Abstract
Diagnosis of a first-time visceral leishmaniasis (VL) infection in Ethiopia is established by use of a rapid diagnostic test (RDT) detecting antibodies against rK39, direct agglutination test (DAT) and microscopy according to the national algorithm. The performance of individual tests and algorithm is variable and depends on several factors, one being HIV status. Limited data are available on the performance of tests in VL-HIV coinfected patients. Assessment of the performance of DAT (ITM-A), rK39 ELISA (Serion) and six RDT (Onsite Leishmania Ab CTK, Antigen ICT Xinjier, IT Leish Biorad, Kalazar Detect Inbios, rK39 IgG1 Coris, rk28 IgG1 Coris) for the diagnosis of VL was done on a panel of 91 stored serum and plasma samples of 'first-episode' suspected VL patients, with HIV coinfection (n = 51) and without (n = 40). A combined reference standard was used: either positive microscopy on tissue aspirates, or in case of negative microscopy, positive PCR results on the aspirate slide. Additionally, endemic healthy controls (n = 20), non-endemic controls (n = 10) and patients with confirmed malaria infection (n = 10) were tested for specificity evaluation. Sensitivities ranged from 69.2% for DAT (applied cut-off ≥ 1/3200) to 92.2% for the Onsite RDT, whereas specificities ranged from 20.0% for Kalazar Antigen ICT to 100% for IT Leish and rK39 IgG1. Sensitivities from all assays decreased upon stratification according to HIV status but was only significantly different for rK39 Serion ELISA (p-value 0.0084) and the Onsite RDT (p-value 0.0159). In conclusion, performance of commercially available assays for VL on samples from Northern-Ethiopian patients varied widely with a substantial decrease in sensitivity in the VL-HIV coinfected group. Clear guidelines on minimal performance criteria of individual tests and algorithms are needed, as well as which reference standard should be used to determine the performance.