Something from nothing: Sensitivity and specificity of Xpert MTB/RIF Ultra on contaminated liquid cultures for tuberculosis and rifampicin-resistance detection

Background: Xpert MTB/RIF Ultra (Ultra) (Cepheid, Sunnyvale, USA) is a widely-used rapid front-line TB and rifampicin susceptibility test. Mycobacterium Growth Indicator Tube 960 (MGIT960) culture is still used as an adjunct for tuberculosis (TB) and drug susceptibility diagnosis but is vulnerable to contamination. Whether Ultra can be used on to-be-discarded contaminated cultures is uninvestigated. Methods: We stored contaminated MGIT960 tubes (growth-positive, acid-fast-bacilli-negative) inoculated to diagnose pulmonary TB in a routine high-volume laboratory in Cape Town, South Africa. Patients who had, at contamination-detection, no positive TB results (smear, Ultra, culture) and another specimen submitted three months post-contaminated specimen submission were selected. We evaluated the sensitivity and specificity of Ultra on contaminated growth from the first culture for 1) TB (next-available non-contaminated culture result reference standard), and 2) rifampicin resistance (vs. MTBDRplus on the later isolate). We calculated potential time-to-diagnosis improvements. We also evaluated MPT64 TBc (TBc). Findings: 2186 patients had a contaminated diagnostic culture. 49% (1068/2186) had no other specimen submitted, despite guidance to the contrary. After 319 ineligible patients were excluded, 799 with at least one repeat specimen submitted remained: 31% (n=246), 54% (n=429) and 16% (n=124) were repeat-specimen culture-positive, -negative, and -contaminated, respectively. When Ultra was done on the initial contaminated growth, sensitivity and specificity were 89% (95% CI 84-94) and 95% (90-98) for TB and 95% (75-100) and 98% (93-100) for rifampicin-resistance. If our approach were performed the day after initial contamination detection, time-to-TB-detection would improve a median (IQR) of 23 (13-45) days and, importantly, provide a result in many patients who had none. TBc had poor accuracy. Conclusion: Ultra on acid-fast-negative growth from contaminated MGIT960 tubes had high sensitivity and specificity; approximating World Health Organization-target product performance sputum test and exceeding drug susceptibility testing (DST) criteria. Our approach could mitigate contamination's negative effects, especially when repeat specimens are not submitted.