Epidemics of mycobacteria due to contamination of medical devices continue to occur. For this reason, we assessed the ability of disinfectants, generally used in hospitals for disinfecting noncritical and semicritical patient care items, to inactivate mycobacteria. A modified Association of Official Analytical Chemists' (AOAC) Tuberculocidal Activity Test, using Middlebrook 7H9 broth as the primary subculture medium and neutralization by dilution, was used to assess the ability of 14 hospital disinfectants to inactivate about 10(6) Mycobacterium tuberculosis and about 10(5) Mycobacterium bovis at 20 degrees C using 10- or 20-minute exposure. All products were tested for each organism using 10 penicylinders (P) and were prepared at the manufacturers' recommended use-dilution. Chlorine dioxide, 0.80% hydrogen peroxide plus 0.06% peroxyacetic acid, and an iodophor achieved complete inactivation (0 + P) of both M. tuberculosis and M. bovis. One quaternary ammonium compound with a tuberculocidal label claim, a quaternary ammonium compound without a tuberculocidal label claim, chlorine (approximately 100 ppm) and 0.13% glutaraldehyde/0.44% phenol/0.08% phenate were not effective (10 + P) against both M. tuberculosis and M. bovis. Another quaternary ammonium compound with a tuberculocidal label claim was tested against only M. bovis and found ineffective (10 + P). Glutaraldehydes (2% alkaline and 2% acid), a phenolic and chlorine (approximately 1,000 ppm) demonstrated complete inactivation of M. tuberculosis (0 + P) and good inactivation of M. bovis (1-3 + P). Two disinfectants, hydrogen peroxide and ethyl alcohol, provided differing results against M. tuberculosis and M. bovis. These studies have important implications for disinfecting semicritical patient care items.