There is no consistently reliable treatment for endocarditis resulting from chronic Coxiella burnetii infection, the causative agent of Q fever. Although certain antibiotics are recommended on the basis of their in vitro bactericidal activities, results of therapy with these antibiotics are often disappointing. To evaluate whether the currently recommended antibiotic susceptibility tests for C. burnetii give misleading results because of continued division of uninfected cells, thereby resulting in the dilution of infected cells and, hence, a false picture of antibiotic efficacy, we blocked cell division during antibiotic susceptibility testing with cycloheximide. Using this new method, we found that the currently recommended antibiotics for the treatment of Q fever, doxycycline, pefloxacin, and rifampin, did not reduce the ratio of infected to noninfected cells (either L929 or P388D1) by 9 days postinfection. To test the hypothesis that this lack of antibacterial activity is due to antibiotic inactivation by the low pH of the phagolysosomes in which C. burnetii is found, we used alkalinizing lysosomotropic agents (chloroquine or amantadine) concurrently with doxycycline. This resulted in the sterilization of C. burnetii infection in P388D1 cells. This finding seems to confirm our suspicion that the acidic conditions of the phagolysosomes in which C. burnetii is located inhibit antibiotic activity. This inhibition can be reversed in vitro when lysosomotropic alkalinizing agents are used.