Find in Library

Diagnostic real-time PCR for the detection of <i>Cyclospora cayetanensis</i> in human stool samples has been applied for two decades. However, recent comparative assessments between in-house and commercial assays suggested room for improvement regarding the agreement of positive signals of the applied real-time PCRs. In order to assess the effect of the choice of the target sequence, 3 inhouse real time PCR assays targeting the 18S rRNA gene (<i>n</i> = 2, one of them later referred to as SSU rRNA gene assay to avoid confusion) and the <i>hsp70</i> gene of <i>C. cayetanensis</i> were compared in a head-to-head comparison with 905 samples with high pretest probability for <i>C. cayetanensis</i> infections from Ghanaian HIV patients in a test comparison without a reference standard. Only slight agreement kappa of 0.095 was observed. In the assays targeting the SSU rRNA gene, the 18S rRNA gene, and <i>hsp70</i>, positive signals were recorded in 63, 45, and 0 instances, respectively, with latent class analysis-based estimation of sensitivity of 32.2%, 23.3%, 0% as well as of specificity of 99.7%, 99.9% and 100%, respectively. High cycle threshold values with an average of about 35 indicated low quantities of target DNA in the samples with similar Ct values in concordantly and discordantly positive samples. In conclusion, the study suggested target-gene-specific differences in the diagnostic accuracy of real-time PCR-based diagnosis of <i>C. cayetanensis</i> as well as an ongoing need for further standardization of this diagnostic approach.