Accepted Manuscript

Title: Combined Implant-Residual Tooth Supported Prosthesis

after Tooth Hemisection: A Finite Element Analysis

Author: Yun He Istabrak Hasan Ludger Keilig Junliang Chen

Qing Pan Yue Huang Christoph Bourauel

PII: S0940-9602(16)30002-4
DOI: http://dx.doi.org/doi:10.1016/j.aanat.2016.01.002
Reference: AANAT 51003

To appear in:

Received date: 18-6-2015

Revised date: 25-11-2015
Accepted date: 17-1-2016

Please cite this article as: He, Y., Hasan, I., Keilig, L., Chen, J., Pan, Q.,
Huang, Y., Bourauel, C.,Combined Implant-Residual Tooth Supported Prosthesis
after Tooth Hemisection: A Finite Element Analysis, Annals of Anatomy (2016),
http://dx.doi.org/10.1016/j.aanat.2016.01.002

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 1 Combined Implant-Residual Tooth Supported Prosthesis after Tooth

2 Hemisection:

3 A Finite Element Analysis

4 Yun He1, 2, Istabrak Hasan2, 3, Ludger Keilig2, 3, Junliang Chen1, Qing Pan1, Yue

5 Huang4, Christoph Bourauel2

t
1
6 Department of Oral and Maxillofacial Surgery, Hospital of Stomatology, Luzhou

ip
7 Medical College, Jianyangnanlu 2, 646000 Luzhou, China

cr
2
8 Endowed Chair of Oral Technology, Rheinische Friedrich-Wilhelms University,

us
9 Welschnonnenstr. 17, 53111 Bonn, Germany
3
10 Department of Prosthetic Dentistry, Preclinical Education and Materials Science,

11 an
Dental School, Rheinische Friedrich-Wilhelms University, Welschnonnenstr. 17,

12 53111 Bonn, Germany

M
4
13 Department of Orthodontics, Hospital of Stomatology, Luzhou Medical College,

14 Jianyangnanlu 2, 646000 Luzhou, China

ed

15 *Corresponding author

16 Dr.med. dent. Dr. rer.nat Istabrak Hasan. Endowed Chair of Oral Technology,
pt

17 Department of Prosthetic Dentistry, Preclinical Education and Materials Science,

ce

18 Dental School, Rheinische Friedrich-Wilhelms University, Welschnonnenstr. 17,

19 53111 Bonn, Germany.

Ac

20 Tel: +49 228 287 22388

21 E-mail address: ihasan@uni-bonn.de

22

23

24

25

26

Page 1 of 38
 28
27

Ac
ce
pt
ed
M
an
us
cr
ip
t

Page 2 of 38
28 Abstract

29 Tooth hemisection preserves partial tooth structure and reduces the resorption of

30 alveolar bone. The aim of this study was to analyze the feasibility of preserving a

31 molar after hemisection and inserting a dental implant with different prosthetic

32 superstructures by means of finite element analysis. Firstly, the distance between the

t
33 root of the mandibular second premolar and the distal root of the first molar were

ip
34 measured in 80 cone beam computed tomography (CBCT) data sets. Based on

cr
35 these data, the lower right posterior jaw segment was reconstructed and the

us
36 geometries of the appropriate implant were imported. Four models were created: (1)

37 Hemi-1: An implant (3.7×9 mm) replaced the mesial root of the molar, and a single

38 an
crown was placed on the implant and residual tooth. (2) Hemi-2: Two separate

39 crowns were generated for the implant and the residual tooth. (3) Single: An implant
M
40 (5.5×9mm) with crown replaced the whole molar. (4) FPD: A 3-unit fixed partial

41 denture combined the distal residual part of the molar and premolar. The results
ed

42 indicated that stresses in the cortical bone and strains in the majority region of the

43 spongious bone were below the physiological upper limits. There were higher
pt

44 stresses in implant with the Hemi-1 and Single models, which had the same
ce

45 maximum values of 45.0 MPa. The FPD models represented the higher values of

46 stresses in the teeth and strains in PDL compared to other models. From a
Ac

47 biomechanical point of view, it can be concluded that a combination of an implant

48 and residual molar after tooth hemisection is an acceptable treatment option.

49 Key words: hemisection; periodontal ligament; implant; stress; displacement

50

51

Page 3 of 38
51 1. Introduction

52 Tooth hemisection refers to sectioning of a molar into two halves after root canal

53 therapy of the healthy root and removal of the diseased root and its coronal portion. It

54 may be a suitable treatment option when decay is restricted to one root or furcation

55 involvement. The preserved root with surrounding periodontal ligament (PDL) not

t
56 only retains its masticatory and sensory function as well as stress transmission

ip
57 pathways into the bone, but it also helps to avoid alveolar resorption (Naveen et al.,

cr
58 2014). Buhler (1994) stated that hemisection should be considered before every

us
59 molar extraction because it provides a good, absolute, and biological cost saving

60 alternative with good long term success. A standard restoration after hemisection is a

61 an
fixed partial denture (FPD), including the residual root with the next adjacent tooth. A

62 FPD has the disadvantage that healthy tooth tissue would be prepared with the risk
M
63 of pulp irritation or even exposure and loss of tooth sensitivity. Moreover, the

64 abutment tooth could be more susceptible to secondary caries and periapical lesions
ed

65 because of the difficulty in preserving optimal oral hygiene in this region.

66 On the other hand, some studies have indicated that hemisected mandibular
pt

67 molars were more prone to complications than implants placed after extraction of the
ce

68 molar (Zafiropoulos et al., 2009; Fugazzotto, 2001). Certainly, the use of dental

69 implants in the rehabilitation of partially and completely edentulous patients is

Ac

70 considered a predictable and successful treatment option with favorable long-term

71 survival rates (Lindth et al., 1998; Holm-Pedersen et al., 2007; Pjetursson et al.,

72 2004). For the implantation in the molar region, Desai et al. (2012) observed that von

73 Mises stress for two implants (Ø 3.75 mm) had 31 %–43 % stress reduction

74 compared to a single implant (Ø 6 mm), and thus suggested that when the

75 mesiodistal space for an implant is more than 12.5 mm, support with two implants

76 should be considered. However, using two implants for one molar undoubtedly

Page 4 of 38
77 increases the expense for the patient. Moreover, sometimes it is impossible to insert

78 a wider implant in the molar region without any auxiliary materials (membrane or

79 bone graft material) and special surgery due to bone atrophy after the loss of tooth.

80 The aim of this study was to analyze the combination of the advantages of

81 implants and tooth hemisection. The purposes of this study were:

t
82 1. Investigation of the feasibility of combining tooth hemisection and a dental

ip
83 implant to restore a molar tooth.

cr
84 2. Approaching a theoretical guidance of restorative methods for a molar with

us
85 lesions restricted to one root or furcation by numerically comparing the biomechanical

86 characteristics of four different restorations.

87 an
M
ed
pt
ce
Ac

Page 5 of 38
 87 2. Material and Methods

88 2.1 Radiological Analysis

89 2.1.1 Radiological Measurement

90 The CBCT data of 80 patients (40 male, 40 female, age: 18-60 years old) taken

91 from January to May 2014 for different diagnostic reasons were randomly chosen for

t
92 evaluation. All scans were performed at 86 kVp, 10 mAs, and 10.8 s exposure, with a

ip
93 resolution of 0.20 mm per slice by KODAk 9500 (Carestream Health, Rochester,

cr
94 USA) at the Hospital of Stomatology, Luzhou Medical College, China. The patients

us
95 were fully informed about the study. This study was approved by the ethics

96 committee of the Luzhou Medical College (No. 2011002).

97 an
The inclusion criteria were as follows: 18 to 60 years old, when a complete

98 detection of the lateral or bilateral mandibular premolars and molar teeth with
M
99 complete root formation can be achieved. The exclusion criteria were: Signs of

100 resorption, posts or coronal restorations or orthodontic treatment is taking place.

ed

101 Measurements were performed with CS 3D Imaging Software 3.1. (Carestream

102 Dental, Atlanta, GA, USA). The minimum distances between the root of the
pt

103 mandibular second premolar and the distal root of the first molar, and the length of
ce

104 the mesial root of the first molar were measured in axial sections and sagittal

105 sections, respectively. For measuring the distance, three reference points on the
Ac

106 distal root of the first molar were chosen: the furcation, the apex and the midpoint

107 between the previous two points (Figure 1).

108 Using these three reference points, the distance between the furcation and the

109 root of the premolar (DFR), between the midpoint and the root of the premolar (DMR)

110 and between the apex and the root of the premolar (DAR) were measured. In cases

111 where two distal roots existed, the distances were measured and recorded

112 respectively.

Page 6 of 38
113 2.1.2 Statistical analysis

114 Data were analyzed using the IBM SPSS statistical package 19.0 (IBM Co.,

115 Chicago, IL, USA), and variables were presented as mean ± standard deviation (SD).

116 One-way ANOVA was used to examine differences among the three test points in

117 measuring the distances. A significance level of 0.05 was chosen.

t
118

ip
119 2.2 Finite element analysis

cr
120 A proper example of CBCT data in DICOM format was selected and imported

us
121 into Mimics research 17.0 (Materialise NV, Leuven, Belgium) for segmentation. For

122 the selected case, the mesial root of the right mandibular first molar suffered from

123 an
vertical root fracture. No signs of bone resorption and periodontal lesions were

124 observed in the image data. The patient is advised to undergo tooth hemisection
M
125 after root canal treatment. The segment included the second premolar, the distal half

126 of the first molar and alveolar bone. Later on, four masks were created by defining
ed

127 different grey value range, including the mask of cortical bone, spongious bone, tooth

128 and pulp (Figure 2). The density of cortical and spongious bone were recorded using
pt

129 grey values of their masks. The segment was further processed in 3-Matic research
ce

130 9.0 (Materialise NV, Leuven, Belgium) to created and converted to a 3D FE model

131 using tetrahedral 4-noded elements. The PDL was separately constructed using the
Ac

132 same software, with a uniform thickness of 0.2 mm. According to the results of CBCT

133 data, the implant with appropriate diameter was selected and insert in the proper

134 position. The geometries of the implant and the abutment (tioLogic©, Dentaurum

135 GmbH &Co.KG, Ispringen, Germany) were constructed from the CAD/CAM data that

136 were generated and provided by the dental implant company. By using measurement

137 tools in Mimics Software, the total thickness of the jaw segment as well as labial and

138 lingual thickness of cortical bone were determined. In detail, the measurements were

Page 7 of 38
139 conducted on four reference positions: distal root of the molar (Hemi-root, HR),

140 implants in Hemi models (Hemi-implant, HI), implant in single model (SI), premolar

141 (PR) (Figure 3). In each reference position, the measurements were performed on

142 three axial slices: alveolar ridge, the apex and the midpoint between the previous two

143 slices. The mean values were calculated and presented as mean ± SD (Table 1). The

t
144 length and width of the distal root was measured as well.

ip
145 As the major goal of this part of the study was not to analyze the biomechanical

cr
146 properties between implant and abutment, the implant and the abutment were

us
147 modeled as one piece. The crowns of the molar and the premolar were modified for

148 use as the prosthesis. Their position and orientation were maintained according to

149 an
the original data in order to define the normal position of the final prosthesis.

150 In order to enable a meaningful comparison, the same segment was used for the
M
151 following four different configurations: (1) Hemisection model-1 (Hemi-1): An implant

152 (3.7×9 mm) replaced the mesial root of the molar after hemisection, and a single
ed

153 crown was placed on the implant and the residual tooth. (2) Hemisection model-2

154 (Hemi-2): An implant (3.7×9 mm) replaced the mesial root of the molar after
pt

155 hemisection and extraction of the mesial half of the tooth, and two separate crowns
ce

156 for the implant and the residual tooth were modeled. (3) Single implant model

157 (Single): An implant (5.5×9 mm) with crown replaced the whole molar after tooth
Ac

158 extraction. This model represented conventional implant restoration for the molar. (4)

159 Fixed partial denture model (FPD): A 3-unit fixed partial denture combined the distal

160 residual part of the molar and premolar. This model represented the traditional

161 clinical treatment for tooth after hemisection. The four models are illustrated in Figure

162 2. The final models consisted of 126,631 (Hemi-1), 124,853 (Hemi-2), 145,187

163 (Single), and 106,993 (FPD) four-noded tetrahedral elements. 3D FE analysis was

Page 8 of 38
164 performed using the software package MSC.Marc/Mentat (MSC. Software, Santa

165 Ana, CA).

166 Material properties: All materials were assumed to be homogeneous, isotropic

167 and linearly elastic. Young’s modulus and Poisson’s ratio of materials used in the

168 analysis were taken from the literatures (Bessone et al., 2014; Eraslan et al., 2005;

t
169 Pratheep et al., 2013) and are listed in Table 2. Because the distal residual part of

ip
170 the molar was assumed to undergo endodontic treatment, material properties of

cr
171 gutta-percha were assigned to the pulp of the distal molar.

us
172 Contact: According to the clinical situation, touching contact was defined between

173 prosthesis and implant, prosthesis and the teeth, and between neighboring teeth.

174 an
The implant was assumed to be completely osseointegrated at the implant/bone

175 interface.
M
176 Boundary conditions and loading: All models were constrained in all directions at

177 the nodes on the mesial, distal and bottom of the segment. For purposes of
ed

178 comparison, a vertical force of 100 N was applied on the crown. For the Hemi-2

179 model, each crown was loaded with 50 N, and 100 N was applied to the pontic with
pt

180 the Single model (Figure ).

ce

181 Displacements, stresses and strains were evaluated and compared for the four

182 models.
Ac

183

Page 9 of 38
183 3. Results

184 3.1 Radiologic Results

185 The distance between the root of second premolar and the distal root of the first

186 molar significantly increased from furcation level (7.6±0.6 mm) to middle level

187 (9.3±0.8 mm) and to apical level (10.6±1.1 mm). The mean length of the mesial root

t
188 of the molar was 10.1±1.0 mm.

ip
189 The density in grey values of cortical bone and spongious bone at the region of

cr
190 interest were 1437-3203 and 1023-1803, respectively. The mean thickness of the jaw

us
191 segment was 3.4±1.5 mm and the mean thickness of cortical bone was 1.9±0.7 mm.

192 The detailed measured values are illustrated in Table 1. The length of the distal root

193 an
was 10.1 mm and its width (middle third) was 5.5 mm.

194 3.2 Numerical results

M
195 3.2.1 Displacement of the Implant and Teeth

196 Implant and tooth displacements are depicted in (Figure 5). The highest values of
ed

197 maximum displacement were observed with the lingual side of the residual molar in

198 the Hemi-2 model (10.0 µm), while the lowest values were observed with the implant
pt

199 in the Single model (4.5 µm). The maximum displacements of the implants in all
ce

200 models were obtained at the lingual side of the abutment. The displacements of the

201 teeth were higher than those of the implant with the Hemi-1 and Hemi-2 models.
Ac

202 Higher displacements for both implant and teeth were obtained with the Hemi-2

203 model compared with the Hemi-1 model. The difference in the maximum

204 displacement between tooth and implant in the Hemi-1 and Hemi-2 models (1.0 µm)

205 was similar to those between the premolar and the molar in the FPD model.

206 3.2.2 Stresses in the Cortical Bone

207 The models with implants indicated higher stresses in the cortical bone compared

208 to the FPD model (4.5 MPa), see (Figure 6). The highest values of the maximum

Page 10 of 38
209 stress were observed with the Hemi-2 model (19.0 MPa), followed by the Single

210 model (13.0 MPa) and Hemi-1 model (12.0 MPa). Additionally, the distribution of the

211 stresses concentrated around the neck of the implants, the Hemi-2 model showed a

212 wider distribution than other models.

213 3.2.3 Stresses in the teeth

t
214 The highest values of the maximum stress in the teeth (Figure 7) were observed

ip
215 at the distal side of the premolar underneath the preparation line with the FPD model

cr
216 (11.6 MPa). The premolar showed a wider and higher stress distribution compared

us
217 with the molar. The maximum stress for the Hemi-1 model (6.0 MPa) was lower than

218 that for Hemi-2 model (9.0 MPa). The stress distribution concentrated around the

219 an
mesial region of the residual molar in the Hemi-2 model. Interestingly, the premolar at

220 Hemi-2 model showed as well a higher stress (3.1 MPa) compared to the Hemi-1
M
221 model (0.6 MPa).

222 3.2.4 Stresses in the Implant

ed

223 Implant stresses are shown in Figure 8. There were higher stresses with the

224 Hemi-1 and Single models, which had the same maximum value of 45.0 MPa. Using
pt

225 two separate crowns for implant and molar after hemisection reduced the maximum
ce

226 stress by around 33 % in the implant. The distribution concentrated around the neck

227 of the implants.

Ac

228 3.2.5 Strains in the Spongious Bone

229 Maximal strains in the spongious bone (Figure 9) for Hemi-1 and Single models

230 were below the physiological upper limit of 3,000 µstrain (Frost, 2003). The highest

231 values of the maximum strain were observed in a small lingual region with the Hemi-2

232 model (5,000 µstrain), while the lowest values were obtained with the Single model

233 (1,800 µstrain). The spongious bone around the premolar had a wider strain

234 distribution in the FPD model in comparison to the molar and at the tip of the

Page 11 of 38
235 premolar a strain of 4,300 µstrain was registered. Moreover, a wider strain

236 distribution with higher values was observed around the molar compared with the

237 implant in the Hemi-1 and Hemi-2 models.

238 3.2.6 Strains in the Periodontal Ligament

239 Strains in the periodontal ligament (PDL) are represented in Figure 10. The highest

t
240 strain was obtained with the FPD model. Strains were concentrated in the PDL of the

ip
241 premolar with the maximum value of 8,000 µstrain, which was two times higher than

cr
242 those around the molar (4,000 µstrain). The maximum strain in the Hemi-1 model

us
243 (5,000 µstrain) was lower than the strain in the Hemi-2 model (7,000 µstrain), and a

244 wider strain distribution was calculated with the Hemi-2 model.

245 an
246
M
ed
pt
ce
Ac

Page 12 of 38
246 4. Discussion

247 A variety of prosthetic techniques can be used to restore the molar with lesions

248 restricted to one root or furcation. This study aimed at evaluating the feasibility of

249 integrating a dental implant and tooth hemisection to preserve and recover a

250 complete molar.

t
251 Nowadays, the finite element analysis (FEA) plays an important role in solving

ip
252 engineering problems in many fields of science and it can be successfully applied in

cr
253 simulations of biomechanical systems (Pessoa et al., 2011; Sakaguchi and

us
254 Borgersen, 1993). The FEA allows the analysis of material mechanical properties and

255 critical regions of very complex geometry of biological structures (Natali, 2002).

256 an
Therefore, it has been widely used in dental implant biomechanics to predict its long-

257 term clinical success and evaluate the effect of various parameters, for example,
M
258 implant geometry, prosthesis design, and stress and strain distributions in peri-

259 implant regions (Hasan et al., 2012; Chun et al., 2012). A definite evaluation of the
ed

260 biomechanical performance of the system can only be carried out using a realistic

261 numerical model, thus requiring the precise reconstruction of the geometry (Rahimi et
pt

262 al., 2005). For this purpose, various imaging programs have been introduced that can
ce

263 generate three dimensional (3D) models by layering image data obtained by CT

264 scanning (Daas et al., 2008; Shigemitsu et al., 2014; Wakabayashi et al., 2010). In
Ac

265 this study, Mimics research 17.0 and 3-Matic research 9.0 were used to reconstruct

266 the models based on an idealized CBCT data. According to the results of bone

267 density and thickness, the bone can be classified in D-2 type described by Lekholm

268 and Zarb (1985). Average occlusal forces in the posterior region ranges from 79 to

269 331 N (Morikawa, 1994). Since, in first approximation, the occlusal force and stress

270 are in a linear relationship in which stress becomes larger with the increase in the

271 occlusal force (Park et al., 2014), and the principal aim of this study is the relative

Page 13 of 38
272 comparison of stresses and strains generated on the bone and implant, the direction

273 of loads was set to be perpendicular to the prosthesis. Again, in order to ensure that

274 for all the four models an equal magnitude of force (100 N) was applied, the metal

275 crowns in Hemi-1 and Single models were loaded. In the Hemi-2 model, each

276 individual metal crown was loaded with 50 N, and finally 100 N was applied to the

t
277 pontic in the FPD model.

ip
278 For the best of our knowledge, this study is the first to evaluate a possible

cr
279 treatment option, combining an implant and a residual molar after hemisection. In this

us
280 case, a controversy would be developed regarding whether implants should be

281 connected to a natural abutment tooth. This complication is due to the different

282 an
mobility patterns of the osseointegerated implant on the one hand and natural teeth

283 on the other hand (Pesun, 1997). An osseointegerated implant is rigidly fixed to the
M
284 bone and can only move around 10 μm in the apical direction, whereas teeth with

285 healthy periodontal ligament can move 25 to 100 μm (Sekine et al., 1986; Pratheep
ed

286 et al., 2013). Consequently, under masticatory load, different patterns of stress and

287 strain can be seen in the bone around an implant and a tooth, which may result in
pt

288 intrusion of the natural tooth, abutment screw loosening, and increased marginal
ce

289 bone loss (Naert et al., 1992). In a meta-analysis study, Muddugangadhar et al.

290 (2015) concluded that the survival rate of implant-supported single crowns was
Ac

291 96.363%, implant-supported FPDs was 94.525% and implant tooth-supported

292 prostheses was 91.27% after 5 years of function. In their opinions, the connection

293 between teeth and implants may not be considered as the first alternative for

294 rehabilitation, and it is preferable to adopt planning of isolated implant supported

295 dentures. However, tooth implant-supported prostheses can be provided if there are

296 certain limitations prohibiting a completely implant-supported prostheses.

297 Nevertheless, many studies showed that the combination between teeth and implants

Page 14 of 38
298 to support a fixed bridge indicated similar success rates to those of fixed implant-

299 supported prosthesis (Gunne et al., 1999; Nickenig et al., 2006). Lanza et al., (2011)

300 suggested that the ideal tooth-implant supported fixed dentures are those in which

301 the space between the abutments is small, including only one tooth and one implant.

302 The present study was different from them, since there was no pontic between the

t
303 implant and tooth.

ip
304 Concerning the displacement of the implant and the teeth, in the present study

cr
305 the teeth showed higher maximum displacement than implants in the Hemi-1 and

us
306 Hemi-2 models. It is noteworthy to mention that the disparity of the values of

307 maximum displacement between the implant and tooth with Hemi-1 and Hemi-2

308 an
models were similar to those between the premolar and the molar with the FPD

309 model, which indicates that the combination of implant and tooth might be
M
310 comparable to combining teeth with different PDL and root to some extent.

311 The FPD model showed the lowest stress in cortical bone, whereas highest
ed

312 values were observed with the Hemi-2 model. This might be ascribed to the fact that

313 the teeth with PDL are capable of transferring and buffering load to the surrounding
pt

314 bone. Moreover, in the Hemi-1 model, the combined tooth may help the implant to
ce

315 share part of the stress, and result in lower stress values compared with the Hemi-2

316 and Single models. As for the maximum stresses in the teeth and strains in the PDL,
Ac

317 a similar situation was obtained. The FPD model represented higher values in the

318 premolar compared with other models, and the values were two times higher than

319 those in the molar. This indicated that the premolar burdened more load and the FPD

320 model can be more dangerous for teeth.

321 This is consistent with the results obtained in a study conducted by Zafiropoulos

322 et al. (2009) in which the outcomes of two therapies for at least 4 years of

323 maintenance were compared: Molars treated by hemisection and implants replaced

Page 15 of 38
324 periodontally involved molars. They concluded that the former therapy had a greater

325 incidence of overall complications (32 %) and non-salvageable complications (13 %)

326 than the implant therapy (10 % and 3 %). The Hemi-1 model also showed lower

327 strains in the PDL of the molar compared with the Hemi-2 model, which may

328 attributed to the combined design. Furthermore, the stresses of the implant were

t
329 much higher than those of the tooth in the Hemi-1 and Hemi-2 models, which was in

ip
330 agreement with the study of de Paula et al. (2012) in which the effect of prosthesis

cr
331 length and implant diameter on the stress distribution in tooth-implant-supported

us
332 prostheses were analyzed by means of the finite element method.

333 Regarding the stresses in the implant, the values were all within the acceptable

334 an
range, below 100 MPa (Bozkaya et al., 2004; Hansson et al., 2000). The lowest

335 values were observed with the Hemi-2 model and the same higher maximum
M
336 stresses were obtained with the Hemi-1 and Single models. This demonstrated that

337 the implant burdened more stress when connected with the tooth, however the
ed

338 magnitude of maximum stress was just similar to restoring the molar with a single

339 wider implant.

pt

340 Large areas of the spongious bone in all models experienced strain below 3,000
ce

341 µstrain, which is commonly indexed as the threshold for bone-fatigue failure (Frost,

342 2003; Hudieb et al., 2010). A small lingual region in the Hemi-2 model (5,000 µstrain)
Ac

343 and the apical region of the premolar (4,300 µstrain) in the FPD model showed

344 higher values. The thinner cortical thickness (1.7 mm around the root and 1.4 mm

345 around the implant) and bone thickness (1.9 mm around the root and 2.0 mm around

346 the implant) were recorded in this higher strain lingual region, which could be one of

347 the possible reasons for higher strain. This result might also explain the following

348 clinical complication that the teeth serving as abutments for a FPD are prone to

349 periapical lesions (Yamazaki, 2013). For a smaller implant, such as the implant

Page 16 of 38
350 (3.7×9 mm) used in this study, higher strains may be induced in the spongious bone

351 compared with wider implants. Additionally, these results are in accordance with the

352 study of Saab et al. (2007). They compared the strain distribution on the bone around

353 an implant (4×13 mm) in the anterior maxilla using straight and angled abutments by

354 applying 178 N oblique loads. The load applied near the cingulum area of the

t
355 prosthesis had an angle of 130 degrees with respect to the long axis of the implant.

ip
356 The results showed that small areas of cancellous bone in straight and angled

cr
357 abutment models experienced strain above 4,000 μstrain. From a clinical point of

us
358 view, combined implant and residual molar to support prosthesis after tooth

359 hemisection also has many advantages. Firstly, tooth hemisection can prevent and

360 an
reduce the loss of alveolar bone by acquiring stimulation from the residual part of the

361 tooth. The less absorption bone facilitates the insertion of the implant. Otherwise, the
M
362 alveolar bone, in some cases, is not sufficient for inserting an implant due to the

363 resorption after tooth extraction. Secondly, hemisection preserves the tooth tissue as
ed

364 much as possible and avoids damage to the neighboring teeth, which is a big

365 problem of a conventional denture. Thirdly, the PDL of the residual molar offers a
pt

366 feedback mechanism to avoid overload on the implant and share the load with the
ce

367 implant.

368 The present study has several limitations and some assumptions. Firstly, the
Ac

369 materials in the models were all assumed to be homogeneous, isotropic and linearly

370 elastic. However, it is well documented that the jaw bone is transversely isotropic and

371 heterogeneous (Cochran, 2000). Despite the difference in stress values between

372 models created with isotropic or anisotropic bone, the stress distribution seems to be

373 similar (Bonnet et al., 2009). Additionally, modeling bone with anisotropic property is

374 complex, problematic and time-consuming in nature. Again, PDL is nonlinear and

375 anisotropic in nature (Provatidis, 2000). But, the biomechanical properties of PDL

Page 17 of 38
376 have not yet been fully investigated. In addition, considering the aim of the present

377 study was to compare different designs of restoration, a 100% implant-bone interface

378 was established, and the implant and the abutment were modeled as one piece,

379 which does not match clinical situations. Thus, the results of FEA should be

380 interpreted with some care. However, the present numerical results can help to

t
381 understand the different biomechanical properties for the different restoration

ip
382 designs. Therefore, the results we obtained could be considered as a reference to

cr
383 choose between different restoration designs after tooth hemisection in the clinical

us
384 treatment. Further research regarding experimental techniques and prospective

385 clinical studies are required to verify the results.

386 an
M
ed
pt
ce
Ac

Page 18 of 38
386 5. Conclusion

387 Within the limitations of this study, the following may be concluded:

388 1. The disparity of the values of maximum displacement between the implant and

389 tooth with Hemi-1 and Hemi-2 models were similar to those between the premolar

390 and the molar with the FPD model, which indicates that the combination of implant

t
391 and tooth might be comparable to combining teeth with different PDL and root to

ip
392 some extent.2. The stress was concentrated around the neck of the implants in

cr
393 Hemi-1, Hemi-2 and Single models.

us
394 3. In FPD model, the premolar was burdened by overloads, which indicates that

395 there is a risk of overloading the premolars with this kind of restoration.

396 an
4. Compared to the Hemi-2 model, lower stress in cortical bone and lower strains

397 in spongious bone and PDL were observed in the Hemi-1 model. Therefore, a single
M
398 crown placed on the implant and distal half of the molar is better than two crowns for

399 them.
ed

400 5. From the biomechanical point of view, the support of a prosthesis by an

401 implant and residual molar after hemisection seems to be an acceptable treatment
pt

402 option.
ce

403 However, prospective clinical studies with long-term follow-up are required for

404 definitive conclusions.

Ac

405

406 Conflict of interest: The authors declare that there is no conflict of interest.

407

408 Acknowledgement

409 This study was conducted in Endowed Chair of Oral and Technology, University

410 of Bonn, Germany, and funded by a grant from the Scientific Department of Luzhou,

411 China (No. 201127) and Luzhou Medical College, China (No. 2012QN-44).

Page 19 of 38
412 The study sponsors have no involvement in the study design, in the collection,

413 analysis and interpretation of data, in the writing of the manuscript, or in the decision

414 to submit the manuscript for publication.

415

416 Declaration

t
417 This study was approved by the Institutional Ethics Committee of Hospital of

ip
418 Stomatology, Luzhou Medical College (certificate number, 2011002).

cr
419

us
420

an
M
ed
pt
ce
Ac

Page 20 of 38
420 References

421 Bessone, L.M., Bodereau, E.F., Cabanillas, G., Dominguez, A., 2014. Analysis of

422 Biomechanical Behaviour of Anterior Teeth Using Two Different Methods: Finite

423 Element Method and Experimental Tests. Engineering. 6, 148-158.

424 Bonnet, A.S., Postaire, M., Lipinski, P., 2009. Biomechanical study of mandible

t
425 bone supporting a four-implant retained bridge: finite element analysis of the

ip
426 influence of bone anisotropy and foodstuff position. Med Eng Phys. 7, 806-815.

cr
427 Bozkaya, D., Muftu, S., Muftu, A., 2004. Evaluation of load transfer characteristics

us
428 of five different implants in compact bone at different load levels by finite elements

429 analysis. J. Prosthet. Dent. 92, 523-530.

430 an
Buhler, H., 1994. Survival rates of hemisected teeth: an attempt to compare them

431 with survival rates of alloplastic implants. Int. J. Periodontics Restorative Dent. 14,
M
432 536-543.

433 Chun, C.M., Huang, H.L., Hsu, J.T., Fuh, L.J., 2012. Influences of internal tapered
ed

434 abutment designs on bone stresses around a dental implant: three-dimensional finite

435 element method with statistical evaluation. J. Periodontol. 83, 111-118.

pt

436 Cochran, D.L., 2000. The scientific basis for and clinical experiences with
ce

437 Straumann implants including the ITI Dental Implant System: a consensus report.

438 Clin Oral Implants Res. 11, 33-58.

Ac

439 Daas, M., Dubois, G., Bonnet, A.S., Lipinski, P., Rignon-Bret, C., 2008. A

440 complete finite element model of a mandibular implant-retained overdenture with two

441 implants: comparison between rigid and resilient attachment configurations. Med.

442 Eng. Phys. 30, 218-225.

443 Desai, S.R., Karthikeyan, I., Singh, R., 2012. Evaluation of Micromovements and

444 Stresses around Single Wide-Diameter and Double Implants for Replacing

445 Mandibular Molar: A Three-Dimensional FEA. ISRN. Dent. 7, 1-10..

Page 21 of 38
446 Eraslan, O., Sevimay, M., Usumez, A., Eskitascioglu, G., 2005. Effects of

447 cantilever design and material on stress distribution in fixed partial dentures--a finite

448 element analysis. J. Oral Rehabil. 32, 273-278.

449 Frost, H.M., 2003. Bone's mechanostat: a 2003 update. Anat. Rec. A Discov. Mol.

450 Cell Evol. Biol. 2, 1081-1101.

t
451 Fugazzotto, P.A., 2001. A comparison of the success of root resected molars and

ip
452 molar position implants in function in a private practice: results of up to 15-plus years.

cr
453 J. Periodontol. 72, 1113-1123.

us
454 Gunne, J., Astrand, P., Lindh, T., Borg, K., Olsson, M., 1999. Tooth-implant and

455 implant supported fixed partial dentures: a 10-year report. Int. J. Prosthodont. 12,

456 216-221. an
457 Hasan, I., Röger, B., Heinemann, F., Keilig, L., Bourauel, C., 2012. Influence of
M
458 abutment design on the success of immediately loaded dental implants: Experimental

459 and numerical studies. Med. Eng. Phys. 7, 817-825.

ed

460 Hansson, S., 2000. Implant-abutment interface: biomechanical study of flat top

461 versus conical. Clin. Implant Dent. Relat. Res. 2, 33-41.

pt

462 Holm-Pedersen, P., Lang, N.P., Müller, F., 2007. What are the longevities of teeth
ce

463 and oral implants? Clin. Oral Implants Res. 18, 15-19.

464 Hudieb, M., Wakabayashi, N., Suzuki, T., Kasugai, S., 2010. Morphologic
Ac

465 classification and stress analysis of the mandibular bone in the premolar region for

466 implant placement. Int. J. Oral Maxillofac. Implants. 25, 482-490.

467 Lanza, M.D, Seraidarian, P.I, Jansen, W.C, Lanza, M.D., 2011. Stress analysis of

468 a fixed implant-supported denture by the finite element method (FEM) when varying

469 the number of teeth used as abutments. J. Appl. Oral Sci. 6, 655-661.

Page 22 of 38
470 Lekholm, U., Zarb, G.A., 1985. Patient selection and preparation. In Tissue-

471 integrated prostheses: osseointegration in clinical dentistry Bränemark PI., Zarb GA.,

472 Albrektsson T., editors. Chicago, pp. 199-209.

473 Lindth, T., Gunne, J., Tillberg, A., Molin, M., 1998. A meta-analysis of implants in

474 partialedentulism. Clin. Oral Implants Res. 9, 80-90.Morikawa, A., 1994. Investigation

t
475 of occlusal force on lower first molar in function. Kokubyo Gakkai Zasshi. Jun, 2):250-

ip
476 74.

cr
477 Muddugangadhar, BC., Amarnath, GS., Sonika, R., Chheda, PS., Garg, A., 2015.

us
478 Meta-analysis of Failure and Survival Rate of Implant-supported Single Crowns,

479 Fixed Partial Denture, and Implant Tooth-supported Prostheses. J Int Oral Health. 9,

480 7-11. an
481 Naert, I., Quirynen, M., van Steenberghe, D., Darius, P., 1992. A study of 589
M
482 consecutive implants supporting complete fixed prosthesis. Part II: Prosthetic

483 aspects. J. Prosthet. Dent. 68, 949-956.

ed

484 Natali, A.N., Pavan, P.G., 2002. A comparative analysis based on different

485 strength criteria for evaluation of risk factor for dental implants. Comput. Methods
pt

486 Biomech. Biomed. Engin. 2, 127-133.

ce

487 Naveen, Y.G., Patel, J.R., Parikh, P., Shah, K., 2014. Alternatives for restoration of

488 a hemisected mandibular molar. BMJ. Case. Rep. 30, 75.

Ac

489 Nickenig, H.J., Schäfer, C., Spiekermann, H., 2006. Survival and complication

490 rates of combined tooth-implant-supported fixed partial dentures. Clin. Oral Implants

491 Res. 17, 506-511.

492 Park, J.M., Kim, H.J., Park, E.J., Kim, M.R., Kim, S.J., 2014. Three dimensional

493 finite element analysis of the stress distribution around the mandibular posterior

494 implant during non-working movement according to the amount of cantilever. J. Adv.

495 Prosthodont. 6, 361-371.

Page 23 of 38
496 de Paula, G.A., da Mota, A.S., Moreira, A.N., de Magahlães, C.S., Cornacchia,

497 T.P., Cimini, C.A Jr., 2012. The effect of prosthesis length and implant diameter on

498 the stress distribution in tooth-implant-supported prostheses: a finite element

499 analysis. Int. J. Oral Maxillofac. Implants. 27, 19-28.

500 Pessoa, R.S., Coelho, P.G., Muraru, L., Marcantonio, E.Jr. Vaz, L.G., Sloten, J.V.,

t
501 Jaecques, S.V., 2011. Influence of implant design on the biomechanical environment

ip
502 of immediately placed implants: computed tomography-based nonlinear three-

cr
503 dimensional finite element analysis. Int. J. Oral Maxillofac. Implants. 26, 1279-1287.

us
504 Pesun, I.J., 1997. Intrusion of teeth in the combination implant-to-natural-tooth

505 fixed partial denture: A review of the theories. J. Prosthodont. 6, 268-277.

506 an
Pjetursson, B.E., Tan, K., Lang, N.P., Brägger, U., Egger, M., Zwahlen, M., 2004.

507 A systematic review of the survival and complication rates of fixed partial dentures
M
508 (FPDs) after an observation period of at least 5 years. Clin. Oral Implants Res. 15,

509 625-642.
ed

510 Pratheep, K.V., Abraham, A., Annapoorni, H., Vigneshwaran, S., 2013.

511 Comparative evaluation of stresses in tooth implant connected fixed partial denture
pt

512 by varying the implant design and position: a 3D finite element study. Indian J. Dent.
ce

513 Res. 24, 439-445.

514 Provatidis, C.G., 2000. A comparative FEM-study of tooth mobility using isotropic
Ac

515 and anisotropic models of the periodontal ligament. Finite Element Method. Med Eng

516 Phys. 5, 359-370.

517 Rahimi, A., Keilig, L., Bendels, G., Klein, R., Buzug, T.M., Abdelgader, I., Abboud,

518 M., Bourauel, C., 2005. 3D reconstruction of dental specimens from 2D histological

519 images and CT-scans. Comp. Meth. Biomech. Biomed. Eng. 8, 167-176.

Page 24 of 38
520 Saab, X.E., Griggs, J.A., Powers, J.M., Engelmeier, R.L., 2007. Effect of abutment

521 angulation on the strain on the bone around an implant in the anterior maxilla: a finite

522 element study. J. Prosthet. Dent. 97, 85-92.

523 Sakaguchi, R.L., Borgersen, S.E., 1993. Nonlinear element contact analysis of

524 dental implant components. Int. J. Oral Maixillofac. Implants 7, 655-661.

t
525 Sekine, H., Komiyama Y., Hotta, H., Yoshida, K., 1986. Mobility characteristics

ip
526 and tactile sensitivity of osseointegrated fixture-supporting system. Oral and

cr
527 Maxillofac Reconstr. 6, 326-332.

us
528 Shigemitsu, R., Yoda, N., Ogawa, T., Kawata, T., Gunji, Y., Yamakawa, Y., Ikeda,

529 K., Sasaki K., 2014. Biological-data-based finite-element stress analysis of

530 an
mandibular bone with implant-supported overdenture. Comput. Biol. Med. 54, 44-52

531 Wakabayashi, N., Kondo, T., Yahagi, R., Suzuki, T., 2010. A patient-based model
M
532 study of fixed splinting of premolars with reduced periodontal support. Int. J. Comput.

533 Dent. 4, 317-330.

ed

534 Yamazaki, S., Arakawa, H., Maekawa, K., Noda, K., Hara, E.S., Minakuchi, H.,

535 Sonoyama, W., Matsuka, Y., Kuboki, T., 2013. A retrospective comparative 8-year
pt

536 study of cumulative complications in teeth adjacent to both natural and implant-
ce

537 supported fixed partial dentures. Int. J. Prosthodont. 3, 260-264.

538 Zafiropoulos, G.G., Hoffmann, O., Kasaj, A., Willershausen, B., Deli, G., Tatakis,
Ac

539 D.N., 2009. Mandibular molar root resection versus implant therapy: a retrospective

540 nonrandomized study. J. Oral Implantol. 35, 52-62.

541

542

543

Page 25 of 38
543 Tabels

544 Table 1: Thickness of the jaw segment and cortical bone (mm). CBT: cortical

545 bone thickness; JST: jaw segment thickness; HR: Hemi-root; SI: single implant;

546 HI: Hemi-implant; PR: premolar.

HR SI HI PR Total
Buccal CBT 2.4±1.4 2.4±1.2 2.0±0.9 1.8±0.7 2.1±1.0

t
Lingual

ip
1.7±0.3 1.8±0.5 1.8±0.4 1.8±0.2 1.7±0.3
CBT
Buccal JST 3.8±2.6 3.6±1.9 3.8±1.4 3.0±1.9 3.5±1.7

cr
Lingual JST 3.3±1.5 3.3±0.9 3.7±1.6 2.7±1.1 3.2±1.2
547

us
548 Table 2: Material properties of the numerical models.

Poisson
Material
Titanium
an
Young’s modulus (MPa)
110,000
ratio
0.30
Cortical bone 14,000 0.30
Spongious bone 1,370 0.30
M
Tooth(dentin) 19,000 0.30
Pulp 20 0.45
Pdl 170 0.45
ed

Metal crown (NiCr

200,000 0.33
alloy)
Guttapercha 0.69 0.45
pt

549

550
ce
Ac

Page 26 of 38
550 Legend of Figures

551 Figure 1: Radiological measurement. (a) Measurement of the length of the mesial

552 root of the molar in the sagittal view. (b) Measurement of the distance between the

553 furcation point of the distal root and the root of premolar (DFR) in the axial view. (c)

554 Measurement of the distance between the middle point (the midpoint of the furcation

t
555 point and apical point) and the root of premolar (DMR) in axial view. (d) Measurement

ip
556 of the distance between the apical point of the distal root and the root of premolar

cr
557 (DAR) in the axial view. F: furcation point; A: apical point; M: the midpoint of the

us
558 furcation point and the apical point.

559 Figure 2: Defining masks by different grey value ranges in Mimics.

560
an
Figure 3: Measuring the total thickness of the jaw segment as well as thickness of
M
561 cortical bone. (a) Hemi-1 model. (b) Single model. HR: Hemi-root; HI: Hemi-implant;

562 PR: premolar; SI: single implant; R: alveolar ridge; A: apex; M: the midpoint.
ed

563 Figure 4: The investigated numerical models and their boundary conditions. (a)

564 Hemisection model 1 (Hemi-1). (b) Hemisection model 2 (Hemi-2). (c) Single implant
pt

565 model (Single).

ce

566 (d) Fixed partial denture model (FPD).

Ac

567 Figure 5: Comparison of the displacement in the implant (the left) and the teeth (the

568 right).

569 Figure 6: Comparison of the stresses in the cortical bone. (a) Obtained maximum

570 values of equivalent stress for the cortical bone. (b) Equivalent stress distribution in

571 four models.

572 Figure 7: Comparison of the stresses in the tooth. (a) Obtained maximum values of

573 equivalent stress for the tooth. (b) Equivalent stress distribution in three models.

Page 27 of 38
574 Figure 8: Comparison of the stresses in the implant. (a) Obtained maximum values

575 of equivalent von Mises stress for implants. (b) Stress distribution in three models

576 with implants.

577 Figure 9: Comparison of the strains in the spongious bone. (a) Obtained maximum

578 values of equivalent of total strain for the spongious bone. (b) Equivalent of total

t
579 strain distribution in four models.

ip
580 Figure 10: Comparison of the strains in the PDL. (a) Obtained maximum values of

cr
581 equivalent of total strain for the PDL. (b) Equivalent of total strain distribution in three

us
582 models.

an
M
ed
pt
ce
Ac

Page 28 of 38
Figure 1

i
cr
us
an
M
ed
pt
ce
Ac

Page 29 of 38
Figure 2

i
cr
us
an
M
ed
pt
ce
Ac

Page 30 of 38
Figure 3

i
cr
us
an
M
ed
pt
ce
Ac

Page 31 of 38
Figure 4

i
cr
us
an
M
ed
pt
ce
Ac

Page 32 of 38
Figure 5

i
cr
us
an
M
ed
pt
ce
Ac

Page 33 of 38
Figure 6

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 34 of 38
Figure 7

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 35 of 38
Figure 8

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 36 of 38
Figure 9

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 37 of 38
Figure 10

t
ip
cr
us
an
M
d
te
p
ce
Ac

Page 38 of 38