The aim of this study was to analyze the effect of the separation of prosthetic crowns from fixed partial dentures by means of stress gradient evaluation. Three photoelastic models were created to examine contiguous implants with varying contact between the crowns (contact point [CP], contact surface [CS], splinted [SP]). The SP group presented the best results, followed by the CS group, indicating that the use of splinted prosthetic crowns and crowns with broad surface contacts is viable when considering the stress values.

The rigid union of prosthetic crowns over multiple adjacent implants has been observed in clinical and in vitro studies.13  This condition was analyzed by Weinberg,4  who affirmed that the movement capacity of osseointegrated implants is microscopic and does not favor an effective occlusal force distribution to multiple implants in the same prosthesis. However, due to the elastic strain of the system components, some load transfer may be possible when splinting adjacent implants. That said, the union of crowns in a structure that does not present a passive framework fit—meaning a perfect adaptation of the prosthesis over the implant platform—may overload the screws or even the implants.5  Thus, some authors have explored the effect of splinting and the intensity of interproximal contact on the passivity of fit and load transfer.6  Photoelastic analysis has been largely applied in dentistry to study the stress distribution around dental implants.7,8  This technique takes advantage of the optical properties of materials that behave in an anisotropic manner under loading, presenting different refraction ratings in the main stress directions.9 

The aim of this study was to evaluate the effect of splinted crowns compared to separate, individual crowns of fixed partial dentures supported by implants in the posterior jaw with various contacts. We performed this investigation using qualitative and quantitative stress analyses based on the plane transmission photoelasticity technique.

Branemark system implants were used (Titamax TI Cortical, Neodent, Curitiba, Brazil) with dimensions of 13.0 mm × 3.75 mm (external hexagon interface) and a 4.1 mm platform, which corresponded to teeth #21, #20, and #19. The samples consisted of three metallic structures (Ni-Cr) that simulated fixed partial dentures over three adjacent implants, varying the interproximal contact. The contact point (CP) group consisted of three crowns separated by points of contact 1.0 mm in diameter, the contact surface (CS) group consisted of three crowns separated by surfaces of contact 3.0 mm in diameter, and the splinted (SP) group consisted of three splinted crowns. The implant spacing was 5.0 mm between the #21 and #20 implants and 5.7 mm between the #20 and #19 implants. A canine tooth (#22) was made with acrylic resin and placed into a 2.0-mm hole adjacent to the #21 implant (Figure 1).

Figures 1 and 2.

Figure 1. (a) Acrylic matrix dimensions; interproximal contact details: point contact (b), surface contact (c) and splinted (d). Figure 2. (a) Acrylic matrix fixed into the articulated box. (b) Photoelastic models contact point (CP), contact surface (CS), and splinted (SP). (c) Adjustable table of x, y, and z directions with a-type loading. (d) a-type loading. (e) b-type loading. (f) c-type loading. (g) Circular polariscope and loading cell.

Figures 1 and 2.

Figure 1. (a) Acrylic matrix dimensions; interproximal contact details: point contact (b), surface contact (c) and splinted (d). Figure 2. (a) Acrylic matrix fixed into the articulated box. (b) Photoelastic models contact point (CP), contact surface (CS), and splinted (SP). (c) Adjustable table of x, y, and z directions with a-type loading. (d) a-type loading. (e) b-type loading. (f) c-type loading. (g) Circular polariscope and loading cell.

Close modal

After the silicon rubber molds were fabricated (Silaex Química LTDA, São Paulo, Brazil) using an articulated acrylic box and an acrylic matrix, three photoelastic models were made with a flexible resin (Polipox, Indústria e Comércio LTDA, São Paulo, Brazil). After 24 hours, the curing of the resin was complete, and the photoelastic model was removed from the mold with smooth movements and taken to a circular polariscope, in the absence of residual stresses resulting from a process called the “edge effect.”9  The optical constant value of the photoelastic resin (Kσ = 0.36) was determined using a calibration process with a compressed disc made with the same photoelastic material.9 

The photoelastic resin used in this study has highly resistant but poorly sensitive to large deformations. Thus, the range of work force applied to this material is small and has low values. The load values in this test were calibrated in such a way to allow better resolution of the fringe orders and a comparative analysis of the phenomenon for the three types of implant crowns evaluated. Three loading types were applied in each model: a, axial force (30 N) directed simultaneously over all three implants; b, force (10.8 N) directed at a 40-degree angle over #19; c, axial force (9.8 N) directed over #20. For group a, three splinted crowns (Ni-Cr) were fabricated, simulating ideal occlusion,10  without premature contact and with load transfer related to the implant angle (Figure 2). Thirty (30) images of each loading type in each group were obtained (n = 270) by the circular polariscope. For each new image, the models were unloaded and loaded again, attempting to replicate the contact of the loading device on the crowns.

Fringe orders and shear stress (τ) values were calculated by Fringes software (MATLAB, MathWorks, Natick, Maine) using the computerized photoelastic analysis determined by the Optical Law of Stress9  (Equation 1).

From Equation 1 and Equation 2,9  it is possible to determine the shear stress from the measured fringe orders, namely:

where σ1 and σ2 are the principal stresses, is the photoelastic constant of the material, N is the fringe order and b is the thickness of the photoelastic model.

A mesh of 27 points of analysis was determined by an external file to standardize the regions of analysis (Figure 3a). To examine critical areas on the crest of bone, a mesh of 12 points was also made for the #19 (groups CPb, CSb, and SPb) and #20 implants (groups CPc and CSc) (Figure 3b). To identify any statistically significant differences between each of the 27 points obtained in the 30 images of the three groups and the 12 points of the crest of bone, a parametric Student t-test (P < .05) was applied. Normalized data was generated for the area under the mean shear stress (τ) for each group, called “pseudo-energy.”

Figures 3–5.

Figure 3. (a) Mesh of 27 points. (b) Mesh of 12 points—zoom of the crestal bone region. Figure 4. Representative images of stress distribution in groups contact point, a (CPa) (a), contact surface, a (CSa) (b), and splinted, a (SPa) (c). Figure 5. Mean values of shear stress (t) for a-type loading.

Figures 3–5.

Figure 3. (a) Mesh of 27 points. (b) Mesh of 12 points—zoom of the crestal bone region. Figure 4. Representative images of stress distribution in groups contact point, a (CPa) (a), contact surface, a (CSa) (b), and splinted, a (SPa) (c). Figure 5. Mean values of shear stress (t) for a-type loading.

Close modal

For group a, the stress distribution around the implants of groups CP, CS, and SP can be seen in Figure 4. There is noticeable similarity among these groups when considering the qualitative analysis. Most points when compared between types of interproximal contact were significantly different (P < .05). The points around the implants in the CSa group had lower levels of stress when compared to the CPa group (except points 18, 24, 26, and 27). For this type of loading, the implants flanking the central implant presented higher stress levels in all groups. In the SPa group, the central implant (#20) had the lowest levels of stress (33.7 kPa) (Figure 5).

Evaluation of the differences between subgroups in group a loading (CPa, CSa, and SPa) was performed, considering all points. The values were normalized by the group that obtained the lowest graph area value (SPa = 568.76) (Table 1).

Table 1

Graph area, a; normalized areas, a (axial force directed simultaneously over all three implants)

Graph area, a; normalized areas, a (axial force directed simultaneously over all three implants)
Graph area, a; normalized areas, a (axial force directed simultaneously over all three implants)

For group b, the stress distribution around the implants of groups CP, CS, and SP can be seen in Figure 6. It can be observed that the #21 and #20 implants were minimally affected in the CP and CS groups. In the SP group, there was a change in the stress distribution given that the mesial point of the #19 implant had lower stress levels (point 20 = 1.31 kPa) and there was a higher stress distribution around the #22 tooth and #21 and #20 implants. The values of shear stress at points around the #19 implant (points 21–27) were lower in the CSb group. When the CPb and CSb groups were compared individually with the SPb group, most points were significantly different (P < .000) (Figure 7). The values of the area under the stress curves were normalized by the group that received the lowest value of the graph area (SPb = 219.32) (Table 2).

Figures 6–8.

Figure 6. Representative images of stress distribution in groups contact point, b (CPb) (a), contact surface, b (CSb) (b), and splinted, b (SPb) (c). Figure 7. Mean values of shear stress (t) for b-type loading. Figure 8. Mean values of shear stress (t) of the bone crest region of #19 implant for b-type loading.

Figures 6–8.

Figure 6. Representative images of stress distribution in groups contact point, b (CPb) (a), contact surface, b (CSb) (b), and splinted, b (SPb) (c). Figure 7. Mean values of shear stress (t) for b-type loading. Figure 8. Mean values of shear stress (t) of the bone crest region of #19 implant for b-type loading.

Close modal
Table 2

Graph area, b; normalized areas, b (force directed at a 40-degree angle over #19)

Graph area, b; normalized areas, b (force directed at a 40-degree angle over #19)
Graph area, b; normalized areas, b (force directed at a 40-degree angle over #19)

In the region of the crest of the bone (zoom) of the #19 implant, when subjected to lateral loading (b), there was a lower stress gradient on the mesial of the implant in the SPb group (point 4 = 4.23 kPa). At the distal aspect of the implant, the stress level among the groups was similar (point 9). The CSb group had a lower stress concentration at all 12 points when compared to the CPb group (Figure 8). Table 3 shows the areas under the curve in the graphs of stress for the groups analyzed with the loading type “b” at 12 points in the region of the crest of bone of the #19 implant. The values shown in Table 3 are normalized by the group that obtained the lowest value of the graph area (SPb zoom = 136.91). It is observed that the CSb zoom group presented levels 98% higher than the SPb zoom group, and the CPb zoom group presented levels that were 106% higher than the SPb zoom group.

Table 3

Graph area, b zoom; normalized areas, b zoom (force directed at a 40-degree angle over #19)

Graph area, b zoom; normalized areas, b zoom (force directed at a 40-degree angle over #19)
Graph area, b zoom; normalized areas, b zoom (force directed at a 40-degree angle over #19)

For group c, the stress distribution of the implants of groups CP, CS, and SP can be seen in Figure 9. A higher stress concentration around the #20 implant was observed for groups CP and CS. Although the stress distribution around the #22 tooth is higher in the SPc group, this group showed a reduction in the stress level around the implants, followed by the CSc group. Multiple points (1–19 and 23) of the CSc group showed lower levels of stress when compared to the CPc group. All of the points analyzed between the CPc and SPc groups were significantly different. The #20 implant in the SPc group presented the lowest stress levels (point 18 = 10.8 kPa) (Figure 10). Table 4 shows the normalized values determined by the group that received the lowest value of the graph area (SPc = 201.3).

Figures 9–11.

Figure 9. Representative images of the stress distribution in groups contact point, c (CPc) (a), contact surface, c (CSc) (b), and splinted, c (SPc) (c). Figures 10. Mean values of shear stress (t) for c-type loading. Figure 11. Mean values of shear stress (t) of the bone crest region of #20 implant for c-type loading.

Figures 9–11.

Figure 9. Representative images of the stress distribution in groups contact point, c (CPc) (a), contact surface, c (CSc) (b), and splinted, c (SPc) (c). Figures 10. Mean values of shear stress (t) for c-type loading. Figure 11. Mean values of shear stress (t) of the bone crest region of #20 implant for c-type loading.

Close modal
Table 4

Graph area, c; normalized areas, c (axial force directed over #20)

Graph area, c; normalized areas, c (axial force directed over #20)
Graph area, c; normalized areas, c (axial force directed over #20)

For the area of crestal bone for the #20 implant, with axial loading (type c), it was noticed that the CSc zoom group showed a lower stress concentration at all 12 points compared to CPc. Higher levels of stress were observed on the superior and proximal areas of bone in relation to the implant, both mesial and distal, with the distal area (point 9) having the highest value of the PC group (25.8 kPa) (Figure 11). Table 5 shows the areas under the curve for the graphs of stress for the groups analyzed with the loading type “c” at 12 points in the crestal bone area of the #20 implant. The SPc group was not analyzed in the area of the crestal bone due to low stress values around the #20 implant, eliminating the need for comparison between other groups. The values shown in Table 5 are normalized by the group that received the lowest value of the graph area (CSc zoom = 196.2). It can be observed that the CPc zoom group showed levels 15.3% higher than the CSc zoom group.

Table 5

Graph area, c zoom; normalized areas, c zoom (axial force directed over #20)

Graph area, c zoom; normalized areas, c zoom (axial force directed over #20)
Graph area, c zoom; normalized areas, c zoom (axial force directed over #20)

Factors such as the occlusal load must be considered during the treatment planning process to ensure that prosthetic design and the number and location of dental implants are not subject to overloading.11  The number of implants required to support a fixed partial denture prosthesis depends on several factors, including the density and dimensions of the available bone, the nature of antagonist teeth, and the location of implants. In this experiment, three implants were positioned in a straight line in the photoelastic model because it is difficult to achieve a significant misalignment of implants due to restrictions typically imposed by bone resorption.12  The bending moment on a three-implant restoration, however, can be reduced 20–60% when there is an offset between the implants of 2–3 mm.13 

Higher success rates for fixed partial denture prostheses are reported in the literature in posterior areas of the mandible.14  In most cases, splinting of fixed prosthetic restorations is recommended during rehabilitation with multiple implants in areas with adjacent missing teeth. This method is necessary if there is less than one implant for each missing tooth, but it may also be applied in adjacent single unit crowns to achieve better biomechanical behavior of the prosthesis.4  In this study, it was observed that splinting adjacent implants led to a lower stress concentration surrounding the implants, especially during eccentric loading. There is a smooth transfer of stress to the #22 tooth in contact with the #21 implant in the “a” loading group. During b-type loading, the stress levels in the SP group were lower with respect to other groups. The CSa group showed values 11.23% higher than the SPa group, and the CPa group showed values 16.32% higher than the SPa group. It should be taken into consideration that functional deformations of the mandible do occur, such as torsion and bending, resulting in higher stress values on posterior implants with splinted crowns.15  Moreover, splinted crowns in a structure that does not have a passive framework fit may cause overload of the implants and, consequently, the surrounding bone.5  One way to avoid jeopardizing bone integrity in this manner is to separate the crowns, individualizing the implant restorations. Furthermore, the use of individual implant-supported restorations can reduce cost and simplify laboratory procedures, facilitating the fabrication of components that provide a passive fit to the prosthetic connections.6,16 

In a qualitative approach, the restorations contacting through points or surfaces in this study showed great similarity in the stress distribution around the implants in a-type loading when compared to the splinted crowns. However, in a study by Nissan,17  single unsplinted restorations were shown to transfer significantly less load to the implants and supporting structures than splinted restorations upon axial loading. Grossmann et al18  complemented this with the finding that when healthy anterior teeth and periodontally healthy canines are present and the patient has vertical occlusal stability, the canines immediately disocclude the posterior implants in lateral movements, and the splinting of posterior implants may be unnecessary. The separation of the crowns may be beneficial in avoiding peri-implant diseases, improving the interproximal contour and avoiding concentration of stress during jaw flexure,19  as well as increasing patient satisfaction. However, in eccentric loading, crown separation can lead to increased overload of implants suffering premature contact. Contacts that generate force that is not transmitted through the long axis of the implant may reduce the strength of the supporting bone.10  The results from groups CPb and CSb showed higher stress concentrations on the #19 implants compared to the SPb group. With all loading types in this study, the CS group presented lower stress values compared to the CP group. Considering the critical area surrounding the implant (zoom), the implants of groups CSb and CSc showed a reduction in the concentration of stress at all points in the crestal bone area by 8% and 15.3%, respectively, when compared to groups CPb and CPc, respectively. These results contradict the idea that a greater interproximal contact leads to a higher stress concentration along the implants with separated crowns.6 

Although it simplifies laboratory procedures, it should be kept in mind that adjusting the interproximal contacts in individual crowns is difficult. The placement of individual implants in the posterior jaw has a high success rate.20  The CSc and CPc groups have similar behavior compared to individual implants, and the highest stress levels in this study are within the range of strength of cortical bone to shearing forces. It is suggested that the separation of crowns may not cause damage in the supporting bone.

An important observation is that bone loss remains within the limits suggested by Albrektsson et al21  of 0.2 mm per year following the placement of the prosthesis; this is a clinically viable value that implantology has supported for many years. However, two other possible causes for loss of crestal bone beyond the standard values observed around integrated implants are reported: (1) local tissue infection caused by bacteria of the oral cavity, colonizing the implant-abutment interface and (2) biomechanical stress acting on the crest of bone around the implants, causing local microfractures and subsequent resorption of the surrounding bone.22 

Considering the levels of stress observed in the CP and CS groups, it is suggested that nonsplinted crowns may not be harmful to the physiological limit of the resistance of the bone, as the range of resistance of the cortical bone to shear forces is 68 MPa.23  According to the results and limitations of this study, crowns that contact each other through a larger surface area have advantages when compared to crowns in contact through a point surface because the distribution of stress in the system is smaller and more homogeneous. Under ideal conditions of occlusion, there is qualitative similarity among the groups (CPa, CSa, and SPa), suggesting that implant crowns may be separated without causing pathological bone resorption.

For the purposes of analysis, the models used in this study simplified some scenarios, such as the clinical conditions to minimize the errors inherent to the loading procedure and preparation of the photoelastic model. However, the method used in this study was correctly indicated, given the complex geometry and loading of the models. According to Dally and Rilley,9  the photoelastic technique has advantages over other methods because it allows a continuous field optical analysis. We recommend that the results of this study be validated by a clinical study, currently in development.

Based on the stress gradient observed in this study, it is determined that splinted crowns show better biomechanical behavior. Crowns contacting through a larger surface have advantages with respect to the crowns with a smaller point of contact because the distribution of stress in the system is smaller and more homogeneous.

Abbreviations

CP

contact point

CS

contact surface

SP

splinted

The Mechanical Projects Laboratory Prof. Henner A. Gomide (LPM)/UFU, HD Post-graduate School of Dentistry, and Neodent for the research incentive. The authors thank CNPq and FAPEMIG for financial support.

1
Weber
HP
,
Sukotjo
C
.
Does the type of implant prosthesis affect outcomes in the partially edentulous patient?
Int J Oral Maxillofac Implants
.
2010
;
22
:
140
172
.
2
Blanes
RJ
,
Bernard
JP
,
Blanes
ZM
,
Belser
UC. A
10-year prospective study of ITI dental implants placed in the posterior region: clinical and radiographic results
.
Clin Oral Implants Res
.
2007
;
18
:
699
706
.
3
Nissan
J
,
Gross
O
,
Ghelfan
O
,
Priel
I
,
Gross
M
,
Chaushu
G
.
The effect of splinting implant-supported restorations on stress distribution of different crown-implant ratios and crown height spaces
.
J Oral Maxillofac Surg
.
2011
;
69
:
2990
2994
.
4
Weinberg
LA
.
The biomechanics of force distribution in implant-supported prostheses
.
Int J Oral Maxillofac Implants
.
1993
;
8
:
19
31
.
5
McGlumphy
EA
,
Mendel
DA
,
Holloway
JA
.
Implant screw mechanics
.
Dent Clin North Am
.
1998
;
42
:
71
89
.
6
Guichet
DL
,
Yoshinobu
D
,
Caputo
AA
.
Effect of splinting and interproximal contact tightness on load transfer by implant restorations
.
J Prosthet Dent
.
2002
;
87
:
528
535
.
7
Bernardes
SR
,
de Araujo
CA
,
Neto
AJ
,
Simamoto
P
, Junior
das Neves
FD
.
Photoelastic analysis of stress patterns from different implant-abutment interfaces
.
Int J Oral Maxillofac Implants
.
2009
;
24
:
781
789
.
8
Tonella
BP
,
Pellizzer
EP
,
Ferraço
R
,
Falcón-Antenucci
R.M
.,
Carvalho
PS
,
Goiato
MC
.
Photoelastic analysis of cemented or screwed implant-supported prostheses with different prosthetic connections
.
J Oral Implantol
.
2011
;
37
:
401
410
.
9
Dally
JW
,
Riley
WF
(
eds.
).
Experimental Stress Analysis
.
New York
:
McGraw-Hill Kogakusha;
1978
.
10
Misch
CE
,
Bidez
MW
.
Implant-protected occlusion: a biomechanical rationale
.
Compendium
.
1994
;
15
:
1330
1342
.
11
Rangert
B
,
Krogh
PH
,
Langer
B
,
Van Roekel
N
.
Bending overload and implant fracture: a retrospective clinical analysis
.
Int J Oral Maxillofac Implants
.
1995
;
10
:
326
334
.
12
Itoh
H
,
Caputo
AA
,
Kuroe
T
,
Nakahara
H
.
Biomechanical comparison of straight and staggered implant placement configurations
.
Int J Periodontics Restorative Dent
.
2004
;
24
:
47
55
.
13
Daellenbach
K
,
Hurley
E
,
Brunski
JB
,
Rangert
B
.
Biomechanics of in-line vs offset implants supporting a partial prosthesis. IADR Abstract #1327
.
J Dent Res
75(spec ed)
:
183
.
14
Lekholm
U
,
Gunne
J
,
Henry
P
,
et al
.
Survival of the Branemark implant inpartially edentulous jaws: a 10-year prospective multicenter study
.
Int J Oral Maxillofac Implants
.
1999
;
4
:
639
645
.
15
Chen
DC
,
Lai
YL
,
Chi
LY
,
Lee
SY
.
Contributing factors of mandibular deformation during mouth opening
.
J Dent
.
2000
;
28
:
583
588
.
16
Vigolo
P
,
Zaccaria
M
.
Clinical evaluation of marginal bone level change of multiple adjacent implants restored with splinted and nonsplinted restorations: a 5-year prospective study
.
Int J Oral Maxillofac Implants
.
2010
;
25
:
1189
1194
.
17
Nissan
J
,
Ghelfan
O
,
Gross
M
,
Chaushu
GL
.
Analysis of load transfer and stress distribution by splinted and unsplinted implant-supported fixed cemented restorations
.
J Oral Rehabil
.
2010
;
37
:
658
662
.
18
Grossmann
Y
,
Finger
IM
,
Block
MS
.
Indications for splinting implant restorations
.
J Oral Maxillofac Sur
.
2005
;
63
:
1642
1652
.
19
van Steenberghe
D
,
Lekholm
U
,
Bolender
C
,
et al
.
Applicability of osseointegrated oral implants in the rehabilitation of partial edentulism: a prospective multicenter study on 558 fixtures
.
Int J Oral Maxillofac Implants
.
1990
;
5
:
272
281
.
20
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
:
381
389
.
21
Albrektsson
T
,
Zarb
G
,
Worthington
P
,
Eriksson
AR
.
The long-term efficacy of currently used dental implants: a review and proposed criteria of success
.
Int J Oral Maxillofac Implants
.
1986
;
1
:
11
25
.
22
Adell
R
,
Lekholm
U
,
Rockler
B
,
Branemark
PI. A
15-year study of osseointegrated implants in the treatment of the edentulous jaw
.
Int J Oral Surg
.
1981
;
10
:
387
416
.
23
Reilly
DT
,
Burstein
AH:
The elastic and ultimate properties of compact bone tissue
.
J Biomech
.
1975
;
8
:
393
405
.