Reply to Parker: Implications of Tuberculosis Sputum Culture Test Sensitivity on Accuracy of Other Diagnostic Modalities.

From the Authors:

We welcome the interest shown by Dr. Parker in our recent publication that described the yield and efficiency of novel intensified tuberculosis (TB) case-finding algorithms for people living with HIV (1). We agree that sputum mycobacterial culture has imperfect sensitivity for active pulmonary TB. Nonetheless, culture is the best available microbiological reference standard for evaluation of novel TB diagnostics. The addition of clinical follow-up can be helpful but also has limitations. For example, in the context of our study, initiation of antiretroviral therapy complicates assessment of whether clinical and radiological improvements are a result of TB treatment. Furthermore, in most TB-endemic areas, culture is not routinely available for TB diagnosis, and Xpert MTB/RIF (Xpert) is used as the confirmatory TB test. In such settings, all patients with a positive Xpert result are in fact regarded as a TB case and initiated on TB treatment.

If we consider all Xpert-positive patients to be true positives, as suggested by Dr. Parker and as occurs in routine clinical practice, our study conclusions would remain unchanged: 1) CRP (C-reactive protein)-based TB screening followed by Xpert confirmatory testing would identify a similar proportion of TB cases as the current intensified case-finding (ICF) algorithm (54% [95% confidence interval (CI), 45–63] vs. 58% [95% CI, 49–66]; difference in yield 4% [95% CI, −9% to +16%]; P = 0.57); 2) CRP-based ICF would use less than half as many Xpert assays (9 vs. 4); and 3) the addition of a single culture would substantially increase ICF yield, detecting ≥77% of all TB cases, regardless of the screening strategy used. Thus, we believe HIV programs should first consider implementation of CRP-based TB screening, which may then enable the routine use of more sensitive confirmatory tests, such as culture, to improve ICF yield.