|<< contents||Section 5.2||Section 5.4|
As others have found, some modification of diagnostic criteria for AE is mandatory for children2. Since early onset and inhalant allergy are likely to be less useful in the young child, early onset was omitted, and personal history of atopic disease was replaced by family history of atopic disease for children under 4. This scheme resulted in a sensitivity of 85% and specificity of 96% when applied to the 38 children age 4 years and under in the hospital paediatric validation study14.
The author is less confident about the validity of our criteria in the first year of life. Although most children at this age were correctly classified, numbers were small, and validation against physician's diagnosis is perhaps not so useful due to the widespread disagreement between experts of what constitutes a case of AE at this age21. One approach of describing AE in the first year of life is simply to focus on recording the prevalence of symptoms such as itchy rash or signs such as flexural dermatitis, analogous to similar conventions in asthma research. Further research using longitudinal designs might then delineate which features best discriminate those individuals who later develop typical AE, although it does not overcome the problem of saying that those who do clear did not originally have AE.
Based on our own data and other studies which have tried to separate seborrhoeic dermatitis of infancy from AE21, we suggest that in order to be classified as a case of AE, children under the age of 1 year should have a history of scratching or rubbing plus three or more of: history of involvement of outer arms/legs, family history of atopic disease in first degree relatives, history of a general dry skin, and visible dermatitis on the cheeks or outer arms/legs with absence of axillary involvement.
Although the effect of potential confounding by age on the usefulness of the six diagnostic criteria has been thoroughly in the development work by i) performing separate regression analyses in children and adults ii) looking for interaction between the criteria and a dummy variable for "age under 16" and iii) by separate analysis of adult and paediatric data in the hospital validation study, further testing of the performance of the criteria in adults in a community setting (eg a factory) is still desirable because of the low numbers of adults with AE in our study. Recall of onset of eczema under the age of 2 is likely to be inaccurate in adults, and this should be replaced "Did your eczema start when you were a child?". Others have expressed concerns about the inclusion of periorbital dermatitis under the sign of visible flexural dermatitis in adults given the higher relative prevalence of contact to atopic dermatitis at this site in adults22. It seems prudent to exclude periorbital dermatitis as an acceptable site for visible flexural dermatitis in adults until further data is available23.