ABSTRACT
Background The inflammatory response within the central nervous system is a key driver of secondary brain injury after hemorrhagic stroke, both in patients with intracerebral hemorrhage (ICH) and aneurysmal subarachnoid hemorrhage (aSAH). In this study, we aimed to characterize inflammatory molecules in the blood and cerebrospinal fluid (CSF) of patients within 72 hours of hemorrhage to understand how such molecules vary across disease types and disease severity.
Methods Biological samples were collected from patients admitted to a single-center Neurosciences Intensive Care Unit with a diagnosis of ICH or aSAH between 2014 and 2022. Control CSF samples were collected from patients undergoing CSF diversion for normal pressure hydrocephalus. A panel of immune molecules in the plasma and CSF samples was analyzed using Cytometric Bead Array assays. Clinical variables, including demographics, disease severity, and intensive care unit length of stay were collected.
Results Plasma and/or CSF samples were collected from 260 patients (188 ICH patients, 54 aSAH patients, 18 controls). C-C motif chemokine ligand-2 (CCL2), interleukin-6 (IL-6), granulocyte-colony stimulating factor (G-CSF), interleukin-8 (IL-8), and vascular endothelial growth factor (VEGF), were detectable in the CSF within the first 3 days after hemorrhage, and all were elevated compared to plasma. Compared with controls, CCL2, IL-6, IL-8, G-CSF, and VEGF were elevated in the CSF of both ICH and aSAH patients (p<0.01 for all comparisons). VEGF was increased in ICH patients compared to aSAH patients (p<0.01). CCL2, G-CSF, and VEGF in the CSF were associated with more severe disease in aSAH patients only.
Conclusions Within 3 days of hemorrhagic stroke, proinflammatory molecules can be detected in the CSF at higher concentrations than in the plasma. Early concentrations of some pro-inflammatory molecules may be associated with markers of disease severity.
INTRODUCTION
Despite significant advances in the acute management of hemorrhagic stroke and its devastating consequences, short– and long-term outcomes remain poor. Mortality remains as high as 40% after intracerebral hemorrhage (ICH) and 35% after aneurysmal subarachnoid hemorrhage (aSAH), and less than 40% of patients regain functional independence.1–5 The inflammatory response driven by the presence of blood within the brain parenchyma, ventricles, and subarachnoid space is a key driver of secondary brain injury.11–15
In animal models of ICH, multiple phases of inflammation occur within the brain.16 Initially, local tissue-resident inflammatory cells are activated, followed by blood brain barrier breakdown, recruitment of circulating pro-inflammatory immune cells, and ultimately activation of tissue repair and anti-inflammatory mechanisms. In aSAH, cerebrospinal fluid (CSF) proinflammatory molecules are associated with vasospasm, delayed cerebral ischemia (DCI), and worse outcomes.17–21 While recent human studies on ICH have reported an initial proinflammatory phase involving central nervous system (CNS)-resident and circulating immune cells followed by an anti-inflammatory reparative phase,22 data on inflammatory processes within the CNS and clinical outcome in patients following ICH and aSAH remains sparse.
Prior studies investigating the inflammatory profile of CSF found variable associations between immune molecules and outcome and focused on ICH or aSAH individually, limiting our understanding of differences in inflammatory responses between the diseases.23,24 In addition, few studies include both plasma and CSF cytokine levels to compare inflammatory responses across tissue compartments. In this study, we sought to more fully characterize CNS inflammation after hemorrhagic stroke in humans by analyzing a panel of inflammatory molecules in biological samples collected within the first 72 hours after symptom onset. To achieve this aim, we compared CSF and plasma levels of various immune molecules; compared CSF levels between ICH, aSAH, and control patients; and investigated relationships between these inflammatory molecules and markers of hemorrhage severity.
METHODS
Study design
In this case-control study, 242 patients with spontaneous ICH or aSAH admitted to the Yale Neurosciences Intensive Care Unit (ICU) between 2014 and 2023 with plasma and/or CSF samples collected for the Yale Acute Brain Injury Biorepository were included. CSF was collected from patients who had external ventricular drains (EVDs) placed as part of their clinical care. Eighteen patients undergoing ventriculoperitoneal shunt placement for normal pressure hydrocephalus (NPH) were enrolled as controls, and CSF was collected during shunt placement. This study was approved by Yale School of Medicine IRB, and consent was obtained from patients or their legally authorized representative. Data reporting adheres to the STROBE guidelines for reporting of observational studies.
Blood and CSF from eligible patients were collected between presentation and 72 hours after symptom onset, and the first sample collected from each patient was included. Patients with plasma and CSF samples collected at the same time (n = 16 ICH patients, n = 17 aSAH patients) were included in our paired samples cohort.
Immunologic Assays
Within 30 minutes of collection, blood and CSF samples were centrifuged, the supernatant was divided into aliquots, and samples were frozen at –80°C. A panel of immune molecules including interleukin-6 (IL-6), granulocyte-colony stimulating factor (G-CSF), IL-8, vascular endothelial growth factor (VEGF), C-C motif chemokine ligand-2 (CCL2), IL-1β, IL-4, macrophage inflammatory protein-1 alpha (MIP-1α), IL-10, tumor necrosis factor alpha (TNF-α), and IL-17 were analyzed using Cytometric Bead Array (BD Biosciences, Franklin Lakes, NJ), a multiplex bead-based immunoassay. Based on prior work in our laboratory showing that VEGF was not detectable in the plasma of hemorrhagic stroke patients at these timepoints (unpublished data), VEGF was not measured in plasma samples. The lower limit of detection for each immune molecule was the concentration of the lowest standard for that assay.
Independent Variables
Independent variables were immune molecules above the lower limit of detection, namely IL-6, CCL2, IL-8, G-CSF, and VEGF. ICU length of stay was extracted from the electronic health record. For patients with multiple ICU admissions during the same hospitalization, the sum of all ICU days was considered their ICU length of stay.
Clinical and Neuroimaging Variables
Demographics, medical history, and hospitalization data was obtained from chart review and follow-up interviews with patients and their families. Clinical and radiographic scores including Glasgow Coma Scale (GCS) scores for all patients and Hunt and Hess scores and Modified Fisher Scale scores for aSAH patients, were extracted from patient charts by trained raters (JY, RC). If disagreement or uncertainty existed between raters, the scores were adjudicated by a board certified neurointensivist (JMB, JK). For ICH patients, imaging was reviewed by a board certified neurointensivist (JMB) who calculated ICH volumes using the ABC/2 method,25 determined whether IVH was present, and classified ICH location. ICH score10 was then calculated using the above data by a trained rater (JY). Patients with primary IVH not attributable to a vascular etiology were characterized as ICH. Patients with a ruptured aneurysm, even if the hemorrhage was largely intraparenchymal, were characterized as aSAH.
Statistical Analysis
Non-normally distributed concentrations of immune molecules were log transformed. In our paired samples cohort, paired t-test was used to compare log CSF and plasma concentrations of the immune molecules. For CSF samples, ANOVA was used to compare log CSF immune molecule concentrations between control, ICH, and aSAH samples. Tukey post hoc test was used to determine differences between groups. Linear regression was used to evaluate for significant associations between log CSF immune molecule concentrations and hemorrhage severity features. There was no missing clinical or demographic data. For samples in which there was missing immune molecule data, those samples were excluded from analysis only of the missing molecule. There was no a priori sample size calculation as this was an analysis of previously collected samples. Based on the 72 CSF samples collected, the study had 80% power to detect a 13% difference or a 95% power to detect a 20% difference in log-transformed immune molecules between control (n = 18), ICH (n = 21), and aSAH (n = 51) samples using a one-way ANOVA with an alpha = 0.05. R software (version 4.3.1) was used for all data management and statistical analysis.
RESULTS
Plasma and/or CSF samples from 260 ICH (n = 188), aSAH (n = 54), and control (n = 18) patients were analyzed. Patients had a median age of 67 years (interquartile range [IQR] 18-95), 51.5% were male, 92.7% were non-Hispanic, and 76.2% were Caucasian (Table 1). In ICH and aSAH patients, the median time from symptom onset to sample collection was 3 [IQR 2] days for CSF and 1 [IQR 2] days for plasma. The measured levels of each inflammatory molecule and lower limits of detection are shown in Supplemental Table 1 for plasma samples and Supplemental Table 2 for CSF samples.
Inflammatory molecules are elevated in the CSF compared to plasma
CCL2, IL-6, G-CSF, IL-8, and VEGF were detectable in CSF of patients within 72 hours after symptom onset of ICH and aSAH. These proinflammatory molecules were present in higher concentrations in the CSF compared to the plasma (Figure 1, p<0.05 for all comparisons).
Heat maps demonstrate higher pro-inflammatory immune molecule in CSF compared to plasma at early timepoints.
In 16 ICH and 17 aSAH patients, paired CSF and plasma samples collected at the same timepoint were available. Mean log concentrations of all detectable inflammatory markers were significantly higher in CSF of ICH patients compared to plasma samples, including CCL2 (CSF 8.16pg/mL vs plasma 4.73pg/mL, p<0.01), IL-6 (CSF 6.53pg/mL vs plasma 3.49pg/mL, p<0.01), G-CSF (CSF 2.95pg/mL vs plasma 1.89pg/mL, p<0.01), and IL-8 (CSF 7.19pg/mL vs plasma 3.96pg/mL, p<0.01; Figure 2A). Additionally, mean log concentrations of these cytokines were also significantly higher in SAH patient CSF compared to plasma samples, including CCL2 CSF 8.36pg/mL vs plasma 4.71pg/mL, p<0.01), IL-6 (CSF 7.07pg/mL vs plasma 3.10pg/mL, p<0.01), G-CSF (CSF 2.79pg/mL vs plasma 1.69pg/mL, p<0.01), and IL-8 (CSF 6.72pg/mL vs plasma 2.57pg/mL, p<0.01; Figure 2B).
In paired samples collected from the same patient at the same timepoint, immune molecule concentrations are higher in the CSF compared to the plasma.
Differences in CSF inflammatory molecules across hemorrhagic stroke type
Compared to CSF collected from controls, patients with both hemorrhagic stroke subtypes had significantly higher CSF concentrations of CCL2 (controls 6.50pg/mL vs ICH 8.24pg/mL vs mean aSAH 8.19pg/mL, p<0.01), IL-6 (controls 1.32pg/mL vs ICH 6.89pg/mL vs aSAH 7.29pg/mL, p<0.01), G-CSF (controls 2.04pg/mL vs ICH 3.46pg/mL vs SAH 3.12pg/mL, p<0.01), and IL-8 (controls 2.23pg/mL vs ICH 7.31pg/mL vs aSAH 6.49pg/mL, p<0.01) (all results reported in mean log concentrations, Figure 3A-D). Patients with hemorrhagic stroke had significantly higher concentrations of CSF VEGF compared to controls, and ICH patients had significantly higher CSF VEGF than aSAH patients (controls 2.25pg/mL vs ICH 5.38 vs SAH 4.02pg/mL, p<0.01; Figure 3E).
Within the CSF, immune molecule concentrations differ between control, ICH, and SAH patients.
Association of inflammatory molecules and features of severity
Patients included in our study demonstrated features associated with severe neurological injury (Table 2).
In CSF collected from ICH patients, a lower log concentration of CCL2 was associated with a larger ICH volume (p=0.037). There was no association between ICH volume and any of the other inflammatory molecules. There were no significant associations between CSF concentrations of the measured inflammatory molecules and admission GCS, ICU length of stay, or ICH score (all p>0.05; Table 3).
For aSAH patients, higher log concentration of VEGF was associated with a lower admission GCS score (p=0.027). Higher log concentrations of G-CSF and VEGF were associated with higher Hunt and Hess scores (G-CSF p=0.020, VEGF p=0.036). Additionally, higher log concentration of CCL2 was associated with higher modified Fisher Scale scores (p=0.017). There were no associations between CSF inflammatory molecules and ICU length of stay in aSAH patients (all p>0.05; Table 3).
DISCUSSION
Inflammation occurs in the CNS after hemorrhagic stroke, and it is possible to detect inflammatory molecules in the CSF of these patients. In this study, we have shown that concentrations of several proinflammatory molecules are higher within the CSF as compared to the plasma of patients in the first 3 days after hemorrhage. These proinflammatory molecules were higher in the CSF in the hemorrhagic stroke patients overall, and within individual patients for whom we had both plasma and CSF samples collected on the same day. Additionally, within the CSF, proinflammatory molecules were higher in ICH and aSAH patients compared to controls and VEGF was significantly higher in ICH patients compared to aSAH patients.
Several clinical and radiographic features are associated with more severe presentation or worse outcomes after ICH and aSAH including admission GCS score and ICU length of stay for both groups,10,26–28 Hunt and Hess score and modified Fisher Scale score for aSAH patients,6,29,30 and ICH score and initial ICH volume for ICH patients.10 In this hypothesis-generating analysis, we found that several of these features of disease severity were associated with CSF inflammatory molecule concentrations. Most notably, higher CSF VEGF and G-CSF were associated with higher Hunt and Hess scores and lower admission GCS, higher VEGF was associated with lower admission GCS score, and higher CCL2 was associated with higher modified Fisher Scale score for aSAH patients. Interestingly, higher CSF CCL2 in this early time window was associated with a smaller ICH volume. We did not find any significant associations between IL-6 or IL-8 concentrations and severity features.
CCL2, a monocyte chemoattractant, IL-8, a neutrophil chemoattractant, and IL-6, a multifunctional proinflammatory cytokine that stimulates production of acute phase proteins, are involved in the early proinflammatory phase that occurs within the CSF after hemorrhage and likely play a role in the recruitment of circulating immune cells to the CNS. In addition to these chemoattractant molecules, cell adhesion molecules responsible for the recruitment of circulating immune cells into the CNS have also been shown to be elevated in the CSF of ICH patients.24 We hypothesize that the presence of these inflammatory molecules early after hemorrhage is not necessarily detrimental, but that increased levels or prolonged elevation beyond the acute period may worsen outcomes and prevent repair mechanisms.
Our findings of elevated proinflammatory molecules early after hemorrhagic stroke is consistent with the prior literature. CCL2 has previously been shown to be higher in the CSF compared to the plasma of aSAH patients within the first 1 to 2 days after hemorrhage, and has also been reported to be associated with elevated IL-6 and correlated with higher risk of vasospasm.31–33 Although we did not assess the relationship between CSF CCL2 and vasospasm in this study, we found an association between early CSF CCL2 and modified Fisher Scale score, a marker of patients at risk for vasospasm and DCI. In ICH patients, CSF CCL2 has been shown to be elevated within the first 1 to 2 days after hemorrhage, and was associated with ICH volume and volume of perihematomal edema.34 In contrast, we find an association of CSF CCL2 with smaller ICH volume, although notably in our study ICH volume did not include the volume of IVH.
Prior studies of CSF IL-6 in hemorrhagic stroke have focused on aSAH patients with studies showing elevated CSF IL-6 compared to patient plasma and control CSF, as well as elevated CSF IL-6 in patients with worse outcome, vasospasm, and DCI.17,18,31,35–40 In ICH patients, elevated early CSF IL-6 is associated with hematoma expansion, larger hematoma volume, worse outcomes, and following intraventricular administration of tissue plasminogen activator.34,41 Our findings are consistent with prior findings of elevated early CSF IL-6 compared to plasma levels and control CSF, but in our small cohort of patients we were not able to assess outcomes or complications, and we did not find an association with disease severity.
CSF IL-8 has been less well studied after hemorrhagic stroke. IL-8 has been found to be elevated in the CSF of aSAH patients in the first week after hemorrhage, along with IL-6.31,42 In ICH patients, CSF IL-8 levels 1 week post-bleed were found to be associated with poor admission GCS and IVH volume.34 In the current study, we found high levels of IL-8 early after hemorrhage in CSF of both ICH and aSAH patients compared to plasma levels and compared to control CSF, although we found no association with disease severity.
In the current study, we report increased levels of VEGF and G-CSF within the CSF after hemorrhagic stroke, and we find associations between these molecules and severity features. There is limited prior investigation of these molecules after hemorrhagic stroke. In aSAH patients, CSF VEGF was elevated compared to patient plasma and control CSF, but was not associated with outcome or vasospasm.31 In ICH, VEGF has been suggested to play an anti-inflammatory role, as it is induced by prostaglandin E2 in human macrophages stimulated with ICH damage-associated molecular patterns.22 Interestingly, CSF VEGF is the only molecule in our study which differed between ICH and aSAH patients, and we showed that it was higher in the CSF of ICH patients despite its classic role in angiogenesis. VEGF may play a dual role in the CNS, as it has been shown to be involved in blood brain barrier permeability, but also appears to play an anti-inflammatory role and is involved in vessel repair.43,44
G-CSF within the CSF of hemorrhagic stroke patients has not been previously investigated, but in a rat model of ICH it was found to be neuroprotective and associated with increased VEGF expression after hemorrhage.45 G-CSF, which is primarily produced by monocytes, is classically involved in the proliferation and differentiation of granulocytes, but has also been shown to be neuroprotective and play a role in plasticity in ischemic stroke.46–48 It is intriguing that early CSF levels of VEGF and G-CSF are associated with features of severity in our study, as these molecules have been previously shown to have some neuroprotective and anti-inflammatory effects, as opposed to CCL2, IL-6, and IL-8 which are predominantly proinflammatory, but this is perhaps not surprising for VEGF which appears to have a time-dependent role in the CNS after injury.
Our study has several limitations and is meant to be hypothesis generating. First, we only collected samples within the first 3 days after ictus, limiting our investigation to the acute inflammatory response. We believe that these early proinflammatory molecules are most likely to be associated with initial disease severity features. Future work with longitudinal CSF samples from hemorrhagic stroke patients will be needed to further understand the temporal changes in these and other inflammatory molecules throughout the disease course. Secondly, we are only able to collect CSF samples from patients with EVDs placed as part of their clinical care, and as such we only have data on more severely affected ICH and aSAH patients. While this is only a subset of ICH and aSAH patients, these patients are at highest risk for poor outcomes so a better understanding their disease course and developing therapeutics for this population is likely to have a substantial impact on their outcome. Our small sample size limits our statistical power but has generated several hypotheses for ongoing investigation into the role of CNS inflammation after hemorrhagic stroke. Finally, our analysis was limited to 11 immune molecules, and thus was not a comprehensive study of all factors that may be associated with hemorrhagic brain injury and outcomes.
Measurement of inflammatory molecules within the CNS is critically important to characterize the inflammatory cascade that occurs after hemorrhagic stroke, and evaluation of inflammatory molecules in the blood is insufficient. By focusing on the temporal trajectory of CNS inflammatory molecules, we aim to better define the kinetics of inflammation in the CNS following hemorrhage. While there is overlap between the inflammation that occurs following ICH and aSAH, given their distinct etiologies there are also likely to be important and informative differences. In this study we have begun to explore the relationship between early inflammatory molecules and disease severity. Future studies on the relationship with clinical outcomes will be important to further define potential therapeutic strategies.
Data Availability
A deidentified version of the data is available upon request.
SOURCES OF FUNDING
Dr. Sansing received support related to this work from NIH/NINDS (R01NS097728 and R21NS108060). Dr. Magid-Bernstein received support related to this work from the Neurocritical Care Society (AWD0007451) and the American Heart Association (23CDA1054469).
DISCLOSURES
All authors report no relevant disclosures.
ACKNOWLEDGEMENTS
The authors would like to thank Manali Phadke for assistance with statistical analysis. Figures were created with BioRender.com
Non-standard Abbreviations and Acronyms
- ICH
- intracerebral hemorrhage
- aSAH
- aneurysmal subarachnoid hemorrhage
- DCI
- delayed cerebral ischemia
- CSF
- cerebrospinal fluid
- EVD
- external ventricular drain
- CCL2
- C-C motif chemokine ligand-2
- IL-6
- interleukin-6
- G-CSF
- granulocyte-colony stimulating factor
- IL-8
- interleukin-8
- VEGF
- vascular endothelial growth factor
- CNS
- central nervous system
- NPH
- normal pressure hydrocephalus
- IL-1β
- interleukin-1 beta
- IL-4
- interleukin-4
- MIP-1α
- macrophage inflammatory protein-1 alpha
- IL-10
- interleukin-10
- TNF-α
- tumor necrosis factor alpha
- IL-17
- interleukin-17
- ICU
- Intensive Care Unit
- GCS
- Glasgow Coma Scale
- IQR
- interquartile range
- ANOVA
- analysis of variance