Rapid and accurate diagnosis of visceral leishmaniasis (VL) is needed to initiate prompt treatment to reduce morbidity and mortality. Here, we evaluated the performance of loop-mediated isothermal amplification (LAMP) assay for the diagnosis of VL from blood in an endemic area in Ethiopia. LAMP was positive in 117/122 confirmed VL cases and negative in 149/152 controls, resulting in a sensitivity of 95.9% (95% CI: 90.69–98.66) and a specificity of 98.0% (95% CI: 94.34–99.59), respectively. The sensitivity of the LAMP assay was 95.0% (95% CI: 88.61–98.34) in HIV-negatives and 100% (95% CI: 85.18–100.0) in HIV-positives. Compared with microscopy, LAMP detected 82/87 (94.3%, 95% CI: 87.10–98.11) of the microscopy+ cases and was negative in 11/27 (40.7%, 95% CI: 22.39–61.20) of the microscopy− cases. Compared with the rK39 serology, LAMP detected 113/120 (94.2%, 95% CI: 88.35–97.62) of the rK39+ cases and was negative in 149/154 (96.8%, 95% CI: 92.59–98.94) of the rK39− cases. However, when compared with microscopy only, rK39 detected 83/87 (95.4%, 95% CI: 88.64–98.73) of the microscopy+ cases and negative in only 12/27 (44.4%, 95% CI: 25.48–64.67) of the microscopy– cases. There was an excellent agreement between rK39 and LAMP (Kappa = 0.91, 95% CI: 0.86–0.96). Furthermore, an algorithm using rK39 followed by LAMP would yield a sensitivity of 99.2% (95%CI: 95.52–99.89) and a specificity of 98.0% (95% CI: 94.34–99.59). The findings demonstrate that LAMP assay is an accurate and rapid molecular assay for VL diagnosis, including in HIV-1 coinfected patients, in an endemic setting.

Background The rapid diagnostic test (RDT) rK39 is currently being used for routine diagnosis of visceral leishmaniasis (VL) in East Africa. However, continuous monitoring of the performance of the assay, in particular its impact on the clinical decision in initiating anti-leishmanial treatment and outcomes remains needed as there are concerns about the diagnostic performance of this test. Methods VL patients prospectively enrolled in a diagnostic trial and with rK39 RDT were included. We evaluated the effect of rK39 testing in guiding treatment initiation and outcome. On the basis of rK39 RDT test result as well as clinical case definition for VL and microscopy examination, the clinicians decide whether to initiate VL therapy or not. Poisson regression models were used to identify factors associated with a decision to initiate VL therapy. In addition, treatment outcomes of those who received VL therapy were compared to those who received non-VL treatment. Results Of 324 VL suspects enrolled, 184 (56.8%) were rK39+ and 140 (43.2%) were rK39‒. In addition, microscopy exam was done on tissue aspirates from a sub-population (140 individuals), which is 43.2% of the suspected cases, comprising of 117 (63.6%) rK39+ and only 23 (16.4%) rK39‒ cases. Of those with microscopy examination, only 87 (62.1%) were found positive. Among 184 (56.8%) patients without microscopy, 67 (36.4%) were rK39+, of whom 83 (65.9%) were positive by microscopy, 21 (16.7%) were negative by microscopy and 22 (17.5%) had no microscopy results. On the other hand, of those who did not receive VL treatment 58/189 (30.7%) were rK39+ and 131 (69.3%) were rK39‒. Of the rK39+ cases who did not receive VL therapy, only 1 (1.7%) patient was microscopy positive, 12 (20.7%) were negative and 45 (77.6%) patients had no microscopy result. Of the rK39‒ cases (n = 131) who did not receive VL treatment, 16 were microscopy negative and 115 without microscopy exams. Whereas positive rK39 result [adjusted Relative Risk (aRR) 0.69; 95% CI: 0.49–0.96, p = 0.029] and positive microscopy results (aRR 0.03; 95% CI: 0.00–0.24, p = 0.001) were independently associated with VL treatment, having confirmed diagnosis other than VL (aRR 1.64; 95% CI: 1.09–2.46, p = 0.018) was independently associated with initiation of non-VL therapy. The proportion of rK39+ patients who received non-VL treatment with no improvement outcome was significantly higher when compared to those who received VL treatment (24.1%, 95% CI: 14.62–37.16 vs. 11.9%, 95%CI: 7.26–18.93; p<0.0001). Conclusion The study shows that a significant proportion of patients with rK39+ results were undertreated. Failure to do microscopy was associated with lack of improved clinical outcome. Including an additional simple point-of-care assay in the diagnostic work-up is urgently needed to correctly identify VL cases and to prevent morbidity and mortality associated with the disease.