Bacterial microleakage and pulp inflammation associated with various restorative materials

dental materials Dental Materials 18 (2002) 470±478

www.elsevier.com/locate/dental

Bacterial microleakage and pulp in¯ammation associated with various restorative materials Peter E. Murray a,*, Abeer A. Hafez b, Anthony J. Smith a, Charles F. Cox c a

Department of Oral Biology, The University of Birmingham, School of Dentistry, St Chad's Queensway, Birmingham B4 6NN, UK b School of Dentistry, University of Southern California, Los Angeles, CA, USA c Department of Restorative Dentistry, The University School of Dentistry, Los Angeles, CA 90033, USA Received 20 September 2000; revised 24 April 2001; accepted 19 June 2001

Abstract Objectives: Many restorative materials are claimed to be successful in preventing bacterial microleakage and minimizing pulp in¯ammation. However, information regarding the in vivo performance of materials in comparison with each other is limited. The aim of this study was to evaluate and compare the pulp response of nine restorative materials when placed in non-exposed monkey cavities. Methods: 279 standardized non-exposed Class V cavities, were prepared into buccal dentin. Cavities were restored with a number of materials in the following categories: Zinc oxide eugenol (ZnOE), Calcium hydroxide [Ca(OH)2], zinc phosphate (ZP), Resin-modi®ed glass ionomer (RMGI), Composite resin (CR), Bonded amalgam (BA), Gutta-percha (GP), Compomer and Silicate. Pulp tissues were collected and evaluated at short, intermediate and long-term intervals according to ISO guidelines; employing histomorphometric analysis, Spearman's rho and ANOVA statistics. Pulp responses were categorized according to FDI, ISO and ADA standards. Bacteria were detected using McKay stains. Results: Pulp in¯ammation was found to be correlated to bacterial microleakage around the restoration (p #0.0001). The frequency of bacterial microleakage was found to vary between restorative materials (p #0.0001). In rank order of preventing bacterial microleakage from best to the worst; RMGI (100%), BA (88%), ZnOE (86%), CR (80%), GP (64%), Ca(OH)2 (58%), compomer (42%), silicate (36%) and ZP (0%). Signi®cance: The most effective restorative materials to prevent bacterial microleakage and pulp injury from in¯ammatory activity were RMGI, BA, ZnOE and CR restorations. q 2002 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. Keywords: Pulp; Microleakage; In¯ammation; Adhesives; Composite resin; Calcium hydroxide; Resin-modi®ed glass ionomer; Silicate; Zinc oxide eugenol; Bonded amalgam

1. Introduction A number of clinical restorations fail, because vital teeth develop hypersensitivity and recurrent pulp in¯ammation symptoms [1]. The clinical skills necessary to prepare and restore cavities to a high technical quality are of extreme importance in reducing these symptoms [2]. However, several surveys have shown how the selection and placement of one type of restorative material in preference to another can make the critical difference between the failure and success of treatment in the long term [3,4]. Various reasons may account for the failure of restorations, such as tooth fracture, marginal fracture and degradation by abra* Corresponding author. Tel.: 144-121-237-2882; fax: 144-121-6258815. E-mail address: [email protected] (P.E. Murray).

sion, attrition, erosion and non-carious defects, as well as patients' dental characteristics, diet and oral hygiene [5,6]. Nevertheless, the most frequent [7,8], and potentially serious postoperative complications can emanate from bacterial microleakage along the restoration interface. Microleakage complications include postoperative sensitivity, marginal discoloration, recurrent caries, pulp in¯ammation, pulp necrosis, periodontal disease, and eventual need for endodontic therapy [9,10]. Overwhelming research data have shown that the microleakage of bacteria is the predominant source of pulp in¯ammation [11]. For over a hundred years, presence of bacteria and their products have been implicated as a prerequisite for induction of the most severe forms of in¯ammatory activity [9]. Severe in¯ammation can often progress to pulp necrosis and periapical lesion development with local bone destruction [12]. Consequently, it is important to ensure that potential sources

0109-5641/02/$20.00 + 0.00 q 2002 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. PII: S 0109-564 1(01)00072-0

P.E. Murray et al. / Dental Materials 18 (2002) 470±478

471

Table 1 Restorative materials No.

Restorative material

1. 3. 6. 7 8. 2. 4. 5. 9. 10. 11. 12. 13. 14. 15. 16.

Composite resin Composite resin Composite resin Composite resin Composite resin Composite resin Composite resin Composite resin Composite resin Compomer Compomer Compomer Zinc oxide eugenol Zinc oxide eugenol Calcium hydroxide/amalgam Bonded amalgam

17. 18. 19. 20.

Silicate Gutta-percha Zinc phosphate Resin modi®ed glass ionomer

a b c d e f g h i j k

Brand name

Testing interval (days)

All bond 24% BPDM a All bondÐwet a All bondÐDry a Liner bond primer 4% 1 2% bond b Liner bond 4% primer 1 bond b 4 META EPIC c Optibond d Concise e Scotchbond 2 e SCA compoglass f HSCA compoglass f PSA Primer g 1Dyract h Zinc oxide eugenol i Intermediate restoration material i Dycal d Scotchbond multi-purpose plus e 1 lumicon j MQ k Gutta-percha i Tenacin k Vitremer 1 vitrebond e

Total teeth

3±21

22±56

57±172

2 6 7 0 0 14 7 8 2 0 0 0 9 4 6 12

2 9 8 0 0 13 5 9 4 0 0 4 7 1 6 12

2 9 10 4 3 9 7 0 6 0 10 9 0 0 7 0

6 24 25 4 3 36 19 17 12 10 10 13 16 5 19 24

6 6 0 0

5 5 7 0

0 0 0 7

11 11 7 7

Bisco, Schaumburg, IL, USA. Kurary Co., Osaka, Japan. Amalgambond, Parkell. Kerr, Orange, CA, USA. 3M Dental Products, St. Paul, MN, USA. Ivoclar Vivadent, Amherst, NY, USA. Amherst, NY, USA. De Trey Dentsply, Konstanz, Germany. L.D Caulk Division, Dentsply Int., Milford, DE, USA. Bayer Dental, Leverkusen, Germany. SS White, Lakewood, NJ, USA.

of injury, which stimulate in¯ammatory reactions, are always minimized. Although there is a clear causative relationship between pulp infection, in¯ammation, and loss of pulp vitality, the question still remains regarding possible differences between the capacity of various restorative materials to prevent bacterial microleakage: such as composite resin (CR) materials which have gained in popularity due to their aesthetic properties, improved longevity, wear resistance, color stability, biocompatibility and anti-microleakage characteristics [13,14]. Resin-modi®ed glass ionomer (RMGI) materials have also gained favor because of their excellent adhesion with tooth surfaces, and ability to inhibit recurrent caries by ¯uoride release which promotes remineralization [15,16]. Compomers combine the major bene®ts of RMGI, with the ease of handling of light-cured CR [17]. Zinc oxide eugenol (ZnOE) has excellent biological sealing characteristics, and an antimicrobial activity due to the effect of eugenol. However, it is unsuitable for use as a long-term luting agent because of its poor physical properties in terms of solubility, setting shrinkage and weakness of tensile

strength [18]. Calcium hydroxide [Ca(OH)2] has a high pH which provides local antibacterial activity and repair activity by promoting dentin formation [19]. However, the therapeutic advantages of Ca(OH)2 appear short-term, as it often fails to provide a permanent barrier to bacterial invasion, allowing pulp in¯ammation and eventual necrosis [20,21]. Zinc phosphate (ZP) is the oldest cementation agent, and in common with silicate, has been associated with pulp in¯ammation in the presence of bacteria [22± 24]. Gutta-percha (GP) remains the most widely used root canal ®lling material due to its well-known lack of cytotoxicity to oral tissues [25]. Nevertheless, despite these advances in our knowledge of restorative-biomaterial interactions, the National Institute of Health estimate that each year in the United States, approximately 200 million ®llings are replaced, in addition to ®rsttime placement of 90 million new ®llings [26]. Some European surveys similarly report that approximately 60% of all restorative treatment involves replacement of existing restorations [5,8]. Clearly, there remains a potential to further increase the long-term success of restorative treatments. This may be achievable by improving our

472

P.E. Murray et al. / Dental Materials 18 (2002) 470±478

Fig. 1. Reactionary dentin secretion beneath the site of cavity preparation.

understanding of the interactions between restorative materials, pulp injury and pulp in¯ammation, together with the incidence, presence or absence of bacterial microleakage. Although many disparate claims are made for the post-operative effects of commonly used restorative materials, few direct comparative data are available. Consequently, the purpose of this study is to address these issues by evaluating and comparing pulpal responses beneath 279 cavities restored with nine restorative materials.

2. Materials and methods 2.1. Specimen preparation Ten adult Rhesus macaca monkeys, approximately 4± 5 years old, each with a complete dentition, were used to study a total of 279 vital teeth. The standardized methods and procedures used in this study have been described elsewhere [10,18,20,27,28]. Brie¯y, each monkey was tranquilized with an intramuscular injection of 10 mg/kg of ketamine hydrochloride (100 mg/ml) (Parke Davis, Rochester, MI, USA). Deeper sedation was with an intramuscular injection of 20 mg/ml of Rompun (Bayer, Pittsburgh, PA, USA). Teeth were cleansed and polished with pumice paste to remove plaque and calculus prior to the operative procedures. Quadrants of teeth were isolated with sterile cotton rolls, and saliva controlled with high-speed evacuation. Class V cavities were prepared with a #330 carbide bur at ultra-high speed under water spray coolant. A new bur was employed on every fourth cavity. Accidental pulp exposure was avoided by checking the thickness of remaining dentin using the UAB-modi®ed Endocater Apex Locator (Hygienic Corp, Akron, OH, USA). Cavities were prepared into dentin leaving a range of remaining dentin thickness (RDT) between 0.011 and 1.862 mm with a mean of

0.579 mm. Class V cavities were restored with various restorative materials according to the manufacturers recommendations (Table 1), as described previously [10,18,20,27,28]. In total, 20 different regimes were used, after postoperative intervals of between 3 and 172 days depending on International Oraganization for Standardization (ISO) testing guidelines, the restored teeth were collected by perfusion ®xation, demineralized, processed and evaluated according to current usage guidelines of the ISO Technical Report (ISO 7405), American Dental Association (ADA) and FeÂdeÂration Dentaire Internationale (FDI). Serial sections were examined by light microscopy, and histometric analysis was conducted as described previously [2,29±31]. Brie¯y, 7 mm serial sections were stained with hematoxylin and eosin, the area of reactionary dentin was estimated histomorphometrically at £ 100 magni®cation using a grid eyepiece graticule. In addition, the RDT of each cavity was also measured. The in¯ammatory response of each pulp was categorized and ranked in order of severity (0, 1, 2, 3 scale) from `absent', `slight', and `moderate' to `severe' pulp, according to published criteria [10,18,20, 27,28]. If `absent,' the tooth pulp contained few in¯ammatory cells, or an absence of in¯ammatory cells associated with cut tubules of the cavity ¯oor; `slight,' the tooth pulp showed some in®ltration of in¯ammatory cells; `moderate,' the tooth pulp had localized polymorphonuclear leukocyte or mononuclear lymphocyte lesions; `severe,' mononuclear lymphocyte and polymorphonuclear leukocyte lesions involved more than one third of the coronal pulp. `Necrosis,' was identi®ed by the localized destruction of pulp cell populations. Bacterial contamination of each restoration was assessed using McKay's stain [32] to detect for the presence of Gram positive and negative microorganisms. Masson trichrome staining was used to assess soft tissue organization and reactionary (reparative) dentin formation. Parametric data were analyzed statistically using Analysis of Variance (ANOVA) tests for multiple comparisons among the means, and nonparametric data were analyzed using Spearman's rho rank correlation tests, at a con®dence level of 95% (STATview software, SAS Inc., Cary, NC, USA).

3. Results 3.1. Identi®cation of reactionary dentin The presence of a tertiary (reparative) dentin matrix was observed beneath 62.8% of the cavities examined within this study. In all cases, a tubular continuity was observed between the normal secondary dentin matrix and the preexisting odontoblasts (Fig. 1). These observations classify the secreted dentin matrix as being reactionary (reparative) in origin [33].

2 bacteria

0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 71 (n ˆ 5) 50 (n ˆ 2) 60 (n ˆ 3) 100 (n ˆ 1) 50 (n ˆ 3) 100 (n ˆ 1) 0 (n ˆ 0) 100 (n ˆ 1) 0 (n ˆ 0) 29 (n ˆ 2) 50 (n ˆ 2) 40 (n ˆ 2) 0 (n ˆ 0) 50 (n ˆ 3) 0 (n ˆ 0) 0 (n ˆ 0) 29 (n ˆ 2) 0 (n ˆ 0) 47 (n ˆ 7) 0 (n ˆ 0) 75 (n ˆ 3) 0 (n ˆ 0) 75 (n ˆ 3) 100 (n ˆ 5) 0 (n ˆ 0) 71 (n ˆ 5) 100 (n ˆ 1) 53 (n ˆ 8) 100 (n ˆ 1) 25 (n ˆ 1) 100 (n ˆ 8) 25 (n ˆ 1) 0 (n ˆ 0) 0 (n ˆ 0) 6 (n ˆ 1) 20 (n ˆ 3) 12 (n ˆ 14) 20 (n ˆ 1) 11 (n ˆ 1) 75 (n ˆ 18) 0 (n ˆ 0) 100 (n ˆ 1) 100 (n ˆ 7) 94 (n ˆ 15) 85 (n ˆ 17) 88 (n ˆ 107) 80 (n ˆ 4) 89 (n ˆ 8) 25 (n ˆ 6) 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 12 (n ˆ 3) 14 (n ˆ 3) 20 (n ˆ 29) 36 (n ˆ 4) 42 (n ˆ 8) 58 (n ˆ 19) 64 (n ˆ 7) 100 (n ˆ 7)

Slight

1 bacteria

100 (n ˆ 7) 88 (n ˆ 21) 86 (n ˆ 18) 80 (n ˆ 117) 64 (n ˆ 7) 58 (n ˆ 11) 42 (n ˆ 14) 36 (n ˆ 4) 0 (n ˆ 0)

Stained bacteria were identi®ed in 28.67% of the 277 restored cavities with McKays stain [32] (Fig. 4). The

Resin-modi®ed glass ionomer Bonded amalgam Zinc oxide eugenol Resin composite Gutta-percha Calcium hydroxide Compomer Silicate Zinc phosphate

3.4. Bacterial microleakage

Absent

The combined ANOVA statistic for the effects of restorative materials on pulp in¯ammation was p # 0.0001 (Table 2), suggesting that the selection of restorative material, and its capacity to prevent bacterial microleakage was more important than other cavity preparation and restoration variables in causing pulp in¯ammation (Table 2). Certain restorative materials sealed the cavities more completely than others, preventing bacterial microleakage; the bacteriometic range of the restorative materials for prevention of bacterial microleakage was from 100% to 0% (Table 3). In rank order of the ability to prevent bacterial microleakage, from the best to worst was RMGI, bonded amalgam (BA), ZnOE, CR, GP, Ca(OH)2, compomer, silicate and ZP (Table 3). Some materials appeared to be correlated with pulp in¯ammation activity (Fig. 2), even in the absence of bacterial microleakage, particularly silicate (Fig. 3). However, the severity of pulp in¯ammation activity always increased once bacteria had in®ltrated the cavity preparation (Fig. 3). In rank order of pulp in¯ammatory activity, from the least to the most severe was RMGI, ZnOE, compomer, CR, ZP, BA, Ca(OH)2, GP and silicate (Fig. 3).

Total In¯ammatory activity

3.3. Restorative materials

Table 3 Percentage (numbers) of cavities showing different grades of in¯ammation in the absence and presence of bacteria

The range of RDTs between the various restorative materials was 0.459±0.917 mm. GP had the lowest mean RDT of 0.459 mm (standard deviation 0.262 mm) and ZP had the greatest mean RDT of 0.917 mm (standard deviation 0.395 mm). The mean RDT of all the cavity restoration materials combined was 0.579 mm (standard deviation 0.360 mm). No statistically signi®cant differences were detected at the p # 0.05 signi®cance level.

1 bacteria

3.2. Cavity remaining dentin thickness

2 bacteria

1 bacteria

0.0409 0.0733 0.2366 0.2980 0.3463 0.5034 0.7831 0.9000

Moderate

ANOVA ANOVA ANOVA ANOVA ANOVA ANOVA ANOVA ANOVA

2 bacteria

0.0001 0.0001 0.0117

2 bacteria

Bacterial microleakage Cavity restoration material Time elapsed since material placement Cavity ¯oor area Cavity ¯oor width Remaining dentin thickness Cavity volume Cavity surface area Reactionary dentin secretion Cavity wall depth Cavity wall area

1 bacteria

Spearman's rho Spearman's rho ANOVA

2 bacteria

Correlation with in¯ammation (P value)

Material

Variable measured

Severe

Test type

1 bacteria

Table 2 Hierarchy of variables correlated to pulpal in¯ammatory activity

473 0 (n ˆ 0) 0 (n ˆ 0) 0 (n ˆ 0) 100 (n ˆ 3) 100 (n ˆ 1) 100 (n ˆ 1) 0 (n ˆ 0) 100 (n ˆ 1) 0 (n ˆ 0)

P.E. Murray et al. / Dental Materials 18 (2002) 470±478

474

P.E. Murray et al. / Dental Materials 18 (2002) 470±478

Fig. 2. In¯ammation of pulp.

Fig. 4. Bacterial microleakage through dentinal tubules at cavity margin.

microleakage of bacteria into cavity preparations appeared to be the most important variable in relation to pulp in¯ammatory activity (ANOVA p # 0.0001) (Table 2). In the absence of bacterial microleakage, pulp in¯ammatory activity was minimal (Fig. 5). Once bacteria had proliferated into the cavosurface space, pulp in¯ammation activity reached a moderate level (Fig. 5). In cases, where bacteria progressed further into the cavity margins, infecting the cut dentinal tubules of the cavity ¯oor, pulp in¯ammatory activity

reached a more severe level (Fig. 5). The progression of bacteria into the pulp chamber further increased the severity of pulp in¯ammatory activity (Fig. 5). Necrosis of pulp tissue (Fig. 6) was only observed in the presence of bacterial microleakage (Table 3). 3.5. Time elapsed since material placement The severity of pulp in¯ammatory response appeared to

Fig. 3. Cavity restoration materials and pulp in¯ammation.

P.E. Murray et al. / Dental Materials 18 (2002) 470±478

475

Fig. 5. Bacterial microleakage and pulp in¯ammation.

be time dependent (ANOVA p # 0.0117) (Table 2). The level of pulp in¯ammation decreased over time, beneath cavities infected with bacteria, and also, beneath cavities in which no bacterial microleakage was detected (Fig. 7). In the absence of bacterial microleakage, very low levels of pulp in¯ammation were detected, and no in¯ammation was detected after 81 days had elapsed (Fig. 7). Higher levels of pulp in¯ammation were observed beneath cavities infected with bacteria, but the severity of in¯ammation reduced substantially after 57 days had elapsed in the absence and presence of bacteria (Fig. 7). 3.6. Cavity preparation variables The only cavity preparation variable observed to be correlated to pulp in¯ammation at the p ˆ 0.05 signi®cance level, was the cavity ¯oor area (ANOVA p # 0.0409) (Table 2).

Fig. 6. Necrosis of pulp.

The ¯oor width, volume, surface area, remaining dentin thickness and wall area of each cavity were not found to be correlated with pulp in¯ammation at the p # 0.05 signi®cance level (Table 2). Increases in the cavity ¯oor area of infected cavity preparations appeared to increase the severity of pulp in¯ammation. However, in the absence of infection, the cavity ¯oor area appeared to be a less important mediator of pulp in¯ammation (ANOVA p # 0.2040). 3.7. Reactionary (reparative) dentin activity Reactionary (reparative) dentin deposition following cavity preparation was not related to pulp in¯ammation, as there were no statistically signi®cant correlation observed at the p # 0.05 signi®cance level (Table 2). 4. Discussion The understanding that pulp in¯ammation and subsequent necrosis are largely a result of the microleakage of bacteria, has led to developments in restorative dentistry which have bene®ted patients over many years. Human postoperative pulpal repair activity [2,29] and pulp cell survival [30,31] have recently been characterized, and it is now understood how issues of technique sensitivity, cavity preparation and restorative material variables interact to permit pulp pathology. Although some evidence is available about pulp responses to medications [34] silicate [35], ZP [36] and glass ionomer cements [37], more complete evidence is lacking in terms of in¯ammatory activity, to explain how cavity preparation and restoration variables interact to cause pulp pathology in the presence and absence of bacteria. Hence our attempt to correlate the variables of cavity preparation with the placement of CR, compomer,

476

P.E. Murray et al. / Dental Materials 18 (2002) 470±478

Fig. 7. Time elapsed since material placement and pulp in¯ammation.

RMGI, BA, ZP, GP, Ca(OH)2, silicate and ZnOE, using morphometric and analytical evaluation. Examining pulp tissue in¯ammation can be problematic, because it is a dynamic activity, which can either subside or increase in severity over time. To overcome this dif®culty, 277 restored teeth were examined between 3 and 172 days, according to current ISO, ADA and FDI guidelines. Cavities were prepared in caries-free teeth, to avoid the possibility of including residual caries effects within the data. Several mechanisms have been postulated for the stimulation of severe pulp in¯ammatory activity. It has previously been alleged that factors such as low material pH [38], supposed chemical cytotoxicity [39] and the acid etching of vital dentin [40] may cause pulp in¯ammation. However, in accord with previous investigations [9,41], higher levels of in¯ammation were associated with the microleakage of bacteria, in comparison with the absence of bacteria. In the absence of bacteria, low level in¯ammatory activity can be attributed to the effects of cavity preparation trauma, and the release of in¯ammatory mediators from sensory nerve ®bers [42], as well as the pulp response to the chemical activity of the restoration materials [43]. Most of the restorative materials tested appear to mediate very low levels of in¯ammation, except silicate, which appears to promote moderate levels of in¯ammatory activity both in the absence and presence of bacterial microleakage. Certain previous investigations have attributed pulp in¯ammation to the phosphoric acid component of silicate [23,44]. Alternatively, the absence of stained bacterial pro®les may be explained by differences in technical processing and staining procedures [9]. The selection of a restorative material, based on its capacity to prevent microleakage is postulated as the most important property in mediating pulp in¯ammation (Table 2). The ability of RMGI, and ZnOE to prevent microleakage, may

be primarily attributed to the antibacterial activity of ¯uoride and eugenol release, and the bacteriometic seal of the material along the cavity walls [9,27]. The microleakage performance of BA and CR restorations can be attributed to their sealing and adhesion with the cavity walls, and an absence of antibacterial activity [21]. This ®nding is in agreement with reports suggesting that new CR products are superior to their predecessors, having improved sealing properties to prevent bacterial microleakage [5,45]. The higher frequency of GP, Ca(OH)2 compomer, silicate and ZP restorations which became infected, can be attributed to their inability to provide a complete seal along the cavity walls, in addition to their lack of antibacterial activity, solubility, and weakness of adhesion to tooth structure [10,21]. The rank order of pulp in¯ammatory activity mediated by restorative materials appears to be almost identical, in the presence and absence of microleakage (Fig. 1). This is because materials mediating lower levels of in¯ammatory activity in the absence of bacteria, also appear to mediate lower levels of in¯ammatory activity when bacterial microleakage is present. This relationship may be explained by the effect of variations in possible chemical cytotoxic activity of restorative materials on pulp tissue. Higher levels of in¯ammatory activity were associated with Ca(OH)2 than with CR restorations. Conversely, the CR One-step was less cytotoxic to pulp tissue than Dycal Ca(OH)2 [46]. Consequently, differences in the chemical cytotoxic injury to pulp tissue caused by ZnOE [47], RMGI [27], CR [48], compomer [17], Ca(OH)2 [21], silicate [23], ZP [22], GP [25] and BA [13] appear to be partially related to the observations of different levels of in¯ammation. Once a restoration becomes infected, the level of in¯ammation appears to be partially mediated by the degree to which the bacterial microleakage has progressed through the cavity towards the pulp. Bacterial microleakage between

P.E. Murray et al. / Dental Materials 18 (2002) 470±478

the restorative material and cavity walls appears to mediate a low level of in¯ammatory activity. Further microleakage progression into cut dentinal tubules of the cavity ¯oor, mediates higher levels of in¯ammatory activity. The most severe in¯ammatory activity is observed once bacteria have penetrated into the pulp chamber. This correlation between the progression of microleakage and stimulation of in¯ammation, may also partially explain why RMGI, ZnOE, compomer and CR stimulated lower levels of in¯ammation when infected. These materials seal the cavity walls more effectively [9,21,27], making it more dif®cult for microleakage to progress through the cavity restorations, irritate pulp tissue and stimulate higher levels of in¯ammation. Conversely, materials, which poorly seal the cavity walls, may allow a faster progression of microleakage, hence, an increased probability of bacteria reaching the cut dentinal tubules and pulp chamber. This explains our observations of higher levels of in¯ammation associated with ZP, silicate, Ca(OH)2, and GP restorations. The accessibility of restorations to the progression of bacterial microleakage, may also explain why cavity preparations with larger cavity ¯oor areas have a tendency to mediate greater in¯ammatory activity. Increases in the area of the cavity ¯oor, will increase the number of transected dentinal tubules beneath the restoration. Consequently, the available routes for microleakage of bacteria and their products towards the pulp tissue would increase. In accord with investigations of pulp tissue reactions following Class V restoration in patients [30,31], other variables, such as cavity volume, cavity surface area, and cavity wall area, appear to have comparatively less effect on pulp activity. Similarly, pulp in¯ammation does not appear to be correlated with the area of reactionary (reparative) dentin in patients [2,29]. The severity of pulp in¯ammation beneath infected and non-infected cavity restorations appears to decrease 57 days after material placement. In the absence of microleakage, reduction in levels of pulp in¯ammation are observed, suggesting that in¯ammation in response to cavity preparation and restoration events are transient, and can be almost completely resolved. Conversely, pulp in¯ammation mediated by bacterial microleakage, was observed to persist until 172 days following placement, although after 57 days it was substantially reduced. This suggests that bacterial microleakage stimulates more chronic forms of in¯ammation. Similar patterns of reduction in in¯ammatory reactions, over time, have been observed in the ferret [23]. In monkeys [28] and humans after 28 days [2,30], reparative processes will commence, including the deposition of reactionary (reparative) dentin at the pulpal end of affected tubules and/or the deposition of sclerotic dentin, or the blockage of tubules by calci®c deposits [49]. As a consequence of these dentin-pulp repair events, the presence of irritants, including microleakage appear less likely to stimulate pulpal in¯ammation. This study provides compelling evidence to encourage

477

further research into re-evaluating the traditional daily use of ZP and Ca(OH)2 for cavity restorations. Our data show these materials and silicate are not as effective as RMGI, BA, ZnOE, and CR restorations in preventing bacterial microleakage, and in doing so, minimizing pulpal in¯ammation. Although the level of in¯ammatory activity mediated by compomer was relatively low, the relatively high percentage of restorations exhibiting microleakage, suggests that more research is required to optimize the placement technique sensitivity issues of these materials, to match the in vivo performance of RMGI. It would seem that bacterial microleakage, in¯ammation and pulp necrosis can be largely avoided, and controlled, provided the most appropriate cavity restorative materials and placement techniques are used.

References [1] ZoÈllner A, Gaengler P. Pulp reactions to different preparation techniques on teeth exhibiting periodontal disease. Journal of Oral Rehabilitation 2000;27:93±102. [2] Murray PE, About I, Lumley PJ, Smith G, Franquin J-C, Smith AJ. Postoperative pulpal and repair responses. Journal of the American Dental Association 2000;131:321±9. [3] Maryniuk GA, Kaplan SH. Longevity of restorations: survey results of dentists' estimates and attitudes. Journal of the American Dental Association 1986;112:39±46. [4] Mertz-Fairhurst EJ, Curtis JW, Ergle JW, Rueggeberg FA, Adair SM. Ultraconservative and cariostatic sealed restorations: results at year 10. Journal of the American Dental Association 1998;129:55±66. [5] Qvist V, Thylstrup A, MjoÈr IA. Restorative treatment pattern and longevity of resin restorations in Denmark. Acta Odontologica Scandinavica 1986;112:39±45. [6] MjoÈr IA, Jokstad A, Qvist V. Longevity of posterior restorations. International Dental Journal 1990;40:11±7. [7] Maupome G, Sheiham A. Criteria for restoration replacement and restoration life-span estimates in an educational environment. Journal of Oral Rehabilitation 1998;25:896±901. [8] Burke FJ, Cheung SW, MjoÈr IA, Wilson NH. Reasons for the placement and replacement of restorations in vocational training practices. Primary Dental Care 1999;6:17±20. [9] Bergenholtz G, Cox CF, Loesche WJ, Syed SA. Bacterial leakage around dental restorations: its effect on the dental pulp. Journal of Oral Pathology 1982;11:439±50. [10] Cox CF. Microleakage related to restorative procedures. Proceedings of the Finnish Dental Society 1992;88(Suppl. 1):83±93. [11] Douglas WH. Clinical status of dentine bonding agents. Journal of Dentistry 1989;17:209±15. [12] Bergenholtz G. Pathogenic mechanisms in pulpal disease. Journal of Endodontics 1990;16:98±101. [13] Tarim B, Suzuki S, Suzuki SH, Cox CF. Marginal integrity of bonded amalgam restorations. American Journal of Dentistry 1996;9:72±6. [14] Suh BI, Cincione FA. All-Bond 2: the fourth generation bonding system. Esthetic Dentistry Update 1992;5:61±6. [15] Donly KJ, Grandgenett C. Dentin demineralization inhibition at restoration margins of Vitremer. Dyract and Compoglass. American Journal of Dentistry 1998;11:245±8. [16] Donly KJ, Segura A, Kanellis M, Erickson RL. Clinical performance and caries inhibition of resin-modi®ed glass ionomer cement and amalgam restorations. Journal of the American Dental Association 1999;130:1459±65. [17] Tarim B, Hafez AA, Suzuki SH, Suzuki S, Cox CF. Biocompatibility

478

[18] [19] [20]

[21] [22] [23] [24] [25] [26] [27] [28]

[29] [30] [31] [32] [33]

P.E. Murray et al. / Dental Materials 18 (2002) 470±478 of compomer restorative systems on nonexposed dental pulps of primate teeth. Operative Dentistry 1997;22:149±58. Cox CF, Keall CL, Keall HJ, Ostro E, Bergenholtz G. Biocompatibility of surface-sealed dental materials against exposed pulps. Journal of Prosthetic Dentistry 1987;57:1±8. Foreman PC, Barnes IE. Review of calcium hydroxide. International Endodontics Journal 1990;23:283±97. Cox CF, Bergenholtz G, Fitzgerald M, Heys DR, Heys RJ, Avery JK, Baker JA. Capping of the dental pulp mechanically exposed to oral micro¯oraÐa 5 week observation of wound healing in the monkey. Journal of Oral Pathology 1982;11:327±39. Cox CF, Suzuki S. Re-evaluating pulp protection: calcium hydroxide liners versus cohesive hybridization. Journal of the American Dental Association 1994;125:823±31. Leirskar J, Helgeland K. Mechanism of toxicity of dental materials. International Endodontics Journal 1981;14:42±8. Tobias RS, Plant CG, Browne RM. Reduction in pulpal in¯ammation beneath surface-sealed silicates. International Endodontics Journal 1982;15:173±80. Schweikl H, Schmalz G. Toxicity parameters for cytotoxicity testing of dental materials in two different mammalian cell lines. European Journal of Oral Science 1996;104:292±9. Pascan EA, Spangberg LS. In vitro cytotoxicity of root canal ®lling materials, gutta-percha. Journal of Endodontics 1990;16:429±33. Arnst C, Carey J. Biotech bodies. Businessweek 1998;July:42±49. Tarim B, Hafez AA, Cox CF. Pulpal response to a resin-modi®ed glass-ionomer material on nonexposed and exposed monkey pulps. Quintessence International 1998;29:535±42. Akimoto N, Momoi Y, Kohno A, Suzuki S, Otsuki M, Suzuki S, Cox CF. Biocompatibility of Clear®l Liner Bond 2 and Clear®l AP-X system on nonexposed and exposed primate teeth. Quintessence International 1998;29:177±88. Murray PE, About I, Franquin J-C, Remusat M, Smith AJ. Restorative pulpal and repair responses. Journal of the American Dental Association 2001;132:482±91. Murray PE, About I, Lumley PJ, Franquin J-C, Remusat M, Smith AJ. Human odontoblast cell numbers after dental injury. Journal of Dentistry 2000;28:277±85. About I, Murray PE, Franquin J-C, Remusat M, Smith AJ. The effect of cavity restoration variables on odontoblast cell numbers and dental repair. Journal of Dentistry 2001;29:109±17. McKay GS. Gram stain modi®ed to improve colour contrast. Journal of Clinical Pathology 1970;23:191. MjoÈr IA. Dentin and pulp. In: MjoÈr IA, editor. Reaction patterns in human teeth, Boca Raton, FL: CRC Press, 1983. p. 63±156.

[34] Kakehashi S, Stanley HR, Fitzgerald R. The exposed germ-free pulp: effects of topical corticosteroid medication and restoration. Oral Surgery, Oral Medicine, Oral Pathology 1969;27:60±7. [35] Brannstrom M, Vojinovic O, Nordenvall KJ. Bacteria and pulpal reactions under silicate cement restorations. Journal of Prosthetic Dentistry 1979;41:290±5. [36] Watts A. Bacterial contamination and the toxicity of silicate and zinc phosphate cements. British Dental Journal 1979;146:7±13. [37] Paterson RC, Watts A. Toxicity to the pulp of a glass-ionomer cement. British Dental Journal 1987;162:110±2. [38] Zander HA, Pejko I. The reaction of dental pulps to silicate cement. Journal of the American Dental Association 1947;34:811±9. [39] Lefkowitz W, Seelig A, Zachinsky L. Pulp response to self-curing acrylic ®lling material. New York Dental Journal 1949;15:373±86. [40] Franquin JC, Brouillet JL. Biocompatibility of an enamel and dentin adhesive under different conditions of application. Quintessence International 1988;19:813±26. [41] Camps J, DeÂjou J, ReÂmusat M, About I. Factors in¯uencing pulpal response to cavity restorations. Dental Materials 2000;16:432±40. [42] Okiji T, Jontell M, Belichenko P, Dahlgren U, Bergenholtz G, Dahlstrom A. Structural and functional association between substance P and calcitonin gene related peptide-immunoreactive nerves and accessory cells in the rat dental pulp. Journal of Dental Research 1997;76:1818±24. [43] Langeland JC, Guttusco J, Jerome DR, Langeland K. Histological and clinical comparisons of Addent with silicate cements and coldcuring materials. Journal of the American Dental Association 1966;72:373±84. [44] Johnson RH, Christensen CJ, Stigers RW, Laswell HR. Pulpal irritation due to the phosphoric acid component of silicate cement. Oral Surgury 1970;29:447±54. [45] van Dijken JWV. Direct resin composite inlays/onlays: an 11-year follow up. Journal of Dentistry 2000;28:299±306. [46] Murray PE, Lumley PJ, Ross HF, Smith AJ. Tooth slice organ culture for cytotoxicity assessment of dental materials. Biomaterial 2000; 21:1711±21. [47] Stanley HR. Design for a human pulp study. Part 1. Oral Surgery Oral Medicine Oral Pathology 1968;25:633±47. [48] Imazato S, Turumi H, Ebi N, Ebisu S. Cytotoxic effects of composite restorations employing self-etching primers or experimental antibacterial primers. Journal of Dentistry 2000;28:61±7. [49] Fish EW. Pathology of the dentin and dental pulp. The British Dental Journal 1932;53:351±63.