journal of dentistry 36 (2008) 10251032

available at www.sciencedirect.com

journal homepage: www.intl.elsevierhealth.com/journals/jden

The effect of several dentin desensitizers on shear bond strength of adhesive resin luting cement using self-etching primer
Jung-Bo Huh a, Jee-Hwan Kim a, Moon-Kyu Chung a, Ho-yong Lee a, Yong-Geun Choi b, June-Sung Shim a,*
a

Department of Prodthodontics, Yonsei University College of Dentistry, 134 Sinchon-dong, Seodaemum-gu, Seoul 120-752, Republic of Korea b Graduate School of Clinical Dentistry, Korea University, 126-1, 5-ka, Anam-dong, Seongbuk-gu, Seoul 136-705, Republic of Korea

article info
Article history: Received 20 May 2008 Received in revised form 25 August 2008 Accepted 27 August 2008

abstract
Objectives: Dentin desensitizers can inhibit the bonding between dentin and resin cements. This study examined the effect of the previous application of desensitizers on the shear bond strength of one resin cement using self-etching primer to dentin. Materials and methods: One hundred and twenty-ve dentin exposed teeth were randomly assigned to four experimental groups and one control group of 25 teeth each. Four dentin desensitizers were applied to the four experimental groups, respectively. The dentin desensitizers used were SuperSeal1 (Phoenix Dental, Inc., USA), MS-Coat1 (Sun Medical

Keywords: Dentin desensitizer Bonding strength Resin cement

Co. Ltd, Japan), Gluma1 (Heraeus Kulzer, Germany), and Copalite Varnish1 (Cooley & Cooley Ltd, USA). Panavia F (Kuraray Co. Ltd, Tokyo, Japan) was attached to the top of each experimental and control group teeth using an Ultradent testing jig (Ultradent Product, Inc., Utah, USA), and the shear bond strength was measured using a Universal testing machine (Model 6022, Instron Co., Canton, MA, USA). The tooth surface was examined by scanning electron microscopy (SEM, JSM-T2000, JEOL, Tokyo, Japan). Results: The control group showed the greatest shear bond strength (14.74 MPa) followed by SuperSeal1 (12.33 MPa), Gluma1 (5.28 MPa), MS-Coat1 (4.44 MPa) and Copalite VarnishTM (3.14 MPa). There was no signicant difference in shear bond strength between the control group and the experimental group treated with Superseal1. The shear bond strength in the other experimental groups treated with Gluma1, Varnish1, and MS-Coat1 was similar to each other but signicantly lower than control or Superseal1. SEM showed revealed resin tags in most of the dentinal tubules in the experimental group treated with the Superseal1. Conclusion: Among the four dentin desensitizers, Superseal1 was the only one that did not interfere with the process of resin bonding. The other dentin desensitizers that contained a resin ingredient interfered with resin retention. # 2008 Elsevier Ltd. All rights reserved.

* Corresponding author. Tel.: +82 2 2228 8713; fax: +82 2 312-3598. E-mail address: jfshim@yuhs.ac (J.-S. Shim). 0300-5712/$ see front matter # 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jdent.2008.08.012

1026 1. Introduction

journal of dentistry 36 (2008) 10251032

Adhesive resin luting cements are widely used in many restorations. The internal bond between the adhesive luting cement and dentin is an important factor that prevents microleakage and the retention of the restoration. However, the application of etch-and-rinse smear layer-removing adhesives is technique sensitive due to the outward ow of dentinal uid via dentin tubules that interferes with resin monomer inltration.1 Post-operative sensitivity is occasionally associated with the use of etch-and-rinse adhesives.2 In an attempt to reduce technique sensitivity, self-etching primer systems have been developed. In these systems, the self-etching primer is applied to the tooth surface, followed by solvent-free hydrophobic bonding resins.3 Yet, self-etching systems that modify the smear layer are also negatively affected by dentinal uid ow during its application.4 Most dentinal tubules in prepared teeth can be initially sealed using an adhesive luting cement during cementation. However, the volatile solvent in adhesive luting cement may evaporate quickly leading to continuous transudation of dentinal uid before polymerization of the adhesive.5 Transudation of dentinal uid and leakage through the hybrid layer may result in the entrapment of water blisters along the adhesivecomposite interface6 that may induce rapid dentinal uid shifts during mastication causing dentin sensitivity.7 Pashley8 stated that a material containing oxalate can seal the dentinal tubules and decrease the sensitization of dentin due to transudation of dentinal uid and leakage after nal cementation of the crown. Occluding dentinal tubules can also reduce internal dentin wetness during the etch-and-rinse bonding procedures.9 And Pashley10 reported that the damage caused by the microleakage of the bonding system could be decreased if permeability could also be reduced by sealing the dentinal tubules. For the above reasons, various dentin desensitizers have been used in cavity preparation or in spontaneous and temporary cold sensitization after abutment preparation. Dentin desensitizers have been developed in various forms. Historically, ranges of methods have been introduced, such as sealing the dentinal tubule by forming a diffusion barrier through medication or through a bonding agent. Most commonly used agents in the treatment of dentine sensitivity can be broadly classied by their modes of action: antiinammatory drugs, protein precipitants, tubule occluding agents, tubule sealants, and recently, laser treatment.4 As stated above, there are many components and application methods for desensitizers. Some of the components may impede the interaction between the dentin and resin cement, which would reduce the bond strength and

crown was cemented with zinc phosphate, glass-ionomer, resin-modied glass-ionomer cement after applying two types of dentin desensitizers, there was no signicant difference in retention compared with the control group. Early studies reported a reduction in bond strength caused by poor resin inltration due to crystal precipitation on the dentin surface.9,12 Several studies have evaluated the effect of oxalate desensitizers on bond strength of etch-and-rinse adhesives to dentin and reported that it did not compromise the early bond strength.9,13 However, there are very few reports on the effect of pretreatment of dentin with desensitizers prior to cementing the restoration with resin adhesive luting cements using self-etching primer. Selfetching adhesives condition and prime enamel and dentin simultaneously without rinsing. So it is possible to remain some component of desensitizer on the dentin surface. If a desensitizer interferes with the bonding system, it will have a negative effect on bonding. For this reason, this study examined the effect of bonding between the resin bonding cement using self-etching primer and dentin after being pretreated with dentin desensitizers using various mechanisms and components.

2.
2.1.

Methods
Fabrication of tooth specimen

Human molars extracted within 1 month were used. The criteria for choosing the teeth are as follows: (1) maxillary and mandibular molars were used to gain a larger area of dentin, and (2) teeth with caries were excluded to reduce the variability. For the convenience of evaluating the shear strength, the 125 teeth selected were embedded in an acrylic resin using custom made cylinder-shaped Teon molds with a diameter and height of 30 mm. They were embedded to let its occlusal surfaces to be as much parallel as possible to the bottom surface of acrylic resin mold in order to allow exposure of transversally cut dentin when preparation of those specimens. The specimens were prepared until the dentin was exposed using a Model Trimmer (Whip Mix Corporation, Louisville, KY, USA). The teeth were carefully prepared so that the exposed dentin surface was parallel to the acrylic resin mold, without exposing the pulp. They were ground off using wet 320 grit silicon carbide abrasive paper. Prepared teeth were washed with pumice with low speed hand piece in order to remove the contaminants acquired during preparation and polishing procedures and they were immersed in distilled water to supply enough moisture to the exposed dentin surface.

Table 1 Composition and manufacturer of the desensitizing agents. Agent

SuperSeal MS-Coat1
1

Composition
Oxalic acid, potassium salt Liquid A: water, copolymer with sulfonic group Liquid B: water, oxalic acid 35% (2-hydroxyethyl) methacrylate, water, 5% glutaraldehyde Copal resin, ether, acetone, alcohol 11

Manufacturer
Phoenix Dental, Inc., Fenton, USA Sun Medical Co. Ltd, Moriyama, Japan Heraeus Kulzer, Dormagen, Germany Cooley & Cooley Ltd, USA

impede sufcient sealing. Swift et al.

Gluma1 Coplalite VarnishTM

reported that when the

journal of dentistry 36 (2008) 10251032

1027

Ultradent company, which surrounded the specimen in the upper area, was used to load a force onto the interface between the dentin and resin cement with a speed of 2 mm/min until fracture occurred. Those specimens that showed denitive fractures within Panavia cement cylinder after loading were excluded from the results of this study.

2.5.

Scanning electron microscopy (SEM)

Fig. 1 Ultradent shear bond test device with Teflon mold (Ultradent Product, Inc., Utah, USA) for bonding resin cement to dentin.

After surface treatment of the specimen, the fracture pattern and microstructure of the tooth surface was examined by SEM (JSM-T2000, JEOL, Tokyo, Japan) at various magnications. Four specimens that had the nearest numerical value to the mean shear bond strength in each experiment group were chosen and examined.

2.6. 2.2. Application of the dentin desensitizer to the tooth specimen

The 125 specimens were randomly divided into ve groups: a control group, and four experimental groups of 25 teeth each. Table 1 lists the medicaments used for surface treatment. The four kinds of dentin desensitizers mentioned in Table 1 were applied to the categorized specimens according to the manufacturers instructions. The control group of exposed dentin surface was not treated.

Statistics analysis

Descriptive data was used to determine the distribution of the experiment and control groups. A test of equal variance was used to analyze the difference in the distribution rate. A Welch ANOVA test was used to conrm the difference between the experimental and control groups. TukeyKramer HSD was used as a post-test to analyze the difference between each experimental group and the control group at a = 0.05.

3.
3.1.

Results
Shear bond strength

2.3.

Bonding of resin cement

After 60 s of applying the ED primer to the dentin surface treated with different desensitizers, the primer solvent was evaporated with a gentle stream of air for 5 s, Panavia F was injected through the Ultradent plastic hole (Fig. 1) and a cemented cylinder of 5 mm in height was made. The specimens bonded with Panavia cement were covered with Oxyguard gel. They were stored in 100% humidity for 24 h prior to measuring their shear bond strength. Table 2 lists the composition of Panavia F (Kuraray Co. Ltd, Tokyo, Japan) (Fig. 2).

The exposed dentin in the control and experimental groups was covered with Panavia F (Kuraray Co. Ltd, Tokyo, Japan). Fig. 3 shows the shear bond strength. The test of equal variances and the degree of freedom in denominator analysis were carried out under a 5% level of signicance. As shown in Fig. 3, there was no signicant difference between the experimental groups and control group treated with SuperSeal1. The experimental groups treated with Gluma1, Copalite VarnishTM, MS-Coat1 showed a signicantly lower ( p < 0.05) shear bond strength than the control group (Table 3).

2.4.

Shear bond strength test 3.2. Scanning electron microscopy

A manufactured specimen was connected to the Universal testing machine (Model 6022, Instron Co., Canton, MA, USA) to test the shear bond strength. A specially made device from the

Fig. 4 shows the microstructure of the surface after treating the dentin surface with the four different types of dentin

Table 2 Components and application of the test luting resin cement system (Panavia F (Kuraray Medical, Inc., Osaka, Japan)) used in this study. Material
Self-etching primer (ED primer)

Component
Liquid A Liquid A Past B Past

Composition
HEMA, MDP, 5-NMSA, accelerator, water B5-NMSA, accelerator, water MDP, co-monomers ller, NaF, BPO Co-monomers ller, NaF, amine, initiator Polyethyleneglycol

Panavia F cement

Oxyguard II

HEMA: 2-hydroxyethyl methacrylate; MDP: 10-methacryloyloxydecyl dihydrogen phosphate; 5-NMSA: N-methacryloxyl-5-aminosalicyclic acid.

1028

journal of dentistry 36 (2008) 10251032

Fig. 2 Schematic diagram showing the method for specimen preparation.

Fig. 3 Means of all pairs using a TukeyKramer HSD.

desensitizers. Each group is labeled AE as follows. The control group (A) had open dentinal tubules. The group treated with Gluma1 (B) showed semi-closed dentinal tubules that were completely plugged, and showed a delicate net form. The application of Gluma1 medicament to the intertubular dentin was observed but it did not form a thick coating. The group treated with MS-Coat1 (C) formed a thick membrane. The membrane did not cover the dentinal tubule totally but most of it extended towards the tubule. In addition, the intertubular dentin was completely covered. The group treated with

Table 3 Means and standard deviations. Number

Control Gluma MS-Coat SuperSeal Varnish 25 25 25 25 25

Mean
14.74 5.28 4.44 12.33 3.14

S.D.
5.96 3.97 1.99 5.06 2.50

SuperSeal1 (D) showed small particles in the dentinal tubules. There were some tubules that were completely covered and some that were not. It was possible to nd particles in the tubules in the opening area. Most of the surface in the group treated with Copalite VarnishTM (E) was covered with a thick layer of resin, as indicated by the asterisk. The image is shown next to the interface that was not treated in order to highlight the difference. In Fig. 5, after the dentin desensitizer had been applied and Panavia F was attached, the shear bond strength was tested and the fractured surfaces were examined. In the control group, (A) resin tag that went into the tubule before the cement polymerized was observed broken off at the dentin surface. Resin tags were observed in group (B) but it was not as consistent as the control group. In addition, a smooth material, Gluma1, was observed between the dentin tubules. Group (C) showed a similar microstructure to that shown in Fig. 4C. Resin tags could not be identied due to the thick layer of medicament plugging the dentin surface. Moreover, there was no evidence of resin penetration due to the smooth

journal of dentistry 36 (2008) 10251032

1029

Fig. 4 Scanning electron microscopic view of the dentin surface after treatment with the desensitizing agents.

surface. Compared with Fig. 4D, group (D) (Fig. 5D) showed the residual resin cement, which explains why there was a fracture. Group (E) did not show any resin tags.

4.

Discussion

In this study, the shear bond strength of the resin cement was measured on dentin surfaces pretreated with four different types of dentin desensitizers that had previously been applied to exposed dentin surfaces. The aim was to determine which dentin desensitizer had the least effect on the bond strength between the dentin and the resin cement using self-etching primer after sealing the dentinal tubule with the dentin desensitizers. Dentin is a tissue containing dentinal tubules with an approximate diameter of 0.62.0 mm.14 More than 2 million dentinal tubules can be exposed per cm2, which once exposed can provide pathways to and from the pulp. Dentinal tubules that are opened during preparation can be plugged by bonding cement during the cementation of a restoration. However, in case of adhesive resin luting cement, the volatile solvent in adhesives may evaporate quickly leading to continuous transudation of dentinal uid before polymerization of the adhesive.5

There is considerable controversy regarding how to seal the dentinal tubules. First, there is the method of using a smear layer. The smear layer can be dened as an amorphous layer of grinding debris slightly bonded to the remnants of the dentin surface after using rotary equipment (bur), and has a thickness of 15 mm.15 Pashley16 reported that the smear layer can decrease the level of dentin permeability by 98% when compared with dentin treated with acid. However, Kerns et al.17 reported that the smear layer may disappear in a few days by the acid produced by the oral environment if there is some microleakage. The second method is using the oxalate particles. When acidic oxalates are applied to the dentin surface, it liberates calcium from the dentin to produce an insoluble calcium oxalate crystals that block dentinal tubules. Pashley and Galloway18 reported that if the dentin surface is treated with an oxalate solution, the hydraulic conductance is decreased, which can efciently prevent dentin sensitivity. The third method is using dentin bonding agents.19,20 The dentin bonding agent applied to the smear layer can increase the mechanical retention and resistance to liquids and acids but it unclear how long the effect will last. There is also a report showing that it is less efcient than forming a smear layer.21

1030

journal of dentistry 36 (2008) 10251032

Fig. 5 Scanning electron microscopic view of the dentin surface after treatment with the desensitizing agents and fracture resin cement (Panavia F).

As mentioned above, there has been considerable effort in desensitization. Pashley8 reported that there is a 90% decrease when using oxalate. In addition, Pashley et al.22 had some doubt regarding whether the effect was long lasting. Previously, the application of oxalate with etch-and-rinse adhesives was shown to lower resindentin bond strength when used before the dentin was acid-etched. This result can be explained by the presence of acid-resistant calcium oxalate crystals on the dentin surface that may prevent etching and the penetration of adhesive resins into dentin surface.12 The MS-Coat1 used in this experiment has the same application theory as the mixture of oxalate and self-curing resin bond introduced by Kerns et al.17 MS-Coat1 contains oxalic acid and a mixture of polymethyl methacrylate and a copolymer of polystyrene sulfonic acid, which helps in the formation of calcium oxalate crystals. This eventually has the same effect as potassium oxalate. Zhang et al.23 reported that MS-Coat1 is sufciently acidic (pH 2.3) in that it does not require etching of the dentin surface and forms calcium oxalate crystals through the release of calcium ions. In addition, oxalate crystals become a mass through the emulsion of the copolymer, which is the component of the resin containing

desensitizer. However, this reaction is a factor that weakens the bond strength of the resin cement using self-etching primer in this experiment. SEM showed that the copolymer was not spread over the collagen matrix but was attached to the dentin surface. Pashley et al.24 reported that smear layer-covered dentin treated with an acidic oxalate desensitizer completely replaces the acid-labile smear layer with a crystal layer that is resistant to acid. This is considered to have an effect on bonding. Gluma1 contains HEMA and glutaraldehyde, which causes the coagulation of the dentin uid protein in the dentinal tubule and plugs the tubule.25 Although it cannot be said positively, because there was no consideration about the dentin uid protein, Gluma1 turned out to give a negative effect on bonding between dentin and the adhesive resin using self-etching primer in this study. There was a recent report by Araha et al.26 suggesting no inuence of Gluma1 desensitizer on bonding strength. However, there were differences in type of adhesive, acid etching method and bond strength test method used in two studies. Hydrophilic calcium oxalate forms an insoluble crystalline layer on intratubular dentin matrix. Such material used this

journal of dentistry 36 (2008) 10251032

1031

study was Superseal1. This material does not contain a resin component. It is acidic enough to remove the smear layer, and replace it with a layer of calcium oxalate crystals. SEM showed that the tubules were opened in the dentinal surface and the oxalate particles had penetrated the tubules. This phenomenon allowed the formation of resin tags and there was no signicant difference between the control group in the shear bond strength test. In other words, it is predicted that in soluble desensitizers, the ED primer reacts directly in dentin but desensitizers containing resin that block the reaction with dentin. Yet, signicantly lower long-term microtensile bond strength and enhanced nanoleakage were produced by the application of oxalate (Superseal, Phoenix Dental) with an etch-and-rinse two-step bonding system in the study of Vachiramon et al.27 Differences in acid etching method and test method must again be considered. Although there was no signicance, shear bond strength of Superseal1 group was smaller than that of control group in this study. Copal varnish has been used for more than 70 years to cover open dentinal tubules. However, it was reported that the varnish shows only transient efciency and has little adhesion.19 The purpose of including Varnish that is not so much applied nowadays is to verify the hypothesis saying resin component of desensitizer weakens resin bonding by blocking the effect of self-etching primer and to nd the similarities with other desensitizers with resin components. The varnish seemed to prevent both ED primer and Panavia cement from interacting with the underlying dentin, resulted in low shear bond strength. SEM revealed the formation of a thick layer of Varnish after application, which completely blocked the bonding between dentin and the resin cement. One limitation of this study was the use of non-vital teeth, which probably caused some loss of dentin uid protein. Such an environment could have prevented Gluma1 from reaction with dentin uid protein. This problem can be solved by lling the pulp chamber with 2% bovine albumin in water under simulated pulpal pressure.28

references

5.

Conclusion

This study examined the effect of several dentin desensitizers on the shear bond strength of an adhesive resin luting cement. Among the four dentin desensitizers examined, Only Superseal group showed similar results with the control group. The others showed statistically low shear bond strength with signicance. Dentin desensitizers with resin components blocked underlying dentin from reacting with self-etching primer in SEM. Based on such results, dentin desensitizer with resin component should better be limited when using them before bonding the restoration with adhesive resin luting cement using self-etching primer.

Acknowledgements
I would like to thank Moon-Ku Jung and June-Sung Shim for their comments and a large number of excellent suggestions during a meticulous and tireless shepherding process.

1. Brannstrom M. Etiology of dentin hypersensitivity. Proceedings of the Finnish Dental Association 1992;88:713. 2. Camps J, Pizaht S, Dejou J, Franquin JC. Effects of desensitizing agents on human dentin permeability. American Journal of Dentistry 1998;11:28690. 3. Van Meerbeek B, De Munck J, Yoshida Y, Inoue S, Vargas M, Vijay P, Van Landuyt K, Lambrechts P, Vanherle G. Buonocore memorial lecture. Adhesion to enamel and dentin: current status and future challenges. Operative Dentistry 2003;28:21535. 4. Ling TY, Gillam DG. The effectiveness of desensitizing agents for the treatment of cervical dentine sensitivity (CDS)a review. The Journal of the Western Society of Periodontology/Periodontal Abstracts 1996;44:512. 5. Jain P, Reinhart JW, Krell KV. Effect of dentin desensitizer and dentin bonding agents on dentin permeability. Americal Journal of Dentistry 2000;13:217. 6. Khan F, Young WG, Shahabi S, Daley TJ. Dental cervical lesions associated with occlusal erosion and attrition. Australian Dental Journal 1999;44:17686. 7. Lambrechts P, Van Meerbeek B, Perdigao J, Gladys S, Braem M. Restoration therapy for erosie lesions. European Journal of Oral Science 1996;104:22940. 8. Pashley DH. Dentin permeability, dentin sensitivity, and treatment through tubule occlusion. Journal of Endodontics 1986;12:46574. 9. Perdigao J, Lopes M. Dentin bondingQuestions for the new millennium. The Journal of Adhesive Dentistry 1999;1:191209. 10. Pashley DH. Consideration of dentine permeability in cytotoxicity testing. International Endodontics Journal 1988;21:14354. 11. Swift EJ, Lloyd AH, Felton DA. The effect of resin desensitizing agents on crown retention. Journal of the American Dental Association 1997;128:195200. 12. Rosenthal MW. Historic review of the management of tooth hypersensitivity. Dental Clinics of North America 1990;34:40325. 13. Renton-Harper P, Midda M. Nd:YAG laser treatment of dentinal hypersensitivity. British Dental Journal 1992; 172:136. 14. Pashley DH. Dynamics of the pulpo-dentine complex. Critical Reviews in Oral Biology and Medicine 1996;7:10433. 15. Brannstrom M. The cause of postrestorative sensitivity and its prevention. Journal of Endodontics 1986;12:47581. 16. Pashley DH. Smear layer; Physiologic consideration. Operative Dentistry 1984;3:139. 17. Kerns DG, Scheidt MJ, Pashley DH, Horner JA, Strong SL, Van Dyke TE. Dentinal tubule occlusion and root hypersensitivity. Journal of Periodontology 1991;62:4218. 18. Pashley DH, Galloway SE. The effect of oxalate treatment on the smear layer of ground surfaces of human dentin. Archives of Oral Biology 1985;30:7317. 19. Ferrari M, Cagidiaco MC, Kugel G, Davidson CL. Clinical evaluation of a one-bottle bonding system for desensitizing exposed roots. American Journal of Dentistry 1999;12:2439. 20. Swift EJ, May KN, Mitchell S. Clinical evaluation of Prime & Bond 2.1 for treating cervical dentin hypersensitivity. American Journal of Dentistry 2001;14:136. 21. Elgalaid TO, Youngson CC, McHugh S, Hall AF, Creanor SL, Foye RH. In vitro dentine permeability: the relative effect of a dentine bonding agent on crown preparations. Journal of Dentistry 2004;32:41321. 22. Pashley DH, Andringa HJ, Eichmiller F. Effect of ferric and aluminum oxalates on dentin permeability. American Journal of Dentistry 1991;4:1236.

1032

journal of dentistry 36 (2008) 10251032

23. Zhang Y, Agee KA, Pashley DH, Pashley EL. The effect of Pain-Free1 desensitizer on dentine permeability and tubule occlusion over time, in vitro. Journal of Clinical Periodontology 1998;25:88491. 24. Pashley EL, Tao L, Pashley DH. Effects of oxalate on dentin bonding. American Journal of Dentistry 1993;6: 1168. 25. Felton DA, Bergenholz G. Evaluation of the desensitizing effect of Gluma Dentin Bond on teeth prepared for complete-coverage restorations. International Journal of Prosthodontics 1991;4:2928.

26. Araha AC, Siquerira Juunior Ade S, Cavalcante LM, Pimenta LA, Marchi GM. Microtensile bond strengths of composite to dentin treated with desensitizer products. Journal of Adhesive Dentistry 2006;8:8590. 27. Vachiramon V, Vargas MA, Pashley DH, Tay FR, Geraldeli S, Qian F, Armstrong SR. Effects of oxalate on dentin bond after 3-month simulated pulpal pressure. Journal of Dentistry 2008;36:17885. 28. Tagami J, Nakajima M, Hosoda H. Inuence of dentine primers on the ow of bovine serum through dentine. Archives Oral Biology 1994;39:146S.