Bayesian assessment of three serological tests for the screening of Trypanosoma cruzi in hunting dogs of Virginia and West Virginia, USA.

Abstract

Trypanosoma cruzi, the causative agent of Chagas disease (CD), is a vector-borne disease that affects several mammalian species, including humans. An endemic cycle in wild animals has been described, where dogs with an outdoor lifestyle are at higher risk. CD diagnosis is challenging, requiring multiple tests. Although there are several serologic tests available, they were originally developed for human diagnosis. Currently, there is limited knowledge about their performance in other species. The objective of this research was to estimate the performance of three serologic assays for CD screening in hunting dogs without clinical CD, to support large T. cruzi surveys. A total of 490 hunting dogs were sampled in VA and WV, USA. Sera were tested in parallel using an in-house ELISA test, an immunochromatographic rapid test (IRT), and an immunofluorescence antibody test (IFAT). Sensitivity (Se) and specificity (Sp) of these tests were estimated using a Bayesian latent class model (BLCM), assuming the absence of a gold standard test. Informative priors were obtained from literature (IRT and IFAT), or from Receiver Operating Characteristic (ROC) curves analysis (ELISA) using an independent panel of 24 canine sera, with known infection status. The BLCM for the ELISA assay estimated a median Se and Sp of 99.3 % (95 % Posterior Probability Interval (PPI): 74.8 %-100 %) and 78.8 % (95 % PPI: 75.2 %-82.1 %), respectively. Similarly, the IRT had medians Se and Sp of 99.1 % (95 % PPI: 83.8 %-100 %) and 80.2 % (95 % PPI: 76.5 %-83.6 %), respectively. On the other hand, the IFAT presented a median Se of 96.7 % (95 % PPI: 90.9 % - 99.3 %) and a median Sp of 98.5 % (95 % PPI: 97.2 % - 99.3 %). A median true prevalence of 1.4 % (95 % PPI: 0.6 %-2.6 %) was estimated. Results suggest that ELISA and IRT are adequate screening tests, given their high Se. However, due to their limited Sp they would require the use of a second highly specific test for CD confirmatory diagnosis.