Seventy-four Mycobacterium tuberculosis clinical isolates from China were subjected to drug susceptibility testing using ethambutol, isoniazid, rifampin, and ofloxacin. The results revealed that the presence of embB mutations did not correlate with ethambutol resistance but was associated with multiple-drug resistance, especially resistance to both ethambutol and rifampin.Ethambutol (EMB), often used in combination with isoniazid, rifampin, and pyrazinamide, is a key drug of first-line antituberculosis treatment. EMB seems to exert its toxic effect by inhibiting the embABC-encoded proteins, and mutations in the embB gene appear to play a major role in the development of EMB resistance in Mycobacterium tuberculosis (13). The marked clinical association between embB codon 306 mutations and EMB resistance in M. tuberculosis at one time led to its proposal as a marker for EMB resistance in diagnostic tests (6,8,11,15). However, discrepancies between the results of genotypic and phenotypic EMB resistance testing have raised questions about the accuracy of molecular assays based on the detection of point mutations in embB codon 306 for prediction of EMB resistance (3,4,7,14). Hazbon et al. (3) has reported that for embB codon 306 mutations, there is "a novel association with broad drug resistance and IS6110 clustering rather than ethambutol resistance."For DNA sequencing of the embB gene (primer set includes forward, 5Ј-CGGCATGCGCCGGCTGATTC, and reverse, 5Ј-TCCACAGACTGGCGTCGCTG) from 141 EMB-resistant and 40 EMB-sensitive clinical isolates from Henan Province, China, ABI Prism Big Dye terminator sequencing kits were used. We found that 45.2% of EMB-resistant isolates harbored embB codon 306 mutations (ATG to ATA, GTG, ATT, CTG, or ATC [five types]) (12). We also found that 15% (6/40) of EMB-susceptible isolates had embB gene mutations (the breakpoint concentration of EMB is 2 g/ml in L-J medium) in 2000. After a previous analysis by denaturing highpressure liquid chromatography (DHPLC) of drug resistance genes in M. tuberculosis in our laboratory (9, 10), we went back to test these isolates and found that the DHPLC results for the embB gene were completely consistent with those of DNA sequencing. Figure 1 shows the DHPLC and DNA sequencing results for the six isolates that were EMB sensitive but harbored embB mutations. Results similar to those we obtained for streptomycin resistance were obtained (R. Shi, J. Zhang, C. Li, Y. Kazumi, and I. Sugawara, unpublished data). We also found that 22% (16/72 EMB-sensitive clinical isolates from an affiliated hospital of the Beijing Tuberculosis and Lung Tumor Research Institute) possessed embB mutations in 2006. We tested the EMB MICs for these 16 isolates, and all were less than 2 g/ml, but the MIC for four isolates was 0.125 g/ml (data not shown).