In the years when Dow Agrosciences had no profit motive to fumigate food with sulfuryl fluoride, it took the following position: “Under no conditions should sulphuryl fluoride be used on raw agricultural food commodities, or on foods, feeds or medicinals destined for human or animal consumption.”

Today, with Dow needing to find a replacement for methyl bromide (an ozone-depleting fumigant that Dow also produced), the company has changed its tune. Dow now claims that fumigating food with sulfuryl fluoride contributes such a small amount to the average Americans’ total fluoride intake that it is a trifling matter for which the law should not be concerned. 

As the following comments explain, sulfuryl fluoride is by no means a mere “de minimis” source of fluoride, and the EPA’s proposal to eliminate it from the food supply is amply justified. (To read the full discussion from which these comments are excerpted, click here.)

1. Acute Fluoride Toxicity Is a Significant Health Threat

Despite claims to the contrary, the sulfuryl fluoride tolerances are not a de minimis source of fluoride exposure. This is evident when considering the risk that the tolerances create for episodes of acute fluoride toxicity. Acute fluoride toxicity causes nausea, stomach pain, vomiting, headache, fatigue, and other flu-like symptoms. As Objectors have long maintained, foods fumigated with sulfuryl fluoride can produce doses of fluoride in a single sitting that are sufficient to produce these effects of acute toxicity.

Although EPA is to be credited for its proposed order to phase out the tolerances, the current challenges that this order faces may be the result of the agency’s failure to fully examine and explain the threat of acute fluoride toxicity that the tolerances create. In this section, therefore, Objectors will reiterate and expand upon their previous discussions concerning the risk of acute toxicity posed by the tolerances.

Our analysis begins first by explaining the errors that led EPA to forego conducting a risk assessment for acute fluoride toxicity. In its March 4, 2005 notice of filing of Dow’s petition to establish tolerances for sulfuryl fluoride, the EPA stated that the lowest dosage that can cause acute fluoride toxicity is 5 mg/kg, but that this dosage only applies “for the lowest third percentile of the infant population.” (EPA 2005). There are two demonstrable errors with this assertion.

First, the 5 mg/kg threshold, known as the “Probable Toxic Dose” (PTD) is not limited to infants. The PTD was established by Gary Whitford in 1987 based on the reported death of a 2-year-old child, and Whitford applied it to children of all ages. (Whitford 1987). In fact, the word “infant” does not appear once in Whitford’s discussion of the PTD.

Second, acute fluoride toxicity can be caused at dosages that are far lower than the PTD. The PTD is the dosage that can cause “death”; thus, as Whitford has explained, “even if it is only suspected that 5 mg F/kg has been ingested, it should be assumed that an emergency exists and that immediate treatment and hospitalization are required.” (Whitford 1990). The dosage that can kill a child is, as one would expect, higher than the dosage that can cause the early symptoms of acute fluoride toxicity, including nausea, stomach pain, vomiting, and headache. Whitford, in fact, specifically warned that dosages below 5 mg/kg should not be considered “innocuous.” (Whitford 1990). Consistent with Whitford’s warning, numerous studies have documented the occurrence of acute fluoride toxicity at dosages well below 5 mg/kg. In 1980, Spoerke showed that acute fluoride toxicity, including nausea, vomiting, and diarrhea, regularly occurred at dosages lower than 1 mg/kg. In 1982, Eichler showed that dosages less than 0.5 mg/kg can cause nausea, vomiting, and fatigue. In 1994, a study in the New England Journal of Medicine showed that a dosage as low as 0.3 mg/kg caused nausea, vomiting, abdominal pain, diarrhea, and headache. (Gessner 1994). And, in 1997, Akiniwa discussed a range of studies where acute fluoride toxicity occurred at dosages as low as 0.1 mg/kg.

A third error underlying EPA’s decision to forego an acute risk assessment is the agency’s assertion that the poisonings periodically caused by malfunctioning water fluoridation equipment involved “extremely high concentrations” of fluoride that “fall far outside the realm of expected exposures” from sulfuryl fluoride fumigation. (EPA 2006). This statement is, again, incorrect. Gessner’s study in the New England Journal of Medicine found that the water fluoridation accident in Hooper Bay, Alaska, caused acute fluoride poisoning at dosages as low as 0.3 mg/kg. (Gessner 1994). In Akiniwa’s comprehensive 1997 review, Gessner’s findings were found to be consistent with the findings of other studies on fluoridation accidents. (Akiniwa 1997). A subsequent study, published in Public Health Reports, further buttresses this conclusion. (Penman 1997). The study, which examined a fluoridation accident in Mississippi, found that adults (average age = 29) who drank less than one glass of water with 48 ppm fluoride suffered “acute nausea, vomiting, abdominal cramps, or diarrhea” as well as “headaches,” and “burning sensations in the throat or chest.” The authors estimated that the total ingested dose producing these symptoms was just 5 to 9 mg. Assuming an average body weight of 70 kg, this represents a toxic dosage ranging from just 0.07 to 0.13 mg/kg, or an average dosage of 0.1 mg/kg.

Based on these studies, it is clear that fluoride can cause acute toxicity at dosages as low as 0.1 to 0.3 mg/kg. This is significant because, as Objectors will demonstrate, these dosages can be produced by consumption of foods fumigated with sulfuryl fluoride. This is particularly evident with respect to both dried eggs and wheat flour. We begin first with the 900 ppm tolerance for dried eggs.

As can be seen in the following table, the consumption of just 2 fumigated dried eggs with allowable levels of fluoride residue produces a dose of 22.5 mg of fluoride, while the consumption of 4 fumigated dried eggs produces a dose of 45 mg/day. When converted into dosages, both of these doses would be more than sufficient to produce acute symptoms in the average weighing adult (70 kg). Specifically, a dose of 22.5 mg would produce a dosage of 0.32 mg/kg, while a dose of 45 mg would produce a dosage of 0.64 mg/kg. The problem, of course, would be even more severe for children. According to the CDC, the average six-year-old weighs about 20 kg. If a six-year-old child were to consume two fumigated dried eggs, the dosage would exceed 1 mg/kg. This is up to 10 times greater than the dosage that can cause acute fluoride toxicity.

Risk of acute fluoride toxicity posed by fumigated dried eggs

Data on which calculations are based

F residue level in dried eggs* 900 ppm or 900 mg/kg
Average weight of one large fresh egg: 50 g (American Egg Board 2005)
Conversion factor from dried egg to fresh egg: 1 part by weight dried egg to 3 parts by weight water (USDA 2003; American Egg Board 2005)
USDA standard serving size: 2 eggs
90th percentile large serving: 4 eggs (FDA 1995; 90th percentile is double the mean)


  • 12.5 g dried egg mixed with 37.5 g water gives 50 g reconstituted egg
  • 12.5 g X 900 mg/kg X 0.001 kg/g = 11.25 mg per fresh egg equivalent
  • 2 egg equivalents X 11.25 mg/egg equivalent = 22.5 mg fluoride per serving
  • 4 egg equivalents X 11.25 mg/egg equivalent = 45 mg fluoride per serving

* Our calculation is based on whole dried eggs.  These are the types of eggs most likely to be used as a direct replacement for fresh eggs in recipes like scrambled eggs and omelets.


The risk of acute fluoride toxicity also exists with the 125 ppm tolerance for wheat flour. As we will demonstrate here, this concentration is sufficient to induce acute fluoride toxicity in children eating just two slices of bread or a single serving of pancakes. Although we recognize that Dow’s field tests have reported the average residue level on wheat flour to be 31.4 ppm, it is inappropriate to assess the risk of acute toxicity on the basis of the average level. According to EPA, “using average values is inappropriate for acute risk assessments . . . because in assessing acute exposure situations it matters how much of each treated food a given consumer eats in the short-term and what the residue levels are in the particular foods consumed.” (EPA 2011). Our calculations here, therefore, focus mainly on EPA’s 125 ppm tolerance level. Nevertheless, we will also demonstrate that the risk of acute fluoride toxicity exists even with respect to wheat flour fumigated at the average level of 31 ppm. We begin first with bread.

According to the USDA, a typical slice of bread contains 14.75 grams of wheat flour. (USDA 2008). If this typical slice is made with fumigated wheat flour, it would contain 1.84 mg of fluoride; two slices would contain 3.68 mg. If consumed by a three-year old child (average weight = 15 kg), these doses would be sufficient to cause acute fluoride toxicity (dosage range = 0.1 – 0.24 mg/kg). The situation would be even worse for breads containing more than 14.75 grams of flour. A quick search on the internet shows that it is not infrequent for bread to contain over 20 grams of wheat flour per slice. In bread containing 20 grams of wheat flour per slice, two slices of the bread would contain 5 mg of fluoride if made with fumigated flour. A 3-year-old child consuming these two slices would receive a dosage of 0.33 mg/kg; a 6-year-old child would receive a dosage of 0.25 mg/kg. Both of these dosages are sufficient to cause acute fluoride toxicity in some children.

A similar poisoning danger exists with other common types of wheat products, including pasta or pancakes. A review of online pasta and pancake recipes shows that it is common for a single serving to use a half-cup of flour. According to the USDA, a half cup of flour equals 60 grams. If these 60 grams were fumigated with sulfuryl fluoride at the allowable level, the resulting dose from a single serving of pasta or pancakes would be 7.5 mg. If consumed by the average weighing 6-year-old child (20 kg), this would produce a dosage of 0.375 mg/kg. Again, this is more than sufficient to cause acute fluoride toxicity.

The risk of acute fluoride toxicity from consuming fumigated flour in pasta or pancakes exists even where the flour contains the average residue level of 31.4 ppm. Sixty grams of this flour would contain 1.88 mg. If this were consumed by the average-weighing 5 year old child (18.2 kg), or the 25th percentile 6-year old child (18.8 kg), the resulting dosage would equal or exceed 0.1 mg/kg.

Based on these calculations, it is evident that food fumigated with legally allowable levels of sulfuryl fluoride can cause food poisoning in the form of acute fluoride toxicity. Such poisoning incidents can hardly be described as de minimis, particularly since children (due to their low bodyweights) will be the primary population at risk for such incidents. Accordingly, because EPA has no authority to carve out a de minimis exception if enforcement of the FQPA produces a gain of more than trivial value, the risk of acute fluoride toxicity provides a sufficient basis, without more, to bar the exercise of the de minimis doctrine in this case.

2. Chronic Fluoride Toxicity Is a Significant Health Threat

The risk of acute fluoride toxicity is not the only reason the EPA should reject the claim in some comments that the tolerances are a de minimis source of fluoride. For example, in considering Dow’s current contention that sulfuryl fluoride can be safely added to our nation’s food supply in small amounts, it is worth noting Dow’s previous statements on this subject. In 1963, Dow stated that:

Under no conditions should sulphuryl fluoride be used on raw agricultural food commodities, or on foods, feeds or medicinals destined for human or animal consumption.” (Bond 1984) (emphasis added).

It is also worth noting that this remains the clear prevailing view of nearly all western, industrialized countries, even those that have approved the use of sulfuryl fluoride as a fumigant in food processing facilities. As summarized by Dow scientists in 2010:

“ProFume is currently approved on emptied flour mills and emptied silos in the following European countries: Austria, Belgium, Germany, France, Greece, Italy, Ireland, Switzerland and UK . . . .”

Not only does Europe require the removal of all food products prior to fumigation, it also requires measures to ensure that fluoride residues on the mill machinery do not contaminate food when production resumes. According to a 2010 risk assessment from the European Food Safety Authority (EFSA):

“Even though uses on the fumigation of food items (dried fruits, nuts) were withdrawn during the peer review procedure and only uses for structural treatments remain, there is still the potential for consumer exposure to inorganic fluoride through contaminated products, such as flour and bran that remained in the mill machinery during fumigation, or grain stored in silos in the mill. Available data show that high fluoride residue levels in flour and bran occurred after the production in a treated mill structure had been taken up again. Therefore, if in practice contamination per se cannot be avoided, then measures to avoid contaminated cereal products entering the food chain are necessary.” (EFSA 2010) (emphasis added).

In accord with the EFSA report, the UK government has imposed the following terms of use on ProFume’s Product Label:

Emptying and cleaning

Mills: The mills should be emptied of flour and cleaned. Mill machinery should be run to remove flour retained within the equipment. Flour storage areas and connecting pipes should be emptied or effectively sealed. Bagged flour must be removed to prevent gas penetration.
Contact of ProFume with flour should be avoided.

Food Processing and Storage Facilities: The food processing facility should be emptied of product and cleaned. Food processing machinery should be free of food material within the equipment. Food storage areas, machinery, equipment and connecting pipes should be opened up and or effectively sealed or emptied. All packaged food material including raw ingredients and finished products must be removed or sealed to prevent gas penetration.
Contact of ProFume with other food material should be avoided.

Re-Starting the Mill

Mills: Flour and bran produced in the first flush (i.e. the 10 minutes after restart of full flow of flour) must be collected and sent for disposal/discarded. The next 50 minutes of production must be collected and recycled through the production process. Discarded flour must not be used for human or animal consumption.

Food Processing Facilities: All production lines must be inspected and any residual food material collected and sent for disposal/discarded. Discarded material must not be used for human or animal consumption.

Canada has taken a similar approach. As EPA noted in its Proposed Order, “Canada has imposed restrictions on the use of sulfuryl fluoride for the fumigation of food processing facilities that are designed to insure that no residues result in food.” (Emphasis added)

The notion, therefore, that contaminating foods with sulfuryl fluoride is merely a de minimis matter has been implicitly, if not explicitly, rejected by most western nations. The consensus among these nations is that the marginal increase in costs associated with removing all food prior to fumigation is not sufficient to justify allowing sulfuryl fluoride to enter the food supply. EPA’s proposed order is consistent with this consensus.

It is particularly important that sulfuryl fluoride not be added to food in the United States because, unlike children in the European countries that disallow food-based tolerances, children in the United States have extensive exposure to fluoride from fluoridated drinking water. Indeed, according to the EPA’s latest document on fluoride exposure, children in the United States are already being chronically overexposed to fluoride. Based on the most recent data, “some children drinking water at the 90th percentile intake level up to about age 7 are being exposed to fluoride on a daily basis at levels at or higher than estimated acceptable intake levels when the concentration of fluoride in their drinking water is at or above 0.87 mg/L.” (EPA, 2010b, p.104). Even if all municipal water systems were to lower the level of fluoride to 0.7 mg/L, as has been suggested recently by the Department of Health and Human Services, the upper limit (UL) for fluoride exposure (IOM, 1997) would still be exceeded by about 10% of children under the age of 4 years, even without any contribution from sulfuryl fluoride. (EPA, 2010b, Table 7-1, p.98 and Table 8-2, p.104).

Reflecting this over-exposure, the latest national survey by the CDC found that 41% of American adolescents now have dental fluorosis (Beltrán-Aguilar et al., 2010). This includes 8.6% with mild dental fluorosis (up to 50% of enamel impacted) and 3.6% with moderate or severe dental fluorosis (100% of the enamel impacted). The latter category is recognized by the NAS as an adverse health effect. (NRC, 2006) The rate of moderate/severe fluorosis is highest among African American and Mexican American children (Beltrán-Aguilar, et al., 2005, tbl. 23), which implicates EPA’s stated policy goals regarding environmental justice. There is a reasonable expectation that even a small additional contribution of fluoride from sulfuryl fluoride residues will lead some children to develop severe dental fluorosis who would not otherwise develop this adverse health effect. Accordingly, there is simply no safe room for the additional exposure to fluoride that will result from sulfuryl fluoride tolerances.

Although the fluoride intake from sulfuryl fluoride may appear small when averaged across all consumers, it can provide a significant source of fluoride exposure for those with high intakes of foods with high fumigation rates. The EPA anticipates, for example, that 100% of cocoa powder, 100% of pinto beans, 100% of dried beans, 99% of walnuts, 69% of dried fruits, 10% of almonds, and 3% of rice will be fumigated with sulfuryl fluoride if the tolerances are not withdrawn. As Dow’s field test data shows, the average fluoride concentration of these crops when fumigated ranges from 1 to 12.5 ppm.

Food Commodity

% Treated (Direct)

Residual F (mg/kg)




All dried




Pinto beans






All dried fruits



All tree nuts






(milled basis)


4.5 (white)

12.5 (brown)

Although the DEEM database shows that the 90th to 95th percentile of the population consumes roughly 2 to 4 times more of any given food item than the average consumer, EPA only assessed exposure from sulfuryl fluoride as a function of age. It is possible, and indeed likely, therefore, that some individuals and populations will receive significantly more fluoride from the tolerances than the estimated “average” exposure. Consider, for example, dry beans. Under the tolerances, 100% of dry beans will be directly fumigated with sulfuryl fluoride, producing an average fluoride content of 4.5 mg/kg. The USDA’s Dietary Guidelines recommend 2 to 3 cups of dry beans per week, or the equivalent of 65 to 97 grams per day, for those with daily nutrient intakes of 2000 calories or more per day. (USDA 2010, p.23, tbl. B2.3). If the sulfuryl tolerances are allowed to stand, the USDA’s recommended guidelines for bean intake would result in a fluoride exposure of 0.29 to 0.43 mg/day from this source alone, hardly a de minimis amount. Although most Americans do not consume the USDA’s recommended intake for beans, it stands to reason that many people do. Hispanics, for example, accounted for 33 percent of U.S. dry bean consumption in 2000, despite representing only 11 percent of the population (Lucier et al., 2000). Hispanics will thus receive significantly more fluoride from this source than the average consumer, which is particularly significant given EPA’s environmental justice policy and the fact that Mexican American children suffer the highest rates of moderate/severe fluorosis in the population. (Beltrán-Aguilar, et al., 2005, Table 23).

The experience with cryolite provides an instructive example of the problems that can result from focusing only on the average intake for the population as a whole. According to EPA’s exposure analysis, the average intake of fluoride from cryolite is less than the average intake from sulfuryl fluoride. (EPA 2010a, Table 1, p.2). Based on the reasoning by Dow AgroSciences and others, cryolite must thus be a de minimis source of fluoride. That, however, is not the case, as fluoride intake from cryolite has been found to be a significant source of fluoride for children who drink appreciable quantities of white grape juice (due to the high quantities of fluoride residues on white grapes). (Stannard 1991). Accordingly, the focus on the overall average intake of fluoride from cryolite or, in this case sulfuryl fluoride, loses sight of the fact that when the data is appropriately analyzed to focus on high-end consumption of specific food items, the intake can become quite meaningful and significant.

Finally, EPA based its proposed order on the premise that dental fluorosis is the only health effect of concern from exposure to fluoride. This premise, however, is an increasingly untenable one. A growing body of evidence indicates that fluoride can cause a range of other adverse effects, including damage to the developing brain. Indeed, a recently published meta-analysis of 27 published studies by a team of Harvard scientists reports that children in high fluoride areas have significantly lower IQ scores than those in low fluoride areas (Choi et al., 2012). One recent study (Ding et al., 2011) found significant negative associations with children’s IQ based on individual-level measures of exposure at relatively low levels of fluoride (0.24-2.84 mg/L). Other studies reviewed by the Harvard team, including many studies that have yet to be considered by the EPA, have found IQ reductions in children drinking water with 0.88 ppm (Lin 1991); 1.75 ppm (Xiang 2003a,b), 1.8-3.9 ppm (Xu 1994); 2.0 ppm (Yao 1996, 1997); 2.1-3.2 ppm (An 1992); 2.38 ppm (Poureslami 2011); 2.45 ppm (Eswar 2011); 2.5 ppm (Seraj 2006); 2.85 ppm (Hong 2001); 2.97 ppm (Wang 2001, Yang 1994); 3.15 ppm (Lu 2000); 4.12 ppm (Zhao 1996). Objectors have recently translated many of the Chinese studies relied upon by the Harvard team and will be providing these studies in a supplemental submission to these Comments. As the Harvard team points out, “dose-response differences in test scores [have] occurred at a wide range of water-fluoride concentrations,” and the “results suggest that fluoride may be a developmental neurotoxicant that affects brain development at exposures much below those that can cause toxicity in adults.” (Choi et al 2012). In light of the FQPA’s clear command to protect infants and children from developmental neurotoxicity, the research on fluoride and IQ calls for greater precaution with current fluoride policies and strongly counsels against EPA’s exercise of the de minimis option.

In this discussion of acute and chronic fluoride toxicity, Objectors have offered arguments and evidence demonstrating that sulfuryl fluoride exposure can cause serious health effects. Therefore, the imposition of a de minimis exception would perpetuate a very real danger to the public health.

Click here to download a pdf copy of the full submission to EPA.