Sputum induction can be used to study airway inflammation in asthmatics. However, it has not been used in patients with corticosteroid-dependent asthma requiring long-term oral corticosteroids. The aim of the study was to assess the number of eosinophils and the levels of eosinophil cationic protein (ECP) in sputum of 17 corticosteroid-dependent asthmatics by comparison with nine mild untreated asthmatics, 10 moderate asthmatics receiving inhaled steroids (ICS) and 11 healthy subjects. In the 17 corticosteroid-dependent asthmatics, we examined sputum eosinophil markers on two occasions and correlated with the control of asthma. Eosinophils were undetectable in controls and were detected in 63.8% of asthmatics. There were no significant differences between the three groups of asthmatics. ECP levels were significantly increased in ICS or corticosteroid-dependent asthmatics by comparison to controls and mild asthmatics. There was no significant difference between ICS and corticosteroid-dependent asthmatics. During follow-up, corticosteroid-dependent asthmatics with a controlled disease had no significant change in eosinophil numbers or ECP levels. On the other hand, corticosteroid-dependent asthmatics with recent exacerbations had a non-significant increase in eosinophil numbers and a significant increase in ECP levels. This study shows that ECP levels may be more accurate than eosinophil numbers in assessing exacerbations in corticosteroid-dependent asthmatics.