Article Text

Download PDFPDF

Drug therapy for children with tuberculosis
  1. L Peter Ormerod1,2,3
  1. 1Chest Clinic, Royal Blackburn Hospital, Blackburn, UK
  2. 2Lancashire Postgraduate School of Medicine, University of Central Lancashire, Preston, UK
  3. 3University of Manchester, Manchester, UK
  1. Correspond toProfessor L Peter Ormerod, Chest Clinic, Royal Blackburn Hospital, Blackburn, Lancs, BBE 3HH; Lawrencve.Ormerod{at}


The scientific basis of drug treatment for both active tuberculosis (TB) disease and TB infection, has been established, with treatment in children being largely extrapolated from adult active disease trials. It is essential that active TB disease is excluded before asymptomatic TB infection is diagnosed and treated. Nearly half of all children with active TB disease are found as asymptomatic tuberculin, or interferon gamma release assay (IGRA), positive contacts on screening by local TB services, usually of sputum TB microscopy positive adult relatives or other index cases, but with evidence of lung infiltrate or mediastinal lymphadenopathy on the child's chest x-ray. New drug regimens for both active disease and latent infection are in development, and also some novel drugs. However, none of these have yet been tested in children, and so again data will need to be extrapolated from adult results. In addition, there are issues regarding pharmacokinetics and dosing for current drugs, particularly isoniazid.

  • Medical Education
  • Medical Journalism

Statistics from

Few, if any clinical trials, on the treatment of active tuberculosis (TB) disease have been carried out in children; studies are (almost) entirely in adults, so we rely on extrapolated data from them.

The scientific basis of chemotherapy

Each of the main anti-TB drugs varies in its capacity to kill bacteria, to sterilise lesions and to prevent the emergence of drug resistance. Killing capacity is measured by early bactericidal activity,1 and sterilising ability by a low relapse rate after cessation of treatment and a low culture-positive rate at 2 months of treatment.2 The efficiency of a drug in preventing the emergence of drug resistance is more difficult to assess and is largely derived from the interpretation of clinical studies.2

Isoniazid is the most potent bactericidal drug and kills more than 90% of bacilli within 7 days, acting on metabolically active bacilli. It is also quite effective at preventing the emergence of drug resistance. Rifampicin is also a good bactericidal drug, with a potent sterilising effect and good ability to prevent the emergence of drug resistance. In addition to acting on rapidly dividing bacilli, it kills so-called ‘persisters’ which remain inactive for long periods but have intermittent periods of metabolism, with only a short drug exposure. This gives it its potent sterilising efficiency. Although bactericidal, pyrazinamide is used mainly for its sterilising ability and is particularly effective at killing intracellular mycobacteria sequestered inside macrophages in an acid environment.1 Streptomycin and ethambutol are less potent drugs, with ethambutol probably being bactericidal only at high concentration. They are less effective at preventing the emergence of resistance to rifampicin and isoniazid.

The efficacy of regimens based on rifampicin, isoniazid and pyrazinamide has been extensively studied in many adult controlled trials of different durations and dosing schedules. With this type of regimen, it is possible to convert over 90% of sputum smear-positive patients to culture-negative status at 2 months, to cure more than 95%, and to have a relapse rate of under 5%. The studies from which various aspects of treatment can be deduced, and from which later chemotherapy recommendations are based, are set out in table 1.3–14

Table 1

Six-month short course regimen trials for the treatment of pulmonary tuberculosis

A duration of 6 months seems to be the shortest period for a regimen based on isoniazid and rifampicin that will give an acceptably low relapse rate. If the duration is reduced below 6 months, higher relapse rates and lower cure rates are found, which are not acceptable for developed countries. Studies in Singapore and Hong Kong comparing 2, 4 and 6 months of pyrazinamide with rifampicin and isoniazid for 6 months showed that pyrazinamide is needed only for the initial 2 months.13 Conversely, however, if pyrazinamide is not used or cannot be tolerated, then a 9-month regimen of rifampicin and isoniazid is required, supplemented by ethambutol for the initial 2 months.

The trials described in table 1 covered a spectrum of dosing schedules from daily treatment throughout, a daily initial phase followed by a twice or thrice weekly continuation phase, or twice or thrice weekly dosing throughout. All of these schedules gave relapse rates of under 5% during periods of follow-up of between 6 and 30 months after cessation of treatment. The dosing schedule used therefore depends on a balance of cost, side effects and tolerance, drug availability and organisational aspects.

When rifampicin was omitted from the regimen in the continuation phase because of cost constraints, a 6-month total duration gave higher relapse rates and inadequate sterilisation. Additionally, if rifampicin is not used in the continuation phase, then there is a significantly increased failure rate if the organism is found to have initial isoniazid resistance.

Regimens of various durations of between 2 and 4 months have been studied in sputum smear-negative TB. For those with positive initial cultures, relapse rates of 32% with 2 months’ treatment,15 7–13% with 3 months’ treatment15 ,16 and 4% with 4 months’ treatment have been reported.15 ,16 The results of the 4-month regimen varied with the initial sensitivities of those with a positive culture. For initially sensitive organisms, the relapse rate was 2% but rose to 8% for those with initial resistance to isoniazid and/or streptomycin.16

Recommended regimens

The recommendations made by various expert bodies are founded on the substantial body of evidence for pulmonary TB, and take into account the likelihood of drug resistance in the target population. The recommendations of the National Institute for Health and Clinical Excellence are set out in table 2.17 Where there are variations between these recommendations and those of the European Respiratory Society,18 the American Thoracic Society19 and the WHO,20 these are also shown in the table.

Table 2

Recommendations from various national and international organisations

In the UK, a 6-month regimen comprising isoniazid, rifampicin, pyrazinamide and ethambutol for the initial 2 months followed by rifampicin and isoniazid for a further 4 months is recommended for adult respiratory TB, including isolated pleural effusion and mediastinal lymphadenopathy, irrespective of the bacteriological status of the sputum.17 A 6-month regimen is therefore recommended for sputum smear-negative disease. Where a positive culture for Mycobacterium tuberculosis has been obtained but susceptibility results are outstanding at 2 months, the initial four-drug phase should be continued until full susceptibility is known, but the total duration of treatment does not need to be extended. Ethambutol is included to cover the possibility of initial isoniazid resistance because of its increasing prevalence.21

If a patient cannot tolerate pyrazinamide, then the treatment required to give a suitably high cure rate is 9 months of rifampicin and isoniazid supplemented by 2 months of initial ethambutol.22 Routine daily pyridoxine is not required but should be given to patients at higher risk of peripheral neuropathy, that is, those with diabetes mellitus, chronic renal failure, alcohol dependency or malnourishment and those who are HIV-positive.17

While there is no realistic prospect of clinical trials to assess these regimens in children, the dosages of the drugs and their pharmacokinetics have been the subject of considerable debate over the last 20 years. In the 1998 Joint Tuberculosis Committee Guidelines,23 a dosage of 5 mg/kg for isoniazid was recommended based on data showing that this dose reached serum levels 60–100 times the minimal inhibitory concentration, and produced satisfactory clinical outcomes.24–26 However, more recent data27 and a re-appraisal of older data28 suggest that higher doses are needed, particularly in younger children and in fast acetylators.29 The WHO has changed its 2003 recommendations30 of 5 mg/kg to 10–15 mg/kg, again with a maximum of 300 mg,20 a move also reflected by a similar recent change in the British National Formulary (BNF) advice of 10 mg/kg dosing. The WHO has now separated recommendations for children from those for adults, which were also revised in 2009.31

A systematic analysis of the safety of ethambutol in children32 confirms its safety, and so on any risk/benefit calculation it should be included in the regimen for active disease, unless or until the presumed source case is shown to be fully drug susceptible. The drug resistance rate in children is the same as that of the adult groups21 who have infected them, and is not significantly different from their population subgroup, so ethambutol should be included until shown otherwise. Visual acuity (Snellen chart) should be checked before using ethambutol where possible, but this is clearly not practicable in small children. However, the safety profile is such that it can be used without such checking unless there is definite renal impairment. The current recommended dose is 15 mg/kg, but some authors recommend 20 mg/kg.33

The practical effects of the dosage changes are that care needs to be taken in dosage calculation. With fixed dose tablets additional isoniazid may now be needed, for example, Rifinah 100/150 two tablets prescribed for a 30 kg child gives the recommended 300 mg (10 mg/kg) of rifampicin but only 200 mg of isoniazid. An additional prescription of isoniazid 100 mg once daily is needed to provide 10 mg/kg isoniazid. Syrups are available for all four first-line drugs, but those for pyrazinamide and ethambutol sometimes have a short shelf life and have to be freshly made up. Crushed tablets, with appropriate dosage adjustments (see above), may be an alternative and an aid to adherence. The once daily treatment regimen is also an aid to adherence, even with multiple medications. The quoted cure rates in adult studies depend on adherence with treatment, and all cases of active TB, including children, should have an assessment of their adherence.17

Treatment of latent TB infection (LTBI)

Some people with asymptomatic TB infection, as judged by a positive tuberculin skin reaction or confirmed by an interferon gamma release assay (IGRA), are thought to have a small number of dormant bacilli, perhaps in the order of 103 or 104. The administration of one or two drugs for a shorter period of time than for disease—chemoprophylaxis or preventive therapy—is likely to kill these organisms and hence reduce the chance of progression to disease. It is therefore very important to differentiate between active TB disease and TB infection.

  • In TB infection, the tuberculin skin test (and IGRA test if available) is positive, the chest radiography is normal, and the patient is asymptomatic and has no clinical findings.

  • In TB disease, the skin test is usually positive and there are clinical signs and symptoms or radiographic changes present (note that NEITHER commercially available IGRA test is currently licensed for use in diagnosis).

This matter is further complicated in HIV-positive individuals, where anergy to tuberculin does not necessarily mean TB infection has not occurred, and it becomes difficult to differentiate between disease and infection. Most LTBI treatment is secondary (ie, after infection, as judged by a positive tuberculin test or IGRA, has occurred). Occasionally it may be given before evidence of infection, to prevent infection occurring (eg, in the neonate of a sputum smear-positive mother); this is on a primary basis.17 ,23

Although multiple-drug regimens are required for the treatment of disease in order to prevent the emergence of drug resistance, it is not illogical to use a one- or two-drug regimen for prophylaxis. Individuals who are receiving chemoprophylaxis are thought to have a low organism burden, in the order of 103 to 104. The chance of spontaneous resistance developing to a single drug is in the order of 10−5 to 10−7. The chance of drug resistance developing is therefore very low, particularly if a two-drug regimen is used. For instance, when mass isoniazid chemoprophylaxis was given to Inuit communities, there was no increase in isoniazid-resistant disease.34

Many randomised placebo-controlled trials of isoniazid chemoprophylaxis have been conducted.35 However, a recent meta-analysis for the Cochrane Review of 77 000 adults in 11 different placebo controlled studies, clearly show no additional protective efficacy with isoniazid durations of over 6 months, only increased toxicity.36

As with chemotherapy for disease, compliance with therapy is important, particularly because individuals on chemoprophylaxis are clinically well and so have less incentive to complete medication: the Alaskan studies also showed that to obtain the full benefit from chemoprophylaxis more than 60% of treatment must be taken.37 Isoniazid hepatotoxicity is not a significant issue in children, unlike in adult preventive treatment, where a rising incidence with age is clear.38

Concerns about the duration of isoniazid treatment have led to a search for alternative and shorter regimens. Rifampicin has been shown to have a greater sterilising effect in animal studies,39 either alone or with pyrazinamide, than has isoniazid. A placebo-controlled prophylaxis study from Hong Kong in patients with silicosis suggested that rifampicin alone for 3 months had similar efficacy to rifampicin and isoniazid for 3 months or isoniazid alone for 6 months.40 In UK studies, the use of rifampicin and isoniazid combined for 3 months showed a significant decrease in childhood TB,41 ,42 an effect which is maintained out to 12 years of follow-up.43

In the USA, either rifampicin for 4 months or isoniazid for 9 months are now recommended,44 despite the meta-analysis on isoniazid preventive therapy showing no additional benefit of regimens of longer than 6 months of isoniazid, only increased toxicity,36 and there being no controlled trial data to support 4 months of rifampicin.

In the UK, either 6 months of isoniazid or 3 months of rifampicin and isoniazid are recommended,17 both having ‘A’ category evidence to support their use. There is also evidence to support 6 months of rifampicin for contacts of isoniazid-resistant disease,45 ,46 and it is recommended for this category in the UK.17 ,23. The most recent development in the USA, but only for children aged 12 years or over, is the recommendation of rifapentine (a long-acting rifamycin) and high dose isoniazid for a 12-dose once weekly treatment for latent infection.47

The rise of drug resistance in TB, with now multi-drug resistant (MDR) and extensively drug-resistant (XDR) TB, has led to research on new uses for current drugs and new drug development, mainly in adults, with little experience of new drugs in paediatric populations, but these drugs may become necessary for selected cases.

Potential future developments

Third generation quinolones

A 6-month regimen of rifampicin and isoniazid, supplemented by an initial 2 months of pyrazinamide and ethambutol remains the ‘gold-standard’ treatment,17–20 with unacceptably high relapse rates, at least regarding developed country standards, if the regimen is shortened in duration,48 certainly for smear and culture-positive disease. Moxifloxacin is highly active against M tuberculosis with activity against M tuberculosis equivalent to that of isoniazid,49 with early bactericidal activity exceeding that of isoniazid between days 2 and 7 of therapy.50 These data and data in the murine model showing that substitution of moxifloxacin for isoniazid dramatically increases sterilising activity during a 4-month regimen,51 have led to extensive investigation of the potential role of it and other third generation quinolones in shortening treatment. Three phase 2B studies have now reported.

However, the 2B study reported in 2006 only showed equivalence to ethambutol in the sterilising ability of the regimen over 2 months,52 using the usual 2-month culture conversion method. A more recent analysis of the phase 2 OFLOTUB study data,53 however, shows that the two newer fluoroquinolones, moxifloxacin and gatifloxacin, may be able to shorten the duration of therapy to 4 months. This study also shows that serial sputum colony counts and the time to culture conversion have the potential to be quantitative predictive ‘endpoints’ in phase 2 trials of the newer anti-TB medications (see below).53 However, the most recent to report54 from Brazil also showed superiority in culture conversion rates for the regimen containing moxifloxacin.

Two phase 3 studies are underway to test the regimen shortening potential of moxifloxacin and gatifloxacin. These are actively recruiting in Africa, and are soon to be expanded to other sites. OFLOTUB55 is comparing the standard 6-month short course chemotherapy with a 4-month regimen of rifampicin, isoniazid and gatifloxacin, supplemented with 2 months of initial pyrazinamide. This study of 1836 patients completed treatment by summer 2009, and should shortly report the 1-year outcome. The REMox Study56 is comparing standard treatment with a 6-month regimen substituting moxifloxacin for ethambutol, and a further 4-month regimen of rifampicin, isoniazid and moxifloxacin, supplemented by 2 months of initial pyrazinamide. This is actively recruiting in multiple sites across the world and is nearing completion of recruitment.

New drugs in development

Novel agents are under development, two of which are in phase 2 development with a view to phase 3 studies. These are:

  • PA-824 (TB Alliance/Chiron), a nitroimidazopyran that inhibits the synthesis of proteins and cell wall lipids. This drug has a narrow spectrum of activity, no cross-resistance with current anti-TB drugs, and a unique mechanism of action.57 It has bactericidal activity against both replicating and static organisms, and has good sterilising activity.58 It is currently in phase 2 testing.59

  • TMC207 (Tibotec), a diarylquinoline (previously R207910) that selectively inhibits mycobacterial F1F0 ATP-synthetase. This is a new class of drugs, without cross-resistance with other classes of anti-TB drug. TMC207 has time-dependent bactericidal activity in vivo and in vitro, and substitution for any of the three main first-line drugs increases potency. It has a particularly impressive synergy with pyrazinamide.60 Although well absorbed orally, one potential drawback is it has a drug–drug interaction with rifampicin, which reduces the bioavailability of TMC207 by 50%.61

The history of the science of the regimens for treatment of TB in children has largely been led by adult studies, with modifications for pharmacokinetics in children as required. This is also likely to be the case for new drugs in development. Treatment of paediatric LTBI has more robust data in children, but newer regimens for example, with rifapentine, also have largely derived efficacy.



  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.