Abstract
Background:
Proximal femur fracture fixation fundamentally alters load transfer, which can trigger adverse bone remodeling, particularly after healing in osteoporotic bone. While previous studies have focused on early fixation stability or acute fracture healing, the long-term mechanical environment after fracture consolidation remains poorly understood. This study examines implant-specific strain behavior in the post-healed state to identify patterns that may predispose to late-stage implant failure or remodeling.
Methods:
Subject-specific finite-element models were created from CT scans of a healthy 22-year-old and an osteoporotic 90-year-old female. Cannulated screws (CS), a dynamic hip screw (DHS), and a femoral neck system (FNS) were virtually implanted in each model and subjected to simulated gait loading. Maximum implant stresses and regional principal strains across the proximal femur were quantified throughout the stance phase.
Findings:
The FNS demonstrated the highest maximum von Mises stresses in both normal and osteoporotic bone, notably at the anti-rotation screw and bolt interface. In osteoporotic bone, the FNS exhibited significantly higher tensile (mean 0.36 % ± 0.22 %) and compressive strains (mean − 0.41 % ± 0.28 %) compared to CS and DHS implants, while in normal bone, FNS strains were comparable to intact femur strains. The CS model showed reduced peak strains and stresses throughout the gait cycle compared to DHS and FNS.
Interpretation:
This study provides a comparative assessment of healed-state strain distributions across common fixation constructs. By characterizing these environments, these data establish a biomechanical framework and highlight the interplay between implant design and bone quality. Further population-based studies are required to refine implant selection.
Keywords: Subject-specific finite element analysis, Internal fixation, Osteoporotic bone, Gait loading
1. Introduction
Femoral neck fractures (FNFs) represent a critical public health challenge due to their high morbidity, mortality, and economic burden. In the United States, age-adjusted incidence rates persist at 793 per 100,000 in women and 369 per 100,000 in men (Brauer et al., 2009; Kazley and Bagchi, 2023; Rogmark et al., 2018). These fractures occur in a biomechanically vulnerable region where compressive forces during gait exceed 2.5 times body weight, creating an environment prone to fracture and secondary complications (Deng et al., 2021; Li et al., 2024). Internal fixation with cannulated screws (CS), dynamic hip screws (DHS), or a femoral neck system (FNS) remains the preferred treatment for stable or minimally displaced fractures, aiming to restore native anatomy as a prerequisite for achieving stable mechanical fixation and promoting healing (Ge et al., 2023).
Despite widespread use, no consensus exists on which fixation construct most effectively restores physiological load transfer after fracture healing, particularly in osteoporotic bone (Kamara et al., 2021; Surgeons, 2021; Wei et al., 2020). This uncertainty is clinically significant because fixation devices inherently alter the mechanical environment of the proximal femur, and these effects may be amplified in osteoporotic bone, where reduced mineral density compromises structural integrity. Persistent alterations in strain after healing can promote bone remodeling and stress shielding. Prior clinical and experimental studies have linked these late-stage mechanical changes to localized bone resorption, periprosthetic fracture risk, and long-term implant loosening (Eijer et al., 2001; Horst et al., 2024; Huiskes et al., 1992a; Wolff, 1986). Additionally, prior studies comparing CS, DHS, and the FNS have primarily assessed early fixation stability, cadaveric loading scenarios, or short-term clinical outcomes, with little attention to the mechanical environment after fracture healing (Ahmed et al., 2023; Florschutz et al., 2015; Kalsbeek et al., 2023; Mahapatra and Pal, 2024; Stoffel et al., 2017; Stoffel et al., 2023; Widhalm et al., 2019; Wu et al., 2022; Xu et al., 2023; Zhao and Liu, 2014). The present study aims to evaluate how each fixation construct alters strain distribution in the post-healed femur across normal and osteoporotic bone. Characterizing these late-stage strain adaptations is essential, as excessive tensile or compressive strains near implant interfaces may accelerate bone resorption and contribute to secondary periprosthetic fractures, while regions of stress shielding may drive localized bone loss through disuse atrophy (Wolff, 1986).
Unlike prior finite element (FE) analyses that focus on acute fracture stability or early healing, this study evaluates how fixation constructs influence post-healed bone strain environments after fracture consolidation. This assumed healed-state modeling assesses late-stage biomechanical adaptation that may affect long-term bone health and implant longevity. Quantifying strain distributions in the healed femur provides a mechanistic link between fixation design and long-term bone adaptation. Elevated tensile or compressive strains can stimulate bone resorption or remodeling, altering load transfer and increasing the likelihood of fixation failure and the eventual conversion to total hip arthroplasty (De Meo et al., 2022; Stockton et al., 2019). Conversely, physiologic strain magnitudes support balanced remodeling and implant longevity, making strain an essential biomarker of post-healed bone health (Wolff, 1986). Even after fracture consolidation, fixation devices continue to bear repeated cyclic loading during daily activities. Therefore, evaluating implant stresses in the healed state is essential for identifying constructs at risk of fatigue failure, late screw or bolt breakage, and long-term mechanical overload that may contribute to hardware loosening or secondary fracture (Spivak et al., 1991). To address the lack of evidence across fixation strategies in this context, we developed subject-specific FE models to quantify implant stresses and regional bone strains under physiological gait loading in normal and osteoporotic bone. We specifically asked: How does implant design alter post-healed strain patterns, and how does bone quality influence these effects? By characterizing healed-state strain behavior, this study aims to clarify how different fixation constructs (CS, DHS, and FNS) influence load transfer after consolidation. These insights help delineate construct-specific mechanical behavior in normal and osteoporotic bone and may assist clinicians in understanding long-term implant performance.
2. Materials and methods
2.1. In silico model creation
Institutional review board approval was not relevant for this study as the models were created from a public, de-identified database. The New Mexico Decedent Imaging Database (NMDID) was queried to identify eligible computed tomography (CT) specimens to use for this study (Edgar et al., 2020). The CT imaging protocol employed for full-body reconstruction involved thin-slice acquisitions at 1 mm intervals with 0.5 mm slice thickness. This approach preserved accurate bone geometry, clearly delineating edges and surfaces for detailed anatomical modeling. We identified two specimens which included a 22-year-old female, with normal, non-osteoporotic bone quality and an atraumatic cause of death. Additionally, we identified a 90-year-old female with a known diagnosis of osteoporosis and an atraumatic cause of death.
Automatic segmentation was conducted using Total Segmentator v2.0, based on the nnU-Net deep learning algorithm, using an NVIDIA RTX 3090 graphics processing unit (Isensee et al., 2021; Wasserthal et al., 2023). The reconstruction was visualized using the 3D Slicer platform to confirm proper segmentation. The right femurs and right hemipelvises were then exported as a surface mesh, 2-dimensional stereolithography (STL) file (Kikinis et al., 2014). MeshLab was used to import the STL file and conduct surface refinement and remeshing (Cignoni M.C. et al., 2008). The Taubin smoothing algorithm was conducted on the bone (lambda = 0.5, mu = −0.53, 10 steps) which minimized rough surfaces on the bone following segmentation (Taubin, 1995). Laplacian smoothing (30 steps) was conducted on only the articular surfaces of the femoral heads and hemipelvises to ensure proper hip joint contact when constructing the final models. Isotropic, explicit, adaptive remeshing was conducted using a mesh size of 1. This mesh size was chosen following a mesh sensitivity analysis, varying the mesh size by 0.25, and identifying the optimal mesh size that minimized stress concentrations and provided a stable value, while keeping in mind computational efficiency (see Supplementary File). The 2D STL models were imported into Abaqus Finite Element Analysis (FEA) (Dassault Systems, Johnston, Rhode Island, USA) where volumetric mesh conversion was applied to the models using the “tri-to-tet” feature to convert the triangular shell surface mesh into solid, 4-node tetrahedral mesh (C3D4). An input file was generated and imported into Bonemat, which mapped elastic material properties on the bone mesh using the aforementioned CT images (Taddei et al., 2004; Taddei et al., 2007; Zannoni et al., 1998). The following CT densitometric equations were derived based on previous phantomless calibration studies (Morgan et al., 2003; Schileo et al., 2007; Schileo et al., 2008; Winsor et al., 2021):
Equation 1: Phantomless CT Densitometric Eq. (Winsor et al., 2021).
Equation 2: Wet conversion equation.
Equation 3: Density-elasticity relationship (Morgan et al., 2003; Schileo et al., 2007).
Hounsfield Unit (HU) values from each CT voxel were first converted to CT density (ρCT) using Eq. 1. This density was then converted to wet apparent density (ρWet) and subsequently to ash-equivalent density (ρAsh) using Eq. 2, following the phantomless calibration methods reported in prior studies. Consistent with Keyak et al. (Keyak et al., 1990; Keyak et al., 1994), the resulting apparent (ash-based) density, ρApp, was the density term used in the density-elasticity relationship (Eq. 3) to compute the elastic modulus (E) for each element. This stepwise conversion ensured that the material properties assigned to the FE models reflected the appropriate physical density corresponding to each CT voxel (see supplementary file for material property distribution for each proximal femur FE model).
Newly modified Abaqus input files were then exported from Bonemat and re-imported into Abaqus CAE, where the femurs and hemipelvises were assembled. Since CT imaging is unable to properly distinguish hip joint cartilage from the femoral heads and to simulate some level of fidelity in hip joint mechanics, we modeled idealized cartilage which was created as a 1.18 mm and 1.26 mm solid mesh (C3D4H) offset for the femoral head and acetabular cartilage, respectively, based on a previous magnetic resonance imaging study conducted by Mechlenburg et al. (Anderson et al., 2008, 2010; Mechlenburg et al., 2007). The cartilage was modeled as an incompressible, Neo-Hookean hyperelastic material (Anderson et al., 2008). Translations and rotations of the femurs and hemipelvises were performed to align the model in a static, standing posture. For each model, the right femoral head was positioned relative to the pelvis per anatomical parameters described by Wu et al. (Wu et al., 2002). The centroids of the best-fit spheres for the acetabulum and femoral heads were used to confirm that the joint center of rotation was properly located within the acetabulum (Rachakonda, 2023). The hip joint was adjusted until the joint space was within reported parameters determined by Goker et al. (Goker et al., 2003). Muscles spanning the hip joint were included according to their physiologic origins and insertions derived from the CT segmentations and were modeled as 1D passive, connector elements with stiffness values based on Phillips et al. (Phillips et al., 2007), providing joint stabilization rather than active force generation. Table 1 includes material properties of the models created (Figs. 1–3).
Table 1:
Material properties of the finite element models.
| Component | Material Properties | Constitutive Relation | Element Type |
|---|---|---|---|
| TAN (Ti-6Al-7Nb) FNS |
E = 105000 v = 0.35 |
Elastic | C3D4 |
| Ti-Alloy (Ti-6Al-4V) CS, DHS |
E = 113800 v = 0.342 |
Elastic | C3D4 |
| Bone | * | Elastic | C3D4 |
| Hip Cartilage (Anderson et al., 2008) | C10 = 6.8 D1= 0.001 |
Neo-Hookean, Hyperelastic | 4 Node Tetrahedral, Hybrid Element (C3D4H) |
| Gluteus Maximus (Phillips et al., 2007) | k = 344 N/mm | Connector Element | |
| Gluteus Medius (Phillips et al., 2007) | k = 779 N/mm | Connector Element | |
| Gluteus Minimus (Phillips et al., 2007) | k = 660 N/mm | Connector Element | |
| Psoas Major (Phillips et al., 2007) | k = 100 N/mm | Connector Element | |
| Adductor Magnus (Phillips et al., 2007) | k = 257 N/mm | Connector Element | |
| Adductor Longus (Phillips et al., 2007) | k = 134 N/mm | Connector Element | |
| Adductor Brevis (Phillips et al., 2007) | k = 499 N/mm | Connector Element |
Note that bone material properties were a range of values due to the heterogeneous material properties applied from the CT Hounsfield units (Schileo, 2007, 2008).
Fig. 1.

Methodology for constructing implanted subject-specific finite element models, including automatic anatomic segmentation, voxel material mapping, virtual surgery, and OpenSim results as boundary conditions.
Fig. 3.

Visualization of load applications and boundary conditions used in the finite element simulations. A) Proximal femur model showing all applied loads and constraints. Light blue vectors represent gluteus medius and minimus muscles (8 fibers in total added for these muscles), the green vector represents the gluteus maximus (2 fibers), and magenta vector represents the abductor muscles (abductor brevis, longus, and magnus, 9 fibers in total). The dark blue vector represents displacement boundary condition generating the joint reaction force, pointing towards the centroid of the femur, based on loading conditions conducted by Anderson et al. and Harris et al. (Anderson, 2010; Harris, 2012). The hemipelvis was fixed at the iliac crest. B–D) Integration of physiological gait mechanics, represented by the OpenSim gait_2392 model with the corresponding FE model. Each panel displays the pelvis-femur configuration aligned to subject-specific gait kinematics, paired with the corresponding musculoskeletal representation of loading response (B), mid-stance (C), and terminal stance (D). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
2.2. Implant model creation
Physical implants of the DePuy Synthes (Johnson & Johnson MedTech, DePuy Synthes, West Chester, Pennsylvania, USA) CS, DHS, and FNS were reconstructed using Solidworks (Dassault Systems, Waltham, Massachusetts, USA) using dimensional specifications and publicly available technical documentation provided in the manufacturer’s implant white papers. The implant models were meshed using a mesh size of 0.5 following a mesh convergence analysis. To ensure proper placement within the bone, the femur and implant models were imported into Blender open-source software where a Boolean operation was performed between the implants and bone, to subtract bony elements for precise implant-bone interface modeling. Following standard clinical practice, the CS were placed in an inverted triangle configuration. All implants were placed to ensure cortical abutment. Volumetric meshing was performed similarly to the steps mentioned above and geometries were imported into Abaqus, where implants were assigned to their associated linear, elastic material properties (Table 1). We assumed a healed, post-operative state where the implants were perfectly bonded and osseointegrated at the bone-implant surface. Based on this, we modeled the interaction between the implants and bone as a “tie” constraint, as global mechanics were of interest, rather than direct implant-bone interface mechanics (MacLeod et al., 2012) (Fig. 2). This modeling approach allows us to characterize the mechanical environment of the assumed healed femur to describe how different fixation constructs influence long-term postoperative load transfer in a fully consolidated, osseointegrated construct.
Fig. 2.

3-dimensional reconstructions of the cannulated screws (CS), dynamic hip screw (DHS), and femoral neck system (FNS) and subsequent implanted finite element models represented in the normal bone quality model. An inverse triangle configuration was used for the CS model, whereas the DHS and FNS cases included the commonly used 2-hole plate configurations.
2.3. Kinematics and kinetics as boundary conditions
The OpenSim gait_2392 model was used to calculate kinematics and kinetics of a walking gait trial, identifying key points of interest in loading response, mid-stance, and terminal stance phases of gait (Delp et al., 2007; Rajagopal et al., 2016) (Fig. 3). These phases were chosen as they represent transitions in load transfer during stance phase of gait, where significant changes in ground reaction forces and hip joint loading occur (Perry, 2010). To ensure proper neutral stance in the FE models to correspond with the geometric planes of the gait_2392 model, the FE models were placed in a pelvic tilt of 120, which corresponds to the neutral stance of this OpenSim model (Rajagopal et al., 2016). We performed a scaling operation on the gait_2392 model for both the young adult and geriatric specimens, matching the height and mass of the virtual models to the OpenSim model. A scaling factor was also used on each muscle fiber’s maximum isometric force.
Equation 4: Scaling factor equation.
Inverse kinematics and joint reaction force (JRF) calculations were performed to obtain pelvis and hip joint kinematics and kinetics at the time points of interest (Supplementary File).
The hemipelvises and femurs of the FE models were adjusted by rotating the body segments around the hip joint center of rotation to match the kinematic joint angles of these key time points. To simulate hip joint loading during these phases of gait, an axial translational boundary condition was placed at the distal femur which pointed towards the hip joint centroid (Anderson et al., 2008; Harris et al., 2012; Kumaran et al., 2024; Kumaran et al., 2025). We systematically adjusted the magnitude of the boundary condition at the distal femur until the FE model produced JRFs that matched the OpenSim-calculated values within 5 % accuracy. This ensured that our models replicated the net effect of the physiological forces (including muscle forces, body weight, gravity, and ground reaction forces) acting on the hip joint during the time points of interest. Once the appropriate load was determined which produced similar JRFs as OpenSim, we applied these same loading conditions to the implanted models (Fig. 3).
2.4. Data analysis
Volumetric strain values were quantified by selecting a ~ 3 mm circular region of elements surrounding the location of peak tensile and compressive strain within each anatomical zone. The average strain across this element cluster was used to represent the regional volumetric strain, following the methodology described by Altai et al. (Altai et al., 2021) for assessing internal strain distributions within the proximal femur. This approach allowed strain characterization beyond the cortical surface, incorporating contributions from the cancellous bone volume. In addition to regional bone strains, maximum von Mises stresses were recorded within each implant, excluding localized, abnormally large stress singularities, to evaluate construct-specific mechanics under gait.
3. Results
3.1. Validation of intact models
Proximal-femur regions (transcervical, basicervical, pertrochanteric, and subtrochanteric) were validated against published in vivo and in silico strain data (Aamodt et al., 1997; Altai et al., 2021; Kersh et al., 2018). Strain magnitudes in all regions fell within reported ranges, with minor deviations (see Supplementary File).
3.2. Intact vs. implanted femurs
Across all simulations of gait (loading response, mid-stance, and terminal stance), introducing any implant altered the global tensile and compressive strain environment relative to the intact femur, with the magnitude of deviation dependent on bone quality (Figs. 5–6). In the normal bone model, the CS, DHS, and FNS constructs each produced modest shifts in strain compared with the intact femur: mean tensile and compressive strains and the overall spread of the distributions remained similar, with slightly wider tensile distributions for the DHS and FNS, with CS retaining the most intact-like strain profile.
Fig. 5.

Overall strain distribution based on implant choice in the normal and osteoporotic models in the four portions of the proximal femur during the entire stance phase of gait (LR, MS, TS). The positive y-axis indicates tensile strains, whereas the negative y-axis represents compressive strains. Each violin plot illustrates the spread and density of strain magnitudes, with thicker sections indicating regions of higher strain frequency. In the normal bone model, both tensile and compressive strains remained relatively balanced across implant types, with only minor variations between fixation constructs. In contrast, the osteoporotic model showed elevated strain magnitudes, both tensile and compressive, across all implant configurations. This amplification indicates that reduced bone stiffness drives greater deformation under equivalent loading, reflecting diminished load-bearing capacity. Among fixation methods in osteoporotic bone, the CS construct retained tensile strains most similar to intact, while the FNS configuration exhibited the widest compressive and tensile strain variability.
Fig. 6.

Tensile (maximum principal) and compressive (minimum principal) strain distributions in the proximal femur for the non-instrumented, intact models.
In the osteoporotic model, the deviation from intact behavior was amplified for all implants. Relative to the osteoporotic intact femur, each construct increased both tensile and compressive strain magnitudes, with the FNS showing large widening of the distributions and highest peak values across gait (mean tensile strain 0.36 % ± 0.22 %, mean compressive strain −0.41 % ± 0.28 %). The DHS demonstrated intermediate increases, whereas CS produced the smallest changes and therefore most closely preserved the intact-like global strain response (Figs. 5–6).
3.3. Variation among implant types in bone strain and implant stress
Across all gait phases, CS consistently produced the most intact-like mechanical behavior in both normal and osteoporotic bone (Figs. 5–6). In the normal femur, tensile and compressive strain distributions remained narrow and closely aligned with those of the intact model, with only modest increases in strain magnitude. In the osteoporotic femur, CS preserved the closest approximation to intact behavior, showing the smallest upward shift in global tensile strain and a moderate increase in compressive strain, substantially lower than that observed with the DHS or FNS. Implant stresses followed the same pattern as CS demonstrated the lowest von Mises stresses across all gait phases in both bone qualities, reflecting a more diffuse transfer of load into the proximal femur (Fig. 4).
Fig. 4.

Maximum von Mises Stress (MPa) on each implant during the stance phases of gait. CS = cannulated screws, DHS = dynamic hip screw, and FNS = femoral neck system. LR = loading response, MS = mid-stance, and TS = terminal stance.
The DHS generated intermediate mechanical responses between CS and the FNS across all simulations. In the normal bone model, DHS-induced tensile strains were only slightly elevated above intact levels, but compressive strains exhibited a wider distribution, particularly in mid-stance and terminal stance, reflecting localized increases in load transfer near the lateral cortex. In the osteoporotic model, these differences became more pronounced as global tensile and compressive strain magnitudes increased beyond those produced by CS, though they remained substantially below the strain elevations produced by the FNS. DHS implant stresses also rose significantly in osteoporotic bone, exceeding CS stresses but consistently lower than those of the FNS (Figs. 5–6). Stress concentrations were observed near the lag screw–barrel interface across gait phases (Fig. 4).
Among the three constructs, the FNS produced the largest alterations to proximal femoral mechanics and the highest implant stresses. In normal bone, the FNS slightly widened both tensile and compressive strain distributions relative to the intact femur, though peak values remained within a physiologic range. In osteoporotic bone, however, the FNS generated large increase in strain magnitude as global tensile strain averaged 0.36 % ± 0.22 %, and compressive strain averaged −0.41 % ± 0.28 %, representing the largest deviation from intact mechanics of all constructs (Fig. 5). These increases were evident across all gait phases, with the basicervical and pertrochanteric regions showing the greatest strain amplification. Implant stresses mirrored this pattern as the FNS consistently exhibited the highest von Mises stresses in both bone qualities, with pronounced peaks located at the interface between the anti-rotation screw and the central bolt (Fig. 4).
3.4. Gait-phase influence on bone strain
3.4.1. Loading response
In both bone qualities, loading response generated the lowest overall strain magnitudes across stance. In normal bone, all implants produced strain distributions only slightly elevated above the intact femur, with CS and the FNS maintaining the closest match. In osteoporotic bone, implant differences became more distinct: the FNS generated the largest early-stance tensile and compressive strains, while CS remained the least deviated from intact behavior. Across regions, early-stance strain patterns remained spatially similar among constructs, with the basicervical region showing variance in strain in the osteoporotic model (Figs. 6–10, 11).
Fig. 10.

Regions of interest in the proximal femur. Strains were measured in the transcervical, basicervical, pertrochanteric, and subtrochanteric regions (Lustosa, 2024).
Fig. 11.

Tensile (positive y-axis) and compressive (negative y-axis) strains during loading response (LR) of gait for each implant.
3.4.2. Mid-stance
Mid-stance produced a clear increase in both tensile and compressive strains across all implants and both bone types. In normal bone, mid-stance differences among CS, DHS, and FNS remained modest, with the DHS showing slightly greater increases in compressive strain. In osteoporotic bone, however, the FNS exhibited the largest strain increase across the transcervical, basicervical, and pertrochanteric regions, while CS again maintained the lowest deviation from intact mechanics. DHS strains were intermediate, reflecting moderate sensitivity to reduced bone stiffness (Figs. 6–10, 12).
Fig. 12.

Tensile (positive y-axis) and compressive (negative y-axis) strains during mid-stance (MS) of gait for each implant.
3.4.3. Terminal stance
Terminal stance produced the highest strain magnitudes across all constructs and bone qualities. In normal bone, strain increases remained contained, with all three implants showing moderate increases relative to intact. However, in osteoporotic bone, terminal stance generated the largest absolute strain values of the stance phase, particularly for the FNS, which demonstrated the widest distributions and highest peak strains. CS continued to show the most intact-like performance, while DHS showed an intermediate response. Across all implants, the subtrochanteric region showed the greatest late-stance variance, particularly in osteoporotic bone (Figs. 6–10, 13).
Fig. 13.

Tensile (positive y-axis) and compressive (negative y-axis) strains during terminal stance (TS) of gait for each implant.
4. Discussion
The goal of this study was to leverage the FE method and create patient-specific virtual models based on CT voxel HU to predict how bone quality differences and internal fixation implant types affect long term, post-operative strains and stresses endured during the gait cycle. Importantly, all simulations represent consolidated, assumed healed fractures rather than the early postoperative period. Overall, these findings indicate that while implant choice causes only modest mechanical disruption in normal bone, osteoporotic bone magnifies the differences among implants, with the FNS producing the greatest strain elevation and CS retaining the most intact-like mechanical response. The FNS recorded the largest maximum von Mises stresses in both normal and osteoporotic bone, concentrated at the interface between the anti-rotation screw and bolt. In normal bone, DHS and CS constructs produced comparable implant stresses, whereas in osteoporotic bone the DHS experienced slightly larger stresses than CS. In the osteoporotic model, the FNS exhibited higher tensile and compressive bone strains than the other implants, while in normal bone FNS strains remained close to the intact case. Elevated tensile and compressive strains with the FNS in osteoporotic bone may heighten the risk of late-stage implant fatigue and loosening, particularly in active patients whose greater walking volumes impose higher cumulative loads. By quantifying these healed-state strain patterns, this study provides mechanical insight into how different fixation constructs continue to influence load transfer after fracture consolidation. Although these findings do not speak to early healing, nonunion risk, or rehabilitation protocols, they clarify how implant choice and bone quality shape the long-term mechanical environment of the proximal femur. These results highlight the effect of osteoporotic bone quality on construct behavior and emphasize the need for further clinical and biomechanical studies focused on long-term implant performance in this high-risk population.
Our findings closely parallel the biomechanical patterns reported by Roytman et al. (Roytman et al., 2022), who evaluated CS, DHS, and FNS in non-fractured osteoporotic Sawbones femora under single-leg stance and lateral-impact loading. Although termed prophylactic fixation, their approach of implanting devices into intact femora without a fracture line is methodologically equivalent to our post-healed configuration, enabling a direct comparison of construct mechanics. Consistent with their observation that the FNS exhibited the lowest stiffness and generated unfavorable transverse subtrochanteric fracture patterns under load, our healed-state models demonstrated that the FNS produced the highest implant stresses and the widest, most amplified strain distributions, particularly in osteoporotic bone. This elevated mechanical demand is likely attributable to the FNS’s fixed-angle design, which transmits load through a single consolidated bolt-screw pathway and concentrates bending and shear stresses at the bolt-anti-rotation screw interface, especially when surrounding osteoporotic bone provides limited structural support (Roytman et al., 2022). Additionally, they found that CS and the DHS provided more favorable load-transfer behavior and resisted external forces more effectively, which aligns with our finding that CS most closely preserved intact-like strain patterns and exhibited the lowest implant stresses across gait phases, with the DHS showing intermediate mechanical performance. Together, their Sawbones experimental results and our CT-based FE simulations provide convergent evidence that bone quality and construct design may influence load sharing in the proximal femur once the bone is already healed, and that fixed-angle devices may impose greater mechanical demand than parallel screw constructs in compromised bone (Mahapatra and Pal, 2025; Roytman et al., 2022).
Bone remodels in response to its mechanical environment, particularly the strain magnitude (Frost, 2003; Huiskes et al., 1992b). Assuming successful fracture healing, an implant that preserves strain distributions akin to the pre-injury (intact) state is likely to minimize stress shielding and reduce bone resorption, thereby mitigating adverse remodeling. In our healed-state models, CS most closely reproduced intact-like strain distributions across both normal and osteoporotic bone, while the DHS demonstrated intermediate behavior with modest deviations from the intact strain environment. Although the FNS preserved near-intact strains in normal bone, it produced the largest tensile and compressive strain amplification in osteoporotic bone, indicating a greater mechanical demand on both the implant and surrounding bone under reduced bone stiffness. These larger strains may predispose patients to periprosthetic or peri-implant fractures, a complication reported most frequently with the FNS in limited clinical series, though similar events have also been described, albeit less commonly, for the DHS and CS in cases of nondisplaced or incomplete femoral neck fractures (Fisher 3rd et al., 2023; Gregory et al., 2024; Müller et al., 2016; Takegami et al., 2022; Yang et al., 2020; Yoon et al., 2024). We suggest that future research incorporate a hybrid approach to assess these mechanical outcomes by employing in-silico trials and prospective clinical studies to assess FNS’s long-term performance in compromised bone (Bischoff et al., 2023; Favre and Bischoff, 2023; Maquer, 2024).
Previous studies have shown the clinical utility and cost-effectiveness in predicting fracture and strain distributions using FE models derived from CT-based voxel modeling methods (Li et al., 2023; Zannoni et al., 1998). Leveraging this approach, we assessed predictability and vulnerability for a peri-implant, or peri-prosthetic fracture, stability of the proximal femur, and potential for bone remodeling issues based on compressive and tensile strains. Beyond acute fixation mechanics, CT-voxel FE models can also serve as a platform to explore long-term bone adaptation. Elevated and asymmetric strain fields indicate regions of increased mechanical demand in the healed femur, but our study does not establish specific remodeling outcomes (Kim et al., 2023; Okike and Hasegawa, 2021; Roels et al., 2014). The FNS was designed to mitigate neck shortening, yet if strain-driven remodeling offsets that advantage, its theoretical benefit could be lost. In contrast, CS fixation, while mechanically less rigid, may better preserve native strain distributions, thereby providing a more physiologic mechanical environment in osteoporotic patients. Future work should integrate remodeling algorithms and longitudinal clinical imaging with FE simulations to clarify how post-fixation mechanical environments shape femoral morphology and to refine implant selection for higher-risk populations.
This study possesses a degree of limitations. Firstly, our methods assume no major bone remodeling as we are simulating an assumed healed, non-displaced fracture in the respective region of interest. This may pose limitations in simulating certain implant models, especially the DHS and FNS in which bone healing and remodeling provides a dynamic environment for the bolt to collapse within the plate in both constructs. Our assessments of the gait cycle were constrained between loading response and terminal stance. Furthermore, the implant-bone interfaces were modeled with “tie” interactions, assuming a perfectly bonded, and osseointegrated environment. As a result, mechanical complications such as screw loosening or implant subsidence were not simulated. However, we were interested in global mechanics, and previous studies have shown minimal differences in overall bone mechanics when simulating with “tie” interactions compared to traditional surface-surface interactions (MacLeod et al., 2012). Furthermore, this provided a computationally inexpensive method to conduct numerous simulations, as each simulation took <1 h to complete on high performance computing clusters. Other simplifications and assumptions include the use of a standard, scaled version of a gait model in OpenSim to simulate walking. If truly patient-specific modeling is to be performed, multiple patient-specific musculoskeletal models could be simulated; however, this may prove to waste computational resources and time, with little clinically meaningful differences. CT images were taken from only two patients for both osteoporotic and normal bone. Future studies will include population-based FE analysis with numerous models and a variety of bone morphologies to assess a variety of bone qualities and proximal femur geometries. Furthermore, direct mechanical testing with physical implants was not performed as part of this study. Instead of a parallel experimental setup, our computational findings were compared against published experimental data from studies that used a methodologically equivalent ‘healed state’ construct, as detailed in the Discussion. In addition to this indirect validation, we conducted a robust, multi-study validation of each component of the intact proximal femurs and assumed that if the intact models were validated, then the implanted FE models would also predict accurate mechanics. Further studies assessing gait can benefit from utilizing motion capture and inverse dynamics to create patient-specific gait models that can be applicable to corresponding CT images/models. Incorporating advanced models that take the anisotropy and non-linearity of soft tissues into account could also produce more clinically relevant results.
5. Conclusion
Using CT-voxel FE models loaded with gait-specific loads, we showed that implant performance may be modulated by bone quality. In our case-specific models, CS generated the lowest peak stresses and most physiological strain distributions in osteoporotic femurs, while the FNS best reproduced intact-bone mechanics in normal bone. While compelling, these findings raise a critical question: are the implants’ design or the patient’s underlying bone quality the primary driver of these altered strain patterns? Future work will focus on decoupling these two variables. Large-scale, population-based computational studies are essential to determine if the observed mechanical differences are consistently attributable to implant choice, or if, as suggested by our osteoporotic model, bone quality is the dominant factor in the post-healed mechanical environment. These insights emphasize the need to incorporate patient-specific bone quality, fracture morphology, surgeon expertise, and resource considerations into implant selection algorithms and post-operative load-bearing considerations. Future work integrating bone-remodeling simulations, both in healed and fractured states, with prospective clinical data will be essential to refine these recommendations and build a knowledge base that could lead to more cohesive implant selection for osteoporotic bone.
Supplementary Material
Fig. 7.

Tensile (maximum principal) and compressive (minimum principal) strain distributions in the proximal femur due to the CS and associated maximum von Mises Stress on the implants for each phase of gait.
Fig. 8.

Tensile (maximum principal) and compressive (minimum principal) strain distributions in the proximal femur due to the DHS and associated maximum von Mises Stress on the implant system for each phase of gait.
Fig. 9.

Tensile (maximum principal) and compressive (minimum principal) strain distributions in the proximal femur due to the FNS and associated maximum von Mises Stress on the implant system for each phase of gait.
Highlights.
Cannulated screws best replicate intact strain in osteoporotic femora during gait.
Femoral neck system concentrates peak stress at the bolt/anti-rotation screw junction.
Elevated femoral neck system strains in osteoporotic bone may drive remodeling and revision risk.
Rehabilitation may consider patient-specific bone quality and implant type.
Funding Acknowledgements
This work was supported by the National Institute on Aging under the Grants for Early Medical/Surgical Specialists’ Transition to Aging Research (GEMSSTAR) grant (R03AG060177) and the Paul B. Beeson Emerging Leaders Career Development Award in Aging (K76AG068435). Support was also provided in part by the NSF Industry/University Cooperative Research Center at the University of California San Francisco, The Ohio State University, and the University of Toledo (NSF 1916636). Additionally, this work was funded by the Orthopaedic Research and Education Foundation (OREF) Mentored Clinician Scientist Grant, made possible by the Ira A. Roschelle Family Foundation.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.org/10.1016/j.clinbiomech.2025.106724.
Footnotes
Declaration of competing interest
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 paper.
CRediT authorship contribution statement
Yogesh Kumaran: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Sophia Soehnlen: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis. Christopher Haritos: Writing – review & editing, Writing – original draft, Investigation. Sanam Jhaveri: Writing – review & editing, Writing – original draft, Visualization, Methodology, Formal analysis. Sudharshan Tripathi: Validation, Methodology. Humza Shaikh: Writing – review & editing, Writing – original draft, Supervision, Investigation. Stephanie Di Stasi: Writing – review & editing, Writing – original draft, Supervision, Investigation, Conceptualization. Carmen E. Quatman: Writing – review & editing, Writing – original draft, Supervision, Resources, Methodology, Investigation, Conceptualization.
This article is part of a Special issue entitled: ‘Finite Element Methods in Orthopaedic Biomechanics’ published in Clinical Biomechanics.
References
- Aamodt A, et al. , 1997. In vivo measurements show tensile axial strain in the proximal lateral aspect of the human femur. J. Orthop. Res. 15 (6), 927–931. [DOI] [PubMed] [Google Scholar]
- Ahmed M, et al. , 2023. A comparison of internal fixation and hemiarthroplasty in the management of un- or minimally displaced hip fractures in patients over 60 years old. Injury 54 (4), 1180–1185. [DOI] [PubMed] [Google Scholar]
- Altai Z, et al. , 2021. Femoral neck strain prediction during level walking using a combined musculoskeletal and finite element model approach. PLoS One 16 (2), e0245121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Anderson AE, et al. , 2008. Validation of finite element predictions of cartilage contact pressure in the human hip joint. J. Biomech. Eng. 130 (5), 051008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Anderson AE, et al. , 2010. Effects of idealized joint geometry on finite element predictions of cartilage contact stresses in the hip. J. Biomech. 43 (7), 1351–1357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bischoff JE, Dharia MA, Favre P, 2023. A risk and credibility framework for in silico clinical trials of medical devices. Comput. Methods Prog. Biomed. 242, 107813. [DOI] [PubMed] [Google Scholar]
- Brauer CA, et al. , 2009. Incidence and mortality of hip fractures in the United States. JAMA 302 (14), 1573–1579. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cignoni MC, P., Corsini M, Dellepiane M, Ganovelli F, Ranzuglia G, 2008. MeshLab: an Open-Source Mesh Processing Tool. In: Sixth Eurographics Italian Chapter Conference, pp. 129–136. [Google Scholar]
- De Meo D, et al. , 2022. Outcome and technical consideration of conversion total hip arthroplasty after failed fixation of intracapsular and extracapsular hip fractures: are they really that different? Injury 53, S23–S28. [DOI] [PubMed] [Google Scholar]
- Delp SL, et al. , 2007. OpenSim: open-source software to create and analyze dynamic simulations of movement. IEEE Trans. Biomed. Eng. 54 (11), 1940–1950. [DOI] [PubMed] [Google Scholar]
- Deng C, Gillette JC, Derrick TR, 2021. Measuring femoral neck loads in healthy young and older adults during stair ascent and descent. PLoS One 16 (1), e0245658. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Edgar HJH, Daneshvari Berry S, Moes E, Adolphi NL, Bridges P, Nolte KB, 2020. New Mexico Decedent Image Database. Office of the Medical Investigator, University of New Mexico. [Google Scholar]
- Eijer H, Myers SR, Ganz R, 2001. Anterior femoroacetabular impingement after femoral neck fractures. J. Orthop. Trauma 15 (7), 475–481. [DOI] [PubMed] [Google Scholar]
- Favre P, Bischoff J, 2023. Identifying the patient harms to include in an in silico clinical trial. Comput. Methods Prog. Biomed. 241, 107735. [DOI] [PubMed] [Google Scholar]
- Fisher JC 3rd, et al. , 2023. Subtrochanteric fracture after femoral neck system of femoral neck fractures: a report of four cases. BMC Musculoskelet. Disord. 24 (1), 749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Florschutz AV, et al. , 2015. Plating of femoral neck fractures: when and how? J. Orthop. Trauma 29 (Suppl. 4), S1–S3. [DOI] [PubMed] [Google Scholar]
- Frost HM, 2003. Bone’s mechanostat: A 2003 update. Anat Rec A Discov Mol Cell Evol Biol 275A (2), 1081–1101. [DOI] [PubMed] [Google Scholar]
- Ge Z, et al. , 2023. Comparison of femoral neck system vs. dynamic hip system blade for the treatment of femoral neck fracture in young patients: A retrospective study. Front. Surg. 10, 1092786. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Goker B, et al. , 2003. The radiographic joint space width in clinically normal hips: effects of age, gender and physical parameters. Osteoarthr. Cartil. 11 (5), 328–334. [DOI] [PubMed] [Google Scholar]
- Gregory D, et al. , 2024. Peri-implant fracture following prior fixation of femoral neck fracture using the femoral neck system: A case report. J Orthop Case Rep 14 (2), 82–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harris MD, et al. , 2012. Finite element prediction of cartilage contact stresses in normal human hips. J. Orthop. Res. 30 (7), 1133–1139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Horst N, et al. , 2024. Stem subsidence in total hip arthroplasty: retrospective investigation of a short stem using a simple measurement approach. Arch. Orthop. Trauma Surg. 145 (1), 40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Huiskes R, Weinans H, van Rietbergen B, 1992a. The relationship between stress shielding and bone resorption around total hip stems and the effects of flexible materials. Clin. Orthop. Relat. Res. 274, 124–134. [PubMed] [Google Scholar]
- Huiskes R, Weinans H, Rietbergen BV, 1992b. The relationship between stress shielding and bone resorption around Total hip stems and the effects of flexible materials. Clin. Orthop. Relat. Res. 274, 124–134. [PubMed] [Google Scholar]
- Isensee F, et al. , 2021. nnU-net: a self-configuring method for deep learning-based biomedical image segmentation. Nat. Methods 18 (2), 203–211. [DOI] [PubMed] [Google Scholar]
- Kalsbeek JH, et al. , 2023. What makes fixation of femoral neck fractures fail? A systematic review and meta-analysis of risk factors. Injury 54 (2), 652–660. [DOI] [PubMed] [Google Scholar]
- Kamara E, Zvi YS, Vail TP, 2021. Treatment of Valgus-impacted and nondisplaced femoral neck fragility fractures in the elderly. J. Am. Acad. Orthop. Surg. 29 (11), 470–477. [DOI] [PubMed] [Google Scholar]
- Kazley J, Bagchi K, 2023. Femoral neck fractures. In: StatPearls. Treasure Island (FL). [PubMed] [Google Scholar]
- Kersh ME, et al. , 2018. Mechanical loading of the femoral neck in human locomotion. J. Bone Miner. Res. 33 (11), 1999–2006. [DOI] [PubMed] [Google Scholar]
- Keyak JH, et al. , 1990. Automated three-dimensional finite element modelling of bone: a new method. J. Biomed. Eng. 12 (5), 389–397. [DOI] [PubMed] [Google Scholar]
- Keyak JH, Lee IY, Skinner HB, 1994. Correlations between orthogonal mechanical properties and density of trabecular bone: use of different densitometric measures. J. Biomed. Mater. Res. 28 (11), 1329–1336. [DOI] [PubMed] [Google Scholar]
- Kikinis R, Pieper SD, Vosburgh KG, 2014. 3D slicer: A platform for subject-specific image analysis, visualization, and clinical support, in intraoperative imaging and image-guided therapy. In: Jolesz FA (Ed.), Springer. New York, NY, New York, pp. 277–289. [Google Scholar]
- Kim H, Kim JS, Sung YB, 2023. Postoperative Valgus deformity and progression of Ostheoarthritis in non-displaced femoral neck fractures. Hip Pelvis 35 (4), 259–267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kumaran Y, et al. , 2025. Phase-specific changes in hip joint loading during gait following sacroiliac joint fusion: findings from a finite element analysis. Clin. Biomech. 122, 106429. [DOI] [PubMed] [Google Scholar]
- Kumaran YMM, Quatman C, Samora J, Soehnlen S, Hoffman B, Nishida N, Tripathi S, Elgafy H, Hanna M, Goel VK, 2024. Pediatric Sagittal Imbalance Contributes To Proximal Femoral Physis Mechanical Shear In Slipped Capital Femoral Epiphysis, in Orthopaedic Research Society Annual Meeting. Long Beach, CA. [Google Scholar]
- Li J, et al. , 2023. Cost-effectiveness Analysis of CT-based finite element modeling for osteoporosis screening in secondary fracture prevention: an early health technology assessment in the Netherlands. MDM Policy Pract 8 (2), p. 23814683231202993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li N, et al. , 2024. Evaluating three internal fixation techniques for Pauwels III femoral neck fractures via finite element analysis. Sci. Rep. 14 (1), 15519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lustosa L, 2024. Regions of the Proximal Femur (Illustration). [Google Scholar]
- MacLeod AR, Pankaj P, Simpson AHRW, 2012. Does screw–bone interface modelling matter in finite element analyses? J. Biomech. 45 (9), 1712–1716. [DOI] [PubMed] [Google Scholar]
- Mahapatra B, Pal B, 2024. Biomechanical analysis of various internal fracture fixation devices used for treating femoral neck fractures: A comparative finite element analysis. Injury 55 (10), 111717. [DOI] [PubMed] [Google Scholar]
- Mahapatra B, Pal B, 2025. From healthy to osteoporotic: Exploring how bone quality alters implant performance in Pauwels type III fracture. In: Proceedings of the Institution of Mechanical Engineers, Part H. 0(0): p. 09544119251333671. [DOI] [PubMed] [Google Scholar]
- Maquer G, et al. , 2024. Developing and validating a model of humeral stem primary stability, intended for in silico Clinical trials. Ann. Biomed. Eng. 52 (5), 1280–1296. [DOI] [PubMed] [Google Scholar]
- Mechlenburg I, et al. , 2007. Cartilage thickness in the hip joint measured by MRI and stereology–a methodological study. Osteoarthr. Cartil. 15 (4), 366–371. [DOI] [PubMed] [Google Scholar]
- Morgan EF, Bayraktar HH, Keaveny TM, 2003. Trabecular bone modulus–density relationships depend on anatomic site. J. Biomech. 36 (7), 897–904. [DOI] [PubMed] [Google Scholar]
- Müller F, et al. , 2016. Peri-implant femoral fractures: the risk is more than three times higher within PFN compared with DHS. Injury 47 (10), 2189–2194. [DOI] [PubMed] [Google Scholar]
- Okike K, Hasegawa IG, 2021. Current trends in the evaluation and Management of Nondisplaced Femoral Neck Fractures in the elderly. J. Am. Acad. Orthop. Surg. 29 (4), e154–e164. [DOI] [PubMed] [Google Scholar]
- Perry JBJ, 2010. Gait analysis: Normal and pathological function. J. Sports Sci. Med. 9 (2). [Google Scholar]
- Phillips AT, et al. , 2007. Finite element modelling of the pelvis: inclusion of muscular and ligamentous boundary conditions. Med. Eng. Phys. 29 (7), 739–748. [DOI] [PubMed] [Google Scholar]
- Rachakonda P, 2023. Sphere Fit. In: MATLAB Central File Exchange: MATLAB Central File Exchange. [Google Scholar]
- Rajagopal A, et al. , 2016. Full-body musculoskeletal model for muscle-driven simulation of human gait. IEEE Trans. Biomed. Eng. 63 (10), 2068–2079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Roels P, et al. , 2014. Mechanical factors explain development of cam-type deformity. Osteoarthr. Cartil. 22 (12), 2074–2082. [DOI] [PubMed] [Google Scholar]
- Rogmark C, et al. , 2018. Hip fractures in the non-elderly—who, why and whither? Injury 49 (8), 1445–1450. [DOI] [PubMed] [Google Scholar]
- Roytman GR, et al. , 2022. Simulating prophylactic fixation methods for osteoporotic femoral neck fracture prevention. Geriatr Orthop Surg Rehabil 13, p. 21514593221141376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schileo E, et al. , 2007. Subject-specific finite element models can accurately predict strain levels in long bones. J. Biomech. 40 (13), 2982–2989. [DOI] [PubMed] [Google Scholar]
- Schileo E, et al. , 2008. An accurate estimation of bone density improves the accuracy of subject-specific finite element models. J. Biomech. 41 (11), 2483–2491. [DOI] [PubMed] [Google Scholar]
- Spivak JM, et al. , 1991. Fatigue failure of the sliding screw in hip fracture fixation: a report of three cases. J. Orthop. Trauma 5 (3), 325–331. [DOI] [PubMed] [Google Scholar]
- Stockton DJ, et al. , 2019. High rate of reoperation and conversion to total hip arthroplasty after internal fixation of young femoral neck fractures: a population-based study of 796 patients. Acta Orthop. 90 (1), 21–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stoffel K, et al. , 2017. Biomechanical evaluation of the femoral neck system in unstable Pauwels III femoral neck fractures: A comparison with the dynamic hip screw and cannulated screws. J. Orthop. Trauma 31 (3), 131–137. [DOI] [PubMed] [Google Scholar]
- Stoffel K, et al. , 2023. Clinical performance of the femoral neck system within 1 year in 125 patients with acute femoral neck fractures, a prospective observational case series. Arch. Orthop. Trauma Surg. 143 (7), 4155–4164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Surgeons TAAOO, 2021. American Academy of Orthopaedic Surgeons Management of Hip Fractures in Older Adults Evidence-Based Clinical Practice Guideline. https://www.aaos.org/globalassets/quality-and-practice-resources/hip-fractures-in-the-elderly/hipfxcpg.pdf. [DOI] [PubMed]
- Taddei F, Pancanti A, Viceconti M, 2004. An improved method for the automatic mapping of computed tomography numbers onto finite element models. Med. Eng. Phys. 26 (1), 61–69. [DOI] [PubMed] [Google Scholar]
- Taddei F, et al. , 2007. The material mapping strategy influences the accuracy of CT-based finite element models of bones: an evaluation against experimental measurements. Med. Eng. Phys. 29 (9), 973–979. [DOI] [PubMed] [Google Scholar]
- Takegami Y, et al. , 2022. Comparison of periprosthetic femoral fracture torque and strain pattern of three types of femoral components in experimental model. Bone Joint Res 11 (5), 270–277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Taubin G, 1995. Curve and Surface Smoothing without Shrinkage. Institute of Electrical and Electronics Engineers. [Google Scholar]
- Wasserthal J, et al. , 2023. TotalSegmentator: Robust segmentation of 104 anatomic structures in CT images. Radiol. Artif. Intell. vol. 5 (5), e230024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wei P, et al. , 2020. Conservative vs surgical treatment of impacted femoral neck fracture in patients 75years and older. J. Am. Geriatr. Soc. 68 (10), 2214–2221. [DOI] [PubMed] [Google Scholar]
- Widhalm HK, et al. , 2019. A comparison of dynamic hip screw and two cannulated screws in the treatment of Undisplaced Intracapsular neck fractures-two-year follow-up of 453 patients. J. Clin. Med. 8 (10). [DOI] [PMC free article] [PubMed] [Google Scholar]
- Winsor C, et al. , 2021. Evaluation of patient tissue selection methods for deriving equivalent density calibration for femoral bone quantitative CT analyses. Bone 143, 115759. [DOI] [PubMed] [Google Scholar]
- Wolff J, 1986. The Law of Bone Remodelling. Springer-Verlag. xii, Berlin; New York, 126 p. [Google Scholar]
- Wu G, et al. , 2002. ISB recommendation on definitions of joint coordinate system of various joints for the reporting of human joint motion–part I: ankle, hip, and spine. Int. Soc. Biomech. J Biomech 35 (4), 543–548. [DOI] [PubMed] [Google Scholar]
- Wu L, et al. , 2022. Should cannulated screws be removed after a femoral neck fracture has healed, and how? A finite element analysis of the femur before and after cannulated screw removal. Int. Orthop. 46 (10), 2393–2403. [DOI] [PubMed] [Google Scholar]
- Xu X, et al. , 2023. Comparison of femoral neck system to multiple cancellous screws and dynamic hip screws in the treatment of femoral neck fractures. Injury 54 (Suppl. 2), S28–s35. [DOI] [PubMed] [Google Scholar]
- Yang YO, Xu S, Wong MK, 2020. Surgical fixation of subtrochanteric fracture in patient with cancellous screw fixation for neck of femur fracture: surgical decision-making and outcome. J Orthop Case Rep 9 (6), 44–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yoon JY, Byun S-E, Cho Y-H, 2024. Fixation of femoral neck fracture with femoral neck system: a retrospective cohort study of 43 patients. BMC Musculoskelet. Disord. 25 (1), 8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zannoni C, Mantovani R, Viceconti M, 1998. Material properties assignment to finite element models of bone structures: a new method. Med. Eng. Phys. 20 (10), 735–740. [DOI] [PubMed] [Google Scholar]
- Zhao W, Liu L, 2014. Effect of dynamic hip system blade on the treatment of femoral neck fractures in elderly patients with osteoporosis. Chin. J. Traumatol. 17 (5), 275–278. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
