Fatigue and FEA of zygomatic implant systems in bilateral maxillary defects Authors: Hao Bai, Wenyan Wu, Xing Wang, Zhen Wang Published: 2025 ABSTRACT PurposeThis study aimed to determine the optimal number of zygomatic implants required to restore bilateral maxillary defects, where conventional implants are contraindicated due to insufficient bone support. It focused on evaluating the mechanical performance and failure mechanisms of different implant configurations.MethodsA controlled in vitro model simulating bilateral maxillary defects was constructed using a representative patient case. Zygomatic implant systems with two, three, or four implants were evaluated under vertical cyclic loading applied at the first molar region. Fatigue testing was performed alongside finite element analysis to assess stress distribution, failure patterns, and load-bearing capacity.ResultsThe four-implant configuration completed all fatigue cycles with minimal displacement and sustained the highest static load (741.7 N). FEA confirmed that this group exhibited the lowest von Mises stress across implants and supporting structures. The three-implant configuration also passed fatigue testing but demonstrated increased displacement, lower load capacity (456.7 N). The two-implant configuration failed prematurely at 33,483 cycles, with FEA showing implant stress exceeding the yield strength of titanium, confirming mechanical insufficiency.ConclusionThis work represents a novel integration of in vitro mechanical testing and FEA in a model simulating bilateral maxillary defects, providing experimentally validated guidance for implant selection. While all configurations exhibited failure modes confined to replaceable components such as screws, reduced implant groups demonstrated significantly lower mechanical thresholds and earlier failure. A four-implant configuration is therefore strongly recommended for clinical use due to its superior biomechanical stability, whereas two-implant configurations are not advisable in functionally demanding scenarios. Background The maxilla is a central structural component of the midface, separating the oral, nasal, and orbital cavities while providing foundational support to the orbit, cheeks, lips, and nose. It plays a critical role in essential functions such as speech, mastication, swallowing, and facial expression [1]. Consequently, bilateral maxillary defects can lead to profound impairments in these functions and cause severe facial deformities. Patients suffering from such defects often experience persistent functional disability, compromised aesthetics, and chronic psychosocial distress [2]. Among these, bilateral maxillary defects require simultaneous attention to functional restoration and aesthetic rehabilitation. Addressing these challenges, zygomatic implants have emerged as a promising surgical alternative. First introduced by Brånemark’s team in 1998 for severe maxillary atrophy, congenital defects, and tumor resection, zygomatic implants have become a widely adopted solution to rehabilitate maxillary defects where conventional implants are infeasible [3, 4]. By anchoring in the dense zygomatic bone, they provide reliable support and offer a graftless approach for bilateral maxillary reconstruction, enabling predictable, minimally invasive surgery [5]. Clinically, three- and four-implant configurations have been reported for supporting maxillary prostheses in bilateral maxillary defects [6, 7]. However, considerable anatomical variation—including limited zygomatic bone volume, soft tissue loss, or the presence of oronasal fistulae—can preclude the placement of two implants per side [8]. Critically, despite widespread clinical use, the biomechanical implications—particularly the minimum number of zygomatic implants required for structural stability—remain poorly understood and lack robust experimental validation. To address this gap, the present study established a controlled in vitro model simulating bilateral maxillary defects reconstructed with zygomatic implants. It combines dynamic fatigue testing and finite element analysis (FEA) to comprehensively evaluate the biomechanical performance of various implant configurations. The analysis focused on identifying differences in stress distribution, failure patterns, and load-bearing capacity among configurations. This study aimed to determine the minimum number of zygomatic implants required to ensure structural stability in bilateral maxillary reconstructions, thereby providing clinically relevant biomechanical evidence to guide implant selection. The null hypothesis was that two zygomatic implants are sufficient to meet functional and mechanical demands. Methods A representative clinical case involving a patient with bilateral maxillary defects was selected for this study (Fig. 1). Zygomatic implant placement was performed under general anesthesia in accordance with standard surgical protocols, and prosthetic rehabilitation was subsequently completed. A total of four zygomatic implants were placed to ensure optimal primary stability.Fig. 1Clinical workflow and outcome of zygomatic implant-supported prosthetic rehabilitation in a patient with a maxillary defect. A Surgical placement of zygomatic implants; B Intraoral view of the defect; C Customized aesthetic prosthesis; D Patient wearing the prosthesis Clinical workflow and outcome of zygomatic implant-supported prosthetic rehabilitation in a patient with a maxillary defect. A Surgical placement of zygomatic implants; B Intraoral view of the defect; C Customized aesthetic prosthesis; D Patient wearing the prosthesis Postoperative impressions were obtained using a splinted open-tray technique to ensure accuracy in capturing implant positions. Digital models were subsequently created using a Medit T710 desktop scanner (Medit Corp) and corresponding 3D ring frameworks were designed with EXOCAD® software based on the actual implant positions. To evaluate the biomechanical performance under different support conditions, specimens were assigned to three experimental groups based on the number of implants used: the Z4 Group (four implants), Z3 Group (three implants), and Z2 Group (two implants). All implants and 3D rings were fabricated from biocompatible, commercially pure grade 4 titanium using a precision turning process. Each implant was fitted with a 3 mm Zygoma Multi-unit straight abutment (Brånemark System®, Nobel Biocare™), which was then connected to a 3D ring to form an integrated structure. To ensure precise and reproducible positioning, a custom 3D-printed guide plate was used to embed each assembly into Plastic Steel® Liquid (B) resin, a rigid material simulating bone support (Fig. 2 A-B). The embedded depth of each implant replicated its intraosseous length measured from postoperative CT scans, thereby simulating clinical placement within the zygomatic bone. Standardized torques of 35 Ncm and 15 Ncm were applied to secure the abutment–implant and ring–abutment interfaces, respectively, ensuring consistent preload across all specimens.Fig. 2Specimen Construction and Fatigue Test Design: showcases the key steps in the specimen preparation and fatigue test process. A Placement of zygomatic implants under the guidance of a 3D-printed surgical guide to ensure precise positioning. B Embedding of the samples in Plastic Steel® Liquid (B) resin to simulate the surrounding anatomical structures and provide stability during testing. C Fixation of the zygomatic implant system using a fixture. D Verification of the impact site to ensure accurate load application at the designated location Specimen Construction and Fatigue Test Design: showcases the key steps in the specimen preparation and fatigue test process. A Placement of zygomatic implants under the guidance of a 3D-printed surgical guide to ensure precise positioning. B Embedding of the samples in Plastic Steel® Liquid (B) resin to simulate the surrounding anatomical structures and provide stability during testing. C Fixation of the zygomatic implant system using a fixture. D Verification of the impact site to ensure accurate load application at the designated location Fatigue testing was conducted using a DYNA-MESS TP 5 LCF servo-pneumatic machine (Aachen/Stolberg, Germany) under room-temperature, dry conditions. Custom-designed fixtures were used to stabilize the specimens, and a sinusoidal vertical compressive load (30–300 N at 10 Hz) was applied directly to the 3D ring at the posterior region of one maxillary side (Fig. 2 C-D). The load was anatomically aligned with the first molar region (Fig. 3). Fatigue testing continued for 5 × 10⁶ cycles or until specimen failure. Failure was detected automatically by a sudden 2 mm displacement or a significant force drop, triggering machine shutdown. Specimens that successfully completed 5 × 10⁶ cycles were considered to have passed the fatigue protocol and were subsequently subjected to static load-to-fracture testing. Collected data included the number of fatigue cycles to failure (if applicable), maximum static load at fracture, and vertical displacement of the loading platen. The fatigue duration was designed to simulate approximately 20 years of clinical service, assuming 1400 masticatory cycles per day [9].Fig. 3Anatomical alignment of the load application site with the first molar region Anatomical alignment of the load application site with the first molar region FEA was performed to evaluate the biomechanical behavior of zygomatic implant system under varying implant configurations (Z4, Z3, and Z2). The analysis focused on stress distribution patterns within the implant system under standardized loading conditions, enabling direct comparison across configurations. To simplify the model and emphasize the mechanical performance of the implant components, the connecting screws were intentionally excluded. The remaining model included the zygomatic implant, abutment, and 3D ring. All Components were modeled in Rhino, meshed in 3-Matic 13.0, and then imported into ANSYS Workbench for simulation. Material properties, loading conditions, boundary constraints, and contact definitions were assigned accordingly. All model components were spatially registered using postoperative CT scans to reproduce clinical implant positioning and anatomical constraints (Fig. 4 A-C). The intraosseous depth of each implant—measured from CT scans—was incorporated into the model as a boundary condition to simulate bone-implant engagement.Fig. 4Assembly of the zygomatic implant-supported system. A Z4 Group; B Z3 Group; C Z2 group Assembly of the zygomatic implant-supported system. A Z4 Group; B Z3 Group; C Z2 group All materials were defined as homogeneous isotropic linear elastic solids with fully bonded interfaces (implant-bone, ring-abutment, abutment-implant), assuming no component motion [10]. To ensure consistency with the in vitro experiment, the zygomatic implants and 3D ring were modeled as commercially pure grade 4 titanium, while the abutments were modeled as titanium alloy. The corresponding material properties, including Young’s modulus and Poisson’s ratio, are listed in Table 1 [11, 12]. The accuracy of the created mesh, the state of the contacts, and the convergence of the model results were analyzed before the created models were considered correct. Table 1Material properties of finite element analysisYoung Modulus(MPa)Poisson Ratio(v)Yield Strength(MPa)Tensile Strength(MPa)ComponentTi(TA4)1.045*1050.357758073D ring、zygomatic implantTi6Al4V1.1*1050.328701070abutment Table 1Material properties of finite element analysisYoung Modulus(MPa)Poisson Ratio(v)Yield Strength(MPa)Tensile Strength(MPa)ComponentTi(TA4)1.045*1050.357758073D ring、zygomatic implantTi6Al4V1.1*1050.328701070abutment Material properties of finite element analysis A static vertical load of 300 N was applied directly to the 3D ring positioned at the posterior region of the unilateral maxilla, corresponding to the in vitro testing location. This standardized load was consistently applied across all configurations to enable direct comparison of stress distribution and mechanical response among models. Results A clear correlation was observed between the number of implants and the biomechanical stability of the reconstruction. Under sinusoidal loading conditions (30–300 N, 5 million cycles), both the Z4 Group and Z3 Group completed fatigue testing without structural failure. Static load testing revealed an ultimate fracture load of 741.7 N for Z4 and 456.7 N for Z3, as presented in Table 2. In contrast, the Z2 Group failed at 33,483 cycles. Notably, no abutment/implant failures occurred across groups. Table 2Presents the fatigue and static test results for each implant configuration under varying load ranges, including the number of cycles until failure (if applicable), the maximum displacement before failure, and the maximum load at failureLoading Force(N)Number of Zygomatic ImplantsCycles Until FailureMaximum of Movement before Failure(mm)Maximum Load before Failure(N)30–30045*1060.14741.735*1062.68456.7233,4832.06302.6 Table 2Presents the fatigue and static test results for each implant configuration under varying load ranges, including the number of cycles until failure (if applicable), the maximum displacement before failure, and the maximum load at failureLoading Force(N)Number of Zygomatic ImplantsCycles Until FailureMaximum of Movement before Failure(mm)Maximum Load before Failure(N)30–30045*1060.14741.735*1062.68456.7233,4832.06302.6 Presents the fatigue and static test results for each implant configuration under varying load ranges, including the number of cycles until failure (if applicable), the maximum displacement before failure, and the maximum load at failure Failure patterns differed across groups. In the Z4 Group, failure was consistently localized to the connecting screw between the 3D ring and the abutment, while the abutment and zygomatic implants remained structurally intact (Fig. 5 A-B). The Z3 Group showed progressive screw fractures, accompanied by visible wear on the abutment surface accumulation of metallic debris, suggesting repeated micromovements under cyclic loading (Fig. 5 C-D). The Z2 Group experienced catastrophic failure prior to completing the fatigue cycles, indicating insufficient mechanical stability under simulated occlusal forces. In this group as well, the failure was confined to the connecting screw, with no deformation or fracture observed in the implant bodies across all configurations.Fig. 5Overview of failure modes and surface changes observed in the Z4 Group and Z3 Group after mechanical testing. A Static failure mode of the Z4 Group under ultimate load: Shear fracture occurred at the screw connecting the ring and abutment, while the implant body exhibited no plastic deformation. B Integrity of the central screw in the Z4 Group: Post-failure examination revealed no fracture in the abutment or central screw, with the threaded structure remaining intact. C Abutment wear and debris accumulation in the Z3 Group: Fatigue testing resulted in titanium alloy debris accumulation at the abutment interface. D Surface morphology of the Z3 Group abutment after dynamic loading: The abutment neck displayed characteristic wear patterns following cyclic loading Overview of failure modes and surface changes observed in the Z4 Group and Z3 Group after mechanical testing. A Static failure mode of the Z4 Group under ultimate load: Shear fracture occurred at the screw connecting the ring and abutment, while the implant body exhibited no plastic deformation. B Integrity of the central screw in the Z4 Group: Post-failure examination revealed no fracture in the abutment or central screw, with the threaded structure remaining intact. C Abutment wear and debris accumulation in the Z3 Group: Fatigue testing resulted in titanium alloy debris accumulation at the abutment interface. D Surface morphology of the Z3 Group abutment after dynamic loading: The abutment neck displayed characteristic wear patterns following cyclic loading Figure 6 illustrates the overall stress distribution patterns observed in each group. The maximum von Mises stress values for each component at direct load-bearing sites are summarized in Table 3.Fig. 6Finite element analysis results obtained under a 300 N vertical load. A Z4 Group; B Z3 Group; C Z2 Group Finite element analysis results obtained under a 300 N vertical load. A Z4 Group; B Z3 Group; C Z2 Group Table 3Maximum von mises stress (MPa) in 3 groupsComponentZ4 GroupZ3 GroupZ2 Group3D ring728.9758.3774.6Abutment826.01007.91881.6Zygomatic implant153.0363.3909.6 Table 3Maximum von mises stress (MPa) in 3 groupsComponentZ4 GroupZ3 GroupZ2 Group3D ring728.9758.3774.6Abutment826.01007.91881.6Zygomatic implant153.0363.3909.6 Maximum von mises stress (MPa) in 3 groups For the 3D ring, stress values remained relatively moderate but progressively increased from 728.9 MPa in the Z4 Group to 774.6 MPa in the Z2 Group. Peak stresses were consistently concentrated at the direct loading site, as shown in Fig. 7, suggesting that a reduction in implant number increases mechanical demand on the ring structure.Fig. 7Distribution of von Mises stress (MPa) in the 3D rings for the three groups. A Z4 Group; B Z3 Group; C Z2 Group Distribution of von Mises stress (MPa) in the 3D rings for the three groups. A Z4 Group; B Z3 Group; C Z2 Group The abutment exhibited the highest stress values among all components. In the Z4 Group, the peak stress was 826.0 MPa, whereas Z3 and Z2 reached 1007.9 MPa and 1881.6 MPa, respectively, indicating that diminished implant support significantly amplifies stress concentration at the abutment level, as illustrated in Fig. 8.Fig. 8Distribution of von mises stress (MPa) in the abutments for the three groups. A Z4 Group; B Z3 Group; C Z2 Group Distribution of von mises stress (MPa) in the abutments for the three groups. A Z4 Group; B Z3 Group; C Z2 Group For the zygomatic implant, von Mises stress increased from 153.0 MPa in the Z4 Group to 363.3 MPa in the Z3 Group, and further to 909.6 MPa in the Z2 Group—surpassing the yield strength of commercially pure grade 4 titanium, which is approximately 775 MPa [11]. This critical stress concentration, visualized in Fig. 9, underscores the biomechanical insufficiency of the two-implant configuration.Fig. 9Distribution of von mises stress (MPa) in the zygomatic implants for the three groups. A Z4 Group; B Z3 Group; C Z2 Group Distribution of von mises stress (MPa) in the zygomatic implants for the three groups. A Z4 Group; B Z3 Group; C Z2 Group Collectively, these results confirm that decreasing the number of implants leads to localized stress intensification in structurally vulnerable components, thereby elevating the risk of fatigue-related failure and compromising the long-term mechanical stability of the prosthetic system. Discussion The present findings indicate that a four-implant zygomatic configuration provides the most reliable biomechanical stability for bilateral maxillary defect reconstruction, whereas three implants represent a functionally compromised but minimally acceptable option, and two implants are mechanically insufficient. This study established a novel in vitro biomechanical model replicating zygomatic implant–supported prosthetic rehabilitation in patients with bilateral maxillary defects. Existing literature on such cases is largely limited to FEA without corresponding experimental validation, and current ISO fatigue testing standards are limited to single-implant restorations rather than complex cross-arch reconstructions [13, 14]. To address this gap, we used a representative clinical case to construct both the in vitro and finite element models, with implant positioning derived from CT data and incorporated prosthetic frameworks incorporating two to four zygomatic implants. All implants, abutments, and 3D rings were modeled using idealized geometries to ensure reproducibility, and the zygomatic implant–bone interface was defined as fully constrained, consistent with the rigid resin embedding used experimentally. To simulate peak physiological loading conditions, vertical cyclic loads were applied directly to the 3D ring at the unilateral posterior maxilla, corresponding to the anatomical location of the first molar—a principal occlusal load-bearing site during mastication [15]. Although clinical mastication also involves oblique and lateral forces, vertical loading represents the dominant functional vector and provides a standardized, reproducible method for biomechanical evaluation. The omission of a full-arch prosthesis further concentrated stress at a single site, thereby creating an extreme condition to evaluate system reliability. Despite these simplifications, the model reproduced the failure modes and displacement patterns observed in vitro, yielding consistent and clinically relevant results. Fatigue testing and FEA consistently showed that Z4 Group provided the greatest mechanical stability. Specimens completed 5 million cycles with minimal displacement (0.14 mm) and sustained a static load of 741.7 N, exceeding average masticatory forces in the male molar region (654–665 N) [16]. FEA further demonstrated that Z4 Group exhibited significantly lower von Mises stress than the Z3 and Z2 Groups, consistent with previous reports that additional implants reduce stress at the bone–implant interface, prosthesis, and components, thereby improving long-term stability [17, 18]. Failure analysis of the Z4 Group revealed damage confined to the connecting screw, while implant bodies and abutments remained intact (Fig. 5 A-B). In screw-threaded systems, cyclic loading typically produces micro-motion at thread surfaces, leading to preload loss and loosening [19]. The “protective failure mode” observed in the Z4 Group, which prioritizes failure of replaceable components such as screws, preserving critical osseointegrated components and reducing the risk of catastrophic structural failure. This multi-level protective mechanism ensures controlled localized failure, interrupting overload transmission and supporting a “fail-safe” clinical design. By contrast, although the Z3 Group also passed the fatigue test, it exhibited clear signs of functional compromise. Displacement reached 2.68 mm, indicating stress accumulation along a unidirectional load path that can precipitate abutment loosening or screw fatigue failure. Fatigue testing also revealed significant wear on the Z3 Group’s abutments, with metal debris accumulation (Fig. 5 C-D), undermining the implant–abutment fit and increasing the risk of loosening or fracture [20]. FEA further confirmed that the stress peak in the Z3 Group was predominantly concentrated at the neck of the abutment, indicating a risk of progressive failure in that region. Previous studies have pointed out that improper biomechanical design or occlusal overload is the primary factor leading to screw loosening and eventual fracture [21]. Clinically, while Z3 Group may meet functional demands, it requires close maintenance to prevent long-term complications [22]. Meanwhile, the Z2 Group failed early at 33,591 cycles and showed the highest stress concentration. FEA showed that von Mises stress in the zygomatic implant reached 909.6 MPa—exceeding the yield strength of Grade 4 titanium (775 MPa)—indicating that two-implant support was mechanically insufficient under peak physiological loading conditions [11]. The convergence of experimental and simulation findings strengthens the conclusion that a two-implant configuration is not advisable for bilateral maxillary reconstruction, particularly in anatomically compromised patients. This study demonstrated that different zygomatic implant configurations exhibit biomechanical behaviors in bilateral maxillary defect reconstruction. Zygomatic implants offer distinct advantages in terms of fixation, structural support, functional restoration, and aesthetic outcomes [23]. However, mechanical complications—including screw loosening, abutment fracture, and peri-implant bone loss—remain common in clinical practice and threaten long-term outcomes [24]. The four-implant configuration demonstrated not only enhanced load distribution but also a favorable failure mode, which helps preserve critical components and potentially reduces the need for secondary interventions. Given these biomechanical complexities and the risk of structural failure, future strategies should integrate intelligent modeling tools to support personalized treatment planning. Recent advancements in computational modeling have introduced machine learning as a powerful tool for optimizing implant-supported prostheses [25–27]. Studies have demonstrated that deep learning-based structural optimization methods, such as BESO-Net, can improve prosthetic design by integrating convolutional neural networks (CNN) with bidirectional evolutionary structural optimization (BESO) [28]. These techniques enable rapid identification of optimal structural configurations, minimizing material use while maintaining mechanical integrity. The application of such intelligent optimization models to zygomatic implant restorations could facilitate personalized prosthetic designs tailored to individual anatomical and functional needs, enhancing stress distribution, reducing complications, and improving long-term success. Despite its contributions, this study is limited by its in vitro nature. The model does not replicate long-term biological remodeling, patient-specific occlusal dynamics, or soft tissue responses. Additionally, only vertical loading was assessed, whereas oblique and lateral forces are also present in functional mastication. Nevertheless, the strong alignment between experimental failure and FEA outputs validates the model’s predictive value. Conclusion In this study, we established a validated in vitro model simulating bilateral maxillary defects and demonstrated that a four-implant zygomatic configuration provides superior biomechanical stability under worst-case occlusal loading. All configurations exhibited a protective failure mode, in which damage occurred in replaceable components such as connecting screws rather than in the implant body itself. However, the four-implant group withstood the highest loads with minimal displacement, whereas reduced configurations showed earlier failure due to increased stress concentration. Despite the use of simplified implant assemblies without full-arch prostheses, the reduced configurations were unable to maintain structural resilience under load. These results provide direct experimental evidence that a four-implant configuration is strongly recommended to ensure mechanical reliability in bilateral maxillary reconstructions, while a three-implant approach may represent a minimally acceptable but higher-risk alternative.