Thread by @realscientists

I meant to tweet more actively this week, but my three-year-old's fever derailed these plans. Luckily, she's better now. This came at the same time as me starting an elimination diet to find out what's problematic for my breastfed baby.

Let's have a look at infant allergies.

This may come as a surprise to most non-parents, but food allergies are surprisingly common in infants. The estimates vary a lot, but fluctuate around 5%: one in twenty infants having a food allergy. You likely know some.

But it's hard to know for sure for reasons we'll look at.

Disclaimer: unlike evolutionary biology and astrobiology, it's not my field of study. But I've raised one child with allergies already, it looks like the younger one may have them too, and I'm planning to tackle the topic as a science writer.

When my older daughter had been three weeks, she developed a peculiar rash and started having more mucosy stool. The pediatrician raised the possibility of cow's milk allergy.
I was surprised! But I stopped eating anything containing milk, and the nursed baby's symptoms improved.

This is called elimination. To make sure you react to some foodstuff, it needs to be followed by exposition - ingesting it again. That can be done at home, or, for severe symptoms, in the hospital under supervision.

This is the most common method of food allergy diagnosis.

With my daughter, we waited with exposition until she was almost a year old, since in the meantime, she likely reacted to other foods with mucus and bits of blood in the stool - egg and soya. These, after cow's milk, are among the most common allergens.

Now you see the milk allergy was an educated guess by the doctor - it could have been something else, but elimination and then exposition showed that allergy to cow's milk was the likely culprit.

It's not a very exact approach, but often the best that can be done.

You can perform a blood test to see if you have antibodies against proteins in some food. However, that can't reveal all types of allergies. It works best for the so-called IgE-mediated allergies... which are also easiest to notice without blood tests. Why?

IgE-mediated reactions occur fast, usually within minutes. They lead to fast release of histamine and you may develop a rash near your mouth, your tongue may burn or swell, you may experience anaphylaxis. It's typically not hard to pinpoint the culprit and then avoid it.

In contrast, nonIgE allergies are mediated by other types of antibodies or cell immunity - it's not a single type of allergy - and the onset is delayed by hours or even days. They usually manifest by diarrhea and other intestinal problems, eczema or both.

Here we get to the diagnosis problem and why we don't know the allergies' prevalence very well.

With nonIgE allergy, tracing the culprit may be hard enough for YOU, an adult well aware of their body's reactions. Try it with a baby.

Why was she so fussy?
What caused this rash?
How much mucus is still normal and how much a sign of pathology?
Is that a bit of blood or not?

It sure feels like divination from tea leaves, except it's from the diaper contents.

Try to do exact science with this 🤷. Another complication is that the potential allergens - fragments of protein from digested food - stay in the mother's body for some time and permeate into breast milk gradually. The symptoms may show after a few hours, or even three days...

Diagnosis is not easy and relies a lot on the caregivers' perceptibility.

Reported prevalences in studies of infant food allergies can vary several times; even by one order of magnitude. The methods vary: some use parents' self-reporting, some more rigorous exposition tests...

This cross-country self-reporting study found prevalences as low as 2% and as high as 12% - within a single study and the same method, only different countries!

Is the difference really so vast, and if so, what's behind it?

karger.com/Article/Abstra…

The real difference is likely smaller, but exists. It seems that a lot of factors might influence the onset of allergy. For instance newborn exposure to formula has been shown to increase the risk of cow's milk allergy... but wait a few months and you might see an opposite trend.

More on the timing of cow's milk protein introduction and the risk of allergy in this review: ncbi.nlm.nih.gov/pmc/articles/P…

The TL;DR is that much remains uncertain and we need more studies.

(I said diagnosing is hard. Correlating it with possible influencing factors is doubly so.)

That's why replication and meta-analysis are so important - AKA a brief science methodology digression! Replication means performing a study with the exact same methods as the original one to see whether you get consistent results. Meta-analyses pool data from more studies.

They allow you to see how much variance there is and whether effects reported in single studies are likely real or flukes, and pinpoint problems within the field.

For food allergy, they can reveal differences between countries, approaches, age groups: jacionline.org/article/S0091-…

The difficulty of studying allergies in small kids may unfortunately lead to some confusion among pediatricians and especially caregivers. You often hear the potential diagnosis, but not the guidelines of what to do. For cow's milk allergy, some mistakenly skip lactose.

Lactose is not the allergy culprit. It's a sugar present in all mammalian milk including human, and lactose intolerance (not an allergy; it's not an immune problem) in infants is extremely rare!

In contrast, reacting to proteins in cow's milk is common: ~2, maybe 5% of infants.

It would also help to see how much and how fragmented proteins from various foodstuffs enter breast milk. Such studies are still very scarce, despite their potential to help. For instance, can exposure to gluten through breast milk trigger celiac disease?

The answer is 'very likely not'. Three studies so far at the amounts of gliadin (the problematic part of the gluten protein complex) in breast milk and found it present, but in amounts thought unable to trigger the disease onset.

(Sorry for another digression from allergy.)

So we've established it's complicated to navigate for researchers, doctors and parents... Is there anything that can do done to *prevent* the development of food allergy to avoid all these problems?

It looks like not too early, but not delayed introduction of foodstuffs (= before approx. 6 months of age) may have some preventative effect, but again, more research is needed to work out the details.

nejm.org/doi/full/10.10…

jacionline.org/article/S0091-…

In any case, good luck to all parents and kids going through this! Luckily, infants' food allergies often fade out. My older one's did after approx. 2 years. We'll see how it goes with the little one!