No previous reports have compared clarithromycin (CAM), rifampicin (RFP) and ethambutol (EB) containing regimens with and without an aminoglycoside antibiotic kanamycin (KM) for the treatment of pulmonary Mycobacterium avium complex (MAC) disease. We conducted a retrospective study to investigate the clinical efficacy of KM using data from 40 patients who received combined chemotherapy for MAC disease with or without KM in the National Hospital Organization Tokyo Hospital from July, 1999 to December, 2005. All patients were administered CAM, RFP and EB for 6 to 12 months, and 20 of the 40 simultaneously received combined chemotherapy with KM. The difference in the backgrounds of the groups was not statistically significant. The improvement rates of clinical symptoms and radiological findings were significantly higher in the KM-treated group than in the KM-untreated group (75% versus 35% and 80% versus 25%). Moreover, the sputum relapse rate was significantly lower in the KM-treated group (18% versus 75%). However, there were no significant differences in the sputum conversion rate (55% with KM versus 40% without KM). As for adverse reactions, there were no significant differences between the groups. Furthermore, we examined time-kill kinetics of KM and streptomycin (SM) against a clinical isolate of M. avium. Most M. avium was killed by KM and SM at concentrations higher than MIC (8 microg/ml), and concentration- and time-dependent killing by KM and SM were almost identical. These observations indicate that KM is effective for treatment of patients with MAC disease.