In order to evaluate the efficacy of macrolides in pediatric patients with recurrent respiratory tract infections (RRTIs), we enrolled 1,706 children (783 females) aged between 6 months and 14 years (median: 4 years) with an acute respiratory infection and a history of RRTIs (> or = 8 episodes per year if aged < 3 years; > or = 6 episodes per year if aged > or = 3 years). The therapies were chosen by the primary care pediatricians and their effects on respiratory relapses were blindly analyzed. Regardless of age and clinical diagnosis, the children treated with macrolides showed a significantly higher rate of short- and long-term clinical success than those receiving beta-lactams (p<0.0001) or symptomatics alone (p<0.0001). These data show that macrolide therapy of acute respiratory infections influences the natural history of RRTIs, probably because of their elective activity on atypical bacteria. They also suggest the possible importance of these pathogens in causing recurrences of respiratory infections in children and show that the infections they cause may have a more complicated course unless treated with adequate antibacterial drugs.