May contain nuts” pops up everywhere now - I found it on a boiled sweet wrapper the other day. It's another tune from the chorus of fusspot Britain, where the most ordinary acts of ingestion seem to need oversight from health and safety. (Exaggeration? I asked for a lamb burger cooked rare in All Bar One - they wouldn't do it unless I signed a waiver form.)

But peanut allergy needs to be taken seriously. I've sat across a gastro-pub table as a peanut-intolerant friend went into full anaphylactic shock over a prawn curry - her throat and lips visibly swelling in front of us. She got to hospital just in time for them to keep her breathing; the pub later admitted that the sauce for the chicken satay might have strayed across the prep area.

Unlike most allergies, it's not curable. With each exposure the reaction gets worse: it can kill, and it does. 40,000 people suffer life-threatening attacks every year. And the problem is spreading at a terrifying rate: numbers affected in Britain doubled between 2001 and 2005. Now nearly two children in 100 have a full peanut allergy - which is why every school, every restaurant and every parent has to be on guard.

Scientists have disagreed for some time over the causes of peanut allergy. Since 1998 the Department of Health has advised parents that if there's some allergic history in the family, they should avoid giving any peanuts to infants. Many researchers noted that since that advice lodged in the worrying public's mind, the incidence graph started to climb strikingly.

Now a fascinating piece of research has emerged; the results of a study of more than 8,000 children in Britain and Israel. Israelis are genetically similar to us, but almost all Israeli children eat a peanut paste as their first solid food. And in Israel the incidence of peanut allergy among children is less than two in 1,000. From eight months old the average Israeli child eats 7g of peanut a month. Most British children eat none.

The report, published in last month's Journal of Allergy and Clinical Immunology, concludes: “These findings raise the question of whether early introduction of peanut during infancy, rather than avoidance, will prevent the development of peanut allergy.”

Surprised? All over the developing world infants are introduced to all sorts of foods that would horrify us. Groundnut and peanut pastes are commonly put in the mouths of babes and sucklings in Africa. In East Asia chewed-up rice spiced with dried fish is fed to babies. And in these societies, allergic diseases are rare to the point of non-existence. This knowledge has at last penetrated the thinking of Western medical researchers: it's now widely thought that early exposure to potential allergens such as cats, wheat, eggs and milk may build up immunities and prove the best way of avoiding allergies.

So should new parents start on the peanut-butter soldiers immediately? Not yet, says Professor Gideon Lack, who is running a six-year study at King's College, London: “There's reasonable doubt about the causes of peanut allergy. But I would urge parents not to feed babies peanuts until the research is complete.” That won't be until 2012. Meanwhile the Department of Health is reviewing its official advice.

While you wait for the wheels of medical science to complete their motions, here's another fact to chew on. One adult in five believes that he or she is intolerant to certain foods. But laboratory testing confirms the existence of a problem in fewer than one in 20.

alex.renton@thetimes.co.uk