National policies on the management of latent tuberculosis infection: review of 98 countries

Abstract Objective To review policies on management of latent tuberculosis infection in countries with low and high burdens of tuberculosis. Methods We divided countries reporting data to the World Health Organization (WHO) Global Tuberculosis Programme into low and high tuberculosis burden, based on WHO criteria. We identified national policy documents on management of latent tuberculosis through online searches, government websites, WHO country offices and personal communication with programme managers. We made a descriptive analysis with a focus on policy gaps and deviations from WHO policy recommendations. Findings We obtained documents from 68 of 113 low-burden countries and 30 of 35 countries with the highest burdens of tuberculosis or human immunodeficiency virus (HIV)-associated tuberculosis. Screening and treatment of latent tuberculosis infection in people living with HIV was recommended in guidelines of 29 (96.7%) high-burden and 54 (79.7%) low-burden countries. Screening for children aged < 5 years with household tuberculosis contact was the policy of 25 (83.3%) high- and 28 (41.2%) low-burden countries. In most high-burden countries the recommendation was symptom screening alone before treatment, whereas in all low-burden countries it was testing before treatment. Some low-burden countries’ policies did not comply with WHO recommendations: nine (13.2%) recommended tuberculosis preventive treatment for travellers to high-burden countries and 10 (14.7%) for patients undergoing abdominal surgery. Conclusion Lack of solid evidence on certain aspects of management of latent tuberculosis infection results in national policies which vary considerably. This highlights a need to advance research and develop clear, implementable and evidence-based WHO policies.


Introduction
Tuberculosis is currently the leading infectious cause of death worldwide. The World Health Organization (WHO) End Tuberculosis strategy aims to substantially reduce tuberculosis incidence by 90% and mortality by 95% compared with the 2015 baselines of 142 cases per 100 000 population and 5.3 to 19 cases per 100 000 (depending on human immunodeficiency virus (HIV) status), respectively 1,2 Achieving this goal requires successful management of latent tuberculosis infection, which serves as a reservoir for new tuberculosis cases. 3 In high-income countries which already have a low incidence of tuberculosis, management of latent infection can contribute to elimination of the disease. 4 A review of treatment regimens found that treatment of latent tuberculosis can reduce the risk of disease reactivation by 60% to 90%. 5 A recent randomized controlled trial in a high tuberculosis burden country showed that the benefits of preventive treatment in people living with HIV can last for more than 5 years. 6,7 The WHO recommends tailored latent tuberculosis infection management based on tuberculosis burden and resource availability. 8 Systematic testing and treatment for latent infection is strongly recommended for people living with HIV and for children younger than 5 years who are household contacts of a pulmonary tuberculosis case, regardless of the country's background tuberculosis burden or resource availability. 9,10 In upper-middle or high-income countries, depending on low tuberculosis burden and availability of resources, systematic testing and treatment of latent tuberculosis is strongly recommended for certain other risk groups: adult household contacts of pulmonary tuberculosis cases; patients with silicosis; pa-tients initiating anti-tumour necrosis factor treatment; patients on dialysis; and organ transplant recipients. 11,12 Despite some progress, particularly over the last decade, the scale-up of tuberculosis preventive treatment remains suboptimal globally. The 161 740 children started on tuberculosis preventive treatment in 2016 represented only 13% of the 1.3 million children estimated to be eligible for treatment. 1 The total number of people living with HIV who were started on tuberculosis preventive treatment in 2016 was at least 1.3 million. 1 Data for other risk groups are not available or very limited.
Barriers to scale-up of tuberculosis preventive treatment include the absence of national policies and a lack of monitoring and evaluation systems. 13 Here we review national policy documents to identify differences in programmatic management of latent tuberculosis infection in high-and low-burden countries.

Methods
The baseline for this descriptive policy review was the 216 countries and territories reporting data to the WHO Global Tuberculosis Programme (194 Member States and 22 associate Member States and territories). Based on the current WHO approach 11 we divided countries into two groups: low burden and high burden. We defined low-burden countries as upper-middle or high-income countries with an estimated annual tuberculosis incidence of less than 100 cases per 100 000 population. 11 High-burden countries were low-to lower-middle-income or other income countries with annual tuberculosis incidence of 100 or more cases Objective To review policies on management of latent tuberculosis infection in countries with low and high burdens of tuberculosis. Methods We divided countries reporting data to the World Health Organization (WHO) Global Tuberculosis Programme into low and high tuberculosis burden, based on WHO criteria. We identified national policy documents on management of latent tuberculosis through online searches, government websites, WHO country offices and personal communication with programme managers. We made a descriptive analysis with a focus on policy gaps and deviations from WHO policy recommendations. Findings We obtained documents from 68 of 113 low-burden countries and 30 of 35 countries with the highest burdens of tuberculosis or human immunodeficiency virus (HIV)-associated tuberculosis. Screening and treatment of latent tuberculosis infection in people living with HIV was recommended in guidelines of 29 (96.7%) high-burden and 54 (79.7%) low-burden countries. Screening for children aged < 5 years with household tuberculosis contact was the policy of 25 (83.3%) high-and 28 (41.2%) low-burden countries. In most high-burden countries the recommendation was symptom screening alone before treatment, whereas in all low-burden countries it was testing before treatment. Some low-burden countries' policies did not comply with WHO recommendations: nine (13.2%) recommended tuberculosis preventive treatment for travellers to high-burden countries and 10 (14.7%) for patients undergoing abdominal surgery. Conclusion Lack of solid evidence on certain aspects of management of latent tuberculosis infection results in national policies which vary considerably. This highlights a need to advance research and develop clear, implementable and evidence-based WHO policies.
per 100 000. Among the high-burden countries, we focused on the top 30 countries in terms of high burden of tuberculosis (both in terms of number of cases and incidence) and on the top 30 countries in terms of high burden of HIV-associated tuberculosis. These countries account for most of the global burden of tuberculosis ( (Fig. 1). We aimed to analyse each country's or territory's national guidelines on the management of tuberculosis, HIV, paediatric tuberculosis, latent tuberculosis infection and HIV-associated tuberculosis, and standard national operating procedures for tuberculosis. We obtained documents by contacting WHO country offices or national programme managers or by downloading them from the official website of the ministry of health or other national health organization.
We selected the information to be extracted a priori based on WHO recommendations for the management of latent tuberculosis infection ( Table 1 and Table 2). [9][10][11] We collected information on the following: (i) at-risk populations targeted; (ii) recommended tests for latent tuberculosis infection; (iii) diagnostic algorithms to exclude active tuberculosis before starting treatment for latent tuberculosis infection; (iv) treatment regimens for latent tuberculosis infection; and (v) presence of monitoring and evaluation systems for the management of latent tuberculosis infection. For high-burden countries, we focused the review only on people living with HIV and children younger than 5 years who have household contact with a tuberculosis case. One researcher collected and entered the data for all low-burden countries and another researcher for all high-burden countries using data extraction forms developed for the study.
Statistical analysis of the data was performed using GraphPad Prism (GraphPad Software Inc., La Jolla, United States of America) and STATA (Stata Corp LLC, College Station, USA) software. Where percentages are indicated, binary indicators (0,1) were created for the absence or presence of each policy item extracted. The means of those binary indicators corresponded to the percentage of countries addressing each policy item. The number and percentage of countries addressing each policy item were calculated and presented.

Results of search
We obtained and analysed copies of policy documents from 98 countries (

Risk groups defined
Of the 30 high-burden countries for which guidelines were obtained, information on the management of latent tuberculosis infection among children with a household contact was available for 25 countries. In four countries the relevant tuberculosis guidelines could not be obtained and in one country the guidelines were written in local languages that we were not able to translate. All 25 countries followed WHO policy (Table 1) recommending treatment for children younger than 5 years with a household tuberculosis contact ( Table 4); 17 of these specifically targeted contacts of smear-positive cases. India and Nigeria recommended preventive treatment for children under 6 years old with a household contact. No policy recommended preventive treatment for contacts of a multidrug-resistant tuberculosis case.
For people living with HIV, 29 countries (96.7%) had recommendations on tuberculosis preventive treatment; only Ghana did not provide any recommendations.

Testing recommendations
For children younger than 5 years with a household contact, 24/25 (96.0%) of the countries analysed did not have recommendations for testing for latent tuberculosis before starting preventive treatment. Only in the Philippines was a tuberculin skin test recommended, with the option to provide preventive treatment without testing when testing was not available. To exclude active tuberculosis before treatment of latent tuberculosis, most countries (24/25, 96.0%) had a policy on symptomatic screening alone. Symptom-based algorithms to exclude active tuberculosis were defined in the guidelines of 12 countries (Table 2).
Of these, 11 countries included cough, fever and weight loss or poor weight gain in their algorithms. The presence of a variety of additional symptoms and signs were also specified: fatigue, wheeze, neck mass, abdominal mass, ascites, diarrhoea, loss of appetite and night HIV: human immunodeficiency virus. a We could not obtain any relevant treatment guidelines for child contacts and we found no recommendations in other available guidelines (4 countries) or the guidelines were written in a local language and we were unable to translate them with confidence (1 country).
National policies on latent tuberculosis Ann Jagger et al.
sweats. The exclusion algorithm was not defined in the remaining countries. For people living with HIV, 86.7% (26/30) of the high-burden countries analysed provided preventive treatment for latent tuberculosis without testing for infection. In South Africa the recommendation was for a tuberculin skin test before starting preventive treatment, but this was not specified by the remaining countries. The majority of the countries (20/30) applied the WHO four-symptom screening rule (current cough, fever, weight loss and night sweats) for excluding pulmonary tuberculosis before starting preventive treatment ( Table 2). Five countries specified a different set of symptoms and another five countries did not specify the symptoms to be used in the exclusion algorithm.
For children older than 12 months living with HIV, 66.7% (20/30) of high-burden countries had a recommendation for symptomatic screening alone before starting preventive treatment. Only India had a policy of doing a tuberculin skin test in addition to symptomatic screening before starting such treatment. In Angola, the recommendations were for chest radiography in addition to symptomatic screening. Only eight (26.7%) countries followed the WHO recommendation to exclude active tuberculosis based on poor weight gain, fever, current cough or contact history with a tuberculosis case ( Table 2).

Treatment recommendations
WHO recommends 6 months of isoniazid monotherapy both for people living with HIV and children with a household contact in high-burden countries (Table 1). Among the high-burden countries reviewed, the majority (18/30) of guidelines recommended 6 months of isoniazid monotherapy, while in six countries (Cambodia, Democratic Republic of Congo, Namibia, Thailand, Viet Nam and Zimbabwe) it was a course of 6-9 months. Central African Republic had a policy of 3 months of rifampicin plus isoniazid, as well as 6 months of isoniazid (Table 4). In Uganda and Pakistan recommendations were for an additional course of prolonged isoniazid treatment (12 and 36 months, respectively) for people living with HIV who have tuberculosis contact history. In South Africa the recommendations were 6-36 months of isoniazid treatment, depending on the results and availability of tuberculin skin testing. In Malawi the policy was continuation of isoniazid treatment for those not receiving antiretroviral therapy but discontinuation once therapy is started.

Monitoring and evaluation indicators
Of the high-burden countries, only Kenya, Malawi, South Africa and Thailand had guidelines that defined indicators to evaluate the coverage of tuberculosis screening and preventive treatment among children younger than 5 years with a household contact. Most countries (18/30) defined an indicator for coverage of preventive treatment in people living with HIV (Table 4). In 2017, 10 of these countries reported data to the Global tuberculosis report 1 about the proportion of patients newly enrolled in HIV care who were provided with tuberculosis preventive treatment (Table 5). Fifteen countries included information on recording and reporting tools for isoniazid preventive treatment in their guidelines (Table 4).

Risk groups defined
The risk groups strongly recommended by WHO to be targeted for latent tuberculosis infection screening (Table 1) were included in the national latent tuberculosis infection policies of between 19 (27.9%) and 54 (79.4%) of 68 lowburden countries (Fig. 2). Specifically, 28 countries (42.1%) had a recommendation to screen children younger than 5 years who are contacts of a tuberculosis case. An additional 49 countries (70.1%) had recommendations to screen all contacts of a tuberculosis case, making no distinction between adults and children. For people living with HIV, the policy in 54 (79.4%) countries was to screen people living with HIV for latent tuberculosis infection and in 23 (33.8%) countries it was to screen immunocompromised individuals, which includes people living with HIV.
In contrast, some of the conditionally recommended categories (such as

Research
National policies on latent tuberculosis Ann Jagger et al.
prisoners and illicit drug users; Table 1) were rarely mentioned in policies (Fig. 2). Notably, some countries included categories that are not recommended by the WHO; nine countries (13.2%) recommended tuberculosis preventive treatment for travellers to high tuberculosis burden countries and 10 (14.7%) for patients undergoing abdominal surgery.

Testing recommendations
The WHO latent tuberculosis infection guidelines indicate that in low-burden countries either a tuberculin skin test or interferon-gamma release assay can be used for diagnosis (Table 1). Of the low-burden countries 33/68 (48.5%) had a recommendation to use tuberculin skin testing as the primary screening method compared with only 2/68 (2.8%) recommending interferon assay (Fig. 3). In 21 countries (30.8%), the policy was either tuberculin skin test or interferon assay as the primary method of screening. In addition, multiple policies specified situations when using one test over the other was preferable. For example, in 21 (30.8%) countries the policy was that interferon assay should be used for individuals vaccinated with bacille Calmette-Guérin (BCG) and in 17 (25.0%) countries that interferon assay and tuberculin skin test should be used sequentially. For some countries, including Costa Rica and Uruguay, there were no explicit recommendations on methods of testing.
An algorithm for excluding active tuberculosis was specified in the policies of 43 (63.2%) low-burden countries, although the content of that algorithm varied greatly from country to country. In Colombia, Ecuador and Uruguay the recommendation was only that active tuberculosis should be ruled out, with no mention of an exclusion algorithm. All other countries required at least a chest X-ray.

Treatment recommendations
The most commonly recommended treatments in low-burden countries were isoniazid for 6 months (55 countries; 80.8%) or 9 months (55 countries, 80.8%) (Fig. 4), which is in line with the WHO guidelines on treatment of latent tuberculosis (Table 1). Alternative treatment options recommended by the WHO were also frequently mentioned in other policies, but to a lesser extent, ranging from 8 (11.7%) to 51 (75.0%) countries.

Monitoring and evaluation indicators
Monitoring and evaluation of latent tuberculosis infection screening was mentioned in the policies of 32 (47.1%) low-burden countries. Even among the countries that mentioned reporting requirements, those were often specific to active tuberculosis, and therefore the form may be inappropriate for latent tuberculosis infection.

Discussion
This review identified that the majority of both high-and low-burden countries had a national policy that addressed latent tuberculosis infection management in people living with HIV and children younger than 5 years with a household contact. Clinical high-risk groups were also covered by most guidelines from low-burden countries. However, the content of the guidelines varied considerable across countries. For example, clear and standard algorithms for excluding tuberculosis before treatment and latent tuberculosis infection testing were not available in many countries, and indicators for monitoring and evaluation were rarely defined. Guidelines are the first step in implementing the programmatic management of latent tuberculosis infection, hence it is essential to provide clear and simple operational National policies on latent tuberculosis Ann Jagger et al.

Fig. 2. Compliance of national policies with World Health Organization guidelines on screening for latent tuberculosis infection among high-risk population groups in low-burden countries
guidance, including evidence-based standardized algorithms and a framework of monitoring and evaluation. 14 The advantage of an evidence-based standardized algorithm was demonstrated by the WHO recommended four-symptom screening rule to exclude active tuberculosis before starting preventive treatment for people living with HIV. 10 This simple algorithm adds to the clarity of the policy and has resulted in a steep rise in implementation of isoniazid preventive treatment among people living with HIV in settings with a high prevalence of tuberculosis and low resources, reaching 1.3 million in 2016. 1 Ensuring that guidelines and algorithms are simple can also facilitate their incorporation into national guidelines. For example, seven out of 10 countries that had algorithms different from the WHO recommendation in a previous policy review 15 have now adopted them (Cameroon, Lesotho, Nigeria, South Africa, Swaziland, United Republic of Tanzania and Viet Nam).
In contrast to the uptake of the screening algorithm for people living with HIV, the corresponding screening algorithm for children was not taken up or defined in national policies. This could be due to the limited evidence about the effectiveness of the algorithm, as it was recommended largely based on expert opinion. 16 Further research is needed to evaluate the performance of the algorithm and identify the optimal approach to exclude active tuberculosis in children before starting preventive treatment.
Consistent with our previous study, 13 we found that the national policies and guidelines in the majority of low-burden countries addressed latent tuberculosis infection specifically or as part of the general tuberculosis policy. The Netherlands has revised its guidelines since the publication of the 2015 WHO latent tuberculosis infection guidelines, 17 which are now mostly consistent with WHO recommendations. A similar revision by other countries would increase alignment between national policies and WHO recommendations. This could lead to more consistent and comprehensive latent tuberculosis infection policies and pave the way for global monitoring and evaluation of the programmatic management of latent tuberculosis infection. Although it may be too early to evaluate the impact of such policy changes on tuberculosis incidence, it is a question that needs to be addressed in the future.
Tuberculin skin testing was the most frequently recommended diagnostic tool. The test requires no laboratory work and is comparably cheaper per unit test than interferon-gamma release assay. That may explain the overwhelming preference for the test over interferon assay in the policies of low-burden countries. Several countries specified additional diagnostic algorithms, such as different tuberculin skin test cut-off points among specific risk groups, sequential use of the two tests, or use of interferon assay for BCG-vaccinated individuals. A systematic review did not show a significant difference in the prediction of progression to active tuberculosis between the two tests in head-to-head analysis. 11 However, there were insufficient data on the predictive utility among specific populations. The diversity of policies across countries calls for more research in how to use interferon-gamma release assay and tuberculin skin testing together among different risk groups based on the underlying tuberculosis epidemiology.
This policy review has limitations. First, determining the latest published guidelines was done through contacting national programmes, WHO offices and through extensive internet searches; however some policies may not have been identified. Even though latent tuberculosis infection monitoring and evaluation indicators may not have been defined in guidelines they may nevertheless exist within a country's national tuberculosis programme or other guidelines. These limitations might have led to misclassification of the findings. Second, a single person was responsible for reviewing policies, extracting relevant information and entering data within each group (highand low-burden countries). While this provided internal consistency, the data collection may have been subject to reviewer bias. In conclusion, our review identified large variations across countries in their national tuberculosis policies. The differences are probably attributable to different country contexts and disease epidemiology and lack of consensus on some aspects of latent tuberculosis infection management. There are unique challenges associated with management of latent tuberculosis infection, such as exclusion of active tuberculosis, testing for latent tuberculosis infection and treatment initiation. It is therefore important to continue to develop clear, implementable and evidence-based WHO policies. An important component of such policies should be monitoring and evaluation, as this is essential to assess progress in the implementation and to make policy decisions. Lack of a monitoring and evaluation component in more than half of the national policies presents a barrier to programmatic management of latent tuberculosis infection. ■ National policies on latent tuberculosis Ann Jagger et al.