Trace Xpert Ultra Results Predict High Two-Year Tuberculosis Risk

Screening for tuberculosis often uses a rapid molecular test called Xpert Ultra, which sometimes returns a very low-level signal labeled a "trace" result. The meaning of that signal has been uncertain: extra tests after a trace result are often negative, but some people may still go on to develop culture-positive tuberculosis. To clarify the risk, Joowhan Sung and colleagues conducted a large community study in Uganda. They screened 31,505 people using sputum Xpert Ultra through event-based and door-to-door approaches. From that effort they enrolled 128 participants with trace-positive sputum (PWTS), 139 Ultra-negative controls into a prospective follow-up group, and 110 Ultra-positive (>trace) controls for a cross-sectional comparison. All enrolled participants had extensive initial evaluations, and those with trace-positive sputum who were not treated, along with the Ultra-negative controls, were followed and re-tested for up to 24 months. The team used two ways to define tuberculosis: a primary definition that included clinician judgment and a secondary definition based strictly on microbiological results, aiming to estimate how many people with trace results eventually developed tuberculosis over two years.

Among the 128 PWTS in the study, 79 (62%) were male and 19 (15%) were HIV-positive. At enrollment 45 (35%) of the PWTS were recommended for treatment right away; eight were lost to follow-up within three months. Seventy-five PWTS who were not treated at baseline were followed for a median of 706 days (interquartile range 344–714), and 19 of these were recommended for treatment during follow-up. The researchers found a substantial risk of tuberculosis among PWTS who were not treated at baseline: the cumulative hazard was 0.24 (95% confidence interval 0.15–0.40) at one year and 0.33 (0.21–0.54) at two years. By contrast, Ultra-negative controls had a cumulative hazard of 0.03 (0.01–0.10) at two years. Using the microbiological definition alone produced similar results (hazard 0.36 [0.22–0.58] at two years). An abnormal baseline chest X-ray was strongly associated with later tuberculosis (hazard ratio 14.6 [3.3–63.8]), whereas baseline symptoms were not predictive.

These findings show that a trace signal on Xpert Ultra is not harmless: many people with trace-positive sputum go on to develop tuberculosis within two years even when initial, more extensive tests do not confirm disease. The strong link with abnormal chest X-ray suggests that imaging can help identify which people with trace results are most likely to progress. Based strictly on these results, the authors conclude that treatment should be considered for most screening participants who have trace-positive sputum and abnormal chest imaging. For screening programs, this means that a trace result should prompt careful follow-up, consideration of early treatment, or both, rather than reassurance after a single set of negative tests. The data also support using clinician judgment alongside microbiological results when deciding whether to start treatment after a trace finding on Xpert Ultra.