Which agents should we use for the treatment of multidrug-resistant Mycobacterium tuberculosis? Giovanni Di Perri* and Stefano Bonora

Journal of Antimicrobial Chemotherapy (2004) 54, 593 602 DOI: 10.1093/jac/dkh377 Advance Access publication 28 July 2004 Which agents should we use for the treatment of multidrug-resistant Mycobacterium tuberculosis? Giovanni Di Perri* and Stefano Bonora JAC Department of Infectious Diseases, University of Torino, Amedeo di Savoia Hospital, Corso Svizzera 164, 10149 Turin, Italy Introduction The inappropriate treatment of drug-susceptible tuberculosis can lead to the selection and transmission of multidrug-resistant tuberculosis (MDR-TB), indicating resistance to at least isoniazid and rifampicin. In the treatment of MDR-TB, residual first-line drugs, such as ethambutol, pyrazinamide and streptomycin must be appropriately combined with additional second-line drugs, guided by individual susceptibility patterns. The clinical pharmacology of these second-line antituberculous drugs is reviewed. Fluoroquinolones represent the only substantial therapeutic advance in the last 20 years. Many factors potentially affect the outcome of MDR-TB. Treatment adherence, prior exposure to antituberculous drugs, the number of drugs to which the infection is still susceptible and the time since the first diagnosis of tuberculosis are the most relevant. The management of MDR-TB requires considerable expertise. When initiating or revising therapy for MDR-TB, the process of selecting drugs should rely on prior treatment history, results of susceptibility testing and an evaluation of the patient s adherence. In making drug selection, we propose to follow a hierarchy based on the intrinsic activity against Mycobacterium tuberculosis and the clinical evidence of efficacy of the available active compounds. Keywords: M. tuberculosis, multidrug-resistant tuberculosis, therapy, second-line agents Infection by Mycobacterium tuberculosis remains a leading cause of death. 1 It is estimated that nearly 9 10 6 new cases of active disease occur each year. 2 The vast majority of the world burden of tuberculosis (TB) is in developing countries, which is one of the main reasons why only 23% of the prevalent active cases are currently estimated to receive an appropriate antituberculous treatment. 1 Inappropriate treatment of drug-susceptible tuberculosis and its consequences on the susceptibility of M. tuberculosis to existing medications Along with unacceptably low cure rates and the continued spread of tuberculosis in the community, a major consequence of inappropriate treatment is the selection of M. tuberculosis isolates that are resistant to antituberculous drugs. 2 Basically two broad categories of inappropriate therapy may be identified: the wrong treatment prescribed (drugs, dosing intervals, duration) or an inadequate intake by patient. 3,4 Although treatment of drug-susceptible TB has changed little over the last 30 years, the incorrect initial prescription of an antituberculous regimen still represents a frequent occurrence. 5,6 A shortage of drugs has been one of the most common reasons accounting for the inadequacy of the initial anti-tb regimen, especially in resource-poor settings, 6 but an ignorance of basic principles of TB management is also a factor. More forgivable but equally harmful errors are made in the choice of drugs for the treatment of drug-resistant TB, often on the basis of an unreliable past medical history and/or in vitro testing of M. tuberculosis drug-susceptibility. 7 However, taking into consideration the high success rate of TB treatment under the DOT policy (Directly Observed Treatment), 8 the principal cause for the generation of drug-resistant TB generally appears to lie in the low degree of patient compliance with treatment. 9 Thus, inappropriate treatment here refers to the patient s decision to interrupt an otherwise effective regimen or to take just part of the regimen prescribed, 9 the latter being the worst situation in terms of probability of promoting the development of resistance. How multidrug resistance is defined in tuberculosis The widely adopted acronym MDR-TB (multidrug-resistant tuberculosis) indicates the presence of M. tuberculosis resistance to, at least, isoniazid and rifampicin, 5 the two fundamental... *Corresponding author. Tel: +39-011740388; Fax: +39-011744971; E-mail: di_perri@dealer.it... 593 JAC vol.54 no.3 q The British Society for Antimicrobial Chemotherapy 2004; all rights reserved.

components of any regimen for the treatment of drug-susceptible TB. Amongst the existing antituberculous drugs, isoniazid has the strongest early bactericidal action and thus significantly contributes to rapidly making patients non-infectious, whilst rifampicin has unique antibacterial properties against bacilli that are no longer in the active phase of replication. 2 The definition of MDR works as a sort of operational threshold in this setting, due to the major consequences that resistance to isoniazid and rifampicin has on the treatment of TB. 10 How resistance to antituberculous drugs is generated The suboptimal use of antituberculous medications creates a selective milieu in the host s tissues where the initially scanty drug-resistant mutants are able to replicate, eventually replacing the initially drug-susceptible M. tuberculosis population. 11 A predictable clinical consequence of these events is disease recrudescence due to drug-resistant bacteria and, unless properly treated, the likely evolution of such a case is towards chronicity. 2 The selection of drug-resistant M. tuberculosis depends on the frequency of the specific drug-resistant mutants in the initially drug-susceptible bacterial population. As a consequence, the chance of selecting such mutants is the highest in the case of monotherapy. 2 Whilst mutants resistant to a single drug may be fairly easily selected by monotherapy, the probability of selecting mutants that are resistant to multiple drugs decreases exponentially by increasing the number of drugs to which M. tuberculosis is simultaneously exposed. 2 The rationale for combination drug therapy was proven by a series of clinical demonstrations that provided unambiguous evidence of how the administration of multiple drugs bears a significantly lower chance of both disease recrudescence and selection of drugresistant strains compared with monotherapy. 2,4 Such clinical research eventually gave the current therapeutic strategy, which consists of three or four drugs in the initial phase of therapy, followed by a consolidation two-drug phase once the initial bacterial biomass has been reduced to such an extent that the chance of selecting residual drug-resistant mutants is exceedingly low. 2 Major differences between the treatment of drug-susceptible and drug-resistant tuberculosis Unlike most bacterial infections, for which several options of comparable efficacy are often available when drug resistance occurs, the cure rate of drug-resistant TB is almost invariably decreased compared with the standard treatment of drug-susceptible disease. 10 Furthermore, MDR-TB requires a two- to fourfold longer period of treatment compared with drug-susceptible TB. 10,12 Since with even the shortest treatment course (6 months) so far validated for drug-susceptible TB, 12 most of the problems from which drug-resistance originates are related to the length of treatment (especially considering tolerability and adherence), the longer time that is required to treat MDR-TB clearly implies an additional risk of poor treatment adherence and, consequently, of treatment failure. 13 Two other major issues significantly contribute to the higher complexity of the treatment of MDR-TB compared to drug-susceptible disease: the increased cost (up to 100 times higher) and the higher frequency of adverse reactions. 10,12 In terms of treatment duration, the greatest difference between drug-susceptible and MDR-TB is the lack of a real substitute for rifampicin, the crucial drug in short-course TB chemotherapy. 3 The introduction of rifampicin allowed the duration of the length of antituberculous treatment to be reduced from 18 24 months down to the currently accepted standard of 6 9 months. 4 A major property that has been attributed to rifampicin is its ability to affect dormant bacilli. 14 Whilst at the outset of anti-tb treatment, the vast majority of organisms are actively replicating and are therefore susceptible to anti-mycobacterial agents, in the subsequent phases of therapy, the residual M. tuberculosis population switches to a virtually inactive metabolic status which make these organisms poorly susceptible to drugs. 2,4 It is thought that these inactive bacilli undergo periodic metabolic reactivation and that during these short periods of activity rifampicin is able to exert its action on the bacterial RNA polymerase, whilst these brief periods of time are insufficient for the other otherwise equally effective bactericidal agents to produce any appreciable effect. 15 Treating MDR-TB: essential clinical pharmacology of second-line antituberculous drugs By definition, chemotherapy of MDR-TB cannot rely upon isoniazid and rifampicin, the two most powerful drugs for the treatment of tuberculosis. Depending on the individual susceptibility pattern, residual first-line drugs must be appropriately combined with additional second-line drugs (Table 1). Streptomycin, other aminoglycosides and capreomycin In clinical practice, an important feature of aminoglycosides and capreomycin is the need for parenteral administration. 16,17 Aminoglycosides exert their effects by binding to the 30S subunit of bacterial ribosomes, thus leading to reduced mrna reading and impaired protein synthesis. 18 These drugs must penetrate into mycobacteria in order to reach their molecular targets. At low ph values, as in cavities or abscesses, 19 drug penetration through bacterial porins is limited and this could be one of the reasons accounting for the clinical ineffectiveness of aminoglycosides as single anti-tb agents. 19 Moreover, these molecules do not penetrate mammalian cells and therefore lack efficacy against intracellular bacteria. 18,19 Aminoglycosides are bactericidal only against rapidly dividing mycobacteria, and have little or no activity against bacilli which are not replicating, 18 such as those persisting for long periods in the stationary phase of growth. This is the pharmacodynamic rationale of using aminoglycosides only in the induction phase, when a large number of rapidly multiplying bacilli at the extracellular level are present, whilst in the maintenance phase agents active against intraphagocytic and slowly dividing mycobacteria are needed. 2 Another crucial issue related to long-term administration of aminoglycosides is toxicity. Ototoxicity and nephrotoxicity are well recognized as dose-related adverse effects of aminoglycosides. 20,21 Amikacin is reputed to be less vestibulotoxic and nephrotoxic than streptomycin and kanamycin and this could be advantageous in clinical practice, especially for long treatment periods. 22 The toxicity profile of capreomycin is similar to that of aminoglycosides, including nephrotoxicity. 23 A distinct clinical entity associated with capreomycin is a form of renal tubulopathy characterized by ion losses with resultant alkalosis. 23 Resistance to streptomycin is more common in those areas where the drug has been more widely used. 24 Ribosomal 594

Table 1. Formulations, dosing, type of activity and evidence of clinical evaluation of agents available for MDR-TB treatment Drug Formulation Average daily dose Type of antimycobacterial activity Evaluation in clinical studies Pyrazinamide tablet (500 mg) 15 30 mg/g 1000 2000 mg daily bactericidal at acid ph + Ethambutol tablet (100, 400 mg) 15 30 mg/kg 800 1600 mg daily bacteriostatic + Streptomycin 1 g vial, iv or im 20 40 mg/kg (1 g) once daily bactericidal against exponential + injection phase bacilli Amikacin/kanamycin 500 mg and 1 g vial, 15 30 mg/kg (1 g) once daily bactericidal against exponential + iv or im injection phase bacilli Capreomycin 1 g vial, iv or im 15 30 mg/kg (1 g) once daily bactericidal against exponential + injection phase bacilli p-aminosalicylic acid granules (4 g packets); 8 12 g/day in two or three doses bacteriostatic + tablet 500 mg Cycloserine capsule (250 mg) 10 15 mg/kg usually 500 750 mg bacteriostatic + in two doses Ethionamide/ tablet (250 mg) 15 20 mg/kg usually bactericidal + prothionamide 500 750 mg/day in single daily or two divided doses Ciprofloxacin tablet (500 and 750 mg) 500 1000 mg daily weakly bactericidal + Ofloxacin tablet (400 mg) 400 800 daily weakly bactericidal + Levofloxacin tablet (500 mg) 500 1000 daily bactericidal + Gatifloxacin tablet (400 mg) 400 mg daily bactericidal Moxifloxacin tablet (400 mg) 400 mg daily bactericidal Co-amoxiclav tablet (1000/250 mg) 3000/750 mg in three doses bactericidal against exponential +/ phase bacilli Clarithromycin tablet (500 mg) 1000 mg bacteriostatic Linezolid tablet and vial (600 mg) 1200 mg in two doses bacteriostatic im, intramuscular; iv, intravenous. Data based on Crofton et al. 10 and Blumberg et al. 48 modifications due to mutations in the region of the S12 interaction with 16S rrna, and particularly changes of the codon 43, are probably the main mechanisms of resistance, 25 although other mechanisms are not excluded. The binding site of each of the aminoglycosides may be different, therefore the ribosomal mutations that mediate resistance to aminoglycosides are likely to be drug-specific. 26 Amikacin is generally active against streptomycin-resistant strains of M. tuberculosis, 27 whilst crossresistance with kanamycin is the rule. 26 On the other hand, strains resistant to amikacin are generally also resistant to streptomycin. Capreomycin, although very expensive for countries with limited resources, is also potentially active against streptomycin-resistant strains. 26 Capreomycin-resistant strains are not usually resistant to amikacin, while the inverse seems to be the case for low-level, but not for high-level, amikacin resistance. 26 p-aminosalicylic acid (PAS) Discovered in the 1940s, the antimycobacterial agent PAS was considered to be a first-line agent, 28 in combination with isoniazid and streptomycin, until it was replaced by ethambutol in the early 1960s. 29 PAS exerts a bacteriostatic effect on M. tuberculosis by competitively blocking the conversion of para-aminobenzoic acid into folic acid. 30 PAS is usually given orally. The granular formulation now available is more easily administered and better tolerated than the original tablet formulation (20 24 tablets per day). 31 However, side effects related to PAS are frequent and include gastrointestinal symptoms, hypersensitivity reactions (up to 10%), hypothyroidism, thrombocytopaenia and intestinal malabsorption. 31 Other side effects include neuritis, convulsion, dizziness and gynaecomastia. Interestingly, isoniazid-resistant bacilli are usually susceptible to these thioamides, although they share the same parent compound, isonicotinic acid. 33 Cross-resistance is complete between ethionamide and prothionamide. 32 Thioamides Following the discovery of isoniazid, numerous pyridine derivatives were tested, with ethionamide and prothionamide being shown to have antimycobacterial activity. 32 The mechanism of action is, like isoniazid, at the level of synthesis of mycolic acids. 33 Both the drugs are bactericidal in vitro but resistance can rapidly emerge. 32 The usual dosage is 500 1000 mg per day in two doses. The most important adverse drug events are gastrointestinal disturbance and hepatotoxicity (hepatitis in 4.3% of patients), with ethionamide slightly less toxic than prothionamide. 34,35 Cycloserine Cycloserine exerts an antimycobacterial bacteriostatic effect by competitively blocking two metabolic steps of the biosynthesis of the bacterial cell wall. 36,37 Clinical studies in the 1950s showed decreased efficacy compared with PAS, and severe dose-related neuropsychiatric toxicity. 38 The latter is frequent (up to 50% of patients at the dose of 1 g/day) and includes convulsive seizures, psychotic episodes, slurred speech, drowsiness and coma. 39 Smaller and divided doses reduce the frequency of adverse events. 40 The oral dose used currently is 250 mg twice or three times a day, and therapeutic drug monitoring is advocated so as not to exceed plasma levels of 30 ng/ml. 40 595

Rifamycins other than rifampicin Rifabutin has considerable cross-resistance with rifampicin, with less than 15% of M. tuberculosis strains resistant to rifampicin retaining susceptibility to rifabutin. 41 In this minor proportion of cases, some of the mutations selected by rifampicin do not modify the RNA polymerase sufficiently as to render this protein resistant to rifabutin. 42 However, the proportion of discordant resistance is too low to make rifabutin a generally useful drug in rifampicin-resistant disease, and no clinical studies have addressed this issue. The pattern and the mechanism of resistance of rifapentine is identical to that of rifampicin. 43 Fluoroquinolones Fluoroquinolones inhibit topoisomerase II (DNA gyrase) of M. tuberculosis. 44 The other molecular target of fluoroquinolones, topoisomerase IV, is absent in M. tuberculosis. 44 A notable property of fluoroquinolones relates to their ability to penetrate into macrophages and to exert intracellular mycobactericidal activity. 45 Although the activity of fluoroquinolones against M. tuberculosis was already evident from the early pre-clinical in vitro screening, 46 their use in the treatment of TB has never been formally pursued by the manufacturers for overt commercial reasons. Most clinical experience has been accumulated with first-generation fluoroquinolones, ofloxacin and ciprofloxacin, 47 which are currently approved as second-line agents for MDR-TB by the WHO, American Thoracic Society, and Centers for Disease Control. 10,48 However, the number of clinical studies examining the role of fluoroquinolones for MDR-TB is limited. Further to the clinical data available on ciprofloxacin 49 53 and ofloxacin, there are few reports regarding the use of sparfloxacin and levofloxacin. 54,55 In most cases, ofloxacin was given at a dosage of 400 mg once a day and ciprofloxacin 500 750 mg twice daily. However, more recently higher dosages have been used such as 800 mg ofloxacin and 1000 mg ciprofloxacin, both once daily. 53,55 In a comparative trial, 56 levofloxacin (which is the S enantiomer of ofloxacin) was found to be more efficacious than ofloxacin when incorporated into regimens used for treatment of MDR-TB. Adverse effects related to the administration of fluoroquinolones include gastrointestinal disturbances, neurological effects, arthropathy and photosensitivity. 44,57 Promising data have been reported for the newer compounds, gatifloxacin 58,59 and moxifloxacin. 60 62 The latter agents have significantly lower MICs for M. tuberculosis than older fluoroquinolones and better pharmacodynamic correlates (Table 2). Sparfloxacin was also shown to be six to eight times more active than ofloxacin (although less active than moxifloxacin and gatifloxacin), but serious photosensitivity limited its clinical use. 63 The pharmacodynamic activity of fluoroquinolones is concentration-dependent. Parameters such as the peak serum concentration (C max ) and the area under the concentration time curve (AUC) to the MIC therefore have to be considered as the best pharmacodynamic correlates of antimycobacterial efficacy. 64 According to these parameters (Table 2), moxifloxacin at the recommended daily dose of 400 mg seems to be the most active fluoroquinolone against M. tuberculosis. 65,66 In patients with active tuberculosis, moxifloxacin was shown to be similar to rifampicin in terms of early bactericidal activity (EBA). 67 Moreover, in a murine model of tuberculosis, a moxifloxacin-containing regimen was found to be effective in a Table 2. Pharmacokinetic/pharmacodynamic characteristics of fluoroquinolones as agents against M. tuberculosis Fluoroquinolone MIC (mg/l) C max (mg/l) C max /MIC AUC/MIC Ciprofloxacin 0.5 4.0 1.5 1 2 10 20 Ofloxacin 1.0 2.0 4.0 2 24 Levofloxacin 1.0 6.21 5 7 40 50 Sparfloxacin 0.2 0.5 1.18 2 40 Gatifloxacin 0.2 0.25 3.42 8.4 68 Moxifloxacin 0.12 0.5 4.34 9 96 Data based on Ginsburg et al. 69 and Berning. 44 shorter time compared with standard therapy. 68 There is currently very limited clinical experience in the use of moxifloxacin or the other newer fluoroquinolones as antituberculous agents, and they are currently not licensed for TB, but should these promising experimental findings be confirmed in field trials then this could represent a significant advance in the treatment of both TB and MDR-TB. The selection of fluoroquinolone-resistant M. tuberculosis has been observed in vivo. 69,70 Mutations of gyra and gyrb genes are required for high-level resistance, 44 although alternative resistance mechanisms, including energy-dependent active efflux, are probably involved. 69 Complete cross-resistance within the class is the currently accepted rule. 69 Potential advantages of high doses of newer compounds against low-level resistance to older fluoroquinolones has to be confirmed in further clinical evaluations. 69 However, decreased susceptibility to fluoroquinolones can develop after short courses of empirical therapy (e.g. 10 15 days), and the risk of selecting resistance before a diagnosis of tuberculosis is established has been described, particularly in HIVpositive subjects. 69,70 Linezolid Linezolid, an agent of the oxazolidinone class mainly used to treat infections caused by Gram-positive bacteria, has good antimycobacterial activity. 71 It appears to inhibit protein synthesis by binding to the 50S ribosomal subunit. 72 The literature currently available is very limited and no clinical data have yet been reported. Clarithromycin This macrolide has a spectrum of activity that includes mycobacteria, 73 and it is frequently employed for the treatment of Mycobacterium avium complex infection. Concentrations which are effective against M. tuberculosis seem to be well above those achievable in serum and lung tissue of patients (MIC 90 >128 ng/ml) 74,75 In animal models, however, clarithromycin showed an antituberculous activity, likely to result from its high intracellular accumulation. 76 Interestingly, in vitro subinhibitory concentrations of cell wall inhibitors were shown to reverse intrinsic M. tuberculosis resistance to clarithromycin, 77 probably by allowing a higher intracellular penetration of the drug. Moreover, clarithromycin was shown to restore, by an unknown mechanism, the in vitro antimycobacterial activity of first-line compounds (isoniazid, rifampicin and ethambutol) 596

against MDR strains. 78 However, no clinical data on the use of clarithromycin in the treatment of MDR-TB are available. b-lactams Co-amoxiclav and ampicillin/sulbactam have in vitro activity against M. tuberculosis. 79 The b-lactamase inhibitor is essential to overcome mycobacterial b-lactamase hydrolysis and to allow penetration of the aminopenicillin through the cell wall. 80,81 The EBA of co-amoxiclav was reported to be comparable to that of ofloxacin, supporting a potential role in the clinical setting. 82 However, bactericidal activity against only exponential-phase and not stationary-phase bacilli suggested a potential clinical usefulness only in the early stages of treatment or as a supportive agent to prevent the selection of resistance against companion drugs. 81 The anecdotal efficacy of co-amoxiclav-containing regimens for treatment of MDR-TB has been described, but definitive clinical evaluation has not yet been carried out. 83,84 Clofazimine Clofazimine, a riminophanazine used against Mycobacterium leprae and M. avium infections, is active against M. tuberculosis. 85 Clofazimine concentrates in macrophages and was reported to be efficacious in a murine model of tuberculosis. 85 No clinical data on its use for treating MDR-TB are currently available. Phenothiazines These compounds are currently used in the management of psychosis. Chlorpromazine has a titrable activity in the inhibition of the growth of intracellular M. tuberculosis in vitro. 86 Thioridazine has also been shown to be active against MDR M. tuberculosis, 87 although the cardiac safety profile is still under investigation. Phenothiazines have not yet been tested for their antimycobacterial activity in humans. Nitroimidazopyrans Newer compounds of this class (which includes metronidazole) were found to display a considerable activity against M. tuberculosis, 88 through the inhibition of both protein and lipid synthesis. They also exert a bactericidal effect against bacilli in the stationary phase. A recently synthesized nitroimidazopyran was shown to act in vitro against MDR M. tuberculosis, 89 and its clinical development seems warranted. Other compounds A series of chemically heterogeneous molecules are undergoing in vitro and in vivo testing, but very limited information on their development is currently available. Tuberactinomycin 90 is a polypeptide active against amikacin-resistant strains that was found to be better tolerated than capreomycin in a preliminary clinical assessment in East Asia. Acetamides 91 belong to a new class of compounds with specific and promising antimycobacterial activity. Pyrrole derivatives 92 were shown to have bactericidal activity against MDR M. tuberculosis strains. Outcome of MDR-TB Most of the data concerning the therapeutic outcome of MDR- TB come from the experience of the last two decades. Following the resurgence of TB in the 1980s, 93,94 most of the attention devoted to MDR-TB was initially focused on its association with HIV infection, 95 particularly following well publicized episodes of inter-human spread among subjects with AIDS. 96,97 The outcome of MDR-TB in the era before the introduction of highly active antiretroviral therapy for HIV infection was found to be very poor, with median survival times barely above 2 months. However, most of these patients were very immunosuppressed at the time of diagnosis of MDR-TB and the latter probably had a relatively minor role in determining outcome. The available knowledge on the outcome of MDR-TB therefore comes from a relatively small number of case series. 98 102 Most of these studies had a retrospective design and there are virtually no randomized clinical studies comparing different regimens for the treatment of MDR-TB. As a consequence, the choice of drug regimens is still mainly based on the application at the individual level of several chemotherapeutic principles rather than upon the results of clinical trials. The heterogeneity of MDR-TB cases is a further complicating factor. Nevertheless, certain variables that influence outcome can be identified. Two major case series of MDR-TB outside the setting of HIV infection have been reported in some detail, 103,104 respectively describing the outcome of 171 and 26 patients. In the study by Goble et al., 103 171 MDR-TB patients with a median age of 46 years who had received a median of six drugs before being retreated, and shed bacilli found to be resistant to a median of six drugs, had an overall response rate of 56%, while in the study by Telzak et al., 104 26 MDR-TB patients with a median age of 37 years and a prior exposure to a median of 3.5 drugs (only 35% had previously received some form of anti-tb treatment) had a 96% rate of treatment success. Taken together, these two studies give an informative picture of the major factors that determine the chances of curing MDR-TB. A lower prior exposure to anti-tb drugs, a higher number of anti-tb drugs to which the individual infection is still susceptible and a shorter time since the first TB diagnosis (indicating a less advanced disease) indicate a greater chance of successful treatment response. These findings have been confirmed by other studies, such as the French national survey by Flament-Saillour et al. 105 Park et al. 106 demonstrated that carefully selected regimens (preferably including four drugs to which the infection was proven to be fully susceptible) led to relatively high cure rates in a series of 107 Korean MDR-TB patients shedding bacilli initially resistant to a mean of four drugs. The importance of the number of antituberculous drugs to which MDR-TB isolates are still susceptible is also emphasized by the reports by Salomon et al., 107 Turett et al., 108 and Park et al., 109 who found that the prompt institution of appropriately selected regimens may even significantly improve the short-term outcome of MDR-TB in patients with HIV infection. In Mexico, Perez-Guzman et al. 110 showed how in particularly favourable conditions, such as with limited baseline resistance and reliable laboratory information on drug susceptibility, MDR-TB in HIV-negative patients may even respond to regimens as short as 12 months. Principles and practice in the management of MDR-TB The current performance of anti-mdr-tb treatment worldwide falls far short of what could be achieved by the most effective use of the existing diagnostic and therapeutic weapons. 597

Nevertheless, MDR-TB may still be considered as a curable disease. 111 The successful management of MDR-TB requires specialized expertise. There are three major issues to be considered: diagnostic techniques, drugs and adherence. A basic starting point is the access to a good diagnostic service. Although most microbiology departments are able to carry out basic microscopy and isolation of mycobacteria, the expertise required to carry out dependable drug-susceptibility testing is variable. When testing for second-line drugs is considered, the number of centres able to provide reliable results on drug susceptibility of M. tuberculosis is reduced to just one or two per country. In these centres, rapid methods to reduce time of diagnosis of MDR-TB and of susceptibility testing could play a crucial role. The radiometric BACTEC 460 technique has been proven to be reliable for second-line drug-susceptibility testing compared with the standard proportion method with solid media, 112 whilst newer fully automated non-radiometric culture methods (e.g. MGIT 960, MB/BACT) have been evaluated only for first-line antituberculous agents. 113,114 Among tools for rapid detection of resistance, a reverse hybridization-based probe assay (INNO-LiPA Rif TB) is available for detection of mutations in the rpob gene for resistance to rifampicin, 114,115 whilst many other amplification-based methods have been proposed for resistance to isoniazid, aminoglycosides and fluoroquinolones. 116,117 When initiating or revising therapy, the general rule is never to add a single drug to a failing regimen: this will simply result in additional resistance. At least three, but preferably four or five, previously unused drugs whose in vitro activity is proven should be administered. In designing a regimen we should not aim to keep drugs in reserve: that is the way to lose the last battle. With adequate information, the choice of an anti-mdr- TB regimen becomes a stepwise process, with preference being given to the residual first-line agents shown to be still active, such as pyrazinamide, streptomycin and ethambutol. Resistance to one of the antituberculous aminoglycosides, most often streptomycin, generally still allows for the selection of another compound from this class. Depending on the local resources, a parenterally administered drug such as amikacin, capreomycin or kanamycin could potentially be included, in association with second-line oral agents (fluoroquinolones, ethionamide, PAS, cycloserine, clarithromycin, co-amoxiclav, linezolid) (Table 3). Table 3. Suggested regimens for patients with different patterns of MDR-TB Pattern of drug resistance Suggested regimen In making the selection of the latter, we propose a hierarchy (Table 4) based on their intrinsic activity against M. tuberculosis and clinical evidence of efficacy. Treatment duration has to be determined on an individual basis, but, as a general rule, it should be continued for at least 18 months after sputum conversion. 10 Second-line drugs are typically less effective and have more side effects than first-line agents. According to baseline M. tuberculosis drug susceptibility, withdrawal of one or more drugs may be possible during treatment without compromising the outcome, but bacteriostatic and not bactericidal agents should be preferentially interrupted, unless intolerable side effects are the reason prompting a regimen revision (e.g. aminoglycosides, cycloserine). Adherence is a major issue. Apart from the case of primary MDR-TB, a lack of treatment adherence in the past is typically the factor that has caused the patient to have MDR-TB. If proper action is not taken to ensure patient compliance, the chance of further treatment failure and increased resistance is high. 118 Referral to adherence-promoting TB services is therefore important. Hospital-based treatment is advisable at least until sputum conversion occurs. After discharge, DOT should be given, particularly for cases of acquired resistance and prior evidence of non-adherence. In the bacteriological and clinical monitoring of response to therapy, it must be anticipated that improvement of MDR-TB is usually slower than with drug-susceptible TB. However, in most case series, patients who eventually became culture negative converted sputum sample after 2 3 months of 98 110 therapy. The role of surgery in the management of patients with extensive pulmonary disease has not been established in randomized studies. However, in some case series, patients with severe MDR-TB appeared to benefit from resection of damaged lung tissue, 118 120 especially when extensive and longstanding fibrosis is present and treatment failure occurred with a large panel of drug resistance. The DOTs Plus Strategy Based on an awareness of the multiple difficulties faced by any large-scale intervention aimed at fighting MDR-TB, a special initiative has been launched within the framework of the Global Minimum duration (months) b Comments Isoniazid and rifampicin (± streptomycin) EMB, PZA, FQN, AMK 18 in extensive diseases, an additional agent may be added Isoniazid, rifampicin and ethambutol PZA, FQN, AMK plus 2 a 18 surgery should be considered (± streptomycin) Isoniazid, rifampicin and pyrazinamide EMB, FQN, AMK plus 2 a 18 24 surgery should be considered (± streptomycin) Isoniazid, rifampicin, ethambutol and pyrazinamide (± streptomycin) FQN, AMK plus 3 a 18 24 surgery should be considered Data based on Crofton et al. 10 and Blumberg et al. 48 EMB, ethambutol; PZA, pyrazinamide; FQN, fluoroquinolones (most experience involves ciprofloxacin, ofloxacin and levofloxacin; moxifloxacin and gatifloxacin have not been yet evaluated in clinical trials); AMK, amikacin (capreomycin may be an alternative in case of resistance). a Additional agents: ethionamide, PAS, cycloserine, b-lactams, clarithromycin, linezolid, clofazimine. b After sputum conversion. 598

Table 4. Proposed ranking of second-line drugs based on intrinsic antimycobacterial activity and clinical evidence of efficacy Drug Levofloxacin Aminoglycosides/capreomycin Ethionamide/prothionamide Ofloxacin/ciprofloxacin PAS Cycloserine b-lactams Clarithromycin, linezolid, clofazimine TB Strategy. 121 This newly conceived programme consists of a comprehensive approach including the major DOTs principles but technically devoted to the intensive diagnostic and therapeutic management of MDR-TB. 121,122 DOTs Plus relies upon a working DOTs Strategy within the activities of the local TB control programme. 121,122 Based on the knowledge of the prevailing resistance patterns in the community, on the individual drug history and on the availability of second-line antituberculous drugs, an initial regimen is chosen (including an injectable drug) and then the patient undergoes monthly monitoring by means of sputum microscopy and culture. Reliance on a reference laboratory able to provide appropriate information on the susceptibility of individual isolates of M. tuberculosis is also an operational requirement, in order to allow for timely changes in the regimen initially selected and for updating the information on the resistance status of M. tuberculosis strains in the community. References Rank 1. Frieden, T. H., Sterling, T. H., Munsiff, S. S. et al. (2003). Tuberculosis. Lancet 362, 887 99. 2. Schluger, N. W., Harkin, T. J. & Rom, W. N. (1996). Principles of therapy of tuberculosis in the modern era. In Tuberculosis, 1st edn (Rom, W. N. & Garay, S., Eds), pp. 751 61. Little, Brown and Company, New York, NY, USA. 3. Iseman, M. D. & Madsen, L. A. (1989). Drug-resistant tuberculosis. Clinics in Chest Medicine 10, 341 53. 4. Rieder, H. A. (2002). Interventions for Tuberculosis Control, 1st edn. International Union Against Tuberculosis and Lung Disease, Paris, France. 5. Espinal, M. A. (2003). The global situation of MDR-TB. Tuberculosis 83, 44 51. 6. Mwinga, A. (2001). Drug resistant tuberculosis in Africa. Annals of the New York Academy of Sciences 953, 106 12. 7. Raviglione, M. C., Gupta, R., Dye, C. M. et al. (2001). The burden of drug-resistant tuberculosis and mechanism for its control. Annals of the New York Academy of Sciences 953, 185 91. 8. Ormerod, L. P. (1999). Directly observed therapy (DOT) for tuberculosis: why, when, how and if? Thorax 54, S42 5. 9. Volmink, J. & Graner, P. (1997). Systematic review of randomised controlled strategies to promote adherence to tuberculosis treatment. British Medical Journal 315, 1403 6. 10. Crofton, J., Chaulet, P. & Maher, D. (1997). Guidelines for the Management of Drug-resistant Tuberculosis, 1st edn. World Health Organization, Geneva, Switzerland. I II III IV V VI No clinical data are currently available for newer fluoroquinolones (gatifloxacin and moxifloxacin). 11. David, H. L. (1970). Probability of distribution of drug-resistant mutants in unselected populations of Mycobacterium tuberculosis. Applied Microbiology 20, 810 4. 12. Chan, E. D. & Iseman, M. D. (2002). Current medical treatment for tuberculosis. British Medical Journal 325, 1282 6. 13. Drobniewski, F. A. & Balabanova, Y. M. (2002). The diagnosis and management of multiple-drug-resistant tuberculosis at the beginning of the new millennium. International Journal of Infectious Diseases 6, S21 31. 14. Dickinson, J. M. & Mitchinson, D. A. (1981). Experimental models to explain the high sterilizing activity of rifampicin in the chemotherapy of tuberculosis. American Review of Respiratory Disease 123, 367 71. 15. Mitchinson, D. A. (2000). Role of individual drugs in the chemotherapy of tuberculosis. International Journal of Tuberculosis and Lung Disease 9, 796 806. 16. Marshall, E. K. (1948). The absorption, distribution and excretion of streptomycin. Journal of Pharmacology and Experimental Therapeutics 92, 43. 17. Doluisio, J. T., Dittert, D. T. & Lapiana, J. C. (1973). Pharmacokinetics of kanamycin following intramuscular administration. Journal of Pharmacokinetics and Biopharmaceutics 1, 253 65. 18. Heifets, L. & Lindholm-Levy, P. (1989). Comparison of bactericidal activities of streptomycin, amikacin, kanamycin, and capreomycin against Mycobacterium avium and M. tuberculosis. Antimicrobial Agents and Chemotherapy 33, 1298 301. 19. Damper, P. D. & Epstein, W. (1981). Role of the membrane potential in bacterial resistance to aminoglycoside antibiotics. Antimicrobial Agents and Chemotherapy 20, 803 8. 20. Smith, C. R., Moore, R. D. & Lietman, P. S. (1986). Studies of risk factors for aminoglycoside nephrotoxicity. American Journal of Kidney Diseases 8, 308 13. 21. Pfuetze, K. H., Pyle, M. M., Hishaw, H. C. et al. (1955). The first clinical trial of streptomycin in human tuberculosis. American Review of Tuberculosis 71, 752 4. 22. Allen, B. W., Mitchinson, D. A., Chan, Y. C. et al. (1983). Amikacin in the treatment of pulmonary tuberculosis. Tubercle 64, 111 8. 23. Hesling, C. M. (1969). Treatment with capreomycin, with special reference to toxic effects. Tubercle 50, 39 41. 24. Cohn, D. L., Bustreo, F. & Raviglione, M. C. (1997). Drugresistant tuberculosis: a review of the worldwide situation and the WHO/IUATLD Global Surveillance Project. International Union Against Tuberculosis and Lung Diseases. Clinical Infectious Diseases 24, 121 30. 25. Finken, M., Kirschner, P., Meier, A. et al. (1993). Molecular basis of streptomycin resistance in Mycobacterium tuberculosis: alterations of the ribosomal protein S12 gene and point mutations within a functional 16S ribosomal RNA pseudoknot. Molecular Microbiology 9, 1239 46. 26. McClatchy, J. K., Kanes, W., Davidson, P. T. et al. (1977). Cross-resistance in M. tuberculosis to kanamycin, capreomycin, and viomycin. Tubercle 58, 29 34. 27. Hoffner, S. E. & Kallenius, G. (1988). Susceptibility of streptomycin-resistant Mycobacterium tuberculosis strains to amikacin. European Journal of Clinical Microbiology and Infectious Diseases 7, 188 90. 28. British Medical Research Council. (1950). Treatment of pulmonary tuberculosis with streptomycin and para-aminosalicylic acid. A Medical Research Council investigation. British Medical Journal 2, 1073 85. 29. British Medical Research Council. (1973). Co-operative controlled trial of a standard regimen of streptomycin, PAS and isoniazid and three alternative regimens of chemotherapy in Britain. Tubercle 54, 99 129. 30. Peloquin, C. A., Berning, S. E., Huitt, G. A. et al. (1999). Oncedaily and twice-daily dosing of p -aminosalicylic acid granules. American Journal of Respiratory and Critical Care Medicine 159, 932 4. 599

31. Akhtar, A. J., Crompton, G. K. & Schonell, M. E. (1968). Paraaminosalicylic acid as a cause of intestinal malabsorption. Tubercle 49, 328 31. 32. Steenken, W. & Montabine, V. (1960). The antituberculous activity of thioamide in vitro and in the experimental animal. American Review of Respiratory Disease 81, 761 3. 33. Hamilton, E., Eidus, L. & Little, E. A. (1962). A comparative study in vivo of isoniazid and a-ethylthioisonicotinamide. American Review of Respiratory Disease 85, 407 12. 34. Pernod, J. (1965). Hepatic tolerance of ethionamide. American Review of Respiratory Disease 92, 39 42. 35. Research Committee of the British Tuberculosis Association. (1968). A comparison of the toxicity of prothionamide and ethionamide. Tubercle 49, 125 34. 36. Steenken, W. & Wolinsky, E. (1956). Cycloserine: antituberculous activity in vitro and in the experimental animal. American Review of Respiratory Disease 56, 196 9. 37. David, H. L., Takayama, K. & Goldman, D. S. (1969). Susceptibility of mycobacterial D-alanyl-D-alanine synthetase to D-cycloserine. American Review of Respiratory Disease 100, 579 81. 38. Storey, P. B. (1960). A comparison of isoniazid-cycloserine with isoniazid-pas in the therapy of cavitary pulmonary tuberculosis. American Review of Respiratory Disease 81, 868 79. 39. Helmy, B. (1970). Side effects of cycloserine. Scandinavian Journal of Infectious Diseases 71, S220 5. 40. Bucco, T., Meligrana, G. & De Luca, V. (1970). Neurotoxic effects of cycloserine therapy in pulmonary tuberculosis of adolescents and young adults. Scandinavian Journal of Infectious Diseases 71, S259 65. 41. Hong Kong Chest Service, British Medical Research Council. (1992). A controlled study of rifabutin and an uncontrolled study of ofloxacin in the retreatment of patients with pulmonary tuberculosis resistant to isoniazid, streptomycin and rifampicin. Tubercle and Lung Disease 73, 59 67. 42. Sintchenko, V., Chew, W. K., Jelfs, P. et al. (1999). Mutations in rpob gene and rifabutin susceptibility of multidrug-resistant Mycobacterium tuberculosis strains isolated in Australia. Pathology 31, 257 60. 43. Jarvis, B. & Lamb, H. M. (1998). Rifapentine. Drugs 56, 607 16. 44. Berning, S. E. (2001). The role of fluoroquinolones in tuberculosis today. Drugs 61, 9 18. 45. Van Caekenberghe, D. (1990). Comparative in-vitro activities of ten fluoroquinolones against Mycobacterium spp. Journal of Antimicrobial Chemotherapy 26, 381 6. 46. Rastogi, N., Lambrousse, V. & Goh, K. S. (1996). In vitro activities of fourteen antimicrobial agents against drug susceptible and resistant clinical isolates of Mycobacterium tuberculosis and comparative intracellular activities against the virulent H37Rv strain in human macrophages. Current Microbiology 33, 167 75. 47. Alangaden, G. J. & Lerber, S. A. (1997). The clinical use of fluoroquinolones for the treatment of mycobacterial diseases. Clinical Infectious Diseases 25, 1213 21. 48. Blumberg, H. M., Burman, W. J., Chaisson, R. E. et al. (2003). American Thoracic Society/Centers for Disease Control and Prevention/ Infectious Diseases Society of America: treatment of tuberculosis. American Journal of Respiratory and Critical Care Medicine 167, 603 62. 49. Sirgel, F. A., Botha, F. J., Parkin, D. P. et al. (1997). The early bactericidal activity of ciprofloxacin in patients with pulmonary tuberculosis. American Journal of Respiratory and Critical Care Medicine 156, 901 5. 50. Mangunnerogoro, H. & Hudoyo, A. (1999). Efficacy of low-dose ofloxacin in the treatment of multidrug-resistant tuberculosis in Indonesia. Chemotherapy 45, S19 S35. 51. Sahoo, R. C. (1993). Ofloxacin in the retreatment of patients with pulmonary tuberculosis resistant to isoniazid, streptomycin and rifampicin, a South Indian experience. Tubercle and Lung Disease 74, 140 1. 52. Berning, S. E., Madsen, L., Iseman, M. D. et al. (1995). Longterm safety of ofloxacin and ciprofloxacin in the treatment of mycobacterial infection. American Journal of Respiratory and Critical Care Medicine 151, 2006 9. 53. Bergstermann, H. & Ruchardt, A. (1997). Ciprofloxacin once daily versus twice daily for the treatment of pulmonary tuberculosis. Infection 25, 227 32. 54. Yew, W. W., Chan, C. K., Chau, C. H. et al. (2000). Outcomes of patients with multidrug-resistant tuberculosis treated with ofloxacin/ levofloxacin-containing regimens. Chest 117, 744 51. 55. Peloquin, C. A., Berning, S. E., Huitt, G. A. et al. (1998). Levofloxacin for drug resistant Mycobacterium tuberculosis. Annals of Pharmacotherapy 32, 268. 56. Yew, W. W., Chan, C. K., Leung, C. C. et al. (2003). Comparative roles of levofloxacin and ofloxacin in the treatment of multidrug-resistant tuberculosis. Chest 124, 1476 81. 57. Hooper, D. C. & Wolfson, J. S. (1985). The fluoroquinolones: pharmacology, clinical uses, and toxicities in humans. Antimicrobial Agents and Chemotherapy 28, 716 21. 58. Zhao, B. Y., Pine, R. & Domagala, J. (1999). Fluoroquinolone action against clinical isolates of Mycobacterium tuberculosis: effects of a C-8 methoxyl group on survival in liquid media and in human macrophages. Antimicrobial Agents and Chemotherapy 43, 661 6. 59. Tomiota, H., Sato, K., Akaki, T. et al. (1999). Comparative in vitro activities of the newly synthesized quinolone HSR-903, sitafloxacin (DU-6859a), gatifloxacin (AM-1155), and levofloxacin against Mycobacterium tuberculosis and Mycobacterium avium complex. Antimicrobial Agents and Chemotherapy 43, 3001 4. 60. Ji, B., Lounis, N., Maslo, C. et al. (1998). In vitro and in vivo activities of moxifloxacin and clinafloxacin against Mycobacterium tuberculosis. Antimicrobial Agents and Chemotherapy 42, 2066 9. 61. Gillespie, S. H. & Billington, O. (1999). Activity of moxifloxacin against mycobacteria. Journal of Antimicrobial Chemotherapy 44, 393 5. 62. Rodriguez, J. C., Riuz, M., Lopez, M. et al. (2002). In vitro activity of moxifloxacin, levofloxacin, gatifloxacin, and linezolid against Mycobacterium tuberculosis. International Journal of Antimicrobial Agents 20, 464 7. 63. Lubasch, A., Erbes, R., Mauch, H. et al. (2001). Sparfloxacin in the treatment of drug resistant tuberculosis or intolerance of first line therapy. European Respiratory Journal 17, 641 6. 64. Wright, D. H., Brown, G. H., Peterson, M. L. et al. (2000). Application of fluoroquinolone pharmacodynamics. Journal of Antimicrobial Chemotherapy 46, 669 83. 65. Lubasch, A., Keller, J. & Borner, M. (2000). Comparative pharmacokinetics of ciprofloxacin, gatifloxacin, grepafloxacin, levofloxacin, trovafloxacin, and moxifloxacin after single dose administration in healthy volunteers. Antimicrobial Agents and Chemotherapy 44, 2600 3. 66. Schentag, J. J., Gilliland, K. K. & Paladino, J. A. (2001). What have we learned from pharmacokinetic and pharmacodynamic theories? Clinical Infectious Diseases 32, S39 46. 67. Gosling, R. D., Uiso, L. O., Sam, N. E. et al. (2003). The bactericidal activity of moxifloxacin in patients with pulmonary tuberculosis. American Journal of Respiratory and Critical Care Medicine 168, 1342 5. 68. Nuermberger, E. L., Yoshimatsu, T., Tyagi, S. et al. (2004). Moxifloxacin-containing regimen greatly reduces time to culture conversion in murine tuberculosis. American Journal of Respiratory and Critical Care Medicine 169, 421 6. 69. Ginsburg, A. S., Grosset, J. H. & Bishai, W. R. (2003). Fluoroquinolones, tuberculosis, and resistance. Lancet Infectious Diseases 3, 432 42. 70. Sullivan, E. A., Kreiswirth, B. N., Palumbo, L. et al. (1995). Emergence of fluoroquinolone-resistant tuberculosis in New York City. Lancet 345, 1148 50. 600

71. Cynamon, M. H., Klemens, S. P., Sharpe, C. A. et al. (1999). Activities of several novel oxazolidinones against M. tuberculosis in a murine model. Antimicrobial Agents and Chemotherapy 43, 1189 91. 72. Rodriguez Diaz, J. C., Ruiz, M., Lopez, M. et al. (2003). Synergic activity of fluoroquinolones and linezolid against Mycobacterium tuberculosis. International Journal of Antimicrobial Agents 21, 354 6. 73. Luna-Herrera, J., Reddy, V. M., Daneluzzi, D. et al. (1995). Antituberculous activity of clarithromycin. Antimicrobial Agents and Chemotherapy 39, 2992 5. 74. Truffot-Pernot, C., Lounis, N., Grosset, J. H. et al. (1995). Clarithromycin is inactive against Mycobacterium tuberculosis. Antimicrobial Agents and Chemotherapy 39, 2827 8. 75. Hoffner, S. V., Gezelius, L. & Olsson-Liljiequist, B. (1997). In vitro activity of fluorinated quinolones and macrolides against drug-resistant Mycobacterium tuberculosis. Journal of Antimicrobial Chemotherapy 40, 885 8. 76. Mor, N. & Esfandiari, A. (1997). Synergistic activities of clarithromycin and pyrazinamide against Mycobacterium tuberculosis in human macrophages. Antimicrobial Agents and Chemotherapy 41, 2035 6. 77. Bosne-David, S., Barros, V., Cabo Verde, S. et al. (2000). Intrinsic resistance of Mycobacterium tuberculosis to clarithromycin is effectively reversed by subinhibitory concentrations of cell wall inhibitors. Journal of Antimicrobial Chemotherapy 46, 391 5. 78. Cavalieri, S. J., Biehle, J. R. & Sandres, W. E. (1995). Synergistic activities of clarithromycin and antituberculous drugs against multidrug-resistant Mycobacterium tuberculosis. Antimicrobial Agents and Chemotherapy 39, 1542 5. 79. Cynamon, M. H. & Palmer, G. S. (1983). In vitro activity of amoxicillin in combination with clavulanic acid against Mycobacterium tuberculosis. Antimicrobial Agents and Chemotherapy 24, 429 31. 80. Kwon, H. H., Tomioka, H. & Saito, H. (1995). Distribution and characterization of b-lactamases of mycobacteria and related organisms. Tubercle and Lung Disease 76, 141 8. 81. Abate, G. & Miorner, H. (1998). Susceptibility of multidrugresistant strains of Mycobacterium tuberculosis to amoxycillin in combination with clavulanic acid and ethambutol. Journal of Antimicrobial Chemotherapy 42, 735 40. 82. Chambers, H. F., Kocagoz, T., Sipit, T. et al. (1998). Activity of amoxicillin/clavulanate in patients with tuberculosis. Clinical Infectious Diseases 26, 874 7. 83. Nakagawa, Y., Shimazu, K., Ebihara, M. et al. (1999). A study of b-lactamase activity of mycobacteria and clinical trial of penicillin/ b-lactamase inhibitor combinations in the treatment of drug-resistant Mycobacterium tuberculosis. Kekkaku 74, 447 52. 84. Nadler, J. P., Berger, J., Nord, J. A. et al. (1991). Amoxicillinclavulanic acid for treating drug resistant Mycobacterium tuberculosis. Chest 99, 1025 6. 85. Jagannath, C., Reddy, M. V., Kailasam, S. et al. (1995). Chemotherapeutic activity of clofazimine and its analogue against Mycobacterium tuberculosis. American Journal of Respiratory and Critical Care Medicine 151, 1083 6. 86. Amaral, L. & Kristiansen, J. E. (2000). Phenothiazines: an alternative to conventional therapy for the initial management of suspected multidrug resistant tuberculosis. A call for studies. International Journal of Antimicrobial Agents 14, 173 6. 87. Ordway, D., Viveiros, M., Leandro, C. et al. (2003). Clinical concentrations of thioridazine kill intracellular multidrug-resistant Mycobacterium tuberculosis. Antimicrobial Agents and Chemotherapy 47, 917 22. 88. Stover, C. K., Warrener, P., VanDevanter, D. R. et al. (2000). A small-molecule nitroimidazopyran drug candidate for the treatment of tuberculosis. Nature 405, 962 6. 89. Murugasu-Oei, B. & Dick, T. (2000). Bactericidal activity of nitrofurans against growing and dormant Mycobacterium bovis BCG. Journal of Antimicrobial Chemotherapy 46, 917 9. 90. Toyohara, M., Nagata, A., Havano, K. et al. (1969). Study of the antituberculous activity of tuberactinomycin, a new antimicrobial drug. American Review of Respiratory Disease 100, 228 30. 91. Jones, P. B., Parrish, N. M., Houston T. A. et al. (2000). A new class of antituberculous agents. Journal of Medicinal Chemistry 43, 3304 14. 92. Deidda, D., Lampis, G., Fioravanti, R. et al. (1998). Bactericidal activities of the pyrrole derivative BM212 against multidrug-resistant and intramacrophagic Mycobacterium tuberculosis strains. Antimicrobial Agents and Chemotherapy 42, 3035 7. 93. Di Perri, G., Cruciani, M., Danzi, M. C. et al. (1989). Nosocomial epidemic of active tuberculosis among HIV-infected patients. Lancet 2, 1502 4. 94. Di Perri, G., Vento, S., Cruciani, M. et al. (1991). Tuberculosis and HIV infection. New England Journal of Medicine 325, 1882 3. 95. Di Perri, G., Vento, S., Cazzadori, A. et al. (1995). Multidrugresistance in patients with AIDS. Archives of Internal Medicine 155, 1452 3. 96. Busillo, C. P., Lessnau, K. D., Sanjana, V. et al. (1992). Multidrug resistant Mycobacterium tuberculosis in patients with human immunodeficiency virus infection. Chest 102, 797 801. 97. Moro, M. L., Gori, A., Errante, I. et al. (1998). An outbreak of MDR tuberculosis involving HIV-infected patients of two hospitals in Milan, Italy. Italian Multidrug-resistant Tuberculosis Outbreak Study Group. AIDS 12, 1095 102. 98. Narita, M., Alonso, P., Lauzardo, M. et al. (2001). Treatment experience of multidrug-resistant tuberculosis in Florida, 1994 1997. Chest 120, 343 8. 99. Tahoglu, K., Torun, T., Sevim, T. et al. (2001). The treatment of multidrug resistant tuberculosis in Turkey. New England Journal of Medicine 345, 170 4. 100. Subbash, H. S., Ashwin, I., Jesudason, M. V. et al. (2003). Clinical characteristics and treatment response among patients with multidrug-resistant tuberculosis: a retrospective study. Indian Journal of Chest Diseases and Allied Sciences 45, 97. 101. Mitnick, C., Bayona, J., Palacios, E. et al. (2003). Communitybased therapy for multidrug-resistant tuberculosis in Lima, Peru. New England Journal of Medicine 348, 119 28. 102. Kim, H. J., Hong, Y. P., Kim, S. J. et al. (2001). Ambulatory treatment of multidrug-resistant pulmonary tuberculosis patients at a chest clinic. International Journal of Tuberculosis and Lung Disease 5, 1129 36. 103. Goble, M., Iseman, M. D., Madsen, L. A. et al. (1993). Treatment of 171 patients with pulmonary tuberculosis resistant to isoniazid and rifampicin. New England Journal of Medicine 328, 527 32. 104. Telzak, E. E., Sepkowitz, K., Alpert, P. et al. (1995). Multidrugresistant tuberculosis in patients without HIV infection. New England Journal of Medicine 333, 907 11. 105. Flament-Saillour, M., Robert, J., Jarlier, V. et al. (1999). Outcome of multidrug-resistant tuberculosis in France. A nationwide case-control study. American Journal of Respiratory and Critical Care Medicine 160, 587 93. 106. Park, S. K., Kim, C. T. & Song, S. D. (1998). Outcome of therapy in 107 patients with pulmonary tuberculosis resistant to isoniazid and rifampicin. International Journal of Tuberculosis and Lung Disease 2, 877 84. 107. Salomon, N., Perlman, D. C., Friedmann, P. et al. (1995). Predictors and outcome of multidrug-resistant tuberculosis. Clinical Infectious Diseases 21, 1245 52. 108. Turett, G. S., Telzak, E. E., Torian, L. V. et al. (1995). Improved outcomes for patients with multidrug-resistant tuberculosis. Clinical Infectious Diseases 21, 1238 44. 109. Park, M. M., Davis, A. L., Schluger, N. W. et al. (1996). Outcome of MDR-TB patients, 1983-93. Prolonged survival with appropriate therapy. American Journal of Respiratory and Critical Care Medicine 153, 317 24. 601