Re- and Demineralization Characteristics of Enamel Depending on Baseline Mineral Loss and Lesion Depth in situ

Transcription

1 Original Paper Received: June 25, 215 Accepted: February 5, 216 Published online: April 5, 216 Re- and Demineralization Characteristics of Enamel Depending on Baseline Mineral Loss and Lesion Depth in situ Richard J. Wierichs Julian Lausch Hendrik Meyer-Lueckel Marcella Esteves-Oliveira Department of Operative Dentistry, Periodontology and Preventive Dentistry, RWTH Aachen University, Aachen, Germany Key Words Demineralization In situ model Remineralization Sodium fluoride Stannous fluoride Abstract Objectives: The aim of this double-blinded, randomized, cross-over in situ study was to evaluate the re- and demineralization characteristics of sound enamel as well as lowly and highly demineralized caries-like enamel lesions after the application of different fluoride compounds. Methods: In each of three experimental legs of 4 weeks, 21 participants wore intraoral mandibular appliances containing 4 bovine enamel specimens (2 lowly and 2 highly demineralized). Each specimen included one sound enamel and either one lowly demineralized (7 days, ph 4.95) or one highly demineralized (21 days, ph 4.95) lesion, and was positioned 1 mm below the acrylic under a plastic mesh. The three randomly allocated treatments (application only) included the following dentifrices: (1) 1,1 ppm F as NaF, (2) 1,1 ppm F as SnF 2 and (3) ppm F (fluoride-free) as negative control. Differences in integrated mineral loss (ΔΔZ) and lesion depth (ΔLD) were calculated between values before and after the in situ period using transversal microradiography. Results: Of the 21 participants, 6 did not complete the study and 2 were excluded karger@karger.com S. Karger AG, Basel /16/52 141$39.5/ due to protocol violation. Irrespectively of the treatment, higher baseline mineral loss and lesion depth led to a less pronounced change in mineral loss and lesion depth. Except for ΔΔZ of the dentifrice with ppm F, sound surfaces showed significantly higher ΔΔZ and ΔLD values compared with lowly and highly demineralized lesions (p <.5, t test). Conclusion: Re- and demineralization characteristics of enamel depended directly on baseline mineral loss and lesion depth. Treatment groups should therefore be well balanced with respect to baseline mineral loss and lesion depth. 216 S. Karger AG, Basel Since 1964 a wide range of in situ caries models have been developed using a variety of hard tissue substrates [Zero, 1995]. Sound surfaces, natural lesions as well as lesions formed in vitro, in situ or in vitro/in situ have been used [Mellberg, 1992]. Although several reviews highlighted the potentials and limitations of the different in situ models from different perspectives [Wefel, 199; Manning and Edgar, 1992; Zero, 1995], only one of them discussed the influence of baseline mineral loss and lesion depth on re- or demineralization characteristics [ten Cate, 1994]. However, this review showed inconclusive results. Dr. Richard J. Wierichs Department of Operative Dentistry, Periodontology and Preventive Dentistry RWTH Aachen University Pauwelsstrasse 3, DE 5274 Aachen (Germany) ukaachen.de

2 142 In fact, only two in situ studies, specifically analyzing the influence of baseline mineral loss and lesion depth on the re- and demineralization characteristics of enamel, could be found in the currently available literature [Strang et al., 1987; Lippert et al., 211]. After the application of dentifrice containing 1,1 ppm F (as sodium fluoride, NaF), a linear relationship between both factors could be shown for predemineralized human enamel specimens, situated in the lingual position of the lower jaws [Strang et al., 1987]. Lesions with higher baseline mineral loss produced a more pronounced remineralization. The second study, also using dentifrice containing 1,1 ppm F (as NaF), revealed different reactions for lesions with different R values (being the ratio of mineral loss, ΔZ, to lesion depth) [Lippert et al., 211]. Highly demineralized lesions (high R value) tended to remineralize, whereas lowly demineralized lesions (low R value) further demineralized. Predemineralized human enamel specimens situated in the vestibular position of the lower jaws were brushed twice daily for 1 min. Additionally, two articles, summarizing data from different in situ studies, revealed similar linear relationships for specimens treated with fluoride dentifrices [Mellberg, 1991b; Schafer et al., 1992] but not for specimens treated with fluoride-free dentifrices [Mellberg, 1991b]. In the first article data from nine fluoride dentifrice groups (1,1 ppm F as NaF) showed an increase in mineral uptake with higher baseline mineral loss, whereas data from eleven fluoride-free dentifrice groups ( ppm F) showed a loss in mineral content in all lesions except the largest ones, with the smallest lesions having the greatest absolute mineral loss [Mellberg, 1991b]. Furthermore, baseline mineral loss did not affect the percentage change when specimens were treated with fluoride dentifrices, whereas it strongly affected the percentage change for fluoride-free treatment. In the second paper data from four in situ studies revealed that small lesions acquired less mineral than large lesions and that baseline mineral loss was related to whether net demineralization or net remineralization occurred [Schafer et al., 1992]. In each of the four included studies human enamel specimens were brushed with dentifrices containing 1,1 1,45 ppm F (as sodium monofluorophosphate, SMFP, or NaF). Although studies have shown an increase in mineral uptake for lesions with higher baseline mineral loss, when testing predemineralized human enamel specimens, none of the articles used bovine specimens or examined sound surfaces compared with predemineralized specimens. No fluoride compounds other than NaF or SMFP (without differentiating between them) have been investigated. Just one of the articles not only analyzed dentifrices containing 1,1 1,45 ppm F (as SMFP or NaF) but also fluoride-free dentifrices [Mellberg, 1991b]. Thus, the purpose of the present study was to investigate the relation between baseline mineral loss and baseline lesion depth on re- and demineralization characteristics of sound surfaces as well as lowly and highly demineralized caries-like enamel lesions using an established in situ caries model. The secondary purpose of this study was to investigate the influence of SnF 2, NaF and fluoridefree dentifrice on this performance. We hypothesized that regardless of the respective fluoride, sound surfaces and lowly demineralized lesions are significantly more prone to demineralization than highly demineralized ones. Materials and Methods Ethical Aspects Ethical approval was given by the local institutional ethical committee (Medical Faculty, RWTH Aachen University; No. EK 136/13). The number of participants was calculated on the basis of previously performed in situ studies [Dijkman et al., 199; Mellberg et al., 1992; Meyer-Lueckel et al., 215a, b]. The α-error was set at 5%. Considering the differences between the toothpastes with 1,1 and ppm F the statistical power calculated was 98% (mean difference of 879, SD = 91). The dropout rate was assumed not to exceed 2%. Approximately 2 participants should have been enrolled into the study for an expected completion of at least 16. However, at the end of the in situ periods the dropout rate was 38% (13 participants completed the study). Since the retrospective power analysis with 13 participants has still provided a power of at least 8%, no additional participant was involved in the study. All participants (21 volunteers, 13 women and 8 men, aged from 21 to 55 years) lived in Aachen, Germany, used tap water with a fluoride concentration of approximately.2 mg/l and signed a written informed consent. They were all in good general health with no signs of active caries or periodontal disease. Exclusion criteria were as follows: pregnancy, current participation in another study, institutionalized patients, periodontal disease, active caries lesions, age <18 years, salivary flow rate <.7 ml/min, no written informed consent and incapability of contracting. Study Design The study design was a double-blinded, randomized, crossover in situ trial with three treatment legs. After screening for general eligibility, dental impressions of the lower jaw were taken and appliances with bilateral flanges were prepared ( fig. 1 B) [Koulourides et al., 1974]. In each of both flanges 2 predemineralized bovine enamel specimens were inserted 1 mm below the acrylic under a plastic mesh (Perfect Splint System; Hager and Werken, Duisburg, Germany) mimicking plaque-retaining surfaces [Meyer-Lueckel et al., 27; Schirrmeister et al., 27; Meyer-Lueckel et al., 215b]. Each specimen included one sound enamel and either one lowly demineralized (7-day, ph 4.95) or one highly demineralized lesion (21-day, ph 4.95; fig. 1 ). Wierichs/Lausch/Meyer-Lueckel/ Esteves-Oliveira

3 A a b c TMR Baseline h d 5 mm 5 mm g Thickness: 1 μm 3.5 mm 3 mm e 5 mm Enamel Dentine f Resin 3.5 mm B H L H L Color version available online i j Demineralization (7 or 21 days, ph 4.95) Resin C S L/H Fig. 1. A Specimen preparation. C = Control surface; S = sound surface; L/H = lowly or highly demineralized lesion; a = frontal view of bovine front tooth; b = separation of crown and root; c, d = cuts perpendicular and parallel to the long axis of the tooth crown; e = obtained specimens ( mm); f = specimen covered with resin; g = predemineralized specimen; h = obtainment of the 1 μm slices for baseline TMR analysis; i = specimen covered with resin for in situ appliance; j = specimens inserted 1 mm below the acrylic under a plastic mesh. B Design of the intraoral mandibular appliances. L = Lowly demineralized lesions; H = highly demineralized lesions. Two specimens with lowly demineralized lesions and 2 specimens with highly demineralized lesions were inserted into the appliances. Each specimen included 1 sound surface. The factors under evaluation were as follows: Baseline substrate condition at three levels: sound, lowly demineralized or highly demineralized Intervention at three levels: application of dentifrice containing: (1) sodium fluoride (1,1 ppm F, ph 7.8) and triclosan (.3%; Colgate Total; Colgate-Palmolive, New York, N.Y., USA; standard treatment, NaF), (2) sodium gluconate-stabilized stannous fluoride (1,1 ppm F, ph 5.6: Procter & Gamble, Weybridge, UK; experimental treatment, SnF 2 ) or (3) fluoride-free dentifrice ( ppm F, ph 7.2; Procter & Gamble; negative control) The participants wore intraoral mandibular appliances for three legs of 4 weeks each, with the in situ exposure only being interrupted during meals and for oral hygiene, resulting in a total wearing time of h per day. Each of the three treatment legs were immediately preceded by a 1-week lead-in period. For each test cycle the participants received a new toothbrush (Oral-B Indicator; Proctor and Gamble, Schwalbach am Taunus, Germany), a nutrition protocol and the dentifrice in a white packaging containing no commercial names. Twice daily (in the morning and evening), after removing the intraoral appliance from the mouth, 1 g of the respective dentifrice was applied on the toothbrush. After brushing the teeth for 3 s, the participants spat out the resultant saliva dentifrice mixture (natural slurry) and applied it extraorally on the specimens for Re- and Demineralization Characteristics of Enamel 2 min. Subsequently, the participants completed brushing their own teeth (total brushing time: 2 min). Specimens were not brushed at any time to allow plaque to grow. Afterwards the appliances were washed with tap water and re-inserted into the oral cavity, and left for at least 3 min without any intake of drink or food during this period. During additional extraoral times the appliances were stored in plastic boxes in a humid environment. These extra times were noted and added at the end of each phase to ensure similar wearing times for each phase for each volunteer. The participants were informed not to use high fluoride-containing products (food and oral hygiene products) and received fluoride-free salt for home use. To trigger demineralization, the appliances were placed in 1% sugar solution for 2 min 3 times daily [Meyer-Lueckel et al., 215a, b]. Randomization After baseline examination a computerized random allocation sequence was generated by the study sponsor who coded the dentifrices with a subject number. The code was provided by a third person not directly involved in this study in a sealed envelope to be broken only in the case of an emergency. Specimen Preparation Bovine incisors were obtained from freshly slaughtered cattle (negative BSE test) and stored in.8% thymol ( fig. 1 A). Teeth 143

4 Table 1. Mean mineral loss, SZ max and lesion depth for sound surfaces as well as lowly and highly demineralized lesions before and after the in situ phase Mineral loss ΔZ Baseline, vol% μm ΔZ Effect, vol% μm ΔΔZ, vol% μm SZ max, Baseline, SZ max, Effect, Lesion depth p vol% vol% L D Baseline, min min μm LD Effect, μm ΔLD, μm p 1,1 ppm F (NaF) Sound a 1,679 (99) 1,679 (99) a 68 a 72 (26) 72 (26) Lowly demineralized 3,31 (95) 4,219 (1,929) 1,189 (1,835) (19) 136 (48) 37 (5).2 Highly demineralized 5,27 (1,91) 5,835 (3,34) 564 (2,995) (22) 178 (48) 27 (68).175 1,1 ppm F (SnF 2 ) Sound a 1,98 (972) 1,98 (972) a 7 a 83 (52) 83 (52 Lowly demineralized (7-day lesion) 3,238 (1,34) 3,466 (2,9) 228 (2,696) (26) 124 (45) 17 (59).312 Highly demineralized (21-day lesion) 5,84 (1,31) 5,431 (2,79) 373 (2,942) (27) 169 (57) 1 (59).552 1,1 ppm F (NaF + SnF 2 combined groups) Sound a 1,794 (1,92) 1,794 (1,92) a 7 a 78 (32) 78 (32) Lowly demineralized (7-day lesion) 3,134 (871) 3,843 (1,755) 79 (1,862) (15) 13 (36) 27 (43).3 Highly demineralized (21-day lesion) 5,537 (882) 5,633 (2,834) 96 (2,666) (17) 174 (57) 18 (56).14 ppm F Sound a 4,176 (4,678) 4,176 (4,678) a 4 a 19 (65) 19 (65) Lowly demineralized (7-day lesion) 2,52 (86) 6,2 (3,77) 3,481 (3,827) (17) 156 (6) 62 (61).3 Highly demineralized (21-day lesion) 4,55 (833) 7,792 (4,881) 3,242 (4,752) (23) 184 (63) 49 (61).14 Data are presented as means (SD). Negative values for ΔZ, ΔΔZ as well as LD and ΔLD indicate demineralization, and positive values indicate remineralization. Values were only included when at least 1 specimen of a duplet of a participant could be analyzed; therefore n = 13. Values of the mean mineral density profiles were used to calculate SZ max. p values indicate significant differences in mineral loss before and after the in situ period. a ΔZ Baseline, SZ max and LD Baseline for sound surfaces were assumed to be zero. were cleaned and approximately 35 enamel blocks ( mm) were prepared (Exakt 3; Exakt Apparatebau, Norderstedt, Germany). After sterilization with ethylene dioxide the enamel blocks were embedded in epoxy resin (Technovit 471; Heraeus Kulzer, Hanau, Germany), ground flat and hand polished (4, grit, silicon carbide, Phoenix Alpha; Wirtz-Buehler, Düsseldorf, Germany). Before demineralizing (.6 μ M methylhydroxydiphosphonate, 3 m M CaCl 2, 3 m M KH 2 PO 4 and 5 m M acetic acid adjusted to ph 4.95 [Buskes et al., 1985]) the specimens for either 7 days (lowly demineralized lesions, n = 175) or 21 days (highly demineralized, n = 175), two thirds of the surface was covered with nail varnish (sound surface and sound control). Thus, each specimen was divided into three parts: one third of the specimen (sound control) was used as control to register a possible damage of the specimen during in vitro demineralization and the in situ period, one third of the specimen (sound surface) was exposed during the in situ period to analyze re- and demineralization characteristics of sound surfaces, and one third was demineralized (either for 7 or 21 days) to analyze re- and demineralization characteristics of lowly and highly demineralized lesions ( fig. 1 ). Thin plano-parallel sections (1 μm) were prepared perpendicular to the surface (Trennschleifsystem Exakt 3; Exakt Apparatebau) and polished (4, grit; Mikroschleifsystem Exakt). Baseline mineral loss and lesion depth were determined by means of TMR (transversal microradiography). For the in situ period 252 specimens with similar mineral loss for lowly demineralized (n = 126) and highly demineralized (n = 126) lesions were chosen ( table 1 ) and inserted into the appliances. 144 Transversal Microradiography After the in situ period thin plano-parallel sections were prepared again. Microradiographs of the enamel specimens were obtained and analyzed as described previously [Meyer-Lueckel et al., 215b]. For lowly and highly demineralized lesions changes in integrated mineral loss (ΔΔZ = ΔZ Baseline ΔZ Effect ) and lesion depth (ΔLD = LD Baseline LD Effect ) were calculated. For a more intuitive reading ΔΔZ and ΔLD for sound surfaces were calculated as well, although integrated mineral loss and lesion depth were measured only after the in situ period. For this ΔZ Baseline and LD Baseline were assumed to be zero. Furthermore, graphics of mean mineral density profiles were prepared for all groups with the TMR/WIM Calculation Program (v5.25; University of Groningen, Groningen, The Netherlands). Values of the mean mineral density profiles were used to calculate the maximum mineral density of the lesion surface zone (SZ max ) [Groeneveld and Arends, 1975; Lippert et al., 215 ]. Free Fluoride Analysis Fluoride concentrations of all dentifrices were measured (Orion Autochemistry System 96; Fisher Scientific, Ulm, Germany) using a calibrated ion-specific electrode (type 96-9 BNC; Fisher Scientific) [Cury et al., 23; Esteves-Oliveira et al., 211]. For low-level measurement four fluoride solutions (3.8, 1.9,.38 and.19 mg/l) were prepared. These concentrations narrowly bracket the expected sample concentrations. The electrode potentials (millivolts) of the four standard solutions were measured and plotted on the linear axis against their concentrations (milligrams/liter) on the log axis. After calibration the fluoride concentrations of the dentifrices were determined. For this, 2 mg of the dentifrices were diluted in 1 ml distilled water at room temperature. Then 4 ml of each solution were centrifuged at 2,5 g and 1 ml of the supernatant was added to 1 ml TISAB II (Fisher Scientific). The percentage of free fluoride in relation to the given total fluoride concentration (manufacturer s information) was determined. Two solutions for each dentifrice were prepared and analyzed in triplicate. Wierichs/Lausch/Meyer-Lueckel/ Esteves-Oliveira

5 A a ab b c c c B a b b a b b Sound Lowly demineralized (7-day lesions) Highly demineralized (21-day lesions) Sound Lowly demineralized (7-day lesions) Highly demineralized (21-day lesions) Color version available online ,1 ppm F ppm F Dentifrice 1,1 ppm F ppm F Dentifrice Fig. 2. Means with 95% confidence intervals of changes in mineral loss (ΔΔZ) and lesion depth (ΔLD). A Mineral loss. B Lesion depth. Sound surfaces presented more changes in ΔZ and lesion depth values than lowly and highly demineralized lesions; however, only for the changes in ΔZ of the ppm F intervention was there no significant difference between the different baseline substrate conditions. For all types of lesions, the toothpaste with ppm F shows significantly more mineral loss than that with 1,1 ppm F. Different letters indicate significant differences between treatments. p <.5, adjusted t test. Electron Microprobe Analysis The contents of MgO, Na 2 O, CaO, Cl, P 2 O 5 and SnO 2 being incorporated in the enamel surface of specimens in group SnF 2 / NaF were measured using a JEOL JXA 8,9 microprobe (JOEL, Eching, Germany) equipped with five WD spectrometers. An accelerating voltage of 15 kv, a probe current of 15 na and a probe diameter of about 5 μm were used. Characteristic X-rays were recorded using a TAP crystal for Mg and Na, a PETJ crystal for Cl and Ca and a PETH crystal for P and Sn. As standard materials spinel (Mg), jadeite (Na), tugtupite (Cl), apatie (P, Ca) and stannous oxide (Sn) were used. After the in situ period 3 specimens of intervention SnF 2 /NaF were prepared and analyzed in triplicate. The percentage of stannous in relation to given total ions was determined. The calculated 1-sigma detection limit of Sn is about 2 ppm (.2 wt%). Re- and Demineralization Characteristics of Enamel Statistical Analysis The data from the participants completing the study (n = 13; per-protocol analysis) were statistically analyzed using the SAS statistical software (SAS 9.2; SAS, Cary, N.C., USA). Variables were tested for normal distribution (Shapiro-Wilk test). For lowly and highly demineralized lesions differences in integrated mineral loss and lesion depth before and after in situ exposure were analyzed using the t test. Analysis of covariance for crossover design (AN- COVA) and the t test were used to detect differences in the changes in mineral loss (ΔΔZ) and lesion depth (ΔLD) between interventions. More technically, the ANCOVA statistical model may be described as a general linear mixed model with baseline, period and treatment as fixed effects and participant as a random effect. Correlations between ΔZ Baseline and ΔΔZ as well as between LD Baseline and ΔLD were assessed using the Spearman s rank correlation coefficient. All tests were performed at a 5% level of significance. Results Regarding the participants consumption of dentifrice no significant differences between NaF (1.6 g), SnF 2 (116. g) and fluoride-free (116. g) toothpastes were observed (p >.5, ANCOVA). Calculus formation or staining was also not detected for any participant. Of the participants, 6 did not complete the study and 2 were excluded due to protocol violation. Out of ethical reasons specified by the ethics committee of the Medical Faculty of the RWTH University, participants who decide to quit the study should be able to do it at any time (even after signing the informed consent) and without mentioning the reasons. A total of 13 specimens broke during preparation, resulting in 143 specimens that could be analyzed after the in situ phase. 145

6 146 Mineral loss (vol%) A Mineral loss (vol%) B Mineral loss (vol%) C ppm F: baseline 1,1 ppm F: baseline Depth (μm) ppm F: effect 1,1 ppm F: effect Fig. 3. Mean mineral density profiles of the enamel specimens before (baseline) and after the in situ period (effect) were assessed with the TMR/WIM Calculation Program. A Sound surfaces. B Lowly demineralized lesions (7-day lesions). C Highly demineralized lesions (21-day lesions). For all 3 substrates net demineralization can be observed after the in situ period. However, when fluoride dentifrice is used, the lowly demineralized lesions show greater mineral loss than the highly demineralized ones. It is also observable that for the highly demineralized lesions the surface of the lesion apparently becomes thicker and less defined when the fluoride dentifrice is used, probably partly obstructing (or slowing down) further demineralization. TMR: Changes in Mineral Loss Mean (SD) baseline mineral loss and lesion depth were 2,93 vol% μm (1,487) and 1 μm (31) for lowly demineralized (7-day) lesions, respectively, and 5,28 vol% μm (1,6) and 149 μm (34) for highly demineralized (21-day) lesions ( table 1 ). In general, independently of the intervention, sound surfaces presented the highest demineralization and highly demineralized lesions the lowest. Both toothpastes with 1,1 ppm F showed similar effects with no significant differences in ΔΔZ and ΔLD for all three baseline substrate conditions (p >.5, ANCOVA). Thus, for further analysis of the influence of the baseline or lesion characteristics on further de- or remineralization, the two toothpastes were combined in a category named 1,1 ppm fluoride intervention ( table 1 ). The reactions and the mean mineral profiles for the three baseline substrate conditions for interventions with (1,1 ppm F) and without ( ppm F) fluoride can be seen in figures 2 and 3. The lowly demineralized lesions tended to demineralize more than highly demineralized lesions, although no significant differences were observed regardless of the intervention (p >.5, adjusted t test ; fig. 2 ). Between sound surfaces and highly demineralized lesions significant differences could only be observed for the intervention 1,1 ppm F (p <.5, adjusted t test; fig. 2 ). Irrespectively of the baseline substrate condition specimens of group ppm F showed a significantly higher change in mineral loss (demineralization) than specimens of group 1,1 ppm F. A similar behavior was observed regarding lesion depth. Regardless of the interventions sound surfaces showed the highest increase in lesion depth, being significantly higher than for lowly and highly demineralized lesions (p <.5, adjusted t test). Highly demineralized lesions revealed the lowest increase in lesion depth. According to Spearman s rank correlation coefficient a low but significant correlation could be found between baseline mineral loss and change of mineral loss (r 1,1 ppm F =.436, p 1,1 ppm F <.1; r ppm F =.246, p ppm F =.8) as well as between baseline lesion depth and change of lesion depth (r 1,1 ppm F =.462, p 1,1 ppm F <.1; r ppm F =.436, p ppm F <.1) for both kinds of intervention. TMR: Mineral Density Profiles After the in situ period specimens of group 1,1 ppm F revealed a surface layer more mineralized than before the in situ period, whereas specimens of group ppm F revealed a surface layer less mineralized than before ( fig. 3 ; table 1 ). Before and after the in situ period specimens with Wierichs/Lausch/Meyer-Lueckel/ Esteves-Oliveira

7 highly demineralized lesions showed the least mineralized surface layer and specimens with sound surfaces showed the most mineralized surface layer ( fig. 3 ; table 1 ). Fluoride Analysis and Electron Microprobe Analysis The percentage of free/soluble fluoride in relation to given total fluoride concentration (SD) was 92.5% (4.6) [1,17.5 (5.6) ppm F] for NaF and 91.7% (7.9) [1,8.7 (86.9) ppm F] for SnF 2. In the fluoride-free dentifrice no free/soluble fluoride was measured. No stannous could be found in the lesions and in the sound surfaces. Discussion The present in situ study evaluated the re- and demineralization characteristics of sound enamel as well as lowly and highly demineralized caries-like enamel lesions in an established in situ caries model. The study hypothesis was partially rejected since for the samples treated with fluoride-free toothpastes no significant differences in the changes of mineral loss were observed for the different baseline substrate conditions (sound, lowly and highly demineralized). However, for the samples treated with the fluoride-containing toothpastes, indeed significantly more demineralization was observed for the sound surfaces than for highly demineralized lesions. Independently of the intervention, a trend could be observed for highly demineralized lesions to demineralize less than lowly demineralized lesions or sound surfaces. A low but significant correlation between baseline substrate condition and demineralization could be observed for all interventions. In our demineralizing in situ model the change in mineral loss and lesion depth decreased with increasing baseline values (more mineral loss). This reflects the increasing potential for demineralization with decreasing baseline mineral loss and lesion depth under demineralizing conditions. Since the design of an in situ model and the environment created by the model will have an overriding impact on its response (net demineralization or net remineralization) [Zero, 1995], the effect observed in the present study is presumed to be reversed under remineralizing conditions. Indeed, increasing potential for remineralization with increasing baseline mineral loss and lesion depth under remineralizing conditions has been observed [Strang et al., 1987]. Thus, in demineralizing in situ models lowly demineralized baseline lesions are presumably more applicable, whereas in remineralizing in situ models highly demineralized baseline lesions seem to be more applicable. Re- and Demineralization Characteristics of Enamel We hypothesized that the effect observed in the present demineralizing in situ study is related to the nature of the subsurface caries lesions. Thus, a more mineralized surface layer could be observed in the presence of fluoride than in its absence. For lowly and highly demineralized lesions this mineralized surfaces layer could presumably be built within a few days, whereas for sound samples this surface layer could only be built after initial demineralization [Arends and Christoffersen, 1986]. Once the mineralized surface layer has been established, it is a barrier firstly for the dissolution of mineral and secondly for the diffusion of acids into deeper parts of the lesions. This inhibiting mechanism increases when lesions becomes deeper [Arends and Christoffersen, 1986]. At the end it might even reach an equilibrium at which neither re- or demineralization occurs [ten Cate and Duijsters, 1982]. The previous findings are confirmed through the present results. In the presence of fluoride significantly more demineralization was observed for the sound surfaces than for highly demineralized lesions. Secondly, fewer minerals could be bound in the surface layer of highly demineralized specimens than in the sound ones, due to the hindered process of dissolution and incorporation of minerals, especially in initially larger lesions [Arends and Christoffersen, 1986]. Hence, a lower SZ max of the lesion surface was observed in highly demineralized specimens. Contrastingly, in the absence of fluoride these effects could not be observed. Here the degree of mineralization of the lesion surface layer for all three types of lesions was in the same order of magnitude, so no protective effect of the surface layer could be established (in the absence of fluoride), resulting in only slight differences (in ΔΔZ, ΔLD and SZ max ) between the various baseline substrate conditions. In addition to the impact of the lesion surface zone on the observed effects, the mineral density may have played a role in the different degrees of de- and remineralization effects observed in the present study. It has recently been shown that lesions with the same ΔZ Baseline but different LD Baseline, and consequently high or low R values, react differently under net remineralizing conditions. Lesions with lower mineral density (high R values) showed the highest reaction to remineralizing solutions [Lynch et al., 27]. Since the observed effect is presumably reversed under demineralizing conditions, this is in agreement with the results of the present study. The lowly demineralized lesions (low R value: 29, SD = 3) demineralized more than the highly demineralized lesions (high R value: 35, SD = 3) under demineralizing conditions. 147

8 Sound enamel as well as lowly demineralized (7-day demineralization; 1 μm) and highly demineralized (21- day demineralization; 15 μm) caries-like enamel lesions were used to analyze re- and demineralization characteristics. From a clinical point of view the lesions classified here as highly demineralized were only 15 μm deep and could be considered subclinical. Some of them would probably be difficult to detect in vivo [ten Cate et al., 28]. However, the initiation and progression of enamel caries lesions seems to be extremely slow [Bjarnason and Finnbogason, 1991], and fluoride seems to be the predominant factor influencing remineralization and arresting the process of subclinical caries lesions [Silverstone, 1982; ten Cate, 1984]. This anticaries effect is assumed to be limited to the outer 15 2 μm of the lesion (at deeper levels no difference was observed compared with placebo) [ten Cate and Rempt, 1986; White and Featherstone, 1987; Bjarnason and Finnbogason, 1991]. However, various types of deeper lesions formed in vitro and in vivo [Al-Khateeb et al., 22] should be analyzed in further studies, firstly to get more information on the tested agent, secondly to get more information on the model being used and thirdly to choose the perfect setting for the scientific issue under investigation (before analyzing the scientific issue under clinical conditions). In the present study the participants were instructed not to brush the specimens at any time to allow plaque to grow and to compare the re- and demineralization characteristics of enamel depending on baseline mineral loss and lesion depth of the caries challenge in the worst-case scenario. This is in contrast to other in situ studies analyzing baseline characteristics [Strang et al., 1987; Mellberg, 1991b; Lippert et al., 211] but might be a realistic sequence for proximal surfaces since many people do not clean these areas on a regular basis. Previous studies have also shown that this model seems preferable to simulating proximal caries [Hara et al., 23; Itthagarun et al., 25; Meyer-Lueckel et al., 27; Cochrane et al., 212; Meyer- Lueckel et al., 215a]. The best way to check the validity of an in situ model is to demonstrate a fluoride dose response similar to the anticipated clinical response. The more likely the model reflects the natural caries process, the more likely the model will be responsive to clinically proven cariostatic agents [Wefel, 1992]. Referring to the known dose-dependent clinical effects of dentifrices containing, 5 and 1,1 ppm F [Walsh et al., 21], this effect could recently be shown for the present in situ model [Meyer- Lueckel et al., 215a]. A fluoride dose-dependent effect was observed for dentifrices containing, 5 or 1, ppm F. Under the conditions tested here, again an effect of the same order could be observed for the dentifrice containing 1,1 ppm F, and significant differences in the change of mineral loss could be observed for all baseline substrate conditions compared with fluoride-free dentifrice. Regarding the lesion size this does not always seem to be the case, as a former in situ study observed much smaller differences in dentifrices containing between and 1,1 ppm F for highly than for lowly demineralized lesions after 14 days in situ [Mellberg, 1991b]. All former in situ studies evaluating baseline mineral loss and lesion depth in in situ re- or demineralization studies used predemineralized human enamel specimens [Strang et al., 1987; Schafer et al., 1992; Lippert et al., 211], except one study presumably using a mixture of predemineralized human and bovine specimens [Mellberg, 1991b]. Contrastingly, in the present study we used bovine enamel. Compared with human enamel bovine teeth are easier to obtain and mineralization patterns show lower variations, resulting in a more consistent experimental response [Mellberg, 1992; Kielbassa et al., 26]. However, bovine enamel is more porous, resulting in faster mineral changes and diffusion rates and consequently faster lesion formation [Edmunds et al., 1988]. Four bovine enamel specimens were inserted in the buccal flanges in the present in situ appliances and plaque accumulation was triggered by recessed specimens 1 mm below the acrylic under a plastic mesh. In contrast to former studies [Meyer-Lueckel et al., 27, 215b], in which 8 bovine enamel specimens were inserted, the size of the buccal flanges of the appliance was reduced in order to increase the wearing comfort. Although the reduced buccal flanges did not result in an increased risk of fracture, no additional comfort was diagnosed (e.g. by fewer pressure sores compared with former studies) or reported by participants. Thus, for further in situ studies buccal flanges with up to 8 specimens seem to be recommendable since more information can be gained simultaneously by using a second (different) substrate (e.g. dentine), a different brushing mode (e.g. brushing) or a different position (e.g. easily cleanable). SnF2 combines therapeutic effects against dental caries [Mellberg, 1991a], gingivitis [Beiswanger et al., 1995], sensitivity [Schiff et al., 26] and dental erosion [Huysmans et al., 211]. Nonetheless, it is very reactive and has to be prevented from being oxidized (to stannous IV) and consequently becoming insoluble and ineffective [Lippert, 213]. Therefore, several approaches to stabilize and to increase the effectiveness have recently been made. It has been combined with, for example, sodium fluoride Wierichs/Lausch/Meyer-Lueckel/ Esteves-Oliveira

9 [Huysmans et al., 211], sodium hexametaphosphate [Stookey et al., 24] or amine fluoride [Huysmans et al., 211], and even stannous chloride was combined with sodium fluoride [Faller et al., 214], generating an SnF 2 complex during toothbrushing. In the present study SnF 2 was stabilized with sodium gluconate, which acts as a chelator. Although former clinical studies showed that sodium gluconate-stabilized stannous fluoride could significantly reduce gingivitis and gingival bleeding compared with a sodium fluoride control [Beiswanger et al., 1995; Perlich et al., 1995], no in situ or clinical study is currently available with respect to its de- and remineralizing properties. In contrast to the first SnF 2 dentifrices, in which the amount of free/soluble fluoride was limited [Tinanoff, 1 995; White, 1995], the tested SnF 2 dentifrice provided the same amount of free/soluble fluoride as the NaF dentifrice. This is in accordance with the result of our TMR analysis, in which no significant difference between SnF 2 and NaF could be observed for sound surfaces or for lowly and highly demineralized lesions. Thus, sodium gluconate-stabilized stannous fluoride dentifrice apparently does not cause any drastic reduction in the total available fluoride of the dentifrice and might be an interesting fluoride compound. Besides reducing demineralization as much as the sodium fluoride-containing dentifrices, SnF 2 might also reduce gingivitis and gingival bleeding, as has previously been reported [Beiswanger et al., 1995; Perlich et al., 1995]. When using transversal microradiography to calculate integrated mineral loss and lesion depth, the quantification may be affected by the incorporation of polyvalent metal compounds (e.g. Sn) [Arndt et al., 26]. Stannous incorporated in the enamel or deposited on the surface as a precipitate could result in a falsely higher mineral content due to a higher absorption coefficient of Sn compared with hydroxyapatite [Arndt et al., 26]. This could be interpreted as enhanced remineralization. Nonetheless, the maximum mineral density of the lesion surface zones (SZ max ) indicated that Sn did not influence mineral quantification in the present investigation. This is confirmed by the results of the electron microprobe analysis. The maximum percentage of stannous in relation to total ions present in enamel was under the detection threshold (.2 wt%). It can be concluded that baseline mineral loss and lesion depth directly affect in situ lesion response. This relationship was not influenced by SnF 2, NaF or fluoridefree dentifrices. Treatment groups should therefore be well balanced with respect to baseline mineral loss and lesion depth. Particularly in demineralizing in situ models, lowly demineralized baseline lesions are presumably more applicable. Acknowledgments The study was financially supported by Procter & Gamble. Procter & Gamble also provided the dentifrices. We wish to thank Mr. M. Stiebritz and Mr. R. Klinghardt for their technical support. We would also like to thank the participants for participating in the study. Author Contributions H.M.-L. and R.J.W. designed and planned the study. R.J.W. and J.L. prepared the samples. R.J.W. performed the measurements and statistical analysis. R.J.W. and M.E.-O. wrote the manuscript. All authors revised the manuscript. Disclosure Statement The authors declare no conflicts of interest. References Al-Khateeb S, Exterkate RA, de Josselin de Jong E, Angmar-Mansson B, ten Cate JM: Light-induced fluorescence studies on dehydration of incipient enamel lesions. Caries Res 22; 36: Arends J, Christoffersen J: The nature of early caries lesions in enamel. J Dent Res 1986; 65: Arndt UW, Creagh DC, Deslattes RD, Hubbell JH, Indelicato P, Kessler EG Jr, Lindrothg E: X-rays; in: International Tables for Crystallography. Chester, IUCr, 26, vol C, chapt 4.2, pp Re- and Demineralization Characteristics of Enamel Beiswanger BB, Doyle PM, Jackson RD, Mallatt ME, Mau M, Bollmer BW, Crisanti MM, Guay CB, Lanzalaco AC, Lukacovic MF, et al: The clinical effect of dentifrices containing stabilized stannous fluoride on plaque formation and gingivitis a six-month study with ad libitum brushing. J Clin Dent 1995; 6: Bjarnason S, Finnbogason SY: Effect of different fluoride levels in dentifrice on the development of approximal caries. Caries Res 1991; 25: Buskes JA, Christoffersen J, Arends J: Lesion formation and lesion remineralization in enamel under constant composition conditions. A new technique with applications. Caries Res 1985; 19: Cochrane NJ, Shen P, Byrne SJ, Walker GD, Adams GG, Yuan Y, Reynolds C, Hoffmann B, Dashper SG, Reynolds EC: Remineralisation by chewing sugar-free gums in a randomised, controlled in situ trial including dietary intake and gauze to promote plaque formation. Caries Res 212; 46:

10 Cury JA, Francisco SB, Simoes GS, Del Bel Cury AA, Tabchoury CP: Effect of a calcium carbonate-based dentifrice on enamel demineralization in situ. Caries Res 23; 37: Dijkman A, Huizinga E, Ruben J, Arends J: Remineralization of human enamel in situ after 3 months: the effect of not brushing versus the effect of an F dentifrice and an F-free dentifrice. Caries Res 199; 24: Edmunds DH, Whittaker DK, Green RM: Suitability of human, bovine, equine, and ovine tooth enamel for studies of artificial bacterial carious lesions. Caries Res 1988; 22: Esteves-Oliveira M, Pasaporti C, Heussen N, Eduardo CP, Lampert F, Apel C: Rehardening of acid-softened enamel and prevention of enamel softening through CO 2 laser irradiation. J Dent 211; 39: Faller RV, Eversole SL, Saunders-Burkhardt K: Protective benefits of a stabilised stannouscontaining fluoride dentifrice against erosive acid damage. Int Dent J 214; 64(suppl 1): Groeneveld A, Arends J: Influence of ph and demineralization time on mineral content, thickness of surface layer and depth of artificial caries lesions. Caries Res 1975; 9: Hara AT, Queiroz CS, Paes Leme AF, Serra MC, Cury JA: Caries progression and inhibition in human and bovine root dentine in situ. Caries Res 23; 37: Huysmans MC, Jager DH, Ruben JL, Unk DE, Klijn CP, Vieira AM: Reduction of erosive wear in situ by stannous fluoride-containing toothpaste. Caries Res 211; 45: Itthagarun A, King NM, Yiu C, Dawes C: The effect of chewing gums containing calcium phosphates on the remineralization of artificial caries-like lesions in situ. Caries Res 25; 39: Kielbassa AM, Hellwig E, Meyer-Lueckel H: Effects of irradiation on in situ remineralization of human and bovine enamel demineralized in vitro. Caries Res 26; 4: Koulourides T, Phantumvanit P, Munksgaard EC, Housch T: An intraoral model used for studies of fluoride incorporation in enamel. J Oral Pathol 1974; 3: Lippert F: An introduction to toothpaste its purpose, history and ingredients. Monogr Oral Sci 213; 23: Lippert F, Churchley D, Lynch RJ: Effect of lesion baseline severity and mineral distribution on remineralization and progression of human and bovine dentin caries lesions. Caries Res 215; 49: Lippert F, Lynch RJ, Eckert GJ, Kelly SA, Hara AT, Zero DT: In situ fluoride response of caries lesions with different mineral distributions at baseline. Caries Res 211; 45: Lynch RJ, Mony U, ten Cate JM: Effect of lesion characteristics and mineralizing solution type on enamel remineralization in vitro. Caries Res 27; 41: Manning RH, Edgar WM: Intra-oral models for studying de- and remineralization in man: methodology and measurement. J Dent Res 1992; 71: Mellberg JR: Fluoride dentifrices: current status and prospects. Int Dent J 1991a;41: Mellberg JR: Relationship of original mineral loss in caries-like lesions to mineral changes in situ. Short communication. Caries Res 1991b;25: Mellberg JR: Hard-tissue substrates for evaluation of cariogenic and anti-cariogenic activity in situ. J Dent Res 1992; 71: Mellberg JR, Petrou ID, Grote NE: A study of the ability of an in situ remineralization model to differentiate between the effects of two fluoride dentifrices that produced significantly different clinical caries results. J Dent Res 1992; 71: Meyer-Lueckel H, Bitter K, Kielbassa AM: Effect of a fluoridated food item on enamel in situ. Caries Res 27; 41: Meyer-Lueckel H, Wierichs RJ, Gninka B, Heldmann P, Dorfer CE, Paris S: The effect of various model parameters on enamel caries lesions in a dose-response model in situ. J Dent 215a;43: Meyer-Lueckel H, Wierichs RJ, Schellwien T, Paris S: Remineralizing efficacy of a CPP-ACP cream on enamel caries lesions in situ. Caries Res 215b;49: Perlich MA, Bacca LA, Bollmer BW, Lanzalaco AC, McClanahan SF, Sewak LK, Beiswanger BB, Eichold WA, Hull JR, Jackson RD, et al: The clinical effect of a stabilized stannous fluoride dentifrice on plaque formation, gingivitis and gingival bleeding: a six-month study. J Clin Dent 1995; 6: Schafer F, Raven SJ, Parr TA: The effect of lesion characteristic on remineralization and model sensitivity. J Dent Res 1992; 71: Schiff T, He T, Sagel L, Baker R: Efficacy and safety of a novel stabilized stannous fluoride and sodium hexametaphosphate dentifrice for dentinal hypersensitivity. J Contemp Dent Pract 26; 7: 1 8. Schirrmeister JF, Seger RK, Altenburger MJ, Lussi A, Hellwig E: Effects of various forms of calcium added to chewing gum on initial enamel carious lesions in situ. Caries Res 27; 41: Silverstone LM: The effect of fluoride in the remineralization of enamel caries and carieslike lesions in vitro. J Public Health Dent 1982; 42: Stookey GK, Mau MS, Isaacs RL, Gonzalez-Gierbolini C, Bartizek RD, Biesbrock AR: The relative anticaries effectiveness of three fluoridecontaining dentifrices in Puerto Rico. Caries Res 24; 38: Strang R, Damato FA, Creanor SL, Stephen KW: The effect of baseline lesion mineral loss on in situ remineralization. J Dent Res 1987; 66: ten Cate JM: The effect of fluoride on enamel deand remineralization in vitro and in vivo; in Guggenheim B (ed): Cariology Today. Basel, Karger, 1984, pp ten Cate JM: In situ models, physico-chemical aspects. Adv Dent Res 1994; 8: ten Cate JM, Buijs MJ, Miller CC, Exterkate RA: Elevated fluoride products enhance remineralization of advanced enamel lesions. J Dent Res 28; 87: ten Cate JM, Duijsters PP: Alternating demineralization and remineralization of artificial enamel lesions. Caries Res 1982; 16: ten Cate JM, Rempt HE: Comparison of the in vivo effect of a and 1,5 ppm F MFP toothpaste on fluoride uptake, acid resistance and lesion remineralization. Caries Res 1986; 2: Tinanoff N: Progress regarding the use of stannous fluoride in clinical dentistry. J Clin Dent 1995; 6: Walsh T, Worthington HV, Glenny AM, Appelbe P, Marinho VC, Shi X: Fluoride toothpastes of different concentrations for preventing dental caries in children and adolescents. Cochrane Database Syst Rev 21; 1:CD7868. Wefel JS: Effects of fluoride on caries development and progression using intra-oral models. J Dent Res 199; 69: ; discussion Wefel JS: Consensus conference on intra-oral models: patient selection, appliance design, and substrate. J Dent Res 1992; 71: 954. White DJ: A return to stannous fluoride dentifrices. J Clin Dent 1995; 6: White DJ, Featherstone JD: A longitudinal microhardness analysis of fluoride dentifrice effects on lesion progression in vitro. Caries Res 1987; 21: Zero DT: In situ caries models. Adv Dent Res 1995; 9: ; discussion Wierichs/Lausch/Meyer-Lueckel/ Esteves-Oliveira