|
Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 97(4) 2002, pp. 477-483 Risk Zones of Human Leishmaniases in the Western Mediterranean Basin. Correlations between Vector Sand Flies, Bioclimatology and Phytosociology Philippe Rispail+, Jacques
Dereure, Daniel Jarry Laboratoire de Parasitologie-Mycologie, Faculté de Médecine de Montpellier-Nîmes et Centre Hospitalier Universitaire de Montpellier, 163 rue Auguste-Broussonet, 34090 Montpellier, France Received 6 June 2001 Code Number: oc02090 Correspondence analysis was applied to sand fly sampling in 865 stations from the Western Mediterranean basin. The position of each of 24 species was determined with respect to the bioclimatic belts. Thus, the multidimensional analyses manifest clear correlations between bioclimatic belts and their expression in the area, the phytosociological groupings, and vector species of visceral and cutaneous leishmaniases. The transfer of these data to usual maps allows to delimit the geographical distribution of these diseases in the Western Mediterranean basin and contributes to the determination, in a rational manner, of the high risk zones.
Key words: sand flies - leishmanial
foci - bioclimatic belts - phytosociological groupings - correspondence analysis
- The leishmaniases are parasitic diseases caused by several species of flagellate Protozoa belonging to the genus Leishmania Ross 1903 and transmitted by the bite of the haematophagous female of a phlebotomine sand fly. Around the Western Mediterranean basin, the animal reservoir hosts of visceral and cutaneous leishmaniases are Canidae and Rodentia. In this geographical area, the vector sand flies are members of the genus Phlebotomus Rondani & Berté 1840, while some species belonging to the genus Sergentomyia França & Parrot 1920 are vectors of Leishmania-like parasites of Old World reptiles. Leishmaniases are an important problem for public health: from 1996 to 1999, about a thousand cases was registered in South Europe and more than 35,000 in North Africa (these data are under-estimated because some countries do not declare leishmaniases); numerous cases were associated with Aids. Moreover, the impact of the climate change induces in Mediterranean countries a risk of increase and an extension of the foci of arthropod-borne diseases (Rhodain 2000). The importance of phlebotomine sand flies as vectors and as determinants of the structure of foci of leishmaniases, has been known for many years. The route-transect method, which has been used for more than thirty years (Rioux et al. 1982a), has enabled the collection and identification of about 200,000 sand flies in some 2,000 stations spread all around the Western Mediterranean basin. Simultaneous study of the distribution of the vector species, their density and natural environmental factors favourable to their development, contributed to the analysis and prediction of leishmanial foci (Rioux et al. 1969). Such a large amount of information could not be processed by monofactorial deterministic methods, but requires multidimensional methods of analysis experimentally before used in North Africa (Rioux et al. 1984b, 1986b, Rispail 1986).
MATERIALS AND METHODS Data collection - Maps and field note-books accurately report the route-transects. These transects, carried out in summer (June to September) from 1972 to 1984, covered numerous leishmanial foci of the Western Mediterranean perimeter, in Algeria (Chott Hodna), France (Cévennes, Corsica, Montpellier area, Nice area, Vallespir), Italy (Grossetto area), Morocco (three long transects from humid North to per-arid South), Spain (Catalonia, Valencia area) and Tunisia (northern and Gafsa areas). Trappings were executed by the team of the Laboratoire de Parasitologie-Mycologie, Faculté de Médecine de Montpellier, with the collaboration of many epidemiologists and entomologists from the above-mentioned areas. A minimum of 30 oiled-paper traps, recovered in a good state after three or four nights (Rioux et al. 1982a), allowed the inclusion of 865 stations. A total of 182,813 specimens of 24 species of Phlebotomus and Sergento-myia were identified (Table). Though phytosociological maps actually exist for the periphery of the entire Mediterranean basin, the only phytogeographical map of Emberger (1936, 1939) and bioclimatic map of Sauvage (Sauvage & Brignon 1963), both available solely in Morocco, are enough large-scaled to precisely locate near stations. The availability of these bioclimatological (five belts, from humid to per-arid) and phytosociological data, wide coverage (132,582 sand flies from 353 stations) and the presence of most of the Mediterranean species (21 from 24), led to the choice of Morocco for a special analysis. Data analysis - The "species" file contained, for each station, the number of traps retrieved and, for each species of sand fly, the number of individuals of each sex. This coding allowed analysis of presence/absence and density of both sexes of each species, together or separately. The "descriptors" file contained the available bioclimatological and phytosociological information. Three tables were analyzed to compare species, descriptors and stations: (1) the species/stations table was either a presence/absence table or the result of the calculation, for each species, of its density (ds) by station corrected according to the formula:
in which use of square root is a mathematical way allowing to reduce the dispersion of the variable without modifying the relative position of the species in the multidimensional space; (2) the stations/descriptors table required a discreet coding of different states of variables; (3) the species/descriptors table recorded the sand flies of each species trapped in the stations distributed in each state of each environmental variable (contingency table). Following the tabulation of data, analyses were carried out using programmes developed from those of the algorithm Analyses de Données of the software Biomeco 2.0 package (Groupe Biométrie CEPE/CNRS, Montpellier). Unlike the standard statistics, the data analysis is multidimensional. It results in an illustration in which degrees of relationship, resemblances or differences between the elements are expressed. Different representations of the data, notably geometrical representations (as opposed to algebric, for instance hierarchical ascending classification), are possible. Among geometrical representations, the correspondence analysis, extensively developed in France under the title "analyse factorielle des cor-respondances" (Legendre & Legendre 1979, Benzécri 1982), is a method of cluster analysis closely related to principal component analysis and particularly well ajusted to qualitative and semi-quantitative data. Linear functions of the indicator variables are chosen in such a way that the multidimensional scatter is most closely approximated by a diagram in a small number of dimensions. The individuals in a particular response category can be represented by a point in this diagram, and it may then be possible to identify interesting clusters of individuals (Armitage & Berry 1994). So, the information contained in the numerous primary descriptors is reduced to a small number of factors. Correspondence analysis consequently allowed a global synthesis, conserving maximum information, notably the qualitative information. This multivariate analysis demonstrated relationships among species, among stations, and between species and stations. Four very scarce species (P. mariae, Sergentomyia sp. belonging to bedfordi group, S. lewisi and S. schwetzi) were treated as supplementary elements (no weight in the analysis; only localization in the cloud of points, compared with others species). When data were available, the position of each species with respect to the bioclimatological and phytosociological variables was determined. RESULTS AND DISCUSSION The Western Mediterranean basin - The correspondence analysis ranks the species by station according to their density, in the form of a continuum, along a clear bioclimatic gradient. So, the analysis indicates the inter-specific relationships and the position of each species within this bioclimatic continuum. The factorial plan I - II organizes the species from the humid to per-arid belts. The second axis probably expresses, within the species of the humid and sub-humid belts, an ancestral biogeographical separation between P. (Larroussius) ariasi Tonnoir 1921 abundant on the Western Mediterranean coast, and P. (Larroussius) perfiliewi Parrot 1930 which has a more oriental situation in the Western Mediterranean basin. The projection of the cloud of points in the plot of axes I and III (Fig. 1), distributes the species along a parabola (artefactual but not disturbing "Guttman effect"), from the upper left humid quarter to the sub-humid, semi-arid and arid lower half, then to the per-arid upper right quarter. Whilst the sand flies of the subgenera Larroussius and Transphlebotomus occupy the humid, sub-humid and semi-arid left quarters, the subgenera Paraphlebotomus and Phlebotomus are distributed in the arid and per-arid right quarters. These results are closely related to the geographical distribution of visceral and cutaneous leishmaniases in the Western Mediterranean basin. Indeed, P. ariasi (Rioux et al. 1984a) and P. (L.) perniciosus Newstead 1911 (Izri et al. 1990) were demonstrated as vectors of visceral and cutaneous leishmaniases due to Leishmania infantum, both from Portugal to Italy and from Morocco to Tunisia, where Canidae are reservoir hosts. L. infantum may be also present in the semi-arid belt in sympatric condition with L. tropica (Dereure et al. 1991), and even in the arid belt (Dereure et al. 1986). In these areas, P. (L.) longicuspis Nitzulescu 1930, the only Larroussius present, could be the vector of L. infantum. P. perfiliewi, proven vector in Algeria (Izri & Belazzoug 1993), is scarce in the South East of France (Izri et al. 1994) and more abundant on the Adriatic coast in Italy (Corradetti 1960). P. (Pa.) sergenti Parrot 1917 was found infected by L. tropica in the semi-arid belt in Morocco (Guilvard et al. 1991), and so was shown to be a vector of anthroponotic cutaneous leishmaniasis in this country. P. (P.) papatasi (Scopoli 1786) was found to be the vector of L. major (Rioux et al. 1986a), agent of zoonotic cutaneous leishmaniasis in the arid and per-arid belts, from rodent reservoir hosts [Meriones spp. in Morocco (Petter 1988), Algeria (Belazzoug 1986) and Tunisia (Rioux et al. 1986c); Psammomys spp. in Algeria (Belazzoug 1983)] to man. The example of Morocco - The results indicate that densities of the 21 species are correlated with the same bioclimatic gradient. In the species / bioclimatological and phytosociological factors analysis (Fig. 2), the species are located according to their average density at the stations corresponding to the bioclimatic belts and their expression in the area, the phytosociological groupings (Fig. 3). Thus, P. ariasi is related to the humid and sub-humid bioclimatic belts with Mediterranean oaks as indicators (Fig. 4A). In these areas, L. infantum is the agent of human visceral and cutaneous leishmaniases (Fig. 4B), from the Mediterranean coast (Rif mountains) to the Souss plain and from Atlantic coast to Moyen-Atlas chain. Epidemiologically, the dog is known as the common reservoir host of L. infantum and P. ariasi as the vector (Rioux et al. 1979). Moreover, Killick-Kendrick and Killick-Kendrick (1999) suggested "that the apparent specialisation in honeydew feeding perhaps limits the distribution of P. ariasi to that of Lachnus roboris", aphid which infests leaves of Quercus ilex. Some other restricted foci, related to adequate microclimate, could be favorable to development of L. infantum. So, in Issafen (Anti-Atlas chain, in the arid bioclimatic area), one autochtonous case of canine visceral leishmaniasis due to L. infantum MON-1 was reported (Dereure et al. 1986). Also, in Taghjicht oasis (per-arid belt) in Southern Morocco, several cases of human cutaneous leishmaniasis were related to L. infantum MON-24 (Rioux et al. 1997). In these two areas, P. longicuspis is present and could be the vector. P. sergenti is related to the semi-arid bioclimatic belt where the vegetation is characterized by arar (Tetraclinis articulata) and junipers (Juniperus spp.) (Fig. 5A), and to a lesser degree, to the arid belt, by argan (Argania spinosa) and common jujube (Ziziphus lotus). Anthroponotic cutaneous leishmaniasis due to L. tropica (Fig. 5B) is common in these areas: in Tanent (northern slope of the Haut-Atlas mountain) and Smimou (at the western extremity of the chain) (Pratlong et al. 1991). The epidemiological cycle is commonly considered as anthroponotic, although several cases of canine cutaneous leishmaniasis due to L. tropica were reported in these two foci (Dereure et al. 1991), where P. sergenti was demonstrated as the vector (Guilvard et al. 1991). It is also important to note that Tanent (Azilal area) is a mixed focus where L. tropica in man and L. infantum in dog are sympatric (Dereure et al. 1991). As above-mentioned, P. longicuspis is present with P. sergenti in this focus, and could be the vector of L. infantum. Moreover, several cases of human cutaneous leishmaniasis due to L. tropica were described in North of Morocco, near Taza, at the junction of the eastern part of Rif mountain and the western part of Moyen-Atlas chain (Guessous-Idrissi et al. 1997b). One case of canine visceral leishmaniasis, also due to L. tropica, was described in this area, where T. articulata is present and signifies semi-arid bioclimatic belt (Guessous-Idrissi et al. 1997a). P. papatasi, proved as the vector of L. major (Rioux et al. 1986a), occurs in the arid and principally per-arid bioclimatic belts, the second one with desertic vegetation and Saharian flora (Atriplex spp., ...) (Fig. 6A). Moreover, Schlein and Warburg (1986) showed that laboratory-bred P. papatasi of both sexes pierce leaves or stems of the per-arid caper plant (Capparis spinosa). This per-arid area has been known for long-time as a focus of zoonotic cutaneous leishmaniasis due to L. major (Fig. 6B), from Guelmime to Foum Zguid (Rioux et al. 1986b) and in the Ouarzazate area (unpublished data). The common reservoir is constitued by rodents of the genus Meriones (Rioux et al. 1982b). In conclusion we can say that one essential feature of each focus of leishmaniasis is the vector which has requirements which determine its distribution. The eco-epidemiological multidimensional analyses are a very useful synthesis tool and show clear correlations between vector species, bioclimatic belts and phytosociological groupings. The bioclimatic and phytological requierements for each sand fly species can be easily transferred to usual maps and therefore, allow to delimit the geographical distribution of cutaneous and visceral leishmaniases in the Western Mediterranean basin. Thus, these analyses contribute to the definition of the structure of the leishmanial foci and to the determination, in a rational manner, of the high risk zones of these infections. The simultaneous presence of a known vector and a known reservoir for the same Leishmania can be considered as a predictional factor of possible occurence of outbreaks of leishmaniasis. Conversely, the verified absence of the vector, even if the reservoir is present in the area, is an absolute factor of security. The implementation of prevention and control are thereby made possible. ACKNOWLEDGMENTS To J-A Rioux who was the head of this laboratory and masterly responsible for the orientation of the eco-epidemiological surveys on leishmaniases from 1972 to 1984. The results presented in this paper have benefitted from the captures in natura and the identifications by many entomologists, particularly: J-C Beaucournu, A Chabaud, H Croset, J-P Dedet, A Delalbre-Belmonte, J and M Gallego, E Guilvard, R Houin, DT Jarry, R and M Killick-Kendrick, G Lanotte, N Léger, R Maazoun, M Maistre, A Martini, G Moreno, J Périères, M Portus, F Pratlong, R Teissier and ID Velez. To RW Ashford (Liverpool School of Tropical Medicine), M Roux (Faculté des Sciences, Marseille), and E Serres (Faculté de Médecine, Montpellier-Nîmes) for their valuable advice.
REFERENCES
Copyright 2002 Instituto Oswaldo Cruz - Fiocruz The following images related to this document are available:Photo images[oc02090f3.jpg] [oc02090f6.jpg] [oc02090f2.jpg] [oc02090e1.jpg] [oc02090f4.jpg] [oc02090t1.jpg] [oc02090f5.jpg] [oc02090f1.jpg] |
|