Performance evaluation of Fujilam in diagnosis of tuberculosis in children under 5 years with presumptive tuberculosis in Uganda

Abstract

Background: Diagnosis of tuberculosis (TB) in children remains challenging due to the paucibacillary nature of the disease, non-specific symptoms of TB, and difficulty in collecting samples from the lower respiratory tract of young children who cannot produce sputum spontaneously. Recently, a next-generation LAM assay was developed: Fujifilm SILVAMP TBLAM test (FujiLAM). However, its performance in children remains to be investigated. Objectives: To assess the diagnostic accuracy of FujiLAM, compared to a microbiological reference standard and Abbott LAM in children under 5 years, with and without comorbidities. Methods: A cross-sectional study nested in the ongoing NOD-pedFEND Cohort study that enrolled 100 children under 5 years with presumptive TB. These children were either attending Baylor Uganda COE, admitted to the paediatric wards in Mulago Hospital, community contacts of adults with confirmed TB, or referred from KCCA clinics. Records were obtained from the NOD-pedFEND study database from August 2021 to August 2023 using a designed data abstraction tool. The novel investigational test under evaluation was the urine FujiLAM and its performance was compared to the microbiological reference standards. FujiLAM results were cross-tabulated against the Microbiological Reference Standard (MRS) using a 2×2 contingency table. Sensitivity and specificity were calculated. Cohen’s Kappa was used to assess the concordance of FujiLAM and other TB tests at 5% level of significance. The diagnostic yield of FujiLAM, and other TB tests was obtained as a percentage of the study population with a positive test. Analysis was done using STATA, College Station TX Version 16. Results: A total of 100 children were included in the study of which 80 had FujiLAM done. Median age was 12months (Inter quartile range: 12-24 months) and 52% were male, 32% had TB contact and 12 were HIV positive. MRS positive TB was found in 18% of the children. Of these none had a positive FujiLAM test resulting in a sensitivity of 0% (95% CI: 0 – 21.8). Out of the 65 children with MRS negative results, 64 had a negative FujiLAM test resulting in a specificity of 98.5% (95%: 91.7-100), Positive Predictive Value of 0% (95% CI: 0-97.5), and Negative Predictive Value of 81.0% (95% CI: 70.6-89). The diagnostic yield of FujiLAM was 1.3% and the concordance with other TB tests was poor (Cohen’s Kappa: -0.03- 0.0). Conclusions and Recommendations: FujiLAM might not serve as a practical standalone diagnostic test for screening or confirming TB in children under 5. However, it could prove beneficial within an integrated diagnostic approach as part of TB composite testing, given its high specificity.