Effect of sodium hypochlorite on dentin bonding with apolyalkenoic acid-containing adhesive system

Raquel Osorio,1 Laura Ceballos,1 Franklin Tay,2 Miguel A. Cabrerizo-Vilchez,3 Manuel Toledano1

1Dental Materials Department, Faculty of Dentistry, University of Granada, Spain2Conservative Dentistry, Faculty of Dentistry, University of Hong Kong, Hong Kong SAR, China3Applied Physics Department, Faculty of Physics Sciences, University of Granada, Spain

Received 31 May 2001; revised 10 August 2001; accepted 5 September 2001

Abstract: The purpose of this study was to evaluate theeffect of sodium hypochlorite (NaOCl) treatment on dentinbonding by means of contact angle (CA), shear bondstrength (SBS), and microleakage (ML) measurements. Ul-trastructure and nanoleakage (NL) of the interfaces wereexamined by transmission electron microscopy (TEM). ForCA, SBS, and TEM evaluation, human molars were sec-tioned to expose dentin surfaces and were either acid-etched(35% H3PO4) or further treated with 5% NaOCl for 2 minbefore the application of Single Bond adhesive. CAs weremeasured using the Axisymmetric Drop Shape Analysistechnique. The Watanabe testing assembly was used for SBSevaluation. ML was assessed by a dye penetration method.NL was examined using a silver-staining technique. The re-sults showed that CA values decreased after acid etching

and even more after NaOCl treatment. NaOCl treatmentproduced lower SBS than acid-etched dentin. Both ML val-ues and NL manifestations were similar for NaOCl-treatedand acid-etched dentin. NaOCl did not completely removethe collagen matrix. NL was manifested along the base ofhybrid layers and within the polyalkenoic acid copolymerin both groups. Adverse chemical interactions could haveoccurred between the remnant collagen matrix and/or min-eralized dentin after NaOCl treatment. There is no addi-tional advantage in using NaOCl treatment with this adhe-sive. © 2002 John Wiley & Sons, Inc. J Biomed Mater Res 60:316–324, 2002; DOI 10.1002/jbm.10074

Key words: dentin; collagen; adhesion; acid etching; sodiumhypochlorite treatment

INTRODUCTION

Phosphoric acid is commonly used to etch hardtooth tissues in attempts to improve adhesives infil-tration and retention. Infiltrating the etched and par-tially demineralized dentin surface with adhesivescontaining hydrophilic adhesive resin monomers isconsidered essential for improving bonding at theresin/tissue interface.1 Micromechanical interlockingof polymerized resins within the network of exposedcollagen fibrils resulted in the formation of the “hy-brid layer.”2

This demineralized dentin surface sometimespresents several characteristics that were thought toplay a negative role in dentin adhesion:

1. After conditioning dentin with acidic agents, the

dense web of collagen fibrils becomes a low-energy surface substrate.3

2. During acid etching, collagen fibrils undergostructural changes that affect the extent of resininfiltration.4 Although collapse and shrinkage ofcollagen fibrils are minimized in the presence ofwater,5 reexpansion of a collapsed demineralizedcollagen matrix remains a considerable problemwith bonding to dentin.6,7

3. Part of acid demineralized dentin collagen re-mains in a destabilized state that is susceptible tohydrolysis and enzymatic degradation.8–11

4. Incomplete resin infiltration within the deminer-alized dentin results in a weak collagen-rich zonesusceptible to microleakage12 or nanoleak-age.13,14

Removal of the collagen fibrils with a deproteiniz-ing agent such as sodium hypochlorite (NaOCl)15 mayfacilitate the infiltration of adhesive resins into a den-tin substrate.16 Depending on the specific compositionof each dentin adhesive, application of NaOCl mayeither increase or decrease dentin bond strengths.17–19

Correspondence to: R. Osorio; e-mail: toledano@platon.ugr.es

Contract grant sponsor: CICYT, Spain; contract grantnumbers: MAT98-0937-C02, MAT01-2843-C02.

© 2002 John Wiley & Sons, Inc.

The purpose of this in vitro study was to evaluate theeffect of NaOCl treatment on the use of a two-step,polyalkenoic acid-containing, self-priming dentin ad-hesive. Comparisons between NaOCl-treated etcheddentin and acid-etched dentin were evaluated bymeans of various attributes, including wettability orcontact angle (CA), shear bond strength (SBS), andmicroleakage (ML) measurements. The ultrastructureof the resin-dentin interfaces was also investigatedwith the use of transmission electron microscopy(TEM). Nanoleakage that occurred along the bondedinterfaces were assessed using a silver-staining tech-nique and examined also by TEM. The null hypothesistested was that there is no difference among theseattributes irrespective of whether acid-etched dentin istreated with NaOCl before the application of the poly-alkenoic acid-containing adhesive.

MATERIAL AND METHODS

CA measurements

The ADSA-CD technique (Axisymmetric Drop ShapeAnalysis—Contact Diameter) was used for contact anglemeasurement. The contact diameter of the drop is deter-mined from a video image of the top view of the sessiledrop. This program permits measurement of the contactangle of a sessile drop even when it lacks a symmetricalprofile. Each 0.3-�L drop was dispensed with a microsy-ringe (Gilmont, Barrington, IL). Images were captured 30 safter a drop of liquid was dispensed from the microsyringe,using a microvideo system (Leica Wild M 32 Microscopy,Wild, Heerbrugg, Switzerland), at a magnification of 10×.The video signals were transmitted to a Data TranslationFrame Grabber Board (Data Translation, Malboro, MA) with256 gray levels and 800 × 600 square pixel resolution. APentium III computer (Intel, London, UK) was used to ac-quire the image from the Frame Grabber and to performcomputations with a software program that enables the con-tact angle of the sessile drop to be derived. The completeset-up was mounted on a pneumatic antivibration table. TheADSA-CD software20 utilized the contact diameter of thesessile drop, the density, and the interfacial tension of theliquid to compute the contact angle.

Twenty caries-free extracted human third molars, refrig-erated in a solution of 0.5% chloramines, for up to 1 monthafter extraction, were used. CA angles were measured onsuperficial and deep dentin according to methods describedin Toledano et al.,21 first with distilled and deionized waterand second with Single Bond adhesive (3M Dental Products;also known as Scotchbond 1).

In the first group (n = 10), each ground dentin surface(600-grit) was acid-etched with 35% phosphoric acid gel(Scotchbond etchant; 3M Dental Products, St. Paul, MN) for15 s, rinsed for 10 s, and then kept moist. In the secondgroup (n = 10), after acid-etching in the manner similar tothe first group, deproteinized dentin surfaces were pro-

duced by the application of 5% NaOCl to the etched dentinfor 2 min. During deproteinization, fresh NaOCl was reap-plied after the first minute, under constant agitation. TheNaOCl-treated surfaces then were rinsed thoroughly for 30s and allowed to remain moist before CA measurement.

SBS testing

Forty noncarious human third molars were used for thispart of the study. The extracted teeth were stored, cleaned,and mounted with cold-cure acrylic resin and randomly di-vided into two groups as previously described (n = 20). Foreach group, half of the teeth were sectioned below the DEJ,and the rest were sectioned 1 mm below the original level, inorder to expose either superficial (subgroup A) or deep den-tin (subgroup B). Bonding surfaces were ground flat underwater irrigation using 600-grit SiC abrasive paper.

SBS testing was performed using the Watanabe test as-sembly and according to the procedure described in Tole-dano et al.22 In group I, the flat unrestricted circular bondingsurface (4 mm in diameter) was etched and bonded withSingle Bond, using a moist bonding technique according tothe manufacturer’s directions. In group II, 5% NaOCl wasapplied to the etched dentin for 2 min as previously de-scribed, before bonding with the dentin adhesive. The ad-hesive was gently air-dried for 2–5 s and light-cured for 10s (Optilux 400; Demetron Research Corp., Danbury, CT). Ahybrid resin composite (Z100 A3 shade; 3M Dental Prod-ucts) was applied in three 1-mm increments and light-activated for 40 s each.

After storing in distilled water at 37°C for 24 h, the speci-mens were thermocycled for 500 cycles between 6 and 60°Cwith a dwell time of 30 s.23 SBS evaluation was conductedusing an Instron 4411 universal testing machine (InstronCorp., Canton, MA) operating at a crosshead speed of 1mm/min. All fractured specimens were examined with astereomicroscope (Model SZ4045TR; Olympus Optical Co.GMBH, Hamburg, Germany) at 20× magnification to deter-mine the mode of failure.

ML evaluation

Twelve freshly extracted, intact, caries-free human thirdmolars were used for this part of the study. Class V cavitieswere prepared on the buccal and lingual surfaces of eachtooth according to the procedure described in Toledano etal.24 The specimens were randomly assigned to two etchedand deproteinized groups as previously described. Eachcavity was bonded with Single Bond and restored with Z100resin composite in two increments. Excess material was re-moved with a No. 170 bur, followed by finishing and pol-ishing with the Sof-lex disc system (3M Dental ProductsDivision). The restored teeth were stored in distilled waterfor 24 h and thermocycled as previously described. The dye(0.5% solution of basic fuchsin) penetration test was per-formed according to the protocol used in Toledano et al.24

For each tooth, sections corresponding to the mesial, central,

317EFFECT OF NaOCl ON DENTIN BONDING

and distal portion of the restoration interfaces were exam-ined along the occlusal and gingival margins with a stereo-scopical microscope at 16× magnification. The extent of dyepenetration along the occlusal and gingival margins of thetwo groups of restorations was evaluated blindly by twoobservers. Scoring was based upon the following criteria: 0,no dye penetration; 1, dye penetration along the interface toone half the depth of the cavity wall; 2, dye penetration tothe full depth of the cavity wall, but not including the axialwall; and 3, penetration to and along the axial wall. Thesection with the maximum leakage value recorded for eachcavity was selected for statistical analysis.

Statistical analyses

CA and SBS measurements from the four designated sub-groups were analyzed using a two-way analysis of variancedesign. The statistical analyses evaluated the effect of thetwo experimental factors: (1) dentin treatment (acid-etchedvs. deproteinization) and (2) dentin depth (superficial vs.deep dentin) as well as the interaction of these two factors onthese measurements; pairwise multiple comparisons wereperformed using Student-Newman-Keuls (SNK) test.

Occlusal and gingival microleakage scores in the four sub-groups were analyzed by the nonparametric Kruskal-Wallisanalysis of variance on ranks. Statistical significance wasalways considered at the 0.05 level.

TEM examination of resin-dentin interfaces

Eight caries-free human third molars were used in thispart of the study. Two dentin surfaces were created accord-ing to the method previously described for each of the fourdesignated subgroups (i.e., acid-etched or deproteinized, su-perficial, and deep dentin). They were bonded using SingleBond and restored with a light-cured, lining resin composite(Protect Liner F; Kuraray Co. Ltd., Osaka, Japan) to facilitatesubsequent ultramicrotomy. One half of each bonded toothwas used for examination of the resin-dentin interfaces.Bonded specimens were fixed in Karnovsky’s fixative (2.5%glutaraldehyde and 2% paraformaldehyde in 0.1M cacodyl-ate buffer, pH 7.3) for 1 h and rinsed thoroughly with 0.1Msodium cacodylate buffer. Ethylene diamine tetra-acetic acid(EDTA)-demineralized, epoxy resin-embedded, 70- to 90-nm-thick ultrathin sections were prepared according to theTEM protocol described in Tay et al.25 Sections were double-stained with uranyl acetate (UA) and Reynold’s lead citrate(LC) to examine the overall status of the modified resin-dentin interfaces. Additional sections were stained withphosphotungstic acid and uranyl acetate to examine the sta-tus of the collagen fibrils within the hybrid layer. They wereexamined with a TEM (Philips EM208S, Eindhoven, TheNetherlands) operating at 80 kV.

Silver-staining technique for nanoleakage evaluation

The other half of each bonded tooth described in the pre-vious section was used for this part of the study. One 0.9-

mm-thick slab was prepared from each tooth by sectioningperpendicularly to the bonded interface with the use of us-ing a slow-speed saw (Isomet, Buehler Ltd., Lake Bluff, IL)under water lubrication. These slabs were coated with fast-setting nail varnish applied 1 mm from the bonded inter-faces. Without allowing these slabs to be grossly dehy-drated, they were immersed immediately in the 50 wt %AgNO3 solution for 24 h, according to the protocol of Sanoet al.26 The silver-impregnated slabs were rinsed in distilledwater and placed in photo-developing solution for 8 h undera fluorescent light to facilitate reduction of the silver ionsinto metallic silver particles. EDTA was used to remove thehydroxyapatite. The slabs were prepared for TEM examina-tion as described previously. Sections were examined eitherunstained or double-stained with UA and LC.

RESULTS

ANOVA, SNK results, and CA measurements onsuperficial and deep dentin after the different dentintreatments are represented in Table I. Regardless ofthe chemical nature of the liquid drop (water or resin),

TABLE IContact Angles Measured on Superficial and Deep

Dentin after the Different Dentin Treatmentsa

Dentin

Water ContactAngle (deg)*

Resin ContactAngle (deg)*

Superficial Deep Superficial Deep

Ground 39.8 ± 13.0a 23.2 ± 10.0b N/A N/AAcid etched 18.8 ± 10.0c 14.6 ± 10.0d 8.9 ± 3.5A 2.7 ± 1.0B

5% NaOCl 3.6 ± 1.0e 4.1 ± 2.6e 2.5 ± 0.7B 1.2 ± 0.3B

aValues are means ± standard deviation.*CA values measured with water and resin were analyzed

separately. Subgroups with the same letter are not statisti-cally significant (p > 0.05). ANOVA for the dependent vari-able water contact angle: p < 0.001; R2 = 0.61. Dentin treat-ment: F = 26.22, p < 0.001. Dentin depth: F = 6.6; p = 0.01.Interactions: F = 3.57, p = 0.03. ANOVA for the dependentvariable resin contact angle: p < 0.001; R2 = 0.77. Dentintreatment: F = 20.96, p < 0.001. Dentin depth: F = 17.77; p =0.001. Interactions: F = 7.83, p = 0.01.

TABLE IIShear Bond Strength Values (MPa) Obtained from

Superficial and Deep Dentin after the DifferentDentin Treatmentsa

Dentin Treatment

Dentin Depth

Superficial* Deep*

Acid-etched 22.5 ± 3.4a 23.4 ± 5.5a

5% NaOCl 15.7 ± 7.4b 12.8 ± 5.0b

aValues are means ± standard deviation.*Subgroups with the same letter are not statistically sig-

nificant (p > 0.05). ANOVA for the dependent variable shearbond strength: p < 0.001; R2 = 0.42. Dentin treatment: F =24.94, p < 0.001. Dentin depth: F = 0.32; p = 0.57. Interactions:F = 1.14, p = 0.30.

318 OSORIO ET AL.

lower mean CA values were recorded for deep dentin,except for the NaOCl-treated subgroups in which nosignificant differences were observed between super-ficial and deep dentin. NaOCl-treated dentin resultedin lower mean CA than etched dentin and etched den-tin in lower CA than grounded dentin.

SBS results are summarized in Table II. Acid-etched

specimens attained higher SBS than NaOCl-treatedspecimens. In both treatment groups, there were nosignificant differences between deep and superficialdentin.

Table III shows ML scores after different dentintreatments. Gingival scores were higher than occlusalin both groups. No differences were found betweenacid-etched and NaOCl-treated specimens.

TEM results are showed in Figures 1 to 4. Apartfrom the distribution of dentinal tubules, ultrastruc-ture of the resin-dentin interfaces in both groups ap-peared similar for superficial and deep dentin. Thehybrid layers in the acid-etched group were palelystained except for the surface portion that was infil-trated by the electron-dense polyalkenoic acid copoly-mer [Fig. 2(A)]. Phase separation of the latter resultedin electron-dense globular formations within the ad-hesive layer as well as the accumulation of a layer ofsimilar electron density over the top of the deminer-alized dentin [Fig. 2(B)]. Banded collagen fibrils within

TABLE IIIMicroleakage Scores after Different Treatments

Occlusal* Gingival*

0 1 2 3 0 1 2 3

Acid-etched 11 1 0 0 a 2 0 1 9 b5% NaOCl 10 1 1 0 a 0 1 4 7 b

*Subgroups with the same letter are not statistically sig-nificant (p > 0.05).

Kruskal-Wallis ANOVA for the dependent variable mi-croleakage: treatment: �2 = 3.57; p = 0.31. Occlusal or gingi-val margin: �2 = 33.97; p < 0.001.

Figure 1. TEM micrographs of Single Bond bonded toacid-etched dentin. Unless specified, sections were stainedwith uranyl acetate (UA) and lead citrate (LC). (A) Low-magnification view of the resin-dentin interface in super-ficial dentin. The polyalkenoic acid copolymer occurred asa separate electron-dense phase and could be seen asglobular structures (pointer) within the adhesive layer anda thin continuous layer (arrow) above the 5-�m-thick hy-brid layer. Bar = 3 �m. (B) High-magnification view of thehybrid layer in (A), showing that it was palely stainedexcept for the superficial 300 nm that was infiltrated by thepolyalkenoic acid copolymer. A diffusion gradient of thepolyalkenoic acid could be vaguely recognized (arrow).Bar = 300 nm. (C) A section that was stained using phos-photungstic acid (PTA) and UA. Banded collagen fibrilswere present within the hybrid layer. However, the fibrilswere swollen and interfibrillar spaces (arrowhead) wereminimal. Bar = 100 nm. C, lining resin composite; A, ad-hesive layer; P, polyalkenoic acid copolymer; H, hybridlayer; D, laboratory demineralized dentin.

319EFFECT OF NaOCl ON DENTIN BONDING

the hybrid layer appeared swollen and separated byminimal interfibrillar spaces [Fig. 2(C)].

Resin-dentin interfaces in NaOCl-treated dentin ap-peared more variable. Incomplete removal of the de-mineralized collagen network by NaOCl frequentlyresulted in remnant hybrid layers that were 1–3 �m

thick [Fig. 3(A,B)]. These hybrid layers were electron-dense along their entire thickness, sometimes contain-ing wide, electron-lucent spaces that were devoid ofcollagen fibrils (not shown). Collagen fibrils that re-mained were partially denatured and loosely ar-ranged, being separated by wide interfibrillar spaces

Figure 2. TEM micrographs of Single Bond bonded to NaOCl-treated, acid-etched dentin, showing the variation in theextent of collagen removal after treatment with 5% NaOCl for 2 min. (A) Low-magnification view of the resin-dentin interfacein deep dentin. The demineralized collagen network was not completely removed, and a remnant, 3-�m-thick hybrid layercould be discerned. Unlike Figure 1, the entire hybrid layer was electron-dense. This is probably caused by either theinteraction of NaOCl with the demineralized collagen fibrils or the partial dissolution of the remnant collagen network thatresulted in increased wettability and better diffusivity of the polyalkenoic acid copolymer. Bar = 3 �m. (B) A higher-magnification view of a representative resin-dentin interface in superficial dentin. Incomplete removal of the collagennetwork resulted in a 1.5-�m-thick, remnant hybrid layer. Collagen fibrils within the entire hybrid layer were intenselyelectron-dense and were separated by clearly recognizable, electron-lucent interfibrillar spaces. Bar = 1 �m. (C) A PTA andUA-stained section showing the status of the collagen fibrils along the surface of the remnant hybrid layer. No collagenbanding could be recognized and loose, denatured subfibrillar strands were present within the wide interfibrillar spaces(arrowhead). Bar = 300 nm. (D) A high-magnification view of a rare occurrence in deep dentin, in which the collagen networkwas almost completely removed by the NaOCl and the hybrid layer was completely nonexistent. The intact dentin wascovered with a continuous layer of polyalkenoic acid copolymer. A few collagen fibrils could be seen along the periphery ofa dentinal tubule (pointer). Bar = 1 �m. C, lining resin composite; A, adhesive layer; H, hybrid layer; D, laboratory demin-eralized dentin.

320 OSORIO ET AL.

[Fig. 3(C)]. Completely removal of the collagen net-work by NaOCl was only rarely observed [Fig. 3(D)].In the absence of a hybrid layer, a thick, continuouslayer of polyalkenoic acid copolymer could be seen ontop of the intact dentin.

Both groups exhibited similar nanoleakage patterns.There was a tendency for thicker silver deposits tooccur between the hybrid layer and the mineralizeddentin [Fig. 4(A)]. Silver-staining also occurred as elec-tron-dense, dendritic deposits within the polyalkenoicacid copolymer [Fig. 4(B)]. Remnant hybrid layersfrom the NaOCl-treated group that contained large,electron-lucent spaces devoid of collagen fibrils werenot particularly susceptible to silver penetration [Fig.4(C)]. Similar to the acid-etched group, dendritic silverdeposits were localized to the polyalkenoic acid co-polymer phase of the adhesive layer. In contrast, silverdeposits were completely absent from the electron-lucent, globular resin phases that were present withinthe polyalkenoic acid copolymer.

DISCUSSION

An adhesive liquid must completely spread on thesubstrate surface in order to come into close contactwith it; its total spreading allows molecular attrac-tions. There are two ways to promote and enhance theadhesion to dentin: the first is to improve the impreg-

nation of monomer into the substrate, and the secondto increase the diffusivity or the ability to penetrateinto the dentinal substrate.27 Good wettability is thefirst prerequisite for optimal adhesion.28 Wettabilitystudies using the contact angle method may be madedirectly on wet dentin that is similar to clinical condi-tions.

Results of our study showed that etched dentin ex-hibited greater wettability (lower CA) than grounddentin. Phosphoric acid conditioning increases dentinroughness and favors capillary action, so a decrease inCA values is expected.5,20 Conversely, the hydropho-bic components of the dentin surface energy are en-hanced by air-drying of the acid-etched dentin.5 Thismay be caused by the relocation of the hydrophobicgroups of the collagen fibrils toward the surface dur-ing dehydration in order to reach a state of lowestsurface energy. As CA measurements were performedon fully hydrated dentin surfaces, lowering of CA wasobtained. The wetting efficacy is further enhanced byusing an adhesive with a resin component that con-tains multiple acidic carboxylic groups (i.e., the hydro-philic polyalkenoic acid copolymer). This is confirmedby the consistent demonstration of a layer of polyalke-noic acid copolymer layer the surface of the etcheddentin in the TEM micrographs. This phenomenon isespecially significant when collagen fibrils were com-pletely removed by NaOCl treatment [Fig. 2(D)]. Add-ing hydroxyethyl methacrylate (HEMA) significantly

Figure 3. TEM images of nanoleakage in Single Bond-bonded, acid-etched dentin. (A) A stained section taken from deepdentin. Nanoleakage was manifested as the occurrence of electron-dense, silver deposits within the 5-�m-thick hybrid layer.In some areas, heavier silver deposits could be seen along the base of the stained hybrid layer (arrows). Nanoleakage withinthe stained polyalkenoic acid copolymer layer could be better appreciated in an unstained section (B). A fine baseline reticularsilver-staining pattern was present throughout the laboratory demineralized dentin. No nanoleakage could be detectedwithin the adhesive-filled dentinal tubules (T). Bar = 1 �m. (B) An unstained section taken from superficial dentin, showingthe presence of silver deposits within the hybrid layer as well as the polyalkenoic acid copolymer layer. In contrast, theelectron-lucent globular adhesive phases (arrowheads) within the latter were devoid of silver deposits. Bar = 1 �m. A,adhesive layer; P, polyalkenoic acid copolymer; H, hybrid layer; D, laboratory demineralized dentin.

321EFFECT OF NaOCl ON DENTIN BONDING

enhances the hydrophilic characteristics of dentin, be-cause of an increase in the concentration of hydroxylgroups on the dentin surface.

CA values derived from deep dentin were lowerthan those obtained from superficial dentin. This is ingood agreement with our previous studies.5,20 Thisdifference may be caused by the morphological varia-tion in the number of exposed tubules and the relativeareas occupied by peritubular and intertubular den-tin29 in the two bonding substrates.30 NaOCl is a well-known nonspecific proteolytic agent capable of re-moving organic components in dentin.17 After NaOCltreatment of dentin, an increase on wettability is ex-pected, because collagen removal produces a hydro-philic surface.20,31,32 Complete removal of organiccomponents of the demineralized collagen matrix byNaOCl also increases the porosity of the intact den-tin.17,30

Adhesion to dentin may be enhanced either by im-proving the diffusivity of adhesion-promoting mono-mers or rendering the bonding substrate more suscep-tible to resin infiltration.27,33 The latter, for example,may be enhanced by optimizing the dimensions of theinterfibrillar spaces within the demineralized collagennetwork.6 Removal of the collagen matrix with a de-proteinizing agent as an alternative method of enhanc-ing adhesion to dentin has led to equivocal results.Some authors reported that removal of unsupportedcollagen fibrils resulted in a more permeable sub-strate16 that enhances the spreading and diffusion ofadhesive monomers through dentin.34 Although den-tin bond strength has been reported to be positivelyenhanced with the use of a NaOCl solution in somestudies,18,35–37 others reported neutral8,36,38,39 or evennegative results.17–,19,36 In the present study, decreasein shear bond strength was consistently observed both

Figure 4. Stained TEM images of nanoleakage in Single Bond-bonded, NaOCl-treated acid-etched dentin. (A) An area in deepdentin that exhibited a nanoleakage pattern that is similar to thatobserved in Figure 3(A). Silver deposits were found predomi-nantly along the surface (pointer) and basal portion (arrow) ofthe remnant hybrid layer. Unlike Figure 3(A), numerous elec-tron-dense spaces (arrowhead) were found within the hybridlayer. Bar = 1 �m. (B) A higher-magnification of an area similarto (A) and taken from superficial dentin. The large electron-lucent, adhesive resin-filled spaces within the hybrid layer (as-terisk) were devoid of silver deposits. As the specimen was im-pregnated with silver nitrate before laboratory demineralizationand epoxy resin embedding, these spaces were filled with theelectron-lucent phase of the adhesive rather than epoxy resin. Weused a UA- and LC-stained section for illustration, as it was notpossible to distinguish collagen fibrils from the collagen-free,resin-filled spaces in the absence of staining. Silver deposits (ar-rowheads) could be easily discerned from the remnant collagenfibrils. Pointer: demineralization front. Bar = 300 nm. (C) Ahigher magnification of the dendritic silver deposits within thepolyalkenoic acid copolymer phase of the adhesive layer. Similar

to the unstained section in Figure 3(B), no silver deposits could be found within the electron-lucent globular resin phases(pointers) within the polyalkenoic acid copolymer. Bar = 1 �m. A, adhesive layer; P, polyalkenoic acid copolymer; H, hybridlayer; D, laboratory demineralized dentin.

322 OSORIO ET AL.

for superficial and deep dentin. It is possible that theeffect of NaOCl treatment is system specific, beingdependent upon the composition of different adhesivesystems.18,36 The substructure of the collagen fibrilsand their chemical composition have been shown tobe chemically altered by different adhesive systems10

as well as NaOCl treatment.15,40 To make bonding toNaOCl-treated demineralized dentin, it is very impor-tant to get hybridized dentin, by means of adhesivesthat contain acidic monomers with high diffusive po-tential.34,36,37 Single Bond does not have such func-tional monomers. It has polyalkenoic acid, which isacid but its molecular weight is too large and it hasdifficulty to diffuse in mineralized dentin (NaOCl-treated demineralized dentin).

SBS method gives stress along the resin-dentin in-terface, but it is difficult to provide it under the hy-bridized dentin, so bonding defects at this level areidentified by ML, NL, and TEM observations. OurTEM results clearly showed that intact hybrid layersin acid-etched dentin exhibited different staining char-acteristics compared with those remnant hybrid layersin NaOCl-treated dentin [Figs. 1(A) and 2(B)]. Theformer could be explained by the restriction of thepolyalkenoic acid copolymer to the superficial portionof the hybrid layer.42 The high molecular weight ofpolyalkenoic acid copolymer (ca. 14,000–20,000), rela-tive to that of HEMA (MW = 130) suggests that thecopolymer has difficulty penetrating the narrow inter-fibrillar spaces of the demineralized dentin matrix.41

Partial dissolution of the collagen matrix by NaOClcould have reduced the diameter of the collagen fibrils[Fig. 2(C)] or created substantial spaces within the ma-trix that were completely devoid of collagen [Fig 4(B)].This, in theory, could promote better infiltration of thepolyalkenoic acid copolymer. However, this did notexplain our TEM observation that only the collagenfibrils but not the interfibrillar spaces or collagen-depleted zones within the hybrid layer in NaOCl-treated etched dentin appeared electron-dense afterstaining. Clearly, this phenomenon could not be ac-counted for by alterations in wettability or diffusivityalone. Chemical changes of the collagen fibrils afterNaOCl treatment may be a more feasible reason forthe altered ultrastructural appearance of these hybridlayers. NaOCl, apart from being an effective depro-teinizing agent, is also a potent biological oxidant.42

In conclusion, although wettability increased andSBS decreased after the use of NaOCl on acid-etcheddentin, both the quantitative microleakage and quali-tative nanoleakage evaluation showed that the use ofNaOCl did not appear to have any advantage overconventional acid-etching in improving the seal indentin that is not protected by enamel. As a result, thenull hypothesis cannot be rejected.

The authors thank Gertrudis Gomez-Villaescusa for tech-nical assistance.

References

1. Gwinnett AJ. Quantitative contribution of resin infiltration/hybridization to dentin bonding. Am J Dent 1993;6:7–9.

2. Nakabayashi N, Kojima K, Masuhara E. The promotion of ad-hesion by infiltration of monomers into tooth substrates. JBiomed Mater Res 1982;16:265–273.

3. Akinmade AO, Nicholson JW. Glass-ionomer cements as ad-hesives. Part I. Fundamental aspects and their clinical rel-evance. J Mater Sci Mater Med 1993;4:95–101.

4. El Feninat F, Ellis TH, Sacher E, Stangel I. Moisture-dependentrenaturation of collagen in phosphoric acid etched human den-tin. J Biomed Mater Res 1998;42:549–553.

5. Rosales JI, Marshall GW, Marshall S, Watanabe L, Toledano M,Cabrerizo MA, Osorio R. Acid etching and hydration influenceon dentin roughness and wettability. J Dent Res 1999;78(9):1554–1559.

6. Pashley DH, Zhang Y, Agee KA, Rouse CJ, Carvalho RM, Rus-sell CM. Permeability of demineralized dentin to HEMA. DentMater 2000;16:7–14.

7. Pashley DH, Zhang Y, Carvalho RM, Rueggeberg FA, RussellCM. H+-induced tension development in demineralized den-tin matrix. J Dent Res 2000;79:1578–1583.

8. Gwinnett AJ. Altered tissue contribution to interfacial bondstrength with acid conditioned dentin. Am J Dent 1994;7:243–246.

9. Nishiyama, N, Asakura T, Suzuki K, Horie K, Nemoto K. Ef-fects of a structrural change in collagen upon binding to con-ditioned dentin studied by 13C NMR. J Biomed Mater Res 1996;29:107–111.

10. Eliades G, Vougiouklakis G, Palaghias G. Effect of dentin prim-ers on the morphology, molecular composition and collagenconformation of acid-demineralized dentin in situ. Dent Mater1999;15:310–317.

11. Agee K, Zhang Y, Pashley DH. Effects of acids and additiveson the susceptibility of human dentine to denaturation. J OralRehabil 2000;27:136–141.

12. Kiyomura M. Bonding strength to bovine dentin with4-META/MMA-TBB resin. Long term stability and influenceof water. J Jpn Dent Mater 1987;6:860–872.

13. Dorfer CE, Staehle HJ, Wurst MW, Duschner H, Pioch T. Thenanoleakage phenomenon: influence of different dentin bond-ing agents, thermocycling and etching time. Eur J Oral Sci2000;108:346–351.

14. Li H, Burrow MF, Tyas MJ. Nanoleakage of cervical restora-tions of four dentin bonding systems. J Adhes Dent 2000;2:57–65.

15. Di Renzo M, Ellis TH, Sacher E, Stangel I. A photoacousticFTIRS study of the chemical modifications of human dentinsurfaces: II. Deproteination. Biomaterials 2001;22:793–797.

16. Inaba D, Duschner H, Jongebloed W, Odelius H, Takagi O,Arends J. The effects of a sodium hypochlorite treatment ondemineralised root dentin. Eur J Oral Sci 1995;103:368–374.

17. Perdigao J, Lopes M, Geraldeli S, Lopes GC, Garcıa-Godoy F.Effect of sodium hypochlorite gel on dentin bonding. DentMater 2000;16:311–323.

18. Saboia VdePA, Rodrigues AL, Pimenta LAF. Effect of collagenremoval on shear bond strength of two single-bottle adhesivesystems. Oper Dent 2000;25:395–400.

19. Frankenberger R, Kramer N, Oberschachtsiek H, Petschelt A.Dentin bond strength and marginal adaption after NaOCl pre-treatment. Oper Dent 2000;25:40–45.

20. Cheng P, Li D, Boruvka L, Rotenberg Y, Neumann AW. Au-

323EFFECT OF NaOCl ON DENTIN BONDING

tomation for axisymmetric drop shape analysis for measure-ments of interfacial tensions and contact angles. Colloids Surf1990;43:151–167.

21. Toledano M, Osorio R, Perdigao J, Rosales JI, Thompson JY,Cabrerizo MA . Effect of acid etching and collagen removal ondentin wettability and roughness. J Biomed Mater Res1999;46(2):198–203.

22. Toledano M, Osorio O, de Leonardi G, Rosales-Leal JI, CeballosL, Cabrerizo-Vilchez MA. Influence of self-etching primer onthe resin adhesion to enamel and dentin. Am J Dent 2001;14:205–210.

23. Crim GA, Garcıa-Godoy F. Microleakage: effect of cycling andstorage duration. J Prosthet Dent 1987;57:574–576.

24. Toledano M, Perdigao J, Osorio R, Osorio E. Effect of dentindeproteinization on microleakage of class V composite resto-rations. Oper Dent 2000;25(6):497–504.

25. Tay FR, Moulding KM, Pashley DH. Distribution of nanofillersfrom a simplified-step adhesive in acid-conditioned dentin. JAdhes Dent 1999;2:103–2117.

26. Baier RE. Adhesion in the biologic environment. Biomater MedDevices Artif Organs 1984;12:133–159.

27. Gordan VV, Vargas MA, Cobb DS, et al. Evaluation of acidicprimers in microleakage of class 5 composite resin restorations.Oper Dent 1998;23:244–249.

28. Sano H, Yoshiyama M, Ebisu S, Burrow MF, Takatsu T, CiucchiB, Carvalho R, Pashley DH. Comparative SEM and TEM ob-servations of nanoleakage within the hybrid layer. Oper Dent1995;20:160–167.

29. Pashley DH. Dentin: a dynamic substrate-a review. ScanningMicrosc 1989;3:161–176.

30. Perdigao J, Thompson JY, Toledano M, Osorio R. An ultra-morphological characterization of collagen-depleted etcheddentin. Am J Dent 1999;12:250–255.

31. Panighi M, G’Sell C. Influence of calcium concentration on thedentin wettability by an adhesive. J Biomed Mater Res 1992;26:1081–1089.

32. Attal JP, Asmussen E, Degrange M. Effects of surface treatmenton the free energy of dentin. Dent Mat 1994;10:259–264.

33. Nakabayashi N, Sami Y. Bonding to intact dentin. J Dent Res1996;75:1706–1715.

34. Prati C, Chersoni S, Pashley DH. Effect of removal of surfacecollagen fibrils on resin-dentin bonding. Dent Mater 1999;15:323–331.

35. Wakabayashi Y, Kondou Y, Suzuki K, Yatani H, Yamashita A.Effect of dissolution of collagen on adhesion to dentin. Int JProsthodont 1994;7:302–306.

36. Inai N, Kanemura N, Tagami J, Watanabe LG, Marshall SJ,Marshall GW. Adhesion between collagen depleted dentin anddentin adhesive. Am J Dent 1998;11:123–127.

37. Phrukkanon S, Burrow MF, Hartley PG, Tyas MJ. The influ-ence of the modification of etched bovine dentin on bondstrengths. Dent Mater 2000;16:255–265.

38. Armstrong SR, Boyer DB, Keller JC, Park JB. Effect of hybridlayer on fracture toughness of adhesively bonded dentin-resincomposite joint. Oper Dent 1997;22:159–166.

39. Coli P, Alaeddin S, Wennerberg A, Karlsson S. In vitro dentinpretreatment: surface roughness and adhesive shear bondstrength. Eur J Oral Sci 1999;107:400–413.

40. Hawkins CL, Davies MJ. Hypochlorite-induced damage toproteins: formation of nitrogen-centred radicals from lysineresidues and their role in protein fragmentation. Biochem J1998;332(Pt 3):617–625.

41. Van Meerbeek B, Conn LJ, Duke ES, et al. Correlative trans-mission electron microscopy examination of nondemineralizedand demineralized resin-dentin interfaces formed by two den-tin adhesive systems. J Dent Res 1996; 75: 879–888.

42. Daumer KM, Khan AU, Steinbeck MJ. Chlorination of pyri-dinium compounds. Possible role of hypoclorite, n-chloramines, and chlorine in the oxidation of pyridinolinecross-links of articular cartilage collagen type II during acuteinflammation. J Biol Chem 2000;275:34681–34692.

324 OSORIO ET AL.