E-test for susceptibility testing of Mycobacterium tuberculosis.

SETTING Initial isolates should be tested for drug susceptibility to confirm the anticipated effectiveness of chemotherapy. OBJECTIVE To evaluate E-test strips for susceptibility testing of Mycobacterium tuberculosis. DESIGN A proportion method using Lowenstein-Jensen medium and the Bactec radiometric system were compared with the E-test (isoniazid [INH], rifampicin [RMP], ethambutol [EMB] and streptomycin [SM]). RESULTS For 73 of the 81 M. tuberculosis isolates (90.1%) the proportion and E-test methods yielded concordant susceptibility results against all four antimicrobial agents tested. Of these 73 strains, 69 were fully susceptible; the four isolates showing resistance to antimicrobial drugs by both methods were also resistant when tested by Bactec 460TB. While the proportion method indicated susceptibility for the eight remaining strains, E-test results showed mono EMB resistance in five strains, INH resistance for two isolates (including one isolate resistant to EMB plus INH), and for one strain E-test yielded resistance to EMB and SM. Using Bactec as the reference method, the E-test resulted in false resistance in eight strains and no false susceptibility. CONCLUSION Due to a substantial rate of false resistance, this method cannot be recommended at present for practical use in clinical laboratories.