The antimicrobial susceptibilities of Legionella pneumophila isolates grown either in U937 human monocytic cells or in Acanthamoeba polyphaga were studied after release from the host cells without further subculture. Time-survival studies showed that exposure of L. pneumophila cells, grown exclusively in vitro, to 5 g of rifampin per ml resulted in at least 99.9% killing after 6 h and no detectable survivors at 24 h. Similar rates of killing were observed for in vitro-grown cells tested by exposure to ciprofloxacin. Conversely, time-survival studies revealed that macrophage-grown and amoeba-grown cells were ca. 1,000-fold more resistant to the activities of both drugs. Macrophage-grown cells treated with 5 g of rifampin per ml showed 70 and 62% survival after 6 and 24 h, respectively. Intracellularly grown legionellae were also highly resistant to erythromycin (8 g/ml). After 24 h of exposure to the drug, there was 70 and 60% survival for amoeba-grown and macrophage-grown legionellae, respectively, whereas in vitro-grown cells showed a 2-log 10 reduction in viable count. When intracellularly grown L. pneumophila cells were subcultured in broth for 48 h, they reverted to the phenotype characteristic of in vitro growth. Morphologically, the cells were larger than their intracellularly grown counterparts and resistance characteristics were lost. The susceptibilities of the subcultured cells to all three drugs were similar to those of Legionella cells grown exclusively in vitro. In view of these findings, the successful treatment of Legionnaires' disease may be related as much to the resistance phenotype induced by intramacrophage growth as to the ability of the antibiotic to enter phagocytic cells.

We investigated 4 cases of legionnaires' disease (LD) reported among workers at an Ohio automotive plant in March 2001. A "confirmed" case of LD was defined as x-ray-confirmed pneumonia and a confirmatory laboratory test. A "possible" case of LD was defined as elevated titers of antibody and respiratory symptoms. Legionella pneumophila serogroup 1 (LP1) was isolated from 1 case patient. Legionella was isolated from 18 (9%) of 197 environmental samples; 3 isolates were LP1 but did not match the case isolate. We conducted a case-control study; 17 case patients with confirmed or possible LD and 86 control subjects (workers with low antibody titers and without symptoms) were enrolled. Visiting a specific cleaning line (odds ratio, [OR], 7.29; 95% confidence interval [CI], 2.31-23.00) and working in the cleaning region of the plant (OR, 3.22; 95% CI, 1.11-9.38) were associated with LD. LD can be transmitted in industrial settings in which aerosols are produced. Clinicians should consider LD when treating persons from these settings for pneumonia.