Clinical evaluation of non-carious cervical lesion restorations using a HEMA-free adhesive: three-year results


Clinical Associate Professor Michael Burrow
Faculty of Dentistry
The University of Hong Kong
Prince Philip Dental Hospital
34 Hospital Road
Sai Ying Pun
Hong Kong SAR


Background:  The use of all-in-one resin-based adhesives in clinical practice has continued to increase. The aim of this study was to evaluate retention and marginal staining of a HEMA-free all-in-one adhesive Go! (SDI, Australia) and Ice resin composite in non-carious cervical lesions (NCCLs).

Methods:  Forty-one restorations were placed in 13 patients (age range 44–72 years). Human Ethics Committee approval from the University of Melbourne and Dental Health Services Victoria was obtained. Restorations were bonded and placed according to the manufacturer’s instructions. Enamel etching was performed on NCCL margins after cleaning with pumice and water. Patients were recalled at six months and one, two and three years. Restorations were evaluated for retention and marginal staining, and photographic records of restorations were obtained.

Results:  At three years, six patients were available for recall with 23 restoration sites reviewed. At the six-month recall, three restorations had been lost, with a further two lost at two years. At three years, an overall cumulative retention rate of 85% was calculated using survival analysis. Fifteen of the 23 restorations showed slight marginal staining.

Conclusions:  At three years, the overall retention rate of 85% indicates a satisfactory result for this new adhesive. Marginal staining was regarded as minimal.

Abbreviations and acronyms:

hydroxyethyl methacrylate


non-carious cervical lesion


Resin-based adhesive use has now become a routine procedure for the placement of tooth-coloured restorations in all locations of the oral cavity. However, there still remains some scepticism in the profession for resin-based restorations to show long-term durability. For any new material or clinical technique to achieve broad acceptance from the profession there has to be a large amount of laboratory and clinical data generated to assess material reliability. This is especially so from the aspect of durable adhesion of resin-based adhesives. Even now, there remains a certain degree of reservation regarding clinical durability,1 possibly because of the rapid introduction of new materials which manufacturers claim have stronger adhesion and are simpler to use. However, this has not always been the case.2 Manufacturers have also attempted to develop simpler systems as a means to theoretically reduce some of the technique sensitivity associated with resin-based adhesives, but this too has not necessarily produced a marked improvement in long-term clinical success.3,4

One of the major changes in resin-based adhesion over the last 15 years has been the move away from using the more aggressive etching of enamel and dentine with phosphoric acid to the incorporation of acidic resin-based etching-priming monomers that usually have a pH of about 2, which is considerably higher than that of phosphoric acid at pH 0.8.5 The reported advantages of these less aggressive self-etching systems is the elimination of the washing and drying step that seems to be one of the major technique sensitive steps for the ‘etch-and-rinse’ systems. In addition, self-etching systems do not completely remove the hydroxyapatite crystals in the dentine, and this has also been reported to enhance bonding by creating a micromechanical ‘lock’, not only to the collagen fibre network but also around hydroxyapatite crystals.6 Recent work has shown that some of the monomers in self-etch systems can also form a salt with the calcium phosphate of a tooth.7 These changes in bonding systems seemed to have reduced technique sensitivity and improve bonding reliability, even for a practitioner with limited clinical experience.8

Accompanying the increased use of resin-based materials has been the concern of sensitization from one of the common monomers used in most resin-based adhesives, namely 2-hydroxyethyl methacrylate (HEMA).9,10 A report claimed dental staff have shown a slight increase in skin reactions to HEMA and there is concern that sensitization could increase with increased use in patients.11 This has led to a small number of manufacturers removing HEMA from their products in order to eliminate a potential problem. However, since HEMA is a hydrophilic monomer and important for bonding to the moist tooth surface, the process of bonding could be more difficult to achieve. This could also affect the durability of the bond. Laboratory data, however, indicate these newer systems, which only use one step for bonding, can achieve bond to strengths to enamel and dentine little different from, although often lower, than other adhesives currently available.2,5,11

It is known that the sclerotic dentine of non-carious cervical lesions (NCCLs) can be more difficult to bond to than the coronal dentine of teeth, hence such a substrate provides a good surface to test the adhesive qualities of an adhesive.12 The aim of the current study was to clinically evaluate one of these new HEMA-free systems (Go!, Southern Dental Industries, Bayswater, Victoria, Australia) when used to restore NCCL.

Materials and Methods

Forty-one restorations were placed in NCCLs in 13 patients (age range 44–72 years). Patients who presented for screening and treatment at the Royal Dental Hospital of Melbourne and exhibited the presence of one or more NCCLs were invited to participate in the current study. Voluntary participation and informed written consent from all subjects was obtained prior to entry into the study. Approval for the clinical evaluation was obtained from the Human Ethics Committee of the University of Melbourne as well as Dental Health Services Victoria. Subjects were excluded from the trial if they had chronic gingivitis, uncontrolled adult periodontitis, high caries risk and untreated lesions, poor oral hygiene or unable to attend recall visits. Teeth with NCCLs were selected for restoration, and varying numbers of lesions were restored per patient (Table 1). The size of restored lesions varied from shallow lesions (less than 1 mm deep, with an occluso-gingival height up to approximately 2 mm), occasionally exhibiting sensitivity to cold, to larger lesions approximately 5 mm occluso-gingivally in height and approximately 2 mm deep.

Table 1.   Distribution of restorations placed
Upper anteriors15
Upper premolars12
Upper molars1
Lower anteriors3
Lower premolars9
Lower molars1

Teeth were restored by one operator using the HEMA-free all-in-one adhesive Go! (SDI) and Ice hybrid resin composite (SDI). Restorations were placed using the following procedure: teeth were isolated using cotton rolls and high velocity evacuation, and then the dentine and enamel of the selected teeth were cleaned with a slurry of pumice and water on a slowly rotating rubber cup in a slow-speed handpiece, washed and dried, but not desiccated. The uncut enamel surface was etched with 37% phosphoric acid (SDI Etch, SDI, Australia) for 20 seconds then washed and dried but ensuring the tooth was not desiccated. A drop of Go! all-in-one adhesive was dispensed and applied to the tooth surface for 20 seconds, blown off with a strong blast of air for five seconds, and light cured using an LED light (Mini LED, Acteon, Mount Laural, NJ, USA) with an output of 1000 mW/cm2 for 10 seconds. Ice resin composite was applied in one increment for the smaller lesions and cured for at least 40 seconds. For larger lesions, the resin composite was placed incrementally, each increment being cured for 40 seconds.

All restorations were contoured with fine composite finishing diamonds in an intermediate-speed handpiece under water spray and finished with KerrHawe Optidisc discs (KerrHawe, Bioggio, Switzerland). Teeth selected for restoration depended on the location of lesions per patient.

Subjects were recalled at six months and one, two and three years after placement. The restorations were checked for presence or absence and for marginal staining. Photographs at 1:1 magnification were taken of the cavities prior to restoration, immediately after, then at six months and one, two and three years. The photographs were also checked for restoration presence or absence, and colour match of the restoration with the surrounding tooth structure. Marginal discolouration was assessed from photographs by comparing test restorations against a standard set of photographs on a nine-point scale whereby 0 represented no staining and 9 represented severe staining. Restoration survival was determined using a survival analysis equation.


From the initial 13 patients recruited to enter the trial, by the three-year recall period only six were available. This allowed the recall of 23 restoration sites for evaluation from the original 41 restorations inserted at baseline. At two years, seven patients attended recall, with 25 restoration sites present for evaluation. At one year, 33 restoration sites (11 patients) were examined which was the same number of patients and sites evaluated at the six-month recall. With respect to loss of restorations, two periods of loss were observed over the three years of the evaluation. One period was at the six-month recall where 33 restoration sites (11 patients) could be evaluated. At this recall it was observed that three restorations had been lost in two patients (92% cumulative retention rate). The second observation of restoration loss was at the two-year recall where a further two restorations had been lost of the 25 restoration sites (seven patients) that were available for evaluation (85% cumulative retention). No further loss of restorations was noted for those patients who attended the three-year recall appointment. At three years, survival analysis calculation for the cumulative retention rate provided an outcome of 85%.

With respect to marginal staining, of the 23 restorations observed at recall, it was noted that 15 of the restorations exhibited slight marginal staining (Fig 1). The degree of staining was at a level that the clinician would note its presence but would be of little or no concern to the patient. Two restorations showed a slight increase of the staining between the two- and three-year recall periods, but again the degree of staining was classified as minimal (Fig 2). No restoration available for evaluation exhibited severe staining that would be obvious to the patient to the extent that a request for replacement is likely to be made.

Figure 1.

 Restoration of teeth 12 and 11 with Go! and Ice resin composite. (a) baseline, immediately after restoration, (b) 6 months, (c) 2 years and (d) 3 years after placement. No marginal staining is observed on any of these restorations.

Figure 2.

 (a) Baseline, (b) 1 year, (c) 3 years – marginal staining can be noticed on the margins of the restorations. Much of this staining could normally be removed at recall visits by judicious polishing of the margins.


This clinical evaluation is one of the few three-year trials of the recently introduced non-HEMA containing resin-based adhesives. The other trials published on such materials are for another product (G-Bond; GC Corporation, Tokyo, Japan),13–15 so this is the first trial evaluating Go! adhesive. The results of this evaluation show satisfactory restoration retention. However, as with many clinical evaluations, the outcomes must be viewed with some degree of caution due to the loss of patients during the evaluation period. The data for the patients not seen cannot be predicted; hence this is one reason why a survival analysis method is used for determining restoration survival, which allows for those restorations that could not be evaluated.

The distribution of restorations in teeth for this trial is perhaps a little different from the typical location of NCCL. It is more common to see NCCL in premolar teeth, but in the current study most restorations were placed in anterior teeth and the upper arch. Whether this would affect outcomes is not known. It reflects the distribution of lesions in this group of patients for this evaluation. All teeth restored in this study were in occlusal contact.

In comparison with the other non-HEMA containing material, the results of the current trial are not as successful for a number of reasons when compared with several studies,15,16 but little or no different from another study evaluating G-Bond.14 The first reason for a difference may be related to the monomer used in each of the materials. The other adhesive has used the well-known and long researched monomer, 4-MET, which has been shown to form a salt when it interacts with pure hydroxapatite.7 The monomer for Go! is new, and there is no evidence indicating whether the same chemical interaction occurs with hydroxyapatite. It would seem that the chemical interaction may be important for establishing durable adhesion to the dentine of NCCL when these simplified resin-based adhesives are used.

The early failures may have been caused by another factor. The three early failures were of restorations that had been recorded as being difficult to place due to the first increment of composite not adhering to the bonded dentine surface before curing. That is, the first increment did not bond to the air-inhibited layer of the resin adhesive and tended to slip off and stick to the instrument, even though a plastic-tipped instrument was used. When this occurred, the initial increment of composite was discarded and a new increment placed in order to achieve the initial bond to the tooth surface. It is possible that with such a stiff material as Ice resin composite, removal of the air-inhibited layer of the bond occurred prior to curing which may have then compromised the adhesion of the new increment of filling material. To date this possible phenomenon has been poorly investigated and could be an important factor ensuring good adhesion to the bonding resin layer, especially for the stiff, so-called ‘packable’ materials. These materials have a lower volume of matrix resin than ‘normal’ composite materials and thus wetting of the bonded surface could be compromised when the air-inhibited layer is lost. This problem could be solved by re-application of the adhesive, light polymerization then placement of the filling material. However, care must be taken to ensure the adhesive layer does not become too thick as this may also compromise long-term adhesion. It seems that apart from this issue, restorations that survive the first 24 months in the oral cavity seem to show stable adhesion thereafter.

In the current study, the uncut enamel of the NCCL was etched as recommended by the manufacturer. Whether this contributed to the retention of the restorations is debatable. A recent clinical study investigating the influence of acid etching on restoration survival over five years concluded that the effect was at best minimal when a two-step self-etching priming adhesive was used.17 The experience of the author comparing results of two clinical evaluations of G-Bond, one with and the other without enamel etching, found no difference in retention over three years (unpublished data). Hence, enamel etching is possibly not so important to the adhesion where the greatest surface area of bonding is not enamel, e.g. the dentine surface of NCCL.

Marginal staining in the current evaluation was minimal, although noted on a larger number of restorations than perhaps in other trials using similar types of adhesive systems. This may possibly be due to the different resin composites employed and the ability to finish and polish the composite. It was noted that the hybrid composite in the current study did not polish to as high a lustre compared with other composites used. Therefore, a slightly rougher surface may increase the possibility of staining. However, in the current study, the degree of staining was minimal and could easily be removed at recall visits by further polishing restoration margins. Additionally, the marginal staining may also have been influenced by the patients selected for this trial. It seems that marginal staining is also a result of the patient’s oral environment such as bacterial species, diet, oral hygiene habits and ability to keep their teeth clean.

In conclusion, the cumulative retention rate over three years when Go! adhesive is used to restore NCCL was satisfactory. It would be interesting to know if the use of a less stiff, less viscous resin composite would have influenced the results.


The project was sponsored by Southern Dental Industries Ltd, Australia. The author would like to thank Dental Health Services Victoria and the Royal Dental Hospital of Melbourne for use of its clinical facilities.