Visceral leishmaniasis (VL) is a widespread protozoan zoonosis caused by the genus Leishmania. Leishmania chagasi (Leishmania infantum) is the reported species present in the New World 1 . Dogs are considered the main epidemiological reservoirs of VL in the domestic environment, and mandatory euthanasia of positive dogs has been the basis of disease control in several developing countries, including Brazil 2 . Because the clinical diagnosis of canine visceral leishmaniasis (CVL) is not reliable due to the wide variety of clinical signs of CVL and the high percentage of asymptomatic dogs, serological methods have been used for the definitive diagnosis of CVL 1 .
Although the detection of amastigote forms of Leishmania sp. may be considered the gold standard for the diagnosis of VL, this procedure is laborious and difficult to routinely apply under field conditions. Serological methods, such as the enzyme linked immunosorbent assay (ELISA) and the indirect immunofluorescent antibody test (IFAT), have been indicated as better, rapid and reliable diagnostic methods that are easily standardized and suitable for mass testing 3 . Despite the high sensitivity and specificity of serological assays for canine leishmaniasis, the presence of cross-reactivity remains controversial. The aim of the present study was to evaluate the cross-reactivity of serum samples from dogs serologically reactive for Trypanosoma cruzi, Ehrlichia canis, Babesia canis, Toxoplasma gondii and Neospora caninum by ELISA, IFAT and an immunochromatographic test (Kalazar Detect™, InBios Inc., Seattle, WA, USA) for canine leishmaniasis. All dogs were from areas that were non-endemic for leishmaniasis, Chagas disease and Trypanosoma evansi, free of vectors and with no single reported case to date.
A total of fifty-seven serum samples were used: 14 dogs experimentally infected with T. cruzi (Colombian strain) with chronic cardiomyopathy and anti-T. cruzi IgG antibodies, as determined by IFAT, with titers ranging from 160 to 640 4 ; 12 dogs naturally infected with B. canis and presenting pale mucous membranes, anemia, intraerythrocytic inclusions of Babesia sp. and anti-B. canis antibodies, as determined by IFAT, with titers ranging from 400 to 800 5 ; 13 dogs naturally infected with E. canis and presenting thrombocytopenia, Ehrlichia morulae within macrophages on capillary blood smears and antibodies, as detected by a commercially available kit (SNAP® 3Dx®, IDEXX Laboratories Inc., Westbrook, ME, USA); 10 dogs naturally infected with T. gondii and presenting clinical signs of the disease and anti-T. gondii antibodies, as detected by IFAT, with titers ranging from 1:128 to 1:1024 6 ; and eight dogs presenting clinical signs of N. caninum infection, with anti- N. caninum antibodies, no detected anti-T. gondii antibodies and IFAT antibody titers ranging from 400 to 800 7 .
The serological detection of anti-Leishmania chagasi IgG antibodies was performed by ELISA as previously described 8 using an L. chagasi total promastigote lysate (strain MHOM/BR/74/PP75) and alkaline phosphatase-conjugated anti-dog IgG (Sigma-Aldrich, St. Louis, MO, USA). A cut-off value of 0.270 was determined using the analysis of serum samples from 50 healthy dogs from an area that was non-endemic for leishmaniasis (mean + 3 SD). The detection of L. chagasi antibodies by IFAT was determined as previously described 9 using a suspension of parasites (MHOM/BR/72) in buffered saline solution (PBS) at a concentration of 8.1 x 106 promastigotes/mL and a fluorescein isothiocyanate-conjugated anti-dog IgG (Sigma-Aldrich, St. Louis, MO, USA). The samples were considered positive when titers were ≥ 1:40. Positive sera were serially diluted and tested to establish the maximum reaction titer. The samples were also tested for Leishmania sp. using an immunochromatographic qualitative antibody assay against L. chagasi recombinant K39 (rK39) antigen (Kalazar Detect™, InBios Inc., Seattle, WA, USA) according to the manufacturer’s protocol.
A total of 24/57 (42.1%) dog samples tested positive using any of the three serological methods for Leishmania sp. ( Table 1 ). Of the positive ELISA samples, the optical density (OD) values of the T. cruzisamples ranged from 0.299 to 0.941, with an average of 0.531 ± 0.226, and the OD value of the E. canissample was 0.385 ( Figure 1 ). Of the positive IFAT samples, the titer values ranged from 40 to 160 and from 40 to 80 for the T. cruzi and T. gondii samples, respectively. Of the positive dipstick samples, the number of samples that cross-reacted using this method was lower than those observed using the ELISA and IFAT methods ( Table 1 ).
Infectious/parasitic agents | Positive dogs (n/%) | ||||||
---|---|---|---|---|---|---|---|
n | ELISA | IFAT | Kalazar Detect™ | ||||
n | % | n | % | n | % | ||
Trypanosoma cruzi | 14 | 9 | 64.3 | 6 | 42.9 | 0 | 0.0 |
Ehrlichia canis | 13 | 1 | 7.7 | 0 | 0.0 | 1 | 7.7 |
Babesia canis | 12 | 0 | 0.0 | 0 | 0.0 | 0 | 0.0 |
Toxoplasma gondii | 10 | 0 | 0.0 | 5 | 50.0 | 1 | 10.0 |
Neospora caninum | 8 | 0 | 0.0 | 0 | 0.0 | 1 | 12.5 |
Total | 57 | 10 | 11 | 3 |
ELISA: enzyme linked immunosorbent assay; IFAT: indirect immunofluorescent antibody test.
Although the ELISA and IFAT tests were conducted using Leishmania crude antigens and although the observed cross-reactivity may be explained by a decrease in test specificity 2 , a previous study also used crude antigens and observed no cross-reactivity with E. canis, T. gondii and B. canis 8 . Antibodies against T. cruzi have been recognized using conventional serological methods as the main cause of cross-reactivity with Leishmania due to the phylogenetic similarity between Leishmania sp. and Trypanosoma cruzi, which poses a problem for overlapping endemic areas 10,11 . In the present study, T. cruzi displayed a high percentage of cross-reactivity with Leishmania when using ELISA and IFAT, both of which are tests that are used as standard diagnostic methods for CVL in Brazil. Samples from five dogs cross-reacted with T. cruzi in both the ELISA and the IFAT. The dipstick test was the only test displaying no cross-reactivity between T. cruzi and Leishmania sp., confirming that rK39 does not cross-react with T. cruzi 12 .
The identification of ELISA cross-reactivity with E. canis (1/13; 7.7%) corroborated previous studies that reported 2/3 (66.6%) cross-reactivity on the IFAT and 1/3 (33.3%) cross-reactivity on the ELISA 13 . Although positive, titers were within 2-fold of the cut-off value (OD value 0.385), and these samples tested negative by IFAT. None of the 12 Babesia canis samples displayed cross-reactivity with Leishmania by either ELISA or IFAT, as previously demonstrated 8 .
In a previous study, cross-reactivity between Leishmania sp. and T. gondii was not identified when dog serum samples were tested by ELISA and IFAT 13 . Although IFAT is considered highly specific 3 , the present study has demonstrated that half of the dogs (5/10; 50%) with anti-T. gondii antibodies were mistakenly considered serologically positive for visceral leishmaniasis.
Despite the fact that the dipstick test did not display cross-reactivity with Chagas disease, dogs seroreactive for E. canis (1/13; 7.7%), T. gondii (1/10; 10%), and N. caninum (1/8; 12.5%) induced false positive results on this test. Our results differed from those of a previous study that reported 3/12 (25%) samples cross-reacting with Chagas disease, and our results differed from those of a previous study, which found 3/12 (25%) cross-reacted with Chagas disease and no false positives with T. gondii, although 1/3 (33.3%) dogs with ehrlichiosis presented false positive results on the dipstick 14 . Although the dipstick test has been developed for the diagnosis of visceral leishmaniasis in humans, previous studies have demonstrated that the sensitivity and specificity of this test ranged from 83% to 91.5% and from 94.7% to 100%, respectively, for disease diagnosis in dogs 8,14 . False positive results were also observed in a rapid commercial ELISA test for dog leishmaniasis, with cross-reactivity found in 3/6 samples (50%) of T. gondii, 1/2 (50%) samples of T. gondii and N. caninum, 3/4 (75%) samples of B. canis and E. canis and 4/7 samples (57.1%) of T. cruzi 15 . Thus, rapid inclinic diagnostic tests should be used in association with other diagnostic tests and should be used carefully when screening for canine visceral leishmaniasis in areas that are endemic for ehrlichiosis, babesiosis, toxoplasmosis and neosporosis. In the present study, with the sole use of the ELISA, IFAT or commercial immunochromatographic tests, approximately 19%, 14% and 11% of the dogs tested would have been considered positive for leishmaniasis, respectively. A recent guideline consensus has concluded that dogs should be considered infected by serology only when antibody titers are 2- to 4-fold higher than the positive threshold value. Otherwise, cytology and polymerase chain reaction (PCR) using bone marrow and/or lymph node samples should be conducted to accurately confirm the diagnosis 11 .
Despite the territorial expansion of VL in Brazil, where the disease has reached even the southern states of the country, when the present study was conducted, the number of municipalities in which vectors and reported cases were present was lower than the current number of such municipalities, and many areas were still considered to be disease-free. Thus, the possibility of dogs having been infected prior to the time of this study by Leishmania chagasi (L. infantum) is remote.
In summary, our results indicate that confounding infectious agents such as Trypanosoma cruzi, Ehrlichia canis, Babesia canis, Toxoplasma gondii and Neospora caninum may cross-react with leishmaniasis serological tests and lead clinicians to falsely diagnose visceral leishmaniasis. Thus, the use of serological methods with low antibody titers for the diagnosis of Leishmania-infected dogs as a criterion for euthanasia should be reassessed.