doi: 10.1111/cea.12127 Clinical & Experimental Allergy, 43, 583–585 EDITORIAL © 2013 John Wiley & Sons Ltd Clinical & Experimental Allergy The hope in redefining atopy This editorial discusses the findings of the paper in this issue by F. L. Garden et al. [1], pp. 633–641. A. J. Lowe 1,2 , S. Zaloumis 1 and S. C. Dharmage 1,2 1 Melbourne School of Population and Global Health, The University of Melbourne, Melbourne, Australia and 2 Murdoch Childrens Research Institute, Melbourne, Australia There is increasing interest in refining allergic pheno- types including asthma. This is driven by the belief that the reason for inconsistent evidence on the aetiology of these conditions is related to lack of clarity in outcome definitions. The study by Garden et al. [1] published in this issue of the Journal is another important step in this process. The epidemic of allergies and asthma in westernized countries over recent decades has been well docu- mented. The rapid increase in the prevalence of these conditions implies that environmental factors are important. However, it has proven remarkably difficult to identify environmental risk factors that are consis- tently related to increased risk of allergic disease. This begs the question ‘if environmental risk factors are important in the aetiology of asthma and allergic disease, why has it proven so difficult to identify con- sistent associations between these conditions and modi- fiable risk factors?’ Identification of such associations is important, to both to understand the aetiology of these conditions, as well as to help lead to effective preven- tive strategies, which is currently sadly lacking. One of the possible reasons that we are not seeing any clear aetiological risk factors for allergic disease is that we are not defining the outcome(s) with sufficient acuity. There has been increasing recognition that what is labelled as ‘asthma’ is really a symptom that has multiple aetiologies [2]. This has lead to calls to better group asthma into specific ‘phenotypes’, based on a range of presenting features. If each phenotype has different underlying patho-physiological processes, it is likely that they also have different underlying risk factors. By grouping multiple types of ‘asthmas’ together, we may be making it impossible to see clear risk factor profiles as the risk factors for each phenotype are distinct. If such a situation is correct, the risk factors that are identified will be those that are most strongly related to the most com- mon ‘asthma’ or to a number of different ‘asthmas’. Research that generates specific ‘asthma’ phenotypes offers the possibility that specific risk factors for each phenotype can be identified and lead to potential preven- tative strategies that target these specific phenotypes. There is already evidence that specific phenotyping of asthma can advance the knowledge in this area. A number of attempts have been made to define wheeze according to age of onset and persistence of wheeze, following on from the seminal work based on the Tuc- son cohort [3]. The risk factors for these wheeze pheno- types seem to be distinct from each other, at least in terms of some genetic polymorphisms [4], even if not for a range of other potential factors [5]. Some groups have proposed separating childhood asthma into multi- trigger vs. virus induced wheeze [6]. We have shown that aetiology of childhood asthma associated with hay- fever seems to differ from childhood asthma that is not associated with hayfever [7]. Similarly, the risk factors for adult atopic asthma seem to differ from that of non-atopic asthma [8]. Recently there has been a move to use data driven (a-theoretical) means of identifying clusters of wheeze outcomes. The benefit of such techniques is that it allows for the natural clustering within the data to be identified, without any preconceptions as to what groups ‘should be’ there. One of the leading forms of these has been latent class analysis (LCA), which may be used where the observed data are consistent with there being a small number of underlying, or latent, classes of the outcome at the population level. LCA has lead to the identification of new wheeze groups includ- ing ‘intermediate onset wheeze’ [9]. Correspondence: Adrian John Lowe, Melbourne School of Population and Global Health, Level 3, 207, Bouverie Street, The University of Melbourne, Melbourne, Vic. 3010, Australia. E-mail: lowe.adrian@gmail.com Cite this as: A. J. Lowe, S. Zaloumis, S. C. Dharmage, Clinical & Experimental Allergy, 2013 (43) 583–585.