in the data extraction process and systematic review, with no exclusions due to overlapping patient populations. The characteristics of these 7 studies evaluating the efficacy of rod characteristics are summarized in Table 4. Among these, 6 studies [20,49,52,57,62,64] were included in the meta-analysis comparing rod alloys (cobalt chromium [CoCr] vs. titanium [Ti]); Charles et al. [36] was excluded for reporting combined PJK/PJF rates.
When comparing CoCr and Ti rods, there were no statistically significant differences in PJK or PJF rates (Supplementary Fig. 1).
7. Level of UIV
Twenty-one studies [20,23,38,39,44,46-48,50,54,55,57,58,61,63,66,68-72] were initially included in the data extraction process; however, the study by Wang et al. [44] was excluded because of an overlapping patient population. The characteristics of the remaining 20 studies [20,23,38,39,46-48,50,54,55,57,58,61,63,66,68-72] included in the systematic review that evaluated the effect of UIV level are summarized in Table 5.
UIV levels were grouped into ranges using cutoffs of T6 (T6 and above vs. T7 and below), T8 (T8 and above vs. T9 and below), T10 (T10 and above vs. T11 and below), and L1 (L1 and above vs. L2 and below). Sixteen studies [20,46,48,50,54,55,57,58,61,63,66,68-72] were included in the meta-analysis. Tian et al. [39] was excluded for reporting combined PJK and PJF rates, and 3 studies [23,38,47] were excluded for using a UIV cutoff at T7.
3) UIV cutoff of T10 (T10 and above vs. T11 and below)
Five studies [54,58,63,68,69] were included in the meta-analysis. When the UIV level was set at T10 and above, PJK rates were lower compared with UIV levels at T11 and below (OR, 0.15; 95% CI, 0.03–0.64; I2=79.9%) (Fig. 5C). However, there was no statistically significant difference in PJF rates when the UIV cutoff was set at T10 (Fig. 6C).
4) UIV cutoff of L1 (L1 and above vs. L2 and below)
Four studies [54,58,68,69] were included in the meta-analysis. There was no statistically significant difference in PJK rates when the UIV cutoff was set at L1 (Fig. 5D). However, when the UIV level was set at L1 and above, PJF rates were lower compared with UIV levels at L2 and below (OR, 0.29; 95% CI, 0.14–0.61, I2=0.0%) (Fig. 6D).
8. Sensitivity Analysis
To examine the robustness of the pooled estimates and the potential influence of bias and individual studies, LOO sensitivity analyses were conducted. LOO analysis was applied to 2 comparisons that showed high or relatively higher heterogeneity compared with other statistically significant results: the effect of prophylactic vertebral augmentation on PJF rates (Supplementary Fig. 2A) and PJF rates according to UIV level with a cutoff at T10 (Supplementary Fig. 2B).
In the analysis of prophylactic vertebral augmentation on PJF rates (Fig. 4B), the pooled OR was 0.58 (95% CI, 0.35–0.95) with an I² of 21.8%. Although this level of heterogeneity is not concerning in absolute terms, it was relatively higher than in other statistically significant pooled results. Across all LOO analyses, the OR remained below 1.0, ranging from 0.47 to 0.70. While omission of some studies led to a loss of statistical significance, the direction of the effect remained consistent, indicating that prophylactic vertebral augmentation reduced the risk of PJF.
Similarly, in the analysis of UIV level with a cutoff at T10 on PJK rates (Fig. 5C), the pooled OR was 0.15 (95% CIm 0.03–0.64) with substantial heterogeneity (I²=79.9%). Across all LOO analyses, the OR remained below 1.0, ranging from 0.07 to 0.26. Excluding the studies by Kang et al. [54] and Lee et al. [58] resulted in 95% CIs that included 1.0, indicating a loss of statistical significance. Despite this, the effect direction remained consistent, suggesting that selecting a UIV at or above T10 reduced the risk of PJK. Notably, heterogeneity dropped markedly (I²=8.8%), with a lower pooled OR (0.07) and a narrower 95% CI (0.02–0.21), when the study by Park et al. [63] was omitted. This suggests that the Park et al. [63] study contributed substantially to the heterogeneity observed among the 4 included studies [54,58,63,68], and that excluding it resulted in a more pronounced reduction in PJK risk and improved precision of the estimate.
DISCUSSION
This meta-analysis evaluated surgical strategies proposed to prevent PJK and PJF. Although the mechanisms underlying these complications remain incompletely understood, several studies have attempted to clarify the biomechanical and anatomical factors contributing to their development.
Among the proposed mechanisms, elevated mechanical stress resulting from the abrupt transition between the rigid instrumented segment and the mobile adjacent segments is one of the most widely accepted explanations [5,22,73,74]. A cadaveric study demonstrated that longer instrumentation increases adjacent segment motion and intradiscal pressure, thereby elevating stress at the adjacent level [75]. Increased segmental hypermobility at the proxjunction has also been implicated as a contributor to mechanical instability. Furthermore, disruption of posterior spinal structures has been shown to decrease flexion stiffness, potentially increasing instability at adjacent motion segments [76]. Accordingly, recent surgical techniques have focused on mitigating these mechanical stresses to reduce the incidence of PJK and PJF. Despite growing interest, the current body of evidence remains limited, with considerable variability in surgical approaches and most studies being retrospective.
The UIV ≥T10 subgroup analysis showed substantial heterogeneity (I²=79.9%), indicating notable variability among the included studies. Although a random-effects model was used, such heterogeneity inherently limits the reliability of the pooled estimate. In addition, the sensitivity analysis revealed a loss of statistical significance when certain studies were excluded; thus, these findings should be interpreted cautiously.
Further exploration identified Park et al. [63] as the primary source of heterogeneity in the analysis of PJK rates using the T10 cutoff. When this study was excluded, heterogeneity markedly decreased (I²=8.8%), and the pooled effect became statistically significant, with a smaller OR and narrower CI. This may be due to the broader definition of PJK in that study, which included vertebral fractures at UIV or UIV+1 and instrumentation failures—criteria more commonly associated with PJF in other literature. Although statistical significance was lost in some LOO analyses, the consistent effect direction supports the robustness of our findings [77].
Although the number of included studies was limited and heterogeneity existed across UIV levels, our meta-analysis showed that spinous process tethering had a preventive effect on PJK. One study investigating sublaminar tethering also reported significantly lower PJF rates in a propensity score-matched cohort [40]. Posterior tethers are thought to reduce mechanical stress at the proximal junction through several biomechanical mechanisms. A finite element analysis by Bess et al. [78] demonstrated that posterior tethers create a more gradual transition in range of motion across the proximal junction, reducing the abrupt stiffness mismatch associated with pedicle screws or hooks. Additionally, at high UIV levels, posterior tethers effectively reduce anterior column compressive forces and mitigate tensile stress on the posterior ligamentous complex [79].
In our meta-analysis, hook fixation was associated with a significantly lower incidence of PJF, although its effect on PJK rates was not significant. A cadaveric study by Metzger et al. [80] showed that supralaminar hooks at the UIV reduced hypermobility at the adjacent segment compared with pedicle screws. Hooks provide a smoother transition of motion between the rigid fused segment and the mobile adjacent levels, thereby reducing stress concentration at the proximal junction and potentially lowering the risk of PJK [81,82]. Even in osteoporotic spines, combining hooks with sublaminar wires may offer biomechanical advantages by reducing proximal junctional stress [83]. However, given their limited effect on PJK, hook constructs may not be sufficient as standalone preventive measures and may be more effective when combined with soft tissue preservation techniques such as tethers.
Osteoporosis is a known risk factor for both PJK and PJF [22,73]. Low Hounsfield unit values at UIV and UIV+1 are specifically associated with increased risk [84,85]. Vertebroplasty is known to increase vertebral body strength [86,87], which may explain our finding that prophylactic vertebral augmentation at UIV or UIV+1 reduced PJF risk. However, this localized increase in stiffness may alter load distribution and heighten stress at adjacent levels, potentially leading to adjacent segment fractures [88]. Because this mechanism parallels that of bony failure in PJF, careful consideration is needed when applying prophylactic vertebral augmentation. In this context, systemic osteoporosis treatment with anabolic agents may serve as a complementary strategy by addressing underlying bone fragility [89,90].
Different rod alloys exhibit varying stiffness properties. While stiffer rods such as CoCr may reduce mechanical failure due to rod breakage, they may increase segmental stiffness at the fused levels, potentially contributing to PJK [52]. However, our meta-analysis found no statistically significant difference in PJK or PJF rates between CoCr and Ti rods. This may be because multiple factors, including rod number, rod diameter, and alignment correction, likely play a more significant role in junctional outcomes than rod stiffness alone [36].
There was no significant difference when the UIV cutoff was set at T6 or T8. Although the rib cage provides additional stability, particularly in the upper thoracic spine, where true ribs are present [91], patients undergoing correction surgery with UIV above T6 often have more severe ASD [92], which may confound outcomes. This may explain why no significant difference was observed in PJK or PJF rates despite the added rib cage stability. A similar confounding effect may exist for the T8 cutoff. Additionally, T8 is often located near the apex of thoracic kyphosis [93], a region subject to increased mechanical stress, which may complicate its suitability as a UIV level [92].
When the UIV cutoff was set at T10, patients with UIV at ≥T10 showed a lower incidence of PJK in our meta-analysis. Although lower thoracic levels (including false ribs) provide less stability than upper thoracic segments, vertebrae above T10 still benefit from rib cage support [94,95].
Moreover, the thoracolumbar junction (TLJ), a transitional zone between the rigid thoracic spine and the mobile lumbar spine, is inherently subjected to mechanical stress [96]. This makes it especially vulnerable to osteoporotic fractures, as reflected in clinical data showing a high prevalence of osteoporotic vertebral fractures at the TLJ, particularly in older individuals [97]. When instrumentation of the mobile lumbar spine extends to the TLJ, stress concentration and hypermobility at the proximal junction may be exacerbated, potentially increasing the risk of PJK or PJF. Consistent with this mechanism, our meta-analysis showed a reduced incidence of PJK when the UIV was placed at T10 or above, thereby avoiding the mechanically vulnerable TLJ. Some studies have suggested that TLJ instrumentation may be considered in selected cases, such as younger patients (<70 years) without osteoporosis [98]. However, based on our findings and the biomechanical considerations outlined above, we recommend avoiding the TLJ as the UIV in vulnerable populations, particularly those with osteoporosis. We further suggest that future meta- analyses and prospective studies incorporate subgroup analyses focusing on osteoporotic patients to clarify the specific risks associated with UIV selection at the TLJ in this population.
Furthermore, UIV at L1 or above showed a lower incidence of PJF in our meta-analysis, which can be explained from multiple perspectives. First, this group includes patients with UIV at T10 or above, levels that have already demonstrated protective effects against PJK, likely because of rib cage support and avoidance of the TLJ. Second, when the UIV ends at L2 or below, particularly in osteoporotic patients, increased weight-bearing stress may lead to mechanical failure, supported by finite element analysis and clinical evidence showing increased adjacent segment stress and degeneration following short lumbar fusion [99,100]. Third, limited correction due to the shorter construct of UIV at L2 or below may contribute to the higher rate of mechanical complications. A recent study reported that short fusion resulted in significant loss of sagittal alignment and a higher incidence of PJK, likely reflecting undercorrection [101]. Therefore, future studies or meta-analyses stratifying outcomes by the degree of sagittal correction, rather than fusion length alone, may provide a more nuanced understanding of risk factors for junctional complications.
Based on our findings, setting the UIV proximal to the TLJ, specifically above T10, may help reduce the incidence of PJK and PJF. Although the stabilizing effect of the rib cage diminishes below the true rib region, even false ribs may provide partial biomechanical support [91,94]. However, selecting a very proximal UIV level should be carefully considered on a case-by-case basis, taking into account the patient’s overall condition, surgical morbidity, and alignment goals. Therefore, a UIV at T10 or slightly higher may provide a balance between achieving mechanical stability and minimizing surgical invasiveness.
Given the heterogeneous and multifactorial nature of PJK and PJF pathogenesis, a single preventive technique may not be universally effective. Additionally, patients vary widely in osteoporosis severity, comorbidities, and alignment goals, further limiting the applicability of a uniform preventive strategy. Moreover, even within the same patient, mechanical stress may differ depending on the UIV level. A biomechanical study by Yagi et al. found that at lower UIV levels, although tethers offer some biomechanical benefit, they do not sufficiently reduce compressive stress, and significant stress concentrations persist, suggesting the need for additional prophylactic measures such as vertebroplasty [79].
From a patient-centered perspective, preventive strategies may be prioritized differently according to individual risk profiles. For example, augmentation may be particularly advantageous for patients with severe osteoporosis, whereas tethers may be more appropriate in cases requiring rigid deformity correction. This aligns with previous comprehensive reviews emphasizing that optimal prevention of mechanical complications requires multifactorial, patient-specific approaches integrating surgical, radiologic, and bone quality factors [102]. However, direct comparative evidence across these subgroups remains limited, and future studies stratified by bone quality, age, and deformity severity are warranted to refine individualized preventive strategies.
This study has some limitations. First, this review focused solely on surgical techniques for preventing PJK and PJF. Therefore, other potential risk factors, such as age, sex, bone mineral density, sagittal alignment, and comorbidities that could act as confounders, were beyond the scope of this analysis.
Second, all included studies were retrospective, and the number of available studies was limited. This also made it impossible to adjust for previously noted nonsurgical confounders, as the small number of studies did not allow meaningful subgroup analyses. Regarding surgical factors, although we performed analyses based on UIV levels in the overall dataset, the limited sample size prevented additional subgroup analyses for each surgical technique (e.g., hooks, tethers) by UIV level or by specific definitions of PJK and PJF used in individual studies. Furthermore, the levels at which preventive techniques (e.g., hooks, tethers, cement) were applied varied across studies, potentially contributing to clinical heterogeneity and limiting the interpretability of the pooled estimates. Nonetheless, sensitivity analyses demonstrated consistent directions of effect, supporting the overall robustness of the findings despite such variability. Additionally, although a random-effects model was applied, this statistical method cannot eliminate inherent between-study heterogeneity.
Third, the retrospective designs and absence of prospective or randomized studies limited the ability to control for bias. In several studies, the decision to apply preventive strategies was left to the surgeon’s discretion. This may have led to the preferential selection of patients with more favorable bone quality, alignment, or overall clinical condition, introducing potential selection bias. Although all included studies scored high on the NOS, such inherent bias cannot be fully controlled and should be considered when interpreting the pooled estimates.
Fourth, definitions of PJK and PJF varied across studies. Because of the limited number of studies in each subgroup, we were unable to perform stratified or sensitivity analyses based on specific definitions. This heterogeneity may have influenced the pooled estimates. However, as PJK and PJF are commonly considered part of a pathological spectrum at the proximal junction, the meta-analysis used reported rates according to each study’s definitions, despite variability.
Fifth, formal assessment of publication bias (e.g., funnel plots or Egger test) was not feasible because each comparison included fewer than 10 studies. Therefore, small-sample publication bias cannot be excluded.
Finally, despite our efforts to minimize duplicate inclusion from overlapping patient populations, some studies may still have included overlapping cohorts, especially those from the same institutions or authorship groups, which may have introduced unintended duplication into the pooled estimates.
CONCLUSION
This meta-analysis suggests that preventive surgical strategies such as tethering, hooks, and prophylactic vertebral augmentation may help reduce the risk of PJK and PJF. However, given the heterogeneity in application levels and patient selection across studies, these results should be interpreted with caution. Rather than dismissing any technique based on limited evidence, we cautiously recommend that these modalities be tailored and combined according to individual patient and surgical contexts. UIV selection should also be made carefully, taking into account each patient’s alignment goals and the stability characteristics of the thoracic spine. While placing the UIV slightly above T10 may offer improved stability, consistent with our findings, care should be taken to avoid unnecessarily proximal fixation.
Despite the inherent limitations of this meta-analysis, including its reliance on retrospective studies, we believe these findings contribute to a more evidence-based approach to preventing PJK and PJF. Nevertheless, as all included studies were retrospective observational cohorts, high-quality prospective and randomized controlled studies are needed to validate these results, reduce confounding, and establish clearer clinical consensus on optimal preventive strategies.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Full text via PMC
Download Citation
Supplement
Print
