Tuberculosis in Children Living With HIV: Ongoing Progress and Challenges

Abstract

There has been much recent progress on control of the tuberculosis (TB) and human immunodeficiency virus (HIV) epidemics globally. However, advances in children have lagged behind, and TB-HIV coinfection continues to be a major driver of pediatric mortality in many settings. This review highlights recent research findings in the areas of prevention, diagnosis, and treatment of HIV-associated childhood TB. Key areas for future research are defined. Current prevention efforts such as vaccination, TB symptom screening, and TB preventive treatment are demonstrated as beneficial but need to be optimized for children living with HIV (CLHIV). Diagnosis of HIV-associated TB in children remains a major challenge, depending heavily on clinicians’ ability to judge an array of signs, symptoms, and imaging findings, but there are a growing number of promising diagnostic tools with improved accuracy and feasibility. Treatment of TB-HIV coinfection has also seen recent progress with more evidence demonstrating the safety and effectiveness of shorter regimens for treatment of TB infection and disease and improved understanding of interactions between antiretrovirals and TB medications. However, several evidence gaps on drug-drug interactions persist, especially for young children and those with drug-resistant TB. Accelerated efforts are needed in these areas to build upon current progress and reduce the burden of TB on CLHIV.

INTRODUCTION

The World Health Organization (WHO) estimates that less than half of children with tuberculosis (TB) < 15 years old are accurately diagnosed and reported [1]. Challenges to TB case finding in children have been exacerbated by the COVID-19 pandemic [2], but pre-pandemic estimates from 2019 show that children living with human immunodeficiency virus (CLHIV) < 15 years old have the highest mortality to incidence ratio (Table 1).

Table 1.

Global Estimates of Incident TB Cases and TB Mortality, 2019

Global Population	Estimated TB Incidence	Estimated TB Mortality	Ratio of TB Mortality to Incidence
All ages [3]	9 960 000	1 400 000	0.14
Living with HIV, all ages [3]	815 000	208 000	0.26
All < 15 years old [3]	1 190 000	230 000	0.19
Living with HIV, <15 years old [4]	62 000	22 000	0.35

Abbreviations: HIV, human immunodeficiency virus; TB, tuberculosis.

CLHIV are a well-known high-risk group for TB [5, 6], and despite progress over the past two decades to address the TB and HIV epidemics globally [2, 7], TB-HIV coinfection in children remains a common clinical challenge in endemic settings. Sub-Saharan Africa is the most affected region, with 22 of the 30 WHO high TB-HIV burden countries [2] and an estimated 36% of child TB deaths [8]. For CLHIV, TB is most aggressive and deadly in the first few years of life [5, 9]. Although data on the underlying cause of death in CLHIV are severely limited, recent well-designed, small studies report TB in 15% [10], 16% [11], and 18% [12] of CLHIV who died.

The objective of this review is to summarize new evidence on prevention, diagnosis, and treatment of TB-HIV coinfection in children published since 2018 [13, 14], and define key areas for future research to eliminate TB-HIV coinfection in children (Table 2).

Table 2.

Research Priorities for the Prevention, Diagnosis, and Treatment of TB in CLHIV

Area	Research Priorities
Prevention	• Development of a safe and effective TB vaccine that can be used for CLHIV, including those with severe immunosuppression • Implementation research to maximize HIV virologic suppression with ART for all CLHIV, and detailing the impact this has on the burden of TB in CLHIV • For screening purposes, in conjunction with currently recommended symptom screening, assessment of the accuracy, feasibility, and cost-effectiveness of tests such as chest radiography (including use of computer aided detection, CAD, software for automated reading), c-reactive protein, and other tests currently used for diagnostic purposes
Diagnosis	• Further detail the accuracy of promising new diagnostic tests such as Ultra and LAM assays, particularly with non-invasively collected specimens such as throat swabs, stool, and urine • Assessment of differences in the accuracy of urine LAM testing by age, collection technique, and immune status • Ongoing development of whole-blood gene-expression signature testing that is accurate and feasible in low-income countries • Development of novel strategies (eg. algorithms and scoring systems based upon local data) to support clinicians in TB treatment decision making
Treatment	• Assessment of how TPT effectiveness varies by TPT duration, immune status, age, and local epidemiology of TB for CLHIV to better determine who most benefits from TPT • Comparison of safety and efficacy of rifamycin-containing TPT regimens with isoniazid • Investigation of barriers to TPT uptake and strategies to improve uptake and good adherence • For CLHIV under six years old, determine the pharmacokinetics of dolutegravir exposure with twice daily dosing while on rifampicin containing ATT • Basic investigations in CLHIV of pharmacokinetics of ART and second-line ATT • Assess effectiveness and safety of four-month ATT regimens for non-severe, drug-susceptible TB in CLHIV • Determine appropriate timing of ART initiation for those being treated for CNS tuberculosis

Abbreviations: ART, antiretroviral therapy; ATT, antituberculosis therapy; CLHIV, children living with HIV; HIV, human immunodeficiency virus; LAM, lipoarabinomannan; TB, tuberculosis; TPT, TB preventive treatment.

PREVENTION

Vaccination

TB vaccines are detailed in a separate article in this Supplement [15]. Vaccination of HIV-negative neonates in high-TB burden settings with bacillus Calmette-Guerín (BCG) is strongly recommended by the WHO [16]. While BCG vaccination has long been known to protect young children from miliary TB and TB meningitis [16, 17], a recent individual-participant data meta-analysis of children exposed to infectious cases of TB demonstrated protection from all forms of TB in children under 5 years old (odds ratio: 0.64, 95% credible interval [CI]: 0.50 to 0.84) [6]. The WHO recommends delaying BCG vaccination of infants known to be living with HIV given the risk for local and disseminated complications of BCG disease until they are established on antiretroviral therapy (ART) and are clinically well [16]. In practice, especially in settings where BCG vaccination of newborns is routine, most infants who acquire HIV through vertical transmission have unknown HIV status in the early neonatal period. Consistent with WHO guidelines, we recommend that providers should not hesitate to give BCG to newborns with unknown HIV status if no clinical evidence suggests HIV infection, given the programmatic challenges of having HIV testing results prior to BCG vaccination of newborns and accessing BCG vaccination later, and given that the benefits of broadly vaccinating newborns with BCG, including those with unknown HIV status, likely outweigh the risk of BCG disease in those with HIV, especially if on ART.

Multiple vaccines to protect against TB are in development. Demonstrating safety and effectiveness of candidate vaccines in CLHIV should be a priority, but many challenges exist: (1) like BCG, live vaccines may be unsafe in immunocompromised CLHIV, (2) immunocompromised CLHIV, who are most vulnerable to severe TB, tend to have the poorest immune responses to vaccination, and (3) early diagnosis of infants living with HIV, due to either delayed testing or delayed turnaround time of testing results, is poor in many settings, making targeted vaccination of this vulnerable population logistically challenging [17]. The most promising TB vaccine in development for healthy CLHIV is the protein-subunit vaccine M72/AS01_E. This vaccine has yet to be specifically studied in CLHIV, but data from phase I and II trials in healthy adults living with HIV and infants and adolescents without HIV demonstrate that it is safe and immunogenic [18]. This vaccine has only been demonstrated to have clinical efficacy for preventing progression to TB disease in adults with TB infection, and large clinical trials will be needed to demonstrate efficacy for the prevention of TB disease in other populations, such as CLHIV [17, 18].

Antiretroviral Therapy

Effective ART dramatically reduces morbidity and mortality for CLHIV, and since 2015 the WHO has recommended that all CLHIV initiate ART regardless of clinical or immunological status [19]. A systematic review and meta-analysis published in 2017 reported that CLHIV are at lower risk for TB when on ART (hazard ratio: 0.30, 95% CI: 0.21 to 0.39) [20]. A recent report from South Africa on declining national TB case notification rates for children and adolescents showed slower decline for CLHIV under 5 years old compared to other groups [21], emphasizing the need for more focus on vulnerable young CLHIV. The steps of early HIV diagnosis, starting effective ART early, and achieving virologic suppression on ART for CLHIV remain major challenges, with only 40% of CLHIV globally estimated to be virologically suppressed [7]. A major priority to reduce the burden of TB in CLHIV is ensuring access to, and support for, safe, palatable, and effective ART.

Screening

Effective screening for TB in CLHIV should (1) promote early diagnosis in those who screen positive, thereby preventing progression to severe disease and (2) promote uptake of TB preventive treatment (TPT) in those who screen negative, thereby preventing incident disease. For people living with HIV, the WHO recommends a symptom-based screening approach. For CLHIV younger than 10 years old, any current cough, fever, history of TB exposure, or poor weight gain constitutes a positive symptom screen requiring further TB investigation. A prospective study of the WHO symptom-based screening approach among 247 South African CLHIV under eight years old reported 57% sensitivity and 97% specificity for TB disease [22]. A larger retrospective study of this screening approach among 20 706 people living with HIV < 19 years old in 6 African countries reported similar performance: 61% sensitivity and 89% specificity, although screening results tended to be less discriminatory for TB disease in children < 7 years [23]. The positive predictive value in that study was only 3% [23], indicating an abundance of false-positive screens that can lead to under-utilization of TPT (discussed further in the “Treatment” section) and unnecessary TB diagnostic testing. More research is needed, but these results emphasize the importance of close clinical follow-up, including repeated screening for symptoms and TB exposure at every clinical encounter.

Ultimately, symptom-based screening alone is unlikely to provide optimal accuracy. Other TB screening methods in people living with HIV recently recommended by the WHO are chest radiography (including the use of computer-aided detection software for automated reading), C-reactive protein, and rapid diagnostic tests (see below) [24]. None of these have been adequately studied in CLHIV [25].

DIAGNOSIS

Definitive diagnosis of TB with microbiological testing in children is challenging. Obtaining respiratory specimens for TB diagnosis in young children traditionally requires invasive, laborious sampling techniques such as gastric aspiration, nasopharyngeal aspirates, or sputum induction. Even when an adequate specimen is obtained, microbiological testing by culture or nucleic acid detection has poor sensitivity as disease is often paucibacillary [26]. Therefore, TB diagnosis in children is often clinical, relying upon assessment of exposure to infectious TB, nonspecific signs and symptoms, and if available, imaging such as chest radiography and tests such as tuberculin skin testing or interferon-gamma release assays. In CLHIV, diagnosis is additionally complicated by other opportunistic lung or systemic infections that clinically present like TB. For example, the PAANTHER study assessing over 400 CLHIV suspected to have TB from four countries found isolation of non-tuberculous mycobacteria from respiratory specimens to be almost as common (10.8%) as microbiological detection of TB (11.7%) [27].

Imaging

There is growing evidence describing the limitations of chest radiography for diagnosing pulmonary TB in children, including inadequate sensitivity and specificity, poor inter-reader reliability, and lack of equipment and expertise for conducting and interpreting chest radiography in some settings [28, 29]. The PAANTHER study estimated chest radiography to have a sensitivity of 71.4% (95% CI: 58.8 to 84.1) and specificity of 50.0% (95% CI: 41.9 to 58.1) for diagnosis of TB even with a consensus of three expert readers [30].

The use of point of care ultrasound (POCUS), performed at the bedside by trained clinicians, is a growing area of interest. POCUS can increase detection of pleural effusions and enlarged mediastinal lymph nodes in children diagnosed with TB, with higher inter-reader agreement with POCUS compared to chest radiography [31]. Formal ultrasonography and POCUS are particularly useful for identifying evidence of abdominal TB in CLHIV [32, 33].

Rapid Diagnostic Tests

In children, the GeneXpert MTB/RIF (Xpert, Cephid Inc, USA) assay has a sensitivity of 65% compared to culture from a respiratory specimen [34]. GeneXpert MTB/RIF Ultra (Ultra, Cephid Inc, USA), the next-generation diagnostic utilizing the same platform as the original Xpert, has superior sensitivity compared to Xpert while maintaining excellent specificity for the diagnosis of TB in children [34]. Increasing evidence describes the utility of Ultra testing on stool [35], a more child-friendly sample compared to respiratory samples. The diagnostic accuracy of Xpert on child stool specimens is similar to respiratory specimens, with possibly higher sensitivity on stool from CLHIV [34].

Lipoarabinomannan (LAM) is a component of the mycobacterial cell wall that can be present in the urine of people with TB. Urine LAM assays are primarily used in people living with HIV, who are often immunosuppressed, leading to a higher occurrence of disseminated TB and the presence of LAM in urine. The Alere Determine TB-LAM (LF-LAM; Abbott, Lake Forest, IL, USA) is a commercially available, inexpensive, point-of-care test recommended by the WHO since 2015 for TB diagnosis in CLHIV with severe immunosuppression or severe TB disease symptoms [36]. LF-LAM predicted mortality in a cohort of hospitalized, ART-naïve CLHIV < 13 years old in Kenya [37]. However, the sensitivity of LF-LAM for detection of TB is suboptimal, ranging from 36% to 65% compared to the reference standard of microbiologic confirmation in 5 diagnostic accuracy studies in CLHIV [38–42].

The Fujifilm SILVLAMP TB LAM (FujiLAM; Fujifilm, Tokyo, Japan) is a novel LAM point-of-care test expected to be commercially available in 2022 but not yet recommended for clinical use by the WHO. Superior sensitivity of FujiLAM compared to LF-LAM for detection of TB in adults living with HIV has been reported from two large studies [43, 44]. Two small studies in CLHIV also suggest better sensitivity of FujiLAM versus LF-LAM (55% vs 37% [38] and 60% vs 36% [39], respectively). More research is needed to better describe the accuracy of urine LAM tests in CLHIV, including subgroup analyses by age and immune status, to determine which CLHIV are most likely to benefit from these tests. Most importantly, though highly specific for TB when positive, its low sensitivity means that urine LAM is not a rule-out test for TB.

Whole-blood gene-expression signature testing for TB continues to be improved, and in adults living with HIV, this type of testing has recently been reported to have high sensitivity and specificity when used for either screening [45] or diagnosis [46] of TB disease. However, this testing is not well studied in CLHIV, and even for adults, these tests are not approved for routine use [26].

Treatment Decision Making

Diagnosis of TB in CLHIV must utilize multiple tools available to a clinician, starting with a detailed history and physical examination. In clinical practice, the decision to treat a child for TB depends not only on the certainty of diagnosis but also on the consequences of a missed diagnosis. Therefore, for children at risk for rapid progression of disease, such as infants and those with advanced HIV, lower certainty of diagnosis may be appropriate to trigger treatment initiation [47].

The decision to treat for TB in CLHIV depends heavily on clinicians’ experience, highlighting the importance of strengthening the capacity of front-line clinicians to suspect and diagnose TB in children [1]. Pragmatic, more objective approaches to support clinicians to diagnose or rule-out TB in children are needed. Using clinical data from the PAANTHER study, a treatment decision score and step-wise algorithm for CLHIV was developed using signs and symptoms of TB and findings from chest radiography, abdominal ultrasound, and Xpert MTB/RIF on sputum, gastric, and stool specimens. Though the reported sensitivity was about 90%, the specificity of up to 62% highlights the need for still better diagnostic approaches in this vulnerable population [33].

For CLHIV enrolled in care for ART, the opportunity to assess those suspected to have TB over time can be leveraged by clinicians to both diagnose TB earlier and more definitively rule out TB for those with alternative diagnoses during follow-up. This highlights the importance of integrated HIV/TB services for children.

TREATMENT

Tuberculosis Preventive Treatment

Since 2011, the WHO has continued to recommended that CLHIV aged 1 year or older in high TB-incidence settings should receive TPT if there are no contraindications and they have a negative symptom screen, regardless of whether the child is on ART and even without known exposure to an infectious TB case [48]. Ongoing symptom and TB exposure screening at each clinical encounter is needed to assess for future TB disease in CLHIV, who remains at risk irrespective of TPT use. The WHO recommends at least 36 months of isoniazid TPT for adolescents and adults living with HIV in high TB-incidence settings [48]. This “conditional recommendation” is based on the results of a meta-analysis of three studies published in 2011 and 2012 with the majority of study participants over 18 years old and not on ART [49]. However, the duration of TPT for CLHIV in high TB-incidence settings is not well defined, and more evidence is urgently needed. Risk stratification based upon ART status and evidence of TB infection may identify CLHIV who would most benefit from longer TPT.

The most common TPT regimen in CLHIV is isoniazid monotherapy. However, there has recently been an expansion of recommended TPT options, including daily rifampicin and isoniazid (RH) and weekly rifapentine and isoniazid (HP), that allow for shorter duration of therapy (eg, 3HR, 3HP, or 4R) compared to isoniazid monotherapy. Rifamycin-containing options must be considered carefully in CLHIV given common drug-drug interactions with ART, as described in the next section. A recent meta-analysis of TPT mainly in adults living with HIV demonstrated that compared to isoniazid-containing TPT regimens, shorter rifamycin-containing regimens were safer and had similar effectiveness [50]. One large study included in that meta-analysis demonstrated that a 1-month regimen of daily rifapentine and isoniazid was non-inferior to 9 months of isoniazid alone for preventing TB in adults and adolescents living with HIV with evidence of TB infection or living in high TB burden settings [51]. Clinical trials in CLHIV are underway to assess shorter-duration TPT regimens. Many CLHIV still do not access and benefit from TPT due to a variety of modifiable barriers [52, 53]. Implementation research is needed to improve uptake of this effective intervention. More accurate, pragmatic strategies for ruling out TB disease in populations eligible for TPT, including CLHIV, would likely increase uptake of TPT by improving clinicians’ confidence that a particular child does not have TB disease and it safe to start them on TPT. In addition, strengthening of healthcare services and improving and standardizing recording and reporting of TPT initiation and completion will also improve TPT delivery.

Treatment of TB Disease

New opportunities and ongoing challenges for treating drug-susceptible, drug-resistant, and central nervous system (CNS) TB in children have recently been reviewed [54], including in this Supplement [55], and also apply to CLHIV. Key considerations for treating TB in CLHIV include adjusting the ART regimen for those already on ART and appropriate timing of ART initiation for those not yet started on ART. Drug-drug interactions between certain antiretrovirals and antituberculosis therapy (ATT), especially rifamycins, which are powerful inducers of the cytochrome P450 enzymes, often complicate co-treatment of TB and HIV. A systematic review of the pharmacokinetics of ATT and ART co-administration was recently published [56]. Data to guide dose adjustments of ART in CLHIV being treated with second-line ATT for drug-resistant TB are severely limited [56]. Recommendations for adjusting key ART drugs while treating CLHIV for TB with rifamycin-based ATT are summarized in Table 3.

Table 3.

Recommended Dosing Adjustments for Key Antiretrovirals in CLHIV Receiving Rifamycin-Containing anti-TB Therapy (ATT) [56]

Antiretroviral	Dosing Adjustment	Comments
Abacavir	No adjustment
Dolutegravir (DTG)	Increase from once to twice daily [57]	Twice daily dosing of DTG while on rifampicin-containing ATT in children 6 years old and above resulted in adequate DTG exposures [58]. Data on DTG exposure while on rifampicin-containing ATT for CLHIV under 6 years old are limited.
Lamivudine	No adjustment
Lopinavir/ritonavir (LPV/r)	“Super-boosting” to achieve 1:1 LPV/r ratio (standard LPV/r ratio is 4:1)	Double-dose lopinavir-ritonavir (LPV/r ratio 8:2) in children < 3 years old [59], and thrice-daily LPV/r (4:1) [60] do not achieve target LPV troughs in a high proportion of children. Super-boosting of LPV/r (1:1) with rifampicin-containing TB treatment resulted in LPV troughs that were non-inferior to standard LPV/r (4:1) without rifampicin in young children [61].
Raltegravir (RAL)	Double each dose	This dosing adjustment safely achieves adequate RAL exposure in children 2 years and above [62], and aged down to 4 weeks with more recent data [63].
Tenofovir disoproxil fumarate	No adjustment
Zidovudine	No adjustment

The SHINE randomized open-label trial compared 4 versus 6 months of standard-dose treatment with rifampicin, isoniazid, and pyrazinamide with or without ethambutol for 1204 children < 16 years old (including 127 CLHIV) with smear-negative, non-severe TB. For all enrolled, 4 months of treatment was non-inferior to the standard 6-month treatment for the primary composite outcome of TB-free survival at 72 weeks after randomization [64]. This study was not powered to assess non-inferiority of 4-month treatment in the sub-group of CLHIV, and this will be an important area of investigation as the WHO has introduced recommendations in 2022 for 4-month treatment of children with non-severe, drug-susceptible TB, which will be rapidly implemented by national TB programs globally.

Timing of ART Initiation

For ART naïve children diagnosed with TB, ATT should be started promptly. The risk of immune reconstitution inflammatory syndrome (IRIS) with initiation of ART is higher early in the course of ATT or if the diagnosis of TB is missed and ATT is not started. IRIS of the CNS can be complex and the potentially severe, with complications such as stroke. The risk of IRIS must be balanced against the risks of delaying ART, which are proportional to the degree of immunosuppression. With these considerations and evidence mostly extrapolated from studies of adults with TB-HIV coinfection, major guidelines [65, 66] recommend that for CLHIV diagnosed with TB, ART should be initiated:

• as soon as possible and within 2 weeks for those who are severely ill;

• within 8 weeks for those who are not severely ill;

• cautiously in those with CNS TB, most expert guidelines recommend delay of approximately 4 weeks

CONCLUSIONS

In conclusion, there has appropriately been prioritization of the “cursed duet” of TB and HIV from the perspectives of clinical and population research, public health and policy, and clinical medicine. However, many of these advances have not been realized for the pediatric population, and CLHIV continue to suffer from poor access to effective TB preventative measures, underdiagnosis of TB, and burdensome ART and ATT regimens that urgently need to be optimized for this population. COVID-19 has reversed many positive gains in the control of TB in CLHIV. Accelerated efforts are needed in these areas to build upon recent progress and minimize the burden of TB on CLHIV.

Note

Supplement sponsorship. This article appears as part of the supplement “What’s New in Childhood Tuberculosis?” sponsored by the Stop TB Partnership.