Role of glial fibrillary acidic protein as a biomarker in differentiating intracerebral haemorrhage from ischaemic stroke and stroke mimics: a meta-analysis

Background: Studies have suggested promising evidence that glial fibrillary acidic protein (GFAP) could be used as a blood biomarker to distinguish between ischaemic stroke (IS) and intracerebral haemorrhage (ICH) in acute stage. Objective: To determine the available evidence for GFAP as a blood biomarker for differentiating ICH from IS and stroke mimics using a meta-analysis approach. Methods: Search terms were used for literature search: (“STROKE” [Mesh] OR “BIOMARKER” [Title/Abstract] OR “GFAP” [Title/Abstract])] OR “SPECIFICITY” OR “SENSTIVITY” at various search engines like PubMed, Google scholar, Trip database, clinicaltrial.gov for articles from 1990 to April 2019 using filter ‘human subjects’. Data were analysed using software STATA version 13. Results: A pooled analysis including 12 studies suggested that GFAP if used as a biomarker to differentiate between different types of strokes (ICH from IS and mimics) had a sensitivity of 78% (95% CI: 67–86%) and a specificity of 95% (95% CI: 88–98%). Positive likelihood ratio (LR) was 14.4 and negative LR was 0.23. SROC with prediction and confidence contours suggests promising area under the curve 0.93, 95% CI ranges from 0.90 to 0.95. Diagnostic odds ratio with 95% CI was observed 63 (31–125). Conclusion: Our meta-analysis suggests that GFAP has a promising diagnostic accuracy for the differentiation of ICH from IS and mimics. Further, phase II and phase III diagnostic test studies are required to validate the findings before using GFAP as a blood based biomarker for clinical use. Trial Registration: This study was registered in OSF registries 10.17605/OSF.IO/B9JP4.