Several reviews have been devoted to the amalgam-bonding technique. Gwinnett and colleagues¹ reviewed the advantages and disadvantages. They cited the most important advantages of bonding amalgam as the conservative preparation it requires, the reduced marginal gap and postoperative sensitivity it produces owing to sealing by resin adhesives, its reduction of secondary caries and its improvement of restored teeth’s resistance to fracture. Its disadvantages are the increased time required to perform amalgam restorations (compared with the time required to perform a conventional restoration), its sensitivity to technique and its higher cost. More recently, Setcos and colleagues^2,3 reviewed studies of bonded amalgam and concluded that although the in vitro studies generally showed amalgam bonding as having the advantage, clinical studies usually show only equal, not superior, results as compared with conventional preparations.

Modern dentin bonding agents generally incorporate some acidic treatment of dentin. This technique has evolved considerably since Fusayama and colleagues⁴ introduced etching with phosphoric acid, a process called "total etching," as being the best treatment of prepared dentin and enamel before application of bonding agents. They demonstrated that this method is biocompatible and significantly increases adhesion of composites to tooth structures because of better penetration of adhesives to dentin tubules and, thus, better protection of the pulp. When amalgam bonding with resins was developed, modern dentin-bonding agents were just becoming popular as a result of improvements, so the amalgam-bonding technique was used with these new adhesives for both dentin and enamel.^2,3

Bonded amalgam restorations may be useful for large lesions in posterior teeth, as well as for teeth that have low gingivalocclusal height, because such restorations eliminate retention pins and their inherent risks (pulp or periodontal perforation). This technique also seals the pulpal-dentinal complex biologically.

We conducted a study to evaluate bonded amalgam restorations and compare them over a five-year period with nonbonded amalgam restorations in terms of anatomical form, marginal adaptation, sensitivity, surface quality and secondary caries. We evaluated these items in a convenience sample of bonded and nonbonded restorations.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Table 1 summarizes the materials used in this study. We chose the adhesive resin All-Bond C&B (Bisco Inc., Schaumburg, Ill.) because it demonstrated high bond strengths in a previous in vitro study.⁵ It is a medium-viscosity luting resin with metal bonding capabilities that works well for amalgam bonding, although it is not specifically formulated or marketed for this purpose. ANA 2000 (Nordiska Dental AB, Angelholm, Sweden) is a dispersed-phase type of high copper amalgam.

The treating dentists acquainted the patients with the aims and course of this clinical study, and the patients signed the statement of informed consent (drafted according to the World Medical Association Declaration of Helsinki⁶). The dentists took pretreatment photographs of each quadrant to be treated; they used local anesthetic in a routine fashion; they prepared the teeth; and they took polyvinyl siloxane (two-step putty-wash) impressions of preparations. They kept replicas of the preparations for future reference, in case there was any question about the mechanical retention of the preparations.

Most of the restorations were placed in conventional carious lesions because they were replacements for unsatisfactory existing restorations, and they were prepared by one of three operators who had no prior knowledge of which restorative technique was to be used. The majority of these were Class II restorations, which included a dovetail and an isthmus, but no accessory retention such as proximal grooves. After preparation, the dentist decided by a coin toss whether the lesion would be restored with or without adhesive. He or she made six preparations for small proximal lesions that did not involve the occlusal surface. These were prepared as proximal slots: flared box forms with no isthmus or dovetail and no retention grooves. Because these preparations contained no undercuts at all, the dentist assigned these lesions to the adhesive group without a coin toss. This is one of the reasons the two groups have different numbers of restorations. The other reason is simply the result of the coin toss, which did not result in an exact 50/50 split.

The lack of undercuts in these preparations was confirmed later by examination of the replicas of the preparations under a dissecting stereo microscope by two dentists. Although these restorations are included in the adhesive group, they were recorded as nonretentive preparations in case of retention failure. However, we observed no retention failures during the five-year period.

Within one month after placement, all restorations were polished and initially clinically evaluated, including making photographs and impressions (baseline). We compared the results of the subsequent evaluations to these baseline data. At each recall examination, the examining dentist cleaned the restorations and the teeth with a rotary brush. Two trained dentists carried out the clinical evaluation with a mirror and explorer, using the modified U.S. Public Health Service, or USPHS, criteria for amalgam restorations.⁷ Both dentists examined each restoration, and they resolved any disagreement of more than one step and reached a consensus. The box lists the criteria used in this study. Photographs and impressions also were taken at each recall visit.

– Nonbonded amalgam restorations (control group, n = 62). After performing a conventional preparation, the dentist partially lined lesions with hard-setting calcium hydroxide, or Ca(OH)₂, -based liner only in places close to pulp. After the liner (if it was used) set, the dentist condensed freshly mixed amalgam into the lesion and carved it into anatomical form.

– Bonded amalgam restorations (experimental group, n = 77). After performing a conventional preparation, the dentist lined lesions with hard-setting Ca(OH)₂-based liner only in places close to pulp. He or she positioned matrix bands with a thin wax coating and wedged them as needed. After the liner (if it was used) set, he or she etched the enamel and the dentin for 15 seconds with 32 percent phosphoric acid, then washed the tooth with air/water spray. The dentist gently air-dried the lesion, leaving the surface visibly moist, then applied five coats of primer and air-dried that for 15 seconds. He or she applied adhesive cement in a thin layer with a brush onto the lesion walls. He then immediately condensed freshly mixed amalgam and carved it into anatomical form.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Table 2 summarizes the numbers of restorations available for evaluation at each recall visit. After two years, one bonded restoration was lost from the study because the tooth became a bridge abutment. After three years, one tooth with a bonded restoration had to be endodontically treated because of secondary caries penetrating into the pulp. The C rating for secondary caries appears in Table 3 and is part of the statistical analysis comparing the rates of secondary caries. This restoration was included in the statistical analysis comparing the rates of secondary caries. After four years, one tooth with a bonded restoration was extracted because of advanced periodontitis, one tooth with a bonded restoration had to be endodontically treated because of prosthetic treatment, and one control tooth was endodontically treated because of primary cervical caries penetrating into the pulp.

Table 3 shows the percentages of A ratings of the bonded (experimental) and nonbonded (control) restorations at the five-year recall. This table includes the bonded restorations in the non-retentive preparations together with the bonded restorations in conventional lesions. This was done because none of them failed catastrophically, and we assumed that lack of mechanical undercuts of the lesion would not affect the other ratings. The data in Table 3 show the arithmetical means (in percentages) of two evaluators for each criterion. Statistical analysis (Fisher exact test) has shown no significant differences (P < .05) between bonded and nonbonded restorations in all the evaluated criteria at the five years recall after placement.

Examples of restorations are shown in Figures 1 and 2. Figure 1A shows two nonbonded (control) restorations at the baseline evaluation, and Figure 1B shows the same restorations at the five-year recall. Both restorations were rated A in all aspects at baseline and the five-year recall, except for the marginal adaptation at the proximal at the five-year recall, which was rated B for both. Figure 2A shows three adhesive restorations at the baseline evaluations, and Figure 2B shows the same restorations at the five-year recall. These restorations were rated A in all aspects at baseline and at the five-year recall.

View larger version (98K): [in this window] [in a new window]   Figure 1. A. Two nonbonded (control) restorations at the baseline evaluation, rated A in all aspects. B. The same restorations at the five-year recall. All ratings were A except for proximal marginal adaptation, which was rated B. (For description of ratings, see box.)

View larger version (56K): [in this window] [in a new window]   Figure 2. A. Three bonded restorations at baseline. All ratings were A. B. Three bonded restorations at the five-year recall. All ratings were A. (For description of ratings, see box.)

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Our study compared bonded and nonbonded amalgam restorations in conventional preparations. The only difference between these two groups was the use of the adhesive, because the decision to use one or the other technique was made after the preparation was completed. Six of the bonded restorations were placed in nonretentive preparations. Although we would have liked to have a large group of such restorations, the limited availability of lesions lending themselves to such preparations and limited resources for the study did not allow it. Ideally, such a group should be tested against a nonretentive, non-bonded control group. However, such restorations have been shown to fail very quickly in laboratory studies.^21,22 The nonretentive occlusal slot preparation has not been used in previous clinical studies, only in clinical technique articles or in laboratory studies such as those of Staninec²¹ and Eakle and colleagues.²² The studies reported that amalgam placed in a nonretentive occlusal slot with adhesive resists dislodgement by simulated occlusal forces as well as or better than amalgam retained by traditional undercuts such as dovetails or proximal grooves. The five-year success of these few nonretentive restorations in our study provides support for the previous in vitro results. When bonding is used, it may be possible to dispense with traditional retention forms such as grooves, dovetails and isthmuses. Our nonretentive preparations were very conservative, removing virtually only caries and little additional healthy tooth structure for access to the caries. More studies with larger numbers of such restorations and longer follow-up times are needed to further validate this concept.

When bonding is used, it may be possible to dispense with traditional retention forms such as grooves, dovetails and isthmuses.

The reported success of the amalgam sealants also supports the idea that undercuts are not necessary to retain bonded amalgam restorations.²³ These were placed with no mechanical preparation at all, and after two years there was no difference between fissure sealing with resin-based composite sealant (All-Bond 2/Liner-F, Bisco Inc.) and fissure sealing with bonded amalgam. The bonding technique used in that study was similar to the present study, except for the choice of resin adhesive.

Bonded amalgam restorations may have other advantages, predicted from in vitro studies^{1–3,5,21,22}—namely, inhibition of leakage, inhibition of secondary caries, decrease in sensitivity, improved marginal integrity and reinforcement of tooth structure. However, the results of our study do not necessarily show any of these advantages. Leakage was not measured directly, but it generally is assumed to be related to sensitivity and secondary caries. Sensitivity was universally low, and we assessed marginal integrity and found it to be similar to that of nonbonded restorations. We found only one true secondary carious lesion, which is not enough to show a significant difference or draw any conclusions about different rates of secondary caries. To test differences in secondary caries rates, we would need to observe larger numbers of teeth in patients with high caries rates and follow them for longer periods.

We did not note any tooth fractures, so no conclusions can be drawn about any possible reinforcement of tooth structure. The lack of differences can be attributed to the fact that the traditional technique for restoring teeth with amalgam works well in conventional preparations, with relatively trouble-free results. Sequelae such as secondary caries or cusp fractures are sufficiently rare that they do not occur in large enough numbers to show a statistically significant difference in reported clinical studies.^{9–16,18–20} A larger study cohort and a longer period of observation would be necessary to show any differences that might exist.

Published clinical studies of bonded amalgam restorations used conventional preparations and generally yielded results similar to ours. The study by Setcos and colleagues¹⁸ of a total of 113 restorations showed no significant differences between control restorations and restorations bonded with ED Primer + Panavia 21, J. Morita USA, Irvine, Calif.). This study evaluated the anatomical form, marginal adaptation and surface quality. Dondi Dall’Orologio and colleagues¹⁰ used four USPHS criteria: retention, marginal adaptation, postoperative sensitivity and presence of secondary caries after 18, 24 and 48 months (N = 260). That study used nonretentive box forms similar to those used in our nonretentive group. They assessed the technique of bonding amalgam to be efficacious. Kennington and colleagues⁹ investigated postoperative sensitivity and found no significant difference between the two kinds of restorations (N = 20). Browning and colleagues¹⁴ reported that bonded restorations (using Optibond, SDS Kerr, Orange, Calif.) and nonbonded restorations were clinically acceptable even after 18 months. They used three criteria: presence of secondary caries, anatomical form and marginal adaptation (N = 30). Ölmez and colleagues¹⁶ compared bonded amalgam restorations (using Amalgambond Plus, Parkell, Farmingdale, N.Y.) with composite restorations made on primary molars, and they found out no significant difference in retention, marginal adaptation, presence of secondary caries and postoperative sensitivity. They indicated Amalgambond Plus as a suitable bonding agent for amalgam in large cavities. Summitt and colleagues¹⁵ compared the bonding technique (using Amalgambond Plus) with restorations using pins on teeth with large lesions including cusps and reported 100 percent retention, with secondary caries, marginal adaptation, marginal discoloration and vitality being equal for restorations accomplished with both techniques after five years. However, they found higher postoperative sensitivity in the pin group within six months after placement. That study also supports the idea that mechanical retention is not necessary when adhesives are used, because many preparations in the adhesive group had no undercuts to retain the amalgam restorations.

Mahler and Engle¹⁹ compared bonded and non-bonded amalgam in conventional preparations and found no differences. They stated that because the results of the two techniques are comparable, the bonding technique is not desirable because it is more expensive and time-consuming. However, they evaluated only two criteria, postoperative sensitivity and marginal adaptation, as evidenced by marginal fracture using the method of comparison with a set of standard photographs. Their study included only conventional, retentive preparations, so they were not able to show any advantages of the adhesive technique in terms of retention.

The adhesive resin used in our study, All-Bond C&B, actually is formulated and marketed as a luting cement, not as an amalgam-bonding resin. We used it because a previous laboratory study showed that it has good bonding strength to both spherical and dispersed-phase amalgam.⁷ We chose the dispersed-phase amalgam used in that study for our own investigation, and we found it to perform well, seeing no material-related problems during the five years.

Our results are likely to apply to strong amalgam adhesives, but may not apply to weaker ones. It is likely that they apply to other amalgams, as a previous study showed that choice of adhesive is more important than choice of amalgam to obtain high bond strengths.⁵

	CONCLUSIONS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Bonded and nonbonded amalgam restorations yielded similar results in conventional preparations after five years of clinical service when evaluated for anatomical form, marginal adaptation, sensitivity, surface quality and secondary caries. Bonded amalgam restorations were clinically successful in a limited number of nonretentive preparations over a five-year period. When placing bonded amalgam restorations, dentists can expect results equal to those of nonbonded restorations in conventional preparations, and they may be able to reduce or eliminate the cutting of healthy tooth structure to provide retention and resistance form for amalgam restorations.

View larger version (92K): [in this window] [in a new window]   Dr. Mach maintains a private practice in general dentistry, Prague, Czech Republic.

View larger version (93K): [in this window] [in a new window]   Dr. Regent maintains a private practice in general dentistry, Prague, Czech Republic.

View larger version (84K): [in this window] [in a new window]   Dr. Staninec is a clinical professor, University of California San Francisco School of Dentistry, 707 Parnassus Ave., San Francisco, Calif. 94143-0758, e-mail "stanin{at}itsa.ucsf.edu". Address reprint requests to Dr. Staninec.

View larger version (97K): [in this window] [in a new window]   Dr. Setcos is a senior lecturer, Restorative Dentistry, Dental School, University of Manchester, Manchester, England.