Findings

Canine leishmaniosis is zoonotic disease caused by L. infantum, a vector-borne protozoan parasite transmitted by phlebotomine sand flies under natural conditions. Clinical manifestations of infection may range from mild to absent to mild, several and even fatal disease. This variability is thought to be the result of cell-mediated immune response from the host which may be influenced by the genetic background of the dog [1]. Individual incubation times range from a few months to several years [2]. The detection of L. infantum antibodies indicates existing or past infections. High antibody levels are often associated with a high parasitic load and disease [3]. Conversely, low antibody levels in clinically normal dogs with a negative result on molecular and/or parasitological tests may indicate exposure or early stages of Leishmania infection [4]. Different confirmatory techniques were used and the diagnosis was based on the clinical manifestation and/or the laboratory abnormalities that were compatible with the disease as well as by the confirmation of L. infantum infection [5].

Serological assays, including enzyme-linked immunosorbent assay (ELISA), indirect immunofluorescence antibody test (IFAT), and rapid tests, are the most common methods for the detection of infection in dogs [6]. Rapid tests have been used as an important first step in diagnostic algorithms, enabling results to be obtained within a short time [7].

The present study evaluated the diagnostic performances, including the sensitivity (probability of a positive test result among those having the target condition), specificity (probability of a negative test result among those without the target condition), accuracy (the ability of a test to differentiate the patient and healthy cases correctly) and area the receiver operating characteristic curve (measure of the usefulness of a test showing the relationship between sensitivity and specificity) of a rapid qualitative immunochromatographic in dogs with different anti-Leishmania antibody levels and compared the results from a combination of two quantitative serological tests, IFAT and ELISA, as reference tests.

Blood samples from 244 dogs of different breeds, ages, and genders that were naturally exposed to L. infantum infection were included in the study from May 2017 to May 2018. These dogs presented at the Veterinary Teaching Hospital of the University of Zaragoza (Spain) for different diagnostic purposes, including the annual screening program for clinically healthy dogs, cases of suspected clinical leishmaniosis, the blood donor screening program and pre vaccination screening for L. infantum infection (Table 1). The serological status was recorded by a retrospective review of sample files through the laboratory database (Tables 2 and 3). Two aliquots of each obtained serum sample were stored at -20 °C until testing. Because samples were collected for the sole intention of determining a diagnosis, ethical approval was not needed. However, the owners were required to sign an informed consent.

Table 1 Details of dogs including purpose of diagnosis and signalment
Full size table
Table 2 Frequency of clinical signs in diseased dogs (n = 121, with IFAT and ELISA result)
Full size table
Table 3 Frequency of hematologic and biochemical alterations in diseased dogs (n = 121, with IFAT and ELISA result)
Full size table

The rapid test (FASTest Leish®, MEGACOR Diagnostik, Hörbranz, Austria) was performed following the instructions of the manufacturer. All tests were stored at room temperature and were performed as described in the instructions supplied with the test kit. This rapid test employs a combination of monoclonal antibodies conjugated with colloidal gold particles and recombinant L. infantum antigens bound to the solid phase of a nitrocellulose membrane that detect anti-L. infantum antibodies in whole blood, plasma or serum from the dog. The antibodies against L. infantum present in the sample react in the conjugate pad with mobile monoclonal antibodies to form antibody complexes. These antibody complexes migrate along the nitrocellulose membrane and bind to fixed Leishmania spp. antigens, producing clear pink-purple colored test line. Particles that do not bind to the conjugate continue their course along the membrane and pass through the control line with membrane-fixed control antibodies. The control line shows that the sample and reagents have been properly applied and migrated through the device. The buffer diluent facilitates the migration and promotes the binding of antibodies to antigens. Samples with a clear test and control line are classified as positive, and samples that show a control line are classified as negative. Prior to the study, technicians were trained to perform and interpret results. All tests were read by two laboratory technicians after 15 min. If discrepancies arose between results, a third observer participated. The examiners were blinded to the results of the quantitative serological tests.

The commercial IFAT (MegaFLUO LEISH test®, MEGACOR Diagnostik, Hörbranz, Austria) was performed as described in the instructions supplied with the test kit. This commercial IFAT has been employed in other comparative studies of diagnostic tests [8]. All samples were examined by two different investigators. They were blinded to the results of the other serological tests. Samples negative at 1:100 were considered negative and the endpoint titer of positive samples was determined by preparing serial twofold dilutions of the serum starting from the cut-off value (1:100).

As a second quantitative serological reference method, an in-house ELISA by L. infantum was performed on all serum samples as previously described, with some modifications [9]. The in-house ELISA test was performed by a different researcher who had no knowledge of the rapid test and IFAT results. The cut-off value was 30 ELISA unit (EU). Sera with an EU ≥ 200 were classified as high positive, with an EU ≥ 100 and < 200 as moderate positive, and with an EU > 30 and < 100 as low positive (Additional file 1).

Data obtained were checked for normal distribution using the Kolmogorov–Smirnov test (P > 0.05). A regression analysis between the two quantitative tests and Pearson´s correlation coefficients were calculated for all results that were concordant among the three serological tests. Statistical analyses were performed using IBM SPSS statistics software version 22 (SPSS Inc., Armonk, NY, USA). Because a single widely accepted reference standard for the diagnosis of L. infantum infection in dogs is not available, samples were classified based on results coinciding in both quantitative reference tests. Samples for which discordant results were obtained with the reference tests were not included in the study. Moreover, dogs that had previously been vaccinated with either of the two vaccines available in Spain for the prevention of canine leishmaniosis were also not included in this study.

Of the 244 sera analysed, 232 samples showed concordant results for IFAT and the in-house ELISA. The remaining 12 samples were considered discordant between reference tests, including 2 samples with a low positive result for the in-house ELISA but a negative result (antibody titer < 1:100) for the IFAT, as well as 10 samples with a positive result (antibody titer of 1:100) for the IFAT but negative result for the in-house ELISA. All discordant samples between the two reference tests were classified as seronegative by the rapid test. Among the 232 serum samples characterized, 121 samples were classified as seropositive for L. infantum infection and 111 samples were classified as seronegative for L. infantum infection (Table 4). Serum samples with different serum anti-Leishmania spp. antibody levels were selected for the evaluation (Table 5). All serum samples classified as positive by both reference assays were also positive in the FASTest LEISH®. Only one seronegative sample classified by the reference assays was discordant and showed a positive rapid test result. There were no additional freeze–thaw cycle or sample handing differences in testing events between the three tests.

Table 4 Comparison of the FASTest LEISH® with IFAT and ELISA results for Leishmania spp. antibody detection
Full size table
Table 5 Number of positive samples and antibody levels (IFAT and ELISA)
Full size table

The rapid test achieved sensitivity of 100% (95% confidence intervals (CI) 96.2–100%) and specificity of 99.1% (95% CI 94.4–100%), and the accuracy was 99.6% (95% CI 97.3–100%). The Cohen’s kappa coefficient between the rapid test and the combined reference assays for the detection of antibodies was 0.99. The area the receiver operating characteristic curve (AUC-ROC) obtained from the curve was 0.995 (95% CI 0.985–1.000) (Fig. 1). Agreement between the two observers in reading the rapid test was 100% and no invalid results were obtained. The regression line was statistically significant (P < 0.001) and Pearson´s correlation coefficient was r = 0.675.

Fig. 1
figure 1

AUC-ROC curve analysis of the rapid test studied. The Area the receiver operating characteristic curve (AUC-ROC) analysis combines sensitivity and specificity into one measurement, and the result is a single global measure of diagnostic effectiveness. Each point on the ROC curve represents a sensitivity/specificity pair corresponding to a particular decision threshold. The blue line represents the AUC-ROC analysis of the FASTest LEISH®, and the green line represents the reference line

Full size image

Sensitivity and specificity are common parameters for evaluating diagnostic methods [10]. Depending on the purpose of the test, sensitivity or specificity should be prioritized. When a test is used to detect L. infantum infection in dogs with suspected clinical leishmaniosis, specificity and sensitivity must be high, but importantly, specificity must not be too low because a confirmatory test should be highly specific to avoid false positives. In contrast, high sensitivity is essential for identifying clinically healthy infected dogs.

When a reliable gold standard is not available, at least two different tests should be performed [11]. In the present study, sample classification was established based on combination of results from ELISA and IFAT, obtaining a better definition of the serological status. When different serological tests are used simultaneously, discrepancies may be observed, and the lack of consistency among tests may depend on their different performances. A possible explanation for these discrepancies may be the inherent differences in the two quantitative assays. Therefore, the differences between the diagnostic performances of the references tests may modify the accuracy of the rapid test analyzed.

Numerous rapid tests are available to detect antibodies in dogs. However, these tests show high specificity and variable sensitivity [12]. Differences in diagnostic results obtained between rapid tests are influenced by the serum panel, technology, type of antigen, test threshold and the different laboratory test considered as a reference assay (Table 6). Standardization and comparison of results between the evaluated test and other similar rapid tests from different studies is not possible due to differences in the parameters used to evaluate a test, such as sample size, definition of the clinical status of the dog involved in the study and the selected reference populations. Therefore, examinations of diagnostic tests should follow the STARD guidelines (Standards for the Reporting of Diagnostic Accuracy Studies) to allow an adequate comparison between results.

Table 6 Concordant results between reference test and results of each reference test for sensitivity and specificity of the FASTest LEISH®
Full size table

Sample inclusion can influence assessment measures with any test. Serum samples with high levels of anti-Leishmania spp. antibodies are more easily detected by a test and contribute to higher diagnostic values [13]. Samples in the present study had different antibody levels ranging from 42 to 372 EU, according to the in-house ELISA. Conversely, antibody titers detected by IFAT ranged from 1:400 to > 1:6400. The main differences between the ELISA and IFAT techniques are the type of antigen used and the technical method performed to obtain the results. In terms of interpreting the results, IFAT is subjective and depends on the operator´s experience, even when two different experienced observers examine the samples. This situation may explain the discordant results obtained.

A potential limitation of the present study was the lack of samples from dogs serologically reactive to other pathogens, particularly different Leishmania spp. or Trypanosoma cruzi, which are both protozoa that belong to the Trypanosomatidae family and share various antigens that may produce cross-reactions [14, 15]. All serum samples were collected from dogs from the city of Zaragoza, a region in Spain where T. cruzi is not present and L. infantum is the parasite responsible for canine leishmaniosis in Europe. Ideally, for a proper clinical sample characterization, the use of techniques of different nature is highly recommended. However, the objective of the study was to evaluated the ability of a rapid test to detect antibodies. For this reason and to characterize the seropositive status of the samples, serological techniques are necessary instead of other methods.

In conclusion, the use of FASTest LEISH® as a screening test, may be particularly useful for identifying clinically healthy, seropositive, infected dogs.