Background: A previous study of homes in Wellington, New Zealand showed that having carpets on floors was the most important determinant of floor Der p 1 levels, but there was much unexplained variability between houses in carpet levels.
Objective: To determine to what extent housing characteristics might explain this variability in Der p 1 levels between houses.
Methods: We returned to a selection of houses with carpets and sampled living room dust from 1 square metre for 1 min and from the whole floor at 5 m(2) per min. Der p 1 levels were estimated by double monoclonal antibody ELISA and are expressed as geometric mean microg/g and microg/m(2) (95% confidence intervals). Questionnaires were used to collect information on housing characteristics.
Results: Der p 1 levels were significantly higher in the 1 square metre sample (40.0, 31.9-50.2 microg/g; 53.4, 41.4-68.9 microg/m(2)) than in the whole room (25.8, 21.3-31.1 microg/g; 5.3, 3.8-7.4 microg/m(2)). However, results from the different sampling methods were correlated (r = 0.51, P = 0.001 for microg/g and r = 0.58, P = 0.001 for microg/m(2)). After controlling for possible confounders, houses with insulation or a room or garage below the living room had approximately half the Der p 1 concentration (P = 0.05 for both samples) and the amount of Der p 1 per m(2) (P = 0.004 for the 1 square meter sample, P = 0.06 for the whole room sample) than houses without these features. Having more than two children was associated with higher levels of Der p 1 in 1 square meter, significant (P = 0.05) for microg/m(2). Carpet underlay less than 8 mm thick was associated with an almost 3-fold increase in microg/m(2) Der p 1 (P = 0.03) and a 1.6-fold increase in microg/g Der p 1 (P = 0.08) in the whole room sample, when compared with thicker carpet underlays.
Conclusion: The presence of insulation is the single most important housing characteristic explaining the between-house variability in Der p 1 levels on carpeted living room floors.