Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562

avai lab le at www.sc iencedi rec t .com

journa l homepage: www. int l .e lsev ierhea l th .com/ journa ls /dema

Micro-tensile bond strength and interfacial characterizationof an adhesive bonded to dentin prepared by contemporarycaries-excavation techniques

Aline de A. Neves, Eduardo Coutinho, Marcio V. Cardoso, Jan de Munck,Bart Van Meerbeek ∗

Leuven BIOMAT Research Cluster, Department of Conservative Dentistry, School of Dentistry, Oral Pathology and Maxillo-Facial Surgery,Catholic University of Leuven, Kapucijnenvoer 7, B-3000 Leuven, Belgium

a r t i c l e i n f o

Article history:

Received 5 May 2010

Received in revised form

24 October 2010

Accepted 15 February 2011

Keywords:

�TBS

Adhesion

Dentin caries

Caries-excavation

Residual caries

a b s t r a c t

Objectives. To evaluate the micro-tensile bond strength (�TBS) and interfacial characteristics

of adhesive–dentin bonds produced after caries-removal with contemporary techniques.

Methods. Carious molars were cut at the base of the fissure, exposing ‘sound’ and ‘carious’

dentin at different spots. After caries-excavation, a composite was bonded using a 2-step

self-etch adhesive. The �TBS was measured and the mode of fracture analyzed using a

stereomicroscope and imaged by Feg-SEM, while additional non-fractured specimens were

histologically analyzed after Masson’s trichrome staining in order to identify potentially

incompletely resin-enveloped collagen.

Results. �TBS to residual caries-excavated dentin was lower than to sound dentin. The differ-

ent caries-removing techniques had a significant effect on the �TBS. Er:YAG laser guided by a

LIF-feedback system (Kavo) resulted in the lowest �TBS (26.8% lower than to ‘sound’ dentin)

and a distinct layer of incompletely resin-enveloped collagen at the interface. Although dif-

ferent degrees of collagen exposure were seen for other caries-removing techniques, such

as a thick layer for CeraBur (Komet-Brasseler), some unprotected collagen areas for Cariex

(Kavo), or completely resin-enveloped collagen for a tungsten-carbide-bur (Komet), the �TBS

appeared not directly affected (10%, 16.6%, and 15.3% lower than to ‘sound’ dentin, respec-

tively). Carisolv (MediTeam) resulted in the highest �TBS (only 1% reduction compared

to that to ‘sound’ dentin), followed by the tungsten-carbide-bur aided by Caries Detector

(Kuraray) (4.8% reduction). Enzymatic caries excavation using the experimental SFC-VIII

(3M-ESPE) aided by a disposable plastic instrument resulted in a 19.4% reduction in �TBS as

compared to that to ‘sound’ dentin.

Significance. The dentin bonding receptiveness depends to a large extent on the caries-

excavation method employed.

© 2011 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

∗ Corresponding author. Tel.: +32 16 337587; fax: +32 16 332752.E-mail address: bart.vanmeerbeek@med.kuleuven.be (B. Van Meerbeek).

0109-5641/$ – see front matter © 2011 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.dental.2011.02.008

Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562 553

1. Introduction

The modern concept of ‘minimal-invasive dentistry’ calls formore conservative elimination of the highly infected and irre-versibly demineralized carious tissue. Such selective cariesexcavation should prevent lesion progression, while maintain-ing the strength and stability of the remaining tooth structurein order to guarantee long-term mechanical resistance againstintra-oral forces [1]. However, defining the actual endpoint ofcaries excavation and thus the start-point of restoration isoften clinically challenging. Since soft and wet dentin cariouslesions harbor significantly more bacteria than hard and drylesions [2], clinicians are commonly advised to remove cariousdentin to the level where it is ‘firm’ [3].

The most conventional method of removing caries involvesthe use of steel or tungsten-carbide burs mounted in alow-speed contra-angle. Although very efficient in terms oftime spent for caries removal [4], the decision to stop cariesremoval using these burs is very subjective, and basicallydepends on the operator’s background and clinical experience.The recently marketed CeraBur (Komet-Brasseler, Lemgo,Germany) is a self-limiting ceramic bur (alumina-based withstabilized zirconia) [5], which according to the manufacturerefficiently cuts infected, soft dentin, while hardly acts on hard,sound tissue.

Another potentially more conservative caries-removingtechnique involves the use of sono-abrasive technology.Diamond-coated oscillating tips coupled to an airscaler havepreviously been investigated for caries removal. They werehowever reported to mostly leave considerable amountsof residual carious dentin [4]. Sono-abrasive technologyusing tungsten-carbide tips (Cariex TC tips, Kavo, Biberach,Germany) has also recently been marketed, but its effective-ness for caries removal has not been evaluated yet.

Besides rotary and oscillating caries-removal methods,chemical agents to selectively dissolve carious dentin cantoday also be applied. The Carisolv system (MediTeam, Göte-borg, Sweden) makes use of a NaOCl-based gel and wasreported to perform well in laboratory [4] and clinical [6]research. Moreover, the dentin substrate left after caries-removal with Carisolv was found to be very compatible withadhesive procedures [7,8]. Recently, an experimental pepsin-based gel was developed (exp. SFC-VIII, 3M-ESPE, Seefeld,Germany), and consists of a moderately acidic buffered solu-tion of pepsin that possesses the ability to cleave denaturedcollagen fibrils. These fibrils are exposed in the carious lesionafter dissolution of the surrounding hydroxyapatite by organicacids, so that carious dentin is specifically targeted [9]. Pre-liminary results have indicated comparable caries-removingproperties of a previous experimental version of this product(exp. SFC-V, 3M-ESPE) with Carisolv [10].

Caries removal using an Er:YAG laser has been intro-duced as a “pain-free” and more tissue-preserving cavity-preparation technique [11]. To overcome the difficulties inestablishing the caries-removal endpoint with this technique,a laser-induced fluorescence (LIF) method was coupled to thelaser equipment. This feedback system activates the Er:YAGlaser to ablade dentin tissue only when the LIF of the tissueis higher than the chosen threshold [12]. The caries-removal

effectiveness of a feedback-equipped Er:YAG laser and theresulting bonding receptiveness of the remaining dentin sub-strate have not yet fully been investigated.

Except in case traumatic tooth injuries need to be restoredor teeth need to be corrected esthetically, adhesive toothrestoration mostly involves bonding to caries-affected dentin.Today’s adhesives bond effectively to sound dentin throughhybridization, but this bonding mechanism remains vul-nerable in the long term. Incomplete resin-envelopmentexposes collagen to oral fluid attack and enzymatic degrada-tion processes that may eventually lead to caries recurrence[13]. Bonding to caries-affected dentin is even less pre-dictable and durable, not only because wider zones ofunprotected collagen [14], but also more cracks and pores arepresent [15].

The aim of the present study was to determine thebonding effectiveness of a ‘gold-standard’ self-etch adhe-sive to residual caries-excavated dentin, as producedfollowing seven different contemporary caries-removing tech-niques. The hypotheses tested were (1) that the �TBSto ‘residual caries-excavated’ dentin is similar as to‘sound’ dentin, and (2) that different caries-removing tech-niques result in dentin substrates equally receptive tobonding.

2. Materials and methods

2.1. Selection of teeth and caries removal

From a bulk of extracted, non-restored molars stored in 0.5%aqueous chloramine, those teeth presenting caries lesions onthe occlusal surface that presumably involve dentin, wereselected. After removing plaque, calculus and other debriswith an airscaler (Sonicflex 2000 equipped with a scaler tip#5: Kavo, Biberach, Germany), the teeth were mounted forease of manipulation in gypsum, leaving the occlusal surfaceexposed.

Digital radiographs were obtained with the aid of a CCDdetector (Vista Ray CCD System, Dürr Dental AG, Bietigheim-Bissingen, Germany) so that teeth without radiographicallydetectable dentin caries could be excluded. All remainingteeth (n = 35) were then divided in 7 groups according to thedifferent caries-removing techniques tested (Table 1).

The occlusal enamel was removed by cutting each toothcrown through the deepest part of the occlusal fissure withthe aid of a 0.3-mm thick diamond cut-off wheel (Struers,Ballerup, Denmark) mounted in an Isomet low-speed saw(Buehler, Lake Bluff, IL, USA). The carious lesion was thenremoved by one of the methods described in Table 1,using the respective caries-removing endpoints indicated.To prepare a standardized smear layer in the sound dentinarea and to produce sound and caries-affected dentin sub-strates at approximately the same depth, each tooth wasfurther ground with a medium-grit (100 �m) diamond bur(842, Komet-Brasseler, Lemgo, Germany) in a water-cooledhigh-speed turbine, mounted in a Micro-Specimen Former(University of Iowa, Iowa City, IA, USA). Extra care wastaken not to touch the bottom of the excavated carieslesion.

Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

554 d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562

Table 1 – Caries-removing techniques, actual caries-excavation procedures and respective caries-removal endpointsused.

Caries-removaltechnique

Manufacturer Excavation procedure Caries-removal endpoint

Tungsten-carbide round bur Komet-Brasseler, Lemgo,Germany

Low-speed contra-angle,approximate speed∼=1500 rpm/min

Hard cavity floor with ablunt explorer

CeraBur (K1SM) Komet-Brasseler Low-speed contra-angle,approximate speed∼=1500 rpm/min

Self-limiting cutting abilityof the instrument

Cariex TC tips Kavo, Biberach, Germany Airscaler (Soniflex 2003L),oscillations <6.5 kHz with watercooling

Hard cavity floor with ablunt explorer

Tungsten-carbide round buraided by Caries Detector

Komet-Brasseler/KurarayMedical, Okayama, Japan

Low-speed contra-angle,approximate speed∼=1500 rpm/min

Light-pink residual cariousdentin staining

Carisolv MediTeam, Göteborg,Sweden

After dispensing using theauto-mix syringe system, a drop ofthe solution was placed on thecarious lesion. After 30 s, theCarisolv metal mace tips (n.2–5)were used to scrape off the carioustissue

Self-limitingcaries-removing ability ofthe solution

Exp. SFC-VIII + plasticinstrument (star v1.3)

3M-ESPE, Seefeld, Germany After mixing the two solutionsaccording to the manufacturer’sinstruction, a drop of the solutionwas placed on the carious lesion.After 30 s, a prototype plasticinstrument (Star v1.3) was used toscrape off the carious tissue

Self-limitingcaries-removing ability ofthe solution

Er:YAG laser (Kavo KEYLaser III)

Kavo Output settings: 250 mJ/pulse witha pulse-repetition rate of4 pulses/s. Ablation was performedwith a non-contact handpiece2060 under water cooling

Feedback system threshold7

2.2. Bonding procedures and micro-specimenpreparation

Each tooth containing both sound and caries-excavated dentinwas next thoroughly washed with water-spray and gently air-dried. A ‘mild’ two-step self-etch adhesive (Clearfil SE Bond,Kuraray, Osaka, Japan) was applied following the manufac-turer’s instructions. A composite crown was built using FiltekZ100 composite (3M-ESPE, Seefeld, Germany) and the locationof the caries-excavated dentin area was marked at the top ofthe composite build-up. After 24-h storage in water at 36 ◦C,the teeth were cut into 1-mm2 stick-shaped micro-specimenswith the aid of the 0.3 mm diamond cut-off wheel (Struers)mounted in an Accutom-50 cutting machine (Struers).

2.3. Laser-induced fluorescence (LIF) measurements

Each micro-specimen was assigned to either the ‘sound’ or‘residual caries-excavated’ dentin group, based in the firstplace on the previously marked area at the top of the com-posite build-up. LIF measurements at each of the four sidesof each micro-specimen were performed using a DiagnodentPen (2190, Kavo, Biberach, Germany). The micro-specimen wasdefinitively assigned to the ‘residual caries-excavated’ dentingroup, if it presented readings similar or above 10 on at leastone of the 4 sides.

2.4. �TBS testing

For �TBS testing, the micro-specimens were fixed to aBIOMAT jig [16] with the aid of a cyanoacrylate-based glue(Model Repair II Blue, Dentsply-Sankin, Ohtawara, Japan), andstressed at a crosshead speed of 1 mm/min until failure usinga universal testing device (LRX, Lloyd Instruments, Hamp-shire, UK), equipped with a load cell of 100 N. The �TBS wasexpressed in MPa, as derived from dividing the imposed force(N) at the time of fracture by the bond area of the individualspecimen (mm2). The occurrence of failure prior to the actualtesting was included in the calculation of the mean �TBSas 0 MPa, with an explicit note of the number of pre-testingfailures (ptf). The mode of failure was determined at a mag-nification of up to 50× using a stereomicroscope (LM fractureanalysis: Wild M5A, Wild-Heerbrugg, Heerbrugg, Switzerland),and recorded either as ‘interfacial’, ‘cohesive in dentin’, ‘cohe-sive in composite’ or ‘mixed failure’ (including failures withinthe adhesive layer and composite). Five teeth were tested foreach caries-removing technique.

2.5. SEM fracture analysis

Following failure analysis in the stereomicroscope, repre-sentative samples were processed for field-emission-gunscanning electron microscopy using secondary electron detec-

Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562 555

Table 2 – ANOVA results after matching each ‘sound dentin’ to a ‘residual caries excavated’ dentin specimen from thesame tooth to detect significant differences in �TBS for type of dentin (‘sound’ × ‘residual caries-excavated’ dentin) andtype of caries-removing technique.

SS Degrees of freedom MS F p

Type of dentin 3615.3 1 3615.3 28.113 0.0001Interaction 1396.1 6 232.7 1.809 0.0979Error 30734.8 239 128.6Caries-removing technique 2945.1 6 490.8 3.304 0.0038Error 35510.2 239 148.6

tion (Feg-SEM; FEI Nova NanoLab 600, Eindhoven, TheNetherlands) in order to more accurately interpret (or confirm)the previously analyzed fracture mode in each group. For thispurpose, samples that failed according to the most represen-tative failure mode and had a �TBS close to the mean of thatparticular group, were selected. Common procedures for SEM-specimen preparation were employed, as described previously[17].

2.6. Interfacial characterization using a Masson’strichrome staining protocol

For each caries-removing technique, two micro-specimens(belonging to the ‘residual caries-excavated’ dentin group)that were not tested for �TBS, were prepared for histolog-ical analysis using a standard Masson’s trichrome stainingtechnique [18]. The stick-shaped specimens were first individ-ually embedded in epoxy resin (Epofix Kit, Struers, Ballerup,Denmark) and sectioned along the longest axis with the aidof the 0.3 mm thick diamond cut-off wheel mounted in theAccutom-50 cutting machine (Struers). The half-sectionedspecimens were glued with cyanoacrylate (Sekunder-kleber,Renfert, Hilzingen, Germany) to a glass plate and ground toa thickness of 20–50 �m (Exakt AW 110, Exakt Technologies,Oklahoma City, OK, USA). Finally, the samples were stainedaccording to the standard Masson’s trichrome technique [18],and imaged under transmitted light. Using this technique,fully resin-enveloped collagen stained ‘green’, unprotectedcollagen ‘red’, composite material ‘beige’ and adhesive ‘yel-low’ due to a reaction with the fixative.

2.7. Statistical analysis

For the statistical analysis, every caries-excavated dentinspecimen was randomly matched to one of the sounddentin (control) specimens from the same tooth. Doing so,246 matched pairs were formed, and analyzed by two-wayANOVA with repeated measurements and Tukey multiple-comparisons in order to identify significant differences in�TBS among the different caries-removing techniques. A chi-square test was used to compare the incidence of the differentfracture modes among the caries-removing techniques. A sig-nificance level of 5% was employed for all analyses.

3. Results

3.1. �TBS testing and LM fracture analysis

Overall, the bond strength to caries-excavated dentin wasabout 15% lower than that to sound dentin (35 MPa versus41 MPa), and this difference was statistically significant(p < 0.0001, Table 2). The type of caries-removing tech-nique significantly affected the �TBS (p = 0.0038), as not alltechniques resulted in equally bonding-receptive dentin sub-strates (Table 3). No statistical difference in �TBS to ‘sound’dentin was found for the samples belonging to the differentcaries-excavation groups. This could be expected, as all con-trol surfaces were prepared by a similar diamond bur (Table 3).All caries-removing techniques tested showed a lower mean�TBS to ‘residual caries-excavated’ dentin than to ‘sound’dentin, but only for the Er:YAG laser, this reduction was sta-tistically significant (Table 3, p-value column). Except whencompared to Cariex, the Er:YAG laser presented a statistically

Table 3 – Mean �TBS values in MPa (SD) for ‘sound’ and ‘residual caries-excavated’ dentin according to thecaries-removing techniques tested.

Caries-removing technique Sound dentin Tukeya Residualcaries-excavateddentin

Tukeya p-Valueb % �TBS from�TBSsound dentin

N

Carisolv 41.7 (11.7) a 41.3 (13.9) a 0.9999 99% 31Caries Detector 39.3 (13.0) a 37.4 (13.2) ab 0.9999 95.2% 31Exp. SFC-VIII + instrument 46 (11.5) a 37.1 (13.7) ab 0.1632 80.6% 28CeraBur 40.1 (12.9) a 36.1 (12.3) bc 0.9146 90% 47Tungsten-carbide bur 39.8 (11.2) a 33.7 (9.2) bc 0.7427 84.7% 29Cariex system 38 (9.5) a 31.7 (10.1) cd 0.3858 83.4% 41Er:YAG laserc 40.7 (12.1) a 29.8 (9.9) d 0.0017 73.2% 39

a Means with the same letter are not significantly different (Tukey multiple comparisons, p < 0.05).b If the p-value is <0.05, the �TBS to ‘residual caries-excavated’ dentin is significantly lower than to ‘sound’ dentin.c This caries-removing technique revealed one specimen that failed prior to testing (= pre-testing failure).

Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

556 d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562

Fig. 1 – Mean �TBS values for the different caries-removingtechniques. Values of all control surfaces were merged, asthey were prepared similarly (with a diamond bur) and nostatistical significant differences were found among thecaries-removing groups. Whiskers denote 95% confidenceintervals. Means connected with a horizontal bar are notsignificantly different.

significantly lower �TBS to ‘residual caries-excavated’ dentinthan the other caries-removal techniques (Table 3, Tukey col-umn; Fig. 1, horizontal bars).

Carisolv-excavated dentin showed the lowest reductionin �TBS when bonded to ‘residual caries-excavated’ dentin,as compared to that of ‘sound’ dentin (1%), followed bytungsten-carbide bur excavation aided by Caries Detector(4.8%), and CeraBur (10%). Tungsten-carbide-bur and Cariexexcavation resulted in a similar bond strength reduction whenbonded to ‘residual caries-excavated’ dentin (15.3% and 16.3%,respectively). The experimental enzymatic caries-excavationmethod SFC-VIII, aided by a plastic instrument, resulted in a19.4% reduction in �TBS, while the Er:YAG laser guided by aLIF-feedback threshold of 7, resulted in the highest and theonly statistically significant reduction in bond strength whenbonded to ‘residual caries-excavated’ dentin (26.8%).

Fig. 2 shows the distribution of fracture type (‘interfacial’,‘cohesive in composite’, ‘cohesive in dentin’ or ‘mixed’) forthe different caries-removing techniques, when bonded eitherto ‘sound’ dentin (Fig. 2A) or to ‘residual caries-excavated’dentin (Fig. 2B). Statistical analysis (Chi Square with Yates’scorrection, p < 0.05) revealed an increase in the percentage ofcohesive failures in dentin when bonded to ‘residual caries-excavated’ dentin versus ‘sound’ dentin (Table 4).

3.2. Feg-SEM fracture analysis

Fig. 3 shows SEM-photomicrographs of ‘residual caries-excavated’ dentin specimens after �TBS testing for eachcaries-removing technique. The fractured Carisolv specimen,shown in Fig. 3A, presents an interfacial fracture at the bot-tom of the hybrid layer. At higher magnification, the dentinaltubules are open and intertubular dentin presented only mildsigns of caries-affected dentin, such as some exposed col-lagen fibrils (arrows, Fig. 3A). For the tungsten-carbide-burexcavation aided by Caries Detector (Fig. 3B), a ‘mixed’ frac-ture pattern was typically seen. At higher magnifications,fractured dentin tubules were occluded by resin tags. Somedemineralized and incompletely resin-enveloped collagen fib-

Fig. 2 – Distribution of type of fractures as seen using thestereomicroscope after �TBS testing for ‘sound’ (A) or‘residual caries-excavated’ dentin (B). A statisticallysignificant increase in cohesive dentin fractures for theresidual caries-excavated dentin was observed (Chi square,p < 0.05).

rils were also evident at intertubular dentin (arrows, Fig. 3B).For the exp. SFC-VIII excavation aided by the plastic instru-ment, also a ‘mixed’ failure pattern was often recorded, fromwhich at higher magnification the cohesively dentin-fracturedarea (Fig. 3C, square) does not show resin tags. On the con-trary, the dentin tubules appear occluded by some kind of

Table 4 – Distribution of fracture types for the ‘sound’and ‘residual caries-excavated’ �TBS specimens (valuesbetween parentheses are the expected Chi squarefrequencies).

Failure mode Sounddentin

Residualcaries-excavateddentin

Total

Interfacial 244 (232) 109 (120) 353Mixed 253 (258) 140 (134) 393Cohesive in composite 7 (7) 4 (3) 11Cohesive in dentin 4* (9) 10* (4) 14Total 508 263 771

∗ Chi Square test with Yate’s correction was statistically significant(p < 0.05).

Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562 557

Table 5 – Mean LIF measurements (SD) for ‘sound’ and‘residual caries-excavated’ dentin according to thecaries-removing techniques tested.

Caries-removingtechnique

Sounddentin

Residualcaries-excavateddentin

Tukeya

Carisolv 4.6 (1.0) 21.0 (16.1) aExp.SFC-VIII + instrument

4.9 (1.4) 14.6 (8.8) ab

Caries Detector 4.2 (1.6) 8.9 (5.0) bCeraBur 4.5 (1.2) 11.5 (8.3) bTungsten-carbide bur 3.4 (1.1) 12.6 (8.3) bCariex system 3.2 (0.9) 13.4 (10.7) bEr:YAG laser 4.3 (1.1) 12.4 (7.1) bTotal 4.2 (1.4) 15.3 (9.2)

a Only ‘residual caries-excavated dentin’ specimens were consid-ered for statistical analysis. Means with the same letter are notsignificantly different (Tukey multiple comparisons, p < 0.05).

intra-tubular deposit, which is typical of caries-affected dentin(arrows, Fig. 3C).

CeraBur-excavated specimens revealed a relatively highnumber of cohesive fractures in dentin, which may be indica-tive of the presence of residual caries. CeraBur-excavatedspecimens typically presented demineralized dentin and opendentinal tubules at the fracture planes (Fig. 3D, arrows). Fortungsten-carbide-bur excavation, fractured specimens typ-ically showed a ‘mixed’ fracture pattern, including areasthat fractured in dentin, composite and within the adhesive.Higher magnification revealed that the dentin tubules wereoften occluded with smear plugs, as could be expected fortungsten-carbide-bur prepared ‘sound’ dentin to which a mildself-etch adhesive was bonded.

Particular structural features common to specimens wheresome residual carious dentin was left, were also remarkablypresent in the fractured surfaces of the Cariex group. The spec-imens often failed at a zone that exhibited dentin fracturesbelow a poorly infiltrated hybrid layer (arrows, Fig. 3F), therebyexposing highly demineralized and thus resin-deficient col-lagen at intertubular dentin (Fig. 3F). Samples representingEr:YAG-laser excavation revealed a ‘mixed’ failure pattern,often also including areas of cohesive failure in dentin (Fig. 3G,white arrows) and at higher magnification tubules occludedby smear plugs. However, collagen fibrils generally appearedexposed and not well enveloped by resin at intertubular dentin(Fig. 3G, black arrows).

Table 5 mentions the mean LIF values for the ‘sound’ and‘residual caries-excavated’ dentin specimens according to thedifferent caries-removing techniques. For the ‘residual caries-excavation’ dentin specimens, Carisolv showed statisticallysignificantly higher LIF values than all other caries-removingtechniques, except when comparing to SFC-VIII.

3.3. Interfacial characterization using a Masson’strichrome staining protocol

All specimens selected for histological analysis after Mason’strichrome staining clearly included a caries-affected dentinlayer at the interface, as was confirmed by a LIF value

higher than 15 at one of the four sides of the speci-mens.

Analysis of the light-microscopy images revealed that thesection thickness significantly affected the intensity of thestaining pattern. As the section thickness in our study var-ied from 17 to 50 �m, correct interpretation was only possibleby directly comparing the ‘residual caries-excavated’ dentinarea (left-sided images, Fig. 4) to the ‘sound’ dentin (con-trol) area (right-sided images, Fig. 4) from the same section.The different caries-excavation methods resulted in differ-ent staining patterns and thus residual caries conditions. Forexample, caries removal with CeraBur resulted in a thickred-stained area, representing unprotected collagen at theadhesive–dentin interface (Fig. 4D). Also Er:YAG-laser excava-tion produced a thin, but homogeneous layer of unprotectedcollagen that is clearly visible just below the interface (Fig. 4G,arrows). Caries removal with a tungsten-carbide bur, on theother hand, resulted in practically no unprotected collagen atthe interface (Fig. 4E). In samples where a tungsten-carbide-bur was used with Caries Detector or where the sono-abrasionCariex system was employed, a low degree of unprotectedcollagen was observed (Fig. 4D and F). Chemo-mechanicalexcavation resulted for both the Carisolv and the experimentalSFC-VIII system in some areas of unprotected collagen, mixedwith areas of fully protected collagen near the interface (Fig. 4Aand C).

4. Discussion

According to the results obtained in the present study, bothhypotheses tested were rejected. Hence, the bonding effective-ness to residual carious dentin was in general lower than to‘sound’ dentin and the different caries-removing techniquesresulted in different bonding-receptive dentin substrates.

Bond-strength testing to dentin always results in consider-able data variance [16] and as a consequence, bond-strengthtesting to a structurally varying substrate, such as ‘residualcaries-excavated’ dentin, must lead to even more spreadingof the data. In the present study, we intended to reduce thisvariability by preparing specimens with areas exhibiting both‘residual caries-excavated’ and ‘sound’ (control) dentin at thesame tooth. Consequently, a paired statistical analysis couldbe applied to eliminate the ‘tooth’ effect when different teethwould have been used to measure the bond strength to bothsubstrates. To better separate the ‘residual caries-excavated’dentin specimens from the ‘sound’ dentin control specimens,LIF values were used after an initial location-based groupassignment. This method was very selective in identifying thespecimens without caries (Table 5), as it also was previouslyused for differentiating ‘sound’ from ‘caries-affected’ micro-specimens. A good agreement was indeed found with themore conventional hardness measurements commonly usedfor this purpose [19].

As the main focus of this study was to measure the bondingeffectiveness to ‘residual caries-excavated’ dentin producedafter different caries-removing techniques, we opted to usea single “gold-standard” adhesive. Clearfil SE Bond is one ofthe most researched adhesives and exhibits good in vitro andclinical bonding effectiveness to dentin [20]. Noteworthy is

Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

558 d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562

Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562 559

the combination of a micro-mechanical and chemical bond-ing capacity, enhancing its short- and long-term performance[21,22].

One of the first studies reporting �TBS to residual cariousdentin, used Caries Detector to establish the caries-excavationendpoint. Using a Clearfil SE Bond precursor (Clearfil LinerBond II, Kuraray), which contained the phenyl-P functionalmonomer, a 50% reduction in bond strength was measuredwhen this mild self-etch adhesive was bonded to residual car-ious dentin, as compared to sound dentin [23]. Recent studiesreporting on bonding effectiveness to residual carious dentin,in which Clearfil SE Bond was used, revealed that the effecton bond strength varied from significantly lower [24], to nodifference [25], or even a significantly higher bond strengthto caries-affected dentin [26]. In the present study, the overallbond strength to ‘residual caries-excavated’ dentin was 15%lower than that to ‘sound dentin’. This reduction was mostevident for the Er:YAG laser excavation, which revealed a 27%decrease (Table 3, p = 0.0017), while other techniques, such asCarisolv, disclosed practically no reduction in �TBS to ‘residualcaries-excavated’ dentin (1%).

According to the Masson’s trichrome staining technique,the ‘red’ staining very likely represents ‘unprotected’ dentincollagen (Type-I collagen constitutes 90% of the organic matrixof dentin) [27]. In case of caries-affected dentin, the red stain-ing may however also disclose the presence of residual caries[14]. Although previous studies have reported that sectionsthrough adhesive–dentin interfaces should be optimally 5 �mthick for staining with Masson’s trichrome [18], the tech-nique employed in this study did not allow thin sections tobe prepared. Demineralized, paraffin-imbedded samples aremaybe more suited for that purpose. The minimum thick-ness obtained was 17 �m (Fig. 4F: Cariex specimen), while themaximum thickness that still allowed qualitative interpreta-tion, was 50 �m (Fig. 4B: Caries Detector specimen). Also, fortungsten-carbide-bur excavation aided by Caries Detector, thered staining produced by the caries-staining dye itself mighthave interfered with the histological staining, and thus havepotentially overestimated the presence of residual caries.

Comparing the Masson’s staining data with the�TBS results, it appears that leaving some unpro-tected/demineralized collagen at the interface does notper se affect the immediate bond strength to dentin (as

for example for CeraBur: Fig. 4B and Table 3). However,it is not known if the long-term bond durability may benegatively affected, as the quality of a hybrid layer in caries-affected dentin is very questionable [28], especially whena thick collagen-rich zone, resulting from incomplete resininfiltration, was clearly seen. The presence of this thicker,collagen-rich and porous zone at the adhesive interface maynot only reduce the�TBS to caries-affected dentin, but it mayalso accelerate the bond-degradation process [18].

In comparison with a previous version of a self-limitingrotary cutting instrument, made of plastic [29], CeraBursrevealed in the present study a higher bond strength to ‘resid-ual caries-excavated’ dentin. This must most likely be ascribedto the apparently enhanced caries-removing ability of thisceramic bur, and this thanks to its better mechanical prop-erties. However, the largest area of ‘unprotected’ collagen wasobserved when caries was excavated with a CeraBur. This zonemay represent residual caries that may have remained despitethe self-limiting properties of the bur. The substrate condi-tion was not checked on its hardness, by which it should beconcluded that a CeraBur may have left some carious tissuebehind in some samples, at least when used without any othercaries-removal endpoint indication.

Regarding the tungsten-carbide tips of the sono-abrasionCariex system, the �TBS results (Table 3) and fractography(Fig. 3D) suggest that residual caries was left in some speci-mens as well. One has previously reported that the oscillationfrom a sono-abrasive diamond tip could be transformed intovibration, and result in a compacting effect on the cariousdentin tissue rather than removing it. This would lead to anoverall apparent increase in hardness of the cavity surface,giving the operator a false indication that the caries-removalendpoint was reached [4].

In this study, no attempt was made to quantify the pres-ence of residual caries in terms of LIF values. The literature,confirmed by recent data, revealed that absolute LIF valuescan be misleading since surface staining of residual dentincan also result in high LIF values [30].

In a previous study using Clearfil SE Bond on Carisolv-excavated carious dentin, only a slightly lower �TBS wasmeasured when compared to sound dentin [7]. In this study,even no difference in bond strength was found (Table 3),thereby corroborating another study [31]. In fact, surfaces

Fig. 3 – Feg-SEM images of fractured �TBS specimens for the different caries-removing techniques. Squared areas aredepicted in higher magnification on the immediate right-sided image. (A) Dentin part after Carisolv excavation, showinginterfacial fracture at the bottom of the hybrid layer. Mild signs of caries-affected dentin are seen as some exposed(unprotected) collagen fibrils (arrows); (B) composite part after tungsten-carbide-bur excavation aided by Caries Detector,showing a ‘mixed’ fracture pattern with dentin fractured at the level of the hybrid layer, revealing typical resin-tagformation and some demineralized dentin with unprotected collagen (arrows); (C) dentin part after exp. SFC-VIIIcaries-excavation aided by a plastic instrument showing a ‘mixed’ fracture pattern with dentin that fractured at the base ofthe hybrid layer and does not show resin tags but some intra-tubular mineral deposition; (D) dentin part after CeraBurexcavation with typical signs of residual caries presence detected in the form of areas that failed cohesively in dentin(arrows); (E) composite part after tungsten-carbide-bur excavation, showing a ‘mixed’ fracture pattern in the hybrid layer,while the tubules are occluded by smear plugs; (F) dentin part after Cariex excavation, disclosing the presence of residualcaries in the form of areas that failed cohesively in dentin (arrows) and exhibited demineralized intertubular unprotectedcollagen dentin; (G) dentin part after Er:YAG laser excavation showing some degree of dentin hybridization as seen by thepresence of smear plugs but a general appearance of an acid-resistant intertubular dentin. D, dentin; Ar, adhesive; C,composite; P, peritubular dentin; Id, intertubular dentin; Cf, ‘unprotected’ collagen fibers; Rt, resin tags; Sp, smear plugs.

Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

560 d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562

Fig. 4 – Representative transmitted light microscopyphotomicrographs of adhesive interfaces of the mildself-etch adhesive bonded to residual caries-excavateddentin after Masson’s trichrome staining. In the leftcolumn, representative photomicrographs of theadhesive–dentin interface are shown, while in the rightcolumn, the corresponding sound dentin area is presented.(A) Interface after Carisolv excavation showing unprotectedcollagen areas mixed within “sound” collagen; (B) Interfaceafter tungsten-carbide bur aided by Caries-Detector,

remaining after Carisolv excavation and bonding with a mildself-etch adhesive seem to be very compatible, allowing deeppenetration of the adhesive and resulting in a thick and homo-geneous hybrid layer [32]. Furthermore, a substantial increasein cohesive dentin fractures was observed, fully in accordancewith previous studies [7,31]. NaOCl may have reduced themechanical properties of dentin, causing the specimens to failmore cohesively in dentin.

Regarding the experimental enzymatic caries-excavationmethod SFC-VIII, the ability to specifically target the irre-versibly denatured collagen in the carious lesion could involvean interesting benefit in contrast to unspecific agents likesodium-hypochlorite. This would render them truly “self-limiting” [9] and avoid some issues regarding the ratheraggressive impact of the chemo-mechanical solution on sounddentin. However, the caries-removal efficiency of this systemis doubtful if it is used with the disposable plastic excavator. Aprevious version of this gel (SFC-V) was also used in conjunc-tion with the plastic excavator, but was not able to removeall demineralized intertubular collagen fibrils from artificialcaries lesions [9]. This fact was corroborated in our study, aslight-microscopy disclosed substantially more exposed col-lagen at the interface when compared to the control sounddentin area (Fig. 4C).

Regarding bond strength, the experimental enzymaticcaries-excavation method SFC-VIII, when used in conjunctionwith the plastic instrument, resulted in a moderate reductionin bond strength (19.4%), as compared to when it was bondedto sound dentin. The SEM-assisted fractography neverthelessrevealed that the dentin tubules were filled with intra-tubularmineral deposits, typical of caries-affected dentin. This proba-bly indicates that the experimental caries-excavation methodSFC-VIII hardly opened tubules to promote resin-tag for-mation (Fig. 3C, arrows). When compared to the promisingbond-strength results obtained by another experimental ver-sion of SFC (SFC-V) and when caries was removed with themace-tip Carisolv instruments [33], it may be that beyondslight differences in the chemical composition between thetwo experimental versions, the plastic instrument itself couldhave impaired the caries-removal ability of SFC-VIII.

In a previous study, less favorable bond-strength resultswere obtained when a self-etch adhesive was bonded todentin, from which caries was excavated with a round steelbur than with an Er:YAG laser [34]. This is in disagreement with

exhibiting some degree of red staining. This could howeveralso result from staining by Caries Detector itself; (C)interface after exp. SFC-VIII caries-excavation aided by aplastic instrument, showing also unprotected collagenareas mixed within “sound” collagen; (D) interface afterCeraBur caries-excavation, exhibiting a thick area ofunprotected collagen; (E) Interface after tungsten-carbidebur excavation, not showing any unprotected collagen atthe interface; (F) interface after caries removal with thetungsten-carbide sono-abrasion Cariex tips, showing someunprotected collagen at the interface; (G) interface afterEr:YAG laser excavation guided a LIF feedback, showing athin, but homogeneous layer of unprotected collagen layerjust below the interface (arrows). C, composite; A, adhesive;D, dentin.

Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562 561

the present results, where the Er:YAG laser resulted in the low-est �TBS and the largest decrease in bond strength to ‘residualcaries-excavated’ dentin (p = 0.0017). This difference might berelated to the caries-excavation endpoint applied, which wasestablished by a tactile method in the first study, and by theLIF-feedback system (with a threshold of 7) in this study. TheLIF-feedback system might have lead to under-preparation, asthe light-microscopy images consistently presented a homo-geneous thin layer of exposed collagen. Moreover, also theeffect of the laser itself on the dentin tissue should be consid-ered. Previously, a bond-strength reduction of more than 50%was observed when intact dentin was prepared with a laserand treated with Clearfil SE Bond [35]. In light of this observa-tion, the results obtained with the Er:YAG laser in this studyare even more favorable, by which one could conclude thata slight under-preparation by the Er:YAG laser might even bebetter than an over-prepared cavity. In the previous study [35],most Er:YAG-abladed specimens presented a cohesive failurein dentin right below the interface, while in the present studymore ‘mixed’ and true ‘interfacial’ failures were observed. Thelower incidence of (sub-)surface damage was corroborated bycareful SEM observation of the fractured surface. This lessdetrimental effect of the laser observed in this study, might,apart from the machine settings, also be related to the higherwater content and the lower mineral content of the carioussubstrate, as compared to sound dentin. Because of the higherwater content, the energy is more adsorbed at the surface,while the lower mineral content makes the substrate less brit-tle. The combination of both may have prevented subsurfacecracking of dentin, typically observed in laser-ablated sounddentin, and thus have improved the bonding effectiveness.

5. Conclusion

Bond strength to ‘residual caries-excavated’ dentin waslower than to ‘sound’ dentin. The various caries-removingtechniques had different effects on the �TBS. Er:YAG-laserexcavation guided by a LIF-feedback system resulted ina statistically significantly lower �TBS as compared tosound dentin and produced a thick layer of demineral-ized/unprotected collagen at the interface with the adhesive.Although different degrees of unprotected collagen wereseen for other techniques (CeraBur, Cariex and tungsten-carbide-bur), the bonding effectiveness seemed not to bedirectly affected by the caries-excavation method used. Cari-solv resulted in the highest �TBS to ‘residual caries-excavated’dentin, followed by the use of a tungsten-carbide-bur inconjunction with Caries Detector. The experimental enzyme-based caries-excavation agent (SFC-VIII), when used inconjunction with a plastic instrument, resulted in only a mod-erate �TBS to ‘residual caries-excavated’ dentin as comparedto ‘sound’ dentin.

r e f e r e n c e s

[1] Tyas MJ, Anusavice KJ, Frencken JE, Mount GJ. Minimalintervention dentistry – a review – FDI Commission Project1-97. International Dental Journal 2000;50:1–12.

[2] Kidd EAM, Joystonbechal S, Beighton D. Microbiologicalvalidation of assessments of caries activity during cavitypreparation. Caries Research 1993;27:402–8.

[3] Kidd EAM. How ‘clean’ must a cavity be before restoration?Caries Research 2004;38:305–13.

[4] Banerjee A, Kidd EAM, Watson TF. In vitro evaluation of fivealternative methods of carious dentine excavation. CariesResearch 2000;34:144–50.

[5] Dammaschke T, Vesnic A, Schäfer E. In vitro comparison ofceramic burs and conventional tungsten carbide bud burs indentin caries excavation. Quintessence International2008;39:495–9.

[6] Ericson D, Zimmerman M, Raber H, Gotrick B, Bornstein R,Thorell J. Clinical evaluation of efficacy and safety of a newmethod for chemo-mechanical removal of caries – amulti-centre study. Caries Research 1999;33:171–7.

[7] Burrow MF, Bokas J, Tanumiharja M, Tyas MJ. Microtensilebond strengths to caries-affected dentine treated withCarisolv. Australian Dental Journal 2003;48:110–4.

[8] Haak R, Wicht MJ, Noack MJ. Does chemomechanical cariesremoval affect dentine adhesion? European Journal of OralSciences 2000;108:449–55.

[9] Ahmed AA, Garcia-Godoy F, Kunzelmann KH. Self-limitingcaries therapy with proteolytic agents. American Journal ofDentistry 2008;21:303–12.

[10] Clementino-Luedemann TNR, Ilie ADN, Hickel R,Kunzelmann KH. Micro-computed tomographic evaluationof a new enzyme solution for caries removal in deciduousteeth. Dental Materials Journal 2006;25:675–83.

[11] Keller U, Hibst R. Effects of Er:YAG laser in caries treatment:a clinical pilot study. Lasers in Surgery and Medicine1997;20:32–8.

[12] Eberhard J, Eisenbeiss AK, Braun A, Hedderich J, Jepsen S.Evaluation of selective caries removal by a fluorescencefeedback-controlled Er:YAG laser in vitro. Caries Research2005;39:496–504.

[13] Spencer P, Swafford JR. Unprotected protein at thedentin–adhesive interface. Quintessence International1999;30:501–7.

[14] Haj-Ali R, Walker M, Williams K, Wang Y, Spencer P.Histomorphologic characterization of noncarious andcaries-affected dentin/adhesive interfaces. Journal ofProsthodontics 2006;15:82–8.

[15] Hsu KW, Marshall SJ, Pinzon LM, Watanabe L, Saiz E,Marshall GW. SEM evaluation of resin-carious dentininterfaces formed by two dentin adhesive systems. DentalMaterials 2008;24:880–7.

[16] Poitevin A, De Munck J, Van Landuyt K, Coutinho E, PeumansM, Lambrechts P, et al. Influence of three specimen fixationmodes on the micro-tensile bond strength of adhesives todentin. Dental Materials Journal 2007;26:694–9.

[17] Perdigao J, Lambrechts P, Van Meerbeek B, Vanherle G, LopesAL. Field emission SEM comparison of four postfixationdrying techniques for human dentin. Journal of BiomedicalMaterials Research 1995;29:1111–20.

[18] Erhardt MCG, Toledano M, Osorio R, Pimenta LA.Histomorphologic characterization and bond strengthevaluation of caries-affected dentin/resin interfaces: effectsof long-term water exposure. Dental Materials2008;24:786–98.

[19] Ceballos L, Camejo DG, Victoria Fuentes M, Osorio R,Toledano M, Carvalho RM, et al. Microtensile bond strengthof total-etch and self-etching adhesives to caries-affecteddentine. Journal of Dentistry 2003;31:469–77.

[20] Scherrer SS, Cesar PF, Swain MV. Direct comparison of thebond strength results of the different test methods: acritical literature review. Dental Materials 2010;26:e78–93.

Author's personal copyJournal Identification = DENTAL Article Identification = 1797 Date: April 22, 2011 Time: 5:18 pm

562 d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 552–562

[21] Inoue G, Tsuchiya S, Nikaido T, Foxton RM, Tagami J.Morphological and mechanical characterization of theacid–base resistant zone at the adhesive–dentin interface ofintact and caries-affected dentin. Operative Dentistry2006;31:466–72.

[22] Van Meerbeek B, Peumans M, Poitevin A, Mine A, Van EndeA, Neves A, et al. Relationship between bond-strength testsand clinical outcomes. Dental Materials 2009;26:E100–21.

[23] Nakajima M, Sano H, Burrow MF, Tagami J, Yoshiyama M,Ebisu S, et al. Tensile bond strength and SEM evaluation ofcaries-affected dentin using dentin adhesives. Journal ofDental Research 1995;74:1679–88.

[24] Xuan W, Hou BX, Lu YL. Bond strength of different adhesivesto normal and caries-affected dentins. Chinese MedicalJournal 2010;123:332–6.

[25] Sengun A, Unlu N, Ozer F, OztUrk B. Bond strength of fivecurrent adhesives to caries-affected dentin. Journal of OralRehabilitation 2002;29:777–81.

[26] Scholtanus H, Purwanta K, Dogan N, Kleverlaan CJ, FeilzerAJ. Microtensile bond strength of three simplified adhesivesystems to caries-affected dentin. Journal of AdhesiveDentistry 2010, doi:10.3290/j.jad.a17545.

[27] Monticelli F, Osorio R, Mazzitelli C, Ferrari M, Toledano M.Limited decalcification/diffusion of self-adhesive cementsinto dentin. Journal of Dental Research 2008;87:974–9.

[28] Nakajima M, Kitasako Y, Okuda M, Foxton RM, Tagami J.Elemental distributions and microtensile bond strength ofthe adhesive interface to normal and caries-affected dentin.Journal of Biomedical Materials Research B: AppliedBiomaterials 2005;72:268–75.

[29] Silva N, Carvalho RM, Pegoraro LF, Tay FR, Thompson VP.Evaluation of a self-limiting concept in dentinal cariesremoval. Journal of Dental Research 2006;85:282–6.

[30] Neves AA, Coutinho E, Vivan-Cardoso M, Van Meerbeek B.Residual caries determination by diagnodent after differentcaries-excavation methods. Journal of Dental Research2010;89, abstr. 2038.

[31] Sonoda H, Banerjee A, Sheriff M, Tagami J, Watson TF. Anin vitro investigation of microtensile bond strengths of twodentine adhesives to caries-affected dentine. Journal ofDentistry 2005;33:335–42.

[32] Cehreli ZC, Yazici AR, Akca T, Ozgunaltay G. Amorphological and micro-tensile bond strength evaluationof a single-bottle adhesive to caries-affected human dentineafter four different caries removal techniques. Journal ofDentistry 2003;31:429–35.

[33] Banerjee A, Kellow S, Mannocci F, Cook RJ, Watson TF. Anin vitro evaluation of microtensile bond strengths of twoadhesive bonding agents to residual dentine after cariesremoval using three excavation techniques. Journal ofDentistry 2010;38:480–9.

[34] Sattabanasuk V, Burrow MF, Shimada Y, Tagami J. Resinadhesion to caries-affected dentine after different removalmethods. Australian Dental Journal 2006;51:162–9.

[35] De Munck J, Van Meerbeek B, Yudhira R, Lambrechts P,Vanherle G. Micro-tensile bond strength of two adhesives toErbium:YAG-lased vs. bur-cut enamel and dentin. EuropeanJournal of Oral Sciences 2002;110:322–9.