This study analyzed resistance of Cambodian MDR-TB strains to major anti-tuberculosis drugs.
Unpublished data showing that 46% of TB relapses in Cambodia are MDR is consistent with previous reports from the Central African Republic , Japan , and South Africa .
RMP-R was confirmed with rpoB core region sequencing in 94% of MDR strains. This agrees with other studies showing that some RMP-R strains carry a wild type rpoB core region sequence . rpoB core region mutations were in the hot spot positions associated with RMP-R strains [12, 13]. We observed a high number of MDR strains carrying several rpoB mutations. This may suggest a stepwise process of additive mutations (as described for FQ-R) that could generate higher levels of resistance to RMP [14–16].
To evaluate the number of XDR strains among MDR strains we identified mutations in gyrA and rrs genes which are associated with FQ resistance and AG/CM resistance respectively. Our study found that 14% of clinical MDR-TB isolates were FQ resistant. This rate is lower than that found in other studies of MDR-TB in Southern Asia: Taiwan 22.2%  or 42.8% using phenotypic DST; Shanghai 25.1%  and the Philippines 17%  or 51.4% . However, these figures are much higher than those found on other continents (4.1% in MDR-TB in the United States and Canada  and 4.3% in MDR-TB in Russia ). Only one of the MDR FQ-R strains was AM-R. Therefore, only 1% of MDR strains are identified as XDR strains. This shows that identification of only gyrA FQ-R mutations is not a strong indication of an XDR phenotype.
Previous studies [23, 24] showed a high level of concordance between MTBDRsl and phenotypic tests: I.E. 90.2% and 75.6% respectively. In this study, we had no access to phenotypic tests for FQ. MTBDRsl and sequencing data of gyrA are 100% concordant and these techniques identified 14% FQ-R in MDR strains. This FQ-R rate may have been underestimated because of the unavailability of FQ phenotypic testing in Cambodia. FQ resistance is most often associated with gyrA mutations [25, 26] and more particularly mutations at codons 94 and 90. This was the case in our study. The percentage of FQ-R M. tuberculosis clinical isolates with gyrA mutations have been reported to be between 70% and 90% . Because we have no access to FQ-R phenotypic tests, our analysis may have missed a number of cases. However, as reported before [23, 24], in the absence of phenotypic tests, molecular tests will detect more than 75% of FQ-R cases. Mutations other than those affecting gyrA and other mechanisms could result in FQ-R, including: decreased cell-wall permeability to drug, efflux pumps, drug sequestration or perhaps even drug inactivation [12, 26]. In a small number of cases, FQ-R could be associated with gyrB mutations and a probable efflux mechanism [27, 28].
FQ has become an essential part of treatment regimens for MDR tuberculosis [25, 26]. Due to their efficacy and safety, the new generation of FQ's is even being evaluated as a first-line medication for tuberculosis [27–29]. Unfortunately, the extensive use of FQ has increased spontaneous acquisition of mutations associated with FQ-R. It has been suggested that routine FQ-R testing in locations where resistant strains are endemic may be clinically useful by showing a significant correlation between development of FQ-R and first-line M. tuberculosis drug resistance . However, extensive use of FQ would likely increase MDR tuberculosis treatment failure. As in other developing countries, problems arise from the uncontrolled use of antibiotics. However, it is interesting to note that only 1 out of 14 MDR FQ-R strains is XDR, thus showing that detection of FQ-R strains cannot be used as a single marker for the detection of XDR cases. Additional tests are required to identify XDR cases, among them molecular tests like MTBDRsl that include the detection of mutations in gyrA associated to FQ-R and rrs associated with AM, KM and CM resistance.
Work by other researchers has demonstrated that the MTBDRsl assay detected 86.7% (39/45) and 100% (5/5) of phenotypically AM and CM resistant TB strains [23, 24]. The mutation a1401 g was the most prevalent. The XDR isolate in our study was confirmed to carry a1401 g mutation in the rrs gene. This result suggests a low rate of XDR amongst MDR TB in Cambodia (1%) as compared to countries like India and Taiwan that have reported rates of 8-15% and 10% respectively [31, 32]. This favorable situation could be linked to the implementation of the DOTS strategy in 1980 by the national TB control program in Cambodia .
Spoligotyping confirmed previous observations that the Beijing family is prevalent in Asia. The higher proportion of Beijing spoligotypes in the MDR group (58%) compared with the non MDR group (28%) is consistent with previous studies demonstrating that strains of the Beijing genotype more readily acquire resistance mutations than non-Beijing strains [33–35]. This percentage is increased in the MDR FQ-R group (71%). Although most of the spoligotypes had the Beijing profile, no major MDR outbreak by this family has been reported in Cambodia. Further studies with new markers are warranted to provide a more accurate picture of the epidemiology of the Beijing strains in Cambodia.