To characterize seasonal variation of three allergens (dust mite, cat, and cockroach) and total culturable fungi and to explore whether residential characteristics were associated with the concentrations of these agents, floor dust was collected from 47 inner-city homes in Minneapolis, Minnesota, over a 1-year period. A longitudinal analysis of allergen and fungal concentrations was carried out using mixed-effect models. Overall, relative humidity was a better predictor of allergen concentrations over time than indoor temperature. Seasonal variation of cat and cockroach allergens was negligible compared with the variability associated with residential characteristics such as race/ethnicity, family income, and the presence of cats. Fungal concentrations showed significant seasonal variation that outweighed the variability associated with residential characteristics. Less than 30% of the dust mite allergen and cockroach allergens concentrations were above limits of detection. Observed cockroach allergen concentrations were higher in Spanish- and Somali-speaking households than in English-speaking households, while English-speaking households had significantly higher cat allergen concentrations compared with the other language groups. The ratios of within-home to between-home variance for total culturable fungi, dust mite, cockroach, and cat allergen concentrations were 2.54, 1.91, 0.55, and 0.24, respectively. This ratio is used to predict the number of repeated measurements of each allergen required to robustly estimate long-term exposure estimates such that exposure misclassification bias is kept within acceptable limits. It is not clear whether repeated measurements of dust mite and cockroach allergens are required for long-term average exposure because of the large fraction of nondetects. It is concluded that a single measurement of cat allergen is a reasonable surrogate for long-term average exposure, since repeated measurements over time were highly correlated. Total culturable fungi will require greater than nine repeated measurements for robust assessment of long-term exposures because of low correlations in fungal measures over time.
October 5th, 2018