Prognostic Utility of Serum Biomarkers in Intracerebral Hemorrhage: A Systematic Review

Abstract

Background. Intracerebral hemorrhage (ICH) accounts for 10–20% of all strokes and is associated with high morbidity and mortality. Recent studies have identified serum biomarkers as a means to improve outcome prognostication in poor grade ICH patients. Poor prognosis of ICH patients and complex pathophysiology of the disease necessitate prognostic serum biomarkers to help guide treatment recommendations. Objective. The objective is to systematically review all biomarkers used to predict long-term functional outcome in patients with spontaneous intracerebral hemorrhage. Results. We identified 36 studies investigating the predictive utility of 50 discrete biomarkers. Data from 4865 ICH patients were reviewed. Inflammatory biomarkers (11/50) were most often studied, followed by oxidative (8/50), then neuron and astrocyte-specific (7/50). S100 calcium binding protein B, white blood cell count, and copeptin were the most often studied individual biomarkers. The prognostic utility of 23 biomarkers was analyzed using receiver operating characteristic curves. Area under the curve (AUC) values for all available biomarkers except neutrophil/lymphocyte ratio were acceptable. Twenty of the 23 biomarkers were characterized by at least one excellent AUC value. Vascular endothelial growth factor, glial fibrillary astrocyte protein, and S100 calcium binding protein B were characterized by outstanding AUC. Conclusions. We identified the inflammatory and neuron and astrocyte-specific biomarker categories as having the greatest number of significant individual biomarker predictors of long-term outcome. Further investigation utilizing cross-validation of prediction models in a second independent group and blinded assessment of outcomes for the predictive utility of biomarkers in patients with ICH is warranted.

Keywords

intracerebral hemorrhage biomarkers prognosis systematic review

Introduction

Intracerebral hemorrhage (ICH) accounts for 10%–30% of all strokes and is associated with high mortality and morbidity.^1-4 Median 30-day mortality ranges from 15% to 50%⁵ and mean functional dependency ranges from 12% to 39%.⁶ Standard of care medical management includes close monitoring and treatment for blood pressure, seizures, increased intracranial pressure, and reversal of anticoagulant and antiplatelet medications.⁷ Decompressive hemicraniectomy is used in severe cases of increased intracranial pressure, and minimally invasive surgical evacuation is under investigation as a promising surgical option.^8-10 Treatment decisions and recovery depend on clinical characteristics including age, hematoma location, hematoma volume, and neurologic examination.¹¹

Recent studies have identified serum biomarkers as a means to improve outcome prognostication in poor grade ICH patients.¹² Some serum biomarkers have diagnostic and prognostic utility in other neurologic disease states as well, such as stroke and subarachnoid hemorrhage.¹³ The application of these biomarkers can improve risk stratification and therefore guide implementation of tailored treatment modalities.

The poor prognosis of ICH patients and complex pathophysiology of the disease emphasize the need for prognostic serum biomarkers to help guide treatment recommendations. In this study, we systematically review the utility of serum biomarkers to predict long-term outcome after ICH.

Methods

We conducted a systematic review according to the Preferred Reporting Items with Systematic Reviews and MetaAnalyses (PRISMA) statement and Prospero (CRD42020176400).

Search Strategy and Selection Criteria

PubMed, Cochrane Library, and Google Scholar were screened using the terms “biomarkers predictive functional outcome intracranial (or intracerebral) hemorrhage” and “intracranial hemorrhage outcome biomarkers.” Queries were restricted to human studies investigating atraumatic spontaneous, supratentorial ICH of ≥30 patients in the English or Spanish language available as of December 31st, 2019. MeSH terms used in our queries were Biomarkers, intracerebral hemorrhage, and intracranial hemorrhage.

Inclusion and Exclusion Criteria

Studies were included in our analysis if they evaluated clinical outcome ≥30 days post ICH and were clinical trials, prospective, or retrospective cohort studies. Studies were excluded if patients were ≤18 years; if the article type was a case report, review article, systematic review, or abstract; or if the outcome of interest was missing. Inclusion/exclusion criteria are reported in Table 1.

Table 1.

Inclusion and Exclusion Criteria.

Category	Inclusion Criteria	Exclusion Criteria
Population characteristics	Spontaneous, supratentorial ICH	Secondary and/or infratentorial ICH
	Human studies	Animal studies
	Age ≥18 years	Nonadult population
Study design	Retrospective and prospective cohort studies and clinical trials	Case reports, reviews, systematic reviews, and abstracts
Measurement variables	Functional outcome measures (mRSs, GOS, Barthel index, and mortality)	Outcome measures not present or not related to functional outcome
	Long-term outcome measures (30 days or more)	Short-term outcome measures (discharge, 1 week)
Language	English and Spanish languages	Non-English or Spanish language

Assessment of Study Quality

Two authors (NM and CR) independently assessed the 36 included studies using a modified version of the evidence methodological quality score adapted from previous systematic reviews on similar topics.^14-18 Internal, statistical, and external validity criteria were scored 0 or 1, with maximum score of 14 (Supplementary Table 1). Studies where ≥80% of criteria were met were considered “high quality,” ≥ 60% but <80% were considered “moderate quality,” ≥ 50% but <60% were considered “low quality,” and <50% were considered “very low quality.”¹⁹ Discrepancies between grading were resolved by a third author (MA).

Statistical Analysis

Descriptive statistics were used to analyze the data. Continuous data are reported as mean ± standard deviation. Categorical data are reported as proportions. We obtained the proportions of biomarkers investigated in each functional outcome group.

We pooled the AUC values of each biomarker category based on outcome used. AUC values were divided into 4 categories: Unreliable (AUC ≤.700), acceptable (.700 ≤ AUC <.800), excellent (.800 ≤ AUC <.900), and outstanding (AUC ≥.900).²⁰ Data were analyzed using Microsoft Excel, and Figures were constructed using Prism 8.3.

Results

Literature Search and Study Characteristics

The results of the electronic search and article selection process are shown in the PRISMA flow diagram (Supplementary Figure 1). In total, 502 articles were identified. Seventeen additional publications from associated clinical trials were manually added, comprising 519 total articles. After initial title and abstract screening, 79 articles were identified and subjected to a full text review. Forty-three studies failed to meet inclusion criteria due to population characteristics (18), study design (10), or outcomes measured (15). Thirty-six studies met the inclusion criteria and were included in our qualitative analysis. In aggregate, data from 4865 ICH patients were reviewed.

Of the 36 studies, 31 were prospective cohort studies, 4 were retrospective cohort studies, and 1 study type was unidentifiable. No clinical trials were identified. The mean study size was 135 ± 92 ICH patients ranging from 40 to 423 patients. Functional outcome was measured by mRS (28 studies), GOS (3 studies), MBI (3 studies), and mortality (28 studies). Regarding the studies using mRS as a functional outcome, 25 characterized a “poor outcome” with a score greater than 2, while 3 of the 28 studies characterized a “poor outcome” with a score greater than 3. Functional outcome was evaluated at 30 (6 studies), 90 (16 studies), 180 (12 studies), and 365 days (3 studies) (Figure 1A).

Figure 1.

Number of biomarkers (A) with the number of studies evaluating their respective biomarkers at each time point, (B) classified by category, and (C) measured by functional outcome type. (D) Frequency distribution of study quality assessment scores.

All 50 biomarkers were tested via blood except D-dimer which was tested via cerebrospinal fluid. Some studies investigated multiple biomarkers using more than one functional outcome measure at more than one time point. In this systematic review, no direct comparisons were made between studies due to the significant variability between data.²¹

Biomarkers Categories

In total, we identified 50 different biomarkers studied related to functional outcome (Figure 2). The breakdown by type of biomarker is shown in Figure 1B. The inflammatory group accounted for the largest number of identified biomarkers (11/50, 22.0%), followed by the oxidative group (8/50, 16.0%), and the neuron-specific group (7/50, 14.0%). The most studied biomarkers (WBC, S100B, and copeptin) were each assessed in 3 studies. Stratified by outcomes, Figure 1C shows that biomarkers were most commonly investigated using mRS as the functional outcome (40/50, 80.0%) followed by mortality (34/50, 68.0%). A summary of the studies included is in Table 2.

Figure 2.

Anatomical schematic illustrating biomarker classification.

Table 2.

Summary of Biomarkers and Associated Outcomes.

Biomarker	Outcome Measure(s)	# of Patients	Summary of Findings	Author, year
Inflammatory markers
Neutrophil count	GOS	104	NS	Zhang et al., 2018
Lymphocyte count	GOS	104	NS	Zhang et al., 2018
Monocyte count	GOS	104	NS	Zhang et al., 2018
White blood cell count (WBC)	GOS	104	NS	Zhang et al., 2018
	GOS and mortality	210	NS	Di Napoli et al., 2011
	Mortality and MBI	423	Peripheral leukocyte count independently predicts poor functional outcome at 3 months	Agnihorti et al., 2011
Neutrophil/lymphocyte ratio (NLR)	GOS	104	Elevated NLR is negatively correlated with 3-month functional outcomes	Zhang et al., 2018
	mRS and mortality	352	NS	Sun, 2017
S100 calcium-binding protein A12 (S100A12)	Mortality	182	S100A12 is an independent predictor of 30-day mortality	Qian et al., 2018
CD163	mRS	41	CD163 is an independent predictor of unfavorable outcome at 3 and 12 months	Garton et al., 2017
sCD40 L	mRS and mortality	110	sCDL40 is an independent predictor of 6-month unfavorable outcome and mortality	Lin et al., 2016
YKL-40	mRS and mortality	172	YKL-40 is an independent predictor of 90-day unfavorable outcome and mortality	Jiang et al., 2014
Gelsolin	mRS and mortality	132	Gelsolin is an independent predictor of 6-month unfavorable outcome and mortality	Zhao et al, 2013
C-reactive protein (CRP)	GOS and mortality	210	CRP is independently associated with 30-day mortality, but not unfavorable outcomes	Di Napoli et al., 2011
	mRS and mortality	103	Maximal CRP level is an independent predictor of unfavorable outcomes but not mortality	Diedler et al., 2009
Oxidative markers
Thioredoxin	Mortality	218	Thioredoxin is an independent predictor of 6-month mortality	Qian et al., 2016
8-iso-prostaglandin F2α	mRS and mortality	128	8-iso-prostaglandin F2α is an independent predictor of 6-month unfavorable outcome and mortality	Du et al., 20-14
Leukocyte 8-hydroxy-2′-deoxyguanosine (8-OHdG)	mRS and MBI	64	Leukocyte 8-hydroxy-2′-deoxyguanosine was independently associated with 30-day functional outcomes	Chen et al., 2011
Erythrocyte glucose-6-phosphate dehydrogenase (G6PD)	mRS and MBI	64	NS	Chen et al., 2011
Erythrocyte glutathione peroxidase (GPx)	mRS and MBI	64	NS	Chen et al., 2011
Plasma malondialdehyde (MDA)	mRS and MBI	64	NS	Chen et al., 2011
Vitamin E	mRS and MBI	64	NS	Chen et al., 2011
Vitamin A	mRS and MBI	64	NS	Chen et al., 2011
Neuron and astrocyte-specific markers
S100 calcium binding protein B (S100B)	mRS and mortality	41	Median S100B is significantly higher in non-survivors vs survivors at 30 days	Lauridsen et al., 2018
	mRS	46	Plasma S100B is significantly elevated in patients with unfavorable vs favorable outcome at 90 days	Zhou et al., 2016
	mRS and mortality	118	S100B is significantly higher in patients with poor vs good functional outcome and in non-survivors vs survivors at 6 months	Yu et al., 2014
Glial fibrillary astrocyte protein (GFAP)	mRS	43	GFAP is an independent predictor of 90-day unfavorable outcome	Xiong et al., 2015
	mRS and mortality	118	GFAP is an independent predictor of 6-month unfavorable outcome and mortality	Yu et al., 2014
Myelin basic protein (MBP)	mRS and mortality	118	MBP is an independent predictor of 6-month unfavorable outcome and mortality	Yu et al., 2014
Neuron-specific enolase (NSE)	mRS and mortality	118	NSE is an independent predictor of 6-month unfavorable outcome and mortality	Yu et al., 2014
Phosphorylated axonal neurofilament subunit H (pNF-H)	mRS and mortality	118	pNF-H is an independent predictor of 6-month unfavorable outcome and mortality	Yu et al., 2014
	mRS and mortality	112	pNF-H is an independent predictor of 6-month unfavorable outcome and mortality	Cai et al., 2013
Ubiquitin carboxyl-terminal hydrolase L1 (UCH-L1)	mRS and mortality	118	UCH-L1 is an independent predictor of 6-month unfavorable outcome and mortality	Yu et al., 2014
Tau	mRS and mortality	118	Tau is an independent predictor of 6-month unfavorable outcome and mortality	Yu et al., 2014
	mRS and mortality	176	Tau is an independent predictor of 90-day unfavorable outcome and mortality	Hu et al., 2012
Molecular adhesion and extracellular matrix markers
Tenascin-c	mRS and mortality	162	High tenascin-c levels can significantly distinguish between survivors and non-survivors and patients with poor vs good outcomes at 90 days	Wang et al., 2018
Galectin-3	mRS and mortality	112	High galectin-3 levels are independently associated with unfavorable outcomes and mortality at 6 months	Yan et al., 2016
Thrombospondin-1	mRS and mortality	110	Thrombospondin-1 is an independent predictor of unfavorable outcome and mortality at 6 months	Dong et al., 2015
Matrix metalloproteinase-3 (MMP-3)	mRS	59	MMP-3 is an independent predictor of 90-day unfavorable outcome	Li et al., 2013
Matrix metalloproteinase-9 (MMP-9)	mRS	59	MMP-9 is an independent predictor of 90-day unfavorable outcome	Li et al., 2013
Coagulation markers
Thromboelastometry (ROTEM)	mRS and mortality	41	NS	Lauridsen et al., 2018
Thrombin generation	mRS and mortality	41	NS	Lauridsen et al., 2018
Thrombin-antithrombin (TAT) complex	mRS and mortality	41	NS	Lauridsen et al., 2018
D-dimer	Mortality	43	CSF D-dimer levels can predict 30-day mortality	Chen et al., 2016
Vascular and angiogenic markers
Signal peptide-Cub-Egf domain-containing protein-1 (SCUBE1)	mRS and mortality	128	SCUBE1 is an independent predictor of unfavorable outcome and mortality at 6 months	Qui et al., 2016
Vascular endothelial growth factor (VEGF)	mRS	92	Higher presenting VEGF concentrations are associated with more favorable functional outcomes	Zheng et al., 2016
	mRS	59	NS	Li et al., 2013
Adrenomedullin	mRS and mortality	114	Adrenomedullin is an independent predictor of 90-day unfavorable outcome and mortality	Wang et al., 2014
Angiopoietin-1	mRS	59	NS	Li et al., 2013
Metabolic markers
Leptin	mRS and mortality	92	Leptin is an independent predictor of 6-month unfavorable outcome and mortality	Zhang et al., 2013
Growth hormone (GH)	mRS, Barthel index, and mortality	40	GH is an independent predictor of 30-day mortality and 90-day unfavorable outcome	Zweifel et al., 2011
Low-density lipoprotein cholesterol (LDL-C)	mRS and mortality	108	LDL is independently related to 90-day mortality but not 90-day unfavorable outcomes	Rodriguez-Luna at al., 2011
Blood glucose (BG)	GOS and mortality	210	NS	Di Napoli et al., 2011
Renal markers
Estimated glomerular filtration rate (eGFR)	mRS and mortality	365	There is a stepwise, dose-dependent relationship between eGFR and unfavorable outcome at 90 days	You et al., 2016
Cystatin C	mRS and mortality	365	High cystatin C level was associated with a 4.01-fold increase in the risk of all-cause mortality at 90 days	You et al., 2016
Copeptin	mRS and mortality	118	Copeptin is an independent predictor of 6-month unfavorable outcome and mortality	Yu et al., 2014
	mRS and mortality	89	Copeptin is an independent predictor of 1-year unfavorable outcome and mortality	Zhang et al., 2012
	mRS and mortality	271	Copeptin is an independent predictor of 90-day unfavorable outcome and mortality	Wei et al., 2014
Other markers
Hepcidin	mRS	41	NS	Garton et al., 2017
	mRS	81	Hepcidin is an independent predictor of 90-day unfavorable outcome	Xiong et al., 2015
Ferritin	mRS	41	Ferritin is an independent predictor of 12-month unfavorable outcome	Garton et al., 2017
N-terminal pro-brain natriuretic peptide (NT-proBNP)	GOS and mortality	132	High NT-proBNP is an independent predictor of 6-month unfavorable outcome and mortality	Li et al., 2017
Caspase-cleaved cytokeratin-18 (CCCK-18)	mRS and mortality	102	Serum CCCK-18 is an independent predictor of unfavorable outcome and mortality at 6 months	Gu et al., 2016

Study Quality Assessment Analysis

The results of the study quality assessment are summarized in Table 3. The mean quality assessment score (QAS) of all 36 studies was 10.78 ± 1.20 ranging from 8 to 13 (maximum score of 14), suggesting that on average the studies included were of moderate quality (Figure 1D). Ten studies (27.8%) had a QAS ≥12, or 80% or more of the total score (Supplementary Table 2).

Table 3.

Summary of Quality Assessment Score Evaluation.

		Internal Validity							Statistical Validity			External Validity
First Author	Year	A	B	C	D	E	F	G	H	I	J	K	L	M	N	Total Score
Inflammatory markers
Zhang	2018	1	1	1	0	0	1	0	1	1	1	1	1	0	0	9
Qian	2018	1	1	1	1	0	1	1	1	1	1	1	1	0	0	11
Sun	2017	1	1	1	1	0	1	1	1	1	1	1	1	0	0	11
Lin	2016	1	1	1	1	1	1	0	1	1	1	1	1	0	0	11
Jiang	2014	1	1	1	1	0	1	1	1	1	1	1	1	0	0	11
Zhao	2013	1	1	1	1	0	1	1	1	1	1	1	1	0	0	11
Agnihorti	2011	1	1	1	1	0	1	0	1	1	1	1	1	0	0	10
Diedler	2009	1	1	1	0	0	1	1	1	1	1	1	0	0	0	9
Oxidative markers
Qian	2016	1	1	1	1	0	1	1	1	1	1	1	1	0	0	11
Du	2014	1	1	1	1	1	1	0	1	1	1	1	1	0	0	11
Chen	2011	1	1	1	1	0	1	0	1	1	1	1	1	0	0	10
Neuron and astrocyte-specific markers
Zhou	2016	1	1	1	1	1	1	0	1	1	0	1	1	0	0	10
Xiong	2015	1	1	1	1	0	1	1	1	1	0	1	1	1	0	11
Cai	2013	1	1	1	1	1	1	1	1	1	1	1	1	0	0	12
Hu	2012	1	1	1	1	0	1	1	1	1	1	1	1	0	0	11
Molecular adhesion and extracellular matrix markers
Wang	2018	1	1	1	1	0	1	1	1	1	1	1	1	0	0	11
Yan	2016	1	1	1	1	1	1	1	1	1	1	1	1	1	0	13
Dong	2015	1	1	1	1	1	1	1	1	1	1	1	1	0	0	12
Coagulation markers
Chen	2016	1	1	1	1	0	1	0	1	1	0	1	1	1	0	10
Vascular and angiogenic markers
Zheng	2016	1	1	1	1	0	1	0	1	1	0	1	1	0	0	9
Qui	2016	1	1	1	1	1	1	1	1	1	1	1	1	1	0	13
Wang	2014	1	1	1	1	1	1	1	1	1	1	1	1	0	0	12
Metabolic markers
Zhang	2013	1	1	1	1	0	1	0	1	1	1	1	1	0	0	10
Zweifel	2011	1	1	1	1	1	1	0	1	1	0	1	1	0	0	10
Rodriguez-Luna	2011	1	1	1	1	0	1	1	1	0	1	1	1	0	0	10
Renal markers
You	2016	1	1	1	1	0	1	1	1	1	1	1	1	0	0	11
Wei	2014	1	1	1	1	1	1	1	1	1	1	1	1	0	0	12
Zhang	2012	1	1	1	1	0	1	1	1	1	1	1	1	0	0	11
Other markers
Li	2017	1	1	1	1	0	1	1	1	1	1	1	1	0	0	11
Gu	2016	1	1	1	1	1	1	1	1	1	1	1	1	0	0	12
Xiong	2015	1	1	1	1	1	1	1	1	1	1	1	1	0	0	12
Combination*
Lauridsen	2018	1	1	1	1	0	1	0	1	0	0	1	1	0	0	8
Garton	2017	1	1	1	1	0	1	0	1	1	1	1	1	0	0	10
Yu	2014	1	1	1	1	0	1	0	1	1	0	1	1	0	0	9
Li	2013	1	1	1	1	0	1	0	1	1	1	1	1	0	0	10
Di Napoli	2011	1	1	1	1	1	1	1	1	1	1	1	1	1	0	13

*Combination denotes studies investigating multiple biomarkers with more than one classification type.

Most studies properly validated predictor variables (94.4%), controlled for patient dropout in the statistical analyses (61.1%), and did not use blinded evaluators to assess functional outcome (63.9%). All other measures of internal validity were adequately met for all studies. Regarding statistical validity, most studies had an adequate sample size (94.4%) and tested the predictor variables for multicollinearity (80.6%). All other studies appropriately tested the relationship between predictor variables and outcome measures for statistical significance.

Regarding external validity, all studies specified relevant patient characteristics for patient selection and all but one study specified inclusion and exclusion criteria. Additional treatment effects on outcome measures were properly controlled for in only 5 studies (13.9%). No studies cross-validated their prediction model in a second independent group (Table 3).

Receiver Operating Characteristic Curve Analyses

Twenty-one studies evaluated the predictive utility of 23 biomarkers using ROC analysis and reported AUC, threshold criterion, sensitivity, and specificity (Table 4). AUC, sensitivity, and specificity are represented as a heat map where lower values are colored blue and higher values are colored red. All AUC values were ≥.700 (acceptable) except for Zhang et al’s study measuring the predictive utility of neutrophil to lymphocyte ratio (NLR) using GOS at 90 days.²² Predictive utility was measured using mRS (82.6%), GOS (8.7%), and mortality (91.3%) as functional outcome measures. Most studies specifically measured mRS and mortality at 180 days (73.9% and 76.0%, respectively).

Table 4.

Receiver Operating Characteristic Curve Analysis.

Biomarker	Outcome Measure	Time (days)	AUC	Criterion	Sensitivity, %	Specificity, %	Author, year	# of Studies with Significant Findings	# of Studies with Nonsignificant Findings
Inflammatory markers
NLR	GOS	90	.688	≥6.46 ng/L	83.10	55.60	Zhang et al., 2018	1	1
S100A12	Mortality	30	.822	≥57.9 ng/mL	69.40	80.80	Qian et al., 2018	1	0
sCD40 L	mRS	180	.808	≥4.4 ng/mL	69.60	77.80	Lin et al., 2016
	Mortality	180	.825	≥4.6 ng/mL	71.90	75.60	Lin et al., 2016
YKL-40	mRS	90	.837	≥229.9 ng/mL	90.80	63.50	Jiang et al., 2014	1	0
	Mortality	90	.845	≥299.0 ng/mL	78.60	75.90	Jiang et al., 2014
Gelsolin	mRS	180	.833	≤58.2 mg/L	69.60	83.00	Zhao et al., 2013	1	0
	Mortality	180	.811	≤50.7 mg/L	87.00	72.10	Zhao et al., 2013
Oxidative markers
TRX	Mortality	180	.826	≥29.6 ng/mL	74.20	76.30	Qian et al., 2016	1	0
8-iso-prostaglandin F2α	mRS	180	.814	≥.53 ng/mL	75.80	80.60	Du et al., 2014	1	0
	Mortality	180	.83	≥.53 ng/mL	84.60	68.50	Du et al., 2014
Neuron- and astrocyte-specific markers
S100B	Mortality	90	.924	≥133 pg/mL	100.00	76.20	Zhou et al., 2016	3	0
	mRS	180	.767	≥400.9 pg/mL	80.90	72.00	Yu et al., 2014
	Mortality	180	.813	≥470.7 pg/mL	83.80	72.80	Yu et al., 2014
GFAP	mRS	90	.936	≥1.04 ng/mL	95.70	80.00	Xiong et al., 2015	2	0
	mRS	180	.756	≥9.5 pg/mL	77.90	68.00	Yu et al., 2014
	Mortality	180	.805	≥11.7 pg/mL	67.60	87.70	Yu et al., 2014
MBP	mRS	180	.733	≥12.3 μg/mL	76.50	60.00	Yu et al., 2014	1	0
	Mortality	180	.787	≥13.3 μg/mL	75.70	70.40	Yu et al., 2014
NSE	mRS	180	.758	≥13.1 ng/mL	73.50	68.00	Yu et al., 2014	1	0
	Mortality	180	.792	≥14.5 ng/mL	78.40	66.70	Yu et al., 2014
pNF-H	mRS	180	.76	≥807.1 pg/mL	60.30	84.00	Yu et al., 2014	2	0
	Mortality	180	.803	≥816.3 pg/mL	73.00	75.30	Yu et al., 2014
	mRS	180	.826	≥547.3 pg/mL	89.60	62.20	Cai et al., 2013
	Mortality	180	.857	≥816.3 pg/mL	73.70	82.40	Cai et al., 2013
UCHL1	mRS	180	.736	≥1876.4 pg/mL	63.20	78.00	Yu et al., 2014	1	0
	Mortality	180	.77	≥2039.9 pg/mL	70.30	75.30	Yu et al., 2014
Tau	mRS	180	.744	≥233.0 pg/mL	77.90	70.00	Yu et al., 2014	2	0
	Mortality	180	.794	≥240.2 pg/mL	86.50	69.10	Yu et al., 2014
	mRS	90	.826	≥91.4 pg/mL	83.60	75.80	Hu et al., 2012
	Mortality	90	.817	≥172.1 pg/mL	78.10	70.50	Hu et al., 2012
Molecular adhesion and extracellular matrix markers
Galectin-3	mRS	180	.827	≥18.9 ng/mL	89.50	65.50	Yan et al., 2016	1	0
	Mortality	180	.832	≥22.4 ng/mL	90.90	64.60	Yan et al., 2016
Thrombospondin-1	mRS	180	.809	≥732.2 ng/mL	69.90	77.80	Dong et al., 2015	1	0
	Mortality	180	.811	≥769.0 ng/mL	75.00	74.40	Dong et al., 2015
Vascular and angiogenic markers
VEGF	mRS	90	.952	≥111.17 pg/mL	91.50	98.70	Zheng et al., 2016	1	1
SCUBE1	mRS	180	.839	≥15.3 ng/mL	84.60	71.40	Qui et al., 2016	1	0
	Mortality	180	.823	≥16.8 ng/mL	84.20	65.60	Qui et al., 2016
Adrenomedullin	mRS	90	.827	≥76.6 pg/mL	89.70	63.00	Wang et al., 2014	1	0
	Mortality	90	.856	≥122.4 pg/mL	66.70	89.30	Wang et al., 2014
Metabolic markers
Leptin	mRS	180	.832	≥13.4 ng/mL	94.10	70.70	Zhang et al., 2013	1	0
	Mortality	180	.817	≥21.5 ng/mL	76.50	75.90	Zhang et al., 2013
Renal markers
Copeptin	mRS	180	.803	≥2369.1 pg/mL	70.60	82.00	Yu et al., 2014	3	0
	Mortality	180	.829	≥2518.2 pg/mL	78.40	74.10	Yu et al., 2014
	mRS	365	.842	≥23.5 pmol/L	76.80	87.90	Zhang et al., 2012
	Mortality	365	.82	≥23.8 pmol/L	81.60	70.60	Zhang et al., 2012
Other markers
NT-proBNP	GOS	180	.838	≥999.85 pg/mL	66.10	98.70	Li et al., 2017	1	0
	Mortality	180	.958	≥999.85 pg/mL	93.80	92.00	Li et al., 2017
CCCK-18	mRS	180	.797	≥192.8 U/L	93.30	54.70	Gu et al., 2016	1	0
	Mortality	180	0.8	≥245.6 U/L	89.30	60.80	Gu et al., 2016

Discussion

Prognostic biomarkers for primary ICH have the potential to identify high-risk patients, guide treatment decisions, inform family members, and aid in clinical trial design. Currently, numerous biomarkers are under investigation. The most commonly investigated blood-based biomarkers for ICH prognosis are inflammatory, oxidative, and neuron and astrocyte-specific.^23-25 This systematic review provides a comprehensive analysis of the available data on biomarkers in ICH. We found that inflammatory biomarkers were the most commonly studied and specifically white blood cell count (WBC), S100B, and copeptin were the most common individual biomarkers studied. All but one study was determined to be of at least moderate quality indicated by the QAS.

Inflammatory Biomarkers

Inflammatory changes in the perihematomal tissue contribute to secondary neuronal injury and poor functional outcomes following ICH.²⁶ Evidence suggests that leukocyte infiltration of adjacent parenchymal tissue elicits an immune response with a subsequent activation of microglia and cytokine release.^27-30 Askenase et al³¹ demonstrated that coordinated functional and metabolic reprogramming in CD14⁺ monocytes/macrophages occurs over time in the hematoma.

Inflammatory biomarkers were most commonly investigated (Figure 1B). We reviewed 4 studies investigating the predictive utility of 4 blood borne leukocytes: neutrophil, lymphocyte, monocyte, total WBC, and the neutrophil–lymphocyte ratio. The predictive utility of WBC count alone was investigated in 3 different studies but was not significantly associated with long-term functional outcomes in two of the studies.^32,22 Other inflammatory biomarkers predicting long-term functional outcome in ICH include CRP,^32,33 S100A12,³⁴ CD163,³⁵ sCD40 L,³⁶ YKL-40,³⁷ and Gelsolin (Table 2).^37,38 Although inflammatory markers are non-neuron-specific, their substantiated predictive utility warrants further investigation.

Tumor necrosis factor-alpha (TNF-a) and interleukin-1 beta (IL-1b) are two well-known proinflammatory cytokines that were not included in this review. TNF-a and IL-1b are thought to exacerbate ICH-induced brain injury. While this has been shown in experimental and animal models, there is a paucity of human studies investigating these ubiquitous proinflammatory cytokines in the setting of primary ICH.²⁹

Neuron and Astrocyte-Specific Biomarkers

Astrocytes serve as the intermediary between the blood vessels and neurons. We identified 7 neuron-specific biomarkers which function as structural proteins, receptors, or enzymes, across 6 studies that increase following neuronal injury and blood–brain barrier disruption. Thus, increased levels are expected to be associated with poor long-term functional outcomes.^25,39-43

Most biomarkers successfully predicted mortality at 90 or 180 days. Yu et al. demonstrated that the ability to predict outcome of several markers (NSE, MBP, GFAP, Tau, S100B, pNF-H, and UCH-L1) was comparable to the predicted capability of the NIHSS. However, none had an additive effect over the predictive capability of NIHSS. The AUCs of S100B and GFAP were outstanding (>.900) in one study each. These data suggest that neuron-specific biomarkers may accurately predict long-term functional outcomes.

Oxidative Biomarkers

Following ICH and subsequent hemolysis, oxidative stress secondary to hemoglobin toxicity and free radical generation causes brain tissue damage.⁴⁴ Of the 8 total oxidative markers investigated, increased levels of TRX, 8-iso-prostaglandin F2α, and 8-OHdG were significantly associated with poor long-term mRS and mortality. TRX and 8-iso-prostaglandin F2α were independent predictors of 180 mortality with excellent AUC values (Table 4).^45,46 Although these two biomarkers had excellent AUC values, the other five oxidative biomarkers were not associated with long-term outcome. Additional studies investigating the predictive utility of oxidative molecules are needed.

Other Biomarkers

One of the most commonly studied biomarkers was copeptin, a surrogate marker for antidiuretic hormone (ADH). ADH levels are increased in syndrome of inappropriate ADH secretion (SIADH), a condition that causes hyponatremia and contributes to worse outcome patients experiencing a stroke.⁴⁷ In critically ill patients, copeptin levels have been shown to correlate with severity of disease.^48,49 Three independent studies identified copeptin as an independent predictor of functional outcome and mortality at 90, 180, and 365 days with excellent AUC values (Tables 2 and 4). In ICH, copeptin is known to correlate with hematoma volume and is thought to be related to edema formation.⁵⁰

Molecular adhesion molecules have been implicated in ICH prognosis due to their roles in tissue remodeling and blood–brain barrier disruption. Degradation of the extracellular matrix by MMP-3 and MMP-9 results in cerebral edema, a secondary neuronal injury associated with poor outcomes.⁵¹ In addition, tenascin-c (TNC) and thrombospondin-1 (TSP-1) have previously been implicated in aneurysmal subarachnoid hemorrhage.⁵²

Vascular and angiogenic biomarker levels change in response to inflammation, hypoxia, or platelet activation and have been studied in SAH, TBI, and ischemic stroke.^51,53-55 However, their predictive utility in predicting long-term outcome in ICH is less clear. VEGF was identified as an outstanding predictor of 90-day mRS (AUC=.95).⁵³ However, these results are from a single study with short follow-up, and variables were not adjusted for in a multivariable model. Additionally, Li et al.⁵¹ did not find a significant association between VEGF and long-term outcome.

There are scarce data investigating coagulation factors in the setting of ICH. Lauridsen et al. demonstrated that in patients without anticoagulant or antiplatelet therapy, local brain injury induced by ICH results in systemically activated coagulation. However, thrombin, ROTEM, and TAT were not significantly associated with long-term outcome in this cohort.³⁹ It is possible that coagulation factors are related to platelet activation in the recovery of ICH.

Leptin was an excellent predictor of mortality and poor outcomes at 6 months (AUC >.8) with increased serum levels predicting higher mortality and poor functional outcome.⁵⁶ Growth hormone levels have also been associated with high mortality and worse functional outcome, but no multivariate analysis was conducted in this particular study.⁴⁴ Similar to VEGF, higher LDL-C levels were observed in patients with low mortality as cholesterol decreases platelet aggregability, contributing to ICH hematoma volume and expansion risk.⁵⁷

Following ICH, hemoglobin released from erythrolysis binds iron in its ferrous (Fe²⁺) state and is then capable of generating reactive oxygen species and subsequent cellular damage. This results in an upregulation of iron regulatory proteins including ferritin and hepcidin, as well as others that have not yet been well-investigated in the setting of ICH such as transferrin and heme oxygenase.³⁵ Garton et al.³⁵ demonstrated that ferritin but not hepcidin was an independent predictor of 12-month functional outcome. This contrasted with Xiong et al.⁴¹ who found that hepcidin was an independent predictor of long-term outcome. Additional investigation of iron regulatory proteins is necessary to ascertain their prognostic utility in ICH patients.

Study Limitations

This systematic review has several limitations. Multiple biomarkers are involved in more than one biological process, precluding uniform categorization. A more detailed analysis of the data is complicated by significant variability as the studies used different outcome measures at different time points.

Our study quality assessment score evaluation showed that the mean quality of all studies was considered moderate (Supplementary Table 2), and that many studies lacked internal and external validity. Only 13 (37.1%) of the 35 studies evaluated used blinded assessment of outcomes, and 21 studies (60%) adequately controlled for patient dropout. No studies cross-validated their prediction model in a second independent group and only 5 controlled for additional treatment effects on outcome measures. Also, 7 (20%) of the 36 studies did not conduct multivariate analyses to determine if the biomarker under investigation was an independent predictor of functional outcome.

Conclusions

Inflammatory as well as neuron and astrocyte-specific biomarkers contained the greatest number of investigated biomarkers, while S100B, WBC, and copeptin were among the most commonly investigated individual biomarkers in primary ICH. S100B, VEGF, and GFAP were the only biomarkers with outstanding AUC values. With the decreased cost and increased availability of proteomics and other omic analyses, multi-omic biomarker studies are likely to produce increasingly accurate multi-omic biomarker signatures in this disease process in the near future. The heterogeneity of the data analyzed in this review highlights the importance of further investigation for the predictive utility of biomarkers in risk stratification and management of primary ICH. Future studies should cross-validate prediction models in a second independent group, use blinded evaluators for outcome measurements, and adjust for treatment effects on outcome measures.

Supplemental Material

sj-pdf-1-nnr-10.1177_15459683211041314 – Supplemental Material for Prognostic Utility of Serum Biomarkers in Intracerebral Hemorrhage: A Systematic Review

Supplemental Material, sj-pdf-1-nnr-10.1177_15459683211041314 for Prognostic Utility of Serum Biomarkers in Intracerebral Hemorrhage: A Systematic Review by Zachary Troiani, Luis Ascanio, Christina P. Rossitto, Muhammad Ali, Shahram Majidi, J Mocco and Christopher P. Kellner in Neurorehabilitation and Neural Repair

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Zachary Troiani

Supplementary Material

Supplementary material for this article is available online on the Neurorehabilitation & Neural Repair website along with the online version of this article.

References

Keep

Hua

. Intracerebral haemorrhage: mechanisms of injury and therapeutic targets. Lancet Neurol. 2012;11(8):720-731. doi:10.1016/s1474-4422(12)70104-7.

Hemphill

3rd Greenberg

Anderson

, et al. Guidelines for the management of spontaneous intracerebral hemorrhage. Stroke. 2015;46(7):2032-2060.

Qureshi

Mendelow

Hanley

. Intracerebral haemorrhage. Lancet. 2009;373(9675):1632-1644. doi:10.1016/s0140-6736(09)60371-8.

Qureshi

Tuhrim

Broderick

Batjer

Hondo

Hanley

. Spontaneous intracerebral hemorrhage. N Engl J Med. 2001;344(19):1450-1460. doi:10.1056/nejm200105103441907.

Sacco

Marini

Toni

Olivieri

Carolei

. Incidence and 10-year survival of intracerebral hemorrhage in a population-based registry. Stroke. 2009;40(2):394-399. doi:10.1161/strokeaha.108.523209.

van Asch

Luitse

Rinkel

van der Tweel

Algra

Klijn

. Incidence, case fatality, and functional outcome of intracerebral haemorrhage over time, according to age, sex, and ethnic origin: a systematic review and meta-analysis. Lancet Neurol. 2010;9(2):167-176.

Kim

Bae

H-J

. Spontaneous intracerebral hemorrhage: management. Journal of Stroke. 2017;19(1):28-39.

Fung

Murek

Z’Graggen

, et al. Decompressive hemicraniectomy in patients with supratentorial intracerebral hemorrhage. Stroke. 2012;43(12):3207-3211.

Kellner

Song

Pan

, et al. Long-term functional outcome following minimally invasive endoscopic intracerebral hemorrhage evacuation. J Neurointerventional Surg. 2020;12(5):489-494.

10.

Pan

Chartrain

Scaggiante

, et al. A compendium of modern minimally invasive intracerebral hemorrhage evacuation techniques. Operative neurosurgery (Hagerstown, Md.). 2020;18(6):710-720.

11.

McDowell

Kellner

Sussman

, et al. The role of admission timing in the outcome of intracerebral hemorrhage patients at a specialized stroke center. Neurol Res. 2014;36(2):95-101.

12.

Senn

Elkind

MSV

Montaner

Christ-Crain

Katan

. Potential role of blood biomarkers in the management of nontraumatic intracerebral hemorrhage. Cerebrovasc Dis. 2014;38(6):395-409.

13.

Boyd

Hayward

Ward

, et al. Biomarkers of stroke recovery: consensus-based core recommendations from the stroke recovery and rehabilitation roundtable. Neurorehabilitation Neural Repair. 2017;31(10-11):864-876.

14.

Schiemanck

Kwakkel

Post

MWM

Prevo

AJH

. Predictive value of ischemic lesion volume assessed with magnetic resonance imaging for neurological deficits and functional outcome poststroke: a critical review of the literature. Neurorehabilitation Neural Repair. 2006;20(4):492-502.

15.

Kim

Winstein

. Can neurological biomarkers of brain impairment be used to predict poststroke motor recovery? a systematic review. Neurorehabilitation Neural Repair. 2017;31(1):3-24.

16.

Chen

S-Y

Winstein

. A systematic review of voluntary arm recovery in hemiparetic stroke. J Neurol Phys Ther. 2009;33(1):2-13.

17.

Hier

Edelstein

. Deriving clinical prediction rules from stroke outcome research. Stroke. 1991;22(11):1431-1436.

18.

Kwakkel

Wagenaar

Kollen

Lankhorst

. Predicting disability in stroke-a critical review of the literature. Age Ageing. 1996;25(6):479-489. doi:10.1093/ageing/25.6.479.

19.

Atkins

Best

Briss

, et al. Grading quality of evidence and strength of recommendations. BMJ. 2004;328(7454):1490. doi:10.1136/bmj.328.7454.1490

20.

Mandrekar

. Receiver operating characteristic curve in diagnostic test assessment. J Thorac Oncol. 2010;5(9):1315-1316. doi:10.1097/jto.0b013e3181ec173d.

21.

Hong

Tosun

Kurland

Gerzanich

Schreibman

Simard

. Biomarkers as outcome predictors in subarachnoid hemorrhage - a systematic review. Biomarkers. 2014;19(2):95-108. doi:10.3109/1354750x.2014.881418.

22.

Association of neutrophil to lymphocyte ratio on 90-day functional outcome in patients with intracerebral hemorrhage undergoing surgical treatment. World Neurosurg. 2018;119:e956-e961.

23.

Tschoe

Bushnell

Duncan

Alexander-Miller

Wolfe

. Neuroinflammation after Intracerebral Hemorrhage and Potential Therapeutic Targets. Journal of Stroke. 2020;22(1):29-46.

24.

Tang

Liu

Zhou

, et al. Role of NADPH oxidase in the brain injury of intracerebral hemorrhage. J Neurochem. 2005;94(5):1342-1350.

25.

W-H

Wang

W-H

Dong

X-Q

, et al. Prognostic significance of plasma copeptin detection compared with multiple biomarkers in intracerebral hemorrhage. Clin Chim Acta. 2014;433:174-178.

26.

Castillo

Dávalos

Alvarez-Sabín

, et al. Molecular signatures of brain injury after intracerebral hemorrhage. Neurology. 2002;58(4):624-629.

27.

Chen

Feng

Zhang

Tang

. NLRP3 inflammasome contributes to inflammation after intracerebral hemorrhage. Ann Neurol. 2014;75(2):209-219.

28.

Keep

Hoff

. Mechanisms of brain injury after intracerebral haemorrhage. Lancet Neurol. 2006;5(1):53-63. doi:10.1016/s1474-4422(05)70283-0.

29.

Wang

Doré

. Inflammation after intracerebral hemorrhage. J Cerebr Blood Flow Metabol. 2007;27(5):894-908.

30.

Agnihotri

Czap

Staff

Fortunato

McCullough

. Peripheral leukocyte counts and outcomes after intracerebral hemorrhage. J Neuroinflammation. 2011;8:160.

31.

Askenase

Goods

Beatty

, et al. Longitudinal transcriptomics define the stages of myeloid activation in the living human brain after intracerebral hemorrhage. Science immunology. 2021;6(56).

32.

Di Napoli

Godoy

Campi

, et al. C-reactive protein level measurement improves mortality prediction when added to the spontaneous intracerebral hemorrhage score. Stroke. 2011;42(5):1230-1236.

33.

Diedler

Sykora

Hahn

, et al. C-reactive-protein levels associated with infection predict short- and long-term outcome after supratentorial intracerebral hemorrhage. Cerebrovasc Dis. 2009;27(3):272-279. doi:10.1159/000199465.

34.

Qian

S-Q

S-R

B-B

Qian

Zheng

X-D

. Serum S100A12 and 30-day mortality after acute intracerebral hemorrhage. Clin Chim Acta. 2018;477:1-6.

35.

Garton

Keep

Hua

. CD163, a hemoglobin/haptoglobin scavenger receptor, after intracerebral hemorrhage: functions in microglia/macrophages versus neurons. Translational Stroke Research. 2017;8(6):612-616.

36.

Lin

X-F

Ten

X-L

Tang

X-B

Chen

. Serum soluble CD40 ligand levels after acute intracerebral hemorrhage. Acta Neurol Scand. 2016;133(3):192-201. doi:10.1111/ane.12445.

37.

Jiang

Y-X

Zhang

G-H

Wang

Z-M

Yang

. Serum YKL-40 levels as a prognostic factor in patients with intracerebral hemorrhage. Clin Biochem. 2014;47(18):302-306.

38.

Zhao

D-Q

Wang

Zhang

H-D

Y-J

. Significant reduction of plasma gelsolin levels in patients with intracerebral hemorrhage. Clin Chim Acta. 2013;415:202-206.

39.

Lauridsen

Hvas

A-M

Sandgaard

, et al. Coagulation profile after spontaneous intracerebral hemorrhage: A cohort study. J Stroke Cerebrovasc Dis. 2018;27(11):2951-2961.

40.

Zhou

Bao

Wang

Pan

. S100β as a biomarker for differential diagnosis of intracerebral hemorrhage and ischemic stroke. Neurol Res. 2016;38(4):327-332.

41.

Xiong

Yang

Zhang

. The use of serum glial fibrillary acidic protein test as a promising tool for intracerebral hemorrhage diagnosis in Chinese patients and prediction of the short-term functional outcomes. Neurol Sci. 2015;36(11):2081-2087.

42.

Cai

J-Y

Chen

M-H

, et al. Predictive value of phosphorylated axonal neurofilament subunit H for clinical outcome in patients with acute intracerebral hemorrhage. Clin Chim Acta. 2013;424:182-186.

43.

S-J

Xuan

H-F

, et al. Predictive value of serum caspase-cleaved cytokeratin-18 concentrations after acute intracerebral hemorrhage. Clin Chim Acta. 2016;452:124-128.

44.

Zweifel

Katan

Schuetz

, et al. Growth hormone and outcome in patients with intracerebral hemorrhage: a pilot study. Biomarkers. 2011;16(6):511-516.

45.

Qian

S-Q

X-C

S-R

B-B

Zheng

X-D

Pan

G-H

. Prognostic value of serum thioredoxin concentrations after intracerebral hemorrhage. Clin Chim Acta. 2016;455:15-19.

46.

W-H

Dong

X-Q

, et al. Plasma 8-iso-Prostaglandin F2α concentrations and outcomes after acute intracerebral hemorrhage. Clin Chim Acta. 2014;437:141-146.

47.

Saleem

Yousuf

Gul

Gupta

Verma

. Hyponatremia in stroke. Ann Indian Acad Neurol. 2014;17(1):55-57.

48.

Dong

X-Q

Huang

W-H

, et al. Change in plasma copeptin level after acute spontaneous basal ganglia hemorrhage. Peptides. 2011;32(2):253-257.

49.

Morgenthaler

Müller

Struck

Bergmann

Redl

Christ-Crain

. Copeptin, a stable peptide of the arginine vasopressin precursor, is elevated in hemorrhagic and septic shock. Shock. 2007;28(2):219-226.

50.

Dobsa

Cullen Edozien

. Copeptin and its potential role in diagnosis and prognosis of various diseases. Biochem Med. 2013;23(2):172-192.

51.

Liu

, et al. Association of molecular markers with perihematomal edema and clinical outcome in intracerebral hemorrhage. Stroke. 2013;44(3):658-663.

52.

Wang

L-G

Huangfu

X-Q

Tao

Zhong

G-J

Z-D

. Serum tenascin-C predicts severity and outcome of acute intracerebral hemorrhage. Clin Chim Acta. 2018;481:69-74.

53.

Zheng

Sun

Yang

Zhao

Fan

. The potential role of vascular endothelial growth factor as a new biomarker in severe intracerebral hemorrhage. J Clin Lab Anal. 2017;31(5):e22076. doi:10.1002/jcla.22076.

54.

Qiu

S-Z

Wang

H-X

Shen

, et al. The prognostic value of serum signal peptide-Cub-Egf domain-containing protein-1 concentrations in acute intracerebral hemorrhage. Clin Chim Acta. 2016;461:103-109.

55.

Wang

C-L

Lin

H-Y

J-W

, et al. Blood levels of adrenomedullin on admission predict outcomes after acute intracerebral hemorrhage. Peptides. 2014;54:27-32.

56.

Zhang

X-M

Huang

L-F

. Prognostic value of leptin: 6-month outcome in patients with intracerebral hemorrhage. Peptides. 2013;43:133-136.

57.

Rodriguez-Luna

Rubiera

Ribo

, et al. Serum low-density lipoprotein cholesterol level predicts hematoma growth and clinical outcome after acute intracerebral hemorrhage. Stroke. 2011;42(9):2447-2452.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.14 MB