Asian Pacific Journal of Tropical Medicine

: 2020  |  Volume : 13  |  Issue : 7  |  Page : 322--327

Leishmania donovani infection in Eastern Sudan: Comparing direct agglutination and rK39 rapid test for diagnosis-a retrospective study

Elfadil Abass 
 Department of Clinical Laboratory Science, College of Applied Medical Sciences, Imam Abdulrahman Bin Faisal University, P.O. Box 2435, Dammam, Saudi Arabia

Correspondence Address:
Elfadil Abass
Department of Clinical Laboratory Science, College of Applied Medical Sciences, Imam Abdulrahman Bin Faisal University, P.O. Box 2435, Dammam
Saudi Arabia


Objective: To compare diagnostic accuracy and agreement between direct agglutination test and rK39 rapid tests for diagnosis of visceral leishmaniasis in an endemic area, the Doka area in Eastern Sudan. Methods: Stored sera of confirmed visceral leishmaniasis cases, unconfirmed visceral leishmaniasis-suspects and negative controls were tested by direct agglutination test and rK39 rapid test. The sera were collected from the Doka area in Eastern Sudan. Diagnostic accuracy of direct agglutination test and rK39 rapid test was assessed in terms of sensitivity, specificity, positive predictive value and negative predictive value, estimated at 95% confidence interval (CI). Agreement between the two tests was determined by the kappa (κ) value. Results: Taking lymph node aspiration of Leishmania as a gold standard, direct agglutination test showed 91.0% sensitivity, 99.3% specificity, resulting in a positive and negative predictive value of 99.3% and 91.0%, respectively. In contrast, the sensitivity of rK39 rapid test was 85.2% and specificity 98.6%, resulting in a positive and negative predictive value of 98.5% and 85.9%, respectively. Most (81.3%) of the confirmed visceral leishmaniasis sera revealed strong antibody titers (≥1:6 400). Some sera (n=5) that were positively tested with rK39 rapid test were negative in direct agglutination test (≤1:800); in contrast, direct agglutination test was positive in 12 confirmed visceral leishmaniasis sera that were negatively tested with rK39 rapid test. There was moderate to good agreement between direct agglutination test and rK39 rapid test for confirmed visceral leishmaniasis patients (κ=0.42, 95% CI=0.21-0.63) and control sera (κ=0.80, 95% CI=0.41-1.00). Conclusions: Both direct agglutination test and rK39 rapid test are satisfactory test systems for visceral leishmaniasis diagnosis in East Sudan. Their simplicity makes them ideal for first healthcare in rural areas. These data are relevant also for other East African endemic countries because of the geographical and overlapping distribution of the Leishmania parasite.

How to cite this article:
Abass E. Leishmania donovani infection in Eastern Sudan: Comparing direct agglutination and rK39 rapid test for diagnosis-a retrospective study.Asian Pac J Trop Med 2020;13:322-327

How to cite this URL:
Abass E. Leishmania donovani infection in Eastern Sudan: Comparing direct agglutination and rK39 rapid test for diagnosis-a retrospective study. Asian Pac J Trop Med [serial online] 2020 [cited 2020 Aug 3 ];13:322-327
Available from:

Full Text

 1. Introduction

Leishmaniases are a group of infectious diseases caused by parasites belonging to the Leishmania genus. Leishmania parasites are transmitted through the bite of an infected female sandfly. The disease occurs in three clinical forms, among which, visceral leishmaniasis (VL) is the most severe form with 100% mortality rate if left untreated[1]. VL is caused by species belonging to Leishmania (L.) donovani complex. The disease is spread over the world with major endemic regions in South-West Asia, East Africa, South America and Mediterranean countries. However, more than 90% of all cases occur in Sudan, South Sudan, Ethiopia, India, Bangladesh and Brazil[2]. East Africa is of particular interest because Leishmaniasis co-exists with other infectious diseases and is expanding into so far non-endemic areas[3],[4],[5].

In Sudan, VL is endemic mainly in the East and Central regions where thousands of cases are reported annually[6],[7]. It is caused by L. donovani and transmitted by the sandfly vectors, Phlebotomus orientalis and Phlebotomus martini[8]. The country has experienced several epidemics, with several thousand fatalities. There is a public health concern about re-occurrence of VL epidemics in Sudan. A prime goal for endemic countries is to establish easy-to-use and reliable VL-diagnostic tests to support early detection of the disease and thus to prevent further outbreaks[9]. Till today, parasitological detection of Leishmania from organ aspirates and in vitro cultivation are considered to be the reference methods of VL diagnosis. However, these methods have several drawbacks: their application in the field is limited; their sensitivity can be affected by type of the organ, from which samples are collected, with high numbers of parasites in spleen and low in lymph node samples. Because of the low sensitivity of most commercially available tests, the absence of parasites in organ aspirates doesn’t rule out the infection[10],[11]. Due to these limitations, parasitological detection is not widely used in many countries.

Sero-diagnostic is an alternative method for VL detection. It is more sensitive than classical parasite isolation and culture methods[12]. Several serological tools are now available and used in endemic countries. In Sudan, VL diagnosis is primarily based on direct agglutination test (DAT) and to some extent, on rK39 rapid tests (RT). These tests, however, may not be specific enough to differentiate between previous exposure and current active disease and may give false positive results in some other diseases such as tuberculosis and malaria[13],[14],[15]. Here, we report on a study which compared the accuracy and concordance of DAT and rK39 rapid tests for VL diagnosis in the Doka area, a major endemic region in Eastern Sudan. The comparison between DAT and rK39 rapid was performed with stored serum samples from a serum bank facility in Sudan.

 2. Materials and methods

2.1. Ethical approval

Human sera used in this study were randomly selected and tested anonymously without referring to the patients’ information. The sera were collected in endemic area of Eastern Sudan-Doka after filling in consent forms. These sera were used previously in other studies[16],[17],[18]. This study was ethically approved by the Ethical Review Committee of the Federal Ministry of Health in Sudan (23-06-2005).

2.2. Characteristics of sera and study design

This is a retrospective study performed with human sera (n=494) obtained from the serum bank at the Laboratory of Biomedical Research of Ahfad University for Women. These samples were collected as part of a previous project at Doka region, Eastern Sudan. VL suspicion was based on clinical signs and symptoms according to clinical practice guidelines of the Sudanese Ministry of Health[19]. Patients from different health centers were sent to Doka hospital for diagnosis and treatment from February 2006 and October 2007. Major signs and symptoms of these patients included: fever, hepatomegaly, splenomegaly, lymphadenopathy and abnormal complete blood count including leukopenia, anemia and thrombocytopenia. Lymph node aspirates and blood samples were collected for VL diagnosis. In this study, we selected 155 sera of confirmed VL cases (positive lymph node samples) and 197 sera of un-confirmed VL suspects (negative lymph node samples). The classification was based on the presence of classical signs and symptoms combining with lymph node results. The patients were not tested for other diseases, as they did not show any other clinical signs. As controls, we selected a group of 142 sera of individuals from the same VL-endemic area that were lacking disease symptoms (apparently healthy), or patients suffering from other diseases rather than visceral leishmaniasis. The control group included VL-endemic controls (n=50), non-endemic controls (n=20), cases with malaria (n=20), pulmonary tuberculosis (n=20), leukemia (n=20) and African trypanosomiasis (n=12). The diagnosis of diseases was performed according to the national guideline of the Sudanese Ministry of Health.

2.3. Serological tests

We used two serological tests, DAT, and rK39 RT based on the recombinant protein of L. infantum.

2.3.1. Direct agglutination test (DAT)

DAT test is based on L. donovani isolated from a VL patient in Eastern Sudan and was prepared at the Laboratory for Biomedical Research, Ahfad University for Women according to the standard procedure, as described previously[20]. Briefly, sera were serially diluted (1:100-1:102 400) in 0.2% gelatin and physiological saline solution (0.9% g/v). Leishmania antigen (50 μL) was mixed with 50 μL diluted sera in V-shaped microtiter plates (Greiner Bio-One, Frickenhausen, Germany) and incubated overnight at room temperature. Antibody titers were determined as the highest dilution of sera showing agglutination reaction with a cut-off titer of 1:3 200. Lower titers were considered borderline (1:1 600) or negative (≤1:800)[20].

2.3.2. rK39 RT

Rapid test kits using the recombinant protein K39[21] was purchased from Bio Rad (France). Tests were performed as recommended by the manufacturer. Briefly, 20 μL of the serum sample was added to the test device. Following incubation the test was read and interpreted. Results were recorded as positive if two lines appeared, or negative when only one line appeared. If no lines appeared, the test was considered invalid and was repeated.

2.4. Statistical analysis

The diagnostic accuracy of DAT and rK39 rapid test was compared in term of sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) and estimated at 95% confidence interval (CI). Sensitivity was presented as percent through the ratio of positive sera divided by confirmed VL cases; specificity in percentage as ratio of all control cases with negative test results. The latter group included sera from patients with malaria, pulmonary tuberculosis, leukaemia, African trypanosomiasis as well as negative controls from VL-endemic and non-endemic areas. PPV was estimated as the number of true positives over the number of true positives and false positives. NPV was estimated as the number of true negatives over the number of true negatives and false negatives. Agreement between the two tests was done by kappa (κ) value using Graph Pad available at kappa1.cfm. The level of agreement between the two tests was done as described earlier[22] as follows: κ<0.4, bad; 0.41≤κ≤0.60, moderate; 0.61≤κ≤0.80, good, and κ≥0.8, excellent.

 3. Results

A total of 494 serum samples including 155 parasitologicaly confirmed VL cases, 197 clinically suspected but un-confirmed VL cases and 142 negative controls were used in this study. Most sera were originally collected at Doka Hospital, Eastern Sudan. Serum samples were tested by DAT produced at Ahfad University and the commercial rK39 RT.

Out of the 155 confirmed VL positive sera, the DAT truly identified 141 positive samples and missed 14 sera. Sensitivity for DAT was 91.0%. Out of the 142 control sera, the test revealed only one false positive result (a VL-endemic negative control) that was negative with the rK39 RT. The test demonstrated 99.3% specificity.

Accordingly, PPV was 99.3% and NPV was 91.0% [Table 1]. For the confirmed VL group, the sensitivity of rK39 RT was 85.2%. Out of 155 VL confirmed positive sera, the test correctly identified 132, but failed in 23 cases. Out of the 142 control sera, rK39 RT identified 140 sera, demonstrating 98.6% specificity. The test was false positive in two cases, one patient suffering from malaria and the other from leukaemia. These two sera showed negative results with DAT. The PPV and NPV were 98.5% and 85.9%, respectively [Table 1].{Table 1}

Out of the 155 confirmed VL sera, 126 (81.3%) showed strong antibody titres (≥1:6 400) with DAT [Table 2]. The test recorded negative antibody titers (≤1:800) in 1 serum positively tested with rK39 RT and in 5 sera that were also negative with rK39 RT. Likewise, DAT recorded weak positive antibody titers in 12 confirmed VL sera that were positive with rK39 RT and 3 sera negative with rK39 RT. There was moderate agreement between DAT and rK39 RT (κ=0.42, 95% CI=0.21-0.63) for confirmed VL patients [Table 2] and control sera (κ=0.80, 95% CI=0.41-1.00) [Table 3].{Table 2}{Table 3}

In the 197 sera of VL-suspected [Table 4], 13 were positive by the two tests and 175 were negative. Four sera were positive by DAT and negative by rK39 RT while five were positive by rK39 and negative by DAT. The agreement between these two tests in the VL-suspected group with unconfirmed diagnosis was also good (κ=0.72, 95% CI=0.54-0.89).{Table 4}

 4. Discussion

In this study, we compared the diagnostic accuracy and agreement between DAT and rK39 rapid tests for VL diagnosis in sera from patients originating from a known endemic region in Eastern Sudan. This region is similar to many other regions in East Africa where the disease has been identified. The assessment of diagnostic tests of VL in these regions is important, as the disease remains a major health problem and many environmental factors favour the disease transmission[23],[24],[25].

DAT and rK39 RT have been previously evaluated in several studies from different countries with considerable variable results[26],[27]. The present study provides evidence of good agreement between DAT and rK39 RT for VL diagnosis in Eastern Sudan. It is essential to validate diagnostic tests in areas where it will find it´s application. Here we show that the sensitivity and specificity of DAT and rK39 are acceptable but not optimal. Previously, we have conducted multicenter study, which revealed that compared to other tests, DAT has a better sensitivity for VL detection in immunocompetent HIV-negative patients in Sudan and India but not in France (94%, 92.4% and 88.5%, respectively). On the other hand, sensitivity of rK39 RT was better in India (96.2%) than Sudan (88%) and France (85%)[18]. Here, we used different DAT (DAT prepared at Laboratory for Biomedical Research-Ahfad University) and rK39 RT obtained from the same source as in the multicenter study (Bio Rad-France). This study confirms variations in the diagnostic accuracy of DAT and rK39 RT, as reported previously[18]. Indeed, HIV can lower sensitivity of the test[27]. This is particularly important, since there is high prevalence of VL/HIV co-infection in East Africa[28],[29],[30]. Heterogeneity in Leishmania parasites is an additional factor that complicates diagnosis of VL[27],[28],[29],[30],[31]. Indeed, combining results of more than one test would increase sensitivity of the final diagnosis.

However, in resource-limited countries like East Africa, this approach seems not to be possible.

To compare accuracy of DAT and rK39 RT with regard to specificity, we extended our study design by also including patients suffering from malaria, pulmonary tuberculosis, leukaemia and African trypanosomiasis, all being confirmed non-VL cases. Of particular importance was the concordance of DAT and rK39 RT in healthy controls and unconfirmed VL-suspected cases. As the incidence of VL is very high in East Sudan[32],[33], serological tests can be used to discriminate between these two important groups without the necessity to perform invasive parasitological diagnostic methods. This approach is used in many health centres in Sudan because it allows quick and easy interpretation of test results[34],[35]. As reported previously, some individuals from endemic areas are positive in serological tests without any signs and symptoms of VL, while some true positive patients don’t have detectable antibody titres[27]. These results may explain the false positive and negative results in our study. It is known that the majority of these patients were not HIV co-infected but other immunocompromising factors cannot be ruled out; e.g., the high prevalence of malnutrition in East Africa may influence the immunity, a question that needs urgent attention in order to reinterpret a low or absent antibody responses in infected patients[36].

All this underscores the need for a combined use of DAT and rK39 RT for diagnosis of VL in rural settings at Eastern Sudan. However, additional methods are required to confirm highly suspected cases with negative serology. ELISA tests based on Leishmania recombinant proteins such as the KLO8 might be good alternatives[18].

It is important to continue developing new diagnostics for VL in Sudan and other endemic countries. Recombinant proteins are very promising to improve sero-diagnosis and discern between false negative and false positive test results. With the high prevalence of asymptomatic cases and co-existence with other diseases such as malaria, care should be taken to identify true VL infected patients to avoid undesired effects of the drugs, including pancreatitis, liver and heart problems and cardiac arrhythmia[37],[38].

The limitations of the study are related to its retrospective nature based on previous findings. The sera of this investigation were already used in previous studies, collected from patients, transported and stored for years under freezing condition. The sera had exposed to freezing and thawing. This may affect the quality and test results.

The accuracy and agreement between DAT and rK39 RT were satisfactory for diagnosis of VL in East Sudan. Their simplicity and fastness make both tests suitable for primary healthcare in rural areas. These data are relevant also for other endemic countries in East Africa because of the overlapping distribution of the L. donovani parasite.

Conflict of interest statement

The authors declare that there is no conflict of interest.


We acknowledge the research team at Laboratory of Biomedical Research (LBR) of Ahfad University for Women in Sudan for providing registered data. Contributors included: Professor Abdallah el Harith, Dr. Durria Mansour, Dr. Mohamed el Mutasim, Dr. Abdelhafeiz Mahamoud, Dr. Hussam Osman.


1Welay GM, Alene KA, Dachew BA. Visceral leishmaniasis treatment outcome and its determinants in Northwest Ethiopia. Epidemiol Health 2017; 39: e2017001.
2World Health Organization. Leishmaniais. [Online]. Available from: [Accessed on 17th November 2019].
3Aschale Y, Ayehu A, Worku L, Testa H, Birhanie M, Lemma W. Malaria-visceral leishmaniasis co-infection and associated factors among migrant laborers in West Armachiho district, North West Ethiopia: Community based cross-sectional study. BMC Infect Dis 2019; 19: 239.
4Burza S, Croft SL, Boelaert M. Leishmaniasis. Lancet 2018; 392(10151): 951-970.
5Diro E, Lynen L, Ritmeijer K, Boelaert M, Hailu A, van Griensven J. Visceral leishmaniasis and HIV coinfection in East Africa. PLoS Negl Trop Dis 2014; 8(6): e2869.
6Mahamoud A, Osman HA, Abass EM, Agib AE, Madi RR, Semiao-Santos SJ, et al. Identification of an area predominantly endemic for childhood and adolescent visceral leishmaniasis in central Sudan. Acta Trop 2018; 178: 142-147.
7Mueller YK, Nackers F, Ahmed KA, Boelaert M, Djoumessi JC, Eltigani R, et al. Burden of visceral leishmaniasis in villages of Eastern Gedaref State, Sudan: An exhaustive cross-sectional survey. PLoS Negl Trop Dis 2012; 6(11): e1872.
8Elnaiem DE. Ecology and control of the sand fly vectors of Leishmania donovani in East Africa, with special emphasis on Phlebotomus orientalis. J Vector Ecol 2011; 36(Suppl 1): S23-S31.
9Harith AE, Mahamoud A, Awad Y, Mansour D, Abass EM, Agib AE, et al. Are we now well prepared for another major visceral leishmaniasis epidemic in Sudan? Open Forum Inf Dis 2019; doi: org/10.1093/ofid/ ofz226.
10Mukhtar M, Ali SS, Boshara SA, Albertini A, Monnerat S, Bessell P, et al. Sensitive and less invasive confirmatory diagnosis of visceral leishmaniasis in Sudan using loop-mediated isothermal amplification (LAMP). PLoS Negl Trop Dis 2018; 12 (2): e0006264.
11Zijlstra EE, AM Siddig, AM El-Hassan, IA El-Toum, M Satti, et al. Kalaazar: A comparative study of parasitological methods and the direct agglutination test in diagnosis. Trans R Soc Trop Med Hyg 1992; 86: 505507.
12Srivastava P, Dayama A, Mehrotra S, Sundar S. Diagnosis of visceral leishmaniasis. Trans R Soc Trop Med Hyg 2011; 105: 1-6.
13Veeken H, Ritmeijer K, Seaman J, Davidson R. Comparison of an rK39 dipstick rapid test with direct agglutination test and splenic aspiration for the diagnosis of kala-azar in Sudan. Trop Med Int Health 2003; 8: 164-167.
14Romero HD, Silva LdeA, Silva-Vergara ML, Rodrigues V, Costa RT, Guimarães SF, et al. Comparative study of serologic tests for the diagnosis of asymptomatic visceral leishmaniasis in an endemic area. Am J Trop Med Hyg 2009; 81: 27-33.
15Pattabhi S, Whittle J, Mohamath R, El-Safi S, Moulton GG, Guderian JA, et al. Design, development and evaluation of rK28-based point-of-care tests for improving rapid diagnosis of visceral leishmaniasis. PLoS Negl Trop Dis 2010; 4(9): e822.
16Abass EM, Mansour D, El Mutasim M, Hussein M, El Harith A. Beta-mercaptoethanol modified ELISA for diagnosis of visceral leishmaniasis. J Med Microbiol 2006; 55: 1193-1196.
17Mutasim M, Mansour D, Abass EM, Hassan WM, El Harith A. Evaluation of a glycerol preserved antigen in the direct agglutination test for diagnosis of visceral leishmaniasis at rural level in eastern Sudan. J Med Microbiol 2006; 55: 1343-1347.
18Abass E, Bollig N, Reinhard K, Camara B, Mansour D, Visekruna A, et al. rKLO8, a novel Leishmania donovani-derived recombinant immunodominant protein for sensitive detection of visceral leishmaniasis in Sudan. PLoS Negl Trop Dis 2013; 7: e2322.
19Ministry of Health, Republic of Sudan, Neglected Tropical Disease Division. Manual for the diagnosis and treatment of leishmaniasis. 2014. [Online]. Available from: Manual_for_the_diagnosis_and_treatment_Leishmaniasis_Guideline_ Sudan_2014.pdf?ua=1. [Accessed on 21st August 2019].
20Harith AE, Chowdhury S, Al-Masum A, Semião-Santos SJ, Karim E, Elsafi S, et al. Evaluation of cleaving agents other than trypsin in direct agglutination test for further improving diagnosis of visceral leishmaniasis. J Clin Microbiol 1995; 33: 1984-1988.
21Burns JM Jr, Shreffler WG, Benson DR, Ghalib HW, Badaro R, Reed SG. Molecular characterization of a kinesin-related antigen of Leishmania chagasi that detects specific antibody in African and American visceral leishmaniasis. Proc Natl Acad Sci U S A 1993; 90: 775-779.
22Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33: 159-174.
23Diro E, van Griensven J, Mohammed R, Colebunders R, Asefa M, Hailu A, et al. Atypical manifestations of visceral leishmaniasis in patients with HIV in north Ethiopia: A gap in guidelines for the management of opportunistic infections in resource poor settings. Lancet Infect Dis 2015; 15: 122-129.
24Oryan A, Akbari M. Worldwide risk factors in leishmaniasis. Asian Pac J Trop Med 2016; 9: 925-932.
25Reithinger R, Brooker S, Kolaczinski JH. Visceral leishmaniasis in eastern Africa-current status. Trans R Soc Trop Med Hyg 2007; 101: 1169-1170.
26World Health Organization. Visceral leishmaniasis rapid diagnostic test performance, [Online]. Available from:; [Accessed on 20th June 2019].
27Abass E, Kang C, Martinkovic F, Semião-Santos SJ, Sundar S, Walden P, et al. Heterogeneity of Leishmania donovani parasites complicates diagnosis of visceral leishmaniasis: Comparison of different serological tests in three endemic regions. PloS One 2015; 10: e0116408.
28Cota GF, de Sousa MR, Rabello A. Predictors of visceral leishmaniasis relapse in HIV-infected patients: A systematic review. PLoS Negl Trop Dis 2011; 5: e1153.
29ter Horst R, Collin SM, Ritmeijer K, Bogale A, Davidson RN. Concordant HIV infection and visceral leishmaniasis in Ethiopia: The influence of antiretroviral treatment and other factors on outcome. Clin Infect Dis 2008; 46: 1702-1709.
30Abass E, Elhussein M. Visceral leishmaniais and HIV co-infection in patients referred to Ahfad Biomedical Research Laboratory of Ahfad University for women-Sudan during 2003-2007. Annals of medicine and healthcare research. Proceedings of the 2009 International Online Medical Conference. Boca Raton, Florida, USA: Universal Publishers; 2009.
31Forgber M, Basu R, Roychoudhury K, Theinert S, Roy S, Sundar S, et al. Mapping the antigenicity of the parasites in Leishmania donovani infection by proteome serology. PLoS One 2006; e40. doi: org/10.1371/ journal.pone.0000040.
32Alvar J, Velez ID, Bern C, Herrero M, Desjeux P, Cano J, et al. Leishmaniasis worldwide and global estimates of its incidence. PLoS One 2012; 7(5) :e35671.
33Al-Salem W, Herricks JR, Hotez PJ. A review of visceral leishmaniasis during the conflict in South Sudan and the consequences for East African countries. Parasites & Vectors 2016; 9(1): 460.
34Veeken H, Ritmeijer K, Seaman J, Davidson R. Comparison of an rK39 dipstick rapid test with direct agglutination test and splenic aspiration for the diagnosis of kala-azar in Sudan. Trop Med Int Health 2003; 8: 164-167.
35Osman HA, Mahamoud A, Abass EM, Madi RR, Santos SJS, Harith AE. Local production of a liquid direct agglutination test as a sustainable measure for control of visceral leishmaniasis in Sudan. Am J Trop Med Hyg 2016; 4: 982-986.
36Bhattacharyya T, Boelaert M, Miles MA. Comparison of visceral leishmaniasis diagnostic antigens in African and Asian Leishmania donovani reveals extensive diversity and region-specific polymorphisms. PLoS Negl Trop Dis 2013; 7: e2057.
37Lage PS, de Andrade PH, Lopes Ade S, Chavez Fumagalli MA, Valadares DG, Duarte MC, et al. Strychnos pseudoquina and its purified compounds present an effective in vitro antileishmanial activity. Evid Based Complement Alternat Med 2013; 2013: 304354.
38Diro E, Lynen L, Mohammed R, Boelaert M, Hailu A, van Griensven J. High parasitological failure rate of visceral leishmaniasis to sodium stibogluconate among HIV co-infected adults in Ethiopia. PLoS Negl Trop Dis 2014; 8: e2875.