j o u r n a l o f d e n t i s t r y 3 5 ( 2 0 0 7 ) 2 6 8 – 2 7 4

The accuracy of electrical method for microleakageevaluation by a three-dimensional analysis

Yukiteru Iwami *, Mikako Hayashi, Fumio Takeshige, Shigeyuki Ebisu

Department of Restorative Dentistry and Endodontology, Osaka University Graduate School of Dentistry, 1-8 Yamadaoka,

Suita, Osaka 565-0871, Japan

a r t i c l e i n f o

Article history:

Received 13 June 2006

Received in revised form

25 August 2006

Accepted 6 September 2006

Keywords:

Electrical method

3D

Analysis

Microleakage

Coronal restoration

a b s t r a c t

Objectives: This in vitro study aimed to investigate the accuracy of an electrical method for

the evaluation of microleakage by a three-dimensional analysis of dye penetration.

Methods: Coronal cavities were prepared on buccal, palatal or lingual surfaces in extracted

human molars. The cavities were then filled with resin composites and were subjected to

10,000 load cycles (425 g). Before cavity preparation and after load cycling, physiological

saline was applied and wiped off, and the change in conductance was measured across the

margin of the restoration in each specimen. After dye penetration, the specimens were

reduced by 100 mm in a direction parallel to the cavity floor, from the surface of the

restoration to the cavity floor. The sequence of reducing the sections by 100 mm and image

taking was repeated. Three-dimensional images of dye penetration were made and the

proportions of the interface showing penetration were calculated.

Results: Pearson’s correlation coefficients between changes in conductance and the surface

area of dye penetration, between these and the rate of dye penetration, and between these

and the depth of dye penetration were 0.932, 0.920 and 0.732, respectively. The correlations

were significant ( p < 0.05).

Conclusions: The results of this electrical method for microleakage evaluation showed

stronger correlations with the three-dimensional amount of marginal leakage (surface area

of dye penetration and rate of dye penetration) than the two-dimensional amount (depth of

dye penetration).

# 2006 Elsevier Ltd. All rights reserved.

avai lable at www.sc iencedi rec t .com

journal homepage: www. int l .e lsev ierhea l th .com/ journals / jden

1. Introduction

Microleakage of cariogenic bacteria along the cavity wall of

restorations is reported to be an important cause of pulpal

inflammation, pulp necrosis1–3 and secondary caries4,5 in

restored teeth. The symptoms of microleakage include

postoperative sensitivity, chronic sensitivity and marginal

discoloration.6,7 However, objective judgement of whether

the cause of these symptoms is microleakage is difficult in

the clinical settings. Previously, in vitro electrical methods for

* Corresponding author. Tel.: +81 6 6879 2927; fax: +81 6 6879 2929.E-mail address: y-iwami@dent.osaka-u.ac.jp (Y. Iwami).

0300-5712/$ – see front matter # 2006 Elsevier Ltd. All rights reservedoi:10.1016/j.jdent.2006.09.003

the evaluation of microleakage were reported and their

clinical use in the future was suggested.8–12 However, these

electrical methods evaluated the margins of restorations

after application of an electrolyte that was not wiped off, so

measurements of conductance or impedance depended on

the area of applied electrolyte.13,14 In addition, these

methods cannot be applied to restorations on dentin or

cementum surfaces, because the influence of the applied

area or amount of electrolyte leads to increased measure-

ment errors.9

d.

j o u r n a l o f d e n t i s t r y 3 5 ( 2 0 0 7 ) 2 6 8 – 2 7 4 269

Fig. 1 – Schema of load cycling. The coronal enamel was

reduced with a standardized cavity preparation device. In

load cycling, the plunger was moved repeatedly down

5 mm and the molar was moved repeatedly in a buccal

lingual direction 1.5 mm. The load was 425 g (plumb:

400 g; plunger only: 25 g) and the cycle was 60 times/min.

To resolve these problems, a new electrical method was

reported in our previous study.15 In brief, after application of

electrolyte to the margin of a restoration and wiping it away to

leave only that penetrating into the marginal gap, microleak-

age of the restoration can be detected by measurement of the

change in conductance across the margin. Using this electrical

method, microleakage in both coronal and root surface

cavities could be detected, and the results of the evaluation

correlated significantly with those of the conventional dye

penetration test.15 Moreover, this method can detect exten-

sions of microleakage in the direction of the pulp chamber of

restorations, and the size of the restorative margin did not

influence the results.16 In our previous in vitro studies,15,16

however, it was difficult to evaluate the actual three-

dimensional microleakage of restorations for comparison

with the results by the electrical method, so dentin bonding

systems were not used in composite filling and the amount of

marginal leakage was determined using the depth of the

prepared cavities. Therefore, previous studies have estab-

lished the principle of this electrical method, but the accuracy

of the results cannot be evaluated clearly when resin

composite restorations are filled with a dentin bonding system

as in the clinical settings. For future clinical use of this method

with composite restorations, it is necessary to evaluate its

results with a dentin bonding system.

Recently, a quantitative three-dimensional method for the

evaluation of contraction gap formation was reported.17 This

three-dimensional method can assess the condition and

amount of microleakage of composite restorations more

accurately than the conventional two-dimensional method.15

Using the three-dimensional method, smaller microleakages

can be detected from composite restorations with a dentin

bonding system than without the bonding system. Therefore,

the purpose of this in vitro study was to investigate the

accuracy of the electrical method for the evaluation of

marginal leakage by the three-dimensional method using

dye penetration, and to evaluate the value of this electrical

method in clinical settings.

Fig. 2 – The prepared cavity and measurement of

conductance. Coronal cavity was prepared on the buccal,

palatal or lingual surface. Each cavity was around 3 mm in

diameter and 2 mm in depth. Conductance was measured

consecutively from the starting point of measurement to

the end point, across the margin.

2. Materials and methods

2.1. Preparation of specimens

In this study, 18 extracted non-carious human molars, which

had been stored in physiological saline at 4 8C, were used

within 1 year of extraction. The molars were obtained from

Osaka University Dental Hospital in accordance with the

protocol approved by the Ethics Committee of Osaka Uni-

versity Graduate School of Dentistry. The molars were divided

into three groups (bonding group, non-bonding group and

non-restorative group, each n = 6). The roots of the molars

were reduced from the apex with a polishing machine (Ecomet

III, Buehler Ltd., Lake Bluff, IL, USA). The pulp tissues of the

molars were then removed using a round bur and a fissure bur.

The coronal enamel was reduced as shown in Figs. 1 and 2

with a standardized cavity preparation device15,16 (Itoh

Engineering Co., Kyoto, Japan) after the molars had been fixed

on the device with self-curing acrylic resin (Uni-Fast II, GC

Corporation, Tokyo, Japan). The device has three micrometers

for controlling the movement of an air-turbine handpiece with

a diamond point. A circular cavity (3 mm in diameter, 2 mm in

depth, Fig. 2) was prepared on the buccal, palatal, or lingual

surface of each molar (bonding group and non-bonding group)

j o u r n a l o f d e n t i s t r y 3 5 ( 2 0 0 7 ) 2 6 8 – 2 7 4270

using the standardized cavity preparation device after

measuring the baseline conductance (the method of measure-

ment is described in the next section). The cavities of the

molars in the bonding group were treated with a dentin

bonding system (Clearfil Liner Bond II S, Kuraray Co. Ltd.,

Osaka, Japan). This binding system included a self-etching

primer, and the procedure was as follows. One drop of Primer

A (Batch No. 00114A) and one drop of Primer B (Batch No.

00111B) were mixed, applied to the cavities for 30 s and air-

dried for 5 s. Bonding Agent (Batch No. 00186B) was applied,

spread with a gentle air-blow and cured for 20 s with a visible

light-curing unit (Optilux 401, Kerr GmbH, Karlsruhe, Ger-

many). The unit was tested with a radiometer attached to the

unit before curing. Then, a hybrid-type resin composite

(Clearfil AP-X, A3, Kuraray, Batch No. 00924A) was filled into

the cavities and cured in bulk for 40 s. The cavities of the

molars in the non-bonding group were filled with the resin

composite without treatment with the bonding system. After

the molars had been stored in physiological saline at 37 8C for

24 h, the surfaces of the restorations were finished and

polished with polishing disks (Soflex, 3M ESPE, St. Paul, MT,

USA). In the non-restorative group, the cavities were not

prepared, were not treated with the bonding system, and were

not filled with the resin composite.

2.2. Load cycling and measurement of conductance

After composite filling, all molars were subjected to mechan-

ical load cycling. The root of the molar was mounted in a metal

tube in a load cycling machine (Yuyama Irika Co., Osaka,

Japan) with self-curing acrylic resin. In this machine, the

plunger was moved repeatedly down 5 mm from the hor-

izontal plane of the molar in the direction of the tooth axis and

the molar was moved repeatedly in a buccal lingual direction

1.5 mm against the plunger (Fig. 1). The load was 425 g (plumb:

400 g; plunger only: 25 g) and the cycle was 60 times/min.

After 10,000 load cycles, the conductance of the margin of

the molar was measured by the following method. An

experimental electrode that can measure conductance under

constant pressure (tip 0.6 mm in diameter at the pointed head)

was used in this experiment; the structure of this electrode is

described in our previous study.15,16 The pulp chamber of each

molar was filled with physiological saline and a copper wire

electrode was put into it. Both of the electrodes were

connected to an impedance meter (HP4263B, Hewlett-Packard

Japan Ltd., Tokyo, Japan) with a personal computer (Endeavor,

NT2500, Epson Direct Corp., Tokyo, Japan) (Fig. 2). The method

of conductance measurement is described in our previous

studies.15,16 In brief, the starting and end points were located

on each composite surface or coronal enamel surface 1.5 mm

from the margin of the restoration. After 10 ml of physiological

saline had been applied on the margin of the restoration with a

micropipette (Varipette 4710, Eppendorf, Hamburg, Germany)

and the excess wiped off, the conductance across the margin

from the starting point to the end point was measured

continuously with the experimental electrode. When this

process was carried out, the surface of the margin was not wet

to the naked eye. The electrical frequency was 100 kHz and the

voltage was 1 V. The change in conductance was defined as the

difference between the conductance at the starting point or

the end point, whichever was larger, and the maximum

conductance excluding the starting and end points. The

change in conductance for each margin was measured 10

times and the mean (CC value) was calculated for data

analysis. As described in the preceding section, before cavity

preparation, the CC value of the enamel surface (at the

location of the cavity) was measured and calculated using the

same method (including measurement after load cycling in

the non-restorative group).

2.3. Three-dimensional analysis

After measurement of conductance and load cycling, the pulp

chambers of the molars were blocked with self-curing acrylic

resin and the molar surfaces were coated with nail varnish,

except for an area 0.5 mm around the margin. The molars

were immersed in 2 wt% methylene blue solution at 37 8C for

24 h, then taken out and the solution wiped off. The molars

were then placed on the standardized cavity preparation

device and were reduced in thickness by 100 mm in a direction

parallel to the cavity floor of the composite restorations, from

the surface of the restoration to the cavity floor. An image of

each reduced surface was taken using an operation micro-

scope (Universal S3, Carl Zeiss, Oberkochen, Germany) with a

CCD camera (Hitachi, Tokyo, Japan). The magnification was

20�. The sequence of reducing the sections by 100 mm and

image taking was repeated for all molars in the bonding group

and the non-bonding group (14–20 images of reduced sections

were taken for each cavity).

After image taking, three-dimensional analysis was carried

out. The method of this analysis has been described

previously.17 The image files were manipulated using a photo

retouching software package (Adobe Photoshop ver.5.0, Adobe

Systems, San Jose, CA, USA). Dye penetration of the sections

was marked over the images of the margins (black lines:

margin penetrated with dye; gray lines: margin not penetrated

with dye). Three-dimensional images were then created, with

black and gray lines marked over the images, using analytical

software (NIH image, National Institutes of Health, USA). The

length of the cavity wall (CL), the area of the cavity floor (CF)

and the length of the margin penetrated by dye (DL) on the

images were measured using the analytical software, and the

surface area of the cavity wall (CA), the surface area of dye

penetration into the cavity wall (DA) and the rate of dye

penetration (DR) were calculated using the following formu-

lae. In addition, the maximal depth of dye penetration (DD)

was obtained by image analysis:

CA ðmm2Þ ¼XðCL� 100 mmÞ þ CF;

DA ðmm2Þ ¼XðDL� 100 mmÞ; DR ð%Þ ¼ DA

CA� 100

2.4. Statistical analysis

For comparison of the three groups after load cycling, CC values

were analyzed by one-way ANOVA and the Games–Howell test.

For comparison between before cavity preparation and after

load cycling, CC values were analyzed by the t-test. In addition,

the relationships between CC values after load cycling and DA,

between these and DR, and between these and DD were

j o u r n a l o f d e n t i s t r y 3 5 ( 2 0 0 7 ) 2 6 8 – 2 7 4 271

Fig. 3 – Example of three-dimensional image of dye penetration after load cycling (bonding group: sample 5). Black portion

(line): cavity wall with dye penetration. Gray portion (line): cavity wall without dye penetration. Gothic line: cavity margin

and line angles. The image data of each section, which consisted of black and gray lines, are shown by the same six black

and gray lines.

analyzed by Pearson’s correlation coefficient and Fisher’s Z

transformation. The significance level was set at 5% in all

analyses.

3. Results

Examples of three-dimensional images of dye penetration in

the molars in the bonding group after load cycling are shown

in Fig. 3. Dye penetration in the bonding group was observed

on the lateral walls only, and not the cavity floor. In the

bonding group, four specimens did not show dye penetration

into dentin, while two specimens did show penetration into

dentin. However, because the direction of section was almost

parallel to the dentino-enamel junction of the specimens, it

Fig. 4 – CC values of the three groups after load cycling and

before cavity preparation. There were significant

differences in the CC values after load cycling among the

three groups ( p < 0.05). In the non-bonding group and the

bonding group, the CC values after load cycling were

significantly larger than those before cavity preparation

( p < 0.05), but in the non-restorative group, there were no

significant difference between them.

was difficult to determine accurately whether the wall was

enamel or dentin. In the non-bonding group, the dye

penetrated all cavity walls, including the cavity floor.

Fig. 4 shows the CC values of the three groups after load

cycling. These increased in the following order: non-restora-

tive group, bonding group, and non-bonding group. There

were significant differences in the CC values after load cycling

among each of the three groups (p < 0.05). In the non-bonding

group and the bonding group, the CC values after load cycling

were significantly larger than those before cavity preparation

(p < 0.05), but in the non-restorative group, there was no

significant difference between the CC values after load cycling

and those before cavity preparation.

Fig. 5 shows the relationship between CC values after load

cycling and DA. The correlation coefficient was 0.932 and there

was a significant correlation (p < 0.05). Fig. 6 shows the

relationship between CC values after load cycling and DR. The

correlation coefficient was 0.920 and there was a significant

correlation (p < 0.05). Fig. 7 shows the relationship between

CC values after load cycling and DD. The correlation coefficient

was 0.732 and there was a significant correlation (p < 0.05).

4. Discussion

In this study, the actual amount of microleakage was

determined by a three-dimensional method and the results

of the electrical evaluation of microleakage were compared

with this amount. In general, the actual amount of micro-

leakage has been determined by tracer penetration methods.

In these methods, after immersion of an extracted tooth with

restorative into a tracer solution (dye,18,19 chemical tracer,20,21

radioactive tracer22,23), the tooth has been sectioned and the

extent of penetration of tracer on one or a few sections

observed. Thus, these methods are not suitable for three-

dimensional evaluation. However, the electrical method for

the evaluation of microleakage is thought to detect three-

dimensional electrolyte penetration into the marginal gap of

composite restoratives. So, the three-dimensional method is

j o u r n a l o f d e n t i s t r y 3 5 ( 2 0 0 7 ) 2 6 8 – 2 7 4272

Fig. 5 – The relationship between CC values after load

cycling and DA (correlation coefficient: 0.932, p < 0.05).

Fig. 7 – The relationship between CC values after load

cycling and DD (correlation coefficient: 0.732, p < 0.05).

suitable for determination of the actual amount of micro-

leakage when the accuracy of electrical method is being

evaluated. In this study, the accuracy of the electrical method

was evaluated by three-dimensional parameters (CA and DR)

using restorations filled with a bonding system, as in the

clinical settings. In addition, for comparison, evaluation using

a two-dimensional parameter (DD) was undertaken. Use of the

three-dimensional method and of the bonding system to

achieve a clearer evaluation of the accuracy of the electrical

method are new in this study, compared with our previous

study.15,16

The three-dimensional method for the evaluation of

microleakage reported in our previous study was also used

in this study.17 We did not previously use dye penetration, but

it was used in this study, as in Gale et al.24 By this method, it

was difficult to evaluate the interface between the cavity wall

and the filling materials, which was parallel to the sectioning

surface.17 Therefore, in this study, when the surface area of

the cavity wall was calculated, the area of the cavity floor,

which was calculated from the actual diameter of the cavity,

was added toP

(CL � 100 mm) to overcome this weak point.

Load cycling has a significant effect on microleakage.25 In

particular, flexural load cycling has a significant effect on the

microleakage of cervical resin composite restorations.26

Flexural load cycling was used in this study before the

Fig. 6 – The relationship between CC values after load

cycling and DR (correlation coefficient: 0.920, p < 0.05).

evaluation of microleakage to increase the amount of

microleakage in each specimen as far as possible and thereby

obtain clear data. However, flexural load cycling may lead to

microfractures on the cervical region of specimens near the

margin of the composite restoration, due to the concentration

of load stresses in this region. It is therefore possible that, in

this experiment, electrolyte penetrating into microfractures

may have been detected as microleakage. As this is a weak

point of this experiment, future investigations addressing this

are necessary.

The CC values of the non-restorative group after load cycling

were significantly smaller than those of the bonding group and

the non-bonding group (p < 0.05, Fig. 4). In addition, in the non-

bonding group and the bonding group, the CC values after load

cycling were significantly larger than those before cavity

preparation (p < 0.05, Fig. 4). From these results, the electrical

method can be seen to detect microleakage. There was a

significant correlation between CC values after load cycling and

DA, DR and DD (p < 0.05), and the correlation between them

was strong (Figs. 5–7). It is therefore considered that the two- or

three-dimensional amount of microleakage can be detected by

measurement of change in conductance with the electrical

method. In some previous electrical methods,13,14 electrolyte

was applied to the whole margin of restorations and was not

wiped off, and the conductance or impedance was then

measured. Consequently, electrolyte other than that which

had penetrated into the marginal gap could have influenced the

results. In this study, because the excess electrolyte on the

margin of the restorations was wiped off, the results were not

influenced by excess electrolyte and showed microleakage

more clearly than previous methods.13,14 The results showed

that this electrical method can detect the degree of micro-

leakage into the interface betweenrestorationsand the dentinal

cavitywall,and that the sensitivity of themethod isveryhigh. In

addition, the correlation coefficient between CC values and the

surface area of dye penetration (DA) and between CC values and

the rate of dye penetration (DR) (three-dimensional amount)

were larger than that between CC values and the depth of dye

penetration (DD) (two-dimensional amount). Accordingly, the

results are thought to depend on three-dimensional penetra-

tion of electrolyte. However, because the depth of dye

penetration in this experiment was measured from many

j o u r n a l o f d e n t i s t r y 3 5 ( 2 0 0 7 ) 2 6 8 – 2 7 4 273

sectional images, compared with one or a few in previous

studies,18,19 the values were maximal on many images. Thus,

there was a significant correlation even between CC values and

the two-dimensional amount of dye penetration (DD) in this

experiment.

The demonstration that this electrical method can detect

microleakage of composite restorations with a dentin bonding

system has possible clinical relevance. In future clinical

applications, it is thought that a buccal clip might be placed

into the oral cavity as an electrode, and the CC values of the

margin of composite restorations might be measured. If this

electrical method is applied clinically, the amount of micro-

leakage of restorations will be determined without their

destruction. In addition, the CC values of margins exposed to

microleakage were correlated with the three-dimensional

amount of dye penetration in this experiment, but the CC

values of margins not exposed to microleakage are thought to

be smaller than those of margins exposed to microleakage. It

may be possible to detect the location of microleakage by this

electrical method in future clinical use. Moreover, as this

method does not involve the destruction of restorations, it is

possible that the progress of microleakage could be monitored

over time. It is thought that this method can be applied to

ceramic inlay restorations and glass-ionomer restorations in

addition to composite restorations, because these are highly

insulated. However, this method cannot be applied to metal

restorations, because the electric current would flow in the

metal itself in addition to the electrolyte.

Before this electrical method can be applied clinically, it

will be necessary to further examine the optimal frequency

and voltage for minimization of external noise. Higher voltage

minimizes external noise,9 but if the voltage is very high,

irreversible or reversible pulpitis may occur.27 Other factors

such as pulp tissue, pulpal pressure, dentin condition, the

moving speed of the electrode, and the optimal electrolyte and

application method must be investigated.

5. Conclusion

In this study, the accuracy of an electrical method for the

evaluation of microleakage was investigated by three-dimen-

sional analysis of dye penetration into the cavity wall of

restorations. The correlation between changes in conductance

(CC values) after load cycling and the surface area of dye

penetration into the cavity wall (DA), and between CC values

and the rate of dye penetration (DR), were stronger than the

correlation between CC values and the depth of dye penetra-

tion (DD). Therefore, the results of this electrical method for

microleakage evaluation were correlated more strongly with

the three-dimensional amount of marginal leakage (DA and

DR) than the two-dimensional amount (DD).

Acknowledgments

This work formed part of the 21st Century COE entitled

‘Origination of Frontier BioDentistry’ held at Osaka University

Graduate School of Dentistry and supported by the Ministry of

Education, Culture, Sports, Science and Technology. This

investigation was also supported, in part, by a Grant-in-Aid for

Scientific Research (17591990) from the Japan Society for the

Promotion of Science.

r e f e r e n c e s

1. Schmeiser R, Gulzow HJ. The influence of luting cements onthe fermentation of sucrose by oral microorganisms. Journalof Dentistry 1999;27:37–41.

2. Camps J, Dejou J, Remusat M, About I. Factors influencingpulpal response to cavity restorations. Dental Materials2000;16:432–40.

3. Murray PE, About I, Franquin JC, Remusat M, Smith AJ.Restorative pulpal and repair responses. Journal of AmericanDental Association 2001;132:482–91.

4. Ben-Amer A, Cardash HS, Judes H. The sealing of the tooth/amalgam interface by corrosion products. Journal of OralRehabilitation 1995;22:101–4.

5. Moreira Jr G, Sobrinho APR, Nicoli JR, Bambirra EA, FariasLM, Carvalho MAR, et al. Evaluation of microbial infiltrationin restored cavities—an alternative method. Journal ofEndodontics 1999;25:605–8.

6. Going RE. Microleakage around dental restorations: asummarizing review. Journal of American Dental Association1972;84:1349–57.

7. Triadan H. When is microleakage a real clinical problem?Operative Dentistry 1987;12:153–7.

8. Martinez CR, Greener EH. Utilization of electricalconductivity as an alternative method of assessing marginalleakage of pit and fissure sealants. Journal of OralRehabilitation 1976;3:69–74.

9. Nakano Y. A new electrical testing method on marginalleakage of composite resin restorations. Japan Journal ofConservative Dentistry 1985;28:1183–98.

10. Momoi Y, Iwase H, Nakano Y, Kohno A, Asanuma A,Yanagisawa K. Gradual increases in marginal leakage ofresin composite restorations with thermal stress. Journal ofDental Research 1990;69:1659–63.

11. Verdonschot EH, Rondel P, Huysmans MCDNJM. Validity ofelectrical conductance measurements in evaluating themarginal integrity of sealant restorations. Caries Research1995;29:100–6.

12. Von Fraunhofer JA, Adachi EI, Barnes DM, Romberg E. Theeffect of tooth preparation on microleakage behavior.Operative Dentistry 2000;25:526–33.

13. Hoppenbrouwers PMM, Scholberg HPF, Borggreven JMPM.Measurement of the permeability of dental enamel and itsvariation with depth using an electrical method. Journal ofDental Research 1986;65:154–7.

14. Huysmans MCDNJM, Verdonschot EH, Rondel P. Electricalconductance and electrode area on sound smooth enamelin extracted teeth. Caries Research 1995;29:88–93.

15. Iwami Y, Yamamoto H, Ebisu S. A new electrical method fordetecting marginal leakage of in vitro resin restorations.Journal of Dentistry 2000;28:241–7.

16. Iwami Y, Yamamoto H, Hayashi M, Takeshige F, Ebisu S.Validity of electrical conductance measurements inevaluating marginal leakage around resin compositerestorations. Operative Dentistry 2002;27:606–12.

17. Iwami Y, Shimizu A, Hayashi M, Takeshige F, Ebisu S. Three-dimensional evaluation of gap formation of cervicalrestorations. Journal of Dentistry 2005;33:325–33.

18. Kubo S, Yokota H, Yokota H, Hayashi Y. The effect oflight-curing modes on the microleakage of cervicalresin composite restorations. Journal of Dentistry2004;32:247–54.

j o u r n a l o f d e n t i s t r y 3 5 ( 2 0 0 7 ) 2 6 8 – 2 7 4274

19. Owens BM, Johnson WW, Harris EF. Marginal permeabilityof self-etch and total-etch adhesive systems. OperativeDentistry 2006;31:60–7.

20. Ramos RP, Chimello DT, Chinelatti MA, Dibb RG, Mondelli J.Effect of three surface sealants on marginal sealing of classV composite resin restorations. Operative Dentistry2000;25:448–53.

21. Tulunoglu o. uctash M, Alacam A, omurlu H. Microleakageof light-cured resin and resin-modified glass-ionomerdentin bonding agents applied with co-cure vs.pre-cure technique. Operative Dentistry 2000;25:292–8.

22. Powis DR, Prosser HJ, Shortall AC, Wilson AD. Long-termmonitoring of microleakage of composites. Part I:Radiochemical diffusion technique. Journal of ProstheticDentistry 1988;60:304–7.

23. Puckett AD, Fitchie JG, Bennett B, Hembree JH. Microleakageand thermal properties of hybrid ionomer restoratives.Quintessence International 1995;26:577–81.

24. Gale MS, Darvell BW, Cheung GSP. Three-dimensionalreconstruction of microleakage pattern using a sequentialgrinding technique. Journal of Dentistry 1994;22:370–5.

25. Jang KT, Chung DH, Shin D, Garcıa-Godoy F. Effect ofeccentric load cycling on microleakage of class V flowableand packable composite resin restorations. OperativeDentistry 2001;26:603–8.

26. Kubo S, Yokota H, Sata Y, Hayashi Y. The effect of flexuralload cycling on the microleakage of cervical resincomposites. Operative Dentistry 2001;26:451–9.

27. Ricketts DNJ, Kidd EAM, Wilson RF. A re-evaluation ofelectric resistance measurements for the diagnosis ofocclusal caries. British Dental Journal 1995;178:11–7.