London School of Hygiene & Tropical Medicine

Tuberculosis (TB), caused by the bacterium Mycobacterium tuberculosis, is a major global health concern, but certain regions, including South Africa, face particularly high burdens of this disease. This is exacerbated by the emergence multidrug-resistant TB (MDR-TB), which is harder to treat, requiring longer, more complex regimens. Bedaquiline, an antibiotic approved in 2012, offers hope, being effective against MDR-TB infections. However, despite its effectiveness, there have been recent reports of increasing drug resistance to bedaquiline in areas of South Africa. 

Prof Keertan Dheda, Professor of Respiratory Medicine at the University of Cape Town and Professor of Mycobacteriology and Global Health at LSHTM, is a world leader in the field of TB diagnosis and transmission. In an interview with Dr Richard Wall, he discusses the impending threat of bedaquiline-resistant TB in South Africa and the steps required to minimise its spread.

Why is bedaquiline resistant-TB becoming an issue in South Africa?

Although bedaquiline resistance is now increasing in South Africa, it’s an emerging problem in many countries. However, the fact that bedaquiline has been used in some parts of South Africa for over 10 years now could be a contributing factor but there are also other countries, like Mozambique, where bedaquiline has been used for much less time where bedaquiline resistance levels are even higher. The long half-life of bedaquiline could also be a contributing factor because the drug lingers in the system for many months after treatment has stopped. Thus, M. tuberculosis might be exposed to sub-optimal levels during relapse or reinfection. Another possibility is inadequate drug penetration of thick wall cavities and other TB lesions, leading to differential exposure of pathogen to drug, resulting in a pharmacokinetic (PK) mismatch.  

What is the occurrence of bedaquiline resistance in South Africa? 

 In unselected rifampicin-resistant isolates from newly diagnosed patients, bedaquiline resistance in South Africa ranges between 5% and 10%. The highest levels are seen in the Province of the Western Cape, where resistance is close to 10%. Perhaps this is not that surprising given that the bedaquiline access programme started in the Western Cape ~12 years ago. Given these alarming levels of resistance, phenotypic bedaquiline resistance testing is now routine in South Africa for all rifampicin-resistant isolates. 

Since bedaquiline is primarily used for MDR-TB, what other drugs are these infections resistant to?

Clinical isolates that are resistant to rifampicin often have concurrent resistance to fluoroquinolones in about 20-30% of cases, and pyrazinamide and ethambutol resistance in individuals with MDR-TB is around 50% or more. There are relatively low levels of resistance to drugs like linezolid and pretomanid (the other group A second-line drugs). Interestingly, we observed high levels of bedaquiline resistance in the absence of fluoroquinolone resistance.  

What causes this drug resistance? 

 Unfortunately, as we all know, resistance is inevitable, and we can only control the rate at which resistance occurs and progresses. For example, the 2012 PETTS multinational study, which evaluated drug-resistance across several countries over a period of time, showed that even when individuals with MDR-TB were near perfectly adherent, fluoroquinolone resistance developed in ~5% of such individuals. For many years we incorrectly blamed patients and healthcare systems for the emerging drug resistance to M. tuberculosis. For example, patients were often blamed because they were not taking their drugs correctly or because health systems no longer stocked the required drugs. However, even if these factors were eliminated, the development of resistance remains inevitable.  

So, what other factors could drive this resistance? 

 More recently we’ve realised that there are several other, more important factors, that are playing a role, including differential host drug metabolism and absorption, which drive pharmacokinetic variability; differential penetration of drugs into TB cavities and lesions; and mycobacterial factors such as efflux pumps that can expel the drug from the mycobacteria. All these factors drive PK mismatch, which essentially is the driver of resistance. Of course, then comes the second phase where there is transmission of these resistant strains, which eventually becomes the dominant mechanism of resistance acquisition. For those interested, we recently published an open access review paper in Nature Reviews Disease Primers, outlining these concepts. 

What is the impact of bedaquiline resistance on the World Health Organisation’s updated standard of care for MDR-TB (bedaquiline, pretomanid, linezolid ± moxifloxacin)? Does the reduced treatment duration lead to a higher likelihood of resistance? 

 The shorter duration of treatment is very effective because these ‘next-generation’ drugs are more potent and mycobactericidal, and it is likely that soon (over the next 5 to 10 years) the length of the regimen could be reduced to as little as 2 months or less. There’s no evidence that the newer, shorter regimens are driving more resistance due to their shorter administration period. However, the reduced treatment duration has been transformational for patients, healthcare workers and healthcare systems. In my own practice, it is amazing how much more tolerable and user-friendly these regimens are. They are well-received by patients and dropout rates have reduced considerably. We are no longer using toxic drugs that can cause hearing loss and patients no longer have to endure painful injections. This also means patients are more compliant with their medication, and these shorter pan-oral regimens are much easier for healthcare systems to manage (rather than injectable drugs). 

How do we reduce the occurrence of bedaquiline-resistant tuberculosis in South Africa and beyond? 

 As long as PK mismatch is minimised, resistance can be minimised, but it cannot be eliminated. The solution is complex and multifactorial. We need to address many health system and patient-level factors, and this will include strengthening healthcare systems, educating healthcare workers, reducing substance abuse amongst patients and so forth. We also need to better understand drug-specific PK variability and mismatch, and the role of therapeutic drug monitoring (especially for drugs like linezolid). Our own research is focusing on the differential penetration of second-line drugs into TB lesions and cavities. We previously demonstrated that for the older-generation of second-line drugs, including the fluoroquinolones, there was significantly poorer penetration into the centre of the cavity (compared to the outer cavity wall) where there was a very high concentration of mycobacteria, and this actually drives drug resistance. Shutting down efflux pumps is an obvious area that needs to be targeted. Finally, very little is known about epigenetic mechanisms that drive drug resistance, and this requires further investigation. 

Should there be more emphasis on testing individuals, diagnostics or bespoke treatment? 

 Yes, diagnostics also play a key role in antibiotic stewardship. At present, it’s somewhat ‘hit and miss’ and treatment is quasi-empiric. We only have rapid readouts for rifampicin, isoniazid and fluoroquinolones in the form of low complexity automated or semi-automated nucleic acid amplification tests. This has already made it much easier to design regimens. However, targeted sequencing technologies like AmPORE and GenoScreen (Deeplex Myc-TB) are the new kids on the block. These technologies are an attractive option for TB endemic countries because they significantly reduce costs associated with phenotypic drug testing and setting up infrastructure, like BSL3 labs. However, there are several potential drawbacks, including lack of readouts in paucibacillary disease and limited positive predictive values for certain drugs like bedaquiline (although the rule-out value is good). More importantly, whether these assays will have a clinical impact at the patient level remains unproven. I have no doubt that more routine and widespread susceptibility testing will reduce the emergence of resistance amplification. 

How important is it to identify asymptomatic cases? 

About 50% of the total burden of TB is asymptomatic (also called subclinical TB), and drug-resistant TB is no different. It is critical that active case finding be dramatically scaled up in all settings, as this will result in earlier diagnosis of drug-resistant TB and will play a major role in reducing amplification of resistance, including primary transmission. Through the LSHTM and UCT, we are undertaking large active case finding studies across four sub-Saharan African countries, where ~16,000 people have been screened using battery-operated Xpert Ultra (XACT-3).  

What should the scientific community do differently? 

Several things come to mind. I will touch on a few of them. Roll-out of community-based active case finding to interrupt transmission early on is critical, as is improved and early diagnosis, developing efflux pump inhibitors, and addressing patient level and healthcare system factors that drive resistance amplification. Ultimately, an effective vaccine will have a major impact on drug-resistant TB. 

How important are new drugs (with novel mechanisms of action) and vaccines for tuberculosis treatment and prevention? 

New antitubercular drugs will no doubt play an important role, but as already outlined, this will only be one piece of the puzzle. If we do not pay attention to the prevention of resistance amplification, these newer drugs will be ‘lost’ very quickly. Encouragingly, there are several new trials going ahead. We are all awaiting the results of the phase 3 M72 trial and the phase 2b MTBVAC study (a live attenuated vaccine). It is unclear which approach will be effective. It remains to be seen whether the phase 3 M72 trial will demonstrate appreciable vaccine efficacy. However, even 50% to 60% efficacy will translate into a huge reduction in disease burden. 

How is your research helping to understand and minimise bedaquiline resistance?

At UCT, in collaboration with the LSHTM, we have developed and optimised, for the first time, a lung challenge model where live BCG is administered to healthy individuals within a broader framework of different clinical susceptibility phenotypes. Firstly, this will serve as an excellent model to test, validate and develop new diagnostics. Secondly, in the future, it could serve as a platform to evaluate new drugs and vaccines thus accelerating their development. In particular, we are working on an inhaled formulation of BCG. This is following the publication of multiple animal studies showing that BCG, when delivered via the lung route, in contradistinction to the intradermal route, is highly effective in preventing TB in monkeys and other animal models. We are currently conducting a UK MRC-funded study comparing intradermal versus intra-lung challenge in humans, and in tandem we are developing an inhaled BCG formulation to be tested in humans. The UCT Lung Institute has a strong collaborative link with LSHTM through the UCT LSHTM capacity development initiative and there’s a great deal of interesting ongoing work that involves new drug trials, basic science studies, and the lung challenge work. However, there are many excellent groups all over Africa that are doing some outstanding and innovative work. There’s a lot to be achieved and we need a lot more people and more funding to be invested in TB research. 

Why is being part of LSHTM helpful for your work and how can it help you to achieve your goals?  

LSHTM can play a critical role in several ways. These include directly driving scientific advances (and there are many excellent groups at LSHTM doing this already), providing data management skill sets for multinational agencies like the WHO to make recommendations, teaching and training the next generation of researchers in TB endemic countries, and better capacitating institutions in TB endemic countries. This is already paying dividends in many countries, including South Africa, which is investing more into research and spending on TB. They have realised that they must also invest in solving their own problems. This not only saves lives but is hugely cost-effective. For example, TB is the most common cause of death in South Africa and sets back its GDP by ~2%, which is more than the entire health spend on all conditions in the country. Thus, it’s a bit of a no-brainer that investment in TB prevention, management, and research pays huge dividends. 

Drug-resistant TB has become a very important component of AMR and contributes to about 25% of the total economic costs attributable to AMR. Recently, the WHO has designated rifampicin-resistant M. tuberculosis as one of the ‘critical’ priority AMR pathogens together with Acinetobacter baumannii and Enterobacterales in terms of R&D and public health measures.