How to Introduce Soy to Your Baby: Tofu, Edamame and Yoghurt

Soy is on every major labelling list. It is one of the nine major allergens in the US, it is on the EU and UK lists of 14, and it is on the Australian and New Zealand list as well. We lay the lists out next to each other in our comparison of the four regions, and soy appears in every column.

For us that settled whether it belonged in the introduction plan. The NHS page on food allergies in babies names soya among the foods that can trigger allergic reactions, says they "can be introduced from around 6 months as part of your baby's diet", and advises introducing them "one at a time so that you can spot any reaction". It also says that once introduced and tolerated, they should become part of the baby's usual diet. That second half is the bit that is easy to lose track of with a food you do not cook often.

Tofu was the first form we used, because it asks almost nothing of the cook. We bought plain firm tofu, cut off a small piece, and pressed it through a fork into a vegetable purée our baby already ate without any fuss. Silken tofu works too and is softer still. There was no marinade and no soy sauce, since soy sauce is salty and not a baby food.

The first taste was about a quarter of a teaspoon, in the morning, on a day with nothing else new on the menu. A morning start meant any reaction would turn up while we were both home and awake. Later in the fortnight we moved to plain, unsweetened soy yoghurt, stirred into porridge or eaten off a preloaded spoon. We read the labels on those tubs, because some flavoured ones carry a surprising amount of sugar.

Edamame looks like an ideal finger food and it is not, at least not whole. A single bean is round and firm, which is the shape choking-hazard lists single out. The CDC's list of choking hazards for young children includes whole beans under proteins, and notes that "cutting food into smaller pieces and mashing foods can help prevent choking."

So we only ever served edamame shelled, boiled until very soft, and then squashed. For the earlier days that meant blending it into a smooth green purée with a little water. Once our baby was managing lumpier food we pressed each bean flat with the back of a spoon so it could not be swallowed as a round piece. We never put the pods anywhere near the high chair. There is more on this in our post on choking hazards, which covers other round foods like grapes that need the same treatment.

Day 1 was a quarter teaspoon of mashed tofu. Day 3 was a full teaspoon, again in purée. Day 7 was a tablespoon of tofu, plus a first try of plain soy yoghurt. By day 14, soy was just part of meals: tofu in a mash, yoghurt with fruit, squashed edamame on the tray.

We kept every other new food out of those two weeks. That is the point of the observation window: if a rash or tummy upset had appeared on day nine, we wanted a short list of suspects, ideally one. We noted the time of each serving and anything unusual afterwards, even when the answer was "nothing".

The same rhythm had already worked for dairy, which mattered, because soy and cow's milk get mentioned together so often. They are different proteins, and we treated them as separate introductions with their own fortnight each.

Often, though not always, and on a slower timeline than parents might hope. The study most often cited on this is Savage and colleagues (2010), which followed 133 children with soy allergy seen at a specialist allergy clinic at Johns Hopkins. Their analysis predicted that the allergy would have resolved in 25% of children by age 4, 45% by age 6 and 69% by age 10. The authors summarised it as roughly half of the children outgrowing soy allergy by age 7. Children with lower soy IgE levels on testing were more likely to outgrow it.

Two things kept us from reading too much into that. The children were a referral group from a specialist clinic, which the authors point out themselves, so they were probably more allergic than the average soy-allergic baby. And a statistic about a clinic population says nothing about any single child. Whether and when a soy-allergic child tries soy again is something an allergist works out with testing, not something a family schedules at home.

Soy-based infant formula came up a few times in our parents' group, usually as a suggestion for a baby with a suspected milk allergy. It is a different question from putting tofu in a purée, and the NHS guidance on types of formula is direct about it: "Only use soya formula if it has been recommended or prescribed by a health visitor or GP." The same page says soya formula is suitable from 6 months, but only under medical supervision, and explains the concern: a baby's low body weight means they take in much higher amounts of phytoestrogens from soya than older children or adults.

So switching formula was never something we would have done on a hunch. A baby who has reacted to cow's milk formula also needs a proper assessment before anything changes, and cow's milk protein allergy symptoms is where we started reading before that appointment.