Nanoparticle capture of urinary lipoarabinomannan for diagnosing childhood tuberculosis

Urinary assays detecting lipoarabinomannan (LAM) as a diagnostic test for tuberculosis (TB) have limited sensitivity in pediatric populations. We aimed to evaluate a urine processing step using the Ceres TB-Nanotrap to concentrate available LAM antigens and augment LAM detection in urine using the Alere lateral flow assay (LF-LAM). This case-control study recruited children with TB and non-TB controls aged 1-18 years from outpatient clinics. The LF-LAM test was performed before and after concentrating 5mL of urine with 400uL of magnetic TB-Nanotrap particles. Band intensity was measured via visual grading and digital quantification. Urine LAM sensitivity by visual grading was 4.5% (95% Confidence Interval [CI]: 0.3 - 18.5) pre-concentration and 50.0% (95% CI 30.0 - 70.0) post-concentration. Sensitivity by digital quantification was 0.0% (95% CI 0.0 - 15.4) pre-concentration and 63.9% (95% CI 42.8 - 81.4) post-concentration. Specificity was high with both methods (visual grading 90.0% [62.8 - 99.4] pre-concentration, 90.0% [62.8 - 99.4] post-concentration; digital analysis 80.0% [44.4 - 97.5], 90.0% [62.8 - 99.4] respectively). For cases, digital analysis showed an increase in median LF-LAM band intensity from 0.0 arbitrary units (AU) (IQR 0.00 - 32.5) in unconcentrated samples to 98.4 AU (IQR 34.9 - 212.2) in concentrated samples. Concentration with TB-Nanotrap greatly increased sensitivity of urine LAM detection without change in specificity. Digital quantitative analysis further increased sensitivity. Use of TB-Nanotrap and digital quantitative analysis of urine LAM improved the diagnostic accuracy of the LF-LAM assay in this pediatric population and should be validated in larger studies.