Analysis of strain at selected bone sites of a cantilevered

implant-supported prosthesis
Esam A. Tashkandi, BDS, MS,a Brien R. Lang, DDS, MS,b and Marion J. Edge,
DMD, MSc
School of Dentistry, The University of Michigan, Ann Arbor, Mich.

Cantilever length of the superstructure of an implant-supported prosthesis can be

affected by the location, number, and dimensions of the implants, arch form, opposing
dentition, and functionally generated forces. In this study, the effects of cantilever
length on load distribution to implants was analyzed. A fresh bovine rib was used to
create a clinical simulation in which three implants were used to support an implant
prosthesis. The implants were connected by means of a cantilevered superstructure.
Six linear strain gauges were placed in strategic locations where it has been reported
that stress is concentrated. The model was loaded with an Instron machine and the
load was applied at 0,5,10,15,20 and 25 mm along the cantilever length. Each loading
test was repeated five times, and loads of 10 and 20 pounds were applied. Dynamic
registration of the strain was recorded at each gauge. Significant differences in the
magnitude of strain between the six strain gauge locations were recorded with the 10
and 20 lb loads and at all cantilevered lengths. The results revealed that the greatest
maximum strain occurred at the strain gauge positioned on cortical bone over the
apex of the most distal implant under both loading conditions. (J Prosthet Dent
1996;76:158-64.)

T he cantilever length of the superstructure of an dentition and the anticipated forces generated by the pa-
implant-supported prosthesis has been considered an im- tient during function and parafunction.
portant factor in the transfer of force that results from an The three-dimensional design of the cantilevered por-
occlusal load through the implant to the supporting b0ne.l tion of the superstructure has been shown to be significant
Equally important is the question: “Is the location of the in stress transfer. To avoid fatigue, Falk et al. 4 recom-
occlusal load on an implant-supported prosthesis a consid- mended that the cantilever section of the framework
eration in the strain distribution pattern in the bone sur- should be at least 5 mm in height and 4 mm in width when
rounding the implant?’ Skalak, in Branemark et a1.,2 cast in type III gold alloy. Dimensions less than those rec-
stated that the “cantilevered ends of a fmed prosthesis in- ommended could result in forces that reach the fatigue
crease loading on the screw nearest to the cantilevered limit of the alloy. Lundgren et a1.5 reported that chewing
end. Moderate extension may be tolerated if the fixtures patterns for edentulous patients with maxillary complete
are sufficiently strong.” Rangert et a1.3 suggested that dentures and mandibular implant-supported prostheses
cantilever lengths should vary with mechanical and bio- were comparable to those of patients with natural denti-
logic conditions, with bone status the most important fac- tion. They also noted larger local forces around the im-
tor in determining the cantilever length. The bone quality plants just anterior to the cantilevers during occluding-
often observed in the mandibular residual arch allows an with complete dentures. Falk et al6 reported that closing
average cantilever extensions of 15 to 20 mm. In contrast, and chewing forces increased bilaterally in the distal
the cantilever length in the more cancellous maxilla should direction for patients with a maxillary complete denture
not exceed 10 mm.3 The location, number, and dimensions and a mandibular implant-supported prosthesis and were
of the implants (length and diameter) and the arch form comparable with tooth-supported cantilever prostheses
are other important factors that can affect cantilever that occluded with natural teeth. On average, 70% of the
length. Other equally important factors are the opposing forces were supported by the cantilevers and 30% by the
implant-supported segments of the prostheses.
Lundgren et a1.7 also reported increased closing and
Research supported by the The American Academy of Fixed Pros- chewing forces distal on the cantilever. When the cantile-
thodontics, Tylman Research Grant. Third Place, Tylman ver occlusal contacts were reduced by 100 pm, the total
Award winner, 1994.
closing and chewing forces decreased over the cantilever
aPhD Student, Oral Health Sciences Program.
bProfessor and Chair, Department of Prosthodontics. segments. However, the leverage at the cantilever joint
CAssistant Professor, Department of Prosthodontics. distal to the most posterior implant remained the same.

158 THE JOURNAL OF PROSTHETIC DENTISTRY VOLUME 76 NLJMRER 2

Teal, LANG, AND EDGE THE JOURNAL OF P~O~~~C DENTISTRY

Gauge 6 Gauge 5 Gauge 4

I

I?&. 1. Model with gauge locations.

Lightening the occlusion over the two most posterior Fig. 2. Inferior surface of bone specimen with strain

implants on each side of the arch by 100 pm resulted in a
pronounced reduction in the total closing and chewing
forces over the entire prosthesesand the cantilever seg-
ments. A marked decreasein the leverage at the cantile-
ver joints was also noted.
There is little information in the literature about load
transfer and the distribution of forcesthrough the implant
complex to the surrounding bone. Therefore, the purpose
of this study was to dete~e the effect of a load on the
distribution of the resultant strain in the surrounding
bone. The load was applied vertically at the terminal im-
plant supporting the framework and at 5 mm increments
alongits cantilever length. This investigation wasinitiated
to addressthe null hypothesisthat there are no differences
in the ~stribution of strain at various bone locations
around implants when a load is applied along the cantile-
ver length of a fmed impIant~supported prosthesisframe-
work.
MATERIf!L AN-D ~THODS
The experimental model designed for this project in-
volved a segmentof bovine bonein which three endosseous
implants (Brbnemark, Nobdpharma USA, Chicago, Ill.)
were placed in a straight line to simulate an osseointegra-
tion condition. The experimental design was not to study
the influence of varying the cantilever length and the load
on different bone specimens.The bovine rib was used asa
vehicle for studying somebasic biomech~i~l questions
and variations among different bony specimenswas not
considered.Strain gaugeswere attached to the surface of
the bone at six diRerent l~ations. A snperstru~tnre was
fabricated to fit the implants with a type III p~ladiu~go~d
alloy (Ney Option, J. M. Ney Co,, BloomGeld,Corm.). The
cantilevered section was designed to permit loading at
several positions along its length.
Three 13 mm self-tapping Branemark implants (SDCA
063, No~lpha~a USA) were surgically placedin the bone
sample by use of the recommended surgical protocol
excluding strict asepsis.The use of self-tapping implants
resulted in intimate approximation and stability. Bicorti-
cal engagement was sought and verified by radiograph.
gaugesapplied at apicesof implants.
Fig. 3. Model on In&on table with stylus contacting at
0.0 mm location.
The distance between the centers of the implants was
m~n~ned at 15 mm to provide space for the strain
gauges. The bone specimenwas frozen in a sealed con-
tainer before experimentation. Three 4.0 mm standard
abutments (SDCA 006, Nobelpha~a USA) were screwed
onto the implant with 20 N/cm torque by use of a marural
torque driver (DIA 250, Nobelpharma USA). A radiograph
was taken to verify that all abutments were completely
seated.
After the abutments were connected,impressionswere
made to obtain a master cast and fab~~a~?a superstrue-
ture with a 30 mm cantilevered sectionfrom the center of
the most distal implant. The cantilever had a width of 6
mm and thickness of 5 mm. The casting was finished and
polished. Care was taken not to damage the machined
surface of the gold cylinder. Starting at the center of the
most posterior implant and directIy above the retaining
screw accesshole, a loading location was maadewith a fine
c~bo~dnrn disk. Loading locations were also prepared
at 5, 10, l&20, and 25 mm in a posterior direction to this
initial loading position.

AWXJST 1996 159

THl3 JOURNAL OF ~R~~TMETIC DENTISTRY TASHKANDI, LANG. AND EDGE

Lod:10lb*
200 IZl Omm
Smm
lS0 IOmm

No 1 No2 No3 No 4 No 5 No6

Strain Gauge

Fig. 4, Mean and SEM of maximum strain at different strain gaugesat different canti-
lever lengths and under different loads,with error bars representing error of multiple ob-
servations from one subject.

Table 1. Three-way ANOVA

source df Sumof squares Mean square F-value P+%&U?

Load 1 3118.0555 3118.0555 423.8366 0.0001

Distance 5 3609.8057 721.9611 98.1360 0.0001
Strain gauge 5 1280319.7274 256063.9455 0.00034807 0.0001
Loadx distance 5 773.5965 154.7193 21.0310 Q.0001
Loadx strain gauge 5 129228.3107 25845.6621 3513.1955 0.0001
Distaucex strain gauge 25 39412.4913 15’76.4997 214.2933 0.0001
Loadx distancex strain gauge 25 4599.6244 183.9850 25.0090 0.0001
Replication 4 40.2047 10.0512 L3F63 0.2458
Residual 284 2089.3139 7.3567
Dependent:Mtimumstiain.

Six open-facedstrain gauges(EA-06-125 m-120, Mea-
surementGroup Inc,, Raleigh, N. C.) were positionedat the
sites (Fig. 1). The strain gaugeshad a resistance of 120.0
ohm c 0.4%. Strain gauge 1 was placedin a direction that
was distal to the most posterior implant, namely nearest
to the cantilevered section(implal~t C). Strain gauge 2 was
positioned anterior to the most posterior implant. Strain
gauge3 wasplacedbetween the middle (implant B) andthe
most anterior (implant A) implants. Strain gauge 4 was
placedin direct oppositionto the apex of implant C. Strain
gauge 5 was positioned in direct opposition to the apex of
implant B. Strain gauge 6 was in direct opposition to the
apex of implant A.
The strain gaugeswere cementedon the prepared sites

160 VOLUME 76 NuMFm‘za


-5 0 5 10 ‘15 20 25 30
Distancein mm
Y = -3.981 - 1.037 *x; RA2 = .969
Fig. 5. Simple regression analysis on maximum strain/
distance at strain gauge 3 nnder 10 lb load.
on the bone with a thin glrn of methyl-Z-~y~oac~late
resin (M-Bond 200 adhesive, Measurement Group Inc.).
The bone specimen was then secured from both ends
to a U-shaped antopol~e~zing acrylic resin mounting
device without supporting the inferior surface of the bone.
The mounting device seenredthe model during loading.
Figure z illustrates the inferior surface of the bone speci-
men with strain gauges applied at the apices of the
implants.
The superst~cture was loaded at six locations.The first
was at the center of the retaining screw accesshole of im-
plant C. This location served as a control because it repre-
sented loading at the long axis of this posterior implant.
The other Jive locations were at 5, 10, 15, 20, and 25 mm
posterior to the initial loading site. Loading forcesof10 and
20 pounds were applied and each load was applied five
times at each Iocation to determine the reproducibility at
the different strain gauges.The bonesamplepositioned in
the mounting devicewas positionedin the In&on machine
(Model 4206, Instron Corp., Canton, Mass.) (Fig. 3). An
aluminum stylus V4 inch in diameter was attached to the
jaws of the loading member of the Instron machine and a
load wasapplied at acrossheadspeedof 0.004inch/minute.
The strain gaugeswere arranged in seriesto form a full
Whetstone bridge. The leads from the strain gaugeswere
connected to a multichannel bridge amplifier to form one
leg of the bridge, An Apple ~aci~~sh computer (Apple
Computers Inc., Cupertino, Calif) was interfaced with the
bridge amplifier to register output signalsof dynamic bone
strains and baselinedeflections from the cortical surfaceof
the bone sample.Superscopedata acquisition system soft-
ware (G. W. Instruments, Albany, N. Y.) was usedto aceu-
mulate the dynamic strain gaugeand baselinedeflections.
RESULTS
The three main factors examined were the load (IO and
20 lb), the cantilever length or distance (0,5,l.O, l&20, and
25 mm), and the six strain gauges.Three-way analysis of
AUGUST 1936
0
-10
-20
-30
-40
-50
-so
-70
-80
-5 0 5 10 15 20 25 30
Distancein mm
Y = -6.442 - 2.401 *X RA2 = .995
Fig. 6. Simple regression analysis on maximum strain/
distance at strain gauge 3 under 20 lb 1oa.d.
variance (NOVA) was performed with full interaction
(Table I). The replication revealed no significant difference
(p = 0.2548), which indicates that the databasewas reli-
able. Furthermore, all three factors and their interactions
indicated statistically significant differences at the
p = 0.0001 level.
The mean and stand~d error of measure(Sag) of max-
imum strain of the three main factors were calculated. Ta-
ble II presentsthe results of the maximum strain that oc-
curred at the di~erent strain gauge lo~ticlns at different
cantilever lengths under the two loads. The maximnm
strain that occnrred under a 20 lb load was statistically
significantly digerent from that associatedwith the 10 Ib
load. For cantilever length, there were no statistically sig-
nificant differences between 5 and 10 mm, and 20 and 25
mm. Regressionanalysis was petiormed to fkther inves-
tigate this correlation. In addition, there wasa statistically
si~~cant difference amongall the strain gangelocations.
The mean maximum strain found at strain gaugelocation
4, which represented the location at the apex of the most
terminal implant, consistently reported the greatest val-
ues.
Figure 4 illustrates the meanand SEM of the interaction
between the three factors of strain gauge location, load,
and cantilever length. Although the maximum strain did
not reveal statistically significant differences at all levels,
the difference in m~mum strain that occurred at strain
gauge 4 was consistently statistically significant.
The form of the relationship betweencantilever distance
and maximum strain was investigated. Tab1.eIII presents
the results of the regression analysis split by load and
strain gauge. In both cases,cantilever length was the in-
dependentvalue and maximum strain was the dependent
value. Both polynomial and simple regression analyses
were pe~o~ed and revealed that the ~ol~.orni~ fit was
not signikkantly a better fit than the linear results. This
indicated that the relationship between the cantilever
length and the maximum strain was linear. Strain gauge
161
THE JOURNAL OF PROSTHBTIC DENTISTRY TA%KANDI, LANG, AND EDGE

Table II. Mean and SEM on maximum strain between the different levels of the strain gauge at the different distances
under the different loads
1Oib 20 lb

Distance StrainNo. Mean SEM Mean SEM

Omm 1 -20.601 0.467 -43.130 0.856
2 -12.207 0.360 -25.566 0.556
3 -4.560 1.438 -6.745 1.138
4 71.238 0.925 150.201 1.315
5 17.379 0.293 36.027 0.359
6 4.239 0.096 9.826 0.193
5mm 1 -21.891 0.707 -38.701 0.496
2 -17.782 0.263 -35.659 0.596
3 -8.740 0.322 -17.861 0.414
4 82.110 2.086 153.348 2.887
5 15,720 0.606 30.266 0.673
6 3.660 0.391 6.839 0.281
10mm 1 -23.297 0.691 -43.225 0.514
2 -23.164 0.440 -43.060 0.495
3 -14.061 0.190 -30.591 0.322
4 94.794 1.712 171.182 1.889
5 14.353 0.293 28.314 0.557
6 2.505 0.354 9.711 4.902
15 mm 1 -24.164 0.736 -45.921 0.118
2 -25.374 0.652 -50.149 0.326
3 19.476 0.498 -42.942 0.242
4 95.843 1.865 182.721 0.970
5 11.326 0.285 24.311 0.391
6 I.541 0.096 3.372 0.305
20mm 1 -25.511 0.761 -47.750 0.782
2 -28.73% 0.929 -62.475 1.333
3 -24.513. 0.831 -53.867 0.831
4 102.805 2.807 195,977 2.912
5 9.373 0.359 19.918 0.359
6 0.289 0.193 0.578 0.281
25 mm 1 -28.303 0.866 -48.905 0.785
2 -33.064 0.824 -74.105 1.191
3 -30.306 0.627 -66.693 0.921
4 112.913 3.092 215.050 2.904
5 7.030 0.548 15.817 0.331
6 -0.193 0.193 -1.445 0.264

3 was selectedas a representa~ve location. Figures 5 and
6 depict the scattergram of cantilever length and maxi-
mum strain that occurred at strain gauge 3 under a 10 Ib
load and under a 20 lb load.
DISCUSSION
The use of the Br&remark implant systemfor this study
wasbasedon the large body of scientific knowledge related
to that system. Long-~~ clinical studies87 g with Brane-
mark implants were also considerations. It is noted th.at
this simulation model was composedof three implants
placed in a straight line in a bovine bone. This modelwas
selectedto obtain baselineinformatian about the distribu-
tion pattern of strain in the absenceof any affset in implant
placement. Osseointe~ation wasnot sought or achievedin
this study; therefore direct transfer of the ma~itude of
data values recordedto an actual clinical patient situation
may not be appropriate. Kowever, the trends in mean
strain values and the locations and the directions of the
strain under the various cantilever conditions tested are
applicable to patients. The useof bone cement was consid-
ered,but that would imply adding an interface betweenthe
bone and the implant, which doesnot occur in the clinical
si~ation that was simulated.
According to Rangert, lo axial loading is more readily ac-
cepted by the design of most implant systems, whereas
transverse loading that produces abending moment has
the greatest potential to induce extremely high stressgra-
dients in the bone surrounding the implants. The ratio of
the lever arms, or leverage, is the factor that dete~ines
TASHKANDI, LANG, AND EDGE TRE JOURNAL OF PROSTHETIC DENTISTRY

this equilibrium of different forces. The product of force Table III. Simple regression analysis of maximum load
times the lever arm must be equal at both sides of the ful- (dependent value) versus distance (independent value)
crum. The product of force and distance to the force axis is Load R-squared The slope
the bending moment.3 In this study, the load applications (lb) Strain gauge value of the fit
were axial over the most terminal implant and transverse
10 S No1 0.742 Y = - 19.992 - 0.3904*x
(bending moments) when the force was applied at 5.0 mm
S No2 0.950 Y = -13.433 - 0.796*X
increments along the cantilever length posterior to the
S No3 0.969 Y = -3.981 - 1.037*x
most distal implant. S No4 0.869 Y = 73.891+ 1.551*x
For this model, bicortical engagement during implant S No5 0.939 Y = -17.802 - 0.422*X
placement was based on the desire to conform as much as s No6 0.904 Y = 4.381- 0.19*x
possible with the “ideal” surgical placement of implants in 20 S No1 0.624 Y = -40.411- 0.336*X
the completely edentulous mandible. The use of the long- S No2 0.981 Y = -24.971 - 139*x
est possible implants are recommended to engage the in- S No3 0.995 Y = -6.442 - 2.401*X
ferior cortical bone. The presence of bicortical engagement S No4 0.939 Y = 144.961+ 2.65*X
has been suggested to achieve better implant stability’l S No5 0.971 Y = 35.497 - 0.778*x
s No6 0.452 Y = 10.633 - 0.466*X
and to distribute the load more evenly.i2 Bicortical an-
chorage of the implant is essential for minimizing the Y vs. X Probability <O.OOl.
stress in the bone.iO In this study, radiographs were taken
to ensure that the implants had bicortical engagement.
The sequence of load application selected for this exper- may be extrapolated that the strain gauges on the superior
iment consisted of a total of 60 loading cycles. The 10 lb load surface were under tension and the gauges on the inferior
was initially applied five times at each of the six loading surface were under compression.
sites, 0, 5, 10, 15, 20, and 25 mm, for a total of 30 cycles. Analysis of dynamic strain that occurred at the six strain
The loading sequence was repeated for the 20 lb load for a gauge locations revealed, regardless of the cantilever
total of 60 cycles. No attempt was made to randomize the length used, a significantly greater strain occurring at the
load cycle or the load itself. This was apparently not a bone surrounding the apex of the most distal implant
problem as the results clearly demonstrated a linear pro- (strain gauge 4). In addition, there was a significant
gression for the strain recordings at the various strain difference in the maximum strain that occurred at all other
gauge sites. If loosening of the nonosseointegrated im- strain gauges when compared with the maximum strain
plants occurred, the linear progression would not have occurring at strain gauge 4. This finding is in agreement
been demonstrated. In this investigation, there were no with the results of Kinni et a1.,r4 who reported stress con-
signs of mobility of any of the implants at any stage dur- centrations at the apex after the application. of a 40 lb load.
ing the experiment. For cantilever length, the results ofthis study agree with
Each strain gauge was strategically positioned to record those of Rodriguez et al., I7 who used a s:imulated bone
the strain in the bone immediately adjacent to the gauge. model of methacrylate resin with six Branemark implants
The gauges placed at the inferior border of the bone spec- and reported that the highest strain was observed at the
imen adjacent to the apex of the implants (implants 4,5, terminal implant. No strain gauges were placed on the in-
and 6) permitted registration of the dynamic increase in ferior surface of the Rodriguez model. Their results also
the maximum strain that occurred at the apex of each im- indicated that there was a significant increase in the strain
plant. Similarly, the gauges placed at the bone surface ad- observed at the distal abutment as the cantilever length
jacent to the framework were strategically positioned to increased.
obtain information about the bone adjacent to the cervix of Tolman and Laney’* demonstrated that one of the com-
the implant. mon mechanical complications has been the fracture of
The selection of the location of the strain gauges was abutment screws. In their study, there were 87 occurrences
based on the results of previous studies13-i5 that reported of fracture of the abutment screw in 77 mandibles. Similar
a concentration of stress around the neck and apex of en- results were reported in the Toronto study by Zarb and
dosseous dental implants. Glantz et all6 demonstrated Schmitt.lg Although no screws fractured in this experi-
that the type of strain regarding tension and compression ment, the magnitude of stress found around the most dis-
was a function of strain gauge orientation. This study fo- tal implant presented the potential for fracture to occur
cused on the relative shift in the maximum strain as a over time under the tested loading conditions.
function of changing the cantilever length. However, Another factor that could potentially affect the strain at
because the three strain gauges on the superior cortex of the distal implant is the precision of fit between the
the bone model (gauges 1,2, and 3) were recording a neg- framework and the abutments. Great effort was made to
ative deflection from the baseline, whereas the three achieve precise fit in this study. The placement of the im-
gauges placed on the inferior cortex (gauges 4, 5, and 6) plants in a straight line as opposed to offsetting the
were recording a positive deflection from the baseline, it implants must also be considered. Under conditions of a

AUGUST 1996
TRE JOURNAL OF PROSTRETIC DENTISTRY TASRKANDI, LANG, AND EDGE

curved arch form, the magnitude of the strain would likely 2. Brdnemark PI, Zarb GA, Albrektsson T. Tissue-integrated prostheses
osseointegration in clinical dentistry. Chicago: Quintessence Publish-
be different; however the direction and location may be the ing Co Inc, 1985128.
same. 3. Rang&B, Jemt T, Jorneus L. Forces and moments on Brdnemark im-
plants. Int J Oral Maxillofac Implants 1989;4:241-7.
CLINICAL IMPLICATIONS 4. Falk H, Laurel1 L, Lundgren D. Occlusal interferences and cantilever
joint stress in implant-supported prostheses occluding with complete
In this study, when a load was applied to the framework, dentures. Int J Oral Maxillofac Implants 1990;5:70-7.
significantly greater strain was recorded at the apex of the 5. Lundgren D, Laurel1 L, Falk H, Bergendal T. Occlusal force pattern
during mastication in dentitions with mandibular fmed partial den-
implant immediately adjacent to the cantilever. As the tures supported on osseointegrated implants, J Prosthet Dent 1987;
load was applied more posteriorly on the cantilever, a sig- 58:197-203.
nificantly greater strain was recorded at all gauge sites 6. Falk H, Laurel1 L, Lundgren D. Occlusal force pattern in dentitions
with mandibular implant-supported fxed cantilever prostheses oc-
and particularly at the apex of the terminal implant. The cluded with complete dentures. Int J Oral Maxillofac Implants 1989;
combined length of an artificial first and second molar was 4:55-62.
18 to 20 mm, and the length of a premolar and first molar 7. Lundgren D, Falk H, Laurel1 L. Influence of number and distribution
of occlusal cantilever contacts on closing and chewing forces in denti-
was 15 to 16 mm. When these replacement teeth are sup- tions with implant-supported tied prostheses occluding with complete
ported by the needed acrylic resin in the implant prosthe- dentures. Int J Oral Maxillofac Implants 1989;4:277-83.
sis, cantilever lengths could approach 20 mm or more in 8. Adell R, Eriksson B, Lekholm U, BrHnemark PI, Jemt T. Long-term
follow-up study of osseointegrated implants in the treatment of totally
length. At the distance of 20 mm or more, the strain edentulous jaws. Int J Oral Maxillofac Implants 1990;5:347-59.
expressed at the bone locations measured may be excessive 9. Albrektsson T, Lekholm U. Osseointegration: current state of the art.
and demonstrate potential for complications and failures Dent Clin North Am 1989;33:537-54.
10. Rangert B. Mechanical and biomechanical guidelines for the use of
of the implant prosthesis. Although determination of the Brbemark system-general principles. Aust Prosthodont J 1993;7
optimum cantilever length was not an objective of this (suppl):39-44.
study, it is apparent that clinicians need to be aware of the 11. Hobo S, Ichida E, Garcia LT. Osseointegration and occlusal rehabilita-
tion. Tokyo: Quintessence Publishing Co Inc, 1989:75.
impact cantilever length has on the magnitude of strain 12. Lum LB, Osier JF. Load transfer from endosteal implants to support-
increase as they design implant restorations. ing bone: an analysis using statics. Part two: axial loading. J Oral Im-
plant01 1992;18:343-53.
CONCLUSIONS 13. Tesk JA, Widera 0. Stress distribution in bone arising from loading on
endosteal dental implants. J Biomed Mater Res 1973;7:251-61.
Within the limits of this investigation, the following 14. Kinni ME, Hokama SN, Caputo AA. Force transfer by osseointegration
conclusions were drawn. implant devices. Int J Oral Maxillofac Implants 1987;2:11-4.
15. Rieger MR, Mayberry M, Brose MO. Finite element analysis of six en-
1. The largest magnitude of maximum strain occurred dosseous implants. J Prosthet Dent 1990;63:671-6.
around the apex of the most distal implant when its apex 16. Glantz PO, Nyman S, Strandman E, Randow K. On functional strain
was in contact with cortical bone. in fixed mandibular reconstructions. II. An in viva study. Acta Odontol
Stand 1984;42:269-76.
2. There was a significant difference in the maximum 17. Rodriguez AM, Aquilino SA, Lund PS, Ryther JS, Southard TE. Eval-
strain that occurred at all strain gauge locations when the uation of strain at the terminal abutment site of a fixed mandibular
cantilever length was increased from 0 to 5,10,15,20 and prosthesis during cantilever loading. J Prosthodont 1993;2:93-102.
18. Tolman DE, Laney WR. Tissue-integrated prosthesis complications.
25 mm. Int J Oral Maxillofac Implants 1992;7:477-84.
3. Although the magnitude of the strain was increased 19. Zarb GA, Schmitt A. The longitudinal clinical effectiveness of osseoin-
when the load was increased from 10 to 20 lb, the pattern tegrated dental implants: the Toronto study. Part III: Problems and
complications encountered. J Prosthet Dent 1990;64:185-94.
of strain for the various gauge locations remained un-
changed. Reprint requests to:
DR. ESAM A. TASHKANDI
ORAL HEALTH SCIENCES PROGR~W
We thank Mr. Rui-Feng Wang for his invaluable assistance with SCHOOL OF DENTISTRY
the statistical analyses. THE UNIVERSITY OF MICHIGAN
hN ARBOR, MI 48109-1078

REFERENCES Copyright 0 1996 by The Editorial Council of The Journal of

1. SkalakR. Biomechanical considerations in osseointegrated prostheses. Prosthetic Dentidy.
J Prosthet Dent 1983;49:843-8. 0022-3913/96/$5.00 + 0. 10/l/73561

164 VOLUME 76 NUMJ3ER 2