COMBINE THE THREE WORDS cancer, chemicals and babies and you are well on the way to creating a scare.

Parents throughout Europe shivered yesterday when the European Food Safety Agency announced that baby foods could contain a chemical that could cause cancer. The villain is called semicarbazide, and it originates from the seals used to keep the contents of bottled food safe from bacterial contamination.

There were plenty of reassurances that the risk, if any, was very small, but in stories like these it is the alarm and not the reassurance that is the take-home message. So what exactly has been found? Is it new? And is it dangerous?

The discovery of low levels of semicarbazide was made several months ago during routine monitoring by a private laboratory on behalf of a food manufacturer. Semicarbazide belongs to a family of chemicals called hydrazines that are known to be carcinogenic in animals, so the discovery was immediately reported to the authorities.

To begin with, it was not clear whether the presence of semicarbazide might be an artefact of the testing procedure. But further tests seem to have eliminated that possibility. The current theory is that semicarbazide is produced from another chemical, azodicarbonamide, which has been used for the past 20 years as a “blowing agent” for plastic production.

Added to plastics during processing, blowing agents form minute cells of gas throughout the material, providing the springiness needed to make an effective seal. The two chemicals are close cousins, so it is reasonable to conclude that semicarbazide is created during the heat-treatment of azodicarbonamide in the manufacture of the seals.

If so, babies fed on food from jars, and anybody else eating products such as fruit juices, jams, honey, ketchup and mayonnaise, pickles, sterilised vegetables and sauces has probably been ingesting tiny quantiities for the past two decades. What has changed is not the food we eat, but the analytical techniques used for detecting tiny impurities in it.

The actual amounts of semicarbazide found in food are minute, ranging from one part per billion up to 25 parts per billion. Baby foods have the highest level because they are packed in smaller jars, so the ratio of seal to food is higher.

One part per billion is roughly equivalent to a single drop of ink in a road tanker full of fuel. A generation ago such small concentrations were beyond detection, so nobody worried about them. Today we do.

The next question is whether semicarbazide in these amounts constitutes any real hazard. In animal experiments, it does not cause cancer in rats, but it does increase the prevalence of cancers of the lung and blood vessels in mice — though only if they are female. A carcinogen that can distinguish the sex of its victim must raise some suspicions about the reliability of the tests.

The doses needed to achieve this effect in female mice are 100 milligrams per kilogram of bodyweight per day throughout a lifetime — making semicarbazide one of the weakest carcinogens among the hydrazines tested.

For the purposes of comparison, the EFSA worked out a “worst-case analysis” for a six-month-old baby given food containing the highest-recorded level of semicarbazide, 25 ppb. Such a baby’s intake would be 2.3 microgrammes per kilogram of bodyweight per day — less than one 40,000th as much as was needed to harm the female mice, and consumed for a much shorter period of time.

Small wonder that the EFSA and the British Food Standards Authority rated the risk “very small” and felt no need to advise parents to stop buying bottled baby food.

In fact, the real risk is not from semicarbazide but from the constant bombardment of messages from guardians of the public health that do more to confuse than they do to safeguard. The media is often blamed for these messages, but they originate from publicly funded organisations such as the FSA or the EFSA which are so terrified of being accused of concealing risks that they broadcast them, passing the responsibility for making a judgment on to the public.

Last year, the scare was acrylamide, a chemical produced by cooking starch-containing foods at high temperatures. According to this scare, foods such as baked potatoes, chips, biscuits and bread are all capable of causing cancer. Before that, the FSA warned that the burning of carcasses of animals after the foot and mouth epidemic could contaminate milk with dioxins. This was a theoretical risk which turned out not to be true, but plenty of people were alarmed and some probably stopped drinking milk.

The danger of trading in theoretical risks is that it may send you stumbling into real ones. The Department of Health elected to use disposable instruments for tonsillectomies to reduce the theoretical risk of transmission of variant CJD, and produced a real risk: the clumsy disposable instruments caused increased levels of bleeding, and two patients died.

The danger with the semicarbazide scare is that effective seals will be replaced with less effective ones, babyfood will become contaminated with bacteria and babies will die of food poisoning. The manufacturers will be blamed.

Once, expert organisations shouldered the burden of making judgments, because it was their job. Sometimes they got it wrong, but far more often they got it right.

Today they see themselves simply as conveyor belts for transmitting the latest scare to the public, simultaneously reassuring them that it is not really anything to worry about.

No wonder people are confused — and worried.

DEBATE

Should we take food scares seriously?

E-mail debate@thetimes.co.uk