The diagnosis of childhood tuberculosis in an HIV-endemic setting and the use of induced sputum.

SETTING

A hospital in the capital city of Uganda.

OBJECTIVES

To determine the proportion of positive induced sputum cultures among probable childhood tuberculosis (TB) cases, to describe the laboratory and radiological features of childhood TB in the context of human immunodeficiency virus (HIV), and to determine the prevalence of HIV infection in these patients.

DESIGN

Cross-sectional descriptive study.

METHOD

Seven hundred and fifty children with suspected TB were evaluated clinically, by chest radiography and by tuberculin test (Mantoux). Probable cases underwent sputum induction or lymph node biopsy, HIV testing and RNA-PCR tests.

RESULTS AND CONCLUSION

The Mantoux test was positive in 55/121 (45%) of the patients; low sensitivity to Mantoux was associated with HIV infection (P = 0.000). Induced sputum yielded 12/101 (12%) positive smears of Mycobacterium tuberculosis and 30/101 (30%) positive cultures. Six of the eight lymph node biopsies were culture-positive, and the histology of seven samples was supportive of TB. Most of the children with probable and confirmed TB presented with similar signs, symptoms and chest radiography patterns, irrespective of HIV status. Elevated ESR had no significant role in the diagnosis of TB in children and even more so in the presence of HIV infection. The HIV infection rate was high among children with TB (49%).