Setting: National Masan Tuberculosis Hospital, Masan, South Korea, a 430-bed tertiary referral hospital specializing in tuberculosis.
Objective: To evaluate the treatment outcomes of standardized, empiric regimens for multidrug-resistant tuberculosis (MDR-TB).
Design: A retrospective analysis of the hospital records of 142 patients with MDR-TB who had failed short-course chemotherapy. Between 1 January 1998 and 30 June 2000, patients were started on one of two standardized, empiric regimens based on previous treatment history. Drug susceptibility testing of the infecting strain was not used to modify the treatment regimen. Treatment was continued for at least 18 months after conversion to a negative culture.
Results: Sixty-three patients (44.1%) were cured and discharged from treatment after at least 18 months of negative cultures; 18 (12.7%) failed treatment, 41 (28.9%) defaulted, four died (2.8%), and 15 (10.6%) were transferred to another institution. One patient is still on treatment. Resistance to ofloxacin was the only risk factor related to poor outcome (death or failure) in univariate or multiple logistic regression analysis.
Conclusions: High levels of resistance to second-line drugs are likely a cause of poor outcome of MDR-TB therapy in Korea. Directly observed therapy and other methods to increase patient compliance should be considered nationwide, as they may improve MDR-TB treatment outcomes.