Comparing accuracy of lipoarabinomannan urine tests for diagnosis of pulmonary tuberculosis in children from four African countries: a cross-sectional study

Lancet Infect Dis. 2021 Mar;21(3):376-384. doi: 10.1016/S1473-3099(20)30598-3. Epub 2020 Dec 11.

Abstract

Background: A sensitive and specific non-sputum-based test would be groundbreaking for the diagnosis of childhood tuberculosis. We assessed side by side the diagnostic accuracy of the urine-based lipoarabinomannan assays Fujifilm SILVAMP TB LAM (FujiLAM) and Alere Determine TB LAM Ag (AlereLAM) for detection of childhood tuberculosis.

Methods: In this cross-sectional study, we tested urine samples from children younger than 15 years with presumed pulmonary tuberculosis. Children were consecutively recruited from four dedicated outpatient childhood tuberculosis clinics in The Gambia, Mali, Nigeria, and Tanzania. Biobanked urine samples were thawed and tested using FujiLAM and AlereLAM assays. We measured diagnostic performance against a microbiological reference standard (confirmed tuberculosis) and a composite reference standard (confirmed and unconfirmed tuberculosis). Sensitivity and specificity were estimated with bivariate random-effects meta-analyses.

Findings: Between July 1, 2017, and Dec 1, 2018, we obtained and stored urine samples from 415 children. 63 (15%) children had confirmed tuberculosis, 113 (27%) had unconfirmed tuberculosis, and 239 (58%) were unlikely to have tuberculosis. 61 children were HIV-positive (prevalence 15%). Using the microbiological reference standard, the sensitivity of FujiLAM was 64·9% (95% CI 43·7-85·2; positive in 40 of 63 confirmed samples) and the sensitivity of AlereLAM was 30·7% (8·6-61·6; 19 of 63). The specificity of FujiLAM was 83·8% (95% CI 76·5-89·4; negative in 297 of 352 unconfirmed and unlikely samples) and the specificity of AlereLAM was 87·8% (79·0-93·7; 312 of 352). Against the composite reference standard, both assays had decreased sensitivity; the sensitivity of FujiLAM was 32·9% (95% CI 24·6-41·9; positive in 58 of 176 confirmed and unconfirmed samples) and the sensitivity of AlereLAM was 20·2% (12·3-29·4; 36 of 176). The specificity of FujiLAM was 83·3% (95% CI 71·8-91·7; negative in 202 of 239 unlikely samples) and the specificity of AlereLAM was 90·0% (81·6-95·6; 216 of 239).

Interpretation: By comparison with AlereLAM, FujiLAM showed higher sensitivity and similar specificity. FujiLAM could potentially add value to the rapid diagnosis of tuberculosis in children.

Funding: German Federal Ministry of Education and Research, the Global Health Innovative Technology Fund, the UK Research and Innovation Global Challenges Research Fund, and the UK Medical Research Council.

Publication types

  • Comparative Study
  • Multicenter Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Child
  • Child, Preschool
  • Cross-Sectional Studies
  • Female
  • Gambia
  • HIV Seropositivity*
  • Humans
  • Infant
  • Lipopolysaccharides / urine*
  • Mali
  • Nigeria
  • Outpatients
  • Sensitivity and Specificity
  • Tanzania
  • Tuberculosis, Pulmonary / diagnosis*
  • Tuberculosis, Pulmonary / urine*

Substances

  • Lipopolysaccharides
  • lipoarabinomannan