11

RESEARCH AND EDUCATION Fit of implant frameworks fabricated by different techniques

Stephen I. Riedy, DDS, MS, ~ Brien R. Lang, DDS, MS, b and Beth E. Lang, BA c School of Dentistry, University of Michigan, Ann Arbor, Mich.

Purpose . This study evaluated the precision of fit betwcen an implant framework and a patient simulation model that consisted of five implant abutments located in the mandibular symphysis area. One-piece cast frameworks were compared with Procera machined and laser-welded frameworks with laser videography. Mater ia l and m e t hods . Five frameworks of each type were measured with a laser digitizer and a graphics computer program to determine a single point represented as the "Centroid" for each framework component and each implant abutment. Differences between the paired centroids for each framework/ abutment interface are reported as x- and y-axis displacements, and z-axis gaps. The direction of the x- and y-axis displacements was determined. Results . There were significant differences (p < 0.05) in the precision of fit between both the one-piece cast frameworks and the Procera frameworks, when compared with the abutments in the patient simulation model. The laser-welded tYamework exhibited a more precise fit than the one-piece casting, with significant differences at four of the five prosthodontic interfaces, when evaluated by the mean z-axis gap at the centroid points. (J Prosthet Dent 1997;78:596-604.)

T h e precision o f fit or the closeness o f the clear- ance between the bearing surfaces of the implant abut- ment and implant componen t housed within a prosthe- sis framework 1 has been questioned as being a signifi- cant factor in: stress transfer, 2 the biomechanics of an implant system, 3,4 the occurrence of complications, ~ 10 and the response o f the host tissues at the biological interface.ll,12 An impor tant question asked by clinicians is: ~'What precision o f fit is achievable in clinical prac- tice, and is the fit different when frameworks are fabri- cated by different techniques?"

I f the precision o f fit or gap between a framework and the abutments is excessive, then the effect o f fit on the biologic interface may become extremely important . There are many factors that can influence the precision of fit achieved, including the manufacture of implant components and the several clinical and laboratory steps involved in the restoration o f the edentulous situation. Impression maldng, product ion o f the master cast, and framework fabrication can accumulatively influence the fit observed by the clinician when the framework is fit- ted to the abutments in the oral environment.

There are two implant framework fabrication tech-

Presented before the Academy of Prosthodontic's Annual Meeting, Orlando, Fla., May 1994.

~Adjunct Assistant Professor, Department of Prosthodontics. bprofessor and Chair, Department of Prosthodontics. CResearch Assistant, Department of Prosthodontics.

niques that are currently used in a majority of clinical situations. One is the conventional lost wax technique, which is used to cast one-piece full-arch implant frame- works. The other involves copy milling sections of an acrylic resin framework pat tern in grade 2 titanium and then laser welding the sections together (Procera sys- tem, Nobel Biocare, AB, G/Steborg, Sweden). The pre- cision of fit achieved with these two techniques has been repor ted by several investigators. Cart and Stewart 13 determined that the conventional lost wax technique, to produce a one-piece full-arch implant framework, was imprecise and inaccurate as judged against their passive fit requirements. However, White 14 has claimed that the cast one-piece Sheffield frameworks satisfy the one-screw fitting test. According to White, .4 corrective soldering has not been required with the Sheffield frameworks, and no implant or implant prosthodontic componen t has broken since 1985. This information is based on retrospective observations and has not been subjected to scientific validation. Jemt and Linden is found that machined and laser-welded titanium frameworks have a better fit to the abutments than do the cast frameworks.

Recognizing the need for additional scientific docu- mentat ion on the precision o f fit achieved by these two framework fabrication techniques, this study was initi- ated to examine thc null hypothesis that there are no differences in the precision of fit between the abutments of a patient simulation model and the prosthetic corn-

596 THE JOURNAL OF PROSTHETIC DENTISTRY VOLUME 78 NUMBER 6

RIEDY, LANG, AND LANG THE JOURNAL OF PROSTHETIC DENTISTRY

Fig. 1. Acrylic resin patient simulation model replicates eden- tulous mandibular arch containing five osseointegrated im- plants in symphysis region. Black reference spheres are essen- tial for computer matching of framework to abutments. (Ar- rows identify three reference spheres essential for computer matching of framework to abutments).

ponents in (a) the one-piece cast framework or (b) the machined and laser-welded titanium framework.

M A T E R I A L A N D M E T H O D S

An acrylic resin patient simulation model replicating an edentulous mandibular arch was designed for this project (Fig. ]) . Five titanium implants (SDCA 062, Nobel Biocare, Inc., Chicago, Ill.), 3.75 x 10 mm, were positioned in 1:he mandibular symphysis region, ante- rior to the mental foremen. Standard 5.5 mm titanium abutments (SDCA 005, Nobel Biocare, Inc.) were joined to the implants with abutment screws and tightened to 20 Ncm.

Ten standardized master casts were made o f the simu- lated mandible', model from 10 separate impressions by using a controlled and repeatable technique. The tray design and the use of an impression splint with an ap- propriate impression material were factors to be consid- ered in controlling the impression technique. Standard stainless steel abutment replicas were joined to square impression copings used in this impression procedure. Die stone was poured into each impression to create 10 master casts.

To control any influence that differences in the mas- ter cast would have on the framework fabrication pro- cess, the master casts were randomly assigned to one o f two groups. For group 1, the master casts were assigned to the Procera laboratory for fabrication of the machined and laser-welded titanium frameworks. For group 2, the master casts were randomly assigned to commercial den- tal laboratories selected in the Midwest United States for fabrication of a cast one-piece framework. Five dif- ferent laboratories, with a minimum of 8 years experi- ence in implant framework fabrication with the lost wax technique, were selected to fabricate the cast frameworks

Fig. 2. Framework positioned on abutments of patient simula- tion model to "best fit" relationship.

as opposed to selecting one specific laboratory. This ap- proach was chosen because it would better represent the reality o f clinical practice. It would also reduce a situa- tion of bias, where the frameworks produced in one com- mercial laboratory would not represent the variables encountered with this fabrication process.

For standardization in the design of the cast one-piece framework, the commercial laboratories were provided with detailed written instructions that included: (a) the type of alloy, which dictated the investment, (b) waxing technique and materials, (c) sprue design, (d) casting technique, and (e) finishing sequence to be followed. Instructions to the Procera laboratory included: (a) type of metal, (b) welding technique, and (c) the finishing sequence to be followed. Photographs o f the exact pat- tern of the framework design to be fabricated on the master cast were also provided to the laboratories. All frameworks were evaluated for compliance with these directions when they were returned to the investigators.

Laser videography ( M i t u t o y o / M T I Corp., Aurora, Ill.) was the method selected to measure the precision o f fit between the abutments and the framework com- ponents. This system combines a laser digitizer with a graphic computer program for both visual and numeri- cal displays of the linear data collected. The optic source is a Gallium Arsenide laser (Mi tu toyo /MTI Corp.) with a wavelength o f 780 nm capable of measuring at the micron level. System software allowed plotting of the collected x-, y-, and z-axis data in a three-dimensional mode. The system accuracy is ±0.001 ram, and repeat- ability tests measuring a calibration cast with five abut- ments during five measurement sessions resulted in a standard deviation of the mean x-axis of 0.010 ram, the y-axis of 0.010 ram, and the z-axis of 0.001 ram.

Three reference spheres essential to the computer matching of the framework to the abutments were in- corporated into the patient simulation model (Fig. 1). One sphere was placed in the area of the tongue space, while the other two were positioned on either side of

DECEMBER 1997' 597

THE JOURNAL OF PROSTHETIC DENTISTRY RIEDY, LANG, AND LANG

Fig. :3. Transfer impression in dental stone records "best fit" orientation of framework to patient simulation model and negatives of reference spheres (arrows).

Z

Fig. 5. Nobel Biocare standard abutment cylinder is digitized with laser videography system, and 1600 x-axis and y-axis measurement points are illustrated in computer graphic ren- dering.

o r e | . . | o , o ~w, ~ ,

j . o j O , w ~ . . . . I i

oe I o ' ~ . . . o " ' ' - - a o e l e , , p ~ ~ ' w ~ " " ' ' ' - o - " (0,0,0)

X

Fig. 4. X-, y-, and z-axis coordinate values for abutment cen- troid point@ (xl, yl, z l ) at a specific prosthodontic interface is compared with centroid point 0 (x2, y2, z2) for framework at this location to calculate linear differences in precision of fit.

the midline, anterior to the mandible. The patient simu- lation model was positioned in the digitizer, and the ar- eas to be measured were identified by the linked com- puter. A 6.0 m m 2 area that covered each abutment was digitized with an x- and y-axis measurement matrix and 0.100 m m between each data point. The area to be digi- tized on each reference sphere was identified. Each bear- ing surface area o f the implant abutment was digitized three times.

To measure the framework components , the frame- work was posit ioned on the abutments o f the patient simulation model , using a technique recommended in the clinic to achieve a "best fit" (Fig. 2). Beginning with

Fig. 6. After series of computer commands to remove hex head, remaining 450 data points are used to determine cen- troid point of abutment cylinder.

the most anterior (center) abutment location and pro- gressing posteriorly, the framework-bearing surfaces were fitted to their respective abutments. When the "best fit" was achieved, guide pins were positioned and t ightened until initial resistance was met. A transfer impression in dental stone was used to record the orientation of the framework to the patient simulation model and to the reference spheres (Fig. 3). The stone impression with the framework and the recorded "negatives" o f the ref- erence spheres were positioned in the digitizer, and ar- eas similar in dimensions to those used for the abut- ments and reference spheres were digitized. Each frame- work component-bear ing surface was digitized three times.

The centroid method previously repor ted by Lie and Jemt, ]6 and Tan et a l ) 7 was used to reduce the x-, y-, and z-axis data collected from the bearing surface for both the abutments and framework components to a single point for fit measurements. This me thod initially locates the center point and long axis o f each compo-

598 VOLUME 78 NUMBER 6

RIEDY, LANG, AND LANG THE JOURNAL OF PROSTHETIC DENTISTRY

l I I l

Fig. 7. Gold cylinder is digitized in manner similar to abut- ment, and 450 data points used to determine centroid point are illustrated in graphic rendering.

nent. The mean z-axis plane of the bearing surface is then calculated, and the center point of the component is projected along its long axis to the z-axis plane as the centroid point. Comparisons between the two centroid points, one for the abutment and the other for the frame- work component, provide the measurements of preci- sion of fit as x-axis, y-axis displacements, and z-axis gaps (Fig. 4).

The laser videography software was capable of defin- ing the centroid points, thus this method was used. The abutment illusllated in Figure 5 represents approximately 1600 x- and y-axis measurement points, collected by the laser videography system. Data points that do not con- tribute to the location of the centroid point are elimi- nated from the abutment data set. The remaining points (approximately 450) and their respective x-, y-, and z-axis coordinate values formed the database used to deter- mine the centroid point (Fig. 6). Data for determining the centroid points for the framework component-bearing sur- faces were managed in a similar manner (Fig. 7).

The center point and long axis of each abutment was located by matching balls of a lr~nown diameter to the recorded data set (Fig. 8). The mean z-axis plane of the bearing surface was computed from the 450 data points, and the centroid point was determined by projecting the center point onto the bearing surface plane (Fig. 9). Differences bewveen the various centroid points for the framework and for the abutments were determined by matching the reference spheres recorded during the mea- surement of the patient simulation with the sphere nega- tives recorded when measuring the framework (Fig. 10, A and B). This; computer function positioned the sets of centroid points in the same orientation that was used in the clinic to achieve a "best fit." Once matched, the centroid poin'cs for the framework could be compared with those of the abutments and the x- and y-axis dis- placements and z-axis gaps calculated (Fig. 11, A and B).

Z

X

I l

I l

Sphere A

of Sphere B

i

C e n t r o i d point ~ ~-~

t ~ J

I I I I I I I I I I

Fig. 8. Centroid point and long axis of abutment (AB) was located by matching spheres of known diameter (spheres A and B) to inner edge of bearing surface.

Fig. 9. Centroid point was determined by projecting center point onto mean bearing surface plane of abutment.

Calculation of a proper sample size was necessary to assure adequate levels of significance and power to de- tect differences in the area of clinical interest. The pur- pose of this study was to estimate the differences in pre- cision of fit between a one-piece cast framework (lain) and those fabricated by copy milling and laser welding (gmw). The null hypothesis was as follows:

DECEMBER 1997 599

THE JOURNAL OF PROSTHETIC DENTISTRY RIEDY, LANG, AND LANG

Fig. 10. A, Centroid points (ar rows) were determined for all abutments (green) and bearing surface components of frame- work (yellow). Reference spheres are also pictured. B, Com- puter-matched reference spheres of patient simulation model with their negative representations recorded in stone transfer impression. This matching procedure produced same relation- ship in computer at each prosthodontic interface that was initially recorded when framework components (yellow) were positioned on abutments (green) to "best fit."

Table I. Establishing the experimental population

a za/2 1 - fJ zf~

0.100 1.645 0.800 0.840

0.050 1.960 0.850 1.030

0.025 2.240 0.900 1.282

0.010 2.576 0.950 1.645

H o : Bm c = ~m w

while the alternative hypothesis was as follows:

Ha: p m c ~ ~m w.

An 0t "a lpha" o f 0.05 and ~ "be ta" o f 0 .10 were used as levels o f statistical significance. I t was est imated that the range in precision o f fit for the one-piece castings wou ld be 0 .020 to 0 .050 m m at the z-axis centroid points, and for the machined and laser-welded frame- work would be 0 .010 to 0.025 ram. Therefore the maxi- m u m range differences wou ld be - 5 (20 to 25) to 40

Fig. 11. A, Fit between prosthetic gold cylinder in framework (yellow) and abutment (green) in patient simulation model is illustrated by computer graphic rendering. B, To demonstrate relationship of centroid point for abutment (AB) to that of framework component, computer graphic representation has been illuminated to illustrate "best fit."

(50 to 10) or 0 .045 mm. The standard deviation (6) for an experimental popu la t ion can be est imated as one four th o f the range; that is sigma = r a n g e / 4 or 0 . 0 4 5 / 4 = 0.011 mm. For this study, an acceptable precision o f fit was established as one that wou ld demons t ra te a z-axis gap between the f ramework and abu tmen t at the centroid point o f less than 0.025 mm. The needed sample size was calculated using the fol lowing formula:

N = 2(Zc~/2 + z[3) 2 x (~2/62

The values for cq Zc~/2, 1 - ]5, and Z[3 are listed in Tab le I . F o r this s tudy , n = 2 ( 1 . 9 6 + 1 . 2 8 ) 2 " ( 0 . 0 1 1 ) 2 / ( 0 . 0 2 5 ) 2 , or n = 4.25, and therefore a sample size o f 5 was needed for each f ramework group. 18 For this study, power = 1 - [3, and 13 = 0.1; therefore, power = 1 - 0.1 = 0.9, mean ing that a difference in the preci- sion o f fit o f the one-piece cast frameworks versus the mach incd and laser-welded framcworks will be detected 90% o f the t ime with a sample size o f n = 5.

The statistical test selected for the analysis o f the data was a two-way analysis o f variance (ANOVA). The level o f significance to reject the null-hypothesis wasp _< 0.05.

600 VOLUME 78 NUMBER 6

RIEDY, LANG, A N D L A N G THE J O U R N A L OF PROSTHETIC DENTISTRY

Table II. Two-way ANOVA on abutment x-, y-, and z-values

Source df Sum of squares Mean square F-value P-value

X-value Framework 1 0.0226 0.0226 11.0807 0.0009

Abutment 4 0.2663 0.0666 32.6092 0.0001

Framework*abutment 4 0.0314 0.0079 3.8507 0.0044

Residual 440 0.8983 0.0020

Y-value Framework 1 0,0137 0.0137 12.7377 0.0004

Abutment 4 0,1360 0.0340 31.7269 0.0001

Framework*abutment 4 0.0651 0.0163 15.1935 0.0001

Residual 440 0.4715 0.0011

Z-value Framework 1 0.0052 0.0052 25.21 76 0.0001

Abutment 4 0.0038 0.0010 4.6709 0.0011

Framework*abutment 4 0.0032 0.0008 3.9066 0.0040

Residual 440 0.0903 0.0002

*interaction between framework and abutment.

Table !11. The mean x-, y-, and z-axis coordinate values and standard deviations in mm for the patient simulation model (PSM) and group 1 (laser-welded) and group 2 (one-piece) frameworks

X Y Z Centroid point PSM FR group 1 FR group 2 PSM FR group 1 FR group 2 PSM FR group 1 FR group 2

/~ean

AB1 -16 ,268 -16 .198" -16 .187" 6.346 6.364 6.323 4.316 4.334* 4.343*

AB2 -9 .297 -9 .185" -9 .198" -2.251 -2.288* -2 ,281" 4.153 4,173" 4 . I70"

AB3 -0 ,249 -0 .139" -0 .131" -5 .165 -5 .210" -5 ,213" 4.275 4.292* 4.301"

A84 8.805 8.892* 8.925* -2.325 -2.356 -2.337 4.361 4.381" 4.387*

AB5 16.930 17.065* 17.097* 5.575 5.592 5.554 4,924 4.945* 4.959*

SD

AB1 0.005 0.032 0.037 0.004 0.035 0.045 0.000 0,015 0.013

AB2 0.008 0.033 0.032 0.005 0,025 0.019 0.002 0.010 0.015

AB3 0.007 0.024 0.044 0.005 0.033 0.020 0.004 0,013 0.015

AB4 0.004 0.034 0.048 0.006 0.037 0.024 0.005 0,014 0.015

AB5 0.005 0.035 0.092 0.003 0.049 0.020 0.003 0.005 0.020

*Level of statistically significance P = 0.05.

R E S U L T S

The data collected consisted of different x-, y-, and z-axis values :For the patient simulation model and the framework centroid points for each of the 10 frameworks (5 in each group) at five abutment locations. Thus a re- peated measm:es ANOVA was performed for frameworks, abutments, and interactions between the frameworlcs and abutments. The factors and interaction results for this ex- perimental design are presented in Table II.

For the x-a~ds values, significant differences were found be tween f rameworks (p = 0 . 0 0 0 9 ) , a b u t m e n t s (p = 0.0001), and the interactions between framework and abutments (p = 0.0044). Significant differences were also found between frameworks (p = 0.0004), abutments (p = 0.0001 ), and the interactions between framework and abutments (p = 0.0001) in the y-axis values. The z-axis values revealed significant differences between frameworks

(p = 0.0001), abutments (p = 0.0011), and the interac- tions between framework and abutments (p = 0.0040).

The mean x-, y-, and z-axis values for the centroid points for the five abutment locations for the patient simulation model and the frameworks in group I (laser- welded) and group 2 (one-piece casting) are presented in Table III. The abutments (AB) were numbered be- ginning AB 1 in the mandibular right canine area of the patient simulation model and proceeding around the arch to AB5 in the left canine region. A significant difference was found between the centroid point mean x- and z-axis coordinate values for frameworks in groups 1 and 2 when compared with the patient simulation model at all abutment locations (p < 0.05). In the y-axis, significant differences were found only at AB2 and AB 3 locations.

The mean differences and standard deviations between the ccntroid point data for the frameworks in groups 1

DECEMBER 1997 601

THE JOURNAL OF PROSTHETIC DENTISTRY RIEDY, LANG, AND LANG

ABI

Y 2O

AB5

-20 --I x

AB4

~ Patlent Simulation Model

~/~\\\\~ Framework group 1

~Fra~work group 2

Fig. 12. X-axis and y-axis mean data for cast one-piece frame- works and machined titanium laser-welded frameworks are plotted with data representing patient simulation model.

and 2 and for the patient simulation model in all three axes are presented in Table IV. When the differences between the framework and the abuunents in group 1 were compared with group 2, significant differences were found in the x- and y-axis arAB4 and AB5. A significant difference was found at AB1 only in the y-axis. In the z-axis, significant differences were found at locations AB1, AB3, AB4, and AB5.

An evaluation o f the raw data for all frameworks in the z-axis demonstrated a range from 0.002 to 0.047 mm for group 1 with 20% of the gaps greater than 0.025 ram. For group 2, the mean z-axis gap data ranged from 0.002 to 0.068 mm with 48% of the gap dimen- sions greater than 0.025 mm (data not provided in this stu#).

The x- and y-axis displacements for both framework groups are illustrated in Figure 12 as determined from an analysis of the x- and y-axis mean data in Table II. For both framework groups, the centroid point for the center framework component was centered at AB3. This com- puter assembly was similar to the process performed when the frameworks were fitted on the patient simulation model. The x-axis distance between the centroid points at locations AB2 and AB4 was less for the laser-welded frameworks (18.077 mm) and greater for the one-piece castings (18.123 ram) than the 18.102 mm for the pa- tient simulation model. For both framework groups, the distance between the centroid points was greater on the x-axis at AB1 and AB5 (group 1 = 33.263 ram; group 2 = 33.284 mm) than the 33.198 mm for the patient simulation model. The centroid points in the y-axis for the laser-welded frameworks were located more poste-

rior than the patient simulation at AB1 (6.346 mm ver- sus 6.364 ram) and AB5 (5.575 mm versus 6.592 mm), whereas the centroid points for the one-piece cast frame- works were positioned more anterior 6.323 mm and 5.554 mm, respectively.

D I S C U S S I O N

The experimental question in this study was to deter- mine the precision of fit o f implant frameworks in the oral environment when fabricated by the two techniques, the conventional lost wax cast method and the machined titanium and laser-welded fabrication process. A patient simulation model was chosen for the experimental de- sign that would permit the measurement of abutments with the laser digitizer.

Every attempt was made to control the variables en- countered in making the impression, pouring the mas- ter cast, and making the stone transfer impression that could potentially alter the fit o f the frameworks to the abutments. Controlling these factors and randomly as- signing the master casts to two groups (one for the cast frameworks and the other for the laser-welded method) eliminated bias related to these variables. Similarly, any differences introduced by positioning of the frameworks on the patient simulation model before making the stone transfer impression were controlled, because the same procedure was used for both framework groups. Differ- ences in the precision o f fit between the abutments and the framework components may be due to the frame- work fabrication process alone, or they may represent the cumulative differences caused by the several steps in the different techniques. In either situation, these dif- ferences are inherent in each fabrication process.

H o o k e 19 demonstrated that two points separated by 1 minute arc, or no closer than 0.100 mm and located approximately 25 cm or 10 inches from the eye, can be seen as two distinct individual points. The explanation for this phenomenon is directly related to the neuro- physiology o f the human eye. Because visual acuity is no better than 0.100 mm when an object is viewed 25 cm away, one could expect a sensitivity of 0.050 mm when the objects are viewed under magnification loops with ×2 magnification. The human optical senses might en- counter difficulty at precise measurements less than 0.050 mm and thus, for this investigation, a z-axis gap size greater than 0.025 mm was the estimate used as a measure o f clinical significance.

In this study, the mean x-, y-, and z-axis coordinate values for both framework groups were measured by la- ser videography, as were the abutments in the patient simulation model. The centroid points for the frame- work components and the patient simulation abutments were computed, and the mean x-, y-, and z-axis differ- ences between the frameworks and the abutment cen- troids were then calculated. These centroid differences were then subjected to a two-way ANOVA to determine

602 VOLUME 78 NUMBER 6

RIEDY, LANG, AND LANG THE JOURNAL OF PROSTHETIC DENTISTRY

Fig. 13. A, Prosthodontic interface (arrow) between AB5 (abutment) and framework component for cast one-piece framework. B, Computer graphic representat ion of prosthodontic interface illustrated in A; reference spheres can be seen.

Fig. 14. A, Prosthodontic interface (arrow) between AB5 (abut- ment) and framework component for machined titanium la- ser-welded framework. B, Reference spheres can be seen in computer graphic representation of prosthodontic interface of A.

whether significant differences existed between the two techniques for fabricating an implant framework.

Comparisons of the mean coordinate values for the x-, y-, and z-axes for both framework groups to the pa- tient simulation model exhibited significant differences (p < 0.05) (Table III). Obviously, something affected the precision o f fit o f the frameworks to the patient simu- lation model. The differences observed demonstrated an influence by the techniques themselves.

In every instance, a significant difference was found in the mean z-axis data for both framework groups, when compared with the mean z-axis measurements for the patient simulation model (Table III). These differences are more precisely determined because of the 0.001 mm resolution used[ to make the measurements and are in- fluenced by the: framework fabrication techniques.

The presence: of a z-axis gap does not mean that there is no contact at this interface. Contact may be occurring somewhere else around the circumference of the bearing surface. Contact may be occurring on the facial, lingual, mesial, or dista][ areas of the bearing surface, or contact may be occurring between the framework component and the inner vertical inclines of the abutment, leading up to

the flat bearing surface. Either of these reasons can result in a gap at the centroid point and other locations at the interface (Figs. 13, A and B, and 14, A and B).

Calculation of the mean x-, y-, and z-axis differences between the framework groups and the patient simula- tion abutments revealed significant differences in the z-axis for framework to abutment interfaces at AB 1, AB 3, AB4, and AB5 (Table IV). For the laser-welded frame- works (group 1), the mean z-axis gaps were 0.018 mm and 0.021 mm for AB 1 and AB5, respectively. The mean z-axis gaps for the one-piece frameworks at AB1 was 0.027 mm and, at AB5, the gap was 0.035 mm. A sig- nificant difference was also found at AB3 or the center a b u t m e n t l oca t ion wi th the mean z-axis gap o f 0.018 mm for laser-welded frameworks (group 1), and the one-piece castings (group 2) being 0.026 mm. For abutment position AB4, the mean z-axis gap o f 0.019 mm was measured for the laser-welded frameworks (group 1), whereas the gap for the one-piece castings (group 2) was determined to be 0.026 mm.

The mean z-axis gaps were greater for the one-piece castings. The magnitude of the z-axis gap at the cen- troid points for the laser-welded frameworks ranged from

DECEMBER 1997 603

THE JOURNAL OF PROSTHETIC DENTISTRY RIEDY, LANG, AND LANG

Table IV. The mean x-, y-, and z-axis differences and standard deviations in mm of the f ramework - abutment PSM

x y z Centroid point Group 1 Group 2 Group 1 Group 2 Group 1 Group 2

Mean

AB1 0.069 0.081 0.018* -0.023 0.018* 0.027

AB2 0.112 0.099 -0.036 -0.029 0.020 0.01 7

AB3 0.110 0.118 -0.045 -0.048 0.01 8* 0.026

AB4 0.087* 0.120 -0.032* -0.012 0.01 9* 0.026

AB5 0.135* 0.166 0.01 7* -0.020 0.021 * 0.035

SD

AB1 0.033 0.037 0.036 0.045 0.015 0.013

AB2 0.034 0.033 0.025 0.020 0.010 0.016

AB3 0.025 0.044 0.033 0.021 0.013 0.015

AB4 0.034 0.048 0.038 0.025 0.015 0.016

AB5 0.035 0.092 0.049 0.020 0.006 0.020

*Level of statistically significance P = 0.05.

0.018 mm at AB3 (center) and AB1 (fight posterior) to 0.021 mm at AB5 (left posterior). The small mean vari- ance (0.003 ram) among the five abutment locations for the laser-welded frameworks seemed to indicate consistency and a more precise fit with this technique. The z-axis gap for the one-piece castings ranged from 0.017 mm at AB2 (fight anterior) to 0.035 mm at AB5 (left posterior) or a variance of 0.018 ram.

C O N C L U S I O N S

Within the limitations of this study, the following con- clusions were drawn.

1. There were significant differences in the precision o f fit between both the machined titanium laser-welded frameworks and the cast one-piece frameworks, when compared with the abutments in the patient simulation model.

2. The machined titanium laser-welded frameworks exhibited a more precise fit than the cast one-piece frame- works, with significant differences at four o f the five prosthodontic interfaces, when evaluated by the mean z-axis gap at the centroid points.

3. The machined titanium laser-welded frameworks exhibited less than a 25 ]am gap in the mean z-axis mea- surement at all five of the framework to abutment inter- faces.

We acknowledge the contributions by Mr. Rui-Feng Wang, Re- search Associate, Department of Prosthodontics, School of Dentistry, University of Michigan, in conducting the statistical analysis of the data.

REFERENCES 1. American National Standards Institute. Industrial engineering terminology.

Revised edition, Norcross (GA): ANSI Z94.0; 1989. p. 07-02, 07-20, 07- 58, 08-02, 08-19, 17-02.

2. Rangert B, Jemt T, Jorneus L. Forces and moments on Br&nemark implants. Int J Oral Maxillofac Implants 1989;4:241-7.

3. Rangert B. Mechanical and biomechanical guidelines for the use of Br&nemark system -- general studies. Aust Prosthodont J 1993;7(Supp):39-44.

4. Rangert B. Mechanical and biomechanical guidelines for the use of Br~nemark

system -- clinical studies. Aust Prosthodont J 1993;7(Supp):45-9. 5. Gregory M, Murphy WM, Scott J, Watson CJ, Reeve PK. A clinical study of

the Br&nemark dental implant system. Br Dent J 1990;168:18-23. 6. Zarb GA, Schmitt A. The longitudinal effectiveness of osseointegrated

dental implants: the Toronto study. Part IIh problems and complications encountered. J Prosthet Dent 1990;64:185-94.

7. Johansson G, Palmqvist S. Complications, supplementary treatment, and maintenance in edentulous arches with implant-supported fixed prosthe- sis. Int J Prosthodont 1990;3:89-92.

8. Jewt T, Linden B, Lekholm U. Failures and complications in 127 consecu- tively placed fixed partial prostheses supported by Br&nemark implants: from prosthetic treatment to first annual checkup, lnt J Oral Maxillofac Implants 1992;7:40-4.

9. Patterson EA, Johns RB. Theoretical analysis of the fatigue life of fixture screws in osseointegrated dental implants. Int J Oral Maxi[Iofac Implants 1992;7:26-34.

10. Carlson B, Carlsson GE. Prosthodontic complications in osseointegrated dental implant treatment. Int J Oral Maxillofac Implants 1994;9:90-4.

11. Hoshaw S, Brunski J, Cochran G. Mechanical loading of Br&nemark fix- tures affects interfacial bone modeling and remodeling. [nt J Oral Maxillofac Implants 1994;9:345-60.

12. Albrektsson T. On long-term maintenance of the osseointegrated response. Aust Prosthodont J 1993;7(Supp):l 5-24.

13. Carr AB, Stewart RB. Full-arch implant framework casting accuracy. Pre- liminary in vitro observation for in vivo testing. J Prosthodont 1993;2:2-8.

14. White GE. Osseointegrated dental technology. Chicago: Quintessence Publishing; 1993. p. 78-90.

15. Jemt T, Linden B. Fixed implant-supported prostheses with welded tita- nium frameworks. [ntJ Periodont Rest Dent 1992;12:177-83.

16. Lie A, Jemt T. Photogrammetric measurements of implant positions. De- scription of a technique to determine the fit between implants and super- structures. J Oral Impl Res 1994;5:1-7.

17. Tan KB, Rubenstein JE, Nicholls JE, Yuodelis RA. Three-dimensional analy- sis of the casting accuracy of one-piece, osseointegrated implant-retained prostheses. Int J Prosthodont 1993;6:346-63.

18. Fleiss JL, Kingman A. Statistical management of data in clinical research. Crit Rev Oral Blot Med 1990;1:55-66.

19. Overheim RD, Wagner DL. Light and color. New York:John Wiley; 1982. p. 163-70.

Reprint requests to: DR. STEPHEN J. RIEDY 44790 CHARNWOdD DR. PLYMOUTH, M[ 48170

Copyright © 1997 by The Editorial Council of The Journal of Prosthetic Den- tistry.

0022-3913/97/$5.00 + O. 10/1/84266

604 VOLUME 78 NUMBER 6