Clinical Applicability of Commonly-Used Exposure Thresholds for Allergen Sensitivity and Asthma Morbidity

Data from the National Cooperative Inner-City Asthma Study (NCICAS) have been used to establish cockroach exposure thresholds for allergic sensitization and asthma morbidity at 2 U/g and 8 U/g, respectively. By definition, a threshold represents a point above which an effect begins. However, these values were cut-points designed to maximize the difference in risk between high and low exposure groups instead of identifying a true threshold. This analysis examined the adequacy of the 2 and 8 U/g cut points as actual thresholds by examining their underlying continuous distributions. Continuous relationships between exposure, sensitivity and morbidity in NCICAS were evaluated using a generalized additive model with generalized cross validation smoothness estimation. The 2 and 8 U/g values mark the start of a slow, steady decline in risk, as opposed to an absolute point where risk begins. Although these cut points do divide the population in to groups with significantly different levels of risk, the curves indicate that the true thresholds for allergic sensitization and asthma morbidity are at lower levels, near the assay's 1 U/g limit of detection. The 2 and 8 U/g cockroach allergen exposure values have generally been accepted as thresholds in the true sense of the word; however, the point at which the risk begins is lower. The continuous relationships detected here indicate that reducing exposure to the established cut-points (or to any discrete value) is valuable, but reducing cockroach exposure to undetectable levels is the best way to minimize the risk of sensitization and asthma morbidity.