Background Adult onset asthma is burdensome to both the individual and state, yet compared to childhood asthma little is known about its epidemiology. This paper examines (1) the extent to which exposure to stressors at different life stages predicts adult onset asthma, and (2) what factors mediate any associations found.
Methods Logistic regression analyses of asthma onset during mid-adulthood on different sets of predictors were conducted in Stata using data from the 1958 British birth cohort, a sample of all persons born in mainland Britain during one week, followed multiple times since birth, and now aged 55. The analyses conditioned on gender. To address (1), separate regressions were estimated of exposure to stressors in different domains, (family and relationship, child loss, material, occupation-related, and trauma), and of cumulative stressors over the life course, and during childhood, the youth-adulthood transition, and early adulthood, tested separately and then mutually adjusted, using Wald tests for trends. Nested models were used to address (2) with hypothesised confounding and mediating variables added in stages to assess the extent of attenuation. Hypothesised confounders were maternal smoking during pregnancy, low and high birth weight, breastfeeding history, and number of older children in the household during childhood. Hypothesised mediators were early respiratory infections, atopic history, residence in an urbanised industrial area, histories of smoking and obesity, internalising type behaviours during childhood and adolescence, and depressive symptoms. Likelihood ratio tests were performed to compare the fits of models to the data. Sensitivity analyses used alternative measures of asthma onset and exposures to stressors. Analyses were conducted on complete cases and using a multiply imputed dataset.
Results Exposures to 3+ stressors during childhood, transition, and early adulthood were associated with increased odds of asthma onset during mid-adulthood by 2.10 (95% confidence interval (CI) 1.26–3.52), 3.80 (95% CI 1.90–7.62), and 2.57 (95% CI 1.55–4.24), respectively. Associations with exposures during early adulthood were explained by exposures during childhood and transition, but the associations were not attenuated by the hypothesised confounding or mediating variables. Findings were robust to sensitivity analyses.
Conclusion This paper, as the only large scale British study to date, contributes to a small body of evidence suggesting that exposures to stressors predict adult onset asthma. The relative importance of stressors in different domains by gender and life stage may inform decisions about the development and targeting of preventive services.