Field performance of three Ebola rapid diagnostic tests used during the 2018-20 outbreak in the eastern Democratic Republic of the Congo: a retrospective, multicentre observational study.

BACKGROUND

The Democratic Republic of the Congo has confronted 13 outbreaks of Ebola virus disease since 1976. Rapid diagnostic tests (RDTs) detecting viral antigens have been developed to circumvent difficulties encountered with RT-PCR for diagnosis in remote low-resource settings, but there is still uncertainty about their performance characteristics and usability during outbreaks. We aimed to assess the field performance of three antigen detection RDTs compared with the gold-standard Cepheid GeneXpert Ebola assay results.

METHODS

We conducted a retrospective, multicentre observational study using complete and de-identified databases of five mobile laboratories (managed by the Institut National de Recherche Biomédicale) to assess the performance of three Ebola virus disease RDTs (QuickNavi-Ebola, OraQuick Ebola Rapid Antigen Test, and Coris EBOLA Ag K-SeT rapid test) run on blood samples of patients with suspected Ebola virus disease in direct comparison with the Cepheid GeneXpert Ebola assay reference test during the 2018-20 outbreak in the eastern Democratic Republic of the Congo. We estimated the sensitivity and specificity of each test through generalised linear mixed models against the GeneXpert Ebola assay reference test and corrected for cycle threshold value and random site effects.

FINDINGS

719 (7·9%) of 9157 samples had a positive GeneXpert Ebola assay result. The QuickNavi-Ebola RDT had a sensitivity of 87·4% (95% CI 63·6-96·8) around the mean cycle threshold value and a specificity of 99·6% (99·3-99·8). The OraQuick Ebola Rapid Antigen Test had a sensitivity of 57·4% (95% CI 38·8-75·8) and specificity of 98·3% (97·5-99·0), and the Coris EBOLA Ag K-SeT rapid test had a sensitivity of 38·9% (23·0-63·6) against the GeneXpert Ebola assay reference and specificity of 97·4% (85·3-99·6). The QuickNavi-Ebola RDT showed a robust performance with good sensitivity, particularly with increasing viral loads (ie, low cycle threshold values), and specificity.

INTERPRETATION

The three RDTs evaluated did not achieve the desired sensitivity and specificity of the WHO target product profile. Although the RDTs cannot triage and rule out Ebola virus infection among clinical suspects, they can still help to sort people with suspected Ebola virus disease into high-risk and low-risk groups while waiting for GeneXpert Ebola assay reference testing.

FUNDING

None.

TRANSLATION

For the French translation of the abstract see Supplementary Materials section.