Ninety-six patients with nocturnal asthma (FEV1 = 60-90% pred, reversibility greater than or equal to 15%, at least 2 awakenings in the week preceding the trial) were included in a double-blind, randomized, crossover, multicenter study which compared the efficacy and side effects of inhaled salmeterol (50 micrograms morning and evening) to the association theophylline-ketotifen (300 mg and 1 mg morning and evening, respectively). The trial included a run-in period of 14 days and 2 periods of successive treatment of 28 days each. The efficacy was expressed in terms of therapeutic success, defined by the total disappearance of nocturnal symptoms during the treatment week. A statistically significant difference (p less than 0.01) was found between salmeterol and the association for this criteria: during the first period, 46% of subjects treated by salmeterol did not present nocturnal awakenings during the last treatment week by comparison with 15% of subjects taking the association; during the second period, corresponding figures were 39% for salmeterol by comparison with 26% for the association. Differences were also significant, favoring salmeterol, for other criteria (lung function tests, extra-need for salbutamol). Side effects were 5 times more frequent for the association (p less than 0.004). Salmeterol is clearly superior to the association theophylline-ketotifen in the treatment of nocturnal asthma.