Comparative study of serologic tests for the diagnosis of asymptomatic visceral leishmaniasis in an endemic area.

Serologic tests have been widely used for the diagnosis of asymptomatic visceral leishmaniasis. This study evaluated five serologic tests used for the diagnosis of asymptomatic infection: enzyme-linked immunosorbent assay (ELISA) using promastigote antigen (ELISAp), ELISA using recombinant K39 (ELISA rK39), and K26 (ELISA rK26) antigens, an indirect immunofluorescence test using Leishmania (Leishmania) amazonensis promastigote antigen (IIFT), and an immunochromatographic test using rK39 antigen (TRALd). As a reference regarding the performance of the tests, patients with classic visceral leishmaniasis originating from Minas Gerais, Brazil (N = 36), were defined as the positive group and samples of healthy individuals from nonendemic areas (Argentina) (N = 127) were used as negative controls. Patients with other diseases such as cutaneous leishmaniasis (N = 53) and malaria (N = 56) were also studied to evaluate the chance of cross-reactivity in these tests. Finally, subjects from an area endemic for visceral leishmaniasis in Brazil (Porteirinha, northern Minas Gerais) (N = 1241) were screened for asymptomatic infection with Leishmania and Chagas disease. The sensitivity of the serologic tests was 50% (18/36), 66.7% (24/36), 69.4% (25/36), 83.3% (30/36), and 88.9% (32/36) for ELISAp, ELISA rK26, ELISA rK39, IIFT, and TRALd, respectively. Specificity, calculated using the truly negative group, was 96% (122/127) for TRALd, 97.6% (124/127) for ELISAp and IIFT, and 100% (127/127) for ELISA rK39 and rK26. Positivity in at least one test employing recombinant antigen was observed in 24 (45%) patients with cutaneous leishmaniasis and 47 (82.4%) with malaria. In the visceral leishmaniasis-endemic area, the positivity of the serologic tests ranged from 3.9% to 37.5%. The enzyme-linked immunosorbent assay (ELISA) tests using recombinant antigens were more frequently positive in subjects with a history of exposure to human or canine visceral leishmaniasis (ELISArK39: 14.6% [149/1017] versus 37.5% [84/224]; ELISA rK26:12.7% [129/1017] versus 21.4% [48/224], P < 0.001 for both). Kappa agreement was low, with a maximum value of 0.449 between ELISAp and IIFT. In addition, among the 112 IIFT-positive subjects, 75 (67%) also presented positive serology for Chagas disease. In conclusion, IIFT and TRALd presented the best performance to diagnose classic cases of visceral leishmaniasis in an endemic area. Cross-reactivity of the tests with Chagas disease, cutaneous leishmaniasis, and malaria should be taken into account. However, the differences in the positivity of the tests used, together with the low agreement between results, do not permit to select the best test for the diagnosis of asymptomatic Leishmania infection.