Drug susceptibility testing of nontuberculous mycobacteria.

Review For reprint orders, please contact: [email protected]

Drug susceptibility testing of nontuberculous mycobacteria

Jakko van Ingen*,1 & Ed J Kuijper2

Abstract Diseases caused by nontuberculous mycobacteria are emerging in many settings. With an increased number of patients needing treatment, the role of drug susceptibility testing is again in the spotlight. This articles covers the history and methodology of drug susceptibility tests for nontuberculous mycobacteria, but focuses on the correlations between in vitro drug susceptibility, pharmacokinetics and in vivo outcomes of treatment. Among slow-growing nontuberculous mycobacteria, clear correlations have been established for macrolides and amikacin (Mycobacterium avium complex) and for rifampicin (Mycobacterium kansasii). Among rapid-growing mycobacteria, correlations have been established in extrapulmonary disease for aminoglycosides, cefoxitin and co-trimoxazole. In pulmonary disease, correlations are less clear and outcomes of treatment are generally poor, especially for Mycobacterium abscessus. The clinical significance of inducible resistance to macrolides among rapid growers is an important topic. The true role of drug susceptibility testing for nontuberculous mycobacteria still needs to be addressed, preferably within clinical trials. Nontuberculous mycobacteria (NTM; synonyms: environmental mycobacteria, mycobacteria other than TB) are emerging pathogens in many settings [1,2] . NTM are environmental organisms that are present in soil and both natural as well as treated water. The environment is the supposed source of human infections, although exact niches remain elusive [3,4] . Most NTM infections are opportunistic in nature and affect patients with impaired local or systemic immunity. Pulmonary NTM disease is the most frequent disease manifestation and has three distinct presentations: TB-like fibrocavitary disease in elderly patients with a history of chronic pulmonary disease; nodular bronchiectatic disease in elderly patients (mainly women; sometimes referred to as Lady Windermere syndrome) without a significant pulmonary history but with phenotypic features that suggest an underlying predisposing condition; and hypersensitivity pneumonitis in patients exposed by aerosols containing NTM (‘hot tub lung’). The most frequent extrapulmonary NTM diseases are lymphadenitis of the cervicofacial nodes in children and localized skin infections by Mycobacterium marinum, Mycobacterium ulcerans or rapidly growing mycobacteria (RGM). Rarely, NTM can cause other extrapulmonary or disseminated infections, typically in the severely immunocompromised [4] . The clinically most frequently encountered species and their associated diseases are detailed in Table 1; currently recommended treatment regimens are detailed in Table 2 [4,5] . The incidence rates of NTM diseases have risenin many settings. In Queensland, Australia, where NTM disease is a reportable condition, the incidence of notified cases of clinically significant

Keywords

• amikacin • clarithromycin • drug susceptibility testing • Mycobacterium abscessus • Mycobacterium avium complex • Mycobacterium kansasii • nontuberculous mycobacteria • rifampicin • treatment

Department of Medical Microbiology, Radboud University Medical Center, PO Box 9101, 6500HB Nijmegen, The Netherlands Department of Medical Microbiology, Leiden University Medical Center, Leiden, The Netherlands *Author for correspondence: Tel.: +31 24 3614356; [email protected]

1 2

10.2217/FMB.14.60 © 2014 Future Medicine Ltd

Future Microbiol. (2014) 9(9), 1095–1110

part of

ISSN 1746-0913

1095

Review van Ingen & Kuijper Table 1. Ten most frequently isolated nontuberculous mycobacteria and recommended drug susceptibility testing practices. Growth rate

Species

Main sites of infection

CLSI-recommended DST platform

Slow

Mycobacterium avium complex (M. avium, Mycobacterium intracellulare, minor species) Mycobacterium kansasii

Pulmonary, lymph nodes

Broth macrodilution in 12B Broth microdilution medium in MH/7H9

CLA, MOX, LNZ, AMI

Pulmonary

Broth microdilution in MH

Mycobacterium xenopi

Pulmonary

Broth microdilution in MH†

Macrodilution, agar proportion Not established

Mycobacterium malmoense (north-western Europe) Mycobacterium simiae

Pulmonary


Not established

RIF, CLA (+ RIB, MOX, LNZ, AMI, STR, INH, EMB, CIP, SXT) RIF, CLA, RIB, MOX, LNZ, AMI, STR, INH, EMB, CIP, SXT RIF, CLA, RIB, MOX, LNZ, AMI, STR, INH, EMB, CIP, SXT

Pulmonary


Not established

Intermediate Rapid

Mycobacterium ulcerans Skin No recommendation (mainly west Africa) Mycobacterium marinum Skin Broth microdilution in MH Mycobacterium abscessus Pulmonary, skin Broth microdilution in MH Mycobacterium chelonae Skin, soft tissues Mycobacterium fortuitum Skin, soft tissues, pulmonary

Broth microdilution in MH Broth microdilution in MH

Alternative(s)

Key drugs to be tested

RIF, CLA, RIB, MOX, LNZ, AMI, STR, INH, EMB, CIP, SXT

No recommendation No recommendation RIF, EMB, CLA, DOX, MIN, SXT Not established CLA, AMI, FOX, IMI, LNZ, CIP, MOX, DOX, TIG, SXT Not established CLA, AMI, TOB, FOX, IMI, LNZ, CIP, MOX, DOX, SXT Not established CLA, AMI, TOB, FOX, IMI, LNZ, CIP, MOX, DOX, SXT

† M. xenopi grows poorly in this medium. AMI: Amikacin; CIP: Ciprofloxacin; CLA: Clarithromycin; CLSI: Clinical Laboratory Standards Institute; DOX: Doxycycline; DST: Drug susceptibility test; EMB: Ethambutol; FOX: Cefoxitin; IMI: Imipenem; INH: Isoniazid; LNZ: Linezolid; MH: Cation-adjusted Mueller–Hinton broth with oleic acid, albumin, dextrose and catalase supplement; MIN: Minocycline; MOX: Moxifloxacin; RIB: Rifabutin; RIF: Rifampicin; STR: Streptomycin; SXT: Co-trimoxazole; TIG: Tigecycline; TOB: Tobramycin. Data taken from [4].

disease rose from 2.2/100,000 in 1999 to 3.2/100,000 in 2005 [1] . In Europe, lower incidences have been recorded; in The Netherlands, the incidence has also increased and has been conservatively estimated at 1.7/100,000 in 2008 [7] . As a result, more patients now receive drug treatment for NTM diseases and more drug susceptibility tests (DSTs) are requested. Sadly, the outcomes of drug treatment remain poor, particularly in pulmonary NTM disease; prolonged culture conversion (i.e., persistently negative cultures during and after treatment) rates of 60–80% have been recorded for Mycobacterium avium complex (MAC) lung disease, but these rates are considerably lower (at 40–50%) for disease caused by Mycobacterium xenopi, Mycobacterium simiae or Mycobacterium abscessus [5] . Despite initial culture conversion, relapse and reinfection rates are considerable, although good-quality data here are limited. Natural (M. abscessus) or acquired resistance (macrolide-resistant MAC,

1096

Future Microbiol. (2014) 9(9)

rifampicin-resistant Mycobacterium kansasii) are important determinants of treatment outcomes [4,5] . This underlines the importance of providing DST results that can predict outcomes of treatment with the tested drugs. In addition, drug susceptibility patterns may also help to distinguish relapse from reinfection. The exact role of DSTs in the design of optimal treatment regimens has, however, not been settled [4] . Discrepancies between in vitro activity and in vivo outcomes of treatment abound, but the underlying reasons for this are not known, although they may be partly due to the current DST practices. Within this article, we describe the currently most used methods and the correlation between the in vitro and in vivo efficacy of tested drugs. In addition, we review the impact of recently emerging data on the pharmacokinetics of antimycobacterial drugs and mechanisms of drug resistance in NTM on DST practices and interpretation of results.

future science group

Drug susceptibility testing of nontuberculous mycobacteria Main methodologies In 1963, George Canetti and colleagues published the first consensus statement on drug susceptibility testing of mycobacteria [8] . Three procedures based on the dilution of the antimycobacterial drugs isoniazid, para-amino salicylic acid and streptomycin in Löwenstein–Jensen medium were described: the absolute concentration method, the resistance ratio method and the proportion method. Along with proposals for standardization of test methodologies, decision rules for interpretation of the results were also provided. Advice on the testing of other antimycobacterial drugs (kanamycin, cycloserine, viomycin, thioacetazone and ethionamide) was provided, but all statements refer to testing the Mycobacterium tuberculosis complex bacteria only, not NTM [8] . At the time, susceptibility testing of NTM mainly served identification purposes (i.e., to exclude the idea that a strain represented M. tuberculosis) [9] . The Löwenstein–Jensen medium was soon replaced by the 7H10 medium developed by Middlebrook and Cohn, on which more strains grow and growth is faster [10] . Over the following decades, a variety of methods for drug susceptibility testing of clinical NTM isolates has been tried. For RGM, test methods used in general bacteriology were adapted, whereas methods designed for M. tuberculosis were also applied to slowly growing mycobacteria (SGM). While several methods have been tried in research settings, very few have been used extensively in the clinical setting. Methods in

Review

current or past clinical use are detailed here and explained in Figure 1. ●●Absolute concentration, resistance ratio

& proportion methods

In the absolute concentration method, the MIC is determined by incubating standardized inocula of mycobacteria on media containing various concentrations of the drug to be tested, including the critical concentration. Bacterial growth that exceeds that of a 1:100 dilution of the inoculum on drug-free medium at and above the critical concentration is interpreted as showing resistance [8] . While this method is still in use for M. tuberculosis and has been widely used for DSTs of NTM, cut-off points for resistance have not been properly defined for NTM [11] . Moreover, none have been clinically validated [12] . The resistance ratio method is methodologically similar to the absolute concentration method, but the MIC is identified and divided by that of the M. tuberculosis H37Rv reference strain, to come to a ratio. Low ratios (8 as resistant. This method is more relevant for M. tuberculosis than for NTM. Nonetheless, this method was applied to NTM in the treatment trial of MAC, M. xenopi and Mycobacterium malmoense pulmonary disease by the British Thoracic Society [13] . The proportion method, as its name suggests, estimates the proportion of bacteria in the inoculum that is resistant to the drug at the tested

Table 2. Currently recommended treatment regimens for pulmonary nontuberculous mycobacteria disease per species. Species

Recommended regimen

Alternative

Mycobacterium avium complex

RIF–EMB–macrolide (± AS) Duration: >12 months CNeg INH–RIF–EMB (± AS) Duration: 12 months RIF–EMB–macrolide Duration: >12 months CNeg RIF–EMB–macrolide (± quinolone) Duration: >12 months CNeg Three or four drugs of: amikacin, cefoxitin, imipenem, tigecycline, linezolid (intensive phase) Duration: >12 months CNeg Macrolide plus two of: amikacin, cefoxitin, imipenem, linezolid (intensive phase) Duration: >12 months CNeg

Clo–EMB–macrolide

Mycobacterium kansasii (RIF susceptible) Mycobacterium malmoense Mycobacterium xenopi Mycobacterium abscessus subsp. abscessus or M. abscessus subsp. bolletii (former ‘Mycobacterium bolletii’): inducible macrolide resistance M. abscessus subsp. bolletii (former ‘Mycobacterium massiliense’): no inducible macrolide resistance

RIF–EMB–macrolide Duration: >12 months CNeg RIF–EMB–quinolone Duration: > 12 months CNeg RIF–EMB–quinolone Duration: >12 months CNeg

Amikacin, cefoxitin (intensive phase), then macrolide, ciprofloxacin, doxycycline [6] Duration: >12 months CNeg

AS: Amikacin or streptomycin; Clo: Clofazimine; CNeg: After conversion to negative cultures; EMB: Ethambutol; INH: Isoniazid; RIF: Rifampicin. Data taken from [5].


www.futuremedicine.com

1097

Review van Ingen & Kuijper

Proportion method on solid medium Drug-free positive control Twofold increasing drug concentrations 1:100 diluted inoculum

Broth microdilution in 96-well format Drug-free positive control Undiluted inoculum

Drug-free positive control Undiluted inoculum

Twofold increasing drug concentrations Twofold increasing drug concentrations

MIC

E-test method MIC i.e., >99% inhibition of growth Broth macrodilution method Drug-free positive control 1:100 diluted inoculum

Antibiotic concentration gradient on a strip MIC Visual margin of growth

Twofold increasing drug concentrations

Figure 1. Major drug susceptibility test methods.

concentration. Drug-containing media as well as drug-free media are inoculated with 2 dilutions of the initial inoculum. If the number of colonies on drug-containing media is >1% of that on the drug-free media, the isolate is considered to be resistant to the drug at the tested concentration [8] . ●●Broth microdilution

Broth microdilution allows the measurement of exact MICs by inoculating small (usually 100-μl) volumes of broth with a standardized inoculum of 5 × 105 CFU, typically in a 96-well plate format (Figure 1) . Growth density is measured optically and compared with growth in drug-free control vials in order to determine MICs; these systems are now available commercially (Sensititre™; Trek Diagnostics, CA, USA) and as in-house assays. Broth microdilution made what was to become a lasting impact in general bacteriology after the report by Ericsson and Sherris in 1971, which effectively rendered it the gold standard [14] . A decade later, in 1982, the first report on the use of broth (Mueller–Hinton) microdilution for drug susceptibility testing of RGM was published [15] . The initial study revealed that not all RGM would grow in the cation-adjusted

1098


Mueller–Hinton broth (CAMHB) medium. This issue could be overcome by supplementing the media with oleic acid, albumin, dextrose and catalase (OADC). A larger follow-up study confirmed the suitability of this method for RGM [16] . Thereafter, the microdilution was set up for use with SGM, applying the Middlebrook 7H9 medium [17] . Microdilution MICs for rifampicin, ethambutol and streptomycin proved lower than in agar, leading to discrepant interpretations for streptomycin [17] . In one multisite reproducibility study, DSTs for MAC by broth microdilution in both Middlebrook 7H9 and Mueller–Hinton medium were compared [18] . End point readings proved easier in 7H9 medium, leading to more reproducible results than in Mueller–Hinton medium [18] ; nonetheless, the current Clinical Laboratory Standards Institute (CLSI) document recommends the use of CAMHB. ●●Broth macrodilution

Akin to the proportion method, MICs can be determined by inoculating vials with liquid media and the antibiotics to be tested in their required concentrations. A 1:100 diluted inoculum is added to a drug-free control vial so the


Drug susceptibility testing of nontuberculous mycobacteria actual MIC can be determined, being the lowest drug concentration that yields less growth than the drug-free vial (thus the lowest concentration that kills >99% of the bacteria in the inoculum). These methods were developed in the late 1970s to determine the drug susceptibility of M. tuberculosis. The first commercially available radiometric broth macrodilution method (BACTEC™ 460; Becton Dickinson, MD, USA) used measurement of 14CO2 produced during the metabolism of 14C-incorporated palmitic acid and was adapted for testing SGM [19–21] . It proved less applicable to RGM because results were difficult to interpret clinically [22,23] . The radiometric BACTEC 460 system is no longer available. Its successor, the Mycobacterial Growth Indicator Tube (MGIT™; BD Biosciences, MD, USA; Figure 1) system, has already become the gold standard for primary culturing as well as DSTs of M. tuberculosis [24,25] . Despite 15 years of clinical use, this platform has not been excessively tested for NTM DSTs. Initial studies revealed largely concordant results between the MGIT and BACTEC 460 methods, except for MICs of ethambutol for the MAC [26] . The major advantage of the MGIT system is that this broth macrodilution methodology is already widely available owing to its success with M. tuberculosis primary isolation and DSTs [12,24–25] . ●●E-tests

E–tests are plastic strips calibrated with a continuous logarithmic MIC scale that covers 15 twofold dilutions of the test drug; these strips are pressed onto a solid medium plate that has been swabbed with a suspension of the mycobacteria with a preset inoculum (Figure 1) . E-tests were introduced in the late 1990s and soon thereafter applied to mycobacteria. The choice of medium, inoculum and duration of incubation are critical steps. Most studies that have assessed this methodology have used Mueller–Hinton blood solid media, although for M. marinum, Middlebrook 7H11 solid medium was preferred, as Mueller– Hinton blood media does not support the growth of all strains of M. marinum [27] . For the SGM M. kansasii, M. marinum and MAC bacteria, as well as for the RGM Mycobacterium chelonae and Mycobacterium fortuitum, MICs determined by E-tests were similar to those measured by the proportion method or absolute concentration method on Löwenstein–Jensen or Middlebrook 7H10 medium for most drugs [28–30] . For


Review

M. marinum, MICs for clarithromycin proved to be two- to three-fold lower by E-test and more than threefold lower for ethambutol, although this was most prominent at very low MICs and did not change their interpretation based on extrapolated cut-offs of M. tuberculosis and those proposed in earlier studies of agar dilution methods [31] . After the initial enthusiasm for these E-tests, multisite reproducibility studies revealed that for RGM, the reproducibility of the E-tests was inferior to that of broth microdilution, particularly for susceptibility testing to amikacin, imipenem and ciprofloxacin, three key drugs [32] . It is important to realize that E-tests were calibrated for MIC readings after 18–24 h of incubation, which is impossible even for the most rapidly growing NTM. ●●Molecular methods

No commercial assays exist for the detection of resistance-conferring mutations in NTM. In-house methods based on sequencing of the target gene and comparisons with wild-type strains of the same species have been developed, mostly for the MAC, M. kansasii and M. abscessus [33–37] . The few published studies have focused on the molecular detection of resistance mechanisms for macrolides, rifamycins and aminoglycosides. Owing to the important role of the macrolides in the treatment of NTM disease and the fact that the MAC and M. abscessus group organisms are the most frequent causative agents, molecular analyses of macrolide susceptibility in MAC and M. abscessus have received most attention [34–37] . Sequencing the 23S rRNA gene has been applied to MAC and rapid growers, and to detect the inducible macrolide resistance that is commonly observed among rapid growers, erm gene mutation analysis is now under investigation [37] . rpoB gene mutation analysis, similar to that conducted for M. tuberculosis, has been applied for the assessment of rifamycin susceptibility in M. kansasii. In order to detect aminoglycoside resistance in the MAC and M. abscessus, 16S rRNA gene mutation analysis can be used [33–36] . In vitro–in vivo correlations The limited data on in vitro–in vivo correlations have fed the myth that DSTs of NTM have no clinical utility. Immediately after NTM were found to be causative agents of human lung disease, discrepancies between in vitro drug susceptibility and in vivo outcomes of drug treatment were observed [4,13,38] . Yet for some key drugs,


1099

Review van Ingen & Kuijper there is clear evidence of an association between in vitro susceptibility and outcomes of treatment. The best evidence for relationships between MICs and outcomes of treatment has been gathered for MAC. This evidence comes mostly from trials of HIV-associated disseminated disease. Trials of monotherapy with rifampicin, ethambutol, clofazimine or clarithromycin for disseminated M. avium disease in HIV-infected patients established that only drug susceptibility testing results for clarithromycin predicted outcomes of treatment with this drug. Chaisson and coworkers showed that follow–up isolates of patients with M. avium bacteremia who experienced failure of clarithromycin treatment had clarithromycin MICs >32 μg/ml, while their pretreatment MICs had been ≤4 μg/ml in broth macrodilution [39] . Such clear relationships between in vitro drug susceptibility and culture conversion rates or symptomatic improvement during monotherapy with the respective drugs could not be proven for rifampicin, ethambutol and clofazimine; the MICs for these drugs did not predict outcomes of monotherapy with these drugs [39,40] . The macrolide antibiotic clarithromycin was also the first drug for which a clear relationship between in vitro activity and in vivo efficacy could be demonstrated in patients with pulmonary MAC disease: case series from the USA [41] and Japan [42] observed that the outcome of treatment with regimens including clarithromycin, rifampicin and ethambutol, with adjunctive aminoglycosides and fluoro quinolones, in terms of the percentage of patients who showed long-term conversion to negative cultures were significantly better in patients in whom baseline and follow-up isolates had low MICs for clarithromycin. For example, in the case series reported from Japan, culture conversion rates were 71.8% overall, but only 25% in patients whose primary isolates were already macrolide resistant [42] . The lack of correlation between in vitro and in vivo activity of the first-line anti-TB drugs isoniazid, rifampicin and ethambutol was also observed in the British Thoracic Society trials of regimens based on rifampicin and ethambutol for the MAC, M. xenopi and M. malmoense lung disease. Single-drug DSTs for isoniazid, rifampicin and ethambutol by the resistance ratio method demonstrated that in vitro resistance, using breakpoints extrapolated from M. tuberculosis DSTs, was commonplace and could not be linked to the outcome of treatment

1100


in vivo [13] . The second trial did not even apply drug susceptibility testing [43] . Amikacin is the most widely used amino glycoside and an important component of treatment regimens for severe MAC lung disease [4– 5,44] . Clinical studies have assessed the benefit of adding streptomycin to the rifampicin–ethambutol–clarithromycin regimen for pulmonary MAC disease, but found that that it only accelerated culture conversion, without improving the clinical outcomes in terms of symptom scores and prolonged culture conversion [45] . In addition, the streptomycin MIC did not predict the outcome of treatment with streptomycin-containing regimens [45] . Similar in vitro–in vivo discrepancies have been noted for classic antiTB drugs, such as ethionamide and cycloserine [4,46–47] , and routine single-drug testing of these drugs is thus not recommended [24] . For M. kansasii, data on in vitro–in vivo correlations stem from large retrospective series. The currently recommended treatment regimen consists of isoniazid, rifampicin and ethambutol for an 18-month duration, or 12 months after culture conversion [4] . The very few patients that experienced treatment failure of this regimen generally had strains isolated after treatment failure that showed rifampicin resistance, while their primary isolates had been tested as being susceptible [4,48–50] . Resistance to isoniazid or ethambutol may also be seen, usually in combination with rifampicin resistance [49] . Hence, rifampicin is the key drug to be tested [24] . In rifampicin-resistant M. kansasii isolates, CLSI guidelines recommend additional tests of isoniazid, amikacin, streptomycin, ciprofloxacin, moxifloxacin, clarithromycin, rifabutin and cotrimoxazole. Clinical validation of these tests has not been performed, although there is again retrospective evidence of the efficacy of sulfonamide-based regimens [49,50] and macrolide-based regimens [51] . The impact of acquired macrolide resistance on outcomes remains to be studied. For clinically significant SGM other than the MAC or M. kansasii, it has been recommended to test rifampicin and the set of drugs tested for rifampicin-resistant M. kansasii [4,24] . Results of these tests should be interpreted with caution, as in vitro activity and in vitro efficacy need not be correlated for any of these drugs. Here, M. simiae warrants specific attention; this species is characterized by high levels of drug resistance and a lack of synergistic activity of rifampicin and ethambutol in vitro [52] . In clinical practice,


Drug susceptibility testing of nontuberculous mycobacteria treatment outcomes in M. simiae disease are poor, which may reflect the high level of drug resistance [52] . For the RGM, in vitro DST results are used to design treatment regimens. As a result, these often feature two to four drug combinations, including a macrolide (depending on whether [inducible] macrolide resistance is measured) with an aminoglycoside, a fluoroquinolone, cefoxitin, imipenem, co-trimoxazole, a tetracycline or linezolid [4] . However, there is a very limited evidence base that proves that relationships between in vitro drug susceptibility and treatment outcomes exist. One study has reported cure rates of 90% after mostly monotherapy with trimethoprim–sulfamethoxazole for M. fortuitum disease and 72% after mostly amikacin with cefoxitin treatment for Mycobacterium chelonei (now separated into M. chelonae and the M. abscessus group) disease for patients whose isolates had proven to be susceptible to the relevant drugs [53] . Nonetheless, in M. abscessus lung disease, the outcomes of treatment using drugs to which susceptibility is noted in vitro still prove very limited. In a recent study of 69 patients, the culture conversion rate was just 48% [6] . Another case series of an equal number of patients followed up in South Korea investigated a fixed regimen designed regardless of in vitro drug susceptibility: clarithromycin, ciprofloxacin and doxycycline combined with intravenous amikacin and cefoxitin for the first 4 weeks. This regimen proved equally successful, with a culture conversion rate of 58% (38/65 patients) [54] . This questions the in vitro–in vivo relationships in pulmonary M. abscessus disease. In the latter study, outcomes were much better in infections by erm gene-defective M. abscessus group strains (‘Mycobacterium massiliense’; 25% [M. abscessus subsp. abscessus] vs 88% [‘M. massiliense’] culture conversion) [54,55] . This may alter the performance and interpretation of macrolide susceptibility test results in the future and is discussed below. For M. marinum, routine drug susceptibility is not currently recommended. Results of monotherapy with clarithromycin, doxycycline or minocycline, rifampicin, co-trimoxazole or clarithromycin, or combinations thereof are uniformly good, and in vitro susceptibility to these compounds is commonplace [56] . Failure of doxycycline treatment with the emergence of doxycycline resistance has been described [57] . DSTs are only warranted in such cases of


Review

treatment failure and should cover all of the aforementioned drugs and a fluoroquinolone [24] . Current recommendations In 2011, the National Committee on Clinical Laboratory Standards (NCCLS; now CLSI) published the second version of its document M24-A2 [24] . These guidelines advise using broth microdilution with CAMHB medium for DSTs of RGM, M. kansasii and nonfastidious SGM. There may also be a role for broth macrodilution only for the MAC; however, the CLSI document proposes using the radiometric BACTEC 460 method with Middlebrook 7H12B medium, which is now out of production. For the MAC, broth microdilution using 10% OADC-enriched, CAMBH medium is the alternative, for which a commercial method (Sensititre) is available. The current CLSI recommendations for NTM are summarized in Table 1. In the absence of guidelines by its European counterpart, the European Committee on Antimicrobial Susceptibility Testing (EUCAST), the CLSI-approved methods can be regarded as the gold standard for now. The current CLSI recommendations are built on a series of comparative studies of the various methods and experiences of a small number of reference laboratories [24] . ●●M. avium complex

Broth macrodilution using the radiometric BACTEC 460 method with Middlebrook 7H12B medium became the recommended DST platform for the MAC because results for all drugs except amikacin and ethambutol were comparable to those of agar dilution, but broth macrodilution added the advantages of a decreased turnaround time and superior interlaboratory reproducibility [18] compared with agar dilution [58,59] . The relative benefits of broth macrodilution and broth microdilution (which was already in use for the RGM) were analyzed in the landmark study by Woods and colleagues [18] . This study compared two broth microdilution assays using either CAMHB or Middlebrook 7H9 liquid medium with broth macrodilution using the commercial BACTEC 460 system using both Middlebrook 7H12B medium at pH 6.8 and a 7H12B medium with its pH raised to 7.3–7.4 [18] . The most important outcome was that all methods yielded similar results in terms of interpretative category (susceptible, intermediate and resistant)


1101

Review van Ingen & Kuijper as was set for each method and medium type. End point readings and interpretations were considered more difficult in the microdilution assays, particularly in CAMHB. As a result of difficulties in end point reading, the reproducibility of macrodilution proved superior to that of microdilution [18] . As stated, the problem with broth macrodilution is that the proposed platform, BACTEC 460, is no longer available (and had always been sparsely available in Europe). Its successor, the MGIT 960 system, is widely available and is thus a likely candidate for becoming the recommended broth macrodilution platform for MAC DSTs. Here, however, the problem is that this method has not been widely used for NTM DSTs and that comparative studies with other recommended methods are few and far between. The two available studies of BACTEC 460 and MGIT 960 DSTs noted largely comparable results for the MAC [26] , including M. avium subsp. paratuberculosis [60] . A comparative study of MGIT and agar dilution DSTs for the MAC revealed mostly concordant results, except for amikacin and ciprofloxacin [61] . Besides an unmet need for proper (clinical) validation of the MGIT platform for MAC, or in fact SGM DSTs, there is also a practical laboratory issue that needs to be overcome. To further automate the testing and interpretation of the results, the MGIT method should be appended with software that can convert the findings into a clinically meaningful result. For M. tuberculosis DSTs, there are software packages and laboratory utensils available to ease the performance of first-line drug testing by the streptomycin, isoniazid, rifampicin and ethambutol (the socalled ‘SIRE’) kit and an even more extensive software package, TB eXiST® (BD Bioscience, Erembodegem, Belgium), for testing multiple concentrations of multiple second-line drugs. There is some experience using this support software for MAC or SGM DSTs [62] . ●●M. kansasii & other SGM

For M. kansasii, broth micro- and macrodilution, agar dilution and agar proportion methods have all been used, albeit mostly without proper clinical validation [11,17,50] . While for most drugs tested the results of these various methods are largely identical, they tend to differ for streptomycin and the aminoglycosides. For these drugs, resistance is recorded by the agar proportion methods on solid media, whereas susceptibility

1102


is measured in broth (Middlebrook 7H9) microdilution; notably, in the absence of established breakpoints for microdilution agar, proportion breakpoints of M. tuberculosis are used for both techniques [17] . This explains part of the observed discrepancies between results of different methods. The susceptibility in broth correlated well with clinical responses in a small series of patients with rifampicin-resistant M. kansasii disease. This clinical correlation was important to confirm broth microdilution as the method of choice for M. kansasii [17,24,50] . Owing to their rarity and the few available studies assessing MICs and outcomes of treatment in vivo, optimal DST methods have not been set for the SGM other than the MAC or M. kansasii. Published laboratory-driven studies have used broth macro- and micro-dilution, absolute concentration methods on Middlebrook 7H10 medium and agar proportion methods. However, none of these studies have performed a proper clinical validation of the used breakpoints for resistance [11,24,52] . These results should thus be interpreted with caution and the consultation of experts is strongly recommended [4] . ●●Rapidly growing mycobacteria

Because classic agar-based methods intended for M. tuberculosis testing did not support the growth of all RGM, particularly M. chelonae, DSTs of RGM have always used methods adapted from general bacteriology. Broth microdilution in Mueller–Hinton medium, for example, was adopted early. The initial studies proved that this medium did need some adjustment, including increased cation concentrations and the use of the OADC supplement in order to ensure support of the growth of all RGM [15,16] . Subsequent studies revealed that broth microdilution DSTs for RGM were superior to macrodilution, agar proportion and agar dilution. This was particularly true for testing the susceptibility to amikacin, for which discrepancies in the interpretation between the various methods were observed; generally, solid medium-based testing reported false resistance to amikacin. In addition, the microdilution platform proved to be more reproducible than E-tests and was validated clinically to a limited extent, rendering it the currently recommended method [15–16,22,24,30,32] . For M. marinum, routine DSTs are not recommended, because this species is susceptible to most classes of drugs and treatment


Drug susceptibility testing of nontuberculous mycobacteria outcomes are generally very good. The use of DSTs is restricted to cases of treatment failure, where acquired resistance may be measured. For this situation, the CLSI recommends to use the broth microdilution platform with CAMBH supplemented with 5% OADC [24] . Both agar dilution methods on Middlebrook 7H10 medium and E-tests have been used in laboratory-driven studies, but their results and interpretations have not been validated clinically [11,63] . The observed disadvantage of the E-tests was that its reproducibility proved to be lower than that of the agar dilution method [63] . However, agar dilution methods tend to measure higher MICs for streptomycin and the aminoglycosides than broth microdilution. This has been suggested to lead to differences in interpretation (resistant vs susceptible), although this is of limited relevance, as breakpoints for both studied methods were extrapolated from M. tuberculosis DSTs [17] , which is unwarranted. Because of the growth characteristics of M. marinum, microdilution trays should be incubated at 28–30°C for 7 days. Recent insights into the pharmacokinetics & pharmacodynamics of key drugs In order to set clinically meaningful breakpoint concentrations for resistance, both the spread in MICs of wild-type strains as well as the pharmacokinetics of the key drugs in patients need to be known (see Figure 2 for explanation of key pharmacokinetic concepts) [64] . In the past year, three studies have explored the pharmacokinetics of key drugs in MAC and M. abscessus lung disease. These studies confirmed that important pharmacokinetic interactions occur in treatment, particularly with regimens that combine rifamycins and macrolides, as the regimens for MAC disease generally do. Simultaneous administration of rifampicin lowers macrolide concentrations in blood by 30% (azithromycin) to 60% (clarithromycin). Rifabutin has a similar, but weaker effect. As a result, 42% (once-daily dosing) to 84% (twice-daily dosing) of the patients met the applied pharmacodynamic index (T50% >MIC [serum drug concentration is above MIC for over 50% of the dosing interval]) of clarithromycin [65] . As expected, this percentage is much higher in patients treated for M. abscessus disease, who do not use rifamycins [66] . These pharmacokinetic data support the current CLSI breakpoints of macrolides for M. abscessus (≤4 mg/l) and MAC (≤8 mg/l) [24,65–66] , but the use of this


Review

time-over-MIC pharmacodynamic index has a limited evidence base. The moxifloxacin pharmacokinetics offer little support to the new CLSI breakpoint for moxifloxacin for the MAC and RGM (≤1 mg/l); with an average area under the concentration–time curve (AUC) of 18.81 mg/h/l in patients with MAC lung disease, only MICs as low as 0.125 mg/l would lead to AUC:MIC ratios of >100, which is associated with the bactericidal activity of fluoroquinolones [65] . Here, too, rifampicin lowers concentrations of moxifloxacin and may contribute negatively to moxifloxacin efficacy in MAC lung disease [65] . Thus, even if a strain is tested as being ‘susceptible’, with an MIC of 0.5 mg/l, moxifloxacin is most likely not an effective drug by itself for the patients infected with this strain. Serum concentrations of rifampicin and ethambutol are generally within the ranges considered adequate in TB treatment; those of rifampicin are actually higher than those observed in TB patients [65,67] . However, NTM show MICs to these agents that are 10–20-fold higher than for M. tuberculosis [11] , and thus pharmacodynamic indices for bactericidal activity of rifampicin against M. tuberculosis (fAUC [free area under the curve]:MIC >24.14) and ethambutol (Cmax:MIC >1.23), as determined in the hollow fiber model, were only attained by 18 and 57%, respectively, of patients with MAC lung disease [65] . Mechanisms of resistance Resistance mechanisms have not been widely studied in NTM; most data pertain to single drugs in single species. In the NTM, protein sequences of key drug targets (rpoB for rifampicin and 16S for streptomycin and aminoglycosides) are similar or identical to those in M. tuberculosis [68] . However, the NTM show much higher MICs in vitro [11] , which implies that in NTM, the impermeability of the cell wall is a stronger driver of drug resistance than polymorphisms in the drug target [12,68] . The broad repertoire of efflux pumps in NTM forms a second line of defense against antimicrobial compounds. The P55 [69] , tap [70,71] , tetV [72] , lfrA [73] and efpA [74] efflux pumps confer tetracycline, aminoglycoside, β-lactam antibiotic and fluoroquinolone, rifamycin and isoniazid resistance to the RGM Mycobacterium fortuitum and Mycobacterium smegmatis. Blocking of these efflux pumps by phenothiazines or verapamil


1103

Review van Ingen & Kuijper

Cmax: peak serum concentration Drug concentration (mg/l)

20

MIC, e.g. 4 mg/l

18 16 14 12 10 8

Time above MIC (T>MIC)

6 4 2 0 0

3

6

9

Area under the time-concentration curve (AUC; in blue)

12 Time (h)

15

18

21

24

Trough serum concentration, prior to new dose of the drug

Figure 2. PK/PD parameters. AUC: Area under the concentration–time curve; T>MIC: Time above the MIC.

has been shown to increase susceptibility to rifampicin, isoniazid, tetracyclines, quinolones and macrolides in vitro in both slowly and rapidly growing NTM [75] . Mutational resistance to macrolides has been observed, mainly in MAC disease. Its emergence significantly lowers the chance of attaining cure by drug treatment [39–42,76] . Mutations in codon 2058 or 2059 of the 23S ribosomal RNA gene (rrl) have been associated with high-level macrolide resistance in both the MAC species and the rapid growers of the M. abscessus group [35–38,77][34–37,76] . The most important risk factors for the development of macrolide resistance are prior macrolide monotherapy or regimens that include only quinolones and macrolides. The use of multidrug regimens that also include rifampicin and ethambutol prevents the emergence of macrolide resistance in MAC bacteria [76] . A recent study has focused on the mutational resistance to aminoglycosides observed in the MAC bacteria. Mutational resistance (i.e., mutations in the 16S rDNA gene) was found in isolates with MICs ≥64 mg/l. Clinical data revealed that these isolates came from patients with previous exposure to aminoglycosides. These high MICs and the mutational resistance observed were considered to be contraindications for further aminoglycoside therapy for patients with MAC disease. Based on these data, a breakpoint concentration of 64 mg/l was proposed [44] .

1104


Mutational resistance to aminoglycosides has also been observed in M. abscessus, particularly in recipients of long-term aminoglycoside treatments, such as cystic fibrosis patients and patients with otomastoiditis [4] . A mutation in codon 1408 of the 16S ribosomal RNA gene (rrs) is responsible for high-level resistance in both M. abscessus and M. chelonae after therapy as well as in vitro selection [77] . As expected, mutational resistance to rifampicin is important in M. kansasii, where rifampicin is the key component of treatment regimens and resistance to rifampicin is associated with treatment failure. The sole available study has recorded mutations in codons 513, 526 and 531 of the rpoB gene of M. kansasii strains with acquired resistance to rifampicin. These mutations are identical to those observed in rifampicin-resistant M. tuberculosis complex isolates [33] . In contrast to general bacteriology, drug resistance in NTM is though not to be related to the spread of plasmids with antibiotic resistance-determining gene variants. NTM do harbor plasmids, but these are species specific and there is no evidence of spread from one species to another [78] . For M. marinum, it is known that its plasmid harbors genes involved in mercury resistance [79] . For most species, the presence or role of plasmids has not been investigated. However, a very recent study has shown


Drug susceptibility testing of nontuberculous mycobacteria that M. abscessus strains in Brazil harbored the pMAB01 plasmid. This 56-kb circular plasmid encodes for several aminoglycoside-converting enzymes and a dihydropteroate synthase type-1 gene involved in susceptibility to sulfonamides. Strains harboring the plasmid were less susceptible to kanamycin [80] , a drug that is not used for NTM infections. Even if newly acquired plasmids yield gene variants that reduce susceptibility to particular antibiotics, it is questionable whether these will have an effect, since the impermeable cell wall and broad-repertoire efflux pumps of NTM already pose a major barrier to the activity of many drug classes. This effect was also seen in the M. abscessus strains in Brazil, where the dihydropteroate synthase type-1 gene in the pMAB01 plasmid did not confer sulfonamide resistance, as the species is already sulfonamide resistant [80] . How to test for inducible macrolide resistance? Inducible macrolide resistance owing to rRNA methylase (erm) genes has now been demonstrated in many RGM, especially in M. abscessus subsp. abscessus [81] ; this inducible resistance is often not reflected in the initial susceptibility results and demands specific testing by laboratories. The relationship between inducible macrolide resistance in M. abscessus and outcomes of treatment with macrolide-based regimens remains uncertain [5,81] , although outcomes seem better in M. abscessus subsp. bolletii (formerly ‘M. massiliense’ ), in which the erm gene is not functional [55] . The CLSI currently recommends that the final reading of clarithromycin susceptibility for RGM should be taken after at least 14 days of incubation, unless resistance (≥8 μg/ml) is recognized earlier; as soon as resistance is recognized, the report can be finalized [24] . This recommendation is based on the findings by Nash et al. [81,82] , who have used a prolonged incubation period and revealed that after 7 and 14 days of incubation, a clear increase in MIC could be observed. Subsequent investigators [37] have shown that inducible resistance is mediated by erm41 activity and that erm41 mutations and deletions, such as those in M. abscessus subsp. bolletii (formerly ‘M. massiliense’), lead to a loss of inducible resistance. Thus, erm41 sequencing could be used as a diagnostic tool instead of – although preferably alongside – microdilution with prolonged incubation.


Review

Conclusion & future perspective DSTs of NTM have proven to be clinically useful in select settings, but most of their role remains unknown. For many drugs, relationships between in vitro activity and in vivo outcomes of treatment have not been studied, particularly in diseases caused by SGM. Such analyses are hampered by the use of multidrug treatment regimens. One way of solving this puzzle could be to initiate early bactericidal activity studies, as in TB [83] . Such studies use monotherapy with the investigational drug for a limited amount of time – typically up to 2 weeks – and then continue using the recommended multidrug treatment. It is not known whether this approach could be feasible for NTM disease. In pulmonary NTM disease, conversion to negative cultures usually takes 3–6 months [4] and it is thus doubtful whether meaningful decreases in NTM load in respiratory specimens can be detected within the first 2 weeks of treatment. On the other hand, prolongation of the monotherapy period may lead to the emergence of drug resistance. The emergence of macrolide resistance after monotherapy or therapy with a macrolide and an ineffective companion drug such as a quinolone has been noted in patients with pulmonary as well as disseminated MAC disease [39,41,76] . These results echo those from the early days of monotherapy for TB. Hence, trials of long-term monotherapy should not be conducted. The current treatment regimens for SGM, especially the MAC, feature combinations of rifampicin and ethambutol; these are based on their synergistic activity [21,52,65,84–85] , as these compounds are inactive alone, both in vitro and in vivo [11,40–41,65] . Numerous studies have measured synergy in vitro, but none have addressed the question of whether low MICs to the combined drugs can predict a favorable outcome of treatment in vivo. One study of only five patients did observe decreases in susceptibility to these two drugs in combination in patients for whom treatment with this drug combination had failed [85] . It has also been observed that this synergistic effect is not seen for all NTM species (rifampicin and ethambutol do not show synergy against M. simiae and M. xenopi [52,84]), nor for all isolates of species for which this effect usually is seen, such as the MAC [84] . Moreover, the MICs measured for the drug combinations are often still in the range that probably cannot be overcome by regular doses of these drugs [52,65,84] .


1105

Review van Ingen & Kuijper The clinical utility of combined DSTs should be the subject of future prospective studies, preferably as part of randomized treatment trials. On the other hand, other than exploiting synergy in order to maximize the efficacy of these drugs that seem so ineffective alone in vitro, we should also reconsider the current dosing of these drugs. In their hollow fiber model, Gumbo and coworkers have already observed that intermittent high doses of ethambutol (i.e., at least 50 mg/kg, rather than the current 15 mg/kg) are most effective [86] . This is in line with the observation that pharmacodynamic parameters for ethambutol are infrequently met in patients receiving current doses, especially if the MICs of the drug tested alone are used for calculations [65] . Such novel dosing schemes should, in turn, lead to new MIC cut-offs to define drug resistance in vitro. A similar scenario could be true for rifampicin. Despite higher serum levels and AUCs than in TB patients, the optimal pharmacodynamic parameters are not met because of the high MICs of rifampicin against NTM in vitro [65] . As higher doses of rifampicin are now actively being pursued in clinical trials of TB [87,88] , the safety and tolerability of this approach is worth testing in patients with SGM disease. In RGM disease, particularly by M. abscessus, the challenge for the coming years will be to truly establish the role of the macrolide antibiotics in the face of the emerging insights in inducible resistance by erm gene activation [5,81–82] . In vitro testing needs to be harmonized and its exact clinical implications need to be tested in prospective trials. In addition, the role of the other frequently used drugs and the predictive values of in vitro susceptibility to these drugs need to be more firmly established. Much of our current thought on the subject is derived from the sole published study by Wallace and colleagues [53] . New insights in taxonomy and new potentially active drugs (e.g., tigecycline and linezolid) need to be combined and should be tested in in vitro models (e.g., hollow fiber models), animal experiments and, ultimately, clinical trials. These should also lead to clinically useful breakpoints for susceptibility testing. The relative rarity of these RGM diseases demands an international collaborative effort that is more likely to be based in academia than in industry. However, this route from in vitro assessments to animal models to pharmacodynamics

1106


models and, ultimately, to clinical trials in order to assess in vitro–in vivo correlations still has many weak points. The effects of medium type, pH and drug stability have not been well studied [12] . An interesting example is the use of the growth-enhancing OADC supplement, which includes albumin that can bind and inactivate antibiotics, as well as oleic acid, which shows some antimycobacterial activity [89] . Our current in vitro DST is based on testing the susceptibility of planktonic bacteria. Mycobacteria are well-known biofilm producers [90] and biofilm formation may play a role in human infection. The activity of currently recommended drugs in biofilms may be very different from their activity against planktonic mycobacteria. While assays to measure drug activity against biofilms exist [91] , these have also not been well studied in terms of their in vitro–in vivo correlations, particularly not for mycobacteria. In addition, mycobacteria are intracellular pathogens and the majority of NTM reside within macrophages [92] . Their intracellular localizations impose another bias to drug susceptibility testing, which is performed against planktonic bacteria in culture media, rather than against mycobacteria residing within macrophages or dendritic cells. Thus, testing the activity of antimycobacterial drugs against mycobacteria within macrophages, as can be done in pharmacodynamic models such as the hollow fiber model, are probably more relevant than current DST practices [86] . Data from these complicated models can be translated back into current DST methodologies by setting new breakpoints for resistance. Animal models of NTM disease exist, but have not been extensively studied. Recent studies have assessed the virtues of different mouse models, mostly using severely immunocompromised mice [93] . These models may not work for all NTM species, although that is a topic that has not been well studied. In addition, these models mimic disseminated NTM disease and may not be good predictors of treatment efficacy in NTM lung disease, which is the most frequent disease type. Animal models for NTM lung disease have been developed; the classical model is a guinea pig model with lung emphysema induced by quartz inhalation [94] , but this model has been little used. This situation is further complicated by the sheer number of drugs and regimens that should be investigated in these models; currently used models may not be


Drug susceptibility testing of nontuberculous mycobacteria suitable for testing all drugs and combinations, owing to animal-specific toxicity issues. Drug susceptibility testing of NTM needs to move from its current status as a mysterious art form that thrives in secluded reference laboratories to a science-based, standardized, transparent and easily interpreted clinical tool that improves the outcome of treatment in the individual patient. Clinicians and microbiologists need each other to reach that goal.

Review

Financial & competing interests disclosure The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. No writing assistance was utilized in the production of this manuscript.

Executive summary ●●

Drug susceptibility testing is an important tool to aid patient management, if properly performed.

●●

Clear correlations between drug susceptibility and outcomes of treatment do exist, but for specific drug classes and specific species.

●●

Broth microdilution is the recommended method of drug susceptibility testing, but many other methods are used without the proper assessment of in vitro–in vivo correlations.

●●

The clinical significance of inducible macrolide resistance among rapidly growing mycobacteria needs further evaluation.

●●

Clinical trials are urgently needed in order to improve patient management as well as drug susceptibility testing practices.

References

microbiologic outcomes in patients receiving treatment for Mycobacterium abscessus pulmonary disease. Clin. Infect. Dis. 52, 565–571 (2011).

Papers of special note have been highlighted as: • of interest; •• of considerable interest 1

Marras TK, Chedore P, Ying AM, Jamieson F. Isolation prevalence of pulmonary nontuberculous mycobacteria in Ontario, 1997–2003. Thorax 62, 661–666 (2007).

2

Thomson RM. Changing epidemiology of pulmonary nontuberculous mycobacteria infections. Emerg. Infect. Dis. 16, 1576–1583 (2010).

3

Falkinham JO 3rd. Surrounded by mycobacteria: nontuberculous mycobacteria in the human environment. J. Appl. Microbiol. 107, 356–367 (2009).

4

Griffith DE, Aksamit T, Brown–Elliot BA et al. An official ATS/IDSA statement: diagnosis, treatment, and prevention of nontuberculous mycobacterial diseases. Am. J. Respir. Crit. Care Med. 175, 367–416 (2007).

•• Essential statement for all diagnostic and treatment decisions. 5

6

van Ingen J, Ferro BE, Hoefsloot W, Boeree MJ, van Soolingen D. Drug treatment of pulmonary nontuberculous mycobacterial disease in HIV-negative patients: the evidence. Expert Rev. Anti Infect. Ther. 11, 1065–1077 (2013). Jarand J, Levin A, Zhang L, Huitt G, Mitchell JD, Daley CL. Clinical and


•• Retrospective study that shows the limited in vitro–in vivo correlations in Mycobacterium abscessus lung disease. 7

van Ingen J. Nontuberculous mycobacteria: from gene sequences to clinical relevance [PhD thesis]. Radboud University Nijmegen, The Netherlands (2009). http://webdoc.ubn.ru.nl/mono/i/ingen_j_ van/nontmy.pdf

8

Canetti G, Froman S, Grosset J et al. Mycobacteria: laboratory methods for testing drug sensitivity and resistance. Bull. World Health Organ. 29, 565–578 (1963).

9

Runyon EH. Anonymous mycobacteria in pulmonary disease. Med. Clin. North Am. 43, 273–290 (1959).

10 Middlebrook G, Cohn ML. Bacteriology of

tuberculosis: laboratory methods. Am. J. Public Health Nations Health 48, 844–853 (1958). 11 van Ingen J, van der Laan T,

Dekhuijzen PNR, Boeree MJ, van Soolingen D. In vitro drug susceptibility of 2275 clinical nontuberculous Mycobacterium isolates of 49 species in The Netherlands. Int. J. Antimicrob. Agents 35, 169–173 (2010).

12 van Ingen J, Boeree M, van Soolingen D,

Mouton J. Resistance mechanisms and drug susceptibility testing of nontuberculous mycobacteria. Drug Resist. Updat. 15, 149–161 (2012). 13 Research Committee of the British Thoracic

Society. First randomised trial of treatments for pulmonary disease caused by M. avium intracellulare, M. malmoense, and M. xenopi in HIV negative patients: rifampicin, ethambutol and isoniazid versus rifampicin and ethambutol. Thorax 56, 167–172 (2001). 14 Ericsson HM, Sherris JC. Antibiotic

sensitivity testing. Report of an international collaborative study. Acta Pathol. Microbiol. Scand. B Microbiol. Immunol. 217 Suppl: 217: 1+ (1971). 15 Swenson JM, Thornsberry C, Silcox VA.

Rapidly growing mycobacteria: testing of susceptibility to 34 antimicrobial agents by broth microdilution. Antimicrob. Agents Chemother. 22, 186–192 (1982). 16 Swenson JM, Wallace RJ Jr, Silcox VA,

Thornsberry C. Antimicrobial susceptibility of five subgroups of Mycobacterium fortuitum and Mycobacterium chelonae. Antimicrob. Agents Chemother. 28, 807–811 (1985). 17 Wallace RJ Jr, Nash DR, Steele LC,

Steingrube V. Susceptibility testing of slowly growing mycobacteria by a microdilution


1107

Review van Ingen & Kuijper antimicrobial susceptibility testing of Mycobacterium marinum clinical isolates. J. Clin. Microbiol. 35, 2083–2086 (1997).

MIC method with 7H9 broth. J. Clin. Microbiol. 24, 976–981 (1986). 18 Woods GL, Williams–Bouyer N,

Wallace RJ Jr et al. Multisite reproducibility of results obtained by two broth dilution methods for susceptibility testing of Mycobacterium avium complex. J. Clin. Microbiol. 41, 627–631 (2003). 19 Snider DE Jr, Good RC, Kilburn JO et al.

28 Fabry W, Schmid EN, Ansorg R. Comparison

Synergistic effects of antimycobacterial drug combinations on Mycobacterium avium complex determined radiometrically in liquid medium. Eur. J. Clin. Microbiol. 6, 530–535 (1987).

of the E test and a proportion dilution method for susceptibility testing of Mycobacterium avium complex. J. Med. Microbiol. 44, 227–230 (1996). Bolmstrom A. Evaluation of Etest for rapid susceptibility testing of Mycobacterium chelonae and M. fortuitum. J. Clin. Microbiol. 32, 1846–1849 (1994). Hoffner SE. Antimicrobial susceptibility of Mycobacterium marinum determined by E–test and agar dilution. Scand. J. Infect. Dis. 33, 585–588 (2001). Multisite reproducibility of Etest for susceptibility testing of Mycobacterium abscessus, Mycobacterium chelonae, and Mycobacterium fortuitum. J. Clin. Microbiol. 38, 656–661 (2000). 33 Klein JL, Brown TJ, French GL. Rifampin

resistance in Mycobacterium kansasii is associated with rpoB mutations. Antimicrob. Agents Chemother. 45, 3056–3058 (2001). 34 Meier A, Heifets L, Wallace RJ Jr et al.

Molecular mechanisms of clarithromycin resistance in Mycobacterium avium: observation of multiple 23S rDNA mutations in a clonal population. J. Infect. Dis. 174, 354–360 (1996).

Susceptibility Testing of Mycobacteria, Nocardiae, and Other Aerobic Actinomycetes (CLSI document M24-A2) (2nd Ed.). Clinical Standards Institute, PA, USA (2011). Antimicrobial susceptibility testing of Mycobacterium tuberculosis (EUCAST document E.DEF 8.1) – report of the Subcommittee on Antimicrobial Susceptibility Testing of Mycobacterium tuberculosis of the European Committee for Antimicrobial Susceptibility Testing (EUCAST) of the European Society of Clinical Microbiology and Infectious Diseases (ESCMID). Clin. Microbiol. Infect. 13, 1144–1156 (2007).

35 Nash KA, Inderlied CB. Rapid detection of

mutations associated with macrolide resistance in Mycobacterium avium complex. Antimicrob. Agents Chemother. 40, 1748–1750 (1996). 36 Wallace RJ Jr, Meier A, Brown BA et al.

Genetic basis for clarithromycin resistance among isolates of Mycobacterium chelonae and Mycobacterium abscessus. Antimicrob. Agents Chemother. 40, 1676–1681 (1996). 37 Bastian S, Veziris N, Roux AL et al.

Assessment of clarithromycin susceptibility in strains belonging to the Mycobacterium abscessus group by erm(41) and rrl sequencing. Antimicrob. Agents Chemother. 55, 775–781 (2011).

26 Piersimoni C, Nista D, Bornigia S, De Sio G.

Evaluation of a new method for rapid drug susceptibility testing of Mycobacterium avium complex isolates by using the mycobacteria growth indicator tube. J. Clin. Microbiol. 36, 64–67 (1998). 27 Flynn CM, Kelley CM, Barrett MS,

Jones RN. Application of the Etest to the

1108

of Mycobacterium avium complex infection: do the results of in vitro susceptibility tests predict therapeutic outcome in humans? J. Infect. Dis. 173, 677–683 (1996). 41 Wallace RJ Jr, Brown BA, Griffith DE,

Girard WM, Murphy DT. Clarithromycin regimens for pulmonary Mycobacterium avium complex: the first 50 patients. Am. J. Respir. Crit. Care Med. 153, 1766–1772 (1996). 42 Tanaka E, Kimoto T, Tsuyuguchi K et al.

Effect of clarithromycin regimen for Mycobacterium avium complex pulmonary disease. Am. J. Respir. Crit. Care Med. 160, 866–872 (1999).

32 Woods GL, Bergmann JS, Witebsky FG et al.

24 Clinical Laboratory Standards Institute.

25 Drobniewski F, Rusch-Gerdes S, Hoffner S.

40 Sison JP, Yao Y, Kemper CA et al. Treatment

31 Werngren J, Olsson-Liljequist B, Gezelius L,

of different test conditions on the susceptibility of Mycobacterium fortuitum and Mycobacterium chelonae to amikacin. J. Antimicrob. Chemother. 33, 483–494 (1994). Chan CY, Cheng AF. In-vitro activity of quinolones and macrolides against mycobacteria. J. Antimicrob. Chemother. 34, 343–351 (1994).

Clarithromycin therapy for bacteremic Mycobacterium avium complex disease. A randomized, double–blind, dose–ranging study in patients with AIDS. AIDS Clinical Trials Group Protocol 157 Study Team. Ann. Intern. Med. 121, 905–911 (1994).

30 Hoffner SE, Klintz L, Olsson-Liljequist B,

22 Hansen KT, Clark RB, Sanders WE. Effects

23 Yew WW, Piddock LJ, Li MS, Lyon D,

39 Chaisson RE, Benson CA, Dube MP et al.

29 Fabry W, Schmid EN, Ansorg R. Comparison

20 Heifets LB, Iseman MD, Lindholm-Levy PJ.

21 Hoffner SE, Svenson SB, Kallenius G.

mycobacterial infection of the lungs. Thorax 23, 94–99 (1968).

of the E test and a proportion dilution method for susceptibility testing of Mycobacterium kansasii. Chemotherapy 41, 247–252 (1995).

Rapid drug–susceptibility testing of Mycobacterium tuberculosis. Am. Rev. Respir. Dis. 123, 402–406 (1981). Determination of MICs of conventional and experimental drugs in liquid medium by the radiometric method against Mycobacterium avium complex. Drugs Exp. Clin. Res. 13, 529–538 (1987).

38 Goldman KP. Treatment of unclassified

•

Important study for the understanding of inducible macrolide resistance in rapid growers.


•

First trial to clearly show that macrolide resistance in vitro is associated with clinical failure in vivo.

43 Research Committee of the British Thoracic

Society. Clarithromycin vs ciprofloxacin as adjuncts to rifampicin and ethambutol in treating opportunist mycobacterial lung diseases and an assessment of Mycobacterium vaccae immunotherapy. Thorax 63, 627–634 (2008). 44 Brown-Elliott BA, Iakhiaeva E, Griffith DE

et al. In vitro activity of amikacin against isolates of Mycobacterium avium complex with proposed MIC breakpoints and finding of a 16S rRNA gene mutation in treated isolates. J. Clin. Microbiol. 51, 3389–3394 (2013). •• Important contribution towards meaningful susceptibility testing to amikacin. 45 Kobashi Y, Matsushima T, Oka M. A double-

blind randomized study of aminoglycoside infusion with combined therapy for pulmonary Mycobacterium avium complex disease. Respir. Med. 101, 130–138 (2007). 46 Kuze F, Kurasawa T, Bando K, Lee Y,

Maekawa N. In vitro and in vivo susceptibility of atypical mycobacteria to various drugs. Rev. Infect. Dis. 3, 885–897 (1981). 47 Kuze F. Experimental chemotherapy in

chronic Mycobacterium avium-intracellulare infection of mice. Am. Rev. Respir. Dis. 129, 453–459 (1984). 48 Ahn CH, Lowell JR, Ahn SS, Ahn SI,

Hurst GA. Short-course chemotherapy for


Drug susceptibility testing of nontuberculous mycobacteria pulmonary disease caused by Mycobacterium kansasii. Am. Rev. Respir. Dis. 128, 1048–1050 (1983). 49 Ahn CH, Wallace RJ Jr, Steele LC,

Murphy DT. Sulfonamide-containing regimens for disease caused by rifampin-resistant Mycobacterium kansasii. Am. Rev. Respir. Dis. 135, 10–16 (1987). 50 Wallace RJ Jr, Dunbar D, Brown BA et al.

Rifampin-resistant Mycobacterium kansasii. Clin. Infect. Dis. 18, 736–743 (1994). 51 Shitrit D, Baum GL, Priess R et al.

Pulmonary Mycobacterium kansasii infection in Israel, 1999–2004: clinical features, drug susceptibility, and outcome. Chest 129, 771–776 (2006). 52 van Ingen J, Totten SE, Heifets LB,

Boeree MJ, Daley CL. Drug susceptibility testing and pharmacokinetics question current treatment regimens in Mycobacterium simiae complex disease. Int. J. Antimicrob. Agents 39, 173–176 (2012). 53 Wallace RJ, Swenson JM, Silcox VA,

Bulen MG. Treatment of nonpulmonary infections due to Mycobacterium fortuitum and Mycobacterium chelonei on the basis of in vitro susceptibilities. J. Infect. Dis. 152, 500–514 (1985). •• Landmark study that reveals associations between in vitro susceptibility and outcomes of treatment in vivo for extrapulmonary disease caused by rapid-growing nontuberculous mycobacteria. 54 Jeon K, Kwon OJ, Lee NY et al. Antibiotic

treatment of Mycobacterium abscessus lung disease: a retrospective analysis of 65 patients. Am. J. Respir. Crit. Care Med. 180, 896–902 (2009). 55 Koh WJ, Jeon K, Lee NY et al. Clinical

significance of differentiation of Mycobacterium massiliense from Mycobacterium abscessus. Am. J. Respir. Crit. Care Med. 183, 405–410 (2011). 56 Aubry A, Chosidow O, Caumes E, Robert J,

Cambau E. Sixty-three cases of Mycobacterium marinum infection: clinical features, treatment, and antibiotic susceptibility of causative isolates. Arch. Intern. Med. 162, 1746–1752 (2002). 57 Ljungberg B, Christensson B, Grubb R.

Failure of doxycycline treatment in aquarium-associated Mycobacterium marinum infections. Scand. J. Infect. Dis. 19, 539–543 (1987). 58 Inderlied CB, Young LS, Yamada JK.

Determination of in vitro susceptibility of Mycobacterium avium complex isolates to antimycobacterial agents by various methods.


Antimicrob. Agents Chemother. 31, 1697–1702 (1987). 59 Woods GL, Witebsky FG. Susceptibility

testing of Mycobacterium avium complex in clinical laboratories. Results of a questionnaire and proficiency test performance by participants in the College of American Pathologists Mycobacteriology E Survey. Arch. Pathol. Lab. Med. 120, 436–439 (1996). 60 Krishnan MY, Manning EJ, Collins MT.

Comparison of three methods for susceptibility testing of Mycobacterium avium subsp. paratuberculosis to 11 antimicrobial drugs. J. Antimicrob. Chemother. 64, 310–316 (2009). 61 Marone P, Bono L, Carretto E, Barbarini D,

Telecco S. Rapid drug susceptibility of Mycobacterium avium complex using a fluorescence quenching method. J. Chemother. 9, 247–250 (1997). 62 Lucke K, Hombach M, Friedel U, Ritter C,

Böttger EC. Automated quantitative drug susceptibility testing of non-tuberculous mycobacteria using MGIT 960/EpiCenter TB eXiST. J. Antimicrob. Chemother. 67, 154–158 (2012). 63 Aubry A, Jarlier V, Escolano S,

Review

phylogenetic position, virulence, drug susceptibility and in vivo response to treatment in mycobacteria interrelated? Infect. Gen. Evol. 12, 832–837 (2012). 69 Silva PE, Bigi F, Santangelo MP et al.

Characterization of P55, a multidrug efflux pump in Mycobacterium bovis and Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 45, 800–804 (2001). 70 Ainsa JA, Blokpoel MC, Otal I, Young DB,

De Smet KA, Martin C. Molecular cloning and characterization of Tap, a putative multidrug efflux pump present in Mycobacterium fortuitum and Mycobacterium tuberculosis. J. Bacteriol. 180, 5836–5843 (1998). 71 Ramon-Garcia S, Martin C, Ainsa JA,

De Rossi E. Characterization of tetracycline resistance mediated by the efflux pump Tap from Mycobacterium fortuitum. J. Antimicrob. Chemother. 57, 252–259 (2006). 72 De Rossi E, Blokpoel MC, Cantoni R et al.

Molecular cloning and functional analysis of a novel tetracycline resistance determinant, tet(V), from Mycobacterium smegmatis. Antimicrob. Agents Chemother. 42, 1931–1937 (1998). 73 Sander P, De Rossi E, Boddinghaus B et al.

Truffot-Pernot C, Cambau E. Antibiotic susceptibility pattern of Mycobacterium marinum. Antimicrob. Agents Chemother. 44, 3133–3136 (2000). 64 Mouton JW, Brown DF, Apfalter P et al. The

role of pharmacokinetics/pharmacodynamics in setting clinical MIC breakpoints: the EUCAST approach. Clin. Microbiol. Infect. 18, E37–E45 (2011). 65 van Ingen J, Egelund EF, Levin A et al. The

pharmacokinetics and pharmacodynamics of pulmonary Mycobacterium avium complex disease treatment. Am. J. Respir. Crit. Care Med. 186, 559–565 (2012). •• Largest pharmacokinetic study in nontuberculous mycobacteria disease to date, including commonly used and rare drugs related to MIC data. 66 Koh WJ, Jeong BH, Jeon K, Lee SY, Shin SJ.

Therapeutic drug monitoring in the treatment of Mycobacterium avium complex lung disease. Am. J. Respir. Crit. Care Med. 186, 797–802 (2012). 67 Magis-Escurra C, Alffenaar JW, Hoefnagels I

et al. Pharmacokinetic studies in patients with nontuberculous mycobacterial lung infections. Int. J. Antimicrob. Agents 42, 256–261 (2013). 68 van Ingen J, Boeree MJ, van Soolingen D,

Iseman MD, Heifets LB, Daley CL. Are

Contribution of the multidrug efflux pump LfrA to innate mycobacterial drug resistance. FEMS Microbiol. Lett 193, 19–23 (2000). 74 Li XZ, Zhang L, Nikaido H. Efflux

pump-mediated intrinsic drug resistance in Mycobacterium smegmatis. Antimicrob. Agents Chemother. 48, 2415–2423 (2004). 75 Rodrigues L, Aínsa JA, Amaral L,

Viveiros M. Inhibition of drug efflux in mycobacteria with phenothiazines and other putative efflux inhibitors. Recent Pat. Antiinfect. Drug Discov. 6, 118–127 (2011). 76 Griffith DE, Brown-Elliott BA, Langsjoen B

et al. Clinical and molecular analysis of macrolide resistance in Mycobacterium avium complex lung disease. Am. J. Respir. Crit. Care Med. 174, 928–934 (2006). 77 Prammananan T, Sander P, Brown BA et al.

A single 16S ribosomal RNA substitution is responsible for resistance to amikacin and other 2-deoxystreptamine aminoglycosides in Mycobacterium abscessus and Mycobacterium chelonae. J. Infect. Dis. 177, 1573–1581 (1998). 78 Picardeau M, Vincent V. Characterization of

large linear plasmids in mycobacteria. J. Bacteriol. 179, 2753–2756 (1997). 79 Stinear TP, Seemann T, Harrison PF et al.

Insights from the complete genome sequence of Mycobacterium marinum on the evolution


1109

Review van Ingen & Kuijper of Mycobacterium tuberculosis. Genome Res. 18, 729–741 (2008). 80 Matsumoto CK, Bispo PJ, Santin K,

Nogueira CL, Leão SC. Demonstration of plasmid-mediated drug resistance in Mycobacterium abscessus. J. Clin. Microbiol. 26, 1727–1729 (2014). 81 Nash KA, Andini N, Zhang Y,

Brown-Elliott BA, Wallace RJ Jr. Intrinsic macrolide resistance in rapidly growing mycobacteria. Antimicrob. Agents Chemother. 50, 3476–3478 (2006). 82 Nash KA, Brown-Elliott BA, Wallace RJ Jr.

A novel gene, erm(41), confers inducible macrolide resistance to clinical isolates of Mycobacterium abscessus but is absent from Mycobacterium chelonae. Antimicrob. Agents Chemother. 53, 1367–1376 (2009). 83 Donald PR, Diacon AH. The early

bactericidal activity of anti-tuberculosis drugs: a literature review. Tuberculosis 88 (Suppl. 1), S75–S83 (2008). 84 van Ingen J, Hoefsloot W, Mouton JW,

Boeree MJ, van Soolingen D. Synergistic activity of rifampicin and ethambutol against slow growing nontuberculous mycobacteria is currently of questionable clinical

1110

significance. Int. J. Antimicrob. Agents 42, 80–82 (2013). 85 Hoffner SE, Heurlin N, Petrini B,

Svenson SB, Kallenius G. Mycobacterium avium complex develop resistance to synergistically active drug combinations during infection. Eur. Respir. J. 7, 247–250 (1994). 86 Deshpande D, Srivastava S, Meek C, Leff R,

Gumbo T. Ethambutol optimal clinical dose and susceptibility breakpoint identification by use of a novel pharmacokinetic– pharmacodynamic model of disseminated intracellular Mycobacterium avium. Antimicrob. Agents Chemother. 54, 1728–1733 (2010). 87 van Ingen J, Aarnoutse RE, Donald PR et al.

Why do we use 600mg of rifampicin in tuberculosis treatment? Clin. Infect. Dis. 52, e194–e199 (2011). 88 Diacon, AH, Patientia, RF, Venter, A et al.

Early bactericidal activity of high-dose rifampin in patients with pulmonary tuberculosis evidenced by positive sputum smears. Antimicrob. Agents Chemother. 51, 2994–2996 (2007). 89 Shojaei H, Daley C, Gitti Z et al.

Mycobacterium iranicum sp. nov., a rapidly


growing scotochromogenic species isolated from clinical specimens on three different continents. Int. J. Syst. Evol. Microbiol. 63, 1383–1389 (2013). 90 Mullis SN, Falkinham JO 3rd. Adherence

and biofilm formation of Mycobacterium avium, Mycobacterium intracellulare and Mycobacterium abscessus to household plumbing materials. J. Appl. Microbiol. 115, 908–914 (2013). 91 Antunes AL, Trentin DS, Bonfanti JW et al.

Application of a feasible method for determination of biofilm antimicrobial susceptibility in staphylococci. APMIS 118, 873–877 (2010). 92 Houben D, Demangel C, van Ingen J. et al.

ESX-1-mediated translocation to the cytosol controls virulence of mycobacteria. Cell Microbiol. 14, 1287–1298 (2012). 93 Lerat I, Cambau E, Roth Dit Bettoni R et al.

In vivo evaluation of antibiotic activity against Mycobacterium abscessus. J. Infect. Dis. 209, 905–912 (2014). 94 Wolinsky E. Chemotherapy and pathology of

experimental photochromogenic mycobacterial infections. Am. Rev. Respir. Dis. 80, 522–534 (1959).


In Vitro Susceptibility Testing of Tedizolid against Nontuberculous Mycobacteria.

Nontuberculous mycobacteria.

Distribution of nontuberculous Mycobacteria strains.

Genomic characterization of Nontuberculous Mycobacteria.

Nontuberculous mycobacteria in diffuse panbronchiolitis.

Nontuberculous mycobacteria and associated diseases.

Nontuberculous mycobacteria immune reconstitution syndrome.

Drug Susceptibility Testing of 31 Antimicrobial Agents on Rapidly Growing Mycobacteria Isolates from China.

The growing challenge of nontuberculous mycobacteria.

Increasing Burden of Nontuberculous Mycobacteria in Korea.

Antimicrobial susceptibility of standard strains of nontuberculous mycobacteria by microplate Alamar Blue assay.

Pulmonary disease caused by nontuberculous mycobacteria.

Nontuberculous Mycobacteria: Epidemiologic, Mycobacteriologic, and Clinical Aspects.

Nontuberculous Mycobacteria in Noncystic Fibrosis Bronchiectasis.

"Multidrug-resistant tuberculosis" may be nontuberculous mycobacteria.

Nontuberculous Mycobacteria: Skin and Soft Tissue Infections.

Gradenigo syndrome caused by nontuberculous mycobacteria.

Tuberculin test conversion during repeated skin testing, associated with sensitivity to nontuberculous mycobacteria.

Increasing nontuberculous mycobacteria infection in cystic fibrosis.

Environmental Nontuberculous Mycobacteria in the Hawaiian Islands.

Prevalence and drug resistance of nontuberculous mycobacteria, northern China, 2008-2011.

Surveillance of Tuberculosis in Taipei: The Influence of Nontuberculous Mycobacteria.

Evaluation of the Speed-Oligo Mycobacteria assay for the identification of nontuberculous mycobacteria.

Surveillance of Extrapulmonary Nontuberculous Mycobacteria Infections, Oregon, USA, 2007-2012.