Background

Visceral leishmaniasis (VL, Kala-azar) is a neglected tropical disease that can be fatal without early diagnosis and proper treatment. VL in east Africa is caused by the Leishmania donovani species complex. The transmission of L. donovani is generally considered anthroponomic. Jambulingam et al. [1] provided definitive evidence incriminating dogs as a L. donovani reservoir in India, but the status in east Africa remains to be substantiated. However, there are reports that associated dogs with L. donovani transmission in the Sudan [2, 3] and Ethiopia [4,5,6]. Also, studies have shown that dogs are among the domestic animals that P. orientalis, the vector of L. donovani in east northern Ethiopia foci, preferentially bites [7, 8].

Visceral leishmaniasis (VL) is reemerging with geographic spread and recurrent outbreaks that have claimed the lives of several hundred Ethiopians over the past 2 decades [9]. Among the factors contributing to its spread and outbreaks are individual, household and socio-geographic risk factors: age, sex, housing conditions, mass movement of temporary laborers, immunosuppression and ecological modifications [9,10,11,12]. The national incidence estimate for Ethiopia based on self-reported cases is up to 4500 new cases per year [13]. The risk model using the geographical information system and statistics showed that about 33% of the total landmass, predominantly within development corridors with significant public health and economic implications, is at high risk for VL [14].

Benishangul-Gumuz regional state is one of the fastest changing development corridors in western Ethiopia. Mega-projects, such as the Great Ethiopian Renaissance Dam, large-scale irrigation and rain-fed commercial agriculture areas and mining activities, have resulted in vast ecological and sociodemographic changes. The rapid assessment by Abera et al. [15] reported a 7.3% (20/275) VL asymptomatic infection prevalence in two kebeles (sub-districts) where a VL patient was reported to have lived [16].

Control strategies for leishmaniasis in Ethiopia rely on case detection and treatment and on vector control. Thus, delineating endemic areas and knowledge about the burden help to attain desired outcomes by targeting resources. Also, knowledge on risk factors associated with exposure is important to design behavioral change communication tools to attain active participation and ownership of programs by affected communities. Therefore, the objective of this study was to assess the epidemiology and risk factors associated with leishmaniasis in humans and dogs in high-risk districts.

Materials and methods

Description of study area

The location of the Benishangul-Gumuz region, western Ethiopia, is 34°10’N, 37°40’E and 09°17’N, 12°06’N. The region is predominantly (75%) lowlands. The total population is around 784,345 with an estimated density of 15.91 people per square kilometer (BGRoHB 2019). The study encompassed six areas at high risk of VL (Fig. 1) as per the environmental factor-based risk model by Tsegaw et al. [14]: Dangur, Guba and Pawi from the Metekel Zone and Banbasi, Kumruk and Sherkole from the Assosa Zone. The region is one of the development corridors with large-scale agricultural, mining and dam projects, which have changed the settlement pattern and caused deforestations and a large influx of people for temporary work and/or permanent settlement.

Fig. 1
figure 1

Map of the study woredas: 1 = Bambasi, 2 = Kurmuk and 3 = Sherkole from the Assosa Zone; 4 = Guba, 5 = Dangur and 6 = Pawi from the Metekel Zone, Benishangul Gumuz regional state, Western Ethiopia

Full size image

Study design and sample size determination

A cross-sectional survey was done from 2018 to 2020 to assess the epidemiology and explore whether there was any zoonotic significance of VL and risk factors associated with exposure to Leishmania infection. Samples were selected using a multi-staged sampling technique. As the primary sampling unit among the three administrative zones of the region, overlaying the environmental factor-based risk map [14], two zones, namely Assosa and Metekel, were selected because they had large areas at high risk of VL. Similarly, within the selected zones, districts with high-risk areas were selected. Then, an operation map was prepared overlaying the risk map of selected districts and the kebele level shapefile to identify high-risk kebeles. Subsequently, study households were randomly selected from each of the kebeles targeting up to 5% of their total population, with overall sample size of 1342 individuals for LST testing.

Following LST-reactive individuals as a focal point, dogs were sampled for serological tests, rK39 and DAT. Also, 253 human blood samples, 67 purposively from LST-reactive and 185 randomly from non-reactive individuals, were tested by DAT and rK39.

After explaining the purpose of the study, a written informed consent form was obtained from each participant or the parents or guardians for minors. Similarly, informed assent was obtained from dog owners to sample dogs. Blood sample were aseptically collected using 5-ml disposable syringes or plain vacutainer tubes from cephalic/saphenous veins of both humans and dogs. Of the collected blood, 20 µl was used to prepare dried blood spots (DBS) on 3MM Whatman paper (Whatman, Maidstone, UK) allowed to fully air dry without exposing to direct sunlight. Sera from both dog and human were used for rK39 ICT and DAT testing.

Leishmanin skin test (LST)

Prior to LST, socio-demographic information was documented from the study participants using a pre-tested semi-structured questionnaire. Then, an intradermal injection of 0.1 ml leishmanin antigen (Pasteur Institute of Iran, Tehran, prepared from L. major) was made at the volar surface of the arm. After 48–72 h, the delayed hypersensitivity reaction was measured with the ballpoint techniques; 5.0 mm and above of the average of the two diameters of an induration was considered positive (Fig. 2).

Fig. 2
figure 2

Leishmanin skin test (LST) procedure. Photos are from this field work on the same participant: a intradermal injection of 0.1 ml of LST (Pasteur Institute of Iran, Tehran, Iran) solution after brief shaking. b Marking of the injection point using permanent marker and c measuring the induration using the ballpoint pen method after 48 to 72 h of injection

Full size image

rk-39 immunochromatographic test (rK39 ICT)

The rK39 ICT (DiaMed- ITLEISH; Bio-Rad Laboratories, Marnes-la-Coquette, France) was done following the supplier’s recommendations. In brief, a 20-μl serum sample was added to the absorbent pad well with 150 μl (2–3 drops) of the chase buffer provided with the kit. Results were read after 10–20 min and recorded as follows: positive when both control and test lines appeared; negative when only control line appeared or invalid when no control line appeared (in such cases tests were repeated) (Fig. 3).

Fig. 3
figure 3

rK39 immunochromatographic test interpretation: top two: positive strips: bottom: negative strip

Full size image

Direct Agglutination Test (DAT)

Sera were transported to Benishangul Gumuz Regional Laboratory in an ice box and stored at −20 ℃. Then, samples were transported to AHRI under cold chain and stored at −20 ℃ until processed. A direct agglutination test was performed according to the manufacturer’s instructions (Institute of Tropical Medicine, Antwerp, Belgium). The presence of antileishmanial antibodies below or at cutoff of 1:3200 titers was used to determine negativity. Both negative and positive controls were run for every batch of kit used. In brief, sera were diluted serially from 1:200 to 1:204800 by transferring 50 μl of diluted serum and discarding the same amount from the last dilution (Fig 4).

Fig. 4
figure 4

A plate showing the DAT test results in sera tested with a starting dilution of 1:200 in column

Full size image

Data analysis

STATA version 13 data software (College Station, TX, USA) was used for data analysis. Descriptive statistics were employed to summarize in terms of frequencies and percentages. Univariate and multivariate logistic regressions were used to determine the association of Leishmania infection with the risk factors and expressed as odds ratio and 95% confidence interval. For all analyses, P < 0.05 was considered a significant difference.

Ethical considerations

The protocol was approved by the AHRI/ALERT ethical review committee (AH01275/0012/18, 19/12/18). Informed consent was obtained from all participants or guardians/parents of minors. For participants between 11 and 18 years, verbal assent was sought in addition to the parental/guardian consent. Similarly, for dogs informed consent was obtained from the owners.

Result

Prevalence of asymptomatic visceral leishmaniasis

Of the total 1342 participants LST tested, 89.2% (801 males and 396 females) were available for result reading. The LST-based prevalence was 6.0% (72/1197). The seroprevalence was 3.2% (8/253) and 5.9% (15/253), respectively, by rk39 and DAT. Three of the 8 rk39 and 7 of the 15 DAT-reactive individuals were LST positive, while 5 of the rk39 and 8 of those DAT reactive were LST negative (Table 1).

Table 1 Prevalence of asymptomatic visceral leishmaniasis in Benishangul Gumuz, western Ethiopia, by age and sex as measured by LST (n = tested positive, N = 1197), DAT (n = tested positive, N = 253) and rK39 (n = tested positive, N = 253), 2018–2020
Full size table

Variation in asymptomatic infection was observed among sites: higher prevalence was detected in kebeles from Guba (16.1 %, 32/199) and Kurumuk (14.0 %, 27/191) districts, respectively. The highest LST positivity was recorded at Abulhorse kebele (28.3 %, 17/60) from Guba district (Table 2).

Table 2 Kebele-level asymptomatic visceral leishmaniasis as measured by the leishmanin skin test (LST), Benishangul Gumuz, western Ethiopia
Full size table

Seroprevalence of asymptomatic visceral leishmaniasis in dogs

Of the 36 dogs owned by households that had LST-reactive member(s), 5 (13.9%) and 2 (5.6 %) were reactive according to rk39 and DAT, respectively (Table 3). The trend in distribution of the seroprevalence in dogs paralleled that observed in humans; more positive dogs were found in sites where there were more LST-positive humans (Tables 2 and 3).

Table 3 Seropositivity of dogs (N = 36) owned by households with LST-positive member(s), Benishangul Gumuz, Western Ethiopia, 2018–2020
Full size table

Factors associated with asymptomatic Leishmania infection

Exposure to Leishmania infection showed a significant gender difference. Females were about 0.4 times less likely to be affected compared to males (AOR = 0.38; 95% CI 0.20, 0.72). Yet age showed no significant association with VL exposure. Presence of a dog in a household was found to increase the likelihood of being LST positive by 2.6-fold (AOR = 2.60; 95% CI 1.54, 4.40). Living in Kurmuk district (AOR 5.85, 95% CI 2.27, 15.09) had the highest risk followed by Guba district (AOR 4.74, 95% CI 1.83, 12.31) (Table 4).

Table 4 Risk factors associated with asymptomatic visceral leishmaniasis as measured by the leishmanin skin test, Benishangul-Gumuz, western Ethiopia, 2018–2020
Full size table

Discussion

Benishangul-Gumuz is one of the crucial development corridors in western Ethiopia. Accompanying the large-scale projects are huge sociodemographic and ecological changes. Large areas in the region were predicted to have a high VL risk based on the environmental factor-based geographical information and statics risk mapping [14]. However, data hardly exist on the epidemiology of VL in the region [15]. To our knowledge, this epidemiological survey is the first to assess the asymptomatic Leishmania infection rate covering wider high-risk districts in the regions with humans and dogs.

The prevalence of Leishmania infection was 6.0% based on LST positivity. The seroprevalence in human Leishmania infection was 3.2% (8/252) by rk39 and 5.9% (15/252) by DAT. The LST positivity rate of 5.4% in our study is in agreement with the previously reported Leishmania infection prevalence reported by Hailu et al. [4] from Aba Roba, southern Ethiopia, and Bsrat et al. [17] from Welkait, northern Ethiopia, of 5.6% and 5.88%, respectively. It is lower than the results reported by Ali et al. [19] from the lower Awash valley, eastern Ethiopia, and Tadese et al. [20] from Raya Azebo, northeastern Ethiopia, who reported reactive rates of 38.3% and 9.08%, respectively. The seroprevalence in Benishangul-Gumuz according to rK39 (3.2%) was lower than that in the report by Alebie et al. [21] from the Gode and Adale districts of the Shebele Zone (12.7%), southeastern Ethiopia. The DAT positivity rate was in agreement with that of Hailu et al. [18], who found 5.4%, and higher than that of Tadese et al. [19], who reported 0.8%. The difference in prevalence is expected as the risk factors or level of risk factors for exposure to sand fly bites differ in different at-risk communities.

Understanding the determinants of VL exposure in an area is important information for designing infection prevention methods. Thus, we examined personal and household factors connected with L. donovani infections. The significant differences in exposure between males and females observed in the present study were supported by Ali et al. [20], Hailu et al. [18] and Bantie et al. [20]. This could be because males are mostly engaged in outdoor activities and stay outdoors, which might increase their chances to have sand fly bites.

In the current study participants who owned dogs had a 2.6 (95% CI 1.54, 4.40) times higher chance of being LST positive, a finding that paralleled the report by Bsrat et al. [4], but the seroprevalence in dogs was higher in the current study. This difference might because we purposely sampled dogs owned by households with LST-positive members and also the small sample size, as our aim was not to determine prevalence in dogs but to generate lead data on whether dogs are implicated in the transmission.

The relatively lower Leishmania infection prevalence detected in this study could indicate the probability that VL is a recent (re)emergence in/introduction to the Benishangul Gumuz region. The lack of significant exposure risk difference between age groups corroborates our argument of VL being a recent phenomenon in the Benishangul Gumuz Region. Furthermore, there was a relatively higher prevalence in districts such as Guba where high rates of socioecological modifications have taken place within the Benishangul Gumuz and Metekel area, which shares borders with the high VL burdened foci in Amhara. Thus, the recent move to previously uninhabited areas and/or influx of people, mostly project employees (civil servants), from all corners to the region might have precipitated the transmission in the community. Lack of sand fly data and limited purposive sampling of dogs hampered reaching a conclusion as to their contribution to the transmission. The risk factors captured in the study were not exhaustive enough because of the lack of experience with what is happening on the ground, which also limited our understanding and ability to make recommendations.

Conclusions

It is noteworthy that our approach to risk modeling to targeted surveillance in areas hitherto not known to be VL endemic proved to be useful. We showed the presence of active VL transmission in a key developmental corridor, Benishangul-Gumuz. We recommend the that the regional health bureau and responsible stakeholders be vigilant and plan early containment measures to avoid possible public health and economic consequences due to VL.