Effect of air pollutants on respiratory symptoms and bronchodilator use among children, considering lags and cumulative exposure over several days (multi-pollutant models). OR's were calculated using mixed models with random intercept for logistic regression, adjusting for sex, severity of asthma, atopy, minimum temperature during the previous day, and chronologic time, and including O3, NO2 and PM2.5 simultaneously in the regression models. Symptom changes are shown with an increase of an interquartile range for each pollutant (1-hr O3 and 1-hr NO2 maximums, and 24-hr average for PM2.5) over several averages. * These models considered record of symptoms at night, and pollution for the same day. The other models considered daily symptoms.