- Email: [email protected]
Dentin smear layer: an asset or a liability for bonding?
Dentin smear layer: an asset or a liability for bonding? G.J. White D.R. Beech M.J. Tyas* Australian Dental Standards Laboratory 240 Langridge Street Abbotsford, Victoria 3067
Australia Received January 6, 1989 Accepted May 6, 1989 *Corresponding author Dent Mater 5:379-383, November, 1989
Abstract-Despite concerns and claims that the smear layer on dentin is undesirable for bonding, supportive evidence is lacking. The clinical efficacy of various agents for smear layer removal and the effect of smear layer removal on the bond strengths of a glass-ionomer cement and three representative dentin bonding agents were examined. For all but one dentin bonding agent (Gluma), a 15-second treatment with 17% EDTA caused a reduction in bond strength. For
Gluma, no significant bond was obtained without EDTA treatment. While Gluma probably bonds via dentinal collagen, the other materials interact primarily with
dentinal calcium. Removal of the smear layer for adhesives reliant on the presence of calcium is therefore undesirable. The clinical effects of some agents proposed for smear layer removal were examined by SEM of replicas.
mear layer is a term used to describe the debris which is generated by the cutting or abrading of tooth substance, especially dentin, and which covers the tooth surface. Eick et al. (1970) found particles r a n g i n g from 0.5 ~m to 15 ~m embedded in an organic layer 0.5 ~m deep, whereas Br~innstr6m (1982) recorded the depth of the layer as varying from 1-5 ~m. The morphology, thickness, and composition of the smear layer vary with the method used for cutting the surface, with coarse diamond abrasives used dry producing the t h i c k e s t deposits (Gwinnett, 1984). In recent years, the smear layer has c o m e u n d e r extensive scrutiny, with attitudes toward its existence ranging from being "a liability in its ability to harbour bacteria beneath restorations" (Gwinnett, 1984) to being a "beneficial, iatrogenically produced cavity liner which reduces dentin permeability far more effectively than cavity varnishes" (Pashley, 1984). Its presence has also been considered to be a major hindrance to the bonding of restorative materials, particularly to dentin (Gwinnett, 1984). Agents for smear layer removal in bonding procedures range from acids and chelating agents which remove plugs in the ends of tubules, to those intended to remove the smear layer, leaving the t u b u l e s occluded (Br~innstrSm, 1982). However, some proprietary dentin bonding agents do not require smear layer removal, and in some cases its removal has an adverse effect on bond strength (Solomon and Beech, 1985). The aims of this study were to ascertain (i) whether the removal of the smear layer from dentin significantly improved the bond strength of a glass-ionomer cement and resin composite with dentin primers, and (ii) the ability of a number of clinical treatments to produce smear-layerfree surfaces.
S
MATERIALS AND METHODS
The crowns of extracted teeth which had been stored in water for approximately six months were embedded in the ends of cylinders of dental stone. Flat, lingual or buccal noncarious surfaces of dentin were exposed by wet-grinding, and were finished with 600-grade SiC paper. The dentin was maintained wet at all times. We defined the dentin area to be used for testing by covering the surface with a strip of PVC tape in which a hole of known diameter had been punched. The materials ar~ listed in Table 1, together with the dentin pretreatments. The mineralizing solution used was that described by Levine et al. (1977), and consisted of equal volumes or solution A (0.5 tool/ L potassium dihydrogen phosphate saturated with calcium hydrogen phosphate dihydrate at pH 4.5) and solution B (0.5 mol/L disodium hydrogen phosphate with 2000 ppm Fat pH 9), mixed for 30 sec. After pre-treatment, the dentin area was dried for 10 sec with clean air, and glass-ionomer cement (P:L 1.8:1) or a dentin bonding agent, followed by resin composite, was applied, followed by a 3.2-mm-diameter stainless steel orthodontic button (Dentaurum, Germany) for glass ionomer or a 3.5-ram-diameter meshbacked orthodontic bracket (A.J. Wilcock, Melbourne, Australia) for composite. In the case of Gluma, the enamel bond supplied with Gluma was applied in accordance with the manufacturer's instructions, immediately before composite placement. The orthodontic buttons and brackets had small flexible wire loops for attachment in the bond strength testing rig and were pressed firmly to fit exactly into the appropriate hole in the tape. The specimens were left on the bench to set for about 10 rain. The glass-ionomer cement specimens
Dental Materials/Nove~nber 1989
379
TABLE 1 MATERIALS USED AND DENTIN PRE-TREATMENTS FOR BOND STRENGTH STUDIES
Brand Name/Manufacturer Fuji IIF (G-C, Tokyo, Japan)
Material Glass-ionomer Cement
Dentin Pre-treatment 1. Distilled water, 60 sec 2. 25% polyacrylic acid, 30 sec; distilled water rinse, 60 sec 3. 25% tannic acid, 30 sec; distilled water rinse, 60 sec
Silar (3M, St. Paul, MN, USA) Gluma (Bayer, Leverkusen, 1. Distilled water, 60 sec West Germany) 2. 17% EDTA, 15 sec; distilled water rinse, 10 sec Two-component Phosphorylated Scotchbond LC (3M, St. Paul, 1. Distilled water, 60 sec Methacrylate, Light-cured MN, USA) 2. 17% EDTA, 15 sec; distilled water rinse, 10 sec 3. Remineralizing solution, 3 days Single-component Phosphory- Prisma Universal Bond (Caulk- 1. Distilled water, 60 sec lated Methacrylate Dentsply, Milford, DE, USA) 2. 17% EDTA, 15 sec; distilled water rinse, 10 sec Composite Restorative Material Aldehyde/hydroxyethylmethacrylate
TABLE 2 MEAN BOND STRENGTH, MPa (__ SD) Material
Fuji IIF
Silar + Gluma Silar + Scotchbond L/C Silar + Prisma Universal Bond
Pre-treatment Distilled water EDTA PAA TA Distilled water EDTA Distilled water EDTA Remineralizing solution Distilled water EDTA
were placed in a closed container over water and the composite specimens immersed in water. (The glass-ionomer cement specimens were not immersed in water because the cemerit is vulnerable to damage by water for a period after setting, unless covered by varnish. Storage in 98% R.H. obviated this potential experimental variable and the need to varnish specimens.) After 24 hours' storage at 37°C, bond strengths were determined by attachment of a chain to the wire of the button/bracket from the tensile load cell of a universal mechanical testing machine (Shimadzu, Japan) and a tensile force applied until fract u r e using 1 mm/min cross-head speed. This bond strength method has been d e s c r i b e d p r e v i o u s l y (Beech, 1972; Beech et al., 1985). At least 10 specimens were tested
No. of Specimens 12 12 12 12 17 17 15 15 14 18 18
Mean Bond Strength 4.5 (0.9) 3.1 (1.5) 4.2 (1.3) 4.0 (1.4) 0.3 (0.5) 4.5 (3.6) 5.7 (2.0) 3.7 (3.5) 5.7 (1.5) 2.9 (2.8) 1.1 (1.3)
from each group, and the results were statistically evaluated by one-way analysis of variance. S c a n n i n g Electron Microscopy ( S E M ) - S c a n n i n g electron micrographs were taken of dentin surfaces which had been prepared by being ground on 600-grade SiC paper and subjected to the pre-treatm e n t regimen of either distilled water, EDTA, polyacrylic acid, or tannic acid. The specimens were dried for 24 hr, sputter-coated with gold, and examined by SEM. Patients with buccal cervical abrasion cavities with exposed dentin were selected and impressions taken of the abraded area by use of addition-cured silicone (Reprosil; CaulkDentsply, Milford, DE, USA). Three impressions were taken: (1) after the lesion was sprayed with air and water, then dried;
380 WHITE et al./SMEAR LAYER AND BONDING
(2) immediately after removal of the first impression, before saliva contamination could occur; and (3) after the application of either Fuji II liquid or EDTA for 15 sec, or 5 sec pumice/water slurry on a rubber cup, followed by being washed and dried. Models from these impressions were cast in epoxy resin (Araldite, CIBA Geigy), and sputter-coated with gold prior to SEM examination. RESULTS
Bond S t r e n g t h - The results of the bond strength evaluation for the materials with the various dentin pretreatments are shown in Table 2. For glass-ionomer cement, the fracture appeared to be partly cohesive and partly adhesive, with fragments of cement remaining on the dentin surface. There was no significant difference b e t w e e n any of the glassionomer values. The most marked change in bond strength was the reduction following the use of EDTA. Silar in conjunction with Gluma had poor bond strength when EDTA was not used; many of the test specimens failed in the handling procedures prior to being tested. When EDTA was used, a significant increase in bond strength was found (P < 0.05). Silar with Prisma Universal Bond showed a significant decrease in bond strength when the smear layer was removed (P < 0.05). Scotchbond with Silar showed a reduction in bond strength when the smear layer was removed, but the difference was not significant because of the high experimental scatter. In the majority of specimens, fracture occurred at the interface between composite and dentin. SEM Examination-Examination by SEM of specimens after bond strength testing of Silar/Gluma with EDTA pre-treatment revealed small tags of resin composite material on the surface and within some of the dentinal tubules (Fig. 1). This was not widespread over the test surface. Dentin examination after specimens were polished in vitro with 600grade SiC paper exhibited the presence of a smear layer (Fig. 2). No underlying morphology could be discerned.
Smear layer was readily removed from the laboratory specimens with either EDTA, polyacrylic acid (PAA), or tannic acid (TA). However, TA did not remove the smear layer as thoroughly as did EDTA or PAA, and t h e r e was evidence of some smear layer both on the dentin surface and within the dentinal tubules following the use of TA (Fig. 3). Models of the initial clinical impressions frequently had the appearance of a surface upon which a layer had deposited by evaporation of a solution, e.g., saliva. This layer was altered in appearance on the replicas from the second impressions. In some isolated instances, it was possible for some underlying morphology to be discerned through the covering layer after the second impression, but this was not a consistent finding. Fuji II liquid and EDTA were effective in clinical cleansing of the tooth surface to the extent of exposing dentinal tubules in the abrasion defects. Several specimens showed the tubules greatly elongated, since they had been exposed at an angle rather than at right angles to the long axis of the tubule. Another specimen showed the tubules to be less numerous p e r unit area (Fig. 4). The pumice and water slurry applied with a rubber cup altered the appearance of the surface but not to the extent of revealing the tubular morphology.
ble 2 for Fuji IIF show that no change occurs by removal of the smear layer, unlike the use of EDTA, which did show a change, albeit not statistically significant, as a consequence of removal of calcium from the dentin surface. Exposure of the dentin surface by effective decalcifying agents (such as 40% solutions of EDTA, citric, lactic, or orthophosphoric acids) causes rapid reduction and eventually complete elimination of bonding
with polyacrylic-acid-based cements (Beech, 1972). The adverse effects of powerful demineralizing agents on bond strength are well-documented (Hood et al., 1981; 0ilo, 1978; Hinoura et at., 1986). However, there is conflict in the literature, based upon laboratory measurements, about the effect of smear layer removal (Powis et al., 1982; Beech et al., 1985; Shalabi et al., 1981). Smear-layer removal p r o c e d u r e s have been
Fig. 1. Tags of composite following tensile testing (Silar/Gluma with EDTA pre-treatment). Bar = 50 ~.m.
DISCUSSION Glass-ionomer C e m e n t s - The mechanism of adhesion of glass-ionomer cement is by ionic attraction between carboxylate ions in the cement and calcium ions in the tooth surface (Beech, 1972), although bonding to collagen has also been s u g g e s t e d (McLean and Wilson, 1977). The adhesive bond may be an intermediate layer formed between the bulk of the cement and the bulk of the tooth substance. Mount (1984) claimed that such a physico-chemical union is dependent upon total cleanliness of the dentin. The application of either PAA or TA provided a substantially clean dentinal surface (Figs. 1,3), and, with such "mild" cleansers, significant demineralization of the surface would not be expected. The results in Ta-
Fig. 2. Smear layer after preparation of dentin with 600-grade SiC paper. Bar --- 50 p.m.
Dextral Materio/s/Nol~e~tber 19~:9 381
Fig. 3. Smear layer in Fig. 2 removed by application of 25% TA for 20 sec. Bar = 50 r~m.
Fig. 4, Cervical abrasion lesion following Fuji II liquid treatment. Bar = 50 i~m.
advocated (Mount, 1984), but there is no evidence that improved clinical retention results. Indeed, retention rates of 97-100% after one year in non-undercut Class V abrasion cavities are easily achievable when only pumice and water prophylaxis are used (Tyas and Beech, 1985; Tyas et al., 1986). Such retention rates would not be improved by chemical removal of the smear layer.
It is likely that the chemical activity of the free polycarboxylic acid liquid in glass-ionomer cement at the time of placement dissolves, penetrates, and incorporates the smear layer, enabling adhesion to the body of the dentin to occur. The considerable variation in nature and access to the smear layer on dentin clinically makes effective removal of the smear layer difficult, and has the po-
382 W H I T E et al./SMEAR L A Y E R A N D BONDING
tential to reduce bonding if decalcification occurs. G l u m a - I n contrast to glass-ionomer cement, Gluma has been postulated to bond to dentin by reaction of the glutaraldehyde in Gluma with basic groups in dentinal collagen such as NH2 and OH (Eliades et al., 1985; Asmussen, 1985). Gluma is unique in being specific to collagen, and therefore a demineralized dentinal surface should be perfectly satisfactory for bonding, provided it is not seriously mechanically weakened. The combination of glutaraldehyde and 2-hydroxyethylmethacrylate (HEMA) in Gluma may also facilitate bonding by dissolution in the dentin surface. From the results in Table 2, those of Munksgaard and Asmussen (1985), and the clinical data of Tyas (1989), it is clear that the smear layer acts as a barrier to bonding of Gluma and must be removed. Gluma is the only material in this study which is not acidic, favors bonding through collagen, and benefits from smear-layer removal. Scotchbond L C and Pvisma Universal B o n d - I n common with glassionomer cement, Scotchbond LC and Prisma Universal Bond show a preference for a surface which has not been subject to s m e a r - l a y e r removal/decalcification (Table 2). This is consistent with some previous reports on Scotchbond (Solomon and Beech, 1985; Hill et al., 1983; Hinoura et al., 1986), although others are less clear-cut in their findings (Retief et al., 1986; Odin and 0ilo, 1986). Scotchbond LC and Prisma Universal Bond are acidic materials with anionic groups designed to interact with calcium ions. Agents which are capable of dissolving mineral in the smear layer will inevitably demineralize the u n d e r l y i n g dentin surface to some extent. The manufacturers of these two dentin bonding agents recognize this by not advocating chemical smear-layer removal. Presumably, like glass-ionomer cement, these acidic monomers can combine with or penetrate the smear layer. Although Scotchbond has a greater bond strength to enamel than to dentin (Solomon et al., 1985), increasing the mineral content of the dentin surface failed to increase the bond strength (Table 2). Dissolution in the dentinal collagen and the
strength/integrity of the dentin surface may be significant factors in optimization of bond strength. Scanning Electron M i c r o s c o p y Scanning electron micrographs revealed that only Fuji II liquid and EDTA will remove the smear layer from cervical abrasion lesions clinically. Duke et al. (1985) reported that in the laboratory most of the smear layer was removed when the surface was subjected to a pumice prophylaxis; however, the present study was unable to confirm that result. Hydrogen peroxide appeared to have no effect upon the smear layer surface, which agrees with the findings of Duke et al. (1985). Powis et al. (1982) also found that polyacrylic acid was effective in laboratory removal of the smear layer; however, in their study tannic acid was less effective than in the present study. Dentin Bonding and the Smear Layer - General ConsiderationsThe variation in the effect of smear layer removal upon the bonding of different adhesives to dentin suggests that routine smear-layer removal is contra-indicated, and should only be done when the need is established. Some dentin bonding agents for resin composites, and the polyacrylic-acid-based cements, have found commercial and clinical viability without the removal of the smear layer. Removal of the smear layer with agents capable of dissolving hydroxyapatite will inevitably reduce the proportion of this mineral in the dentin surface. The polyacrylic-acidbased cements and many dentin bonding agents favor a highly calcified surface. The decalcified surface after smear layer removal will be soft and proteinaceous. This ,surface is clearly preferred for Gluma, which is somewhat exceptional in bonding via dentin collagen. Removal of the plugs at the cut ends of dentin tubules as a consequence of smear layer removal may allow access of bacteria and chemicals to the pulp. Exudation of proteinaceous fluid from the tubules may also interfere with bonding, since many dentin bonding agents are or-
ganic hydrophobic compounds which are unable to compete with dentinal fluid. Furthermore, there is no evidence that penetration of bonding agent into dentinal tubules is advantageous to bonding; it does not give retention of resins clinically and is not analogous to resin tags in enamel. In the less-than-ideal environment of a busy dental practice, uniform removal of the smear layer without unblocking the dentinal tubules will be difficult to achieve. Adhesive techniques which can cope with the smear layer may be more likely to meet with clinical success.
REFERENCES ASMUSSEN, E. (1985): Clinical Relevance of Physical, Chemical and Bonding Properties of Composite Resins, Oper Dent 10:61-73. BEECH, D.R. (1972): Improvement in the Adhesion of Polyacrylate Cements to Human Dentine, Br Dent J 135:442445. BEECH, D.R.; SOLOMON, A.; and BERNIER, R. (1985): Bond Strength of Polycarboxylic Acid Cements to Treated Dentin, Dent Mater 1:154-157. BRANNSTR~M, M. (1982): Dentin and Pulp in Restorative Dentistry. London: Wolfe Medical Publication Ltd. DUKE, E.S.; PHILLIPS, R.W.; and BLUMERSHINE, R. (1985): Effects of Various Agents in Cleaning Cut Dentine, J Oral Rehabil 12:295-302. LICK, J.D.; WILKO, R.A.; ANDERSON, C.H.; and SORENSON, A.E. (1970): Scanning Electron Microscopy of Cut Tooth Surfaces and Identification of Debris by Use of the Electron Microscope, J Dent Res 49:1358-1368. ELIADES, G.L.; CAPUTO, A.A.; and VOUGIOUKLAKIS, G.J. (1985): Corhposition, Wetting Properties and Bond Strength with Dentin of Six New Dentin Adhesives, Dent Mater 1:170-176. GWINNE'Fr, A.J. (1984): Smear Layer: Morphological Considerations, Oper Dent (Suppl) 3:3-12. HILL, G.L.; JENSEN, M.E.; and ZIDAN, O. (1983): Shear-bond Strengths of a New Dentinal Bonding Agent: Pretreatment Effects, J Dent Res 221: Abst. No. 469. HINOURA, K.; MOORE, B.K.; and PHILLIPS, R.W. (1986): Influence of Dentin Surface Treatments on the Bond Strengths of Dentin-lining Ce-
merits, Oper Dent 11:147-154. HOOD, J.A.A.; CHILDS, W.A.; and EVA_~S, D.F. (1981): Bond Strength of Glass Ionomer and Polycarboxylate Cements to Dentine, N Z Dent J 77:141-144. LEVINE, R.S.; BEECH, D.R.; and GARTON, B. (1977): Improving the Bond Strength of Polyacrylate Cements to Dentine, Br Dent J 143:275277. McLEAN, J.W. and WILSON,A.D. (1977): The Clinical Development of the Glassionomer Cement. III. The Erosion Lesion, Aust Dent J 22:190-195. MOUNT, G.J. (1984): Glass Ionomer Cements: Clinical Considerations. In: Clinical Dentistry, James W. Clark, Ed., Philadelphia; Harper and Row. MUNKSGAARD, E.C. and ASMUSSEN, E. (1985): Dentin-Polymer Bond Mediated by Glutaraldehyde/HEMA, Scand J Dent Res 93:463-466. ODI~N, A. and 0ILO, G. (1986): Tensile Bond Strength of Dentin Adhesives, Dent Mater 2:207-213. 0ILO, G. (1978): Adhesive Bonding of Dental Luting Cements: Influence of Surface Treatment, Acta Odontol Scand 36:263-270. PASHLEY, D.H. (1984): Smear Layer: Physiological Coasiderations, Oper Dent (Suppl) 3:13-29. Powls, D.R.; FOLLER~tS, T.; MERSON, S.A.; and WILSON, A.D. (1982): Improved Adhesion of a Glass-ionomer Cement to Dentin and Enamel, J Dent Res 61:1416-1422. RETIEF, D.H.; GROSS, J.D.; BRADLEY, E.L.; and DENYS, F.R. (1986): Tensile Bond Strengths of Dentin Bonding Agents to Dentin, Dent Mater 2:7277. SHALABI, H.S.; ASMUSSEN, E.; and JORCENSEN, K.D. (1981): Increased Bonding of a Glass Ionomer Cement to Dentin by Means of FeC13, Scand J Dent Res 89:348-353. SOLOMON, A. and BEECH, D.R. (1985): Bonding of Composites to Dentin using Primers, Dent Mater 1:79-82. TYAS, M.J. (1989): Clinical Evaluation of Three Dentine Bonding Agents, Aust Dent J (in press). TYAS, M.J. and BEECH, D.R. (1985): Clinical Performance of Three Restorative Materials for Non-Undercut Cervical Abrasion Lesions, Aust Dent J 30:260-264.
TYAS, M.J.; BURNS, G.A.; BYRNE, P.F.; CUNNINGHAM,P.J.; DOBSON, B.C.; and WIDDOP, F.T. (1986): Clinical Evaluation of Scotchbond: One Year Results, Aust Dent J 31:159-164.
Dental Materials~November 1989 383
Australia Received January 6, 1989 Accepted May 6, 1989 *Corresponding author Dent Mater 5:379-383, November, 1989
Abstract-Despite concerns and claims that the smear layer on dentin is undesirable for bonding, supportive evidence is lacking. The clinical efficacy of various agents for smear layer removal and the effect of smear layer removal on the bond strengths of a glass-ionomer cement and three representative dentin bonding agents were examined. For all but one dentin bonding agent (Gluma), a 15-second treatment with 17% EDTA caused a reduction in bond strength. For
Gluma, no significant bond was obtained without EDTA treatment. While Gluma probably bonds via dentinal collagen, the other materials interact primarily with
dentinal calcium. Removal of the smear layer for adhesives reliant on the presence of calcium is therefore undesirable. The clinical effects of some agents proposed for smear layer removal were examined by SEM of replicas.
mear layer is a term used to describe the debris which is generated by the cutting or abrading of tooth substance, especially dentin, and which covers the tooth surface. Eick et al. (1970) found particles r a n g i n g from 0.5 ~m to 15 ~m embedded in an organic layer 0.5 ~m deep, whereas Br~innstr6m (1982) recorded the depth of the layer as varying from 1-5 ~m. The morphology, thickness, and composition of the smear layer vary with the method used for cutting the surface, with coarse diamond abrasives used dry producing the t h i c k e s t deposits (Gwinnett, 1984). In recent years, the smear layer has c o m e u n d e r extensive scrutiny, with attitudes toward its existence ranging from being "a liability in its ability to harbour bacteria beneath restorations" (Gwinnett, 1984) to being a "beneficial, iatrogenically produced cavity liner which reduces dentin permeability far more effectively than cavity varnishes" (Pashley, 1984). Its presence has also been considered to be a major hindrance to the bonding of restorative materials, particularly to dentin (Gwinnett, 1984). Agents for smear layer removal in bonding procedures range from acids and chelating agents which remove plugs in the ends of tubules, to those intended to remove the smear layer, leaving the t u b u l e s occluded (Br~innstrSm, 1982). However, some proprietary dentin bonding agents do not require smear layer removal, and in some cases its removal has an adverse effect on bond strength (Solomon and Beech, 1985). The aims of this study were to ascertain (i) whether the removal of the smear layer from dentin significantly improved the bond strength of a glass-ionomer cement and resin composite with dentin primers, and (ii) the ability of a number of clinical treatments to produce smear-layerfree surfaces.
S
MATERIALS AND METHODS
The crowns of extracted teeth which had been stored in water for approximately six months were embedded in the ends of cylinders of dental stone. Flat, lingual or buccal noncarious surfaces of dentin were exposed by wet-grinding, and were finished with 600-grade SiC paper. The dentin was maintained wet at all times. We defined the dentin area to be used for testing by covering the surface with a strip of PVC tape in which a hole of known diameter had been punched. The materials ar~ listed in Table 1, together with the dentin pretreatments. The mineralizing solution used was that described by Levine et al. (1977), and consisted of equal volumes or solution A (0.5 tool/ L potassium dihydrogen phosphate saturated with calcium hydrogen phosphate dihydrate at pH 4.5) and solution B (0.5 mol/L disodium hydrogen phosphate with 2000 ppm Fat pH 9), mixed for 30 sec. After pre-treatment, the dentin area was dried for 10 sec with clean air, and glass-ionomer cement (P:L 1.8:1) or a dentin bonding agent, followed by resin composite, was applied, followed by a 3.2-mm-diameter stainless steel orthodontic button (Dentaurum, Germany) for glass ionomer or a 3.5-ram-diameter meshbacked orthodontic bracket (A.J. Wilcock, Melbourne, Australia) for composite. In the case of Gluma, the enamel bond supplied with Gluma was applied in accordance with the manufacturer's instructions, immediately before composite placement. The orthodontic buttons and brackets had small flexible wire loops for attachment in the bond strength testing rig and were pressed firmly to fit exactly into the appropriate hole in the tape. The specimens were left on the bench to set for about 10 rain. The glass-ionomer cement specimens
Dental Materials/Nove~nber 1989
379
TABLE 1 MATERIALS USED AND DENTIN PRE-TREATMENTS FOR BOND STRENGTH STUDIES
Brand Name/Manufacturer Fuji IIF (G-C, Tokyo, Japan)
Material Glass-ionomer Cement
Dentin Pre-treatment 1. Distilled water, 60 sec 2. 25% polyacrylic acid, 30 sec; distilled water rinse, 60 sec 3. 25% tannic acid, 30 sec; distilled water rinse, 60 sec
Silar (3M, St. Paul, MN, USA) Gluma (Bayer, Leverkusen, 1. Distilled water, 60 sec West Germany) 2. 17% EDTA, 15 sec; distilled water rinse, 10 sec Two-component Phosphorylated Scotchbond LC (3M, St. Paul, 1. Distilled water, 60 sec Methacrylate, Light-cured MN, USA) 2. 17% EDTA, 15 sec; distilled water rinse, 10 sec 3. Remineralizing solution, 3 days Single-component Phosphory- Prisma Universal Bond (Caulk- 1. Distilled water, 60 sec lated Methacrylate Dentsply, Milford, DE, USA) 2. 17% EDTA, 15 sec; distilled water rinse, 10 sec Composite Restorative Material Aldehyde/hydroxyethylmethacrylate
TABLE 2 MEAN BOND STRENGTH, MPa (__ SD) Material
Fuji IIF
Silar + Gluma Silar + Scotchbond L/C Silar + Prisma Universal Bond
Pre-treatment Distilled water EDTA PAA TA Distilled water EDTA Distilled water EDTA Remineralizing solution Distilled water EDTA
were placed in a closed container over water and the composite specimens immersed in water. (The glass-ionomer cement specimens were not immersed in water because the cemerit is vulnerable to damage by water for a period after setting, unless covered by varnish. Storage in 98% R.H. obviated this potential experimental variable and the need to varnish specimens.) After 24 hours' storage at 37°C, bond strengths were determined by attachment of a chain to the wire of the button/bracket from the tensile load cell of a universal mechanical testing machine (Shimadzu, Japan) and a tensile force applied until fract u r e using 1 mm/min cross-head speed. This bond strength method has been d e s c r i b e d p r e v i o u s l y (Beech, 1972; Beech et al., 1985). At least 10 specimens were tested
No. of Specimens 12 12 12 12 17 17 15 15 14 18 18
Mean Bond Strength 4.5 (0.9) 3.1 (1.5) 4.2 (1.3) 4.0 (1.4) 0.3 (0.5) 4.5 (3.6) 5.7 (2.0) 3.7 (3.5) 5.7 (1.5) 2.9 (2.8) 1.1 (1.3)
from each group, and the results were statistically evaluated by one-way analysis of variance. S c a n n i n g Electron Microscopy ( S E M ) - S c a n n i n g electron micrographs were taken of dentin surfaces which had been prepared by being ground on 600-grade SiC paper and subjected to the pre-treatm e n t regimen of either distilled water, EDTA, polyacrylic acid, or tannic acid. The specimens were dried for 24 hr, sputter-coated with gold, and examined by SEM. Patients with buccal cervical abrasion cavities with exposed dentin were selected and impressions taken of the abraded area by use of addition-cured silicone (Reprosil; CaulkDentsply, Milford, DE, USA). Three impressions were taken: (1) after the lesion was sprayed with air and water, then dried;
380 WHITE et al./SMEAR LAYER AND BONDING
(2) immediately after removal of the first impression, before saliva contamination could occur; and (3) after the application of either Fuji II liquid or EDTA for 15 sec, or 5 sec pumice/water slurry on a rubber cup, followed by being washed and dried. Models from these impressions were cast in epoxy resin (Araldite, CIBA Geigy), and sputter-coated with gold prior to SEM examination. RESULTS
Bond S t r e n g t h - The results of the bond strength evaluation for the materials with the various dentin pretreatments are shown in Table 2. For glass-ionomer cement, the fracture appeared to be partly cohesive and partly adhesive, with fragments of cement remaining on the dentin surface. There was no significant difference b e t w e e n any of the glassionomer values. The most marked change in bond strength was the reduction following the use of EDTA. Silar in conjunction with Gluma had poor bond strength when EDTA was not used; many of the test specimens failed in the handling procedures prior to being tested. When EDTA was used, a significant increase in bond strength was found (P < 0.05). Silar with Prisma Universal Bond showed a significant decrease in bond strength when the smear layer was removed (P < 0.05). Scotchbond with Silar showed a reduction in bond strength when the smear layer was removed, but the difference was not significant because of the high experimental scatter. In the majority of specimens, fracture occurred at the interface between composite and dentin. SEM Examination-Examination by SEM of specimens after bond strength testing of Silar/Gluma with EDTA pre-treatment revealed small tags of resin composite material on the surface and within some of the dentinal tubules (Fig. 1). This was not widespread over the test surface. Dentin examination after specimens were polished in vitro with 600grade SiC paper exhibited the presence of a smear layer (Fig. 2). No underlying morphology could be discerned.
Smear layer was readily removed from the laboratory specimens with either EDTA, polyacrylic acid (PAA), or tannic acid (TA). However, TA did not remove the smear layer as thoroughly as did EDTA or PAA, and t h e r e was evidence of some smear layer both on the dentin surface and within the dentinal tubules following the use of TA (Fig. 3). Models of the initial clinical impressions frequently had the appearance of a surface upon which a layer had deposited by evaporation of a solution, e.g., saliva. This layer was altered in appearance on the replicas from the second impressions. In some isolated instances, it was possible for some underlying morphology to be discerned through the covering layer after the second impression, but this was not a consistent finding. Fuji II liquid and EDTA were effective in clinical cleansing of the tooth surface to the extent of exposing dentinal tubules in the abrasion defects. Several specimens showed the tubules greatly elongated, since they had been exposed at an angle rather than at right angles to the long axis of the tubule. Another specimen showed the tubules to be less numerous p e r unit area (Fig. 4). The pumice and water slurry applied with a rubber cup altered the appearance of the surface but not to the extent of revealing the tubular morphology.
ble 2 for Fuji IIF show that no change occurs by removal of the smear layer, unlike the use of EDTA, which did show a change, albeit not statistically significant, as a consequence of removal of calcium from the dentin surface. Exposure of the dentin surface by effective decalcifying agents (such as 40% solutions of EDTA, citric, lactic, or orthophosphoric acids) causes rapid reduction and eventually complete elimination of bonding
with polyacrylic-acid-based cements (Beech, 1972). The adverse effects of powerful demineralizing agents on bond strength are well-documented (Hood et al., 1981; 0ilo, 1978; Hinoura et at., 1986). However, there is conflict in the literature, based upon laboratory measurements, about the effect of smear layer removal (Powis et al., 1982; Beech et al., 1985; Shalabi et al., 1981). Smear-layer removal p r o c e d u r e s have been
Fig. 1. Tags of composite following tensile testing (Silar/Gluma with EDTA pre-treatment). Bar = 50 ~.m.
DISCUSSION Glass-ionomer C e m e n t s - The mechanism of adhesion of glass-ionomer cement is by ionic attraction between carboxylate ions in the cement and calcium ions in the tooth surface (Beech, 1972), although bonding to collagen has also been s u g g e s t e d (McLean and Wilson, 1977). The adhesive bond may be an intermediate layer formed between the bulk of the cement and the bulk of the tooth substance. Mount (1984) claimed that such a physico-chemical union is dependent upon total cleanliness of the dentin. The application of either PAA or TA provided a substantially clean dentinal surface (Figs. 1,3), and, with such "mild" cleansers, significant demineralization of the surface would not be expected. The results in Ta-
Fig. 2. Smear layer after preparation of dentin with 600-grade SiC paper. Bar --- 50 p.m.
Dextral Materio/s/Nol~e~tber 19~:9 381
Fig. 3. Smear layer in Fig. 2 removed by application of 25% TA for 20 sec. Bar = 50 r~m.
Fig. 4, Cervical abrasion lesion following Fuji II liquid treatment. Bar = 50 i~m.
advocated (Mount, 1984), but there is no evidence that improved clinical retention results. Indeed, retention rates of 97-100% after one year in non-undercut Class V abrasion cavities are easily achievable when only pumice and water prophylaxis are used (Tyas and Beech, 1985; Tyas et al., 1986). Such retention rates would not be improved by chemical removal of the smear layer.
It is likely that the chemical activity of the free polycarboxylic acid liquid in glass-ionomer cement at the time of placement dissolves, penetrates, and incorporates the smear layer, enabling adhesion to the body of the dentin to occur. The considerable variation in nature and access to the smear layer on dentin clinically makes effective removal of the smear layer difficult, and has the po-
382 W H I T E et al./SMEAR L A Y E R A N D BONDING
tential to reduce bonding if decalcification occurs. G l u m a - I n contrast to glass-ionomer cement, Gluma has been postulated to bond to dentin by reaction of the glutaraldehyde in Gluma with basic groups in dentinal collagen such as NH2 and OH (Eliades et al., 1985; Asmussen, 1985). Gluma is unique in being specific to collagen, and therefore a demineralized dentinal surface should be perfectly satisfactory for bonding, provided it is not seriously mechanically weakened. The combination of glutaraldehyde and 2-hydroxyethylmethacrylate (HEMA) in Gluma may also facilitate bonding by dissolution in the dentin surface. From the results in Table 2, those of Munksgaard and Asmussen (1985), and the clinical data of Tyas (1989), it is clear that the smear layer acts as a barrier to bonding of Gluma and must be removed. Gluma is the only material in this study which is not acidic, favors bonding through collagen, and benefits from smear-layer removal. Scotchbond L C and Pvisma Universal B o n d - I n common with glassionomer cement, Scotchbond LC and Prisma Universal Bond show a preference for a surface which has not been subject to s m e a r - l a y e r removal/decalcification (Table 2). This is consistent with some previous reports on Scotchbond (Solomon and Beech, 1985; Hill et al., 1983; Hinoura et al., 1986), although others are less clear-cut in their findings (Retief et al., 1986; Odin and 0ilo, 1986). Scotchbond LC and Prisma Universal Bond are acidic materials with anionic groups designed to interact with calcium ions. Agents which are capable of dissolving mineral in the smear layer will inevitably demineralize the u n d e r l y i n g dentin surface to some extent. The manufacturers of these two dentin bonding agents recognize this by not advocating chemical smear-layer removal. Presumably, like glass-ionomer cement, these acidic monomers can combine with or penetrate the smear layer. Although Scotchbond has a greater bond strength to enamel than to dentin (Solomon et al., 1985), increasing the mineral content of the dentin surface failed to increase the bond strength (Table 2). Dissolution in the dentinal collagen and the
strength/integrity of the dentin surface may be significant factors in optimization of bond strength. Scanning Electron M i c r o s c o p y Scanning electron micrographs revealed that only Fuji II liquid and EDTA will remove the smear layer from cervical abrasion lesions clinically. Duke et al. (1985) reported that in the laboratory most of the smear layer was removed when the surface was subjected to a pumice prophylaxis; however, the present study was unable to confirm that result. Hydrogen peroxide appeared to have no effect upon the smear layer surface, which agrees with the findings of Duke et al. (1985). Powis et al. (1982) also found that polyacrylic acid was effective in laboratory removal of the smear layer; however, in their study tannic acid was less effective than in the present study. Dentin Bonding and the Smear Layer - General ConsiderationsThe variation in the effect of smear layer removal upon the bonding of different adhesives to dentin suggests that routine smear-layer removal is contra-indicated, and should only be done when the need is established. Some dentin bonding agents for resin composites, and the polyacrylic-acid-based cements, have found commercial and clinical viability without the removal of the smear layer. Removal of the smear layer with agents capable of dissolving hydroxyapatite will inevitably reduce the proportion of this mineral in the dentin surface. The polyacrylic-acidbased cements and many dentin bonding agents favor a highly calcified surface. The decalcified surface after smear layer removal will be soft and proteinaceous. This ,surface is clearly preferred for Gluma, which is somewhat exceptional in bonding via dentin collagen. Removal of the plugs at the cut ends of dentin tubules as a consequence of smear layer removal may allow access of bacteria and chemicals to the pulp. Exudation of proteinaceous fluid from the tubules may also interfere with bonding, since many dentin bonding agents are or-
ganic hydrophobic compounds which are unable to compete with dentinal fluid. Furthermore, there is no evidence that penetration of bonding agent into dentinal tubules is advantageous to bonding; it does not give retention of resins clinically and is not analogous to resin tags in enamel. In the less-than-ideal environment of a busy dental practice, uniform removal of the smear layer without unblocking the dentinal tubules will be difficult to achieve. Adhesive techniques which can cope with the smear layer may be more likely to meet with clinical success.
REFERENCES ASMUSSEN, E. (1985): Clinical Relevance of Physical, Chemical and Bonding Properties of Composite Resins, Oper Dent 10:61-73. BEECH, D.R. (1972): Improvement in the Adhesion of Polyacrylate Cements to Human Dentine, Br Dent J 135:442445. BEECH, D.R.; SOLOMON, A.; and BERNIER, R. (1985): Bond Strength of Polycarboxylic Acid Cements to Treated Dentin, Dent Mater 1:154-157. BRANNSTR~M, M. (1982): Dentin and Pulp in Restorative Dentistry. London: Wolfe Medical Publication Ltd. DUKE, E.S.; PHILLIPS, R.W.; and BLUMERSHINE, R. (1985): Effects of Various Agents in Cleaning Cut Dentine, J Oral Rehabil 12:295-302. LICK, J.D.; WILKO, R.A.; ANDERSON, C.H.; and SORENSON, A.E. (1970): Scanning Electron Microscopy of Cut Tooth Surfaces and Identification of Debris by Use of the Electron Microscope, J Dent Res 49:1358-1368. ELIADES, G.L.; CAPUTO, A.A.; and VOUGIOUKLAKIS, G.J. (1985): Corhposition, Wetting Properties and Bond Strength with Dentin of Six New Dentin Adhesives, Dent Mater 1:170-176. GWINNE'Fr, A.J. (1984): Smear Layer: Morphological Considerations, Oper Dent (Suppl) 3:3-12. HILL, G.L.; JENSEN, M.E.; and ZIDAN, O. (1983): Shear-bond Strengths of a New Dentinal Bonding Agent: Pretreatment Effects, J Dent Res 221: Abst. No. 469. HINOURA, K.; MOORE, B.K.; and PHILLIPS, R.W. (1986): Influence of Dentin Surface Treatments on the Bond Strengths of Dentin-lining Ce-
merits, Oper Dent 11:147-154. HOOD, J.A.A.; CHILDS, W.A.; and EVA_~S, D.F. (1981): Bond Strength of Glass Ionomer and Polycarboxylate Cements to Dentine, N Z Dent J 77:141-144. LEVINE, R.S.; BEECH, D.R.; and GARTON, B. (1977): Improving the Bond Strength of Polyacrylate Cements to Dentine, Br Dent J 143:275277. McLEAN, J.W. and WILSON,A.D. (1977): The Clinical Development of the Glassionomer Cement. III. The Erosion Lesion, Aust Dent J 22:190-195. MOUNT, G.J. (1984): Glass Ionomer Cements: Clinical Considerations. In: Clinical Dentistry, James W. Clark, Ed., Philadelphia; Harper and Row. MUNKSGAARD, E.C. and ASMUSSEN, E. (1985): Dentin-Polymer Bond Mediated by Glutaraldehyde/HEMA, Scand J Dent Res 93:463-466. ODI~N, A. and 0ILO, G. (1986): Tensile Bond Strength of Dentin Adhesives, Dent Mater 2:207-213. 0ILO, G. (1978): Adhesive Bonding of Dental Luting Cements: Influence of Surface Treatment, Acta Odontol Scand 36:263-270. PASHLEY, D.H. (1984): Smear Layer: Physiological Coasiderations, Oper Dent (Suppl) 3:13-29. Powls, D.R.; FOLLER~tS, T.; MERSON, S.A.; and WILSON, A.D. (1982): Improved Adhesion of a Glass-ionomer Cement to Dentin and Enamel, J Dent Res 61:1416-1422. RETIEF, D.H.; GROSS, J.D.; BRADLEY, E.L.; and DENYS, F.R. (1986): Tensile Bond Strengths of Dentin Bonding Agents to Dentin, Dent Mater 2:7277. SHALABI, H.S.; ASMUSSEN, E.; and JORCENSEN, K.D. (1981): Increased Bonding of a Glass Ionomer Cement to Dentin by Means of FeC13, Scand J Dent Res 89:348-353. SOLOMON, A. and BEECH, D.R. (1985): Bonding of Composites to Dentin using Primers, Dent Mater 1:79-82. TYAS, M.J. (1989): Clinical Evaluation of Three Dentine Bonding Agents, Aust Dent J (in press). TYAS, M.J. and BEECH, D.R. (1985): Clinical Performance of Three Restorative Materials for Non-Undercut Cervical Abrasion Lesions, Aust Dent J 30:260-264.
TYAS, M.J.; BURNS, G.A.; BYRNE, P.F.; CUNNINGHAM,P.J.; DOBSON, B.C.; and WIDDOP, F.T. (1986): Clinical Evaluation of Scotchbond: One Year Results, Aust Dent J 31:159-164.
Dental Materials~November 1989 383