Xpert MTB/RIF diagnostic test for TB: a global update

Key points

• The Xpert MTB/RIF assay is a new test which allows the diagnosis of tuberculosis within two hours from a single sputum sample.
• Its use as the initial diagnostic test is recommended by the World Health Organization for all people with suspected multidrug-resistant TB (MDR-TB) and for all people with HIV who have suspected TB.
• Its use is also recommended as a follow-on test after smear microscopy where MDR-TB or HIV is of lesser concern, especially in suspected smear-negative TB.
• The roll-out of the Xpert MTB/RIF test is being supported by funding from several donors across a wide range of countries with a high prevalence of MDR-TB and/or HIV.
• The test improves the rate of diagnosis of both drug-sensitive and drug-resistant TB.
• The test also supports active case finding, by both improving the diagnostic yield and by limiting the waiting time for results. This reduces loss to follow-up.
• Very rapid initiation of TB treatment has been achieved by use of Xpert MTB/RIF supported by Community Health Facilitators, who provide the link between the labs, the clinics and the community health workers.
• Use of Xpert MTB/RIF has also been shown to reduce the delay in treatment initiation in smear-negative TB patients.
• While Xpert MTB/RIF and culture testing may have similar positivity rates, Xpert MTB/RIF significantly decreases the turnaround time for diagnosis of rifampicin resistance and can be placed in decentralised settings.
• Testing for susceptibility to isoniazid and second-line drugs should be done for all patients identified as rifampicin-resistant by the GXP.

This edition of HATIP was kindly supported by the HIV/AIDS Department of the World Health Organization. Thanks to Dr Linh Nhat Nguyen and Dr Wayne Van Gemert of the Stop TB department (WHO) and Dr Eyerusalem Negussie of the HIV/AIDS department (WHO) for their review of this edition.

Background: Improving TB diagnosis through use of the Xpert MTB/RIF assay

Only 66% of the estimated number of incident TB cases were diagnosed and notified to WHO in 2011. The low rate of case detection is due in part to an overreliance on sputum smear microscopy, a diagnostic test with low sensitivity, to diagnose TB in many low- and middle-income countries where a high HIV prevalence makes the use of smear-microscopy ineffective.

Smear microscopy is a less sensitive means of detecting TB bacteria in people living with HIV. This is because people with more advanced HIV disease tend to have lower levels of TB bacteria in their sputum. As a result, a larger proportion of people with HIV will test smear negative than in the general population. In order to diagnose TB in these smear-negative suspects bacterial culture has been used, but bacterial culture can take weeks or months to deliver a result, by which time a person with HIV may have died.

Diagnosis of drug resistance also remains a particular challenge, given the infrastructure, biosafety requirements and human resources needed to conduct traditional culture and drug susceptibility testing (DST).

The Xpert MTB/RIF test is able to diagnose tuberculosis (TB) in a single test from a sputum sample. It can also detect resistance to rifampicin, which in many settings is a reliable proxy for multidrug-resistant TB (MDR-TB). However, culture and conventional drug susceptibility testing against second-line anti-TB drugs are still necessary to confirm or exclude XDR-TB and to monitor response to MDR-TB treatment.

One of the main benefits of this novel diagnostic is that it gives results in two hours, allowing for rapid diagnosis and an earlier start of treatment whereas TB diagnosis from microscopy takes up to two days, and culture and DST take weeks for detection of TB drug resistance.

The Xpert MTB/RIF test also has a sensitivity similar to culture on solid media, far superior to that of conventional microscopy. The high sensitivity of Xpert MTB/RIF makes this diagnostic useful in the diagnosis of TB in people with HIV co-infection, where the sensitivity of smear microscopy alone is low.

The development of the Xpert MTB/RIF assay for the GeneXpert platform was completed in 2009 and endorsed by the World Health Organization (WHO) in 2010. 

The successes and challenges of implementing the Xpert MTB/RIF test received considerable attention at the 43^rd Union World Conference on Lung Health held in Kuala Lumpur in November 2012.

Previous editions of HATIP have reported on the Xpert MTB/RIF test and its benefits in active case finding (HATIP 191, May 2012), as well as how it can be used with other TB diagnostic tools (HATIP 193, May 2012).

This edition examines data and lessons learnt from programmes where Xpert MTB/RIF has already been implemented in high-burden TB settings in countries such as South Africa, Kenya, India, Cambodia and Botswana. These findings were presented at the Union conference and also at the Conference on Retroviruses and Opportunistic Infections (CROI) in March 2013.

World Health Organization recommendations on use of Xpert MTB/RIF

The World Health Organization recommends that:¹

• Xpert MTB/RIF should be used as the initial diagnostic test in individuals suspected of having MDR-TB or HIV-associated TB. (Strong recommendation).

“These are the two groups of people who will benefit most from this diagnostic. HIV-infected individuals co-infected with TB are more difficult to diagnose with smear microscopy, resulting in delays in diagnosis, which can lead to death in HIV-infected individuals. We also want to diagnose people suspected of having MDR-TB as quickly as possible to ensure that they are started on effective treatment without delay, which in turn reduces the risk of transmission of MDR-TB in the community,” Dr Christopher Gilpin of the WHO Stop TB Department told the conference.²

• Xpert MTB/RIF may be considered as a follow-on test to microscopy in settings where MDR-TB or HIV is of lesser concern, especially in further testing of smear-negative specimens. (Conditional recommendation acknowledging major resource implications.)

These recommendations apply to the use of Xpert MTB/RIF in sputum specimens only, as data on the utility of Xpert MTB/RIF in extra-pulmonary specimens are still limited. Based on the generalisation of data from adults, these recommendations also apply to children. Although studies have shown that multiple sputum specimens increase the sensitivity of the Xpert MTB/RIF, WHO recommends the use of one sputum specimen for diagnostic testing, given the resource implications of using multiple specimens.

WHO acknowledges the need for a review of the accuracy of GXP in detecting extra-pulmonary TB and paediatric TB, and organised an Expert Group meeting to evaluate the evidence base in May 2013.

Global update on the use of Xpert MTB/RIF

Globally, 966 GXP instruments, comprising of 5017 modules and 1,891,970 Xpert MTB/RIF cartridges had been procured in the public sector, as of 31 December 2012. The majority of the modules (2,049) and cartridges (1,108,140) have been procured by South Africa, where Xpert MTB/RIF testing as the initial diagnostic is national policy. Other countries which are investing heavily in the procurement of the GXP platform include Brazil, Cambodia, the Democratic Republic of Congo, India, Kenya, Tanzania and Zimbabwe.³

A consultative meeting was held in April 2012 for implementers and technical partners in the Xpert MTB/RIF roll out, with the aim of sharing current information and experiences on the use of the diagnostic, developing a greater understanding of implementers' needs and to plan improved co-ordination between country programmes, donors and partners. 

The major recurrent points raised by the implementers included:⁴

• User satisfaction: Implementers indicated an overall high level of satisfaction, describing Xpert MTB/RIF as “fast, easy-to-use, modern and much less cumbersome than conventional TB diagnostic techniques.”

• Need for price reduction: The price of the test cartridge was repeatedly stated as the main obstacle to an accelerated and sustainable roll-out of the test (but it should be noted that this consultation took place before the price reduction to USD 9.89).

• Diagnostic and clinical algorithms: Investment in time and resources to develop and implement effective diagnostic and clinical management algorithms incorporating the new technology is needed.

• Errors and invalid results: In a small but significant number of sites, errors and invalid results in the initial phase of implementation raised anxiety regarding use of Xpert MTB/RIF.

• Scaling up treatment capacity to match diagnostic capacity: Treatment of rifampicin-resistant TB cases that are diagnosed by Xpert MTB/RIF was a major concern in many sites which did not have the effective second-line drugs to treat newly diagnosed patients. Correct diagnosis in the absence of appropriate treatment nevertheless allows for patients to protect the health of their families.

• Need for innovation: Technical innovation is needed to allow for Xpert MTB/RIF to be used in more settings at levels closer to the point of care, where lack of reliable electricity and high temperatures may otherwise prevent the reliable use of the test. Implementation of electronic information management systems and mHealth initiatives through the use of mobile telephones will facilitate stronger links between diagnosis and follow-up care.

• Private sector access: Adoption of Xpert MTB/RIF by the large private sector in many high-burden countries would be highly beneficial for increasing patient access to rapid and reliable diagnosis, and at the same time replacing technologies that are not endorsed by WHO. The establishment of collaborations between private providers and national TB control programmes would be mutually beneficial, allowing for private providers to access concessional prices and for national TB control programmes to ensure that patients whose TB is diagnosed in the private sector are duly reported and subsequently registered for appropriate treatment.

Addressing funding constraints

The high price of the GeneXpert platform and test is the main barrier to the sustainable scale-up of the technology. Since December 2010, a single Xpert MTB/RIF cartridge cost USD 16.86 at the concessionary price for the public sector, including non-governmental organisations, in 145 eligible countries.

In August 2012 the price of cartridges was reduced to USD 9.98 for the public sector in the 145 countries due to a ‘buy down’ by PEPFAR, USAID, UNITAID and the Bill and Melinda Gates Foundation, from Cepheid, the only manufacturers of the GeneXpert technology. This concessionary price is not available to the for-profit private sector.⁵ (See a full list of eligible countries on the FIND website.)

Approximately 1.4 million Xpert MTB/RIF test cartridges and over 200 GeneXpert instruments will be made available from 2013 -2015 through the TBXpert Project to 21 countries including Bangladesh, Belarus, Cambodia, the Democratic Republic of Congo, Ethiopia, India, Indonesia, Kenya, Kyrgyzstan, Malawi, Mozambique, Myanmar, Nepal, Pakistan, Philippines,  Moldova, Swaziland, Uganda, Tanzania, Uzbekistan and Vietnam.

The USD 25.9 million project is funded by UNITAID and implemented by the WHO Stop TB Department and the Stop TB Partnership, including partners such as the Global Laboratory Initiative (GLI), TB REACH, the EXPAND-TB Project, Interactive Research and Development (IRD) and the African Society for Laboratory Medicine (ASLM). By linking this broad network of partners and existing initiatives for TB laboratory strengthening and innovative approaches to expand access to vulnerable populations, the TBXpert Project hopes to build the capacity of countries to receive and effectively use the GXP technology.

Information on the availability of the GXP technology, country by country, is available through a WHO-managed mapping website.

Xpert MTB/RIF roll-out

WHO Stop TB Department’s Laboratories, Diagnostics and Drug Resistance Unit has been collecting data from early implementers of Xpert MTB/RIF since 2011. The data collection is intended to monitor laboratory workload, numbers of tests being performed and results, and main logistical and operational problems being encountered, on a facility basis.

Evidence regarding the impact of the new diagnostic continues to be gathered. A register of Xpert MTB/RIF research projects is also being compiled. Twenty-six studies were either underway or already completed in February 2013, including a large randomised trial to evaluate the use of Xpert MTB/RIF in primary health clinics for TB diagnosis in people living with HIV. The TB Neat study is recruiting participants in South Africa, Tanzania, Zambia and Zimbabwe.

In a recent editorial introducing a Cochrane Collaboration Diagnostic Test Accuracy Review, Professor Elizabeth Corbett of the London School of Hygiene and Tropical Medicine, and Dr Danielle Cohen of the Malawi-Liverpool-Wellcome Clinical Research Programme, raised a number of questions regarding the use of Xpert MTB/RIF that still need to be addressed by research.

• How can Xpert MTB/RIF be integrated into systematic screening activities, particularly in populations at very high risk of TB, such as people living with HIV?

• At what level of the health system should access to Xpert MTB/RIF be prioritised? How does it perform in peripheral health centres?

• How can TB programmes best use this valuable new tool for detection and management of rifampicin resistance?

• How can we effectively evaluate this test in routine practice, where access to culture is generally non-existent and where patient-important outcomes are more important?⁶

Research presented at the 2012 World Lung Health meeting began to address some of these questions, relating to diagnostic yield, to use of Xpert MTB/RIF to support TB and drug-resistant TB diagnosis in peripheral health centres, and in relation to diagnosis of drug-resistant TB.

Diagnostic yield

Xpert MTB/RIF implementation is intended to improve case detection of both drug-sensitive and rifampicin-resistant TB cases.

Studies presented from South Africa, India and Cambodia showed that in settings with a high burden of TB and DR-TB, Xpert MTB-RIF resulted in encouraging improvements in diagnostic yield. The studies also showed that increases in diagnostic yield could be achieved when testing was decentralised to lower levels of the health system.

South Africa

In Cape Town, South Africa, the phased introduction of Xpert MTB/RIF across five of the city’s eight sub-districts resulted in an increase of approximately 20% in diagnostic yield in the first quarter after introduction, although this increase was not sustained in subsequent quarters.⁷

All TB suspects were screened according to the following algorithm:

• Smear and culture sub-districts: two smears in low-MDR risk patients, and two smear plus culture and drug susceptibility testing in high-MDR risk patients.

• Xpert MTB/RIF sub-districts: two sputum samples collected from all suspected TB cases. If sample 1 was negative and the patient was HIV negative the second sample was discarded. If sample 1 was negative and the patient was HIV positive the second sample was submitted for culture. If sample 1 was positive but rifampicin-sensitive, a confirmatory smear was carried out on sample 2. If sample 1 was positive and rifampicin-resistant, sample 2 was submitted for smear, culture, LPA and second-line drug susceptibility testing.

The 2011 fourth-quarter mean diagnostic yield across the sub-districts for DS-TB was found to be 20.3% for the Xpert MTB/RIF algorithm compared to 16.9% for the standard-of-care smear microscopy/culture algorithm. However, this difference was not sustained for the second time period. For the second quarter of 2012, the Xpert algorithm’s mean diagnostic yield decreased to 16.8% compared to 16.3% for the smear/culture algorithm.

The mean MDR-TB yield for the Xpert MTB/RIF algorithm was 7.3% and 5.3% in the first and second time periods respectively compared to 4.7% and 5.3% for the smear/culture algorithm. “The increase in the MDR-TB diagnostic yield suggests some benefit for DR-TB, but extra time points would be needed,” said Dr Naidoo.

In this study, it is important to note that the yield refers to the algorithm and that the number of sub-districts implementing the Xpert MTB/RIF testing algorithm increased from two in the fourth quarter of 2011 to three in the first quarter of 2012. This may have had an effect on the result. In addition, not all suspected TB cases in the districts implementing the Xpert MTB/RIF algorithm underwent Xpert MTB/RIF testing, making the decline in yield in the second quarter of 2012 hard to interpret.

India

A 7.2% increase in the diagnostic yield when using Xpert MTB/RIF for TB case finding compared to smear microscopy was found in a feasibility and impact study of using the test at decentralised laboratories in India, as reported by Dr Naraj Raizada, Project Manager of Xpert MTB/RIF projects for FIND in India.⁸

All patients with suspected pulmonary TB (PTB) and drug-resistant TB (DR-TB) were offered a single Xpert MTB/RIF test situated at 18 sites in varying settings across India serving a population of 8.7 million people. 22,345 pulmonary TB and 1,738 DR-TB suspected cases were tested between March and October 2012.

The study compared the diagnostic yield of TB before and after the establishment of the Xpert MTB/RIF testing facilities. Of the 9124 suspected TB cases tested using smear microscopy at baseline, 1312 confirmed cases of TB were detected (a 14.3% smear positivity rate). This resulted in an average of 29 TB cases per site per month. This increased to 47 TB cases per site per month when using Xpert MTB/RIF. 4422 of the 22,345 TB suspected cases were confirmed as having TB when the Xpert MTB/RIF tests were used (a 20% positivity rate).

However, they also conducted an internal comparison of diagnostic yield by performing smear microscopy on the same sputum specimens used in the Xpert MTB/RIF testing. Smear microscopy yielded a positivity rate of only 12.8%, compared to the 20% positivity rate using Xpert MTB/RIF on the same sputum specimens. 1559 cases of TB were missed using smear microscopy. Xpert MTB/RIF also diagnosed 569 cases of rifampicin-resistant TB (293 among the suspected TB cases and 276 among suspected DR-TB cases).

This study also looked at the feasibility of deploying Xpert MTB/RIF at decentralised level and found that each of the 18 sites could do 5000 tests per month with minimally modified infrastructure and human resources. The study found that if two hours of back-up electricity could be supplied, it is sufficient to avoid major cartridge wastage. Ninety-three per cent of patients had interpretable results from the first test and this could be increased to 99% on a repeat test.

Cambodia

The Cambodian National TB Programme (NTP) also found that the use of Xpert MTB/RIF increased the diagnostic yield of detecting TB in operational settings. When the records of Xpert MTB/RIF assays were examined for 2011 and 2012, TB was confirmed in 348 of the 1375 (25.31%)  tests done, compared to the national average of 10% in cases of standard-of-care use of smear microscopy to detect TB.⁹ 39 of the 348 (11.2%) TB cases were found to be rifampicin resistant.

The Xpert MTB/RIF tests had been conducted on sputum samples from a variety of settings including 284 samples from prisons, 73 as part of active case finding for DR-TB among high-risk groups, 223 in routine MDR-TB finding and 795 among close contacts of TB cases.

Using Xpert MTB/RIF in active case finding

Some of the most exciting results presented at the conference on active case finding concerned the impact of implementing the Xpert MTB/RIF test in South Africa and Kenya.

South Africa

Professor Harry Hausler of the TB/HIV Care Association reported on TB REACH-funded case finding in the Sisonke district of Kwazulu-Natal province, South Africa. The district in which the project is taking place was the first in South Africa to use Xpert MTB/RIF to test all people with suspected TB.¹⁰

“Xpert has really facilitated case finding in this deeply rural district of Kwazulu-Natal,” said Professor Hausler.

Five sub-district labs are equipped to run the test, and mobile teams routinely screen for TB symptoms and collect sputa for testing at these labs. The results are then communicated to clinics and community health workers, who are responsible for recalling patients with a positive Xpert MTB/RIF result, as well as organising the screening of contacts. There are approximately 800 community health workers in Sisonke district.

Between October 2011 and October 2012, 21,712 tests were carried out, diagnosing 1471 TB cases, 9.9% of which were rifampicin-resistant. Smear testing is carried out at baseline in order to monitor treatment; one-third of all these cases were smear-negative. 93% of the TB cases had a recorded HIV test result, and 72.4% of these patients were HIV-positive.

During 2012 active case finding almost doubled the number of samples referred for Xpert MTB/RIF testing compared to the last quarter of 2011, while the proportion of TB cases detected among those tested fell from 12 to 4% in the last quarter.

Although mobile health teams play an important role in the drive to improve active case finding, Prof. Hausler noted that the vast majority of TB cases continue to be identified through passive case finding within health facilities (91%), and that door-to-door case finding produces a very low yield of TB cases. Just 1% of those referred for Xpert MTB/RIF testing as a result of door-to-door case finding were found to have active TB, compared with 9% of presumptive TB cases referred for testing within health facilities.

One of the most important reasons for implementing Gene Xpert testing is because it has the potential to reduce the delay between identification of a presumptive TB case, delivery of a positive result and the initiation of TB treatment. The longer that treatment is delayed the greater the risk of death and transmission of TB to others. In a sample of patients referred for testing between January and June 2012, Prof. Hausler showed that 40% of those diagnosed with active TB started TB treatment within 48 hours of providing a sputum sample, and 27% within two to five days.

“We are hoping that this is going to translate into reduced TB transmission in the community,” said Dr Hausler. “What’s even more exciting is that around 40% of drug-resistant TB patients are started on an MDR regimen within 48 hours.” At one site, Pholela community health clinic, Prof. Hausler showed that TB treatment initiation within five days of giving sputum had risen from 10% of patients in the April to June 2011 period to 89% of patients in the same quarter in 2012, and 61% of patients diagnosed with active TB at that clinic started TB treatment less than 24 hours after giving their sputum sample, an impressive vindication of the argument that Xpert MTB/RIF scale up has the potential to dramatically limit the time during which people can transmit TB to others.

This rapid turnaround is underpinned by the involvement of Community Health Facilitators, lay people who form the link between the labs and the clinics. They are responsible for checking Xpert results on a daily basis and passing that information back to the clinics as quickly as possible. Community health workers are then tasked with getting people diagnosed with TB onto treatment as quickly as possible by following them up in the community. Anyone who tests HIV-positive will also be initiated onto ART as quickly as possible.

Kenya

In cases where TB is suspected but microscopy yields a smear-negative result, the delay to treatment that results from referral for culture means that patients will continue to deteriorate. Smear-negative TB is more likely to be diagnosed in people living with HIV, who will also be in need of antiretroviral therapy if they are found to have TB. Xpert MTB/RIF has the potential to greatly reduce the delay before treatment initiation in smear-negative patients.

Dr Jane Carter, the TB/HIV technical advisor for USAID AMPATH presented on the results of using the Xpert MTB/RIF on smear-negative patients in an existing TB REACH programme at 200 primary health care facilities in the north west of Kenya.¹¹

“We found that the Gene Xpert can be rolled out swiftly and effectively to peripheral health units to aid in TB diagnosis,” said Dr Carter.

Five thousand patients with smear-negative TB are to be offered culture diagnosis as part of the project. The culture laboratory is centralised in Eldoret, which is six to eight hours drive from some of the sites. Xpert MTB/RIF testing is to be offered to 2000 smear-negative TB patients. Gene Xpert facilities were established at three peripheral sites in three sub-districts, with the aim of bringing diagnostic services closer to patients. A hub transport system was established to transport specimens, request forms and result forms to and from the health facilities to the testing sites which were between two and 26km apart.

If a patient had a smear negative result, the same sputum specimen would be used for both the culture and Xpert MTB/RIF test. This was to avoid the inconvenience and cost to the patient of needing to return to the facility to provide a second sample. Although TB care and diagnostic tests are free in Kenya, patients are charged a minimum of KSh100 (equal to one day’s wage) to register at the health facility.

Of the 1171 culture tests performed on non-contaminated specimens between December 2011 and July 2012, 9.4 % (n=99) were found to be culture positive. The culture arm had to be stopped prematurely due to poor supply chain management which did not allow for reliable supply of the culture reagent, leading all smear-negative samples to receive Xpert MTB/RIF testing.

The Xpert MTB/RIF test found a similar positivity rate for drug-resistant (DR)TB. Of the 824 smear-negative patients screened using Xpert MTB/RIF between December 2011 and September 2012, 8.6 % (n=71) were found to have TB, while two cases of rifampicin resistant TB  were found.

One of the main advantages of using Xpert MTB/RIF over the culture testing was that the time it took for the clinician to receive results after sending the sputum sample decreased from five weeks to five days.

Xpert MTB/RIF for detection of rifampicin resistance

Rifampicin resistance as a proxy for MDR-TB

According to WHO guidelines on the use of Xpert MTB/RIF, resistance to rifampicin, as identified by Xpert MTB/RIF, should be used as a proxy for MDR-TB. However, a retrospective patient-based analysis which assessed the proportion of isoniazid-susceptible TB cases among rifampicin-resistant TB cases in Botswana, has found that a high proportion of rifampicin-resistant TB isolates are still susceptible to isoniazid.¹²

The study by Dr Valentina Anisimova of KNCV in Botswana gathered data on isoniazid susceptibility among rifampicin-resistant TB isolates by examining the results from DR-TB surveys conducted in the country in 1996, 1999, 2002 and 2007 to 2008. It also used drug-sensitivity testing results from the Botswana National Reference Laboratory from 2006 until 2010.

32.2% and 23.8% of new rifampicin-resistant TB cases were found to be susceptible to isoniazid in the DR-TB surveys and routine data respectively. Among rifampicin-resistant retreatment cases 43.1% and 25.2% were found to be susceptible to isoniazid in the DR-TB surveys and routine data respectively. Combined, 37.9% of rifampicin-resistant cases in the DR-TB and routine surveys were found to be isoniazid-susceptible . The study also found considerable fluctuation in the proportion of patients with isoniazid-susceptible TB between surveys at different time points.

“Testing for susceptibility to isoniazid should be done for all patients identified as rifampicin-resistant by the Gene Xpert,” said Dr Anisimova. “The proportion of isoniazid-susceptible rifampicin-resistant isolates is changing over time and should be closely monitored to inform management of patients with DR-TB.”

Dr Anisimova also recommended that isoniazid should be included in the treatment regimen until isoniazid resistance is confirmed. This study has important implications for the current WHO guidance of using rifampicin resistance as a proxy for MDR-TB which should be reviewed. Although it is recommended to send rifampicin-resistant TB samples for further DST, including isoniazid in the treatment regimen until resistance is confirmed may result in better treatment outcomes for patients.

Xpert MTB/RIF and detection of DR-TB

TB/HIV Care Association’s intervention in Sisonke district in KwaZulu-Natal, South Africa¹³ (described previously) detected 140 cases of rifampicin-resistant TB (a diagnostic yield of 9.9%) compared to 8.7% in the rest of the province. DR-TB patients have been initiated on treatment at a decentralised hospital at sub-district level, compared to the standard approach of being hospitalised at a centralised facility, 150 km from Sisonke district. This has alleviated the burden on the patient’s family members to visit their relatives.

DR-TB patients are also discharged after two consecutive negative smears, as opposed to the policy in the province of only discharging patients at the end of the six-month injectables period. TB/HIV Care Association is also using outreach teams made up of an enrolled nurse and a counsellor to provide patients with injections at home.

The use of Xpert MTB/RIF has also allowed for the time from sputum collection to treatment initiation to be dramatically reduced, rendering patients uninfectious sooner. Between January and June 2012, 36% of MDR-TB patients were initiated on treatment less than 48 hours after providing a sputum specimen, while 18% were initiated on treatment between two and five days after providing sputum.

FIND has also established 12 new Xpert MTB/RIF testing sites in difficult-to-reach areas in India for rifampicin drug-sensitivity testing to supplement the capacity of the existing reference laboratory network of the National TB Programme.¹⁴ It aims to conduct 24,000 rifampicin-resistance tests and expects to detect 7,000 cases of MDR-TB annually. Six of the 12 Xpert MTB/RIF testing sites had been established and used between August and October 2012. Of the 2900 suspected DR-TB cases tested using the Xpert MTB/RIF, 462 (16%) were diagnosed with MDR-TB.

The use of Xpert MTB/RIF in combination with other diagnostic methods

The benefits of using a range of TB diagnostics such as smear microscopy, urinary lipoarabinomannan (LAM) assays and Xpert MTB/RIF have been discussed in previous editions of HATIP. The advantage of these diagnostic tools is that they are more sensitive and specific. The sensitivity of a test is the percentage of results that will be correctly positive when a disease is actually present. Lower rates of sensitivity will produce more false negative results. The specificity of a test is the percentage of results that will be correctly negative when a disease is not present. Lower rates of specificity will produce more false positive results.

These more sensitive and specific diagnostic tools for TB are especially important for diagnosing TB in people with HIV, whose smear microscopy TB test results are often false negatives. Despite being more sensitive and specific than smear microscopy alone, Xpert MTB/RIF and urinary lipoarabinomannan (LAM) assays are rapid diagnostics that have suboptimal sensitivity when used alone in people with HIV who may have TB. Until recently the diagnostic yield of these tests used in combination for the diagnosis of active TB was unknown.

However, a study conducted in Uganda in people with HIV found the sensitivity of the combination of Xpert plus LF-LAM was 87% (n=88 of 101, 95% CI 0.79-0.93), which was superior to either test alone (p<0.05) and was similar to the sensitivity of liquid culture (94%, 95% CI 0.88-0.98, p = 0.17)^.15

The study, which was presented at the 20th Conference on Retroviruses and Opportunistic Infections in Atlanta in March 2012, compared the diagnostic accuracy of sputum smear microscopy, sputum culture (solid and liquid), mycobacterial blood culture, Xpert MTB/RIF testing of archived sputum pellets, and urinary testing for LAM using a point-of-care lateral flow test (LF-LAM) in people with HIV with signs or symptoms of TB.

Participants included 101 individuals with culture-confirmed TB and 107 people with suspected TB in whom TB was excluded clinically and microbiologically. The study found that among the 101 individuals with confirmed TB, the sensitivity of Xpert MTB/RIF was 77% (95% CI, 0.68-0.85), and was superior to that of LF-LAM which was found to have a sensitivity of  50% (95% CI, 0.39-0.60, p <0.001), and smear-microscopy with a sensitivity of 31% (95% CI, 0.22-0.41, p <0.001). Specificity of the Xpert MTB/RIFand LF-LAM among the 107 individuals without TB was more than 97%.

Urinary LAM testing with Xpert MTB/RIF was also highly cost-effective compared to using either smear microscopy or Xpert MTB/RIF alone. The study found that the incremental cost-effectiveness ratio (ICER) when comparing the combination of Xpert MTB/RIF and LAM to smear alone was USD50 per disability-adjusted life year (DALY) and USD29 when compared to Xpert MTB/RIF alone.

Conclusion

While the introduction of Xpert MTB/RIF has been hailed as one of the greatest achievements for the diagnosis of TB and MDR-TB, the price of the machine and the cartridges is the main obstacle which implementers face, despite the decrease in the price. Innovations to reduce the price are necessary to ensure that this useful technology can be scaled up and used effectively globally. As the India study showed, the use of this technology is feasible when certain measures are put in place.

Xpert MTB/RIF has demonstrated significant gains in intensified case finding of TB and MDR-TB, as shown by numerous studies presented at last year’s Union World Lung Health Conference. In every study conducted in operational settings, there was a significant increase in diagnostic yield, demonstrating the potential for Xpert MTB/RIF to assisting in detecting all TB cases. The Xpert MTB/RIF’s superior sensitivity in detecting TB, especially in smear-negative cases is also greatly beneficial to people with TB/HIV co-infection.

Most importantly, the two-hour turnaround time from sputum specimen to result means that the time to treatment initiation can be decreased significantly, as shown in the Sisonke District study by TB/HIV Care Association.

However, challenges remain such as ensuring reliable electricity supply and giving staff adequate training and supervision to avoid uninterpretable results. These are aspects which will need to be addressed going forward.

Further resources

• WHO Stop TB Department Xpert MTB/RIF roll out monitoring website

http://www.who.int/tb/laboratory/mtbrifrollout/en/index.html

• Mapping of Xpert roll out research projects

http://xrmt.treattb.org/

• Data collection on the use of Xpert MTB/RIF by early implementers since 2011

https://extranet.who.int/xpertmtbrif/content/home