Asthma as an outcome: Exploring multiple definitions across birth cohorts in the children's respiratory and environmental workgroup
Visness C, Gebretsadik T, Jackson DJ, Gern JE, Biagini Myers J, Havstad S, Lemanske RF, Hartert TV, Khurana Hershey GK, Zoratti EM, Martin L, and Johnson CC. Asthma as an outcome: Exploring multiple definitions across birth cohorts in the children's respiratory and environmental workgroup. Am J Respir Crit Care Med 2018; 197:A4594.
Am J Respir Crit Care Med
Rationale: A 2010 review of asthma birth cohort studies found 60 different definitions for the outcome of asthma among 122 studies. Reports regarding the risk of developing asthma may be sensitive to the outcome definition, which can result in inconsistent findings across birth cohorts. Methods: We use data from 5 longitudinal US birth cohorts that are part of the Children's Respiratory and Environmental Workgroup (CREW) consortium within the ECHO program to examine how the prevalence of asthma in middle childhood (age 6-10 years) varies by applying individual cohort outcome definitions across the other cohorts, where possible. Two cohorts (Wayne County Health Environment Allergy and Asthma Longitudinal Study [WHEALS], Childhood Allergy/Asthma Study [CAS]) were recruited from the general population and determined asthma status by parental report of physician diagnosis. The other three cohorts (Childhood Origins of Asthma Study [COAST], Urban Environment and Childhood Asthma [URECA], Cincinnati Childhood Allergy and Air Pollution Study [CCAAPS]) recruited children with an elevated risk for developing asthma based on a family history in at least one parent. CCAAPS verified asthma status using bronchodilator reversibility and methacholine challenge. COAST and URECA used a combination of reported diagnosis, symptoms, and medication use to determine asthma status, but followed different algorithms. Results: Asthma prevalence using the original cohort definitions ranged from 10 to 28%, and was strongly influenced by whether the cohort population was selected to be at high risk for asthma development or not (23% vs. 16%, p<0.001). Using doctor diagnosis of asthma across 4 cohorts resulted in an overall prevalence of 18% (range 10-33%, see Figure). The modified COAST definition resulted in an overall asthma prevalence of 21% (range 10-37%). The modified CCAAPS definition could be applied to 3 cohorts and resulted in an overall prevalence of 24% (range 11-39%). Unexpectedly, asthma prevalence in COAST was reduced from 23% to 11% following the CCAAPS criteria, whereas prevalence in URECA increased from 28% to 39%. Conclusions: Harmonization of outcome definitions across birth cohorts is challenging, as not all cohorts collected the requisite data for all definitions and procedures for data collection vary. Findings using pulmonary function measures are likely influenced by protocols for when and how measurements are made. A common definition using available data is being developed and will be applied across ECHO cohorts in order standardize questions and procedures for determining the asthma outcome.