Review articles Leishmaniasis emergence in Europe

Leishmaniasis emergence in Europe is reviewed, based on a search of literature up to and including 2009. Topics covered are the disease, its relevance, transmission and epidemiology, diagnostic methods, treatment, prevention, current geographical distribution, potential factors triggering changes in distribution, and risk prediction. Potential factors triggering distribution changes include vectorial competence, importation or dispersal of vectors and reservoir hosts, travel, and climatic/environmental change. The risk of introducing leishmaniasis into the European Union (EU) and its spread among Member States was assessed for the short (2-3 years) and long term (15-20 years). There is only a low risk of introducing exotic Leishmania species because of the absence of proven vectors and/or reservoir hosts. The main threat comes from the spread of the two parasites endemic in the EU, namely Leishmania infantum, which causes zoonotic visceral and cutaneous leishmaniasis in humans and the domestic dog (the reservoir host), and L. tropica, which causes anthroponotic cutaneous leishmaniasis. The natural vector of L. tropica occurs in southern Europe, but periodic disease outbreaks in Greece (and potentially elsewhere) should be easily contained by surveillance and prompt treatment, unless dogs or other synanthropic mammals prove to be reservoir hosts. The northward spread of L. infantum from the Mediterranean region will depend on whether climate and land cover permit the vectors to establish seasonal biting rates that match those of southern Europe. Increasing dog travel poses a significant risk of introducing L. infantum into northern Europe, and the threat posed by non-vectorial dog-to-dog transmission should be investigated.


Leishmaniasis
Leishmaniasis (or 'leishmaniosis') is a complex of mammalian diseases caused by parasitic protozoans classified as Leishmania species (Kinetoplastida, Trypanosomatidae) [1,2].Natural transmission may be zoonotic or anthroponotic, and it is usually by the bite of a phlebotomine sandfly species (order Diptera, family Psychodidae; subfamily Phlebotominae) of the genera Phlebotomus (Old World) and Lutzomyia (New World) [2,3].Primary skin infections (cutaneous leishmaniasis) sometimes resolve without treatment, with the host developing acquired immunity through cellular and humoral responses [4], but the infection can spread to produce secondary lesions in the skin (including diffuse cutaneous leishmaniasis), the mucosa (muco-cutaneous leishmaniasis) and the spleen, liver and bone marrow (visceral leishmaniasis, which is usually fatal if untreated) [1].Worldwide, at least 20 Leishmania species cause cutaneous and/or visceral human leishmaniasis (HumL) [1,5].Most foci occur in the tropics or subtropics, and only zoonotic L. infantum is transmitted in both the eastern and western hemispheres [5] (Table 1).

Worldwide and European relevance of leishmaniasis
The World Health Organization (WHO) reports that the public health impact of leishmaniasis worldwide has been grossly underestimated for many years [1].In 2001 and 2004, Desjeux reported that in the previous decade endemic regions had spread, prevalence had increased and the number of unrecorded cases must have been substantial, because notification was compulsory in only 32 of the 88 countries where 350 million people were at risk [5,6].About two million new cases of HumL (half a million visceral) are considered to occur every year in the endemic zones of Latin America, Africa, the Indian subcontinent, the Middle East and the Mediterranean region [1].
Risks of emergence or re-emergence of leishmaniasis in Europe are associated with three main scenarios: 1) the introduction of exotic Leishmania species or strains into Europe via the increasing worldwide travelling of humans [6] and domestic dogs [7], 2) the natural spread of visceral and cutaneous leishmaniasis caused by L. infantum and L. tropica from the Mediterranean region of Europe, where these species are endemic [1,8,9], to neighbouring temperate areas where there are vectors without disease [2], 3) the re-emergence of disease in the Mediterranean region of Europe caused by an increase in the number of immunosuppressed people.
The high prevalence of asymptomatic human carriers of L. infantum in southern Europe [10][11][12][13] suggests that this parasite is a latent public health threat.This was demonstrated by the increase of co-infections with human immunodeficiency virus (HIV) and Leishmania that has been observed since the 1980s [14], with leishmaniasis becoming the third most frequent opportunistic parasitic disease after toxoplasmosis and cryptosporidiosis [15].

Disease transmission and epidemiology
Visceral leishmaniasis (VL) is usually fatal if untreated, and so it is distinguished from cutaneous leishmaniasis (CL) in all sections of the current review.If untreated, uncomplicated CL is often disfiguring, but not fatal.
In contrast, muco-cutaneous and diffuse cutaneous disease can lead to fatal secondary infections even if treated.Patient immunodeficiency is one factor for this, but in Latin America these diseases are associated with regional strains of the L. braziliensis and L. mexicana species complexes [1,5].
A female sandfly ingests Leishmania while blood-feeding, and then transmits the infective stages (usually accepted to be the metacyclic promastigotes) during a subsequent blood meal [16].The infective promastigotes inoculated by the sandfly are phagocytosed in the mammalian host by macrophages and related cells, in which they transform to amastigotes and often provoke a cutaneous ulcer and lesion at the site of the bite.
There are only two transmission cycles with proven long-term endemism in Europe [2,17]: zoonotic visceral and cutaneous HumL caused by L. infantum throughout the Mediterranean region; and, anthroponotic cutaneous HumL caused by L. tropica now occurring sporadically in Greece (Table 1, Figure 1).
Worldwide, most transmission cycles are zoonoses, involving reservoir hosts such as rodents, marsupials, edentates, monkeys, domestic dogs and wild canids [2,5,6,18] (Table 1).However, leishmaniasis can be anthroponotic, with sandflies transmitting parasites between human hosts without the involvement of a reservoir host.Anthroponotic transmission is characteristic of species of the L. tropica complex and, except for L. infantum, of the L. donovani complex.One species (L.donovani sensu stricto) or two species (L.donovani and L. archibaldi) cause periodic epidemics of anthroponotic visceral leishmaniasis ('Kala-azar') in India and northeast Africa, respectively [19].Sandfly vectors of both complexes (L.donovani and L. tropica) are abundant in southern Europe.
The domestic dog is the only reservoir host of major veterinary importance, and in Europe there is a large market for prophylactic drugs and treatment of canine leishmaniasis (CanL) caused by L. infantum [2].Domestic cats might be secondary reservoir hosts of L. infantum in southern Europe [13], because they are experimentally infectious to sandflies [20] and natural infections can be associated with feline retroviruses [21].
Fewer than 50 of the approximately 1,000 species of sandflies are vectors of leishmaniasis worldwide [3].This is due to the inability of some sandfly species to support the development of infective stages in their gut [16] and/or a lack of ecological contact with reservoir hosts [22].Our understanding of the fundamentals of leishmaniasis epidemiology has been challenged in the last 20 years.Firstly, HIV/Leishmania co-infections were recorded in 35 countries worldwide, and widespread needle transmission of L. infantum was inferred in southwest Europe [15], where Cruz et al. demonstrated Leishmania in discarded syringes [23].
Secondly, leishmaniasis has become more apparent in northern latitudes where sandfly vectors are either absent or present in very low densities, such as in the eastern United States (US) and Canada [24] as well as in Germany [25][26][27].Most infections involve CanL, not HumL, and this might be explained by dog importation from, or travel to, endemic regions, followed by vertical transmission from bitch to pup or horizontal transmission by biting hounds [24].Vertical transmission of HumL from mother to child has rarely been reported [28].

Diagnostic methods
Most diagnoses are only genus-specific, being based on symptoms, the microscopic identification of parasites in Giemsa-stained smears of tissue or fluid, and serology [18,29].Consequently, the identity of some causative agents has only been known relatively recently, following typing performed during limited ecoepidemiological surveys.For example, it was thought that all cutaneous leishmaniasis cases in Europe were caused by L. tropica, until Rioux and Lanotte reported L. infantum to be the causative agent in the western Mediterranean region [30].
Rioux and Lanotte used multi-locus enzyme electrophoresis (MLEE) to identify Leishmania species and strains [30], which remains the gold standard [1,18].However, MLEE requires axenic culture [31] in which one strain can overgrow others in mixed infections.It is therefore more practical to identify the isoenzyme strains (or zymodemes) by directly characterising the enzyme genes [32].Other molecular tests have been used to identify Leishmania infections in humans, reservoir hosts and sandfly vectors [33], including in the Mediterranean region [34], but there has been no international standardisation [29].However, PCR of the internal transcribed spacer of the multi-copy nuclear ribosomal genes is often used [34,35].A set of carefully chosen criteria must accompany PCR-based diagnosis, especially for immunocompromised patients [14,15].Monoclonal antibodies have long been available for the identification of neotropical species [36] and the serotyping of Old World species [37] but they are not widely used.
Most sensitive molecular techniques indicate only the presence of a few recently living Leishmania, not that the parasites were infectious.Therefore, serology is often more informative [29].However, antigens prepared in different laboratories can cause test variation for the frequently used methods [29]: the indirect fluorescent antibody test (IFAT), the enzyme-linked immunosorbent assay (ELISA), the indirect haemagglutination assay (IHA) and the direct agglutination test (DAT).Some antigens are stable and produced commercially, such as the recombinant (r) K39 for a dipstick or strip test.Multi-centre studies of 'Kala-azar' diagnostics [38] showed that both the freeze-dried DAT and the rK39 strip test could exceed the 95% sensitivity and 90% specificity target, but only for the strains found in some regions.Antibody detection tests should complement other diagnostic tests, because they do not usually distinguish between acute disease, asymptomatic infections, relapses and cured cases [38].
Delayed hypersensitivity is an important feature of all forms of human leishmaniasis [4] and is often measured by the leishmanin skin test (or Montenegro reaction) [29].False-positivity is approximately 1% in otherwise healthy people.Other problems with this test include the absence of commercially available leishmanins, that there is complete cross-reactivity among most species and strains of Leishmania, and that for VL its applicability is limited to the detection of past infections, because a complete anergy is found during active disease.

Treatment
Pentavalent antimonials were the first-choice drugs for leishmaniasis worldwide [39,40].Miltefosine, Paromomycin and liposomal Amphotericin B are gradually replacing antimonials and conventional Amphotericin B in some regions [40,41], especially where there is drug resistance or the need to develop combination therapy to prevent the emergence of resistance to new drugs [41].
Highly active anti-retroviral therapy (HAART) treatment has reduced the incidence of co-infections with Leishmania and HIV by preventing an asymptomatic infection with L. infantum from becoming symptomatic, but unfortunately it is not good at preventing visceral leishmaniasis relapses.The benefits of treatment are not as clear-cut as they are for other opportunistic diseases [42].

Prevention
Most research on vaccines is strategic, not applied, for example targeting secretory-gel glycans of Leishmania [43] and some sandfly salivary peptides [44], both of which are injected into the mammalian host by the female sandfly during blood feeding.Therapeutic vaccine trials continue to use killed cultured parasites (often with BCG as adjuvant) in combination with antileishmanial drugs but with only 0-75% efficacy [45].
One second generation recombinant vaccine contains a trifusion recombinant protein (Leish-111f), and some of its epitopes are shared by L. donovani and L. infantum [46].
Research and development of vaccines against CanL has been stimulated by the economic importance of dogs and their role as reservoirs of HumL caused by L. infantum in the Americas and the Mediterranean region.Leishmune is the first licensed vaccine against CanL.It contains the fucose-mannose ligand (FML) antigen of L. donovani and has a reported efficacy of 76-80% [47].The industrialised formulation of FMLsaponin underwent safety trials in Brazil [48].The vaccine LiESAp-MDP (excreted/secreted antigens with adjuvant) was reported to have an efficacy of 92% when tested on naturally exposed dogs in the south of France [49,50].More recently, a modified vaccinia virus Ankara (MVA) vaccine expressing recombinant Leishmania DNA encoding tryparedoxin peroxidase (TRYP) was found to be safe and immunogenic in outbred dogs [51].
One means of controlling transmission is to reduce the biting rate of peri-domestic sandfly vectors of visceral HumL and CanL.This has been effective locally, by using repellents [52], insecticide-impregnated nets and bednets [52], topical applications of insecticides [53] and deltamethrin-impregnated dog collars [54,55].The latter are favoured by many dog owners in southern Europe.

Outside the European Union
Table 1 (updated from Ready [2]) relates the distributions of each form of HumL to causative species and known reservoir hosts [1,5,6,17,18].Most VL foci occur in India and neighbouring Bangladesh and Nepal, and in Africa (Sudan and neighbouring Ethiopia and Kenya), where anthroponotic 'Kala-azar' is caused by L. donovani and in north-eastern Brazil and parts of Central America, where zoonotic infantile visceral leishmaniasis is caused by L. infantum.Most CL foci occur in Latin America, North Africa, and the Middle East, and muco-cutaneous and diffuse cutaneous disease are frequent in South America [56].

Inside the European Union: main biomes
Only two transmission cycles have been endemic in the European Union (EU) for a long time, and both are widespread in the adjoining Middle East and in North Africa: zoonotic cutaneous and visceral leishmaniasis caused by L. infantum throughout the Mediterranean region and anthroponotic cutaneous leishmaniasis caused by L. tropica, which occurs sporadically in Greece and probably neighbouring countries and poses a high risk of introduction by migrants and travellers into the rest of the EU [2,6,17] (Table 1, Figure 1).The former is endemic and sandfly-borne only in the Mediterranean region of the EU ('Mediterranean forests' biome), where its epidemiological significance is clear from published serological surveys [7,8].However, the vectors of L. infantum [57] (Figure 2, Table 2, updated from Ready [2]) are also abundant in the adjoining parts of the temperate region (Temperate broadleaf forests' biome), in northern Spain [58] and central France [59], and small numbers occur as far north as Paris [60] and the upper Rhine valley in Germany [26].The occurrence of 'vectors without disease' poses a significant risk for the emergence of leishmaniasis in temperate regions of Europe [2].

Climate change
Most transmission of Leishmania is by the bite of permissive sandflies, and so climate change might affect leishmaniasis distribution directly, by the effect of temperature on parasite development in female sandflies [16], or indirectly by the effect of environmental variation on the range and seasonal abundances of the vector species.Female sandflies seek sheltered resting sites for blood meal digestion, and in southern Europe the temperatures of these micro-habitats are buffered but vary significantly with the external air temperature [2].
Based on molecular markers, European vectors of leishmaniasis have extended their ranges northward since the last ice age (approximately 12,000 years ago) [61,62], and the mapping of statistical measures of climate has permitted transmission cycles to be loosely associated with some Mediterranean bioclimates [63].However, bioclimate zones and their vegetation indicators vary regionally, and ongoing climate change may alter the patterns of land cover and land use.The geographic information system (GIS)-based spatial modelling of the Emerging Diseases in a changing European Environment (EDEN) project is permitting an analysis of changes in climate and land cover [64] and their effects on sandflies.
The project 'climate Change and Adaptation Strategies for Human health in Europe' (cCASHh) concluded: "There is no compelling evidence, due to lack of historical data, that sand fly and VL distributions in Europe have altered in response to recent climate change" [9].
There is now a published analysis of the northward spread of CanL and its vectors in Italy [65], but an association with climate change was only surmised.

Capacity and competence of vectors in Europe
Vectorial capacity has only been calculated indirectly.
The average number of gonotrophic cycles (i.e.egg development following a bloodmeal) completed by P. ariasi in the south of France was only a little greater than one [66].Therefore, relatively small changes in temperature could have a large effect on vectorial capacity, because transmission occurs only during the second or subsequent bloodmeals and temperature affects the level of activity of the sandfly and therefore the frequency of the bloodmeals.

Alone, PCR detection of a natural infection of
Leishmania in a sandfly does not identify a vector.It only indicates that the sandfly has fed on an infected mammalian host [35] because many parasites do not survive in a non-permissive sandfly after bloodmeal defecation [16].
Vectorial competence has been tested [67] or inferred based on finding naturally infected females of the more abundant human-biting species [3,57,68], from which it  is concluded that the principal vectors of L. infantum in the Mediterranean region are members of the subgenus Larroussius (Table 2, Figure 2).The vectorial competence of Phlebotomus (Transphlebotomus) mascittii should be tested because this species is now known to be widespread in northern France, Belgium and Germany [69].However, low rates of biting humans and autogeny (the ability to produce eggs without a bloodmeal) cast doubt on its epidemiological importance [2].
Based on distribution and vectorial competence elsewhere, P. sergenti sensu lato is likely to be the main vector of L. tropica in southern Europe [3].

Importation or dispersal of vectors and reservoir hosts
The importation or inter-continental dispersal of vectors is unlikely because sandflies are not as robust as some mosquitoes and are not known to be wind-dispersed [3].Any importations are unlikely to be significant for several reasons: The natural vectors of Old World leishmaniasis are already abundant in Mediterranean Europe (Table 2, Figure 2); most American sandflies are believed to be poor vectors of Old World Leishmania species [3]; and Leishmania species native to the Americas have hosts that do not occur in Europe [56].
The vector of L. major in North Africa and the Middle East is P. papatasi, which is locally abundant in southern Europe.However, the natural reservoir hosts of this parasite are usually gerbil species not present in EU countries [18] and the risks of them dispersing into southern Europe or surviving accidental/deliberate release by humans have not been assessed.

Importance of travel within Europe (mainland and overseas territories) and internationally
Travel has led to increasing numbers of HumL cases that need to be treated, e.g. in France [70], Germany [25], Italy [71] and the United Kingdom [72].Leishmaniasis in Guyana (overseas region of France) is a major source of exotic cases imported to mainland France, and L. infantum has travelled in the reverse direction in a dog [73].Isoenzyme [12] and molecular markers [32,34] can sometimes identify the origins of Leishmania strains.
Travel poses the risk of the emergence in southern Europe of anthroponotic L. donovani [74] and L. tropica (see above), and the introduction to northern Europe of L. infantum in dogs taken to the Mediterranean region on holiday or rescued from there as strays [7].

The logic of visceral leishmaniasis control
Based on compartmental mathematical (R 0 ) models, Dye [75] concluded that insecticides can be expected to reduce the incidence of HumL caused by L. infantum even more effectively than they reduce the incidence of CanL, but only where transmission occurs peridomestically and the sandfly vectors are accessible to treatment, as in parts of Latin America.For control of HumL and CanL in Europe, Dye [75] concluded that a dog vaccine is highly desirable, because sandfly vectors here are less accessible to insecticide treatment.In Europe, CanL is a veterinary problem with socio-economic importance and a vaccine is more likely to be afforded than elsewhere.

Risk assessment of introduction, establishment and spread in the European Union (EU) for the short term (2-3 years)
'Oriental sore' caused by L. tropica is usually anthroponotic, and it is sporadically endemic in Greece and endemic in neighbouring countries to the EU.However, all these parasites pose a significant risk of introduction to Europe by intravenous drug users (IVDUs) and the establishment of local transmission by syringe needles, especially if these patients have HIV co-infections.This is based on the experience with endemic visceral leishmaniasis caused by L. infantum [15,42].

Risk assessment of introduction, establishment and spread in the European Union for the long term (15-20 years)
Leishmania infantum is currently the only significant causative agent of visceral and cutaneous HumL endemic in Europe.Its high prevalence in asymptomatic humans and in the widespread reservoir host (the domestic dog) means there is a high risk of emergence in parts of Europe further north, as demonstrated in northern Italy [65].In addition to risk factors [78] and statistical models [64] with associated risk maps, EDEN is producing R0 mathematical models as part of research to explain why large regions of temperate Europe have sandflies without HumL in the presence of imported CanL.Some of the key data come from questionnaires to veterinary clinics, validated by prospective serological surveys of CanL, from northern and southern areas with a wide range of disease prevalence.
Increasing dog travel poses a significant risk of introduction of L. infantum into northern Europe from the Mediterranean region.There is also a risk of establishment of non-vector transmission and spread as has been observed in North America [24].Non-vector transmission might explain the autochthonous cases of CanL in Germany [25,26].
L. tropica has been isolated from both the domestic dog and the black rat [5,8], and so the risk of introduction and spread of CL caused by this parasite in the EU should be re-assessed if either these mammals or related synanthropic species were found to be reservoir hosts (rather than dead-end hosts) in the disease foci in North Africa and southwest Asia [35].

Assessment of whether the existing data sources are adequate and, if not, identification of missing key data needed for conducting risk assessment studies
Research data about leishmaniasis and its spatial distribution in Europe and the Mediterranean region are being enhanced [79] and made accessible online by EDEN and another EU-funded project, LeishRisk, which has collaborated with the WHO to produce an E-compendium, a compilation of peer-reviewed literature on leishmaniasis epidemiology [1,80].
However, public health and veterinary surveillance data are more fragmentary, which undoubtedly caused the public health impact of leishmaniasis to be underestimated for many years in Europe as well as worldwide [1].The WHO has concluded that more surveillance is necessary in Europe to assess an emergence of leishmaniasis [9], but the partners of the EDEN leishmaniasis sub-project have stressed the need for better coordination of existing surveillance, including linking human health and veterinary data for the zoonotic disease.Currently (EDEN partners, personal communications), HumL is notifiable in Greece, Italy, Portugal and Turkey, and in endemic autonomous regions in Spain.CanL is notifiable in Greece and at municipality level in the endemic regions of the four other countries mentioned above.Neither disease is notifiable in France.At international level, WHO organised a meeting of Eurasian countries in 2009 (J.Alvar, personal communication), aimed at standardising surveillance and reporting, and CanL is reported as a listed disease ('other diseases') of the World Organisation for Animal Health (OIE) [81].
The monitoring of dog travel [7] should continue to be improved and standardised.

Figure 1
Figure 1 Distribution by country of Leishmania species transmitted by phlebotomine sandflies in Europe up to 2009

Figure 2
Figure 2 Distribution of vectors of leishmaniasis in European countries up to 2009

Table 1
European distribution of parasites causing most human leishmaniasis worldwide up to 2009 change: the impact on the epidemiology and control of animal diseases (S. de la Roque, ed.).Rev Sci Tech Off Int Epiz.2008;27(2):399-412.

Table 2
European distributions of sandfly vectors of human leishmaniasis up to 2009 (unproven role throughout range) risks of emergence as sandfly-borne diseases in southern Europe in the short and long term because of the absence of their exotic vectors and mammalian reservoir hosts.
borne disease in southern Europe in the short and long term, even though its principal vector (P.papatasi) is locally abundant, because its main gerbil reservoir hosts are absent.Cutaneous leishmaniases caused by the American parasites of the L. braziliensis and L. mexicana complexes have low