The bactericidal activities of several once-weekly rifapentine (P)-containing combination regimens against Mycobacterium tuberculosis, and their ability to prevent the selection of rifampin (R)-resistant mutants, were compared with those of the standard six-times-weekly regimen consisting of R, isoniazid (H), and pyrazinamide (Z) in a mouse experiment. Mice were infected intravenously with 1.3 x 10(7) cfu of M. tuberculosis strain H37Rv, and 8 wk of treatment began on Day 14 after infection, when mice were randomly allocated to an untreated control group and nine treatment groups of 30 mice each. At the end of 8 wk of treatment, all the tested regimens showed promising bactericidal activities. Once-weekly P alone was less bactericidal than six-times-weekly R alone; likewise, the once-weekly P-containing combined regimens were less bactericidal than the six-times-weekly standard regimen. However, the difference in killing was about 1 log10, which represented only a fraction of the overall 4 log10 to 5 log10 magnitude of killing effects. The addition of streptomycin (S) improved the bactericidal effect of once-weekly PHZ, and the effect of once-weekly PHZS was further enhanced when it was preceded by 2 wk of daily HZS. The latter regimen achieved the same level of activity as the standard six-times-weekly regimen. All of the once-weekly P-containing combined regimens were able to prevent the selection of R-resistant mutants, whereas monotherapy with R or P selected resistant mutants in approximately 50% of animals.