Intended for healthcare professionals

Clinical Review State of the Art Review

Current and future treatments for tuberculosis

BMJ 2020; 368 doi: (Published 02 March 2020) Cite this as: BMJ 2020;368:m216
  1. Anthony Lee, medical student1 2,
  2. Yingda Linda Xie, assistant professor2 3,
  3. Clifton E Barry, senior investigator2,
  4. Ray Y Chen, associate research clinical2
  1. 1Medical Research Scholars Program, National Institutes of Health, Bethesda, MD, USA
  2. 2Tuberculosis Research Section, Laboratory of Clinical Immunology and Microbiology, Division of Intramural Medicine, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD, USA
  3. 3Division of Infectious Diseases, Department of Medicine, Rutgers New Jersey Medical School, Newark, NJ, USA
  1. Correspondence to RY Chen ray.chen{at}


Guidelines on the treatment of tuberculosis (TB) have essentially remained the same for the past 35 years, but are now starting to change. Ongoing clinical trials will hopefully transform the landscape for treatment of drug sensitive TB, drug resistant TB, and latent TB infection. Multiple trials are evaluating novel agents, repurposed agents, adjunctive host directed therapies, and novel treatment strategies that will increase the probability of success of future clinical trials. Guidelines for HIV-TB co-infection treatment continue to be updated and drug resistance testing has been revolutionized in recent years with the shift from phenotypic to genotypic testing and the concomitant increased speed of results. These coming changes are long overdue and are sorely needed to address the vast disparities in global TB incidence rates. TB is currently the leading cause of death globally from a single infectious agent, but the work of many researchers and the contributions of many patients in clinical trials will reduce the substantial global morbidity and mortality of the disease.


Tuberculosis (TB) is the leading cause of death globally from a single infectious agent, even surpassing HIV.1 To achieve the World Health Organization’s End TB Strategy (a 90% decrease in TB incidence and 95% decrease in TB mortality by 2035 compared with 2015) requires shorter and more effective treatment regimens for drug sensitive (DS) and multidrug resistant (MDR) TB (disease resistant to the two first line drugs isoniazid and rifampin). Treatment for latent TB infection also needs to be shorter in duration and targeted to individuals at risk of progression to be practical and effective globally. The current poor global control of TB is due in part to the lack of research innovation over the past few decades; current DS-TB treatment guidelines have been essentially unchanged for 35 years and treatment still takes a minimum six months. Randomized controlled clinical trials for TB treatments are often challenging to complete because of the extended duration of treatment (six months for DS-TB and up to 24 months for MDR/extensively drug resistant (XDR)-TB) and follow-up times (often one year after treatment completion) required to confirm cure in standard of care control arms. The good news is that multiple clinical trials for DS-TB, MDR-TB, XDR-TB, and latent TB are ongoing, and early results have led to important changes in the MDR-TB treatment guidelines. Unlike the last 30+ years, the next five years should continue to bring about major changes in the treatment of TB. This review summarizes the current state of TB treatment with an emphasis on the ongoing clinical trials to anticipate how guidelines may change in the coming years.


According to WHO estimates, in 2018 10 million people developed TB globally, for an incidence of 132/100 000 people. This global average, however, hides the vast disparities between developed and developing countries. Almost all cases are concentrated in South East Asia (44%), Africa (24%), and the western Pacific (18%) regions. The eastern Mediterranean accounts for 8.1%, Americas 2.9%, and Europe 2.6% of cases. The highest income countries have incidence rates of <10/100 000 people, while the countries with the highest rates are >50-fold higher, exceeding 500/100 000 people. Although global rates of DS-TB are slowly decreasing, rates of MDR-TB are decreasing less quickly, affecting 3.4% of new TB cases and 18% of previously treated cases. Even more challenging to treat is XDR-TB: MDR-TB with additional resistance to fluoroquinolones and injectable aminoglycosides. An estimated 6.2% of MDR-TB cases were actually XDR-TB in 2018.1

Sources and selection criteria

We searched PubMed for clinical trials of treatments for active and latent TB, including host directed therapies (HDT), through September 2019 and included only English language peer reviewed articles. We used terms including “active tuberculosis clinical trial,” “latent tuberculosis clinical trial,” “drug sensitive,” “drug resistant,” and “HDT.” We prioritized randomized controlled trials, but we also included some notable non-controlled interventional trials. Case reports, case series, and observational trials were generally excluded. We also searched the references used in current TB treatment guidelines and relevant systematic review articles, prioritizing the most recent articles. In addition, because ongoing trials do not yet have published results, we searched clinical trials registries, primarily We conducted a general internet search to look for additional information on registered clinical trials or to identify clinical trials registered with other clinical trials registries.

Active TB treatment

History of TB treatment

Treatment of tuberculosis has relied on multidrug chemotherapy to achieve three objectives:

  • To rapidly reduce mycobacterial burden to decrease disease morbidity, mortality, and transmission

  • To eradicate persistent mycobacterial populations to prevent relapse

  • To prevent acquisition of drug resistance.

Before the development of effective chemotherapy, TB treatment was essentially palliative care in sanatoriums with fresh air and sunlight. The age of TB chemotherapy began with the discovery of anti-tuberculosis compounds, beginning with streptomycin and para-aminosalicylic acid (PAS) in 1944.

Reviews of historical TB treatment trials date back to the 1940s, including the initial trial of PAS versus streptomycin to treat TB in 1944, which was one of the first randomized controlled chemotherapy trials ever conducted.2 These landmark studies encompassed dozens of randomized controlled trials of a few hundred participants each across Africa, India, Hong Kong, and the UK, and investigated the most optimal drug doses and combinations. These trials resulted in an initial “triple therapy” combination of isoniazid, PAS, and streptomycin for 24 months in the 1950s that cured 90-95% of patients.

In the 1980s results from subsequent trials led to formation of the currently recommended regimen of isoniazid, rifampin, pyrazinamide, and ethambutol for six months. The duration of treatment was established after increased relapse rates were seen in trials of less than six months’ duration.34567 Incredibly, despite the advent of newer drugs, this first line regimen has not changed for more than 35 years.

The coming decade is likely to change this for DS-TB and MDR-TB, with many clinical trials expected to be completed in the next several years (fig 1). However, the factors that allowed testing every combination of drug, dose, and duration viable in the 1940s through 1970s—limited numbers of drugs, abundance of patients, and relatively few concurrent trials competing to enroll—are no longer true today. To optimize the limited resources and increase the chance of clinical trials success, better methodologies are needed, possibly based on optimized biomarkers such as the NexGen EBA trial (described below), to determine the best combinations of drugs to take forward into advanced stage clinical trials.

Fig 1
Fig 1

Timeline of ongoing active TB clinical trials with start dates and projected end dates

Trials of shortening treatment for drug sensitive TB

Recent efforts to shorten duration of chemotherapy for DS-TB have focused on incorporating a fluoroquinolone or using higher doses of a rifamycin. The fluoroquinolone based strategies faced a major setback with the failure of three large trials in 2014,8910 despite higher rates of sputum culture conversion at two months in the fluoroquinolone arms in two of these trials.89 There is only one ongoing trial to our knowledge, conducted by the Beijing Chest Hospital, which is still attempting a fluoroquinolone strategy by continuing all four intensive phase drugs with or without levofloxacin for 4.5 months (table 1). Data from trials conducted in the 1970s and 1980s by the British Medical Research Council suggested that the sterilizing activity of pyrazinamide was limited to the initial eight weeks of treatment, owing to the acidic environment required for pyrazinamide activation.11 It remains to be seen whether the extension of intensive phase treatment and adding levofloxacin to 4.5 months will be sufficient for successful shortening of treatment.

Table 1

Ongoing trials of shortening treatment for drug sensitive, multi-drug resistant, and extensively drug resistant TB

View this table:

Two other ongoing trials attempt to shorten treatment by increasing the dose of the rifamycin, rifampin, or rifapentine. There is substantial evidence to support that rifampin dosed at the recommended 10 mg/kg is well below its therapeutic threshold and that much higher doses, even ≥35 mg/kg, may be safe and effective.1213 These studies suggest that high dose rifampin may shorten treatment based on faster rates of sputum culture conversion achieved.141516

The caveat is that earlier sputum culture conversion may not result in earlier cure, as shown by the failure of the fluoroquinolone treatment shortening trials. The RIFASHORT trial takes the surrogate results to a phase III trial, testing a standard of care (SOC) arm with rifampin 600 mg (10 mg/kg) daily against two experimental four month treatment arms using rifampin 1200 mg or 1800 mg. The TBTC Study 31 trial tests a SOC arm against two experimental rifapentine 1200 mg daily arms (instead of rifampin), with the second arm also adding moxifloxacin 400 mg daily. Results are expected in 2020 and, if successful, would be the first successful four month DS-TB trials to date.

In contrast to these trials which evaluate new or modified drug regimens, two trials are investigating treatment shortening strategies—PredictTB and TRUNCATE-TB. The presentation of TB is quite variable, so some individuals will likely need longer therapy than others no matter what regimen is used. Prospectively identifying this higher risk group can help target treatment duration appropriately.

PredictTB is based on a previously conducted trial that shortened therapy to four months among 394 pulmonary TB patients with less severe disease at baseline, defined as no cavity on baseline chest radiograph, and with sufficient early treatment response, defined as conversion of month 2 sputum culture to negative.17 Although this trial failed to achieve its predefined 5% non-inferiority margin, the researchers’ algorithm for patient stratification increased the treatment success rate in the four month arm from around 80% seen in other non-stratified four month treatment trials to 93% in their trial. The PredictTB trial revises the baseline disease stratification from chest radiograph to FDG-positron emission tomography/computed tomography (PET/CT) scan. Appropriate disease response is also changed from month 2 sputum culture conversion to reduction in PET/CT disease burden at month 1 and sputum GeneXpert cycle threshold increase at month 4.18 If the trial is successful, this strategy of shortening treatment only among those less severely diseased could also be evaluated among MDR-TB cases. In addition, investigation will focus on blood, sputum, urine, and other imaging biomarkers that correlate with PET/CT results and are more amenable to being scaled up globally.

In contrast to PredictTB, which only uses the four first line TB drugs, TRUNCATE-TB is a multi-arm, multi-stage trial that incorporates high dose rifamycins and various combinations of linezolid, clofazimine, levofloxacin, or bedaquiline into four two month treatment regimens. Patients who relapse are retreated with standard four drug therapy for six months with the hypothesis that this overall strategy will be non-inferior to standard six month treatment outcomes at two years.19 Both PredictTB and TRUNCATE-TB build on previously gathered data that show most patients are successfully treated before six months.3 The PredictTB strategy identifies these patients prospectively, then shortens treatment only for this cohort. The TRUNCATE-TB strategy shortens treatment for everyone to two months using a more intense intensive phase regimen, assuming that those who relapse will not develop resistance and can still be cured with subsequent standard six month therapy. In the end, a combination of these two strategies may be the most successful in maximizing cure rates with minimal treatment burden.

Drug resistant TB treatment shortening trials

An estimated 8% of TB patients globally have isoniazid resistant but rifampin susceptible TB. In 2018, WHO updated its treatment guidelines for isoniazid resistant TB to recommend levofloxacin, rifampin, pyrazinamide, and ethambutol for six months.20 More worrisome, however, is rifampin resistance. WHO estimates that only 55% of the estimated 558 000 incident cases of rifampin resistant (RR-TB) and MDR-TB in 2017 successfully completed treatment.1

The landscape of RR-TB and MDR-TB treatment was, for many years, defined by prolonged treatment (20-24 months) with older drugs that had substantial side effects and required daily injections for the initial eight months. This began to change with the publication of the “Bangladesh regimen” in 2010, which touted a successful nine month treatment regimen with daily injections only for the initial four months.21 However, because the trial was an uncontrolled observational study of sequential treatment arms, questions were raised about the validity of these results.

A slightly modified Bangladesh regimen was therefore tested within a rigorous randomized controlled trial (Standard Treatment Regimen of Anti-Tuberculosis Drugs for Patients with MDR-TB [STREAM]) comparing the nine month experimental arm with a 20 month SOC control arm that followed the 2011 WHO guidelines. This trial enrolled 424 participants and showed that the shortened nine month arm was non-inferior (using a 10% margin) to the standard 20 month arm.22 WHO had recommended the shortened nine month treatment regimen for uncomplicated MDR-TB (no additional resistance to fluoroquinolones or injectable agents) in 2016 and affirmed this recommendation with the results of the STREAM trial. For patients ineligible for the nine month regimen, the recommended treatment duration remains 18-20 months.

WHO 2019 guidelines now recommend an all oral, bedaquiline-containing regimen to replace injectable agents as the preferred treatment regimen.2324 WHO’s move toward shorter and all oral regimens for MDR-TB is also reflected in the ongoing MDR-TB clinical trials. Almost all experimental regimens being tested (except for Opti-Q, which began in 2015) are six to nine months long, and most of the experimental arms are completely oral (except for Opti-Q and STREAM stage 2 six month arm intensive phase; table 1). The change in the recommended WHO SOC regimen in duration and composition has implications for ongoing trials using the now obsolete 20-24 month regimens for their control arm (Opti-Q, MDR-END, NEXT, endTB, and TB-PRACTECAL). If enrollment is ongoing, these trials face the decision of whether to update their control arm treatment regimen or duration mid-trial.

Even more problematic are the trials that have experimental regimens that are now obsolete, either because of their duration or inclusion of injectable agents. The Opti-Q trial is one of these but, notably, its primary endpoint is not final treatment outcome but the surrogate endpoints of time to culture conversion by month 6 and safety of higher levofloxacin dosing. The only other trial that still uses an injectable agent is the STREAM stage 2 trial six month arm. If the other trials with all oral six month arms are successful, particularly in the XDR-TB trials, then injectable agents will phase out of first line therapy regardless of the treatment duration and level of resistance (eg, MDR-TB plus fluoroquinolone resistance).

Among the MDR/XDR-TB trials, there are two other major differences between the trials: 1) whether the trial incorporates a concurrent control arm compared with a historical control; and 2) the treatment regimen and duration (six versus nine months). The MDR (non-XDR) TB trials all have a WHO standard control arm. Of the three XDR-TB trials, however, only endTB-Q has a concurrent WHO standard control arm. Both Nix-TB and ZeNix only include experimental arms with no concurrent SOC control arm, with the primary outcome being relapse or failure at 12 months after enrollment. The benefits of not including a control arm are a smaller sample size and a much shorter trial, since the WHO SOC control was 20-24 months at the time these trials began. However, these studies also incur considerable risk, because if the treatment success rate in the experimental arm is below 95%, the benefit of the experimental arm may not be clear. In the three contemporaneous fluoroquinolone DS-TB treatment shortening trials8910 for example, the per protocol favorable outcomes of the control arms, despite using the same regimen, varied from 88.7% (OFLOTUB trial) to 92.0% (REMoxTB trial) to 95.1% (RIFAQUIN trial; table 2). If the experimental arm favorable outcome was 85%, whether this was a success or failure may differ depending on which control arm was used. Applying control arm results of a separately conducted study can interject misleading comparisons owing to varying populations and methodologies, and thus comparing with a historical control increases the risk of unrecognized biases and confounding factors in the data.25

Table 2

Study control arm favorable outcome

View this table:


On 14 August 2019, the US Food and Drug Administration approved pretomanid, its second novel anti-tuberculous drug in 40 years (bedaquiline was approved in 2012). While additional agents against drug resistant TB are critically needed, the basis on which pretomanid was approved is somewhat controversial.2627

Pretomanid was approved only in combination with bedaquiline and linezolid to treat XDR-TB and non-responsive MDR-TB because this is how it was used in the Nix-TB trial. In that study, 95/107 (89%) patients had successful outcomes at one year and six months after the end of treatment.28 As previously discussed, the Nix-TB trial was a small single arm trial that relied on a historical control for comparison, which is problematic given that bedaquiline and linezolid were not included in many historical controls. In addition, all patients received the same three drugs, two of which are already known to be very active against drug resistant TB. For example, when single drug linezolid was added to a failing treatment regimen in XDR-TB patients, 34/39 (87%) converted their sputum culture to negative by six months of treatment.29 Thus, the individual potency of pretomanid against TB is unknown and more studies are needed to determine this.

Unfortunately, it is unlikely that any of the currently ongoing trials using pretomanid (SimpliciTB, TB-PRACTECAL, Nix-TB, ZeNix) will reveal pretomanid’s potency because of the lack of an appropriate comparison arm for this purpose. TB-PRACTECAL, for example, has a WHO SOC control arm but all three experimental arms contain the Nix-TB regimen just with or without moxifloxacin or clofazimine (table 1), leaving the same conundrum where the individual effect of pretomanid cannot be differentiated from that of bedaquiline and linezolid. Pretomanid is a bicyclic 4-nitroimidazole, which is the same class as delamanid, another novel anti-TB agent that is still trying to find its proper place in the landscape of treating drug resistant TB.30

Adjunctive and host directed therapies (HDT)

Augmentation of the host immune response is an additional strategy of growing interest to improve treatment efficacy. Current HDT candidates are largely repurposed drugs with pre-clinical evidence for counteracting Mycobacterium tuberculosis pathogenesis and survival, augmenting protective immune responses, and/or modulating detrimental inflammatory responses that exacerbate disease. These therapies encompass agents used to treat metabolic and psychiatric conditions, parasitic infections, cancer, and inflammatory conditions, and span various stages of pre-clinical and early clinical development.

Treatment endpoints of culture conversion and durable cure remain relevant to clinical trials of most HDTs. However, further considerations include how and when to utilize HDTs to complement antibacterial treatment, and how to incorporate clinical biomarkers to investigate or validate therapeutic mechanisms. For example, the IMPACT-TB study (NCT03891901) evaluating safety and pharmacokinetics of imatinib at varying doses with and without isoniazid and rifabutin in healthy adults, will measure myelopoietic effects, proposed as a key therapeutic mechanism of imatinib (table 3).

Table 3

Host directed therapy (HDT) TB treatment trials

View this table:

Stage 2 of the StAT-TB trial will evaluate pravastatin as an adjunct to HRZE against time to sputum culture conversion as well as the secondary endpoint of improvement of pulmonary function among individuals with drug sensitive pulmonary TB (NCT03456102). Relevant study populations and biomarkers will also depend on the targeted intervention. HDTs targeted to modulate detrimental inflammatory responses, such as dexamethasone (NCT03092817), may include or target individuals with TB meningitis where uncontrolled inflammation is closely associated with morbidity and mortality.

Finally, HDTs that act directly or indirectly via host responses to disrupt M tuberculosis pathogenesis may have the greatest impact in treating patients with high burden disease or DR-TB, where cure rates are lower and disease chronicity and treatment duration are longer.

Several HDT clinical trials, including a completed phase II study of adjunctive ibuprofen (NCT02781909; no results posted) and an ongoing phase II/III study of recombinant human IL-2 (NCT03069534)31 have targeted sputum culture conversion in MDR/XDR-TB. Vitamin D has been evaluated as an adjunct to standard therapy in multiple trials with conflicting results, suggesting that any effect relative to anti-tuberculous chemotherapy is modest at best. This topic has been reviewed, including the ongoing randomized trials.32 If results support robust efficacy endpoint data, we may anticipate future trials evaluating HDTs as a complementary treatment shortening strategy.

Capturing heterogeneity of drug responses in early phase clinical evaluation

When viewing the TB trials together (fig 2), it appears that almost all possible combinations of the newer drugs are tested for either six or nine months without a clear rationale for how one combination or duration was selected compared with another. With the number of new compounds and HDTs in development33 and the more limited number of patients available to enroll in trials, testing every possible combination as done previously is impractical. Improved methodologies are needed to allow rational selection of the optimal combinations and durations to take forward into clinical trials.

Fig 2
Fig 2

Ongoing active TB treatment shortening clinical trials with duration and drug composition of the experimental regimens tested

Current methods rely heavily on early bactericidal activity studies,34 which measure the decline in serial sputum colony forming units (CFU) over the initial 14 days of treatment as a measure of sterilizing potency. This approach is endorsed by the Global Alliance for TB Drug Development35 and the US FDA36 but EBA does not appear to reflect durable cure.37 Certain drugs, such as rifampin, pyrazinamide, and linezolid, have relatively low EBA but have been shown repeatedly to be instrumental in killing TB.

Other drugs, like ethambutol and the fluoroquinolones, have high EBA but are not assumed to be very potent clinically (ethambutol at currently used doses) or have not been successful in contributing to treatment shortening (fluoroquinolones). One possible explanation for this is that sputum based methodologies, like EBA, only sample changes in bacillary populations on airway or cavity surfaces rather than deep necrotic lesion compartments which harbor recalcitrant TB populations that are killed more slowly.

A drug’s ability to shorten treatment may be more related to how well it penetrates into necrotic, caseous lesions than airway surfaces.3839 For example, both pyrazinamide and rifampin diffuse into caseous lesions well whereas moxifloxacin does not.40 which may explain why pyrazinamide and rifampin have both contributed to DS-TB regimen treatment shortening but moxifloxacin has not. Combining drugs with different lesion pharmacokinetic profiles is therefore an important consideration in rational regimen design. Furthermore, relying only on sputum measurements to reflect the myriad changes occurring throughout the lungs on treatment is likely to be imprecise.

We are conducting a trial called Radiologic and Immunologic Biomarkers of Sterilizing Drug Activity in Tuberculosis (NexGen EBA; NCT02371681) that attempts to augment a traditional EBA study with PET/CT imaging biomarker data to additionally capture sterilizing activity of first line drugs and their combinations which have previously been clinically and pharmacokinetically characterized. While sputum based measurements average all changes across the lungs into one result, a PET/CT image of the lungs can quantitate differential changes in different lesions. We hypothesize that, owing to different lesion pharmacokinetics, different treatment effects—including early bactericidal activity important for initial culture conversion and sterilizing activity important for durable cure—can be captured in different lesion types. This trial has been completed and data analyses are ongoing, with results expected in 2020. With this knowledge, rational combinations of drugs and HDTs can be selected and taken forward to later phase clinical trials with better chances of success.

Preventive TB treatment

Updates to treatment of latent TB infection (LTBI)

Unlike DS-TB treatment guidelines, LTBI treatment guidelines have changed in the last decade. The biggest change was the introduction of isoniazid/rifapentine once weekly for 12 weeks as a successful LTBI treatment regimen.414243 In addition, rifampin dosed daily for four months was confirmed to be as effective as isoniazid for nine months but with better treatment completion rates and fewer side effects.44

Additional latent TB clinical trials are ongoing and span the scope of DS and MDR latent TB treatment, as well as HIV negative and HIV positive populations (table 4, figs 3and 4). Three DS latent TB trials are ongoing (2R2, ASTERoiD, and SCRIPT-TB), all of which test shorter, regimens of one to two months of higher dose rifamycins compared with the SOC regimen and vary in target LTBI populations and endpoints of completion/safety versus efficacy. Even if rifapentine based trials are successful in shortening LTBI treatment duration, uptake of rifapentine based regimens is currently limited by cost and availability,45 with only one manufacturer making this drug globally. This situation will hopefully change as generic manufacturers enter this market.46 Three ongoing trials are testing LTBI treatment regimens with MDR-TB exposure (table 4, fig 3, fig 4). Two trials, V-QUIN MDR and TB-CHAMP, both test levofloxacin for six months compared with placebo. The third trial, PHOENIx MDR-TB, compares six months of isoniazid against six months of delamanid.

Table 4

Ongoing latent TB treatment shortening trials

View this table:
Fig 4
Fig 4

Ongoing latent TB treatment shortening clinical trials with duration and drug composition of the experimental regimens tested

Fig 3
Fig 3

Timeline of ongoing latent TB clinical trials with start dates and projected end dates

Targeted treatment of incipient TB infection

While TB treatment and control efforts are currently dichotomized around “latent” and “active” TB infection, neither of the two available methods to identify latent TB infection (tuberculin skin test and interferon gamma release assays) distinguish the few asymptomatic individuals who will develop active TB (estimated to be 5-10% lifetime risk) from the majority of individuals who will never develop active TB. Treating all M tuberculosis latently infected individuals for months is not feasible in most areas of the world. Rather, evidence is emerging of a spectrum of TB disease activity, and new approaches attempt to predict asymptomatic individuals who will progress to active TB disease (incipient TB infection).

Many retrospective studies of biomarker signatures, particularly transcriptomic, have been published identifying incipient or active TB in different cohorts47 but these signatures generally do not work well across populations and more work is needed to identify consistent signatures across cohorts. In addition, retrospective correlations of biomarker signatures ultimately need to be confirmed prospectively.

The Correlate of Risk Targeted Intervention Study (CORTIS; NCT02735590) is currently evaluating a host transcriptional signature previously found to have 66% prognostic sensitivity and 81% specificity for incident TB disease within 12 months48 to see how well treatment with three months of weekly isoniazid/rifapentine reduces the incidence of TB in individuals who test positive.49 The ability to both shorten and target preventive TB therapy to the few asymptomatic individuals at greatest risk of TB progression will greatly increase the global feasibility of treating LTBI.

Treatment of HIV/TB co-infection

Active TB

Many of the ongoing TB clinical trials include individuals who are HIV positive (table 1, table 4). The recent WHO recommendation of nine months of treatment for MDR-TB also applies to individuals infected with HIV.23 The treatment of active TB in HIV co-infected patients is generally the same as in patients uninfected with HIV, but with additional attention to overlapping toxicities and drug-drug interactions between commonly used HIV and TB drugs, particularly rifampin. Rifampin is a potent inducer of cytochrome P4503A4, uridine diphosphate glucuronosyltransferase 1A1, and P-glycoprotein enzymes, and thus will typically reduce the levels of drugs metabolized by these enzymes. Some drugs should not be used in conjunction with rifampin, while others can be used with dose adjustments. Rifabutin is less of an enzyme inducer than rifampin and can substitute for rifampin to allow certain antiretroviral drugs to be used. Drug-drug interactions are complex, and current guidelines continuously update tables on recommended combinations and dosing.5051 The recommended first line treatment is to use a rifampin based TB regimen with an efavirenz based antiretroviral therapy regimen. Options include using raltegravir instead of efavirenz or a ritonavir boosted protease inhibitor in conjunction with rifabutin. More studies are needed to define better drug-drug interactions for the newer anti-TB drugs, such as pretomanid.

The other major consideration in HIV-TB co-infection is the timing of treatment. When HIV is diagnosed first, earlier initiation of antiretroviral therapy and isoniazid preventive therapy independently reduce the development of active TB and death.525354 When TB is diagnosed first, there is a clear mortality benefit to starting antiretroviral therapy shortly after anti-TB therapy, despite the increased risk of developing immune reconstitution inflammatory syndrome.555657 According to current guidelines,58 ART should begin within two weeks for those with CD4 cell counts <50 cells/mm3. For patients with higher CD4 cell counts, antiretroviral therapy should begin no later than eight weeks after starting anti-TB therapy because of known benefits of earlier initiation of antiretroviral therapy.545960

Latent TB

The treatment of LTBI in individuals infected with HIV is also like that of uninfected individuals, with attention paid to rifamycin drug-drug interactions and the possibility of an additional, even shorter one month regimen. Rifapentine is a key component of shortened LTBI treatment regimens, with rifapentine/INH once weekly×12 weeks currently recommended for HIV infected individuals if drug-drug interactions are manageable. The recent BRIEF TB/A5279 trial showed that in HIV infected individuals, weight based rifapentine and isoniazid 300 mg daily for one month was non-inferior in preventing TB compared with the standard isoniazid daily for nine months.61 A similar trial, HIV-NAT 225, started in Thailand in 2019 (table 4). Rifapentine/INH once daily for one month may become a recommended regimen in guidelines, but few antiretroviral drugs are recommended to be co-administered with rifapentine.50 Protease inhibitors and the CCR5 inhibitor maraviroc are contraindicated. Among the non-nucleoside reverse transcriptase inhibitors, only efavirenz is approved for co-administration, although a phase I drug interaction study with doravirine is ongoing (NCT03886701). Another trial is testing the steady state pharmacokinetic effects of once-weekly rifapentine with tenofovir alafenamide fumarate, a preferred backbone nucleoside reverse transcriptase inhibitor (NCT03510468). Among the integrase strand transfer inhibitors, only raltegravir is currently recommended to be dosed with rifapentine only at the once weekly dose. Another trial (DOLPHIN) tested the safety and pharmacokinetics of dolutegravir with rifapentine/isoniazid once weekly in HIV infected adults and presented preliminary results showing this combination was well tolerated and that dolutegravir 50 mg daily maintained HIV viral suppression.62 Additional trials testing similar combinations with rifampin are also ongoing.

Genotypic and phenotypic drug resistance testing

Substantial resources in TB drug development, including most current active and latent TB clinical trials, are directed against drug resistant tuberculosis (table 1, table 4).33 The diagnosis of drug resistance traditionally has been done phenotypically, based on laboratory culture of M tuberculosis in the presence of varying concentrations of antibiotics. This technique requires substantial laboratory infrastructure and, most problematically, returns weeks to months after an empiric treatment decision has already been made. Genotypic resistance testing, which is considerably faster, reproducible, and operable within more automated, less biohazardous formats, are more compatible with guiding prompt treatment decisions outside of reference laboratories.

Commercial kits have been available since the mid-2000s but still necessitated considerable laboratory infrastructure until the rollout of the Cepheid GeneXpert MTB/RIF (Xpert; Sunnyvale, CA, USA) in 2010.63 Xpert was the first fully integrated, automated genotypic test and provided a direct readout of TB and rifampin resistance within two hours. Recent advances to the GeneXpert platform include the Xpert MTB/RIF Ultra that is more sensitive 64 and the Xpert XDR that detects genotypic resistance to isoniazid, fluoroquinolones, and aminoglycosides.65 However, detection of point mutations in specific target regions does not capture all phenotypic resistance, will continue to lag identification of new genetic resistance mechanisms, and may become irrelevant with the changing landscape of preferred second line drugs.

To that end, whole genome sequencing (WGS) provides more comprehensive information and has high sensitivity and specificity for isoniazid, rifampin, pyrazinamide, and ethambutol compared with phenotypic resistance testing.66 WGS can also provide useful information to characterize mixed infections, differentiate relapsed infections from re-infections, and contribute to transmission tracing. Governmental programs in England, the Netherlands, and New York are shifting resources from phenotypic testing to WGS to save time and money.67

Recent advances in highly portable, miniature sequencing platforms may facilitate access to TB WGS in resource limited settings.68 Although WGS is typically done from a cultured sample of M tuberculosis, which adds two to four weeks to the result time, efforts are under way to perform analyses directly on sputum. Sputum, however, is complicated by human and other bacterial DNA.69 The interpretation of genotypic results is also not always straightforward and large standardized databases with user friendly software are needed.6870 Finally, although overall sequencing costs continue to decline, startup capital costs for the machines are still prohibitive.71 The notably faster time to results of genotypic over phenotypic testing make it far more clinically relevant and the test of choice for the future. As more advances are made and costs decline, WGS will likely become more and more accessible globally.


The international TB community jointly released DS-TB treatment guidelines in 2016.72 This effort was sponsored by the American Thoracic Society, the Centers for Disease Control and Prevention (CDC), and the Infectious Diseases Society of America and was endorsed by the European Respiratory Society and the US National Tuberculosis Controllers Association. The American Academy of Pediatrics, the Canadian Thoracic Society, the International Union Against Tuberculosis and Lung Disease, and WHO also participated. These guidelines note that treatment by directly observed therapy (DOT) by trained personnel compared with self-administered therapy (SAT) is the standard of care in most TB programs in the US and Europe. WHO released another version in 2017 with some updates primarily focused on patient care aspects.73 The WHO guidelines allow for SAT but conditionally recommends either DOT in the community or home by trained providers or video observed therapy over SAT. For drug resistant TB, the primary guidelines are from WHO, which released its latest version in 2019.2324 Recent reviews have been published for the management of drug resistant TB.7475

For the treatment of LTBI, WHO and the CDC have released updated treatment guidelines in 2020.7677 Both guidelines recommend targeted testing and treatment of LTBI with slight differences in the recommended treatment regimens. The regimens recommended by both guidelines are: 1) isoniazid with rifapentine once weekly for three months, and 2) isoniazid with rifampin daily for three months. In addition, WHO also recommends isoniazid daily for six or nine months as a primary regimen, whereas the CDC now lists these as alternative rather than preferred regimens. In contrast, the CDC recommends rifampin daily for four months as a preferred regimen, whereas WHO lists this as an alternative. The WHO guidelines also include two other regimens that the CDC guidelines do not: 1) isoniazid with rifapentine daily for one month, as an alternative regimen; and 2) for HIV infected individuals in high TB transmission settings, isoniazid daily for 36 months.

Emerging treatments

Owing to the number of ongoing TB clinical trials, the next several years will hopefully be a time of marked changes in treatment guidelines to include newer drugs, shorter regimens, and perhaps even HDTs. Future trials being planned for DS-TB will not only include treatment shortening but will also include newer drug such as bedaquiline. One recent result for MDR-TB treatment is the ACTG5343 (DELIBERATE) trial that tested the safety of combining bedaquiline and delamanid to treat MDR-TB, focusing on cardiac QTc prolongation since both drugs are known to have this side effect. Early results showed patients had a mean QTcF prolongation from baseline of 11.9 ms in the bedaquiline arm, 8.6 ms in the delamanid arm, and 20.7 ms in the combined arm. Thus, the combination of bedaquiline and delamanid appears to be safe from a cardiac standpoint, with only a modest additive effect on QTcF prolongation, in patients with normal QTcF interval at baseline.78


With the recent advances in TB treatment and diagnostics, and many ongoing trials, the next five years will likely see major changes to TB treatment approaches. However, the future trend is clear and is already arriving for drug resistant TB—shorter, all oral regimens incorporating newer drugs. The question is which drugs, how short, and what patients. Will treatment of TB continue to be a one size fits all with the same regimen and duration recommended for all patients? Will patients be stratified into risk categories with treatment regimens and durations tailored for each? How will all the novel TB compounds currently in clinical development be incorporated effectively and efficiently to create optimal treatment combinations? What role will HDTs play in shortening treatment and reducing functional impairment? Is a one month or even a two week treatment regimen for TB realistic in the next 10 years? The results of ongoing and future trials will help answer these questions. However, even with these biomedical advances toward shorter, oral regimens, we continue to face fundamental challenges along the path to comprehensive, accessible, effective, and reliable delivery of treatment. Concerted efforts of community advocates, researchers, drug companies, TB programs, healthcare delivery systems, and innovative technologies around patient centered strategies will be critical to implement these advances and end the TB epidemic.798081

Research questions

  • What biomarkers will predict the development of active TB among LTBI patients to enable targeted treatment that is globally scalable?

  • What biomarkers will predict treatment outcomes early during treatment of active TB that will allow for more personalized treatment algorithms rather than the current one-size-fits-all approach?

  • What biomarkers will predict the development of relapsed TB after completion of treatment for active TB?

  • How should novel and existing drugs be combined to determine the best regimens to bring forward into phase III clinical trials testing of shortened treatment regimens for drug sensitive and drug resistant TB?

How patients and the public were involved in the creation of this article

This article was reviewed by two people who had been successfully treated for TB, one from the US and one from South Africa. The patient from the US received treatment by DOT, which “felt both like a hassle (because I had two young children and was busy caring for them at home) and like a good safeguard against forgetfulness (as one prone to getting caught up in caring for the kids).” She was interested in research on how current treatments could be implemented more successfully The patient from South Africa stated, “I don’t have any comments and any questions or comments I thought of while reading were inevitably answered further down.”


  • Series explanation: State of the Art Reviews are commissioned on the basis of their relevance to academics and specialists in the US and internationally. For this reason they are written predominantly by US authors

  • Acknowledgment: This research was supported by the Intramural Research Program of the US National Institutes of Health, National Institute of Allergy and Infectious Diseases.

  • Contributorship: YLX, CEB, and RYC developed the idea for this review. AL, YLX, and RYC performed the literature search. AL, YLX, CEB, and RYC wrote the review. RYC is the guarantor. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.

  • Competing interests: The BMJ has judged that there are no disqualifying financial ties to commercial companies. The author declares the following other interests: none.

  • Further details of The BMJ policy on financial interests are here:

  • Provenance and peer review: commissioned; externally peer reviewed.


View Abstract

Log in

Log in through your institution


* For online subscription