Fluoride Concentrations in Typical Brazilian Foods and in Infant Foods

OBJECTIVE: To determine fl uoride concentrations in the typical Brazilian meal (rice with beans) and in processed infant foods, and to estimate their contribution towards dental fl uorosis. METHODS: The foods were purchased at supermarkets in the cities of Piracicaba and Campinas, Southeastern Brazil. The processed infant foods were bought in 2001 and the rice and beans in 2003, and they were analyzed immediately. Three brands of rice, three brands of beans and 36 samples of infant foods were analyzed, divided into fi ve groups: ready-to-eat, porridges, formulated foods, powdered milk and others. For the rice and beans, fl uoride concentrations were determined in the raw grains and after they were cooked with fl uoridated (0.7 ppm) or distilled water. All the fl uoride analyses were performed using a specifi c electrode. A dose of 0.07 mg/kg/day was considered to be the upper limit of fl uoride exposure in terms of fl uorosis risks. RESULTS: The fl uoride concentrations found in the grains of rice and beans were low. However, they increased 100 to 200-fold after cooking in fl uoridated water. Even so, they were lower than what is found in some processed foods. A meal of rice and beans prepared with fl uoridated water would be responsible for 29% of the threshold dose for fl uoride intake in terms of acceptable fl uorosis; the contribution from some processed foods reaches 45%. CONCLUSIONS: The typical Brazilian food, even when prepared with fl uoridated water, is safer in terms of the risk of dental fl uorosis than are some processed infant foods.


INTRODUCTION
Fluoride use has been the main public health measure for controlling dental caries both in developed and in developing countries. 3,15 However it has led to an increase in dental fl uorosis. 1,5 Dental fl uorosis is a defi ciency in enamel mineralization due to daily fl uoride intake during tooth development. 7 Since the dose-effect relationship is not precisely known, the dose of 0.07 mg F/day/kg of body weight has been accepted as the upper limit in terms of the clinically acceptable risk of dental fl uorosis. 2 The main sources of fl uoride that have been associated with increased dental fl uorosis are fl uoridated water, fl uoridated supplements, fl uoridated dentifrices and processed baby foods consumed before the age of six years. 14 Although the consumption of processed foods is low in developed countries, their contribution towards the risk of dental fl uorosis has been extensively studied in these countries, 11 but little is known in developing countries. 4 Besides the consumption of processed foods, the contribution of the typical diet of each country towards the risk of dental fl uorosis should be taken into account. As regards diet, data from two Brazilian cities with fl uoridated water has shown that diet contributes 31-44% to the daily intake of fl uoride by children aged 2-3 years. 17 Rice with beans (in the approximate proportion of 2:1) is the typical Brazilian dish, and its contribution towards daily fl uoride consumption, per se, or when cooked with fl uoridated water, is unknown. A low fl uoride concentration was found in Brazilian beans bought on the Japanese market, 16 but the sample chosen was unrepresentative of the product that the Brazilian population consumes, and the accompanying rice was not analyzed.
Although dental fl uorosis is the only side-effect attributed to exposure to low fl uoride concentrations such as water fl uoridation, 13 other effects due to exposure to multiple sources should not be ignored. Therefore, the present study aimed at analyzing the fl uoride concentrations in rice, beans and processed baby foods sold in Brazil, and to estimate their contributions towards the development of dental fl uorosis.

METHODS
Six 1 kg packs of rice grains and beans of the three brands most sold in the stores in Piracicaba, State of São Paulo, Brazil, were bought at supermarkets. The processed baby foods were bought in the city of Campinas, State of São Paulo, Brazil, and were grouped into fi ve categories: ready-to-eat foods (N=9), foods for porridge (N=11), baby formulas (N=9), powdered milk (N=4) and other foods (N=3). For each food, within each category, two to six samples were bought, in up to three commercial stores. The brand names of the products were omitted and were replaced by codes in the results tables, since the aim of this study was not to compare the products within the same category but between categories.
For analysis of the raw grains, these were ground up in a domestic food processor until homogeneous powders were obtained. Triplicate samples of 600 mg were weighed (± 0.01 mg) on petri dishes (Falcon 1007), to which 2.4 mL of deionized water was added.
To analyze the fl uoride in cooked rice and beans, six samples of each brand of grain were cooked in deionized or fl uoridated water (0.7 ppm F). The cooking conditions were standardized, particularly the ratio between grain weight and water volume (weight/volume, w/v), which was 1:2 and 1:4, respectively for the rice and beans. The cooked foods were homogenized in a blender with deionized water in the proportions (w/v) of 1:2 for the rice and 1:1 for the beans. A volume of 3.0 mL of each homogenized food was transferred to petri dishes.
Triplicate samples of each processed foods were weighed (± 0.01 mg) in petri dishes to which 3.0 ml of deionized water was added.
The samples of rice and beans (raw and cooked) and the samples of processed foods placed in the petri dishes were subjected to microdiffusion facilitated by hexamethyldisiloxane (HMDS), 21 to separate and subsequently analyze the fl uoride present.
Before sample weighing, a cap from a plastic tube (Falcon 2030) was placed using vaseline in the center of a plastic petri dish. After sample weighing, 0.10 mL f 1.65N NaOH 12 was placed in the cap. The petri dish was sealed with vaseline and 1.0 mL of 6N HCl saturated with HMDS added to the sample through a hole made in the petri dish cap. The hole was sealed with vaseline and the assemblage was shaken at room temperature. After 12-18h the Falcon plastic cap containing fl uoride diffused from the sample was put to dry at 60°C. This cap was then put into a plastic assay tube (Falcon 2017) containing 0.40 mL of 0.66N acetic acid. The tube was inverted and vigorously shaken to dissolve the fl uoride crystals present in the cap. The fl uoride concentration was determined by means of a specifi c electrode (Orion Research Inc., Model 96-09; Boston, United States) coupled to an ion analyzer (Orion Research Inc., Model EA 940, Boston, United States). Standard solutions (Orion #940907) in triplicates, at concentrations from 0.05 to 2.5 mg F/mL, were also acid-diffused. Likewise, blanks were subjected to this procedure.
With the aim of ensuring accurate determination, increasing quantities of powdered grains were subjected to acid-diffusion and the degree of linearity between the weight and fl uoride concentration was determined. Thus, it was found that, for the raw foods, 600 mg was the ideal weight of powdered grains for use in the analysis, since this quantity in 3.0 mL of water enabled adequate homogenization during microdiffusion and the quantity of fl uoride extracted satisfi ed the sensitivity limit of the electrode.
For the cooked foods, standardization was done with homogenized preparations using different proportions of water (1:1 to 1:9; w/v). The best results were 1:1 for the beans (r=1.00) and 1:2 (r=0.99) for the rice.
Due to the very low concentrations of fl uoride in the group of processed foods, the maximum amount of sample to be weighed was also standardized, which would enable sample homogenization and accurate fl uoride determination. It was found that a mean of 103% of the fl uoride present in the sample was recovered.
Taking into consideration the upper limit of 0.07 mg F/kg/day for the risk of fl uorosis, 3 the amount of food intake per day was estimated. The amount of rice and beans consumed by 27 children, aged from 20 to 30 months, who were attending the São Vicente de Paulo public daycare center in Piracicaba, State of São Paulo, was determined by the duplicated-diet technique. 12 The mean amounts (±SD) found were 119.1 ± 17.0 and 104.6 ± 14.8 g/day of rice and beans, respectively.
For the group of processed foods, the estimates were based on the recommended intake indicated on the product package, considering two-year-old children weighing approximately 13.2 kg.
The results found from the raw rice and bean samples and from the cooked beans were analyzed using the Kruskal-Wallis test and the cooked rice data were analyzed using the Tukey test. For all statistical analyses, the Bioestat * software was used and the signifi cance limit was set at 5%. The results from the processed foods were analyzed descriptively.

RESULTS
The fl uoride concentrations in the rice grains of the different brands analyzed differed statistically (p<0.05), and were highest in the A brand rice (Table 1), but among the bean samples no signifi cant difference was found (p>0.05).  The fl uoride concentrations in the powdered milk porridges and other foods (Table 5) were low, the porridge A presented the higher fl uoride concentration (1.63-0.04 mg F/kg). Their contribution towards undesirable dental fl uorosis doses would be lower than 10%.

DISCUSSION
The solid and liquid diet is a signifi cant source for the daily intake of fl uoride in both developed 19 and developing countries. 17 It must therefore be considered to be risk factor for dental fl uorosis. 14 Fluoride has widespread distribution in nature, and this makes it diffi cult to identify its origin in the diet. Its presence in the diet depends on regional characteristics, such as the typical dish of each population. There are meals were prepared with fl uoridated water, the fl uoride concentrations were much higher, and it was estimated that a meal with rice and beans would contribute 11% to 13% and 16% to 17%, respectively, to the threshold dose for the risk of dental fl uorosis ( Table 2).
The fl uoride concentrations in the ready-to-eat meals were very low, and their contributions towards the risk of dental fl uorosis were less than 2% (Table 3). Table 4 shows that the fl uoride concentrations in the formulated baby foods and milk-based foods would contribute 6 to 18% towards the threshold dose for the risk of dental fl uorosis. On the other hand, in the soybean-based formulas, higher concentrations of fl uoride were found, and some products could contribute up to 45% of the threshold dose of daily fl uoride intake in terms of the clinically acceptable dental fl uorosis.  some foods that are recognized as being rich in fl uoride, such as black tea-based beverages. 8 The fl uoride concentrations found in raw grains of rice and beans were low, and those in beans confi rmed what had been observed in Brazilian seeds sold on the Japanese market. 16 Investigation of the reasons for the variability between brands was not the purpose of this study, but such variability may be related to fertilization, cultivation medium, plant type or irrigation with water containing natural fl uoride.
Although the analysis of fl uoride concentration in the grains is important, it is even more important to correlate the effect of the cooking water towards increasing the fl uoride concentrations in foods to be consumed, since some foods may uptake fl uoride more than others. In addition, some foods are consumed in a more liquid form than others. Therefore, the amount of fl uoride intake will depend on the fi nal concentration in the product and on the conditions under which the food is cooked.
The fl uoride concentrations determined from raw rice and bean samples (Table 1) showed that, when these grains are cooked in non-fl uoridated water (Table 2), they do not cause concern regarding the suggested limit for an acceptable level of dental fl uorosis. 2 Thus, the mean fl uoride concentrations in rice and beans prepared with non-fl uoridated water were 0.044 ± 0.047 μg F/g and 0.149 ± 0.131 μg F/g respectively, so that a rice and bean-based meal would subject children to a dose of only 0.0020 mg F/kg of weight per day (Table 2). However, when these foods were prepared with fl uoridated water, the dose increased tenfold (to 0.020 mg F/kg of weight per day). On the other hand, the quantity of fl uoride consumed by children through rice and beans cooked in optimally fl uoridated water is completely safe in terms of fl uorosis, since it corresponds to 29% of the upper limit of the acceptable dose in relation to the risk of dental fl uorosis. 2 However, some of the processed baby foods (Table 4) are of concern in terms of the risk of dental fl uorosis. 14 The data from the present study have confi rmed that soybean-based foods 22 could contribute up to 50% of the maximum limit of the dose in terms of clinically acceptable fl uorosis. This concern about baby formulas dates back to 1979. There have been attempts in the United States to reduce the fl uoride concentration in these foods, which however have had little infl uence on reducing the risk of developing dental fl uorosis, 18 particularly in regions with fl uoridated water.
Low fl uoride concentrations were found in vegetable and fruit-based ready-to-eat foods (Table 3), thus concording with fi ndings from products on the American market. 20 In beef or chicken-based products (data not shown), lower concentrations than in products on the American market were found. 10 High fl uoride concentration are often found in foods because of the inclusion of bone fragments in the product, which are relatively rich in fl uoride. Neither the fruit-based products nor the vegetable-based products containing chicken or meat were of concern with regard to being a source of fl uoride and a risk of dental fl uorosis (Table 3).
In the formulated foods (Table 4), the fl uoride concentrations ranged from 0.47 to 3.84 μg F/g, and this variability could be explained mainly by the water that was used in the processing. 22 The mean concentration found for milk-based formulas was 0.87 ± 0.48 μg F/g, and this result was higher than what was found by other authors. 10,22 For the soybean-based foods (Table 3), the mean fl uoride concentration found in the products on the Brazilian market (2.368 ± 1.206 μg F/g) was lower than in products in the United States. 22 The fl uoride concentrations found in the cereals group (Table 5) were in agreement with what has been observed on the Brazilian market. 4 The fl uoride concentrations in products for preparing porridge (Table 5), which are basically cereals, were between 0.0 and 2.11 μg F/g. This mean was lower than what was found in Canadian products, 6 which may be explained by the fl uoride concentration in the water used for processing these products. 9,20 The low fl uoride levels found in the samples from various brands of powdered milk, which were mostly less than 0.05 μg F/g (Table 5), are plausible, since milk is a food that naturally contains extremely low fl uoride concentrations, without any risk whatsoever of dental fl uorosis. The same is valid for the groups of "other foods" analyzed (Table 5).
It is concluded that the majority of the foods studied, including typically Brazilian foods, do not present any cause for concern with regard to the chronic toxicology of fl uoride, taking into consideration the concentration in the foods, the amount consumed, and the dose that is considered to be the safe threshold for clinically acceptable dental fl uorosis.