Abstract
This study was performed to comparably assess two commercial real-time PCR assays for the identification of Trypanosoma cruzi DNA in serum. A total of 518 Colombian serum samples with high pre-test probability for infections with either T. cruzi or apathogenic Trypanosoma rangeli were assessed. The assessment comprised the NDO real-time PCR (TIB MOLBIOL, ref. no. 53-0755-96, referred to as the TibMolBiol assay in the following) with specificity for T. cruzi and the RealStar Chagas PCR Kit 1.0 (altona DIAGNOSTICS, order no. 611013, referred to as the RealStar assay in the following) targeting a kinetoplast sequence of both T. cruzi and T. rangeli without further discrimination. To discriminate between T. cruzi- and T. rangeli-specific real-time PCR amplicons, Sanger sequencing results were available for a minority of cases with discordant real-time PCR results, while the amplicons of the remaining discordant samples were subjected to nanopore sequencing. The study assessment indicated a proportion of 18.1% (n = 94) T. cruzi-positive samples next to 24 samples (4.6%) containing DNA of the phylogenetically related but apathogenic parasite T. rangeli. The observed diagnostic accuracy as expressed by sensitivity and specificity was 97.9% (92/94) and 99.3% (421/424) with the TibMolBiol assay and 96.8% (91/94) and 95.0% (403/424) with the RealStar assay, respectively. Reduced specificity resulted from cross-reaction with T. rangeli in all instances (3 cross-reactions with the TibMolBiol assay and 21 cross-reactions with the RealStar assay). DNA from the six discrete typing units (DTUs) of T. cruzi was successfully amplified by both real-time PCR assays. In summary, both assays showed a comparable diagnostic accuracy for the diagnosis of T. cruzi from human serum, with a slightly higher specificity seen for the TibMolBiol assay. The pronounced co-amplification of DNA from apathogenic T. rangeli according to the RealStar assay may be a disadvantage in areas of co-circulation with T. cruzi, while the test performance of the two compared assays will be quite similar in geographic settings where T. rangeli infections are unlikely.