Impact of HIV-1 co-infection on presentation and hospital-related mortality in children with culture proven pulmonary tuberculosis in Durban, South Africa.

BACKGROUND: Diagnosis of tuberculosis (TB) in childhood is difficult and is compounded by HIV-1, as both diseases often co-exist and have many similar features. Most studies from developing countries have included subjects in whom the diagnosis of TB is suspected but not proven. We therefore compare the findings in HIV-infected and non-HIV-infected children with culture-proven TB.
METHODS: Records were obtained from the laboratory at King Edward VIII Hospital, Durban, South Africa, between January 1998 and December 1999. Children aged 0-12 years with proven pulmonary tuberculosis (sputum, gastric washing or endotracheal aspirate culture for Mycobacterium tuberculosis) from the paediatric medical wards and intensive care unit were included in the study. A retrospective chart review of demographic data, clinical presentation, diagnostic modalities for TB, HIV-1 result, management and outcome were evaluated.
RESULTS: Of 138 culture-proven cases of TB identified during the study period, the medical records of 118 (86%) could be traced. Of these, 57 (48%) were HIV-1 infected, 44 (37%) non-HIV-1-infected, and in 17 (14%) HIV-1 status was not determined. In contrast to previous studies, this study has shown that TB-HIV co-infection in children is common (48% of all culture-proven cases), the presentation of tuberculosis may be acute (43%), and supportive tests are individually only reliable in confirming a diagnosis in a third of cases. All culture evaluations for M. tuberculosis were positive by 8 weeks. Where other diseases often co-exist with TB and HIV infection and the pressure for hospital in-patient admissions are excessive, the diagnosis of tuberculosis could easily be missed (21.2%). Clubbing and age over 2 years were the most reliable indicators of underlying HIV-1 disease in a child with tuberculosis, while clinical features, radiology and supportive tests were found to be similar between HIV-infected and non-HIV-infected TB cases. Hospital-related mortality, all causes, was higher (17.5%) in the HIV-1-infected than the non-infected group (11.4%).
CONCLUSION: The changing pattern of presentation of childhood tuberculosis and the high prevalence of TB in HIV endemic areas has made it imperative to maintain a high index of suspicion, with culture evaluation being an important part of clinical practice.