Probit analysis applied to the allergen dose-response curve: a method for epidemiologic surveys

The degree of skin sensitivity to five common allergens (grass, Dermatophagoides pteronyssinus, mugwort, birch, and Parietaria) was determined by the threshold dilution technique in all the skin test reactors of a random sample of 295 schoolchildren (142 male and 153 female subjects, age range 11 to 14 years), and the frequency distribution of responders at each concentration was analyzed by probit analysis. The potency of each allergen was presented in terms of median effective dose (ED50), and comparison between different allergens and between symptomatic and asymptomatic subjects was made by computing the relative potency. The ED50 of the allergens was found to be nearly identical (grass, 13.5; D. pteronyssinus, 12.3; mugwort, 9.6 activity units by RAST (AUR)/ml) with the exception of birch (22 AUR/ml). Grass and D. pteronyssinus demonstrated a lower ED50 in symptomatic subjects (3.3 and 2.8 AUR/ml, respectively) than in asymptomatic subjects (85 and 27 AUR/ml, respectively). The lower fiducial limits of ED50 in symptomatic subjects demonstrated to be a cutoff point, since they included only 5% and 12% of asymptomatic reactors to grass and D. pteronyssinus, respectively. We conclude that probit analysis applied to the distribution of threshold doses of allergen extracts is a useful method to evaluate skin sensitivity in epidemiologic surveys. We believe that the ED50 is a practical and reliable allergy index.