Patients with MDR-TB receiving either levofloxacin or moxifloxacin along with other second-line anti-TB medication from January 2002 through December 2008 were included. Patients were treated in one of three hospitals affiliated with Seoul National University College of Medicine in Korea: Seoul National University Hospital, Seoul Metropolitan Government Seoul National University Boramae Medical Center, or Seoul National University Bundang Hospital. Mycobacterium tuberculosis was identified by sputum culture, and all patients showed resistance to at least isoniazid and rifampicins in in vitro drug-susceptibility testing. Patients < 18 years of age, those treated with both levofloxacin and moxifloxacin for TB, and patients who received less than 3 months of levofloxacin or moxifloxacin were excluded from analysis. This study was approved by the ethics review committees of all three hospitals. Previous studies by our group were also based on part of this population of patients with MDR-TB [12-15].

The choice of FQNs was based on the preference of the attending physician, according to the drug-susceptibility test results. The other combined drugs included aminoglycosides, prothionamide, cycloserine, pyrazinamide, rifabutin, ethambutol, p-aminosalicylic acid, amoxicillin-clavulanic acid, clarithromycin, sulfamethoxazole, and trimethoprim. Drugs used before the diagnosis of MDR-TB were not included in the analysis.

Records were reviewed for age, gender, body mass index (BMI), underlying comorbidities, smoking history, family history of TB, primary drug resistance, laboratory test results, radiographic findings, FQN dose, the combination of extrapulmonary TB, nontuberculous mycobacteria (NTM) colonization, other anti-TB medications, duration of treatment, results of drug-susceptibility testing, and adverse events.

Treatment outcomes were classified into the following groups in accordance with the suggested criteria of Laserson et al. [16]: cure, treatment completed, failure, death, default, and transferred out. Additionally, if patients were diagnosed with bacteriologically confirmed MDR-TB after being cured or after treatment was completed, they were considered relapse cases. Based on these classifications, treatment outcomes were further categorized into treatment success (cured and treatment completed) or adverse treatment outcome (death, failure, and relapse) to identify predictors of poor treatment response.

Extensively drug-resistant tuberculosis (XDR-TB) was defined as laboratory-confirmed resistance to all of the following: isoniazid, rifampins, any FQNs, and second-line injectable agents such as capreomycin, kanamycin, and amikacin. Although some patients with a poor response to treatment had subsequent resistance testing performed, they were classified according to the initial drug-susceptibility test results.

Data are expressed as median values with interquartile ranges (IQRs) or means ± standard deviations. The demographic characteristics, laboratory results, radiographic findings, and treatment outcomes were compared between the levofloxacin group and the moxifloxacin group using Pearson's χ² test or Fisher's exact test for categorical variables and Student's t test for continuous variables. To understand the impact of choice between the levofloxacin and moxifloxacin group on treatment outcomes, we compared selected clinical variables between treatment success and failure through a univariate comparison and subsequent multiple logistic regression. All statistical analyses were performed with SPSS version 11.0 (SPSS Inc., Chicago, IL, USA).

Between January 2002 and December 2005, 171 patients received either levofloxacin or moxifloxacin to treat MDR-TB. In total, 123 patients (71.9%) were treated with levofloxacin and 48 patients (28.1%) received moxifloxacin. One-hundred nine patients were from Seoul National University Hospital, 26 were from Seoul Metropolitan Government Seoul National University Boramae Medical Center, and 36 patients were from Seoul National University Bundang Hospital. Except for one 38-year-old Chinese man, all 170 patients were Korean.

All subjects had radiographic and bacteriologic evidence of pulmonary TB, and 20.5% had combined extrapulmonary involvement. The most common extrapulmonary TB was tuberculous lymphadenopathy found in 17 patients (10.5%). The median age of the 171 patients was 42 years (IQR, 28 to 52). Ninety-six patients (56.1%) were male, and 84 patients (49.4%) had a history of TB treatment. Among the previously treated patients, 64.3% had been treated once, 28.6% twice, and 7.1% more than three times. No significant difference was found between patients treated with levofloxacin or moxifloxacin in terms of BMI, smoking history, family history of TB, primary drug resistance, proportion of XDR-TB, and radiographic findings (Table 1). However, serum creatinine was higher in the moxifloxacin group (0.9 ± 0.2 vs. 1.0 ± 0.2; p = 0.025) than in the levofloxacin group. Coinfection with the human immunodeficiency virus was rare, occurring in only one patient in the levofloxacin group. NTMs colonization was found in 20 patients (11.7%), and the most common NTM identified was Mycobacterium abscessus.

The median duration of treatment was 594 days (IQR, 481 to 772) in the levofloxacin group and 673 days (IQR, 530 to 778) in the moxifloxacin group. The usual prescribed daily dose of moxifloxacin was 400 mg, and the daily dose of levofloxacin varied from 300 to 1,000 mg. Sixty-eight patients (39.8%) were hospitalized at the initiation of treatment, and 20 patients had at least one surgical resection of a diseased lung as an adjunctive treatment for TB.

A median of five drugs (range, 4 to 6) were used in the MDR-TB treatment, and this was similar between the two groups (p = 0.244). However, the number of susceptible drugs used was significantly lower in the moxifloxacin group (4 vs. 5; p = 0.048) than in the levofloxacin group. Furthermore, according to the drug-susceptibility test results, the number of resistant drugs (4 vs. 5; p < 0.001) and the number of patients with ofloxacin resistance (14.6% vs. 35.4%; p = 0.005) was significantly higher in the moxifloxacin group than in the levofloxacin group. The use of ethambutol was more common in the levofloxacin group than in the moxifloxacin group (43.9% vs. 22.9%; p = 0.011) and the use of amoxicillin-clavulanate (24.4% vs. 45.8%; p = 0.006) was more common in the moxifloxacin group than the levofloxacin group (Table 2). The use of other TB drugs and the rates of adverse drug reactions were not different between the two groups. Isoniazid was used in 7.6% and rifamycins were used in 6.4% of the patients after being diagnosed with multidrug resistance. Among 171 patients, 137 (80.1%) were considered treatment successes and 16 (9.4%) were considered treatment failures. Eighteen patients (10.5%) were classified as default or transfer-out (Table 3). Treatment success was achieved among 97 patients (78.9%) in the levofloxacin group and 40 (83.3%) in the moxifloxacin group. Based on the variables included in the univariate comparison between the treatment success and failure groups, the final multiple logistic regression model included serum creatinine level, resistance to ofloxacin, and the number of susceptible drugs used. The patients in the moxifloxacin group were not more likely to have treatment success than those in the levofloxacin group (adjusted odds ratio, 0.76; 95% CI, 0.24 to 2.43; p = 0.65; Table 4).