We wished to assess the efficacy of inhaled salmeterol (SML; 50 micrograms b.i.d.) compared to a combination of slow-release theophylline and ketotifen p.o. (TK; T 300 mg+K 1 mg b.i.d.) for the treatment of nocturnal asthma. Ninety six patients with nocturnal asthma, (forced expiratory volume in one second (FEV1) 60-90% of predicted value, reversibility > or = 15%, at least two nocturnal awakenings per week) were eligible for a multicentre, double-blind, double-dummy cross-over study (14-day run-in, two successive 28-day treatment periods). Efficacy was assessed as success/failure, success being defined as the complete disappearance of nocturnal symptoms/awakening during the last week of each treatment period. There was a statistically significant difference between SML and TK for this criterion: 46% and 39% success with SML during periods I (first 28-day period) and II (following the cross-over), compared to only 15% and 26% with TK, respectively (p < 0.01). SML was also significantly better for the other criteria (lung function, rescue salbutamol intake during day and night). Side-effects were five times less frequent in SML-treated patients (p < 0.004). Efficacy and tolerance of SML were obviously far better than those of TK in patients with nocturnal asthma.