Background At present, diagnosis of Ebola virus disease requires transport of venepuncture blood to field biocontainment laboratories for testing by real-time RT-PCR, resulting in delays that complicate patient care and infection control efforts. Therefore, an urgent need exists for a point-of-care rapid diagnostic test for this disease. In this Article, we report the results of a field validation of the Corgenix ReEBOV Antigen Rapid Test kit. Methods We performed the rapid diagnostic test on fingerstick blood samples from 106 individuals with suspected Ebola virus disease presenting at two clinical centres in Sierra Leone. Adults and children who were able to provide verbal consent or assent were included; we excluded patients with haemodynamic instability and those who were unable to cooperate with fingerstick or venous blood draw. Two independent readers scored each rapid diagnostic test, with any disagreements resolved by a third. We compared point-of-care rapid diagnostic test results with clinical real-time RT-PCR results (RealStar Filovirus Screen RT-PCR kit 1·0; altona Diagnostics GmbH, Hamburg, Germany) for venepuncture plasma samples tested in a Public Health England field reference laboratory (Port Loko, Sierra Leone). Separately, we performed the rapid diagnostic test (on whole blood) and real-time RT-PCR (on plasma) on 284 specimens in the reference laboratory, which were submitted to the laboratory for testing from many clinical sites in Sierra Leone, including our two clinical centres. Findings In point-of-care testing, all 28 patients who tested positive for Ebola virus disease by RT-PCR were also positive by fingerstick rapid diagnostic test (sensitivity 100% [95% CI 87·7-100]), and 71 of 77 patients who tested negative by RT-PCR were also negative by the rapid diagnostic test (specificity 92·2% [95% CI 83·8-97·1]). In laboratory testing, all 45 specimens that tested positive by RT-PCR were also positive by the rapid diagnostic test (sensitivity 100% [95% CI 92·1-100]), and 214 of 232 specimens that tested negative by RT-PCR were also negative by the rapid diagnostic test (specificity 92·2% [88·0-95·3]). The two independent readers agreed about 95·2% of point-of-care and 98·6% of reference laboratory rapid diagnostic test results. Cycle threshold values ranged from 15·9 to 26·3 (mean 22·6 [SD 2·6]) for the PCR-positive point-of-care cohort and from 17·5 to 26·3 (mean 21·5 [2·7]) for the reference laboratory cohort. Six of 16 banked plasma samples from rapid diagnostic test-positive and altona-negative patients were positive by an alternative real-time RT-PCR assay (the Trombley assay); three (17%) of 18 samples from individuals who were negative by both the rapid diagnostic test and altona test were also positive by Trombley. Interpretation The ReEBOV rapid diagnostic test had 100% sensitivity and 92% specificity in both point-of-care and reference laboratory testing in this population (maximum cycle threshold 26·3). With two independent readers, the test detected all patients who were positive for Ebola virus by altona real-time RT-PCR; however, this benchmark itself had imperfect sensitivity. Funding Abundance Foundation.
Bibliographical noteFunding Information:
Funding was provided by a gift from the Abundance Foundation (Stephen Kahn). Corgenix provided the test kits only. Neither source had any role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit for publication. The corresponding author had full access to all the data in this study and had final responsibility for the decision to submit for publication.
Funding for this study was provided by a gift from the Abundance Foundation (Stephen Kahn). Corgenix provided test kits for the study but did not provide any monetary support; Corgenix personnel were not involved in data acquisition or analysis. We thank John Schieffelin for his assistance with protocol submission in Sierra Leone and with provision of test kits. We acknowledge the entire staff of the Public Health England Port Loko laboratory and the Partners In Health clinical sites at Maforki and Kambia for their contributions to study execution. We thank Kevin Savage for his contributions to protocol submissions, Sid Atwood, David Williams, Chuan-Chin Huang, and Farzad Noubary for help with statistical analysis, and Neill Keppie for his help with Public Health England database queries. We thank the leadership of the Department for International Development for their support of Public Health England laboratory study activities.