SUMMARYOBJECTIVE: To evaluate the diagnostic performance of Xpert® MTB/RIF on stool samples from children with clinical suspicion of pulmonary tuberculosis (PTB) at primary care clinics.DESIGN: A cross-sectional diagnostic evaluation enrolling 5–16 year olds from whom one induced sputum (IS) sample was tested for microbiological TB confirmation. Results of a single stool sample tested using Xpert were compared against microbiologically confirmed TB, defined as a positive result on sputum microscopy and/or culture and/or IS Xpert.RESULTS: Of 222 children enrolled, 218 had complete microbiological results. The median age was 10.6 years (interquartile range 8–13). TB was microbiologically confirmed in 19/218 (8.7%) children. Of these, respectively 5 (26%), 9 (47%) and 15 (79%) were smear-, culture- and IS Xpert-positive. Stool Xpert was positive in 13/19 (68%) microbiologically confirmed cases and 4/199 (2%) microbiologically negative cases. Stool Xpert detected 76.9% (10/13) of human immunodeficiency virus (HIV) infected and 50% (3/6) of non-HIV-infected children with microbiologically confirmed TB (P = 0.241).CONCLUSION: Stool Xpert is a potential alternative screening test for children with suspected TB if sputum is unavailable. Strategies to optimise the diagnostic yield of stool Xpert assay need further study.

Background:Chronic immune activation is a feature of HIV infection associated with accelerated HIV disease progression. There is conflicting data on the association of biomarkers of immune activation with traditional markers of HIV disease progression; CD4 counts and viral load (VL).Objective:The study aimed to determine the association of biomarkers of immune activation; interferon (IFN)-γ-induced protein 10 (IP-10) and soluble cluster of differentiation 14 (sCD14) in chronic HIV infection with traditional markers of HIV disease progression.Methods:We collected demographic data, enumerated CD4 counts and quantified VL in 183 antiretroviral therapy (ART)-naive adults with chronic HIV infection. Plasma concentrations of IP-10 and sCD14 were quantified in the ART-naive adults with chronic HIV infection and 75 HIV-uninfected controls.Results:IP-10 concentrations were significantly higher in the HIV-infected group (median; 257.40pg/ml, IQR; 174.08-376.32) than in the HIV-uninfected (median; 86.19pg/ml, IQR; 67.70-116.39) (P<0.001). Similarly, sCD14 concentrations were significantly higher in the HIV-infected (median; 1.45µg/ml, IQR; 1.02-2.16) group than in the controls (median; 0.89µ/ml, IQR; 0.74-1.18) (P<0.001). High log10 IP-10 concentrations were positively correlated with high log10 viral loads (Spearman’s correlation coefficient [R]=0.21, P=0.003) and inversely correlated with low CD4 counts (R= -0.19, P=0.011). In contrast, log10 sCD14 was not significantly associated with either log10 viral loads (R=0.03, P=0.707) nor CD4 count (R=-0.04, P=0.568).Conclusion:We conclude that plasma sCD14 and IP-10 were elevated in the HIV-infected patients compared to HIV-uninfected individuals possibly due to on-going immune activation. In addition, plasma high concentrations of IP-10 but not sCD14 concentrations are associated with high VL and low CD4 count.