When my older daughter was born, in 2000, I was told not to feed her any of the most common allergens until she was a year old: no eggs, seafood, chocolate, nuts, or peanuts. Especially no peanuts.
We followed the rules as we understood them, both with her and with our younger daughter. I have friends with food-allergic kids; this is a danger I take very seriously, although allergies didn’t particularly run in either family so we didn’t wait until two or three or five to introduce peanut products, like some of my friends did. But, the people who were delaying further and further noted that the number of kids with dangerous allergies was still increasing, so…
…yeah, it turns out we were doing everything wrong. According to a new study published last month delaying the allergenic foods was the worst possible thing to do.
This wasn’t entirely new news. Back in 2008, a study came out comparing children in the UK (where peanuts were verboten until age three) to children in Israel. Israel either missed or ignored the memo about withholding peanuts, and peanut-laden foods are a staple of toddlerhood. The researchers compared a group of Israeli kids to a group of Jewish kids in the UK, and discovered that peanut allergy rates were ten times higher in the UK.
But the new study making news this week actually did a side-by-side randomized comparison — though not placebo-controlled for practical reasons.
The researchers recruited 640 babies at high risk for peanut allergy. I should explain, actually, that the way you know a baby is at high risk for developing food allergies is if they have atopy, or atopic syndrome, which is a cluster of allergy symptoms such as eczema. All the babies were given a skin test to see if they had peanut allergy already. Those with a mild allergy1 continued in the study. Participants were randomly divided into two groups: one avoided peanuts, and one was deliberately fed a peanut product.
The peanut-eating babies were fed Bamba, a snack that’s sort of the Israeli version of Cheetos, except that comparing it to Cheetos doesn’t make clear how ridiculously popular the stuff apparently is among Israelis. These peanut-enhanced corn puff thingies are the top selling snack in Israel; 90% of Israeli families buy the stuff regularly. Also, it sounds a little more like Pirate’s Booty than Cheetos, if Pirate’s Booty tasted like peanut butter instead of like Styrofoam packing material. They apparently kind of disintegrate in your mouth, so you can feed them to your infant as an early finger food.
The kids in the study were tested regularly throughout infancy and toddlerhood to see if they’d developed peanut allergy. The results were decisive. When they were five years old, of the kids who’d tested negative on the initial test, 13.7% of the peanut-avoiders were now allergic to peanuts. (Remember, these were high-risk kids to begin with.) Just 1.9% of the peanut-consumers had developed a peanut allergy.
Of the kids who had that initial mild positive reaction, 35.3% of the peanut avoiders were now solidly allergic compared to 10.6% of the peanut eaters. That also goes against the conventional wisdom that prevailed among parents for years — that a child with even a very mild reaction should be very careful to avoid the trigger food, to avoid making the allergy worse with increased exposures. Like the advice to delay the introduction of peanuts, this seems to have been exactly backwards.
This got me wondering why, exactly, we collectively concluded that peanuts should be treated like arsenic for the first one, two, three, or five years of a child’s life.
Was there a study? Was there anything that this rule was based on?
In 1989, there was a paper in the British Medical Journal on the development of atopic eczema in infants, and whether restricting the diet of breastfeeding mothers or giving the babies hypoallergenic formula could improve things.2 The researchers found that eczema was least common in breastfed babies whose mothers were following restricted diets, and in formula-fed babies who were drinking hypoallergenic formula. Next best was breastfeeding without a restricted diet, then soy formula, and finally cows’ milk formula, although I just want to note that there were fewer than 50 babies in each group so the difference between the soy and the cow’s milk formula was two babies who developed eczema.
This and a few other studies had created an ongoing debate, by the early 1990s, about whether mothers of babies with a strong family history of allergies should be counseled to avoid the most common allergenic foods while pregnant and breastfeeding. The problem is that pregnant women already have a long list of foods they’re not supposed to eat, and peanuts are a cheap, convenient source of protein and fat. Doctors are generally hesitant to tell breastfeeding mothers to restrict their diets because so many women will respond by switching their babies to formula, and the hypoallergenic formulas recommended in this study are shockingly expensive (and also so nasty that one of the studies commented that babies sometimes refuse to drink them).
The first half of the 1990s was also when concern about peanut allergies really surged. This was when schools first started restricting PBJs in lunch boxes; newspapers ran stories about children and teenagers dying tragically due to accidental exposures, as well as articles about concerned parents explaining that food allergies were no joke. Among medical researchers, everyone knew that peanut allergies were increasing, but no one knew why. (They still don’t know why – telling everyone to avoid peanuts for the first one to five years of a child’s life made the problem worse, but the initial rise predates the advice. I like the “hygiene hypothesis,” the theory that it’s because we’re all too clean now, because it transforms a messy house into virtuous, responsible parenting.)
Then, in 1996, there were two studies, both done in the UK and published in the BMJ, on peanut allergies, that seem to have set off a lot of worried coverage in the newspapers. In April, “Clinical Study of Peanut and Nut Allergy in 62 Consecutive Patients“, by Dr. Pamela Ewan focused on 62 patients she’d seen in her practice as an allergist: their allergies and reactions, when they developed their allergies, and what non-food allergies and atopic symptoms were seen in each patient. She didn’t exactly say straight out that young children shouldn’t eat peanuts, but that idea leaked in around the edges. She observed that “the incidence of allergy to each type of nut seemed broadly related to the relative amounts of each ingested in the population” – peanut was common, cashew not. She went on to note that a lot of peanut allergies developed very early and this “probably relates to early introduction of peanuts into the diet.” Later, she asserted (with no source) that “the diet of children one to two generations ago was much simpler, and peanuts or nuts were used less and introduced much later.” (Americans have been eating peanuts in quantity since the early 1900s, with peanut butter being a staple sandwich filling since WWII.) She did recommend withholding peanuts from high-risk children: “Avoidance of the allergens during the period when sensitization seems common, possibly to the age of 7 years, would be justified. There is a case for considering avoidance of peanuts — which are consumed mainly as peanut butter — in these children.”
That same issue also had an editorial from American allergist Hugh A. Sampson, and he was a lot less reticent with the recommendations. He called for doctors to identify high-risk infants and advise their parents to eliminate peanuts until the child was at least three; he wanted breastfeeding mothers counseled to eliminate peanuts from their own diets. He thought children being allergy-tested for any reason should be checked for peanut allergy, and if they showed up as peanut-allergic — even on a blood test — they should eliminate all peanut and nut products for three to five years.3
Later in 1996, the BMJ published a paper about a study done on the Isle of Wight to try to determine peanut allergy prevalence.4 All babies born between January 1989 and February 1990 were followed, with the consent of their parents, and monitored for certain allergies; the children were followed until they were four. One detail I found particularly interesting: the researchers noted that although they gathered data throughout the study on factors ranging from whether the parents smoked to whether there was a pet in the house, “questions about eating nuts were asked only at age 4 years,” which suggests strongly to me that no one considered this question important until the study was almost over. This could also mean they got really bad data. How many parents of four-year-olds can remember, out of the blue, when their kid first ate nuts? Unless it was super memorable because the kid had a dramatic allergic reaction.
Unlike the 1989 study, the Isle of Wight researchers found that it didn’t matter if mothers ate nuts while pregnant or breastfeeding. They also didn’t make any sweeping recommendations in the article, although another paper published in the same issue found that if you have a sibling with a peanut allergy, you’re more likely to develop one, and blamed mothers: “Peanut allergy is presenting earlier in life, possibly reflecting increased consumption of peanut by pregnant and nursing mothers.” I am beginning to think that “possibly” is medical journal code for “I have no evidence for this hypothesis whatsoever and certainly can’t offer you a citation, but it fits with my pre-existing biases and assumptions very nicely.”
“PARENTS WARNED OF PEANUT RISK TO CHILDREN” was the headline in The Independent when they reported on Ewan’s study in April 1996. Even though Ewan’s study restricted recommendations to children at high risk of developing peanut allergies, the very first sentence of the article was, “Parents were warned yesterday not to give very young children peanuts or other nuts as new research revealed that many were developing potentially life-threatening allergies.” The article went on to claim that the study’s findings “highlight the danger of giving peanut butter to babies. More than a third of the nut allergy cases reported occurred in children under two, most of whom were given the spread in the first 12 months of life,” and that (unnamed) allergy specialists “say that avoidance of peanuts/nuts by young children is the only safeguard.” It finished by noting that things were even worse in the U.S. and saying that “the growing range of peanut products, and peanuts as an ingredient in convenience and junk foods, is believed to be behind the surge in cases there.” Note the use of passive voice. Believed by who? It doesn’t say. Believed why? It doesn’t say that, either.
“YOUNG CHILDREN SHOW BIG INCREASE IN DEADLY PEANUT ALLERGY” was the Sunday Times of London headline in March of 1996; they were talking about the (not-yet-published) Isle of Wight study. The third paragraph in the article asserted that the study “has confirmed fears that the growing use of peanut extracts, especially in baby foods and skin creams, could have sensitised thousands of children to the nuts, giving them an allergy that will be with them for life.” Note: the study confirmed nothing of the kind. It confirmed that peanut allergies were indeed becoming more common, but not that the “growing use of peanut extracts” was any sort of cause. ((I mean, it did discuss where children might have encountered peanuts prior to eating them officially for the first time, since there were children in the study who reacted to peanut butter the first time they ever had it, and that meant they’d encountered peanuts before somewhere.)) The article also cited the lead scientist, David Hide, saying that the “the increase in the allergy, almost unknown 20 years ago, coincided with growing use of peanuts.” Hide died suddenly of a heart attack between interview and publication, so he wasn’t available for any clarifications later on.
Somewhat better reporting was provided by Science News (written for the science-y layperson), which at least grasped that Ewan’s study was making recommendations for children at high risk. They had the headline, “Family allergies? Keep nuts away from baby.” But they also focused on the finding that nut and peanut allergies usually developed by the age of three, and quoted Dr. Sampson at length: “We think the maturation of the immune system and its ability [to tolerate novel substances] is not well-established at a very young age. That’s why more food allergies occur in the first couple years of life.” They also reiterated his recommendation that at-risk children avoid peanuts and nuts. An opposing viewpoint was included, but it wasn’t anyone questioning the science behind the recommendation; instead, the argument was entirely pragmatic, pointing out that “because peanut butter is so tasty and nutritious…most American kids get fed a peanut butter and jelly sandwich long before they’re a year old.”
The advice here didn’t go completely unchallenged. In the letters column a few months after Ewan’s paper ran, Doctors Gideon Lack and Jean Golding, from the Royal Hospital for Sick Children, took issue with it: “Regrettably, [Ewan] does not provide any evidence to back her recommendation that ‘young allergic children should avoid peanuts and nuts to prevent the development of this allergy’ and her extraordinary suggestion that avoidance should be practised until the age of 7…[In fact,] exposure to peanuts and other food allergens during lactation and childhood may be important in the development of immunological tolerance and may prevent allergic sensitisation to these foods.” Dear Doctors Lack and Golding: it would probably be unprofessional to call up your colleague and say, “I TOLD YOU SO. IN 1996. I TOLD YOU.” But I hope you are at least relishing a sense of satisfaction that the current research suggests that you were absolutely right.
In 2000, the American Academy of Pediatrics issued a recommendation that for high-risk infants, dairy be avoided until 12 months, eggs until 2 years, and peanuts, nuts, and fish until 3 years. I am not sure whether they ever officially said anything about children more broadly, but certainly many of the mothers I knew seemed well aware that peanuts were off-limits until long after most other foods could be introduced.
Although the “wait, giving kids peanuts prevents food allergies!” news hit the media last month, the “avoid peanuts” recommendation has been questioned for a while, in part because of anecdotal reports from Israelis that despite feeding their kids Bamba like Americans feed their kids rice cereal, they seemed to know a lot fewer kids with food allergies than Americans did. In 2007, Science had a piece about whether avoidance might actually be causing problems rather than solving them, and quoted Dr. Sampson (author of the clarion-call “ban! all! the nuts!” editorial in 1996). “I’m still nervous about peanut,” he admitted, even though by 2007 the guidelines were being broadly questioned. “I don’t have proof,” he added, “I just have this sort of sense that there’s something different about it.”
That really seems like a mic drop moment. He had sort of a sense.
In a lot of ways, that almost underlines the problem and puts exclamation points around it. It would be nice — it would be great — if scientists were able to disregard the conventional wisdom about whatever it is they’re studying. It would be lovely if they could look at the evidence and base their recommendations on what they actually found. If they could start by studying the question of whether withholding peanuts would reduce peanut allergies, rather than embracing the dogma that this was the way forward and not even questioning it for a decade.
A few years ago, the AAP dropped their recommendations about restricting certain foods, saying that there didn’t seem to be any evidence for them. There’s talk already of revised guidelines that would recommend introducing peanut before 11 months.
I’ll note that there are, in fact, peanut-allergic children in Israel; the rates of peanut allergy were increasing rapidly before the recommendation to restrict peanut intake entered the conventional wisdom; I had a peanut-allergic classmate in the 1980s, there are no guarantees. But oh my god, these recommendations, which were apparently based on gut feelings and conventional wisdom – these recommendations caused allergies, and when peanut allergies kept right on going up, it took twenty years to actually produce the research saying, “oh, wait. Oops.”
The reporting on the study focused on the effects of withholding peanuts from non-allergic children, but it’s the slightly sensitive kids I keep thinking about. It’s clear from the paper that in Hide’s study, when a child had a positive allergy test but no actual reaction, the parents were told to stop giving the kid peanut products. They mention one child, “from a vegetarian family,” with a strong reaction on the skin patch test who continued eating peanuts with no ill effects; they do not actually say “because she had terrible parents,” but the disapproval radiates through the page. It’s got to be scary, as a parent, to have a child with a positive allergy test who could start reacting to peanuts at any moment, but continuing to give them peanuts (while keeping antihistamines close to hand) turns out to be exactly the right thing to do.
Follow up note: I found Bamba at a grocery store near my house that has a large Kosher section and bought a small package to try them. They taste like Styrofoam packing material that has been made out of wood shavings and the dust left behind in a package of airline peanuts after you eat them. They would make an easily pincer-grasped, gummable baby snack but why would any adult eat these? Do 95% of Israelis really buy these regularly? Why?
Is this how the rest of the world feels about Cheetos? Weird.
[Angry peanut via Shutterstock]
showing a 1 mm to 4 mm wheal from the skin exposure ↩
Mothers were told not to consume milk and other dairy products, eggs, fish, peanuts, or soy. ↩
Blood tests are known to identify a lot of allergies that don’t show up clinically, which is to say, the test might say you’re allergic, but you don’t actually react to the food. Skin tests can also have false positives, but blood tests even more so. ↩