British <scp>HIV</scp> Association guidelines for the management of tuberculosis in adults living with <scp>HIV</scp> 2019

Abstract

These guidelines are dedicated to the memory of Professor Stephen Lawn, a scientist and clinician whose pioneering work helped transform the management of TB in people living with HIV. The overall purpose of these guidelines is to help physicians manage adults with tuberculosis (TB)/human immunodeficiency virus (HIV) co-infection. Recommendations for the treatment of TB in HIV-positive adults are similar to those in HIV-negative adults. Of note, the term "HIV" refers to HIV-1 throughout these guidelines. The British HIV Association (BHIVA) fully revised and updated the Association's guideline development manual in 2011. Further updates have been carried out subsequently 1. Full details of the guideline development process, including conflict of interest policy, are outlined in the manual. BHIVA has adopted the modified Grading of Recommendations Assessment, Development and Evaluation (GRADE) system for the assessment, evaluation and grading of evidence and development of recommendations (see below and Appendix 1) 2, 3. The scope, purpose and guideline topics were agreed by the writing group. Questions concerning each guideline topic were drafted and a systematic literature search was undertaken by an information scientist. Details of the search questions and strategy (including the definition of populations, interventions and outcomes) are outlined in Appendix 2. BHIVA guidelines for the treatment of TB/HIV co-infection were last published in 2011 4. For the 2017 guidelines, Medline, EMBASE and the Cochrane Library were searched between August 2015 and January 2016. Abstracts from selected conferences (see Appendix 2) were searched between August 2015 and January 2016. For each topic and healthcare question, evidence was identified and evaluated by writing group members with expertise in the field. Using the modified GRADE system, writing group members were responsible for assessing and grading the quality of evidence for predefined outcomes across studies and developing and grading the strength of recommendations. An important aspect of evaluating evidence is an understanding of the design and analysis of clinical trials, including the use of surrogate marker data. Decisions regarding the clinical importance of difference in outcomes were made by the writing group. Before final approval by the writing group, the guidelines were published online for public consultation and an external peer review was commissioned. BHIVA views the involvement of people living with HIV (PLWH) and community representatives in the guideline development process as essential. The writing group included two representatives appointed through the UK Community Advisory Board (UK-CAB) and community groups are specifically invited to participate in the public consultation process. The GRADE Working Group 2 has developed an approach to grading evidence that moves away from initial reliance on study design to consider the overall quality of evidence across outcomes. BHIVA has adopted the modified GRADE system for its guideline development (see Appendix 1). The advantages of the modified GRADE system are: (i) the grading system provides an informative, transparent summary for clinicians, PLWH and policymakers by combining an explicit evaluation of the strength of the recommendation with a judgement of the quality of the evidence for each recommendation, and (ii) the two-level grading system of recommendations has the merit of simplicity and provides clear direction to PLWH, clinicians and policymakers. The strength of recommendation is graded as 1 or 2 as follows: The strength of a recommendation is determined not only by the quality of evidence for defined outcomes but also by the balance between desirable and undesirable effects of a treatment or intervention, differences in values and preferences and, where appropriate, resource use. Each recommendation concerns a defined target population and is actionable. The quality of evidence is graded from A to D and for the purpose of these guidelines is defined as the following: In addition to graded recommendations, the BHIVA writing group has also included good practice points (GPPs), which are recommendations based on the clinical judgement and experience of the group. GPPs emphasise an area of important clinical practice for which there is not, nor is there likely to be, any significant research evidence. They address an aspect of treatment and care that is regarded as such sound clinical practice that healthcare professionals are unlikely to question it and where the alternative recommendation is deemed unacceptable. It must be emphasised that GPPs are not an alternative to evidence-based recommendations. The following measures have been or will be undertaken to disseminate and aid implementation of the guidelines: The guidelines will be fully updated and revised in 2021. However, the writing group will continue to meet regularly to consider new information from high-quality studies and publish amendments and addenda to the current recommendations before the full revision date where this is thought to be clinically important to ensure continued best clinical practice. These guidelines update the previously published BHIVA guidelines on the treatment of TB/HIV co-infection from 2011 1 and are designed to provide a clinical framework applicable to adults living with HIV in the UK who have TB. They do not include management of HIV-positive children with TB. The guidance is based on the evidence available, although some recommendations necessarily rely on expert opinion until further data become available. These guidelines should be used in conjunction with: The WHO reported in 2018 the following data for 2017 7: The incidence of TB in England is higher than in most Western European countries 8. Although it was declining during most of the 20th century, a steady increase was observed from the late 1980s to 2005. The annual incidence rates of TB among adults living with diagnosed HIV in England and Wales declined from 17.5 per 1000 (420/23,990) in 2008 to 4.4 per 1000 (300/68,350) in 2011 9. This trend is largely due to a decline in new HIV diagnoses among men and women born in countries of sub-Saharan Africa where the prevalence of both HIV and TB is high, as well as to an increase in total number of PLWH 10. TB incidence varies by demographic characteristics with rates among people born outside the UK of 7.7 per 1000 population (in 2011), women (6.7), those aged 25–39 years (10.5) and people of black African ethnicity (7.7). The risk of developing TB is estimated to be between 26 and 31 times greater in PLWH than among those without HIV infection. Thus, all individuals with TB, regardless of their perceived risk of HIV infection, should be offered an HIV test. In HIV co-infection, the clinical and radiographic presentation of TB may be atypical. Compared with the immune-competent population, TB/HIV-positive individuals with active pulmonary TB are more likely to have normal chest radiographs or sputum that is smear negative but culture positive 11, 5 (see Section 5 and Appendix 3). The clinician caring for HIV-positive individuals therefore needs to have a high index of suspicion for TB in symptomatic individuals, especially those who have lived in TB-endemic parts of the world. As the investigation and treatment of both TB and HIV infection is complex, it is mandatory to involve specialists in HIV, respiratory and/or infectious diseases. Treatment of TB benefits the individual and also the community. The aims of treatment are 1: Microscopic smear of clinical specimens remains an essential part of TB diagnosis. The quality of any investigation is related to the quality of the specimen and the clinical detail provided with the request. There must therefore be close liaison with the mycobacteriology laboratory. Results should be available within 1 working day. Use of molecular biology allows for early identification of mycobacteria and of genotypic (rifampicin/isoniazid) drug susceptibility. The Xpert MTB/RIF is an automated molecular test for identification of M. tuberculosis and of rpoB mutations conferring resistance to rifampicin. It is very specific (99%) and its sensitivity for smear-positive, culture-positive TB approaches 98%, compared with a sensitivity of 65% for microscopy 1. The sensitivity for rifampicin resistance is slightly lower (95%) than the sensitivity for M. tuberculosis identification (see Appendix 3). In smear-positive samples, its use can allow rapid confirmation that AFB are not M. tuberculosis, potentially avoiding unnecessary treatment and infection-control measures 2. The newer Xpert MTB/RIF Ultra (Cepheid) has been shown to have improved sensitivity but lower specificity in HIV-positive individuals compared with Xpert MTB/RIF 3 and is recommended by WHO for sputum and selected extrapulmonary samples 4. Despite the high sensitivity and specificity, molecular biology tests have to be performed together with cultures and phenotypic drug susceptibility testing. All specimens, even those negative for M. tuberculosis on polymerase chain reaction (PCR), still require culture because a negative PCR does not exclude M. tuberculosis and a positive PCR does not currently indicate the full drug-susceptibility profile 5, 6. Whole-genome sequencing (WGS) is available in the UK and is currently being used to identify clusters and to detect genotypic resistance but it requires a culture isolate; Xpert MTB/RIF can be performed on a primary sample (without the need for a positive culture), for example a sputum sample, and detects M. tuberculosis and mutations associated with rifampicin resistance more quickly. The sensitivity and specificity of IGRAs in HIV-positive people is suboptimal when used alone to "rule in" or "rule out" active TB disease 7-10. IGRAs should not be used to diagnose or exclude active TB (see Appendix 3). Identification of mycobacteria is performed at reference centres, and is based on molecular techniques, morphology, growth and biochemical characteristics. Liquid culture medium provides more rapid results than solid medium and M. tuberculosis can usually be grown in 7–28 days. Drug-susceptibility tests using WGS and phenotypic assays are usually available within 10–21 days from receipt of isolates by the laboratory. The commonest presentation of TB in the CNS is tuberculous meningitis (TBM), which is the most severe form of TB with the highest mortality (between 20% and 50%) and morbidity, as diagnosis and treatment are often delayed 11. Less commonly it can manifest as tuberculous encephalitis, intracranial tuberculomas or tuberculous brain abscess(es) 11. Early diagnosis is challenging due to the non-specific symptoms of TBM, such as fever, headache and vomiting, with gradual onset and duration, often lasting for weeks. Meningism, with or without focal neurological deficits, behavioural changes and alterations in consciousness are also features of TBM. The main investigations are cranial imaging (magnetic resonance imaging) and lumbar puncture for CSF analysis. Significant CSF findings in TBM include a mainly mononucleate cell (lymphocytic predominant) pleocytosis in 60–85% of patients, in which the total white count ranges between 100 and 500 cells/mm3. In advanced HIV, CSF can be acellular. Low CSF glucose levels (usually less than 2.5 mmol/L) and high protein levels, typically between 1 and 5 g/L, are also suggestive of TBM. Identification of M. tuberculosis in CSF by culture remains the "gold standard", but has a limited sensitivity (ranging between 10% and 60%). Microscopy with Ziehl–Neelsen staining for AFB detection has a low sensitivity in the CSF (10–60%), due to the small number of tubercle bacilli usually present. Large volumes (minimum 6 mL) of CSF should be examined to enhance the sensitivity 12, 13. The WHO recommendation is to use Xpert MTB/RIF as the preferred initial test for diagnosis of TB meningitis instead of conventional tests (see Appendix 3). However, a negative Xpert MTB/RIF result on a CSF sample does not exclude TB meningitis. Where available, use of Xpert MTB/RIF Ultra is preferred as it has a higher sensitivity than Xpert MTB/RIF in diagnosing TB meningitis 14. Adenosine deaminase (ADA) (a predominant T lymphocyte enzyme, which catalyses the conversion of adenosine and deoxyadenosine to inosine and deoxyinosine, respectively) measurement can also be of use in the diagnosis of TB meningitis. Levels in CSF are significantly elevated in TBM with a sensitivity and specificity ranging from 60–90% and 80–90%, respectively 12. However, the ADA assay has not been standardised and the "cut-off" level that defines a positive result has not been determined, and consequently it is not recommended as part of routine investigation for TB meningitis 15, 16. Where HIV is endemic, TB pleuritis is the most common cause of a lymphocytic effusion, thought to result from primary infection in 30% of patients 18. In individuals with a suspected TB pleural effusion it is important to obtain cultures on pulmonary (generally sputum or bronchoscopic) samples, including in the absence of obvious parenchymal involvement as, even in individuals with normal underlying lung parenchyma on chest radiography, the yield of sputum culture in induced samples approaches 55% 17. The diagnosis of TB pleuritis is also made by detection of M. tuberculosis in pleural fluid or pleural biopsy specimens, or by assumption if M. tuberculosis is identified in sputum and there is co-existent pleural effusion, either by microscopy and/or culture or by the histological demonstration in the pleura of caseating granulomas together with AFB. Microscopy for AFB in the pleural fluid can identify M. tuberculosis in approximately 20% of HIV-positive individuals with pleural TB, though the yield can be up to 50% 19 if the patient's CD4+ cell count is less than 100 cells/mm3 19, 20. TB PCR has a low sensitivity for diagnosis of pleural TB. A pooled analysis of data from 20 studies that assessed the use of pleural fluid molecular diagnostic tests showed a high specificity (97% for commercial and 91% for in-house tests) but a generally poor and variable sensitivity (62% for commercial and 76.5% for in-house tests) 6. Where available, medical thoracoscopy may be useful in the diagnosis of pleural TB. In settings of low TB incidence, thoracoscopy has proved to be an effective diagnostic tool in HIV-negative patients, with a pooled sensitivity for TB on culture and histology of 93%, in combination with ADA, and a specificity of 100% 18. Measurement and quantification of ADA in pleural fluid may also be useful. Individuals who present with a lymphocytic predominant exudative pleural effusion and raised ADA level have a high probability of having pleural TB (see Appendix 3). Data on the accuracy of molecular biological tests for diagnosis of TB in non-respiratory specimens have been reported in two systematic reviews (SRs), which both support their use in diagnosis of extrapulmonary TB 21, 22 (see Appendix 3). The urine lateral flow lipoarabinomannan (LF-LAM) assay is a point-of-care test for active TB (see Appendix 3). Its sensitivity is highest in individuals with a CD4+ cell count <100 cells/mm3 23. Therefore it represents a useful adjunctive diagnostic for individuals with CD4+cell counts <100 cells/mm3 and in those who present with serious illness of unknown cause. Mycobacterial blood culture has also proven useful in diagnosis of disseminated TB in patients with low CD4+ cell counts (sensitivity 20–40%) 24. The cytopathological diagnosis of TB is based on finding AFB on Ziehl–Neelsen staining of tissue or a cytological preparation (e.g. a lymph node aspirate). Supplementary supportive evidence is provided by the finding of macrophage granulomas with or without necrosis. The finding of AFB in a cytopathological specimen should be critically interpreted in the context of a patient's presentation, their imaging findings and results from other laboratory investigations. It is important to precisely identify AFB where possible, using culture and molecular diagnostic techniques. The classical lesions of TB include epithelioid cell granulomas with or without Langhans giant cells and caseation necrosis, and AFB. Other diseases, infectious and non-infectious, have similar granuloma morphology as TB, and fungal staining must always be undertaken to exclude mycosis (e.g. histoplasmosis) as the relevant agent. If TB is diagnosed histopathologically, but standard treatment appears ineffective, non-tuberculous mycobacterial infection should be considered. Other differential diagnoses that can mimic TB include: sarcoidosis, histoplasmosis, nocardiosis, leishmaniasis, granulomatous reaction to local tumour, common variable immunodeficiency syndromes, vasculitis syndromes, autoimmune diseases and Gram-negative infections (e.g. brucellosis and melioidosis). In difficult cases, multidisciplinary consultation is invaluable, where all the information – clinical, radiological, pathological, molecular diagnostics and results of treatment – can be critically reviewed. Because the presence of granulomas is regarded as typical of TB, differential diagnoses should be considered, especially if response to treatment is not progressing as expected. MDR-TB definition: resistance to at least isoniazid and rifampicin. Pre-extensively drug-resistant (XDR)-TB definition: resistance to isoniazid and rifampicin and either a fluoroquinolone or second-line injectable agent but not both. XDR-TB definition: resistance to isoniazid and rifampicin and quinolones and at least one second-line injectable (e.g. amikacin). The number and proportion (1.7%) of TB cases with initial rifampicin-resistant/MDR-TB in England has been relatively stable since the peak in 2011 (89, 1.8%). Public Health England reported that in England in 2017, 3.3% (3/90) of patients with both HIV and culture-positive TB had rifampicin-resistant/MDR-TB while 7.8% (7/90) had isoniazid resistance without MDR-TB 25. The presence of the following risk factors should always raise suspicion of possible drug-resistant TB: Molecular tests for rifampicin resistance are useful when MDR-TB is suspected (e.g. in a recent immigrant from an area with a high prevalence of rifampicin-resistant disease), as a large proportion of rifampicin-resistant strains have isoniazid resistance as well 26 (see Appendix 3). In the UK, the majority of cases of TB occur in individuals from high- and medium-incidence settings 1, suggesting a substantial role for reactivation of latent infection. Individuals with LTBI are at increased risk of developing active TB, especially if they have recently acquired M. tuberculosis or are immunocompromised 1. HIV-positive individuals from countries with a high TB incidence, especially from sub-Saharan Africa, often present with TB as the first manifestation of immunosuppression, and mortality among HIV-positive with TB remains high 2. high and medium TB incidence as and respectively 3 (see 5 for TB incidence by LTBI for new to the UK from countries with high TB incidence is an effective as well as public 6 and is recommended by 3. WHO guidelines for countries with a low TB for LTBI in all HIV-positive However, it has recently been shown that this approach is unlikely to be in the UK 8. The risk of to active TB in the population is highest within the first years following M. tuberculosis infection and HIV-positive individuals with LTBI are more likely to to active TB than HIV-negative individuals 9. incidence of active TB is associated with low CD4+ cell including while on and with on the risk of TB among HIV-positive individuals, although it should be that in from countries of high TB incidence, such as those of sub-Saharan Africa, the risk of TB of HIV is high 2 (see For clinical a positive result in an individual with clinical or evidence of active TB Before for or active TB should be with a and The advantages of IGRAs include the of a blood test with need for to the These assays are more than although the may be for healthcare and possible specificity to individuals being for LTBI 14. Although the proportion of individuals with a positive result treatment for active TB with 15, a positive result even years treatment still indicate previously In that population, treatment for LTBI may be only if there has been significant new for LTBI with an and in HIV-positive individuals 3. However, in of and of a sensitivity among those with low CD4+ cell and results due to and to non-tuberculous limited data of using and to identify LTBI among those with low CD4+ cell the use of in a UK (see Appendix The study may a more evidence-based individuals born in including the UK, will be at greater risk of developing TB than for and LTBI in those from countries (e.g. the who have risk factors such as to a TB should be identified through routine or to or of in countries 3. factors of to HIV-positive individuals include: a of working in medical settings in TB drug receipt of for following and biological disease for In to the guidelines, that local for the increase in for on of patients and that it is to and to those at risk at their routine In women diagnosed with HIV, and LTBI in the as in individuals, including use of chest if clinically In this recommendation, have the risk of from treatment for in is associated with other risk factors (see using an similar to that by the WHO to exclude active TB 17. Other investigations may be for example chest or lymph node biopsy is clinically or through It is important to consider the of TB to because of the risk of among those with low CD4+ cell counts (see Section and Appendix There have been in HIV-positive individuals a of treatment for LTBI with an ranging from to (see Appendix that be considered, on individual and and for which there is evidence of 2 and 3 have been shown to be to 1 in of 19 and in the of TB treatment and of 3 or A of rifampicin has been shown to be effective in active TB, but there is evidence that cause more than isoniazid alone and they are therefore not recommended must be to with for LTBI treatment are not in this the of of in the non-specific in up to 20% of individuals but most of this is and only by elevated levels of (usually <100 24. isoniazid for clinical symptomatic is but can be if associated with other such as (e.g. or disease is and isoniazid can be continued with clinical and laboratory The risk of severe Group 3 or associated with isoniazid for LTBI is to studies should also be with due to treatment for the needs to be the risk of Individuals for LTBI should be of symptoms of such as vomiting, or or these those aged should be to their healthcare and if there is a in should treatment of isoniazid treatment for LTBI have included the risk of active TB as a Although there are concerns that isoniazid the of drug-resistant TB evidence from has shown significant between drug resistance and use of isoniazid and/or for LTBI 19, in of high TB incidence have shown that isoniazid in rates of TB and that isoniazid in TB incidence among HIV-positive a strategy may in which is more common than reactivation in such settings For the isoniazid need to be continued or at least until the CD4+ cell count had on and there are data to support such an in settings of lower TB It is clear that TB. It should be if not in and for those with active and LTBI (see Section and Appendix of TB treatment of active TB is therefore not recommended in the UK but should be For HIV-positive individuals with a of to drug-resistant TB to one or more there are limited data to support any of help management of such cases an management be from between the the HIV a in the management of drug-resistant TB, and public include: and the regarding early presentation with any symptoms of possible use a treatment for LTBI to which the patient's is to be and use a standard LTBI if there is thought to have been LTBI before the with drug-resistant disease The treatment of TB through an clinical to the current standard of of 2 of and by of rifampicin and isoniazid 1, 2 (see the use of where available. recent to TB to (e.g. by using have proved with high rates of TB should be during the in HIV-positive individuals, as this strategy has been associated with acquired resistance There is evidence that individuals with disseminated TB should more there is CNS clinicians use treatment to as per guidance for CNS TB TB even though may be is a with similar to rifampicin M. tuberculosis although have been in individuals The main of is that it allows the of (see Section and Appendix should be used as an to TB to the response to M. tuberculosis in those with meningitis. An of individuals with TB meningitis showed a in mortality among those who adjunctive during