Multidrug-resistant tuberculosis (MDR-TB) is defined as one that is resistant to both isoniazid and rifampicin regardless of its resistance to any other antituberculosis drugs. According to "Study on the incidence of drug resistance for new admissions" reported by TB Sanatorium Council in 1992 in Japan, no increase of incidence was observed in either the first treatment or re-treatment as compared with former reports. However, U.S.A. study indicates a significant increase of MDR-TB, which is supposed to have been caused by a primary drug resistance prevailed in an infected area, or an acquired (secondary) drug resistance due to incomplete and/or faulty treatment for active TB. Many incidences were also reported for mass nosocomial infection of MDR-TB with HIV patients. In spite of these serious issues in U.S.A., MDR-TB has not yet been a major concern in Japan, while Japan should work out countermeasures in advance with careful observation of its tend. One of the causes of mass nosocomial infection of MDR-TB observed in U.S.A. is reported due to a delayed treatment after long procedures of TB identification and susceptibility tests followed specimen sampling. Rapid tests of identification and susceptibility for TB, MDR-TB in particular, are long expected. The introduction of recent molecular genetics technology will help to develop new rapid tests. While a relationship between drug resistance and TB gene is recently known to certain extent, total mechanism of TB resistance cannot be fully explained with only certain gene identified in the connection with drugs. Early treatment is critical for MDR-TB with HIV patient, as their prognosis is far worse than MDR-TB with non-HIV. Aside from HIV infection, very limited drugs are available for the treatment of MDR-TB. Drugs should be carefully selected based on the resistance patterns of each strain as well as its side effects anticipated.