Facebook
Twitter
LinkedIn
Email
 |
Checking for direct PDF access through Ovid | |
Excerpt
Although a body of research is developing regarding the role of air pollutant exposures on severity, exacerbations, or prevalence of asthma, 12–15 there is relatively little research on asthma initiation. For that reason, the study by Ponsonby and colleagues in this issue should attract considerable attention. 16 Their study creates a cohort of children with prospective data from infancy by linking a cross-sectional survey for asthma occurrence at age 7 with an infant cohort originally assembled to study sudden infant death syndrome (SIDS). This clever approach provides the opportunity to examine associations with asthma for some important risk factors, including certain indoor air pollutants. Notably, their results point toward the possible role of exposure to indoor combustion products on asthma initiation.
Ponsonby and colleagues provide some provocative insights into these questions by virtue of their cohort design, but our confidence in their conclusions suffers from the limitations inherent in the combination of two studies. Because of its initial focus, the infant data collection reflected the goals of the original SIDS study; and questions targeting asthma risk factors were limited. Apparently there were no biological or environmental samples collected on the original cohort. Assuming a dose-response effect of environmental tobacco smoke (ETS) on asthma onset, the cotinine validation study from a separate cohort of infants points to a different exposure pattern than the one suggested by the asthma onset pattern in the original cohort—the lowest cumulative incidence occurred in smoking homes where infant ETS exposures were limited rather than in nonsmoking homes. The available ETS, indoor gas combustion, and ventilation predictors are suggestive of an effect of combustion dose, but the results are not consistent. Exposure misclassification is one likely candidate for these inconsistent results. Another possible explanation is bias due to the interplay between exposure-related selection in the original SIDS cohort and the potential for response-related selection in the cross-sectional asthma survey. 17 Importantly, the identified associations only account for a small attributable fraction in this population. Looking beyond these concerns, we believe this study provides good justification for investing in a much more targeted asthma study beginning in infancy.
Most air pollution epidemiology studies of asthma health effects are opportunistic studies of public use datasets (eg, ecologic time-series studies of short-term ambient air pollutant effects on morbidity outcomes), relatively inexpensive observational studies such as asthma panel studies, or large cross-sectional studies. While these studies have been successful in identifying influences of criteria air pollutants on lung function changes or worsening of asthma symptoms, they have two substantial drawbacks.
