STUDY OBJECTIVES: To compare the efficacy, safety, and effects on sleep quality of salmeterol and extended-release theophylline in patients with nocturnal asthma. DESIGN: Randomized, double-blind, double-dummy, three-period crossover. SETTING: Outpatients at a single center. Patients spent 1 night during screening and 2 nights during each study period in a sleep laboratory for completion of sleep studies. PATIENTS: Male and female patients who were at least 18 years old with nocturnal asthma (baseline FEV1, 50 to 90% of predicted) and who required regular bronchodilator therapy. Patients on inhaled corticosteroids, cromolyn, and nedocromil were allowed into the study if their dosing remained constant throughout the study. INTERVENTIONS: Inhaled salmeterol (42 microg per actuation), extended-release oral theophylline (titrated to serum levels of 10 to 20 microg/mL), and placebo taken twice daily. MEASUREMENTS AND RESULTS: Efficacy measurements included nocturnal spirometry, nocturnal polysomnography, sleep questionnaires, and daily measurements of lung function and symptoms. Salmeterol was superior to theophylline (p < or = 0.05) in maintaining nocturnal FEV1 levels and was superior to placebo (p < or = 0.05) in improving morning and evening peak expiratory flow (PEF) and in decreasing nighttime albuterol use. The use of salmeterol significantly increased the percentage of days and nights with no albuterol use and decreased daytime albuterol use compared with theophylline and placebo (p < or = 0.05). Sleep quality global scores significantly improved with salmeterol and placebo (p < 0.001) but not with theophylline. The effects on sleep architecture were similar across treatment groups. CONCLUSIONS: Salmeterol (but not theophylline) was associated with sustained improvements in morning PEF, protection from nighttime lung function deterioration, reductions in albuterol use, and improvements in patient perceptions of sleep. No differences were seen in polysomnographic measures of sleep quality.