Preventing and treating TB in children – more baby steps

Polly Clayden, HIV i-Base

Children with tuberculosis (TB) are usually not infectious, so they are rarely considered to be a public health priority.

Where children’s needs are not neglected, prevention and treatment practice is mostly guided by findings extrapolated from adult research, so might not always be appropriate.

Pharmacokinetics (PK) of all drugs can vary hugely between babies, children and adults because of physiological differences, immaturity of enzyme systems and other mechanisms involved in drug metabolism. There is also great variability across age groups.

For example, drugs given to pre- and full-term neonates tend to have considerably longer half-life than that of adults. Between two and six months of age this difference disappears. Conversely, after this period, half-life in children can be shorter than in for specific drugs and pathways.

Children with TB have a wider spectrum of disease than adults. Although a larger proportion of young children below three years old have disseminated TB, primary childhood TB is typically more benign than that of adults. Adult forms of TB only emerge during adolescence. Like adults, children can be infected with or develop multi-drug resistant (MDR)TB.

As with HIV, young children who are unable to swallow tablets need child friendly formulations. Ideally – and where there are sufficient data to guide dosing – these should be in solid fixed dose combination (FDC) forms that are dispersible in liquids and can facilitate dosing across different weight bands.

Currently the only fixed dose formulations available for first-line TB treatment in children use old doses and there are scant data for second-line TB drugs. There is also data very little to guide use of even first-line ones in neonates and infants with low birth weight.

Several abstract and symposium sessions at the 44th Union Conference on Lung Health, and ongoing research first presented at this meeting last year, showed glimmers of progress in prevention and treatment of TB in children.

New first-line formulations

The doses of first-line TB drugs in children have mostly been scaled down from those of adults, using milligram per kilogram (mg/kg) ratios. The current first-line treatment for drug-sensitive TB is a combination of isoniazid, rifampicin, pyrazinamide, and ethambutol.

In 2010 the WHO revised its dosing recommendations for first-line drugs for children, after several PK studies found suboptimal levels with previously recommended doses. [1] See Table 1. Currently there are no quality-assured medicines available in the formulations recommended by WHO to treat children with TB. This means that children are treated with far-from-simple regimens of FDCs formulated with previous ratios, and a mix of divided and single tablets to make up the dosing shortfall.

Table 1. WHO-suggested simplified table for FDCs for TB treatment in children <25kg
Weight bands Intensive phase Continuation phase
75R /50H /150Z 100 E 75R /50H
4.0 – 7.9 kg 1 tablet 1 tablet 1 tablet
8.0 – 11.9 kg 2 2 2
12.0 – 15.9 kg 3 3 3
16.0 – 24.0 kg 4 4 4

Note: Children weighing 4.0-7.9, 8.0-11.9, 12.0-15.9 and 16.0-24.0kg should receive daily doses of 75, 150, 225 and 300mg of rifampicin (R); 150, 300, 450 and 600mg of pyrazinamide (Z); and 50, 100, 150 and 200 mg of isoniazid (H), respectively.

New formulations for first-line treatment are an urgent priority, so news during the conference that TB Alliance – a not for profit development partnership – has entered into collaboration with Svizera Europe to develop and distribute new dispersible FDCs with the recommended doses is welcome. [2] The FDC tablets produced with the old doses by this company are approximately 30% smaller than those from other brands and this is important for children.

A proposed trial – SHINE (SHorter treatment for mINimal TB in children) – sponsored by the UK Medical Research Council (MRC), in collaboration with South African, Zambian, Ugandan and Indian researchers, will use these new FDCs. [3] The principal research question for this trial is whether the standard 6-month regimen can be reduced with similar efficacy to 4 months in HIV positive and negative children with minimal TB, using revised dosing guidelines.

Other key questions are: Do the new doses, given as new FDCs and prescribed according to weight bands, result in appropriate drug exposures compared to historical paediatric and adult PK data and to dosing with single products? For children coinfected with HIV, can currently recommended adjusted strategies and doses of antiretrovirals appropriately overcome the effect of rifampicin at higher doses?

The trial is planned to start in 2014.

PK for second-line dosing

There is virtually no data on second-line TB drug dosing in children. Child friendly formulations are not usually available and the doses using divided and/or crushed tablets are uncertain. PK data on which to base optimal dosing is lacking. Also, second-line drugs are more toxic than those used in first-line treatment and adverse events are hard to monitor in children. Second-line TB drugs are also frequently used with antiretrovirals in coinfected children.

At the Union meeting last year Annekke Hessling from the Department of Paediatrics and Child Health, University of Stellenbosch, Cape Town, described a large ongoing study to characterise PK and toxicity of second-line TB drugs for treatment and prevention of drug resistant TB in HIV positive and negative children, by age and HIV status. [4] She presented preliminary data for three second-line TB drugs: ethionamide, amikacin and ofloxacin.

This ambitious study will be running over the next five years with an enrollment target of about 300 children. Age matched HIV positive children not on TB treatment will be enrolled as controls (42 receiving efavirenz and 22 lopinavir/r). The drugs under evaluation are: ethionamide, terizidone, ofloxacin, levofloxacin, moxifloxacin, amikacin, high dose isoniazid (INH), PAS, linezolid and capreomycin (the group will be looking at delamanid and bedaquiline and novel TB drugs as they become available for children).

The study includes intensive PK sampling, clinical follow up until treatment completion for children with active TB, cross sectional PK data from children receiving prophylaxis and toxicity monitoring.

Dr Hessling presented data from HIV positive and negative children, receiving routine treatment or prophylaxis for MDR TB, from December 2011 to September 2011. Children with severe anaemia (Hb <8g/dL) and/or weighing <5 kg were excluded.

Directly observed, exact doses were administered using the upper limit of the recommended doses following a standard breakfast: ethionamide 20 mg/kg (recommended dose 15-20 mg/kg/day), amikacin 20 mg/kg (15-22.5 or 30 mg/kg/day) and ofloxacin (15-20 mg/kg/day). Intensive sampling was performed at 0, 1, 2, 4, 6 and 8 hours post dose and C-max, T-max, AUC0-8 and t1/2 were compared to adult targets.

Seventy children (46 with TB disease and 24 receiving prophylaxis) were in the study group. Respectively, 12, 15 and 19 children in the disease group were age <2, 2-5 and 6-15 years. Only children <5 years are given prophylaxis for TB routinely, of these 6 were <2 and 18 were 2-5 years old. Overall, 12 (26.7%) children in the TB disease group were HIV positive and receiving ART. About 70% of the children with TB disease had pulmonary TB and the remainder had extra pulmonary TB or both.

The PK evaluation for ethionamide, revealed HIV negative children with higher C-max in the 0-2 years age group than the other two groups (median 7.66 vs approximately 5 ug/mL) but this was not significant; T-max peaked sooner and achieved higher target levels earlier (mean 1.80 hours vs 3.15 in the oldest age group, p=0.001), although overall exposure (AUC) was similar across age groups. HIV positive children had lower levels than HIV negative ones (median 4.86 vs 6.37 ug.h/mL, p=0.051). Dr Hessling noted that a larger sample size would probably show lower AUC as well. She described the finding that younger children peaked higher and earlier as “quite surprising” as the only other study that has looked at ethionamide PK in children by age group showed the opposite, she suggested that this might be due to crushing the tablets. The lower levels seen with HIV positive children compared to negative is consistent with previous observations. Although adult targets are unclear, the MIC achieved in children was similar or above that of adults.

For amakacin, Cmax was lower in the youngest group than the other two (median 43.65 vs approximately 49 μg/mL), T-max was lower (mean 1.00 vs approximately 1.13 hours) and AUC lower (median 103.85 vs 159.25 ug.h/mL in the oldest group, p=0.016). Levels did not differ by HIV status. At a dose of 20 mg/kg per day all children exceeded the adult target (Cmax 35-40 ug/mL). Dr Hessling suggested that perhaps 15 mg/kg, less frequent dosing and TDM should be evaluated particularly with relation to toxicities (amakacin can cause irreversible deafness). Interim data at a median of just over five months follow up showed hearing loss in 3/28 children, all with levels exceeding the adult target C-max. She also noted its low early bacterial activity, although it is given for MDR-TB, this compounded with its high toxicity, make it, “not such a wonderful drug”.

Fluoroquinolone antibiotics are potent anti-TB medications. They are the most important drugs in current treatment regimens for MDR TB and are likely to be key components of future regimens for the treatment and prophylaxis of drug-susceptible and DR-TB in adults and children.

Concerns about arthropathy in juvenile animal studies have historically limited the use of fluoroquinolones in children. Other potential toxicity concerns are QT interval prolongation and CNS toxicity. There are limited data on their PK in children with TB disease and latent TB infection. There is particularly limited paediatric data on fluoroquinolone safety and tolerability for prolonged courses required for DR-TB treatment (especially about QT prolongation).

Levofloxacin MIC is approximately half of that of ofloxacin. Oral bioavailability is good for both drugs: 85-95% ofloxacin and 99% levofloxacin. Elimination is primarily unchanged in urine.

Because the newer fluoroquinalones have improved efficacy, South African guidelines recommend moxifloxicin in children above and levofloxacin in children below 8 years of age for treatment and prevention of MDR-TB.

Target concentrations for fluoroquinalones are: AUC0-24/MIC primary PD index (target 100), Cmax/MIC (target 8-10) and Cmax (target 8-12 ug/mL).

Dr Hessling reported that giving ofloxacin achieved higher Cmax in the youngest versus oldest groups (median 9.4 vs 7.16 ug/mL), higher and earlier mean peak in Tmax (1.42 vs 2.60 hours, p=0.39) and similar overall exposure. This drug is given routinely as prophylaxis for MDR-TB and levels were higher in this group but this might be an age effect as it is given only to younger children. There was no difference by HIV status and adult targets were achieved.

At this years Union meeting, Anthony Garcia-Prats from the same research group showed more data from the study for ofloxacin in a symposium session: MDR TB in children and adolescent issues. [4]

This analysis was in 85 children, of which 55 had MDR-TB disease and 30 were receiving MDR prophylaxis. In the disease group there were approximately 30% of children in each of the 0-2 and 2-5 year old age groups and 40% in the 6-15 years group. Almost 75% of children in the prophylaxis group were 2-5 years and the remainder 0-2. Almost 75% of children in the disease group had pulmonary TB and the rest had extra pulmonary or both; 20% of the TB disease group had HIV. There were a substantial amount of underweight and stunted children in both groups but very little acute malnourishment.

The youngest age groups had a higher mean C-max of 10.4 ug/mL and this occurred earlier at 1.3 hours, these values were 8.5 and 8.1 ug/mL and 1.8 and 2.5 hours in the 2-5 and 5-16 age groups respectively, p<0.001. There were no differences in AUC by age. The median t1/2 was shorter in the youngest age group, 3.2 ug.h/mL vs 3.5 in children 5 years and older, p=0.001.

There were no differences in PK parameters by HIV status or weight for age Z-score (WAZ). There was a trend towards higher Cmax in children who received crushed tablets; it also occurred slightly earlier and this was significant, p=0.03.

The Cmaxs reported in this study compared favourably to those in adults receiving an 800 mg dose, but elimination was twice as rapid and AUC less than half.

The investigators used these PK data to generate pharmacodynamic indices with assumed ofloxacin MIC of 1.0 and 2.0. With an MIC of 1.0 the Cmax/MIC would be within the proposed targets for most children but below with an MIC of 2.0; AUC/MIC was far below proposed targets. See Table 2.

Table 2. Estimated pharmocodynamic parameters for ofloxacin
Oflox MIC 1.0 Oflox MIC 2.0 Proposed targets
Mean Cmax/MIC (SD) 8.97
Mean AUC/MIC (SD) 44.17

Dr Garcia-Prats noted that despite low AUCs, 92% of children had successful outcomes with MDR-TB treatment in their cohort with this dose of ofloxacin.

Safety and tolerability analyses from 44 children in the disease group receiving 6 to 8 drug regimens with a median time of 7.2 months follow up revealed no ofloxacin-related grade 3 or 4 events, or discontinuations. There was no arthralgia or arthritis.

Stephanie Thee from the Stellenbosch group showed PK data from a cross over study of ofloxacin and levofloxacin in an oral abstract session: Novel concept in the diagnosis and treatment of tuberculosis and children. [5]

The study aims were to investigate the PK and characterise the cardiac effects of ofloxacin and levofloxacin in HIV positive and negative children aged 0-8 years.

The study is nested in the ongoing larger MDR-PK study with the same inclusion/exclusion criteria as described above. It was a cross over design with intensive PK evaluation (T0, 1, 2, 4, 6, 8) on first flouroquinalone at steady state (2-8 weeks) followed by switch with intensive PK on second flouroquinalone at steady state (1-8 weeks).

Ofloxacin was dosed at 20 mg/kg and levofloxacin at 15 mg/kg, using crushed or whole tablets, given on an empty stomach.

Cardiotoxicity was assessed with 12-lead ECGs, baseline and during PK sampling at 3 hours post dose.

A total of 23 children were included in the analysis; 12 were being treated for MDR disease and 11 were receiving prophylaxis for MDR prevention. A third of the children being treated were in the 0-2, 2-6 and >6 year old age groups and 5 children age 0-2 and 6 children aged 2-6 were receiving flouroquinalones for MDR prevention. Four children in the disease group were HIV positive and all were over 6 years old. Five in the disease group and one in the prevention group were malnourished (WAZ <-2).

Table 3. PK of ofloxacin and levofloxacin in 23 children
Parameter Ofloxacin Levofloxacin
Cmax (ug/mL) 9.67
Tmax (h) 1.61
kel (1/h) 0.22
t1/2 3.2
AUC0-8 (ug*h/mL) 43.34

Note: All parameters median and IQR except Tmax mean and SD.

Results available from 22 patients for ofloxacin kel and T1/2.

Table 4. Ofloxacin by clinical characteristics
Cmax (ug/mL) T1/2 (h) AUC0-8 (ug*h/mL)
Sub group n Median (IQR) p-value n Median (IQR) p-value n Median (IQR) p-value
MDR disease 11 10.20
10 3.13
11 46.53
MDR prevention 11 8.88
0.77 11 3.18
0.89 11 43.34
0-2 years 9 10.90
9 2.91
9 46.53
2-6 years 9 8.78
9 3.21
9 44.34
>6 years 4 7.69
0.02 3 3.44
0.29 4 39.01
HIV status
HIV positive 4 7.69 (6.21-9.39) 3 3.44
HIV negative 18 9.86
0.25 18 3.16
0.48 44.67
Weight for age Z-score (WAZ)
> -2 18 10.05
18 3.20
< -2 4 6.63
0.04 4 2.64
0.42 29.37
Table 5. Levofloxacin by clinical characteristics
Cmax (ug/mL) T1/2 (h) AUC0-8 (ug*h/mL)
Sub group n Median (IQR) p-value n Median (IQR) p-value n Median (IQR) p-value
MDR disease 11 7.00
11 3.24
11 32.50
MDR prevention 11 6.32
0.53 11 2.98
0.31 11 29.89
0-2 years 9 7.00
9 2.79
9 29.89
2-6 years 9 6.86
9 3.22
9 31.69
>6 years 4 4.98
0.40 4 3.37
0.18 4 27.49
HIV status
HIV positive 4 4.98
4 3.37
4 27.49
HIV negative 18 6.88
0.39 18 3.09
0.23 18 31.38
Weight for age Z-score (WAZ)
> -2 17 6.86
17 3.14
17 45.77
< -2 5 6.71
0.91 5 3.23
0.97 5 29.24
Table 6. Pharmacodynamic parameters using published MICs of M.tb oxofloxacin and levofloxacin in children (n=23)
Oxofloxacin 20mg/kg Levofloxacin 15 mg/kg p-value
MIC for M.tb/published (ug/mL) 2.0 1.0 1.0 0.5
Mean AUC0-inf/MIC (SD) (target 100) 23.13
Mean Cmax/MIC (SD) (target 8-10) 4.5

Dr Thee showed very detailed analyses of PK parameters by subgroups – see tables 3 to 6. For ofloxacin there was no difference between the disease and prevention groups. Cmax was higher in the older children and AUC similar. There was no difference by HIV status but malnourished children had a much lower Cmax and AUC than those who were not.

None of the comparisons for levofloxacin were statistically significant.

Following a dose of ofloxacin 20 mg/kg and levofloxacin 15 mg/kg drug levels in children are less than half those of adults receiving standard oral doses of 800 mg and 750 mg respectively. PD indices favoured levofloxacin over ofloxacin. In Dr Garcia Prats’ presentation he suggested that 15-20 mg levofloxacin might be more appropriate and the group are currently looking at this dose.

No QTc prolongation (defined as QTc >450 ms) was reported in this study; mean QTc was 361 and 369 ms for ofloxacin and levofloxacin respectively.

Dr Thee noted that more data are urgently needed on fluoroquinalones in children in combination with optimised background regimen as well as novel anti tuberculosis drugs. The large PK study of anti-TB drugs for treatment and prevention of MDR-TB is ongoing and will result in a very important data set.

Three drug MDR-TB prevention

In the MDR symposium, James Seddon from Imperial College, London presented data on effectiveness, tolerability and adherence to a 3-drug MDR preventive therapy regimen on behalf of the Stellenbosch group and investigators from the UK. [6]

Once again, there are limited data to guide the management of children exposed to MDR. This study looked at the tolerability and toxicity of a standard preventive therapy regimen, given to children exposed to infectious MDR-TB, and explored the risk factors for poor outcome.

It was a prospective cohort study conducted in the Western Cape. Children <5 years old, or HIV positive children <15, were recruited from May 2010 through April 2011 if they had been exposed to someone with ofloxacin-susceptible MDR-TB.

Children were given a preventive therapy regimen according to local guidance: ofloxacin, ethambutol and high-dose isoniazid for six months. Adherence and adverse events were evaluated; poor outcome was defined as incident TB or death from any cause. The children were followed for 1-2 years.

The study included 186 children with a median age of 34 months (IQR 14–47). Nine children (5.0%) out of 179 tested for HIV, were positive; 73/183 (40%) were TST positive.

The majority (75.8%) of children had good adherence. Only 6 (3.2%) children developed Grade 3 adverse events. Of these, 3 events were associated with inadvertent overdosing of ofloxacin. Very few side effects were associated with muscle, joint or bone pain. No children had their regimen changed or modified. One child (0.5%) died – thought to be sudden infant death syndrome – and 6 (3.2%) developed incident tuberculosis during 219 patient-years of observation.

The investigators observed 31.9 (95% CI 12.8-65.9) per 1000 patient years: poor outcome was associated with: young age <1 year, HIV-positive status, multiple source cases and poor adherence. See Table 7.

Table 7. Outcome of children on 3-drug MDR-TB prevention therapy
Sub-group Events Years of observation Incidence (95% CI) Rate ratio p-value
Age >12 months 2 175.5 11.4
0-12 months 5 43.5 114.9
Source cases Single 2 152.4 13.1
Multiple 5 56.4 88.6
HIV status Negative 5 201.5 24.8
Positive 2 7.6 263.8
Adherence Good 2 164.3 12.2
Poor 5 54.8 91.3

This is the largest observational study of children on 3-drug preventive therapy regimens – Dr Seddon noted that an RCT is needed. He concluded that the regimen was well tolerated and few children who were adherent to therapy developed TB or died. He added that the provision of preventive therapy to vulnerable children following exposure to MDR-TB should be considered.

A novel approach for the evaluation of new TB drugs in children

A novel approach for accelerating access to new TB drugs and regimens, in infants and young children, was presented at the Union meeting last year; this was discussed again in the MDR-TB symposium.

Researchers from the TB Alliance and the Stellenbosch group have been looking at ways to speed up access to new TB drugs and regimens in infants and young children. Stephen Murray presented the proposed framework for this evaluation.

It will soon be appropriate for the TB Alliance to begin trials in children for some of the drugs that they are currently studying in adults. Their approach is to enrol people who are sensitive to the drugs in the regimens being studied regardless of DS/MDR classification. The proposal would do the same in children.

TB trials with efficacy as the primary endpoint are not required for children as power would be prohibitive and at least similar efficacy to adults is assumed. Matching PK parameters to that in adults has proven to be safe and effective and regulators recognise that it is possible to extrapolate efficacy in children from adult data. But trials in children cannot begin until the adult dose has been established and safety and efficacy demonstrated in this population – the question is when is this?

The traditional approach to collecting PK and then safety data in children is sequentially in de-escalated age bands: 12-16, 6-12, 2-6 and 0-2 years. But experience in older children might not mitigate the risk in younger ones as differences are caused by changes in metabolism at different ages. Drugs to be used mainly by children are not developed in this way.

The TB Alliance plan proposes hospitalised TB patients in all age groups simultaneously receive single dose for initial PK (based on adult dose and modeling) on top of background therapy, which is a small and manageable risk. Next step would be 14 day multiple dose PK also in hospitalised TB patients.

This approach means that approval for the youngest children would not be delayed – 0-2 is a critical age for TB and has a huge and unmet need for treatments. It is important though that studying the older group is not delayed if paediatric formulations are not available for the younger ones.

This approach could provide faster information for registration and evidently that both the FDA and EMA have indicated that they are open to considering it.


Unless otherwise stated, references from the 44th Union Conference on Lung Health, 30 October – 3 November 2013, Paris.

See also the HTB report from last years Union meeting: Paediatric TB: glimpses of PK data and a potential new approach to drug development.

  1. World Health Organisation. Rapid advice: Treatment of tuberculosis in children. 2010. (PDF)
  2. TB Alliance press release. New collaboration will develop, deliver needed childhood TB medicines. Svizera Europe and TB Alliance commit to fulfil urgent need. 30 October. 2013.
  3. Personal communication from Diana Gibb. (November 2013).
  4. Hessling et al. Pharmacokinetics of second line TB therapy in children. Tuberculosis treatment considerations for neonates and infants. 43rd Union Conference on Lung Health. 13 – 17 November 2012. Kuala Lumpur, Malyasia. Symposium 19.
  5. Garcia Prats AJ et al. The pharmacokinetics and safety of fluoroquinalones for the treatment and prevention of drug-resistant tuberculosis in HIV-infected and -uninfected children. Symposium 10.
  6. Thee S et al. Pharmacokinetics of oxofloxacin and levofloxacin in children with multidrug-resistant tuberculosis. Oral abstract OP-212-02.
  7. Seddon JA et al. Effectiveness, tolerability and adherence to a 3-drug MDR preventative therapy regimen. Symposium 10.
  8. Mendel CM et al. A framework for the evaluation of new TB drugs in children. 43rd Union World Conference on Lung Health, 13-17 November 2012, Kuala Lumpur, Malaysia.
  9. Mendel CM et al. A framework for the evaluation of new TB drugs in children. Symposium 10.

Links to other websites are current at date of posting but not maintained.