Deciding on the safety of certain food additives is no piece of cake, especially when the additive is used widely and the food industry is heavily dependent on it. A case in point are two food preservatives, BHA (butylated hydroxyanisole) and BHT (butylated hydroxytoluene).

In 1983, scientists at the Food and Drug Administration received some bad news: A Japanese researcher reported that BHA caused cancerous tumors in the forestomach of rats.

Since then, forestomach tumors have shown up in other rat studies involving BHA. The law -- specifically the Delaney clause of the Food, Drug and Cosmetic Act -- requires FDA to ban food additives that cause cancer in animals. Yet FDA officials have decided, at least for now, to keep BHA on the market.

Why hasn't FDA taken action? "We don't believe there is likely to be human risk from ingesting low levels of BHA," says Robert Scheuplein, deputy director of the Office of Toxicological Sciences in FDA's Center for Food Safety and Applied Nutrition.

Speaking last month at a three-day conference sponsored by the International Life Sciences Institute-Nutrition Foundation, an industry-funded research group whose members are BHA users, Scheuplein pointed out that the key issue in a safety evaluation of BHA is "the relevance of the animal data to human risk."

BHA and BHT belong to a class of food additives called antioxidants, which suppress the natural reaction of fats and oils in foods to combine with oxygen and get rancid. So ubiquitous are these compounds that most Americans consume several milligrams of them each day. They extend the shelf life of vegetable oils, nuts, spices, processed meats, chewing gum, breakfast cereals, shortening, snack foods like potato chips, cake mixes and many other foods.

Rancidity is not only undesirable from an esthetic point of view. The oxidation process can damage vitamins.

In addition, the byproducts of the oxidation process are believed to be toxic themselves. Paul Addis, a food scientist at the University of Minnesota, noted at the conference that coronary artery disease may in part be caused by the consumption of these oxidation products, especially the cholesterol oxides. Some of the oxides are thought to be carcinogenic. (One supplement manufacturer is advertising its Vitamin C and E pills as protectors against the so-called "free radicals" -- the chemical intermediaries involved in the oxidation process. Vitamins C and E have antioxidant properties and are used by the food industry in some products for that purpose.)

Here are some of the complex and conflicting issues that FDA toxicologists must consider in their safety evaluation of BHA:

*Cancerous tumors in the rat studies appear to be confined to the forestomach and show up only at the highest doses. People do not have forestomachs, and in dog, pig and monkey studies on BHA, stomach tumors were not detected. However at the meeting there was some discussion that cell changes may have been detected in the esophagus of some of the monkeys and pigs.

*BHA has been shown to be a cancer "promoter." Under certain experimental conditions, when rats are given a known carcinogen, followed by BHA, there are more tumors than when the carcinogen is administered by itself.

BHA also appears to inhibit some tumors, thus acting in some situations as a cancer preventer. In studies reported at the meeting by Dr. Nobuyuki Ito, the Japanese researcher who first detected the forestomach tumors, BHA inhibited liver and mammary gland tumors in rats who had been given potent carcinogens.

*BHA may interact with BHT, which is often used with BHA in the same foods. Michael Thrush of Johns Hopkins University told the meeting that the toxicity of BHT increased when mice were also given BHA. Another researcher, Harry Conacher of Canada's Health Protection Branch, reported that more BHA was retained in rats' tissues when the animals had been fed BHA and BHT in combination than when BHA was fed alone.

If researchers could discover how BHA is working in rats to produce the forestomach tumors, they might be able to determine if that same mechanism could produce tumors in humans. At the meeting, some of the researchers said they believed the cancer is really a consequence of a toxic dose of the substance to the animal -- in other words, the dose is so high it poisons the animal, and the cancer is an outgrowth of that poisoning.

BHT also has been under a cloud, with safety studies presenting mixed results. It does appear to "promote" liver tumors in laboratory animals, although in some animal studies it also inhibits tumor growth.

Many manufacturers are now switching to TBHQ (tertiary butylhydroquinone), a chemical relative of BHT. However, the United States is one of only nine countries in the world that currently permits TBHQ for use in any foods -- most believe the data is inadequate to show safety. The United States is planning full-blown safety tests on the additive.

Ironically, some health foods stores are selling BHT in supplement form, based on unproven claims that the pills will serve as "life extenders" or will treat genital herpes. "There is a population of people exposing themselves to BHT in other than the normal route of diet," noted Thrush at the meeting, suggesting that this will have to be considered in evaluating human risk.

Where does this all leave the average consumer who, according to a Canadian study, probably has some BHA and BHT accumulated in his tissue? Michael Jacobson, executive director of the Center for Science in the Public Interest, says the risk posed by BHA and BHT is "very small," but one the government should eliminate. "It doesn't make sense to have that kind of question mark in our food supply," he insists.

He says consumers can avoid the additives by staying away from the highly processed and often-times high-fat foods to which they are added. "Usually you can find a competing brand without it," Jacobson says, pointing out that this raises questions about the real necessity of the preservatives.

However, as John Findley of Nabisco Brands Inc., noted at the meeting, "Is 'natural' necessarily better when it's rancid?"