Reliability of techniques used in the diagnosis of canine visceral leishmaniasis by the national control program in Brazil: A survey in an area of recent transmission.

One of the key components of the Brazilian Program for the Control of Visceral Leishmaniasis (PCLV) is the euthanasia of Leishmania-infected canine reservoirs, the detection of which depends on a screening procedure involving a Dual Path Platform® (DPP) immunoassay and a confirmatory enzyme-linked immunosorbent assay (ELISA). The aims of the present study were to evaluate the reliability of these techniques in a region of recent transmission of canine VL, to follow up the seroconversion 3-4 months after the initial diagnosis of DPP reactive but ELISA indeterminate or non-reactive dogs, and to identify the species of Leishmania in circulation in the area. Each animal was submitted to DPP under field conditions, performed by municipal health workers using peripheral blood (DPP-field), to DPP under laboratory conditions using serum (DPP-lab) and to ELISA using serum. The agreements between the tests were determined using McNemar's χ2 test, Cohen's kappa coefficient (k) at the 95% confidence interval and prevalence-adjusted bias-adjusted kappa (PABAK). Of the 1130 dogs examined, 74.2% were non-reactive in all three tests applied. Based on the PCLV positive-infection criterion, seroprevalence was 8.9% (101/1130) with 83.2% (84/101) of infected animals showing reactivity in all three tests while 7.8% (8/101) were reactive in DPP-field and ELISA and 8.9% (9/101) in DPP-lab and ELISA. The proportions of disagreements were substantial in all comparisons. Inter-rater reliability between DPP-field and ELISA (k=0.55; PABAK=0.78) and DPP-lab and ELISA (k=0.59; PABAK=0.81) were considered moderate, while that between DPP-field and DPP-lab (k=0.61; PABAK=0.79) was classified as marginally good. The proportion of seroconversions in DPP reactive animals that were initially ELISA indeterminate was significantly higher than in those that were DPP reactive but initially ELISA non-reactive. Restriction fragment length polymorphism analysis revealed the presence of Leishmania infantum, the etiologic agent of VL, in bone marrow samples from VL-infected animals. Our data showed that the techniques and protocols currently employed in the PCLV screening approach are not entirely reliable. Further consideration should be given to monitoring dogs with undetermined results in ELISA and a better training should be provided for health workers responsible for performing DPP tests applied under field conditions.