Interlaboratory drug susceptibility testing of Mycobacterium tuberculosis by a radiometric procedure and two conventional methods.

A total of 224 recent isolates of Mycobacterium tuberculosis from 163 patients selected to have multidrug resistance were tested against streptomycin (SM), isoniazid, rifampin, and ethambutol (EMB) by the rapid radiometric BACTEC method and two conventional proportion methods: the World Health Organization (WHO) method, using Lowenstein-Jensen medium; and the Veterans Administration reference laboratory for mycobacteria (VA) method, using Middlebrook 7H10 agar medium. The results were compared, focusing on the concentrations of the drugs in all three methods. Among the four drugs tested, most of the discrepancies in measured activity were observed with SM and EMB, generally because of differences in the drug concentrations used by the three methods. A 4-micrograms amount of SM in the BACTEC method was found to be slightly less active than 10 micrograms in the VA method and significantly more active than 4 micrograms of dihydrostreptomycin in the WHO method. With EMB, 2.5 micrograms in BACTEC was similar to 5 micrograms in the VA method and 2 micrograms in the WHO method, while 10 micrograms in the BACTEC method was found to be more active than 10 and 2 micrograms in the VA and WHO methods, respectively. To attain close agreement, drug concentrations used in the BACTEC method should be carefully selected when a comparison is to be made with any conventional method employed in a laboratory. Standardization of in vitro susceptibility testing is greatly needed to achieve uniformity among the test methods used to evaluate tuberculosis therapeutics.