Bias due to methods of parasite detection when estimating prevalence of infection of Triatoma infestans by Trypanosoma cruzi.

The study aimed to quantify the bias from parasite detection methods in the estimation of the prevalence of infection of Triatoma infestans by Trypanosoma cruzi, the agent of Chagas disease. Three common protocols that detect T. cruzi in a sample of 640 wild-caught T. infestans were compared: (1) the microscopic observation of insect fecal droplets, (2) a PCR protocol targeting mini-exon genes of T. cruzi (MeM-PCR), and (3) a PCR protocol targeting a satellite repeated unit of the parasite. Agreement among protocols was computed using Krippendorff Kα. The sensitivity (Se) and specificity (Sp) of each protocol was estimated using latent class models. The PCR protocols were more sensitive (Se > 0.97) than microscopy (Se = 0.53) giving a prevalence of infection of 17-18%, twice as high as microscopy. Microscopy may not be as specific as PCR if Trypanosomatid-like organisms make up a high proportion of the sample. For small T. infestans, microscopy is not efficient, giving a prevalence of 1.5% when PCR techniques gave 10.7%. The PCR techniques were in agreement (Kα = 0.94) but not with microscopy (Kα never significant with both PCR techniques). Among the PCR protocols, the MeM-PCR was the most efficient (Se=1; Sp=1).