Nocturnal cough and wheeze are common in asthma and often treated with beta2 agonists or theophyllines. As nocturnal asthma and these therapies may affect sleep and cognition, we compared 50 microg salmeterol inhaled every 12 h with individually dose-titrated sustained-release oral theophylline on sleep quality and cognitive performance in 15 patients with stable nocturnal asthma (overnight peak expiratory flow rate [PEFR] fall > or = 15%, > or = 1 asthmatic awakening/week) using a double-blind, double-dummy, crossover design with 14-d therapy limbs. Cognitive testing and polysomnography were performed on Nights 13 and 14. Trough plasma theophylline concentration after Night 14 on theophylline was median 11.1 (interquartile range 8.3, 15.2) microg/ml. Overnight PEFR falls were similar [salmeterol 2.3 (0, 10.6), theophylline 3.5 (-0.3, 9.6)%, p = 0.4] but on salmeterol there were more nights without awakenings [median difference 1 (0, 2), p < 0.01], fewer nocturnal arousals [difference -3 (-7, 2) h(-1), p < 0.05] and improved quality of life (p = 0.05). Sleep architecture did not otherwise differ. Visual vigilance improved on salmetrol (p < 0.05), but otherwise daytime cognition was unaffected. There was no patient preference for either therapy. Hence in patients with nocturnal asthma, we demonstrate no major clinical advantage, but a small benefit in sleep quality, quality of life, and daytime cognitive function with salmeterol.