OpiniãoThe current guidelines for treatment of DR-TB stress the need for access to reliable, quality-assured drug susceptibility testing (DST), to be provided by national TB programmes (NTPs) and associated laboratories, to inform the use of the WHO-recommended regimens. Rapid molecular testing is making it increasingly feasible for NTPs to detect MDR/RR-TB and other types of resistance quickly, and to use the results to guide treatment decisions (8, 9). Hence, rapid molecular testing should be made available and accessible, to ensure DST for at least rifampicin, isoniazid and fluoroquinolones, given that DST for these drugs is essential for selecting the most appropriate initial DR-TB regimen. If capacity does not exist, NTPs must promptly build rapid molecular testing capacity, and all efforts must be made to ensure universal access to all patients initiating a regimen for any form of TB, including both drug-susceptible and drug-resistant forms. While NTPs build capacity to ensure routine performance of DST for medicines used in the clinical management of all patients, building surveillance systems to determine the local prevalence of DR-TB strains will be important to guide programmatic planning. This is especially important for drugs where resistance testing is not routinely performed or when the resistance prevalence of the drug is expected to be low initially (e.g. pretomanid) and needs to be monitored over time. Drug-resistance surveillance (DRS) can be based on data from routine diagnostic DST in TB patients (i.e. continuous surveillance) or from special surveys representative of the entire TB patient population (i.e. drug-resistance surveys) (10). DST for rifampicin is mandatory for all cases, and DST for fluoroquinolones is mandatory in cases of demonstrated rifampicin resistance. DST for the drugs used in the newly recommended regimens is now increasingly critical. Data from local TB DRS may provide baseline estimates of the prevalence of resistance including among relevant subgroups of recurrent and re-registered individuals with TB disease (e.g. recurrence or new episode of TB or return after loss to follow-up). It can also provide monitoring trends to inform DST algorithms and inform broader local policy decisions (1, 11).
WHO recommends the use of the approved rapid molecular test as the initial test to detect TB disease as well as resistance to several anti-TB agents before the initiation of appropriate therapy for all TB patients. The increased recognition of drug resistance and improved access to rapid molecular testing have led more NTPs to test for at least rifampicin resistance at the start of TB treatment. Almost all WHO-recommended rapid tests for the initial diagnosis of TB include rifampicin-resistance testing; also, the most recent class of nucleic acid amplification tests (NAATs) include initial testing for rifampicin and isoniazid resistance. There are several other WHO-recommended molecular tests available that offer manual or automated DST for isoniazid, fluoroquinolones, pyrazinamide, ethionamide and injectable agents. The most recent recommendation of a low complexity assay for isoniazid and fluoroquinolone testing will ensure decentralized access to essential DST (12). Results can be available the same day (automated tests) or within a few days (manual tests) and can thus be used to decide on the initial regimen for treatment of Hr-TB or some other forms of DR-TB. Apart from their rapidity, some of these tests can also provide information on mutation patterns, which can influence the choice of treatment. If the inhA mutation is the only mutation present, it is likely that isoniazid can still be effective at a high dose, whereas if the katG mutation alone or both inhA and katG are present, isoniazid is no longer effective, even at a high dose. Some tests report “low-level” isoniazid resistance, which implies the presence of inhA mutations only. If rifampicin resistance is detected, rapid molecular tests for resistance to fluoroquinolones should be performed promptly, to inform the decision on which regimen to use for the treatment (13). If rifampicin resistance is not detected, rapid molecular testing for resistance to isoniazid is recommended, to inform the decision on whether the regimen to treat Hr-TB needs to be used.
Rapid molecular testing for rifampicin, isoniazid and fluoroquinolones is widely available; countries have accumulated experience in using these rapid tests, and access is also supported by the main donors where necessary. No rapid molecular testing is currently available for ethambutol, bedaquiline, clofazimine, linezolid, pretomanid and delamanid. WHO is currently evaluating tests for targeted nextgeneration sequencing solutions, which may provide an opportunity for the rapid molecular testing for multiple anti-TB medicines. Commercially available rapid molecular methods detect 86% (manual) to 93% (automated) of fluoroquinolone-resistant strains (12). Resistance to the fluoroquinolones can be inferred from the results of molecular testing for the presence of certain gyrA and gyrB mutations; however, only some of the genetic mutations are known or can be detected. Culturebased DST for fluoroquinolones should be considered when possible, especially when the prevalence of resistance to these drugs is high, or when resistance is suspected despite the molecular tests being negative. Further details on diagnostic tests recommended can be found in the relevant WHO consolidated guidelines (12) and operational handbook (14) in Module 3: Diagnosis – rapid diagnostics for tuberculosis detection.
Country programmes need to work towards the establishment of phenotypic DST for all TB medicines for which there are now agreed reliable and reproducible methods (e.g. bedaquiline, clofazimine, delamanid, fluoroquinolones, isoniazid, linezolid and rifampicin). The critical concentrations for various drugs were either established for the first time (bedaquiline, clofazimine, delamanid and linezolid) or revised (fluoroquinolones) in a WHO technical consultation in 2017 (15). Resistance to ethionamide/ prothionamide may be inferred from the results of molecular testing for isoniazid resistance (i.e. presence of mutations in the inhA promotor region) using either automated or manual molecular tests. However, susceptibility to ethionamide/prothionamide cannot be inferred purely based on the absence of inhA promotor gene mutation in commercially available NAATs, because resistance can be conferred by other mutations in the inhA gene and its promotor and by mutations in the ethA gene that are not detected by these NAATs. A standardized phenotypic DST method for pretomanid is being developed and is likely to be available soon. Phenotypic DST for cycloserine/terizidone, ethambutol, ethionamide/prothionamide, imipenem/meropenem or p-aminosalicylic acid is not routinely recommended because results may be unreliable (13).
The inability to undertake DST routinely in all patients despite all possible efforts should not be a barrier to starting patients on a potentially life-saving MDR-TB regimen; however, treatment should always be considered in a context of the potential risk of prescribing ineffective treatment and amplifying drug resistance, with a subsequent decrease in the likelihood of treatment effectiveness. If DST for bedaquiline and linezolid is not yet available, the clinician or the TB programme manager needs to estimate the likelihood of effectiveness of the medicines used, informed by the patient’s history of use of TB medicines, the drug-resistance pattern of the contact or index case, and recent representative DRS data. A reliable clinical history of exposure to bedaquiline and linezolid should thus be considered when designing a treatment regimen; however, this should be the main source of evidence to guide regimen design only in situations where phenotypic DST is not yet available. For paediatric patients, it is not always possible to obtain a DST result, owing to the difficulty of obtaining an adequate specimen or the lack of bacteriological confirmation; hence, the treatment regimen design is usually based on the drug-resistance pattern of the index case. In the absence of individual DST, relevant population surveillance data are essential to inform the choice and design of MDR-TB treatment regimens. In addition to TB DRS findings, it is important for practitioners to know which medicines have been in frequent use in each geographical setting or patient groups. If DST is not routinely available for individual patients where there is treatment failure, storage of M. tuberculosis isolates collected at baseline or during treatment monitoring can be considered for performing phenotypic DST or whole genome sequencing at reference laboratories.
New regimen adoption and implementation can and should proceed while the DST capacity is being established.
Despite some of the uncertainties about DST, NTPs should strive to test for resistance to a wide set of TB drugs and offer the most appropriate treatment regimen. The patient’s clinical response to treatment should always be carefully monitored. If there is a poor treatment response, undiagnosed drug-resistance or the coexistence of susceptible and resistant organisms in the same patient should be considered, as should alternative explanations for failure to respond to treatment (e.g. poor or erratic adherence to treatment, malabsorption, inadequate patient education or support, immune reconstitution inflammatory syndrome or the presence of comorbidities) (16).