Discovery and validation of a 3-gene signature to distinguish COVID-19 and other viral infections in emergency infectious disease presentations; a case-control then observational cohort study
File(s)1-s2.0-S2666524721001452-main.pdf (997.92 KB) Supplementary_LANCETMICROBE.docx (2.73 MB)
Published version
Supporting information
Author(s)
Type
Journal Article
Abstract
Background: Emergency admissions for infection often lack initial diagnostic certainty. COVID-19 has highlighted a need for novel diagnostic approaches to indicate likelihood of viral infection in a pandemic setting. We sought to derive and validate a blood transcriptional signature to detect viral infections including COVID-19 among adults with suspected infection presenting to the Emergency Department (ED).
Methods: Blood RNA sequencing was performed on a discovery cohort of adults attending the ED with suspected infection who had subsequently-confirmed viral, bacterial, or no infection diagnoses. Differentially expressed host genes were subjected to feature selection to derive the most parsimonious discriminating signature. RT-qPCR validation of the signature was then performed in a prospective cohort of ED patients presenting with undifferentiated fever, and a second case-control cohort of ED patients with COVID-19 or bacterial infection. Signature performance was assessed by calculating area under receiver-operating characteristic curves (AUC-ROCs), sensitivities, and specificities.
Findings: A 3-gene transcript signature was derived from the discovery cohort of 56 bacterial and 27 viral infection cases. In the validation cohort of 200 cases, the signature differentiated bacterial from viral infections with an AUC-ROC of 0.976 (95% CI: 0.919-1.000), sensitivity 97.3% and specificity of 100%. The AUC-ROC for C-reactive protein (CRP) and leucocyte count (WCC) was 0.833 (95% CI: 0.694-0.944) and 0.938 (95% CI: 0.840-0.986) respectively. The signature achieved higher net benefit in decision curve analysis than either CRP or WCC for discriminating viral infections from all other cases. In the second validation analysis the signature discriminated 35 bacterial infections from 34 SARS-CoV-2 positive COVID-19 infections with AUC-ROC of 0.953 (95% CI: 0.893-0.992), sensitivity 88.6% and specificity of 94.1%.
Interpretation: This novel 3-gene signature discriminates viral infections including COVID-19 from other emergency infection presentations in adults, outperforming both WCC and CRP, thus potentially providing significant clinical utility in managing acute presentations with infection.
Methods: Blood RNA sequencing was performed on a discovery cohort of adults attending the ED with suspected infection who had subsequently-confirmed viral, bacterial, or no infection diagnoses. Differentially expressed host genes were subjected to feature selection to derive the most parsimonious discriminating signature. RT-qPCR validation of the signature was then performed in a prospective cohort of ED patients presenting with undifferentiated fever, and a second case-control cohort of ED patients with COVID-19 or bacterial infection. Signature performance was assessed by calculating area under receiver-operating characteristic curves (AUC-ROCs), sensitivities, and specificities.
Findings: A 3-gene transcript signature was derived from the discovery cohort of 56 bacterial and 27 viral infection cases. In the validation cohort of 200 cases, the signature differentiated bacterial from viral infections with an AUC-ROC of 0.976 (95% CI: 0.919-1.000), sensitivity 97.3% and specificity of 100%. The AUC-ROC for C-reactive protein (CRP) and leucocyte count (WCC) was 0.833 (95% CI: 0.694-0.944) and 0.938 (95% CI: 0.840-0.986) respectively. The signature achieved higher net benefit in decision curve analysis than either CRP or WCC for discriminating viral infections from all other cases. In the second validation analysis the signature discriminated 35 bacterial infections from 34 SARS-CoV-2 positive COVID-19 infections with AUC-ROC of 0.953 (95% CI: 0.893-0.992), sensitivity 88.6% and specificity of 94.1%.
Interpretation: This novel 3-gene signature discriminates viral infections including COVID-19 from other emergency infection presentations in adults, outperforming both WCC and CRP, thus potentially providing significant clinical utility in managing acute presentations with infection.
Date Issued
2021-11-01
Date Acceptance
2021-05-28
Citation
The Lancet Microbe, 2021, 2 (11), pp.594-603
ISSN
2666-5247
Publisher
Elsevier
Start Page
594
End Page
603
Journal / Book Title
The Lancet Microbe
Volume
2
Issue
11
Copyright Statement
2021 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license
License URL
Sponsor
Imperial College Healthcare NHS Trust- BRC Funding
Medical Research Council (MRC)
Imperial College Healthcare NHS Trust- BRC Funding
National Institute for Health Research
UK Research and Innovation
European Commission
Identifier
https://doi.org/10.1016/S2666-5247(21)00145-2
Grant Number
RDA02
MR/R502376/1
RDF04
RP-2016-07-012
9815274 MC_PC_19025
Horizon 2020
Publication Status
Published online