RESUMO
Bolivia has the highest incidence of Chagas disease (CD) worldwide. Caused by the parasite Trypanasoma cruzi, CD is generally a chronic condition. Diagnosis is logistically and financially challenging, requiring at least two different laboratory-based serological tests. Many CD cases are missed; in Bolivia it is estimated just 6% of individuals chronically infected with T. cruzi get diagnosed. Achieving control on the way to elimination of CD requires a radical simplification of the current CD testing pathways, to overcome the barriers to accessing CD treatment. We aimed to generate unbiased performance data of lateral flow assays (LFAs) for T. cruzi infection in Bolivia, to evaluate their usefulness for improving T. cruzi diagnosis rates in a precise and efficient manner. This retrospective, laboratory-based, diagnostic evaluation study sought to estimate the sensitivity/specificity of 10 commercially available LFAs for T. cruzi, using the current CD diagnostic algorithm employed in Bolivia as the reference test method. All tests were blinded at the study site and performed by three operators. In total, 470 serum samples were tested, including 221 and 249 characterized as CD-positive/-negative, respectively. The LFAs were scored according to their relative importance using a decision-tree-based algorithm, with the mean decrease in Gini index as the scoring metric. The estimates of sensitivities ranged from 62.2-97.7% (95% confidence interval (CI) lower bound 55.0-94.7%); for specificities the range was 78.6-100% (95% CI lower bound 72.0-97.5%); 5/10 and 6/10 tests had sensitivity >90% and specificity >95%, respectively. Four LFAs showed high values of both sensitivity (93-95%) and specificity (97-99%). The agreement between 6 LFAs and the reference tests was almost perfect (Kappa 0.83-0.94). Most LFAs evaluated thus showed performances comparable with current laboratory-based diagnostic methods.