Superior facet joint violation after lumbar pedicle screw placement: a scoping review of prevalence, biomechanics, and implications for adjacent segment disease

Abstract

Facet joint violation (FJV) is a known complication of pedicle screw fixation that may contribute to adjacent segment disease (ASD) by altering spinal biomechanics and increasing loading of the cranial facet joints. However, the prevalence, biomechanical effects, and long-term consequences of FJV remain unclear. A scoping review was conducted using the PubMed, Embase, and Scopus databases. Clinical studies reporting the number of FJVs in relation to screws placed during lumbar fusion or assessing the frequency of ASD in patients with FJV were included. Biomechanical studies evaluating segmental kinematics and loading after FJV were also included. Data regarding study characteristics, surgical techniques, FJV rates, severity grading, and outcomes were extracted. Bayesian statistical models were applied for pooled prevalence estimates. Fifty studies met the inclusion criteria (39 clinical and 11 biomechanical studies). The prevalence of FJV varied, with robotic-assisted percutaneous placement associated with the lowest risk (4.79%; 95% credible interval [CrI], 3.88–5.79), and freehand percutaneous placement associated with the highest risk (19.45%; 95% CrI, 18.15–20.73). FJV rates were highest at L2 (14.5%; 95% CrI, 10.4%–19.0%) compared to lower levels. Biomechanical studies indicate that minor FJV may destabilize the superior segment by disrupting the facet capsule, while severe FJV involving full joint traversal might stabilize the segment due to the screw’s interaction with both articulating processes. Limited evidence suggests an association between FJV and ASD, though methodological limitations, selection bias, and reporting errors limit conclusions. FJV is common during lumbar fusion, especially with percutaneous screw placement. While biomechanical evidence suggests differing grades of FJV may variably alter segmental stability, its role in ASD development remains uncertain. Given its prevalence, biomechanical implications, and potential mitigation through navigation or robotics, determining whether FJV is a clinically significant driver of ASD is a key research imperative.

GRAPHICAL ABSTRACT

Introduction

Lumbar fusion relies on internal fixation to stabilize the spine while arthrodesis develops, preventing prolonged bed rest or external bracing. Pedicle screw (PS) fixation, introduced in the 1970s, has become the gold standard for internal fixation due to its ability to stabilize all three spinal columns, improving fusion rates and clinical outcomes compared to earlier techniques like translaminar fixation [1]. Despite its advantages, PS placement remains technically demanding, particularly with minimally invasive (MI) approaches where direct anatomical visualization is limited. One well-recognized complication is superior facet joint violation (FJV), a phenomenon that has long been implicated in the development of adjacent segment disease (ASD) [2].

The rise of MI spine surgery has been driven by its potential to reduce perioperative morbidity. Traditional open PS fixation requires extensive paraspinal muscle dissection, which can lead to postoperative pain, muscle atrophy, and prolonged recovery. By contrast, percutaneous PS placement minimizes soft tissue trauma and preserves the posterior tension band, theoretically reducing the risk of iatrogenic instability and pain [3]. However, the lack of direct visualization during percutaneous insertion may render screw placement challenging and increase the risk of FJV [4].

FJV is of particular concern at the upper level of the construct, where screws breach the joint communicating with the cranial vertebra, as this may accelerate ASD [2]. This represents the most common long-term complication after lumbar fusion, occurring in up to 4% of patients per year, depending on diagnostic criteria [5]. Despite its prevalence, the etiopathogenesis of ASD remains debated, with theories focusing on postoperative alterations in sagittal balance [6], altered facet/disc loading due to instrumentation [7], and preexisting intervertebral disc degeneration [8]. While biomechanical models suggest that FJV alters segmental kinematics and increases superior facet loading, a clear consensus regarding its role in ASD development remains elusive [9].

Since 2008, several clinical [10], cadaveric [2,11], and computational [12] studies have investigated the relationship between FJV, segmental biomechanics, and ASD, yielding heterogeneous findings and disparate conclusions. This scoping systematic review aims to synthesize existing literature on FJV’s prevalence, biomechanical consequences, and potential role in ASD development after lumbar fusion for degenerative spinal pathology.

Methods

Search strategy and selection criteria

This review adhered to PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines. PubMed, Embase, and Scopus databases were searched on January 23, 2025, using the key terms “FJV” OR (“Facet” AND “Violation”) without date restriction. Two reviewers (C.M. and D.K.) independently screened titles and abstracts, resolving disagreements by consensus. Inclusion criteria were clinical studies reporting FJVs associated with a specified, or calculable, number of screws inserted for internal fixation during lumbar fusion or the frequency of ASD versus FJV, as well as primary research on biomechanical changes after FJV. Exclusion criteria included studies on traumatic or neoplastic pathologies, revision surgery, non-English reports, and review articles. Data were extracted using a standardized form, capturing details such as author(s), year of publication, study design, sample size, patient characteristics, surgery type, number of screws inserted, number of FJVs, level, and FJV severity. As a scoping review, a formal risk of bias assessment was not conducted, following established methodological frameworks. However, methodological limitations of included studies were considered in the synthesis. This study was not preregistered.

Data synthesis

Rates of FJV were pooled for crude comparison using hierarchical binomial and multinomial Bayesian models without random effects in Python 3 with PyMC. Stratification was by surgical technique, upper instrumented level, and FJV severity (graded per Babu et al. [4]) (Fig. 1). Standard procedures were used to assess model convergence (R-hat values <1.01, effective sample size at bulk and tail, visual inspection of sampling output) [13]. In figures showing pooled outcomes, the y-axis represents the probability of the event (0%–100%). Raw values from included studies are shown as points. The credible interval (CrI) is used analogously to the traditional confidence interval (CI), representing the range within which the true value is presumed to lie with 95% certainty.

Fig. 1.

Babu grading of facet joint violation (FJV). (A) Grade 0: no violation. (B) Grade 1: FJV into cortical bone without involvement of articular cartilage. (C) Grade 2: cartilage breach without full joint traversal. (D) Grade 3: complete joint traversal. Adapted from Zhao et al. Global Spine J 2023;13:730-6 [14], under the Creative Commons Attribution License (CC BY).

Results

Search results

The literature search yielded 1,152 articles, of which 482 duplicates were removed. After screening the titles and abstracts of the remaining 670 articles, 107 full texts were sought, but full texts of six articles could not be accessed. After applying the inclusion and exclusion criteria, 50 studies were included in the review (Fig. 2).

Fig. 2.

Flow chart of search strategy. FJV, facet joint violation; CBT, cortical bone trajectory.

Overview of clinical studies

Thirty-nine studies reported the in-vivo clinical prevalence of FJV per screw following lumbar spine screw fixation using various techniques [14–52]. Across studies, the mean patient age was 58.5 years (range, 24.7–73.2 years), and the mean body mass index (BMI) was 26.0 kg/m² (range, 21.2–31.9 kg/m²). FJV grading systems varied, with the Babu grading system most commonly used (18 studies), followed by the Yson grade (eight studies) and Modified Park’s (four studies). Other studies had used various bespoke and analogous systems. Fusion approaches varied widely, with posterior techniques like transforaminal lumbar interbody fusion (TLIF) and posterior lumbar interbody fusion (PLIF) most frequently described.

Prevalence of Facet Joint Violation

The prevalence of FJV was synthesized from 39 studies reporting FJV per screw. Fig. 3 illustrates the variation in FJV risk per screw across navigation technologies and screw fixation techniques. Robotic-assisted (RA) percutaneous PS placement demonstrated the lowest risk (4.79%; 95% CrI, 3.88–5.79), while freehand percutaneous PS fixation had the highest (19.45%; 95% CrI, 18.15–20.73). Intermediate risks were observed for robotic open PS (22.20%; 95% CrI, 17.69–26.71), robotic cortical bone trajectory (CBT) screws (15.17%; 95% CrI, 11.27–19.48), navigated open (16.71%; 95% CrI, 13.87–20.07), navigated percutaneous (6.37%; 95% CrI, 5.56–7.23), navigated CBT (11.22%; 95% CrI, 9.12–13.69), freehand open (16.40%; 95% CrI, 15.15–17.67), and freehand CBT (10.94%; 95% CrI, 8.93–13.28).

Fig. 3.

Rate of facet joint violation (FJV) versus surgical technique. Open and percutaneous pedicle screw insertion show in blue and orange respective. Cortical bone trajectory (CBT) screw fixation shown in green. Estimates from individual studies are shown as points while the pooled mean and 95% credible interval is displayed via the thick horizontal line and the thin horizontal lines above and below.

Comparisons of posterior distributions consistently favored RA percutaneous placement, with at least 99.3% probability of a lower FJV risk compared to all other techniques. Among freehand approaches, CBT demonstrated a lower risk than percutaneous and open techniques (>99.9% probability). Freehand percutaneous screw placement was associated with a slightly higher risk than freehand open placement (>99.9% probability).

The estimated proportion of patients expected to experience at least one FJV is 9 per 100 with RA-percutaneous PS, 12 per 100 with navigated percutaneous PS, 30 per 100 with freehand open PS, and 35 per 100 with freehand percutaneous PS placement.

Severity of facet joint violation

Fourteen studies reporting FJV severity using the Babu grading system or directly analogous classifications were included to estimate the distribution of FJV grades [14–16,18,20,24,25,28,33,35,36,40,45,49]. The Babu system categorizes FJV from grade 0 (no violation) to grade 3 (complete joint traversal), with grade 1 indicating screw placement within the facet joint without breaching the articular cartilage and grade 2 indicating cartilage breach without full joint traversal (Fig. 3). All navigation and screw fixation types were aggregated. The distribution of FJV grades was as follows: grade 0, 79.8% (95% CrI, 78.6%–81.1%); grade 1, 13.3% (95% CrI, 12.3%–14.4%); grade 2, 4.66% (95% CrI, 3.99%–5.32%); and grade 3, 2.19% (95% CrI, 1.74%–2.64%) (Fig. 4).

Fig. 4.

All techniques aggregate frequency of facet joint violation (FJV) versus Babu grading of severity. Frequency of FJV stratified by severity as measured via the Babu classification. Height of bars indicates mean estimate and horizontal lines indicate boundaries of 95% credible interval.

Facet joint violation rate versus upper level of fusion construct

To assess the relationship between FJV rates and the upper level of the fusion construct, data from 25 studies reporting the number of screws implanted per level were synthesized [4,14,15,17,20,21,23,24,26,29–36,38–40,43,45,48,50,51]. The risk of FJV was highest at the L2 level, at 14.5% (95% CI, 10.4%–19.0%). Other levels were associated with lower risks: L3, 8.9% (95% CI, 6.6%–11.2%); L4, 10.7% (95% CI, 9.3%–12.2%); and L5, 9.7% (95% CI, 8.2%–11.4%) (Fig. 5).

Fig. 5.

All techniques aggregate frequency of facet joint violation (FJV) versus upper instrumented level. Frequency of FJV stratified by upper instrumented level. Height of bars indicates mean estimate and horizontal lines indicate boundaries of 95% credible interval.

Biomechanics

Five studies assessed the biomechanical impact of FJV on the superior adjacent segment [2,11,12,53,54]. These included ex vivo and computational investigations, with two studies using fresh-frozen human specimens and two studies employing formaldehyde-fixed specimens (Table 1). All human cadaveric studies assessed changes in range of motion (ROM) using dedicated testing rigs capable of rotating and translating the upper segment in the sagittal, axial, and coronal planes. Reference markers and optical tracking were used to measure differences in ROM when known forces were applied. One cadaveric study also measured changes in surface strain of the intervertebral disc, while one finite element model evaluated contact pressures in the superior disc and facet joints after FJV. None of the physical experiments incorporated axial preload to replicate in vivo loading conditions.

Table 1.

Summary of included studies investigating the biomechanical effect of FJV

Author	Levels investigated	Model	Type of FJV	Control group	Outcome measure(s)	Key findings	Notes
Human cadaveric
Cardoso et al. [2]	L1–sacrum sequentially from L5	-	Facet capsule denuded + cortical violation with rongeur after PS placement, both unilateral and bilateral defects induced.	PS placed with intact facet capsule/cortex	ROM in axial rotation, flexion/extension, and lateral bending	↑ Axial rotation with bilateral FJV No change in flexion/extension or lateral bending	Method of FJV may clinically resemble partial facetectomy rather than in-vivo FJV. No record of preload applied to specimens.
Xu et al. (Part 1) [54]	L4–S1	-	Bilateral impingement of head of PS into facet joint	PS placed without impingement of facet joint	ROM in axial rotation, flexion/extension, and lateral bending	↓ Flexion-extension No change in axial rotation or lateral bending	No preload applied to specimens
Xu et al. (Part 3) [11]	L4–S1	-	Bilateral violation of facet joint articulation by PS	PS placed without violation of facet joint	ROM in axial rotation, flexion/extension, and lateral bending	↓ Flexion-extension and axial rotation. No change in lateral bending	No preload applied to specimens
Wangswatwong et al. [53]	L3–S1	-	Lateral surface of SAP and capsule violated with rongeur then medial entry PS placed bilaterally	More lateral PS entry point chosen with care to avoid damage to facet capsule/joint	ROM in axial rotation, flexion/extension, and lateral bending. Surface strain of intervertebral disc	↑ Flexion, axial rotation, and lateral bending, No change in extension FJV may alter surface strain patterns; no significant relationship identified.	Change in lateral bending found only on right side despite symmetrical arrangement of screws and bilateral violation. No record of preload applied to specimens. Degree of joint violation not recorded
Computational
Kim et al. [12]	L2–L5	Finite elements	Bilateral L4 screws: (1) traversing joint space; (2) within 1 mm of joint space; and (3) preserving facet joint	Intertransverse fusion of L4/L5 with and without PS fixation without facet joint violation	Facet contact force Adjacent segment disc pressure	↑ Facet contact pressure & disc pressure at L3/L4 with FJV, greatest with articular surface violation; effects not transmitted to L2/L3 (two levels above fusion mass)	-

FJV, facet joint violation; PS, pedicle screw; ROM, range of motion; SAP, superior articular process.

Range of motion

Cardoso et al. [2] used 10 cadaveric lumbar spines (L1–sacrum) to assess ROM before and after L5–S1 PS placement. Initially, the facet joint capsule was preserved, and then unilateral and bilateral FJV were sequentially induced using a Leksell rongeur to denude the capsule and violate the external cortical surface of the superior articular process. The study found that unilateral FJV had no significant effect on axial rotation, while bilateral FJV increased axial rotation by >10%. Neither unilateral nor bilateral FJV altered ROM in flexion-extension or lateral bending at any tested level from L5 to L1.

Wangsawatwong et al. [53] divided their sample into two groups of seven segments including the L3–S1 vertebrae and compared a lateral converging screw trajectory, preserving the facet joints, versus a medial entry point which incurred bilateral FJV. In the FJV group, a Leksell rongeur was used to bite the lateral most surface of the superior articular process and violate the joint capsule while exposing the screw entry point between the inferolateral aspect of the facet joint and the mid-transverse process line. A biomechanical setup, similar to Cardoso et al. [2], was used to test ROM, with additional strain analysis of the rostral L3/L4 intervertebral disc. They found that bilateral FJV significantly increased ROM in flexion, axial rotation, and lateral bending, with the greatest change observed in axial rotation.

Xu et al. [11,54] conducted a series of experiments assessing the impact of screw head impingement and severe facet violation at L4–S1 levels in cadaveric segments from 24 individuals split into groups of six. Compared to controls, screw head impingement significantly reduced flexion-extension ROM (89.7%±5.6%, p=0.007) but did not alter axial rotation (103.5%±9.6%, p=0.36) or lateral bending (103.5%±9.6%, p=0.83). Severe facet violation significantly reduced ROM in flexion-extension (56.8%±5.0%, p=0.001) and axial rotation (61.9%±16.4%, p=0.02) but did not affect lateral bending (95.5%±22.4%, p=0.71). Noting the contrasting results, the authors suggested that the FJV by Cardoso et al. [2] may better reflect partial facetectomy rather than clinical FJV.

Force transmission

Kim et al. [12] conducted a finite element analysis of biomechanical changes following FJV. They used a three-dimensional model based on a computed tomography (CT) scan of a healthy 46-year-old man, simulating decompression by removing supraspinous and interspinous ligaments between L4–L5 and partially resecting spinous processes, while preserving the posterior ligament complex between L3 and L4. The model tested three scenarios: (1) PSs traversing the L4/L5 facet joint, (2) screws preserving the facet joint, and (3) screws placed 1 mm from the joint surface. A 150 N preload was applied with 10 Nm flexion, extension, torsion, and lateral bending.

Under extension, FJV increased L3/L4 facet contact pressure, with severe violation causing a 166.7% rise and 1 mm impingement increasing it by 102.6%. This effect was reversible upon screw removal. Torsional moments also increased L3/L4 facet contact force by 132.1% and 96.4% for severe and 1 mm FJV, respectively. Similarly, disc pressures at L3/L4 increased by 68.1% (extension) and 76.5% (torsion) for severe FJV, and by 55.9% (torsion) and 51.1% (extension) for 1 mm FJV. These effects were not transmitted two segments above. All fusion models, including those without PSs, elevated adjacent segment disc and facet pressures beyond baseline.

The strain analysis by Wangsawatwong et al. [53] found that FJV reduced anterior disc bulging in flexion but increased cranial-caudal stretch posteriorly. However, these differences were inconsistent due to substantial variation, making their clinical implications uncertain.

Facet Joint Violation and Subsequent Development of Adjacent Segment Disease

Six studies investigated the relationship between FJV and ASD [10,55–59] (Table 2). These retrospective analyses used varying definitions of radiographic ASD (rASD) and clinical ASD (cASD). Sample sizes ranged from 87 patients [59] to 658 patients [55], and follow-up duration varied from 2 [57,58] to 13.7 years [59].

Table 2.

Summary of included studies investigation the relationship between FJV and ASD

Author	Fusion type	Definition of ASD	Significant relationship with FJV		Primary limitations
			Univariate analysis	Multivariate analysis
Bagheri et al. [55]	Lumbar fusion with pedicle screw fixation (not further described)	rASD: disc height collapse >20%, olisthesis >4 mm, instability >10°, spinal stenosis or disc herniation	Yes (p<0.01)	Yes (OR, 7.48; p<0.01)	No detailed surgical description Imaging and grading used to diagnose FJV not described
Shin et al. [56]	MIS and open TLIF	rASD: instability (≥10° angulation, >4 mm olisthesis) or disc degeneration cASD: VAS ≥3, ODI ≥15	Yes (p<0.05 but reported CIs included null value)	Yes, but only for open TLIF (OR, 56.4; 95% CI, 1.03–999)	Reporting/calculation error (CIs inconsistent with significance)
Ouchida et al. [57]	LLIF with posterior instrumentation	rASD: slippage ≥3 mm, intravertebral narrowing ≥3 mm or opening ≥5 mm	No (p=0.86)	Not performed	FJV not included in multivariate analysis. Heterogeneous surgical techniques.
Oh et al. [59]	Open PLIF and PLF	ASD: anterolisthesis ≥4 mm, angular motion ≥10°, >50% disc height loss, progressive of facet/disc degeneration	Yes (p<0.05)	Not performed	Surgical technique not described in detail.
Levin et al. [10]	Lumbar fusion with pedicle screw fixation (not further described)	Reoperation at 1, 2, and 3 years postoperative	No at 1 year (p=0.57) Yes at 2 and 3 years (both p=0.02)	Yes (2 yr OR, 2.89; 95% CI, 1.15–7.27; 3 yr OR, 2.53; 95% CI, 1.16–5.51)	Likely significant selection bias (1,080 cases screened, only 240 included) Sagittal balance not included in multivariate model
Wang et al. [58]	TLIF and PLIF (no further details)	cASD: disc degeneration causing radiculopathy, stenosis, or instability	Yes (75% vs. 19%, p<0.01)	Yes (OR, 2.02; 95% CI, 2.89–18.41)	MLEs fall outside CIs, indicating errors in reporting or calculation

FJV, facet joint violation; ASD, adjacent segment disease; rASD, radiographic ASD; OR, odds ratio; MIS, minimally invasive surgery; TLIF, transforaminal lumbar interbody fusion; cASD, clinical ASD; VAS, Visual Analog Scale; ODI, Oswestry Disability Index; CI, confidence interval; LLIF, lateral lumbar interbody fusion; PLIF, posterior lumbar interbody fusion; PLF, posterolateral fusion; MLE, maximum likelihood estimate.

A retrospective cohort study by Bagheri et al. [55] assessed radiographic criteria for ASD in 658 patients who received fusion with PS fixation for degenerative lumbar pathology. rASD was defined as loss of disc height, spondylolisthesis, instability on flexion/extension radiographs, or stenosis/disc herniation on follow-up CT/magnetic resonance imaging. After excluding 28 patients due to insufficient follow-up (<4 years) or missing imaging, the study found that 76 patients (12.1%) developed rASD after posterior lumbar fusion at 4 years. Univariate analysis showed that the rASD group had lower preoperative (32.34°±12.1° vs. 40.41°±3°, p=0.02) and postoperative lumbar lordosis (31.35°±10.1° vs. 38.72°±23°, p=0.03), more frequent FJV (p<0.01), superior segment disc degeneration (p<0.01), longer fusion constructs (p<0.01), rigid rather than dynamic instrumentation (p<0.01), higher baseline BMI (27.86 vs. 23.15, p=0.03), and greater preoperative L1–S1 sagittal plumb line (p=0.04). Multivariate logistic regression identified superior FJV as an independent predictor of rASD, with an odds ratio (OR) of 7.48 (p<0.01), adjusting for lumbar lordosis, fusion length, superior segment degeneration, and BMI.

Shin et al. [56] compared ASD after minimally invasive surgery (MIS) and open TLIF employing equivalent diagnostic criteria for rASD while also assessing differences in cASD, defined as new back/leg pain or worsening moderate/severe disability. At 1 year, the incidence of rASD was 6.6% in the MI-TLIF group and 16.7% in the open TLIF group (p=0.61). However, cASD was significantly less common after MI-TLIF than open TLIF (12.3% vs. 28.6%, p=0.039). At five years, the incidence of rASD was significantly higher after open TLIF (35.2% vs. 46.0%, p=0.02), and return of symptoms was marginally more common after open TLIF (42.2% vs. 54.1%, p=0.23). By 10years, there was no significant between-group difference in terms of rASD (58.1% MI-TLIF vs. 60.6% open TLIF, p=0.13) or symptomatic ASD (60.7% vs. 78.8%, p=0.08). Reoperation rates at 10 years were higher after open TLIF (21.4%) than MI-TLIF (13.0%), but the between-group difference was not statistically significant. Instability was more frequently seen at 5 and 10 years after MI-TLIF than open TLIF (5 years: 20.5% vs. 8.1%; 10 years: 35.5% vs. 18.2%) while disc degeneration was more common after open TLIF (24.2% vs. 9.7%).

Univariate analysis revealed that superior FJV predicted ASD in both groups, with preoperative adjacent disc degeneration significant only in the open TLIF groups. Although all three p-values were under 0.05, the CIs did not exclude zero, indicating a potential reporting or calculation error. Multivariate analysis found cranial FJV independently associated with ASD after open TLIF (OR, 56.4; 95% CI, 1.03–999) but not MI-TLIF (OR, 1.21; 95% CI, 0.29–5.1).

Ouchida et al. [57] retrospectively studied ASD after lateral lumbar interbody fusion (LLIF) with posterior instrumentation for unstable degenerative spondylolisthesis. The surgical procedures varied, with some patients undergoing open LLIF with direct decompression via partial laminectomy, while others received percutaneous PS insertion and indirect decompression. The use of intraoperative navigation was not specified. FJV was defined as screw involvement within the superior facet cortex and rASD as ≥3 mm vertebral slippage, ≥3 mm intervertebral space narrowing, or ≥5 mm intervertebral opening. Among 97 LLIF patients, 19 (24.3%) developed rASD at 2 years, with two symptomatic cases requiring revision. Univariate analysis showed no significant differences in lumbar lordosis (38.8°±14.9° vs. 44.5°±12.2°, p=0.09), pelvic incidence-lumbar lordosis mismatch >10° (52.6% vs. 47.4%, p=0.68), or FJV (15.8% vs. 14.1%, p=0.86). However, adjacent level decompression was more common in rASD+ patients (21.1% vs. 3.8%, p=0.035). No multivariate analysis incorporating FJV was performed.

Oh and Seo [59] retrospectively analyzed rASD following open PLIF or PLF with PSs. They defined ASD using the same criteria as Shin et al. [56] and Bagheri et al. [55]. The study required postoperative CT scans within 1 year, but the number of patients excluded due to missing imaging was not reported. FJV was graded as no violation, suspected violation, or definite violation. Among 56 ASD cases and 31 controls (mean follow-up: ASD 13.7 years, controls 11.7 years), FJV was significantly more frequent in ASD patients (47.3% vs. 30.6%, p<0.05), with bilateral violations also being more common (23.2% vs. 9.7%). Definite violations occurred in 24% of screws in the ASD group versus 8% in controls. The surgical techniques were not described in detail.

Levin et al. [10] retrospectively assessed FJV in 112 patients and 128 controls, excluding trauma and revision cases. Out of 1,080 lumbar fusions (not further described), only 240 met the inclusion criteria due to strict follow-up and imaging requirements, indicating potential selection bias. Three-year reoperation rates were higher in the FJV group (19.6% vs. 9.4%, p=0.02). Multivariate regression analysis confirmed FJV as an independent predictor of reoperation at 2 and 3 years (2 years OR, 2.89; 95% CI, 1.15–7.27; 3 years OR, 2.53; 95% CI, 1.16–5.51). Decompression at the superior level was not a significant factor, and sagittal balance measures were not included in the regression model.

Wang et al. [58] retrospectively analyzed lumbar degeneration after TLIF and PLIF (not further described) in 237 patients. Patients without 2-year follow-up or postoperative imaging were excluded, although the exact numbers were not specified. ASD was defined as disc degeneration causing clinical symptoms, such as radiculopathy, stenosis, or instability. After a mean follow-up of 2.6 years, 15 patients (6.3%) developed symptomatic ASD. Univariate analysis revealed that superior FJV was significantly more frequent in ASD patients (75% vs. 19.3%, p<0.01). However, no differences were found in lumbar lordosis, adjacent segment motion, or paraspinal muscle fat infiltration. Multivariate regression identified BMI (OR, 1.36; 95% CI, 1.68–6.43), preoperative superior segment disc degeneration (OR, 4.30; 95% CI, 5.93–31.55), and superior FJV (OR, 2.02; 95% CI, 2.89–18.41) as independent predictors of ASD. However, the reported maximum likelihood estimates (e.g., OR 2.02 for FJV) fell outside their respective CIs, indicating a reporting or calculation error that raises concerns about the validity of these findings.

Discussion

The review suggests that FJV is a common issue after lumbar fusion, especially percutaneous PS insertion without intraoperative navigation or robotic assistance. Available evidence suggests that up to one-third of patients may experience at least unilateral FJV under such conditions. Conversely, RA PS placement had the lowest FJV rates, with approximately 5% of screws or 9% of patients affected, while navigated approaches were associated with an intermediate risk.

When using freehand techniques, open approaches tend to better preserve facet integrity compared to percutaneous placement. However, there is limited data for direct comparisons between open and percutaneous techniques in navigation or RA settings. Similarly, the medial entry point of CBT screws conceptually reduces the risk of FJV and this is suggested in freehand comparisons. Yet, the impact of robotic or navigated CBT screw placement remains unclear, with only a single data point available for each technique.

The most common pattern of FJV is partial cortical encroachment. When considering all techniques, over 6% of screws are expected to partially or fully transect the facet’s articulating surface, which translates to approximately 13% of patients having at least one screw penetrating the joint surface. The L2 vertebra appears to be at highest risk of FJV, potentially due to the medial orientation of upper lumbar pedicles, which may predispose them to a medial screw trajectory similar to that intentionally used in the biomechanical study by Wangsawatwong et al. [53].

A 2025 survey found that fewer than half of surgeons routinely use intraoperative navigation [60], suggesting FJV is likely common in practice. Despite its potential link to ASD, the biomechanical consequences of FJV remain poorly understood due to limited research. The pioneering study by Cardoso et al. [2] is often cited as evidence that FJV destabilises the cranial segment, yet as noted by Xu et al. [11], the methodology by Cardoso et al. [2] did not involve direct screw penetration of the facet joint. Instead, screws were placed within the pedicle, and the facet capsule was deliberately excised and the superior articular process violated using a rongeur. Given the critical role of the facet capsule in resisting flexion and axial rotation [61], the findings by Cardoso et al. [2] might reflect the effect of capsular disruption rather than true FJV.

A similar pattern can be seen in the study by Wangsawatwong et al. [53], which found increased flexion, axial rotation, and lateral bending ROM following bilateral FJV. However, their methodology also involved rongeur preparation of the entry point, compromising the facet capsule and cortex before screw insertion—the authors did not quantify the extent to which screws invaded the joint. This raises the question about whether the observed instability was due to capsular disruption rather than screw placement alone. These findings suggest that FJV may represent two distinct pathological entities. Minor violations, involving only the facet cortex, may cause instability in flexion and rotation due to capsular disruption. In contrast, severe violations, where the screw fully transects the articular surface, may function as an extension of the construct, akin to transfacet fixation, and could theoretically stabilize the superior level in the acute setting. This hypothesis is supported by the finding by Xu et al. [11] that severe bilateral FJV significantly reduced motion in the superior segment.

Interestingly, the review of ASD by Shin et al. [56] following MIS and open TLIF found that MIS was associated with greater segmental instability. Their use of freehand percutaneous screw placement likely resulted in a high incidence of FJV, predominantly grade 1 or 2 violations that penetrate the cortex and capsule while sparing the articulation. This aligns with the notion that minor FJV may be destabilizing, whereas more extensive violations could confer mechanical support. If this conceptualization is correct, varying grades of FJV may have disparate effects on fusion rates and ASD.

There is a lack of robust evidence supporting the assumption that FJV increases forces across the cranial disc and facet joints. The primary study suggesting this, i.e., the finite element model by Kim et al. [12], has not been externally validated in animal, cadaveric, or human studies. Similarly, the assessment by Wangsawatwong et al. [53] of surface strain alterations following FJV failed to reach statistical significance, limiting definitive conclusions.

Clinical evidence implicating FJV as a significant contributor to ASD remains inconclusive. There is ongoing debate about whether ASD is a distinct pathological entity or just the natural progression of degenerative disc disease [8]. While sagittal alignment is often cited as a key factor in ASD, the evidence is inconsistent, with several studies failing to replicate significant associations [8,62,63]. Five out of six studies included in this review reported a significant relationship between FJV and ASD, but methodological limitations need to be considered. Two studies contained clear calculation or reporting errors [56,58], while the results of Levin et al. [10] were likely compromised due to selection bias. Only one study employed a multivariate model, controlling for sagittal balance, preexisting disease, and FJV [55]. While this constitutes the strongest evidence supporting FJV as a risk factor for rASD, it does not demonstrate a causal linkage with symptomatic disease. Reverse causation should also be considered: it is plausible that preexisting facet arthropathy distorts joint architecture, leading to FJV while also constituting the initial insult forbearing ASD.

Despite frequent references to FJV as a risk factor for ASD, this remains a theoretical concern requiring further validation. Future research should prioritize biomechanical studies to determine whether FJV alters force transmission in the superior segment, validating or refuting the finite element model by Kim et al. [12]. Experimental studies should investigate how different degrees of FJV impact spinal stability, particularly whether minor cortical violations cause instability in flexion and rotation, and whether full joint penetration stabilizes the superior segment. Further research should explore how different grades of FJV affect screw pullout strength, particularly in elderly and osteoporotic cohorts, and its impact on fusion development. Clinical studies should investigate superior segment degeneration in FJV and non-FJV cohorts, controlling for baseline degeneration and sagittal balance measures. Importantly, studies should differentiate between FJV involving only the cortex and FJV involving complete penetration of the articular surface to further improve their contribution to the discourse.

Conclusions

FJV is a common issue in clinical practice, especially with freehand percutaneous PS placement. Published reports suggest screws placed into the second lumbar vertebra are more likely to violate the facet than those placed in more caudal lumbar levels, which may be all of largely equivalent risk. Biomechanical results suggest that minor FJV (cortex and capsule violation) has a net destabilizing effect on the upper vertebra, while more severe FJV (transecting the joint articulation) might have a stabilizing effect. Limited evidence indicates that facet violation may increase force transmission through the cranial disc and facet joints, but a causal link to ASD has not been established. Clinical studies evaluating the in vivo effect of FJV on the superior segment are rare and limited by possible selection bias and reporting errors. Further research is necessary to elucidate the biomechanical and clinical consequences of both minor versus severe FJV on the superior adjacent segment before FJV can be confidently cited as a significant contributor of ASD.

Key Points.

• Facet joint violation (FJV) is most common after freehand percutaneous screw placement and lowest after robotic assisted placement.

• Most violations are minor cortical breaches, but >6% may fully or partially traverse the articular surface.

• Minor FJV may destabilise the cranial segment while severe FJV may acutely stabilise it.

• Most studies report an association between FJV and adjacent segment disease but there is a significant risk of bias.

• Future studies should validate if FJV increases loading of the superior segment and should distinguish between minor and severe violations.