ABSTRACT
The aim of the study was to design a standardized mechanical test setup and a corresponding finite element analysis to assess the stability and strength of both patient‐specific and conventional implants for posterior wall acetabular fractures. Ten synthetic hemi‐pelves with posterior wall fractures were biomechanically tested with two types of implants: a patient‐specific implant (PSI) and a seven‐hole plate conventional implant. 3D‐printed guides ensured reproducibility. The models were tested using an Instron machine. The protocol involved 10,000 cyclic load cycles with static tests at 3200 N before and after to simulate early postoperative weightbearing conditions. Construct stiffness, stiffness over cyclic loading and fracture gapping were measured and compared. A finite element analysis was created with similar conditions to investigate stresses within the synthetic bone and fixation materials. The mechanical tests showed comparable stiffness for PSI (1.75 kN/mm) and the conventional implant (1.71 kN/mm, p = 0.47). Stability over 10,000 cycles was similar, and fracture gapping remained minimal (0.0–0.8 mm) without significant differences. No failure or plastic deformation occurred under 3200 N loading. Finite element analysis confirmed that von Mises stresses remained below the yield stress. This study introduces a reproducible workflow for biomechanical testing of acetabular fractures using synthetic bone models and 3D‐printed guides. It serves as a step‐by‐step guideline and standard reference for pelvic biomechanical testing. Both patient‐specific and conventional implants, using a seven‐hole plate construct with one or two screws through the plate in the fracture fragment, provide stable fixation for large posterior wall fragments.
Keywords: 3D technology, acetabular fractures, biomechanics, FEA, patient‐specific implants
1. Introduction
Posterior wall fractures account for about 20% of all acetabular fractures [1, 2], making it one of the most common acetabular fractures. Surgery is often indicated when the fracture is unstable and surgical treatment should increase joint stability, improve functional outcomes, and reduce the risk of total hip arthroplasty (THA). Open reduction and internal fixation (ORIF) surgery is typically performed using a combination of commercially available conventional implants and screws. In recent years, patient‐specific implants (PSIs) have been introduced for acetabular fracture surgery [3, 4]. These implants are designed based on the patient's anatomy and fracture lines, allowing for a precise fit and optimal screw placement during surgery.
Posterior wall acetabular fracture surgery with conventional implants involves manual bending of the implant to match the patient's reconstructed bone surface (i.e., posterior wall and column). Getting a close match between the implant surface and the surrounding bone, as well as obtaining ideal screw placements, remains challenging, even for experienced trauma surgeons. Furthermore, the required repetitive manual bending introduces plastic deformation (i.e., surpassing the yield stress) and generates residual stresses within the implant, increasing the likelihood of implant failure and compromising fatigue resistance [5]. PSIs, designed based on the patient's anatomy and fracture lines, can provide a more precise fit, are ready‐to‐use and have predetermined preferred screw positions during surgery. While the clinical benefits are still under investigation—and preliminary results suggest promising outcomes [3, 4], its biomechanical strength, compared to the current standard treatment with conventional implants, remains unknown. It is hypothesized that this precise fit and ready‐to‐use implant shape offer greater stability of the whole construct, potentially enabling quicker loadbearing postoperatively. This could result in an accelerated rehabilitation period and improved patient care. However, to our knowledge, comparative mechanical studies comparing these two types of implants for posterior wall fractures of the acetabulum are not yet available.
Therefore, the aims of our study were (1) to design a standardized mechanical test setup to investigate the stability and weight‐bearing capabilities of two different implants (the PSI and the manually bent conventional implant) for an isolated posterior wall acetabular fracture and (2) to compare these results with a finite element analysis (FEA) of a similar construct to investigate and validate computational analysis of the setup for further implant design assessment.
2. Materials and Methods
2.1. Materials
A total of 10 fourth‐generation composite bone hemi‐pelves (Sawbones; Pacific Research Laboratories, Vashon, WA, USA) were utilized for biomechanical testing. Synthetic bones were chosen to eliminate uncontrollable variables, such as variations in shape, bone quality, and size, which arise with cadaveric bones. Two fixation methods were tested (1): a patient‐specific posterior wall implant developed previously by our group [3, 4] and (2) a conventional off‐the‐shelf implant; the seven‐hole straight plate (DePuy Synthes, Raynham, MA, USA). For both implants, 3.5 mm cortical screws were used.
A computed tomography (CT) scan of the synthetic hemipelvis model was performed, and the data were segmented using Mimics Medical software (version 24.0, Materialise, Leuven, Belgium) to generate three‐dimensional (3D) virtual models. A pre‐set threshold for the synthetic bone (≥ −200 Hounsfield Units, HU) was applied to select the hemipelvis. The 3D virtual models were then imported into 3‐Matic software (version 18.0, Materialise, Leuven, Belgium) to create saw guides, drill guides, implant designs, and customized inserts for the universal setup. The shape of the fracture was derived from a clinical case treated in our hospital (Figure 1a). A 3D saw guide was produced to ensure consistent cutting of a posterior wall fracture in each hemipelvis. The guide featured slots on both sides of the bone through which a fretsaw could be inserted to saw the synthetic bone (Figure 1b,c), which resulted in a single posterior wall fragment (Figure 1d) with a maximum length of 53 mm and a width of 32 mm.
Figure 1.

Preparation of the synthetic bone models before biomechanical testing. A posterior wall fragment (a) was virtually modeled to design a 3D sawing guide (b), which was 3D‐printed and used with a fretsaw (c) to create a single posterior wall fragment (d). A patient‐specific implant (PSI) (e) was designed along with a drilling guide (f), which was also 3D‐printed (g). After drilling, the PSI was fixated with 3.5 mm cortical screws (h). The conventional implant was manually bent and then optically scanned and virtually placed on the model (i), and a drilling guide was designed (j) and 3D‐printed (k). The conventional implant was then fixated with 3.5 mm cortical screws (l).
The screw trajectories of the PSI were determined in collaboration with an experienced trauma surgeon and planned using the 3D software 3‐Matic. The PSI workflow allows for considerable freedom in the positioning and direction of the screws. Therefore, two screws were placed in the bone fragment. A PSI was designed using 3‐Matic, Solidworks Professional (version 2020, Dassault Systèmes Solidworks Corp., Waltham, MA, USA), and the Geomagic for Solidworks package (3D systems, Rock Hill, SC, USA) (Figure 1e). The workflow for designing and manufacturing a PSI tailored to acetabular fractures has been previously introduced and detailed by our group [3, 4]. The PSIs were fabricated from medical‐grade titanium alloy grade 5 using a 5‐axis milling machine at Witec Medical B.V. (Stadskanaal, the Netherlands). For comparison, a conventional seven‐hole straight plate was manually contoured to fit the synthetic bone model, which stimulated intraoperative implant bending. The conventional implant was optically scanned with the 3Shape Trios intraoral scanner (3Shape Trios A/S 2018, Copenhagen, Denmark), and its placement was virtually mapped on the 3D model (Figure 1i) to determine the screw trajectories start points. The remainder of the screw trajectories was determined in collaboration with an experienced trauma surgeon and mapped in 3D. Drill guides were designed for both the PSI and conventional implants to ensure consistent screw trajectories across all specimens (Figure 1f,g for the PSI, and Figure 1j,k for the conventional implant). All guides (e.g., the saw and drill guides) were 3D‐printed using polyamide (PA12) powder and selective laser sintering techniques (Oceanz, Ede, The Netherlands). Both implants were fixated with 3.5 mm cortical screws (Figure 1h for the PSI and Figure 1l for the conventional implant). Due to the simplicity of the clear cut and the freedom of no surrounding soft tissue, no reduction tool was used. The aim of this study was to compare the PSI, along with its workflow which allows considerable freedom in screw placement, to the conventional implant approach. Therefore, the number and location of screws may differ between the two implants. However, this approach provides a comparison that reflects the clinically relevant treatment strategies for both implants.
2.2. Biomechanical Test Set‐Up
Mechanical tests were conducted using an Instron 4300 Series testing system (Instron, England), equipped with a 5 kN capacity load cell. A customized biomechanical setup was designed to fit the Instron machine and held the hemipelvis in a consistent orientation across all specimens. The pelvis was positioned to align the applied load with the direction of peak contact force during walking for the second gait step, as reported by Bergmann et al. [6]. The sacroiliac and pubic joints were fixated in all three planes and clamped in plastic (PA12) moulds (white parts in Figure 2a–c), which were virtually designed and 3D‐printed. These plastic moulds fitted precisely into custom‐designed, milled aluminum holders (gray parts in Figure 2a–c). The holders were secured within the overall test setup frame, and the entire physical test setup is illustrated in Figure 2d,e. Axial force was applied via a custom‐made spherical load‐applicator with a diameter of 50 mm (Figure 2f).
Figure 2.

The biomechanical test set‐up. The digitally designed parts (a–c), illustrating the plastic moulds in white and the aluminum milled frame holders in gray. The physical test set‐up in the Instron machine (d), with the PSI‐model (e), and illustration of the three points for measurements of fragment gapping (f).
2.3. Testing Protocol
The testing protocol was designed to compare the fixation strength between two groups (PSI vs. conventional implants). The large hemipelvis sawbone model represented a 91 kg male [7]. Initially, a ramp force of 3200 N (approximately 3.5 times body weight) was applied at an increase rate of 20 N/s, the static test. Following this initial load, the model was unloaded, and a test protocol consisting of 10,000 load cycles at 3200 N was applied at a frequency of 0.5 Hz. This represents the first 4 weeks postoperatively, before 80% of the load bearing capacities can be taken over by the reconstructed bone [8]. This means 1 week of recovery due to wound closure and 3 weeks of 1000 steps a day. After cyclic loading, the initial 3200 N test was repeated to assess whether any system weakness occurred during the cyclic loading phase. The final 3200 N tests were recorded using Logitech webcam (C920 HD Pro) and analyzed with Tracker software (version 6.2), an open‐source video analysis tool, to investigate fracture displacement. Fracture gapping was quantified as the change in distance between predefined reference points over time, with positive values indicating an increased distance and negative values indicating decreased distance.
2.4. Finite Element Modeling
In addition to the physical mechanical tests, a computational model was created using FEA. Through comparison of the computational results to the physical test the FEA model was validated, ensuring its accuracy and reliability. The FEA used a similar workflow and conditions as previous research from our group [9]. The 3D models of the virtual bone fragments, the implant, cylinders representing simplified screws, and the load applicator were meshed with 10‐node tetrahedral elements (C3D10) in 3‐Matic software and imported into Mimics software. The CT‐scan was used to assign material properties for the synthetic bone. HU > −200 represents the cortical layer, while the mesh within represents the cancellous synthetic bone. The number of elements and the material properties are shown in Table 1. The screws were modeled as cylinders, as this simplification saves significant time and is considered sufficiently accurate according to the literature [11, 12]. The areas in contact with the inserts of the setup were constrained from movement in all directions as boundary conditions. Contacts and constraints were modeled similarly to a previous study [9]. FEA was performed in Abaqus software (version 2022, Dassault Systèmes, Vélizy‐Villacoublay, France). The screws were tied to the bone, assuming proper grip [11]. Penalty contacts with a friction coefficient of 0.3 were applied between the fracture line surfaces [13] and between the bone and the implant [14]. The interaction between the implant and the screw heads were assumed to be tied [15]. The load applicator was restricted to move only along its longitudinal axis. The load applicator was first moved along its longitudinal axis for 0.7 mm, to assure contact with the synthetic bone. Subsequently, the force was applied at the top of the load applicator in the form of a pressure, with a total resulting force of 3200 N.
Table 1.
The number of elements per FEA‐part with their corresponding Young's modulus (E) and Poisson ratio (v).
| Part | Number of elements | Material properties |
|---|---|---|
| Base part hemipelvis |
PSI‐model: 287,082 CONV‐model: 197,423 |
E = 16.7 GPa, v = 0.3 [10] |
| Fragment part hemipelvis |
PSI‐model: 21,504 CONV‐model: 20,641 |
E = 157 MPa, v = 0.3 [10] |
| Screws (stainless steel) |
PSI‐model: 119,361 CONV‐model: 208,550 |
E = 200 GPa, v = 0.33 |
| Conventional plate (stainless steel) | 166,701 | E = 200 GPa, v = 0.33 |
| Patient‐specific implant (titanium) | 50,675 | E = 114 GPa, v = 0.3 |
| Load applicator (aluminum) | 8859 | E = 68.7 GPa, v = 0.3 |
2.5. Outcome Variables
-
1.Mechanical testing
-
1.1.Construct stiffness (kN/mm) measured between 1.0 and 3.0 kN during the static test, both before and after cyclic loading.
-
1.2.Construct stiffness (kN/mm) calculated as the mean stiffness between the 5000th and 10,000th cycle during the cyclic loading, between 1.0 and 3.0 kN, to investigate stability over time.
-
1.3.Fragment gapping [mm] at three points (Figure 2f), measured using video tracking software during the static tests.
-
1.1.
-
2.FEA results
-
2.1.Von Mises stresses (MPa) in the synthetic bone and implant for both fixation constructs, to investigate stress distributions and suggest potential implant failure.
-
2.2.Displacement (mm) distributions to compare deformation patterns of both constructs and to investigate FEA predictions and mechanical test results to validate the model.
-
2.3.Fragment gapping (mm) at three points (Figure 2f), based on the coordinates of the nodes during the simulation.
-
2.1.
2.6. Statistical Analyses
The overall data set is divided in two groups, five samples each: the models with PSI and those with the conventional implant (CONV). The continuous variables (construct stiffness of static and cyclic loading and the fracture gapping measurements) are summarized using median and interquartile range (IQR). A Mann–Whitney U‐test is conducted to evaluate differences between the two groups, with a significance level of 0.05. Statistical analysis is performed using IBM SPSS Statistics (version 28.0.1.0, IBM, Chicago, IL, USA).
3. Results
3.1. Mechanical Testing
The results for the mechanical testing were summarized in Table 2. First, Figure 3 shows the construct stiffness in the displacement–loading graph for the static loading mechanical test (Figure 3). The median construct stiffness shows no significant difference between the implants prior and after cycling loading (p = 0.47 and p = 0.22, respectively). Both implants showed higher stiffness during cyclic loading when the speed is increased and do not lose stability over 10,000 cycles (Figure 4). No significant differences were found between the two groups (p = 0.08). Some fragment gapping (range 0.0–0.8 mm), measured around three points along the fracture line, were seen in both groups with no significant difference between the groups for all three points (p = 0.46, p = 0.46, and p = 0.60).
Table 2.
Results from mechanical testing.
| Construct stiffness (kN/mm)—before cyclic loading‐between 1.0 and 3.0 kN | Construct stiffness (kN/mm)—after to cyclic loading‐between 1.0 and 3.0 kN | Mean stiffness during cyclic loading (kN/mm) between 1.0 and 3.0 kN | Fracture gapping (mm) | |||
|---|---|---|---|---|---|---|
| Point 1 | Point 2 | Point 3 | ||||
| PSI1 | 1.70 | 2.40 | 2.90 | 0.0 | 0.7 | 0.3 |
| PSI2 | 1.56 | 1.89 | 3.00 | 0.0 | −0.3 | 0.2 |
| PSI3 | 2.33 | 1.96 | 2.72 | 0.1 | −0.4 | 0.2 |
| PSI4 | 1.89 | 1.96 | 2.87 | 0.0 | −0.6 | 0.2 |
| PSI5 | 1.75 | 1.77 | 2.73 | 0.0 | −0.6 | 0.2 |
| Median (IQR) | 1.75 (1.70–1.89) | 1.96 (1.83–2.18) | 2.87 (2.73–2.90) | 0.0 (0.0–0.0) | −0.4 (−0.6 to −0.4) | 0.2 (0.19–0.25) |
| CONV1 | 1.81 | 1.89 | 2.77 | 0.1 | −0.8 | 0.3 |
| CONV2 | 1.42 | 1.83 | 3.02 | 0.0 | −0.6 | 0.1 |
| CONV3 | 1.58 | 1.59 | 2.97 | −0.2 | −0.3 | 0.0 |
| CONV4 | 1.71 | 2.14 | 2.93 | 0.0 | −0.4 | 0.2 |
| CONV5 | 1.76 | 1.66 | 3.21 | −0.2 | −0.3 | 0.0 |
| Median IQR | 1.71 (1.58–1.76) | 1.83 (1.63–2.02) | 3.02 (2.83–3.21) | 0.0 (−0.2 to 0) | −0.4 (−0.6 to −0.3) | 0.1 (0–0.2) |
| p‐value | 0.47 | 0.21 | 0.08 | 0.46 | 0.46 | 0.60 |
Note: Construct stiffness before cyclic loading with 20 m/s and during cyclic loading (1 Hz), both measured between 1.0 and 3.0 kN. Fracture gapping at three measurement points along the fracture line. All results are reported for the patient‐specific (PSI) and conventional (CONV) implant, including medians, IQRs, and statistical comparisons (p‐values).
Figure 3.

Force‐displacement graphs measured during the single loading with rate of 20 N/s. Blue lines represent the patient‐specific implant models, while red/pink lines represent the conventional implant models. Left graph illustrates the tests before cyclic loading, and the right graph show measurements after 10.000 cycles at 3200 N.
Figure 4.

Stiffness measured between 1.0 and 3.0 kN during each cycle of the 10,000 cyclic loading at 0.5 Hz. The solid line represents the median with the interquartile range (IQR) around the line. Blue lines represent the patient‐specific implant models, while red lines represent the conventional implant models.
3.2. FEA
The von Mises stresses in the bone and fixation materials were calculated using FEA under a compressive load of 3200 N (Figure 5). Both models had stress concentrations near the sacroiliac joint, which could be attributed to the application of a rigid boundary condition. These areas were outside the scope of our investigation; therefore, no additional efforts were made to refine the boundary conditions. Peak stresses of 442 MPa and 499 MPa were found at the acetabular region for the PSI and conventional models, respectively. These stress concentrations corresponded to the locations where small fractures occur during mechanical testing, likely caused by the rigid aluminum loading applicator (see File S1). Within the PSI model, a peak stress of approximately 55 MPa was observed. For the conventional implant model, the corresponding peak stress was higher, at approximately 165 MPa. In both models, peak stresses in the fixation screws were observed at the first screw superior to the acetabular region. In the PSI, the peak stress in this region is approximately 197 MPa, comparable to the conventional implant with a peak value of 209 MPa. All stresses in the fixation materials were below their respective yield stresses (250 MPa for stainless steel, 875 MPa for titanium [16]) and thus confirm the biomechanical testing where no failure or plastic deformation is found.
Figure 5.

Contour plots of the von Mises stress (MPa) of the synthetic bone of the PSI model (a) and conventional implant model (b) and the von Mises stress (MPa) of the PSI (c) and conventional implant model (d) under 3200N loading.
Displacement contour plots in Figure 6 illustrate a total displacement of 0.87 and 0.86 mm for the load applicator, however, it should be noted that 0.7 mm of this was contributed due to the first step of the contact in the simulation. Six points alongside the fracture line were measured, at similar locations as those for the tracker software (Figure 2f). The fragment gapping at the PSI FEA model were 0.08, 0.00, 0.05 mm for point 1, 2, 3, respectively. The fragment gapping at the conventional implant FEA model were 0.11, 0.01, 0.05 mm for point 1, 2, 3, respectively.
Figure 6.

Displacement contour plots in mm model with patient‐specific implant (a) and model with conventional implant (b) under 3200N loading. The initial movement until contact was 0.7 mm.
4. Discussion
A standardized workflow was developed to enable reproducible mechanical testing of posterior wall acetabular fractures, supported by the 3D preparation process. The mechanical tests revealed no significant differences between a patient‐specific and conventional implants for posterior wall fixation in terms of construct stiffness during a static 3200 N load test, stability over 10,000 cycles, or fracture gapping. FEA supported these findings, showing no plastic deformation or signs of implant failure. These initial results suggested that both the PSI and the conventional implant, using a seven‐hole plate construct with one or two screws through the plate in the fracture fragment, are viable options for treating large posterior wall fragments.
The biomechanical behavior of posterior wall fractures of the acetabulum has been explored in previous studies, but comparisons are challenging due to variations in test setups and loading conditions. For example, Zhang et al. [17] examined displacement across three fixation constructs using cadaveric bones, including the entire pelvis and both femurs, but only simulated six cycles of 1500 N, finding no significant differences. Wu et al. [18] used synthetic bones like in our study, assessing four different osteosyntheses constructs under 350 N for 500 cycles, followed by load‐to‐failure testing. Their failure load, defined as an abrupt change in the load‐displacement curve, was significantly lower than the forces applied in our study (2332–2942 N), however, their loading direction targeted the posterior wall directly, with the hip flexed at 75 degrees and is thus not comparable to our study and had less clinical relevance. Pease et al. [19] investigated three different fixation methods, using the same synthetic bone, with a total of 12,000 cycles under 1000 N and 1500 N loads. Their reported construct stiffness was significantly higher than ours (11.0, 5.6, and 4.9 kN/mm compared to our range of 2.7–3.1 kN/mm). This difference may be attributed to variations in loading direction and fixation method. Their iliac crest was mounted in Woods metal, whereas our model is clamped only at the SI joint and the pubic symphysis. This likely allows for greater movement in our model, which is representative for clinical practise. With the lack of comparison abilities currently available, our study contributes to the existing literature by providing a comprehensive description of the entire workflow, offering a detailed protocol that can be applied in future research to enable more consistent and comparative studies on biomechanical testing of pelvic and acetabular fractures.
FEA is a powerful computational technique that enables the comparison of various loading conditions and design alternatives. The results of this study demonstrated that the peak stresses within the fixation material remains below their respective yield stresses, indicating that no plastic deformation or failure would occur under the given loading conditions. High peak stresses were observed primarily within the acetabular region, likely due to the loading applied by the load applicator. This finding was validated by our mechanical tests, where no failure or plastic deformation occurs within the fixation material, and only small cracks are observed at the surface of the acetabular region, which aligns with the peak stress locations identified in the FEA. In a 2022 literature review [20], only 2 out of the 36 biomechanical studies on acetabular testing utilized both a physical mechanical test setup and a FEA [21, 22]. Unfortunately, comparison with these studies was not feasible. Aziz et al. [22] investigated a different type of acetabular fracture, specifically a posterior hemi‐transverse fracture, and also included the presence of a THA. Additionally, they secured the synthetic bone to the testing setup with multiple screws and used a lower loading condition (2207 N). On the other hand, Shim et al. [21] did focus on a posterior wall fracture, but their testing setup used a foam cortical shell model, which has distinct material properties compared to the Sawbones fourth‐generation models employed in our study. This FEA was a simplified model of the biomechanical test setup, focusing only on the applied load. However, the advantages of FEA are great, and models like this can be expanded to include various fracture types and PSI‐designs. Detailed patient‐specific models, incorporating real fracture types and material properties as recently introduced by our group [9], can be utilized to investigate personalized biomechanics. Ultimately, after thorough validation, these models could be applied in clinical settings to predict load‐bearing capacities and improve patient care.
Both conventional and PSIs, consisting of a seven‐hole plate construct with one or two screws through the plate in the fracture fragment, were able to withstand a peak force of 3.5 times bodyweight, with no significant differences observed between groups. In our study, we applied a relatively high load of 3200 N, exceeding the median load of 800 N (range: 250–2450 N) reported in previous studies [20]. However, high forces above 3000 N are realistic and can easily occur during daily activities, including walking [23, 24, 25, 26]. Although minimal fragment movement (range 0–0.8 mm) is observed, the fragments remained secured, with no fractures or loss of stiffness detected after 10,000 cycles. According to ASTM International standards, one million cycles is considered a suitable benchmark for assessing the endurance limit in biomechanical testing [20, 27]. Additionally, Bergman et al. [28] reported a median of almost 1.4 million cycles per year for normal patients. Olsen et al. [29] suggest that mechanical support from the device is no longer required once the bone has healed, typically after approximately 200,000–250,000 cycles of mechanical loading. These higher numbers of cycles would require 111–780 h of testing with 0.5 Hz per cycle, which is not feasible in our study. Instead, we selected 10,000 cycles to simulate the first four postoperative weeks, a clinically relevant timeframe. This period includes 1 week of rest for wound healing, followed by 3 weeks of partial recovery, with patients performing approximately 1000 steps per day. At 4 weeks, reconstructed bone takes on up to 80% of the load‐bearing capacity [8]. Future studies could explore higher cycle counts to further validate these findings. Our biomechanical results suggest that, for posterior wall fragments, both implants provide sufficient stability to support weight‐bearing immediately after surgery. However, achieving this stability depends on proper fracture reconstruction and adequate bone grip, and should be assessed during surgery. A previous pilot study illustrated that the precise fit of the PSI facilitated easier fracture reconstruction by acting as a template, which could represent a potential advantage [3, 4]. Nevertheless, both implant types were biomechanically capable of achieving comparable stability. Its shape reflects the intended fracture reduction, which may offer a potential advantage in achieving accurate alignment during surgery.
This study had some limitations. First, we used synthetic bones rather than human cadaveric bones. Synthetic bones are commonly used in biomechanical testing because they mimic the mechanical behavior of human bone while exhibiting less variance, making them well‐suited for comparative studies. Their use also means that soft tissue interactions were not accounted for. Additionally, small discrepancies were observed in their geometry, leading to our second limitation. These discrepancies may have slightly influence on the results, since the inserts are universal, so the bones are able to slightly move around. Nevertheless, as they were present in both groups, we consider their impact negligible. During testing, superficial cracks and fractures were noted in the acetabular region precisely where the load is applied, likely due to load application by the metal sphere. In real‐life scenarios, the femur would not exhibit the same rigidity or rounded structure, and a cartilage layer would distribute the load into the joint more evenly. As a result, peak forces concentrate at the crack sites. Since these cracks would not typically occur under physiological conditions such as walking, we attribute them to the test setup design and excluded them as indicators of failure (see File S1A). Peak stresses at these regions are also observed in the FEA. Another limitation concerns the fracture gapping measurements, which were only measured during one test. More precise and comprehensive measurements in future studies would be valuable and could provide better insight into the fracture gap behavior. Lastly, the FEA had some limitations. It was an initial‐stage model with simplifications, meaning it does not fully replicate an in‐situ scenario including other tissues (e.g., muscles). Only one synthetic bone was scanned using a CT scanner, and the material properties and geometry of this single model were applied across all simulations. This uniformity could have introduced minor discrepancies in the results. A mesh convergence study was not performed in this study and mesh sizes were taken from a previous study [9]. Finally in this model, screws were represented as simple rods rigidly attached to the bone, assuming firm fixation. Studies have shown that this simplification affects outcomes by less than 5% compared to models incorporating screw threads or friction‐based interactions [11, 12]. However, in clinical practice, factors like decreased bone quality due to aging or osteoporosis can lead to weaker screw fixation or slippage and should be noted. Future perspectives include the investigation of more fracture types and different implant and screw configurations.
5. Conclusion
This study introduces a reproducible workflow for biomechanical testing of acetabular fractures using synthetic bone models, supported by the 3D preparation process of the customized surgery saw and drill guides and the 3D printed inserts. The extensive workflow demonstrates potential for investigating clinically relevant questions. Initial results indicate that both a PSI and the tested conventional implant are viable options for the fixation of large posterior wall fragments. When these implants are combined with at least two screws on either side of the bone fragment and one screw within the fragment, adequate reconstruction is achieved, and stable screw fixation is ensured, a stable construction appears to be achieved. Further clinical trials are needed to investigate whether partial load bearing might be safely allowed, potentially enhancing the rehabilitation process.
Author Contributions
Miriam G. E. Oldhoff: conceptualization, methodology, investigation, formal analysis, writing – original draft. Anne M. L. Meesters: conceptualization, methodology, writing – original draft. Joep Kraeima: writing – review and editing, supervision, funding acquisition. Mehool R. Acharya: conceptualization, writing – review & editing, Kaj ten Duis: investigation, writing – review and editing. Jean‐Paul P. M. deVries: writing – review and editing, supervision. Bram J. Merema: conceptualization, methodology, investigation, formal analysis, writing – original draft. Frank F. A. IJpma: conceptualization, methodology, investigation, writing – original draft, supervision, funding acquisition. All authors have read and approved the final submitted manuscript.
Supporting information
Figure SD1: Photograph of the inner surface of the acetabulum of a sawbone model, with drawn‐on fracture lines representing the cracks occurred during testing of both the PSI and CONV models.
Acknowledgments
We extend our sincere gratitude to the UMCG Instrumental Research Facility for their work in manufacturing the mechanical setup. We thank the UMCG Innovation Center, particularly Elena Merlo, Vincent Huisman, and Jan Willem Veldsink, for their support. We also wish to thank the Department of Biomaterials and Biomedical Technology at UMCG for their testing machine, especially Joop de Vries for his assistance with the testing facilities and prof. dr. Helder Santos for his supervision. Finally, we would like to acknowledge Witec Medical B.V. (Stadskanaal, The Netherlands) for their contribution in manufacturing the patient‐specific osteosynthesis material. This study was supported, within the Just Transition Fund program IMPACT‐implants, co‐financed by the European Union, with co‐financing from the Ministry of Economic Affairs and Climate and coordinated by Samenwerkingsverband Noord‐Nederland (SNN).
Contributor Information
Miriam G. E. Oldhoff, Email: m.g.e.oldhoff@umcg.nl.
Frank F. A. IJpma, Email: f.f.a.ijpma@umcg.nl.
References
- 1. Kelly J., Ladurner A., and Rickman M., “Surgical Management of Acetabular Fractures–A Contemporary Literature Review,” Injury 51, no. 10 (2020): 2267–2277. [DOI] [PubMed] [Google Scholar]
- 2. Giannoudis P. V., Grotz M. R., Papakostidis C., and Dinopoulos H., “Operative Treatment of Displaced Fractures of the Acetabulum: A Meta‐Analysis,” Journal of Bone and Joint Surgery. British Volume 87, no. 1 (2005): 2–9. [PubMed] [Google Scholar]
- 3. IJpma F., Meesters A., Merema B., et al., “Feasibility of Imaging‐Based 3‐Dimensional Models to Design Patient‐Specific Osteosynthesis Plates and Drilling Guides,” JAMA Network Open 4, no. 2 (2021): e2037519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Merema B. J., Kraeima J., ten Duis K., et al., “The Design, Production and Clinical Application of 3D Patient‐Specific Implants With Drilling Guides for Acetabular Surgery,” Injury 48, no. 11 (2017): 2540–2547. [DOI] [PubMed] [Google Scholar]
- 5. Lin A. S., Fechter C. M., Magill M., Wipf F., Moore T., and Guldberg R. E., “The Effect of Contouring on Fatigue Resistance of Three Types of Fracture Fixation Plates,” Journal of Orthopaedic Surgery and Research 11 (2016): 107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Bergmann G., Graichen F., Rohlmann A., et al., “Realistic Loads for Testing Hip Implants,” Bio‐Medical Materials and Engineering 20, no. 2 (2010): 65–75. [DOI] [PubMed] [Google Scholar]
- 7. Zdero R., Brzozowski P., and Schemitsch E. H., “Biomechanical Properties of Artificial Bones Made by Sawbones: A Review,” Medical Engineering & Physics 118 (2023): 104017. [DOI] [PubMed] [Google Scholar]
- 8. Green D. P., Rockwood and Green's Fractures in Adults (Lippincott Williams & Wilkins, 2010). [Google Scholar]
- 9. Oldhoff M., Kamal Z., Ten Duis K., et al., “Semi‐Automated Finite Element Analyses of Surgically Treated Acetabular Fractures to Investigate the Biomechanical Behaviour of Patient‐Specific Compared to Conventional Implants,” Journal of Orthopaedic Surgery and Research 19, no. 1 (2024): 541. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Elenes E. Y., Roan E., Marinescu R. C., and Janda H. A., “Editors. Development and Validation of a 2D/3D Finite Element Model of a Composite Hemipelvis,” ASME International Mechanical Engineering Congress and Exposition (2010): 401–409. [Google Scholar]
- 11. Inzana J. A., Varga P., and Windolf M., “Implicit Modeling of Screw Threads for Efficient Finite Element Analysis of Complex Bone‐Implant Systems,” Journal of Biomechanics 49, no. 9 (2016): 1836–1844. [DOI] [PubMed] [Google Scholar]
- 12. MacLeod A. R., Pankaj P., and Simpson A. H., “Does Screw–Bone Interface Modelling Matter in Finite Element Analyses?,” Journal of Biomechanics 45, no. 9 (2012): 1712–1716. [DOI] [PubMed] [Google Scholar]
- 13. Fu S., Zhao Y., Lian W., et al., “Comparison of the Risk of Breakage of Two Kinds of Sacroiliac Screws in the Treatment of Bilateral Sacral Fractures,” European Spine Journal 23 (2014): 1558–1567. [DOI] [PubMed] [Google Scholar]
- 14. Ji T., Guo W., Tang X. D., and Yang Y., “Reconstruction of Type II+ III Pelvic Resection With a Modular Hemipelvic Endoprosthesis: A Finite Element Analysis Study,” Orthopaedic Surgery 2, no. 4 (2010): 272–277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lei J., Liu H., Li Z., Wang Z., Liu X., and Zhao L., “Biomechanical Comparison of Fixation Systems in Posterior Wall Fracture of Acetabular by Finite Element Analysis,” Computer Assisted Surgery (Abingdon, England) 21, no. 1 (2016): 117–126. [DOI] [PubMed] [Google Scholar]
- 16. Rack H. and Qazi J., “Titanium Alloys for Biomedical Applications,” Materials Science and Engineering: C 26, no. 8 (2006): 1269–1277. [Google Scholar]
- 17. Zhang Y., Tang Y., Wang P., Zhao X., Xu S., and Zhang C., “Biomechanical Comparison of Different Stabilization Constructs for Unstable Posterior Wall Fractures of Acetabulum. A Cadaveric Study,” PLoS One 8, no. 12 (2013): 82993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Wu X., “A Biomechanical Comparison of Different Fixation Techniques for Fractures of the Acetabular Posterior Wall,” International Orthopaedics 42 (2018): 673–679. [DOI] [PubMed] [Google Scholar]
- 19. Pease F., Ward A. J., Stevenson A. J., et al., “Posterior Wall Acetabular Fracture Fixation: A Mechanical Analysis of Fixation Methods,” Journal of Orthopaedic Surgery (Hong Kong) 27, no. 3 (2019): 2309499019859838. [DOI] [PubMed] [Google Scholar]
- 20. Hinz N., Dehoust J., Münch M., et al., “Biomechanical Analysis of Fixation Methods in Acetabular Fractures: A Systematic Review of Test Setups,” European Journal of Trauma and Emergency Surgery 48, no. 5 (2022): 3541–3560. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Shim V., Böshme J., Vaitl P., Josten C., and Anderson I., “An Efficient and Accurate Prediction of the Stability of Percutaneous Fixation of Acetabular Fractures With Finite Element Simulation,” Journal of Biomechanical Engineering 133 no. 9 (2011): 094501. [DOI] [PubMed] [Google Scholar]
- 22. Aziz M., Dessouki O., Samiezadeh S., Bougherara H., Schemitsch E. H., and Zdero R., “Biomechanical Analysis Using FEA and Experiments of a Standard Plate Method Versus Three Cable Methods for Fixing Acetabular Fractures With Simultaneous THA,” Medical Engineering & Physics 46 (2017): 71–78. [DOI] [PubMed] [Google Scholar]
- 23. Bergmann G., Bender A., Dymke J., Duda G., and Damm P., “Standardized Loads Acting in Hip Implants,” PLoS One 11, no. 5 (2016): 0155612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Damm P., Graichen F., Rohlmann A., Bender A., and Bergmann G., “Total Hip Joint Prosthesis for In Vivo Measurement of Forces and Moments,” Medical Engineering & Physics 32, no. 1 (2010): 95–100. [DOI] [PubMed] [Google Scholar]
- 25. Layton R., Messenger N., and Stewart T., “Characteristics of Hip Joint Reaction Forces During a Range of Activities,” Medical Engineering & Physics 108 (2022): 103894. [DOI] [PubMed] [Google Scholar]
- 26. Lu T.‐W., Taylor S. J., O'connor J. J., and Walker P. S., “Influence of Muscle Activity on the Forces in the Femur: An In Vivo Study,” Journal of Biomechanics 30, no. 11–12 (1997): 1101–1106. [DOI] [PubMed] [Google Scholar]
- 27. ASTM , F384‐17. Standard Specifcations and Test Methods for Metallic Angled Orthopedic Fracture Fixation Devices (ASTM International, 2017). [Google Scholar]
- 28. Bergmann G., Graichen F., Rohlmann A., et al., “Realistic Loads for Testing Hip Implants,” Bio‐Medical Materials and Engineering 20, no. 2 (2010): 65–75. [DOI] [PubMed] [Google Scholar]
- 29. Olson S. A., Marsh J. L., Anderson D. D., and Latta Pe L. L., “Designing a Biomechanics Investigation: Choosing the Right Model,” Journal of Orthopaedic Trauma 26, no. 12 (2012): 672–677. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure SD1: Photograph of the inner surface of the acetabulum of a sawbone model, with drawn‐on fracture lines representing the cracks occurred during testing of both the PSI and CONV models.
