Arch Dis Child Educ Pract Ed doi:10.1136/archdischild-2012-302876
  • Picket

Isoniazid prophylaxis started at 3–4 months of life does not prevent tuberculosis disease or infection in both HIV-infected and uninfected children

Setting: South Africa and Botswana

Patients: The study looked at infants who were born to HIV-infected women. To be included in the study, the infants had to be between 91 and 120 days of life, not infected with or exposed to a case of tuberculosis, without other chronic illness, or immunosuppressed from an infection other than HIV.

A total of 1351 children were included in the analysis and were split into two groups based on their HIV status after testing: HIV positive and HIV negative (table 1).

Intervention: The infants in both groups were randomly assigned to receive either isoniazid at a daily dose of 10–20 mg/kg or placebo. The study drug was initiated four days into randomisation, and continued for 96 weeks in total. All HIV-infected children continued to have access to antiretroviral therapy. Compliance was based on self-reporting.

Outcomes: The study protocol determined four tuberculosis infection or disease categories to include definite, probable, possible and latent tuberculosis. This was based on a combination of microbiological testing, radiological and clinical features and results of a tuberculin skin test.

The primary endpoint was tuberculosis disease and death in both groups, and also, latent tuberculosis in the HIV-uninfected group within 2 years after randomisation.

Secondary outcomes were the reduced risk of HIV disease progression or death in the HIV-infected group, and the improvement of disease-free survival in the uninfected group.

Follow-up period: Children were followed for 96–108 weeks. Three hundred and ninety-one children in the infected group, and 297 children in the uninfected group were not included as either lost to follow-up, or for some reason did not meet the set standards, and therefore, did not complete the study.

Table 1

Summary of first endpoint met toward primary outcome measures in children randomly assigned to isoniazid/placebo

Conclusion: Primary isoniazid prophylaxis did not improve tuberculosis disease-free survival among HIV-infected children, or tuberculosis infection-free survival among HIV-uninfected children immunised with the Bacillus Calmette-Guérin (BCG) vaccine.

Abstracted from: [CrossRef][Medline][Web of Science]Google Scholar


  1. Nick Brown2
  1. 1Department of Paediatrics, Queen Alexandra Hospital, Portsmouth, UK
  2. 2Department of Child Health, Salisbury District Hospital, Salisbury, UK
  1. Correspondence to Dr Mildred Adaku Iro, Department of Paediatrics, Queen Alexandra Hospital, Portsmouth PO6 3LY, UK; adakuiroh{at}

This is an unapologetically negative Picket. HIV is a huge global health issue, and tuberculosis coinfection is a major factor for morbidity and mortality.1 Despite a number of weaknesses, notably the looseness of the definition of tuberculosis and the high loss to follow-up, the results from this trial in infants with HIV infection are intriguing and compatible with adult trials.2 The children under study were tuberculosis-naïve at the outset as far as one can pragmatically tell, and gained no additional protection from isoniazid with which compliance was excellent.

The only other paediatric study in the literature (Zar et al)3 was also South African, and comparing efficacy is of interest. In Zar's study, isoniazid reduced all-cause mortality by 54% and tuberculosis infection by 72%, and as a result, the trial was stopped early. If only the children who were assigned isoniazid are compared by study (index study and the present one), the relative risk of tuberculosis (RR 0.40, 95% CI 0.18 to 0.92) was significantly lower in Zar's study. Mortality, however, was very similar (RR 1.06, 95% CI 0.64 to 1.75). The main difference in the populations is the much greater baseline/placebo mortality in Zar's study (16% vs 5.5%), so one inference might be that efficiency of the intervention, in terms of tuberculosis prevention is related to the underlying population all-cause mortality risk. This must be taken into consideration given the additional risk of isoniazid resistance in those starting treatment early.4

Perhaps different timing for starting isoniazid, or initiation based on laboratory results like CD4 count will improve efficacy. Perhaps other forms of chemoprophylaxis may prove more efficacious. In the meantime, of course, work on new candidate vaccines continues apace in the hope that there is a worthy successor to BCG on the near horizon.5


  • Competing interests None.

  • Provenance and peer review Not commissioned; internally peer reviewed.


Register for free content

Free sample
This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of ADC Education & Practice.
View free sample issue >>

Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.

Navigate This Article