I.M. Saleh, I.E. Ruyter, M. Haapasalo & D. Orstavik
NIOM, Scandinavian Institute of Dental Materials.
Department of Endodontics, Institute of Clinical Dentistry, University of Oslo, Norway.

Introduction.
The standard method of obturation of the root-canal system uses a core material in combination with a rootcanal sealer. Despite ongoing research and recent developments in endodontic materials, complete sealing of the root-canal system with currently accepted materials and obturation techniques is not a predictable procedure. Microleakage, whether from an apical (Dow&Ingle 1955, Strindberg 1956) or a coronal direction (Madison &Wilcox1988), remains a clinical problem and a possible source of failure (Saunders&Saunders1994). Adesirable property of a root-canal sealer, therefore, is to have good sealing ability (Branstetter & von Fraunhofer 1982). In addition, a good sealer must have adhesive strength, both to the dentine and to the core material, which usually is gutta-percha (SpDngberg 1998). The sealer must also have cohesive strength to hold the obturation together. Although good adhesion of the sealer might be expected to improve the sealing ability, such a relationship has not yet been proven.
The sealing ability of endodontic sealers has traditionally been evaluated by leakage tests. Dye leakage (Stewart1958, Ainley1970), bacterial leakage (Goldman et al. 1980) and liquid pressure techniques (Wu et al. 1993) are amongst the most frequently used methods. Endodontic leakage tests may be difficult to standardize (Al-Ghamdi & Wennberg 1994) and the results are difficult to reproduce and compare (Wu & Wesselink 1993).
Only a few studies have attempted to evaluate the adhesive properties of root-canal sealers by measuring their bond strengths (Wennberg & Irstavik 1990, Gettleman et al. 1991, De Gee et al. 1994, Fidel et al. 1994, Lalh et al. 1999a, Pecora et al. 2001, Timpawat et al. 2001). These studies have investigated the effect of smear layer removal on sealer adhesion with controversial results. It has been suggested that the endodontic smear layer acts as a physical barrier interfering with adhesion and penetration of sealers into dentinal tubules, which may affect the sealing efficacy of rootcanal obturation (Sen et al.1995). The majority of these studies used the same dentine pretreatment for different sealers, irrespective of the chemical composition of the sealer. Sealers available to the profession today are zinc oxide-eugenol based, calciumhydroxide based, resin based, glass ionomer based or silicone based. The mechanism of adhesion amongst this wide range of chemical compositions cannot be expected to be exactly the same.
The purpose of this investigation was therefore to study the adhesion of root-canal sealers of different chemical composition to dentine and gutta-percha by tensile bond strength measurements. The effect on adhesion of various dentine pretreatments and the effect of an experimental primer on the adhesion of a silicone-based sealer were also investigated. Another aim was to describe the type of bond failure at the dentine-sealer and the sealer-gutta-percha interfaces for the debonded surfaces.

Materials and methods.
The technique for testing adhesion of the root-canal sealers to dentine and gutta-percha was a modification of a previously described method (Wennberg & Irstavik 1990).
A total of 104 extracted human single-rooted teeth were used in this study. The teeth were stored in 0.01% NaOCl at 4 8C. Before performing the experiments, the teeth were thoroughly rinsed with distilled water. Root dentine cylinders, 4 mm in diameter, were cut in a bucco-lingual direction at a right angle to the tooth’s long axis, using a water-cooled carbide trephine bur. The cylinders were mounted in brass holders using zinc phosphate cement. After setting, the dentine surfaces were ground flat against 500-grit silicone carbide abrasive paper (Struers, Copenhagen, Denmark) under running water. A Planopol grinding machine (Struers, Copenhagen, Denmark) was used to control the angle of the ground surface. The prepared specimens were stored in distilled water at 4 8C. Gutta-percha cylinders of 4 mm diameter were prepared from heat-softened gutta-percha (Roeko, Langenau, Germany).The cylinders were mounted in brass holders and secured mechanically by holes in the holders. Their end surfaces were ground flat in a similar manner to the dentine cylinders.
The dentine specimens were randomly divided into four equal groups and their surfaces conditioned with either 37% H3PO4 for 30 s, 25% citric acid for 30 s, 17% disodium ethylene diamine tetraacetic acid (EDTA) for 5 minor10 mL distilled water (control).The conditioned dentine surfaces were then rinsed with 10 mL distilled water and dried with an air stream for 5 s. The surfaces thus produced were further characterized by scanning electron microscopy. Each group of conditioned specimens was further divided into six equal subgroups (n = 4) according to the type of sealer used.
Five sealers of different chemical compositions were tested (Table 1). An experimental primer supplied with RoekoSeal Automix (RS) was also investigated. The sealers were mixed according to the manufacturer’s instructions. The dentine and gutta-percha surfaces were coated with a thin layer of the freshly mixed sealer and the cylinders were immediately pressed together by means of a spring. To ensure alignment of the dentine and gutta-percha surfaces during the setting of the sealer, the two brass holders with their respective dentine and gutta-percha cylinders were placed in a special device consisting of a plastic block with a semicircular groove of the same diameter as the brass holders (Fig.1). Excess sealer was carefully removed with a cotton pellet. The brass holders were then fixed in that position with a clamp.
The test specimens were kept in an incubator at 37 8C and a relative humidity of 90 _5%. The sealers were allowed to set for1.5 times the manufacturer’s stated setting time. The test specimen, still fixed to the plastic block, was then mounted in a universal testing machine (Instron, Instron Limited, Bucks, UK). After mounting, the plastic block was removed and the test specimen was subjected to a tensile load at a constant cross-head speed of 1mm min_1 (Fig. 2). The force (N) required to rupture the bond was recorded and used to calculate the bond strength (MPa).
After adhesion testing, the fractured surfaces were examined under a stereomicroscope at x25 magnification to describe the nature of bond failure: cohesive within the sealer, adhesive at the sealer-gutta-percha interface and/or adhesive at the sealer-dentine interface (Fig. 3).
Statistical analysis was performed using one-way anova followed by the Bonferroni test for multiple comparisons to compare the mean tensile bond strength of the six sealers at the different dentine pretreatment conditions. Significance was established at the 5% level.

Table 1. Root-canal sealers tested.



Figure 1. A test specimen with gutta-percha (left), sealer (middle) and dentine (right) in plastic holder.



Figure 2. A test specimen mounted in the testing machine.

The mean tensile bond strengths for the various subgroups are illustrated in Fig. 4. Mean tensile bond strengths ranged from 0.07 MPa (Apexit (AP)) to 1.19 MPa (AH Plus (AH)). AH had the highest bond strength amongst the sealers investigated. The differences in bond strength between the other sealers were not statistically significant.
Removal of the smear layer with EDTA significantly reduced, rather than increased, the adhesion of AH (P < 0.01) and of AP (P < 0.05). The adhesion of Roeko-Seal Automix (RS) and Ketac-Endo (KE) was not influenced by dentine pretreatments significantly. The use of a primer with RS significantly enhanced its adhesion incase of untreated dentine (P < 0.001) (Fig. 5).However, no significant difference was found when the primer was applied to dentine pretreated with EDTA or acids. The bond strength of Grossman’s sealer (GS), on the other hand, was significantly increased when dentine had been pretreated with phosphoric acid (P < 0.01).
Inspection of fractured surfaces revealed the bond failure to be mainly adhesive to dentine for GS and RS, and mainly adhesive to gutta-percha for KE. The failure appeared to be predominantly cohesive with in the sealer for AP, with some areas debonding in the sealer-dentine interface. For AH, the failure was mainly cohesive within the sealer (Fig. 3A).When the primer was applied, the failure for RS in case of untreated dentine was adhesive to gutta-percha (Fig. 3C), whilst that to pretreated dentine was adhesive to dentine.

Figure 3. Stereomicrographs of debonded surfaces (magnification x25). Left, gutta-percha surface; right, dentine surface.
(A) Cohesive Failure with sealer left on both surfaces; AH, no pretreatment.
(B) Adhesive failure to dentine; AP, EDTA pretreatment.
(C) Adhesive failure to gutta-percha; RP, no pretreatment.



Figure 4. Mean tensile bond strengths (MPa) of root-canal sealers when applied as a thin layer between gutta-percha and dentine pretreated with different solutions.



Figure 5. Mean tensile bond strengths (MPa) of RS when applied as a thin layer between gutta-percha and dentine pretreated with different solutions.

Discussion.
The ability of root-canal sealers to adhere to dentine and gutta-percha is expected to result in superior sealing ability, which in turn should reduce leakage in clinical situations. Adhesion should also improve the stability of the root filling, e.g. during preparation for post space. Adhesion depends on a multitude of interacting factors including the surface energy of the adherend (dentine or gutta-percha), the surface tension of the adhesive (sealer), the adhesive’s ability to wet the surfaces and the cleanliness of the adherend surface (Bayne 2001).Moreover, stresses caused by differences in thermal expansion coefficients and dimensional changes during setting of the adhesive may affect adhesive bonds. Two mechanisms of adhesion may be distinguished: chemical and mechanical. In case of chemical bonding a smooth surface generally results in better adhesion. Micromechanical bonding, on the other hand, requires the presence of irregularities on the surface of the adherend into which the adhesive can penetrate.
The endodontic smear layer forms over the surface of dentinal walls when the root canals are instrumented (McComb & Smith 1975). The significance of the smear layer in endodontics has been the subject of extensive debate since it was first described. Certain factors would indicate removal of the smear layer prior to root-canal filling. Not only can the smear layer act as a reservoir or substrate for microorganisms (Pashley 1984), but it will also obstruct the extension of sealer tags into the dentinal tubules and thereby decrease adhesion by micromechanical forces (Kouvas et al.1998). On the other hand, it has been argued that removal of the smear layer increases dentine permeability with possible diffusion of noxious substances from the root canal to the external root surface (Galvan et al. 1994). Presence of the smear layer may also discourage bacterial penetration and colonization of the dentinal tubules (Michelich et al. 1980, Drake et al. 1994). Furthermore, the constituents of the smear layer may improve adhesion by altered wet ability and chemical bonding. For such reasons, retaining the smear layer on the root-canal walls has been considered to be beneficial.
No adhesion test has yet been generally accepted. Most previous studies have measured the bond strength of endodontic sealers in bulk and to dentine only. In the present study, tensile bond strength measurement was performed by testing the sealers in a thin layer between dentine and gutta-percha rather than in bulk, as they would be used in a clinical context (Wennberg&Irstavik 1990). Since tensile bond strengths of low magnitude are highly susceptible to transverse forces (Irstavik et al. 1983), special care was used when handling the specimens during mounting in the testing machine to avoid influences from such forces.
All the sealers tested showed measurable adhesive properties. The resin-based sealer (AH) had the highest bond strength amongst the sealers investigated, whilst the calcium hydroxide-based sealer (AP) demonstrated the lowest values. These findings are in general agreement with most other studies (Wennberg & Irstavik 1990, Gettleman et al.1991, Webster et al. 2001).
Phosphoric acid, citric acid and EDTA were used as dentine pretreatments in the present study. The pretreatment of dentine with17% EDTA is known to remove the smear layer, whereas distilled water leaves it intact (Goldman et al.1981, Yamada et al.1983). Both phosphoric and citric acids are similarly known to be effective in removing the smear layer (Tidmarsh1978,Baumgartner et al.1984).Pretreatment of dentinewith35%phosphoric acid or 6% citric acid was found to be more effective than EDTA in removing the smear layer (Timpawat et al. 2001).
In the present study, the effect of smear layer removal on adhesion varied amongst the different sealers investigated. For the resin-based sealer, AH, dentine pretreatment resulted in no improvement in adhesion. A stronger bond might have been expected, as it has been suggested that the exposure of the dentinal tubules creates a much more irregular surface allowing for penetration of resinous tags (Gettleman et al. 1991). However, we found that pretreatment of dentine, with EDTA in particular, significantly decreased the bond strength as reflected in the change of the failure pattern towards debonding at the sealer-dentine interface. Similar findings have previously been reported (Wennberg & Irstavik1990, Webster et al.2001). A possible explanation may be that the weak demineralization by EDTA leaves a relatively smooth surface of the organic dentine structure, which does not offer an increased area for adhesion. By contrast, dentine treated with EDTA showed greater adhesion to the resin-based sealers than dentine that received no treatment in other studies (Gettleman et al. 1991, Pecora et al. 2001). With all pretreatments other than EDTA, inspection with the stereomicroscope revealed the bond failure to be predominantly cohesive within the sealer (Fig. 3A). It has been suggested that the very slow setting of resin-based sealers might allow sufficient time for the development of adhesion to dentine, but that shrinkage stresses might fracture the still weak unset sealer cohesively (DeGee et al.1994).The present findings suggest that improving the cohesive strength of AH may further promote its sealing ability.
Similar to AH, the adhesion of the calcium hydroxide based sealer, AP, was significantly decreased by EDTA treatment. These findings are in disagreement with previous studies (Wennberg & Irstavik 1990, Fidel et al. 1994) where a significant increase in bond strength was reported for calcium hydroxide-based sealers when the dentine surface was pretreated with EDTA. Other studies, however, indicated that the tensile bond strength of calcium hydroxide-based sealers did not significantly increase when the smear layer was removed (Gettleman et al. 1991). AP failed cohesively within the sealer, with some areas debonding in the sealer-dentine interface. When the dentine was pretreated with EDTA, failure became predominantly adhesive to dentine (Fig. 3B), reflecting that smear layer removal reduced adhesion. Debonding at the sealer-dentine interface (Gettleman et al.1991), as well as cohesive failure within the sealer (Wennberg & Irstavik 1990), has been observed for calcium hydroxide-based sealers.
In contrast to adhesion of the majority of sealers, which mostly relyona mechanical type of bonding, glass ionomer cements also bind by chemical interaction (Wilson et al.1983). It is generally accepted that in case of chemical bonding, a smooth surface is desirable for adhesion (McComb&Smith1976).The present investigation supports this suggestion, as the highest bond strength( although not statistically significant) resulted when the smear layer was kept. The mean tensile bond strengths recorded for the glass ionomer-based sealer (KE) ranged from 0.35 to 0.41MPa. A wider range (0.1- 1.4 MPa) has been previously reported (Timpawat et al. 2001). They found pretreatment of dentine with 35% phosphoric acid or6%citric acid to result insignificantly higher bond strength for KE than treatment with EDTA. In the present study, the effect of etching dentine with phosphoric and citric acids could not be evaluated as KE debonded at the gutta-percha interface. The effect of pretreatment with EDTA, however, could be shown by the mode of failure being partly adhesive to dentine. The shear bond strength of glass ionomer-based sealers has been found to be higher with the smear layer present (Lalh et al. 1999a, Webster et al. 2001). The finding that the failure of KE was mainly adhesive to gutta-percha may be attributed to the high setting contraction of the fast setting glass ionomers leading to premature adhesive failure (De Gee et al. 1994). Improving the bond between glass ionomer sealers and gutta-percha may be a step towards achieving a better seal.
The bond strength of the recently introduced silicone (polydimethylsiloxane)-based sealer (RS) has not been previously reported. In the present study, no significant difference in bond strength was found, whether dentine was pretreated or not. However, the application of an experimental primer prior to RS significantly increased its bond strength. This occurred only when the smear layer was not removed. The ability of this primer to enhance the adhesion of RS to untreated dentine was reflected in the change of the pattern of failure of the specimens. The failure changed from a predominantly adhesive failure to dentine to a predominantly adhesive failure to gutta-percha, following application of the primer (Fig. 3C).These results suggest a possible interaction with the smear layer.
For the zinc oxide-eugenol-based sealer, GS, the decrease in bond strength when dentine was pretreated with EDTA was of no statistical significance. The sealer debonded at the dentine interface, whether dentine was pretreated with EDTA or left untreated, as reported in another study (Gettleman et al. 1991). In contrast to the present findings, pretreatment with EDTA was found to result in significant increase intensile bond strength for zinc oxide-eugenol-based sealers (Wennberg & Irstavik1990). In another study (Gettleman et al.1991), the increase in bond strength was not statistically significant. The use of phosphoric or citric acid resulted in a cohesive failure in the sealer. This was reflected in the significant increase in bond strength recorded under these conditions and may be related to chemical effects of the acids, causing precipitation of zincphosphate aggregates at the interface.
The present investigation confirms that the smear layer may play an important role in the bonding of endodontic sealers to dentine (Lalh et al. 1999b).Whether it is beneficial to recreate the smear layer under aseptic conditions prior to root-canal obturation or not remains to be answered. In further studies, scanning electron microscopy and energy dispersive X-ray spectroscopy will be used for analyses of effects of various dentine pretreatments on the smear layer.

References.

Ainley JE (1970) Fluorometric assay of the apical seal of rootcanal fillings. Oral Surgery, Oral Medicine and Oral Pathology 29, 753-62.
Al-Ghamdi A, Wennberg A (1994) Testing of sealing ability of endodontic filling materials. Endodontics and Dental Traumatology10, 249-55.
Baumgartner JC, Brown CM, Mader CL, Peters DD, Shulman JD (1984) A scanning electron microscopic evaluation of rootcanal debridement using saline, sodium hypochlorite, and citric acid. Journal of Endodontics 10, 525-31.
Bayne S (2001) Bonding to dental substrates. In: Craig RG, Powers JM, eds. Restorative Dental Materials, 11the dn. St. Louis: Mosby Inc., pp. 260-2.
Branstetter J, von Fraunhofer JA (1982) The physical properties and sealing   action of endodontic sealer cements: a review of the literature. Journal   of Endodontics 8, 312-6.
De Gee AJ, Wu M-K, Wesselink PR (1994) Sealing properties of Ketac-Endo glass   ionomer cement and AH26 root-canal sealers. International Endodontic Journal   27, 239-44.
Dow PR, Ingle JI (1955) Isotope determination of root-canal failure. Oral Surgery, Oral Medicine and Oral Pathology 8, 1100-4.
Drake DR, Wiemann AH, Rivera EM, Walton RE (1994) Bacterial retention in   canal walls in vitro: effect of smear layer. Journal of Endodontics 20,   78-82.
Fidel RA, Sousa N, Spano JC, Barbin EL, Pecora JD (1994) Adhesion of calcium   hydroxide-containing root-canal sealers. Brazilian Dental Journal 5, 53-7.
Galvan DA, Ciarlone AE, Pashley DH, Kulild JC, Primack PD, SimpsonMD (1994)   Effect of smear layer removal on the diffusion permeability of human roots. Journal of Endodontics 20, 83-6.
Gettleman BH, Messer HH, ElDeebME (1991) Adhesion of sealer cements to dentin   with and without the smear layer. Journal of Endodontics17, 15-20.
Goldman LB, Goldman M, Kronman JH, Letourneau JM (1980) Adaptation and porosity   of poly-HEMA in a model system using two microorganisms. Journal of Endodontics   6, 683-6.
Goldman LB, Goldman M, Kronman JH, Lin PS (1981) The efficacy of several   irrigating solutions for endodontics: a scanning electron microscopic study. Oral Surgery, Oral Medicine and Oral Pathology 52, 197-204.
Kouvas V, Liolios E,Vassiliadis L, Parissis-Messimeris S, Boutsioukis A (1998)   Influence of smear layer on depth of penetration of three endodontic sealers:   an SEM study. Endodonticsan d Dental Traumatology14, 191-5.
Lalh MS, Titley K, Torneck CD, Friedman S (1999a) The shear bond strength   of glass ionomer cement sealers to bovine dentine conditioned with common endodontic   irrigants. International Endodontic Journal 32, 430-5.
LalhM S, Titley KC,Torneck CD, Friedman S (1999b) Scanning electron microscopic   study of the interface of glass ionomer cement sealers and conditioned bovine   dentin. Journal of Endodontics 25, 743-6.
Madison S, Wilcox LR(1988) An evaluation of coronal microleakage in endodontically   treated teeth. Part III. In vivo study. Journal of Endodontics14, 455-8.
McComb D, SmithD C (1975) A preliminary scanning electron microscopic study   of root canals after endodontic procedures. Journal of Endodontics 1, 238-42.
McComb D, Smith D C (1976) Comparison of physical properties of polycarboxylate-based   and conventional root-canal sealers. Journal of Endodontics 2, 228-35.
Michelich VJ, Schuster GS, Pashley DH (1980) Bacterial penetration of human   dentin in vitro. Journal of Dental Research 59, 1398-403.
Irstavik D, Eriksen HM, Beyer-Olsen EM (1983) Adhesive properties and leakage   of root-canal sealers in vitro. International Endodontic Journal16, 59-63.
Pashley DH (1984) Smear layer: physiological considerations. Operative Dentistry   3 (Suppl.),1 3-29.
Pecora JD, Cussioli AL, Guerisoli DM, Marchesan MA, Sousa- Neto MD, Brugnera   JA (2001) Evaluation of Er: YAG laser and EDTAC on dentin adhesion of six   endodontic sealers. Brazilian Dental Journal 12, 27-30.
Saunders WP, Saunders EM(1994) Coronal leakage as a cause of failure in root-canal   therapy: a review. Endodontics and Dental Traumatology 10, 105-8.
Sen BH, Wesselink PR, Turkun M (1995) The smear layer: a phenomenon in root-canal   therapy. International Endodontic Journal 28, 141-8.
SpDngberg LSW (1998) Instruments, materials and devices. In: Cohen S,   Burns RC, eds. Pathways of the Pulp, 7th edn. St.Louis: Mosby Inc., pp.510.
Stewart GG (1958) A comparative study of three root-canal sealing agents. Oral Surgery, Oral Medicine and Oral Pathology 11, 1029-41.
Strindberg LZ (1956) The dependence of the results of pulp therapy on certain   factors. Acta Odontologica Scandinavica 14, 1-175.
Tidmarsh B G (1978) Acid-cleansed and resin-sealed root canals. Journal   of Endodontics 4, 117-21.
Timpawat S, Harnirattisai C, Senawongs P (2001) Adhesion of a glass-ionomer   root-canal sealer to the root-canal wall. Journal of Endodontics 27, 168-71.
Webster CN, Goldbeck AP, Khajotia SS, Duncanson MG (2001) Bond strengths   of endodontic sealers to dentin and guttapercha. Journal of Dental Research   80, 258 (AADR abstract 1778).
Wennberg A, Irstavik D(1990) Adhesion of root-canal sealers to bovine dentine   and gutta-percha. International Endodontic Journal 23, 13-9.
Wilson AD, Prosser HJ, Powis DM (1983) Mechanism of adhesion of polyelectrolyte   cements to hydroxyapatite. Journal of Dental Research 62, 590-2.
Wu M-K, De Gee AJ, Wesselink PR, Moorer WR (1993) Fluid transport and bacterial   penetration along root-canal fillings. International Endodontic Journal   26, 203-8.
Wu M-K, Wesselink PR (1993) Endodontic leakage studies reconsidered. Part   I. Methodology, application and relevance. International Endodontic Journal   26, 37-43.
Yamada RS, Armas A, Goldman M, Lin PS (1983) A scanning electronmicroscopic   comparison of a high volume final flush with several irrigating solutions. Part   3. Journal of Endodontics 9, 137-42.