Results of a South African study in this week’s issue of The LANCET show how a diagnosis of tuberculosis in young children can be confirmed by a straightforward sputum test, rather than the conventional and invasive procedure of gastric lavage.
Diagnosis of tuberculosis is difficult in infants and young children and can be complicated by HIV infection. Heather Zar (School of Child and Adolescent Health, University of Cape Town, South Africa) and colleagues studied 250 children aged 1 month to 5 years who were admitted for suspected pulmonary tuberculosis in Cape Town. Sputum induction and gastric lavage were done on three consecutive days according to a standard procedure.
Samples from induced sputum and gastric lavage were positive in 87% and 65% of children, respectively. The yield from one sample from induced sputum was similar to that from three gastric lavages. In addition, almost half of all culture positive sputum samples were also smear positive, enabling rapid diagnosis and initiation of treatment. There was no difference in the reliability of diagnosis between HIV-positive and HIV-negative children. Sputum induction was useful even in young infants, with almost 40% of children with a positive sputum culture being less than one year of age.
Professor Zar comments: “In children with suspected pulmonary tuberculosis, sputum induction, not gastric lavage, should be the standard technique for microbiological diagnosis. One sample is sufficient, but if resources allow and if the child is in hospital, two or three specimens can increase microbiological yield . . . The important clinical usefulness of sputum induction for diagnosis of tuberculosis in this study raises possibilities for its use in primary care, and for diagnosis of other respiratory diseases in infants and young children.”
In a Research Letter (p 150), Daniel Vargus and colleagues present a preliminary comparison of the string test—where the removal of swallowed string from the upper gastrointestinal tract induces sputum— with conventional sputum induction for the diagnosis of tuberculosis in 228 HIV-infected patients. These patients were under investigation for tuberculosis and were either unable to produce an adequate specimen or they had a previous negative specimen. 52 HIV-positive controls had the same procedure. The use of the string test followed by sputum induction detected more cases of tuberculosis on culture of the specimens than did sputum induction alone, including the diagnosis of one patient with tuberculosis in a group of asymptomatic controls.