Health Risk Assessment Benzene in Beverages

Chemical Health Hazard Assessment Division
Bureau of Chemical Safety
Food Directorate
Health Canada
May 19, 2006

NOTE: This health risk assessment was prepared before the results for all of the products were available and was based on results for 67 samples representing 63 unique products. The final table of results shows concentrations for all products analysed. Since the final table of results shows the average concentration for all samples, there may be some differences between the concentrations shown in that table and those appearing in Appendix 2 of this assessment. The additional results that became available after this assessment was conducted do not affect the conclusion of the assessment.

Summary

Health Canada has investigated soft drinks and other beverages for the presence of benzene following reports that trace levels of benzene had been found in such products. Benzene can form in beverages when ascorbic acid combines with either sodium or potassium benzoate, with certain additional conditions present, such as heat, ultraviolet light and metallic ions in the mixture.

Samples of 118 different soft drink and beverage products containing benzoates (used as a preservative) were analyzed for levels of benzene, and results are provided in Survey of Benzene in Soft Drinks and other Beverage Products. The results from 67 of the samples (representing 63 unique products) were available for assessment at the time of writing this report, and are shown in Appendix 2. Five samples of cherry-flavoured Kool-Aid Jammers "10" (Kraft) and one sample of tropical punch-flavoured Kool-Aid Jammers "10" (Kraft) contained elevated levels of benzene (14.0 to 17.6 ng/mL). The manufacturer has reformulated these products. Of the remaining 14 positive samples, three were concentrates that require dilution prior to consumption, and contained between 6.3 and 10.3 ng/mL benzene, three were samples of various flavours of Kool-Aid Jammers (sweetened with glucose-fructose as opposed to Kool-Aid Jammers "10" which are sweetened using sucralose and acesulfame-potassium) that contained between 3.1 and 3.8 ng/mL benzene, and the remaining positive samples contained less than 3 ng/mL.

Based on the observed levels of benzene in soft drinks and reported actions by manufacturers, Health Canada scientists assessed the potential exposure to these products over a short-term period relative to the doses associated with negative health effects from benzene, and concluded that consumption of these products would constitute a class 3 health risk. Had the potential exposure to this product been projected to be over an extended period of time, this would have been considered a low class 2 health risk.

Health Canada has concluded that soft drinks and other beverages available for sale in Canada are safe, based on these findings and on the actions taken by the producers to reformulate products where necessary.

Background

Benzene is a known human carcinogen. It is naturally occurring but is also manufactured for use in the industrial sector. Human exposure to benzene is mainly through inhalation (e.g., vehicle exhaust and cigarette smoke) and to a much lesser extent through the ingestion of food and water.

A survey of beverages available on the Canadian retail market was conducted as a result of reports of elevated levels of benzene in soft drinks in the United States. Samples of various types of beverages and beverage mixes were gathered in the spring of 2006, and analysed for the presence of benzene.

A similar issue had arisen in the early 1990s. At that time, investigations revealed that ascorbic acid can react with either sodium benzoate or potassium benzoate under certain conditions of heat and/or ultraviolet light to form benzene. Benzoates are employed as preservatives to prevent bacterial growth. Ascorbic acid may be added as Vitamin C or as an food additive (antioxidant). Other components can either enhance (e.g. transition metals, intense sweeteners, ascorbic acid naturally present in fruit juices) or inhibit (e.g. nutritive sugars, calcium disodium ethylenediaminetetraacetic acid) benzene formation. Benzaldehydes (e.g. in flavourings) may also react with ascorbic acid to form benzene. Based on these findings, manufacturing processes were reported to have been adjusted in order to eliminate or greatly reduce the formation of benzene.

Hazard characterisation

A risk assessment of benzene ingested in the diet was first conducted in the Bureau of Chemical Safety (BCS) in 1991 based on a two-year carcinogenicity study in rats and mice conducted by the U.S. National Toxicology Program (NTP). This study indicated that benzene was a multi-potential carcinogen in both mice and rats and a Tolerable Daily Intake (TDI) of 3.6 (micrograms per kg body weight per day (µg/kg bw/day) was calculated initially, based on a Lowest Observed Adverse Effect Level (LOAEL) of 25 mg/kg bw/day (5 days/week) in the mouse for both non-cancer endpoints and for multiple tumour sites. To calculate the TDI, the LOAEL was recalculated to 18 mg/kg bw/day to reflect the dose for a 7-day week, and a safety factor of 5000-fold was applied (10-fold each for interspecies and intraspecies extrapolation, 10-fold for severity of the effect (cancer) and 5-fold to account for the lack of a No Observed Adverse Effect Level or NOAEL). In order to account for the fact that exposure to benzene through a number of routes would need to be included under the 3.6 µg/kg bw/day level, an additional 10-fold safety factor was applied to give a TDI for benzene ingested (from food and beverages) of 0.36 µg/kg bw/day.

To determine whether the basis for the previous hazard assessment of benzene required updating, a survey of current regulatory guidelines for benzene was conducted (See Appendix 1). The U.S. Environmental Protection Agency (US-EPA) and the Netherlands National Institute for Public Health and the Environment (RIVM) have derived guidelines for exposure to benzene through the oral route using human occupational inhalation exposure studies by Rothman et al. 1996 and Rinsky et al., 1987, respectively. In addition, the BCS evaluated reports from Health Canada (HC), WHO and U.S. EPA on the derivation of guidelines for benzene in drinking water. The NTP 1986 rodent study which was the basis of the BCS risk assessment, continues to be cited as the pivotal study for the WHO and Health Canada drinking water risk assessments.

The US-EPA established a chronic oral Reference Dose (RfD) based on benchmark dose modelling of a non-carcinogenic endpoint (decreased lymphocyte count) in the Rothman et al. study and applied route-to-route extrapolation (inhalation to oral) and safety factors. Furthermore, the EPA also established drinking water guidelines based on the Rinsky et al. study with leukemia as the end point and using a cancer oral slope factor and route-to-route extrapolation. The Rinsky et al. analysis was also selected as the basis for the RIVM risk assessment for oral exposure.

To validate the route-to-route extrapolation method for the non-cancer endpoint, the US-EPA conducted comparisons with the NTP mouse data by calculating a RfD using both a LOAEL and Benchmark Dose Level (BMDL) with the addition of safety factors. This procedure generated RfDs of 6.0 and 0.7µg/kg bw/day, respectively, with the BMDL modelling of NTP data being the most conservative and the LOAEL method being the least conservative. These values were comparable to that of 4 µg/kg/bw/day generated by the US-EPA using the non-cancer endpoint from the human data (Rothman et al., 1996), and that calculated by Health Canada in 1991.

Risk assessments for benzene in drinking water (U.S. EPA, HC, WHO) used linear extrapolation methods applied to carcinogenic data to derive a concentration or range of concentrations that were associated with a defined risk level or range of risk levels (e.g. 10^-6 or 1 in a million). When we used the US-EPA range of oral slope factors (cancer risk per unit dose) derived by the U.S. EPA from the Rinsky et al. data to calculate the oral exposure associated with a cancer risk level of 10-6, a range of 0.018 to 0.066 µg/kg/bw/day was obtained. This range is about two orders of magnitude less than the values obtained for a TDI/RfD using the safety factor method. The TDI established by BCS in 1991 for exposure from food and beverages (0.36 µg/kg/bw/day) therefore falls midway between the range of values calculated using linear extrapolation from the human cancer endpoint, and those calculated using the safety factor approach with non-cancer endpoints from human or experimental animal data.

Because of the increased preference for using a margin of exposure approach in conducting risk assessments for compounds that are both genotoxic and carcinogenic, the possibility of calculating a bench mark dose (BMD) using the data from the 1986 NTP rodent assay with the US-EPA software was considered. While benzene induced an increase in tumour incidence in a number of different tissues/organs in rats and mice, none of these were consistent with the cancer endpoint in humans, i.e. acute non-lymphocytic leukaemia. While these data had been used in mathematical models in the past, they were considered inappropriate due to the lack of a dose response. Consequently, in the short term, the existing TDI and the oral slope factor from the US-EPA based on a human cancer endpoint were used for the risk assessment.

Also of note, a thorough evaluation of available data on benzene was conducted by Toxicology Excellence for Risk Assessment (TERA) in March 2006. The quantitative risk assessment is pending as TERA has tentatively scheduled a peer consultation meeting June 15-16, 2006 at the Northern Kentucky University Metropolitan Education and Training Services (METS), Erlanger, Kentucky.

Exposure Assessment

Analytical Results

The results of the analyses of 67 samples (representing 63 unique products) were available for this assessment. Samples included juices, soft drinks, low-alcohol coolers, syrups (e.g. grenadine), and cocktail mixes (See Appendix 2). Benzene concentrations were reported in ng/mL but were converted to ng/g using the density values shown. The density of beverages not requiring dilution before consumption (ready to consume) ranged from 0.9868 to 1.0638 g/mL. The density of the syrups and cocktail mixes ranged from 1.0674 to 1.340 g/mL.

Of the 67 samples analysed, 47 were below the analytical detection limit of 1.0 ng/mL. In the 20 positive samples, benzene concentrations ranged from 1.1 ng/mL to 17.6 ng/mL:

• Six samples from the same manufacturer (five cherry-flavoured Kool-Aid Jammers "10" and one tropical punch-flavoured Kool-Aid Jammers "10") contained between 14.0 and 17.6 ng/mL. All contained benzoate (and sorbate) as a preservative. The manufacturer, Kraft, has indicated that they had reformulated the product to address the issue, and no longer uses benzoate (or sorbate) but uses aseptic hot filling of the bottles to achieve microbial stability. [Samples of the reformulated product have been analysed, and benzene was not detected.]
• Three concentrates (which require dilution prior to consumption as a beverage) that contained between 6.3 and 10.3 ng/mL benzene. The manufacturer (Cadbury Beverages Canada) of one of these products, Rose's Cocktail Infusion Cranberry Twist Mix, indicated that this product has been reformulated and this has addressed the issue of benzene formation. There is no indication that the other two concentrated products have been reformulated. [Note: Subsequent to this assessment, the analysis of additional samples of Amaretto syrup led to the calculation of an average benzene concentration of 5.0 μg/L. A concentration of 2.5 μg/L benzene was found in a later sample of Mott's Mr & Mrs T Margarita Mix.]
• Three samples of Kool-Aid Jammers (products distinct from Kool-Aid Jammers "10"), one each of the blue raspberry, orange, and grape flavours, that contained between 3.1 and 3.8 ng/mL benzene.
• The remaining eight positive samples contained less than 3 ng/mL.

For concentrated products, appropriate dilution factors, based on the product label instructions, were derived to obtain the benzene concentration in the product "as consumed" (see Appendix 2).

Consumption Figures

Appropriate concentration values for beverages as consumed, as shown in Table 1, were used to create exposure scenarios for the consumption of:

1. non-alcoholic beverages (excluding the cherry and tropical punch-flavoured Kool-Aid Jammers "10");
2. alcoholic beverages (low-alcohol coolers and cocktails); and
3. only the old formulation of cherry or tropical punch-flavoured Kool-Aid Jammers "10".

Mean concentration values were used (rather than higher percentile figures) in consideration of the chronic nature of the potential health effects of exposure to benzene.

For "alcoholic beverages", Nutrition Canada Eaters Only mean consumption figures were employed. For "soft drinks", the Nutrition Canada Eaters Only mean consumption figures were employed for 5 to 11 year old children and teens 12 to 19 years of age. The Nova Scotia mean Eaters Only consumption figure for "soft drinks, regular" was employed for adults. In all but one case, the frequency of consumption was assumed to be daily. For the consumption of alcoholic beverages by 12 to 19 year olds, frequency of consumption was assumed to be two days of every seven.

For a comparison of intake values readily available from various sources, see Appendix 3.

Body weights

The following mean body weights were employed in the assessment: 60 kg for an adult; 26.4 kg for a 5 to 11 year old; and 53.8 kg for a 12 to 19 year old.

Probable Daily Intake Values

The "probable daily intakes" (PDI) of benzene from the consumption of the following were calculated:

1. all non-alcoholic beverages (excluding the Kool-Aid Jammer "10" products that have since been reformulated);
2. alcoholic beverages (low-alcohol coolers and cocktail mixes);
3. the six fruit beverage samples that contained elevated levels of benzene (and which have since been reformulated but some of which still remains on the market);
4. beverages containing a theoretical level of 5 ng/mL; and
5. beverages containing a theoretical level of 10 ng/mL.

The six Kraft Kool-Aid Jammers "10" products that contained 14.0 to 17.4 ng/mL benzene and which have since been reformulated were not included with the other non-alcoholic beverages because they are no longer entering the market. It is understood that the old product should be depleted by mid-June, 2006 and so continued long-term consumption of the "old formulation" will not be possible.

The deterministic exposure assessment used to calculate the PDI values is shown in Appendix 4. A summary of the PDI values is shown in Table 2.

Other Sources of Exposure to Benzene

There have been various estimations of human exposure to benzene from air, food and water. Inhalation (e.g. of cigarette smoke, automobile exhaust, and gasoline vapour) has been shown to be the largest route of exposure to benzene (Environment Canada and Health and Welfare Canada, 1993; Wallace, 1996). Exposure from food and water is generally much less than that from air (WHO, 2003).

Low levels of benzene have been measured in a wide variety of foods, and it has been reported that benzene is present naturally and from environmental contamination.

Researchers with the U.S. Food and Drug Administration (US-FDA) (Fleming-Jones and Smith, 2003) gathered samples of 70 different types of foods and prepared them for consumption as they would be in a domestic kitchen. They detected benzene in 68 of the 70 foods at levels ranging from 1 ng/g, in milk-based infant formula and raw strawberries, to 190 ng/g, in a sample of fully cooked ground beef (the average concentration in twelve samples of ground beef was 40 ng/g). Relatively high levels were also found in at least one sample each of a cola (138 ng/g), raw bananas (132 ng/g), and cole slaw with dressing (102 ng/g). Results for all food samples were not provided in the publication. It has been stated elsewhere that benzene is naturally occurring in certain foods, with reference to specific publications (citations noted in McNeal et al., 1993).

An earlier publication, also by researchers from the US-FDA (McNeal et al., 1993) analysed a limited number of foods (most of which were not represented in the survey of Fleming-Jones and Smith). Low levels (2 ng/g or less) of benzene were found in the samples with no added benzoates (e.g. fruit beverages, sodas, fruit preserves, brewed instant coffee, baked potato, hard-boiled egg, smoked fish, roasted peanuts) compared to those in some, but not all, samples containing added benzoates (e.g. 3 ng/g in a cocktail mix, not detected in a fruit punch or in diet sodas, 38 ng/g in an imitation fruit preserve, 5 ng/g in an imitation grape jelly, 7 ng/g in a BBQ sauce, and 22 ng/g in a taco sauce). An exception was liquid smoke with no added benzoates. One of the liquid smoke samples contained 121 ng/g benzene.

The authors noted that their data contradicted earlier reports of high benzene levels in foods such as eggs and roasted peanuts and suggested that the earlier reports may have been due to laboratory contamination.

The following summarizes various estimates of overall dietary exposure to benzene from foods:

• Wallace (1996), in a review of environmental exposure to benzene, stated that the 1980 U.S. EPA Total Exposure Assessment Methodology studies on benzene, and subsequent studies, "found no evidence of food contributions to body burden of participants."
• A 2004 U.S. EPA research abstract (2004) describes more recent work on human exposure to benzene in an urban environment using a probabilistic exposure model. The research indicated that the majority of exposure was via inhalation (exposures in the home and in vehicles and during refueling of vehicles). Seven percent of total exposure was determined to be dietary.
• An assessment report prepared under authority of the Canadian Environmental Protection Act (EC & HC, 1993) described Canadian exposure to benzene from various routes. The highest exposure was through inhalation (approximately 98% of total exposure to benzene among non-smokers). Dietary intake of benzene by adults was estimated to be 20 ng/kg bw/day.
• In 1992, the Chemical Health Hazard Assessment Division (CHHAD) of Health Canada estimated exposure to benzene from a variety of foods including butter, beef, irradiated beef (if it were available on the market), boiled eggs, haddock, coffee, oranges, mangoes, various fruits and fruit juices, beans, split peas, and rum. The mean exposure ranged from 120 to 325 ng/kg bw/day, depending on the food consumption scenario.
• The supporting documentation for the benzene guideline for Canadian drinking water quality (HC, 1987) cites a 1980 U.S. National Research Council publication that apparently reported a "rough estimate of dietary intake in the United States" of 250 μg/day (4,200 ng/kg bw/day for a 60 kg adult).
• A total dietary intake value of 180 μg/day (3000 ng/kg bw/day for a 60 kg adult) is described by WHO (2003), which cited a 1988 publication from the Netherlands in support of this figure.

Small amounts of benzene are also produced in the body as a result of amino acid metabolism.

While these values are not consistent, it is generally considered that dietary intake of benzene is a minor source of total benzene exposure.

Risk Characterisation

Risk Characterisation of the "old" formulation of cherry and tropical punch-flavoured Kool-Aid Jammers "10" that could remain available at retail (estimated time for complete depletion of this product from retail is June, 2006).

The Kraft Kool-Aid Jammers "10" samples that contained elevated levels of benzene (between 14.0 and 17.6 ng/mL) are no longer being manufactured under the old formulation. Although not verified with the manufacturer, it appears that these particular products first entered the market in Canada in early 2004. The manufacturer stopped distribution to the retail level and estimated that the old formulation would be depleted from retail shelves by mid-June 2006. [Note: Subsequent to the completion of this assessment, samples of the reformulated product were analysed and benzene was not detected.]

The PDI value for the consumption of the old formulation of cherry and tropical punch-flavoured Kool-Aid Jammers "10" by a consumer with strong preference for this product, such that they consume it on a daily basis, was approximately 180 ng/kg bw/day for children 5 to 11 years of age. It was approximately 130 ng/kg bw/day for teens and adults. These PDI values are lower than the Tolerable Daily Intake (TDI) and Reference Dose (RfD) values that have been derived by various regulatory groups. They are approximately three times higher than the U.S. EPA-derived dose associated with a cancer risk of 10^-6 and would correspond to a lifetime risk of 10^-5. Had the potential exposure to benzene from these products been projected to be over an extended period of time, the exposure would have been considered of low concern to human health and a low class 2 health risk. However, since retail level stocks of the old formulation should be depleted by mid-June 2006, ensuring that continued long-term exposure to these products will not occur, risk from consumption of these products would be considered of negligible concern to human health. Therefore, it is considered that this exposure scenario constitutes a class 3 health risk.

It is difficult to place these PDI values in the context of total dietary exposure to benzene in consideration of the range of reported dietary intakes. An intake of 180 ng/kg bw/day is significant relative to the intakes estimated by CHHAD in 1992 and by EC & HC in 1993. Relative to the estimation from the Netherlands (as cited by WHO) and from the U.S. NRC in 1980 (cited by HC, 1987), this PDI value would be an order of magnitude lower than total dietary intake.

Risk Characterisation of the remaining products (other than the old formulation of cherry and tropical punch-flavoured Kool-Aid Jammers "10").

With respect to exposure scenarios that were based on products other than the old formulation of cherry and tropical punch-flavoured Kool-Aid Jammers "10", the highest PDI value was approximately 9 ng/kg bw/day. This value is lower than all TDI and RfD values derived by other regulatory groups and would be associated with an estimated cancer risk of 10-7, based on the U.S. EPA's oral slope factor. Therefore, these products would be of negligible concern to human health. As a result, it is considered that the remaining products would also constitute a class 3 health risk.

As noted previously, it is difficult to place these PDI values in the context of total dietary exposure to benzene in consideration of the range of reported dietary intakes. An intake of 9 ng/kg bw/day is insignificant relative to the intake estimated by CHHAD in 1992 but is roughly half that estimated by EC & HC in 1993. Relative to the estimation from the Netherlands (as cited by WHO) and from the U.S. NRC in 1980 (cited by HC, 1987), this PDI value would be four orders of magnitude lower than total dietary intake.

Risk Mitigation

Other Jurisdictions

The United States Food and Drug Administration (FDA) has analysed benzene levels in various products (data not available) and has indicated that as of March 21, 2006, their surveys have revealed that the "vast majority of beverages containing both benzoate preservative and ascorbic acid contained either no detectable benzene or levels below 5 ppb" (U.S. FDA, 2006). Health Canada is not aware of any other action taken by the U.S. FDA (other than anecdotal reports of communications with the soft drink industry in the U.S.). The U.S. FDA publically released new data on benzene in beverages on May 19, 2006.

The Food Standards Australia and New Zealand reported in April 2006 (FSANZ, 2006) that it is conducting a survey of soft drink and juice products containing benzoates and ascorbic acid in response to the findings in the United States. The results were not yet available at the time of compiling this report. Health Canada has exchanged information on assessing and addressing potential risks related to benzene in beverages with FSANZ.

The United Kingdom Food Standards Agency (U.K. FSA, 2006) released the results of its survey of benzene in soft drinks on March 31, 2006. Of the 150 samples analysed, four contained benzene levels greater than the WHO guideline for drinking water. The FSA does not have a "statutory limit" for benzene in soft drinks. However, the FSA requested that the four companies producing these beverages withdraw them from sale and that the soft drinks industry ensure that levels of benzene are kept as low as practicable. The FSA also noted in communications that the "levels of benzene reported in this survey will only make a negligible impact on people's overall exposure to benzene and so any additional risk to health is therefore likely to be minimal."

Drinking Water Guidelines

The Guidelines for Canadian Drinking Water Quality include a maximum acceptable concentration (MAC) for benzene in drinking water of 0.005 mg/L (5 μg/L). This MAC was derived based on the results of experimental (animal) studies, which led to the calculation of an estimated lifetime cancer risk of 3.1 × 10^-6 to 3.4 × 10^-5 from the daily consumption of 1.5 L of drinking water containing 5 μg/L benzene.

The World Health Organization established, in 1984, a drinking water quality guideline value of 0.01 mg/L benzene. This value was also derived in consideration of the results of experimental studies. The value was reaffirmed in 1993 and again in 2003 (WHO, 2003).

Considerations Arising from the Risk Assessment

• Children constitute a group that is consuming many of these products, with certain exceptions such as the low-alcohol coolers and alcohol-containing cocktails.
• Benzene is considered genotoxic and carcinogenic (although it is not known whether there is a genotoxic mode of action with respect to its carcinogenicity). Its presence in these products is preventable. An ALARA (as low as reasonably achievable) approach has been applied by Health Canada, with consideration given to the lifetime cancer risk, in the case of genotoxic carcinogens that have not been directly added to the food but are present for some other reason (e.g. environmental or processing contaminants).
• The potential exposure to benzene from soft drink consumption would constitute a relatively small portion of overall lifetime dietary exposure to benzene. The old formulation of the cherry and tropical punch-flavoured Kool-Aid Jammers "10" would also constitute a relatively small portion of overall lifetime dietary exposure to benzene, considering that the distribution of these products (the old formulation) was halted on May 19^th, 2006.
• Several estimations of total dietary exposure to benzene were located, although an exhaustive investigation was not conducted. The values were not consistent. However, it is generally considered that dietary intake of benzene is a minor source of total benzene exposure.
• With respect to the product remaining on the market, samples of which contained elevated levels of benzene (14.0 to 17.6 ng/mL), these products would be considered of relatively low concern to human health over the long term and of negligible concern to human health over the short term. The ALARA principle has, in effect, been applied in that the manufacturer has reformulated the product to prevent the formation from benzoate of benzene, a genotoxic and carcinogenic compound, and has frozen further shipments of the old formulation to retail.
• Another consideration is whether or not this issue can be expected to arise again. Regulatory authorities believed that the issue had been addressed after the first case of elevated benzene in soft drinks in the early 1990's. The recent finding of a small number of beverages containing elevated levels of benzene suggests that it is possible that new beverages could enter the retail market without due consideration to the formulations and whether benzene could be generated in situ. [Note: Refreshments Canada, a trade association representing a number of companies that manufacture and distribute non-alcoholic liquid beverages in Canada, has prepared a guidance document to assist the industry in mitigating the potential for benzene formation in beverages.

Risk Management Options for Canada

While the beverage industry has already taken action on this issue, any further risk management actions will be determined by the Canadian Food Inspection Agency, in consultation with Health Canada and in consideration of the classification of health risk.

References

Environment Canada and Health and Welfare Canada, 1993,  Benzene - Canadian Environmental Protection Act, Priority Substances List Assessment Report

Fleming-Jones, M.E., and Smith, R.E., 2003, Volatile Organic Compounds in Foods: A Five Year Study, Journal of Agricultural and Food Chemistry, 51(27):8120-8127.

Health Canada (HC), 1987, Benzene - Supporting Document for the Guidelines for Canadian Drinking Water Quality

Food Standards Australia and New Zealand (FSANZ), 2006, Food Standards News 56 April 2006 ("Benzene in soft drinks")

National Research Council (NRC), 1980, Drinking water and health. Volume 3, National Academy Press, Washington, DC; as cited in Health Canada (HC), 1987, Benzene - Supporting Document for the Guidelines for Canadian Drinking Water Quality

National Toxicology Program (NTP), 1986, Toxicology and Carcinogenesis Studies of Benzene (CAS No. 71-43-2) in F344/N Rats and B6C3F1 Mice (Gavage Studies). NTP, Research Triangle Park, NC.

McNeal, T.P., Nyman, P.J., Diachenko, G.W., Hollifield, H.C., 1993, Survey of Benzene in Foods by Using Headspace Concentration Techniques and Capillary Gas Chromatography, Journal of AOAC International, 76(6):1213-old.

Netherlands National Institute of Public Health and the Environment (RIVM), 2001, Re-evaluation of human-toxicological maximum permissible risk levels. RIVM report 711701025.

Rinsky, R.A., Young, R.J. and Smith, A.B., 1981, Leukemia in benzene workers. American Journal of Industrial Medicine. 2:217-245.

Rinsky, R.A., Smith, A.B., Horning, R.W, Filloon, T.G., Young, R.J., Okun, A.H. and Landrigan, P.J., 1987, Benzene and leukemia: an epidemiologic risk assessment. New England Journal of Medicine. 316 (17):1044-1050.

Rothman, N., Li, G.L., Dosemeci, M., Bechtold, W.E., Marti, G.E., Wang, Y.Z., Linet, M., Xi, L.Q., Lu, W., Smith, M..T., Titenko-Holland, N., Zhang, L.P., Blot, W., Yin, S.N. and Hayes, R.B., 1996, Hematotoxicity among Chinese workers heavily exposed to benzene. American Journal of Industrial Medicine. 29:236-246.

Toxicology Excellence for Risk Assessment (TERA),Voluntary Children's Chemical Exposure Program (VCCEP), 2006, Tier 1 pilot Submission for Benzene (Cas No. 71-43-2).

United Kingdom Food Standards Agency (U.S. FSA), 2006,  Survey of benzene levels in soft drinks

United States Food and Drug Administration (U.S. FDA), 2006, Letter Regarding Benzene Levels in Soft Drinks (letter from Dr. Robert Brackett, Director, Center for Food Safety and Applied Nutrition, to Mr. Richard Wiles, Environmental Working Group).

United States Environmental Protection Agency (U.S. EPA), Office of Research and Development, 2004, Human Exposure and Dose Modeling for Benzene: An Urban Area Case-Study, National Exposure Research Laboratory Research Abstract

United States Environmental Protection Agency (U.S. EPA), Integrated Risk Information System (IRIS), 2003,  Benzene (CASRN 71-43-2).

Wallace, Lance, 1996, Environmental Exposure to Benzene: An Update, Environmental Health Perspectives, 104(Supplement 6):1129-1136.

World Health Organization (WHO), 2003,  Benzene in Drinking-water - Background document for development of WHO Guidelines for Drinking-water Quality, Geneva, Switzerland.
And:  GUIDELINES FOR DRINKING-WATER QUALITY312 WHO (2003) Bentazone in drinking-water. Background document for preparation of WHO Guidelines for drinking-water quality. Geneva, World Health Organization

Appendix 1: International Guidelines for the Ingestion of Benzene

HC/CFIA Health Risk Definitions - Applied to Microbiological, Nutrition and Chemical Hazards, May 13, 2005

Health Risk 2: The health risk identified represents a situation where there is a reasonable probability that the consumption/exposure to a food will lead to temporary or non-life threatening health consequences, or that the probability of serious adverse consequences is considered remote.

>Health Risk 3: This represents a situation where there is a reasonable probability that the consumption/exposure to a food is not likely to result in any adverse health consequence. The situation identified may be an indication of a breakdown in Good Manufacturing Practices (e.g., sanitation, quality issues, etc.); in Good Agricultural Practices (e.g. pesticide residue in food above the established MRL); in Good Practices in Veterinary Medicine (e.g. animal drug residue in food above the MRL) or some other relevant factor (e.g., food containing non-permitted nutrients or food additives above the permitted levels, nutrients that do not meet label claim, health-related labelling infractions, etc.).