After the introduction of anti-tubercular drugs, pulmonary tuberculosis (TB) became a medical disease. The treatment outcomes for drug-susceptible pulmonary TB are highly favorable if drug compliance is good. However, the incidence and degree of TB drug resistance are increasing worldwide. Recently, a definition of extensively drug-resistant TB (XDR-TB) was formally adopted, and XDR-TB patients have been found across the globe. The drugs used to treat drug-resistant TB, particularly multidrug-resistant tuberculosis (MDR-TB), are highly toxic, costly, and require treatment durations of up to 2 years. Significantly, these treatments often lead to disappointing outcomes.

In order to increase treatment success rates, the availability of second-line drugs and education and social support systems are important. To further increase success rates, a rapid and accurate diagnostic method for drug-resistant TB is needed, along with the development of new drugs with novel mechanisms of action that can be given over shorter durations. In this review, we discuss the current state of medical treatment for pulmonary MDR-TB.

Anti-TB agents have traditionally been classified as first- or second-line drugs. However, as the incidence of MDR-TB has increased, these agents have been categorized into 5 groups to assist with drug selection for MDR-TB [1].

Group 1 includes first-line oral drugs such as rifabutin. Group 2 comprises injectable agents including streptomycin (traditionally included as a first-line drug) and kanamycin, capreomycin, and amikacin (previously regarded as second-line drugs). Group 3 consists of fluoroquinolones. However, among these agents, ciprofloxacin is no longer considered to be an anti-TB drug and gatifloxacin is no longer used because of the risk of dysglycemic adverse events. Moxifloxacin and levofloxacin are the newer fluoroquinolones in this group. In an earlier retrospective analysis of 106 patients with MDR-TB, the treatment success rate was higher for the levofloxacin-containing regimen (90.0%) than the ofloxacin-containing regimen(79.7%) [2]. In another retrospective comparison of levofloxacin and moxifloxacin in MDR-TB, both drugs showed equivalent efficacy [3]. In a prospective study of these agents, the choice of levofloxacin or moxifloxacin for the treatment of MDR-TB did not affect sputum culture conversion rates after 3 months of treatment [4]. Group 4 drugs include the conventional second-line anti-TB drugs: prothionamide, cycloserine, and para-aminosalicylic acid. Group 5 includes drugs with unproven efficacy in MDR-TB. These drugs can be used only for patients for whom a regimen with confirmed anti-TB activity is not available.

Updated guidelines now recommend that 4 or more drugs with definite or highly-probable effectiveness against TB should be included in treatment regimens [5]. The selection of TB drugs is structured such that treatment options proceed from group 1 through to group 5. Among the group 1 drugs, pyrazinamide or ethambutol can be selected for MDR-TB cases. However, the World Health Organization (WHO) does not recommend ethambutol, citing its limited efficacy against MDR-TB [5]. The efficacy of rifabutin for MDR-TB is unclear, and this agent is thus not routinely included in MDR-TB regimens.

Having selected the group 1 drugs, 1 injectable group 2 drug should then be selected as part of the regimen. The concomitant use of 2 or more injectable drugs is not recommended. In addition, the WHO does not recommend streptomycin for MDR-TB because streptomycin-resistance rates are high, as is the frequency of ototoxicity, compared with other injectable drugs [5]. Capreomycin is not available in Korea and can only be purchased from the Korea Orphan Drug Center. Capreomycin is also very expensive. Kanamycin and amikacin have nearly 100% cross-resistance. Amikacin is usually injected intravenously (even though intramuscular injection is possible), and therefore kanamycin is the drug of choice for MDR-TB in Korea. The WHO recommends kanamycin as the first injectable drug for MDR-TB [5].

Among group 3 drugs, the newer fluoroquinolones are recommended for MDR-TB, with moxifloxacin or high-dose levofloxacin typically being the drugs of choice. A recent Korean randomized study has revealed that moxifloxacin and levofloxacin have comparable efficacy in terms of 3-month culture conversion rates and adverse events [4]. The concomitant use of 2 or more fluoroquinolones is not recommended.

At this stage in the regimen design, only 3 or fewer agents have been selected (pyrazinamide, 1 injectable drug or one fluoroquinolone). Hence, group 4 drugs should also be included in the regimen. In Korea, the available group 4 drugs include prothionamide, cycloserine,and para-aminosalicylic acid. These can be selected on the basis of drug susceptibility results, previous treatment history, and the resistance pattern in the region. Prothionamide is the preferred choice, followed by cycloserine then para-aminosalicylic acid [5]. The WHO has recommended an empirical regimen for MDR-TB comprising pyrazinamide, 1 injectable agent (kanamycin), 1 fluoroquinolone (moxifloxacin or levofloxacin), prothionamide, and either cycloserine or para-aminosalicylic acid (if cycloserine cannot be used). Ethambutol maybe used but is not included in the count of effective drugs [5].

An appropriate regimen cannot be formulated for some MDR-TB patients on the basis of previous anti-TB treatment history, drug susceptibility results, and drug adverse events. In these cases, group 5 drugs should be included in the regimen. Because the efficacy of group 5 drugs is unclear, more than 2 of these agents can be included.

The most important and effective second-line drugs for MDR-TB are the fluoroquinolones, and the most potent and well-known group 5 drug, linezolid, is therefore recommended in cases of fluoroquinolone-resistant MDR-TB or XDR-TB.

Traditionally, when solid media have been used for culture and drug susceptibility tests (DST), the DST results only became available 3-4 months after anti-TB treatment initiation. In Korea, DST panels include both first- and second-line drugs. Hence, clinicians can select anti-TB drugs according to DST results as well as the previous anti-TB treatment history. However, the WHO has recently recommended the use of commercial line probe assays for the rapid detection of MDR-TB [6], and clinicians can therefore obtain the DST results for isoniazid and rifampicin within 2-3 days without knowing the results for second-line drugs. In this situation, clinicians should formulate an anti-TB regimen on the basis of previous history of anti-TB treatment, the local area DST pattern, or the index case DST pattern. If the previous history of anti-TB treatment is not available, the empirical regimen should comprise pyrazinamide, 1 newer fluoroquinolone, 1 second-line injectable drug, prothionamide, and cycloserine or para-aminosalicylic acid. In cases of rifampicin mono-resistance detected using a rapid molecular test, the chance of MDR-TB development still exists (because the sensitivity of isoniazid resistance detection by molecular testing is lower than that of rifampicin resistance detection). Thus, rifampicin mono-resistant cases should be regarded as MDR-TB. In these instances, clinicians can add isoniazid to the regimen, but it should not be counted as an effective drug.

With regard to the molecular line probe assay to detect second-line TB drug resistance, too few data on the direct testing of sputum specimens are available to develop policy guidance on its use. Hence, a WHO expert group concluded that because our knowledge of the degree of cross-resistance between second-line injectables for TB is incomplete, the MTBDRsl assay (Hain Lifescience, Nehren, Germany) cannot be used to identify individual drugs to be used for treatment [7].

Among the group 5 drugs, linezolid and clofazimine are worthy of mention. Linezolid is a member of the oxazolidinone family of antibiotics and inhibits protein synthesis [20]. It is active against a variety of gram-positive organisms, anaerobes, and atypical microbes. Linezolid has also been found to be active against M. tuberculosis in vitro, in animal studies, and in some clinical trials. The usual effective anti-bacterial dose of linezolid is 600 mg twice daily over a maximum recommended duration of 28 days. The problem with its use in anti-TB treatment is the risk of the development of long-term adverse events: the most frequent is peripheral neuropathy, with optic neuropathy and cytopenia also being reported. A recent prospective trial has shown that linezolid is effective for achieving culture conversion among patients with treatment-refractory XDR-TB [21]. Most patients (34 of 39 [87%]) in that study had a negative sputum culture within 6 months of linezolid being included in their treatment regimen. However, of the 38 patients who received linezolid, 31 (82%) had clinically significant adverse events that were possibly or probably related to this antibiotic, and 3 patients discontinued therapy as a result. Patients in this earlier study cohort who received a 300 mg/day linezolid dose after a second randomization had fewer adverse events than those who continued with the 600 mg/day schedule [21]. In another report, treatment of intractable MDR/XDR-TB with linezolid at a daily dose of 300 mg was found to be effective (favorable treatment outcomes, including treatment successor still on treatment after culture conversion, occurred in 78% of cases), and was associated with fewer neuropathic side effects than daily dosages of 600 mg/day or 1200 mg/day [22].

Clofazimine is a fat-soluble riminophenazine dye used in the treatment of leprosy [23]. However, even though clofazimine is listed as a group 5 drug for MDR-TB, its efficacy against this disease is uncertain. In a murine model of MDR-TB, a clofazimine-containing regimen achieved a higher degree of culture conversion after 5 months of second-line drug treatment than a control regimen [24]. Furthermore, in a report from Bangladesh, short-course regimens containing clofazimine showed high efficacy in the treatment of MDR-TB [25]. Skin discoloration is the principal adverse event associated with this antibiotic, and this reaction recovers very slowly owing to clofazimine's long half-life (approximately 70 days) [26]. The optimal dose of clofazimine is not yet established, but in previous clinical studies, it has been used at 100 mg/day in the majority of patients (84.1%). In a Korean study, 32 patients with MDR-TB and additional resistance to ofloxacin (11 with XDR-TB) were treated with clofazimine-containing regimens [27]. The treatment success rate in this series was 48.4% (15/31), and multivariate analysis revealed that both male sex and linezolid use were significant indicators of treatment success. Only 4 patients in this cohort achieved treatment success without linezolid or surgical resection. Nine patients stopped receiving clofazimine owing to adverse events, including skin discoloration (n = 3), hepatotoxicity (n = 3), and gastrointestinal disturbance (n = 3). The authors of that study concluded that clofazimine may be an effective treatment for MDR-TB but that concurrent or subsequent use of linezolid may contribute to treatment success.

Six drugs with antimicrobial activity (phenothiazine, metronidazole, doxycycline, disulfiram, tigecycline, and co-trimoxazole) are not included in the WHO guidelines for MDR-TB treatment but have been proposed as potential candidates for evaluation against M. tuberculosis [28]. The optimal doses are not yet known, however (Table 1) [1].

Traditionally, the recommended treatment duration for MDR-TB has been 18 to 24 months, including a period of 12-18 months after culture conversion. According to a recent meta-analysis of 9,153 MDR-TB patients, an intensive phase of at least 8 months and a total duration of at least 20 months is recommended in patients without any previous MDR-TB treatment [5]. However, such durations present problems in terms of costs and adverse events. Recently, Van Deun et al [25]. reported successful treatment outcomes in an MDR-TB cohort from Bangladesh using a 9-month regimen. In that study, 206 patients with MDR-TB who had not been treated previously with second-line drugs were enrolled. They received an intensive treatment phase of at least 4 months' duration involving gatifloxacin, clofazimine, ethambutol, pyrazinamide, kanamycin, and high-dose isoniazid, and a 5-month maintenance phase with gatifloxacin, clofazimine, ethambutol, and pyrazinamide. The relapse-free cure rate was 87.9% in this patient series. Several subsequent studies have also reported successful treatment outcomes for MDR-TB with 9-12-month short-course treatments.

The STREAM trial (Standardized Treatment Regimen of Anti-TB drugs for patients with MDR-TB) has been initiated, and will test a 9-month MDR-TB treatment regimen. In the aforementioned Bangladesh study, the drug resistance rates for ofloxacin and kanamycin were 10.3% and 0%, respectively; the rates were slightly higher (16.6% and 13.3%, respectively) among 1,407 patients with MDR-TB in a Korean series [29]. Recent guidelines for MDR-TB therapy published by the International Union Against Tuberculosis and Lung Disease recommend the 9-month Bangladesh regimen for MDR-TB cases showing susceptibility to fluoroquinolones and injectable drugs [30].

Theoretically, patients with XDR-TB are resistant to at least 1 fluoroquinolone and 1 second-line injectable drug. However, our knowledge of the cross-resistance patterns between drugs in a given class is not complete. The acquisition of resistance to streptomycin and second-line injectable drugs is derived from different gene mutations. Streptomycin susceptibility was found to be an important predictor of favorable long-term survival in patients with pre-XDR-TB [31, 32]. Hence, other second-line injectable drugs or streptomycin could be selected in TB regimens. Because newer generation fluoroquinolones may be effective against ofloxacin-resistant strains [19], these agents can also be selected for the treatment of XDR-TB if there is no previous history of treatment failure with this class of drug. Currently, the most potent drug in group 5 is linezolid, and it should thus always be included in XDR-TB regimens (clofazimine should be included when linezolid is not available). Surgical resection should also be considered to maximize the effects of drug treatment. Two Korean studies have reported treatment success rates for XDR-TB of 67% and 53.5%, respectively [33, 34].

In December 2012, the US FDA approved diarylquinoline as a new drug for MDR-TB [35]. Delamanid is already in phase II trials [36]. However, as anti-TB drugs are given in combination to overcome drug resistance, further studies of optimal companion drugs are needed. A 2013 pipeline report by the Treatment Action Group has highlighted drugs with anti-TB activity from various clinical studies, and these are listed in Table 2 [37].

Even though treatment success rates for MDR-TB are currently disappointing, the appropriate and considered use of current TB drugs can achieve 70-80% treatment success rates in MDR-TB. In selected patients with MDR-TB, rates can even be increased up to 80-90% by combining surgical resection with medications. In intractable cases of MDR-TB and in patients with XDR-TB, group 5 drugs such as linezolid and clofazimine may improve treatment success rates. It is hoped that new drugs with novel mechanisms of action, and that are effective even short durations of therapy, may significantly ameliorate the worldwide TB epidemic.