Fluoride release from fluoride varnishes exposed to commonly consumed beverages: An in vitro study

Background The aim of this study was to compare the in vitro fluoride release from fluoride varnishes exposed to commonly consumed beverages. Material and Methods One hundred and twenty acrylic blocks were divided randomly into ten experimental groups (n= 12 per group). For the experiment, 24 blocks were prepared for each fluoride varnish (Duraphat®, Duofluorid XII®, Clinpro™, MI Varnish™ and Profluorid®). The blocks were placed into artificial saliva for 30 minutes and in carbonated beverage or fruit juice for up to 24 hours. Artificial saliva and beverages were evaluated to determine the fluoride release using an ion-selective electrode. Data were analyzed using ANOVA F, Friedman and Kruskal Wallis test for bivariate analysis; and three-way ANOVA (fluoride varnish, beverages, exposure time). Results When the fluoride varnishes were compared according to each exposure time, a statistically significant difference was found between all the fluoride varnishes for each evaluation time on carbonated beverage and fruit juice. MI Varnish™ presented the highest fluoride release in carbonated beverage (94.44±5.47ppm) and fruit juice (126.16±8.89ppm) at 8 hours. Duraphat® presented the lowest fluoride release at baseline (0.44±0.08ppm) for carbonated beverage group. A three-way comparison between fluoride release, exposure time and fluoride varnish were statistically significant (p<0.001). When evaluating the effect of the three independent variables together on fluoride release, it was found that the variables fluoride varnish (p<0.001) and exposure time (p<0.001) contributed to the release of fluoride. Conclusions The type of fluoride varnish and the time after the application contribute to the fluoride release model. Key words:Fluorides, topical, sodium fluoride, beverages.


Introduction
Sodium fluoride is used in dentistry in various fluoride materials as a sensitivity treatment, as well as in dental caries prevention through topical applications, especia-lly in children and teenagers. In 2018, the U.S. Food and Drug Administration (FDA) approved fluoride varnishes (FV) for anticaries treatment (1). Also, the recommendation of FV for individuals at risk of dental caries has been described in clinical practice guidelines and health care programs around the world (2)(3)(4). In addition, FV are a modality of topical fluoride, a term that describes delivery systems that provide fluoride to dental surfaces exposed to them, at high concentrations for a local protective effect (5). Since FV have an important role in caries prevention and is commonly used in dental practice, it is important to determine the concentration of fluoride that varnishes released. On the other hand, it is highly common the massive consumed of beverages and food with acidic pH level from children and teenagers (6). Moreover, manufacturers do not mention the type of diet the patient should consumed after FV application. Manufacturer's instructions suggest time in which food should not be consumed and the recommended hours of a soft diet however, there are no instructions related to the pH of the diet post application (7). Hence, information that describes if an acidic condition could affect fluoride release (FR) it is necessary. The condition to which FV is exposed could affect their FR. However, while FR has been extensive studied from different varnishes, in vitro and in vivo (8)(9)(10)(11)(12)(13)(14), there is not enough information whereas if the acidic condition modifies the release (7). Moreover, there has not been studies to determine if non-alcohol beverages could affect FR, since these commercial drinks have as part of their composition citric acid, phosphoric and malic acid (15). Thus, the aim of this study was to compare the in vitro release of fluoride from FV exposed to commonly consumed beverages.

Material and Methods
Details about the composition of evaluated FV and their manufacturer´s dietary instructions related to beverages are on Table 1  juice was used conducted at 5 % significance level and a 95% power, from the pilot study. The pilot study was based on the final sample size of a previous study (n=3) (7) obtaining a sample of 1 specimen per group. One hundred and twenty acrylic (methyl methacrylate) blocks (25mmx25mmx9mm) were used as study substrates and divided randomly into ten experimental groups (n= 12 per group): G1: 5% NaF (gold standard: Duraphat® Colgate-Palmolive, Hamburg, Germany) exposed to carbonated beverage G2: 6% NaF + 6% CaF2 (Duofluorid XII® FGM, Joinville, Brazil) exposed to carbonated beverage G3: 5% NaF + TCP (Clinpro™ 3M ESPE™, Saint Paul, USA) exposed to carbonated beverage G4: 5% NaF + CPP-ACP (MI Varnish™ GC®, Tokyo, Japan) exposed to carbonated beverage G5: 5% NaF (Profluorid® Voco, Cuxhaven, Germany) exposed to carbonated beverage G6: 5% NaF (gold standard: Duraphat®) exposed to fruit juice G7: 6% NaF + 6% CaF2 (Duofluorid XII®) exposed to fruit juice G8: 5% NaF + TCP (Clinpro™) exposed to fruit juice G9: 5% NaF + CPP-ACP (MI Varnish™) exposed to fruit juice G10: 5% NaF (Profluorid®) exposed to fruit juice The blocks were cleaned with deionized water and the FV were applied in a smooth surface of each acrylic block. The amount of FV applied in each block was determined after measuring the unit-dose packages using an analytical balance (ED 224S, Sartorius GA, Göttingen, Germany). It was determined to use 40mg in each acrylic block using the manufacture's applicator for Clinpro™ and MI Varnish™, and a similar brush for the other three FV according to the manufacturer's instructions. For the FV in individual unit-dose package, they were thoroughly mixed prior to use with the manufac-ture's applicator for ten seconds to homogenize the FV since components of sodium fluoride varnishes can separate during storage (7). For the experiment, 24 blocks were prepared for each FV: half of the acrylic blocks for carbonated beverage (Coca-Cola®) and the other half for fruit juice (Pulp®). To measure the pH level of both beverages, a pH benchtop meter (inoLab® pH 7310, WTW™ Xylem Analytics, Weilheim, Germany) was used and the measurements were in triplicate immediately after opening at 20°C, the mean of those lectures were recorded as the pH level (15). The main characteristics of the beverages are showed in Table 2.

3.61
Artificial saliva Calcium (0.1169 g of calcium hydroxide/L of deionized water); 0.9 mM of phosphorus and potassium (0.1225 g potassium phosphate monobasic/L of deionized water); 20 mM TRIS buffer (2.4280 g TRIS buffer/L of deionized water). After the FV application, each block was submerged individually in 30ml of artificial saliva in a 100ml plastic beaker (16). The beakers were placed onto an orbital shaker at 100 rpm for 30 minutes. Then, the blocks were removed from the beakers and the artificial saliva was preserved for the measure of fluoride concentration (7). Immediately after, each block was fully submerged in the beakers with the beverages. Half of the blocks were exposed to 30ml of carbonated beverage (Coca-Cola®; pH 2.57) and the other half were exposed to 30ml of fruit juice (Pulp®; pH 3.61), the beakers were placed onto the orbital shaker at 100 rpm for 30, 420 (7 hours) and 960 minutes (16 hours). The beverages were decanted in a plastic container and renewed at 60 minutes (1 hour) and 480 minutes (8 hours) the FV application (7,13).

7.0
-Measure the fluoride release The FR was determined using an ion-selective electrode (9609 BNWP-Orion Research Inc., Beverly, MA, USA) coupled with a benchtop meter (Orion Star A214, Orion Research Inc.). The calibration of the electrode was performed using a calibration curve with standard solutions of known fluoride concentrations: 0.1, 1,10, 40 and 100 ppm obtained from a standard solution of 100 ppm F-(Thermo Scientific Orion) (13,17).
The values for the standard curve were obtained from 750 µL of Total Ionic Strength Adjustor Buffer (TISAB II, Thermo Scientific Orion) and 750 µL of the fluoride standard solutions. The fluoride concentration was calculated using a linear regression equation between de electrical conductivity (mV) and the logarithm of the fluoride concentration of the standard solutions (10,17).
A new calibration was performed before reading the solutions (mean r2=0.9999) and checked every hour with solutions of know fluoride concentrations: 1, 2 and 10 µg F-/mL (Thermo Scientific Orion) with a 2.41% of coefficient of variation between the expected and the calculated concentrations through the whole experiment. The aliquots of each solution: artificial saliva, carbonated and fruit juice were mixed with TISAB II in a 1:1 proportion and the concentrations were obtained compared to the standard curve, by transforming the mV to ppm. The samples were read duplicate and the mean of the two reading was recorded as the fluoride concentration (ppm) (7,17

Results
For the carbonated beverage, the FR increased from 30 minutes at baseline on artificial saliva to 8 hours for Du-raphat®, Duofluorid XII® and MI Varnish™; time after which the FR decreased. On the other hand, the FR increases numerically from 30 minutes on artificial saliva to 24 hours for Clinpro™. Profluorid® also had a different tendency, with a decrease on FR from 30 minutes on artificial saliva to 1 hour, an increased from 1 to 8 hours and a final decrease on FR to 24 hours (Fig. 1). On fruit juice, FR for MI Varnish™ was similar to carbonated beverage. The FV Duraphat®, Clinpro™ and Profluorid® showed a decreased on FR from 30 minutes at baseline on artificial saliva to 1 hour, then an increased was showed from 1 to 24 hours. Duofluorid XII® presented a similar tendency than the one presented by Profluorid® exposed to carbonated beverages (Fig. 2). When the FV were compared according to each time of exposure, a statistically significant difference was found between all the FV for each of the evaluation times on carbonated beverages (Table 3) and fruit juice ( Table  4). A three-way comparison between FV, exposure time and commonly consumed beverage FV was statistically significant (p<0.001). When evaluating the effect of the three independent variables (FV, beverages and exposure times) together on the release of fluoride, it was found that the variables FV (p<0.001) and exposure time (p<0.001) were those that contributed to the release of fluoride, whereas the type of beverage was not statically significant (p=0.458).

Discussion
The present in vitro study shows that each FV vary its FR in artificial saliva and commonly consumed beverages depending on the exposure time.   On the other hand, Clinpro™, had their peak fluoride release at 24 hours on carbonated beverage (5.12±1.13 ppm F-). The same happened on fruit juice to Dura-phat® (1.86±0.21 ppm F-), Clinpro™ (4.18±1.05 ppm F-) and Profluorid® (8.20±0.48 ppm F-). In an in vitro study by Chhatwani et al., one of the studied varnishes, Fluor Protector S, showed a peak fluoride release at time 1 (after 20 days of thermal cycling). It can be assumed that Fluor Protector can deliver a high initial level of fluoride release to the surrounding oral cavity environment, but thereafter the fluoride levels fall to a minimum (23). This is similar to our findings where the FR was statistically different for exposure time according to each varnish. In spite of the results from FR, it cannot be concluded that this is a measure to determine cariostatic effect of a FV. This because the primary action of fluoride in caries prevention is more closely related to its presence in the fluid phases of the oral cavity, where fluoride must be constantly present at low concentrations (13). Furthermore, based on their results, Dehailan et al. described that FR cannot be used as a predictive measure for enamel fluoride uptake which is why one should be careful in relating FR to FV efficacy (20). Also, an in situ study by Attin et al. made clear that within the first day after application of a highly concentrated fluoride varnish (Duraphat® -KOH soluble) fluoride deposition in close vicinity to the fluoridation occurs. This means that the preventive action of fluoride varnish is limited to the area to which varnish has been applied (24). Lippert concluded in a previous in vitro study that the pH level of the solutions in which FV were submerged affects the FR. According to this study, some FV were prone to low pH fluoride loss compared to FR in artificial saliva (pH level = 7.0) (7). However, in the present study the beverages used, with a pH level of 2.57 and 3.61, did not influenced in the FR of the tested FV. This could be related to the solutions used, Lippert used anhydrous citric acid with pH levels of 2.27 y 3.75 (7), while we used commonly consumed commercial beverages. The ingredients of the beverages, such as phosphoric acid and acidity regulators, could also influenced in this difference as well as the exposure time evaluated. Clinically, FV post-application instructions advice to not brush the teeth surface for approximately 24 (14).
Low pH beverages, such as carbonated beverages and fruit juices, are highly consumed, especially by children and teenagers (15,25,26). That is why it is necessary a better understanding of FR from FV under various acidic conditions. Despite, the lack of statistically difference in this study, more experimentations must be done to determine if other formulas or a bigger difference between the pH level of the beverages might modify the outcomes. Furthermore, fluoride can be toxic only in highly elevated concentrations (5mg/kg) (27) for this reason, FV does not pose a risk of toxicity (7). However, the fluoride present in other products such as toothpaste, juices, cereal, salt and more (28) should also be taken into consideration in the risk of fluoride exposure. Though, it is necessary to determine if beverages different from the ones tested could affect the FR. This is essential, since, undoubtedly, it will be the patient who will ingest these concentrations. In addition, excessive fluoride intake has been linked to detrimental health effects such as the development of dental and skeletal fluorosis, increased bone fractures, and deficits in cognitive development in children (28)(29)(30).

Conclusions
The present study shows that the in vitro FR had a significant difference related to each type of FV and expo-sure times. Thus, it is concluded that the type of FV and the time after the first FV application contribute to the fluoride release model. Nevertheless, this study presents limitations due to the fact that it is an in vitro experimentation and it only intended to mimic what occurs clinically. For this reason, clinical studies are necessary to confirm the results of this in vitro research.