13.1. Recommendations

13.1.1. Isoniazid preventive therapy

1. All HIV-infected infants and children exposed to TB through household contacts, but with no evidence of active disease, should begin isoniazid preventive therapy (IPT).
   (Strong recommendation, very low quality of evidence)
2. Children living with HIV (>12 months of age and including those previously treated for TB), who are not likely to have active TB and are not known to be exposed to TB, should receive 6 months of IPT as part of a comprehensive package of HIV care.
   (Strong recommendation, moderate quality of evidence)
3. Infants living with HIV, who are unlikely to have active TB and are not known to be exposed, should not receive IPT as part of a comprehensive package of HIV care.
   (Conditional recommendation, very low quality of evidence)
4. The recommended dose of isoniazid (INH) for preventive therapy in HIV coinfection is 10 mg/kg daily for 6 months (maximum 300 mg/day). (For simplified dosing see Table 18)

Table 18

Simplified, weight-based dosing for isoniazid 10 mg/kg/day.

13.2. Background

Tuberculosis represents a significant threat to child health. HIV infection increases the susceptibility to infection with M. tuberculosis, the risk of rapid progression to TB disease, and reactivation of disease in older children with latent TB. Increasing levels of coinfection with TB and HIV in children have been reported from resource-limited countries [133], with the prevalence of HIV in TB-infected children ranging from 10% to 60% [134-139].

13.3. TB screening and prevention for HIV-infected infants and children

All HIV-infected infants and children should be evaluated for contact with a TB source case at every visit to a health-care facility. Those presenting with poor weight gainⁱ, current cough or fever should be evaluated for TB, and those with active TB disease should be treated [140].

Provide HIV-infected infants and children who present with no evidence of active TB, but who are a household contact of a TB patient, with IPT at a dose of of 10 mg/kg/day for 6 months (not to exceed a maximum daily dose of 300 mg) (see Table 18). The child should be seen monthly and provided one month's supply of INH at each visit. Overall, there is a paucity of evidence concerning the use of IPT in infants. Some experts believe that IPT should be provided for all HIV-infected infants (as is recommended for TB-exposed infants) in order reduce their risk of acquiring TB, which has high rates of morbidity and mortality [141-144]. However, research to date does not support this [145].

HIV-exposed infants who are identified as HIV-infected within 3 – 4 months of age should be monitored monthly, including an assessment for TB and for possible contact with an active TB case, but should not be prescribed IPT. However, if contact with an active TB case has occurred, TB infection should be strongly considered and once active TB disease has been excluded, IPT for 6 months must be given. In many settings, infants presenting late for HIV diagnosis have a high likelihood of TB coinfection.

Isoniazid preventive therapy is recommended for all HIV-infected children older han one year of age who are unlikely to have active TB, even in the absence of documented exposure to an active TB source case [146]. A randomized study showed that the provision of IPT to HIV-infected children reduces the incidence of TB by 72% and all-cause mortality by 64% [143], confirming the beneficial effect of IPT observed in the adult population [147]. However, findings from a randomized controlled trial performed in South Africa [145] showed that there is no benefit from IPT when HIV-infected infants with no known exposure to an active TB case are identified in the first 3 to 4 months of life, and started immediately on ART and carefully monitored for new TB exposure or disease on a monthly basis.

(See the graphic description for TB screening and initiation of IPT in Annex J, Figure 11.)

Additionally, all HIV-exposed infants and children should benefit from co-trimoxazole preventive therapy (see http://www.who.int/hiv/pub/paediatric/co-trimoxazole/en/index.html). Co-trimoxazole preventive therapy is particularly important for children coinfected with TB and HIV [148, 149]. Studies in adults and children indicate better survival rates in patients coinfected with HIV and TB who received co-trimoxazole preventive therapy compared with those who didn't [150]. All eff ort must be made to identify HIV/TB coinfected infants and children in order to prevent TB disease [143].

13.4. Diagnosis of TB

Primary TB disease in children presents with a broad range of non-pulmonary and pulmonary manifestations.

In paediatric TB, bacteriological confirmation should be sought whenever possible, even if it is difficult and the results frequently negative. Appropriate clinical samples may include sputum (by expectoration, gastric aspiration or sputum induction), fine-needle aspiration of enlarged lymph nodes, pleural fluid or ear swab from chronically discharging ears. Gastric aspiration is advisable only when culture is available, as smear positivity on microscopy is low and the procedure is stressful for the child. Alternatively, induced sputum can be used for young children as this does not require hospital admission and the yield is greater than that for gastric aspirate. The disadvantage is the need for infection control to protect health-care personnel.

In many cases, particularly in young children, diagnosis is presumptive and based on a number of factors. These include; a constellation of clinical signs and symptoms, known contact with a household member with TB disease, a positive tuberculin skin test (TST)ⁱⁱ and radiological findings on chest x-ray. Different scoring systems have been proposed to assist the clinician; unfortunately, they are difficult to validate and perform poorly in HIV-infected children [151, 152]. A “trial of TB treatment” should not be used as a diagnostic test for TB in children.

Encouraging data show that interferon-gamma release assays (IGRAs) are more sensitive than TST in detecting TB in HIV-infected children, including those with a low CD4 count and/or malnutrition [153-155]. In addition, excellent specificity for M. tuberculosis infection has been reported and unlike TST, IGRAs are unaffected by prior BCG vaccination or exposure to environmental mycobacteria. However, more evidence is needed and range of implementation issues that are relevant in most TB/HIV endemic settings (e.g. cost, specific laboratory equipment and the need for a venous blood sample) must be solved. Therefore, the use of IGRAs is currently not recommended outside research settings with laboratory-validated procedures.

13.5. Treatment of TB in HIV-infected infants and children

The underlying principles for the treatment of TB in HIV-infected children are the same as for children who are not HIV-infected. However, the co-management of TB and HIV, and the treatment of HIV infection, is complicated by drug interactions, particularly between rifampicin and the NNRTI and PI classes of ARVs. These drugs have similar routes of metabolism and co-administration may result in sub-therapeutic drug levels. ART should not be interrupted but dose adjustments of ART may be needed when taken with the rifamycins, especially rifampicin. The potential use of rifabutin, considered in adults to overcome drug – drug interactions, is not recommended due to insufficient data and lack of an available formulation in children. In addition, the choice of ART regimen in TB/HIV coinfected children is complicated by the relatively limited number of available paediatric ARV formulations and the lack of dosing information for some ARVs (particularly for children less than 3 years of age).

13.6. Choice of first-line ARV regimens in children receiving rifampicin-containing TB treatment

Pharmacokinetic data are limited regarding the concomitant use of rifampicin and ARVs in children, though some studies are under way as these guidelines are being developed. Updated evidence-based guidance will be made available on the WHO website (www.who.int) as data become available. The current treatment recommendations for children with TB/HIV coinfection are as follows:

In adults, both standard (600 mg) and increased (800 mg) EFV doses have been used with rifampicin. However, adequate virological and immunological response with standard 600 mg dosing has been documented [156], and higher doses are associated with a higher incidence of toxicity and are not recommended. NVP levels are reduced with concurrent administration of rifampicin, with reductions in the area under the curve (AUC) of 31–37% [157, 158]. The use of higher doses of NVP with rifampicin has not been evaluated. Additionally, NVP, like rifampicin and INH, has potential hepatotoxicity. As with EFV, some clinical reports indicate adequate virological and immunological responses and acceptable toxicity with standard doses of NVP administered concomitantly with rifampicin [159]. Additionally, in early 2009, new evidence emerged based on small studies in children, suggesting that co-administration of rifampicin may not result in lowered NNRTI concentrations in paediatric populations. Therefore, prescribing the maximum dose of 200mg/m² is currently recommended as a safer approach to avoid sub-therapeutic drug levels, but more data are urgently needed to address the exact dose requirements. A regimen of two NRTIs plus NVP should be considered only when careful clinical and laboratory monitoring for potential liver toxicity can be assured.

Alternatively, a triple NRTI regimen can be used, especially in infants previously exposed to NVP. A previous study of HIV-infected adults demonstrated that a regimen of AZT/3TC/ABC has lower virological potency than an EFV-based regimen (79% versus 89% efficacy at 32 weeks) [59]. However, recent findings (NORA substudy of the DART trial) provide reassurance on the triple NRTI regimen and show that patients on triple NRTIs do not appear to have any clinical disadvantage [162, 163].

13.7. When to start ART following initiation of rifampicin-containing TB treatment

ART is a priority for children with WHO clinical stages 3 and 4. Because the degree of immunodeficiency in TB/HIV coinfected children is highly correlated with mortality [133], earlier initiation of ART is critical in coinfected children with low CD4 values.

In HIV-infected children with TB disease, initiation of TB treatment is the priority. However, the optimal timing for the initiation of ART during TB treatment is not known. The decision on when to start ART after starting TB treatment involves a balance between the child's age, pill burden, potential drug interactions, overlapping toxicities and the possible development of IRIS versus the risk of further progression of immune suppression with the associated increase in mortality and morbidity.

Early initiation of ART is advocated for any patient with TB disease regardless of clinical stage and degree of immunosuppression. A randomized controlled trial (SAPIT) provides moderate evidence for early initiation of ART in terms of reduced all-cause mortality and improved TB outcomes in adults [164]. Trial participants were grouped into “integrated” (immediate and end of initiation phases combined) and “sequential” treatment arms. Mortality was 55% lower in the integrated treatment arm compared with the sequential treatment arm. Similar data are not available in the paediatric population and additional research is urgently needed to address this issue. However, the expected benefits in terms of reduction in mortality and TB transmission outweigh the potential concerns related to onset of IRIS due to TB treatment and drug – drug interactions. Moreover, harmonization with adult recommendations will likely facilitate programme uptake.

Therefore, the current recommendation is that any child with active TB disease should begin TB treatment immediately and begin ART as soon as tolerated (2 to 8 weeks into TB therapy), irrespective of clinical stage and degree of immunosuppression.

The potential for IRIS (see below) should be considered in all children starting ART, particularly in those with low CD4 values.

13.8. Considerations for children diagnosed with drug-resistant TB

There are few data on the care of HIV-infected children who are exposed to multidrug-resistant TB (MDR-TB) or extensively drug-resistant TB (XDR-TB). These children will require referral to local TB experts familiar with the regional drug resistance profile. Treatment of the drug-resistant source case must be a priority. Initiation of ART at the earliest possible opportunity should also be ensured.

More detailed guidance is being developed by the WHO department of TB to address infection control for TB in infants, as well as guidance on managing XDR- and MDR-TB in HIV-infected children. The most up-to-date information on TB/HIV co-infection can be found at http://www.who.int/tb.

13.9. Considerations for children diagnosed with TB while on first-line ARV regimens

ART should continue in children already on a first-line ARV regimen who are subsequently diagnosed with TB. However, the ARV regimen should be reviewed and may need adjustment in order to ensure optimal treatment of both TB and HIV and to decrease the potential for toxicities and drug – drug interactions. In children on a standard NNRTI-based first-line regimen who develop TB due to either primary infection or unmasking of existing TB (see Section 13.11), consider substituting NNRTI-based therapy with a triple NRTI regimen. Alternatively, the children could remain on their standard regimen of two NRTIs + one NNRTI, which is preferred in children receiving EFV-based regimens. In children and adolescents for whom EFV is not recommended and who are taking NVP, an increased dose (up to a maximum of 200 mg/m²) should be administered. Because of possible overlapping toxicities and drug – drug interactions, children who are given rifampicin and NVP concomitantly should be followed up more frequently and laboratory parameters, if available, should be checked.

Where TB is being considered as a sign of treatment failure of the first-line regimen, switching to a second-line regimen should be considered if the child has taken ART for an adequate time (i.e. more than 24 weeks), has initially responded to it, and has not responded to anti-TB treatment. Consultation with a paediatric TB expert is suggested for the construction of a second-line regimen.

13.10. Considerations for children diagnosed with TB while on RTV-boosted PI ARV regimens

For children who are receiving a second-line regimen with RTV-boosted PIs and who are diagnosed with TB, the choice of ART regimen is more difficult because of likely resistance to first-line NRTI and NNRTI drugs and varying interactions between rifampicin and the PIs. Unboosted PIs are not recommended to be administered with rifampicin because of the decrease in PI drug levels.

Although there are few published data relating to children, for those coinfected with TB/HIV whose anti-TB treatment includes rifampicin, the dosage of RTV in the LPV/r regimen can be increased to a ratio of 1:1 in order to achieve adequate LPV exposure. Where available, monitoring of therapeutic drug levels should be considered when these medications are co-administered. Further research on the toxicity and efficacy of super boosted LPV/r is, however, still needed [165]. The presence of a cold chain should be ensured if heat-stable formulations are not available. One study of this regimen used by healthy adults demonstrated a high incidence of adverse reactions; in another study, the only one published to date of TB/HIV coinfected children, this regimen was generally well tolerated [165].

The use of SQV with higher-dose (i.e. full-dose) RTV boosting has been suggested but, because of significant hepatocellular toxicity in adults receiving this combination with rifampicin, concomitant administration of rifampicin with RTV-boosted SQV as part of ART is not recommended.

NFV should not be administered with rifampicin and the use of other boosted PI combinations is discouraged until further data become available [166, 167]. Reassessment and referral for the construction of a salvage regimen, as appropriate, are indicated.

Table 19, Table 20 and Annex J, Figure 11 summarize the WHO recommendations for ART in HIV-infected children diagnosed with TB. Research is urgently needed to evaluate the pharmacokinetics and clinical outcomes of administration of NNRTIs and PIs with rifampicin in children so that evidence-based recommendations can be made.

Table 19

Recommendations for the timing of ART following the initiation of TB treatment with a rifampicin-containing regimen in HIV-infected infants and children.

Table 20

Recommendations for co-management of TB and HIV in infants and children diagnosed with TB while on ART.

13.11. IRIS in the context of co-therapy for TB/HIV

IRIS has been observed in children receiving anti-TB therapy who have initiated ART. This syndrome is primarily reported in adults, but is also seen in children (see Section 8.6 for more on IRIS).

A clinical case definition of paradoxical TB-associated IRIS has recently been proposed (Table 21). Where available, an increase in CD4 count and a decrease in viral load since start of ART would further support the diagnosis of IRIS.

Table 21

Proposed clinical case definition of paradoxical TB-associated IRIS.

Some cases of IRIS in HIV-infected children may in fact be TB. Other cases may be localized or disseminated BCG disease in children who have received a BCG vaccination. HIV-infected children suspected of having disseminated BCG disease should be referred to an appropriate expert for management, as the diagnosis of BCG disease is difficult and the treatment is specialized [141, 168].

Footnotes

i

“Poor weight gain” is defined as: 1. reported weight loss, or 2. very low weight (weight-for-age less than -3 z-score), or 3. underweight (weight-for-age less than -2 z-score), or 4. confirmed weight loss (>5%) since the last visit, or 5. growth curve flattening.

ii

In an HIV-infected child, induration of >5 mm diameter is read as a positive TST; however, coinfected children may also present with a negative TST.