Journal of

Oral Rehabilitation

Journal of Oral Rehabilitation 2010 37; 658662

Analysis of load transfer and stress distribution by splinted and unsplinted implant-supported xed cemented restorations
J. NISSAN*, O. GHELFAN*, M. GROSS* & G. CHAUSHU

Departments of *Oral Rehabilitation and

Oral and Maxillofacial Surgery, The Maurice and Gabriela Goldschleger School of Dental Medicine, Tel Aviv University, Tel Aviv, Israel

Controversy remains over the rehabilitation of implant-supported restorations regarding the need to splint adjacent implant-supported crowns. This study compared the effects of simulated occlusal loading of three implants restored with cemented crowns, splinted versus unsplinted. Three adjacent screw-shaped implants were passively inserted into three holes drilled in a photo-elastic model. Two combinations of cemented restorations were fabricated; three adjacent unsplinted and three adjacent splinted crowns. Strain gauges were connected to the implant necks and to the margins of the overlaying crowns. Fifteen axial static loads of 20-kg loadings were carried out right after each other via a custombuilt loading apparatus. Strain gauges located on the implant neck supporting splinted restoration
SUMMARY

demonstrated signicantly (P < 0001) more strain (sum of strains = 334854 microstrain) compared with the single crowns (sum of strains = 98857 microstrain). In contrast, signicantly (P < 0001) more strain was recorded on the strain gauges located on the restoration margins of the single crowns (sum of strains = 75632 microstrain) when compared with splinted restorations (sum of strains = 18612 microstrain). The concept of splinting adjacent implants to decrease loading of the supporting structures may require re-evaluation. The clinical relevance of these ndings needs further investigation. KEYWORDS: load, stress distribution, splinted, unsplinted Accepted for publication 19 March 2010

Introduction
Occlusal load and its distribution is considered to be one of the principal components that inuences the success and failure of implant-supported restorations over time (14). The traditional rationale for splinting teeth was to increase retention and resistance resulting in decreased stresses, improving prosthesis stability (5). The rationale of splinting in implant dentistry to minimize stress by increasing the resistance area over, which the load is distributed, is controversial (6). The biomechanical advantages following splinting restorations are still unclear (7). Evidence-based data to support splinting are largely missing for teeth and even more for implants (7, 8).
2010 Blackwell Publishing Ltd

Some authors have maintained that occlusal loads transferred to implants supporting splinted restoration are larger than those applied to implants supporting unsplinted restorations because of the development of moments (6, 9). Clinical studies on the successful restoration of unsplinted adjacent implant-supported restorations in partially edentulous individuals have been reported (1014). Implant splinting did not signicantly improve implant success rates for implant-supported xed partial dentures (971%) compared to single-implant restorations (943%) (14). Furthermore, splinting did not have an effect on crestal bone loss (15). This study examined load transfer and stress distribution of simulated axial occlusal loading on adjacent
doi: 10.1111/j.1365-2842.2010.02096.x

LOAD TRANSFER AND STRESS BY SPLINTING AND UNSPLINTING

implant-supported xed restorations. The effects on splinted and unsplinted implant-supported restorations are compared using strain gauge analysis. the restorations were fabricated on attached xed abutments with 2 mm gingival height. Two groups of restorations were cast in Remanium CoCrMo Model Casting Alloy-GM 380: 1 Three unsplinted crowns (Fig. 1). 2 Three splinted crowns (Fig. 2). The restorations were fabricated with the occlusal anatomy of upper rst molars with the same mesiodistal and bucco-lingual dimensions using a silicone index. Screw-retained abutments were placed at a controlled torque of 35 Ncm. During each loading session, the restorations were cemented onto the implant abutments using temporary cement. No residual stress was apparent in the photo-elastic model as veried by visual inspection. Contact points of the individual restorations were fabricated so that dental oss passed with slight difculty according to standard clinical procedure. An additional third strain gauge was cemented horizontally onto each cast restoration at the cervical margin parallel to the margin in the mid-buccal dimension (EA-06-032DE-350*). These were designed to measure the peripheral strain in the margins of each casting. Fifteen static loadings were carried out right after each other with 20 kg weights via a custom-built loading apparatus. Load was applied simultaneously through three individual pins to the inner inclines of the buccal cusps of each set of restorations at 0 to the vertical axis (Fig. 3).

659

Materials and methods

A photo-elastic block model (PLM-4B*) with modulus of elasticity 450 ksi [range of human bone (16)] was constructed. The model dimensions were 158 352 32 mm. Three holes were drilled vertically in a straight line in the mid-axis of the photo-elastic model at predetermined locations to lengths of 12 mm. Drilling was carried out according to surgical protocols with successive drill diameters in sequence to minimize residual stresses in the model. Stresses introduced into the photo-elastic model by the drilling process were relieved by placement of the model in an oven on a teon surface for 120 min at 70 C. The model was cooled in the closed oven. Stress relief was veried with the aid of a circular polariscope, and the model was found to be stress free. Implants were located 78 mm from the edge of the model and each implant separated by 4 mm. Three external hex, screw type titanium implants of diameter 38 mm, of length 12 mm were inserted into the model. The 2-mm implants neck protruded forms the superior surface. Two strain gauges (Strain-gauge EA-06-015EH-120*) were cemented (M-Bond 200*) horizontally onto the neck of each implant on the buccal and lingual aspects at a 180-inclination to each other prior to abutment and crown placement. These strain gauges measured the bending components (tension compression) created from both vertical and horizontal vectors arising from applied forces. Strain gauges were connected to a strain indicator (Strain Indicator System 5000*) that provided a simultaneous direct reading of strain in microstrain units of all model components for each loading session. The described design has been used in previous studies (17). Impressions were taken by the open tray technique with acrylic splinted transfer copings using custom acrylic trays. Polyether impression material was used. A master working model was fabricated on which all

*Vishay Measurement Group Inc., Raleigh, NC, USA. Nobel Biocare, Zurich, Switzerland. Duralay Reliance Dental Mfg Co., Worth, IL, USA. Impregum F; ESPE, Seefeld, Germany. 2010 Blackwell Publishing Ltd

Fig. 1. Unsplinted crowns.

Temp Bond, NE Kerr, CA, USA.

660

J . N I S S A N et al.
were compared by the use of the one-way parametric analysis of variance (ANOVA). P values of <005 were considered statistically signicant.

Results
Strain gauges located on the implant neck supporting splinted restoration demonstrated signicantly (P < 0001) more strain (sum of strains = 334854 microstrain) compared with the single crowns (sum of strains = 98857 microstrain) (Table 1). In contrast, signicantly (P < 0001) more strain was recorded on the strain gauges located on the restoration margins of the single crowns (sum of strains = 75632 microstrain) when compared with splinted restorations (sum of strains = 18612 microstrain) (Table 2).

Fig. 2. Splinted crowns.

Table 1. Microstrain values on implant necks Microstrain values Implant 1 Strain Gauge location B* PNS 2 BNS P* 3 B* P* Fig. 3. A custom-built loading apparatus. Restoration modality Single Splint Single Splint Single Splint Single Splint Single Splint Single Splint M 16264 188440 1027 4840 4853 3113 52920 96501 11713 23360 12100 18600 SD 6941 5962 162 486 266 125 421 1928 396 401 720 1177

NS, Not Signicant; B, Buccal; P, Palatinal. *P < 0.001.

For each loading, strain gauge recordings were made. Strain gauges measure electrical resistance. During extension or contraction, the strain gauge records changes in electrical resistance. The degree of distortion of the strain gauge is recorded in calculated microstrains values, where strain = = DL (change in length of the strain gauge) L (lm m) = DR (change in electrical resistance in the strain gauge) R. Statistical analysis Descriptive analysis consisted of mean and standard deviation of microstrain values for each group. Groups

Table 2. Microstrain values on crown margins Microstrain values Crown 1* 2* 3* Restoration modalities Single Splint Single Splint Single Splint Mean 13426 8700 40746 2126 21460 7786 SD 116 633 350 103 1123 091

*P < 0.001.
2010 Blackwell Publishing Ltd

LOAD TRANSFER AND STRESS BY SPLINTING AND UNSPLINTING

unsplinted restorations, signicantly (P < 0001) less stress was generated in the implant neck when compared to splinted restorations (98857 versus 334854 microstrain). For each single restoration, there is inherent inaccuracies because of component mist (crown abutment and abutment implant) resulting in preload stresses. When several adjacent implant restored crowns are joined, there is a summation of these mist inaccuracies, and signicantly increased moments because of splinting, resulting in transfer of increased loads to the implants and supporting structures (6). This can also explain the disparity of microstrain values on implant necks exhibited between the implants. In contrast, signicantly (P < 0001) more strain was recorded on the strain gauges located on the restoration margins of the single crowns when compared with splinted restorations (75632 versus 18612 microstrain). Therefore, cemented splinted restorations exhibited better load sharing than non-splinted storations, however, they transferred more forces towards the implant neck because of bending moments. In addition, in this study, implants were loaded in a vertical inclination, while in the clinical setup, nonaxial loads are also generated. Moreover, occlusal contacts are most often lateral (lateral function and parafunction, and asymmetric contraction of the jaw closing muscles combine). As a result, non-axial loading is generally the rule. These factors will all combine to greatly increase the bending moments observed in this study. Consequently, additional moments yield signicantly higher strain values in the implant neck and restorations than seen in the study. The present discusses axial loading of splinted and unsplinted implant-supported restorations. Future studies should asses whether same results will be still obtained following non-axial loading. Within the limitations of this experimental model, the following conclusions may be drawn: 1 Single unsplinted restorations transfer signicantly less load to the implants and supporting structure than splinted restorations. 2 Splinted restorations transfer signicantly less load to the crown margins than unsplinted restorations. 3 The concept of splinting adjacent implants to decrease loading of the supporting structures may require re-evaluation. 4 The clinical relevance needs to be investigated with controlled long-term clinical studies.

661

Discussion
Occlusal loads on osseointegrated implants are cited as a signicant factor in the long-term success of implantsupported restorations (1, 6). It is common clinical practice to join adjacent implant-supported restorations in the restoration of the partially edentulous. Resistance and retention forms may be a major indication for splinting. An additional rationale of splinting implant crowns together is to favourably distribute the nonaxial loads, minimize their transfer to restoration and the supporting bone and to increase the total load area (18). This practice is taken from concepts of splinting teeth, where the assumption is that joined linear and non-collinear units improve the collective resistance to forces and alters the centre of rotation of the joined units (19). Several in vitro studies reported conicting results. Guichet et al. (20) in a 3D photo-elastic study support this concept reporting that cemented splinted restorations exhibited better load sharing than non-splinted restorations. Brunsky et al. (6) maintained that loading of splinted implant-supported crowns generates moments resulting in greater forces on the implants when compared to the applied force. Kim et al. (21) compared provisional and permanent cement retained, and screw-retained 2-unit splinted restorations using a photo-elastic and strain gauge bench model. A single provisionally cemented restoration showed the least stress compared to splinted and cantilevered modalities. On the contrary, clinical studies do not seem to support splinting. Glantz et al. (9) reported on unexpectedly high functional bending moments on the implants in vivo, on maximum biting and chewing in a conventional cross arch splinted restoration. Bender (10) in a 4-year clinical study reported higher success rates for adjacent unsplinted cemented restorations when compared to adjacent splinted cemented restorations. He maintains that non-splinted restorations allow the optimal transfer of stress to the supporting structures. In another clinical study consisting of 199 implants and 74 partially edentulous patients, splinted implants showed greater crestal bone loss (02 mm more) than non-splinted ones. These differences were statistically signicant. They concluded that splinted implants appeared to favour greater crestal bone loss (22). This study compared cemented single versus splinted congurations. The results showed that in single
2010 Blackwell Publishing Ltd

662

J . N I S S A N et al.

References
1. Skalak R. Biomechanical consideration un osseointegrated prothesis. J Prosthet Dent. 1983;49:843848. 2. Isidor F. Loss of osseointegration caused by occlusal load of oral implants. A clinical and radiographical study in monkeys. Clin Oral Implants Res. 1996;7:143152. 3. Leung KC, Chow TW, Wat PY, Comfort MB. Peri-implant bone loss: management of a patient. Int J Oral Maxillofac Implants. 2001;16:273277. 4. Isidor F. Inuence of forces on peri-implant bone. Clin Oral Implants Res. 2006;17:818. 5. Serio FG. Clinical rationale for tooth stabilization and splinting. Dent Clin North Am. 1999;43:16. 6. Brunski JB, Puleo DA, Nanci A. Biomaterials and biomechanics of oral and maxillofacial implants : current status and future developments. Int J Oral Maxillofac implants. 2000;15:1546. 7. Grossmann Y, Finger IM, Block MS. Indications for splinting implant restorations. J Oral Maxillofac Surg. 2005;63: 16421652. 8. Gross M, Laufer BZ. Splinting osseointegrated implants and natural teeth in rehabilitation of partially edentulous patients. Part I: laboratory and clinical studies. J Oral Rehabil. 1997;24:863870. 9. Glantz PO, Rangert B, Svensson A, Stafford GD, Arnvidarson B, Randow K et al. On clinical loading of osseointegrated implants. A methodological and clinical study. Clin Oral Implants Res. 1993;4:99105. 10. Bender MF. Unsplinted crowns on implants in subantral augmented region: an evolution. J Oral Implantol. 1995; 2:121131. 11. Lindh T, Gunne J, Tillberg A, Molin M. A meta-analysis of implants in partial edentulism. Clin Oral Implants Res. 1998;9:8090. 12. Naert I, Koutsikakis G, Duyck J, Quirynen M, Jacobs R, van Steenberghe D. Biologic outcome of implant-supported restorations in the treatment of partial edentulism. part I: a longitudinal clinical evaluation. Clin Oral Implants Res. 2002;13:381389.

13. Simon RL. Single implant-supported molar and premolar crowns: a ten-year retrospective clinical report. J Prosthet Dent. 2003;90:517521. 14. Weber HP, Sukotjo C. Does the type of implant prosthesis affect outcomes in the partially edentulous patient? Int J Oral Maxillofac Implants. 2007;22(Suppl):140172. 15. Blanes RJ, Bernard JP, Blanes ZM, Belser UC. A 10-year prospective study of ITI dental implants placed in the posterior region. II: inuence of the crown-to-implant ratio and different prosthetic treatment modalities on crestal bone loss. Clin Oral Implants Res. 2007;18:707714. 16. Van Staden RC, Guan H, Loo YC. Application of the nite element method in dental implant research. Comput Methods Biomech Biomed Engin. 2006;9:257270. 17. Nissan J, Gross M, Shifman A, Assif D. Stress levels for welltting implant superstructures as a function of tightening force levels, tightening sequence, and different operators. J Prosthet Dent. 2001;86:2023. 18. Kim Y, Oh TJ, Misch CE, Wang HL. Occlusal considerations in implant therapy: clinical guidelines with biomechanical rationale. Clin Oral Implants Res. 2005;16:2635. 19. Faucher RR, Bryant RA. Bilateral xed splints. Int J Perio Rest Dent. 1983;3:837. 20. Guichet DL, Yoshinobu D, Caputo AL. Effect of splinting and interproximal contact tightness on load transfer by implant restorations. J Prosthet Dent. 2002;87:528535. 21. Kim WD, Jacobson Z, Nathanson D. In vitro stress analyses of dental implants supporting screw-retained and cementretained prostheses. Implant Dent. 1999;8:141150. 22. Rokni S, Todescan R, Watson P, Pharoah M, Adegbembo AO, Deporter D. An assessment of crown-to-root ratios with short sintered porous-surfaced implants supporting prostheses in partially edentulous patients. Int J Oral Maxillofac Implants. 2005;20:6976.
Correspondence: Dr Joseph Nissan, Department of Oral Rehabilitation, School of Dental Medicine, Tel Aviv University, Tel Aviv, Israel. E-mail: nissandr@post.tau.ac.il

2010 Blackwell Publishing Ltd