Treatment of Myocardium avium complex infection: does the beige mouse model predict therapeutic outcome in humans?

Abstract

To determine the predictive value of a standard murine model in the treatment of disseminated Myocardium avium complex (MAC) infection, beige mice were infected with MAC strains isolated from human immunodeficiency virus-infected patients and treated with the same antibiotic (ethambutol, clofazimine, or rifampin) that had been administered to the subject from whom that strain had been recovered. While ethambutol had the greatest bacteriologic efficacy in humans (mean decrease +/-SD, 1.0+/-0.5 log 10 cfu/mL of blood), clofazimine had the greatest bacteriostatic efficacy in mice (mean decrease +/- SD, 2.8 +/- 0.7 log(10) cfu/g of tissue). A linear correlation was not observed between bacteriostatic activity in mouse liver or spleen and the degree of bacteriologic response in humans (P > or = to .1). Odds ratios for a response in humans based on a bacteriologic response in mice were not significant for each agent (P > or = to .1, all cases).