Finite Element Analysis of Zygomatic Implants in Intrasinus and Extramaxillary Approaches for Prosthetic Rehabilitation in Severely Atrophic Maxillae

P with severely resorbed edentulous maxilla can be treated using conventional removable dentures, implant-supported !xed prostheses, or overdentures. An implant-supported !xed prosthesis or overdenture requires a minimum of four implants, two in the posterior region and two in the anterior region. The placement of the two implants in the posterior region can be achieved using sinus bone grafting or zygomatic implants.1–2 Bone grafting is normally chosen as the standard procedure for the treatment of severely atrophic maxillae before the placement of conventional implants. However, this procedure is resource demanding and requires a relatively long treatment time and a longer healing period for the patients.3–4 In addition, the harvest of bone grafts could cause morbidity at the donor site.1 Based on the literature, the implant survival rate is lower for grafted maxillae compared with non-grafted maxillae, especially in the posterior region.1,5 In the anterior region, the success rate of the implant depends mostly on the bone volume before the treatment.5 The Brånemark System (Nobel Biocare) has introduced an alternative system utilising zygomatic implants to overcome these problems.1,3–4 The original purpose of zygomatic implants was to rehabilitate patients who 1 Postgraduate Student, Medical Implant Technology Group, Faculty of Biomedical Engineering and Health Sciences, Universiti Teknologi Malaysia, Johor Bahru, Malaysia. 2 Associate Professor, Medical Implant Technology Group, Faculty of Biomedical Engineering and Health Sciences, Universiti Teknologi Malaysia, Johor Bahru, Malaysia. 3 Senior Lecturer, Department of Conservative Dentistry, Faculty of Dentistry, University of Malaya, Kuala Lumpur, Malaysia. 4 Associate Professor, Department of Conservative Dentistry, Faculty of Dentistry, University of Malaya, Kuala Lumpur, Malaysia.

7][8] However, the function of this implant had been expanded for the rehabilitation of patients with severely resorbed edentulous maxillae. 3 zygomatic implant for the treatment of edentulous maxillae comes in various sizes, lengths, diameters, and thread distributions.[10] The selection of approach depends on the degree of bone resorption in the maxilla. 90][11] Several surgical approaches have been available in practice since the zygomatic implant was introduced, such as the intrasinus approach, sinus slot approach, extrasinus approach and extramaxillary approach.] This approach was originally de ned by the Brånemark System in 1988, which involved the insertion of a long implant (between 35 mm to 55 mm) anchored to the zygomatic bone, following an intrasinusal trajectory.The penetration of the implant body through the maxillary sinus cavity needs to be considered because the soft tissue may be highly a ected.Stella and Warner 12 described a variant of the intrasinus technique in which the implant is positioned through the sinus via a narrow slot, following the contour of the malar bone and introducing the implant into the zygomatic process.In this way, the need for fenestration of the maxillary sinus is avoided, and the implant is on course to emerge over the alveolar crest at rst molar level with a more vertical angulation. 2,12The extrasinus approach, on the other hand, is mainly used to treat patients who have pronounced buccal concavity. 23] The extramaxillary technique is the latest surgical approach introduced for the treatment of edentulous atrophic maxillae using zygomatic implants. 9This technique is signi cantly di erent compared with other approaches because the implant body is only anchored to the zygomatic bone.The coronal part of the implant body is placed externally to the maxilla and then covered with soft tissue using a zygomatic implant with a di erent thread distribution.The emergence of the implant head will be more prosthetically correct as it is located on or close to the alveolar ridge.The major di erence between all approaches is the di erence in implant path insertion.
2] Feedback from patients regarding discomfort was identi ed as the main problem because the bulky prosthesis may a ect dental hygiene and increase mechanical resistance. 9omplications of peri-implant soft tissue bleeding and increases in probing depth may occur due to the inappropriate position of the zygomatic implant head and abutment. 1For the extramaxillary approach, implant mobility and fracture of abutment screw are among the complications that have been reported...It can be concluded that most complications are mainly caused by insu cient primary stability of the zygomatic implant in supporting the prosthesis.
A key factor for dental implant success or failure depends on stress transmission to the surrounding bone. 13Inappropriate loading may result in the concentration of stress in the bone around the implant, which could lead to bone resorption.It is known that the vertical load component plays a major role in masticatory loading. 14Conversely, the role of the horizontal load component cannot be compromised although its value is minimal, especially when an angled implant is used.
No consensus exists on the ideal approach for the placement of zygomatic implants in regard to the degree of bone anchorage and implant inclination. 2ittle is known about the quantity of bone that accumulates around zygomatic implants through di erent techniques on the e ects of mechanical implant stability.To the best of our knowledge, no comparative studies have examined these two surgical approaches via the nite element method.In this paper, the stress distribution and micromotion for two of the approaches -the intrasinus and extramaxillary -was investigated.

MATERIALS AND METHODS
A series of computed tomography (CT) image datasets of a real complete denture wearer with a high degree of maxillary bone resorption was utilized to develop a three-dimensional (3D) model of the bones, prosthesis, and soft tissue using Mimics/Magics 10.01 (Materialise).The selected regions of interest were the left maxilla and zygomatic bone (Figs 1a and 1b), where the craniofacial model was assumed to be symmetric for the analysis. 15The maxilla and zygomatic bones consists of two layers, the cortical and cancellous with thicknesses ranging from 1.4 mm to 2.2 mm for cortical layers. 16The height and width of the atrophic maxilla were measured to determine a suitable approach for treatment, either through the use of zygomatic implants alone or in conjunction with conventional implants. 9Based on our measurements, the average height of the anterior and left posterior maxilla sections was 8.1 mm and 5.8 mm, respectively. 17The width of the alveolar ridge in the molar region was 9.7 mm. 18Hence, according to the Cawood & Howell edentulous jaw classi cation, 17 the types of this patient's edentulism can be classi ed as Class III for the anterior maxilla and Class V for the posterior maxilla.Therefore, the patient could be treated with a zygomatic implant placed bilaterally in conjunction with two conventional implants in the anterior region. 9To determine the length of zygomatic implant to be used, the distance from the jugale point of zygomatic bone to the alveolar crest was measured.The angulation of zygomatic implant was determined between the implant body length and the plane through the infraorbital foramen. 19Results from this measurement were 48.9 mm for the length and 45.7 degrees for the angle.
A prosthesis superstructure with ange was modeled based on the original patient's complete denture design from the CT dataset.The design and geometry of the model were assumed to be symmetric at 1.5 mm to 3.4 mm in thickness, 12.5 mm to 19.1 mm in width and 15.4 mm to 18.3 mm in height.A single xedconnection type was employed using two screws to secure the prosthesis to the implants. 20The prosthesis and its framework were modeled as one piece and assumed to be made of gold alloy for the analysis. 21he construction of a soft tissue model was also carried out based on the patient's CT dataset.The thickness of the soft tissue was set to a range of 0.4 mm to 5.5 mm.This is crucial as it a ects the positioning of the prosthesis and must re ect the measurements of the actual model as closely as possible.
A 3D computer-aided design (CAD) software, SolidWorks 2009 (Dassault Systèmes SolidWorks) was utilized to develop the implant models. 21The construction of the implant model required a matched abutment model to connect the implant body to the prosthesis.One zygomatic implant, 46.5 mm in length, and a straight multi-unit abutment from the Brånemark System (Nobel Biocare) were used in both surgical approaches. 20The design of the zygomatic implant in the extramaxillary approach is slightly di erent compared with the one in the intrasinus approach because the diameter of the implant body is larger and only the apical part is threaded.For the conventional implant, a 4.0 × 10.0 mm implant with an angled multi-unit abutment at 30 degreees was chosen from the same manufacturer 22 as shown in Figs 2a and 2b.The 3D solid implant designs from the CAD software were then transferred to other software, Abaqus/CAE 6.9-1 (Dassault Systèmes Simulia) to generate surface triangular elements before the virtual surgery simulation.All models were meshed with 0.5-mm triangular element mesh.
The conventional implant was virtually placed in the anterior region of the maxilla adjacent to the lateral incisor, whereas the zygomatic implant model was placed in the posterior region according to the earlier described techniques: adjacent to the rst molar and second premolar for the intrasinus and extramaxillary approaches, respectively.The extramaxillary approach was found to have increased the distal cantilever length almost two times longer than the intrasinus.The reverse was seen for the buccal cantilever length where the intrasinus prosthesis was 1.3 times longer than the one in the extramaxillary (Figs 3a and 3b).All models were converted into four nodes of the tetrahedral element type in nite element analysis software, MARC 2007 (MSC Software).The total number of elements for the intrasinus model was 390,899 while the extramaxillary model had a total of 394,091 tetrahedral elements.The friction coe cient, µ, for all contact-ing surfaces were set at 0.3 to simulate an immediate loading condition.The threaded part of the implant body for both approaches was simulated via contact properties accordingly and was assigned a friction coe cient of 0.5 to represent the strong attachment to the bones.All materials were assumed to be isotropic, homogenous, static and linearly elastic.The material properties of all models are shown in Table 1.There were two types of loading applied to thenite element models -static occlusal and masseter loadings.Simulated occlusal loading of 230 N 23 and 50 N 20 were applied separately as vertical and lateral loadings, respectively (di erent loading cases), on the top surface of the prosthesis in the rst molar region.For the masseter loading, a 300-N load with force components of -62.1 N in the x-axis, 265.2 N in the z-axis, and -125.7 N in the y-axis [15][16]20 was applied at the muscle attachment area on the zygomatic arch to represent the action of the masseter muscles.
For the boundary conditions, the mid-sagittal, posterior, and top cutting planes were constrained in the x, y, and z directions to prevent any movements. 24All loading and boundary conditions as shown in Fig 4.
In this study, three indices were used to verify the result of this nite element study.The rst index was the total contact area between the zygomatic implant body with the surrounding bone tissues.The second index was a comparison of equivalent von Mises stress (EQV) distribution and its magnitude to assess the behaviour of bones and implants under simulated loadings.The results were presented in colored contour plots with blue representing low-magnitude stress and grey representing high-magnitude stress.The third index was the comparison of the displacement value of the zygomatic implant body for both approaches.All mentioned indices provided signi cant information regarding the in uence of di erent surgical approaches on mechanical criteria where high contact area and low-magnitude stress and displacement are favorable for an encouraging interpretation of the results.

RESULTS
Figure 5 illustrates the value of total contact area between zygomatic implant body and bone tissues, both cortical and cancellous.It was clearly shown that the extramaxillary approach had higher bone-to-implant contact area compared with the intrasinus approach.The total contact area was increased by 40.3% through the extramaxillary approach, nearly a two-fold increase in the area of contacting surfaces.For the intrasinus approach, about 23.9% of the implant body surface that had contact with the cortical and cancel- lous bone compared with 39.9% for the extramaxillary approach.
Tables 2 and 3 summarize the maximum and average values of EQV measured at the bone-implant interface, zygomatic bone, and zygomatic implant body under vertical and lateral loading, respectively.Figures 6a to 6d show that the stresses within the bone model concentrated around the region of the implant and at the simulated loading location for both approaches.Under vertical loading, a high stress value of 35.9 MPa was generated at the edge of maxilla and de ected to the coronal part of the implant body and prosthesis for the extramaxillary approach.Similar observations were seen in the intrasinus approach; however, the magnitude of EQV was higher, 50.6 MPa.The stress was uniformly distributed in the zygomatic bone in two main directions, towards the temporal and frontal processes of the zygoma for both approaches.Maximum stresses of 270.3 MPa and 286.6 MPa were generated for the intrasinus and extramaxillary approaches, respectively, within the bone under both types of loading.Stresses were more gradually redistributed at the maxillary sinus wall for the extramaxillary approach as compared to the intrasinus.When loaded laterally, the EQV value increased at the bone-implant interface in the extramaxillary approach to a maximum of 106.5 MPa, whilst the stress value at a similar interface in the intrasinus approach was merely 42.9 MPa.
Figures 6a to 6d also shows the distribution of EQV on the zygomatic implant body.Under vertical load-ing, stress concentrated at the coronal part of the implant body on the distal side with a maximum value of 86.8 MPa for the extramaxillary approach.However, a larger stress concentration region was observed at the abutment-implant connection towards the middle part of the implant body on the buccolingual side with a peak stress of 370.9 MPa for the intrasinus approach.The apical portion of the implant seemed to have a smaller EQV distribution than the coronal portion.The EQV distribution under lateral loading was less widespread within the implant body for the intrasinus approach, with the maximum value concentrated at the abutment-implant joint (121.3MPa).Stresses higher than 100 MPa were found to be more widespread for the extramaxillary approach.
In terms of implant displacement, Figs 7a to 7d show that the zygomatic implant body in the intrasinus approach had a similar displacement value of 0.010 mm as the extramaxillary approach under vertical loading.Lateral loading signi cantly increased the displacement of the implant body in both surgical approaches.A higher displacement magnitude was found in the extramaxillary approach (0.024 mm) compared with the intrasinus approach (0.012 mm), which increased by 58% and 17%, respectively, as a result of lateral loading.Figures 7a to 7d also show that the coronal part of the implant body deformed more than the apical part for both intrasinus and extramaxillary approaches.In comparison, the coronal part of the implant body in the intrasinus approach showed the most signi cant bending effect in the buccal direction under vertical loading, and the reverse is seen under lateral loading.All images are t to scale.

DISCUSSION
As mentioned above, several types of surgical approaches exist to treat edentulous atrophic maxillathe intrasinus, sinus slot, extrasinus, and extramaxillary approaches.[26] Comparative study between the most popular, the intrasinus approach and the latest, the extramaxillary approach, can highlight their strengths and weaknesses and provides crucial information to improve clinical outcomes as very few studies have addressed these issues. 20,27Through nite element analysis, a simulation can be made to determine the type of surgical approach that may provide better stress distribution and implant stability in the maxilla and zygoma.
In this study, the concept of immediate loading was simulated through nite element analysis since it has become popular among dental surgeons for the treatment of fully and partially edentulous patients.9][30][31] The types of surgical techniques, implant designs, implant surface roughness, bone quality, and bone quantity are some of the factors that contribute to the primary stability of implants in immediate loading cases. 28Primary or initial stability is de ned as the strength of anchorage or engagement of the implant body to the bone site without any critical movement that can cause implant failure after implantation. 28,323 According to Aparicio et al, 1 the cumulative failure rates of zygomatic implants and conventional implants were 1.6% and 5.2%, respectively, with follow-up time periods of six months to 12 years.4][35][36] Ahlgren at al 37 reported a 100% success rate for zygomatic implants achieved for a follow-up period of 11 to 49 months.Another study by Aparicio et al 32 found similar results when 69 patients treated with 131 zygomatic implants within a period of 6 months to 5 years followup.There were also cases of 2 zygomatic implants placed bilaterally without additional retention from anteriorly placed conventional implants.A study conducted by Duarte et al 31 showed that out of 48 zygomatic implants, 2 implants had failed after 30 months.Among the complications identi ed through the use of intrasinus approach were bleeding of peri-implant soft tissue, increased probing depth and sinusitis. 1,26,38hese problems could be caused by inappropriate positioning of the zygomatic implant body and abutment resulting from the chosen surgical technique.The design of prosthesis also plays an important role for successful clinical outcomes.
Malo et al 9 reported 98.5% and 100% cumulative survival rates for implants (conventional and zygomatic) and prosthetics, respectively, in their 1-year follow-up study.They investigated the application of the extramaxillary approach for the treatment of atrophic maxillae using a new zygomatic implant design with immediate loading.Their results showed that only one zygomatic implant failure was caused by implant mobility due to a disconnection between the implant and prosthesis.
Our results showed that the intrasinus approach increased the stress magnitude at the bone-implant interface and zygomatic implant body under vertical loading approximately 1.41-and 4.27-fold higher, respectively, compared to the extramaxillary approach.Although the applied loading is close to the implant head, it was more towards the buccal aspect (the implant head was positioned palatal to the ridge), thereby creating buccal cantilever force.The implant body had to sustain higher loads to counter the bending moment from vertical loading, and this resulted in a high concentration of stress at the bone-implant interface and the coronal part of the implant body.For the extramaxillary approach, even though the location of implant head was on the alveolar ridge, it was further away from the loading point, which produces the distal cantilever e ect.The prosthesis transferred a large amount of stress to the maxilla and zygomatic bone, which resulted in a wider stress distribution within the bone at the implant site (premolar region).However, the maximum stress value is lower than that of intrasinus approach.The stress peaked at the abutment-implant body connection for both approaches due to the connection between the two parts.
These ndings could also be attributed to the total contact area between implant and bone tissues.According to Javed et al, 28 the threaded implant design could increase the primary stability by reducing the micromotion of the implant.The implant used in the intrasinus approach is likely to have high contact area since the implant surface area increases due to the thread along the implant body.However, the implantbone contact area only occurred at the alveolar ridge, slightly in the palatal aspect and at the jugale point of the zygoma, which resulted in a smaller mating surface compared to the extramaxillary approach.The percentage of bone-implant contact area for the extramaxillary approach was higher (40.3%) due to the placement of the implant body external to the maxilla as well as at the maxillary sinus wall.It is noteworthy that the coronal part of the implant body had no threads to avoid in-fections of the soft tissue. 9However, the insertion path of the zygomatic implant increased the contact area of the implant body to the bone and therefore reduced the stress at the bone-implant interface.
Lateral loading contributed more to the increase in the magnitude of stress at the maxilla around the neck of implant and a wider stress dispersion area on the implant body in the extramaxillary approach due to large rotational e ects caused by the torque generated.Since the distance from the implant head to the loading point is longer than in the intrasinus approach, the magnitude of torque produced increases proportionally, resulting in a high stress level on the surrounding bone (about 2.48-fold higher).The high stress value in the maxilla is a concern because it could lead to marginal bone loss in the long-term. 39nder both the vertical and lateral loadings, the maximum stress was generated on the zygomatic bone for both approaches with a higher value found in the extramaxillary approach.Most of the stress produced from simulated occlusal loading was borne by the zygoma and appears to be independent of the maxillary anchorage.These results were in agreement with the ndings of Ujigawa et al, 20 who reported that the applied loading will be transferred through the infrazygomatic crest and directed into the temporal and frontal processes of zygomatic bone.
In all models tested, the highest stress value was recorded within the implant bodies as compared to the bone model.It seems possible that the results are due to the high modulus of elasticity of titanium alloy (Ti6Al4V), 110,000 MPa compared to 13,400 MPa for cortical bone.The maximum stress values generated within the implant bodies in all loading conditions for both approaches have no tendency to cause implant failures since titanium alloys are known to be able to tolerate stress up to 900 MPa. 39However, the reverse was observed for the bone, as the peak stress magnitude exceeded the yield strength of cortical bone, 69 MPa. 40The percentage of nodal stress higher than 69 MPa, however, was only about 0.1% for both extramaxillary and intrasinus approaches.
All displacement magnitudes of the zygomatic implant body found in the intrasinus and extramaxillary approaches under both loading conditions are lower than the threshold motion limit reported in the literature.The maximum value of 24 µm generated by the extramaxillary approach under lateral loading was relatively higher than the intrasinus approach (12 µm); however, the implant bodies have a low tendency for failure as the value of micromotion between 50 µm and 150 µm may negatively in uence osseointegration and bone remodelling at the bone-implant interface. 28Moreover, the zygomatic implant has a higher tendency to displace and bend under lateral loading due to its increased length to width ratio. 20The present study showed that more deformations occurred in the buccal direction for both approaches.The ndings also demonstrated that the implant body deformed from the coronal to the middle of the implant body without signi cantly a ecting the apical part.This observation was parallel with the implant stress and displacement contour plots.It shows that adequate strength for anchorage in the zygomatic bone is achievable for both approaches without critical deformation in the apical part.This is in agreement with the work of Stievenart et al 4 who reported that the success rate of treatment using the zygomatic implant depended mostly on the strength of the zygomatic cortical bone.Anchorage within the trabecular bone was less important since the strength of the anchorage in the cortical layer was able to retain the prosthesis successfully. 4The implant body also penetrates through a small volume of cancellous bone depending on the zygoma's anatomy.Another important reason for the placement of implant within the zygomatic bone as it has very little tendency towards resorption or regeneration. 41imitations of the present study were the following: (1) the simpli cation of material properties used in the analysis, which were assumed to be homogenous, isotropic and linearly elastic; and (2) only unilateral modelling was conducted, whereas bilateral simulation would be more clinically relevant.

CONCLUSIONS
In conclusion, the intrasinus approach generated 1.41-and 4.27-fold higher stress at the bone-implant interface and the zygomatic implant body, respectively, under vertical loading than the extramaxillary approach.However, the reverse was seen under lateral loading where the extramaxillary approach showed an increased stress level at the bone-implant interface by 2.48-fold.The zygomatic implant body in the extramaxillary approach also exhibited micromotion with a magnitude two-fold higher than those with the intrasinus approach under lateral loading.Both techniques may be used for the treatment of severely atrophic maxillae; however, the intrasinus approach is more favorable if lateral loading is a major concern.

Figs
Figs 1a and 1b (a) The reconstruction of a three-dimensional craniofacial model from CT images as viewed from sagittal plane.(b) Region of interest shown in yellow.a b

Figs
Figs 2a and 2b Three-dimensional model of (a) conventional implant and (b) zygomatic implant used in FEA. a b

Figure 4 :Figure 5 :
Figure 4: Boundary conditions and loading con gurations as shown in frontal and sagittal views.
Figs 6a to 6d Equivalent von Mises stress distribution within bone (front view) and implant model of intrasinus approach under (a) vertical loading and (b) lateral loading, and for the extramaxillary approach under (c) vertical loading and (d) lateral loading.All images are t to scale.

Table 1
Material Properties Used in the Finite Element Analysis

Table 2
Magnitude of EQV (MPa) rRecorded at Different Positions Under Vertical Loading for both Approaches