1. Introduction
Acetabular fractures are relatively uncommon, with an incidence rate ranging from 3 to 8.1 cases/100,000 person/year. These fractures have a high impact on patient quality life and are mostly seen in younger patients and persons aged over 50 years old [
1,
2]. Traditionally, acetabular fractures are classified according to Letournel classification into ten major fracture patterns, which consist of five simple patterns and five complex patterns [
3,
4]. Due to the complexity of the pelvic anatomic structure, acetabular fractures represent a challenging procedure for orthopedic surgeons.
At present, a well-established surgical procedure in the treatment of acetabular fractures is represented by the anterior intrapelvic approach (AIP) [
5,
6,
7]; this technique allows for the minimization of operation times and risks related to blood loss during surgical intervention [
5,
6]. At any rate, the purpose of the surgical treatment is the achievement of anatomical reduction and stable fixation. In transverse and T-shaped fractures there is no bony structure able to oppose to medial subluxation of the femoral head since the entire medial wall is detached from the superior aspect of the dome. Therefore, systems able to fix both columns are traditionally used in these fracture patterns. Recently, a new pre-contoured anatomic plate design has been developed aiming at fixation of both column and quadrilateral plate [
8]. Its anatomic shape was designed to be implanted with its native curvature, but this latter could not be always adequate to perfectly match to the bone shape. Thus, further in-situ bending is often indispensable [
6], with a consequent increase in surgical times and modification of the performance of the native device.
In this framework, a lack of biomechanical data regarding these plates in operative conditions emerges. Many examples are present in the literature about transforming the qualitative assessment of clinical problems to quantitative by exploiting numerical modeling [
9,
10,
11,
12,
13,
14,
15]. Here, finite element modeling was used to biomechanically examine different bone-plates configurations.
The aim of this study is therefore to compare the stability guaranteed by two different suprapectineal plates: (1) a standard pelvic plate (SPP) and (2) a suprapectineal quadrilateral surface buttressing plate (SQBP).
3. Results
Results were at first analyzed focusing on Von Mises stresses and displacements distributions. In the regions of plate-bone contact, highest stresses are located at the pubic tubercle (
Figure 6a). Only exception is the SQBP C1 plate in the T-shaped fracture, which shows a maximum stress of 174.4 MPa near screw hole number 15 (
Figure 6b). For both fracture types fixed with the SQBP C2 plate, a clear reduction of the stresses at the pubic tubercle induced by the bone-plate larger contact is visible, whilst in proximity of the screw hole number 9, middle-low values of stresses (almost 20 MPa) can be noted (
Figure 6b). Conversely, comparing the SQBP C1 and the SPP C2 fixation systems within the acetabular cup region (
Figure 6c), in both fracture types the most solicited part consists in the antero-superior dial, condition determined by the direction of the external applied load, as already observed by Dalstra and Huiskes [
41]. Here in fact, the maximal Von Mises stresses are on average 55 MPa with peaks up to 70 MPa; in the acetabulum remaining part the values range between 4 MPa (in the center of the acetabulum) and 23 MPa. From a qualitative evaluation, it was noted how the transverse fractures create more uniform stress distributions, suggesting a more homogeneous load transfer through the bone; this is probably due to the absence of the additional fracture of the T-shaped configuration, that represents an obstacle to the stress propagation.
Comparing the SPP C2 and the SQBP C2 configurations, no macroscopic difference is visible in the stress distribution pattern within the same fracture type: stresses seem indeed to follow the same propagation curves.
Considering the hemipelvis globally, the highest value of stress among the combinations is found in the T-shaped fracture models: 382 MPa for the SQBP C1, 412 MPa for the SPP C2 and 337.8 MPa for the SQBP C2. This phenomenon is directly related to the conformation of the fracture which permits a rigid motion of the bone stumps decreeing frictional contacts. In
Table 6 the maximal Von Mises stresses found in the specific regions of the pelvis are reported.
The main outcome of the displacement analysis (
Figure 7) is related to the differences between the two fracture types. In particular, all the T-shaped fracture models show the less stable behavior. In relation to the described situation, the SQBP C1 model has the highest magnitude of displacement equal to 3.1 mm, conversely in the SPP C2 model the maximum value is 2.33 mm. Observing the C2 models (both SPP and SQBP plates), similar behaviors arise. In fact, changing the plate, the differences are negligible: in both configurations the transverse fracture reports a maximal value of 0.96 mm.
3.1. Interfragmentary Movement Analysis (IFM)
3.1.1. Axial Strain
Comparing the configurations that differ both in terms of plate and screws configuration (SQBP C1 and SPP C2), it is possible to note two main differences regarding the elementary transverse fracture:
The median direction of the incremental distances (
Figure 8b) of SPP C2 is mainly oriented towards a minimum axial expansion (median 1.57%) of the fracture gap; conversely SQBP C1 mainly brings the fracture gap towards minimum compressions (median −0.61%).
The slightly greater quantity in the SPP C2 model of observations satisfying the optimal range, ascertained by the percentage histogram in
Figure 8a (36% against 31% of SQBP C1).
Comparing the T-shaped fracture, in its transverse component (
Figure 8b) SQBP C1 shows a better performance, both related to the number of points couples belonging to the optimal range (45.5% vs. 41%) and to the acceptability range (68% vs. 55%). As far as the vertical component of the T shaped fracture is concerned, a very high axial compression is visible in both models, which generates an almost closure of the fracture gap. The minimal values are indeed respectively −90% (median −68%) for the SQBP C1 model and −69% (median −48%) for the SPP C2 model, which both result outside the optimal and the acceptable range (
Figure 8b).
Comparing the C2 screw configuration models (SPP C2 and SQBP C2), the same behavior emerges according to the histogram percentages (
Figure 8a): 36% of the couples belong to the optimal range for both the transverse elementary fractures, 41% of the occurrences satisfy the optimal range for both the transverse components of T-shaped fractures and, eventually, only almost 7% of occurrences are within the transition range as regards the two vertical components of T-shaped fractures.
3.1.2. Shear Strain
The distribution of shear strain among the various fractures shows that none of the configurations report an optimal behavior in blocking sliding motions in the fracture plane (0% shear strain). Comparing the models SQBP C1 and SPP C2, when the elementary transverse fracture is considered the former has lower shear strain values and almost 82% of the values are within the acceptable range (0–15)%, while the latter is characterized by a greater variability of shear strain in a range between 4% and 65% and with an acceptability percentage of 18% (
Figure 8c). As far as the T-shaped fracture is concerned, in its transverse component a similar behavior between the two models is visible: for the SQBP C1 the shear strain values range between 5% and 77%, whilst for the SPP C2 range between 10% and 35%. Respectively, the percentage of couples of points within the acceptable range is 9% and 14%. For the vertical component of the T shaped fracture, the SQBP C1 values range between 18% and 27%, meaning that no points satisfy the acceptable range; in SPP C2 configuration, on the other hand, the shear strain is wider, between 10% and 50%, but with 20% of the points within the acceptable range.
If the C2 screws configurations models are compared, it can be noted that for the T-shaped fracture the differences are minimal: in the transverse fracture the range of shear strain distribution (10%, 36%) is almost identical (
Figure 8d), and it is feasible to assess the same value of median equal to 26% with the same percentage of values belonging to the acceptable range (16.64%). In the vertical component of the T shaped fracture, a range of distribution equal to (10%, 50%) is reached, with median equal to 22%, for the SPP plate. The SQBP plate reaches a very similar range of values (9%, 47%) with median 21%. For both plates, the same percentage of 20% of the points within the acceptable range was computed. The substantial difference between the two plate is in the elementary transverse fracture: SPP plate has an acceptable percentage of 18% while the SQBP plate the percentage increases to 90%; in this latter, indeed, the range of shear strain is confined between 5 % and 16% with a median of 9%.
4. Discussion
The acetabular fractures treatment aims at the obtainment of an anatomical reduction, in order to minimize the risk of post-traumatic osteoarthritis, and of a stable fixation, able to maintain this reduction during the healing process. In transverse and T-shaped fractures, the entire medial portion of the acetabulum is detached from the upper portion so there are no bone structures capable of preventing medial subluxation of the femoral head. For this reason, traditionally, systems capable of fixing both columns are used in these fracture patterns. With this aim, plates that allow fixation of both columns and the quadrilateral plate through the anterior intrapelvic access have been developed. However, literature on the biomechanical stability guaranteed by these plates is limited.
The aim of this study was to compare, from a biomechanical point of view, the suprapectineal plate (SQBP) and the traditional reconstruction plate (SPP) in the treatment of transverse and T-shaped fractures of the acetabulum.
Observing the generated plate geometries, their property of following the articulate shapes of the hip stands out. The possibility to deform a 3D CAD plate design on the patient geometry, which immediately could fit on the patient’s pelvic bone, would allow the surgeon to further minimize operation times and risks related to blood loss during surgical intervention with respect to the standard plane plates. Although rapid manufacturing techniques are emerging in the orthopedic devices production, the diffusion of these methodologies is still limited, especially in the field of load bearing plates [
42]. Therefore, if the direct production of this patient specific plate through standard manufacturing technologies would comport realizability difficulties, an alternative approach could be implemented: thanks to the generation of a mold CAD model on the basis of the patient-adapted 3D CAD design, obtained in accordance with the procedure described in this work, the surgeon can model the plate on the physical mold manufactured in plastic materials through a low-cost 3D printing [
43] thus obtaining a patient-adapted plate. This preoperative planning procedure, above all, would abolish any in-situ bending, reducing intervention time and preserving the patient from surgical risks as bleeding and infections [
43].
Our plates comparison should be discussed in details. Focusing on the stress distribution, the models with the most solicited bone-plate contact zones are the ones with highest stress detected on the plate: the SPP plate showed the highest stresses manifestation. Furthermore, the SPP plate showed a more flexible behavior with respect to the SQBP plate because of its ability to adapt to the bone shape and thus it transfers more easily the load to the bone. The SQBP C2 configuration would result conversely stiffer and with lower strains present in the suprapectineal portion than the others, decreeing lower induced stresses to the bone. Even with equal screws configuration, the SQBP plate allows the reduction of stresses on the bone. Taking into account the results of the IFM, the infrapectineal portion of the SBQP plate, although without screws, brings improvements in the shear strain reduction (four times less than the standard plate: 65% vs. 16%). The same results have not been noticed on the axial compressions: on this axis, the two plates seem to similarly behave.
The SQBP plate with the screw configuration C1 achieves the best results in transverse fracture: it imposes an axial compression to the fracture gap (although the optimal range is only reached in 31% of points couples) and at the same time it minimizes the shear strain (which is within the acceptable range in 82% of the measurements).
The same behavior was not demonstrated in T shaped fractures: although the SQBP plate with the screw configuration C1 produces an optimal axial compression in 45.5% of the occurrences, shear strains are generated, reaching up the 77% of the average fracture gap wideness. The latter may compromise the secondary stability.
Regarding the axial strain in the vertical component of T-shaped fractures, all the plates and screw configurations are unable to prevent such a high compression of the fracture gap (the SPP C2 combination remains below 70% in compression whilst the SQBP C1 combination comports compression up to 90%).
From the shear strain point of view, the combination SQBP C1 has the most restrict distribution (reaching the value of 28%), probably due to the infrapectineal portion that would hold ischium stump with a better adhesion.
To the authors’ knowledge, no computational study has involved the SQBP plate performance assessment, therefore, direct comparisons with literature are arduous. From a clinical point of view, our results support the findings by Kistler et al. [
8], who biomechanically compared through cyclic mechanical testing five different fixation methods in transtectal transverse fractures. According to their results, authors stated that quadrilateral surface buttress plates are biomechanically comparable and, in some cases, superior to traditional forms of fixation in this synthetic hemipelvis model. Similar results have been reported by Chen et al. [
44], that evaluated the anterior column-posterior hemitransverse fractures. Chen and colleagues concluded that, although the suprapectineal pelvic brim plate with three periarticular long screws remains the gold standard in this fracture, the suprapectineal and infrapectineal QLS buttress plate are at least comparable with standard forms of fixation at resisting fracture motion and medial subluxation.
The results reported herein should be considered in the light of two major limitations, that could be addressed in future research. At first, fixed constraints at the pubic symphysis and at the iliac crest helped the authors to lighten the simulation, but represent a more critical situation than the physiological condition, characterized by flexible joints in these anatomical areas. Moreover, a single loading condition was investigated, neglecting muscle forces which could locally influence the stress state of the system. In any case, the comparative nature of the study makes the results obtained reliable, albeit not generalizable. Secondly, although according to literature two cartilages (one acetabular and one femoral) are present, a unique acetabular cartilage was here considered, attributing to its thickness the sum of the of the two original cartilages. This choice was considered reasonable given the purpose of the study. Lastly, the bonded contact imposed between bone and screws is a simplification of the real solicitation at the bone-screw interface, which could have influenced the distribution of stresses around the screws. However, fixations stability was assessed through the IFM analysis, which relies on the fracture fragments displacements. It is reported in the literature that, despite locally the bonded constraint does not allow relative motion between the screw and the bone, thus developing tension where separation would occur in reality, the interface modeling strategy has almost no influence on the global load-deformation behavior [
18]. Therefore, the chosen approach did not influence, in the authors’ opinion, the considerations made regarding the stability of the constructs.