Abstract
Objective: Objectives of this study were to (1) compare the mean shear-peel bond strength of orthodontic bands luted to porcelain molar denture teeth with glass ionomer cement (GIC), resin-modified glass ionomer cement (RMGIC), or compomer cement; (2) assess the amount of cement remaining on the teeth after debanding; and (3) compare the survival times of the cemented bands subject to mechanical fatigue.
Materials and Methods: Sixty banded denture teeth (20 per cement group) were used to determine shear-peel bond strength, and 30 banded denture teeth (10 per cement group) were used to determine fatigue survival time. Shear-peel bond strength was determined with a universal testing machine, and groups were compared by one-way analysis of variance. The amount of cement remaining on the teeth after band removal was scored, and a chi-square test was used to compare groups. Fatigue testing was conducted in a ball mill, and a log-rank test was used to compare differences in survival times.
Results: No differences were found in mean shear-peel bond strength among the three groups. The amount of cement remaining on the teeth varied between the compomer and GIC groups (P = .01), with more compomer cement remaining relative to GIC. The mean survival times of bands cemented with compomer or RMGIC were longer than for bands cemented with GIC (P < .001).
Conclusion: The findings show that on porcelain teeth the band cements have comparable mean shear-peel bond strengths, but that band retention with RMGIC and compomer cement are superior to GIC when subjected to simulated mechanical fatigue.
INTRODUCTION
Orthodontic fixed appliances frequently include stainless steel bands secured to posterior teeth by a combination of close adaptation to teeth and an interposing layer of luting cement. Numerous laboratory studies have been used to investigate the bond strength and the fatigue strength of band cements on natural teeth.1–7 A review of the literature by Millett and associates3 found that the clinical failure rate of bands cemented on natural teeth with glass ionomer cement (GIC) ranged from 6% to 26% over observation periods ranging from 12 to 24 months. With more adults seeking orthodontic treatment, many of whom have full-coverage porcelain crowns, the performance of band cements when used to secure bands to porcelain molar crowns needs evaluation.
GIC is particularly useful in dentistry because of the release of fluoride.8 Relative to zinc phosphate, GIC has higher tensile and compressive strengths, decreased brittleness, and lower oral solubility.9 More recently developed cements include resin-modified glass ionomer cement (RMGIC) and compomer cement. Fluoride release and reuptake occurs with RMGIC in a range similar to that of conventional glass ionomer.10,11 Compomers, also known as acid-modified composite resin,12 modified composites,6 or polyacid-modified composite resins,5 are essentially resin composites that slowly release fluoride and have physical properties very similar to resin composite.13
Laboratory investigations of force values required to remove cemented bands from natural teeth show that RMGIC and GIC are generally equivalent, whereas fatigue studies show RMGIC has superior fatigue properties.6,7 Studies of the retentive strength of bands cemented with compomer cements have demonstrated mixed results, with some investigators finding that compomers perform equivalent to GIC and RMGIC5,7 and others finding lower retentive strength than either RMGIC or GIC.6 The site of cement failure has been shown to occur primarily at the band-cement interface with GIC,1,4,7 whereas failure is more common at the cement-enamel interface for RMGIC4,6,7 and compomers.7
The aims of the present study were to (1) compare the shear-peel bond strength of microetched orthodontic bands cemented to porcelain crowns with one of three orthodontic band cements: GIC (Ketac-Cem, 3M ESPE, St Paul, Minn), RMGIC (Multi-Cure, 3M Unitek, Monrovia, Calif), and compomer (Transbond Plus, 3M Unitek); (2) evaluate the site of bond failure with each cement type and to compare the amount of cement remaining on the crown after debanding; and (3) compare the survival time of bands with each cement type after simulating mechanical fatigue stress in a ball mill.
MATERIALS AND METHODS
To study shear-peel bond strength, 60 anatomic porcelain mandibular first molar denture teeth (Trubyte bioform 33°, size 34 M, Dentsply, York, Pa) were equally divided into three groups. The apical base of the teeth were prepared for mounting by using a bur to increase the undercut of the diatoric hole, removing the glaze with 9.5% buffered hydrofluoric acid, and applying silane (Ultradent Porcelain Etch and Silane, Ultradent Products, South Jordan, Utah) and an adhesive primer (Transbond XT Light Cure Adhesive Primer, 3M Unitek). The teeth were mounted on blocks of Triad Colorless TruTray Material (Dentsply). The denture teeth had a large undercut along the lingual-gingival surface that was embedded in the Triad material to prevent the undercut from acting as a mechanical lock for the cements.
First molar bands with microetched fitting surfaces (3M Unitek) were adapted to each tooth. Ketac-Cem, Multi-Cure, and Transbond Plus (Table 1) were used for band cementation according to manufacturers' instructions. Bands cemented with Transbond Plus and Multi-Cure were light cured at 450 mW/cm2 from the occlusal surface for 30 or 40 seconds, respectively. Once cured, the specimens were transferred immediately to containers with 100% humidity and placed in an oven maintained at 37°C for 24 hours before testing.
The specimens were subjected to a shear-peel debanding force with an Instron testing machine (Model TT-B, Instron Engineering Corporation, Canton, Mass) in tensile mode with a crosshead speed of 12.5 mm/ min. Two 0.9-mm (0.036-inch) stainless steel wire loops engaging the buccal tube and the lingual cleat were held in grips attached to the load cell with a universal joint. The universal joint allowed for movement in three planes of space (Figure 1). Testing proceeded until the band was removed completely from the tooth. The maximum debanding force (in pounds) recorded on the force/time curve was converted to a shear-peel bond strength value (MPa) by dividing the maximum force by the band surface.
A sample secured in the Instron testing machine for a shear-peel bond strength test
A sample secured in the Instron testing machine for a shear-peel bond strength test
The surface area provided by the manufacturer (128.387 mm2) was adjusted because of the short lingual surface of one of the denture teeth crowns, resulting in a portion of the band being exposed above the occlusal surface of the tooth. To determine percentage of surface area contacting the crowns, for five bands the buccal and lingual attachments were removed and the bands were weighed. Band material above the occlusal surface was removed with a bur, and the mean change in weight of the bands was measured.
Immediately after the deband testing, specimens were visually assessed to determine the site of cement failure and were classified by the Adhesive Remnant Index (ARI) as originally proposed by Årtun and Bergland14 for bonded brackets and adapted for cemented bands by Millett et al.6 Scoring was as follows: 0, no cement remains on the tooth surface; 1, less than half of the tooth surface is covered by cement; 2, more than half of the tooth surface is covered by cement; 3, all of the tooth surface under the band is covered by cement.
To test fatigue survival time, 30 porcelain teeth were divided into three groups of 10. With each group, one of the cements was applied to the entire inner surface of the band, which was then seated flush with the crown's mesial and distal marginal ridges. The banded teeth were stored for 24 hours at 37°C in 100% humidity before testing according to the protocol of Abu Kasim et al.15 For each group, all specimens were placed in 140-mL capacity ceramic containers with 130 g of 1.2- × 1.2-cm ceramic pellets and 70 mL of water (37°C). The containers were rotated on the ball mill for consecutive 1-hour intervals (Figure 2; Norton Plastics and Synthetics Division, Akron, Ohio). The teeth were inspected after each hour; those with loose bands were removed, and testing was resumed until all cemented bands had loosened.
The ball mill with rollers that rotated two ceramic cylinders containing bands cemented on denture teeth, water, and ceramic pellets. The inset photo shows the relative size of a ceramic pellet to a banded tooth specimen
The ball mill with rollers that rotated two ceramic cylinders containing bands cemented on denture teeth, water, and ceramic pellets. The inset photo shows the relative size of a ceramic pellet to a banded tooth specimen
Mean shear-peel bond strengths of the three cements were compared by one-way analysis of variance. A chi-square test was used for comparison of ARI scores. For the ball mill experiment, Kaplan-Meier estimates of the survivor functions and the log-rank test were used to compare fatigue survival time distributions. Statistical tests were conducted by using SPSS (SPSS Inc, Chicago, Ill). All tests were made at the significance level of α ≤ .05.
RESULTS
Shear-peel bond strength tests showed no significant differences among cement groups (P = .214). Weight measurements of the bands with the lingual area extending beyond the occlusal surface removed showed the mean surface area in contact with the tooth was 85.5% ± 3.4% of the original surface area. Adjusted values for mean shear-peel bond strengths are shown in Table 2. During the tests, several teeth pulled out of the mounting before the band was removed, including three in the Ketac-Cem group, two in the Multi-Cure group, and two in Transbond Plus group, leaving sample sizes of 17, 18, and 18 teeth, respectively.
Sheer-Peel Bond Strength Values for Mandibular Molar Bands Cemented With Ketac-Cem, Multi-Cure, or Transbond Plus

The most common site of failure for all cement groups was at the porcelain-cement interface (ARI scores of 0 or 1; Table 3). A significant difference in the distribution of ARI scores was found between the Ketac-Cem and Transbond Plus groups (P = .011), with more cement remaining on the teeth with Transbond Plus than with Ketac-Cem. Ketac-Cem samples failed exclusively at the porcelain-cement interface with no residual cement remaining on any teeth (ARI score of 0). Multi-Cure and Transbond Plus samples received an ARI score of 1 with considerably less than 50% of the crown surface covered by cement (Figure 3). No significant differences in ARI scores were found between Ketac-Cem and Multi-Cure groups or between Multi-Cure and Transbond Plus groups.
Distribution of Adhesive Remnant Index (ARI) Scores for Bands Cemented with Ketac-Cem, Multi-Cure, and Transbond Plus

Lingual (L) and buccal (R) views of a representative Transbond Plus sample showing cement remaining after a shear-peel bond strength test (circled). The sample received an Adhesive Remnant Index score of 1
Lingual (L) and buccal (R) views of a representative Transbond Plus sample showing cement remaining after a shear-peel bond strength test (circled). The sample received an Adhesive Remnant Index score of 1
Ball mill test results showed the mean fatigue survival time of bands cemented with Transbond Plus and Multi-Cure was significantly longer (4.6 hours and 5.4 hours) than with Ketac-Cem (2.3 hours; P < .001; Figure 4). There was no significant difference in mean survival time between bands cemented with Transbond Plus or Multi-Cure. It was not possible to determine the location of cement failure resulting from the fatigue test because the cement was dislocated from both the teeth and bands as the specimens tumbled.
Fatigue test results showing hourly survival distribution of the molars bands in each of the three cement groups
Fatigue test results showing hourly survival distribution of the molars bands in each of the three cement groups
DISCUSSION
In the present study, the mean shear-peel bond strength did not differ significantly between cement groups. Although we are unaware of other studies that have investigated the cements on porcelain teeth, the cements have been investigated on natural teeth. Similar to our results, Millett and associates7 found no significant differences in shear-peel bond strength of bands cemented with RMGIC (Multi-Cure) vs GIC (Ketac-Cem), and Aggarwal and collegues5 found no difference in mean shear-peel bond strengths when comparing RMGIC (Multi-Care) with compomer (Transbond Plus). In contrast to our results, Millett and associates6 found that bands cemented on natural teeth with compomer (Transbond Plus) had a significantly lower mean shear-peel bond strength compared with GIC (Ketac-Cem).
The force per unit area values (MPa) for all three cements in this study were generally lower than values obtained in investigations on natural teeth.4,6,7 Mean values for GIC (Ketac-Cem) on natural teeth as determined in three studies ranged from 1.27 to 1.65 MPa,4,6,7 or about two times greater than in the present study (0.74 MPa). Millett and associates7 determined the shear-peel bond strength of bands cemented with RMGIC (Multi-Cure) to be 1.63 MPa, again approximately two times greater than in the present study (0.69 MPa). In contrast, Millett and associates6 found a lower value for compomer (Transbond Plus) on natural teeth (0.415 MPa) in comparison with the present study (0.89 MPa). The standard deviations in all studies included the present range between 20% and 50% of the mean shear-peel bond strength value, indicating wide individual sample variation.
Multiple factors likely contributed to the lower mean shear-peel bond strength values measured on porcelain teeth compared with values found on natural teeth. One factor may be the shape of the denture teeth where the buccal and lingual surfaces of the denture tooth had little if any convexity, thus providing minimal mechanical retention. Other considerations are the surface characteristics of porcelain such as greater smoothness and lower surface tension relative to enamel. In contrast to porcelain, chemical bonding to enamel may play a role in the adhesion of GIC and RMGIC to natural teeth.8 This factor may also explain why no differences were found among the cements tested in this study as opposed to the results of investigations on natural teeth showing bands cemented with compomer (Transbond Plus) had lower shear-peel bond strength relative to RMGIC and GIC.6 Nevertheless, Millett and associates7 did not find differences in bond strength when comparing the compomer Ultra Band Lok with both GIC (Ketac-Cem) and RMGIC (Multi-Cure), suggesting that the composition of the different brands of compomer cement may also influence the results. Although it is typical in dentistry to enhance adhesion of composite based cements such as compomers to porcelain with hydrofluoric acid etching, this typically is not performed during orthodontic banding.
Visual assessment of the teeth after band removal showed that minimal amounts of cement were retained on the surface of the teeth and that the microetched surface of the band was more retentive. It is likely that failure occurred almost exclusively at the porcelain-cement interface in the present study because (1) microetched bands were used, (2) there is no chemical adhesion by these cements to porcelain, and (3) the highly polished surface of porcelain teeth provides minimal surface area for mechanical adhesion.
Use of fatigue survival time testing in conjunction with shear-peel bond strength testing provides further information on how cemented bands hold up to externally applied mechanical stress.2–4,6,7,15 The method of using a ball mill to fatigue test has been shown to provide reproducible results in a short period of time that are consistent with clinical performance of cements.3,15 Although the method provides a simplified model of clinical conditions, a limitation is that the fatigue test does not provide a quantitative assessment of the loads developed. The most likely process resulting in failure is slow development of a crack through prechipped cement and ongoing destabilization resulting from the mechanical action of the pellets.15
In the present study, though ceramic pellet size, shape, and the setting of the ball mill (in revolutions per minute) varied somewhat from other investigations, generalized comparisons can be made regarding survival time. Our results, showing that RMGIC and compomer cement had approximately twice the survival time of GIC, are comparable with other studies where GIC was consistently found to have the shortest survival time.4,6,7 However, one study comparing compomer (Transbond Plus) with GIC (Ketac-Cem) found no statistically significant difference in survival times,6 but the survival time of 9.9 hours for Ketac-Cem was approximately three times longer than what was found in the present or other investigations.4,7 Other investigators have also noticed this discord in results for GIC relative to the other two types of cements in shear bond vs fatigue tests.2,7 It is possible that fatigue testing results could be improved if GIC were tested after greater lengths of time than the 24-hour interval as used in the present investigation, where the added time may allow for further setting.16
Results of the present study suggest that with porcelain crowns, RMGIC (Multi-Cure) and compomer (Transbond Plus) can be expected to have lower clinical failure rates for band cementation than GIC (Ketac-Cem). Further studies would provide additional data that orthodontists could use for cement selection. Clearly, as mentioned above, a prospective clinical trail would be ideal for comparing cements. In addition, clinical studies could be used to determine if porcelain surface treatments, such as with hydrofluoric acid, would be justified. Regarding in vitro studies such as the present study, because bands are used clinically for months to years, it would be useful to follow cement retention characteristics beyond the initial 24 hours. Last, a limitation in the present study was that molar bands were cemented on dry teeth, an environment that is often difficult to achieve clinically.
Investigations of band cements have shown that moisture contamination by saliva can affect adhesion and failure site. For example, Aggarwal and colleagues5 studied the shear-peel bond strength of bands in the presence or absence of saliva contamination during cementation and found that saliva was detrimental to band retention when using compomer (Transbond Plus), whereas their results suggested that some salivary moisture enhances RMGIC and GIC adhesion. Thus, effects of moisture on strength of the cements would provide additional useful clinical information.
CONCLUSIONS
No significant difference was found between the mean shear-peel bond strengths of bands cemented on porcelain teeth with GIC (Ketac-Cem), RMGIC (Multi-Cure), or compomer (Transbond Plus).
The amount of cement remaining on porcelain teeth after debanding was significantly lower for GIC (Ketac-Cem) than for compomer (Transbond Plus), but both were equivalent to RMGIC (Multi-Cure).
Mean fatigue survival times of bands cemented on porcelain teeth with either compomer (Transbond Plus) or RMGIC (Multi-Cure) were equal and significantly longer than for bands cemented with GIC (Ketac-Cem).
Results of the present study relative to previous studies using natural teeth suggest that in a clinical setting the cements on porcelain teeth would provide less favorable retention of orthodontic bands.
REFERENCES
Author notes
Corresponding author: David A. Covell Jr, PhD, DDS, Department of Orthodontics, Oregon Health & Science University, 611 SW Campus Dr, Portland, OR 97239-3097 ([email protected])