Recently, "low-dose and long-term" erythromycin treatment has been reported as effective on diffuse panbronchiolitis (DPB), but its mechanism is still obscure. Patients with DPB were found to have significantly higher percentages of neutrophils in the pre-erythromycin treatment bronchoalveolar lavage fluid (BALF) than healthy nonsmoking volunteers (p < 0.001). They showed a significant reduction in BALF neutrophil percentages after erythromycin treatment (p < 0.01). The neutrophil chemotactic activity (NCA) was significantly elevated in BALF obtained from 19 patients with DPB compared with that from healthy volunteers (p < 0.001). A significant reduction in the NCA was observed in post-erythromycin treatment BALF of 11 patients with DPB (p < 0.001). Additionally, there was a significant correlation between the reduction of NCA and neutrophil percentage in pre- and post-erythromycin treatment BALF (r = 0.726, p < 0.05). Finally, we investigated the effect of erythromycin on the intrapulmonary influx of neutrophils by intratracheal injection of lipopolysaccharide (LPS) and interleukin-8 (IL-8) in mice. The intrapulmonary influx of neutrophils was significantly suppressed (p < 0.001) in mice intraperitoneally injected with erythromycin at 5 mg per animal 2 h before intratracheal injection of LPS (control group: 6.5 +/- 1.6 x 10(5) versus erythromycin-treated group: 1.7 +/- 0.5 x 10(5)), but not 10 h before lung challenge. This inhibition was observed at 6 h after lung challenge and became maximal with 84% suppression at 24 h. Week-long administration of erythromycin did not alter the intrapulmonary influx of neutrophils. The number of neutrophils in the peripheral blood was not affected by erythromycin, indicating that the drug was not toxic.(ABSTRACT TRUNCATED AT 250 WORDS)