Wheezing phenotypes in childhood.

There can no longer be any doubt that, within the spectrum On this dubious basis, a third phenotype with a peak prevalence at six years was identified between the group of of wheezing disorders of childhood, several distinct phenoearly transient wheezers and later atopic asthmatic subjects. types can be recognised. Although clinicians have been The distinction between children who wheezed in the first aware of this for decades, it is only through the painstaking three years and those who only wheezed in their sixth year recording and analysis of population cohorts over many is dependent simply on the time points which were chosen years that convincing evidence has emerged. by the Tucson group. Bearing in mind the continual switchIt is worth reminding ourselves here of three important ing which occurs between the wheezing and non-wheezing issues in long term studies of the natural history of disease. sets in a population over the years, a full breakdown of Firstly, studies which set out to test explicit hypotheses are the groups would be needed to make a judgement on the especially important. Although valuable information can numbers of phenotypes. Our own estimate from tables 3 be gleaned from massive information gathering projects, and 4 shows that atopy (by skin prick tests at the age of this is often more by luck than design. Secondly, a long 11) was about 60% in currently non-wheezing children at term approach to organisation and funding is needed to all ages, while the proportion of atopic children in the ensure the best gains from early investment. We detect a wheezing groups rose steadily from 71% at three years of reluctance to plan far ahead in the UK, perhaps driven age to 76% of those wheezing at six years and to 90% at by the four yearly research assessment exercises in UK age 11. This suggests a steady enrichment of atopic children universities and by project grants of 2–3 years duration. within (or a loss of non-atopic children from) the declining Thirdly, data must be stored in an accessible format. wheezing population. A further phenotype perhaps? Written records are bulky but they are durable and reRather than force each child into a particular phenotype, trievable over long periods. With advances in technology, is it not more useful and logical to consider the risk factors will electronic archives be equally accessible in 60 years which may be operating over different time periods in time, or will the decipherment of Linear B be re-enacted the population, and to which individual children may be each time we try to analyse old data sets? variously susceptible? 11 Figure 1 (adapted from WenPapers in this issue of Thorax from two of the most nergren and Wilson) is an attempt to illustrate this point. influential recent cohort study groups – from Aberdeen, The implications are that clinical phenotypes are not static Scotland and Tucson, USA – address the subject of so that transient viral wheeze can, for instance, occur in childhood wheezing phenotypes, their classification and subsequently atopic wheezers. The cohort data can then be heredity. 3 While neither provides unambiguous results, used to examine questions relating to interactions between both have important messages, methodological lessons, these risk factors. and interesting data. It is important to subject them to The latest instalment concerning the highly informative public scrutiny and debate. What can we make of them? Aberdeen cohort 13 14 considers their offspring 30 years The Tucson cohort is younger but has the advantage of after recruitment. Attrition, small and highly selected detailed information collected during early childhood – an study groups, and statistically marginal outcomes present age when many formative events occur, when the rate of problems. For example, the conclusions that prepubertal developmental change is at its greatest, and when wheezing male non-atopic offspring of probands with a childhood phenotypes are changing. The original Aberdeen cohort history of wheezy bronchitis have smaller spirometric values was recruited 30 years ago from children aged 10–14 years, than controls, and that these children are less likely to the age at which the (current) Tucson data end. It is suffer current (but not past) wheeze than the children tempting – despite a generation time gap and 5000 miles of non-atopic controls, is based on tiny numbers and – to treat them as providing a continuous account. This could be rash. The report by Stein and colleagues from Tucson deals with clinical features of the 60% of the original birth cohort still living in the area at age 11, complementing data collected at birth, three years and six years. These data points arbitrarily divide childhood into age periods and we need to remind ourselves that changes occur gradually during childhood, not at fixed time points. A comprehensive set of clinical and physiological data was collected. The population may be atypical in that, at 11, 25% had a current history of wheezing and 60% were atopic on skin prick testing. Methacholine responsiveness, positive skin tests, and male sex were strongly linked with current wheeze at 11, but not with wheeze at younger ages. Again, this confirms previous observations in high risk populations. In contrast to other studies which found PEF Birth