Trypanosome Infection in Dromedary Camels in Eastern Ethiopia: Prevalence, Relative Performance of Diagnostic Tools and Host Related Risk Factors

Veterinary Parasitology 211 (2015) 175–181

Contents lists available at ScienceDirect

Veterinary Parasitology

journal homepage: www.elsevier.com/locate/vetpar

Trypanosome infection in dromedary camels in Eastern Ethiopia: Prevalence, relative performance of diagnostic tools and host related risk factors

Regassa Fikru a,b,c,∗, Yimer Andualem d, Terefe Getachew a, Joris Menten e, Epco Hasker f, Bekana Merga a, Bruno Maria Goddeeris c, Philippe Büscher b a College of Veterinary Medicine and Agriculture, Addis Ababa University, PO Box 34, Debre Zeit, Ethiopia b Institute of Tropical Medicine, Department of Biomedical Sciences, Nationalestraat 155, B-2000 Antwerp, Belgium c KU Leuven, Faculty of Bioscience Engineering, Department Biosystems, Kasteelpark Arenberg 30, B-3001 Leuven, Belgium d School of Veterinary Medicine, Wollo University, PO Box 1145, Dessie, Ethiopia e Institute of Tropical Medicine, Department of Clinical Sciences, Nationalestraat 155, B-2000 Antwerp, Belgium f Institute of Tropical Medicine, Department of Public Health, Nationalestraat 155, B-2000 Antwerp, Belgium article info abstract

Article history: A cross-sectional study was conducted in Chifra and Dewe districts of Afar region, Eastern Ethiopia, Received 29 March 2014 to determine the prevalence, agreement between diagnostic tests and host related risk factors of try- Received in revised form 27 March 2015 panosome infection in camel. An overall prevalence of 2%, 24.1%, 21.3%, 9.5% and 7.8% was recorded with Accepted 2 April 2015 respectively Giemsa stained thin blood smear, CATT/T. evansi, RoTat1.2 PCR, 18S PCR and ITS-1PCR in a cohort of 399 animals. Only one T. vivax infection was confirmed by TvPRAC PCR indicating T. evansi as Keywords: the predominant species affecting camels in the study area. No single animal was positive when tested Dromedary camel with T. evansi type B specific EVAB PCR. There was slight agreement between the CATT/T. evansi and Ethiopia Prevalence the molecular tests. Among the PCR methods, RoTat 1.2 PCR yielded a significantly higher positivity rate Risk factor compared to 18S PCR and ITS-1 PCR. There was no significant difference in the positivity rate observed Trypanosomosis in each gender of camels (p > 0.05). The positivity rate was significantly higher in camels with poor body Trypanosoma evansi condition and in older animals when tested using the CATT/T.evansi or RoTat 1.2 PCR (p > 0.05). Camels Trypanosoma vivax that tested positive with all tests had significantly lower PCV’s (p<0.05). This study provides further evidence that T. evansi is endemic in the Afar region of Ethiopia. The latent class analysis indicated an estimate overall prevalence of 19% (95% CI: 13–28). Moreover, the model indicated low sensitivity of CATT/T. evansi (43%) and the PCR tests (39–53%) but higher specificity of the PCR tests (86–99%) and low specificity of CATT/T. evansi (80%). This study suggests that improved sensitivity and reliability of the tests would help diagnosis of trypanosomosis. Further studies are required to determine the prevalence of clinical disease and losses due to trypanosomosis. © 2015 Elsevier B.V. All rights reserved.

1. Introduction surra (Röttcher et al., 1987). T. evansi is mechanically transmitted by blood sucking flies, such as Tabanidae and Stomoxys sp (Hoare, Trypanosomosis is one of the major health problems with high 1972). A number of researchers have investigated the epidemiol- morbidity and mortality in camels in Ethiopia (Demeke, 1998; ogy of camel trypanosome infection in different parts of Ethiopia. Tekle and Abebe, 2001). The disease is caused by infection with These studies were mainly based on parasitological and serolog- hemoflagellated parasites belonging to the genus Trypanosoma. ical tests (Richard, 1979; Elias, 2003; Hailemariam et al., 2008; Trypanosoma evansi belongs to Trypanozoon subgenus and is the Hagos et al., 2009; Kassa et al., 2011; Tadesse et al., 2012). Para- most commonly reported cause of camel trypanosomosis called sitological examination suffers from limited sensitivity even when using the haematocrit centrifugation technique. Serological tests are unable to distinguish past and current infections as the antibod- ∗ ies persist in the circulation (Büscher, 2014). The serological tests Corresponding author at: Addis Ababa University, College of Veterinary Medicine and Agriculture, Debre Zeit, PO Box 34, Debre Zeit, Ethiopia. Tel.: +251 911907056. also are not fully specific due to the possibility of cross reactions E-mail address: fi[email protected] (R. Fikru). with antibodies produced against other infections. For example http://dx.doi.org/10.1016/j.vetpar.2015.04.008 0304-4017/© 2015 Elsevier B.V. All rights reserved. 176 R. Fikru et al. / Veterinary Parasitology 211 (2015) 175–181

Desquesnes et al. described cross reactivity of T. evansi antigen with season (kerma) taking place between Mid-June and Mid-September anti T. cruzi antibodies (Desquesnes et al., 2007). Control of camel and the short rainy season (sugum) occurring between March and trypanosomosis depends mainly on the use of curative and prophy- April. The average annual rainfall is between 400 mm and 600 mm. lactic drugs and their application ideally should take into account actual information of the disease. However to properly understand 2.2. Study animals and design the epidemiology of a disease, up-to-date and high-quality data are ≥ required through application of accurate diagnostic tools (Ali et al., Dromedary camels 6 months old were considered for blood 2009). In Ethiopia, a number of studies indicated that T. evansi is the sampling. Age, gender, and previous treatment history were sole pathogenic trypanosome infecting camels (Hailemariam et al., recorded. Body condition was assessed according to Faye et al. 2008; Kassa et al., 2011; Tadesse et al., 2012). These findings need (2001). Animal age groups were defined as follows: 6 months to also to be verified with more accurate diagnostic techniques like 1 year, >1–5 years, >5–10 years and >10 years. PCR to assess the possibility of infection with other mechanically A cross-sectional study was conducted on 399 camels (199 transmitted trypanosomes like T. vivax (Birhanu et al., 2013). There- camels from Chifra and 200 camels from Dewe) from November fore, this study was undertaken to assess the prevalence of camel 2011 to April 2012. A combination of convenience, purposive and trypanosome infection in the Afar region using different diagnostic multistage stratified sampling methods was applied according to techniques, to assess the relative performance of different diag- Toma et al. (1999). First, the two study districts were selected based nostic PCRs and to identify host related risk factors associated with on their camel population and reported cases of camel trypanoso- trypanosome infection. mosis. A complete list (sampling frame) of the pastoral villages was obtained from which five pastoral associations (PAs) within each district were selected based on their accessibility. From these 2. Materials and methods PAs, three to five herds were selected from volunteer herders. Pas- toralists keep their camel herd in groups based on age, gender 2.1. Study area and reproductive status (lactating) in separate stables which were considered as stratum. However, some dominant lactating females The present study was conducted in Chifra and Dewe districts of could not be caught. An average of 10 camels (8–12) was randomly Afar Region, located in the North-Eastern part of Ethiopia between selected from each herd depending on the number of camels avail- 39◦34 and 42◦28E longitude and 8◦49 and 14◦30N latitude in able from each stratum. The samples size required to estimate the the rift valley (Fig. 1). The districts have similar arid and semi-arid prevalence of infection in the camel population was 384 based on agro-ecology, where livestock production is the main occupation of the formula given by Thrusfield with a design prevalence of 50%, the community. The average elevation of the districts is 802–825 m a precision of 5% and a 95% confidence interval (Thrusfield, 2005). above sea level with annual temperature ranging from 25 ◦Cto The sample size was increased to 399 to provide an allowance for 33 ◦C. The rainfall is bimodal with erratic distribution, the long rainy sample loss.

Fig. 1. Map of Ethiopia with the two districts (green) Chifra and Dewe in the Afar Region (orange border). (For interpretation of the color information in this ﬁgure legend, the reader is referred to the web version of the article.) R. Fikru et al. / Veterinary Parasitology 211 (2015) 175–181 177

Table 1 Primer sequences, target sequences and expected amplicon size of the different PCRs used to test blood from 399 camels from Eastern Ethiopia.

Target group Target sequence Primers Primer sequence Amplicon (bp) Reference

Trypanosomatidae 18S rRNA 18S F 5-CGCCAAGCTAATACATGAACCAA-3 110 (Deborggraeve et al., 2006) 18S R 5-TAATTTCATTCATTCGCTGGACG-3 Trypanozoon ITS-1 rRNA ITS-1 F 5-TGTAGGTGAACCTGCAGCTGGATC-3 450 (Desquesnes et al., 2001) ITS-1 R 5-CCAAGTCATCCATCGCGACACGTT-3 T. evansi type A RoTat1.2 RoTat1.2F 5-GCGGGGTGTTTAAAGCAATA-3 205 (Claes et al., 2004) RoTat1.2R 5-ATTAGTGCTGCGTGTGTTCG-3 T. evansi type B minicircle EVAB-1 5-ACAGTCCGAGAGATAGAG-3 436 (Njiru et al., 2006) EVAB-2 5-CTGTACTCTACATCTACCTC-3 T. vivax Proline racemase TvPRAC F 5-CGCAAGTGGACCGTTCGCCT-3 239 (Fikru et al., 2014) TvPRAC-R 5-ACGCGGGGCGAACAGAAGTG-3

2.3. Specimen collection and diagnostic test 2.5. Data analysis

Blood samples were collected from the jugular vein into 10 ml Data were analysed using SPSS (Statistical Package for Social heparinised vacutainer tubes. From the collected blood, thin smears Sciences, version 20) and STATA (Stata Statistical Software: Release were prepared, air dried, fixed with methanol and stained with 12. StataCorp LP). McNemar chi-square (2) test and odds ratios Giemsa for parasitological examination. Packed cell volume (PCV), (OR) were calculated to compare the prevalence of trypanosome expressed in percentage, was estimated after 5 min centrifugation infection in the two study areas. The level of agreement between at 12,000 rpm of blood in two capillary tubes. diagnostic tests was determined using Cohen’s kappa coefficient About 250 ␮l of blood were preserved in an equal volume of interpreted following Landis and Koch (Cohen, 1960; Landis and AS-1 buffer (QIAgen) and stored at ambient temperature for DNA Koch, 1977). ANOVA was used to assess the difference in mean PCV extraction. The remaining blood was centrifuged and plasma was of parasitemic and aparasitemic camels. 95% binomial confidence collected and preserved at −20 ◦C until the antibody detection was interval and p < 0.05 was set to decide on statistical significance. performed with the Card Agglutination Test for Trypanosomiasis For an exploratory assessment of the diagnostic accuracy of the (CATT/T. evansi). CATT/T. evansi was performed as per the man- different tests, allowing for the imperfect nature of all diagnos- ufacturer‘s instructions (Institute of Tropical Medicine, Antwerp, tic methods used, we performed latent class analysis (LCA). This Belgium). Briefly, 25 ␮l of camel plasma, diluted 1:4 in CATT-buffer, model-based approach allows an unbiased estimation of the sensi- were dispensed onto a reaction zone of a plastic test card. After tivity and specificity of diagnostic methods when a gold standard adding one drop (about 45 ␮l) of CATT reagent, the reaction mix- test that is 100% sensitive and 100% specific is unavailable. We ture was spread by a stirring rod and allowed to react on a CATT performed LCA using WinBUGS (Stata Statistical Software: Release rotator for 5 min at 70 rpm. A specimen was considered positive 12. StataCorp LP) and R, assessing several models allowing for when blue agglutinates were visible (Bajyana Songa and Hamers, dependence between test results within infected and non-infected 1988; Verloo et al., 2000). subjects, as well as the standard conditional independence model (Menten et al., 2008; R Core Team, 2014). The best fitting models presumed 100% specificity of the thin blood smear Giemsa staining technique and allowed false negative test results on the 3 PCR tests 2.4. Molecular analysis to be correlated. This model showed a good fit to the data (Bayesian lack-of-fit p-value = 0.238). For the molecular analysis, DNA was extracted with the QIAamp mini blood kit (Qiagen) from 200 ␮l blood/AS1 buffer into 200 ␮l elution buffer according to the manufacturer’s instructions as 3. Results described elsewhere (Fikru et al., 2012). Extracted DNA was stored ◦ at −20 C until tested by the respective PCRs. The PCRs with their The observed prevalences of trypanosome infection in camel target sequence, primer sequences and amplicon lengths are repre- in both districts, estimated using the diagnostic tests are given in sented in Table 1. Interpretation of the results after electrophoresis Table 2. in 2% agarose gels and staining with ethidium bromide (EtBr), is The overall prevalence of camel trypanosome infection was sig- based on the characteristic amplicon lengths. The different PCRs nificantly (p < 0.05) lower in Giemsa stained thin smear (Table 2) were conducted on all the samples to assess their performance than in the other tests. The serological test, CATT/T. evansi and the to detect and identify the different trypanosome species affecting molecular test, RoTat1.2 PCR (T. evansi type A-specific), revealed a camels. significantly higher prevalence than ITS-1 and 18S PCRs (p < 0.05). Furthermore, comparison of the analytical sensitivity of the A number of specimens were shown to contain T. vivax based on RoTat 1.2 PCR, 18S PCR and ITS-1 PCR was carried out on purified the results of the ITS-1 PCR. The parasites observed by thin blood DNA of two T. evansi isolates, from Kazakhstan and Morocco. The smear examination had morphology characteristic of T. evansi with DNA was tested as fivefold dilution series in water, ranging from a small subterminal kinetoplast at the posterior end and a flagel- 1 ng/␮l down to 0.064 pg/␮l. lum (Hoare, 1972). With the T. vivax proline racemase PCR, only

Table 2 Observed prevalence in % and with 95% conﬁdence interval of T. evansi infection in camels in two districts of Eastern Ethiopia using different diagnostic tests.

District Thin blood smear CATT/T. evansi RoTat1.2 PCR 18S PCR ITS-1 PCR Total

Chifra 2 (0.06–3.94) 20.6 (15–26.2) 21.1(15.3–26.8) 8 (4.2–11.8) 8 (4.2–11.8) 199 Dewe 2(0.05–3.95) 27.5(21.3–33.7) 21.5 (15.8–27.2) 11 (6.7–15.3) 7.5 (3.9–11.2) 200 Total 2(0.6–3.4) 24.1(19.9–28.3) 21.3 (17.3–25.3) 9.5 (6.6–12.4) 7.8 (5.2–10.4) 399 178 R. Fikru et al. / Veterinary Parasitology 211 (2015) 175–181

Table 3 Frequency distribution (N) and probability of being infected for each outcome pattern observed in ﬁve diagnostic tests for T. evansi infection performed on 399 camels from Eastern Ethiopia, as estimated with latent class analysis.

CATT/T. evansi PCR Thin blood smear N Probability infected

RoTat 1.2 18S ITS1

+ + + + + 2 100 ++++− 6 100 ++−− + 1 100 + −−−+ 4 100 − +++− 18 100 −−−−+ 1 100 + − ++ − 1 96.8 +++−− 3 94.1 −−++ − 1 92.2 − ++−− 4 85.3 ++− + − 1 84.7 + − + −− 1 60.5 −−+ −− 2 38.2 −−−+ − 2 19.6 ++−− − 13 18.4 + −−−− 64 17.8 − + −− − 37 7.2 −−−−− 238 6.7 one T. vivax infection could be confirmed. No single T. evansi type B Table 4 infection was detected with EVAB PCR. No statistically significant An exploratory assessment of the diagnostic accuracy of the different diagnostic tests performed on 399 camels from Eastern Ethiopia, estimated by latent class difference in prevalence of trypanosome infection was recorded analysis with a prevalence of 19.4% (95% CI: 12.8–28.0). between the study districts by any of the diagnostic tools. Out of 96 CATT/T. evansi positives samples, respectively 26, 13 Tests Sensitivity (95% CI) Specificity (95% CI) and 10 samples were also positive in RoTat 1.2 PCR, 18S PCR and CATT 42.5 (32.4–52.9) 80.4 (77.8–84.6) ITS-1 PCR (Table 3). On the other hand, out of 303 CATT/T. evansi PCR RoTat1.2 52.9 (39.3–67.9) 85.9 (84.1–88.1) PCR 18S 47.7 (33.0–64.2) 99.2 (98.0–100) negative samples, respectively 59, 25 and 21 samples were pos- PCR ITS-1 39.1 (26.4–52.8) 99.4 (98.8–100) itive for RoTat1.2 PCR, 18S PCR and ITS-1 PCR. There was slight Giemsa stained thin smear 10.8 (7.3–14.6) 100 (fixed) agreement between CATT/T. evansi and 18S PCR (K = 0.07), CATT/T. evansi and RoTat1.2 PCR (K = 0.08), and CATT/T. evansi and ITS-1 PCR (K = 0.05). Out of the 85 RoTat1.2 PCR positives, 52 were negative PCR (0.32 pg/␮l) and 18S PCR (0.32 pg/␮l). However, at the lower in 18S PCR and 58 were negative in ITS1 PCR. 18S PCR and ITS- DNA concentrations, aspecific amplicons start to appear besides the 1 PCR picked respectively 5 and 4 positive specimens out of 314 Trypanozoon-specific 450 bp in the ITS-1PCR (Fig. 2). RoTat 1.2 PCR negative specimens (Table 3). There was slight agree- None of the PCR tests showed a significant (p > 0.05) difference ment between RoTat1.2 PCR and Giemsa stained thin blood smear in positivity rate between gender groups and between animals (K = 0.03) but there was a moderate agreement between RoTat 1.2 with poor and moderate body condition (Table 5). However a sig- PCR and 18S PCR (K = 0.47) and a fair agreement between RoTat1.2 nificantly higher prevalence in female animals (␹2 = 5.4, p < 0.05; PCR and ITS-1 PCR (K = 0.39). There was a substantial agreement OR = 0.6, CI 0.3-0.9 by CATT/T. evansi) and in camels with poor body between 18S PCR and ITS-1 PCR (K = 0.79) though out of 38 posi- condition (␹2 = 20.3, p < 0.05; OR = 26.8, CI 3.3–220.9 by thin smear tives for 18S PCR, 10 specimens were negative for ITS-1 PCR and examination, ␹2 = 31.4, p < 0.05; OR = 4.1, CI 2.4–6.8 by RoTat1.2 out of 31 ITS-1 PCR positives, 3 were negative for 18S PCR (Table 3). PCR) was recorded. No significant difference in positivity rate was There was slight agreement between Giemsa stained thin blood recorded between age groups with ITS-1 PCR (␹2 = 0.7, p = 0.3; smear and 18S PCR (K = 0.05). OR = 1.7, CI 0.5–6.0) and 18S PCR (␹2 = 0.2, p = 0.4; OR = 1.3, CI The frequency distribution of tests results (N) and probability of 0.4–4.6) but in CATT/T. evansi (␹2 = 3.8, p < 0.05;, OR = 2.3, CI 1.0–5.2) being infected for each outcome pattern as estimated with LCA is and RoTat1.2 PCR (␹2 = 5.6, p < 0.05;, OR = 2.7, CI 1.2–6.2) there were presented in Table 3. Further results from LCA are summarised in significantly more positives in older camels. History of trypanoci- Table 4. The estimated prevalence from the model was 19% (95% dal treatment had no significant (p > 0.05) impact on positivity rate CI: 13–28). The model indicated for CATT/T. evansi a low sensitivity by all the diagnostic tests. of 43% with relatively low specificity of 80%. The PCR tests showed The overall mean PCV in the studied camels was low (23.6%) higher specificity (86% for RoTat 1.2, >99% for 18S and ITS-1) with indicating that the animals were generally in poor condition. It sensitivity similar to CATT/T. evansi in the 40–50% range (53% for is noteworthy that in our study, no camels were observed with RoTat 1.2 PCR, 48% for 18S PCR, and 39% for ITS-1 PCR). The LCA good body condition. In all diagnostic tests PCV was significantly results confirmed the very low sensitivity (11%) of Giemsa stained (p < 0.05) lower in test positive animals compared to test negative thin blood smear in this setting. animals (Table 6). The latent class model classified all those with Giemsa stained thin blood smear positive as infected, as well as all those with at 4. Discussion least 2 PCRs positive, or CATT/T. evansi and at least one of 18S PCR or ITS-1 PCR positive. Those with only CATT/T. evansi or at most a In this study, Giemsa stained thin blood smear examination, single PCR positive were classified as not infected, as were those antibody detection by CATT/T. evansi and molecular tests by RoTat with only CATT/T. evansi and RoTat 1.2 PCR positive. 1.2, 18S and ITS1-PCRs were used to assess prevalence and host When tested on purified DNA of a Kazakhstan and a Moroc- related risk factors of trypanosome infection in camels in Afar can T. evansi strain, the analytical sensitivity of the ITS-1 PCR was region, Eastern Ethiopia. Only 2% of the examined camels were 0.064 pg/␮l with both T. evansi strains and lower than the RoTat1.2 found infected by T. evansi by microscopy. This corresponds with R. Fikru et al. / Veterinary Parasitology 211 (2015) 175–181 179

Table 5 Host related variables and positivity rate obtained with different tests for T. evansi infection diagnosis, performed on 399 camels from Eastern Ethiopia.

Number Thin blood smear CATT/T. evansi RoTat1.2 PCR 18S PCR ITS-1 PCR Positive Positive Positive Positive Positive

Gender Female 251 2.8* (0.8–4.8) 27.9 (22.6–33.5)* 23.1 (17.9–28.3) 8.8 (5.3–12.3) 6.8 (3.7–9.9) Male 148 0.7 17.6 (11.5–23.7) 18.2 (12.0–24.4) 10.8 (5.8–15.8) 9.5 (4.8–14.2)

Body condition score Moderate 311 0.3 17.7 (13.5–21.9) 22.5 (17.9–27.1) 10 (6.7–13.3) 7.4 (4.5–10.3) Poor 88 8* (2.3–10.7) 46.6* (36.4–56.8) 17 (9.1–24.9) 7 (1.7–12.3) 9.1 (3.1–15.1)

Treatment history No 200 2.0 (0.1–3.9) 27.5 (21.3–33.7) 21.5 (15.8–27.2) 11 (6.7–15.3) 7.5 (3.8-11.2) Yes 199 2.0 (0.0–4.0) 20.6 (15.0–26.2) 21.1 (15.4–26.8) 8 (4.2–11.8) 8 (4.2–11.8)

Age category ≤1 year 50 0 20 (8.9–31.1) 22.0 (10.5–33.5) 12.0 (13.0–31.0) 8.0 (0.5–15.5) 110 year 25 8 (0.0–18.3) 40.0* (20.8–59.2) 40.0* (20.8–59.2) 12.0 (0.0–24.7) 12.0 (0.0–24.7)

*Signiﬁcant difference.

previous findings in camels in different parts of Ethiopia, includ- densities of Tabanid and Stomoxys flies during that season (Löhr ing two other districts in Afar Region, Gowani and Awash Fentale et al., 1985; Luckins, 1988; Hailemariam et al., 2008). (Getahun, 1998; Hailemariam et al., 2008; Kassa et al., 2011; In the absence of a golden standard test, the LCA revealed an Tadesse et al., 2012) and in other African countries, for instance overall estimate prevalence of 19.4%, with low sensitivity of CATT/T. 2.3% in Kenya (Ngaira et al., 2004) and 2.5% in Nigeria (Egbe-Nwiyi evansi and the PCR tests (43–53%) but better than the sensitivity of and Chaudry, 1994). However, some other studies conducted in Giemsa stained thin blood smear examination (11%). When con- Ethiopia reported higher prevalences 6.5–12.1% (Richard, 1979; sidering microscopy, one should keep in mind that the observed Abebe, 1991; Hailu, 2000; Hagos et al., 2009). This could be due prevalence reflects only a fraction of the real prevalence since most to differences in the management and husbandry regimens of the microscopic techniques are poorly sensitive (OIE, 2012). In case camels under study, or to seasonal effects as well as to study design. of Giemsa stained thin smear examination, the lower detection Previous studies have shown that the prevalence of surra in camels limit is greater than 500,000 trypanosomes/ml of blood (OIE, 2012). is higher during the rainy season and could be explained by higher Therefore, indirect diagnosis, e.g. by detection of T. evansi specific antibodies, may give a better estimation of the infection burden as has been shown in previous studies (Diall et al., 1994; Gutierrez et al., 2000). An overall seroprevalence of 24.1% was recorded with CATT/T. evansi and corresponds with the CATT seroprevalence observed in Bale zone (24.9%) (Hagos et al., 2009). The observed seroprevalences may be overestimates of the actual prevalence since antibody detection tests have the inher- ent shortcoming that past infections cannot be distinguished from current infections (Luckins, 1988; Ngaira et al., 2003; Zeyed et al., 2010). Molecular diagnostics targeting parasite DNA are considered good surrogates for parasite detection (Büscher, 2014). An overall positivity rate of 21.1%, 9.5% and 7.8% was observed respectively by RoTat 1.2 PCR, 18S PCR and ITS-1 PCR, which is significantly higher than recorded by thin smear examination (2%). CATT/T evansi and RoTat1.2 PCR yielded comparable positivity rates that were significantly higher than in 18S and ITS-1 PCRs. Similar observations were made in Uganda where very low molecular prevalence of Try- panozoon was recorded with the ITS-1 PCR as compared to a PCR for the single copy gene phospholipase C (GPI-PLC) (Ahmed et al., 2013). The higher positivity rate recorded with RoTat1.2 PCR needs

Table 6 Mean PCV (%) with standard deviation (SD) and 95% conﬁdence interval of 399 camels from Eastern Ethiopia tested using different tests to detect T. evansi infection.

Test PCV SD 95% Conﬁdence interval

Thin blood smear Negative 23.7 2.7 23.4–24.0 Positive 20.7* 2.1 19.0–22.5 CATT/T. evansi Negative 24.3 2.6 24.0–24.6 Positive 21.5* 1.8 21.1–21.9 RoTat1.2 PCR Negative 23.8 2.7 23.5–24.1 Positive 23.0* 2.7 22.4–23.5 Fig. 2. Analytical sensitivity of the three PCRs: Panel A = RoTat1.2 PCR; Panel B = 18S 18S PCR Negative 23.7 2.7 23.5–24.0 PCR, Panel C = ITS-1PCR. Lanes 1: Gene Ruler 100 bp DNA ladder (Fermentas). Lanes Positive 22.6* 2.7 21.8–23.5 2–8: DNA of the T. evansi isolate from Kazakhstan at decreasing concentrations ITS-1 PCR Negative 23.7 2.7 23.4–24.0 ␮ (1000, 200, 40, 8, 0.16, 0.032, 0.0064 pg/ l). Lanes 9–15: DNA of the T. evansi Positive 22.6* 2.7 21.7–23.6 isolate from Morocco at decreasing concentrations (1000, 200, 40, 8, 0.16, 0.032, * 0.0064 pg/␮l). Significant difference. 180 R. Fikru et al. / Veterinary Parasitology 211 (2015) 175–181 special attention but could partly be explained by the low RoTat1.2 be positive in Giemsa stained thin smear and CATT/T. evansi, com- PCR specificity estimate by LCA. Our data show that the analytical pared to those with moderate body condition. Nutritional stress or sensitivity of RoTat 1.2 PCR based on the single copy gene is lower other diseases, whether infectious or not, could render camels more than the ITS-1 and 18S PCR based on multi copy sequences. Yet, susceptible to infection with T. evansi (Eyob and Matios, 2013). On RoTat 1.2 PCR unexpectedly yielded a higher positivity rate than the other hand, camel trypanosomosis by itself is one of the major ITS-1 and 18S PCR. The lowest positivity rate obtained by ITS-1 PCR causes resulting in poor body condition (Röttcher et al., 1987). Sig- remains unexplained but it could be related to limited specificity of nificantly lower PCV values were recorded in seropositive and PCR the ITS-1 primers. We observed that when the concentration of tar- positive camels which correspond with other studies using sero- get DNA decreases in the presence of host DNA, the primers become logical, parasitological and molecular diagnostic tests (Diall et al., less specific and tend to bind to the host DNA. On the other hand, we 1994; Atarhouch et al., 2003; Tadesse et al., 2012). cannot rule out that the RoTat1.2 PCR is less specific in the presence In this study, we used LCA to estimate the accuracy of the differ- of host DNA as it is supported by the LCA where RoTat1.2 PCR is less ent diagnostic tests in the absence of a gold standard. LCA revealed specific than ITS-1 PCR and 18S PCR. Noteworthy, only one out of 8 lower sensitivity for CATT/T. evansi (43%) than expected from pre- thin blood smear positive animals was negative for CATT/T. evansi vious studies (80–92%) while the specificity (80.4%) falls within the whereas out of 303 CATT/T. evansi negative samples 19.5%, 8.3% and expected range (69–100%)(Diall et al., 1994, 1997; Verloo et al., 6.9% respectively, were positive for RoTat1.2, 18S and ITS-1 PCRs. It 1998; Ngaira et al., 2003). Unfortunately, for reasons of specimen is possible but unlikely that all these seronegative animals carried import regulations, we were not able to perform immune trypanol- early infections and had not yet formed detectable antibody levels. ysis with T. evansi RoTat 1.2 as reference test for T. evansi type A The number of RoTat 1.2 PCR positives that could not be detected by specific antibodies (Verloo et al., 2001). The low sensitivity of the CATT/T. evansi is significantly higher than with 18S and ITS-1 PCRs. CATT/T. evansi in this study may be due to unusual low amounts This could be explained by the presence but not yet the expression of specific antibodies circulating in the infected camels that might of the RoTat1.2 gene in the circulating trypanosomes (Verloo et al., have been detected in immune trypanolysis, or to infections with 2001). On the other hand there are a number of RoTat1.2 PCR and T. evansi type B that remained under the lower detection limit of CATT/T. evansi negatives that are positive in 18S and ITS-1 PCR that the EVAB PCR. could be explained by the absence of the RoTat1.2 gene like it has In conclusion, this study confirms that trypanosome infection been described for the T. evansi type B strains isolated in Kenya is a potential threat affecting the health and productivity of camel, and suspected to circulate also in Ethiopia but not confirmed in our in particular adult female animals, in Afar, Ethiopia. T. evansi is the study (Elsaid et al., 1998; Davison et al., 1999; Ngaira et al., 2003; prominent trypanosome species affecting camels since only one Hagos et al., 2009; Zeyed et al., 2010; Salim et al., 2011). case among 399 camels was confirmed infected with T. vivax by Even though 18S PCR and ITS-1 PCR cannot differentiate TvPRAC PCR. There are significant discrepancies in sensitivity of between the taxa of the Trypanozoon group, the positive results different diagnostic tests used in this epidemiological study neces- reported in this study are most likely due to T. evansi, since most of sitating improvement of the available diagnostic tests to provide them are positive in the T. evansi specific RoTat1.2 PCR and there are high quality epidemiological data. no records of tsetse flies in the Afar region excluding the presence of T. brucei but not of non-RoTat 1.2 T. evansi (Njiru et al., 2006). In Ethiopia, tsetse flies and tsetse-transmitted trypanosomes prevail Acknowledgement in the Southwestern and Northwestern parts of the country following major river basins between longitude 33o and 38o E and The Ph.D. fellowship of Fikru Regassa was financed by the Bel- latitude 5o and 12o N while the rest of the country, the North, gian Directorate General for Development Cooperation. Southeast and East are considered tsetse free (National Tsetse and Trypanosomosis Investigation and Control Center (NTTICC), 2004; References Sinshaw et al., 2006). Our study revealed, for the first time in Ethiopia, a T. vivax infection in camels. That there are no previ- Abebe, W., 1991. Practices and major health problems of cames in the Ogaden, ous reports of camels infected with T. vivax in Ethiopia is probably Ethiopia. Eura. J. Agric. Environ. Sci. 8, 633–642. due to the fact that the commonly used serological and parasito- Ahmed, H.A., Picozzi, K., Welburn, S., MacLeod, A., 2013. A comparative evaluation of PCR-based methods for species-specific determination of African animal logical tests fail to detect T. vivax or to distinguish it from T. evansi. trypanosomes in Ugandan cattle. Parasit. Vectors 6, 316. Mixed infections of T. congolense, T. vivax and Trypanozoon have Ali, I., Shafi, M., Chaudhry, Q., Faroo, Q., 2009. Camel rearing in Cholistan desert of been reported in camels in Nigeria but were based on microscopic Pakistan. Pakistan Vet. J. 29, 85–92. Atarhouch, T., Rami, M., Bendahman, M.N., Dakkak, A., 2003. Camel trypanosomo- examination while in Sudan, using the ITS-1 PCR, only T. evansi (Try- sis in Morocco 1: results of a first epidemiological survey. Vet. Parasitol. 111, panozoon) has been detected in camel (Mbaya et al., 2010; Salim 277–286. et al., 2011). Pathogenicity of T. vivax to camels is not well described Bajyana Songa, E., Hamers, R., 1988. A card agglutination test (CATT) for veterinary use based on an early VAT RoTat 1/2 of Trypanosoma evansi. Ann. Soc. Belg. Méd. but infected animals could act as reservoir and spread the para- Trop. 68, 233–240. site to the more susceptible livestock species (Enwezor and Sackey, Berlin, D., Nasereddin, A., Azmi, K., Ereqat, S., Abdeen, Z., Baneth, G., 2010. Lon- 2005). gitudinal study of an outbreak of Trypanosoma evansi infection in equids and dromedary camels in Israel. Vet. Parasitol. 174, 317–322. Whatever diagnostic technique used, the overall observed Birhanu, H., Fikru, R., Kidane, W., Gebrehiwot, T., Tola, A., Goddeeris, B.M., Büscher, prevalence of surra was similar in both study sites which can be P., 2013. Epidemiology of Trypanosoma evansi in domestic animals in Tigray and explained by comparable camel husbandry practices and ecolog- Afar regions, Northern Ethiopia. In: 32nd Meeting of the International Scientific ical biotopes in both districts. On the other hand, a significantly Council for Trypanomiasis Research and Control (ISCTRC), Khartoum, Sudan, September 8–12. higher prevalence was recorded in female animals by CATT/T. evansi Büscher, P., 2014. Diagnosis of African trypanosomiasis. In: Magez, S., Radwan- and Giemsa stained thin smear. Similar findings were reported in ska, M. (Eds.), Trypanosomes and Trypanosomiasis. Wien, Springer-Verlag, pp. Kenya and in Israel (Singh et al., 2004; Berlin et al., 2010). The fact 189–216. Claes, F., Radwanska, M., Urakawa, T., Majiwa, P., Goddeeris, B., Büscher, P., 2004. that seropositivity in CATT increases with age is in agreement with Variable surface glycoprotein RoTat 1.2 PCR as a specific diagnostic tool for the previous studies and probably reflects time of exposure to biting detection of Trypanosoma evansi infections. Kinetoplastid Biol. Dis. 3, 1–6. flies and persistence of circulating antibodies, even after success- Cohen, J., 1960. A coefficient of agreement for nominal scales. Educ. Psychol. Meas. 20, 37–46. ful treatment (Dia et al., 1997; Gutierrez et al., 2000; Atarhouch Davison, H.C., Thrusfield, M.V., Muharsini, S., Husein, A., Partoutomo, S., Rae, P.F., et al., 2003). Camels with poor body conditions were more likely to Masake, R., Luckins, A.G., 1999. Evaluation of antigen detection and antibody R. Fikru et al. / Veterinary Parasitology 211 (2015) 175–181 181

detection tests for Trypanosoma evansi infections of buffaloes in Indonesia. Epi- Hoare, C.A., 1972. The Trypanosomes of Mammals. Blackwell Scientific Publications, demiol. Infect. 123, 149–155. Oxford, 749 pp. Deborggraeve, S., Claes, F., Laurent, T., Mertens, P., Leclipteux, T., Dujardin, J.C., Kassa, T., Eguale, T., Chaka, H., 2011. Prevalence of camel trypanosomosis and Herdewijn, P., Büscher, P., 2006. Molecular dipstick test for diagnosis of sleeping its vectors in Fentale district, South East Shoa zone, Ethiopia. Vet. Arh. 81, sickness. J. Clin. Microbiol. 44, 2884–2889. 611–621. Demeke, G., 1998. Prevalence of camel trypanosomes and factors associated with Landis, J.R., Koch, G.G., 1977. The measurement of observer agreement for categorical the disease occurrence in Leben district, Borena zone, Oromia region Ethiopia. data. Biometrics 33, 159–174. M.Sc. Thesis. Addis Ababa University and Free University of Berlin. Löhr, K.F., Pholpark, S., Srikitjakarn, L., Thaboran, P., Bettermann, G., Staak, C., 1985. Desquesnes, M., Bosseno, M.-F., Brenière, S.F., 2007. Detection of Chagas infections Trypanosoma evansi infection in buffaloes in north-east Thailand. I. Field inves- using Trypanosoma evansi crude antigen demonstrates high cross-reactions with tigations. Trop. Anim. Health Prod. 17, 121–125. Trypanosoma cruzi. Infect. Genet. Evol. 7, 457–462. Luckins, A.G., 1988. Trypanosoma evansi in Asia. Parasitol. Today 4, 137–142. Desquesnes, M., McLaughlin, G., Zoungrana, A., Dávila, A.M., 2001. Detection and Mbaya, A., Ibrahim, U., Apagu, S., 2010. Trypanosomosis of the dromedary camel identification of Trypanosoma of African livestock through a single PCR based on (Camelus dromedarius) and its vectors in the tsetse-free arid zone of North East- internal transcribed spacer 1 of rDNA. Int. J. Parasitol. 31, 610–614. ern Maiduguri, Nigeria. Vet. Parasitol. 124, 187–199. Dia, M.L., Diop, C., Aminetou, M., Jacquiet, P., Thiam, A., 1997. Some factors affecting Menten, J., Boelaert, M., Lesaffre, E., 2008. Bayesian latent class models with condi- the prevalence of Trypanosoma evansi in camels in Mauritania. Vet. Parasitol. 72, tionally dependent diagnostic tests: a case study. Stat. Med. 27, 4469–4488. 111–120. National Tsetse Trypanosomosis Investigation and Control Center (NTTICC), 2004. Diall, O., Bajyana Songa, E., Magnus, E., Kouyate, B., Diallo, B., Van Meirvenne, N., Annural report on tsetse and trypanosomosis survey. Bedele, Ethiopia. Hamers, R., 1994. Evaluation d’un test sérologique d’agglutination directe sur Ngaira, J.M., Bett, B., Karanja, S.M., Njagi, E.N.M., 2003. Evaluation of antigen and carte dans le diagnostic de la trypanosome caméline à Trypanosoma evansi. Rev. antibody rapid detection tests for Trypanosoma evansi infection in camels in Sci. Tech. Off. Int. Epizoot. 13, 793–800. Kenya. Vet. Parasitol. 114, 131–141. Diall, O., Diarra, B., Sanogo, Y., 1997. Evaluation of the antigen ELISA as a tool for Ngaira, J.M., Njagi, E.N.M., Ngeranwa, J.J.N., Olembo, N.K., 2004. PCR amplification of assessing the impact of tsetse control programmes on the incidence of try- RoTat 1.2 VSG gene in Trypanosoma evansi isolates in Kenya. Vet. Parasitol. 120, panosome infections in livestock. In: IAEA (Ed.), Joint FAO/IAEA Division of 23–33. Nuclear Techniques in Food and Agriculture and the IAEA Department of Tech- Njiru, Z.K., Constantine, C.C., Masiga, D.K., Reid, S.A., Thompson, R.C., Gibson, W.C., nical Co-operation held in Addis Ababa, Ethiopia, 1995. Ethiopia, April, 17–28, 2006. Characterization of Trypanosoma evansi type B. Infect. Genet. Evol. 6, pp. 57–64. 292–300. Egbe-Nwiyi, T., Chaudry, S., 1994. Trypanosomosis: prevalence and pathology of OIE, 2012. Manual of diagnostic tests and vaccines for terrestrial animals. camel of arid zone of north eastern Nigeria. Trop. Vet. 20, 30–34. http://www.oie.int/fileadmin/Home/eng/Health standards/tahm/2.04.18 Elias, G., 2003. Study of the Prevalence of Camel Trypanosomosis in Selected Sites of TRYPANOSOMOSIS.pdf East Shoa, Arsi and Bale Lowlands of Ethiopia. Addis Ababa University, Faculty R Core Team, 2014. R: A language and environment for statistical computing. In: of Veterinary Medicine, Debre Zeit, Ethiopia. Foundation for Statistical Computing (Ed.), Vienna, Austria. Elsaid, H.M., Nantulya, V.M., Hilali, M., 1998. Diagnosis of Trypanosoma evansi infec- Richard, D., 1979. The disease of the dromadary (Camelus dromedarius) in Ethiopia. tion among sudanese camels imported to Egypt using card agglutination test Ethiop. Vet. Bull. 2, 46–67. (CATT) and antigen detection latex agglutination test (SURATEX). J. Protozool. Röttcher, D., Schillinger, D., Zweygarth, E., 1987. Trypanosomiasis in the camel Res. 8, 194–200. (Camelus dromedarius). Rev. Sci. Tech. Off. Int. Epizoot. 6, 463–470. Enwezor, F.N.C., Sackey, A.K.B., 2005. Camel trypanosomosis—a review. Vet. Arh. 75, Salim, B., Bakheit, M.A., Kamau, J., Nakamura, I., Sugimoto, C., 2011. Molecular epi- 439-425a. demiology of camel trypanosomiasis based on ITS1 rDNA and RoTat 1.2 VSG Eyob, E., Matios, L., 2013. Review on camel trypanosomosis (surra) due to Try- gene in the Sudan. Parasit. Vectors 4, 31. panosoma evansi: epidemiology and host response. J. Vet. Med. Anim. Health Singh, N., Pathak, K.M.L., Kumar, R., 2004. A comparative evaluation of parasito- 5, 334–343. logical, serological and DNA amplification methods for diagnosis of natural Faye, B., Bengoumi, M., Cleradin, A., Tabaran, A., Chilliard, Y., 2001. Body condition Trypanosoma evansi infection in camels. Vet. Parasitol. 126, 365–373. score in dromedary camel: a tool for management of reproduction. Em. J. Food Sinshaw, A., Abebe, G., Desquesnes, M., Yoni, W., 2006. Biting flies and Trypanosoma Agric. 13, 1–6. vivax infection in three highland districts bordering lake Tana, Ethiopia. Vet. Fikru, R., Goddeeris, B.M., Delespaux, V., Moti, Y., Tadesse, A., Bekana, M., Claes, F., Parasitol. 142, 35–46. De Deken, R., Büscher, P., 2012. Widespread occurrence of Trypanosoma vivax Tadesse, A., Omar, A., Aragaw, K., Mekbib, B., Sheferaw, D., 2012. A study on camel in bovines of tsetse- as well as non-tsetse-infested regions of Ethiopia: a reason trypanosomosis in Jijiga zone, eastern Ethiopia. J. Vet. Adv. 2, 216–219. for concern? Vet. Parasitol. 190, 355–361. Tekle, T., Abebe, G., 2001. Trypanosomosis and helminthoses: major health problems Fikru, R., Hagos, A., Rogé, S., Reyna-Bello, A., Gonzatti, M.I., Merga, B., Goddeeris, of camels (Camelus dromedarius) in the southern rangelands of Borena, Ethiopia. B., Büscher, P., 2014. A proline racemase based PCR for indentification of Try- J. Camel Pract. Res. 8, 39–42. panosoma vivax in cattle blood. PLoS ONE 9, e84819. Thrusfield, M.V., 2005. Veterinary Epidemiology. Blackwell Publishing Professional, Getahun, D., 1998. The prevalence of camel trypanosomosis and factors associated Ames Iowa, USA, pp. 610. with the disease occurrence in Leben district, Borena zone, Oromiya region, Toma, B., Dufour, B., Sanaa, M., Benet, J.-J., Ellis, P., Moutou, F., Louza, A., 1999. Applied Etiopia. Addis Ababa University, Faculty of Veterinary Medicine and Freie Uni- Veterinary Epidemiology and Control of Disease in Populations. Maison Alfort, versität Berlin, Faculty of Veterinary Medicine, M.Sc. Thesis. Paris, France, pp. 576. Gutierrez, C., Juste, M.C., Corbera, J.A., Magnus, E., Verloo, D., Montoya, J.A., 2000. Verloo, D., Holland, W., My, L.N., Thanh, N.G., Tam, P.T., Goddeeris, B., Vercruysse, J., Camel trypanosomosis in the Canary Islands: assessment of seroprevalence and 2000. Comparison of serological tests for Trypanosoma evansi natural infections infection rates using the card agglutination test (CATT/T.evansi) and parasite in water buffaloes from north Vietnam. Vet. Parasitol. 92, 87–96. detection tests. Vet. Parasitol. 90, 155–159. Verloo, D., Magnus, E., Büscher, P., 2001. General expression of RoTat 1.2 variable Hagos, A., Yilkal, A., Esayass, T., Alemu, T., Fikru, R., Feseha, G., Goddeeris, B.M., Claes, antigen type in Trypanosoma evansi isolates from different origin. Vet. Parasitol. F., 2009. Parasitological and serological survey on trypanosomosis (surra) in 97, 183–189. camels in dry and wet areas of Bale Zone, Oromyia Region, Ethiopia. Rev. Méd. Verloo, D., Tibayrenc, R., Magnus, E., Büscher, P., Van Meirvenne, N., 1998. Perfor- Vét. 12, 569–573. mance of serological tests for Trypanosoma evansi infections in camels from Hailemariam, L., Desalegn, L., Ibrahim, H., 2008. Prevalence and distribution of camel Niger. J. Protozool. Res. 8, 190–193. trypanosomosis in the semi-arid and arid Awash Valley of Ethiopia. Ethiop. J. Zeyed, A., Habeeb, S., Allam, N., Ashry, H., Mohamed, A., Ashour, A., Taha, H., 2010. Anim. Prod. 8, 1–9. A critical comparative study of parasitological and serological differential diag- Hailu, D., DVM Thesis 2000. The prevalence of camel trypanosomosis in the salt nostic methods of Trypanosoma evansi infection in some farm animal in Egypt. convey routes of Afar-Tigray. Addis Ababa University, Faculty of Veterinary Vet. World 1, 325–328. Medicine, Debre Zeit, Ethiopia.