Methodological difficulties mean that interpretation of these trials is difficult. RCTs that employ a parallel design with an unblinded normal control diet may introduce bias in favour of the active group. In addition, those trials that place all participants on exclusion diets and then re-introduce the suspected offending food compared with a control, risk introducing another allergen (for example soya) or introducing the suspected allergen (for example cows' milk) in a way that does not mimic real life. Poor concealment of randomisation allocation, lack of blinding and high dropout rates without an ITT analysis all mean that the above studies should be interpreted with great caution. Future studies should ideally be longer term, more pragmatic and ensure that randomisation is concealed.
Uncontrolled elimination diets followed by double-blind, placebo-controlled challenges with foods suspected to aggravate symptoms have also been tried in atopic eczema.17-21 These studies are not the same as RCTs of food elimination. Instead they try to answer the question: "Does food X make a particular child's atopic eczema worse?". The precise relationship between such food challenge studies and long-term benefits of exclusion of the suspected foods to atopic eczema sufferers is not clear. Blood and skin prick tests are usually only helpful in predicting clinical response if they are negative.22,23 It should also be borne in mind that this high negative predictive value has only been shown in relation to provocation of symptoms after double-blind challenge, and not clinical response following food elimination, which are not necessarily the same thing. The relationship between atopic eczema and food sensitivity is a complex one and readers are referred to a clear evidence-based work by David for further information.24
Was this article helpful?