Novel biomarkers for distinguishing bacterial from non-bacterial infection: a systematic review

Research output: Contribution to conferencePosterpeer-review

27 Downloads (Pure)


Background: Distinguishing between serious and invasive bacterial infections and non-bacterial aetiologies is challenging, particularly in high-risk groups such as febrile infants under 90 days old. As such, clinicians tend to adopt a cautious approach, potentially leading to the overuse of antibiotics. Existing biomarkers, e.g. C-reactive protein (CRP) and procalcitonin have significant limitations in differentiating bacterial infection and a number of novel biomarkers are emerging to aid this diagnostic process. We present a systematic review of novel biomarkers of bacterial infection.

Methods: A literature search was performed in Medline Ovid using predefined terms. Included articles reported the diagnostic utility of novel host blood plasma or serum biomarkers for discriminating bacterial infection from other aetiologies, with no age restrictions. Following published guidelines for minimal diagnostic performance requirements, biomarkers with a sensitivity ≥90% and specificity ≥80%, or an area under the curve (AUC) ≥0.9 were selected. Studies assessing these biomarkers were required to demonstrate minimal bias.

Results: Our initial search identified 2,236 publications. Following screening and full-text review, 47 articles met the inclusion criteria and were subjected to final analysis. Three individual biomarkers and two grouped signatures met the specified thresholds. The individual biomarkers include interferon-gamma (IFN-γ), interferon-alpha (IFN-a) and lipocalin-2 (LCN2). One grouped signature comprises tumour necrosis factor-related apoptosis-inducing ligand (TRAIL)+interferon gamma-induced protein 10 (IP-10)+CRP and the other grouped signature comprises E-selectin (SELE)+interferon-18 (IL18)+neural cell adhesion molecule 1 (NCAM1)+galectin-3-binding protein (LG3BP)+LCN2+IFN-γ.

IFN-γ, studied in six publications encompassing 962 subjects, exhibited an AUC ≥0.9 (0.94) in one study. IFN-a, in one publication with 101 subjects, showed an AUC of 0.93. LCN2 had AUCs of 0.97 and 0.91, meeting the threshold in two out of the nine relevant publications and was studied in a total of 2,220 subjects. The TRAIL+IP-10+CRP signature, investigated in eight publications in a total of 2,638 subjects, exceeded the threshold in six publications, with AUCs of 0.94 and 0.90, sensitivities ranging from 93.5% to 98.1% and specificities between 88.0% and 94.3%. The SELE+IL18+NCAM1+LG3BP+LCN2+IFN-γ signature, studied in a single publication comprising 306 subjects had a sensitivity of 90.4% and a specificity of 89.6%.

Conclusion: Our review identified three promising individual biomarkers and two grouped signatures for differentiating bacterial from non-bacterial infection. Among these, the TRAIL+IP-10+CRP signature demonstrated promise across the greatest number of subjects. To assess the diagnostic efficacy of the novel biomarkers and grouped signatures in discriminating bacterial infection, further validation is necessary in specific cohorts such as febrile infants.
Original languageEnglish
Number of pages1
Publication statusPublished - 16 May 2024
EventInfectious Diseases Society of Ireland, Annual Meeting 2024 - Royal Marine Hotel, Dun Laoghaire, Dublin, Ireland
Duration: 16 May 202417 May 2024


ConferenceInfectious Diseases Society of Ireland, Annual Meeting 2024
Abbreviated titleIDSI
Internet address

Cite this