Background: There are limited data on the prevalence of multi-drug resistant tuberculosis (MDR-TB), estimated at 0.6-6.7%, in African children with tuberculosis. We undertook a retrospective analysis of the prevalence of MDR-TB in children with Mycobacterium tuberculosis (MTB) at two hospitals in Johannesburg, South Africa.
Methods: Culture-confirmed cases of MTB in children under 14 years, attending two academic hospitals in Johannesburg, South Africa during 2008 were identified and hospital records of children diagnosed with drug-resistant TB were reviewed, including clinical and radiological outcomes at 6 and 12 months post-diagnosis. Culture of Mycobacterium tuberculosis complex (MTB) was performed using the automated liquid broth MGIT™ 960 method. Drug susceptibility testing (DST) was performed using the MGIT™ 960 method for both first and second-line anti-TB drugs.
Results: 1317 children were treated for tuberculosis in 2008 between the two hospitals where the study was conducted. Drug susceptibility testing was undertaken in 148 (72.5%) of the 204 children who had culture-confirmed tuberculosis. The prevalence of isoniazid-resistance was 14.2% (n = 21) (95%CI, 9.0-20.9%) and the prevalence of MDR-TB 8.8% (n = 13) (95%CI, 4.8-14.6%). The prevalence of HIV co-infection was 52.1% in children with drug susceptible-TB and 53.9% in children with MDR-TB. Ten (76.9%) of the 13 children with MDR-TB received appropriate treatment and four (30.8%) died at a median of 2.8 months (range 0.1-4.0 months) after the date of tuberculosis investigation.
Conclusions: There is a high prevalence of drug-resistant tuberculosis in children in Johannesburg in a setting with a high prevalence of HIV co-infection, although no association between HIV infection and MDR-TB was found in this study. Routine HIV and drug-susceptibility testing is warranted to optimize the management of childhood tuberculosis in settings such as ours.