Parents ask me this in almost every ADHD consultation, usually after reading that a diet change transformed someone's child. The research on elimination diets in ADHD is real, it is decades old, and it says something more specific than either "food doesn't matter" or "it's all about diet".
What it says, roughly: a minority of children — likely a small one — have attention and behaviour that genuinely responds to removing particular foods, and there is no reliable way to identify them in advance except by carefully trying. The cost of trying is not zero, particularly for a child.
What has actually been studied
Artificial food colours and preservatives. The most robust finding in the field. Several controlled trials, including a well-known UK study using mixtures of artificial colourings and sodium benzoate, found small but measurable increases in hyperactive behaviour in children — not only children with ADHD. The effect sizes were modest, and the response was not uniform: some children reacted, most did not. This work contributed to European labelling requirements for certain colourings.
Few-foods / oligoantigenic diets. The most aggressive approach: reduce the diet to a small set of unlikely-to-react foods (typically lamb, rice, pear, certain vegetables) for two to five weeks, then reintroduce foods one at a time. Trials of this method have reported meaningful behavioural improvement in a substantial proportion of the children who completed them — but the studies were small, mostly unblinded during the reintroduction phase, and conducted with intensive dietetic support. Meta-analyses that weight for blinding find real but smaller effects.
Sugar. The most persistent belief with the least support. Controlled challenge studies have repeatedly failed to show that sugar causes hyperactivity, including studies where parents were told their child had received sugar when they had not — parental ratings went up, measured behaviour did not. Sugar still matters for dental and metabolic health, and a sugary breakfast with no protein sets up a rough morning through blood glucose swings rather than through hyperactivity.
Omega-3 fatty acids. Not an elimination approach but frequently discussed alongside it. Meta-analyses find a small benefit for ADHD symptoms, considerably smaller than stimulant medication. Reasonable as an adjunct in some children; not a substitute.
Food sensitivity testing: what to avoid
This is where families lose the most money. IgG food antibody panels — sold directly to consumers and by some practitioners — are not validated for diagnosing food sensitivity. IgG antibodies to foods reflect exposure, not intolerance, and the major allergy societies advise against using these panels to guide dietary restriction. A panel will typically return a long list of "reactive" foods, and acting on it can restrict a child's diet substantially for no benefit.
What is worth testing: coeliac serology if there are gastrointestinal symptoms, poor growth or a family history, and true IgE allergy testing where there is a history suggesting genuine allergy. Both answer specific questions.
If you want to try an elimination trial properly
Done badly, an elimination diet is weeks of family conflict with an uninterpretable result. Done properly it is a structured experiment:
- Establish a baseline first. Two weeks of a standardised rating scale completed by both a parent and a teacher, before changing anything. Without a baseline you are comparing against memory, which is unreliable and biased toward wanting it to work.
- Start narrow, not broad. Removing artificial colours, artificial flavours and benzoate preservatives is a low-cost, low-risk first step with the best evidence behind it. It mostly means reading labels rather than removing food groups.
- Give it three to four weeks, then re-score with the same scales and the same raters.
- Reintroduce deliberately. If there was improvement, add back one item at a time for several days each and watch. Improvement that survives reintroduction of everything was probably not the diet.
- Involve a dietitian for anything wider. A few-foods diet in a growing child risks inadequate calcium, iron, protein and calories, and it can entrench a fear-based relationship with food. It should not be run without professional supervision, and it should not be run indefinitely.
What I weigh before recommending it
Two considerations that rarely make it into the articles:
Restriction has psychological costs for children. Being the child who cannot eat what everyone else eats at a birthday party is not neutral. In children who already have social difficulties — common in ADHD — an unnecessary restriction can cost more than it returns.
Delaying effective treatment has costs too. Stimulant medication has among the largest effect sizes in child psychiatry. A family that spends eight months on sequential elimination trials while a child struggles academically and socially has paid for that time. Running a dietary trial alongside proper treatment is reasonable; running it instead of treatment usually is not.
The dietary factors that matter more, for more children
Less exciting, better supported:
- Protein at breakfast and regular meals, which steadies the morning and reduces the mid-afternoon crash. This matters more once a stimulant is on board, since appetite suppression is a common side effect.
- Iron status. Low ferritin is associated with more severe ADHD symptoms and with restless sleep. It is measurable and correctable, and worth checking.
- Sleep. Sleep-deprived children look inattentive and impulsive. Nothing dietary competes with fixing a bedtime.
- Adequate total intake in medicated children, where weight and growth need monitoring.
In our child and adolescent assessments we take a full dietary and sleep history alongside the ADHD evaluation, and where the picture warrants it we check iron and coeliac status. Where objective measurement helps distinguish attention difficulties from other explanations, computerised ADHD testing gives a baseline that is not subject to the same expectation bias as parent report — which is genuinely useful when you are also trialling a dietary change.
If you want to try one, do it with a baseline, a narrow first step and a defined endpoint. For families interested in the broader nutritional picture, that is what a nutrition consultation is for.
FAQs
Do artificial food colours make ADHD worse?
Controlled trials have found small increases in hyperactive behaviour with mixtures of artificial colourings and sodium benzoate, in children generally rather than only those with ADHD. The effect is modest and not uniform — some children react and most do not. Removing them is a low-risk first step to trial.
Does sugar cause hyperactivity?
Controlled challenge studies have repeatedly failed to show it, including studies where parents wrongly believed their child had been given sugar and rated behaviour as worse anyway. Sugar still matters for dental and metabolic health, and a protein-free sugary breakfast can make mornings harder through blood glucose swings.
Are IgG food sensitivity tests useful for ADHD?
No. IgG food panels are not validated for diagnosing food sensitivity — the antibodies reflect exposure rather than intolerance — and allergy societies advise against using them to guide dietary restriction. Coeliac serology and true IgE allergy testing answer specific clinical questions and are appropriate when indicated.
How long does an elimination diet take to show results?
Three to four weeks on the elimination phase, with rating scales completed before and after by the same people. Then foods are reintroduced one at a time over several days each. Anything shorter than that cannot distinguish a real effect from a good week.
Can diet replace ADHD medication?
For most children, no. Stimulant medication has one of the largest effect sizes in child psychiatry, and dietary interventions — including omega-3 supplementation — show considerably smaller effects. Dietary work is a reasonable addition alongside treatment; substituting it usually costs the child time.
