How is Xpert MTB/RIF being implemented in 22 high tuberculosis burden countries?

To the Editor:

Accurate and rapid diagnosis is crucial for tuberculosis control by ensuring a timely start to treatment and reducing transmission. In 2012, almost one third of tuberculosis cases were not diagnosed and/or reported to national tuberculosis programmes (NTPs), and <25% of estimated incident multidrug-resistant (MDR) cases were diagnosed [1]. Xpert MTB/RIF (Cepheid, Sunnyvale, CA, USA), a nucleic acid amplification test, was recommended in 2010 by the World Health Organization (WHO) for detection of HIV-associated pulmonary tuberculosis and rifampicin resistance [2]. In 2013, the test was recommended for detection of paediatric tuberculosis and some forms of extrapulmonary tuberculosis (EPTB), as well as an initial test to replace smear microscopy [3].

Following these recommendations, modules and cartridges have been procured in increasing numbers. As of June 30, 2014, 15 846 Xpert modules and 7.5 million cartridges were procured by 104 countries at concessional prices [4], yet the potential market is much larger [5]. Although general policies regarding Xpert in the 22 high-burden countries (HBCs) have been summarised [1] and some experiences from early Xpert implementers are available [6, 7], a more comprehensive analysis of NTPs’ policies and implementation of Xpert has not been performed.

To assess the current landscape of implementation of Xpert, we designed a standardised questionnaire that was sent to NTPs in 22 HBCs that account for 80% of tuberculosis cases globally. We contacted NTP managers and representatives with responsibilities relating to Xpert. Questionnaires were completed from January to July 2014, with follow-ups to ensure completion and clarify any ambiguities. Questions covered the following topics: funding sources, instrument placement, access in the private sector, testing algorithms, result reporting and treatment decisions for rifampicin-resistant results. Additionally, to better assess the scale of implementation, we analysed publicly available Xpert procurement data [4].

As shown in table 1, of the 22 HBCs, 19 (86%) reported an existing national plan or policy pertaining to Xpert. Seven (32%) of the 22 countries reported the use of domestic funding for Xpert procurement. However, only Brazil and Russia currently fund all Xpert testing with domestic resources, while the majority of HBCs rely on some of the 16 international donor groups identified. As many as six external donors were reported in some countries, suggesting a strong need for in-country coordination.

Until June 2014, of the 7.5 million cartridges procured through public sector pricing, HBCs procured 6.4 million (85%). Of those, 4.2 million (66%) of cartridges were procured by South Africa alone, which along with China, India and Brazil, account for 80% of total HBC procurement. The ratio of smear volumes for initial diagnosis [5] to the number of Xpert cartridges procured during a roughly similar time period was used as an approximate index of Xpert market penetration in the public sector. The ratio in South Africa was 1.6, significantly lower than most other HBCs where approximately 40–70 smears were performed for each Xpert. Evidently, wide-scale implementation of Xpert has only occurred in South Africa, while other HBCs continue to rely heavily on smear microscopy.

While all countries reported deployment of Xpert in the public sector, only five (23%) reported public–private partnerships around Xpert testing, the initiatives to promote the collaboration between private and public health providers in the delivery of tuberculosis care; an additional eight (36%) use Xpert in other private-sector settings. As Xpert was initially recommended for use at district and subdistrict laboratories [8], eight (36%) countries reported the deployment of Xpert at microscopy or peripheral health centres, showing promising progress. 18 (82%) reported deployment at district and subdistrict levels, and 17 (77%) reported deployment at reference or centralised laboratories. Although a previous study showed that Xpert implementation is feasible in some primary care facilities [9], the current infrastructure in HBCs might not be adequate for wide-scale coverage [10].

With respect to testing algorithms, only South Africa, Brazil and Russia recommend Xpert for all people suspected of having tuberculosis. Additionally, Brazil reported plans to replace smear microscopy with Xpert in 92 cities across the country. Although all HBCs recommend Xpert as an initial test for drug-resistant tuberculosis (DR-TB), eligibility criteria vary among them. Four countries recommend Xpert only for patients with suspected drug resistance, although in Pakistan and Bangladesh, Xpert is also being used for general tuberculosis case finding at selected sites [7]. The remaining 19 HBCs recommend Xpert among HIV-infected patients, although in Thailand and Uganda, Xpert is recommended only after negative smear results, against WHO recommendations. However, given the limited number of cartridges procured outside South Africa, actual application of these algorithms is likely to be limited. Testing strategies focusing on the detection of drug resistance among retreatment cases only identify a fraction of total new MDR cases in most countries and will limit the ability to scale-up DR-TB treatment programmes. Ultimately, countries have to work towards universal drug susceptibility testing (DST) as outlined in the Global Plan and Post-2015 Global TB Strategy [11, 12], but this will require greater resources.

While updated policy guidance on Xpert for the diagnosis of paediatric tuberculosis and EPTB was only issued in October 2013, 14 (59.1%) countries already reported recommending Xpert in children suspected of having tuberculosis. The use of Xpert for EPTB diagnosis was recommended in four (18%) countries.

WHO developed new recording and reporting recommendations in 2013 largely in response to the introduction of new molecular tests [13]. 14 (64%) countries recommended recording Xpert-positive results as bacteriologically positive, while three (14%) reported having no standards for reporting at this time. These findings demonstrate progress after some early implementers documented challenges around unclear and inconsistent reporting [7].

Initial WHO guidance for treatment decisions for patients with rifampicin resistance but not at risk for DR-TB recommended follow-up DST using another method, citing poor positive predictive values for Xpert [2]. Recent evidence suggests that using phenotypic DST as the reference standard misses some rifampicin-resistant cases [14]. Currently, WHO recommends that a rifampicin-resistant Xpert result for persons suspected of having DR-TB is sufficient to initiate second-line treatment (SLT) [3]. Most countries initiate SLT for those with risk factors for drug resistance (without confirmation or while waiting for confirmation of Xpert results), while three (14%) require confirmatory DST prior to SLT initiation. For patients at low risk of drug resistance, 13 (59%) countries require confirmatory DST before initiating SLT. A number of countries reported that current guidelines are under review and likely to change as more evidence becomes available.

Overall, we found the uptake of WHO guidelines on Xpert has been relatively quick compared with other guidelines on new tuberculosis diagnostics, such as light-emitting diode microscopy or same-day smear diagnosis. However, previous studies [7] suggest the implementation of Xpert in the field may deviate from stated national policy, and we found current Xpert testing is mainly donor-funded, mostly limited to district or reference laboratories, and primarily used in patients suspected of having DR-TB, and to a lesser extent among persons suspected of HIV-associated tuberculosis. Models suggest that more restrictive implementation strategies might limit the impact of Xpert [15]. Therefore, we hope these results will serve to raise awareness about the need for more ambitious testing algorithms (e.g. universal DST) and implementation for greater impact, acknowledging this will only be possible with much greater investments in improved tuberculosis diagnosis and care from both donors and domestic funding.