Abstract
Typhoid is an endemic in Fiji with increases observed since the early 2000s and frequent outbreaks reported. We assessed the diagnostic accuracy of currently available typhoid rapid diagnostic tests (RDTs) (TUBEX, Typhidot Rapid, and Test-It assay) to establish their performance against blood culture in Fiji and to examine their suitability for rapid typhoid outbreak identification. The performance of RDTs was assessed in the public health reference laboratory in Suva, Fiji, according to the manufacturers' instructions. A simulation was used to examine the potential use of RDTs for attribution of a febrile illness outbreak to typhoid. For the diagnostic evaluation, 179 patients were included; 49 had blood culture-confirmed typhoid, 76 had fever as a result of non-typhoid etiologies, and 54 were age-matched community controls. The median (interquartile range) age was 29 (20-46) years. Of the participants, 92 (51.4%) were male and 131 (73.2%) were indigenous Fijians. The sensitivities of the tests were 77.6% for TUBEX, 75.5% for Typhidot Rapid, and 57.1% for Test-It assay. The Test-It assay had the highest specificity of 93.4%, followed by Typhidot Rapid 85.5% and TUBEX 60.5%. Typhidot Rapid had the best performance in the simulation for attribution of a febrile illness outbreak to typhoid. Typhoid RDTs performed suboptimally for individual patient diagnosis due to low sensitivity and variable specificity. We demonstrate that RDTs could be useful in the field for rapid attribution of febrile illness outbreaks to typhoid. Typhidot Rapid had the best combination of sensitivity, specificity, positive and negative predictive values, cost, and ease of use for this purpose.