Although Hanifin and Rajka's criteria have been used for the diagnosis of atopic eczema (AE), there is no instrument destined for epidemiological studies on AE that actually uses them. Written questionnaires (WQ) have generally been used, but when translated into another language they must be validated. The Intemational Study of Asthma and Allergies in Childhood (ISAAC) WQ was previously validated in a comprehensive study, but its validation in Brazil had not been done. Our objective was to validate the eczema component of the self-applicable ISAAC's WQ following its translation into Portuguese. The group of 10 pediatricians and 10 pediatric allergologists graded the questions from 0 to 2 and established the maximum score for each question. The WQ was answered by parents or guardians of children with atopic dermatitis (AE), aged 6-7 years (n = 23) and of non-AE control children of the same age (n = 46) as well as by AE (n = 24) and non-AE (n = 48) adolescents, aged 13-14 years. In order to evaluate the reproducibility of the ISAAC WQ, half of these individuals answered the same questionnaire after 2 to 4 weeks. The maximum possible global scores were 13 for the children aged 6-7 years and 11 for the adolescents, and the cutoff level for both groups was 3. In both age periods the WQ was reproducible (Kappa and McNemar tests) in a significant way (6-7 years, Kw = 0.79; 13-14 years, Kw = 0.73). The prevalence of AE, using the validated WQ, was then studied. The WQ was applied to the parents of 3,005 children aged 6-7 years and to 3008 children aged 13-14 years. Response rates were 72% and 94% for the 6-7-year-old children and the 13-14-year-old children, respectively. There was a slight predominance of male children in the studied population. In the group of the 6-7-year-old children, the cumulative prevalence of AE was 13.2% for boys and for girls; in the group of the 13-14-year-old children, it was 12.5% and 15.4%, respectively. AE severity was similar for both age groups. Using the criteria of global cutoff score, in the group of the 6-7-year-old children, the prevalence of AE was 12.6% for boys and 13.8 for girls; in the group of the 13-14-year-old children, it was 11.7% and 12.4%, respectively. There were no significant differences between them. In conclusion, the AE component of the ISAAC WQ proved to be reproducible, adequate, and able to discriminate between AE and control children. A significant concordance was observed between the criteria utilized in this study (ISAAC x global cutoff score).