In vivo degradation of resin–dentinbonds produced by a self-etch vs. atotal-etch adhesive system

K. Koshiro1, S. Inoue2, T. Tanaka1,K. Koase1, M. Fujita1, M. Hashimoto1,H. Sano1

1Department of Oral Health Science, HokkaidoUniversity Graduate School of Dental Medicine,Sapporo, Japan; 2Division for General Dentistry,Hokkaido University Dental Hospital, Sapporo,Japan

Dental adhesives have been evolving at a rapid rate overthe past decade. A large part of this success is attributedto the significant advances in dentin bonding technology,at least in the short term (1, 2). One of the most recentdevelopments for dentists and researchers is the simpli-fication of the multistep bonding process, using one oftwo different approaches: the �total-etch� or �etch-and-rinse� system (2); and the �self-etch� system (2). Anotheraspect of this study is the durability of resin–dentinbonds under function for a long time period in an oralenvironment.To date, a controlled clinical study has shown that the

retention rates of restorations bonded with contempor-ary adhesives had improved to almost 100% for obser-vations over a 2–3-yr time-period (3). On the other hand,a number of in vitro and in vivo long-term studies on thedurability of the bond have reported that degradation ofthe resin–dentin interface occurs when they are exposedto oral fluids or water for months or years (4–12). Severalin vitro studies have shown that long-term water storageresulted in an increase in adhesive joint failures (5, 6, 9).It has also been reported that indirect exposure (i.e.enamel-protected) of resin–dentin interfaces to water for4 yr, produced by two-step �etch-and-rinse� adhesives,caused less degradation of the interface than directexposure. These results suggest that enamel protects theresin–dentin bond against degradation (6). In vivo studies

on this subject are still few (8, 10, 12), but they offerimportant clues to clarify the degradation process. Sanoet al. (10) reported that the in vivo bond strength ofresin–dentin bonds produced by a self-etching primingsystem was low, but stable, over a 1-yr time-period inmonkeys. Scanning electron microscopy (SEM) obser-vation of the fractured surfaces after bond strengthtesting revealed increased porosity over time in the upperportion of the hybrid layer and within the adhesive resin.Furthermore, an examination of resin–dentin bonds inhumans over a 1–3 yr time-period, using a three-steptotal-etch adhesive, showed that the bond strengthdecreased over time and was clearly related to failuremodes, suggesting that the hybrid layer degraded overtime (8). Moreover, the results of another 1-yr in vivostudy (12) using monkey teeth showed that resinousmaterial within the hybrid layer produced by a self-etching priming system seemed to increase in porosity, asobserved by SEM, indicating that the hybrid layerdegraded after 1-yr in the oral environment.The purpose of this study was to evaluate the durab-

ility of in vivo bonds between dentin and two adhesivesystems – a two-step self-etch and a two-step etch-and-rinse adhesive system – using the microtensile testingmethodology and SEM observation, in order to gainmore information about the degradation phenomenonwhich may occur at the interface. The tested hypothesis

Koshiro K, Inoue S, Tanaka T, Koase K, Fujita M, Hashimoto M, Sano H. In vivodegradation of resin–dentin bonds produced by a self-etch vs. a total-etch adhesive sys-tem. Eur J Oral Sci 2004; 112: 368–375. � Eur J Oral Sci, 2004

The purpose of this study was to evaluate the long-term durability of in vivo bondstrengths and the morphological changes of interfaces between dentin and twoadhesive systems. Class V preparations were prepared on the facial surfaces of 14intact teeth of two monkeys and restored with a combination of Unifil Bond/Z250 orSingle Bond/Z250. One year later, 10 additional teeth were restored with the samematerials and the monkeys were killed after 24 h. All of the restored teeth were sub-jected to microtensile bond strength (lTBS) testing. The debonded surfaces of thedentin sides were morphologically observed using Fe-scanning electron microscopy(SEM), as were the polished cross-sections of resin–dentin interfaces. For both UnifilBond and Single Bond, the lTBS at 24 h was significantly higher than that at 1 yr.Fe-SEM observations of polished cross-sectioned and fractured surfaces showed thatporosity within the hybrid layers produced by Single Bond increased over time.However, the interface produced by Unifil Bond revealed no noticeable changes inmorphology between 24-h and 1-yr specimens. It is concluded that even though thebond strengths of both adhesive systems declined over time, the bonding interfaceusing self-etching primers was relatively stable over time compared to the wet bondingsystem.

Dr Kenichi Koshiro, Department of Oral HealthScience, Hokkaido University Graduate Schoolof Dental Medicine, Kita 13 Nishi 7, Kita-Ku,Sapporo 060–8586, Japan

Telefax: +81–11–7064878E-mail: rgv250sp@dab.hi-ho.ne.jp

Key words: durability; degradation; total-etch;self-etch; microtensile bond test

Accepted for publication April 2004

Eur J Oral Sci 2004; 112: 368–375Printed in UK. All rights reserved

Copyright � Eur J Oral Sci 2004

European Journal ofOral Sciences

was that the adhesive interfaces would show morpholo-gical and physico-mechanical changes in vivo over time.

Material and methodsThe Ethics Committee of Hokkaido University GraduateSchool of Dental Medicine approved the animal experi-ment. The distribution of teeth used in this study is listed inTable 1. Two monkeys (Macaca fascicularis) were placedunder general anesthesia by intramuscular injection of22 mg kg)1 ketamine (Veterinary Ketalar 50; Sankyo, To-kyo, Japan). The surfaces of the 14 intact teeth were ultra-sonically cleaned to remove calculus prior to preparation.Very shallow, saucer-shaped Class V dentin preparationswere prepared on the facial surfaces of the 14 intact teethusing a high-speed tapered diamond bur (440; GC, Tokyo,Japan) under water spray. The bur was replaced for eachpreparation. All cavosurface margins were in enamel, butthe enamel was thinner near the gingival extension. Sevenpreparations (the first molars in maxilla and mandible) weretreated with Unifil Bond (GC), according to the manufac-turer’s instructions, restored with Z250 (3M ESPE, St Paul,MN, USA), and light cured for 20 s (Optilux 500, Deme-tron/Kerr, Danbury, CT, USA). The other 7 teeth (secondmolars in maxilla and mandible) were filled with SingleBond/Z250 (3M ESPE). Chemical formulations and appli-cation instructions of adhesives and resin composites testedare listed in Table 2. One year later, the monkeys werere-anaesthetized and 10 other intact teeth (central incisors

and canines in maxilla and mandible) were restored witheither Unifil Bond/Z250 or Single Bond/Z250 in a mannersimilar to that described above.The next day, the monkeys were killed by injection of

125 mg kg)1 suxamethonium chloride (Succin; YamanouchiPharmaceutical, Tokyo, Japan) and 3.8 mg kg)1 thiopentalsodium (Ravonal; Tanabe Seiyaku, Osaka, Japan). Then,the 24 restored teeth (14 treated for 1 yr and 10 treated for1 d) were surgically extracted from the maxilla or mandiblewith care taken to avoid placing any stress near the restoredsurfaces.

SEM observations of polished cross-sections

The 10 resin-bonded specimens (five each for the twoproduct combinations, six 1-yr and four 1-d restorations)were sectioned into slabs � 0.7 mm thick, perpendicular tothe adhesive interface using a low-speed diamond saw underwater coolant. The slabs were then fixed (for 12 h at 4�C) in0.1 m sodium cacodylate buffer, pH 7.4, containing 2.5%glutaraldehyde, rinsed three times (20 min each rinse) with0.2 m sodium cacodylate buffer, pH 7.4, followed by dis-tilled water for 1 min, and were then embedded in epoxyresin (SpeciFix-20; Struers, Tokyo, Japan). The embeddedspecimens were polished with 6-, 3-, and 1-lm diamondpastes (DP-Paste; Struers) under DP-Lubricant Blue cool-ant. The specimens were then immersed in 6 m HCl (WakoPure Chemical Industries, Osaka, Japan) for 30 s and in 1%NaOCl (Wako Pure Chemical Industries) for 10 min. Theywere then ultrasonically rinsed in distilled water for 30 s.

Table 1

Distribution of teeth

Cross-sectional SEM observation Microtensile bond strength test

Unifil Bond / Z250 Single Bond / Z250 Unifil Bond / Z250 Single Bond / Z250

1 d 1 yr 1 d 1 yr 1 d 1 yr 1 d 1 yr

U.I.LS. U.1M.R. U.I.R. U.2M.LS. U.I.LS. U.1M.R. U.I.R. U.2M.R.L.C.LS. U.1M.LS. L.I.R. L.2M.R. U.C.LS. U.1M.LS. U.C.R. U.2M.LS.

L.1M.LS. L.2M.LS. L.I.LS. L.1M.R. L.I.R. L.2M.R.L.1M.LS. L.2M.LS.

Two monkeys were used in this study. C, canine; I, central incisor; L, lower; LS, left side; 1M, 1st molar; 2M, 2nd molar; R, right side;U, upper. SEM, scanning electron microscope.

Table 2

Chemical formations and application procedures of adhesives and resin composites tested

AdhesivesResin composites Manufacturer Components Lot no. Composition Application procedures

Unifil Bond GC Primer 0011101 HEMA, 4-MET, ethanol,water

Apply for 20 s, air-dry

Adhesive resin 0011101 UDMA, HEMA, filler Apply and 10 s light-cureSingle Bond 3M ESPE Etchant

Primer &adhesive resin

2000121620001216

35% phosphoric acidBis-GMA, Vitremer copolymer,HEMA, ethanol, initiator, water

Apply for 15 s, 10 s rinseApply, blot-dry and 10 slight-cure

Z250 3M ESPE Shade A3 0AFL BIS-GMA, UDMA,BIS-EMA (ethoxylated bis-GMA),zirconia silica filler

Apply and 20 s light-cure

Bis-GMA, bisphenyl glycidyl methacrylate; DMA, dimethacrylate; HEMA, hydroxyethyl methacrylate; 4-MET, 4-methacryloyl-oxyethyl trimellitic acd; UDMA, urethane dimethacrylate.

Degradation of resin–dentin interface in vivo 369

Subsequently, specimens were dehydrated in ascendinggrades of ethanol (25–100%), immersed in hexamethyldisi-lazane (HMDS) for 10 min and placed on a filter paperinside a covered glass vial underneath a hood for 12 h.Ultimately, the specimens were placed on an aluminumstub, sputter-coated with gold, and observed using an Fe-SEM (S-4000; Hitachi, Tokyo, Japan) at an acceleratingvoltage of 10 kV.

Microtensile bond strength test

The 14 surfaces of resin composite restorations (seven res-torations each for the two product combinations, eight 1 yrand six 1 d) were etched with 37% phosphoric acid gel(Scothbond Etchant; 3M ESPE, London, ON, Canada) for40 s, rinsed with water and air-dried. The surfaces were thentreated with a silane-coupling agent (Activator in ClearfilPorcelain Bond; Kuraray, Osaka, Japan), followed byapplication of an adhesive resin (Clearfil Mega Bond;Kuraray), which was light cured for 10 s. Next, three or fourlayers of resin composite (Clearfil AP-X; Kuraray) wereincrementally placed on the treated surface up to 4 mmthick; each layer was light cured for 20 s. After storage inwater at 37�C for 24 h, the specimens were cut into� 0.7 mm slices, parallel to the long axis of the tooth, using alow-speed diamond saw (Isomet; Buehler, Lake Bluff, IL,USA) under water cooling. The number of slices obtainedvaried depending on the tooth size. Each slice was trimmedinto an hourglass shape to create a small bonding surfacearea, of � 1 mm2, using a super-fine diamond point (c-16ff;GC) in a high-speed hand-piece with copious water spray.The width and thickness of the bonded interface of eachspecimen was measured using digital calipers (CD-15C;Mitsutoyo, Kawasaki, Japan) prior to testing. Specimenswere mounted in a Ciucchi’s jig (13) using a cyanoacrylateadhesive (Model Repair II Blue; Dentsply/Sankin, Otahara,Japan). The specimens were loaded in tension, at a cross-head speed of 1 mm min)1, in a universal testing machine(EZ test; Shimadzu, Tokyo, Japan). Bond strength datawere analysed by one-way analysis of variance (anova)and Scheffe’s test at a 5% level of significance. Whenspecimens failed before testing, a bond strength of 0 MPawas included in the calculation of the mean lTBS. Thenumber of pretesting failures was explicitly noted.The mode of failure was determined at a magnification of100· using Fe-SEM, and recorded as �cohesive failure indentin�, �adhesive failure� or �mixed adhesive and/or cohe-sive failure�.

Fractured surface observation

After the microtensile testing, the debonded surfaces ofdentin sides were observed microscopically to evaluate themorphological differences between 1-yr and 1-d debondedsurfaces. The specimens were air-dried, mounted on analuminum stub, gold sputter-coated and examined using anFe-SEM at an accelerating voltage of 10 kV.

Results

SEM observation of polished cross–sections ofinterface

For Unifil Bond, the interfaces of both the 1-d and 1-yrspecimens showed no noticeable change in morphology.

Both showed thin hybrid layers, of <1 lm width, thatare typically produced by self-etching primer systems(Fig. 1A,B). In 1-yr specimens restored with SingleBond, gaps in the interface were seen. Specifically, por-tions of the hybrid layer had disappeared (Fig. 1D) and/or there was a large increase in porosity within the hybridlayer (Fig. 1F), when compared with the 1-d controlspecimens that exhibited thicker hybrid layers and longresin-tags (Fig. 1C,E).

Microtensile bond strength test

Microtensile bond strengths (lTBS) at 1 d and 1 yr, forspecimens bonded with the two adhesive systems, areshown in Table 3. For both Unifil Bond and SingleBond, a high lTBS at 1 d (38.7 and 48.1 MPa, respect-ively) decreased significantly (P < 0.05, 14.4 and11.7 MPa, respectively) after 1 yr of function. In addi-tion, a greater number of pretesting failures occurred inthe 1-yr specimens of Single Bond compared with UnifilBond, while no pretesting failures occurred in any of the1-d specimens.The results of the failure mode analysis are summar-

ized in Table 4. For Unifil Bond, at 1 d, almost half ofthe specimens exhibited adhesive failures. In contrast, at1 yr, the rate for this failure mode had decreased to 10%and cohesive failure in dentin was the predominant modeof failure. No pretesting failures occurred in specimensbonded with Unifil Bond. With respect to the specimensof Single Bond, no adhesive failure was detected and nopretesting failures were observed at 1 d. One year later,adhesive failures, including pretesting failures, increasedand mixed failures, had decreased by approximately one-third (Table 4).

Fractured surface observation

Fig. 2 shows the fractured surface morphologies classi-fied as cohesive failure in resin composite after lTBStesting. For both adhesives, at 1 d, filler particles wereenveloped by resin matrix (Fig. 2A,C). However, after1 yr of function, many voids were observed that ap-peared to be where filled particles had been, and thedetached filler particles were not covered with resinmatrix (Fig. 2B,D).Fig. 3 shows the surfaces of fractured dentin sides of

1-d and 1-yr specimens bonded with Unifil Bond, clas-sified as mixed adhesive failure and cohesive failurewithin the bottom of the hybrid layer. Similar imageswere seen in both specimens, showing a hybrid layer withresin-encapsulated collagen fibrils (Fig. 3C,D).Fig. 4 shows the surfaces of fractured dentin sides of

1-d and 1-yr specimens bonded with Single Bond, clas-sified as mixed adhesive failure between the top of theadhesive layer and the composite, and cohesive failure ofthe hybrid layer. Fig. 4A shows the fractured surface ofdentin side at 1 d, demonstrating cohesive failure withinthe hybrid layer that was densely penetrated by resinmatrix. However, the fractured surfaces of the 1-yrspecimens revealed the large spaces around the collagenfibrils within and/or at the bottom of the hybrid layer

370 Koshiro et al.

(Fig. 4B,C). In Fig. 4B, some collagen fibrils were cov-ered with an amorphous material, which could be eitherresidual resin or gelatinized collagen.

Discussion

The durability of bonds between resin and tooth sub-strates is of significant importance for the clinical lon-gevity of adhesive restorations. Bonding to enamel isthought to be reliable and durable, especially using etch-and-rinse adhesives (14). It has recently been shown thatbonding to enamel seals and protects the more vulner-able resin–dentin bond against water degradation(6,9,15). On the contrary, the long-term stability of resin-bonded dentin remains questionable (6, 9). In vitrolaboratory studies reported decreases in dentin bondstrength after long-term storage in water (6, 7, 15, 16).Although in vivo long-term durability studies reportedthe degradation of hybrid layers over time, they are stilllimited with respect to number and adhesive type. Thus,in the present study, currently available adhesive sys-tems, a two-step self-etch system (Unifil Bond) and atwo-step etch-and-rinse (total-etch, so called �wet bond-ing�) system (Single Bond), were used.In the present study, the lTBS for Unifil Bond signi-

ficantly decreased, within 1 yr, to a value of only 37% ofthe 1-d bond strength value (Table 3). Takahashi et al.(12) reported that, after functioning for 1 yr in themouth of monkeys, the bond strength of the sameadhesive decreased to 70% of the 1-d value (not a sta-tistically significant difference). The number of 1-yrspecimens in their study was less than half of the numberused in ours. This might account for some difference in

Table 3

Microtensile bond strength to dentin

Unifil Bond / Z250 Single Bond / Z250

Mean (MPa) SD Ptf/n Mean (MPa) SD Ptf/n

1 d 38.7 9.0 0/17 48.1 7.4 0/131 yr 14.4 9.3 1/18 11.7 9.3 5/17

Ptf/n, number of pretesting failures/total number of specimens.

Fig. 1. Observations of polished cross-sections of the resin–dentin interface by Fe-scanning electron microscopy (SEM).(A) Fe-SEM photomicrograph of a 1-d functioned resin–dentin interface produced by Unifil Bond. A thin hybrid layer (H) of<1 lm width is observed. BR, bonding resin, UD, unaffected dentin. (B) Fe-SEM photomicrograph of a resin–dentin interfaceproduced by Unifil Bond that was in function for 1 yr. It had a similar appearance to the 1-d specimen shown in Fig. 1A. BL,Bonding layer, UD, unaffected dentin, H, hybrid layer. (C) Fe-SEM photomicrograph of 1-d resin–dentin interface of Single Bond.A thicker hybrid layer (H) � 3 lm deep and long resin-tags (RT) can be seen. RC, resin composite, BL, bonding layer, UD,unaffected dentin. (D) Fe-SEM image of a 1-yr resin–dentin interface of Single Bond (different part of the interface from Fig. 1C).Note that there is a place where the hybrid layer disappeared (white arrow) and a gap formed between the resin composite andadhesive resin (asterisk). RC, resin composite; BR, bonding resin; UD, unaffected dentin. (E) High magnification of the hybrid layerof a 1-d Single Bond specimen (Fig. 1C). Hybrid layer (H) of � 1 lm thickness are observed with collagen fibrils sufficientlyinfiltrated by the resin matrix. No gap or detached filler particles were observed. BR, bonding resin; UD, unaffected dentin.(F) Higher magnification of the hybrid layer of the 1-yr Single Bond specimen (Fig. 1D). Porosity (white arrows) of hybrid layer(H) was increased compared with Fig. 1E. BR, bonding resin; RT, resin tag.

Table 4

Failure mode analysis

Cohesivefailure

in dentin

Adhesive failure(+ pretesting

failure)

Mixed adhesiveand/ or

cohesive failure Total

Unifil Bond1 d 5 (29.4%) 8 + 0 (47.1%) 4 (23.5%) 171 yr 14 (77.8%) 2 + 0 (11.1%) 2 (11.1%) 18

Single Bond1 d 5 (38.5%) 0 + 0 (0.0%) 8 (61.5%) 131 yr 5 (29.4%) 3 + 5 (47.1%) 4 (23.5%) 17

Degradation of resin–dentin interface in vivo 371

the results, but it should be noted that both studiesshowed reductions in resin–dentin bond strengths. Theresult of another in vivo report, using a self-etch adhesive(Clearfil Liner Bond II; Kuraray), showed that the bondstrength was stable (at � 19 Mpa) throughout the 1-yrtesting period (10). One possible reason for the reportedstability in bond strength may have been a lower initialbond strength from that expected. In human teeth, thisadhesive has typically shown a resin–dentin bondstrength of � 40 MPa, for both normal superficial anddeep dentin, 1 d after bonding (17, 18).For Single Bond, a �wet bonding� system, the lTBS

decreased significantly over time in a manner similar tothat seen with Unifil Bond. This result was in accordancewith another in vivo study using a wet bonding system.Hashimoto et al. (8) reported that the resin–dentin bondstrength of Scotchbond Multi-Purpose (3M ESPE) wasreduced from 28.9 MPa to 9.1 MPa over 1–3 yrs in vivo,using human deciduous teeth. It has been reported thatindirect exposure of the resin–dentin interface to waterfor 4 yr did not significantly decrease the lTBS ofScotchbond 1 (the same material as Single Bond inEurope; 3M ESPE), but direct exposure of resin–dentinbonds to water caused the lTBS of this adhesive todecrease significantly (6). In the present study, prepar-ation margins were surrounded with enamel. Theexpectation was that the resin–dentin interface should beprevented from direct water invasion by the protectiveseal of the surrounding enamel margins. However, thelTBS of both adhesive systems decreased significantly

over time. Repeated occlusal forces, configuration fac-tors, outward fluid flow from dentinal tubules in vitalteeth, fluid permeation through the adhesive interfaceand/or thermal stress in the mouth of monkeys, mighthave contributed to a more rapid degradation of resin–dentin bonds in vivo. Although the lTBS values of bothadhesive systems were similarly decreased over time,material-specific morphological differences wereobserved at resin–dentin interfaces and debonded surfa-ces of dentin by Fe-SEM observation. The SEM imagesof the resin–dentin interface produced by Unifil Bondwere unchanged over time (Fig. 1A,B), whereas degra-dation of the interface occurred in the case of SingleBond. SEM images of 1-d specimens (Fig. 1C,E) indi-cated good penetration of resinous components intointerfibrillar spaces and the formation of a thick hybridlayer and long resin tags at the interface of Single Bondand dentin. Exposed collagen fibrils appeared to be wellenveloped with resinous component. However, 1 yr later,gaps at the interface, indicating disappearance of thehybrid layer (Fig. 1D) and increasing porosity (Fig. 1F),were found. The wet bonding system uses a strong acid,such as phosphoric acid, to etch the tooth, resulting incomplete dissolution of hydroxyapatite crystals andexposure of naked collagen fibrils. As long as the colla-gen fibrils are enveloped by adhesive resin, the resin mayprotect the fibrils from acid, base, and enzymatic attack.Cyclic masticatory function in the oral environment mayfatigue the integrity of resin–enamel bonds, thereby per-mitting micro- or nano-leakage of the peripheral enamel

Fig. 2. Fe-scanning electron microscopy (SEM) observation of fractured surface after lTBS testing. The fracture was classified ascohesive failure in composite resin. (A) Dentin side of a 1-d specimen produced by Unifil Bond. Filler particles were denselyenveloped by resin matrix; no gap was observed. (B) Dentin side of a 1-yr specimen produced by Unifil Bond. Detached fillerparticles (white arrow) and many voids (asterisk) were observed. Detached filler particles are not covered with matrix resin.(C) Dentin side of a 1-d specimen produced by Single Bond. Filler particles were densely enveloped by resin matrix and no gap wasobserved, similar to Fig. 1A. (D) Dentin side of a 1-yr specimen produced by Single Bond. Detached filler particles (white arrow) andmany voids (asterisk) were observed. Detached filler particles are not covered with matrix resin.

372 Koshiro et al.

seal. This, in turn, could extract resinous materials and/orlead to degradation of both resin and exposed collagenfibrils. Observation of fractured interfaces after micro-tensile bond testing, in the present study, seemed to con-firm this speculation. The fractured surface of 1-dspecimens showed a dense hybrid layer that consisted ofresin-enveloped collagen fibrils and resinmatrix (Fig. 4A).One year later, the porosity of the hybrid layer hadincreased. Almost all resin components, which hadoccupied interfibrillar spaces, may have been extracted,and the resin-coated collagen fibrils also disappeared(Fig. 4C). Residual resin and gelatinized collagen wereobserved (Fig. 4B,C). Thus, with respect to morpho-logical change of resin–dentin interfaces observed inthis study, SEM images suggested that the degradationof the bonding interface was dependent upon the typeof system used.Takahashi et al. (12) reported, using the same adhe-

sive (Unifil Bond), that the surface porosity of the 1-yrfractured specimens seemed to have increased within thehybrid layer when compared to that of the 1-d speci-mens. They speculated that it might be a result of therelease of uncured water-soluble monomers over a longperiod of time. However, degradation at the fracturedinterface over time was not detected in the current study(Fig. 3), as the cross-sectional morphology revealed nonoticeable changes (Fig. 1A,B), even though the lTBS

was reduced over time (Table 3). One possible explan-ation is the fact that Unifil Bond contains 4-MET as afunctional monomer, which has been reported tohave a chemical bonding capacity to hydroxyapatitethat remained around the collagen fibrils, even after self-etching treatment, as a result of its relatively high pHvalue (� 2) (19). Moreover, Yoshida et al. claimthat interactions between hydroxyapatite crystals andfunctional monomers of self-etching primers maycreate insoluble calcium salts (20). These insoluble saltsmay prevent the loss of calcium from the matrix overtime. This might contribute to the stability in morphol-ogy over a period of 1 yr, using the self-etching systemcompared with the wet bonding system. A further in-depth study should be performed to clarify this.In the present study, degradation occurred within the

resin composite used (Fig. 2). At 1 d, filler particles werenot detached from the resin matrix and no voids wereobserved between the filler particles and resin matrix forZ250 (Fig. 2A,C, respectively). However, after 1 yr offunction within an oral preparation, the filler particlesbecame detached from the resin matrix and many voidsappeared in the matrix where the particles had beenlocated (Fig. 2B,D). This phenomenon agrees with theresults of previous reports (8, 10). The silane-couplingagent on the surface of the filler particles might degrade,over time, through hydrolysis (21, 22).

Fig. 3. Fe-scanning electron microscopy (SEM) photomicrographs of fractured surface after lTBS testing of Unifil Bond.(A) Dentin side of a 1-d specimen. Failure was observed within the adhesive (BR) to the bottom of the hybrid (BHL). The hybridlayer consists of resin-enveloped collagen fibrils and surrounding apatite crystals. (B) Higher magnification of the bottom of thehybrid layer (BHL) of Fig. 3A. A dense hybrid layer is observed. Adhesive resin is seen to penetrate well into the bottom of thehybrid layer. (C) Dentin side of a 1-yr specimen produced by Unifil Bond. Mixed failure occurred with the adhesive (BR) to thebottom of the hybrid (BHL). No degradation of adhesive resin or the hybrid layer, over time, was observed. (D) Higher magnifi-cation of the bottom of the hybrid layer (BHL) of Fig. 3C. No morphological differences were found compared with the 1-d specimen(Fig. 3B).

Degradation of resin–dentin interface in vivo 373

It is concluded, within the limits of this study, thatbonding interfaces formed using self-etching primersappeared to be more stable over time compared to thoseformed with the wet bonding system, even though thebond strengths of both adhesive systems decreased overtime. Further in-depth study is needed to determine howthe degradation occurs, in order to develop long-lastingdentin-bonding systems.

Acknowledgments – This work was supported, in part, byGrant-in-Aid for Science Research from the Japan Society forPromotion of Science no. 11470401 and no. 15390573. Wethank all manufacturers for the generous donation of adhesivesand composites investigated.

References1. Tay FR, Pashley DH. Dental adhesives of the future. J

Adhesive Dent 2002; 4: 91–103.2. Van Meerbeek B, De Munck J, Yoshida Y, Inoue S, Vargas

M, Vijay P, Van Landuyt K, Lambrechts P, Vanherle G.

Adhesion to enamel and dentin. Current status and futurechallenges. Oper Dent 2003; 28: 215–235.

3. Van Meerbeek B, Vargas M, Inoue S, Yoshida Y, Peu-

mans M, Lambrechts P, Vanherle G. Adhesives andcements to promote preservation dentistry. Oper Dent 2001;26: 119–144.

4. Akimoto N, Tkamizu M, Kawano A. Clinical evaluation of aself-etching system. J Jpn Conserv Dent 2000; 43: 1271–1280.

5. Armstrong SR,Keller JC, BoyerDB. The influence of waterstorage and C-factor on the dentin–resin composite microten-sile bond strength and debond pathway utilizing a filled andunfilled adhesive resin. Dent Mater 2001; 17: 268–276.

6. De Munck J, Van Meerbeek B, Yoshida Y, Inoue S, VargasM, Suzuki K, Lanbrechts P, Vanherle G. Four-year waterdegradation of total-etch adhesives bonded to dentin. J DentRes 2002; 82: 136–140.

7. Gwinnett AJ, Yu S. Effect of long-term water storage ondentin bonding. Am J Dent 1995; 8: 109–111.

8. Hashimoto M, Ohno H, Kaga M, Kudo K, Sano H, Oguchi

H. In vivo degradation of resin–dentin bonds in humans over 1–3 years. J Dent Res 2000; 79: 1385–1391.

9. Hashimoto M, Ohno H, Sano H, Tay FR, Kaga M, Kudou

Y, Oguchi H, Araki Y, Kubota M. Micromorphologicalchanges in resin–dentin bonds after 1 year water storage.J Biomed Mater Res 2002; 63: 306–311.

Fig. 4. Fe-scanning electron microscopy (SEM) photomicrographs of the fractured surface after lTBS testing of Single Bond.(A) Dentin side of the 1-d specimen. Fracture occurred from the top of the hybrid layer to the bottom. The hybrid layer is denselyformed. MHL, middle of hybrid layer; BHL, bottom of the hybrid layer. (B) Dentin side of a 1-yr specimen. This image shows thearea in the middle of the hybrid layer (MHL). Compared with Fig. 4A, the resin matrix within the hybrid layer seems to be extracted.Note that only a few collagen fibrils covered with resin matrix are observed. (C) Dentin side of the 1-yr specimen. This image showsthe area of the demineralized dentin zone (DD) and/or the bottom of the hybrid layer. Note the increase in the size of the interfibrilarspaces within loose collagen network. Resinous components that filled the hybrid layer and enveloped collagen fibrils are only presentin a few sites (arrows).

374 Koshiro et al.

10. Sano H, Yoshikawa T, Pereira PNR, Kanemura N, Mori-

gami M, Tagami J, Pashley DH. Long-term durability ofdentin bonds made with a self-etching primer, in vivo. J DentRes 1999; 78: 906–911.

11. Swift EJ Jr, Perdigao J, Heymann HO, Wilder AD Jr,Bayne SC, May KN Jr, Sturdevant JR, Roberson TM.Eighteen-month clinical evaluation of a filled and unfilleddentin adhesive. J Dent 2001; 29: 1–6.

12. Takahashi A, Inoue S, Kawamoto C, Ominato R, Tanaka T,Sato Y, Pereira PNR, Sano H. In vivo long-term durability ofthe bond to dentin using two adhesive systems. J Adhesive Dent2002; 4: 151–159.

13. Paul SJ,WelterDA,GhaziM, PashleyDH. Nanoleakage atdentin adhesive interface vs m-tensile bond strength. Oper Dent1999; 24: 181–188.

14. Frankenberger R, Kramer N, Petschelt A. Long-termeffect of dentin primers on enamel bond strength and marginaladaptation. Oper Dent 2000; 25: 11–19.

15. Kiyomura M. Bonding strength to bovine with 4-META/MMA-TBB resin: long term stability and influence of water.J Jpn Soc Dent Mater Dev 1987; 6: 860–872.

16. HashimotoM, Ohno H, Sano H, KagaM, Oguchi H. In vitrodegradation of resin–dentin bonds analyzed by microtensile

bond test, scanning and transmission electron microscopy.Biomaterials 2003; 24: 3795–3803.

17. Inoue S, Vargas M, Abe Y, Yoshida Y, Lambrechts P,Vanherle G, Sano H, Van Meerbeek B. Microtensile bondstrength of eleven comtemporary adhesives to dentin. J Adhe-sive Dent 2001; 3: 237–245.

18. Yoshikawa T, Sano H, Burrow MF, Pashley DH. Effect ofdentin depth and cavity configuration on bond strength. J DentRes 1999; 78: 898–905.

19. Inoue S, Vargas M, Abe Y, Yoshida Y, Lambrechts P,Vanherle G, Sano H, Van Meerbeek B. Microtensile bondstrength of eleven contemporary adhesives to enamel. Am JDent 2003; 16: 329–334.

20. Yoshida Y, Nagakane K, Fukuda R, Nakayama Y,Okazaki M, Shintani H, Inoue S, Tagawa Y, Suzuki K,Demunck J, Van Meerbeek B. Comparative study on ad-hesive performance of functional monomers. J Dent Res 2004;83: 454–458.

21. Drummond JL, Botsis J, ZhaoD, Samyn J. Fracture propatiesof aged and post-prossed dental composite. Eur J Oral Sci 1998;106: 661–666.

22. Soderholm KJ. Water sorption in a bis (GMA) /TEGDMAresin. J Biomed Mater Res 1984; 18: 271–279.

Degradation of resin–dentin interface in vivo 375