Accuracy of three corrective techniques for implant bar fabrication

Giancarlo G. Romero, DDS, MS,a Robert Engelmeier, DMD, MS,b John M. Powers, PhD,c and
Andrew A. Canterbury, MA, CDTd
University of Texas-Houston Dental Branch, Houston, Texas
Statement of the problem. Numerous articles emphasize the importance of passivity of implant-
prosthetic component interfaces. Nonpassive interfaces can lead to bone loss, abutment fracture, and con-
necting screw breakage.
Purpose. The purpose of this study was to evaluate 3 postcasting techniques for the correction of non-
passive fit between a cast bar superstructure and its interface with an implant abutment.
Material and methods. Thirty implant Hader bars were fabricated based on a metal model composed
of two 3.8/4.5 HL PME titanium implant abutments. Initial measurements were collected on the y-axis
of the left implant abutment-bar interface by using a M2001ARS toolmaker microscope. Means were cal-
culated from buccal, distal, and lingual measurements on each specimen. Ten specimens were sectioned,
indexed, and corrected by casting the same alloy (group 1). Ten specimens were sectioned, indexed, and
corrected by soldering (group 2). The last 10 specimens were submitted to 2 cycles of electrical discharge
machining on a MedArc M-2 EDM machine (group 3). Postcorrection measurements were collected on
the 3 groups. A 1-way ANOVA and a Tukey-Kramer test at a .05 significance level were performed on
the 3 groups after the corrective techniques.
Results. Initial gap means were 192 µm for group 1, 190 µm for group 2, and 198 µm for group 3.
There was a significant difference (P<.05) in gap means between group 1 (15 µm) and group 2 (72 µm)
as well as between group 2 and group 3 (7.5 µm) after each correction technique. No difference was
detected between group 1 and group 3.
Conclusion. The electrical discharge machining group resulted in the smallest mean gap distance of 7.5
µm, thus meeting the criteria of passive fit (within 10 µm) described in the literature. (J Prosthet Dent
2000;84:602-7.)

CLINICAL IMPLICATIONS
The use of implant superstructures that splint 2 or more implants require a passive fit
to avoid possible complications such as bone loss, abutment fracture, and connecting
screw breakage. Of the correction techniques tested, electrical discharge machining may
result in the best fit.

T he attachment of implant-retained, removable

dental restorations to supporting fixtures can be
accomplished in 2 ways. The first is with the use of
resilient attachments on freestanding implant abut-
ments. The second is with the use of resilient
attachments to attach the prosthesis to a rigid bar
assembly that interconnects the implant fixtures.
Retention derived from a bar-retained prosthesis can
be impressive; however, a minimum of 12 to 13 mm of
interarch space is required for this approach.
aResident,
Graduate Prosthodontic Program.
bDirector, Graduate Prosthodontic Program.
cProfessor and Vice-Chairman, Department of Restorative Dentistry
and Biomaterials.
dAdvanced Prosthodontic Technical Specialist.
Protocols for the restoration of dental implant
patients are well documented in the dental litera-
ture.1-5 Most authors agree on the requirement for
a passive fit between the prosthesis framework and
the implant fixtures.6-8 The resiliency of the peri-
odontal membrane encountered in the natural
dentition is absent in the case of osseointegrated
dental implants. A nonpassive fit can lead to bone
loss, abutment fracture, and connecting screw
breakage; the last of these is by far the most com-
mon complication of a nonpassive fit.9,10 To achieve
optimum distribution of the loading forces through
the screw joint, there must be a precise, passive fit
between the implant abutment and the implant fix-
ture. The fit at this interface is considered to be
passive if the gap between the fixture and the abut-
ment is kept at 10 µm or less.11

602 THE JOURNAL OF PROSTHETIC DENTISTRY VOLUME 84 NUMBER 6

ROMERO ET AL THE JOURNAL OF PROSTHETIC DENTISTRY

Conventional dental laboratory techniques do not

allow the fabrication of a rigid bar assembly with an
acceptable degree of accuracy of fit. The error is due
mostly to the inconsistency of volumetric and linear
expansion of the fabrication materials used, which
include impression material, gypsum products, waxes,
investments, and casting metal. Potential distortion
can be generated at any step of the fabrication
process.12-18
Several different postcasting techniques have been
developed to correct inaccuracies of fit resulting from
the fabrication process.19-23 Among them are cutting
and soldering, cast-to, Kulzer abutment luting, and
electrical discharge machining (EDM). Even though
these techniques may be used to correct inaccuracies of
fit, they coincidentally may introduce additional inac- Fig. 1. Metal master model with titanium implant abutments.
curacies because of the dimensional linear changes of
the indexing materials and soldering investing material,
structural changes in the metal after heat treatment,
and wear of the components used for the execution of
the techniques.24-28
The purpose of this study was to evaluate 3 post-
casting techniques for the correction of nonpassive fit
between a cast bar superstructure and its interface with
implant fixtures. The techniques evaluated were cut-
ting and soldering, cast-to, and EDM. Evaluation was
based on the measurement and comparison of the
implant abutment-bar gold coping interface gap dis-
tances after correction by the 3 techniques just A B
mentioned.
Fig. 2. Verification jig on master (A) and working (B) models.
MATERIAL AND METHODS
Master and working model fabrication
Thirty implant bars were fabricated on a metal mas-
ter model. The master model was composed of 2 fabrication, a verification jig was used to verify the
titanium implant abutments (3.8/4.5 HL PME trans- reproducibility of the master model with the work-
mucosal abutment, 6 mm Steri-Oss, Dental Implant, ing models (Fig. 2).
Yorba Linda, Calif.) permanently screwed and
Bar fabrication
cemented into a tightly fitted hole. The abutments
were 18 mm apart and marked left (L) and right (R) Two gold alloy bar copings (Steri-Oss, Dental
on the metal block (Fig. 1). Implant) were torqued to the abutment replicas using
Thirty custom acrylic resin trays (Fastray, Bosworth 15 N·cm. A prefabricated Hader bar pattern, 18 mm
Co, Skokie, Ill.) were fabricated at least 24 hours in length and 3.8 mm in height (Attachments
before making the impressions. Perforations holes and International, Inc, San Mateo, Calif.), was secured to
polyether adhesive (Impregum F, ESPE America, the bar coping with pattern resin material (GC
Norristown, Pa.) were used as retentive mechanisms America, Inc, Chicago, Ill.). A silicone matrix was
for the polyether impression material (Impregum F, made for the relocation of the remaining patterns for
ESPE America). Thirty impressions were made by the specimens in this study. The bars were invested
using an open tray impression technique and 2 guide with a phosphate-bonded investment (Power-Cast,
pin-retained square impression copings (Steri-Oss, Whip Mix Corp) and cast with a type IV gold alloy
Dental Implants) attached to the implant abutments (Ney-Oro CB; J.M. Ney, Bloomfield, Conn.) by using
on the metal master model. Thirty working models a 2-stage burnout cycle: first stage, 11°C/min up to
were poured in type IV high-strength dental stone 427°C held for 30 minutes; second stage, 11°C/min
(Hard Rock, Whip Mix Corp, Louisville, Ky.), and up to 704°C held for 75 minutes. Great care was
2 abutment replicas were used (Blue Aluminum, taken to avoid damaging the interface surfaces of the
Steri-Oss, Dental Implants). Previous to the bar gold cylinders; the bulk of investment was removed

DECEMBER 2000 603

THE JOURNAL OF PROSTHETIC DENTISTRY
Fig. 3. Graphic illustration of gap measurement at interface
between implant bar assembly and implant abutment. Inset
is M2001ARS Toolmaker microscope.
A B
Fig. 4. Bar cut (A) and edges undercut (B) before correction
by cast-to technique.
A B
Fig. 5. Bar cut (A) and indexed for correction (B) by cut-and-
solder technique.
with an air chisel, followed by ultrasonic cleaning sub-
merged in a hydrofluoric acid substitute solution. Air
abrasion and polishing were avoided. Thirty bars were
fabricated.
Initial measurements
The initial measurements were made at the left
implant abutment-gold coping interface with a
M2001ARS Toolmaker Microscope with a rotary
604
ROMERO ET AL
stage (Gaertner Scientific Corporation, Chicago,
Ill.). Only the right abutment screw was torqued to
15 N·cm by using a new screw on each bar and leav-
ing the left abutment without a screw. Point “A” was
marked at the buccal, distal, and lingual surfaces of
the master cast abutments. Point “B” was marked at
the corresponding points on the opposing edge of
the gold copings. After careful alignment of the
specimens at each point A, gap distances were mea-
sured by rotating the linear scale on the drum
micrometer on the y-axis until the 90-degree filar
reticle intercepted point B at that position. Three
measurements were made of the gaps at the left
abutment of each specimen (Fig. 3). After the initial
measurements, the specimens were sorted into 3
equal groups for the postcasting modification tech-
niques.
Cast-to technique (group 1)
Bars corrected with the cast-to technique were sec-
tioned 5 mm from the right gold cylinder using a
0.45-mm separating disk. Mechanical undercuts were
milled into each cut surface before indexing (Fig. 4). The
segments of the bar were aligned with a Hader clip on
the metal master model and torqued using 15 N·cm.
They were then luted together with GC pattern resin.
A 12-mm, 10-gauge wax sprue was attached to the
junction. The assembly then was invested by using the
same investing technique previously described and cast
with the same type IV alloy from which the bar had
been fabricated.
Soldering technique (group 2)
Bars corrected by the soldering technique were sec-
tioned and indexed as those in group 1. No undercuts
were milled into the cut edges of the bars; the cut sur-
faces were polished before soldering. All the soldering
blocks were standardized with a silicone matrix and a
phosphate-bonded investment (Fig. 5). These blocks
were placed in a cold burnout oven, heated at the rate
of 17°C/min to 649°C, and then held for 20 minutes.
Borax flux (B flux, Ney Company, Bloomfield, Conn.)
was applied to the connector joint space. A 6-mm
length of a narrow strip of solder (0.615 alloy, Ney
Company) was sliced between the cut edges at the
junction and melted with a gas-air torch. The same
divesting and cleaning procedure was used for groups
1 and 2.
Electrical discharge machining (group 3)
Electrical discharge machining (EDM) working
models were made with mounting stone in which 2
copper abutment replicas were embedded. Ten trans-
fer jigs were fabricated from the metal master model in
which the copper abutment replicas were secured with
a catalyst-activated cyanoacrylate cement (Zapit,
VOLUME 84 NUMBER 6
ROMERO ET AL THE JOURNAL OF PROSTHETIC DENTISTRY

Table I. Means and standard deviations for the 3 groups

before and after technique
Group Status No. of bars Mean (µm) SD (µm)

Cast-to As-cast 10 192 70

Solder As-cast 10 190 70
EDM As-cast 10 198 78
Cast-to Modified 10 15 12
Solder
EDM
Modified
Modified
10
10
72
7.5
15
3.3
A B
EDM = Electrical discharge machining.
Fig. 6. A, Verification jig fabricated on master model. B,
Transfer jig with copper abutment analogs and wires
attached before pouring EDM working model.
Dental Ventures of America, Inc, Anaheim Hills,
Calif.) (Fig. 6). A silicone matrix was used to stan-
dardize the boxing of the working models, which were
poured in mounting stone. Two models were fabricat-
ed for each bar specimen.
A model 2-dimensional EDM machine (MedArc,
Pilot Industry, Inc, Grass Lake, Mich.) was used for
the electrical discharge machining of the bars in this
group. The machine settings were as follows: 25 to 40
microseconds (feed rate), 30 to 60 volts (gap voltage),
1 to 3 amperes (power level), and positive polarity.
The gap voltage meter and ampere meter provided a
visual indication of the efficiency of the operation. A
steady reading indicated an efficient cut, whereas a A B
highly erratic reading indicated an inefficient cut. The
criterion for stopping the EDM process was seeing Fig. 7. A, Working model ready for EDM process. B, Bar
continuous sparks around both copper abutments (workpiece) undergoing EDM process.
when settings were at their lowest readings. This cycle
was run twice for each bar, once on each EDM work-
ing model (Fig. 7).
Postcorrection measurements were collected on all
3 groups at the same locations and in the same man-
ner as the initial measurements. The gap dimension
data were analyzed by 1-way analysis of variance
(Super-ANOVA, Abacus Concepts, Berkeley, Calif.)
for the 3 groups. The means were compared by Tukey-
Kramer interval at a .05 significance level.
RESULTS
Precorrection measurements were made at the
interface gap of the left abutment of each specimen. A
mean gap distance was calculated for each specimen by
measuring the gap at the mesial, distal, and lingual
aspects with a M2001ARS Toolmaker microscope.
Postcorrection measurements were made at the gap
of the left abutment of each specimen as previously Fig. 8. Graphic representation of as-cast and corrected
described. The smallest average gap distance found means for 3 tested groups.
after correction was 7.5 µm (group 3). The specimens
in group 1 had the second greatest decrease in average
gap distance (15 µm). Group 2, corrected by cutting 3 groups following the corrective techniques. There
and soldering the connector joint, followed with was a statistical difference of P<.05 among the groups.
72 µm. Table I and Figure 8 show the graphic repre- The Tukey-Kramer test showed a significant differ-
sentations of these mean values and their standard ence between groups 1 and 2 and between groups 2
deviations. A 1-way ANOVA was performed on the and 3. There was no statistically significant difference

DECEMBER 2000 605

THE JOURNAL OF PROSTHETIC DENTISTRY
between group 1 (the cut-and-cast-to technique
group) and group 3 (the EDM technique group).
There was a correction of horizontal discrepancies
in all specimens in the cast-to group (group 1), which
means that there was no horizontal misalignment
between the right and left implant abutments and gold
copings after they were corrected by the cast-to proce-
dure. Seven of the 10 specimens in the EDM group
(group 3) were horizontally out of alignment; howev-
er, this misalignment did not interfere with the screw
connection procedure. Only 2 specimens in the cut-
and-solder group (group 2) demonstrated horizontal
misalignment.
DISCUSSION
The fabrication of an implant bar assembly involves
a variety of clinical and laboratory steps, including
impression development, master cast fabrication, bar
assembly wax-up, sprue and investment techniques,
and finally, casting and finishing procedures. The
potential exists at each of these steps to generate a dis-
tortion that may result in a nonpassive fit of the
restoration. Even though the techniques used in this
study strictly followed the guidelines established in the
literature, the 30 implant bar assemblies evaluated all
yielded initial gap distances that were beyond accept-
able accuracy.11 Inconsistencies in the linear and
volumetric expansion of the materials used makes such
inaccuracy unavoidable.12-18
The cast-to technique produced a more accurate fit
than the cut-and-solder procedure. White24 has sug-
gested that a casting metal must have adequate flow
properties to liquefy and that the alloy must be heated
to a temperature above its melting temperature (super-
heated). The initial contraction of the metal mass
occurs when the superheat is lost. The greater the dif-
ference between the superheat temperature and the
melting temperature of the metal, the greater the ulti-
mate contraction of the casting alloy as it solidifies.
Higher temperatures are reached in the area of a sol-
der joint than throughout the rest of the cast bar
assembly because of the concentration of the torch
flame in this area. This concentration is needed to melt
the solder and enable it to flow.
The acrylic indexes used to relate the cut pieces of
the bar assembly samples introduced more uncontrol-
lable error in both the cast-to and cut-and-solder
techniques. Shrinkage of the resin created stresses in
the indexes as the resin polymerized. This resulted in a
reduced distance between the coping portion of the
assemblies after the indexes were removed from the
metal master model.25
Similar results were found in a previous study26 that
compared postcasting correction of fit techniques.
According to the criteria proposed by Rangert and Jemt
et al, referenced in Carr et al,11 a gap of 10 µm or less is
606
ROMERO ET AL
necessary for the fit to be passive. Neither the cast-to
technique nor the cut-and-solder technique resulted in
an acceptable (passive) fit of the bar assemblies evaluat-
ed in this study. Only 50% of the cast-to specimens fell
below the 10 µm standard; none of the specimens cor-
rected by the cut-and-solder technique met this
standard. The technique of casting the connector
joints has several advantages over soldering the same
joints: There is less distortion of the position of the
framework parts during casting than during soldering;
the casting alloy used to fabricate the framework is the
same alloy used to cast the connector joint; and when
the connector joint is waxed, invested, and cast, there
is better control of the volume, position, and flow of
the metal. In addition, mechanical undercuts can be
used in the joint.
All specimens in the EDM group demonstrated an
overall mean gap distance of 7.5 µm, which was the
smallest error of fit among the 3 test groups. No
investing or casting procedures were used when cor-
recting this group. There were virtually no mechanical
forces at work on these samples. The cutting elec-
trodes never contacted the workpieces. The resulting
error in gap distance may be due to the addition of
error from analog wear during the EDM process,
dimensional changes that occurred at the time of the
transfer jig fabrication, and/or expansion of the gyp-
sum material used for model fabrication.
The metal removed from the electrode during
EDM usually is a fraction of that removed from the
workpiece. However, the loss of metal from the elec-
trode may be significant if the machining process
stretches over a long period. This problem can be con-
trolled in part by regulating the polarity in the EDM
computer. It is best to select positive polarity when
there is a need for fine detail or a 3-dimensional cut.
In this study, all EDM cycles were performed with
positive polarity to minimize electrode wear. Despite
this conservative approach, none of the copper abut-
ment analog electrodes showed “no wear” at the end
of the machining cycles. Similar wear was observed in
a previous study by Linehan et al.27 Two EDM mod-
els were used for each sample in this group to
minimize the error caused by metal removal from the
electrodes.
No heat was produced during the EDM process.
This eliminated 1 heat treatment of the specimens dur-
ing the correction procedure. Tan et al28 concluded
that the heat cycle of the casting process on metal
alloys might alter the configuration of the cast metal.
This EDM group of test samples was never submitted
to a heat process; the same is true of the samples in the
other groups. Fewer step errors were accumulated by
the bar assemblies corrected by the EDM process.
The EDM test samples demonstrated the most hor-
izontal discrepancy of fit among all the samples
VOLUME 84 NUMBER 6
ROMERO ET AL
evaluated in this study. The process of sectioning and
indexing the separate pieces of the test samples in the
other 2 groups allowed the copings to realign with
their respective implant abutments. No segments were
cut in the case of the machined samples in group 3.
These samples were submitted to the EDM process in
the as-cast state.
Further research is needed to determine whether
the cast-to technique provides a stronger connector
joint than the cut-and-solder technique. If that is
the case, further research will be needed to evaluate
the result of first using the cast-to technique to cor-
rect the horizontal discrepancy of fit, then using
EDM to correct any remaining error.
CONCLUSIONS
Within the limitations of this study, the following
may be concluded about the 3 postcasting techniques
to correct the nonpassive fit of cast implant bar assem-
blies:
1. The cut-and-solder technique yielded the great-
est gap distances. None of these gap distances met the
standard for passive fit of a cast implant bar assembly.
2. Only those samples corrected by EDM met the
standard established in the literature for passive fit of a
cast implant bar assembly.
3. Although the EDM process produced the small-
est gap distances, there were no statistically significant
differences between this group and the group correct-
ed with the cast-to technique.
4. The cast-to technique best corrected horizontal
misalignment of the coping portions of the sample
bar assemblies. The cut-and-solder technique was
slightly less accurate. EDM was the least effective
technique for the correction of this type of error.
Horizontal misalignment did not interfere with the
implant screw connection process.
The following companies donated components and materials
necessary for this study: Steri-Oss, Dental Implants, Yorba Linda,
Calif.; Ney, Bloomfield, Conn.; ESPE America, Norristown, Pa.; and
Whip Mix, Louisville, Ky.
REFERENCES
1. Branemark PI, Hansson BO, Adell R, Breine U, Lindstrom J, Hallen O, et
al. Osseointegrated implants in the treatment of the edentulous jaw:
experience from a 10-year study. Scand J Plast Reconstr Surg Suppl
1977;16:1-132.
2. Zarb GA, Schmitt A. The longitudinal clinical effectiveness of osseointe-
grated implants: the Toronto Study. Part II: the prosthetic results. J Prosthet
Dent 1990;64:53-61.
3. Branemark PI, Zarb GA, Albrektsson T. Tissue-integrated prostheses,
osseointegration in clinical dentistry. Chicago: Quintessence; 1988.
4. Lundqvist S, Carlsson GE. Maxillary fixed prostheses on osseointegrated
dental implants. J Prosthet Dent 1983;50:262-70.
5. Lindqvist LW, Carlsson GE, Glantz PO. Rehabilitation of the edentulous
mandible with a tissue-integrated fixed prostheses: a six-year longitudi-
nal study. Quintessence Int 1987;18:89-96.
6. Zarb GA, Symington JM. Osseointegrated dental implants: preliminary
report on a replication study. J Prosthet Dent 1983;50:271-6.
DECEMBER 2000
THE JOURNAL OF PROSTHETIC DENTISTRY
7. Sones AD. Complications with osseointegrated implants. J Prosthet Dent
1989;62:581-5.
8. Skalak R. Biomechanical considerations in osseointegrated prostheses. J
Prosthet Dent 1983;49:843-8.
9. Adell R, Lekholm U, Rocker B, Branemark PI. A 15-year study of osseoin-
tegrated implants in the treatment of the edentulous jaw. Int J Oral Surg
1981;10:387-416.
10. Adell R, Eriksson B, Lekholm U, Branemark PI, Jemt T. A long-term fol-
low-up of osseointegrated implants in the treatment of totally edentulous
jaws. Int J Oral Maxillofac Implants 1990;5:347-59.
11. Carr AB, Gerard DA, Larsen PE. The response of bone in primates around
unloaded dental implants supporting prostheses with different levels of
fit. J Prosthet Dent 1996;76:500-9.
12. Carr AB, Stewart RB. Full-arch implant framework casting accuracy:
preliminary in vitro observation for in vivo testing. J Prosthodont
1993;2:2-8.
13. Carr AB. Comparison of impression techniques for a five-implant
mandibular model. Int J Oral Maxillofac Implants 1991;6:448-55.
14. Linke BA, Nicholls JI, Faucher RR. Distortion analysis of stone casts made
from impression materials. J Prosthet Dent 1985;54:794-802.
15. Inturregui J, Aquilino SA, Ryther JS, Lund PS. Evaluation of three impres-
sion techniques for osseointegrated oral implants. J Prosthet Dent
1993;69:503-9.
16. Assif D, Marshak B, Schmidt A. Accuracy of implant impression tech-
niques. Int J Oral Maxillofac Implants 1996;11:216-22.
17. Humphries RM, Yaman P, Bloem T. The accuracy of implant master casts
constructed from transfer impressions. Int J Oral Maxillofac Implants
1990;5:331-6.
18. Gettleman L, Ryge G. Accuracy of stone, metal and plastic die material.
J Calif Dent Assoc 1970;46:28-31.
19. Sutherland JK, Hallam RF. Soldering technique for osseointegrated
implant prostheses. J Prosthet Dent 1990;63:242-4.
20. Hobo S, Ichida E, Garcia LT. Osseointegration and occlusal rehabilita-
tion. Chicago: Quintessence; 1989. p. 176.
21. Sellers GC. Direct assembly framework for osseointegrated prosthesis. J
Prosthet Dent 1989;62:662-8.
22. Votik AJ. The Kulzer abutment luting: Kal technique. A direct assembly
framework method for osseointegrated implant prostheses. Implant
Society 1991;2:11-20.
23. Parel SM. Modified casting technique for osseointegrated fixed prosthe-
sis fabrication: a preliminary report. Int J Oral Maxillofac Implants
1989;4:33-40.
24. White GA. Osseointegrated dental technology. Carol Stream, IL:
Quintessence; 1993. p. 70-80.
25. Cho GC, Chee WW. Efficient soldering index materials for fixed partial
dentures and implants substructures. J Prosthet Dent 1995;73:424-7.
26. Zoidis PC, Winkler S, Karellos ND. The effect of soldering, electroweld-
ing, and cast-to procedures on the accuracy of fit of cast implants bars.
Implant Dent 1996;5:163-8.
27. Linehan AD, Windeler AS. Passive fit of implant-retained prosthetic
superstructures improved by electric discharge machining. J Prosthodont
1994;3:88-95.
28. Tan K, Rubenstein JE, Nicholls JI, Yuodelis RA. Three-dimensional
analysis of the casting accuracy of one-piece, osseointegrated implant-
retained prostheses. Int J Prosthodont 1993;6:346-63.
Reprint requests to:
Dr Robert Engelmeier
Director, Graduate Prosthodontic Program
University of Texas-Houston Dental Branch
6516 John Freeman Ave, DB-429C
Houston, TX 77030
FAX: (713)-500-2353
E-MAIL: rengelme@mail.db.uth.tmc.edu
Copyright © 2000 by The Editorial Council of The Journal of Prosthetic
Dentistry.
0022-3913/2000/$12.00 + 0. 10/1/111494
doi:10.1067/mpr.2000.111494
607