Measurement of eosinophil percentages and ECP concentration in induced sputum may be useful in the diagnosis and assessment of the variability of airway inflammation in bronchial asthma (BA). To evaluate the usefulness of sputum eosinophil counts and ECP concentrations in the diagnosis of BA, we measured these parameters in 68 patients with respiratory complaints. In addition, we followed-up 14 BA patients with variable airflow limitation for 45.4 +/- 10.4 days. The BA group (n = 41) showed a higher percentage of sputum eosinophilia (24.5 +/- 7.6 vs. 2.2 +/- 2.9%, p < 0.001) and a higher level of sputum ECP (198.2 vs. 90.6 micrograms/L, p < 0.05) than those in the nonasthmatic group (NBA, n = 27). The sensitivity and specificity of sputum eosinophilia (> or = 5%) for the diagnosis of BA were 85.4% and 92.6%, respectively, which were better than the sensitivity (68.3%) and specificity (55.5%) of the increased level of sputum ECP (> or = 100 micrograms/L). Patients with moderate-to-severe persistent BA had a higher percentage of sputum eosinophil (n = 23, 34.6 +/- 10.6%) than those of mild persistent BA (n = 18, 10.7 +/- 5.2%, p < 0.01), but we could not find significant difference in ECP levels between mild persistent and moderate-to-severe persistent asthma. The percentages of sputum eosinophilia showed a moderate correlation with ECP (r = 0.4358, p < 0.01) and with the peak expiratory flow rate (PFR, r = -0.4746, p < 0.01) but sputum ECP did not correlate with PFR. In 14 BA patients who were followed, there was a relationship between changes of PFR and the percentage of sputum eosinophil (r = -0.7238, p < 0.01), but the change of PFR did not correlate with the change of sputum ECP levels. These results suggest that the sputum eosinophil count and sputum ECP level could be helpful in the diagnosis of BA, but that sputum ECP is not satisfactory for the assessment of variability of airway eosinophilic inflammation during the initial anti-inflammatory management of BA.