Differences in drug resistance profiles of Mycobacterium tuberculosis isolates causing pulmonary and extrapulmonary tuberculosis in a medical centre in Taiwan, 2000–2010

Few studies have investigated the drug resistance profiles of Mycobacterium tuberculosis (MTB) isolates recovered from different sites of infection. A total of 4521 non-duplicate MTB isolates, including 3723 (82.3%) from respiratory specimens and 798 (17.7%) from non-respiratory sources, were recovered from patients treated at a medical centre in Taiwan from 2000 to 2010. Trend analysis showed a significant decrease (P < 0.05) in the rates of resistance to isoniazid, rifampicin and ethambutol, a decrease in resistance to any one of four agents, namely isoniazid, rifampicin, ethambutol or streptomycin, and a decrease in resistance to both isoniazid and rifampicin (multidrug resistance) amongst pulmonary MTB isolates. A similar decrease in resistance to isoniazid and ethambutol was noted amongst non-pulmonary isolates. Rates of drug resistance were significantly higher amongst MTB isolates recovered from respiratory specimens than amongst those from non-respiratory specimens to 0.2 μg/mL isoniazid (15.3% vs. 9.4%; P < 0.0001), 1 μg/mL rifampicin (5.5% vs. 3.3%; P = 0.0108), 5 μg/mL ethambutol (7.3% vs. 3.8%; P = 0.0004), and both isoniazid and rifampicin (4.8% vs. 2.5%; P = 0.0051). Resistance rates amongst isolates causing tuberculous lymphadenitis were significantly lower than amongst those causing genitourinary tuberculosis (TB) to isoniazid (3.5% vs. 19.4%, P = 0.0012) and to isoniazid, rifampicin, ethambutol or streptomycin (9.6% vs. 22.6%, P = 0.0003). In conclusion, the rates of resistance to first-line anti-TB agents and to multiple agents differed amongst MTB isolates obtained from different infectious sources. Continuous monitoring of resistance of MTB isolates from various sites is necessary in order to establish an effective TB surveillance programme.