Introduction
Patients with diffuse idiopathic skeletal hyperostosis (DISH) often develop spinal instability after fractures due to ankylosis and bone fragility. Because conventional fixation frequently fails in poor-quality bone, long-range fusion is typically recommended [1]. We developed a novel percutaneous pedicle screw (PPS) insertion technique, known as the single or double endplate penetrating screw (SEPS/DEPS) technique, and have reported its effectiveness. This technique enhances fixation strength by penetrating the endplate, the cortical bone of the vertebral body [2]. Our clinical studies found that the pedicle screw (PS) insertion torque (IT) with DEPS was approximately 1.34 times higher than the traditional trajectory (TT) [2]. Our finite element method (FEM) study further demonstrated that the SEPS trajectory (SEPST) provided approximately 1.4 times higher pullout strength (POS) than TT, and the DEPS trajectory (DEPST) provided approximately 1.7 times higher POS than TT [3]. Additionally, our cadaver validation studies demonstrated that SEPST/DEPST achieved 32%–70% higher IT and 45%–82% higher POS than TT [4]. Our previous FEM studies were largely consistent with the results of these cadaver studies, supporting the reliability of FEM analysis.
However, additional indications for the use of the DEPS technique have been identified which has expanded its use to spinal disorders other than DISH. For instance, distal screws are inserted with TT or SEPS when fixation extends into the lumbar spine because DEPS is not suitable for mobile segments. This may increase the risk of implant failure (IF) due to the mechanical demands placed on the mobile fixation points [2,5].
We focused on three key aspects to identify methods for maximizing the fixation strength of SEPS/DEPS. First, the optimal screw diameter for DEPS remains unknown [2]. Second, while longer screws generally improve fixation [6–8], DEPS already uses long screws, and further extension may cause bilateral tip interference or anterior wall penetration. Targeting the stronger lateral rim of the endplate with shorter screws may be advantageous [9]. Third, we hypothesized that a screw with smaller pitch and larger core diameter, an endplate thread screw (ETS), might be optimal based on the study of Weidling et al. [10]. However, our previous cadaver study did not show a significant advantage [4].
Accordingly, this study aimed to identify the optimal SEPST/DEPST modifications to maximize fixation strength using FEM. We conducted three analyses: (1) screw diameter optimization in DEPST, (2) a fixation comparison between conventional DEPS trajectories (C-DEPST) and short DEPST (S-DEPST), and (3) a fixation strength comparison between conventional cancellous thread screw (CTS) and ETS.
Materials and Methods
This study was conducted with the approval of the Ethical Committee of Kyorin University School of Medicine (R03-188). Written informed consents were obtained from all participants in this study.
Finite element model creation
The finite element (FE) bone model used in this study was based on our previously reported model [3]. A total of 10 FE models of the T12 and L1 vertebrae were constructed from computed tomography (CT) DICOM data of 10 osteoporotic patients (two males, eight females; mean age, 74.7±4.6 years; range, 70–87 years), each with femoral neck bone mineral density <70% (mean 62.4%±5.6%), measured using dual-energy X-ray absorptiometry. CT scans were obtained using a Revolution EVO scanner (GE Healthcare, Chicago, IL, USA) with a 64-row detector, 120 kV, Smart mA (80–400 mA), noise index 20, 0.625 mm slice thickness, and 0.35 mm pixel width. CT data were processed in Mechanical Finder ver. 12.0 (Research Center of Computational Mechanics Inc., Tokyo, Japan) to generate three-dimensional FE models. The mesh consisted of first-order tetrahedral elements, with a global mesh size <1 mm and a refined mesh of 0.075 mm around the screw threads to capture local stress distribution. The mesh density (450,000–650,000 solid elements, 80,000–140,000 nodes, 50,000–60,000 shell elements) was determined based on a mesh convergence study to balance computational cost and numerical accuracy. Material properties were assigned based on the CT Hounsfield units, and Young’s modulus was calculated using Keyak’s method [3,11,12]. The Poisson ratio was set at 0.4. Shell elements representing the cortical bone were modeled with a Young’s modulus of 10 GPa. The screws were modeled as a linear elastic titanium alloy (Ti6Al4V) with a Young’s modulus of 108,853.8 MPa, a yield stress of 824.7 MPa, a critical stress of 899.3 MPa, and a Poisson ratio of 0.28. The T12–L1 intervertebral disc was excluded, and facet joints were modeled as fused. Pedicle screw models were based on the Saccura screw (TEIJIN, Osaka, Japan), with a diameter of 6.5 mm and length of 55 mm. The ETS had a thicker core and smaller pitch, based on the design described by Weidling et al. [10] (Fig. 1, Table 1). The screw was inserted into the vertebral model using computer-assisted modeling techniques. Contact conditions were defined on the screw–bone interfaces, and the coefficient of friction was set to zero based on previous studies [11]. A static nonlinear analysis was performed to simulate loading conditions.
For the pullout strength (POS) and four-directional loading tests (4DLT), the upper endplate of T12 and the lower endplate of L1 were fully constrained. In the lower instrumented vertebra (LIV) model, the upper endplate of T12 and the bilateral screw heads were constrained to simulate anchorage.
Analysis 2: Comparison of fixation strength between C-DEPST and S-DEPST
The C-DEPST model used a 6.5-mm screw from Analysis 1. The S-DEPST model involved a straight screw trajectory penetrating the posterolateral endplate without inward tilt (Fig. 2). The screw insertion angles in the sagittal, coronal, and axial planes, and the screw lengths within the vertebral body were measured. The cranial endplate of T12 and the caudal endplate of L1 were fully restrained for POS evaluation, and a pullout force was applied along the screw axis (Fig. 3A). For the 4DLT, the screw tip was loaded cranially, caudally, medially, and laterally to assess directional strength (Fig. 3B) [3]. The failure point was defined as the point at which the load–displacement curve changed abruptly. To evaluate motion suppression at the LIV, we inserted bilateral screws at L1 using both C-DEPST and S-DEPST. The cranial endplate of T12 and the screw heads were restrained. A crossbar was attached to the L1 caudal endplate through a disc-like structure. A 7.5 nm moment was applied to the crossbar to simulate flexion/extension (Fig. 3C), lateral bending, and axial rotation (Fig. 3D) [3,13]. The vertebral motion angle (VMA) was recorded.
Analysis 3: Comparison of fixation strength between CTS and ETS
CTS and ETS with diameters of 6.5 mm were placed in the C-DEPST model created in Analysis 1, and POS and LIV VMA were compared.
Statistics
Differences between groups were tested using the Mann-Whitney U test for two-group comparisons and one-way analysis of variance for multiple groups. Statistical analysis was conducted using EZR (https://www.r-project.org/), a modified version of R Commander with additional statistical functions [14].
Results
Analysis 1: Comparison of POS with different screw diameters in DEPST
The POS values for each screw diameter were as follows: 292.6±89.3 N with a 4.5 mm screw, 548.3±206.7 N with a 5.5 mm screw, 738.3±227.5 N with a 6.5 mm screw, and 999.1±271.4 N with a 7.5 mm screw. The differences among the four screw sizes were significant, with larger diameters exhibiting higher POS (p<0.01) (Fig. 4).
Analysis 2: Comparison of fixation strength between C-DEPST and S-DEPST
The screw insertion angles of C-DEPST and S-DEPST were as follows: sagittal plane (C-DEPST/S-DEPST, 33.8°±4.1°/46.8°± 3.7°), coronal plane (C-DEPST/S-DEPST, 23.5°±5.3°/1.4°±2.2°), and axial plane (C-DEPST/S-DEPST, 16.1°±3.3°/1.5°±2.3°), with statistically significant differences in all planes (p<0.01). The length of screw insertion within the vertebral body was 23.8±1.9 mm for C-DEPST and 17.6±2.5 mm for S-DEPST, with S-DEPST being significantly shorter (p<0.01) (Fig. 5A).
The POS values were 738.3±227.5 N in C-DEPST and 1,081.1±248.3 N in S-DEPST, with S-DEPST showing significantly higher fixation strength (p<0.01) (Fig. 5B). In the 4DLT, the C-DEPST and S-DEPST values were as follows: cranial (C-DEPST/S-DEPST, 3,384±2,000 N/7,443±3,999 N), caudal (C-DEPST/S-DEPST, 3,191±1,812 N/6,722±4,198 N), medial (C-DEPST/S-DEPST, 2,798±1,923 N/5,954±2,493 N), and lateral (C-DEPST/S-DEPST, 2,351±1,337 N/5,420±2,037 N). S-DEPST demonstrated significantly higher fixation strength in all directions (p<0.05) (Fig. 5C). However, there were no significant differences in the VMA of LIV in all directions: flexion (C-DEPST/S-DEPST, 0.46°±0.2°/0.64°±0.37°; p=0.18), extension (C-DEPST/S-DEPST, 0.45°±0.2°/0.49°±0.24°; p=0.72), lateral bending (C-DEPST/S-DEPST, 0.22°±0.16°/0.17°±0.09°; p=0.45), and axial rotation (C-DEPST/S-DEPST, 0.15°±0.07°/0.16°±0.06°; p=0.67) (Fig. 5D).
Analysis 3: Comparison of fixation strength between CTS and ETS
The POS values of 738.3±227.5 N (CTS) and 799.3±204.8 N (ETS) showed no significant difference (p=0.33) (Fig. 6A). The values of VMA of LIV were as follows: flexion (CTS/ETS, 0.46°±0.2°/0.44°±0.18°; p=0.92), extension (CTS/ETS, 0.45°±0.2°/0.43°±0.22°; p=0.91), lateral bending (CTS/ETS, 0.22°±0.16°/0.22°±0.1°; p=0.99), and axial rotation (CTS/ETS, 0.15°±0.07°/0.15°±0.04°; p=0.99), with no significant differences in any direction (Fig. 6B).
Discussion
We previously reported that the actual surgical IT of DEPST was approximately 1.34 times higher than that of TT, and that IF was also significantly lower in DEPST [2]. Recent Japanese studies support the effectiveness of endplate penetrating screws for DISH-related spinal disorders [5,15]. Our previous FEM and cadaver studies demonstrated that SEPST and DEPST significantly improved IT and POS compared with TT by 40%–71% in FEM [3] and by 1.3–1.8 times in cadaver tests [4], highlighting the consistency between both methods and supporting the reliability of FEM analysis.
The initial analysis demonstrated a positive correlation between screw diameter and the POS of SEPS/DEPS. Similarly, prior studies on TT demonstrated that POS increased with larger-diameter screws due to the higher ratio of outer screw diameter to inner pedicle diameter [16,17]. Furthermore, Matsukawa et al. [18] reported that larger-diameter screws provided a greater contact area with cortical bone, resulting in higher POS in cortical bone trajectory using the FEM study [18]. These findings suggest that the higher POS observed with larger-diameter SEPS/DEPS screws may be attributed to both the increased percent fill and the contact area with cortical bone. Our previous FEM study demonstrated a POS of 431.7±197.4 N for the 6.5-mm TT, while the current 4.5-mm SEPS/DEPS screw demonstrated 292.6±89.3 N, suggesting that at least a 5.5 mm screw is needed to achieve adequate strength.
Second, S-DEPST, specifically designed to target the endplate rim with higher bone strength, demonstrated 1.46 times higher POS than C-DEPST and 2.5 times higher POS than TT. Segami et al. [9] reported that the outer rim of the endplate is approximately 30% stronger than the center, with greater strength observed in the lower endplate. Based on these reports, we hypothesized that fixation strength would be greater if the screw penetrated the lateral rim of the endplate, even if the screw length within the vertebral body was shorter. This hypothesis was supported by the higher POS of S-DEPST in the current FEM analysis. However, no significant difference in VMA at LIV was observed, likely due to the shorter screw length within the vertebral body, which may have reduced the lever arm and limited motion suppression. This is similar to findings with CBT [19]. Furthermore, the sagittal plane insertion angle for S-DEPST exceeded 45°, which was larger than the 30° in C-DEPST. Because typical multi-axial screws allow a head movement range of approximately 30° [20], inserting PS with a steep angle of 45° could increase the risk of PS loosening or vertebral fracture during final rod fixation. Therefore, it is critically important to modify the insertion trajectory to maintain the sagittal angle within 30° when using currently available screw systems. Alternatively, the development of a novel pedicle screw system with an expanded head angulation range is strongly recommended to accommodate steep trajectories, such as those used in S-DEPST. Implementation of these solutions would significantly enhance the clinical feasibility and safety of this technique.
Finally, the impact of the screw design was evaluated. Previous studies on screw design have reported using polyurethane foam as a substitute for bone material [10,21,22], cadavers [10,23], or FEM [24–26]. Recently, dual-threaded screws with proximal cortical threads and distal cancellous threads have been widely used [16,27,28]. Weidling et al. [10] reported that reducing the pitch of screws increased the flank overlap area and contact surface in bone with low density, whereas larger core diameters enhance compaction in bone with high density. We hypothesized that ETS would be the most effective DEPS technique, particularly in osteoporotic vertebral bodies and cortical endplates. However, our FEM study failed to demonstrate any significant differences. Similarly, Weidling et al. [10] showed higher fixation strengths with larger core diameters and smaller pitches only with polyurethane foam, but not in cadaver specimens. Therefore, our FEM result may be due to the patient CT data used in this FEM model which was inhomogeneous, with partial osteosclerosis and degeneration similar to cadaver tissue, unlike the homogenous nature of polyurethane foam.
The current study has several limitations. First, we adopted a two-vertebrae model to evaluate the isolated mechanical effects of screw shape and insertion trajectory. However, this simplification limits the ability to fully replicate the complex load-sharing and biomechanical behavior observed in multilevel spinal fixation. Therefore, to improve clinical applicability, future studies should incorporate multi-level constructs, including multiple vertebrae, pedicle screws, and rods.
Second, S-DEPST involves extremely steep insertion angles of approximately 45° in the sagittal plane, which may be impractical during surgery. Third, this study focused on the initial fixation strength using static FEM analysis, including POS, 4DLT, and VMA. Clinically important failure modes, such as fatigue, toggling, and adjacent segment issues, were not assessed. These analyses are partially possible with advanced FEM techniques but were not feasible with the software used in this study. To improve clinical relevance, future studies should include cyclic loading or cadaver testing. Finally, the Keyak formula used in this study was derived from the femur, not the vertebra.
Conclusions
This study demonstrated that fixation strength increased with larger screw diameters in the DEPS technique. The optimal trajectory for the DEPS technique is to avoid inward tilting and to penetrate the posterolateral side of the endplate, even if the screw length within the vertebral body is shortened. Although there were no significant differences in POS and VMA between ETS and CTS, ETS may provide greater fixation strength and warrants further clinical validation.
Key Points
Larger screw diameters in double endplate penetrating screw trajectory (DEPST) significantly increased pullout strength.
Short DEPST (S-DEPST) showed 1.46 times higher fixation than conventional DEPST and 2.5 times higher than traditional trajectory.
No significant difference in vertebral motion angle between S-DEPST and conventional DEPST.
Endplate thread screws (ETS) demonstrated slightly higher fixation than cancellous thread screws, but without statistical significance.
Strong fixation can be achieved with larger diameters and S-DEPST, while the role of ETS requires further validation.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
