Asymptomatic Leishmania infantum infections in humans living in endemic and non-endemic areas of Croatia , 2007 to 2009

K Šiško-Kraljević (katarina.sisko.kraljevic@mefst.hr)1,2,3, A Jerončić3,4, B Mohar1, V Punda-Polić1,5 1. Department of Medical Microbiology, Split University School of Medicine, Split, Croatia 2. Department of Microbiology, Public Health Institute of Split and Dalmatia County, Split, Croatia 3. These authors contributed equally to this work and share first authorship 4. Department for Research in Biomedicine and Health, Split University School of Medicine, Split, Croatia 5. Department of Clinical Microbiology, Split University Hospital, Split, Croatia


Introduction
Leishmaniasis is caused by Leishmania spp., hemoflagellate protozoa belonging to the family Trypanosomatidae.Infected phlebotomine sandflies serve as vectors for the transmission of all Leishmania species.Human leishmaniasis can be divided into three main disease manifestations: (i) visceral leishmaniasis (VL), (ii) cutaneous leishmaniasis (CL) and (iii) mucocutaneous leishmaniasis (ML).The strain of the infecting organism and the host's immunologic status greatly influence clinical manifestations.
Zoonotic VL caused by Leishmania infantum is endemic in Mediterranean countries of Europe and domestic dog is the main reservoir [1][2][3].During the last decade, there have been reports of spreading leishmaniasis northward into previously non-endemic areas of central and northern Europe [1][2][3].It is well known that, besides clinical cases of VL, asymptomatic infections are common in endemic areas [2,3].
Croatia encompasses 56,538 km2.According to the 2011 census, it had a population of 4,284,889 [4].Geographically, Croatia is composed of three areas: the Adriatic coastal zone with islands in the south, the Pannonian Plain in the north, and the mountainous region in-between.The Croatian littoral comprises a relatively narrow land belt with islands along the eastern coast of the Adriatic Sea (Figure).The littoral is traditionally divided into two large areas based on geography, ecology and cultural heritage: the northern (Istria and Primorje counties) and the southern part (Dalmatia) which is further subdivided into northern, central and southern Dalmatia.The Adriatic islands and the coastal zone are characterised by Mediterranean climate which provides good living conditions for sandfly vectors of Leishmania [5,6].Continental Croatia has temperate continental or continental climate and is usually considered free of phlebotomine vectors of Leishmania.
Human VL and canine VL has been reported in central and southern coastal and insular Dalmatia (from Split to Dubrovnik) since 1930 [7].From 1931 to 1957, 398 human VL cases were diagnosed in this region [8].A case of CL was firstly recorded in 1945 and 201 CL cases were recorded by 1957.After the late 1950s, the number of VL cases declined, probably because of mass spraying with antimalarial insecticides [8].Since 1990, studies have identified re-emerging foci of both human and canine VL in central and southern Dalmatia [9][10][11].
In Croatia notification is compulsory for both VL and CL, although these diseases are not included in the decisions of the European Parliament for reporting communicable diseases [12,13].Medical practitioners are notifying human leishmaniasis cases through Croatia's health information system or directly to the epidemiologist in the regional public health institutes.Cases are defined as probable which is a case clinically compatible in endemic region or with epidemiological connection to a confirmed case, and confirmed that is a case laboratory-confirmed by positive parasitological (microscopy or cultivation) and serological (IFA, ELISA) tests.Case definitions for CL are similar, excluding the serological tests [13].According to data periodically published by the Croatian National Institute for Public Health there were between one and four new cases of VL reported each year in Croatia in the last decade [14].The estimated mean annual incidence of human leishmaniasis is 0.4 per 100,000 population [5].Leishmaniasis in Croatia is described as predominately paediatric disease: almost half of VL patients are children up to the age of 10 years and the disease is more often found in men [9,10].Most of the reported cases occurred among inhabitants of the Croatian coast and islands [9,10].Besides this, few VL cases were diagnosed in Austrian [15] and Hungarian [16] tourists after returning from the Dalmatian littoral.
Recently, three Phlebotomus species, known to serve as vectors for L. infantum were found in central and southern coastal and insular Dalmatia [5].A veterinary seroepidemiologic survey conducted in central Dalmatia among apparently healthy dogs using dot-ELISA, found canine seroprevalence ranging from zero to 42.85%, which was in accordance with previous recognition of central and southern coastal and insular Dalmatia as a high endemic foci of L. infantum for dogs [11].In contrast to high seroprevalence ratios in dogs, no information about the prevalence of the infection in otherwise healthy human inhabitants of different Croatian regions is available.
The aim of this study was to assess the prevalence of IgG antibodies to L. infantum among healthy people living in different regions of Croatia, and to compare the seroprevalence in endemic regions with that in non-endemic regions.This is the first investigation on asymptomatic leishmaniasis in residents of Croatia.

Methods
The target population for our study was the apparently healthy general population in previously recognised endemic and non-endemic areas for leishmaniasis in Croatia.The studied areas known as endemic areas: insular Croatia, and central and southern Dalmatia, as well as areas previously considered nonendemic but with favourable Medieranean climate: northern Dalmatia, Istria and Primorje.Two counties in northern Croatia with continental climate where sandflies were considered rare or absent: Brod-Posavina County (centre Slavonski Brod) and the most north-western Međimurje County (centre Čakovec), were also included.

Serological survey
Serum samples were collected from 2007 to 2009 to examine the prevalence of IgG antibodies to L. infantum in Croatia.We used 'residual' sera of apparently healthy, immunocompetent individuals who did not show any symptoms of leishmaniasis and came to hospitals or clinics for routine laboratory check-ups or for blood donation.Participating hospitals' or clinics' laboratories were selected according to their geographic location: 14 served population along the Adriatic coast and two were from non-endemic parts of Croatia.Laboratories provided data on age, sex and site of residence for the study participants.To determine the target sample size for each region we used published statistical tables, which provide minimal sample sizes that are necessary for given combinations of precision, confidence levels, and variability.
A minimum of 1,100 individuals were selected as the target sample size for the coastal region with previously recognised endemic foci in order to reach the precision level of ±3% under assumption of the confidence level of 95%, p=0.5 and the size of population greater than 100,000 inhabitants.For islands, the targeted precision on population with more than 100,000 inhabitants was ±5% resulting in a minimum sample size of 400, whereas for the reference non-endemic continental region, under the assumption of maximum seroprevalence of 2%, the targeted sample size was 84 to reach the ±3% precision level.
Convenience sampling was used to select respondents in the most economically, technically, and operationally feasible method [17].
To estimate whether the study sample represents the research setting population, we compared the percentage of men estimated from the study sample with the 2011 census data (Table 1).
We classified sites based on their seroprevalence as 'high endemic focus' and 'moderate seroprevalence'.We derived these definitions combining the data from the literature with the results of this study.In particular, in a study by Federico et al. a seroprevalence of L. infantum infection of 4% in the Rome and Caltanissetta area was marked as 'moderate' [18] whereas Marty et al. detected a seroprevalence of 38% in the high endemic focus of Alpes-Maritimes, using Western blot technique, which is more sensitive than the technique used in this study [19].The seroprevalence rate in the majority of our coastal areas was around 7% and defined as 'moderate seroprevalence'.The one area with a seroprevalence of 22% had a significantly higher seroprevalence in comparison with surrounding areas, all of which were classified as 'moderate seroprevalence'.Therefore, that site was defined as a 'high endemic focus' and the definition was further supported by epidemiological data since the majority of cases of VL in Croatia are reported from 'high endemic foci' [5,[9][10][11].
Sera were tested for the presence of IgG antibodies to L. infantum by commercial enzyme-linked immunosorbent assay (NovaLisa Leishmania infantum IgG, NovaTec Immunodiagnostica GmbH, Dietzenbach, Germany), according the manufacturer's instructions.Declared specificity and sensitivity of the test were 85% and 91% respectively.A serum was considered positive when the ratio between the optical density (OD450) value of the serum and the cut-off was >1.1.
The study was approved by the ethics committees of the Split University School of Medicine and of the Split University Hospital.Individual informed consent was not required according to the ethical committee.
To analyse the spatial distribution pattern of Leishmania infection in more detail, we estimated strength of associations of seropositivity with the site of residence in particular regions: coastal, continental, or islands (Table 3).Only in Adriatic coastal counties seroprevalence of sites differed significantly from the average regional rate.While central Dalmatia inhabitants had

Discussion
To our knowledge, this is the first study of the prevalence of asymptomatic Leishmania infection in the general population of Croatia.In order to determine the Leishmania infection distribution, residents of various geographical and ecological areas were included in a serological survey.Primorje) where cases of VL have not been reported yet, suggests that Leishmania transmission toward north may have occurred.
A highly endemic focus in central Dalmatia was confirmed in accordance with previous reports of both canine and human leishmaniasis in central and southern Dalmatia's coast and islands [5,[7][8][9][10][11].However, the northwestern part of the Adriatic littoral, including Istria, Primorje, and northern Dalmatia, was considered a non-endemic region [5,[7][8][9][10][11].In 2002, a case of VL was diagnosed in a patient who had no history of travel to known endemic regions and apparently had contracted the infection during his stay on the Velebit Mountain in northern Dalmatia [20].Our results confirm this observation as they indicate that asymptomatic Leishmania infection is found throughout the eastern Adriatic coast and islands.Depending on the geographical location, a moderate to high (4%-22%) prevalence of asymptomatic infection has been observed.
Data from seroepidemiological studies conducted in other Mediterranean countries have also shown a variable prevalence (from 0.5 to 56%) of Leishmania infections depending on the geographic regions studied and on the test used for detection [3,21].In our study we used a commercial ELISA as a relatively simple method for testing a large number of sera.Convenience sampling was used for selecting respondents.Despite some of its limitations, in a study by Kelly et al. it was shown that a convenience sample of sera from diagnostic laboratories was an appropriate sampling strategy to provide population immunity data to inform country's health policies [17].Although the healthy donor effect bias was present in Brod-Posavina County, blood donors represented only a minority in other sites and are not expected to have influenced results.
The highest prevalence of 22.2%, significantly higher than in other coastal zones, was found in central Dalmatia.In comparison with other coastal areas, central Dalmatia has a high background of canine leishmaniasis and sandflies [5,6,11] which are likely to be associated with high prevalence of asymptomatic Leishmania infection among residents in this area.Central Dalmatia is also an active focus with the highest number of human VL cases in Croatia [9,10,14].Most of VL cases diagnosed in Croatia during the study period from 2007 to 2009 occurred in central Dalmatia: seven cases were diagnosed with a mean annual incidence of 0.5 per 100,000 population (unpublished data).In southern Dalmatia, Leishmania exposure was higher than expected (8% seroprevalence), despite the lack or small number of clinical cases and apparently lower risk of infection than observed for central Dalmatia.
As expected, the seroprevalence was significantly lower in residents of continental areas of Croatia; in fact all but one person (a 40 year-old woman) were seronegative.As these regions are not considered to be endemic for sandflies, we cannot exclude the possibility that the one seropositive person could have been infected while traveling to the Croatian littoral where people from continental areas often spend their summer vacations.
The same possibility of travel-acquired infection with consecutive seroconversion has to be taken in account when the seemingly northward spread of asymptomatic leishmaniasis is interpreted.One could also speculate that autochthonous Leishmania infection may occur due to the observed spread of sandflies to some regions at the North of Croatia [1,2], especially in light of the recent finding of Phlebotomus spp. in southern Hungary near the Croatian border [16].
In agreement with other reports [21][22][23][24] our study shows that there is no difference in the prevalence of asymptomatic Leishmania infection between men and women.
In respect to age, our prevalence results differ from findings in other studies.A study in Brazil, using a non-commercial ELISA found that 28.5% sera of 638 tested children aged between 0 and 5 years were positive, and concluded that infection is associated with the age of ≥2 years [27].In our study, seropositives were most likely to be aged 0 to 9 years.Since in Croatia VL is still predominantly paediatric disease [6,7] it cannot be excluded that some of the seropositives in this age group might become symptomatic, therefore further differently designed studies with follow up of such participants are needed.
It can be concluded that compared to other parts of Croatia, seroprevalence is significantly higher in central and southern coastal as well as insular areas.This indicates the presence of asymptomatic L. infantum infection in humans and confirms the endemicity of these areas.This finding may be of particular importance in light of the increasing popularity of the Croatian coast and islands, from Istria to Dubrovnik, as a holiday destination for travellers from Leishmania-free areas or countries.Our data should alert physicians to consider leishmaniasis in the differential diagnosis of conditions such as unexplained febrile illness especially in immunocompromised subjects returning from these endemic areas.In addition, seropositivity observed in non-endemic areas and the higher seroprevalence in children should be investigated in the future.Therefore, further studies including clinical, parasitological, epidemiological and entomological investigation are required for elucidating the cycle of transmission, the maintenance and the role of Leishmania in human and animal health in different Croatian regions.

Table 1 Sex distribution in the study population (n=2,035) compared with that of the total Croatian population
CI: confidence interval.a Healthy donor effect bias present in Brod-Posavina county as blood donors constituted the majority of the sample.
Leishmania infection, including age as categorical variable, sex, and region of residence.Significance level was set at 0.05.In case of multiple testing, we adjusted the p-values with the Bonferroni correction.
of seroprevalence for anti-Leishmania IgG in asymptomatic healthy individuals,Croatia, 2007Croatia,  - 2009 (n=2,035)    (n=2,035) In contrast to sex, we found significant differences in the median age of seropositive and seronegative persons: 40 years (IQR: 16-58), and 42 years (IQR: 22-60), respectively (Mann-Whitney U test, p=0.039).This indicated age as a risk factor for Leishmania infection.Additionally, in the multivariate model with age, sex and geographical region as covariates, age Age and sex and seropositivity for anti-Leishmania IgG of study population by area of residence,Croatia, 2007Croatia,  -2009 (n=2,035)    (n=2,035) The findings of this study reveal a strong association of seropositivity with geographic region (east Adriatic coastal and islands areas) and age group.The presence of seropositive people in northwestern coastal and island regions (Istria and Table2

Table 3
Seroprevalence of anti-Leishmania IgG by residence compared with unweighted average in the associated region,Croatia,  2007Croatia,   -2009, (n=2,035)   , (n=2,035) b Site samples were too small for reliable comparison.

Table 4
Seroprevalence of anti-Leishmania IgG by age, sex and region of residence, Croatia, 2007-2009 (n=2,035) a Significant at level p<0.05.