Secondary injury and inflammation after intracerebral haemorrhage: a systematic review and meta-analysis of molecular markers in patient brain tissue

Abstract

Background: Inflammatory responses to intracerebral haemorrhage (ICH) are potential therapeutic targets. We aimed to quantify molecular markers of inflammation in human brain tissue after ICH compared with controls using meta-analysis.

Methods: We searched OVID MEDLINE (1946-) and Embase (1974-) in June 2020 for studies that reported any measure of a molecular marker of inflammation in brain tissue from five or more adults after ICH. We assessed risk of bias using a modified Newcastle-Ottawa Scale (mNOS; mNOS score 0-9; 9 indicates low bias), extracted aggregate data, and used random effects meta-analysis to pool associations of molecules where more than two independent case-control studies reported the same outcome and Gene Ontology enrichment analysis to identify over-represented biological processes in pooled sets of differentially expressed molecules (International Prospective Register of Systematic Reviews ID: CRD42018110204).

Results: Of 7501 studies identified, 44 were included: 6 were case series and 38 were case-control studies (median mNOS score 4, IQR 3-5). We extracted data from 21 491 analyses of 20 951 molecules reported by 38 case-control studies. Only one molecule (interleukin-1β protein) was quantified in three case-control studies (127 ICH cases vs 41 ICH-free controls), which found increased abundance of interleukin-1β protein after ICH (corrected standardised mean difference 1.74, 95% CI 0.28 to 3.21, p=0.036, I²=46%). Processes associated with interleukin-1β signalling were enriched in sets of molecules that were more abundant after ICH.

Conclusion: Interleukin-1β abundance is increased after ICH, but analyses of other inflammatory molecules after ICH lack replication. Interleukin-1β pathway modulators may optimise inflammatory responses to ICH and merit testing in clinical trials.

Keywords: immunology; meta-analysis; neuropathology; stroke; systematic reviews.

Figure 1

Flow diagram of the study selection. *Exclusion of ineligible analyses only. Other eligible analyses from the same study remain included. GO, Gene Ontology; ICH, intracerebral haemorrhage.

Figure 2

Chord diagram of replicated findings. This demonstrates which studies conducted analyses of the same gene symbols and indicates that many replicated analyses were derived from two studies. Segments indicate studies reporting analyses of a gene symbol that has been analysed more than once. Chords link segments reporting the same molecule. Chord width is weighted by risk of bias, with larger width indicating lower risk of bias.

Figure 3

Forest plot of pooled independent associations of IL-1β protein with ICH. Tissue analysed by immunohistochemistry. Studies of surgically resected perihaematomal tissue compared with healthy-appearing tissue that was resected on approach to the haematoma or from controls undergoing surgery for non-haemorrhagic disease. ICH, intracerebral haemorrhage; IL-1β, interleukin-1β; SMDc, corrected standardised mean difference.

Figure 4

Significantly enriched GO biological process terms according to direction of association with ICH. Top 15 most statistically significantly enriched terms in sets of gene symbols that were subject to at least one, two or three analyses by controlled studies included in our review and found to be increased or decreased after ICH. Ranked by number of genes annotated for each go term. GO, Gene Ontology.