From November 2008 through May 2010, M. tuberculosis isolates were cultured during routine care of adults >18 years of age with primary TB and no history of treatment. The patients were from 2 regional referral centers, the Irkutsk Regional TB-Prevention Dispensary and the Research Practice Center for Phthisiatry (Yakutia); the study was approved by the institutional review boards at the University of Virginia and Irkutsk State Medical University.

Initial pretreatment isolates were grown on Lowenstein-Jensen agar slants and identified to species in accordance with World Health Organization recommendations. Drug susceptibility was tested by absolute concentration method on agar slants; drugs tested were rifampin (critical concentration 40 µg/mL), isoniazid (1 µg/mL and 10 µg/mL), ethambutol (2 µg/mL), streptomycin (10 µg/mL), ethionamide (30 µg/mL), and kanamycin (30 µg/mL). Susceptibility to a fluoroquinolone and pyrazinamide was not routinely tested. DNA extraction was performed on all isolates, followed by 12-loci mycobacterial interspersed repetitive unit–variable number tandem repeat (MIRU-VNTR) analysis (7) and further lineage definition by region of difference deletions, or for Ural strains as described (5). Phylogenetic tree construction was based on the MIRUVNTRplus database (8), and VNTR international type numbers were confirmed on the SITVIT database (9). DNA from MDR isolates was amplified and sequenced for the known drug-resistance determining regions katG, inhA, rpoB, embB, gyrA, rrs, and eis by using methods described by the Centers for Disease Control and Prevention (10). For pncA, the entire open reading frame and upstream promoter region were amplified. Sequences were compared with published sequences for M. tuberculosis H37Rv by using GeneDoc version 2.7.0.

Among 235 patients with primary TB (130 from Yakutia, 105 from Irkutsk), isoniazid monoresistance was found in isolates from 16 (12%) from Yakutia and 19 (18%) from Irkutsk (p = 0.27). Multidrug resistance was found for 61 patients (36 [28%] from Yakutia and 25 [24%] from Irkutsk) (p = 0.55). Mean age (± SD) for these 61 patients was 33 (± 12) years, 40 (66%) were male, and these characteristics did not differ significantly between patients from Irkutsk and from Yakutia. However, no HIV-infected patients were identified from Yakutia compared with 11 (44% with MDR TB) from Irkutsk (p<0.001). Twelve MDR TB patients from Irktutsk died (outcome unknown for the other 13 patients), including all with HIV, compared with 4 (11%) from Yakutia who died (p = 0.002). Follow-up varied and was limited mostly to inpatients.

Among all 235 patients with primary TB, strains of the Beijing family were significantly more common among those from Irkutsk (70 [67%]) than from Yakutia (40 [31%]) (p<0.001). However, strains found in Yakutia (S 256 [11%], T 8 [7%], and Ural 171 [5%]) were not found in Irkutsk (Table 1). The cluster of S 256 (MIRU profile 233325153325) was the most common among primary MDR TB isolates from Yakutia and was fully 86% MDR (Table 1; Technical Appendix).

Among isolates from patients with primary MDR TB, 51 (84%) were available for DNA sequencing: 27 from Yakutia and 24 from Irkutsk (Table 2). Among isoniazid-resistant isolates, the mutation in codon 315 of katG was present in 91%. Among rifampin-resistant isolates, mutations in the resistance-determining region of rpoB (codons 511–533) were present in only 79%. The pncA mutation was common across genotypes from both sites, occurring in 62% of isolates amplified. Notably, both isolates with mutation in eis from Yakutia occurred in MDR strains with the S 256 genotype and without rrs mutation.

In eastern Siberia, >25% of primary TB was MDR, equivalent to the highest proportion reported from the Russian Federation (2). However, regionally specific genotypic patterns and resistance mutations were identified. As expected, in Irkutsk primary MDR TB was driven by strains of Beijing lineage (5,6). Yet in the more geographically isolated population of Yakutia, a strain previously unidentified in the Russian Federation, S 256, had a MIRU profile recently found among Canadian Aboriginal populations (11). In Yakutia, S 256 was highly drug resistant and was the most common genotype among patients with primary MDR TB.

Although rpoB mutations were found in only 79% of rifampin-resistant isolates, these findings are consistent with those in a recent report from Novosibirsk oblast, which similarly included non-Beijing and S-family strains and found a sensitivity of only 63% for the rpoB mutation (12). Lack of phenotypic correlation can result from alternate mechanisms of resistance or imperfect conventional susceptibilities in Lowenstein-Jensen medium or from use of old drug stock. Such discrepancy necessitates urgent clarification because substitution of conventional susceptibility testing with molecular probe–based methods such as GeneXpert MTB/RIF (Cepheid, CA, USA) has been strongly advocated but would lead to dramatically different results and treatment regimens (13). Of note, isolates of the S 256 strain accounted for a proportion of the cases in which mutation in the promoter region of eis was associated with kanamycin resistance, but rrs was wild type. Commercial assays have focused on the rrs locus, which has greater sensitivity for amikacin, as the sole target for the class of injectable agents (14), yet in Eastern Siberia, the injectable agent available is kanamycin. Furthermore, we found a range of reported and unreported mutations across the entire pncA gene; most were point mutations resulting in amino acid substitution, but some strains had mutations that resulted in deletion or frameshift. Phenotypic methods and assays of functional pyrazinamidase activity should be performed in this region because results might have major implications for novel MDR TB drugs that work best with pyrazinamide (15).

Study limitations include selection bias of isolates from passive surveillance. We were unable to obtain detailed clinical information about all patients with primary TB, thus preventing adequate comparison of nongenotypic risk factors for MDR TB or establishment of definitive epidemiologic links among clustered isolates. Furthermore, lack of conventional fluoroquinolone or pyrazinamide susceptibility testing limited comparison with gyrA and pncA mutations, respectively. Despite these limitations, this work characterizes severe isoniazid monoresistant and MDR TB in eastern Siberia among patients with no history of TB treatment. The regionally distinct phylogenetic patterns and certain drug-resistance mutations necessitate careful application of novel diagnostics and empiric therapeutic strategies.