Background: Diagnostics are key for the detection and containment of filovirus disease outbreaks. Beyond lab-based nucleic acid and immunoassays, the 2013-2016 West Africa Ebola viral disease (EVD) outbreak saw a growth of antigen-based lateral flow rapid diagnostic tests (Ag RDTs) for use in the field. Despite concerns about their lower sensitivity, the COVID-19 pandemic demonstrated that widespread use of Ag RDTs can identify more cases and support patient isolation and end-of-quarantine activities. Considering that EVD outbreaks continue to occur in Central and Eastern Africa, this thesis aims to understand the utility of current EBOV Ag RDTs for outbreak response.
Findings: We first performed a systematic review and meta-analysis on the performance of EBOV Ag RDTs, searching for studies published between 1976 and 2023. We estimated the pooled sensitivity and specificity were 82.1% (95%CI: 75.2 -88.0) and 97.0% (95%CI: 95.1 – 98.2), respectively, with a high positive predictive value (94%). Next, we evaluated eight EBOV Ag RDTs to assess their ability to detect EBOV and SUDV. Analytical panels using viral cell slurries of EBOV, Sudan virus (SUDV), and Crimean-Congo Hemorraghic Fever virus (CCHFV) were used to assess limit of detection; clinical samples from previous EBOV and SUDV outbreaks were included to determine sensitivity and specificity. We found five Ag RDTs that detected EBOV and three that detected SUDV, although clinical sensitivity was low (EBOV: 20-40%, SUDV: 33%), improving only with higher viral loads. All Ag RDTs demonstrated 100% clinical specificity with no cross-reactivity to CCHFV. Lastly, we developed a mathematical compartmental model calibrated to the 2013-2016 EVD outbreak in Sierra Leone to identify trade-offs in Ag RDT accuracy, time-to-isolation, and accessibility for EVD outbreak response. Model parameters included test accuracy, accessibility, and time-to-isolation and were varied to evaluate their impact on total epidemic size in isolation and in combination. We found that reductions in EBOV Ag RDT test sensitivity or specificity alone increases the expected number of cases between 11.7-223%, while any decrease in time-to-isolation alone or increase in testing rate alone would decrease the expected number of cases by 47.7%-87.7%. When combing the three factors, we found a combined net reduction of mean cases between 71.6-92.3%.
Discussion: The picture that emerges is one where current, commercially available Ag RDTs continue to perform suboptimally. However, modeling suggests that wider testing access with Ag RDTs, which can provide rapid results and be rolled out more broadly compared to PCR testing, can compensate for lower clinical performance. While confirmatory testing is still needed as current Ag RDTs cannot be used to rule out disease in symptomatic individuals, the findings from this thesis supports a more strategic integration of Ag RDTs into EVD outbreak response activities, and suggests that under the right conditions, Ag RDTs could be deployed for initial triage in high-incidence zones (e.g. Ebola treatment centers), cadaver screening, surge capacity during periods of diagnostic overload, and decentralized surveillance in high-risk areas.
