Induced sputum differential cell counts have been advocated as a method of non-invasively assessing airway inflammation in asthma and other airway diseases. Since sputum induction usually involves delivering hypertonic saline via a high output ultrasonic nebulizer there have been concerns about its safety in asthma. There are relatively little data on the effects of sputum induction in large numbers of patients. We have examined the success rate and effect of sputum induction on forced expiratory volume in 1 sec (FEV1) in 100 inductions performed on 79 patients using a low output nebulizer. Thirty-seven patients had asthma, 29 had miscellaneous conditions (mainly chronic cough) and 13 were subjects without respiratory symptoms. Sputum was induced 10 min after 200 micrograms of inhaled salbutamol by sequential 5-min inhalations of 3, 4 and 5% saline delivered via a Fisoneb ultrasonic nebulizer and FEV1 was measured after each inhalation. Sputum induction resulted in a sample suitable for analysis in 92% of asthmatics, 90% of those with miscellaneous conditions and 100% of normal subjects. The mean (SEM) maximum per cent fall in FEV1 was 5.4% (0.1), 4.3%, (1.0) and 2.6% (1.1) in subjects with asthma, miscellaneous conditions and in asymptomatic subjects respectively. Only 13 inductions resulted in a > 10% fall in FEV1, and only three of these resulted in a > 20% fall. The maximum per cent fall in FEV1 did not correlate with baseline FEV1 % predicted (r = -0.17), the log sputum eosinophil count (r = -0.12), or the methacholine PC20 (r = -0.14). We conclude that sputum induction using a relatively low output ultrasonic nebulizer with premedication with salbutamol is successful and safe in the majority of patients with asthma and other airway conditions.