Comparative performance of four rapid Ebola antigen-detection lateral flow immunoassays during the 2014-2016 Ebola epidemic in West Africa

BACKGROUND: Without an effective vaccine, as was the case early in the 2014–2016 Ebola Outbreak in West Africa, disease control depends entirely on interrupting transmission through early disease detection and prompt patient isolation. Lateral Flow Immunoassays (LFI) are a potential supplement to centralized reference laboratory testing for the early diagnosis of Ebola Virus Disease (EVD). The goal of this study was to assess the performance of commercially available simple and rapid antigen detection LFIs, submitted for review to the WHO via the Emergency Use Assessment and Listing procedure. The study was performed in an Ebola Treatment Centre laboratory involved in EVD testing in Sierra Leone. In light of the current Ebola outbreak in May 2018 in the Democratic Republic of Congo, which highlights the lack of clarity in the global health community about appropriate Ebola diagnostics, our findings are increasingly critical. METHODS: A cross-sectional study was conducted to assess comparative performance of four LFIs for detecting EVD. LFIs were assessed against the same 328 plasma samples and 100 whole EDTA blood samples, using the altona RealStar Filovirus Screen real-time RT-PCR as the bench mark assay. The performance of the Public Health England (PHE) in-house Zaire ebolavirus-specific real time RT-PCR Trombley assay was concurrently assessed. Statistical analysis using generalized estimating equations was conducted to compare LFI performance. FINDINGS: Sensitivity and specificity varied between the LFIs, with specificity found to be significantly higher for whole EDTA blood samples compared to plasma samples in at least 2 LFIs (P≤0.003). Using the altona RT-PCR assay as the bench mark, sensitivities on plasma samples ranged from 79.53% (101/127, 95% CI: 71.46–86.17%) for the DEDIATEST EBOLA (SD Biosensor) to 98.43% (125/127, 95% CI: 94.43–99.81%) for the One step Ebola test (Intec). Specificities ranged from 80.20% (158/197, 95% CI: 74.07–88.60%) for plasma samples using the ReEBOV Antigen test Kit (Corgenix) to 100.00% (98/98, 95% CI: 96.31–100.00%) for whole blood samples using the DEDIATEST EBOLA (SD Biosensor) and SD Ebola Zaire Ag (SD Biosensor). Results also showed the Trombley RT-PCR assay had a lower limit of detection than the altona assay, with some LFIs having higher sensitivity than the altona assay when the Trombley assay was the bench mark. INTERPRETATION: All of the tested EVD LFIs may be considered suitable for use in an outbreak situation (i.e. rule out testing in communities), although they had variable performance characteristics, with none possessing both high sensitivity and specificity. The non-commercial Trombley Zaire ebolavirus RT-PCR assay warrants further investigation, as it appeared more sensitive than the current gold standard, the altona Filovirus Screen RT-PCR assay.