Pooling sputum samples for the Xpert MTB/RIF Ultra assay: A sensitive and effective screening strategy

Abstract

The sensitivity of Xpert MTB/RIF (Xpert) pooled testing is limited for diagnosing patients with paucibacillary tuberculosis (TB). We assessed whether pooled testing with Xpert MTB/RIF Ultra (Ultra) can be a sensitive and effective approach for mass TB screening. Conserved, frozen sputum samples, previously confirmed as positive or negative for Mycobacterium tuberculosis by individual Xpert assays, were mixed in pools of 4, 8, and 16 and then tested using Ultra. Each pool contained a single positive sample with varying mycobacterial loads. We then simulated TB screening at prevalence ranges of 0.2–1.0 % and calculated the cartridges required per case detected at different pool sizes. The overall sensitivity of Ultra pooled testing was high (88.9 %, 75.9–96.3). Sensitivity was greater in pools in which the positive sample had a high mycobacterial load compared to those with scant bacilli. As prevalence increased, the optimal pool size and benefits of pooled testing declined, but a pool size of 8 resulted in at least 80 % cartridge savings with the highest simulated prevalence. Sputum pooling using Ultra could be a sensitive and effective strategy for TB screening. However, broad TB screening in communities with limited resources will require new, lower-cost, high-throughput screening tools, perhaps based on non-sputum specimens.