OR05
Point-of-care lung ultrasound for the detection of pulmonary tuberculosis in Benin: the TrUST study
V Suttels(1) J D Du Toit(2) P A Wachinou(3) A R Hada(3) A A Fiogbe(3) B Guendehou(3) F Alovokpinhou(3) G Makpemikpa(3) M Rafiou(3) O Goudjanou(3) C Bessat(1) A Roux(1) E Garcia(6) T Brahier(1) J Vignoud(5) J Doenz(5) G Agodokpessi(3) D Affolabi(7) M A Hartley(4) N Boillat-Blanco(1)
1:Lausanne University Hospital, dpt of infectious diseases, Lausanne, Switzerland; 2:MRC Wits Rural Public Health and Health Transitions Research Unit, Acornhoek, South-Africa; 3:National Teaching Hospital for Tuberculosis and Respiratory Diseases (CNHU-PPC), Cotonou, Benin; 4:Yale School of Medicine, Yale Institute for Global Health, New Haven, CT 06519, USA; 5:Intelligent Global Health Research Group, Swiss Intitute of Technology (EPFL), 1015 Lausanne, Switzerland; 6:Lausanne University Hospital, dpt of emergency medicine, Lausanne, Switzerland; 7:Laboratoire Supranational de Référence des Mycobactéries (LRM), Cotonou, Bénin
Non-sputum triage tests are prioritised by the WHO to rule out tuberculosis (TB) and identify individuals that require further testing. We investigate the diagnostic performance of lung ultrasound (LUS) in a tertiary outpatient consultation in Benin. This 2-year prospective cohort included adult patients presenting with a lower respiratory tract infection according to the treating physician (October 2021- August 2023). Standardized LUS images and videos were collected. All images were reviewed according to pre-specified categories. TB was defined as a positive GenXpert MTB/RIF® or Xpert Ultra® on sputum. All patients were screened for HIV (Alere Determine® HIV-1/2 ). We evaluated the association of these ultrasound categories with TB using univariate logistic and multivariate logistic regression. Out of 504 patients included, 192 (38%) were TB positive (TB+) and 312 (62%) TB negative (TB-). Overall, 78 (15%) patients had documented HIV (median CD4 count/mm³ of 92 [IQR43-358]). TB+ was associated with consolidations larger than 1 cm in any quadrant (p<0.001, odds ratio [OR] 15 95%CI 9.7-23.8), apical consolidations larger than 1 cm (p<0.001, OR 19 95%CI 11.2-32.8) and apical subpleural lesions <1cm (p<0.001, OR 3.6 95%CI2.5-5.3). Multivariate logistic regression score showed an area under the receiver-operating curve of 0.88 to detect TB (sensitivity 90%, specificity 64%, negative predictive value 91% and positive predictive value 60%). LUS is a promising triage tool to exclude TB in the outpatient setting for symptomatic patients in a TB endemic region. Next steps are external validation and computer assisted diagnosis.