- Email: [email protected]
Effects of thermocycling, load-cycling, and blood contamination on cemented implant abutments
Effects of thermocycling, load-cycling, and blood contamination on cemented implant abutments Daniel J. GaRey, DDS, MS,n Anthony H. L. Tjan, DrDent, DDS, PhD,b Robert A. James, DDS, MS,” and Angelo A. Caputo, PhDd Loma Linda University, School of Dentistry, Loma Linda, and University of California, School of Dentistry, Los Angeles, Calif. This study compared the effects of thermocycling, load-cycling, and human blood contamination on the retentive strength of five diRerent cements for luting posts to root-form implants. For each cement, 10 specimens (controls) were stored in an incubator, 10 specimens were thermocycled, 10 specimens were subjected to cyclic compressive loading, 10 specimens were subjected to a combination of thermocycling and cyclic compressive loading, and 10 specimens were contaminated with blood before cementation, then exposed to a combination of thermocycling and compressive loading. After 70 hours, retentive tests were performed on the Instron machine, and data were recorded in kilograms. Significant retentive differences were identified among the cements and with load-cycling, but minimal effect on the retentive strength was demonstrated from thermocycling. Blood contamination in combination with thermocycling and load-cycling adversely affected the retentive strengths of all of the cements and could be a major cause of abutment failure in dental implants. (J PBOSTAET DENT 1994;71:124-32.)
R -
oot form implants sometimes cannot be restored with commercially available screw-retained transmucosal abutments because of lack of parallelism of the implants or dentition, or because of damage to the internal thread in the implant. One approach to this problem is cementing a custom cast abutment or a machined cementable abutment in the implant. The recommended luting agent for cementing abutments to implants is a composite resin. Japanese researchers have developed two adhesives based on (1) 4-META monomer and (2) phosphate monomer. The 4-META/MMA-TBB cement, marketed as C & B Metabond resinous cement (RC) in the United States (Table I) was first reported by Takeyama et a1.l Masuhara et al2 observed that when 4-META (4-methacryloxyethyl trimellitic anhydride) and tributyl borane oxide (TBO) (catalyst) were combined with polymethvl methacrylate (MMA), the resulting composite resin retained impressive
*Assistant Professor, Oral Implantology Section, Department of Restorative Dentistry, Loma Linda University, School of Dentistry. bProfessor and Director, Biomaterials Research, Department of Restorative Dentistry, Loma Linda University, School of Dentistry. cProfessor and Director, Oral Implantology Clinic, Department of Restorative Dentistry, Loma Linda University, School of Dentistry. dProfessor and Director, Biomaterials Research Section, University of California, School of Dentistry. Copyright @ 1994 by The Editorial Council of THE JOURNAL OF PROSTHETIC DENTISTRY. 0022-3913/94/$1.00 + .lO. 10/l/61176
124
shear strength after thermocycling3 and recorded greater bonding strength to nonprecious metals than Panavia (Table I) and BIS-GMA resinous cements.4y5 However, Wada demonstrated diminished bond strength with its submersion in water. The other newer cement is Panavia EX, a BIS-GMA composite resin and phosphate monomer, lo-methacryloyloxydecyl dihydrogenphosphate (MDP). Panavia EX RC was introduced by Omura et al.’ as a highly filled (76%) powder/liquid cement system consisting of BIS-GMA composite resin, a phosphorylated methacrylate monomer (MPD) and a unique catalyst of BPO, t-amine, and sulfinate, that requires an anaerobic environment to harden. Early research suggested that Panavia EX RC exhibited low solubility in saliva, good physical properties, low polymerization shrinking, and low water sorption.6 Omura et al.7 believed that the bond strength of Panavia EX RC did not deteriorate after water storage, but Diaz-Arnold et al8 demonstrated decreased tensile strength after thermocycling and water storage. Blood contamination can occur when abutments are cemented in implants that are placed at the crest of alveolar bone, but no reports have been published describing the effects of blood contamination on resinous cements. This experiment investigated the effect of simulated oral stresses on post retention in commercially available implants with five commercial cements. The stresses resembled intraoral conditions when (1) the cemented post-implant specimen was exposed to heat and cold; (2) the cemented specimen endured compression as during swallowing and chewing; and (3) blood contaminated the implant before cementation of the post.
VOLUME71
NUMBER2
THE JOURNAL
GAREY ET AL
Fig.
1. Acrylic ring.
Fig.
OF PROSTHETIC
DENTISTRY
2. Plexiglas acrylic plastic implant positioner.
Table I. Luting agents Batch number Luting
agent and brand
Code
Base/powder
Catalyst/liquid
Manufacturer
distributor
Resin Panavia
Pn837
2116
Resiment with fluoride
011191
011491
C + B Metabond
01203
102031
Thin Film crown cementation
558007
557009
B98120790
H99042690
znPo4 Flecks zinc cement
J. Morita Tustin, Calif. Septodont New Castle, Del. Parke11 Biomaterials Farmingdale, N. Y. Den-Mat Santa Maria, Calif. Mizzy Corp. Cherry Hill, N. J.
Five different cements were selected because they have different chemical and physical characteristics and are used for cementing posts to implants (Table I). All of the cements were mixed according to the manufacturers’ instructions, which ranged from brief directions for Den-Mat thin film RC to detailed instructions for C & B Metabond RC. Elecause Dent-Mat Thin Film RC is not distributed with a standard manipulation technique, it was mixed for 15 seconds with linear (not rotary) strokes pressing the cement to the mixing pad. Investigators have shown tensile strength variations with different powder-to-liquid ratios of zinc phosphate cement; therefore 0.4 gm powder to 1.5 ml liquid dispensed from a calibrated ICCsyringe was mixed according to manufacturer’s directions.g
Corp., Anaheim, Calif.), 8 to 10 mm long and 3.8 mm diameter that had been rejected after machining because of external imperfections. The implants were internally and externally threaded, and their internal dimensions were in compliance with manufacturers’ specifications except for minor external defects. Each implant was secured with epoxy resin (Fig. 1) in an acrylic resin ring. These acrylic resin rings were 3/ inch long cross sections cut from l-inch outside diameter extruded clear acrylic resin tubing (S & W Plastics, Colton, Calif.). Plexiglas acrylic plastic (Rohm and Haas Co., Philadelphia, Pa.) was glued to the bottom of each ring with acrylic resin cement, resulting in a small watertight container. A custom-made Plexiglas acrylic plastic positioner suspended eight implants at uniform height in the center of eight rings (Fig. 2). Then epoxy resin (Buehler, Lake Bluff, Ill.) was mixed according to the manufacturer’s directions, poured in the rings, and allowed to set.
Implants
Abutments
This experiment selected 250 threaded commercially pure (cp) titanium root-form implants (Steri-Oss, Denar
straight
MATERIAL
A.ND METHODS
Cements
FEBRUARY
1994
The experiment reused 50 cementable, cp titanium, cementable
abutments
(Steri-Oss,
Denar
Corp.
125
THE
JOURNAL
OF PROSTHETIC
DENTISTRY
GAREY
ET AL
blood before the cement was placed. This was accomplished by expressing a drop of blood from a syringe with a 20-gauge needle, placing the drop inside the implant body, and spreading it over the entire internal surface with a blunt dental explorer. This was verified by visual inspection and the posts were then cemented.
Testing
Fig. 3. Abutment heads before and after machining. Anaheim, Calif.). The post portion was 5 mm long and the abutment 8 mm long. The tapered portion of the abutments were lathe-modified to zero-degree convergence to limit slippage during retentive testing (Fig. 3). Before these reused posts were cemented, they were soaked 24 hours in methylene chloride to soften residual cement. This residue was then removed with a modified dental chisel. The posts were then sandblasted for 5 seconds with 60 pm aluminum oxide (Ney-Brasive, J.m. Ney Co., Bloomfield, Conn.) at 1 hour before cementing, cleaned ultrasonically in acetone for 10 minutes, rinsed three times in deionized water, and air dried. The cemented areas of the posts were not touched before cementation, to avoid contamination.
Ceruentation Groups of 10 implants for one type of cement were cemented three or four at a time. One researcher mixed all of the cements and seated all of the posts. Immediately before cementing posts in the implants bodies, the researcher irrigated three or four implant bodies for 10 seconds with a stream of deionized water and air dried them for 10 seconds with filtered compressed air from an air syringe. Cement was inserted in the implant bodies with a 0.5 mm diameter plugger instrument until visual inspection verified all internal implant walls were coated. Then the post was entirely coated and seated in the implant body and rotated 180 degrees.The excesscement was wiped from the outside of the post, and the assembly was placed under a 5 kg load at 100% relative humidity and room temperature for 10 minutes. The cemented specimen was removed from the load and stored in an incubator (Model 2, Precision Scientific Co., Columbus, Oh.) at 37’ C in 100% relative humidity. For the blood-contaminated segment, each implant group was washed and air dried as previously described and the implants were then coated with fresh whole human
126
procedures
When specimens were not being thermocycled or loaded compressively, they were stored at 37OC in 100 % relative humidity. The cemented posts were subjected to the following five conditions. 1. Storage (control). Fifty post/implants specimens (10 for each cement) were cemented, then stored 70 hours and subjected to retentive tests. 2. Thermocycling. Fifty cemented specimens (10 for each cement) were stored 28 hours at 37’ C in 100 % relative humidity, then thermocycled between 8” C to 60” C for 400 cycles (2.5 minutes per cycle). Seventy hours after cementation, each specimen was subjected to testing for retention. 3. Compression loading. Fifty cemented specimens (10 for each cement) were stored 20 hours at 37O C in 100% relative humidity, then 25 were compression-loaded 27,500 percussions over 4 hours. The remaining 25 were compression-loaded 27,500 percussions over the following 4 hours. All 50 were stored overnight in the incubator and the same sequence was repeated for a total of 55,000 percussions on each specimen, Seventy hours after cementation, each specimen was tested for retention. 4. Compression loading and thermocycling. Fifty specimens (10 for each cement) were cemented, stored 20 hours at 37” C in 100% relative humidity, then load-cycled in compression and thermocycled as described. Seventy hours after cementation, each specimen was subjected to retentive tests. 5. Blood contamination, thermocycling, and compression loading. Fifty specimens (10 for each cement) were contaminated with blood, posts were cemented and stored 20 hours at 37“ C in 100 % relative humidity, then thermocycled and compressive-loaded as in group 4. Seventy hours after cementation, each specimen was subjected to retentive testing.
Test apparatus A custom compressor apparatus was designed to apply repetitive 4.6 kg vertical compressive loads at 35.5’ C f 2.2” C in 100% humidity (Fig. 4). It consisted of a Plexiglas acrylic plastic jig to hold the implants, a water bath to maintain correct temperature and humidity, and a variable speed drill press to generate repetitive load. The compressor jig was constucted from two l-inch thick square Plexiglas plastic plates (12 inch on side) secured with plated steel bolts and separated by plated steel washers (Fig. 5). The superior and inferior plates were notched
VOLUME
71
NUMBER
2
GAREY
THE
ET AL
JOURNAL
OF PROSTHETIC
DENTISTRY
Fig. 5. Compression jig with counter.
Fig. 4. Compressor apparatus. for correct realignment after disassembly. The jig contained 25 specimens at one time with 25 E-clip retained stainless steel rods (% inch in diameter) positioned in the superior plate directly over each specimen. This jig was wedged by gravity in the Plexiglas acrylic water bath that was sustained by gravity and Plexiglas acrylic wedges on the drill press table. The water bath temperature was maintained at 35.5” C 1 2.2” C with a modified submergible heater (ModeI LZ, Ebo-Jager, EL Segundo, Calif.). The variable speed drill press generated load by rolling two round 4.5 kg weights (SportMart Corp, Colton, Calif.) around the circle outlined by the stainless steel rods in the superior plate of the compressor jig (Fig. 4). A mechanical counter tallied the revolutions. The weights were fixed to a V8 inch round stainless steel axle by machined brass bushings stabilized in the weights with methyl methacrylate. E-clips maintained the positions of the weights at the ends of the axle. A drive shaft attached the drill press motor to this axle and E-clips maintained the drive shaft in the center of the axle. The compressive apparatus imparted only vertical load (4.6 kg) by passing the load through the stainless steel rods in the superior plate of the jig to the specimens in the inferior plate. Because the rods only rested on the specimens and were unattached, no horizontal force component was transferred. Twenty-five specimens were loaded at a time for 4-hour periods at 120 loads per minute for 2 consecutive days. Specimens not being t.hermocycled or compressive-loaded were stored at 37’ C in 100 % relative humidity. Addition of the weight of the drill press to the compressive load was avoided by machining a %sinch difference in the end hole of the drive shaft and the 3/s inch axle
FEBRUARY
1994
connecting the weights. This looseness of fit allowed vertical and horizontal movement of the axle during rotation and compressive loading and negated possible drill press loads. The drive shaft of the drill press was verified as perpendicular to the Plexiglas acrylic compressive jig by alignment of a % inch hole in the center of the jig with a straight % inch rod positioned in the drill press chuck. Smooth, unbinding interaction between the rod and the compressive jig demonstrated a go-degree alignment. The cemented specimens were arranged around the compressor jig in this sequence: Panavia EX, C & B Metabond, Thin Film, Resiment, and zinc phosphate cement. The height of the rods was standardized to ensure that the loads contacted each rod for identical periods. This was accomplished by placing specimens in the jig, then positioning the superior plate over the inferior plate in proper alignment. The heights of the rods were verified by an aluminum bar 1.5 cm by 1 cm x 10 cm placed over three rods simultaneously and measurement of the middle rod with a 0.012 inch feeler gauge. Rods that were shorter than the adjacent. rods were adjusted by use of circular polypropylene shims placed under the corresponding specimens to a 0.012 inch tolerance. Rods taller than adjacent rods had the bottom of the acrylic ring shortened slightly by sanding with a bench-mounted belt sander. By these techniques, the loads were held constant and vertical to each cemented specimen.
Retentive
tests
Components and posts were subjected to tensile or retentive tests on the universal testing machine (Model 1125, Instron Engineering, Canton, Mass.) at the end of each test period. Retentive test loads were parallel to the long axis of the post/implant specimens and continuous at lmm/minute crosshead speed until fracture. The loads at failure were recorded in kilograms.
127
THE
JOURNAL
OF PROSTHETIC
DENTISTRY
GAREY
ET AL
80
Kgf.
80
12345 R
12345 P
12345 2
12345 T
12345 C
Fig. 6. Summary of Scheffe multiple comparison test: R, Resiment; P, Panavia; 2, ZnP04, T, Thin Film; C, C & B Metabond. 1, Control; 2, thermocycle; 3, load-cycle; 4, thermocycle and load-cycle; 5, blood, thermocycle, load-cycle. Horizontal lines connect statistical means that are not significantly different at p < 0.05. II. Mean dislodging forces and standard deviations in kilograms of force
Table
Resiment
Panavia
Thii
Film
C&BMETA
Exper. Group No.
n
x
SD
n
TL
SD
II
%
SD
n
iI
SD
n
%
SD
1 2 3 4 5
10 10 9 9 8
100.8 98.6 94.2 92.6 63.3
8.0 6.4 6.7 4.1 8.7
10 10 9 10 8
99.6 96.2 98.0 86.9 21.1
13.0 7.9 8.6 8.8 9.2
10 9 10 10 10
81.8 78.8 72.1 76.8 44.6
7.4 9.3 7.5 8.6 17.4
10 8 10 10 10
73.3 77.3 70.7 68.0 52.2
10.7 9.6 3.1 6.5 7.0
10 10 8 10 8
91.4 90.8 82.8 88.3 30.2
12.0 6.8 9.7 15.3 11.0
znPo4
1, Control; 2, thermocycle; 3, compression; 4, thermocycle and compression; 5, thermocycle, compression, and blood. ra, Number of specimens.
Because complete cement deterioration did not occur after thermocycling and compressive loading during preliminary tests, specimens were not examined for looseness during compressive loading.
Interface
failure
mode
One investigator visually inspected each specimen after separation and estimated the percent of the surface of the post without cement. Data for each cement were tabulated and one-way, twoway, and three-way analysis of variances (ANOVA) and Scheffe multiple comparison F-tests (p < 0.05) were computed by use of the ABsurv statistical program (AndersonBell Corp., Parker, Colo.).
RESULTS Table II lists for each cement the loads required to separate the posts from the implants after the specimens were subjected to the various tests. The data were analyzed by 128
two-way ANOVA with cement and treatment conditions as the independent variables. The differences among cements were significant at p < 0.0001 and the differences among test conditions were significant at p < 0.0001. The interaction was significant at p < 0.0001. In general, the cements diminished in strength with treatment; however, the rate of decline was amplified by the addition of blood. The rate of decline was not congruent among cements and this may explain the significant interaction. Tables III and IV summarize the ANOVA for each cement and test conditions, and Fig. 6 shows the results of the Scheffe multiple comparison tests comparing the retentive strength of each cement after each test condition. A three-way ANOVA comparing dislodging forces for the treatment groups not contaminated with blood was performed because the effect of blood contamination was great, and it was impossible to see smaller statistical differences between other variables. The analysis was constructed with type of cement, thermocycling (yes/no), and VOLUME
71
NUMBER
2
GAREY
ET AL
THE
load-cycling (yes/no) as independent variables. This indicated statistically significant effects due to type of cement (p < 0.0001) and load-cycling (p < 0.0001) but no thermocycling effect 0, := 0.42). There were no significant interactions. This three-way analyses indicated that Panavia RC and Resiment RC required significantly greater dislodging forces than Thin Film RC and C & B Metabond RC and that zinc phosphate cement was intermediate between the two groupings and not significantly different from either (Table V).
Type of failure For all of the cements and all of the test conditions, minimal cement adhered to the posts aft,er dislodgement. Thus, cement failure was primarily adhesive or at the post/ cement interface.
DISCUSSION The three-way ANOVA demonstrated a small but statistically significant difference in the various cements’ reaction to load, suggesting that load affects the strength of FEBRUARY
1994
OF PROSTHETIC
DENTISTRY
Table III. Summary of one-way ANOVA (test conditions variable) Cement
Source
Panavia
A
Resiment
B A B
Thin Film C&B Metabond
Effects of test; conditions Of the control groups that were simply cemented and stored in the incubator, those luted with Panavia RC, Resiment RC, and zinc phosphate cements required substantially greater loads to dislodge the posts than those cemented with the Thin Film and C & B Metabond RCs (Fig. 6). The control-group means were the greatest values recorded except for the C & B Metabond RC. The mean retentive strengths of the thermocycled specimens were similar to the controls but (except C & B Metabond RC) were slightly less but not significantly less at p < 0.05 (Fig. 6). After compressive loading, the retentive strength of zinc phosphate cement diminished, but it remained more retentive than the Thin Film or C & B Metabond RCs. Panavia and Resiment RCs were substantially stronger (p < 0.05) than the other cements after compressive loading (Fig. 6). The combined thermocycled/compressive loaded specimens cemented with Panavia RC, Resiment RC, and zinc phosphate cement again separated at significantly greater loads than the Thin Film and C & B Metabond RCs. Combining blood contamination with thermocycling and compressive loading dramatically weakened all the cements (Table II & Fig. 6). Resiment RC was significantly stronger than Panavia RC, Thin Film RC, and zinc phosphate cement. C & B Metabond RC was the next most retentive and significantly better than Panavia RC and zinc phosphate cement. Thin Film RC recorded significantly greater strength than Panavia RC (Table III). Blood contamination reduced the retentive strength of all of the cements more than either thermocycling or compressive loading individually or combined, as can be seen by comparing Figs. 7 and 8.
JOURNAL
&PO4
DF
Sum of squares
Mean of squares
A B A
4 3’7292.00 9322.99 42 3965.77 94.42 4 7758.00 1939.50 41 1965.65 47.94 4 9023.05 2255.76 44 5062.06 115.05 4 3534.37 883.59
B A B
43 2587.30 4 22906.30 41 5343.10
60.17 5726.58 130.32
ProbF
ability
98.7362
0.0000
40.4545
0.0000
19.6074 0.0000 14.6850 0.0000
43.9426
0.0000
A, Between groups; B, within groups.
the cements (Table V). This may be misleading in a clinical situation because the difference found was small and the data were pooled from the three treatment groups (compression only, thermocycle only and compression and thermocycle only). In the oral environment, only the compression/thermocycle group exists. The small difference discovered between dislodging forces required for loaded (83 kgf) and nonloaded units (89 kgf) may not be clinically relevant. This effect may increase with more load-cycling, but additional study would be necessary for a conclusion. The one-way ANOVAs comparing individual cements and test conditions are more definitive.‘O The control group’s means were generally the greatest values recorded (Table II). Previous investigator.@ cemented steel disks to teeth, stored them for 7 days in 37’ C water, and reported retentive strengths approximately 50% less than our findings for identical or similar cements, but differences in surface areas, in test specimens, and in storage times could account for the dissimilarities. Thermocycling resulted in a slight reduction in retentive strength compared with the control means (except C & B Metabond RC),12 but the difference was not statistically significant (Table III). Uchiyama13 cemented shallow crowns to teeth and, after 24 hours and 300 thermocycles, reported retentive strengths of 73 to 112 kg for Panavia RC. He also investigated the effects of thermocycling on dye penetration with Panavia RC, 4-META RC, and zinc phosphate cement and discovered dye penetration on both enamel and dentin with zinc phosphate cement but only slightly with Panavia or 4-META resinous cement. Other investigators have reported minimal dye penetration after thermocycling with Panavia and 4-META resinous cements.14*I5 D&-Arnold et aL8 reported no significant difference in retentive strengths of Ni-Cr-Be alloy disks cemented with Superbond C & B, Panavia, or Comspan RCs and thermocycled 300 times. In the current study, the posts and implants are both made from grade 2 cp titanium with a coefficient of thermal expansion of 27 10-6/oC (Pa129
THE JOURNAL
OF PROSTHETIC
GAREY ET AL
DENTISTRY
120 100 80 60 40 20 0
Control *R
-e-P
ThermoKomp B-Z
-CT
*C
Fig. 7. Effect of load-cycling and thermocycling on dislodging forces.
cific Titanium, San Diego, Calif.). The coefficients of thermal expansion for the cements (25-35 10e6/“C, per personal communication with manufacturers) were similar to cp titanium, which suggests that minimal dimensional change would occur during thermal shock tests. This may explain why no significant differences were found between control and thermocycled groups. Kerby et al.12 measured cemented post retention in Steri-Oss implants that had been thermocycled 1000 times and reported slightly smaller retentive values for many of the same luting agents observed in this study. The posts used in their study were not cleaned and abraded with aluminum oxide as were the posts in this study, which may account for the differences in retentive values. In addition, the crosshead speed of the Instron machine they used to dislodge the posts was 0.5 mm/minute compared with 1 mm/minute crosshead speed used in this study and this may contribute to the discrepancies. All thermocycled groups in the current study were thermocycled at 28 hours after cementation as stated in Material and Methods. Load-cycling began at 20 hours after cementation; therefore, thermocycling and load-cycling did not occur simultaneously. Group 2 was stored at 37O C, 100% humidity, for 28 hours before thermocycling and groups 3, 4, and 5 were load-cycled 20 hours after cementation, then thermocycled if required (4 and 5). Hence, group 2 was stored 8 hours longer than other groups before treatment. This design was choosen so that all of the themocycling was performed at the same time period in the curing cycle of the resins. Matsumura et a1.3considered thermocycling a technique to accelerate water-aging deterioration. However, their 130
Therm0 + Comp +R
4p
Therm0 + Comp + Blood -cz
+T
+C
Fig. 8. Effect of blood on dislodging forces.
study used 20,000 to 50,000 cycles and did not demonstrate a significant difference in retention between lesser and greater thermocycled specimens. This study selected 400 thermocycles because Crim and Franklin16 demonstrated no difference in dye penetration of crowns luted with composite resin cement when the crowns were thermocycled either 100 or 1500 cycles. Composite resin deterioration with prolonged water storage has been reported. Studies by Soderholm,r7 DiazArnold et al.,8 and Arai et al.ls suggested that tensile and transverse strength of composite resin diminished slowly and proportionally to the time immersed in water. Because posts cemented in implants are intended to remain cemented for the life of the implant, water deterioration is particularly undesirable. The inevitable cement failure should be planned for by keying the post to the implant body and making prosthetic appliances detachable. The decrease in tensile strength after compressive loading was greater than the deterioration after thermocycling, and the three-way ANOVA demonstrated a statistical difference. Thus, this reduction suggested that compressive loading may influence cement failure more than thermocycling does, especially if identical metals are luted with cements. Uchiyama13 recorded 10% to 20% reductions in shear bond for disks cemented on teeth with both 4-META RC and Panavia EX RC and then shear-loaded to failure. This study averaged 1% to 5% reduction in retention with thermocycling alone and 2% to 10% reduction with compressive loading alone. Uchiyama’s study differed from the current study in that nonprecious alloy disks were cemented to facial tooth surfaces, then the disks were loaded to failure along the long
VOLUME
71
NUMBER
2
THE JOURNAL
GAREY ET AIL
Table IV.
OF PROSTHETIC
DENTISTRY
Summary of one-way ANOVA (luting agent variable) Test condition
Source
DF
A B A B A B A B A B
4 45 4 42 4 41 4 44
Control Thermocycle Compression Compression, Thermocycle Blood, Comprewion, Thermocycle
4
39
Sum of squares
Mean of squares
5549.68 4918.10 3517.34 2676.66 5859.86 2208.06 3865.59 3992.82 9320.93 5128.55
1387.42 109.29 879.34 63.73
F
1464.96 53.86 966.40 90.75 2330.23
Probability
12.6947
0.0000
13.7979
0.0000
27.2020
0.0000
10.6495
o.oooo
17.7213
0.0000
131.49
A, Betweengroups;II, within groups.
Table V. Summary of three-way ANOVA means and standard deviation Variable Thermocycle
Cement
Load
P
R
Z
T
C
Yes
NO
Yes
No
Mean
95.10
96.71
88.61
77.33
72.05
85.58
86.34
82.66
89.20
SD
10.68 3!)
7.08 38
11.50 38
8.65 39
8.31 38
12.62 96
14.33 96
13.02 95
13.18 97
Cell N
P, Penavia; R, Resimenlt; Z’, Thin Film; C, C & B Metahond; 2, zinc phosphate. Horizontal lines connect means that are not significant statistically at p < 0.05. This analysie does not include blood contamination data.
axis of the teeth. The current study used titanium posts cemented inside titanium implants, and the separation forces were primarily shear. However, because the posts were cemented inside the implants, friction between the posts and implants increased the force required for cement failure. With all treatments except blood contamination, Panavia and Resiment RCs were significantly stronger than other cements in the current experiment (Table III). After compressive loading, the retentive strength of zinc phosphate cement diminished, but it continued to be more retentive than Thin Film RC or C & B Metabond cement. Kakigawa et al.lg demonstrated that 66,400 cyclic loads with 7.5 kg decreased the tensile strength of Panavia EX RC by 36% (78 kg to 50 kg) on crowns cemented to prepared teeth. They reported that Superbond C & B RC exhibited greater retention than Panavia EX RC. After the combination of thermocycling and compressive loading, the specimens in the current study cemented with Panavia RC, Resiment RC, and zinc phosphate cement required substantially greater loads to be separated than specimens cemented with Thin Film and C & B Metabond RCs. Compared with the controls, a 3 % to 13 % reduction was recorded in the tensile strengths of the cements studied, but the deterioration was not statistically significant. Nakamura et aLI compared resinous and conventional ce-
FEBRUARY
1994
ments for luting inlays to natural teeth with 4-META cement and Panavia EX RC and discovered a similar nonsignificant reduction in strengths. Using cyclical loading (0 to 30 kg) of approximately lo7 times and thermocycling (4” to 60” C) lo3 times, they deceased inlay retention by approximately 10 % , similar to the values in this experiment, suggesting that elevation of the number of stresses may not significantly weaken 4-META or Panavia EX RCs. Combining cyclic compressive loading and thermocycling appeared to diminish the retentive strength of the cements more than either cyclic loading or thermocycling alone (Table II), which suggests that there may be an additive effect of compressive loading and thermocycling to weaken cements, particularly Panavia Ex, Resiment, and C & B Metabond resinous cements. Blood contamination with thermocycling and compressive loading substantially weakened all of the cements (Tables II and III). Resiment RC was significantly stronger than Panavia RC, Thin Film RC, and zinc phosphate cement. C & B Metabond RC was the next most retentive and was significantly stronger than Panavia RC and zinc phosphate cement; and Thin Film RC was significantly stronger than Panavia RC (Table III). Blood contamination reduced the retentive strength of all of the cements more than thermocycling or compressive loading individually or combined. In this study, this variable was the most 131
TEE
JOURNAL
OF PROSTHETIC
important factor affecting retentive strength of cements to lute posts to implants, with Panavia RC and zinc phosphate cement the most affected (Table III). However, all of the cements tested showed significant decreases in tensile strength with blood contamination. C & B Metabond RC displayed the least percentage of change, but it was consistently the weakest cement tested, except with blood contamination. Resiment RC recorded the greatest retentive strength, but it was notstatistically better than C & B Metabond RC. Cement interface failure is defined as the surface or area of cement failure. Adhesive failure is cement fracture at the post/cement interface or surface, and cohesive failure implies fracture within the cement. In the nonblood-contaminated specimens, the cement failure occurred primarily at the post/cement interface, indicating that the bond of cement to the post was weaker than the physical strength the cement. Nearly all blood-contaminated samples reflected cohesive cement failure. Presumably, blood contamination decreased the physical properties of the cements. Further study is recommended to determine the exact role of blood on the resinous cements. The cement/implant interface was deliberately omitted in the discussion because the implant was threaded for mechanical retention and did not require adhesion, The posts in the current study were reused. A pilot study suggested no significant differences in retentive values of posts that were repeatedly reused when the posts were cleaned and abraded as mentioned in Material and Methods. Posts not treated in this manner may have different retentive values. The clinical implication of this study is that blood contamination during cementation of implant posts could drastically reduce the strength of the cement and should be avoided. This suggests that blood contamination during cementation of any FPD would result in a weakened cement and the amount of weakening depends on the cement. CONCLUSIONS This study using an Instron machine indicated that thermocycling did not significantly reduce retentive strength of the test cements, but that cyclical compressive loading did. However, the decrease is small and may not be clinically apparent. The combination of thermocycling, cyclical load-stressing, and blood contamination substantially reduced the retentive strengths for all of the cements. This suggests that blood adversely affects the retentive strength of the cements tested more than other variables. Resiment RC exhibited the greatest mean retentive strength with blood contamination but it was significantly less than that of the Resiment RC control.
132
CAREY
DENTISTRY
ET AL
REFERENCES 1. Takeyman M, Kashibuchi N, Nakabayashi N, Mashuhara E. A research on self-curing resins (no. 17). Adhesive resin to tooth substance and dental alloys. J Jpn Sot Dent Appar Mater 1978;19:179-84. 2. Masuhara E, Kojima K, Kimura T. Studies on self-curing resins (2). Effect of alkylboron on the polymerization of methyl with benzoly peroxide. J Res Inst Dent Mater, Toyoko Med Dent Univ 1962;2:369-74. 3. Matsumura H, Yoshida K, Tanaka T, Atauta M. Adhesive bonding of titanium with a titanate coupler. J Dent Res 1990;69:1614-6. 4. Brazilay I, Myers ML, Cooper LB, GraserGN. Mechanical and chemical retention of laboratory cured composite to metal surfaces. J PROSTHET DENT 1966;59:131-7.
5. Imhery TA, Burgess JO, Naylor WT. Tensile strength of resin cements following various alloy surface treatments [Abstract]. J Dent Res 1991;70(special issue):399 6. Wada T. Properties of an adhesive resin. Dent Di,~ond 1984;9:102-9. 7. Omura I, Yamauchi J, Wada T. Adhesive and mechanical properties of anew dental adhesive [Abstract]. J Dent Rea 19&1,63(specialissue):233. a. D&-Arnold A, Williams WV. Aguilino S. Tensile strengths of three luting agents for adhesion fixed partial dentures. Int J Proethodont 1989;2:115-22. 9. Wacker DR, Tjan AHL. Effect of variation in powder-to-liquid ratio of zinc phosphate cement on the retention of posta. J PROWHET DENT 1988;60:49-52.
10. Conover WJ. Practical non-parametric statistics. New York: John Wiley & Sons, 1980;168. 11. Akau MN, Lorey RE. Tensile retention of posts using various cementa/ bonding agents [Abstract]. J Dent Res 1999;69:86?. 12. Kerhy R, McGlumphy EA, Holloway J. Some physical properties of implant abutment luting cements. Int J Prosthodont 1992;5:321-5. 13. Uchiyama Y. Clinical study on the retentive force of adhesive resins. J Jpn Prosthodont Sot 1986;30:198-294. 14. Nakamura K, Takatau T, Hoaoda H. Retentive characteristics and durability of adhesive resin luting cements for internal inlays. In: Okabe T, Takahashi S, eds. Transactions of International Congress on Dental Materials. Honolulu: The Academy of Dental Materials and The Japanese Society of Dental Materials and Devices, 1989;5. 15. Tjan AHL, Dunn JR, Grant BE. Microleakage of cast crowns cemented with a non adhesive resin [Abstract]. J Dent Res 1990,69:123. 16. Crim GA, Franklin G. Microleakage: the effect of storage and cycling duration. J PROSTHETDENT 1987;57:574-6. 17. Soderholm KJ. Degradation of glass 6ller in experimental composites. J Dent Ree 1981;69:1867-75. 18. Arai K, Fujii K, Hashimoto H. Influence on physical and mechanical properties and cause of disintegration for posterior composite resins by long-term water immersion. In: Okabe T, Takahaahl S, eds. Transactiona-International Congress on Dental Materials. Honolulu: The Academy of Dental Materials and The Japanese Society of Dental Materials and Devices, 198953. 19. Kakigawa H, Yokoyama Y, Tajima K, et al. Effect of repeated loading on retention of the crown fixed with adhesive resin cements. In: Okabe T, Takahashi S, eds. Transactions-International Congress on Dental Materials. Honolulu: The Academy of Dental Materials and The Japanese Society of Dental Materials and Devices, 1989,6. Reprint requests to: DR. DANIEL J. GAREY SCHOOL OF DFSTISTFIY DEPARTMENT OF RESTORATIVE DENTISTRY LOMA LINDA UNNEIWTY LOMA LINDA, CA 92350
Contributing Author Steven Prim, CDA, Instructor, Oral Implantology Laboratory, Department of Restorative Dentistry, Loma Linda University, School of Dentistry.
VOLUME
71
NUMBER
2
R -
oot form implants sometimes cannot be restored with commercially available screw-retained transmucosal abutments because of lack of parallelism of the implants or dentition, or because of damage to the internal thread in the implant. One approach to this problem is cementing a custom cast abutment or a machined cementable abutment in the implant. The recommended luting agent for cementing abutments to implants is a composite resin. Japanese researchers have developed two adhesives based on (1) 4-META monomer and (2) phosphate monomer. The 4-META/MMA-TBB cement, marketed as C & B Metabond resinous cement (RC) in the United States (Table I) was first reported by Takeyama et a1.l Masuhara et al2 observed that when 4-META (4-methacryloxyethyl trimellitic anhydride) and tributyl borane oxide (TBO) (catalyst) were combined with polymethvl methacrylate (MMA), the resulting composite resin retained impressive
*Assistant Professor, Oral Implantology Section, Department of Restorative Dentistry, Loma Linda University, School of Dentistry. bProfessor and Director, Biomaterials Research, Department of Restorative Dentistry, Loma Linda University, School of Dentistry. cProfessor and Director, Oral Implantology Clinic, Department of Restorative Dentistry, Loma Linda University, School of Dentistry. dProfessor and Director, Biomaterials Research Section, University of California, School of Dentistry. Copyright @ 1994 by The Editorial Council of THE JOURNAL OF PROSTHETIC DENTISTRY. 0022-3913/94/$1.00 + .lO. 10/l/61176
124
shear strength after thermocycling3 and recorded greater bonding strength to nonprecious metals than Panavia (Table I) and BIS-GMA resinous cements.4y5 However, Wada demonstrated diminished bond strength with its submersion in water. The other newer cement is Panavia EX, a BIS-GMA composite resin and phosphate monomer, lo-methacryloyloxydecyl dihydrogenphosphate (MDP). Panavia EX RC was introduced by Omura et al.’ as a highly filled (76%) powder/liquid cement system consisting of BIS-GMA composite resin, a phosphorylated methacrylate monomer (MPD) and a unique catalyst of BPO, t-amine, and sulfinate, that requires an anaerobic environment to harden. Early research suggested that Panavia EX RC exhibited low solubility in saliva, good physical properties, low polymerization shrinking, and low water sorption.6 Omura et al.7 believed that the bond strength of Panavia EX RC did not deteriorate after water storage, but Diaz-Arnold et al8 demonstrated decreased tensile strength after thermocycling and water storage. Blood contamination can occur when abutments are cemented in implants that are placed at the crest of alveolar bone, but no reports have been published describing the effects of blood contamination on resinous cements. This experiment investigated the effect of simulated oral stresses on post retention in commercially available implants with five commercial cements. The stresses resembled intraoral conditions when (1) the cemented post-implant specimen was exposed to heat and cold; (2) the cemented specimen endured compression as during swallowing and chewing; and (3) blood contaminated the implant before cementation of the post.
VOLUME71
NUMBER2
THE JOURNAL
GAREY ET AL
Fig.
1. Acrylic ring.
Fig.
OF PROSTHETIC
DENTISTRY
2. Plexiglas acrylic plastic implant positioner.
Table I. Luting agents Batch number Luting
agent and brand
Code
Base/powder
Catalyst/liquid
Manufacturer
distributor
Resin Panavia
Pn837
2116
Resiment with fluoride
011191
011491
C + B Metabond
01203
102031
Thin Film crown cementation
558007
557009
B98120790
H99042690
znPo4 Flecks zinc cement
J. Morita Tustin, Calif. Septodont New Castle, Del. Parke11 Biomaterials Farmingdale, N. Y. Den-Mat Santa Maria, Calif. Mizzy Corp. Cherry Hill, N. J.
Five different cements were selected because they have different chemical and physical characteristics and are used for cementing posts to implants (Table I). All of the cements were mixed according to the manufacturers’ instructions, which ranged from brief directions for Den-Mat thin film RC to detailed instructions for C & B Metabond RC. Elecause Dent-Mat Thin Film RC is not distributed with a standard manipulation technique, it was mixed for 15 seconds with linear (not rotary) strokes pressing the cement to the mixing pad. Investigators have shown tensile strength variations with different powder-to-liquid ratios of zinc phosphate cement; therefore 0.4 gm powder to 1.5 ml liquid dispensed from a calibrated ICCsyringe was mixed according to manufacturer’s directions.g
Corp., Anaheim, Calif.), 8 to 10 mm long and 3.8 mm diameter that had been rejected after machining because of external imperfections. The implants were internally and externally threaded, and their internal dimensions were in compliance with manufacturers’ specifications except for minor external defects. Each implant was secured with epoxy resin (Fig. 1) in an acrylic resin ring. These acrylic resin rings were 3/ inch long cross sections cut from l-inch outside diameter extruded clear acrylic resin tubing (S & W Plastics, Colton, Calif.). Plexiglas acrylic plastic (Rohm and Haas Co., Philadelphia, Pa.) was glued to the bottom of each ring with acrylic resin cement, resulting in a small watertight container. A custom-made Plexiglas acrylic plastic positioner suspended eight implants at uniform height in the center of eight rings (Fig. 2). Then epoxy resin (Buehler, Lake Bluff, Ill.) was mixed according to the manufacturer’s directions, poured in the rings, and allowed to set.
Implants
Abutments
This experiment selected 250 threaded commercially pure (cp) titanium root-form implants (Steri-Oss, Denar
straight
MATERIAL
A.ND METHODS
Cements
FEBRUARY
1994
The experiment reused 50 cementable, cp titanium, cementable
abutments
(Steri-Oss,
Denar
Corp.
125
THE
JOURNAL
OF PROSTHETIC
DENTISTRY
GAREY
ET AL
blood before the cement was placed. This was accomplished by expressing a drop of blood from a syringe with a 20-gauge needle, placing the drop inside the implant body, and spreading it over the entire internal surface with a blunt dental explorer. This was verified by visual inspection and the posts were then cemented.
Testing
Fig. 3. Abutment heads before and after machining. Anaheim, Calif.). The post portion was 5 mm long and the abutment 8 mm long. The tapered portion of the abutments were lathe-modified to zero-degree convergence to limit slippage during retentive testing (Fig. 3). Before these reused posts were cemented, they were soaked 24 hours in methylene chloride to soften residual cement. This residue was then removed with a modified dental chisel. The posts were then sandblasted for 5 seconds with 60 pm aluminum oxide (Ney-Brasive, J.m. Ney Co., Bloomfield, Conn.) at 1 hour before cementing, cleaned ultrasonically in acetone for 10 minutes, rinsed three times in deionized water, and air dried. The cemented areas of the posts were not touched before cementation, to avoid contamination.
Ceruentation Groups of 10 implants for one type of cement were cemented three or four at a time. One researcher mixed all of the cements and seated all of the posts. Immediately before cementing posts in the implants bodies, the researcher irrigated three or four implant bodies for 10 seconds with a stream of deionized water and air dried them for 10 seconds with filtered compressed air from an air syringe. Cement was inserted in the implant bodies with a 0.5 mm diameter plugger instrument until visual inspection verified all internal implant walls were coated. Then the post was entirely coated and seated in the implant body and rotated 180 degrees.The excesscement was wiped from the outside of the post, and the assembly was placed under a 5 kg load at 100% relative humidity and room temperature for 10 minutes. The cemented specimen was removed from the load and stored in an incubator (Model 2, Precision Scientific Co., Columbus, Oh.) at 37’ C in 100% relative humidity. For the blood-contaminated segment, each implant group was washed and air dried as previously described and the implants were then coated with fresh whole human
126
procedures
When specimens were not being thermocycled or loaded compressively, they were stored at 37OC in 100 % relative humidity. The cemented posts were subjected to the following five conditions. 1. Storage (control). Fifty post/implants specimens (10 for each cement) were cemented, then stored 70 hours and subjected to retentive tests. 2. Thermocycling. Fifty cemented specimens (10 for each cement) were stored 28 hours at 37’ C in 100 % relative humidity, then thermocycled between 8” C to 60” C for 400 cycles (2.5 minutes per cycle). Seventy hours after cementation, each specimen was subjected to testing for retention. 3. Compression loading. Fifty cemented specimens (10 for each cement) were stored 20 hours at 37O C in 100% relative humidity, then 25 were compression-loaded 27,500 percussions over 4 hours. The remaining 25 were compression-loaded 27,500 percussions over the following 4 hours. All 50 were stored overnight in the incubator and the same sequence was repeated for a total of 55,000 percussions on each specimen, Seventy hours after cementation, each specimen was tested for retention. 4. Compression loading and thermocycling. Fifty specimens (10 for each cement) were cemented, stored 20 hours at 37” C in 100% relative humidity, then load-cycled in compression and thermocycled as described. Seventy hours after cementation, each specimen was subjected to retentive tests. 5. Blood contamination, thermocycling, and compression loading. Fifty specimens (10 for each cement) were contaminated with blood, posts were cemented and stored 20 hours at 37“ C in 100 % relative humidity, then thermocycled and compressive-loaded as in group 4. Seventy hours after cementation, each specimen was subjected to retentive testing.
Test apparatus A custom compressor apparatus was designed to apply repetitive 4.6 kg vertical compressive loads at 35.5’ C f 2.2” C in 100% humidity (Fig. 4). It consisted of a Plexiglas acrylic plastic jig to hold the implants, a water bath to maintain correct temperature and humidity, and a variable speed drill press to generate repetitive load. The compressor jig was constucted from two l-inch thick square Plexiglas plastic plates (12 inch on side) secured with plated steel bolts and separated by plated steel washers (Fig. 5). The superior and inferior plates were notched
VOLUME
71
NUMBER
2
GAREY
THE
ET AL
JOURNAL
OF PROSTHETIC
DENTISTRY
Fig. 5. Compression jig with counter.
Fig. 4. Compressor apparatus. for correct realignment after disassembly. The jig contained 25 specimens at one time with 25 E-clip retained stainless steel rods (% inch in diameter) positioned in the superior plate directly over each specimen. This jig was wedged by gravity in the Plexiglas acrylic water bath that was sustained by gravity and Plexiglas acrylic wedges on the drill press table. The water bath temperature was maintained at 35.5” C 1 2.2” C with a modified submergible heater (ModeI LZ, Ebo-Jager, EL Segundo, Calif.). The variable speed drill press generated load by rolling two round 4.5 kg weights (SportMart Corp, Colton, Calif.) around the circle outlined by the stainless steel rods in the superior plate of the compressor jig (Fig. 4). A mechanical counter tallied the revolutions. The weights were fixed to a V8 inch round stainless steel axle by machined brass bushings stabilized in the weights with methyl methacrylate. E-clips maintained the positions of the weights at the ends of the axle. A drive shaft attached the drill press motor to this axle and E-clips maintained the drive shaft in the center of the axle. The compressive apparatus imparted only vertical load (4.6 kg) by passing the load through the stainless steel rods in the superior plate of the jig to the specimens in the inferior plate. Because the rods only rested on the specimens and were unattached, no horizontal force component was transferred. Twenty-five specimens were loaded at a time for 4-hour periods at 120 loads per minute for 2 consecutive days. Specimens not being t.hermocycled or compressive-loaded were stored at 37’ C in 100 % relative humidity. Addition of the weight of the drill press to the compressive load was avoided by machining a %sinch difference in the end hole of the drive shaft and the 3/s inch axle
FEBRUARY
1994
connecting the weights. This looseness of fit allowed vertical and horizontal movement of the axle during rotation and compressive loading and negated possible drill press loads. The drive shaft of the drill press was verified as perpendicular to the Plexiglas acrylic compressive jig by alignment of a % inch hole in the center of the jig with a straight % inch rod positioned in the drill press chuck. Smooth, unbinding interaction between the rod and the compressive jig demonstrated a go-degree alignment. The cemented specimens were arranged around the compressor jig in this sequence: Panavia EX, C & B Metabond, Thin Film, Resiment, and zinc phosphate cement. The height of the rods was standardized to ensure that the loads contacted each rod for identical periods. This was accomplished by placing specimens in the jig, then positioning the superior plate over the inferior plate in proper alignment. The heights of the rods were verified by an aluminum bar 1.5 cm by 1 cm x 10 cm placed over three rods simultaneously and measurement of the middle rod with a 0.012 inch feeler gauge. Rods that were shorter than the adjacent. rods were adjusted by use of circular polypropylene shims placed under the corresponding specimens to a 0.012 inch tolerance. Rods taller than adjacent rods had the bottom of the acrylic ring shortened slightly by sanding with a bench-mounted belt sander. By these techniques, the loads were held constant and vertical to each cemented specimen.
Retentive
tests
Components and posts were subjected to tensile or retentive tests on the universal testing machine (Model 1125, Instron Engineering, Canton, Mass.) at the end of each test period. Retentive test loads were parallel to the long axis of the post/implant specimens and continuous at lmm/minute crosshead speed until fracture. The loads at failure were recorded in kilograms.
127
THE
JOURNAL
OF PROSTHETIC
DENTISTRY
GAREY
ET AL
80
Kgf.
80
12345 R
12345 P
12345 2
12345 T
12345 C
Fig. 6. Summary of Scheffe multiple comparison test: R, Resiment; P, Panavia; 2, ZnP04, T, Thin Film; C, C & B Metabond. 1, Control; 2, thermocycle; 3, load-cycle; 4, thermocycle and load-cycle; 5, blood, thermocycle, load-cycle. Horizontal lines connect statistical means that are not significantly different at p < 0.05. II. Mean dislodging forces and standard deviations in kilograms of force
Table
Resiment
Panavia
Thii
Film
C&BMETA
Exper. Group No.
n
x
SD
n
TL
SD
II
%
SD
n
iI
SD
n
%
SD
1 2 3 4 5
10 10 9 9 8
100.8 98.6 94.2 92.6 63.3
8.0 6.4 6.7 4.1 8.7
10 10 9 10 8
99.6 96.2 98.0 86.9 21.1
13.0 7.9 8.6 8.8 9.2
10 9 10 10 10
81.8 78.8 72.1 76.8 44.6
7.4 9.3 7.5 8.6 17.4
10 8 10 10 10
73.3 77.3 70.7 68.0 52.2
10.7 9.6 3.1 6.5 7.0
10 10 8 10 8
91.4 90.8 82.8 88.3 30.2
12.0 6.8 9.7 15.3 11.0
znPo4
1, Control; 2, thermocycle; 3, compression; 4, thermocycle and compression; 5, thermocycle, compression, and blood. ra, Number of specimens.
Because complete cement deterioration did not occur after thermocycling and compressive loading during preliminary tests, specimens were not examined for looseness during compressive loading.
Interface
failure
mode
One investigator visually inspected each specimen after separation and estimated the percent of the surface of the post without cement. Data for each cement were tabulated and one-way, twoway, and three-way analysis of variances (ANOVA) and Scheffe multiple comparison F-tests (p < 0.05) were computed by use of the ABsurv statistical program (AndersonBell Corp., Parker, Colo.).
RESULTS Table II lists for each cement the loads required to separate the posts from the implants after the specimens were subjected to the various tests. The data were analyzed by 128
two-way ANOVA with cement and treatment conditions as the independent variables. The differences among cements were significant at p < 0.0001 and the differences among test conditions were significant at p < 0.0001. The interaction was significant at p < 0.0001. In general, the cements diminished in strength with treatment; however, the rate of decline was amplified by the addition of blood. The rate of decline was not congruent among cements and this may explain the significant interaction. Tables III and IV summarize the ANOVA for each cement and test conditions, and Fig. 6 shows the results of the Scheffe multiple comparison tests comparing the retentive strength of each cement after each test condition. A three-way ANOVA comparing dislodging forces for the treatment groups not contaminated with blood was performed because the effect of blood contamination was great, and it was impossible to see smaller statistical differences between other variables. The analysis was constructed with type of cement, thermocycling (yes/no), and VOLUME
71
NUMBER
2
GAREY
ET AL
THE
load-cycling (yes/no) as independent variables. This indicated statistically significant effects due to type of cement (p < 0.0001) and load-cycling (p < 0.0001) but no thermocycling effect 0, := 0.42). There were no significant interactions. This three-way analyses indicated that Panavia RC and Resiment RC required significantly greater dislodging forces than Thin Film RC and C & B Metabond RC and that zinc phosphate cement was intermediate between the two groupings and not significantly different from either (Table V).
Type of failure For all of the cements and all of the test conditions, minimal cement adhered to the posts aft,er dislodgement. Thus, cement failure was primarily adhesive or at the post/ cement interface.
DISCUSSION The three-way ANOVA demonstrated a small but statistically significant difference in the various cements’ reaction to load, suggesting that load affects the strength of FEBRUARY
1994
OF PROSTHETIC
DENTISTRY
Table III. Summary of one-way ANOVA (test conditions variable) Cement
Source
Panavia
A
Resiment
B A B
Thin Film C&B Metabond
Effects of test; conditions Of the control groups that were simply cemented and stored in the incubator, those luted with Panavia RC, Resiment RC, and zinc phosphate cements required substantially greater loads to dislodge the posts than those cemented with the Thin Film and C & B Metabond RCs (Fig. 6). The control-group means were the greatest values recorded except for the C & B Metabond RC. The mean retentive strengths of the thermocycled specimens were similar to the controls but (except C & B Metabond RC) were slightly less but not significantly less at p < 0.05 (Fig. 6). After compressive loading, the retentive strength of zinc phosphate cement diminished, but it remained more retentive than the Thin Film or C & B Metabond RCs. Panavia and Resiment RCs were substantially stronger (p < 0.05) than the other cements after compressive loading (Fig. 6). The combined thermocycled/compressive loaded specimens cemented with Panavia RC, Resiment RC, and zinc phosphate cement again separated at significantly greater loads than the Thin Film and C & B Metabond RCs. Combining blood contamination with thermocycling and compressive loading dramatically weakened all the cements (Table II & Fig. 6). Resiment RC was significantly stronger than Panavia RC, Thin Film RC, and zinc phosphate cement. C & B Metabond RC was the next most retentive and significantly better than Panavia RC and zinc phosphate cement. Thin Film RC recorded significantly greater strength than Panavia RC (Table III). Blood contamination reduced the retentive strength of all of the cements more than either thermocycling or compressive loading individually or combined, as can be seen by comparing Figs. 7 and 8.
JOURNAL
&PO4
DF
Sum of squares
Mean of squares
A B A
4 3’7292.00 9322.99 42 3965.77 94.42 4 7758.00 1939.50 41 1965.65 47.94 4 9023.05 2255.76 44 5062.06 115.05 4 3534.37 883.59
B A B
43 2587.30 4 22906.30 41 5343.10
60.17 5726.58 130.32
ProbF
ability
98.7362
0.0000
40.4545
0.0000
19.6074 0.0000 14.6850 0.0000
43.9426
0.0000
A, Between groups; B, within groups.
the cements (Table V). This may be misleading in a clinical situation because the difference found was small and the data were pooled from the three treatment groups (compression only, thermocycle only and compression and thermocycle only). In the oral environment, only the compression/thermocycle group exists. The small difference discovered between dislodging forces required for loaded (83 kgf) and nonloaded units (89 kgf) may not be clinically relevant. This effect may increase with more load-cycling, but additional study would be necessary for a conclusion. The one-way ANOVAs comparing individual cements and test conditions are more definitive.‘O The control group’s means were generally the greatest values recorded (Table II). Previous investigator.@ cemented steel disks to teeth, stored them for 7 days in 37’ C water, and reported retentive strengths approximately 50% less than our findings for identical or similar cements, but differences in surface areas, in test specimens, and in storage times could account for the dissimilarities. Thermocycling resulted in a slight reduction in retentive strength compared with the control means (except C & B Metabond RC),12 but the difference was not statistically significant (Table III). Uchiyama13 cemented shallow crowns to teeth and, after 24 hours and 300 thermocycles, reported retentive strengths of 73 to 112 kg for Panavia RC. He also investigated the effects of thermocycling on dye penetration with Panavia RC, 4-META RC, and zinc phosphate cement and discovered dye penetration on both enamel and dentin with zinc phosphate cement but only slightly with Panavia or 4-META resinous cement. Other investigators have reported minimal dye penetration after thermocycling with Panavia and 4-META resinous cements.14*I5 D&-Arnold et aL8 reported no significant difference in retentive strengths of Ni-Cr-Be alloy disks cemented with Superbond C & B, Panavia, or Comspan RCs and thermocycled 300 times. In the current study, the posts and implants are both made from grade 2 cp titanium with a coefficient of thermal expansion of 27 10-6/oC (Pa129
THE JOURNAL
OF PROSTHETIC
GAREY ET AL
DENTISTRY
120 100 80 60 40 20 0
Control *R
-e-P
ThermoKomp B-Z
-CT
*C
Fig. 7. Effect of load-cycling and thermocycling on dislodging forces.
cific Titanium, San Diego, Calif.). The coefficients of thermal expansion for the cements (25-35 10e6/“C, per personal communication with manufacturers) were similar to cp titanium, which suggests that minimal dimensional change would occur during thermal shock tests. This may explain why no significant differences were found between control and thermocycled groups. Kerby et al.12 measured cemented post retention in Steri-Oss implants that had been thermocycled 1000 times and reported slightly smaller retentive values for many of the same luting agents observed in this study. The posts used in their study were not cleaned and abraded with aluminum oxide as were the posts in this study, which may account for the differences in retentive values. In addition, the crosshead speed of the Instron machine they used to dislodge the posts was 0.5 mm/minute compared with 1 mm/minute crosshead speed used in this study and this may contribute to the discrepancies. All thermocycled groups in the current study were thermocycled at 28 hours after cementation as stated in Material and Methods. Load-cycling began at 20 hours after cementation; therefore, thermocycling and load-cycling did not occur simultaneously. Group 2 was stored at 37O C, 100% humidity, for 28 hours before thermocycling and groups 3, 4, and 5 were load-cycled 20 hours after cementation, then thermocycled if required (4 and 5). Hence, group 2 was stored 8 hours longer than other groups before treatment. This design was choosen so that all of the themocycling was performed at the same time period in the curing cycle of the resins. Matsumura et a1.3considered thermocycling a technique to accelerate water-aging deterioration. However, their 130
Therm0 + Comp +R
4p
Therm0 + Comp + Blood -cz
+T
+C
Fig. 8. Effect of blood on dislodging forces.
study used 20,000 to 50,000 cycles and did not demonstrate a significant difference in retention between lesser and greater thermocycled specimens. This study selected 400 thermocycles because Crim and Franklin16 demonstrated no difference in dye penetration of crowns luted with composite resin cement when the crowns were thermocycled either 100 or 1500 cycles. Composite resin deterioration with prolonged water storage has been reported. Studies by Soderholm,r7 DiazArnold et al.,8 and Arai et al.ls suggested that tensile and transverse strength of composite resin diminished slowly and proportionally to the time immersed in water. Because posts cemented in implants are intended to remain cemented for the life of the implant, water deterioration is particularly undesirable. The inevitable cement failure should be planned for by keying the post to the implant body and making prosthetic appliances detachable. The decrease in tensile strength after compressive loading was greater than the deterioration after thermocycling, and the three-way ANOVA demonstrated a statistical difference. Thus, this reduction suggested that compressive loading may influence cement failure more than thermocycling does, especially if identical metals are luted with cements. Uchiyama13 recorded 10% to 20% reductions in shear bond for disks cemented on teeth with both 4-META RC and Panavia EX RC and then shear-loaded to failure. This study averaged 1% to 5% reduction in retention with thermocycling alone and 2% to 10% reduction with compressive loading alone. Uchiyama’s study differed from the current study in that nonprecious alloy disks were cemented to facial tooth surfaces, then the disks were loaded to failure along the long
VOLUME
71
NUMBER
2
THE JOURNAL
GAREY ET AIL
Table IV.
OF PROSTHETIC
DENTISTRY
Summary of one-way ANOVA (luting agent variable) Test condition
Source
DF
A B A B A B A B A B
4 45 4 42 4 41 4 44
Control Thermocycle Compression Compression, Thermocycle Blood, Comprewion, Thermocycle
4
39
Sum of squares
Mean of squares
5549.68 4918.10 3517.34 2676.66 5859.86 2208.06 3865.59 3992.82 9320.93 5128.55
1387.42 109.29 879.34 63.73
F
1464.96 53.86 966.40 90.75 2330.23
Probability
12.6947
0.0000
13.7979
0.0000
27.2020
0.0000
10.6495
o.oooo
17.7213
0.0000
131.49
A, Betweengroups;II, within groups.
Table V. Summary of three-way ANOVA means and standard deviation Variable Thermocycle
Cement
Load
P
R
Z
T
C
Yes
NO
Yes
No
Mean
95.10
96.71
88.61
77.33
72.05
85.58
86.34
82.66
89.20
SD
10.68 3!)
7.08 38
11.50 38
8.65 39
8.31 38
12.62 96
14.33 96
13.02 95
13.18 97
Cell N
P, Penavia; R, Resimenlt; Z’, Thin Film; C, C & B Metahond; 2, zinc phosphate. Horizontal lines connect means that are not significant statistically at p < 0.05. This analysie does not include blood contamination data.
axis of the teeth. The current study used titanium posts cemented inside titanium implants, and the separation forces were primarily shear. However, because the posts were cemented inside the implants, friction between the posts and implants increased the force required for cement failure. With all treatments except blood contamination, Panavia and Resiment RCs were significantly stronger than other cements in the current experiment (Table III). After compressive loading, the retentive strength of zinc phosphate cement diminished, but it continued to be more retentive than Thin Film RC or C & B Metabond cement. Kakigawa et al.lg demonstrated that 66,400 cyclic loads with 7.5 kg decreased the tensile strength of Panavia EX RC by 36% (78 kg to 50 kg) on crowns cemented to prepared teeth. They reported that Superbond C & B RC exhibited greater retention than Panavia EX RC. After the combination of thermocycling and compressive loading, the specimens in the current study cemented with Panavia RC, Resiment RC, and zinc phosphate cement required substantially greater loads to be separated than specimens cemented with Thin Film and C & B Metabond RCs. Compared with the controls, a 3 % to 13 % reduction was recorded in the tensile strengths of the cements studied, but the deterioration was not statistically significant. Nakamura et aLI compared resinous and conventional ce-
FEBRUARY
1994
ments for luting inlays to natural teeth with 4-META cement and Panavia EX RC and discovered a similar nonsignificant reduction in strengths. Using cyclical loading (0 to 30 kg) of approximately lo7 times and thermocycling (4” to 60” C) lo3 times, they deceased inlay retention by approximately 10 % , similar to the values in this experiment, suggesting that elevation of the number of stresses may not significantly weaken 4-META or Panavia EX RCs. Combining cyclic compressive loading and thermocycling appeared to diminish the retentive strength of the cements more than either cyclic loading or thermocycling alone (Table II), which suggests that there may be an additive effect of compressive loading and thermocycling to weaken cements, particularly Panavia Ex, Resiment, and C & B Metabond resinous cements. Blood contamination with thermocycling and compressive loading substantially weakened all of the cements (Tables II and III). Resiment RC was significantly stronger than Panavia RC, Thin Film RC, and zinc phosphate cement. C & B Metabond RC was the next most retentive and was significantly stronger than Panavia RC and zinc phosphate cement; and Thin Film RC was significantly stronger than Panavia RC (Table III). Blood contamination reduced the retentive strength of all of the cements more than thermocycling or compressive loading individually or combined. In this study, this variable was the most 131
TEE
JOURNAL
OF PROSTHETIC
important factor affecting retentive strength of cements to lute posts to implants, with Panavia RC and zinc phosphate cement the most affected (Table III). However, all of the cements tested showed significant decreases in tensile strength with blood contamination. C & B Metabond RC displayed the least percentage of change, but it was consistently the weakest cement tested, except with blood contamination. Resiment RC recorded the greatest retentive strength, but it was notstatistically better than C & B Metabond RC. Cement interface failure is defined as the surface or area of cement failure. Adhesive failure is cement fracture at the post/cement interface or surface, and cohesive failure implies fracture within the cement. In the nonblood-contaminated specimens, the cement failure occurred primarily at the post/cement interface, indicating that the bond of cement to the post was weaker than the physical strength the cement. Nearly all blood-contaminated samples reflected cohesive cement failure. Presumably, blood contamination decreased the physical properties of the cements. Further study is recommended to determine the exact role of blood on the resinous cements. The cement/implant interface was deliberately omitted in the discussion because the implant was threaded for mechanical retention and did not require adhesion, The posts in the current study were reused. A pilot study suggested no significant differences in retentive values of posts that were repeatedly reused when the posts were cleaned and abraded as mentioned in Material and Methods. Posts not treated in this manner may have different retentive values. The clinical implication of this study is that blood contamination during cementation of implant posts could drastically reduce the strength of the cement and should be avoided. This suggests that blood contamination during cementation of any FPD would result in a weakened cement and the amount of weakening depends on the cement. CONCLUSIONS This study using an Instron machine indicated that thermocycling did not significantly reduce retentive strength of the test cements, but that cyclical compressive loading did. However, the decrease is small and may not be clinically apparent. The combination of thermocycling, cyclical load-stressing, and blood contamination substantially reduced the retentive strengths for all of the cements. This suggests that blood adversely affects the retentive strength of the cements tested more than other variables. Resiment RC exhibited the greatest mean retentive strength with blood contamination but it was significantly less than that of the Resiment RC control.
132
CAREY
DENTISTRY
ET AL
REFERENCES 1. Takeyman M, Kashibuchi N, Nakabayashi N, Mashuhara E. A research on self-curing resins (no. 17). Adhesive resin to tooth substance and dental alloys. J Jpn Sot Dent Appar Mater 1978;19:179-84. 2. Masuhara E, Kojima K, Kimura T. Studies on self-curing resins (2). Effect of alkylboron on the polymerization of methyl with benzoly peroxide. J Res Inst Dent Mater, Toyoko Med Dent Univ 1962;2:369-74. 3. Matsumura H, Yoshida K, Tanaka T, Atauta M. Adhesive bonding of titanium with a titanate coupler. J Dent Res 1990;69:1614-6. 4. Brazilay I, Myers ML, Cooper LB, GraserGN. Mechanical and chemical retention of laboratory cured composite to metal surfaces. J PROSTHET DENT 1966;59:131-7.
5. Imhery TA, Burgess JO, Naylor WT. Tensile strength of resin cements following various alloy surface treatments [Abstract]. J Dent Res 1991;70(special issue):399 6. Wada T. Properties of an adhesive resin. Dent Di,~ond 1984;9:102-9. 7. Omura I, Yamauchi J, Wada T. Adhesive and mechanical properties of anew dental adhesive [Abstract]. J Dent Rea 19&1,63(specialissue):233. a. D&-Arnold A, Williams WV. Aguilino S. Tensile strengths of three luting agents for adhesion fixed partial dentures. Int J Proethodont 1989;2:115-22. 9. Wacker DR, Tjan AHL. Effect of variation in powder-to-liquid ratio of zinc phosphate cement on the retention of posta. J PROWHET DENT 1988;60:49-52.
10. Conover WJ. Practical non-parametric statistics. New York: John Wiley & Sons, 1980;168. 11. Akau MN, Lorey RE. Tensile retention of posts using various cementa/ bonding agents [Abstract]. J Dent Res 1999;69:86?. 12. Kerhy R, McGlumphy EA, Holloway J. Some physical properties of implant abutment luting cements. Int J Prosthodont 1992;5:321-5. 13. Uchiyama Y. Clinical study on the retentive force of adhesive resins. J Jpn Prosthodont Sot 1986;30:198-294. 14. Nakamura K, Takatau T, Hoaoda H. Retentive characteristics and durability of adhesive resin luting cements for internal inlays. In: Okabe T, Takahashi S, eds. Transactions of International Congress on Dental Materials. Honolulu: The Academy of Dental Materials and The Japanese Society of Dental Materials and Devices, 1989;5. 15. Tjan AHL, Dunn JR, Grant BE. Microleakage of cast crowns cemented with a non adhesive resin [Abstract]. J Dent Res 1990,69:123. 16. Crim GA, Franklin G. Microleakage: the effect of storage and cycling duration. J PROSTHETDENT 1987;57:574-6. 17. Soderholm KJ. Degradation of glass 6ller in experimental composites. J Dent Ree 1981;69:1867-75. 18. Arai K, Fujii K, Hashimoto H. Influence on physical and mechanical properties and cause of disintegration for posterior composite resins by long-term water immersion. In: Okabe T, Takahaahl S, eds. Transactiona-International Congress on Dental Materials. Honolulu: The Academy of Dental Materials and The Japanese Society of Dental Materials and Devices, 198953. 19. Kakigawa H, Yokoyama Y, Tajima K, et al. Effect of repeated loading on retention of the crown fixed with adhesive resin cements. In: Okabe T, Takahashi S, eds. Transactions-International Congress on Dental Materials. Honolulu: The Academy of Dental Materials and The Japanese Society of Dental Materials and Devices, 1989,6. Reprint requests to: DR. DANIEL J. GAREY SCHOOL OF DFSTISTFIY DEPARTMENT OF RESTORATIVE DENTISTRY LOMA LINDA UNNEIWTY LOMA LINDA, CA 92350
Contributing Author Steven Prim, CDA, Instructor, Oral Implantology Laboratory, Department of Restorative Dentistry, Loma Linda University, School of Dentistry.
VOLUME
71
NUMBER
2