A follow-up study was undertaken to determine the predictors of the course of recurrent wheezing episodes in children (between 0 to 8 years). In 1991, 600 children with wheezing episodes had been recorded by physicians who participated in a sentinel practice network. On each consultation with the same physicians from October 1994 until June 1995, parents of these children were questioned again about the course of their respiratory symptoms (n = 218). Associations between the course of disease and predictors (recorded by physicians in 1991) were analysed using polytomous logistic regression. The following factors were significantly associated with the frequency of asthmatic episodes (odds ratio (OR) > 1 indicates an unfavourable course of disease in comparison with the reference category, [95% confidence interval]): indoor cigarette smoking: OR=1.7; [1.0-3.0], older than 1 year of age (reference: < 1 year): OR=3.0; [1.1-8.5], more than 5 asthmatic episodes during the year before the first registration: OR=2.7; [1.3-5.6], infect- associated asthma: OR=0.4 [0.2-1.0], pediatrician as recording physician (reference: general practitioner): OR=0.4 (0.2-0.8). No significant association with the course of disease was found for sex, education of the parents, region, parental asthma, hospital admissions because of obstructive symptoms. In correspondence with other studies, the majority of children showed a favourable course of their obstructive respiratory symptoms: for only 7% the frequency of episodes increased during 3 years after the first contact. Indoor smoking and severity of asthma are known as predictors of the course of the disease. The better prognosis of infect-associated obstructive symptoms supports the thesis that the majority of infants with asthmatic symptoms have narrow, infect-mediated airways obstructions, but no increased risk for bronchial asthma in their later lives. An early identification of children at risk may allow a specific and intensified therapy to improve the course of disease.