The diagnosis of occupational asthma (OA) needs to be made with as much objective evidence as possible. If there is airway inflammation, measurement of this should be an asset. The objective of this study was to investigate whether there is an increase in induced sputum and blood eosinophils and eosinophil cationic protein (ECP) in OA after work exposure. Patients were assessed after a 2-4 week period at work and away from work with cell counts and ECP assays performed blind to the clinical data. They were considered to have OA if symptoms were worse at work and there was a fall in forced expiratory volume in one second (FEV1) > or =20% or in the provocative concentration of methacholine causing a 20% fall in FEV1 (PC20) of four-fold or more compared with away from work. Patients whose symptoms were worse at work but had a change in FEV1 of <20% and in methacholine PC20 of less than four-fold were considered as controls. Sixteen patients were studied. Ten had OA and six were controls. Patients with OA had a significant increase in median (interquartile range) sputum eosinophils and ECP when at work compared with the periods out of work, 10.0 (17.05) versus 0.8 (1.6)% (p=0.007) and 3,840 (6,076) versus 116 (180) microg x L(-1) (p=0.01). They also had a higher blood eosinophil count, 0.3 (0.5) x 10(9) versus 0.2 (0.1) x 10(9) x L(-1) (p=0.013), and a trend towards higher serum ECP levels, 44.0 (20.0) versus 32.0 (18.5) microg x L(-1) (p=0.07). In conclusion, the proportion of eosinophils and levels of eosinophil cationic protein in sputum are particularly high at work in patients with occupational asthma, suggesting that the measurement of these factors can supplement other physiological outcomes in establishing the diagnosis of occupational asthma.