Comparison of vacuuming procedures for reservoir dust mite allergen on carpeted floors.

Allergen concentration in dust collected from a single square meter of carpet has been widely used as a proxy for aeroallergen exposure in a room. However, little is known about the horizontal distribution of allergens in carpeting or whether sampling several smaller sites might better approximate the average level. Eight bedrooms and 4 lounge rooms were divided into squares of 0.25 m(2), and the vacuumed dust was assayed for mite allergen (Der p 1). The proximity of each square to windows, doors, and beds was determined. In each room, the total allergen for each of 3 randomly chosen sets of 4 adjacent squares (forming a 1-m(2) area) was compared with that for each of 3 randomly chosen sets of 4 nonadjacent squares. The reliability of measuring allergen concentration from a single 0.25-m(2) area was modest (intraclass correlation coefficient [ICC] = 0.52), indicating wide-ranging allergen concentrations across short distances within each room. The reliability of the measurements based on 4 nonadjacent squares (ICC = 0.782) was higher than that of the measurements based on 4 adjacent squares (ICC = 0.67). Der p 1 concentrations decreased with distance from doors (paired partial correlation coefficient, -0.1217; P = 0.03). We recommend sampling 4 nonadjacent 0.25-m(2) areas that are not equidistant from the door when reservoir concentrations of Der p 1 are being measured.