Evaluation of Stress Distribution on Mandibular Implant-Supported Overdentures With Different Bone Heights and Attachment Types: A 3D Finite Element Analysis

Most edentulous patients who use complete dentures suffer from functional, aesthetic, and psychosocial problems. Implant-retained overdentures are utilized to overcome those issues.^1–5  However, the placement of implants may be challenging in severely resorbed edentulous arches and is a common problem with the long-term use of complete dentures.⁶ 

Implant treatment modalities may vary as the stress distribution associated with the mechanical behavior of the implant-bone complex changes.^6,7  Finite element analyses have been used in implant dentistry to predict the effects of clinical situations on the success rates of the implants.^3,8  Prior studies on the stress distribution of implant-supported overdentures have evaluated various contributing factors.³  Accordingly, the factors contributing to the distribution of stress can be considered, such as the thickness and resiliency of the mucosa, the attachments of the overdentures, alveolar bone height, and functional loads.^3,9  The distribution of forces is affected by the number of implants, the shapes and sizes of the individual components of the implant-prosthetic structure, and the quality and quantity of the surrounding bone.^4,10  It has been reported that the stress values in the cortical bone around the implant platforms were higher than those in other locations.¹¹ 

Occlusal forces are transmitted directly to the surrounding bone by the dental implants.⁵  It is difficult to place the implants in edentulous arches with significant bone loss. Therefore, several techniques have been described to manage the compromised mandibular bone anatomy. Different attachment systems can be possible options for implant-retained prostheses, such as for bar, magnet, locator, and ball attachments.¹² However, overdentures using two implants with locators and/or ball attachments have been a viable treatment option for edentulous arches.⁴  When two unsplinted implants are used, retentive attachments with the same height should be used to achieve a favorable stress distribution.¹²  However, if the amounts of resorbed bone on the right and left sides of the mandible are not the same, then the occlusal forces may distribute unevenly. The heights of retentive attachment can be either changed, or surgical procedures can be used—such as bone graftings or alveoloplasty—to compensate for the bone height differences.^12,13 

The aim of this study is to compare the stress distributions of implant-supported overdenture models with different attachments when bone heights are altered in the left and right sides of the mandible owing to resorption. Stresses were compared between models using three dimensional finite element analysis (3D-FEA).

A cone-beam computerized tomography (CBCT) image of a completely edentulous mandible was used for 3D modeling. Five 3D models of the completely edentulous mandible with various bone heights were designed and fabricated with software (3D Doctor, Able Software Corp, Lexington, Mass). The following five groups of models were studied (Figure 1).

1. Model 1 (1–1): Control model; attachment height of 1 mm on both the right and the left sides

2. Model 2 (1–2): Attachment height of 1 mm on the right side and 2 mm on the left side

3. Model 3 (1–4): Attachment height of 1 mm on the right side, and 4 mm on the left side

4. Model 4 (2–2): Attachment height of 2 mm on both the right and the left sides

5. Model 5 (4–4): Attachment height of 4 mm on both the right and the left sides

Two titanium implants (3.75 mm × 11.5 mm) and a locator, a ball attachment, and their other components along with a prosthesis were scanned with an optic scanner (Smart Optics Activity 880 Sensortechnik GmbH, Bochum, Germany). Both the locator (Zest Anchors, Escondido, Calif) and the ball attachments were composed of three parts: 1) an abutment, 2) a plastic male part, and 3) a metal housing. All three parts were accepted as metal alloys, while the prostheses and denture teeth were treated as acrylic resin.

The models and data from the scanner were reformatted and recorded in standard triangle language (.stl) format. The models were imported into 3D modeling software (Rhinoceros, Robert McNeel & Associates, Seattle, Wash) and trabecular bone, cortical bone, and mucosal properties were formed. The implants were virtually inserted into the canine regions, and the prosthetic components were placed in all 3D models.

All models were conjoined and transferred into the FEA software (Algor Fempro, Algor, Pittsburgh, Pa) to evaluate the stress distribution. The material properties of the bone, the mucosa, and the prosthetic components were determined from information obtained from the literature (Table 1).

In the control model, the cortical bone was created with a thickness of 2 mm and a height of 1 mm. In order to simulate bone resorption, the bone heights were decreased using Boolean operations, while the thicknesses of the cortical bone and mucosa were kept constant, as determined above, which ultimately resulted in the various attachment levels.

All 3D models and components were transferred in a software (Algor Fempro) for preprocessing to create the solid models. The models were conjoined with eight-node cubic elements. All materials used in this study were assumed to be isotropic, homogenous, and linearly elastic. Moreover, the osseointegration was considered to be 100% at the bone-implant contact. The numbers of elements and nodes are listed in Table 2.

The models were constrained at the nodes on the cross-sectional areas behind the retromolar pads in all degrees of freedom (Figure 2). Vertical and oblique occlusal loads of 100 N were used in this study, as suggested in the literature.^1,11,20  The loads applied on the overdentures were assumed static and vertical loads were directed on the central fossa of the right and the left first molar teeth, unilaterally and bilaterally. An oblique load of 100 N was applied at an angle of 30° with respect to the long axis of the implants from the center of the buccal cusp of the first molar of the overdenture unilaterally in buccolingual direction. Additionally, the overdentures and attachments were assumed to be in perfect contact, and the friction coefficient was ignored.

The negative principal stress (minimum principal stress) and positive principal stress (maximum principal stress) were evaluated in the cortical bone around the implant necks in all five models. The minimum principal stress values were larger than those of the maximum principal stresses in all models. The von Mises stresses (maximum equivalents stresses) were also determined at the abutment-implant structure for vertical and oblique loadings. Stress patterns were obtained at all loading conditions, and vertical loading conditions revealed lower stress values than oblique loadings for cortical bone and implant structures. Principal stress values were evaluated for brittle materials, such as bone, and von Mises stresses were analyzed for ductile materials like titanium.^21,22 

The maximum principal stresses were observed around the mesial side of the contralateral implant, and the minimum principal stresses were noted around the distal side of the ipsilateral implant during unilateral loading. The von Mises stresses were observed on the abutment neck area at the abutment-implant structure at the side of the loading.

The values of tensile stresses (maximum principal stresses) of all the models for vertical loading are illustrated in Figure 3. The maximum principal stress values decreased with bilateral loading compared to unilateral loading. The ball attachment models yielded lower stress values than those of the locator models in the presence of bilateral loading. However, all the models (except Model 3) resulted in similar stress values independent of the attachment types in the presence of unilateral loading. Unilateral loading caused increased stress values at the left side than the right side of Model 3 (Figures 4 and 5).

The values of compressive stresses (minimum principal stress) of all the analyzed models in the presence of vertical loading are presented in Figure 3. The compressive stress values were higher in the case of bilateral loading compared to those created with unilateral loading in the ball attachment models except for Models 1 and 5. However, the compressive stress values of all the models with locators were found to be higher than in the case of bilateral compared to unilateral loading. Unilateral loading caused higher stress values at the left side than the right side of the mandible in Models 2 and 3, which had different bone heights at the left and right sides of the mandible (Figures 6 and 7).

The ball attachment models yielded lower tensile and compressive stress distributions in the case of bilateral than in the case of unilateral loading. The models that had the same bone levels yielded higher compressive and tensile stress values than the models that had different bone levels on the left and the right sides. Model 5 yielded less tensile and compressive stress values than Models 1 and 4, which had the same bone level at both sides of the mandible at all loading conditions. The principal stress values in the case of oblique loadings are illustrated in Figure 8.

Maximum and minimum principal stress patterns on cortical bone are presented in Figures 9 and 10, respectively. The differences of maximum principal stress values between the models with the same bone height (Models 1, 4, and 5), and the models with unilateral bone resorption (Models 2 and 3) were higher in the case of oblique than in the case of vertical loading. However, the minimum principal stresses of all the models in the case of oblique loading were similar. The ball attachment models yielded smaller principal stress values than the locator models for both vertical and oblique loading conditions.

The von Mises stress values on the abutments for oblique loadings are shown in Figure 11. Ball attachments yielded higher von Mises stress concentrations than locators. However, the von Mises stress values were similar among the ball attachment models and also similar among the models with locators both for vertical and oblique loadings.

In this study, FEA was used to evaluate the biomechanical behaviors of two implant-retained mandibular overdentures with different bone heights and attachment types. Various locator heights and ball attachments were used in the analyses of the five mandibular models. Different amounts of bone resorptions were assumed on the left and the right sides of the mandible except the control model. The maximum and the minimum principal stresses on cortical bone around the implant necks and von Mises stresses on the implant abutment structures were investigated following vertical and oblique loadings.

Irrespective of the attachment type, bilateral loading exhibited more favorable tensile stress distribution with lower values than those induced by unilateral vertical and oblique loadings. However, oblique loading generated higher stresses than vertical loading in all models. The compressive stresses were more intense than the tensile stresses which may be attributed to the frictionless, tight connection between the attachments and the prostheses, and to the fact that the prostheses were mostly supported by the mucosa. The highest stress concentrations were found at the distal side of the cortical bone around the implant neck owing to the vertical force that was applied on the first molar tooth. However, the highest tensile stresses were found on the mesial sides. Nevertheless, the stress patterns were reversed in the case of oblique loading.

The compressive stresses for bilateral loading were higher than those for unilateral vertical loading for the ball attachment models, except for Models 1 and 5. The stress distribution and favorable force absorption may be explained by the sufficient bone support and the thick mucosa in these models.

Models with unilateral bone resorption (Models 2 and 3) yielded higher stress levels on cortical bone in the presence of oblique loading than models with uniform bone height levels (Models 1, 4, and 5). This may be explained by the moment force and lever arm effect, and may be more evident in the presence of oblique forces than in the presence of vertical forces on the attachments at increased heights. This may be also related to the smaller amount of bone on left side of the mandible. Barao et al²  reported that oblique forces generated more stress levels and simulated the clinical situation. Stress differences among the models on the bone and on the attachments following oblique loads were higher than those developed following vertical loads. This was also similar with the study of Chun et al²³  which compared the attachment overdenture types.

Assunçao et al¹  used implants (3.75 mm × 11.5 mm) with O-ring attachments in edentulous mandibular models with different mucosal thicknesses and resiliencies in their study. They found that a softer mucosa increased the stress values on the implants and the need for a larger denture-bearing area. Their findings were similar to the findings of our study regarding the stress areas that were mostly observed around the neck of the implants. However, the longer attachments with the thicker mucosa caused lower stress values at the cortical bone around the implant neck in this study, which is in contrast with the findings of Assunçao et al.¹ 

Ebadian and coworkers¹⁵  assessed the stress distribution of the overdentures which were supported by two implants and bar attachments. They noted that when a unilateral load was applied, maximum stress was found on the crestal bone around the implants on the ipsilateral side. These findings are in agreement with our findings. When the attachment height increased, it was observed that the maximum stress values around the implants were increased in the unilateral loaded models but decreased in the cases of bilateral loadings. Bilateral loading caused the stresses to distribute more evenly than unilateral loading, which was in line with the results of our study. Therefore, the bilateral balanced occlusion may be more favorable for bone in implant-retained overdentures.

Ozan and Ramoglu¹²  studied the stress distribution of mandibular implant-supported overdentures with locator and ball attachments and concluded that the higher attachment levels yielded less stress distributions, which is in agreement with the findings of our study. By contrast, the same investigators found that the stresses on the locator attachments yielded lower values than those of the ball attachments. This difference may have resulted from the use of plastic male parts in the ball attachments in our study, which were not used in the study by Ozan and Ramoglu.¹² 

Khurana et al²⁴  also investigated the stress distributions on the implant retained overdentures with different attachment heights. It was concluded that the locator system revealed lower stress in the bone and prosthetic components than those of the ball attachments. The von Mises stresses on the attachments concentrated around the neck of the ball attachment area in the current study. Additionally, ball attachments yielded increased stress levels compared to those seen at the locator. This may be attributed to the lower height and wider design of the locator compared to the ball attachment, in agreement with Khurana et al.²⁴ 

Conversely, principal stresses on cortical bone were lower in models with ball attachments. It was assumed that stresses were absorbed by ball attachment neck that caused less stress to be transmitted to the bone. Both principal stresses on bone and the von Mises stresses on attachments could not reach the yield stress value limits of bone and titanium.^21,22  Therefore, the results of this study may not be detrimental to the bone and prosthetic components.

In the present study, the thickness of the mucosa influenced the stress distribution. The attachments at increased heights acted as lever arms and led to increased stress concentrations around the cortical bone and around the implant necks compared to those of shorter attachments, independently from the attachment type.^12,25  The results did not support this fact, which may be attributed to the shock absorption feature of the thicker mucosa compared to the thin mucosa. It was also observed that the bone height level was not a critical factor on the stress distribution.

Regardless of the clinical situation, the results of this study showed that thicker and softer mucosa may provide better stress absorption responses. For this reason, careful selection of the attachment system and the type of mucosa may be important for better stress distributions. FEA allows the determination of the stress distribution on biological and prosthetic structures, even if there is a small change in the shape of the components. Favorable predictions about the clinical situations could be made about the selection of implants, attachments, and prostheses.⁴ 

It should be kept in mind that osseointegration was assumed to be 100% of the bone-implant contact in the present study. This is not indicative of the clinical reality. In addition, the frictional coefficient was ignored and the forces that acted on the overdenture were assumed static. Nevertheless, this does not realistically reflect the situation during mastication. However, this study may provide an insight on the overdenture design when different bone heights are present in the quadrants of the mandible.

1. The tensile stresses that can be detrimental to crestal bone were lower than the compressive stresses in all the analyzed models.

2. The stress values were distributed more uniformly in the case of bilateral loading. Therefore, bilateral balanced occlusion can be preferable in the cases of the two implant-retained overdentures

3. The ball attachment models yielded lower minimum principal stress values than the locator models. The findings were opposite for the von Mises stresses for the attachments. Even so, the ball attachments may be preferable for different bone heights in the mandible for the preservation of the crestal bone.

4. Bone adjustments may not be necessary for the stress distributions at different attachment heights.

Abbreviations

3D

three-dimensional

CBCT

cone-beam computerized tomography

FEA

finite element analysis

The study was presented as an oral presentation at CED-IADR/NOF Oral Health Research Congress, Vienna, Austria, September 21–23, 2017.

The authors participated in this study have no affiliations with or involvement in any organization or entity with any financial interest or nonfinancial interest in the subject matter or materials discussed in this article.