Detection of Mycobacterium tuberculosis in clinical specimens from children using a polymerase chain reaction

We evaluated the usefulness of the polymerase chain reaction (PCR) using the insertion sequence IS6110 as the target for DNA to detect Mycobacterium tuberculosis in clinical specimens from children.

This was a prospective, controlled, blinded study comparing PCR on clinical specimens, mycobacterial culture, and clinical diagnosis.

Sixty-five hospitalized children were evaluated, 35 with tuberculosis disease and 30 controls. Cases were defined by culture and/or specific clinical criteria. Controls included patients with tuberculosis infection but no detectable disease as well as patients free of tuberculosis infection and disease.

Polymerase chain reaction had a sensitivity of 40% and a specificity of 80% compared with clinical diagnosis. Mycobacterial culture had a sensitivity of 37%. The combination of culture and PCR identified 19 of 35 children (54%) with clinically diagnosed tuberculosis. There were six children with false-positive PCR results: One had tuberculosis infection without disease, two had Mycobacterium avium lymphadenitis, and three had diagnoses unrelated to tuberculosis.

The sensitivity of PCR is comparable to that of culture for detecting M tuberculosis in children, and may strengthen and hasten the clinical diagnosis in culture-negative patients. However, because of the limitations in specificity, the results of PCR alone are insufficient to diagnose tuberculosis in children. Although ongoing refinements in PCR techniques should improve the specificity of this test, epidemiologic and clinical information continue to be the most important consideration in the diagnosis of tuberculosis in culture-negative children.