Estimation of Trypanosoma cruzi infection in the main vector Triatoma infestans: accounting for imperfect detection using site-occupancy models.

BACKGROUND

Vector infection prevalence is a key component of vectorial capacity and transmission risk. Optical microscopy observation (OM) of fecal drops has been the classic method for detecting Trypanosoma cruzi infection in triatomine bugs until the advent of polymerase chain reaction (PCR)-based techniques. However, agreement among OM- and PCR-based techniques has been highly heterogeneous.

METHODS

We used hierarchical site-occupancy models accounting for imperfect detection to estimate method-specific detection probabilities of T. cruzi infection in field-collected Triatoma infestans and to assess whether T. cruzi infection varied with triatomine developmental stage and collection ecotope. We also performed a scoping review of the literature on comparisons between OM and PCR for T. cruzi infection diagnosis in triatomines. Triatomines were collected before vector control interventions in Pampa del Indio houses (Argentine Chaco) and examined by OM. We amplified the variable regions of the kinetoplastid minicircle genome (vkDNA-PCR) in DNA extracted from the rectal ampoules of 64 OM-positive and 65 OM-negative T. infestans.

RESULTS

vkDNA-PCR detected T. cruzi infection in 59 (92.2%) OM-positive bugs and in 19 (29.2%) OM-negative triatomines in blind tests. The overall prevalence of infection, as determined by a positive test result by either vkDNA-PCR or OM, was 64.3% [95% confidence interval (95% CI) 55.8-72.1%]. Detection probability of T. cruzi infection by vkDNA-PCR (92%, 95% CI 83-97%) was substantially higher than for OM (76%, 95% CI 65-84%). Infection was minimal (26.2%) in peridomestic nymphs and maximal in domestic adult triatomines (81.7%). In the literature review encompassing 26 triatomine species from 11 countries, inter-method agreement ranged from 28.6% to 100%. The lowest agreement was observed in Rhodnius sp. and Panstrongylus lutzi and the highest among Triatoma sp., with wide variability in the protocols and outcomes of molecular diagnosis in comparison with OM.

CONCLUSIONS

Our study provides a synthesis on the different sources (both biological and technical) of variation of the outcomes of OM- and PCR-based diagnosis of T. cruzi infection in triatomines and identifies new research needs for diagnostic improvement.