Diagnostic performance of rK28- and rK39-based rapid diagnostic tests for Visceral Leishmaniasis in Kenya: Implications for the national diagnostic algorithm

Read the full article See related articles

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

Background

Visceral Leishmaniasis (VL) is a neglected tropical disease (NTD) caused by Leishmania infantum and L. donovani and transmitted by a bite of an infected female phlebotomine sandfly. The disease disproportionately affects impoverished and underserved populations, where inadequate access to healthcare contributes to delayed diagnosis, treatment, and increased morbidity and mortality. In Kenya, 11 out of the 47 counties are known to be endemic for VL. Visceral Leishmaniasis is fatal if it’s not timely diagnosed and treated. Parasitological methods remain the gold standard for diagnosing and confirming VL, but they are invasive and require specialised training and equipment. Serological tests in the form of Rapid Diagnostic Tests (RDTs) are widely used for initial screening and detection but can still report positive results for long periods after curative treatment due to the persistence of anti-VL antibodies. It is recommended that VL diagnosis follows an algorithmic approach. This study evaluated the diagnostic performance of rK28 and rK39 RDTs (Kala-azar Detect and IT Leish) in various algorithmic combinations against a composite reference standard in Kenya.

Methodology

We evaluated rK28 and Kala-azar Detect along with rK39 IT-Leish as index tests, and Direct Agglutination Test (DAT) as the reference standard test. Participants were recruited from seven health facilities located in V L endemic regions of Marsabit and Turkana counties, Kenya. Blood samples were collected, as well as patient demographic data. Treatment decisions followed the national diagnostic algorithm and health facility routine procedures. Diagnostic test performance was assessed individually, in parallel, and based on sequential algorithmic combinations. Specificity, sensitivity, balanced accuracy, and negative and positive predictive values were statistically estimated.

Results

A total of 398 participants were recruited for the study, with fever and splenomegaly being the most reported clinical symptoms. Individually, the rK28 kit demonstrated a sensitivity of 96%, a specificity of 86%, and a balanced accuracy (BA) of 91.2%. The rK39 IT-Leish kit showed a sensitivity of 87%, a specificity of 93%, and a BA of 90.4%, while the rK39 Kala-azar Detect showed a sensitivity of 67%, a specificity of 91%, and a BA of 79%. The rK39 IT Leish to rK28 algorithm based on negatives had the highest balanced accuracy of 90% with a sensitivity of 92% and a specificity of 88%, while the rK28 to rK39 Kala-azar detect algorithm based on positives showed the lowest balanced accuracy of 80% with a sensitivity of 65% and a specificity of 94.1%.

Discussion

The rK28 RDT demonstrated consistently high sensitivity (96–98%) and moderate specificity (72–86%), findings broadly consistent with reports from other East African settings. In contrast, the rK39-based tests showed a higher specificity (91–93%) but lower sensitivity (67–87%), which limits their effectiveness for elimination-focused programs that lean towards not missing a positive case. Parallel testing algorithms that combine the rK28 and rK39 RDTs improved sensitivity (up to 97.7%), whereas sequential (based on confirming negative results) approaches increased specificity at the expense of sensitivity when compared to the rK28 RDT alone. These approaches were more balanced as compared to sequential (based on confirming positive results), which increased the overall specificity but greatly impacted the sensitivity. This flexibility in combinations provides more options for programs based on risks and overall goals. The performance of the rK28 RDT gives it an advantage over the rK39 kits, especially with rK39 IT Leish kits facing stockouts. While the study faced limitations of lack of PCR data that led to the reliance on DAT as the reference standard, as well as underpowered secondary analyses, the findings point towards strong preliminary evidence that can be used to guide updating of the national algorithm, supply chain strengthening, and policy decisions that may accelerate VL elimination in the country and the region.

Conclusion

The findings demonstrate that diagnostic algorithms incorporating rK28 provide robust performance for VL diagnosis. These results support consideration of the rK28 diagnostic kit as a valuable component of VL diagnostic strategies, with the potential to enhance case detection and contribute to VL control and elimination efforts in Eastern Africa.

Article activity feed