BACKGROUND There is a controversy over the occurrence of a late asthmatic response during exercise-induced asthma. While some workers have documented such a response as a genuine phenomenon, others have attributed this to drug withdrawal. OBJECTIVE We carried out the present study to investigate whether a late asthmatic response occurs during exercise-induced asthma as a genuine event and, if so, what are the factors which determine its occurrence. METHODS Sixteen, clinically stable asthmatic patients with laboratory-proven exercise-induced asthma underwent a standardized exercise challenge on a bicycle ergometer. The airway response to exercise was studied by spirometry to measure FEV1. Spirometry was carried out before the exercise, at 4, 8, 15, 30, and 60 minutes, and then hourly for the next 7 hours. Spirometry was also done repeatedly as above on a non-exercise control day, four to seven days earlier. RESULTS Eight (50%) subjects developed a second fall of greater than 10% in FEV1, three to eight hours after recovery from the early response. The late fall in FEV1 after exercise was significantly greater than the spontaneous decay of lung function at the corresponding clocktime on the non-exercise control day. The dual responders did not differ from those with isolated early responses with respect to age, duration and severity of asthma, treatment requirements, peripheral blood eosinophilia, and atopic status. Baseline FEV1 and maximum fall in FEV1 during the early response, and the rate of its development and recovery from it were also similar. Among the dual responders, the late response was not related to the baseline FEV1 or to the intensity of the early response. CONCLUSIONS A late asthmatic response is a genuine phenomenon in exercise-induced asthma. Its occurrence cannot however be predicted by any clinical or physiologic factors.