Feline leishmaniosis in the Mediterranean Basin: a multicenter study

Background Cats are now recognized as competent hosts for Leishmania infantum and a blood source for sand fly vectors. Although canine leishmaniosis (CanL) is endemic in Mediterranean Basin countries, large-scale epidemiological studies are lacking for feline leishmaniosis (FeL). This study aimed to assess the prevalence of L. infantum infections, associated risk factors, clinical signs, and clinicopathological abnormalities in domestic cat populations from six Mediterranean Basin countries. Methods From 2019 to 2022, blood and serum samples of cats (n = 2067) living in Italy (n = 300), Greece (n = 297), Portugal (n = 295), France (n = 231), Israel (n = 313), and Spain (n = 631) were collected along with animal data (i.e., age, sex, breed, housing conditions, and geographical origin), clinical signs, and laboratory blood test parameters. Cats were grouped according to their age as kittens (up to 1 year), young (older than 1 and younger than 7 years), mature (between 7 and 10 years), and senior (older than 10 years). Serum samples were tested for L. infantum by immunofluorescence antibody test (IFAT) and enzyme-linked immunosorbent assay (ELISA), and blood samples of seropositive cats were tested for L. infantum kinetoplast deoxyribonucleic acid (kDNA). Viral infection by feline immunodeficiency virus (FIV) and feline leukemia virus (FeLV) was molecularly addressed in all cats enrolled. Statistical analysis was performed to evaluate the association between the risk of L. infantum infection and independent variables, and among co-infection of L. infantum with FIV and/or FeLV, clinical signs, and clinicopathological abnormalities. Results Overall, 17.3% (358/2067) of cats scored positive for L. infantum by serological tests. Specifically, 24.7% were from Portugal, 23.2% from Greece, 16.6% from Israel, 15% from Spain, 13.3% from France, and 12.6% from Italy. Leishmania infantum DNA was detected in 15 seropositive animals. Housing condition and FIV infection proved to be risk factors for FeL. Leishmania seropositivity was significantly associated with weight loss, lymphadenomegaly, gingivostomatitis, and oral ulcers, as well as with reduced albumin and albumin/globulin ratio, increased total globulins and total proteins, leukocytosis, and thrombocytosis. Conclusions This study provides, for the first time, a large-scale epidemiological survey on FeL and its clinical presentation, revealing that L. infantum circulates among domestic cats, especially shelter/free-roaming and FIV-infected animals, living in CanL endemic countries of the Mediterranean Basin. Graphical Abstract


Background
Leishmania infantum (Kinetoplastida: Trypanosomatidae) is one of the most important zoonotic vector-borne pathogens transmitted by sand flies (Diptera: Psychodidae) [1] and is widespread in countries of the Mediterranean Basin, Middle East, western Asia, and Brazil [2,3].Dogs, the main reservoirs of this protozoan parasite, are fundamental for the circulation of the infection in endemic areas, with a significant proportion of them not showing any clinical sign or presenting few and often non-specific symptoms [4][5][6].Cats have long been considered less susceptible hosts for L. infantum; however, they are now recognized as competent hosts for this parasite and a blood source for its sand fly vectors [7][8][9].Accordingly, in recent decades, many cases of feline leishmaniosis (FeL) have been described in regions endemic for canine leishmaniosis (CanL) (e.g., Italy, Spain, Greece, Portugal, Cyprus, Israel, Brazil), with prevalence ranging from 0.7% to 70%, depending on animal lifestyle, number of animals tested, and diagnostic technique employed [10,11].In addition, in Brazil, there are reports on L. infantum-infected cats in 12 out of 27 states [9], highlighting a potential reservoir role of these felids [12,13].Under specific ecological habitats (e.g., animal shelters) and in the presence of CanL, cats can be significantly more exposed to L. infantum infection (i.e., 75%) than dogs (i.e., 37%).Notwithstanding, they present lower parasitemia [6].Overall, studies performed in endemic regions, point out a lower prevalence of FeL when compared with CanL from the same area [14].For example, in the Aeolian Islands (Sicily, Italy), a Mediterranean region highly endemic for CanL, a 25.8% seroprevalence of FeL was reported, being about half of the prevalence recorded in dogs (i.e., 41.8%) [15].
Although scientific information about FeL has increased in the past few years, many clinical and diagnostic challenges remain unsolved, hindered by a broad spectrum of clinical signs and laboratory abnormalities described in the literature [16][17][18].Among FeL diseased animals, unspecific clinical observations (i.e., lymphadenomegaly, weight loss, and pale mucous membranes) are often associated with cutaneous (i.e., exfoliative and ulcerative dermatitis, nodules, onychogryphosis, nasal/ footpad hyperkeratosis) and ocular lesions (i.e., keratoconjunctivitis, uveitis), alongside hyperglobulinemia and mild/moderate non-regenerative anemia [11,18].In addition, infection with the feline immunodeficiency virus (FIV) and/or feline leukemia virus (FeLV) may predispose cats to becoming co-infected with Leishmania and developing FeL [8,19].Nonetheless, the subclinical presentation of FeL may challenge its diagnosis, which usually requires a combined approach using direct and indirect Leishmania-specific laboratory tests [20,21].Qualitative/quantitative immunofluorescence antibody test (IFAT) and enzyme-linked immunosorbent assay (ELISA) are the most common serological techniques used for both diagnosis and epidemiological studies [22][23][24].Similarly, molecular investigations (e.g., real-time polymerase chain reaction [qPCR], conventional polymerase chain reaction [cPCR]) are widely employed in research and clinical practice [12,25].
Given the endemicity of CanL in the Mediterranean Basin, the increasing number of FeL cases recorded, and the absence of epidemiological studies carried out on a large scale, this study aimed to assess the prevalence of L. infantum infection, associated risk factors, and clinicalhematological abnormalities in domestic feline populations from six countries of the Mediterranean Basin.

Study locations
From 2019 to 2022, domestic cats (n = 2067) were sampled by six veterinary academic institutions in southern European countries, specifically Italy (n = 300), Greece (n = 297), Portugal (n = 295), France (n = 231), Israel (n = 313), and Spain (n = 631).Approximately 300 cats were sampled per country, using the same enrollment criteria.One exception was Spain, where the number of enrolled animals was double that in the other countries, in proportion to country's geographical dimensions, and given the fact that it presents one of the areas with the highest CanL endemicity in the studied region [26,27].Serum and blood samples were sent to the Parasitology Unit of the Department of Veterinary Medicine, University of Bari (Italy), for IFAT and molecular analysis.Individual aliquots of each serum sample (approximately 20 μl) were sent to the School of Veterinary Medicine, Hebrew University, Israel, for in-house ELISA testing.Whenever serum volume was insufficient, only IFAT was performed.

Inclusion criteria
All the animals included in the study had a history of outdoor access and had not been treated with ectoparasiticides or repellents for at least 6 months prior to sampling.At enrollment, signalment data (i.e., age, sex, breed), housing conditions, and the geographical area (zip code, city/town) for each cat were recorded in individual files.In addition, health status and laboratory parameters, including complete blood cell count (CBC) and serological biochemical parameters (i.e., creatinine, urea, alanine aminotransferase [ALT], albumin, total proteins, total globulins, and albumin/globulin [A/G] ratio, were recorded, when available, in individual clinical cards.Clinical data were summarized as follows: general signs (i.e., fever, pale mucous membranes, hepatomegaly, weight loss, jaundice, asthenia, anorexia, lymphadenomegaly), skin signs (i.e., ulcers/crusts/scales, dandruff, nodules/hemorrhagic cysts, alopecia, squamous cell carcinoma), ocular signs (i.e., blepharitis, keratitis, conjunctivitis, uveitis), oral signs (i.e., gingivostomatitis, mouth ulcers), gastrointestinal signs (i.e., vomiting, diarrhea), respiratory signs (i.e., dyspnea, nasal discharge, ocular discharge), and renal sign (i.e., chronic kidney disease [CKD]).Specifically, at least 10% of the cats sampled from each geographical area had clinicopathological information.Cats enrolled were grouped according to their age as kitten (up to 1 year), young (older than 1 and younger than 7 years), mature (between 7 and 10 years), and senior (older than 10 years) [28].

Serological testing
Serum samples were tested for antibodies against anti-L.infantum by means of IFAT and ELISA.The IFAT was carried out at the Parasitology Unit, Department of Veterinary Medicine of Bari University (Italy), following a previously described protocol [24].Positive and negative controls included serum samples from a cat infected with L. infantum, previously diagnosed using IFAT and molecular assays (i.e., qPCR), and serum samples from healthy cats previously tested negative by the methods described, respectively.A sample was scored positive whenever it produced a clear promastigote fluorescence at a cut-off dilution of 1:80, as currently recommended by the Leish-Vet guidelines [20].Positive serum samples were titrated by serial dilutions until negative results were obtained.All IFAT tests were read in a double-masked manner by two different operators.
Serum samples were tested by ELISA at the School of Veterinary Medicine, Hebrew University, Rehovot, Israel, using crude leishmanial antigen, as described previously [6].Each plate was read when the absorbance (lambda = 405 nm) of the positive cat reference serum reached a value between 1.1 and 1.2.A titration of positive and negative reference cat sera was included on each plate to monitor inter-assay variation.The serological cut-off of optical density (OD) = 0.6 was calculated based on three standard deviations above the mean OD value of readings from 13 control serum samples from seronegative and PCR-negative cats living in a non-endemic region.

Molecular testing
Deoxyribonucleic acid (DNA) was extracted from 200 μl of whole blood using a commercial kit (QIAamp DNA Blood & Tissue Kit, Qiagen, Hilden, Germany) according to the manufacturer's instructions and analyzed for the detection of FIV and FeLV proviral DNA by PCR using primers and protocol described previously [29,30].Whole blood DNA samples of seropositive (by IFAT and/ or ELISA) cats were further tested by qPCR for the detection of a fragment (120 base pairs) of the L. infantum kinetoplast deoxyribonucleic acid (kDNA) minicircle, using primers, probes, and protocols described elsewhere [31].For all qPCR runs, positive (i.e., DNA extracted from a blood sample of a cat molecularly and serologically positive for L. infantum) and negative controls (DNA extracted from blood samples of negative healthy cats) were included.Samples were scored as positive for L. infantum kDNA when a threshold cycle of less than 37 was recorded.

Mapping and statistical analysis
The cumulative prevalence of Leishmania infection was calculated by summing the seropositivity by IFAT and ELISA.The 95% confidence interval (CI) values were calculated for each prevalence recorded using Epitools -Epidemiological Calculators software [32].Agreement between IFAT and ELISA test results was evaluated by Cohen's kappa test.The location of L. infantum-positive cats was georeferenced using a geographical information system (GIS) program (QGIS software, Buenos Aires version).
Feline categorical data were summarized as count and percentage.Comparisons between independent groups were performed by the chi-squared test.Differences in cumulative serological positivity in cats of different sexes, ages, housing conditions, and countries, as well as of individuals infected or not by FIV and FeLV, were analyzed using a generalized linear model with a binomial error structure (1 = serological positivity, 0 = negativity) and a fixed factor: y = Xß + ε, where y is the vector of the observation (i.e., the cat serological positivity for L. infantum), X is the incidence matrix, ß is the vector of fixed effects (i.e., housing condition, sex, age, country of origin, FIV or FeLV infection), and ε is the vector of the random residual effects.A probability level of P < 0.05 was used to assess the significance of differences among values.

Discussion
The data presented indicate that L. infantum circulates within domestic feline populations living in countries of the Mediterranean Basin, where CanL is endemic.Indeed, this study collectively represents the first largescale epidemiological survey on FeL conducted using the same procedures and diagnostic protocols.Hence, the overall seroprevalence of L. infantum recorded herein (17.3%) is not comparable with that derived from previous epidemiological studies conducted in individual countries, and often at a regional level [10,11].Nonetheless, it is known that sand fly vectors of L. infantum (i.e., Phlebotomus ariasi, Phlebotomus neglectus, Phlebotomus perfiliewi, Phlebotomus perniciosus) are present in the investigated areas [6,[33][34][35][36][37], and they may have cats as a blood source.Accordingly, a statistical positive association between Leishmania infection and antibody response to P. perniciosus saliva was described in cats [38], and cat blood was detected in P. perniciosus specimens from Spain [39][40][41] and Italy [42].The L. infantum seroprevalence herein recorded in cats from each individual country is higher than in most of the previous studies, probably because it derives from the combination of IFAT and ELISA results and not just one serological test, as before [11].Nonetheless, the seropositivity rates detected in Israel (16.6%),Spain (15%), and Italy (12.6%) are consistent with those retrieved in other studies from the same areas [43][44][45][46][47][48][49].Likewise, the prevalence recorded in France (i.e., 13.3%) is similar to that reported in the only epidemiological investigation available from this country (i.e., 12%) [50], although the pathogen circulation was confirmed in many clinical case reports [51][52][53].On the other hand, the rates of Leishmania-positive cats from Greece (i.e., 23.2%) and Portugal (i.e., 24.7%) are higher than those of the other investigated countries.At least for Portugal, the above picture may be related to the feline population sampled, which was mostly represented by shelter or free-roaming animals (i.e., 98.9%) (Table 2).Indeed, an outdoor lifestyle favors the animal's exposure to sand fly bites, compared to individual housing [54,55], as also suggested in a longitudinal study conducted in the Aeolian Islands [15].The scattered-coastal distribution of FeL herein described may be related to the humid climate conditions of the investigated areas, which are suitable for sand flies to thrive [33].This indicates the importance of regular use of repellents in cats living in seaside environments, to reduce the risk of sand fly bites [56].In Israel, the distribution of FeL overlaps that of CanL in the central region of the country [57].Overall, the seroprevalence of L. infantum in cats from each country is about half of that in dogs from the same areas (i.e., Italy: 12.6% in cats vs. up to 29.6% in dogs; Spain: 15% vs. up to 57%; Portugal: 24.7% vs. up to 56%; France: 13.3% vs. up to 29.6%; Israel: 16.6% vs. up to 36%; Greece: 23.2% vs. up to 50%) [6,26,36,[58][59][60].This ratio was previously discussed in different epidemiological contexts [15,48], alluding to the possibility of cats playing a less important role than dogs in the perpetuation of the L. infantum life cycle.However, these data should be interpreted cautiously, considering the difficulties in comparing animal populations, speciesspecific host-parasite interactions, and the accuracy of the available diagnostic tests.Indeed, one of the major challenges encountered in critically assessing Leishmania seroprevalence lies in the lack of consensus among techniques used in laboratories worldwide (i.e., IFAT, ELISA, western blot, direct agglutination test [DAT]) and cut-off values [14,24,61,62].Given the absence of standardized FeL screening tests, both IFAT and ELISA were used, having a fair agreement (κ agreement = 0.246), as previously suggested [62][63][64].Accordingly, IFAT has a higher sensitivity than ELISA in the detection of subclinical/ asymptomatic feline infections [62].Therefore, the first technique is recommended for epidemiological purposes aiming to detect the exposure of cats to Leishmania, even if clinically healthy.Conversely, ELISA should be preferred in animals presenting signs suggestive of disease [62,64].As more than half of the IFAT-positive cats (i.e., 54%) had low antibody titers (1:80), these animals may have been exposed to L. infantum and/or may have generated a protective (Th1) immunity towards the parasite [65].This is also supported by the high percentage of seropositive cats (i.e., 74%) with no clinical signs.Moreover, the detection of L. infantum DNA in blood samples of only a few individuals confirmed that blood is not a proper biological sample for Leishmania diagnosis, given the low parasitemia in the feline host [5,6,66].Overall, future research should investigate the immune response of cats to L. infantum infection and validate the application of molecular tools with non-invasive feline samples (i.e., conjunctival swabs), as has been done for the diagnosis of CanL [67][68][69].
Shelter/free-roaming cats had a higher risk of infection than owned cats (P = 0.003), reflecting a higher exposure to phlebotomine sand flies, but also to several adverse environmental conditions (e.g., adverse weather conditions, poor nutritional state) that may impair their health status and favor Leishmania infection.In addition, FIV infection was found to be a risk factor for FeL (P = 0.002), as immunocompromised cats might be more prone to developing FeL clinical signs [8,[70][71][72].A similar picture was found in human patients with HIV and co-infected with Leishmania [73].Therefore, FIV infection should be considered within the diagnostic algorithm of Leishmania infection and for evaluating the prognosis of the disease.The clinical signs associated with seropositivity for L. infantum (i.e., weight loss, P = 0.023; lymphadenomegaly, P = 0.002; gingivostomatitis, P = 0.017; and oral ulcers, P = 0.002) agree with the FeL clinical presentation described previously [11].Conversely, dermatological lesions, although defined as common FeL clinical signs [17,74], were not predominant among seropositive cats examined in this study.Thus, FeL should be considered as a polysymptomatic disease that may or not present dermatological signs.Furthermore, the laboratory abnormalities statistically related to seropositive animals (i.e., hypoalbuminemia, P = 0.002; reduced A/G ratio, P < 0.0001; increased total globulins, P < 0.0001; increased total proteins, P = 0.004; leukocytosis, P = 0.028; thrombocytosis, P = 0.006) are consistent with the pathogenesis of infection and reflect the pathological findings typically observed in CanL [18].These outcomes are supported by previous data [11] and underscore the importance of investigating hypergammaglobulinemia as a possible laboratory alteration indicative of FeL.Nevertheless, clinical signs and hematological abnormalities reported here should be interpreted prudently considering the possible occurrence of concomitant diseases and/or co-infections within the studied feline population.

Conclusions
This study provides, for the first time, a large-scale epidemiological survey on FeL, highlighting the circulation of L. infantum among domestic cats, especially shelter/ free-roaming and FIV-infected animals, living in CanL endemic countries of the Mediterranean Basin.Some clinical signs and clinicopathological abnormalities were statistically related to Leishmania infection in the studied cat populations.Therefore, FeL should be included in differential diagnoses for feline patients with suggestive clinical signs and clinicopathological abnormalities when they live in or have traveled to regions where the disease is endemic.Under the above circumstances, the data substantiate the need for preventive measures using proper repellents for cats during the sand fly transmission season.

Fig. 1
Fig. 1 Map of study areas (i.e., A Portugal, B Spain, C France, D Italy, E Greece, F Israel) indicated by provinces, showing the sample size of cats enrolled, according to their Leishmania infantum seropositivity

Fig. 2
Fig. 2 Correspondence analysis showing the association between clinicopathological abnormality scores [A albumin; B total globulins; C total proteins; D white blood cells, WBC; E platelet count, PLT; F albumin/globulin (A/G) ratio] and L. infantum seropositivity in cats.Each clinicopathological abnormality score for selected parameters is indicated by a red cross, whose distance from 0 and closeness to "yes" or "no" serological cumulative positivity indicate an association

Table 1
Comparison of serological and molecular prevalence of Leishmania infantum infections with animal data

Table 2
Serological (IFAT, ELISA) prevalence of Leishmania infantum infections in shelter/free-roaming and owned animals per geographical area

Table 3
Health status, laboratory parameters, and retrovirus infections in the feline population studied

Table 4
Risk factors associated with Leishmania infantum infection in cats