ARCHIVED - Survey of Benzene in Soft Drinks and other Beverage Products

Background

Benzene, a known carcinogen to humans, is naturally occurring (for example, in crude oil) but is also manufactured for use in the industrial sector. Benzene is released to the air from various sources, including industrial emissions, fires, vehicle exhaust and tobacco smoke. Because benzene readily evaporates into the air from gasoline, gas stations are also a source of benzene to the air. Accidental spills and industrial waste water, for example, may introduce benzene to water and soil. Since benzene is an environmental contaminant that can be absorbed by plants and animals, food products may contain trace levels of benzene.

However, in the early 1990s, it was found that benzene can also be formed in soft drinks containing certain food preservatives and nutrient additives. It was shown that benzoate salts, used as an anti-microbial agent in certain soft drink products, can react with ascorbic acid (Vitamin C) to form benzene, especially in the presence of light and elevated temperatures. Ascorbic acid may be either naturally present from a fruit juice ingredient in the soft drink or added as an antioxidant food additive.

Shortly after the finding that benzoates and ascorbic acid can react to form benzene, Health Canada conducted a survey of benzene in fruits, fruit juices, fruit drinks, and soft drinks (Page et al., 1992). Benzene levels in the drinks were below the Canadian guideline (maximum acceptable concentration) of 5 microgram per liter (µg/L) benzene in drinking water. Benzene levels in the fruit samples were also below 5 µg/L. These levels were determined to be of very low concern to human health.

Recently, a private laboratory in the United States informed the U.S. Food and Drug Administration (FDA) that levels of benzene in a small number of soft drinks containing benzoate salts and ascorbic acid exceeded the U.S. Environmental Protection Agency (EPA) maximum contaminant level (MCL) of 5 µg/L for benzene in drinking water. Immediately after learning of the U.S. results, Health Canada initiated a survey of benzene in soft drinks as well as in some low alcohol drinks and cocktail mixers in order to assess benzene levels in products available in Canada. Samples of 118 products were analysed for benzene. Benzene was found in samples of four of the products at levels above the Canadian guideline of 5 µg/L benzene in drinking water. In these samples, average concentrations ranged from 6.0 to 23.0 µg/L. Benzene in most of the remaining products was either not detected or was detected at levels below 5 µg/L.

Sampling Plan and Analytical Methodology for Soft Drinks and other Beverage Products

Samples of a total of 118 soft drink and other beverage products were collected (two of the products were collected from different lots). Only those products that contained benzoate salts and/or ascorbic acid were gathered. Most of the products are either carbonated or non-carbonated soft drinks, with a few low alcohol (0.5% alcohol vol.) drinks and cocktail mixers included. Several cranberry drink products were also included since cranberries are known to contain benzoic acid as a natural constituent. At least four samples of each of 118 products were purchased in April 2006 in a local grocery store in Ottawa. Additional samples of one of the products and samples of five additional (reformulated) products were provided directly by the producers in May 2006 after being notified by Health Canada of the results from the samples collected in April. The five additional products had been prepared using new formulations that do not require the addition of benzoate salts or replace an ingredient that naturally contained benzoic acid. Microbial stability in reformulated product that eliminated the use of benzoate salts is achieved using hot (aseptic) packing of the soft drink product.

The samples were analysed by two laboratory methods based on headspace gas chromatography-mass spectrometry (GC-MS) analysis.

The Data on Benzene Levels in Soft Drink and other Beverage Products

Table 1(a) summarises the levels of benzene determined in samples of ready-to-consume products. Table 1(b) summarises the levels in concentrated products that require dilution before drinking. Products that were analysed subsequent to the initial survey of 118 products are also included in the tables.

The limitations of the data from this survey should be noted. In most cases, samples of each product were taken from only one production lot. Therefore, variations from one production lot of a product to another lot were not addressed in this survey. The data may not represent the distribution of benzene in soft drinks in the general Canadian food supply. It should also be noted that if a product brand does not appear in either Table 1(a) or 1(b), it means only that samples of that product were not collected.

Benzene was either not detected or its levels were below the method reporting limits (the lowest concentration that can be reliably measured) in approximately 80% of the 118 products. Samples of 16 products contained benzene at levels equal to or less than the Canadian guideline of 5 µg/L for benzene in drinking water. There were four products (about 3%) with benzene levels greater than 5 µg/L. Two of these products contained benzene at average levels above the WHO guideline for benzene in drinking water (10 µg/L). The average benzene levels in cherry flavoured Kool-Aid Jammers "10" drinks (old formulation) from Lot 1 were 15 µg/L and from Lot 2, 19 µg/L. The average benzene levels in tropical punch flavoured Kool-Aid Jammers "10" drinks (old formulation) from Lot 1 were 18 µg/L and from Lot 2, 23 µg/L. Benzene levels in a variety of flavours (blue raspberry, grape, orange, kiwi strawberry) of Kool-Aid Jammers (a unique product that is not the same as Kool-Aid Jammers "10") were all below 5 µg/L. Benzene was not detected in any samples, provided by the producer, of reformulated Kool-Aid Jammers or reformulated Kool-Aid Jammers "10" products.

The average benzene level in Mr & Mrs T Margarita Mix was 9.9 µg/L, but the benzene level dropped to 2.5 µg/L in a sample of the same product provided by the producer.

The benzene level in Rose's Cocktail Infusion - cranberry twist mix was 6.0 µg/L, but it was not detected in the reformulated product provided by the producer.

Conclusions

Samples of soft drink products in this survey were collected from only one grocery store in Ottawa . This store had a supply of a sufficient variety of soft drink products to permit an assessment of benzene levels in soft drinks in Canada in a very short period of time. Variations from one production lot of a product to another lot were not fully addressed in this survey. Samples of only two products were collected from two different lots, while the remaining products were collected from a single lot. Thus, consumers should be aware of the limitations of the data from this survey and that the data may not represent the distribution of benzene in soft drinks in the general Canadian food supply.

The data demonstrate that for the majority of product samples analysed, benzene could either not be detected or was present at levels that were too low to reliably measure. For those few products in which benzene levels were first found to be in excess of the Canadian guideline of 5 µg/L benzene in drinking water, significant reductions of benzene levels were observed in those that were reformulated.

Notes for the Tables

• "n.d." means that benzene was not detected. Where benzene was detected, confirmation of benzene identity was based on three criteria: (1) the retention time (within ±0.01 min of the standard); (2) the ion ratio of m/z 77 divided by m/z 78; and (3) the ion ratio of m/z 51 divided by m/z 78 (both ion ratios have to be within ±25% of the standard). Benzene was not considered to be detected if any of the above three criteria were not met.
• "< 1" and "< 2" mean that benzene was detected in the products, but the levels were less than the reporting limit of either one of the two GC-MS based methods, which are 1 µg/L for method 1 and 2 µg/L for method 2 (the reporting limit is the lowest concentration that can be reliably measured).
• For most of the products, only one of the four samples collected was analysed, either in duplicate or triplicate, and their results are shown as the average of the replicate analyses. For some other products, multiple samples from the same lot were analysed, and their results are shown as both the average and, in parenthesis, the range of values.
• Same products from different lots were treated as different samples, and their results are shown separately.

† These products were provided directly by the producers. The producers of Kool-Aid Jammers, Kool-Aid Jammers "10", and Rose's Cocktail Infusion - cranberry twist mix have indicated that these products have been reformulated to eliminate the formation of benzene.

* Some product labels provide instructions for use as a dessert topping, beverage, etc. Only instructions pertaining to consumption of the product as a beverage are provided in the table. Where more than one instruction for serving as a beverage is given on the label, only one set of instructions is provided as an example in the table.

† These products were provided directly by the producers. The producers of Kool-Aid Jammers, Kool-Aid Jammers "10", and Rose's Cocktail Infusion - cranberry twist mix have indicated that these products have been reformulated to eliminate the formation of benzene.

References

Page, BD; Conacher, HBS; Weber, D; Lacroix, G (1992) A survey of benzene in fruits and retail fruit juices, fruit drinks, and soft drinks. Journal of AOAC International, Vol. 75, No. 2, 334-340.