An inhaled glucocorticoid is currently the medication of choice for long-term control of persistent asthma in children. The role of long-acting beta2-adrenergic-receptor agonists, such as salmeterol, needs to be defined. We conducted a randomized, double-blind, placebo-controlled, parallel-group, one-year study of 241 children (mean [+/-SD] age, 9.3+/-2.4 years) with clinically stable asthma and less than one month of prior glucocorticoid use. We compared inhaled beclomethasone dipropionate (200 microg twice daily) with salmeterol xinafoate (50 microg twice daily) and placebo (lactose). The primary outcome measure, airway responsiveness (as assessed with a methacholine challenge) was evaluated before treatment; after 3, 6, 9, and 12 months of treatment (12 and 36 hours after study medications had been withheld); and 2 weeks after the end of treatment. Spirometry, symptoms, use of rescue medication (200 microg of albuterol inhaled as needed), and adverse effects were also assessed. During months 1 through 12 overall, beclomethasone was associated with significantly less airway hyperresponsiveness than salmeterol (P= 0.003) or placebo (P<0.001). This effect was lost two weeks after treatment had been stopped. As compared with placebo, beclomethasone was associated with less variability between morning and evening in the peak expiratory flow (P=0.002), as was salmeterol (P=0.02). Beclomethasone was also associated with a reduced need for albuterol as rescue therapy (P<0.001) and fewer withdrawals because of asthma exacerbations (P=0.03), but salmeterol was not (P=0.09 and 0.55, respectively). During months 1 through 12, linear growth was 3.96 cm in the children receiving beclomethasone, as compared with 5.40 cm in the salmeterol group (P=0.004) and 5.04 cm in the placebo group (P=0.018). Height was not measured after treatment ended. Beclomethasone was effective in reducing airway hyperresponsiveness and in controlling symptoms of asthma, but it was associated with decreased linear growth. Salmeterol was not as effective as beclomethasone in reducing airway hyperresponsiveness or in controlling symptoms; however, it was an effective bronchodilator and was not associated with rebound airway hyperresponsiveness, masking of symptoms, or adverse effects.