Introducing solid foods, recommendations that turned twice

Around the turn of the 2000s, several professional bodies recommended delaying the introduction of foods considered allergenic, on a plausible line of reasoning: less early exposure, less sensitisation.

Those recommendations rested mainly on observational studies and on physiological reasoning, without any controlled trial having tested them directly. Nor were they uniform: age thresholds, the list of foods concerned and the definition of at-risk children varied by country and body, which made practice itself heterogeneous.

The British trial published in 2015 under the name LEAP randomly allocated high-risk infants between early introduction of peanut and avoidance until age five.

The trial, led by Gideon Lack's team, recruited infants aged four to eleven months with severe eczema, egg allergy, or both, the criteria used to define high risk. Its conclusions apply first and foremost to that population.

Results showed a markedly lower frequency of allergy in the early introduction group. Recommendations were revised in several countries in the following years, in the opposite direction to the previous shift.

A recommendation founded on coherent reasoning and observational data was contradicted by a trial. The configuration is classic in public health, and it is the main argument for trials where they are feasible.

It also illustrates the cost of delay: more than a decade passed between the adoption of a recommendation and its direct test, during which the advised practice ran the other way. Nothing rules out current recommendations evolving in turn in light of new trials.

The optimal starting age for solid foods is the subject of converging recommendations around four to six months, with different wordings across bodies and a moderate level of evidence.

Questions of sequence, texture and quantity remain little tested. Recommendations on those points rest largely on expert consensus, as official documents explicitly state.

Wordings do not always coincide from one body to another: the World Health Organization, for instance, recommends exclusive breastfeeding until six months, while other professional bodies state a range. These differences stem partly from different contexts of application.

Practices differ sharply between neighbouring countries, including where official recommendations are close. The first foods introduced, their order and their form vary with culinary traditions more than with data.

That variation is worth knowing: it indicates that many rules presented as health-based are in fact stable local conventions, which makes them neither dangerous nor necessary.