Prediction and prevention of tuberculosis in contacts - Authors' reply

Abstract

We thank Sandra Arend and Jonathan Uzorka for their interest in our study. They argue that because our risk score predicts tuberculosis independently of the contact tuberculin skin test (TST) results and index case smear grade, it is likely that subsequent re-exposure to tuberculosis was an important risk factor. Therefore, they debate the justification for our calculation of numbers needed to treat (NNT) with preventive therapy at the time of known exposure to prevent each tuberculosis case. Although we agree that re-exposure might have been an important, unmeasured risk factor in our population, the highest rate of tuberculosis among contacts was in the first year after known exposure, and nearly all the cases occurred during the first 3 years. Furthermore, prolonged exposure to the index case was an important risk factor. Taken together, these findings suggest that the known exposure was by far the most important factor in determining tuberculosis risk. However, we appreciate the potential importance of subsequent re-exposure and therefore calculated the NNT to prevent each tuberculosis case within 5 as well as 10 years. Importantly, we calculated NNT principally to illustrate practical differences between risk groups. Because the cutoffs used to define risk groups were arbitrary and our risk score is a continuous variable, these NNTs inherently change depending on how a high-risk contact is defined...