Challenges in the use of tests to diagnose tuberculosis infection

Dear Editors

Congratulations on the publication of the “Tuberculosis in NSW” edition (2013; 24(1)). It makes excellent reading on a disease that continues to smoulder in this country. However, there is one topic on which some authors appear unclear: the limitations of immunological tests for both tuberculosis (TB) disease and infection. I refer to the tuberculin skin test (TST) and interferon-gamma release assays (IGRAs), in particular the QuantiFERON-Gold In Tube^® (QFT-Gold IT). Certainly Britton et al are correct in stating the unreliability of such tests in infants under the age of 2. Indeed, these tests have little application in diagnosing disease, but are the only diagnostic agents we have for infection. It is not good enough for authors to tell us that the TST is negative or positive since this means nothing in an investigation where we have to balance sensitivity against specificity. A TST threshold of 10 mm induration might be said to achieve this balance, but the QFT-Gold IT shows that it does not. Across the world, tuberculins are produced that are of different potency, are recommended to be given in different doses (not always 10 units), and a “positive induration” may be less than 10 mm. Therefore authors should be encouraged to tell us what dose of which tuberculin has achieved what degree of induration. Although IGRAs are reported as positive and negative, this is dependant on an arbitrary cut-off point. The criterion we use for a “positive” TST in Australia may be sensitive, but has poor specificity and for QFT-Gold IT, mediocre sensitivity, if good specificity. I am surprised that none of the articles dealing with TB infection mention the use of both tests being used together (except in the BCG-vaccinated), a strategy that we in Australia can surely afford.

John E. Thompson

Editor, The Australian Tuberculosis Review