20th International AIDS Conference - Melbourne, Australia


MOAB0205LB - Oral Abstract Session

Performance of Xpert MTB/RIF testing for M.tuberculosis (TB) detection in HIV+ and HIV- pulmonary TB suspects in low versus high TB prevalence settings: the ACTG 5295/TBTC 34 study

Presented by Anne Luetkemeyer (United States).

A. Luetkemeyer1, C. Firnhaber2, M.A. Kendall3, X. Wu3, D. Benator4, G.H. Mazurek5, B. Metchock5, P. Johnson6, S. Swindells7, I. Sanne2, D.V. Havlir1, B. Grinsztejn8, D. Alland9, ACTG A5295/TBTC 34 Study Teams

1San Francisco General Hospital, University of California, San Francisco, HIV/AIDS, San Francisco, United States, 2University of Witswatersrand, Clinical HIV Research Unit, Johannesburg, South Africa, 3Harvard School of Public Health, Center for Biostatistics in AIDS Research, Boston, United States, 4Washington DC Veterans Affairs Medical Center, Infectious Diseases Clinic, Washington, United States, 5Centers for Disease Control and Prevention, Division of Tuberculosis Elimination, Atlanta, United States, 6Cepheid, Sunnyvale, United States, 7University of Nebraska Medical Center, Omaha, United States, 8Instituto de Pesquisa Clinica Evandro Chagas Fiocruz, Rio de Janiero, Brazil, 9New Jersey Medical School, Rutgers Biomedical and Health Sciences, Newark, United States

Background: Xpert MTB/RIF is a rapid test for identification of TB and rifampin (RIF) resistance with limited data from low TB prevalence settings and among HIV infected persons. Xpert performance was compared in low vs. high TB prevalence regions, HIV+ vs. HIV-, induced vs. expectorated sputum, and unprocessed vs. sedimented (digested, decontaminated, concentrated) sputum.
Methods: Pulmonary TB suspects from one low TB prevalence region (US: target ≥70%) and two high prevalence sites (Brazil/South Africa) had 1 unprocessed or sedimented sputum tested by Xpert (G4 cartridges) and compared to 2 sputa, each evaluated with AFB smear and mycobacterial culture on liquid and solid media. All had HIV testing. Fisher''s exact test compared sensitivity and specificity. RIF susceptibility was by agar proportions method.
Results: Required test results were available for 720 of 994 participants(median age 46; 63% male; 71% from US, 19% South Africa, 10% Brazil; 48% HIV­+, median CD4+ 157 (IQR 44, 369)).109(15%) had ≥1 culture positive for M.tuberculosis (TB+), of which 69 (63%) were AFB+ (≥1 positive AFB smear). Of TB+, 42 (39%) were HIV+. Sensitivity of Xpert was 100% among AFB+/TB+ regardless of HIV status, region, sputum processing, or induced/expectorated collection. Among AFB-/TB+, sensitivity was 61.5%. Sensitivity among AFB-/TB+ was higher for sedimented (69.7%, 23/33) vs. unprocessed sputum (16.7%, 1/6)(p= 0.02); but similar in HIV+ vs. HIV- (57.9% vs. 65.0%)(p=0.75), low vs. high prevalence regions (58.8% vs. 63.6%)(p= 0.99), and induced vs. expectorated sputum (63.6% vs. 56%)(p=0.73). Overall specificity was 98.9% (100% AFB+/TB+, 98.8% AFB-/TB+) and did not differ significantly by HIV status, sputum type, or TB prevalence (low 99.3% vs. high 97.4%)(p=0.08). Xpert was positive for 1 of 64 AFB- subjects from whom nontuberculosis mycobacteria was recovered (NTM+), and 0/3 AFB+/NTM+. Xpert detected 3/3 RIF-resistant specimens with one false positive RIF resistance (98.8% specificity, 81/82).
Conclusions: In a low TB prevalence setting, a single Xpert identified all AFB+/TB culture+ and 58.8% of AFB-/TB culture+, with 99.3% specificity. Performance was not significantly impacted by region, HIV status or sputum collection method. Xpert identifies TB with high specificity in low prevalence settings and may facilitate timely evaluation and treatment in HIV+ and HIV-.

Back to the Programme-at-a-Glance