Chun-Li Lin Jen-Chyan Wang Wen-Jen Chang

Biomechanical interactions in tooth implant-supported xed partial dentures with variations in the number of splinted teeth and connector type: a nite element analysis

Authors afliations: Chun-Li Lin, Department of Mechanical Engineering, Chang Gung University, Tao-yuan, Taiwan Jen-Chyan Wang, Department of Prosthodontics, Faculty of Dentistry, Kaohsiung Medical University, Kaohsiung City, Taiwan; Department of Dentistry, Division of Prosthodontics, Kaohsiung Medical University Hospital, Kaohsiung City, Taiwan Wen-Jen Chang, Department of Industrial Engineering and Management, Ta-Hwa Institute of Technology, Hsiung-Lin, Hsinchu County, Taiwan Correspondence to: Jen-Chyan Wang DDS, Associate Professor Department of Prosthodontics Faculty of Dentistry Kaohsiung Medical University Kaohsiung, Taiwan Tel.: 886-7-3121101 ext 2157 Fax: 886-7-3210637 e-mail: jechwz@kmu.edu.tw

Key words: biomechanics, connector, nite element analysis, splinting Abstract Objective: The aim of this study was to investigate the biomechanical interactions in tooth implant-supported xed partial dentures (FPDs) under several loading conditions with different numbers of splinted teeth and connector types (rigid and non-rigid) by adopting the three-dimensional (3D) non-linear nite element (FE) approach. Material and methods: A 3D FE FPD model was constructed containing one Frialit-2 implant in the mandibular second-molar region splinted to the rst and second premolars. Frictional contact elements were used to simulate realistic interface conditions within the implant system and the non-rigid connector function. The main effects for each level of the three investigated factors (loading condition, number of splinted teeth and connector type) in terms of the stress values and dissimilar mobility of the natural teeth and implant were computed for all models. Results: The results indicated that load condition was the main factor affecting the stress developed in the implant, bone and prosthesis when comparing the type of connector and the number of splinted teeth. The stress values were signicantly reduced in centric or lateral contact situations once the occlusal forces on the pontic were decreased. However, the prosthesis stress for the non-rigid connections was increased more than 3.4-fold relative to the rigid connections. Moreover, the average tooth-to-implant displacement ratios (RTID) with a non-rigid connection were obviously larger than those for rigid connections under axial loading forces. Adding an extra tooth to support a three-unit toothimplant FPD only exploited its function when the prosthesis withstood lateral occlusal forces. Conclusions: The load condition is the main factor affecting stress distribution in different components (bone, prosthesis and implant) of toothimplant-supported FPDs. Minimizing the occlusal loading force on the pontic area through selective grinding procedures could reduce the stress values obviously. A non-rigid connector may more efciently compensate for the dissimilar mobility between the implant and natural teeth under axial loading forces but with the risk of increasing unfavorable stresses in the prosthesis.

Date: Accepted 15 June 2006

To cite this article: Lin C-L, Wang J-C, Chang W-J. Biomechanical interactions in toothimplant-supported xed partial dentures with variations in the number of splinted teeth and connector type: a nite element analysis. Clin. Oral Impl. Res. 19, 2008; 107117 doi: 10.1111/j.1600-0501.2007.01363.x

The high success rate achieved for dental implants with respect to osseointegration has become an accepted clinical reality (Laufer & Gross 1998; Goodacre & Kan 1999; Becker et al. 2000; Sahin et al. 2002; Cehreli et al. 2004). A controversy has developed, however, regarding whether im-

plants should be connected to natural abutments or be self-supporting (Gross & Laufer 1997; Laufer & Gross 1998; Becker et al. 2000; Lin & Wang 2003). While splinting the implant and tooth is a rational alternative in some clinical situations, the dissimilar mobility patterns of the osseoin-

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

107

Lin et al . Biomechanical interactions in toothimplant splinting system

tegrated implants and natural teeth make the biomechanical behavior of the entire system complicated (Pesun 1997). An osseointegrated implant is rigidly xed to bone and can move only 10 mm in the apical direction, whereas teeth with healthy periodontal ligaments can move 25100 mm (Sekine et al. 1986; Nyman & Lang 1994). This movement disparity may cause relative motion of the implant-tooth superstructure when the splinted system is loaded by occlusal force. During loading, higher bending moment induced by the mismatch between the implant and tooth may result in abutment screw loosening (screw-retained) or fracture of implants or prosthesis. Loss of osseointegration and increased marginal bone loss may also occur around the implant as a result of overload (Naert et al. 2001b). To compensate for the dissimilar mobility of natural teeth and implant systems, several specic methods have been suggested. Non-rigid connectors act as stress breakers with the ability to separate the splinted units, thus compensating for the variance in mobility between the implant and tooth. However, several reports have explored the use of non-rigid connectors and the association with abutment tooth intrusion (Rieder & Parel 1993; Pesun 1997; Sheets & Earthman 1997; Becker et al. 2000). Speculative etiology for this phenomenon of tooth intrusion includes disuse atrophy, mechanical binding, and impaired rebound memory, etc. Indeed, clinical observations have also cast doubt with respect to function, with most revealing non-signicant or minimal differences between rigid and non-rigid toothimplant connections (McGlumphy et al. 1989; Mathews et al. 1991; Nishimura et al. 1999). By contrast, also toothimplant prostheses with rigid connection, advocated by certain authors, owing to the prosthesis and implant possess the inherent exibility to accommodate dissimilar mobility characteristics (Gross & Laufer 1997; Naert et al. 2001a; Menicucci et al. 2002; Tangerud et al. 2002). However, the results of long-term radiographic follow-up evaluation of toothimplant prostheses indicate more bone loss around implants that are rigidly connected to teeth as compared with non-rigid toothimplant connections (Naert et al. 2001b). Accordingly, consensus regarding the proper connector design

for implant-teeth splinting systems has not been attained. Lang et al. (2004) conducted a systematic review to assess the survival and complication rates of combined toothimplantsupported conventional xed partial dentures (FPDs) after an observation period of at least 5 years; the results indicated that the survival rates of both implants and reconstructions combined with tooth implant-supported FPDs were lower than those reported for solely implant-supported FPDs (Pjetursson et al. 2004). Intrusion was detected in 5.2% of abutment teeth among non-rigid connections after a 5-year period. Hence, treatments without connecting implants to teeth may be preferred. For reasons of anatomical structures or patient-centered preferences, FPDs supported by combination of implants and two abutment teeth with a rigid connection have been suggested. To reduce the risk of tooth intrusion, Becker et al. (2000) also advocated a minimum of two natural teeth connected (double-abutted) when an implant was considered for connection with natural teeth using a non-rigid connection. Indeed, connecting teeth together in a splint system is one of the signicant methods used to decrease mobility in compromised periodontal dentitions for FPDs. The inuence of the number of splinted teeth has been investigated in several biomechanical studies, with the authors pointing out that tooth splinting may decrease both displacement and stress concentrations, especially in periodontally compromised dentition. (Wylie & Caputo 1991; Yang et al. 1996, 1999) However, the effects and biomechanical aspects of the number of splinted teeth in toothimplant connected systems remain unclear and clarication is required. To further elaborate the mechanical responses of the splinting system, the occlusal force (magnitude, frequency, direction and location) is the predominant factor, directly affecting the load-transfer and stress distribution to the implant and teeth. Although higher bending moments on implants have been observed where rigid connectors were used (Van Oosterwyck et al. 1998; Wang et al. 1999), unfortunately, there has been insufcient focus on the mechanical interactions of bone, prosthesis and implants used for the splinting system where rigid/non-rigid connec-

tors are coupled with different numbers of splinted teeth under various loading situations. Experimental approaches or clinical observations reported to date have not provided sufcient information to allow determination of the biomechanics in complicated toothimplant-supported systems. The nite element (FE) method provides mechanical responses and alters parameters in a more controllable manner, driving its common use as an analytical tool in dental biomechanical studies (Lin et al. 2001; Tangerud et al. 2002; Lin & Wang 2003). Accordingly, nonlinear FE stress analysis has been used in this study to investigate: (1) the effects of loading condition, number of splinted teeth and rigid/ non-rigid connectors on the mechanical responses of bone, prosthesis and implant; (2) capabilities with respect to compensating for dissimilar mobility with rigid/ non-rigid connectors under several loading conditions; and (3) interactions between rigid/non-rigid connectors and number of splinted teeth in a tooth-implantsupported FPD.

Methods
FE model of a toothimplant-connected system

One Frialit-2 root-form implant, 4.5 mm in diameter and 13 mm in length, with a screw-retained MH-6 abutment (Friadent, Mannheim, Germany) was used as the investigated implant system. The detailed dimensions of the implant, abutment and abutment screw were obtained from our previous study (Lin & Wang 2003), and a solid model of the implant system was constructed using the CAD system (Pro/Engineering, Parametric Technology Co., MA, USA). The implant system was designed for the second molar position in an adult-human partially dentate mandibular model with optimal periodontal support, but edentulous distal to the second premolar. For the solid-model construction of teeth, prosthesis and alveolar bone, two freshly extracted intact rst and second premolars were cleaned to strip the periodontal ligaments. The roots were embedded into an epoxy resin block up to 1 mm below the cementumenamel junction (CEJ). After tooth preparation, wax-up

108 |

Clin. Oral Impl. Res. 19, 2008 / 107117

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

Lin et al . Biomechanical interactions in toothimplant splinting system

and casting procedures, a type II gold-alloy four-unit FPD from the rst premolar to the second molar was fabricated and xed onto the natural teeth with adhesive. The FPD portion of the system was then re-embedded to expose the toothresin prosthesis sections parallel to the mesial distal axis using a low-speed diamond saw (Isomet, Buehler, IL, USA). Dentine, pulp and prosthesis boundaries from each sectioned digital image were detected using an image-processing program developed in house (Lin et al. 1999). The coordinates for each point on the boundaries was entered into the FE program (ANSYS, Rel. 8.0; ANSYS Inc., Houston, TX, USA) to generate solid models of the tooth and prosthesis. On the basis of the root-form geometry of the premolars, simplied 0.25-mm periodontal ligaments (PDL) and alveolar bone (cortical and cancellous) volumes were then created according to the literature (Stegaroiu et al. 1998; Tada et al. 2003). Also, a non-rigid keyway attachment was designed between the second premolar and pontic. The implant system was then imported into ANSYS and placed in the second molar position to complete the solid model of the splinting system. A mesh model was generated using a mapping approach with eight-node isoparametric brick elements (solid 45); nonlinear frictional contact elements (contact49; dened as node to surface) were used to simulate the adaptation between the various components within the implant system (abutment/xture, abutment screw/abutment and abutment screw/xture). The frictional contact allowed tangential node slip without interpenetration between the different materials. This conguration permitted transfer of compressive and tangential forces without tension in the contact zone. A friction coefcient of 0.5 was assumed for all contact surfaces (Merz et al. 2000). It has been proved that using the contact mode to mimic intercomponent friction within the implant system provides a more realistic representation of interface xation, which then provides greater delity with respect to the relative component micro-motion (Merz et al. 2000; Lin & Wang 2003). Our FE model consisted of 45,364 bricks, 4517 contact elements and 50,320 nodes (Fig. 1).

Fig. 1. A three-dimensional nite element model of a tooth/implant-supported system was constructed for analysis in this study. (a) abutment, (b) implant screw, (c) xture, (d) premolars with keyway, (e) implant/ pontic with key portion and (f) alveolar bone (cortical and cancellous), (g) full model with two free-standing implant restorations.

Connector designs, teeth splinting and loading conditions

To understand the inuence of various connection designs and the number of splinted teeth used in the splinting systems, a previous teethimplant-connected FE model was modied and extended to match the parametric analyses. Two connector designs, non-rigid (keyway stress breaker) and rigid (one-piece), were considered the calculated modes to understand their inuences. When a rigid connector was used in the FE models, the nodes

attached to the male/female components at the same location needed to merge to modify the original interfacial xation (contact), becoming a bonded condition. This did not allow relative micro-motion, and the displacement was continuous between the different materials. Splinting of one or two natural teeth permitted examination of the interactions by deleting the connector elements between the rst and second premolars in the FE model. Under the frictional condition between various components within the implant

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

109 |

Clin. Oral Impl. Res. 19, 2008 / 107117

Lin et al . Biomechanical interactions in toothimplant splinting system

Table 1. Detailed loading positions, connecting types, splinted teeth and sequence of simulated nite element (FE) models in this study
Loading types (6) Contact positions Splinting teeth (splinting one tooth: 1; splinting two teeth: 2) 1 1 2 2 Connector type (non-rigid connector: 1; rigid connector: 2) 1 2 1 2 Sequences

Uniform multiple axial contacts on premolars (200N), pontic (200N) and molar (200N)

1 2 3 4

1 Uniform multiple oblique contacts on premolars (200N), pontic (200N) and molar (200N) 1 1 2 2 1 2 1 2 5 6 7 8

2 Only axial contacts on the rst and second premolars (200N) 1 1 2 2 1 2 1 2 9 10 11 12

3 Only oblique contacts on the rst and second premolars (200N) 1 1 2 2 1 2 1 2 13 14 15 16

4 Uniform multiple axial contacts on premolars (200N) and molar (200N) with a reduced force on pontic (40N) 1 1 2 2 1 2 1 2 17 18 19 20

5 Uniform multiple oblique contacts premolars (200N) and molar (200N) with reduced force on pontic (40N) 1 1 2 2 1 2 1 2 21 22 23 24

system, six loading types with two connector designs (non-rigid/rigid) and two splinting situations (one/two teeth) were considered the calculated modes to under-

stand the mechanical interactions. Table 1 provides details of the loading positions, connector types and splinting teeth for these FE models. The mesial and distal

exterior nodes of the bony segment were xed in all directions as the boundary conditions. Linear elastic, homogeneous and isotropic material properties of the

110 |

Clin. Oral Impl. Res. 19, 2008 / 107117

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

Lin et al . Biomechanical interactions in toothimplant splinting system

Table 2. Material properties assigned to implant material, dental tissues, prosthesis and alveolar bone
Materials Titanium (implant system) Dentin Periodontal ligament (PDL) Gold alloy (prosthesis) Cortical bone Cancellous bone Youngs modulus (MPa) 110,000 18,600 69 90,000 15,000 1500 Poissons ratio 0.35 0.31 0.45 0.3 0.3 0.3 References Benzing et al. (1995), Van Rossen et al. (1990) Middleton et al. (1996) Ko et al. (1992) Benzing et al. (1995), Moffa et al. (1973) Stegaroiu et al. (1998) Stegaroiu et al. (1998)

Table 3. Contact-simulated results of maximal von Mises stresses of the implant (rI, max), prosthesis (rP, max) and alveolar bone (rAB, max) for all simulated models with 6 loading types
Loading types 1 Splinting teeth (splinting one tooth: 1; splinting two teeth: 2) 1 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2 Non (two free-standing implant restorations) Non (two freestanding implant restorations) Connector type (non-rigid connector: 1; rigid connector: 2) 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 Non (two free-standing implant restorations) Non (two freestanding implant restorations) Implant system (sI, max) (MPa) 481.5 470.4 426.6 421.6 1329 1335 1200 1207 22 54.3 30.7 38.6 171.4 219.2 178.2 229.8 290.7 303.7 269.9 304.7 953.7 944.5 905.3 882.8 302.8 1101 Prosthesis (sP, max) (MPa) 352.1 73.6 303.5 75.2 386.8 141 506 132.9 31.6 20.5 48.3 19.8 239.2 53 222.7 70.5 93.2 35.4 84.7 36.7 256.6 91.6 333.3 81.4 35.2 114.7 Alveolar bone (sAB, max) (MPa) 34.8 36 45.7 46.3 154.3 150.4 153.8 145.4 7.5 7.5 7.8 8.7 45.6 45.2 44.3 44 26.2 26.5 31 31.2 112.4 108.6 112.5 105.6 24.8 130

1 2

dental tissue, PDL, prosthesis and implant system were assumed in the simulations, as adopted from the relevant literatures (Table 2). Maximum von Mises stress values and the stress distributions in the implant system, alveolar bone and prosthesis were recorded for all models. To reduce the complexity of the results, the main effect of each level of the three investigated factors (load condition, number of splinted teeth and connector type) on the mechanical response (stress) was computed based on statistical methods (Dar et al. 2002). In addition, the data from simulated results were analyzed using a general linear model analysis of variance (ANOVA) test by MINITAB commercial statistical package (version 12.23, MINITAB Inc., PA, USA).

The analysis gave the percentage contribution that each investigated factor makes to the total sum of squares (TSS) and could determine the factor levels that minimize the stress in the implantbone system (Dar et al. 2002). Furthermore, the average displacements (in occlusal, bucco-lingual and mesio-distal directions) of the natural teeth and implant, dened as the average nodal crown displacements of premolars and implant, were calculated to investigate the non-rigid (keyway stress breaker) and rigid connector compensation. In order to avoid the incidence of tooth intrusion and other possible complications, many dentists favor a treatment concept without connecting implants to neighboring teeth in such a situation.

Therefore, a 3-D FE model with two freestanding implant restorations in rst and second molar region was constructed and tested as a reference for comparison (Fig. 1g).

Results
The maximal von Mises stress values for six loading types of toothimplant FPD and two loading types of freestanding implant restorations in the implant (sI, max), alveolar bone (sAB, max) and prosthesis (sP, max) are listed in Table 3. For free-standing implant-supported restorations, all the stress values of the implant, prosthesis and alveolar bone are smaller than those

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

111 |

Clin. Oral Impl. Res. 19, 2008 / 107117

Lin et al . Biomechanical interactions in toothimplant splinting system

a
Main effects plot for implant

MPa
1400 1200 1000 800 600 400 200 0 1 2 3 4 5 6

1400 1200 1000 800 600 400 200 0 1 (Tooth) 2 (Teeth)

1400 1200 1000 800 600 400 200 0 1 (Non-rigid)

Loading

Splinting
160 140 120 100 80 60 40 20 0 160 140 120 100 80 60 40 20 0 1 (Tooth) 2 (Teeth)

Connector

2 (Rigid)

b
Main effects plot for bone

MPa
160 140 120 100 80 60 40 20 0 1 2 3 4 5 6

1 (Non-rigid)

2 (Rigid)

Loading Main effects plot for prosthesis

Splinting
300 250 200 150 100 50 0 300 250 200 150 100 50 0 1 (Tooth) 2 (Teeth) 1 (Non-rigid)

Connector

300 250 200 150 100 50 0

MPa

2 (Rigid)

Loading

Splinting

Connector

Fig. 2. Main effects of the six loading types (corresponding to Table 1), two splinting situations (splinting one tooth: 1; splinting two teeth: 2) and two connector types (non-rigid connector: 1; rigid connector: 2) at each level for peak von Mises stress on (a) implant, (b) alveolar bone and (c) prosthesis.

Table 4. Summary of the analysis of variance (ANOVA) showing the statistical results of maximum stress with respect to implant
Source Loading type Splinting situation Connector type Loading type Splinting situation Loading type Connector type Splinting situation Connector type Error Total DF 5 1 1 5 5 1 5 23 SS 628.09 0.23 2.37 0.34 4.99 0.28 1.90 638.20 MS 126.62 0.23 2.37 0.07 1 0.28 1.90 %TSS 98.42 0.04 0.37 0.05 0.78 0.04 0.30 100 P-value 0 0.47 0.06 0.96 0.16 0.43

DF, degrees of freedom; SS, sum of squares; MS, mean square; TSS, total sum of squares.

Table 5. Summary of the analysis of variance (ANOVA) showed the statistical results of maximum stress with respect to bone
Source Loading type Splinting situation Connector type Loading type Splinting situation Loading type Connector type Splinting situation Connector type Error Total DF 5 1 1 5 5 1 5 23 SS 1663.66 2.68 0 5.34 0.58 0 0.27 1672.53 MS 332.73 2.68 0 1.07 0.12 0 0.05 %TSS 99.47 0.16 0 0.32 0.03 0 0.02 100 P-value 0 0 0.88 0 0.20 0.85

DF, degrees of freedom; SS, sum of squares; MS, mean square; TSS, total sum of squares.

of rigid or non-rigid connected toothimplant FPD in axial and oblique loadings. For toothimplant-supported FPD, to determine the relative importance of the investigated factors and their interactions, ANOVAs were used and are shown in Tables 46 for implant, alveolar bone and prosthesis stresses, respectively. The main effect of the loading types, number of splinted teeth and connector types at each level for maximal stress in implant, alveolar bone and prosthesis can be found in Fig. 2. The loading condition for the splinting system was the main factor affecting stress development in the implant, alveolar bone and prosthesis. The results showed that the loading condition signicantly (Po0.05) determined the magnitude of the stress values and the percentage contributions were 98.42%, 99.47% and 60.57% for the implant, bone and prosthesis, respectively (Tables 46). In general, lateral occlusal forces (loading types 2, 4 and 6) obviously increased the stress values

112 |

Clin. Oral Impl. Res. 19, 2008 / 107117

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

Lin et al . Biomechanical interactions in toothimplant splinting system

Table 6. Summary of the analysis of variance (ANOVA) showed the statistical results of maximum stress with respect to prosthesis
Source Loading type Splinting situation Connector type Loading type Splinting situation Loading type Connector type Splinting situation Connector type Error Total DF 5 1 1 5 5 1 5 23 SS 962.33 1.86 575.66 3.45 33.40 0.81 11.18 1588.69 MS 192.47 1.86 575.66 0.69 6.68 0.81 2.24 %TSS 60.57 0.12 36.23 0.22 2.10 0.05 0.70 100 P-value 0 0.40 0 0.89 0.12 0.57

DF, degrees of freedom; SS, sum of squares; MS, mean square; TSS, total sum of squares.

relative to the axial analogs (loading types 1, 3 and 5) (Fig. 2a). When the occlusal forces acting on the pontic were adjusted so that they were less than those for other areas (proportion 1 : 5, loading types 5 and 6), the values for sI, max, sAB, max and sP, max decreased obviously. The maximum stress for loading types 5 and 6 was 2669% lower than for loading types 1 and 2, regardless of the connector used and the number of splinted teeth. The factor of splinting situation was found to inuence the stress values in bone signicantly (Po0.05); however, the percentage contribution was only 0.16% (Table 5). The main effect plot also showed that the number of splinted teeth did not affect the stress values for the implant, alveolar bone and prosthesis (Fig. 2b). For the connector used, the ANOVA results showed that it did signicantly (Po0.05) affect the stress values in prosthesis, and the percentage contribution was 36.23% (Table 6). The main effect plot showed that the type of connector appeared not to inuence the implant and alveolar bone and use of non-rigid connectors increased the average stress on the prosthesis more than 3.4-fold (238 vs. 69 MPa) relative to the rigid connectors (Fig. 2c). In addition, the cross-interaction of loading type and splinting situation was also found to be a signicant factor affecting the stress value in bone but with a slight percentage contribution of 0.32%. With respect to stress distribution, the sI, max positions were located at the mesial buttjoint interface between the abutment and implant for all simulated models, with sI, max values over 1000 MPa when uniform multiple oblique forces (loading type 2) acted on the splinting system (Fig. 3a and b). The sAB, max positions were asso-

ciated with the loading type, and were located on the lingual cervical areas in the cortical bone of the natural teeth and implant side where occlusal forces acted only on the premolar (loading types 3 and 4) and the entire splinting system (loading types 1, 2, 5 and 6), respectively (Fig. 3c and d). Depending on the specic connectors used, the sP, max positions were located on the concave surface of the male component and bottom area for non-rigid and rigid connectors, respectively (Fig. 3e and f). In order to understand the effects of the various connector and splinting-teeth options in terms of their effectiveness of compensating for the dissimilar mobility of the natural teeth and implant, the ratios (RTID) for average teeth displacements with healthy periodontal ligaments in the mesio-distal, bucco-lingual and occlusal directions were compared with those of an osseointegrated dental implant (Table 7). Comparing the ratios of the non-rigid and rigid connections (S1NR vs. S1R or S2NR vs. S2R), those for the former were obviously larger (S1NR4S1R or S2NR4 S2R) under loading types 1, 3 and 5 (vertical occlusal forces). Their variations were greater than 23% for loading types 1, 3 and 5 but less than 9% for loading types 2, 4 and 6. Comparing the different splinting situations (S1NR vs. S2NR and S1R vs. S2R), the values of RTID were obviously larger with single-tooth splinting relative to the dual-tooth variant under loading types 2, 4 and 6 (oblique occlusal forces), with the corresponding variations higher than 30%.

Discussion
For free-standing implant-supported restorations, all the stress values of implant,

prosthesis and alveolar bone are much smaller than those of rigid or tooth-implant FPD subjected to simulate axial and oblique loadings was found in present FEA study. Not only does it prevent possible tooth intrusion and other complications, solely implant-supported FPDs also yield a better stress distribution under functional loading. Hence, it is suggested as a primary treatment option in a distal free-end partial edentulous situation. Although a series of long-term clinical reports have involved a large number of patients treated with tooth-to-implant connected prostheses, controversies still exist concerning the type of connection between the teeth and the implants. In 2001, Naert et al. (2001b) addressed a dilemmatic conclusion that a rigid connector could prevent intrusion of abutment teeth but signicantly greater bone loss could be observed around the connected implants. These results suggest that other factors such as occlusal forces, number of splinted teeth and condition of the periodontal support may play a primary role in inuencing the biomechanical complications of the implantteeth (splinting) system. Unfortunately, biomechanical investigation is problematic where only clinical or experimental approaches are used with limited information and sample variations. In many biomechanical investigations, therefore, FE analysis (FEA) has generally been accepted as a signicant complementary tool to enhance understanding of the detailed mechanical responses. In order to investigate the mechanical interactions of splinted systems involving multi-factorial parameters (loading conditions, number of splinted teeth and connector type) with different levels, three-dimensional (3D) FE analysis and reasonable interfacial conditions were undertaken to simulate the relative micro-motion between different components and the discontinuous stress transfer. In our previous work (Lin & Wang 2003), non-linear contact analysis was validated as a more realistic interfacial condition for simulation of the compensative mechanism within the implant system and keyway sliding function and, thus, it was used for all the simulations in this study. Further, full factorial analysis permitted exploration of every possible level-factor combination, with the main effect on maximum von-Mises stress also

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

113 |

Clin. Oral Impl. Res. 19, 2008 / 107117

Lin et al . Biomechanical interactions in toothimplant splinting system

a
MPa
0 134.111 268.222 402.333 536.444 670.556 804.667 938.778 1073 1207

b
MPa
0 134.111 268.222 402.333 536.444 670.556 804.667 938.778 1073 1207

d
MPa MPa
1.311 17.32 33.33 49.339 65.348 81.358 97.367 113.377 129.386 145.396 1.311 17.32 33.33 49.339 65.348 81.358 97.367 113.377 129.386 145.396

e
MPa
0 44.444 88.889 133.333 177.778 222.222 266.667 311.111 355.556 > 400

f
MPa
0 44.444 88.889 133.333 177.778 222.222 266.667 311.111 355.556 > 400

Fig. 3. (a, b) von Mises stress distribution of the abutment and implant in the splinting system under loading type 2 with a rigid connector and splinting two teeth (model 8). The positions of sI, max were found at the mesial butt joint interface between the abutment and implant. (c) von Mises stress distribution of the cortex bone in the splinting system under loading type 4 with a rigid connector and splinting two teeth (model 16). The positions of sAB, max were found on the lingual cervical area at the premolar side. (d) von Mises stress distribution of the cortex bone in the splinting system under loading type 2 with a rigid connector and splinting two teeth (model 8). The positions of sAB, max were found on the lingual neck area at the implant side. (e) von Mises stress distribution of the prosthesis in the splinting system under loading type 2 with a non-rigid connector and splinting two teeth (model 7). The positions of sP, max were found at the concave surface of the male component. (f) von Mises stress distribution of the prosthesis in the splinting system under loading type 2 with a rigid connector and splinting two teeth (model 8). The positions of sP, max were at the bottom area of the distal connector.

computed for each level of the three investigated factors to elucidate the level interactions. It was found in a series studies that the force distribution depends not only on the occlusal contact pattern and the length of the cantilever beam but also the type of prosthetic reconstruction in the opposite jaw occluding with the cantilever segment (Falk et al. 1989, 1990; Lundgren et al. 1989). While the magnitude of the loading

forces and the contact locations may alter the mechanical responses in an implanttooth splinting system, the six loading types modeled in this study can only simulate some of the possibilities of axial and oblique occlusal contacts encountered in clinical situations. Loading types 1 and 2 represent uniform multiple vertical and lateral forces acting on an implanttoothsupported prosthesis adjusted to even contact in maximal intercuspation and lateral

working positions, respectively. Loading types 3 and 4 simulate natural teeth subjected to centric and lateral contact. Loading types 5 and 6 were used to compare the changes in biomechanical behavior, simulating a relatively light load magnitude on the pontics in the centric and lateral contacts following selective grinding procedures. The results of the simulations indicate that, compared with connector type and

114 |

Clin. Oral Impl. Res. 19, 2008 / 107117

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

Lin et al . Biomechanical interactions in toothimplant splinting system

Table 7. Contact-simulated results of the ratios of average tooth displacements with healthy periodontal ligaments to those of an osseointegrated dental implant (RTID) for all models and variations between non-rigid/rigid connectors or one/two teeth splinting
Loadings Ratio (RTID) of average tooth-to-implant displacement S1NR 1 2 3 4 5 6 3.46 2.35 16.58 19.95 2.8 2.2 S1R 2.03 2.22 12.5 18.88 1.92 2 S2NR 3.09 1.45 15 12.19 2.4 1.5 S2R 1.8 1.36 11 11.14 1.86 1.41 Ratio variation between non-rigid/ rigid connectors Ratio variation between one/ two teeth splinting

(S1NR-S1R)/ (S2NR-S2R)/ (S1NR-S2NR)/ (S1R-S2R)/ S1NR S2NR S1NR S1R 0.41 0.06 0.25 0.05 0.31 0.09 0.42 0.06 0.27 0.09 0.23 0.06 0.11 0.38 0.10 0.39 0.14 0.32 0.11 0.39 0.12 0.41 0.03 0.30

S1NR, Splinting one tooth with a non-rigid connector; S1R, Splinting one tooth with a rigid connector; S2NR, Splinting two teeth with a non-rigid connector; S2R, Splinting two teeth with a rigid connector.

number of the splinted teeth, loading type (both in terms of occlusal force and contact location) had greater percentage contributions on stresses on the implant, alveolar bone and prosthesis (Tables 46). Oblique occlusal forces produce a lateral bending moment that signicantly increases stress when compared with axial occlusal forces (loading types 2, 4, 6 and 1, 3, 5, respectively; Table 3). When the occlusal forces acting on the pontic were reduced (loading types 5 and 6) to reevaluate the stress distribution in the splinting system, the maximal stress values for the implant, alveolar bone and prosthesis decreased 2669% (compared with loading types 1 and 2) owing to the smaller bending moment acting on the splinted system. This result is in agreement with a previous report (Falk et al. 1990) that the occlusal design had a major inuence on the difference force distribution between a cantilever segment and implant-supported area. It is implied that serial occlusal adjustments should be performed to obtain an evenly distributed load between implant and teeth for tooth-implant-supported FPDs under heavy bites. A lighter contact on the pontic may be more benecial for an implanttooth-supported FPD-bearing axial or oblique loading in terms of preventing the potential overloading on implant. With regard to the inuence of the number of splinted teeth, we found that addition of an extra tooth had no effect on the stress values in the toothimplant-supported FPD. This result indicates that the use of double periodontal healthy abutment

teeth combined with an implant in clinical situation routinely is not justied. The validity of applying this double abutted concept to teeth with reduced periodontal support needs to be further investigated. In terms of the effect of the connector used, the ANOVA results showed that it was also the primary factor (36.23%) affecting the stress values in prosthesis. The main effect plot showed that the maximal von Mises stresses on the prosthesis (sP, max) for non-rigid connections were increased more than 3.4-fold compared with the rigid analogues. This was due to the transfer of occlusal forces from the prosthesis to the alveolar bone by implant abutment only with the non-rigid connector, which acted as a stress breaker within the splinted system. Conversely, stress may be transferred from the prosthesis to the alveolar bone through both the natural teeth and the implant when a rigid connector (one piece) is used in the implanttooth splinted system. The maximal stress values were concentrated within the implant system at the contact buttjoint interface between the abutment and implant xture and the base of the internal hexagon joint of the abutment. (Fig. 3a and b) Our results indicate that potential engineering problems, such as xtureabutment interface failure, screw loosening and fracture may occur in a toothimplant-supported FPD after long-term clinical service. The largest stresses on the alveolar bone occurred at the lingual side around the implant neck due to the primary bending moment acting on these areas when occlusal forces

attempt to distort the entire splinting system (loading types 1, 2, 5 and 6; Fig. 3d). In addition, with the non-rigid connectors stress concentrations were identied at the concave surface of the male component due to geometrical complications (Fig. 3e). With the rigid connection, the corresponding concentrations were located at the base of the distal connectors as this region was a branch point for the prosthesis (Fig. 3f). In both connector types, these stress concentrations may profoundly increase the risk of failure after long-term dynamic loading. Based on the concept of compensating for the dissimilar mobility between the natural tooth and implant, RTID values for all models, and the variations between nonrigid/rigid connectors and one/two teeth splinting were calculated to determine variations in the effectiveness of different connectors used and splinting situations. Where the splinting-system RTID values were higher with the non-rigid connectors compared with the rigid connectors (S1NR4S1R or S2NR4S2R), it indicated successful compensation for the dissimilar mobility of the natural tooth and implant side with the key and keyway sliding function. This phenomenon was obvious with loading types 1, 3 and 5 (vertical occlusal forces), with the higher variations (423%) also reecting the greater degree of compensation for the dissimilarities in mobility achieved with the non-rigid connector. However, small differences in RTID and corresponding variations were found with oblique occlusal forces (loading types 2, 4 and 6). This indicates that the non-rigid connector is inefcient in terms of compensating for the variations in mobility between the natural teeth and implant due to the effects of mechanical binding, where the sliding function is constrained axially. Furthermore, the lower RTID for the two-tooth splinting system relative to the one-tooth variant (S2NRoS1NR and S2RoS1R) indicates that additional splinting may exploit its function. This phenomenon was more obvious for loading types 2, 4 and 6, with the corresponding variations above 30%. It seems that splinting with an extra tooth may provide a greater extent of compensation for the dissimilarities in the relative mobility of teeth and implant when the system is subjected to oblique forces.

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

115 |

Clin. Oral Impl. Res. 19, 2008 / 107117

Lin et al . Biomechanical interactions in toothimplant splinting system

3D non-linear FEA was applied in this study to investigate the basic mechanical toothimplant-supported system interactions under various occlusal forces with rigid/non-rigid connectors and for different splinting situations. However, the current investigation was certainly limited by the theoretical assumptions, such as loading conditions, material properties and geometry of the alveolar bone, which probably affected the accuracy of the results obtained with respect to the mechanical responses and stress states. As mentioned previously, loading types 14 are simple representations of possible occlusal contact encountered in clinical situations. Loading types 5 and 6 (reducing axial and oblique occlusal forces on the pontic) were simulated to compare the inuences on biomechanical behavior in the splinting system. Linear elastic (homogeneous and isotropic) properties were adopted for all materials due to numerical-convergence considerations and the larger variance in physical properties noted in the literature, especially the PDL characterization (Yoshida et al. 2001; Toms et al. 2002). In addition, only simplied bony segments were modeled for parameter studies of the splinting system; the bending of the mandible during mastication has not been taken into account. Therefore, the FEA modeling results provide only a general insight into the biomechanical aspects of the splinting system under average conditions. Because some parameters that may inuence the clinical situation are excluded, further long-term clinical studies to examine correlation between stress analysis and clinical outcome of these

tooth-implant-supported are needed.

reconstructions

Conclusions
The following conclusions were derived from outcomes of the current extensive 3D non-linear numerical analyses: 1. For free-standing implant-supported restorations, all the stress values of implant, prosthesis and alveolar bone are smaller than those of rigid or nonrigid connected toothimplant FPD in simulated axial and oblique loadings. Loading condition is the main factor signicantly affecting stress distribution in the implant, alveolar bone and prosthesis of a toothimplant-supported FPD in comparison with connector type and number of splinted teeth. Selective grinding procedures can reduce the cantilever effect and redistribute stress in the maximum intercuspation or lateral working position for a tooth implant-supported prosthesis. Use of a non-rigid connector may be more efcient in terms of compensation for the dissimilar mobility between the natural teeth and implant under axial loading forces. However, it should be used with caution as it breaks the stress transfer and increases the unfavorable stress values in prostheses.

2.

3.

4.

Acknowledgements: This research was supported by grant NSC93-2213-E182-009 from the National Science Council, Taiwan.

References
Becker, C.M., Kaiser, D.A. & Jones, J.D. (2000) Guidelines for splinting implants. Journal of Prosthetic Dentistry 84: 210214. Benzing, U.R., Gall, H. & Weber, H. (1995) Biomechanical aspects of two different implant-prosthetic concepts for edentulous maxillae. International Journal of Oral & Maxillofacial Implants 10: 188198. Cehreli, M.C., Sahin, S. & Akca, K. (2004) Role of mechanical environment and implant design on bone tissue differentiation: current knowledge and future contexts. Journal of Dentistry 32: 123132. Dar, F.H., Meakin, J.R. & Aspden, R.M. (2002) Statistical methods in nite element analysis. Journal of Biomechanics 35: 11551161. Falk, H., Laurell, L. & Lundgren, D. (1989) Occlusal force pattern in dentitions with mandibular implant-supported xed cantilever prostheses occluding with complete dentures. International Journal of Oral & Maxillofacial Implants 4: 5562. Falk, H., Laurell, L. & Lundgren, D. (1990) Occlusal interferences and cantilever joint stress in implant-supported prostheses occluding with complete dentures. International Journal of Oral & Maxillofacial Implants 5: 7077. Goodacre, C.J. & Kan, J.Y.K. (1999) Clinical complications of osseointegrated implants. Journal of Prosthetic Dentistry 81: 537552. Gross, M. & Laufer, B.Z. (1997) Splinting osseointegrated implants and natural teeth in rehabilitation of partially edentulous patients. Part I: laboratory and clinical studies. Journal of Oral Rehabilitation 24: 863870. Ko, C.C., Chu, C.S., Chung, K.H. & Lee, M.C. (1992) Effect of posts on dentin stress distribution in pulpless teeth. Journal of Prosthetic Dentistry 67: 421427.

116 |

Clin. Oral Impl. Res. 19, 2008 / 107117

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

Lin et al . Biomechanical interactions in toothimplant splinting system

Lang, N.P., Pjetursson, B.E., Tan, K., Bragger, U., Egger, M. & Zwahlen, M. (2004) A systematic review of the survival and complication rates of xed partial dentures (FPDs) after an observation period of at least 5 years II. Combined toothimplant-supported FPDs.. Clinical Oral Implants Research 15: 643653. Laufer, B.Z. & Gross, M. (1998) Splinting osseointegrated implants and natural teeth in rehabilitation of partially edentulous patients. Part II: principles and applications. Journal of Oral Rehabilitation 25: 6980. Lin, C.L., Chang, C.H., Cheng, C.S., Wang, C.H. & Lee, H.E. (1999) Automated nite element mesh generation for maxillary second premolar. Computer Methods and Programs in Biomedicine 59: 187195. Lin, C.L., Chang, C.H. & Ko, C.C. (2001) Multifactorial analysis of an MOD restored human premolar using Auto-Mesh nite element approach. Journal of Oral Rehabilitation 28: 576 585. Lin, C.L. & Wang, J.C. (2003) Nonlinear nite element analysis of a splinted implant with various connectors and occlusal forces. International Journal of Oral & Maxillofacial Implants 18: 331340. Lundgren, D., Falk, H. & Laurell, L. (1989) Inuence of number and distribution of occlusal cantilever contacts on closing and chewing forces in dentitions with implant-supported xed prostheses occluding with complete dentures. International Journal of Oral & Maxillofacial Implants 4: 277283. Mathews, M.F., Breeding, L.C., Dixon, D.L. & Aquilino, S.A. (1991) The effect of connector design on cement retention in an implant and natural tooth-supported xed partial denture. Journal of Prosthetic Dentistry 65: 822827. McGlumphy, E.A., Campagni, W.V. & Peterson, L.J. (1989) A comparison of the stress transfer characteristics of a dental implant with a rigid or a resilient internal element. Journal of Prosthetic Dentistry 62: 586593. Menicucci, G., Mossolov, A., Mozzati, M., Lorenzetti, M. & Preti, G. (2002) Tooth-implant connection: some biomechanical aspects based on nite element analysis. Clinical Oral Implants Research 13: 334344. Merz, B.R., Hunenbart, S. & Belser, U.C. (2000) Mechanics of the implant-abutment connection: an 8-degree taper compared to a butt joint connection. International Journal of Oral & Maxillofacial Implants 15: 519526. Middleton, J., Jones, M. & Wilson, A. (1996) The role of the periodontal ligament in bone modeling:

the initial development of a time-dependent nite element model. American Journal of Orthodontics and Dentofacial Orthopedics 109: 155162. Moffa, J.P., Lugassy, A.A., Guckes, A.D. & Gettleman, L. (1973) An evaluation of nonprecious alloys for use with porcelain veneers. Part I. Physical properties. Journal Prosthetic Dentistry 30: 424431. Naert, I.E., Ducky, J.A.J., Honsy, M.M.F., Quirynen, M. & van Steenberghe, D. (2001b) Freestanding and tooth-implant connected prostheses in the treatment of partially edentulous patients. Part II: an up to 15-years radiographic evaluation. Clinical Oral Implants Research 12: 245251. Naert, I.E., Ducky, J.A.J., Honsy, M.M.F. & van Steenberghe, D. (2001a) Freestanding and toothimplant connected prostheses in the treatment of partially edentulous patients. Part I: an up to 15years clinical evaluation. Clinical Oral Implants Research 12: 237244. Nishimura, R.D., Ochiai, K.T., Caputo, A.A. & Jeong, C.M. (1999) Photoelastic stress analysis of load transfer to implants and natural teeth comparing rigid and semirigid connectors. Journal of Prosthetic Dentistry 81: 696703. Nyman, S. & Lang, N.P. (1994) Tooth mobility and the biological rationale for splinting teeth. Periodontology 2000 4: 1522. Pesun, I.J. (1997) Intrusion of teeth in the combination implant-to-natural-tooth xed partial denture: a review of the theories. Journal of Prosthodontics 6: 268277. Pjetursson, B.E., Tan, k., Lang, N.P., Bragger, U., Egger, M. & Zwahlen, M. (2004) A systematic review of the survival and complication rates of implant supported xed partial dentures (FPDs) after an observation period of at least 5 years I. Implant-supported FPDs. Clinical Oral Implants Research 15: 625642. Rieder, C.E. & Parel, S.M. (1993) A survey of natural tooth abutment intrusion with implantconnected xed partial dentures. International Journal of Periodontics and Restorative Dentistry 13: 334347. Sahin, S., Cehreli, M.C. & Yalcin, E. (2002) The inuence of functional forces on the biomechanics of implant-supported prostheses-a review. Journal of Dentistry 30: 271282. Sekine, H., Komiyama, Y., Hotta, H. & Yoshida, K. (1986) Molility characteristics and tactile sensitivity of osseointegrated xture-supporting system. In: van Steenberghe, D., ed. Tissue Integration in Maxillofacial Reconstruction, 326332. Amsterdam: Excerpta Medica.

Sheets, C.G. & Earthman, J.C. (1997) Tooth intrusion in implant-assisted prostheses. Journal of Prosthetic Dentistry 77: 3945. Stegaroiu, R., Kusakari, H., Nishiyama, S. & Miyakawa, O. (1998) Inuence of prosthesis material on stress distribution in bone and implant: a 3dimensional nite element analysis. International Journal of Oral & Maxillofacial Implants 13: 781790. Tada, S., Stegaroiu, R., Kitamurs, E., Miyakawa, O. & Kusakari, H. (2003) Inuence of implant design and bone quality on stress/strain distribution in bone around implants: a 3-dimensional nite element analysis. International Journal of Oral & Maxillofacial Implants 18: 357368. Tangerud, T., Gronningsater, A.G. & Taylor, A. (2002) Fixed partial dentures supported by natural teeth and Branemark system implants: a 3-year report. International Journal of Oral & Maxillofacial Implants 17: 212219. Toms, S.R., Dakin, G.J., Lemons, J.E. & Eberhardt, A.W. (2002) Quasi-linear viscoelastic behavior of the human periodontal ligament. Journal of Biomechanics 35: 14111415. Van Oosterwyck, H., Duyck, J., Vander Sloten, J., Van der Perre, G., De Cooman, M., Lievens, S., Puers, R. & Naert, I. (1998) The inuence of bone mechanical properties and implant xation upon bone loading around oral implants. Clinical Oral Implants Research 9: 407418. Van Rossen, I.P., Braak, L.H., de Putter, C. & de Groot, K. (1990) Stress absorbing elements in dental implants. Journal of Prosthetic Dentistry 64: 198205. Wang, L., Sadler, J.P., Breeding, L.C. & Dixon, D.L. (1999) An in vitro study of implant toothsupported connections using a robot test system. Journal of Biomechanical Engineering 121: 290 297. Wylie, R.S. & Caputo, A.A. (1991) Fixed cantilever splints on teeth with normal and reduced periodontal support. Journal of Prosthetic Dentistry 66: 737742. Yang, H.S., Chung, H.J. & Park, Y.J. (1996) Stress analysis of a cantilevered xed partial denture with normal and reduced bone support. Journal of Prosthetic Dentistry 76: 424430. Yang, H.S., Lang, L.A. & Felton, D.A. (1999) Finite element stress analysis on the effect of splinting in xed partial dentures. Journal of Prosthetic Dentistry 81: 721728. Yoshida, N., Koga, Y., Peng, C.L., Tanaka, E. & Kobayashi, K. (2001) In vivo measurement of the elastic modulus of the human periodontal ligament. Medical Engineering Physics 23: 567 557.

c 2007 The Authors. Journal compilation c 2007 Blackwell Munksgaard

117 |

Clin. Oral Impl. Res. 19, 2008 / 107117