We studied the sensitivity and specificity of guanine environmental te
sts in the evaluation of the two mite allergen levels, i.e. 2 mug/g an
d 10 mug/g of Der p I and Der f I, considered to be risk factors for s
ensitization or for the development of acute asthma. We examined 239 h
ouse-dust samples for Der p I and Der f I levels (ELISA) and guanine c
ontents (semiquantitative guanine test and quantitative assays). All h
ouse-dust samples with class 2 or 3 guanine tests contained more than
2 mug/g of Der p I and Der f I. The probability that house-dust sample
s of class 2 contained more than 10 mug/g of mite allergens was 88%; i
t was 100% for house-dust samples of class 3. The probability that a h
ouse-dust sample of class 0 contained less than 2 mug/g of mite allerg
en was 87%. For each level of mite allergen, a ROC curve was construct
ed with the true positive rates and the false positive rates calculate
d by different cutoffs of guanine concentration. The cutoff point whic
h gave the best compromise between sensitivity (76%) and specificity (
89%) was 2100 mug/g for the threshold of 2 mug/g of Der p I and Der f
I. For detection of a mite allergen > 10 mug, a guanine content of 300
0 mug/g gave the best compromise between sensitivity (86%) and specifi
city (93%). In conclusion, the guanine test represents a satisfactory
environmental test, inexpensive and simple, for predicting mite allerg
en levels. Semiquantitative and quantitative guanine measurements are
more accurate in predicting an exposure level of 10 mug/g than of the
level 2 mug/g of Der p I and Der f I.