|
NON-TECHNICAL SUMMARY
BACKGROUND
Whether dietary factors can affect children’s behaviour remains a controversial issue. Health professionals and researchers have generally downplayed the idea that certain foods or food additives may contribute to child behaviour problems. By contrast, many parents (and some teachers) are convinced that there is a link.
In the 1970s the US paediatrician Ben Feingold suggested that ADHD symptoms in many children could be reduced by eliminating from the diet various artificial food colourings, naturally occurring salicylates (substances found naturally in many fruits and vegetables) and some preservatives. A number of controlled trials – mainly published in the 1970s and 1980s - appeared to support this hypothesis, but the evidence from these was mixed. The conclusions of several qualitative reviews were also mixed, and served only to polarise debate.
Only one statistical summary (or ‘meta-analysis’) of the different trials has previously been published, and this was over 20 years ago. It concluded that the ‘Feingold Diet’ had only a small and non-significant effect on ADHD symptoms. According to the researchers who carried out the current review, an updated meta-analysis was necessary for several reasons. - The earlier statistical summary included an overly broad range of studies (involving different dietary interventions), which could have reduced the observed association between diet and ADHD symptoms.
- It failed to include some relevant studies that were available at the time
- Several more good quality trials have been published since then.
METHODS
For this review, the authors first searched systematically for all published studies reporting the results of placebo-controlled, double blind trials assessing the effects of artificial food colourings (AFCs) on behaviour in children with ADHD. They found 15 trials involving a total of 219 participants.
All of these were double-blind ‘cross over’ trials, which means that children received two different diets in turn, one that included AFCs and one that did not. While they were following each diet, the children's behaviour was rated by parents, and/or teachers, and/or clinicians. In all cases, neither the children nor the people providing the ratings knew which diet the child was receiving at the time. (This ‘double-blind’ procedure is essential to rule out the effects of expectations or pre-existing biases; and trials of this kind are commonly used to assess the effects of medical treatments.)
The researchers then ‘pooled’ the findings across all the different studies in a ‘meta-analysis’ - a special type of analysis designed to examine the overall pattern of results. Although it has some drawbacks, this approach is considered the most reliable basis for drawing conclusions about the balance of evidence in a particular research area.
RESULTS
The primary result from the meta-analysis was that children's behaviour did improve significantly when AFCs were eliminated from their diet. The average improvement from eliminating AFCs from the diet was around one third to one half of the size of the improvement typically associated with medication treatment for ADHD.
This study also considered some secondary questions.
- Did the ‘quality’ of the studies affect the results obtained?
Yes. When the two least rigorous studies were excluded, the significance of the results was reduced, but it still remained significant.
- Did eliminating AFCs lead to greater behavioural improvements in children who were previously thought to be responsive to AFCs?
Yes. If children had responded to AFCs in a non-blinded (open) trial, or if their parents believed they were affected by diet, then they showed more substantial improvements when AFCs were removed during the double-blind study.
- Were the effects noticed equally by parents, teachers and health professionals?
No. Significant results were found from all the different types of rating combined, but when each of these groups were considered separately, the impact on children's behaviour of eliminating AFCs was only significant for the parent ratings.
- Were the effects stronger for children who had a rigorous official diagnosis of hyperactivity / ADHD?
No. Although the researchers thought that this might be the case (and some previous studies have suggested that ‘hyperactive’ and ‘normal’ children might react differently to chemical additives), the results did not support this hypothesis.
DISCUSSION
A systematic review of the evidence from controlled trials has shown that the behaviour of hyperactive children does improve when artificial food colourings (AFCs) are eliminated from their diet. These results therefore broadly support the theories first put forward by Feingold in the 1970s (although the full ‘Feingold Diet’ also involved other dietary changes). The mechanism by which AFCs may adversely affect children's behavior is still not known, although the authors note that it could reflect either allergic reactions, or actual pharmacological effects that AFCs induce.
On average, the removal of AFCs alone yielded only modest benefits for these children (around one third to one half of the improvement that would be expected from conventional medication), suggesting that dietary changes may need to be supplemented by other forms of intervention. The current treatments of choice for ADHD are medication treatment and/or behaviour therapy, but the findings from this study support the idea that dietary treatments are also worthy of consideration.
The authors also point out that the current results could underestimate the actual effect of AFCs. Several of the trials analysed used doses that were well below children's true likely daily exposure; and because the impact of AFCs on behaviour and learning may occur within several hours of ingestion, some studies may have allowed too much time between the consumption of AFCs and the measurement of outcomes to fully capture their possible effects. In addition, it is possible that in some children, greater behavioural changes would result from eliminating other items as well (e.g., artificial flavors and preservatives), but for consistency, the studies included in this meta-analysis only examined the impact of eliminating AFCs from children's diet.
Individual differences are clearly a key issue in the debate surrounding ADHD and diet. The evidence reviewed here showed that negative effects on behaviour following consumption of AFCs were greatest in children previously identified as ‘responders’, either by non-blind testing or by parents. This indicates that parents’ observations and beliefs can be helpful in identifying children who may respond well to dietary interventions.
Parent ratings were more sensitive to the effects of AFCs than teacher or clinician ratings. By contrast, more significant results from teacher ratings are often found in studies of medication treatment in ADHD (although this may also reflect issues of dosage and timing of such medications). Differences between parent and teacher ratings are well known in ADHD research, and can arise for various reasons. Parents and teachers see children in different settings, at different times, and have different relationships with children as well as different expectations from them. The authors of this study did note that some of the apparent effects of artificial food colourings (such as sleep problems and irritability) might be expected to have more impact on parents than teachers.
It is important to note that not all of the trials reviewed here used ratings that specifically related to the ‘core’ ADHD symptoms of restlessness and inattention. Some focused on specific symptoms thought to be triggered by AFCs such as sleeplessness and irritability, while others used scales customized for each child based on parents' reports of how they thought diet affected their child. These were all blinded trials, so the use of these alternative outcome measures does not invalidate the finding that parents, on average, saw an improvement in their children's behavior when AFCs were eliminated. It does, however, leave open the question of how specific the effects of AFCs may be in relation to the core diagnostic symptoms of ADHD.
In this respect (and re the greater sensitivity of parent ratings), the results of this systematic review are consistent with those of Bateman et al, 2004 who studied behavioural effects of similar food additives in a large group of 3-year-old children. The negative effects that they found (on parent ratings) from a mixture of AFCs and one common preservative were not confined to subgroups with ADHD, allergies, both or neither, but were significant across the whole of the study sample.
The authors of this review are understandably cautious about the clinical implications of their findings. They note that the restrictive nature of AFC free diets may place a burden on children and families, and suggest that "imposition of the diet should be done reluctantly" until more certain methods have been developed to identify AFC-responsive children. They do, however, emphasise the need for additional research in this area.
In conclusion, this systematic review provides strong evidence that the behaviour of children with ADHD can be made worse by dietary factors, and that eliminating AFCs from their diets will, on average, result in behavioural improvements. The authors note that these findings are consistent with accumulating evidence that neurobehavioral toxicity may result from a wide variety of distributed chemicals.
|
