Project Denver and the Soft Drink Industry's Benzene Problem

Dr. Glen Lawrence, who was science advisor to the FDA's New York labs in 1990 and 1991, recently advised me that he is shocked to see that there are drinks that still contain the ascorbic acid-benzoate combination that is known in the industry to lead to benzene formation. The Professor had published a lucid explanation of the chemical interaction involved in a peer reviewed journal cited below. Dr. Lawrence noted that benzene is associated with leukemia as a carcinogen, and it can take many years before the leukemia develops. About 30 percent of cancers in children ages 0-14 years are leukemia. He explained that school children exposed to benzene in drinks may not develop leukemia until they are in their 20s.

I had just been given internal soda companies about the formation of benzene in soft drinks by an industry whistleblower who had been part of a secret research project. I knew that that the situation was much more serious than even Dr. Lawrence realized. The problem is especially dire in warm climates or where the technical fix to avoid the formation of benzene in soft drinks is not being used.

Last week, there was a recall by Nestle after inked from the packaging leaked into the milk served babies. The incident was reminiscent of the benzene first discovered in Perrier's carbonated bottled water in 1990 and then in some other fruit-flavored water (McKesson) and (Koala Springs) later in the year. Perrier's benzene contamination was due to contaminated carbon dioxide (from the ground) . The others were due to the breakdown of the artificial preservative benzoate (and there were a number of regional recalls that year involving bottled waters). In testing, the FDA was also finding benzene, for example, in orange soda.

The soda pop contaminated with benzene in 1990 escaped the public's notice. But it most definitely was known by company officials by December 1990. The companies commenced frantic, secret research projects. Coca-Cola, Pepsi, and Cadbury did not recall any products that tested above the limits accepted for water supplies. At Cadbury, the research project was known as Project Denver. While Perrier had been recalled globally at 11-18 parts per billion (ppb), Diet Orange Crush, for example, tested at 25 ppb before exposure to heat and 82 after exposure. (RSSL, a global leader in analytical testing serving food companies, did the testing) Under European regulations the standard for benzene in drinking water is now 1 ppb. The regulators in Europe I've contacted are not aware of the tendency of benzene to form from the mere combination of certain ingredients. As a result, no regular testing is done. For analytical chemists worldwide unfamiliar with the chemical process involved, please consult the peer-reviewed published article by the FDA consulting scientist "Benzene Production from Decarboxylation of Benzoic Acid in the Presence of Ascorbic Acid and a Transition-Metal Catalyst." Journal of Agricultural and Food Chemistry (May 1993) .

Coca-Cola issued a dramatic recall of Dasani from Britain for the carcinogen bromate last year. Five years earlier, many schoolchildren got sick after drinking soft drinks in 1998 and 1999 after the carcinogen benzene was found in many of its products. There were massive recalls in Belgium, UK and France followed by explanations deemed unconvincing by investigators. "Coke's contamination story 'highly unlikely,' " BBC News, August 17, 1999. Coca-Cola's CEO pointed to phenol (a derivative of benzene) as due to fungicide on some wooden pallets in France. Similarly, in Israel, the year before, a company spokesman pointed to paint thinner absorbed from products in various stores in Israel. (In Bloomberg at the time, it was reported that it was benzene that was found). There must never be a repeat of that experience or tolerance of the industry's failure to fully disclose the nature of the underlying benzene problem. In lay terms, the formation of benzene is caused by the breakdown of sodium benzoate in the presence of ascorbic acid and some other ingredients. Only then other competing explanations be fairly judged like contaminated carbon dioxide, fungicide on pallets, gear lubricants, cleaning solvents, paint thinner in stores etc.

Pepsi's experience in the US is illustrative of the need for testing. For example, in 1996, Pepsi recalled approximately 30,000 cases in bottles and cans, to include Welch's sparkling grape soda, due to an off odor and taste. In 1997, PepsiCo South recalled 137,000 cases (all sizes and package types) of a wide variety of drinks -- mainly diet drinks and citrus flavors -- due to an off odor and taste. In 2002, a recall of 7200 cases of Mountain Dew was suggested as possibly due to contamination from equipment cleaning fluid. Not to be outdone, in 1997, Coca-Cola recalled 300,000 bottles and cans due to a off odor and flavor. Often a recall is attributed to a gear lubricant, such as recalls by Coca-Cola bottling companies in the US in 1990, 1992 and 1994 or a recall in Australia in 2002. Benzene is a common ingredient in non-food grade lubricants. According to a 2004 Shell Oil comment at the FDA continued to be widely used in the beverage industry. So one question that arises is: was the determination that the contamination due to a gear lubricant in many of the cases just based on testing that showed benzene?

When sued in 2001 by someone who has drank a soda laden with lubricant in 2001, Coca-Cola declined to list the ingredients of the lubricant or describe how it got there. Thus, even where a particular incident has nothing to do with the benzene formation at issue, candor with consumers as to the causes of contamination is not a trademark. When a consumer complains that a soft drink has an off odor and taste, someone out in the field who may not have been in a food factory for years, may then go to try to pinpoint the problem. The inspector has been trained by the FDA in diplomacy in dealing with managers who don't even want to hear the word "recall". Faced with only an off odor and taste, they are left to the good intentions of the company. Testing for benzene should be done not only upon complaint, but routinely as is done in the case of municipal drinking water and bottled water.

France has banned all vending in schools. England just announced it will go soda free all grades. Public schools in Australia, India, Scotland and Wales likely will soon go soda-free. In the US, Schwarzenegger signed a law that will make California schools soda free (and Massachusetts and Arizona are likely next). Maine and New Jersey took decisive administrative action with less fanfare. From an insider's perspective, however, diabetes, obesity, osteoporosis, dental decay, caffeine addiction, and behavioral problems are just some of the health problems raised by the consumption of soda by a captive audience of children at school. Health advocates just are not yet informed by the documents now being made available by an industry whistleblower that I represent about the tendency of the known carcinogen benzene to form in soft drinks.

The internal soda company documents show that in late 1990, the benzene was found to be from the breakdown of sodium benzoate in the presence of ascorbic acid, rather than from contaminated carbon dioxide (such as had led to the Perrier recall earlier that year). Benzoate converts to benzoic acid in an acidified environment such as most soft drinks. The benzoic acid then interacts with the ascorbic acid in some situations and forms benzene. Instead of disclosing to consumers the nature of the underlying benzene problem -- at the very least on the occasion of the massive recalls in 1998 and 1999 -- Coca-Cola and its main bottler CCE offered up unconvincing explanations about fungicide on wooden pallets and paint thinner absorbed from products at the various stores. They then emphasized the political fix throughout the late 1990s rather than the technical fix in marketing to a captive audience of schoolchildren. Benzene is a nasty carcinogen that is taken seriously by environmental authorities. It also should be taken seriously when children are encouraged to pour it down their gullet instead of water by some popular teen celebrity.

As far back as 1990, some well-known products tested far above accepted limits for drinking water. Some products tested above even the 20 parts per billion ('ppb"), the highest level of the 1998 recalls of numerous soda products and nationwide bans in Europe -- and far higher when exposed to heat and light. Yet, there was no recall and no disclosure to consumers of the 1990 testing results by the major companies. See J. M. Packman (Coca-Cola's Food Law Counsel) , "Civil and criminal liability associated with food recalls," Food and Drug Law Journal 53 (1998), pp. 437–452. (Indeed, some US soft drink products tested just this past week tested well above that level, even before exposure to heat or sunlight). What did the NSDA representatives tell the FDA at the meeting on December 17, 1990? (What testing data precisely was disclosed?)

I am awaiting a response on a FOIA request submitted to the FDA and have a lot of confidence in the agency's good faith and their continued efforts to safeguard the public health. The agency is faced with budgetary constraints, massive responsibilities, and limited enforcement powers. The FDA will continue to need candor and prompt action by industry in responding to problems as they arise.

An uploaded internal document explains that the benzene generated in U.S. Diet Crush before exposure was 25 ppb and after exposure to 16 hours of UV ~ 30 degrees C was 82 ppb. That's 16 times the level in the US at which newspapers have to be notified (to tell consumers not to drink the water) for municipal water supplies. After exposure to the heat and UV, Diet Slice was tested by RSSL at 41.5 or 8 times the maximum level of 5 ppb permitted for water supplies in the US.

In the US, if there are 5 ppb benzene detected in a water supply, radio and newsapers have to be notified. (EPA Consumer Factsheet on: BENZENE.) That level is not merely not "fit for its intended purpose" under the product liability laws. Upon drinking over a prolonged period of time, that can be cancer in a can. The cause addressed by Project Denver, rather than contaminated carbon dioxide, was due mainly to the interaction of sodium benzoate and ascorbic acid and was greatest when the product was exposed to heat or ultraviolet light. But hundreds of pages of Project Denver documents, to include testing of competitor products, could speak for themselves. Calcium disodium EDTA (Calcium disodium ethylenediaminetetraacetate) is one technical fix that is intended to avoid the problem. Under the Code of Federal Regulations, it is permitted for canned carbonated drinks; Dow, the manufacturer, advises that its approval was related to the concern for degradation of the metal can. There are other possible fixes. Why are there so many products worldwide with no technical fix apparent from the ingredients? Is it because calcium disodium EDTA is not deemed safe and not permitted in drinks in EU and Australia? Under European regulations the standard for benzene in drinking water is now 1 ppb under the Quality of Water Intended for Human Consumption. Worldwide, outside of the US and Europe, the WHO recommended guideline level of 10 ppb in drinking water is useful in assessing what maximum level might be acceptable for soft drinks.

Legislators should rid schools of soda. (Editorial, "Junk food out of British schools," The Post and Courier, Nov. 19, 2005; Editorial, "Youth obesity far bigger than Coke," Atlanta Journal Constitution, Dec. 1, 2005.) But if it is allowed, state legislators should require testing for benzene given the number of instances of schoolchildren getting sick . Water supplies are regularly tested and bottled water in the US is subject to regulations relating to permissible benzene levels. If soda is being urged by soda companies and some school administrators and legislators as a substitute for water, then soft drinks should be tested also. If the American Beverage Association ("ABA") President Susan Neely truly wants to be "proactive," then the ABA should support regular, transparent and independent testing such as is done for tap water and bottled water. Given Coca-Cola and Pepsi have actively opposed disclosing the levels of pesticides in their drinks in India, the companies cannot be relied upon to ensure that no product above the safety guidelines is sold. We should hold our beverage companies who provide the beverages for profit to the same safety standard as our ground water and drinking water.

Regulators around the world: Priority in government testing by agencies such as the Food Standards Agency in the UK should be given to diet drinks with benzoate and ascorbic acid but not any of the technical fixes. As explained in the internal memorandum written by the Cadbury Vice-President at the time, the effect was greatest in diet drinks. This likely would have been because of the absence of the insulating effect of the sugar. According to my client who is available to take a call from you and who used to work on this research relating to benzene when the issue first arose, the beverage with the greatest risk would contain the combinations as follows: (1) diet or reduced sugar, (2) benzoate, (3) ascorbic acid or its sister erythorbic acid, (4) juice, and (5) citrus or cherry flavor. It is also important to test for undisclosed calcium disodium EDTA and ensure that any use of calcium disodium EDTA is approved as safe for the use. The strategy that should be taken in each country will vary with the products. For example, the analysis in testing in the United Kingdom might start with the list of 85 or so soft drinks marketed to children in the UK that contain sodium benzoate. Then testing generally could be limited to those that contain ascorbic acid (especially, for example, citrus flavors) that do not have any apparent technical fix.

In terms of developing factual information about the discovery of the benzene problem at Coca-Cola, Dr. Michael E. Knowles from Coca-Cola , Director Scientific & Regulatory Affairs was head of the UK Fisheries and Food's Food Science Division from 1986 to 1989. From 1989-1991, Dr. Knowles was Chief Scientist (Fisheries and Food) and Head of the Food Science Group. After the 1990 benzene crisis involving Perrier (and the problem was discovered concerning soft drinks), he was hired by Coca-Cola in 1991. He likely would know both what was disclosed to the regulatory authorities and what was known by Coca-Cola. At Pepsi, Louis Imbrogno is a key senior technical executive who was with the company in 1990 and is still there. He might be able to shed light on the testing results shared with regulatory authorities.