Facet joint loading after Kevlar-reinforced disc replacement vs. lumbar fusion. A cadaveric model study

Abstract

Background

Lumbar spinal fusion, while effective for degenerative disc disease, eliminates motion and may accelerate adjacent segment degeneration. Total disc replacement (TDR) with anterior annular reinforcement may offer a motion-preserving alternative.

Objective

To compare spinal kinematics and facet joint pressures between (1) total lumbar disc replacement with anterior Kevlar band reinforcement (TDR group) and (2) standard 360° lumbar fusion (fusion group) in a cadaveric model.

Methods

Six fresh-frozen human lumbar spines (L2–S1) were tested in a custom motion simulator under both surgical conditions at L4–L5. Kinematic parameters included range of motion (ROM), peak angles, and motion harmony (angle–acceleration correlation). Zygapophyseal joint pressures were measured using Tekscan sensors. Data were analyzed using repeated-measures ANOVA.

Results

The TDR group preserved significantly greater flexion (−25.7° vs −21.8°, p < 0.001) and left axial rotation (p = 0.025), while the fusion group showed greater extension (p < 0.001). TDR demonstrated superior motion harmony across all axes (p < 0.01) and better preserved ROM at adjacent levels. No significant differences in mean maximum facet joint pressures were observed (p > 0.05), though slightly higher values were noted in TDR during axial rotation.

Conclusion

Anterior annular reconstruction with Kevlar bands during total disc replacement restores near-physiological motion and improves kinematic control without increasing facet joint stress. This technique may offer a biomechanically favorable alternative to lumbar fusion.

Graphical abstract

1. Introduction

Degenerative lumbar disc disease is a leading cause of chronic low back pain and disability, especially in aging populations.¹ Lumbar fusion (LF), the traditional surgical treatment, eliminates motion at the operated segment and increases stress at adjacent levels, contributing to adjacent segment degeneration (ASD).^{2, 3, 4, 5}

Total disc replacement (TDR) was developed to preserve motion, maintain disc height, and reduce biomechanical (BM) stress on adjacent segments.^{6, 7, 8} Studies show that TDR maintains near-normal intradiscal pressure and facet joint loading, unlike fusion, which often increases both.^9,10 However, TDR may elevate facet joint (ZJ) stress at the index level, particularly during extension and axial rotation.^{11, 12, 13} These effects depend on the implant design, placement, and alignment.^14,15

Moreover, during most TDR procedures, the anterior longitudinal ligament and a large portion of the annulus fibrosus must be excised to allow prosthesis insertion. This resection disrupts the natural restraint mechanisms of the spine, particularly those related to axial rotation and extension.^16,17 These anterior structures play a crucial role in maintaining segmental stability and controlling motion.¹⁸

Most TDR procedures require resection of the anterior longitudinal ligament (ALL) and the anterior annulus fibrosus, compromising the natural restraint mechanisms critical for stability and motion control.^{16, 17, 18} To address this, synthetic materials such as Kevlar have been proposed for reconstructing the anterior column. Kevlar bands mimicking the annular lamellae have shown promise in restoring motion harmony and near-physiological kinematics.¹⁹

Yet, the impact of such reinforcement on ZJ pressure remains unclear. This study compares spinal kinematics and ZJ pressures in cadaveric spines treated with (1) TDR reinforced by anterior Kevlar bands and (2) standard 360° LF. We hypothesize that Kevlar reinforcement improves motion control and reduces abnormal ZJ stress compared to fusion.

2. Materials

2.1. Study design and specimens

This cadaveric BM study used six fresh-frozen human lumbosacral spine specimens (L2–S1) from the University of Valencia. Donor ages ranged from 37 to 50 years (mean 43.3 ± 6.1). All specimens were white Spanish individuals, reflecting the local availability of donations, which may limit generalizability. Strict inclusion criteria were applied. Disc quality was assessed using the Pfirrmann grading system, with grades I–II at all levels except L4–L5 (grades III–IV were allowed). This selection ensured consistent baseline disc integrity and minimized degenerative variability. Surgical sequence and mechanical preconditioning were standardized. Specimens with prior spinal surgery, trauma, malignancy, osteoporosis (DEXA Z-score < −2.5),²⁰ or inflammatory disease were excluded. MRI confirmed disc status and absence of facet OA (Pathria grade 1).^21,22

CT (slice thickness, 0.625 mm) and 1.5T MRI were used to assess vertebral and disc morphology. Specimens were stored at −25 °C and thawed before preparation.

2.2. Test conditions

Each specimen was tested in two sequential surgical configurations at the L4–L5 level (Supplementary Fig. 1):

• •KC condition: TDR with anterior Kevlar band reinforcement in an X-shape.
• •X condition: 360° LF with PEEK cage and pedicle screws.

Details are in Supp. Table 1. All procedures were performed by the senior author (VV), a board-certified spine surgeon, ensuring consistency and minimizing operator bias. The testing order was fixed: KC first, then X. A standardized preconditioning protocol (5 reps per axis) and a 30-min washout period were used to minimize viscoelastic and hysteresis effects.

2.3. Implant details

The TDR implant consisted of CoCrMo endplates with a titanium coating for osseointegration and a polycarbonate urethane (Bionate 80A) core for dual articulation and shock absorption.^23,24

2.4. Experimental setup

Testing utilized a custom motion simulator (Instituto de Biomecánica de Valencia, IBV), which enabled flexion-extension, lateral bending, and axial rotation via motor-driven torque (±8 N m). Motion was recorded at 100 frames per second (fps) using a FLIR camera and ChArUco markers. Tracking accuracy was refined via bundle adjustment algorithms (positional error <0.14 mm, angular error <0.25°).²⁴

2.5. Kinematic measurements

Five optical markers were placed on L3, L4, L5, and the fixed ends (L2, sacrum). Variables included (Supplementary. Table 1):

• •Maximum/minimum angles per axis.
• •Range of motion (ROM): 95th – 5th percentile of angular displacement.
• •Motion harmony (MH): Pearson correlation between angle and angular acceleration (values near −1 = higher harmony).

Each cycle consisted of 8 reps at 10°/s, followed by 5 conditioning cycles.²⁵

2.6. Facet joint pressure measurement

Six Tekscan A201-100 sensors (max 445 N, Ø9.53 mm) were implanted bilaterally in L3–S1 ZJs. Sensors were fixed with plastic fins and bone screws post-capsulotomy (Fig. 1, Supp. Fig. 2). Outputs were calibrated to N using known loads per IBV protocols. Two biomedical engineers supervised all tests, verifying calibration and equipment status before each cycle. Any irregularities prompted recalibration. Pressure data were recorded for all motion cycles under both conditions.

Fig. 1.

Pressure sensor insertion protocol. (A) Tekscan pressure sensor; (B) sensor with plastic fins; (C) sensor placement in L3–S1 facets; (D) fixation using cancellous bone screws.

2.7. Data analysis

Repeated-measures ANOVA compared BM variables using GNU Octave, Python, and R (Deducer GUI) [Fellows, 2012]. Significance: p < 0.05. Mean ± SD for ROM, MH, and max ZJ forces are reported. Error bar graphs show 95 % CI (Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Supp. Fig. 3–11; Table 1, Table 2, Table 3, Supp. Table 2–14).

Fig. 2.

Error bar plots showing flexion-extension kinematics (ROM, flexion, extension, harmony) with 95 % CI across KC and X conditions. KC: Kevlar-reinforced disc replacement; X: 360° lumbar fusion.

Fig. 3.

Error bar plots of lateral bending kinematics with 95 % CI in KC vs. X groups. KC: Kevlar-reinforced disc replacement; X: 360° lumbar fusion.

Fig. 4.

Error bar plots of axial rotation angles and harmony across spinal levels.

Fig. 5.

Mean torque-angle curves for flexion-extension: negative values denote flexion, positive values denote extension.

Fig. 6.

Mean torque-angle curves for lateral bending: negative for left bend, positive for right.

Fig. 7.

Mean torque-angle curves for axial rotation: negative for left rotation, positive for right.

Table 1.

Descriptive statistics of whole-spine flexion-extension (mean, SD for ROM, flexion, extension, harmony). KC: Kevlar-reinforced disc replacement; X: 360° lumbar fusion.

Variable	Condition	Mean	SD	Maximum	Minimum	p-value
EXTENSION (°)	KC	15.553	7.411	24.728	6.772	0.000
	X	19.217	5.126	27.259	14.418
FLEXION (°)	KC	−25.714	6.055	−16.868	−30.351	0.000
	X	−21.773	10.402	−5.569	−31.983
RANGE (°)	KC	41.267	3.365	45.546	36.436	0.812
	X	40.991	6.077	48.323	32.828
HARMONY	KC	−0.227	0.058	−0.156	−0.292	0.005
	X	−0.201	0.028	−0.168	−0.235

Table 2.

Repeated-measures ANOVA results comparing KC and X for flexion-extension. KC: Kevlar-reinforced disc replacement; X: 360° lumbar fusion.

Variable	Condition	Mean	SD	Maximum	Minimum	p-value
EXTENSION (°)	KC	15.553	7.411	24.728	6.772	0.000
	X	19.217	5.126	27.259	14.418
FLEXION (°)	KC	−25.714	6.055	−16.868	−30.351	0.000
	X	−21.773	10.402	−5.569	−31.983
RANGE (°)	KC	41.267	3.365	45.546	36.436	0.812
	X	40.991	6.077	48.323	32.828
HARMONY	KC	−0.227	0.058	−0.156	−0.292	0.005
	X	−0.201	0.028	−0.168	−0.235

Table 3.

Intersegmental ROM (flexion-extension) at L2–S1 in KC vs. X. KC: Kevlar-reinforced disc replacement; X: 360° lumbar fusion.

	Condition	Mean	SD	Maximum (°)	Minimum (°)	p-value
L2-L3	KC	12.030	4.606	18.713	8.554	0.088
	X	10.933	2.552	15.142	8.970
L3-L4	KC	6.838	0.727	7.939	6.105	<0.001
	X	7.776	0.936	10.121	6.262
L4-L5	KC	8.086	3.236	12.710	5.136	0.089
	X	9.645	4.935	16.866	2.998
L5-S1	KC	15.368	6.857	24.859	9.334	0.001
	X	13.134	3.545	18.043	9.645

2.8. Ethical considerations

All specimens were donated with prior consent under protocols approved by the University of Valencia. PRIM, S.A. (Madrid, Spain) provided the LF implants (X condition) without financial compensation. The ADDISC TDR remains investigational and is not yet commercialized.

3. Results

3.1. Whole-spine kinematics

3.1.1. Flexion-extension

All specimens completed flexion–extension testing. Max flexion was greater in KC (−25.7° ± 6.1°) vs. X (−21.8° ± 10.4°, p < 0.001). Extension was greater in X (19.2° ± 5.1°) vs. KC (15.6° ± 7.4°, p < 0.001). ROM did not differ significantly (p = 0.812) (Table 1, Fig. 2).

MH was better in KC (−0.23 ± 0.06) vs. X (−0.20 ± 0.03, p = 0.005), indicating smoother motion (Table 2).

3.1.2. Lateral bending

Right bending was greater in KC (−14.9° ± 3.9°) vs. X (−10.2° ± 3.5°, p < 0.001); left bending showed no difference (p = 0.267). ROM was similar (p = 0.812) (Supp. Table 3, Fig. 3).

MH was better in X (−0.38 ± 0.03) vs. KC (−0.30 ± 0.14, p = 0.005), possibly due to increased stiffness in fused specimens.

3.1.3. Axial rotation

Left rotation was greater in KC (10.4° ± 4.6°) vs. X (9.8° ± 4.8°, p = 0.025). Right rotation and total ROM showed no significant difference (p > 0.05). MH favored KC (−0.60 ± 0.01) vs. X (−0.58 ± 0.01, p < 0.001) (Supp. Table 4, Fig. 4).

3.2. Segmental ROM

Segmental analysis revealed that (Supplementary. Table 5–6, Supplementary. Fig. 3–7):

• •The L3–L4 and L5–S1 segments showed significantly different ROM in flexion-extension between conditions (p < 0.001 and p = 0.001, respectively), with more motion preserved in KC (Table 3, Fig. 5).
• •All levels showed significant differences in ROM in lateral bending (Supplementary. Table 3 and 12, Fig. 6). Notably, L3–L4 and L5–S1 were more mobile in KC.
• •In axial rotation, ROM at L3–L4 and L4–L5 was greater in KC compared to X (p ≤ 0.002), while L5–S1 also showed significant differences (p < 0.001) (Supp. Table 4, Fig. 7).

3.3. Torque-angle curves

KC allowed more motion in flexion and axial rotation under equal torque.

X showed stiffer behavior, especially in extension and lateral bending (Fig. 5, Fig. 6, Fig. 7).

3.4. Zygapophyseal joint pressure

Across all planes, no significant differences in mean max ZJ pressure between KC and X (p > 0.05) (Supp. Table 6–14, Figs. 6–11).

Slightly higher pressures in KC during axial rotation:

• •Left: KC 26.6 ± 31.3 N vs. X 24.3 ± 29.9 N
• •Right: KC 22.5 ± 30.2 N vs. X 21.1 ± 29.4 N

Sensor-specific data showed high inter-specimen variability and asymmetrical loading. Sensors at L4–L5 (positions 5 and 6) recorded the highest loads during axial rotation in both conditions.

3.5. Summary of key findings

• •KC preserved more motion in flexion and left axial rotation.
• •KC showed better MH across all axes.
• •X increased extension but reduced motion smoothness.
• •KC preserved mobility at adjacent levels.
• •ZJ loads were similar, with slightly higher values in KC during rotation.

4. Discussion

This BM study provides comparative insights into spinal motion dynamics and ZJ pressures following two surgical interventions at L4–L5: TDR using the ADDISC prosthesis with anterior Kevlar reinforcement (KC) and standard 360° lumbar fusion (X). KC preserved segmental mobility, improved MH, and did not increase ZJ pressures significantly.

4.1. Motion restoration and control

Our results affirm that KC maintains ROM closer to physiological norms compared to LF. KC specimens showed superior motion in flexion and left axial rotation, while X showed increased extension. These findings align with prior cadaveric and computational studies emphasizing the importance of preserving motion to prevent BM imbalances in adjacent segments.^9,10

One of the most notable findings is the increased MH in KC, defined as a more linear and predictable correlation between angular displacement and acceleration. This smoother motion profile is significant because motion quality—not just ROM—is increasingly recognized as a marker of spinal health and prosthesis success.¹⁶

Segmental analysis confirmed that mobility was better preserved at adjacent levels (L3–L4 and L5–S1) in KC. This fact is essential because LF has been repeatedly linked to ASD, which may result from altered motion patterns and increased BM loading on unfused segments.^3,5

While MH is a BM parameter, its clinical relevance lies in reflecting the smoothness and predictability of spinal movement. A stronger negative correlation between angular displacement and acceleration (values near −1) indicates more harmonic motion, which may reduce abrupt loading and microinstability. This smoother kinematic behavior could lower BM stress on adjacent segments and ZJs, potentially improving long-term outcomes and reducing the risk of ASD or prosthesis-related complications.

Although KC showed overall improvement in MH, asymmetries in lateral bending and axial rotation—especially greater right bending and left rotation—could reflect subtle anatomical variations, such as ZJ orientation or ligament tension. The retroperitoneal approach and Kevlar band placement may have introduced minor directional bias in reinforcement tensioning. These factors, along with the limitations of cadaveric models, may explain the discrepancies. Importantly, these asymmetries did not result in significant differences in ZJ loading, suggesting Kevlar reinforcement maintained balanced load distribution despite directional variability.

4.2. Facet joint load distribution and implant safety

Contrary to concerns from earlier studies about TDR-induced ZJ overload,^11,12 our results showed no significant increase in ZJ pressures in KC vs. X across any motion plane. Although KC showed slightly elevated pressures during axial rotation, these differences were not substantial and likely reflect a more physiologic load distribution through the preserved motion segment.

ZJ integrity plays a key role in load transmission, especially during rotation and extension. In most TDR systems, resection of the ALL and anterior annulus shifts the motion control burden posteriorly.¹⁴ Re-establishing anterior column restraint with crossed Kevlar bands in KC may redistribute forces more evenly, avoiding posterior overload while enabling controlled mobility.

The pressure sensors provided quantitative data supporting the hypothesis that anterior annular reconstruction can maintain ZJ protection. These findings align with other experimental models, which show that restoring anterior restraint reduces non-physiologic kinematics.^17,18

Notably, sensors at L4–L5 consistently registered the highest ZJ pressures during axial rotation in both conditions. This fact likely reflects the BM prominence of the index level, where surgery directly alters load transmission and motion dynamics. Elevated loads at L4–L5 may stem from its central role in segmental rotation and disruption of native stabilizers. These findings highlight the importance of reconstructing anterior restraint to modulate posterior joint stress at the operated level.

4.3. Biomechanical relevance of anterior Kevlar reinforcement

Lumbar spine stability depends not only on bony and ZJ alignment but also on the integrity of annular and ligamentous structures. The ALL and annulus fibrosus act as primary stabilizers during hyperextension and axial rotation.²⁶ In TDR, these are typically excised, removing a key stabilizing component. Kevlar bands in an X-shape aim to replicate the native lamellar annulus architecture, which resists torsional and flexural loads.

This study supports the concept biomechanically: Kevlar reinforcement was linked to reduced kinematic irregularities and did not increase ZJ load. Prior in vivo work has confirmed Kevlar's biocompatibility and mechanical resilience in orthopedic applications.¹⁹

Importantly, this approach addresses a known deficiency in current TDR systems—namely, the inability to reconstruct the anterior tension band without adding excessive rigidity, as seen in LF. Functional anterior reinforcement may represent a paradigm shift in TDR design, enabling surgeons to restore anterior stability dynamically rather than statically.

4.4. Clinical translation and future directions

Although our results are derived from a cadaveric model, they provide strong preclinical evidence supporting the viability of annular reconstruction during TDR. In clinical practice, a key barrier to adoption may be the technical challenge of deploying Kevlar bands in the retroperitoneal space, especially near major vessels. However, advances in minimally invasive techniques and robotic assistance may facilitate such procedures.

Our study also highlights the potential to combine prosthetic design evolution with bio-inspired reinforcement strategies. Instead of viewing prosthesis motion and stability as opposing forces, future TDRs might integrate elastic anterior structures capable of modulating segmental BM.

Regarding intraoperative feasibility, the insertion of a Kevlar band is safe and practical. The staple deployment instrument is compact and ergonomic, allowing precise placement via a retroperitoneal approach. In our clinical practice, we routinely collaborate with a vascular surgeon to access lumbar IVDs and manage any vascular complications. This teamwork ensures secure anterior spine exposure and minimizes the risk of injury near major vessels. While anterior access requires specialized surgical expertise, it does not introduce additional barriers beyond those of standard procedures.

4.5. Limitations

This study has limitations. The small sample size (n = 6), despite statistical adjustments, limits the generalizability of the findings. Cadaveric models lack biological remodeling and muscular dynamics seen in vivo. MH, while useful biomechanically, is a proxy and should be linked to clinical outcomes in future research.

Findings apply specifically to the unconstrained TDR with Kevlar reinforcement. Other implant designs may behave differently, so the results may not be applicable to all TDR systems. Future studies should assess how implant type affects ZJ loading and motion preservation. Ongoing research includes comparisons with standalone TDR to clarify the role of anterior reinforcement.

4.6. Clinical significance

This study highlights the biomechanical advantages of combining total disc replacement with anterior Kevlar reinforcement. By restoring anterior column integrity and preserving near-physiological motion, this technique may reduce adjacent segment degeneration and avoid excessive facet joint loading.

If validated clinically, it could offer spine surgeons a motion-preserving alternative to fusion, potentially improving long-term outcomes and reducing the need for revision surgery.

5. Conclusions

This biomechanical study shows that anterior annular reinforcement with crossed Kevlar bands, combined with the ADDISC total disc prosthesis, preserves physiological spinal motion, enhances kinematic harmony, and maintains facet joint loading within normal ranges compared to 360° lumbar fusion.

Kevlar-reinforced disc replacement allowed greater flexion and axial rotation, particularly at adjacent levels, without increasing zygapophyseal joint stress. These findings support the potential of anterior ligamentous reconstruction in TDR to restore natural spinal biomechanics.

Further in vivo studies are needed to confirm clinical applicability and long-term outcomes.

Guardian/patient consent

This study was conducted using cadaveric human lumbar spines provided by the Department of Anatomy of the Medical School of the University of Valencia, Spain. All specimens were donated with prior consent for scientific use in accordance with institutional ethical protocols.

Credit author statement

Vanaclocha, Amparo: Conceptualization, Methodology, Investigation, Writing – Review & Editing.

Vanaclocha, Vicente: Conceptualization, Supervision, Writing – Original Draft, Writing – Review & Editing.

Atienza, Carlos: Methodology, Investigation, Data Curation.

Jorda-Gomez, Pablo: Software, Validation, Formal Analysis.

Peñuelas, Andres: Resources, Investigation.

Gomez-Lopez, David: Visualization, Data Curation.

Saiz-Sapena, Nieves: Writing – Review & Editing, Project Administration.

Vanaclocha, Leyre: Supervision, Writing – Original Draft, Writing – Review & Editing.

Ethical approval and consent statement

All cadaveric specimens used in this study were donated with prior consent for scientific use. The study protocol was reviewed and approved by the institutional ethics committee at the University of Valencia. All procedures were conducted in accordance with relevant guidelines and regulations governing the ethical use of human anatomical material for research purposes.

Funding statement

This research was conducted using resources provided by the Instituto de Biomecánica de Valencia (IBV) and the Department of Anatomy at the University of Valencia. No external funding was received for this study. All cadaveric specimens were donated with prior consent for scientific use in accordance with institutional ethical protocols.

Conflict of interest statement

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this manuscript.

Appendix A. Supplementary data

The following are the supplementary data to this article:

Supplementary Table 1

Description of experimental test conditions (KC: ADDISC + Kevlar; X: 360° fusion).

Supplementary Table 2

Interpretation of max/min kinematic values per axis.

Supplementary Table 3

ANOVA for lateral bending variables of the whole spinal specimen.

Supplementary Table 4

ANOVA for axial rotation of the whole spinal specimen.

Supplementary Table 5

ANOVA results for measurements of the inter-segmental range of motion in lateral bending.

Supplementary Table 6

ANOVA results for measurements of the variable range of inter-segmental motion in axial rotation.

Supplementary Table 7

Overall facet joint pressures (N) across all movements.

Supplementary Table 8

Sensor-specific joint pressures (N) for all motion planes.

Supplementary Table 9

Effect of study conditions on the maximum strength recorded in the zygapophyseal joints during extension motion.

Supplementary Table 10

Effect of the study conditions on the maximum force recorded in the zygapophyseal joints for the flexion movement.

Supplementary Table 11

Effect of study conditions on maximum strength recorded in the zygapophyseal joints for the right lateral bending movement.

Supplementary Table 12

Effect of study conditions on maximum strength registered in the zygapophyseal joints for the left lateral bending movement.

Supplementary Table 13

Effect of the study conditions on the maximum force recorded in the zygapophyseal joints for the left rotational movement.

Supplementary Table 14

Effect of study conditions on maximum force registered in facet joints for the right axial rotation motion.

Supplementary Figure 1.

Surgical setup and instrumentation for biomechanical testing. (A) Kevlar-reinforced ADDISC disc replacement (KC condition); (B, C) 360° lumbar fusion (X condition). PEEK cage in superior surface view (D), front view (E), and side view (F).

Supplementary Figure 2.

Schematic illustration of zygapophyseal pressure sensor placements (positions 1–6).

Supplementary Figure 3.

Error bar diagram with the confidence interval of the mean value of the range (°) of movement between segments in flexo-extension movement.

Supplementary Figure 4.

Error bar diagram with the confidence interval of the mean value of the range (°) of the movement between segments in lateral bending motion.

Supplementary Figure 5.

Error bar diagram with the confidence interval of the mean value of the range (°) of the movement between segments in the rotational motion.

Supplementary Figure 6.

Mean maximum zygapophyseal joint pressure (N) during extension movement in KC vs. X, with 95 % confidence intervals.

Supplementary Figure 7.

Error bar diagram with the confidence interval of the mean value of the force in the zygapophyseal joints in the flexion movement.

Supplementary Figure 8.

Error bar diagram with the confidence interval of the mean value of the force in the zygapophyseal joints in the right lateral bending movement.

Supplementary Figure 9.

Error bar diagram with the confidence interval of the mean value of the force in the zygapophyseal joints in the left lateral bending movement.

Supplementary Figure 10.

Error bar diagram with the confidence interval of the mean value of the force in the zygapophyseal joints in the right axial rotation movement.

Supplementary Figure 11.

Error bar diagram with the confidence interval of the mean value of the force in the zygapophyseal joints in the left axial rotation movement.