When the edge of certainty is close enough

Compelling evidence of a causal link between food coloring and the aggravation of attention deficit hyperactivity disorder in certain children has prompted a UK behavioral psychologist to invoke "the precautionary principle” and call for a ban.

Jim Stevenson, professor emeritus in Southampton University's Brain Behaviour Laboratory, likens the state of knowledge of the effects of E-number coloring to what was known about lead and children’s IQ in the 1980s.

In that case, a sequence of statistical adjustments to take into account "confounding social differences" reduced the effect size to 0.17 – a figure "at the margins of certainty".

Sir Michael Rutter announced in 1983: “A marked reduction in the level of environmental lead is likely to make an important difference to some children. Moreover it is important to recognize that a small change in mean IQ or average behavior of the population as a whole will have a much greater effect at the extremes of the distribution […]. Accordingly, actions to cut down the amount of lead pollution of the environment should be worthwhile; there is sufficient justification for action now”.

Action duly followed, Stevenson writes, and it turned out to be a prudent use of the precautionary principle. The current evidence on food colors was comparable and the effect size data similar, but hedged health warnings were the extent of the response.

He reviews the argument in the latest edition of the Journal of Children's Services.

The first to raise concerns, he writes, was the US pediatric allergist Ben Feingold, who suggested in Why Your Child is Hyperactive? that many children with symptoms of attention deficit hyperactivity disorders would benefit from being put on a diet free of naturally occurring aspirin-like salicylates and azo dye food colors.

Research over the next two decades tended to confirm his theory but not consistently across the general population nor with sufficient statistical force to warrant any outright ban.

Just this year, the UK National Institute for Health and Clinical Excellence (NICE) ruled that “the elimination of artificial coloring and additives from the diet is not recommended as a generally applicable treatment for children and young people with ADHD”.

NICE recommended case by case appraisal. If it appeared that certain food and drink influenced the behavior of particular children, they should keep a record and seek a referral to a dietitian.

Jim Stevenson's case for a ban focuses on the distinction in public health policy between the onus on an "at risk" individual to take proper care to avoid exposure, and the duty on government to deal with known risks to the general population.

The Southampton University Research Group's studies were designed to investigate that relationship.

The first, on the Isle of Wight, used a double blind placebo-controlled design with 277 three-year-olds.

Based on parent ratings of behavior, it found that the levels of hyperactivity were significantly higher among those on the additive cocktail than on placebo – but the effect could not be detected on the behavior measures obtained in the clinic .

The mixed-age Southampton investigation was broadly similar. It too found elevated levels of hyperactivity among those ingesting the additives – but the results did not reach the 5% level of significance in all analyses.

The UK Committee on Toxicity concluded that the work provided supporting evidence to suggest that certain mixtures of artificial food colors together with the preservative sodium benzoate were associated with an increase in hyperactivity in children from the general population. “If causal, this observation may be of significance for some individual children across the range of hyperactive behaviors, but could be of more relevance for children towards the more hyperactive end of the scales.”

On the other hand, the European Food Standards Authority panel described the effects in the Southampton study as small and their significance for children's development and education uncertain – to which the research team rejoined that since the colors being tested were of no nutritional value, removing them from children’s diets would do no harm and some good.

Out of which exchange came a ruling that manufacturers should label foods containing the six colours with the warning: “may have an adverse effect on activity and attention in children”.

“In appraising what action is appropriate,”Stevenson writes, “there is a need to consider hazard, exposure and risk. In terms of exposure, these food additives are widely present in foods ingested by children, including confectionery, cakes, biscuits and soft drinks.

“If the hazard is low but the exposure is high, what does that mean for risk? In formulating policy on public health harm, consideration has to be given to the putative hazard, in other words the degree of exposure, and the risk to future well-being likely to arise from such an exposure.

“Therefore, in terms of a public health formulation of harm, there is evidence that food colors produce a small hazard with high exposure that presents an educationally significant long-term risk to the child.”

Failure to impose a mandatory ban on the six food colors in the Southampton study (sunset yellow FCF (E110), quinoline yellow (E104), carmoisine (E122), allura red (E129), tartrazine (E102) and ponceau 4R (E124) was inadequate.

See: Stevenson J "Food additives and children’s behaviour: evidence-based policy at the margins of certainty” Journal of Children’s Services, 4, 2 or visit the the publishers’ website [1]

Reference: Rutter, M. (1983) “Low level lead exposure: sources, effects and implications”. In Rutter M and Russell Jones R (eds.) Lead versus Health: Sources and effects of low lead exposure. Chichester, John Wiley.