Journal of Dental Research http://jdr.sagepub.

com/

Fluoride Release from Restorative Materials and Its Effects on Dentin Demineralization
C. Francci, T.G. Deaton, R.R. Arnold, E.J. Swift, Jr, J. Perdigao and J.W. Bawden J DENT RES 1999 78: 1647 DOI: 10.1177/00220345990780101001 The online version of this article can be found at: http://jdr.sagepub.com/content/78/10/1647

Published by:
http://www.sagepublications.com

On behalf of:
International and American Associations for Dental Research

Additional services and information for Journal of Dental Research can be found at: Email Alerts: http://jdr.sagepub.com/cgi/alerts Subscriptions: http://jdr.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://jdr.sagepub.com/content/78/10/1647.refs.html

>> Version of Record - Oct 1, 1999 What is This?

Downloaded from jdr.sagepub.com by guest on October 30, 2011 For personal use only. No other uses without permission.

J Dent Res 78(10): 1647-1654, October, 1999

Fluoride Release from Restorative Materials and Its Effects on Dentin Demineralization
C. Francci1, T.G. Deaton2, R.R. Amold3, E.J. Swift, Jr.4*, J. Perdigao4, and J.W. Bawden2
'Department of Dental Materials, University of Sao Paulo, Sao Paulo, Brazil; 2Department of Pediatric Dentistry, University of North Carolina, Chapel Hill, North Carolina; 3Department of Diagnostic Sciences, University of North Carolina; 4Department of Operative Dentistry, School of Dentistry, University of North Carolina, CB#7450, 302 Brauer Hall, Chapel Hill, North Carolina 27599-7450; *to whom correspondence should be addressed
Abstract. As the use of adhesive restorative materials has increased during the last several years, interest in adhesive materials that release fluoride has also grown. The purpose of this study was to measure fluoride release from several adhesive restorative materials and to evaluate its effect on dentin resistance to demineralization and on bacterial metabolism in a modified in vitro system. Standardized cavities (1.8 mm in diameter) were prepared in bovine teeth that had been ground to dentin. One cavity in each tooth was restored with one of the following restorative systems: (a) Single Bond and Z100; (b) Single Bond and Tetric Ceram; (c) Fuji Bond LC and Z100; (d) Fuji Bond LC and Tetric Ceram; (e) Fuji II LC; or (f) Fuji IX GP. The other cavity in each tooth was "restored" with wax as a control. For each restorative system, 12 specimens were evaluated for fluoride release during the first 24 hrs after restoration placement. Dentin adjacent to the restored sites was subjected to lactic acid challenge (pH 4.3) for 3 hrs, after which calcium release was measured. Another 12 specimens in each group were stored for 24 hrs in de-ionized water, and were exposed to an S. mutans suspension (1:1 THB/de-ionized water and 50 mM glucose, A660 = 0.2) for 6 hrs, followed by calcium release and pH measurement. Bulk specimens of each material were also made and stored in water. Fluoride released from Fuji Bond LC, Fuji IX GP, and Fuji II LC in bulk was significantly greater than from the other materials. In the restored dentin specimens, increased resistance to demineralization from a lactic acid challenge was directly related to fluoride release. The same effects were seen as a result of the S. mutans challenge. While fluoride release from restorative materials increased the resistance of dentin to demineralization in this system, the clinical relevance of the findings is not known. Key words: dentin bonding agents, resin composites, glassionomer cements, fluorides, dentin demineralization.
Received November 4, 1998; Last Revision February 10, 1999; Accepted February 12, 1999

Introduction
Almost 75% of all restorative procedures involve replacement of existing restorations (Kidd et al., 1992). The reason most commonly cited for restoration replacement is secondary caries, which accounts for about 40% of such replacements (Maclnnis et al., 1991). The caries process is a dynamic balance between demineralization and remineralization of the hard dental tissues that may eventually result in cavitation (Fejerskov, 1997). The metabolic products of dental plaque bacteria, including lactic, acetic, and citric acid, reduce pH in the micro-environment of the tooth surface and demineralize the dental hard tissues. Thus, any mechanism that inhibits such acid production, increases resistance to demineralization, and/or facilitates remineralization is of considerable clinical interest. Some restorative materials release fluoride, which may help to inhibit recurrent caries (Hatibovic-Kofman and Koch, 1991; Serra and Cury, 1992; Benelli et al., 1993; Dijkman et al., 1994; Gilmour et al., 1997; Dionysopoulos et al., 1998; Pereira et al., 1998). It appears that the primary mechanisms by which fluoride prevents caries are by improving resistance to demineralization and by facilitating remineralization of hard tissues. In high concentrations, fluoride may also directly affect cariogenic bacteria by inhibiting key metabolic enzymes, thereby reducing the ability of bacteria to produce acid (Marquis, 1995; Guha-Chowdhury et al., 1997). The use of restorative materials that not only release fluoride, but also are adhesive to tooth structure, could be especially effective in preventing recurrent caries. The first fluoride-releasing adhesive restorative materials were the silicate cements. These materials are no longer used because they had very poor physical properties. However, in 1972, Wilson and Kent introduced the glass-ionomer cements, which are derived from silicates. These materials bond to enamel and dentin and release fluoride (Palenik et al., 1992). In vitro studies suggest that fluoride release from glass ionomers may prevent recurrent caries around restorations (Geiger and Weiner, 1993; Tam et al., 1997)). They may also reduce the risk of
1647

Downloaded from jdr.sagepub.com by guest on October 30, 2011 For personal use only. No other uses without permission.

1648

Francci et al.

J Dent Res 78(10) 1999

ground with wet 120-grit silicon carbide abrasive paper on a mechanical grinder (Buehler, Ltd., Lake Bluff, IL, USA). The lingual surfaces were ground to a flat surface, exposing the pulp chamber so that pulpal tissues could be removed (Fig., b). The teeth were sectioned longitudinally into halves by means of a diamond-saw microtome (Gillings-Hamco, Rochester, NY, USA). To seal the pulp chamber, a resin bonding agent (Scotchbond MultiPurpose Adhesive, 3M Dental Products Division, St. Paul, MN, USA) was applied with a clean brush tip to the pulp chamber walls and the cut lingual surface, and was light-activated for 10 sec. Facial dentin surfaces were polished by hand with wet 240-, 400-, and 600-grit silicon carbide abrasive papers to produce standardized smear layers (Pashley et al., 1988). The specimens were stored in de-ionized water at 4C for 24 hrs. The cut lingual surface of each tooth half was bonded to a frosted 75 x 50 x 1 mm glass microscope slide (Fisher Scientific, Pittsburgh, PA, USA) with cyanoacrylate (Fig., c). Each glass microscope slide held 6 teeth, yielding a total of 12 sections. Two cavities were prepared in the facial dentin surface of each tooth (i.e., 1 in each paired section) (Fig., d). Both cavities were at the same location inciso-gingivally. A round carbide bur (#018, Brasseler USA, Savannah, GA, USA) was used at high speed under air/water coolant. The cavities were circular, with a diameter of the carbide bur (1.8 mm) and a depth of 0.8 mm. We ensured reproducibility of the bur inclination and cavity preparation depth by mounting the handpiece in a modified surveyor that permitted us to control its vertical movement. A coupled caliper was used for the measurement of handpiece movements. The restorative materials used in the study and their respective batch numbers are listed in Table 1. The teeth were randomly divided into 6 experimental restorative groups, as shown in Table 2. In the paired sections from each tooth, one of the cavities was restored with one of the selected restorative systems. The other cavity was "restored" with wax (Modern Materials, Heraeus Kulzer, South Bend, IN, USA) and served as a control for variations in dentin resistance to demineralization among different teeth (Fig., e). The dentin/enamel bonding agents Single Bond and Fuji Bond LC are non-fluoride-releasing and fluoride-releasing materials, respectively. Z100 (shade B2) is a resin composite that is identified by the manufacturer as not containing fluoride, while Tetric Ceram (shade B2), also a resin composite, contains both a ytterbium trifluoride radiopaquer and barium-aluminum-fluorosilicate glass fillers and is defined by its manufacturer as a fluoride-releasing material (Ivoclar North America, 1997). Fuji II LC (shade A2) is a light-cured resin-modified glass ionomer, and Fuji IX GP (shade A2) is a conventional glass ionomer. As glass ionomers, these materials are known to release fluoride. Each adhesive system was used according to its manufacturer's instructions. For Single Bond, the dentin surface was conditioned for 15 sec with Scotchbond Etching Gel (35% phosphoric acid) and was rinsed. Excess water was blotted, leaving the surface moist. Two consecutive coats of Single Bond were applied, air-dried for 2 sec, and light-cured for 10 sec. For Fuji Bond LC, the dentin surface was conditioned for 15 sec with Fuji Bond LC Conditioner (20% polyacrylic acid, 3% aluminum chloride) and rinsed. Excess water was blotted, leaving the surface moist. The adhesive was mixed, applied, and light-cured for 20 sec. The composites and the glass ionomers were placed in bulk and were light-cured for 40 sec (except for Fuji IX, which is a self-curing

caries on the adjacent surfaces of other teeth (Derkson et al., 1989). However, conventional glass ionomers suffer from a number of disadvantages, including short working time, long setting time, high sensitivity to early loss or gain of water, poor wear resistance, and surface roughness (Sidhu and Watson, 1995). Recently, two types of hybrids of glass ionomers and resin composites were introduced. The first is the resin-modified glass ionomer (RMGI), or a glass ionomer modified by the addition of methacrylate resins. RMGIs offer longer working and controlled setting times, rapid development of strength, and lower sensitivity to environmental moisture changes, and can be finished and polished immediately after being lightcured (Sidhu and Watson, 1995; Musa et al., 1996). The second hybrid is the polyacid-modified resin composite (or "compomer"), which contains some components of a glass ionomer but lacks the typical glass-ionomer acid/base reaction during the initial setting process (McLean et al., 1994). Most tooth-colored restorations are still prepared with resin composites. Modem composites have excellent physical properties and are clinically proven, but generally do not release fluoride. A few composites are reported to include fluoride that may be released by ion exchange or hydrolysis (Steinmetz et al., 1997), and there appears to be much interest among clinicians in fluoride-releasing composites. Adhesion of resins to enamel, which is 92% inorganic by volume (Jenkins, 1978), is considered highly reliable. However, dentin, because of its complex structure and variable composition, can give less predictable adhesion (Swift et al., 1995). It is 45 to 50% mineral, approximately 30% organic, and 20% water by volume (enkins, 1978). The lack of a good seal on dentin substrates jeopardizes the durability of cervical restorations (Dionysopoulos et al., 1998). Furthermore, dentin is more susceptible to acid attack than enamel (Arends et al., 1992). Its critical pH for demineralization is 6.0 to 6.8, compared with enamel at 5.2 to 5.7 (Hoppenbrouwers et al., 1986). Little is known about the caries-inhibitory effects of fluoridereleasing adhesive materials on dentin. The purpose of this in vitro study, therefore, was to evaluate the effects of fluoride release from several adhesive materials on dentin resistance to demineralization in response to lactic acid and Streptococcus mutans acid challenges. S. mutans has been identified as a principal etiologic agent for recurrent caries lesions (Kidd, 1989). It is homofermentative and produces significant quantities of lactic acids from carbohydrate substrates. The specific null hypotheses tested in this study were: (1) that fluoride released from restorative materials does not cause dentin to be more resistant to demineralization under standardized conditions of acid challenge; and (2) that fluoride is not released from these materials in sufficient quantities to reduce S. mutans acid production.

Materials and methods

General description of the experimental system In compliance with institutional guidelines, we obtained and cleaned 144 recently extracted bovine teeth (Randolph Packing, Asheboro, NC, USA). The crowns were sectioned from the roots (Fig., a). The facial enamel of each crown was removed by being

Downloaded from jdr.sagepub.com by guest on October 30, 2011 For personal use only. No other uses without permission.

J Dent Rcs 78(10) 1999

Fliuoride Release anid Resistaince to Deniineralizatin

1 64 649

Table 1. Materials used in the study material) by means of a Demetron 401 curing unit (Demetron/Kerr, Danbury, CT, USA). A dental Manufacturer Material Batch # radiometer (Demetron/Kerr, Model 100) was used to ensure that light intensity was at least 450 mW/cm2 Single B _)nd 7AD 3M Dental Products Division, St. Paul, MN throughout the study. The conventional glass Z100 19970616 ionomer Fuji IX was condensed and covered with an Tetric Ceeram 905563 lvoclar North America, Amherst, NY acetate sheet to avoid desiccation. Fuji Bon LC GC America, Chicago, IL. 110461 After specimens were stored in an incubator Fuji II L( 062796B under 100', humidity for 20 min at 37C, they were Fuji IX C;P 111161 polished with wet 600-grit silicon carbide abrasive paper for 30 sec. The purpose of polishing was to simulate clinical conditions and to remove excess restorative materials from the adjacent dentin. Table 2. Restorative groups and respective materials Borosilicate glass rings (4 mm inner diameter x 6 mm height) (Kimble Glass Company, Vineland, NJ, Restorative MaterialIs Group Material Types USA) were fitted to the dentin surface around each restoration. The purpose of these rings was to form Z()( Single Bond DBA + Composite 1 reservoirs for solutions to be placed in contact with Tetric Ceram + Bond Single (F-) DBA Composite 2 Laboratory dentin. the restorations and surrounding Zl(( Fuji Bond LC DBA (F-) + Composite 3 film (Parafilm M, American Can Company, -retr-ic Ceram Fuji Bond LC DBA (F-) + Composite (F-) 4 Greenwich, CT, USA) was fitted around the bottom II LC Fuji (resin-modified) Glass ionomer 5 of each glass ring to seal the contact with the dentin Fuji IX Gl' Glass ionomer (convenitional) 6 surface. After placement of the glass rings on the dentin surfaces, a layer of nail varnish (Super Shine, Del Laboratories, Farmingdale, NY, USA) was (aqueous solution of 0.31N,, HMDS, 4.48, 36 N H,SO4) was pipetapplied to secure them and to prevent leakage or contamination ted into each dish. The dishes were rocked gently on a rocker of the surrounding dentini by the solutions placed inside the rings platform (Bellco Biotechnology, Vineland, NJ, USA) for 2 hrs so (Figs. e, f). that possible fluoride contamination would be removed. A 2(011_L aliquot was taken from each sample reservoir and dissolved in Preliminary studies the 4 mL of diluted HMDS in each dish, and a 25-1LL droplet of 0.05 N NaOH was suspended under each lid as a fluoride trap. Three preliminary studies were conducted in an effort to control The dishes were carefully sealed vwith a thin ribbon of petroleumn for critical variables and to establish certain parameters of the jelly along the inside edge of each lid, and the vent hole was experimental protocol. The first preliminary study was consealed with a small square of laboratory film (Parafilm M) and ducted to ensure that the borosilicate glass wells did not release petroleum jelly. The fluoride was diffused from the HMDS/samor absorb fluoride. Ten wells were sealed to a flat laboratory film (Parafilm M) surface, and 30 IL of I ppm (part per million) fluoride (as NaF) solution was added to each well. To prevent I,,, .......... I.......evaporation, the top rim of each M. well was coated with a thin layer of petroleum jelly, and CI b Parafilm was applied to seal the top of the well. The wells were a incubated in 100'%, humidity at 37C for 24 hrs. Rcstorative The solutions were collected imaterial cover slip niarafilin wax and assayed for fluoride concend tration by means of the microdiffusion method originally aP t X Variiih laraf;; Ca. described by Taves (1968) and modified by Bawden et al. Dentin (1989). I'lastic Petri dishes (60 x 15 mm) (Falcon, Becton d e f Dickinson, Lincoln Park, NJ, USA), with 5-mm holes in their Figure. Procedural steps in test system design: (a) bovine tooth crown sectioned from root; (b) flattened lids, were used as diffusion facial and lingual surfaces, with line showing plane of section; (c) sections glued to microscope slide, w ith chambers. A 4-mL quantity of cavities prepared in each section; (d) superior view of cavity preparations; (e) restored cavities, reserx oirs, diluted hexamethyldisilazane and varnish seal; and (f) lateral view of glass vvell-filled with solution overlying restored cavity.

ci

......

__

4064.4

Downloaded from jdr.sagepub.com by guest on October 30, 2011 For personal use only. No other uses without permission.

1650

Francci et al.

J Dent Res 78(10) 1999

cured for 40 sec by the Optilux 401 device. The conventional glass-ionomer discs (Fuji IX GP) were protected with an acetate sheet to avoid desiccation. Borosilicate glass rings (4 mm inner diameter x 6 mm height) were fitted and sealed onto each specimen surface. These reservoirs were filled with 60 FL de-ionized water, covered, and stored in an incubator at 100% humidity for 24 hrs to permit fluoride to be released. We took 20-,uL samples for fluoride assay according to the method previously described. Data were submitted to one-way ANOVA and Bonferroni's post hoc test at a 95% confidence level. The statistical analyses were carried out with the SPSS/PC software system, version 8.0 (SPSS, Inc., Chicago, IL, USA).

ple solution to the NaOH droplet over an 18-hour period. The dishes were opened, the bottoms discarded, and 10 pLL of 0.2 N acetic acid mixed with the dried NaOH droplet. This solution was submitted to fluoride assay by means of the fluoride-specific electrode (#9409, Orion Research, Boston, MA, USA) attached to a calomel reference electrode (#41239, Beckman Instruments, Irvine, CA, USA) in such a manner that the tips of the electrodes could be bridged by the sample droplet. The electrodes were allowed to stabilize for 60 sec before the millivolt reading was recorded. Standards bracketing the readings were read before and after the samples were read. Between measurements, the electrodes were placed in TISAB (glacial acetic acid, sodium citrate, sodium chloride in water; Orion Research) until a baseline reading was achieved. The results were expressed as ng/well. Because the diffusion process concentrates the fluoride, the
method has a lower detection limit of 0.005 ppm on the linear part of the curve, with a measurement error of 5%. Diffusion also results in a clean solution for measurement by the electrode, thus eliminating the possibility of undetected electrode interference by other ions. No changes were detected in the fluoride concentration of the 10 solutions sampled, confirming that the borosilicate glass wells neither released nor absorbed measurable amounts of fluoride. In a second preliminary study, wherein we monitored pH changes over time, we tested a viable suspension of S. mutans strain NCTC 10449 in the mid-exponential phase of growth (optical density at 660 nm = 0.4), cultivated in Todd-Hewitt Broth (Difco Laboratories, Livonia, MI, USA), for its ability to metabolize glucose in the system. Starting at a near-neutral pH (7.4), we added 50 mM glucose to the suspension contained in the control wells (cavity preparations filled with wax). A pH of 4.3 was achieved after 3 hrs as a result of bacterial acid production. In a third preliminary experiment, 27 bovine teeth were used as described in the main experimental design, except that the teeth were not cut in half. Only one cavity was prepared on each tooth, and that was filled with wax. A borosilicate glass ring was fixed around the restored cavity. A volume of 60 ~iL of lactic acid solution at pH 3.4 was placed in each reservoir. Rings were sealed with Parafilm and stored at 37C and 100% humidity. Three rings, randomly assigned, were opened at two-hour intervals. We pipetted 30-,uL aliquots of each solution for calcium assay using a Zeeman 5100 Spectrophotometer with an AS-51 Auto Sampler (Perkin-Elmer, Norwalk, CT, USA). We used the data to establish a calcium saturation curve in the solution at the maximum demineralization rate to be observed during the experiments. It was determined that the curve was linear for 4 hrs, at which time a significant bend in the curve was first observed. It was necessary for the experiments to be conducted over a period of time during which the curve was nearly linear; otherwise, the experimental curves would begin to approach each other and eventually converge, resulting in misleading observations. An experimental demineralization time of 3 hrs was selected. This time period also allowed measurable amounts of calcium to be
released into the solution.

Effect of fluoride release on dentin demineralization under lactic acid challenge Seventy-two bovine teeth that had been sectioned, prepared, and restored as described (12 per restorative system) were used. Each of the paired reservoirs was filled with 30 FL de-ionized water. Specimens were stored in an incubator at 37C and 100% humidity to permit fluoride to be released from the adhesive and filling materials into the water and diffused into the surrounding dentin. After 24 hrs, the solution was pipetted from each reservoir, and any excess water inside the control and paired experimental wells was removed by being blotted with a laboratory tissue paper (Kimwipes EX-L, Kimberly-Clark, Roswell, GA, USA). Both reservoirs (test and control) for each tooth were then filled with 60 ,uL of lactic acid solution at pH 4.3. After 3 hrs, the control and experimental wells were sampled (30 FLL) and assayed for Ca++ according to the method described above. The calcium release values observed in the control wells were used to adjust for variations in dentin demineralization rates from tooth to tooth by application of the following formula:
CaA = CaCE/CaCT x CaE

CaA = adjusted mean Ca2+ from the test wells for a single restorative system. CacE = mean Ca2+ released from the control wells for a single restorative system. CacT = mean Ca2+ released from the control wells for all six restorative systems. CaE = mean Ca2+ released from the paired experimental wells for the same restorative system.
The remaining solution in each well was used for the fluoride assay according to the method described above. Data were analyzed by a paired t test and ANOVA.

Evaluation of fluoride effects

on bacterial metabolism and pH changes Seventy-two of the sectioned, prepared, and restored bovine teeth (12 per group) were used in this test. The reservoirs were filled with de-ionized water (30 ,uL) for 24 hrs to permit fluoride to be released from those materials that contained fluoride. Thirty

Fluoride release from dental materials in bulk Specimens of each dental material were fabricated in glass cylinders (6 mm internal diameter x 2 mm height). These matrices were filled in bulk, and, where indicated, materials were light-

pLL of Todd-Hewitt Broth containing S. mutans NCTC 10449 in the mid-exponential phase of growth (optical density at 660 nm = 0.4 from an initial OD 660 nm < 0.01) was added to each well. Fifty mM glucose (0.54 ,ug) was added to each reservoir, and the

Downloaded from jdr.sagepub.com by guest on October 30, 2011 For personal use only. No other uses without permission.

j Dent Res 78(10) 1999

Fluoride Release and Resistance to Demineralization

1651

reservoirs were sealed and incubated for 6 hrs at 37C. When a 30-pLL quantity of the bacteria-containing broth was added, the volume in the wells was increased to 60 ,uL, and the resulting fluoride concentrations in the solutions were similar to those observed in the solutions in the wells during the lactic acid challenge experiment. The solutions were left in the wells for 6 hrs. At the end of the incubation period, a 30-,uL aliquot of solution was retrieved from each reservoir (test and control for each tooth), and calcium was assayed by the method previously described. The six-hour exposure time allowed for the time required to obtain a minimum pH (4.3) and resulted in release of reliably assayed amounts of calcium that were in the desired portion of the saturation curve. The pH of the solution in each reservoir (test and control) was measured after 6 hrs by means of a pH microprobe (MI-410, Microelectrodes, Bedford, NH, USA). The recorded pH values in test and control reservoirs were compared by paired t test for each restorative group.

Table 3. Fluoride release from bulk specimens of the tested dental materials Material
Fuji Bond LC Fuji IX GP Fuji II LC
Type DBA Conventional GIC RMGI DBA Composite Composite

Fluoride Release?a
yes yes yes no yes no

Fluoride Release

p.g SDb
31.5 4.6* 18.1 2.Ot 12.3 1.6t 0.8 0.4+ 0.7 + 0.4+ 0.5 + 0.3+

Single Bond Tetric Ceram Z100

a

As reported by the manufacturer. Means with the same symbol are not significantly different (p >

0.05).

Results
The mean ( SD) fluoride release values for bulk specimens of each material after exposure to de-ionized water for 24 hrs are shown in Table 3. Fuji Bond LC released the most fluoride, followed by the conventional glass ionomer Fuji IX GP, and the resin-modified glass ionomer Fuji II LC. Differences between these materials were statistically significant, as shown by Bonferroni's multiple-range test (p < 0.05) following one-way ANOVA (p < 0.0001). The remaining materials released significantly less fluoride, and the values were not significantly different from each other. Two of the materials (Single Bond and Z100) reportedly contain no fluoride, but did release detectable amounts. Tetric Ceram, which reportedly does contain fluoride, released only trace amounts. The mean fluoride and calcium release in specimens under lactic acid attack are summarized in Table 4 for each of the restorative systems and its control. Under the conditions of this phase of the experiment, exposed dentin surrounded each restored cavity. Fuji Bond LC, which released the most fluoride in bulk, did not release as much fluoride into the lactic acid solution as either Fuji IX GP or Fuji II LC. However, it was applied in a thin layer as a bonding agent, which is its intended purpose. Of the restorative materials, Fuji IX GP released the most fluoride and Fuji II LC released the second most, the same relative order as observed for

bulk specimens. The fluoride release value for Fuji IX GP was the only one that was statistically different from the other values. Calcium release from test sites subjected to lactic acid challenge and from control sites were compared by means of a paired t test. The groups restored with the conventional glass ionomer Fuji IX GP and with the resin-modified glass ionomer Fuji II LC had significantly less calcium release than their controls (p < 0.001) (Table 4). ANOVA and Bonferroni's multiple-range test for adjusted calcium release values showed that the specimens restored with conventional ionomer Fuji IX GP specimens had the greatest increase in dentin resistance to demineralization (as indicated by lower calcium release), followed by the resin-modified glass ionomer, Fuji II LC. The remaining groups had similar adjusted calcium release values (p > 0.05). A strong inverse relationship was noted between fluoride release from the restorative systems and calcium release from dentin (Pearson correlation coefficient, r = -0.75, p < 0.0001). The mean calcium release values and the mean broth pH values resulting from exposure to the S. mutans challenge are summarized in Table 5. Because the exposure time to the S. mutans challenge was twice as long as the lactic acid challenge, all of the restorative groups except for the conventional glass ionomer (Fuji IX GP) had greater calcium release when exposed to S. mutans than when exposed to lactic acid (compare adjusted calcium release values in Tables 4 and 5). None of the test pH values was significantly different from its controls or each other, suggesting that none of the restorative systems released enough fluoride to affect S. mutans metabolism.

Table 4. Fluoride release and calcium release from specimens subjected to lactic acid challenge Restorative Group (n = 12)

Fluoride Release ng SDa

Test

Calcium Release SD (ng)a Control Adjusted 180.0 19.8

181.8 9.0

Adhesive + Composite

Single Bond
Fuji Bond LC

Z100 Tetric Ceram

0.00* 0.00*

Zioo

Resin-modified Glass lonomer Conventional Glass lonomer

a

Tetric Ceram Fuji II LC Fuji IX GP

0.00* 10.9 17.1* 25.8 14.1* 144.0 68.4t

180.0 18.0 173.4 10.8 151.8 12.0 136.2 10.8 126.6 + 19.8b 114.6 13.8b

166.8 12.0*
156.0 12.0*

156.6 22.2 140.4 7.8

151.2 + 15.6b

160.2 16.8* 159.0 16.2*

136.8 17.4t

171.0 20.4b

109.8 10.8t

Means with the same symbol are not significantly different (p > 0.05). Test and control means of the same restorative group are significantly different (p < 0.001).
Downloaded from jdr.sagepub.com by guest on October 30, 2011 For personal use only. No other uses without permission.

1652

Francci et al.

J Dent Res 78(10) 1999

Table 5. Calcium release and pH of specimens subjected to S. mutans challenge

Restorative Group

Test

Calcium Release SD (ng) Control Adjusteda

227.4 60.0 190.8 49.8 264.6 31.8 243.0 50.4

Test

pH + SD Control 5.11 0.15 5.19 0.19 5.12 0.12 5.18 0.11 5.11 0.14 5.08 0.17

Adhesive + Composite

Single Bond

Zi1O

Resin-modified Glass Ionomer Conventional Glass lonomer

a

Tetric Ceram Z100 Fuji Bond LC Tetric Ceram Fuji II LC Fuji IX GP

194.4 67.8 219.6 60.6 244.2 24.0 222.6 37.8 169.2 43.2b 105.0 33.0b

198.6 70.8* 277.2 103.8k 211.8 30.6* 214.2 46.8* 225.0 46.8b 182.4 76.2t 210.0 49.2b 115.8 34.8+

5.22 0.24 5.21 0.18 5.10 0.11 5.17 0.07 5.14 0.16 5.18 0.14

Means with the same symbol are not significantly different (p > 0.05). Test and control means of the same restorative group are significantly different (p < 0.01 for Fuji II LC; p < 0.001 for Fuji IX GP).

The calcium release values resulting from the S. mutans challenge showed considerably more variability than was observed in the lactic acid challenge experiment. Only the groups restored with conventional (Fuji IX GP) or resin-modified glass ionomer (Fuji II LC) had calcium release values significantly lower than their respective controls when submitted to a paired t test (p < 0.0001 and p < 0.01, respectively). Using ANOVA and Bonferroni's multiple-range test, we found that only Fuji IX GP had an adjusted calcium release rate that was significantly less than that for the other restorative systems. Overall, there was an inverse relationship between fluoride release and calcium release from dentin when exposed to an S. mutans suspension (Pearson correlation coefficient, r = -0.44, p < 0.0001).

Discussion
We undertook this study with two purposes in mind. The first was to develop an improved system for the in vitro testing of dental materials with respect to their ability to prevent demineralization of adjacent tooth structure. The second was to apply the method to six restorative systems to estimate their capacity to prevent demineralization of adjacent dentin as related to the amount of fluoride release from each of the systems. Leakage studies (Reeves et al., 1995; Castelnuovo et al., 1996) indicate that bonding of a material to tooth structure reduces the potential for in vitro demineralization of the enamel or dentin by reducing penetration of cariogenic bacteria and acids through the material/tooth structure interface. Other in vitro studies indicate that fluoride release from certain restorative materials can reduce the severity of recurrent caries (Hatibovic-Kofman and Koch, 1991; Serra and Cury, 1992; Benelli et al., 1993; Dijkman et al., 1994; Gilmour et al., 1997; Dionysopoulos et al., 1998; Pereira et al., 1998). This effect may be the result of increased resistance to demineralization and enhanced remineralization (ten Cate and van Duinen, 1995). While there is an abundance of such studies reporting on these effects with respect to restorative materials and enamel, fewer have been conducted on dentin (Purton and Rodda, 1988; ten Cate and van Duinen, 1995). The method used in this study involved a number of controls and precision measurements not previously included in any single in vitro method. The first preliminary experiment

was conducted to document that no measurable amount of fluoride was released or absorbed by the borosilicate glass wells or the materials used to affix them to the dentin samples. Second, the S. mutans suspension was shown to be consistent for phase, optical density, and its capacity to achieve pH 4.3 in 3 hrs when in contact with exposed dentin in the wells. During the S. mutans challenge experiment, the pH of the solutions in the wells was similar among the restorative groups at the end of the exposure time (6 hrs), indicating that the challenge was consistent among the restorative systems. However, it was not feasible to monitor pH values during the experiment, and there may have been variations among systems at some time during the exposure. Third, dentin calcium release values would be misleading if the time of exposure of the dentin to lactic acid or bacterial suspension exceeded the point where the rate of uptake into the solution was approaching stability. It was determined that a departure from linearity in the curve was detectable at 4 hrs. The acid challenge experiment was conducted for 3 hrs to ensure that solution saturation effects did not measurably distort the data and yet released sufficient calcium to facilitate accurate assay. The experiment exposing dentin to the bacterial challenge was run for 6 hrs, sufficient time for pH decrease after the addition of glucose and the release of reliably assayed calcium concentrations. In addition, all experiments were controlled for variability in acid resistance from one tooth to another. After the dentin surface was exposed and prepared, each tooth was sectioned into two halves. Identical cavities were prepared in each of the paired halves as closely as possible to the sectioned surface, allowing for space to affix the glass well. Each test cavity, restored with one of the selected material systems, was matched with its paired control, which was filled with wax. Test measurements of calcium release from dentin as a result of exposure to lactic acid (pH 4.3) or a standardized suspension of S. mutans were adjusted to the calcium release rate observed in the paired controls. There was considerable variation in calcium release rates from tooth to tooth, and adjustment of the test values to account for this variability had clear effects on the data. As seen in Tables 4 and 5, the rank order of some test results changed as a result of controlling for this variable. Also, the calcium release values for the restorative systems that released undetectable amounts of fluoride in these phases of the study became more similar when controlled for variations in dentin resistance to demineralization.

Downloaded from jdr.sagepub.com by guest on October 30, 2011 For personal use only. No other uses without permission.

j Dent Res 7800) 1999

Fluoride Release and Resistance to Demineralization

1653

The micro-method used for the fluoride assay has been described previously (Bawden et al., 1989). It has a low detection limit (0.005 ppm) and measurement error ( 5%). It eliminates the possibility of undetectable electrode interference by competing ions. The calcium assay method was also able to measure small amounts of calcium release from the dentin (50 -qg) with a measurement error of 3%. The measurement of calcium release is a quantitative and sensitive way to measure dentin demineralization. The comparison of data presented in Tables 3 and 4 concerning fluoride release from the various materials and restorative systems is of interest. In both cases, the materials or systems were exposed to de-ionized water for 24 hrs. However, fluoride release from the bulk materials Fuji IX and Fuji II LC was at least two orders of magnitude greater than fluoride release from restorations in bovine dentin. The exposed area of the bulk materials was about five times that of the restorations, which would account for some of the difference. The remainder must be accounted for by uptake of fluoride by dentin surrounding the restorations. It appears that 95% of the fluoride released from the materials diffused into dentin and only 5% into the water. Dentin apparently acts as a relatively powerful sink for fluoride. While Fuji Bond LC released the most fluoride when exposed in bulk to water (Table 3), it released less fluoride into solution than Fuji IX GP or Fuji II LC in restored specimens (Tables 4 and 5). This discrepancy is easily accounted for, because Fuji Bond LC was used as a bonding agent in restored specimens and thus was applied in a thin layer. The findings are consistent with those of other in vitro studies that have reported an inverse relationship between fluoride release from dental materials and increased resistance to demineralization in adjacent enamel or dentin (Hattab et al., 1989; Swift, 1989; Tam et al., 1997). Although two of the six materials used in the study (Single Bond and Z100) were reported by the manufacturer (3M Dental) as not containing fluoride, both of them released trace amounts of fluoride. Personal correspondence with the manufacturer revealed that fluoride in small amounts is, in fact, added as a catalyst. In contrast, Tetric Ceram, which is described by its manufacturer as a fluoride-releasing material, released only trace amounts of fluoride. While this experimental system appears to be well-controlled, sensitive, and quantitative, as with similar in vitro methods, its relevance to the clinical situation is not established. Controlled clinical trials will be required to document or discredit its clinical usefulness. Until that is done, the findings from this system cannot be extrapolated as clinically relevant. In summary, an in vitro method for quantitative measurement of the effects of selected restorative materials on the resistance to acid etching of adjacent dentin is presented. The method combines a combination of controls and micro-analytical methods not previously reported for a single protocol. There was an inverse relationship between fluoride release from the restorative systems and the demineralization rate of the adjacent dentin as a result of lactic acid or bacterial challenge. The clinical relevance of the method has not been

reject the second null hypothesis. Specifically, fluoride release did increase dentin resistance to demineralization, but did not inhibit S. mutans acid production.

Acknowledgments
Materials for this project were donated by 3M Dental Products Division, GC America, and Ivoclar North America. Some financial support was provided by the Carolina Fund of the University of North Carolina and the Coordenagao de Aperfeiqoamento de Pessoal de Nivel Superior. The authors thank Dr Jose Fortunato Ferreira Santos for his role in facilitating the faculty exchange that resulted in this work.

References
Arends J, Christoffersen J, Buskes JA, Ruben J (1992). Effects of fluoride and methanehydroxydiphosphate on enamel and on dentine demineralization. Caries Res 26:409417. Bawden JW, Deaton TG, Koch GG, Crawford BP (1989). Effect of an acute maternal dose on fetal plasma and enamel uptake in response to an acute maternal fluoride dose in guinea pigs. J Dent Res 68:1167-1172. Benelli EM, Serra MC, Rodrigues AL Jr, Cury JA (1993). In situ anticariogenic potential of glass ionomer cement. Caries Res
27:280-284. Castelnuovo J, Tjan AH, Liu P (1996). Microleakage of multi-step and simplified-step bonding systems. Am J Dent 9:245-248. Derkson GD, Richardson AS, Jinks GM (1989). Clinical evaluation of a restoration containing fluoride: two-year results. Pediatr Dent 11:286-290. Dijkman GE, de Vries J, Arends J (1994). Secondary caries in dentine around composites: a wavelength-independent microradiographical study. Caries Res 28:87-93. Dionysopoulos P, Kotsanos N, Pagadogiannis Y, Konstantinidis A (1998). Artificial secondary caries around two new F-containing restoratives. Oper Dent 23:81-86. Fejerskov 0 (1997). Concepts of dental caries and their consequences for understanding the disease. Community Dent Oral Epidemiol 25:5-12. Geiger SB, Weiner S (1993). Fluoridated carbonatoapatite in the intermediate layer between glass ionomer and dentin. Dent Mater 9:33-36. Gilmour AS, Edmunds DH, Newcombe RG (1997). Prevalence and depth of artificial caries-like lesions adjacent to cavities prepared in roots and restored with a glass ionomer or a dentin-bonded composite material. J Dent Res 76:1854-1861. Guha-Chowdhury N, Clark AG, Sissons CH (1997). Inhibition of purified enolases from oral bacteria by fluoride. Oral Microbiol Immunol 12:91-97. Hatibovic-Kofman S, Koch G (1991). Fluoride release from glass ionomer cement in vivo and in vitro. Swed Dent J 15:253-258. Hattab FN, Mok NY, Agnew EC (1989). Artificially formed carieslike lesions around restorative materials. J Am Dent Assoc
118:193-197.

established. The study rejected the first null hypothesis, but failed to

Hoppenbrouwers PM, Driessens FC, Borggreven JM (1986). The vulnerability of unexposed human dental roots to demineralization. J Dent Res 65:955-958.

Downloaded from jdr.sagepub.com by guest on October 30, 2011 For personal use only. No other uses without permission.

1654

Francci et al.

J Dent Res 78(10) 1999

Purton DG, Rodda JC (1988). Artificial caries around restorations in roots. J Dent Res 67:817-821. Reeves GW, Fitchie JG, Hembree JH Jr, Puckett AD (1995). Microleakage of new dentin bonding systems using human and bovine teeth. Oper Dent 20:230-235. Serra MC, Cury JA (1992). The in vitro effect of glass-ionomer cement restoration on enamel subjected to a demineralization and remineralization model. Quintessence Int 23:143-147. Sidhu SK, Watson TF (1995). Resin-modified glass ionomer materials-A status review for the American Journal of Dentistry. Am I Dent 8:59-67. Steinmetz MJ, Pruhs RJ, Brooks JC, Dhuru VB, Post AC (1997). Rechargeability of fluoride releasing pit and fissure sealants and restorative resin composites. Am J Dent 10:36-40. Swift EJ Jr (1989). In vitro caries-inhibitory properties of a silver cermet. J Dent Res 68:1088-1093. Swift EJ Jr, Perdigao J, Heymann HO (1995). Bonding to enamel and dentin: a brief history and state of the art. Quintessence Int 26:95-110. Tam LE, Chan GPL, Yim D (1997). In vitro caries inhibition effects by conventional and resin-modified glass-ionomer restorations. Oper Dent 22:4-14. Taves DR (1968). Determination of submicromolar concentrations of fluoride in biological samples. Talentia 15:1015-1023. ten Cate JM, van Duinen RN (1995). Hypermineralization of dentinal lesions adjacent to glass-ionomer cement restorations. I Dent Res 74:1266-1271. Wilson AD, Kent BE (1972). A new translucent cement for dentistry: the glass ionomer cement. Br Dent J 132:133-135.

Ivoclar North America (1997). Tetric Ceram Tetric Flow scientific documentation. Amherst, NY: Ivoclar, pp. 9-10. Jenkins GN (1978). Chemical composition of teeth. In: The physiology and biochemistry of the mouth. 4th ed. Jenkins GN, editor. Oxford: Blackwell Scientific Publications, p. 60. Kidd EAM (1989). Caries diagnosis within restored teeth. Oper Dent 14:149-158. Kidd EAM, Toffenetti F, Mjor IA (1992). Secondary caries. Int Dent

J 42:127-138.
MacInnis WA, Ismail A, Brogan H (1991). Placement and replacement of restorations in a military population. J Can Dent Assoc 57:227-231. Marquis RE (1995). Antimicrobial actions of fluoride for oral bacteria. Can J Microbiol 41:955-964. McLean JW, Nicholson JW, Wilson AD (1994). Proposed nomenclature for glass-ionomer dental cements and related materials. Quintessence Int 25:587-589. Musa A, Pearson GJ, Gelbier M (1996). In vitro investigation of fluoride ion release from four resin-modified glass polyalkenoate cements. Biomaterials 17:1019-1023. Palenik CJ, Behnen MJ, Setcos JC, Miller CH (1992). Inhibition of microbial adherence and growth by various glass ionomers
in vitro. Dent Mater 8:16-20.

Pashley DH, Tao L, Boyd L, King GE, Homer JA (1988). Scanning electron microscopy of the substructure of smear layers in human dentine. Arch Oral Biol 33:265-270. Pereira PN, Inokoshi S, Tagami J (1998). In vitro secondary caries inhibition around fluoride releasing materials. J Dent 26:505-510.

Downloaded from jdr.sagepub.com by guest on October 30, 2011 For personal use only. No other uses without permission.