Visceral leishmaniasis
Up to 2010, 2784 serum samples were addressed to our laboratory for suspected VL cases. Anti-Leishmania antibodies were demonstrated in 944 sera. FAT and rK39 DST were positive in 944 (99.2%) and 822 (87.1%) cases respectively. In all positive patients, VL was confirmed by the demonstration of amastigotes in bone marrow and/ or cytocentrifuged blood and/or favourable outcome under antimonial treatment. In 8 confirmed VL cases, both serological tests were negative.
Thus, the total number of VL cases diagnosed over the 25 year period is 952. Three hundred and eight blood samples were investigated by the cytoconcentration technique. Among them, 201 were from confirmed VL patients, and amastigotes were demonstrated in 88 (43.7%) of them. In six cases, bone marrow examination was also negative. Apart from the confirmed VL FAT negative cases, all the remaining negative sera were from patients suffering from diseases other than VL, mainly leukaemia and various haematological disorders. Nine hundred and thirty four of the VL patients were children, aged 2 months to 14 years (median: 26.8 months). Eighteen were adults aged 19 to 39 years (median: 25.9 years); one of them was a HIV+ woman. The number of annually diagnosed cases ranged from 14 cases in 2004 to 83 cases in 2005. The highest proportion of VL cases (60%) was from Kairouan governorate, known to include the most active foci of VL in Tunisia, followed by Sousse governorate and, to a lesser extent from other areas in central Tunisia (Mahdia, Monastir, Kasserine).
All confirmed VL patients were treated with glucantime®, usually a single 21 to 28 day course. The outcome was favourable in the majority of them. Death occurred in 18 (1.9%) patients. All were children. In 4 of them, no response to antimonials was observed despite an adequate treatment regimen. They were assumed to be resistant to antimonials even though the resistance could not be assessed by in vitro testing of the strain. In an additional child, initially treated by three successive courses of glucantime®, healing was ultimately obtained with ambisome® (liposomal amphotericin B).
Typing of Leishmania strains
Over the 25 year study period 103 Leishmania strains were typed by either IEE or PCR or both. More than 150 isolates were in vitro cultured. Fifteen of them were from bone-marrow aspirates in VL cases. All were typed by IEE and found to be L. infantum. Most cultured isolates were from CL patients. Eighty one of them could be maintained in culture media, cryopreserved and typed by IEE. The remaining isolates were either lost after repeated subculture or contaminated during sampling or subculturing. It should be mentioned that strains addressed to the laboratory of Monastir, were typed and identified at the zymodeme level, whereas those typed in the LEEP were only identified at the species level. Sixty four isolates were typed by molecular techniques: 4 were typed by K-DNA-PCR and PCR-RFLP, and 60 by multiplex PCR.
Overall, 88 CL strains were typed by either IEE or PCR technique: 7 were L. infantum, 3 were L. killicki and 78 were L. major. All L. major strains that were typed at the zymodeme level were found to be L. major MON 25.
Discussion
One of the main characteristics of the four leishmaniasis forms occurring in Tunisia is their changing epidemiological pattern, which has continuously been described over the last 30 years. These changes mainly consisted of the spread of all forms of the disease from their initial endemic foci towards neighbouring or farther areas.
The VL which has long been confined to the northern parts of Tunisia showed a noticeable spread towards the central and southern parts of the country, where the disease has never been previously reported. The Kairouan governorate where VL was unknown till 1980, has progressively emerged as a highly endemic area and by the early 1990s, it became the most active focus in terms of incidence (Anonymous; Bel Hadj et al., 1996; Besbès et al., 1994; Bouratbine et al., 1998). These findings contrast with earlier periods where VL was mainly reported from Zaghouan, North-western and Tunis governorates and to a lesser extent from the central coastal region (Sousse) (Anderson et al., 1938; Ben Ismail et al., 1986; Ben Rachid et al., 1983; Chadli et al., 1968). In the same time, VL progressively spread further south to southern and central western governorates (Sfax, Sidi Bouzid, Kasserine, Tozeur) (Ayadi et al., 1991; Chargui et al., 2007). While spreading, the VL incidence markedly increased from a median of 12 to 68 annual recorded cases, 50 % of them being from Kairouan foci. Similar trends were described in SCL which, up to 1980, was very rarely reported south to the Tunisian ridge ("dorsale"). From the early 1980s, and up to date, SCL has continuously been reported from central and southern parts of the country (Aoun et al., 2000, 2008; Ben Saïd et al., 2006; Kallel et al., 2005, 2008a).
ZCL was first reported on the end of the 19th century causing large outbreaks in Gafsa and neighbouring areas. Later, it has nearly disappeared for half a centrury; but reappeared again in 1982 causing a widespread epidemic that is still going on and constitutes a major public health problem in extended areas of the country, with up to date more than 150000 cases (Anonymous; Ben Abda et al., 2009; Ben Ismail & Ben Rachid, 1989).
CCL, known to be confined to limited areas in the presaharian region of Tataouine, has itself shown spread towards areas far from its original endemic foci, up to Kairouan governorate (Aoun et al., 2008; Haouas et al., 2005).
The described changes are supposed to mainly originate in the agricultural development and the ecological transformations that occurred in the concerned areas and led to a marked increase in irrigated surfaces that in turn helped the reservoirs (mainly rodents) and vectors to abundantly proliferate, and created biotopes very suitable for the Leishmania cycles to establish and amplify (Aoun et al., 2008; Ben Abda et al., 2009; Ben Salah et al., 2007). In addition, urbanization of previously rural areas made the populations’ access to medical facilities easier, and consequently more and more leishmaniasis cases could be detected and diagnosed. The impact of climate change and variability on leishmaniasis (due to el Niño) was shown in Colombia and is associated to shifts in insect and animal distributions (Cardenas et al., 2008). This phenomenon was also shown to impact on the incidence of VL in Brazil and models were established to predict high risk years for VL (Franke et al., 2002). In Tunisia, the relationship between climate variability and leishmaniasis is not well studied. However, incidence rate of VL was shown to be positively correlated with mean yearly rainfall and continentality index; a rainy year is followed 2 years later by an increase in VL cases likely modulated by the intensity of transmission to dogs and by the influence on sandfly abundance (Ben-Ahmed et al., 2009). Distribution of sandflies classically associated to VL was also shown to be dependent on bioclimate (Zhioua et al., 2007), likewise for the distribution of Phlebotomus papatasi, the vector of ZCL (Chelbi et al., 2009). Knowledge on the epidemiological patterns and trends in leishmaniasis has much increased over the last 3 decades mainly because of a better identification of the Leishmania species involved in the natural cycles of the parasite. This knowledge greatly benefited of the availability of techniques used in typing isolates obtained from humans or from reservoirs and sandflies. In this respect, IEE or molecular techniques allowed a more precise identification of the Leishmania at the species, zymodeme and genomic (schizodeme) level (Ben Hammouda et al., 2000; Ben Ismail et al., 1992; Ben Said et al., 2006; Guerbouj et al., 2007; Guizani et al., 1993, 1994a, 1994b; Guizani et al., 2002; Hanafi et al., 2001; Kallel et al., 2008b).
As far as zymodemes are concerned, it was shown that L. major was much more homogeneous than L. infantum. Indeed, all L. major strains obtained from humans, rodents and sandflies proved to be identical to the L. major MON 25 reference strain (Aoun et al., 2008 ; Ben Ismail et al., 1986; Haouas et al., 2007 ; Kallel et al., 2005). In contrast, at least three L. infantum zymodemes occur in Tunisia : i) the MON 1, mainly causing VL and to a lesser extent SCL ; ii) the MON 24, the most common agent of SCL, that also causes VL ; iii) the MON 80, which is very rarely isolated (Aoun et al., 1999, 2008; Bel Hadj et al., 2000, 2002; Gramiccia et al., 1991; Haouas et al., 2007). These findings show that a single zymodeme may cause quite different diseases; and raise the as yet poorly documented question of pathogenesis and virulence of strains in cause. In addition, it was shown that the distribution of both MON 1 and MON 24 L. infantum zymodemes was different according to the geographical areas the patients originated from. Indeed, L. infantum MON 24 zymodeme was more frequently reported in VL cases from Kairouan governorate as compared to northern VL foci where MON 1 zymodeme is predominant. This was attributed to the emergent character of Kairouan foci which may favour and select atypical or rare variants (Aoun et al., 2008; Ben Abda et al., 2009). Involvement of different sandfly species in transmission of different L. infantum parasites can not be excluded (Guerbouj et al. 2007). It is well established that different sandfly species belonging to the sub-genus Larroussius are involved in the transmission of L. infantum; in Tunisia clear association between sandfly species and bioclimate was demonstrated and presence of species of this subgenus in the different bioclimatic zones allowed explaining the extension of VL (Zhioua et al., 2007).
The reservoir of L. killicki is still debated. Some findings argue for its zoonotic origin as the disease is hypoendemic and frequently reported in rural populations. In addition, L. killicki in Kenya was isolated from hyracoides (Sang et al., 1994). However, CCL has also been reported from urban areas and from regions where L. major ZCL is highly prevalent and additional CCL cases were probably misdiagnosed. So, the CCL may be more frequent than previously thought. On the other hand, it was hypothesized that competition between L. major and L. killicki may lead to some degree of pressure exerted by the first species with a subsequent reduction in the incidence of the second one (Aoun et al., 2008). However, parasite identification still remains circumstantial; parasite isolation from cutaneous lesions is often contaminated by overgrowing microorganisms. In this situation, isolating and typing of much more additional strains is highly needed, in order to get further insight in the knowledge of the concerned Leishmania cycles.
Even though IEE has been demonstrated to be quite efficient in identifying isolates and still constitutes a reference tool, molecular techniques would be more relevant in assessing variability of strains, as even minor genomic variations may have important consequences on the epidemiological and the clinical levels. An additional advantage of the molecular techniques is their rapidity in that culture of isolates is no more needed. On the other hand, they are much adapted for both diagnostic (PCR, Multiplex PCR, Real time- PCR) and epidemiological (PCR-RFLP, Multiplex PCR, RAPD) purposes (Aoun et al., 2008; Bel Hadj et al., 2002; Ben Ismail et al., 1992; Ben Said et al., 2006; Chargui et al., 2005; Guerbouj et al., 2007; Guizani et al., 1994a, 1994b; Guizani et al., 2002; Hanafi et al., 2001; Saadi et al., in preparation). In addition, molecular techniques are specific and highly sensitive, allowing identification of parasite species hard to isolate or to maintain in vitro in culture or too scarce to be detected or cultured. It has been supposed that the dog is the reservoir of L. infantum MON 24 zymodeme and P. perfiliewi the phlebotomine vector. However, isolation of this zymodeme from either the dog or the sandfly has been unsuccessful, making the hypothesis questionable (Aoun et al., 2008). In this situation, molecular techniques would be much more suitable for this purpose. Using DNA tools, P. langeroni was found infected with L. infantum in an active transmission focus of SCL (Guerbouj et al., 2007). It may be concluded from all these findings that identification and typing of isolates are much needed in order to best assess the epidemiology and to investigate the diversity and changing patterns of leishmaniasis.
The incidence of leishmaniasis as reflected by the annually reported cases at the national level together with our own findings was shown to highly vary and fluctuate over the 25 year period study. These fluctuations were attributed first to the dynamics of the reservoirs’ populations and sandflies, as a consequence of climatic variations, environmental and ecological changes occurring in the endemic areas (see section 2); and second, to the immunization, through the repeated outbreaks, of the infected human populations, that progressively makes previously infected people in the endemic area less receptive to the infection. This was best demonstrated in the ZCL form (Ben Salah et al., 2007).
In Tunisia, human activities, agricultural development and the subsequent ecological changes mainly benefited to rodents which dramatically proliferated in areas where their populations were previously too low for the Leishmania cycle to be completed. This can explain that ZCL is by far the predominant and the most widespread form of leishmaniasis in Tunisia, in terms of both incidence and geographical distribution. In addition, no sustained control programme for ZCL was carried out in most endemic areas, because of the difficulties that arose in the organization and the feasibility of such programmes. This made the ZCL epidemics still going on. It is worth mentioning however that control projects may be successful, as demonstrated by the results of those launched in 1992 in Sidi Bouzid area and again in the mid 2000s in Sidi Bouzid and Sidi El Heni, that led to an obvious decrease in the ZCL incidence over the mid 1990s in Sidi Bouzid and the last few years in both regions.
As compared to SCL, the spread of VL has been more important and obvious. This finding was related, among other factors, to the ecology of the phlebotomine vectors as it was shown that P. perniciosus (sub-genus Larroussius), the vector of MON 1 L. infantum zymodeme is more resistant to dry climate as compared to P. perfiliewi, the presumed vector of MON 24 zymodeme (Aoun et al., 2008 ; Ben Abda et al., 2009). Given the fact that presence and distribution of the different sandfly species generally associated to L. infantum transmission explained extension of VL (Zhioua et al., 2007) and that the extensions seem to concern in majority MON 24 parasites (Aoun et al., 2008; Ben Abda et al., 2009; Kallel et al., 2008a), explanations may well relate to parasite features not accounted for by the mere attribution to zymodemes and/ or to differential distribution of reservoirs yet to identify. Even though molecular techniques namely PCR and its variants have been increasingly used in the diagnosis of all forms of leishmaniasis, it is to be pointed out that conventional techniques should not be neglected. In VL, bone-marrow aspirate examination was shown to be sensitive enough, even though not optimal. Examination of cytoconcentrated peripheral blood was demonstrated to be positive in nearly 50 % of VL cases and was proposed to be carried out at first in order to avoid the painful bone-marrow puncture (Ben Said et al., 1998, Chemli et al., 2006). In serodiagnosis, FAT was found to be very efficient and suitable as its sensitivity exceeded 95 % and rK39 DST, even though less sensitive, very useful because it is highly specific, easy to perform and adapted to epidemiological investigations (Saghrouni et al., 2009). PCR is much more adapted to the diagnosis of the asymptomatic and subclinical forms of the disease because of its sensitivity and the scarcity of parasites in such situation. In CL, parasitological diagnosis by demonstrating amastigotes in dermal specimens is again very adequate, even though PCR was demonstrated to be more sensitive (Chargui et al., 2005). In contrast, serodiagnosis is much less useful in CL diagnosis as antibodies’ amounts are often very low and results difficult to interpret. One of the main objectives of this study was to assess whether the activity of our laboratory reflects the actual situation of the leishmaniases at the country level, and whether it could constitute an indicator of public health relevance. As far as VL is considered, most patients originating from Kairouan and Sousse governorates, where are located the most active foci, were investigated for serodiagnosis in our laboratory. So, we may conclude that our findings reflect the actual incidence of VL in the region. This is further confirmed by the comparison of our results to the statistics on the annually recorded cases of VL and edited by the primary health care direction of the Tunisian Ministry of Health (Anonymous). Obviously, our findings match these reports as highest number of registered cases at the national level was in 1992 (n = 130), 1993 (n = 122) and 2005 (n = 122); and highest number of cases diagnosed in our laboratory was as follows: 74 cases in 1992, 72 cases in 1993 and 83 cases in 2005. This is not surprising as nearly 50 % or more of VL cases originate from Kairouan governorate.
Concerning CL, we have to point out that first SCL cases from central Tunisia were diagnosed in our laboratory in patients with clinical presentation found to be evocative of this form, and amastigotes, very suggestive of L. infantum as compared to L. major because of their small size (Ben Saïd et al., 2006). As mentioned in section 2, additional cases were further reported and the spread of SCL towards central and southern Tunisia confirmed. Similarly, the first CCL case originating from outside the original foci of Tataouine area was again diagnosed in our laboratory (Haouas et al., 2005); and CCL spread was further confirmed by additional reports leading to the revised geographical distribution of the disease (Bouratbine et al., 2005; Aoun et al., 2008). In our laboratory, we used to diagnose ZCL since 1986. Up to the late 1990s, the number of diagnosed cases was too moderate as compared to that of ZCL reported cases in the endemic regions (Figures 7 & 9). This, for at least, the two following reasons: i) at the beginning of the outbreak, many patients suffering from ZCL were only clinically diagnosed, so that only a few of them were addressed to the laboratory for parasitological confirmation, ii) from the 1990s, a great number of families originating from ZCL endemic areas migrated and settled in Sousse city and suburbs, where they progressively constituted a large community. Most of them used to return back to their home of origin for summer holidays where, because continuously exposed to phlebotomine bites, many of them contract leishmaniasis which is later addressed to us for diagnosis. It is worth mentioning that a high proportion of this community originates from Sidi Khelif in Ouled Haffouz delegation, Sidi Bouzid; and and many of them were contaminated there and were later diagnosed in our laboratory. From 2000 and onwards, the number of ZCL patients who attended our laboratory for diagnosis dramatically increased. We consider that since this date the activity of our laboratory indirectly but adequately reflects the actual status of ZCL in central Tunisia, namely in Sidi Bouzid, Mahdia, Kairouan and Sousse governorates. This is best illustrated by our findings in 2004 where > 400 ZCL cases were diagnosed in our laboratory. In the same year, the highest incidence of ZCL (> 15000 cases) was registered at the national level, because of the three epidemic peaks that occurred in Sidi Bouzid (Sidi Bouzid city, Regueb, Menzel Bouzayène), Gafsa (Gafsa city, Sned, Mdhilla) and Kairouan (Nasrallah, Hajeb Layoun, Chrarda) governorates (Anonymous). The peak we registered in 2001 was due to an outbreak that arose in Chahda, a small locality in Chorbène delegation (Mahdia governorate). Indeed, more than 50 % of cases we diagnosed in 2001 were from this locality. A similar peak in Sidi El Heni delegation (Sousse governorate) was registered in 2006, with 107 cases. This peak was followed by a noticeable decrease as soon as control activities were carried out in the area. Similarly the decrease in ZCL incidence shown in our study over the last 5 years can be attributed to an important decrease in the transmission in Sidi Bouzid region where control activities were launched in the mid 2000s (Figure 9). This decrease was reflected on the national level over the same period (Figure 7). Last, our results confirm the seasonal occurrence of ZCL.
Our findings and results show that the activity of our laboratory reflects the situation of leishmaniasis at the regional level and at least in part at the national level, and may be an indicator of public health relevance. It may thus contribute to early alert of health authorities on changing epidemiological trends or emergence of leishmaniasis in the region. However, our activities may not reflect the actual situation and epidemiological changes occurring in farther endemic areas like Gafsa, Tozeur, Kébili, Médenine and Gabès known to include highly active foci because very few patients from these regions seek medical facilities in Sousse hospital. Work is in progress to identify and characterize the Leishmania parasites actually circulating in central Tunisia using molecular DNA techniques.
