Cerebral Hemorrhage: Pathophysiology, Treatment, and Future Directions

Abstract

Intracerebral hemorrhage (ICH) is a devastating form of stroke with high morbidity and mortality. This review article focuses on the epidemiology, etiology, mechanisms of injury, current treatment strategies, and future research directions of ICH. Incidence of hemorrhagic stroke has increased worldwide over the past 40 years, with shifts in the etiology over time as hypertension management has improved and anticoagulant use has increased. Preclinical and clinical trials have elucidated the underlying ICH etiology and mechanisms of injury from ICH including the complex interaction between edema, inflammation, iron-induced injury, and oxidative stress. Several trials have investigated optimal medical and surgical management of ICH without clear improvement in survival and functional outcomes. Ongoing research into novel approaches for ICH management provide hope for reducing the devastating effect of this disease in the future. Areas of promise in ICH therapy include prognostic biomarkers and primary prevention based on disease pathobiology, ultra-early hemostatic therapy, minimally invasive surgery, and perihematomal protection against inflammatory brain injury.

Introduction

Intracerebral hemorrhage (ICH) is a devastating form of stroke characterized by bleeding into the brain parenchyma. While this form of stroke accounts for only 10% of all strokes in the United States¹ and 6.5–19.6% worldwide,² mortality from ICH remains as high as 50% at 30-days.^3–5 Over the past decade there have been significant advances in the understanding of ICH risk, potential treatments, and outcomes.

Epidemiology

Since the 1980s, the incidence of ICH-related hospitalizations in the United States has largely remained stable at around 20 per 100,000 persons per year,^{3, 6} but global incidence increased 47% between 1990 and 2010, primarily driven by low-income countries.⁷ In high income countries, despite improved hypertension management, it is hypothesized that overall incidence has remained stable due to increased use of antithrombotics, resulting in more anticoagulant-related ICHs.⁸ This idea is supported by a study investigating the incidence of ICH in Cincinnati from 1988 through 1999. This may in part be due to improved access to imaging and a five-fold increase in anticoagulant-associated ICH during the 1990s.⁹

ICH is more common in men, occurs more frequently in the winter months, and incidence increases with age.^{8, 10} In-hospital mortality following ICH decreased between the 1970s and 1980s, but has since remained roughly stable around 35–40%.^{3, 6} One study investigating both early and late case fatality found that from 1985 to 2011 30-day case fatality fell from 40–33%, while 48-hour case fatality did not change.¹¹ Longer-term case fatality rates at 1 year and 5 years remain high with rates of roughly 55% and 70%, respectively.^{8, 12}

Population-based studies of ICH in different racial and ethnic groups have consistently found a higher incidence of ICH in Blacks and Hispanics in the U.S. population when compared to whites.^{10, 13–15} Incidence of ICH in whites increases with age, while this age-related difference does not occur in Blacks.¹³ The increased incidence of ICH in Blacks and Hispanics is attributed to more hypertension in these populations.^{10, 14, 15} Additionally, in-hospital and 30-day mortality are higher in Blacks, and young and middle-aged patients are more commonly affected.^{6, 13} Internationally, the incidence of ICH and the proportion of ICH out of total stroke burden in Japan is nearly 40%, more than twice that in Western countries,^{8, 16} although case-fatality is lower.³

Etiology and Risk Factors

The primary pathoetiologies of spontaneous ICH are chronic hypertension and cerebral amyloid angiopathy (CAA). Secondary pathoetiologies are bleeding from imageable vasculopathies and tumors, and hemorrhagic conversion of ischemic stroke or venous thrombosis. Coagulopathy, platelet dysfunction, and illicit drug use can contribute to ICH or its severity (Table 1).^{17, 18}

Table 1.

Risk factors for spontaneous intracerebral hemorrhage (ICH)

ICH Risk factors
Modifiable	Non-modifiable
○ Hypertension ○ Coagulopathy (medication-related, acquired) ○ Current smoking ○ Excessive alcohol intake ○ Diabetes Mellitus ○ Sympathomimetic/Illicit Drugs	○ Prior ICH ○ Advanced age ○ Male Sex ○ Non-White ethnicity ○ Cerebral amyloid angiopathy ○ Chronic Kidney Disease ○ Coagulopathy (congenital) ○ Tumors ○ Vascular lesions (both genetic and spontaneous): cavernous angiomas, Moyamoya disease or syndrome, AVMs, aneurysms

Traumatic ICH is a distinct etiology from spontaneous ICH. It is beyond the scope of this review but should be considered and ruled out as potential cause of ICH.

Risk factors for primary ICH

Hypertensive microangiopathy

Untreated arterial hypertension increases the risk of stroke 2- to 4-fold. Chronic hypertension is the primary cause of ICH in patients under 70 years old.¹⁹ High arterial pressures cause vascular remodeling at the cellular level resulting in lipohyalinosis and true arteriolar dissections (Charcot-Bouchard aneurysms), which rupture and allow high pressure extravasation of blood into the deep parenchyma.^20–22 Hypertensive bleeds occur in the deep areas of the brain that are supplied by small vessels, including the putamen, caudate, thalamus, brainstem and deep cerebellar nuclei.^{18, 23}

Preclinical studies in murine models suggest that increased arterial stiffening leading to hypertensive vasculopathies and ICH may result from dysregulation of MMPs, elastin, and collagen.^{24, 25} Several other pathological mechanisms have also been proposed to explain these age-related structural and mechanical changes including metabolomic syndrome, neurohormonal and inflammatory disorders, and dysregulation of the renin-angiotensin system.^26–29 An increase in the pro-inflammatory cytokines TNF-α and IL-6, as well as leukocyte infiltration into the perivascular spaces, may explain the dysfunction of both endothelium and extracellular matrix (ECM).³⁰

Cerebral Amyloid Angiopathy (CAA)

CAA is also an age-related disease, where amyloid builds up within the arterial walls, resulting in loss of smooth muscle cells and significantly increases the risk for lobar ICH.³¹ This results in fibrinoid necrosis and weakening of the vessel wall with ultimate rupture into the brain parenchyma. CAA occurs primarily in lobar locations or in the subarachnoid space over the cortical convexities and is associated with occult cerebral microbleeds (CMBs) or superficial siderosis.³² A familial form of CAA occurs earlier in life and is caused by a mutation in the amyloid precursor protein gene.³³ The sporadic form of CAA is characterized by a later age of onset and is broken down is characterized by build-up of amyloid in the cortical capillaries, arteries, arterioles, veins and venules.³⁴ Several studies have suggested that APOE epsilon 2 or 4 alleles are at a greater risk of developing CAA and further ICH.^35–37

The progression from amyloid deposition weakening the blood vessels to CMBs, and further to primary ICH has not been fully elucidated. However, recent studies have suggested that Aβ₄₀, a breakdown product of amyloid precursor protein, disrupts pericytes and contributes to vascular permeability.^{38, 39} Moreover, the breakdown of the ECM may be triggered by the release of various inflammatory cytokines by astrocytes, which causes instability of the tight junctions and increases the release of MMPs.^{40, 41} E4FAD mice, engineered to study the physiopathogenesis of CAA, exhibit CMBs after 2 months of age with similar phenotypic features observed in humans.⁴² This mouse model is generated with three mutations in the Amyloid Precursor Protein (K670N/M671L + I716V + V717I) and two mutations in the Presenilin 1 gene (M146L + L286V), as well as with a targeted replacement of the mouse endogenous APOE allele with the APOE4 allele.⁴³ Both hypertensive angiopathy and CAA are thought to be synergistic, contributing to ICH risk and neurocognitive impairment.

Coagulopathy

Anticoagulation and antiplatelet medications, which are used to treat multiple hematological or vascular disorders, are a known risk factor for spontaneous ICH. Historically, the use of warfarin was associated with an 8- to 19- fold increased risk of ICH.^{44, 45} Newer generation direct acting anticoagulant agents (DOACs) are associated with a significant risk of ICH⁴⁶ but this risk is lower than that seen with warfarin and is associated with smaller hematoma volumes and less severe stroke syndromes.⁴⁷ Whether DOAC-related ICH is associated with lower mortality or better outcomes compared to warfarin-associated ICH remains controversial.^{47, 48} Along with a higher incidence of ICH, patients on anticoagulation also suffer from larger hematomas and elevated risk of mortality.^{49, 50} Antiplatelet therapy contributes to the risk of ICH, greater with clopidogrel and other agents than with aspirin alone,⁵¹ but is lower than that with anticoagulants.

Coagulopathies associated with congenital (most notably hemophilia) or acquired (liver failure, diffuse intravascular coagulation) thrombotic factor deficiencies or thrombocytopenia can cause or exacerbate ICH.⁵² Recently, COVID-19 infection has been associated with small increased risk of ICH resulting from endothelial dysfunction, diffuse intravascular coagulation (DIC), and/or anticoagulation use.^{53, 54}

Drug use and other risk factors

The use of sympathomimetic agents such as amphetamine and cocaine have been implicated given the temporal relationship of use and hemorrhage. Phenylpropanolamine, previously found in appetite suppressants and cough medications, increases the risk of ICH.⁵⁵ It is believed that transient supraphysiologic increases in blood pressure leads to vessel rupture. Cocaine has also been implicated in causing aneurysmal arteriopathy.⁵⁶ Other illicit drugs have not been well studied, but are thought to induce vasculitic changes, vessel wall weakening, and rupture. Current smoking, excessive alcohol consumption, diabetes, and prior vascular disease such as prior ICH or myocardial infraction all augment the risk of ICH.^57–59 Although the pathophysiologic reason is unclear, many studies have found a protective effect of hyperlipidemia on ICH and have in fact reported an association between low serum cholesterol and increased ICH risk.^60–62

Risk factors for secondary ICH

Several vascular lesions have been associated with an increased risk of ICH (Figure 1). These are often referred to as secondary ICH.

Figure 1. Cerebrovascular etiologies of intracerebral hemorrhage. Several vascular pathologies increase the risk for intracerebral hemorrhage (ICH).

Hypertension microangiopathies (top left) present as microhemorrhages primarily localized in the deep brain structures following chronic hypertension due to structural and mechanical age-related changes within arterial walls. Microhemorrhages caused by cerebral amyloid angiopathy (top right) on the other hand, are generally found within lobar regions, and are caused by amyloid deposit which weakens vascular integrity. Arteriovenous malformations (middle left) are abnormal connections between arterial and venous vasculature. Cavernous Angioma disease (middle right) is a neuro-vascular disorder characterized by dilated leaky capillaries following dysfunctional endothelium. Moyamoya (bottom left) is a rare condition causing vascular constriction within the arteries at the Circle of Willis and ensuing abnormal vessel formation compensating for the blockage of normal blood flow to this area. Finally, aneurysms (bottom right) are dilations of blood vessels caused by a weakening in the parent artery angioarchitecture.

Cavernous angioma disease

Cavernous angioma (CA) disease is an autosomal dominant neurovascular disorder characterized by endothelial cell dysfunction and dysregulated neurovascular unit angioarchitecture.^{63, 64} Two forms of the disease exist, a familial form and a sporadic form, both harboring mutations in one of three genes, CCM1, CCM2, and CCM3.⁶⁵ Cerebral cavernous malformation (CCM) proteins form a complex that closely interacts with tight junction proteins and various signal transduction pathways.⁶⁶ Murine and zebrafish models with mutations in any of these three genes have been engineered to study the physiopathogenesis of CAs, and also the phenotypic features including hemorrhage, lesion development, and response to therapeutics.^{67, 68} Recently, several dysregulated pathways associated with endothelial cell senescence, loss of cell-cell junctional integrity, neuro-inflammatory processes, and increased endothelial layer permeability were commonly identified between CA disease and the aging brain.⁶⁹ These overlapping mechanisms suggest that CA disease may help to define an investigative paradigm to elucidate biological processes involved in CMBs and further ICH.

Moyamoya

Moyamoya is a rare abnormal hyperplasia of smooth-muscle cells and luminal thrombosis in arteries within the basal ganglia.⁷⁰ This dysplasia exists in two forms (i.e., disease or syndrome), with similar angiographic findings but different physiopathologic mechanisms, leading to the idiopathic or acquired underlying basal arterial occlusive disease and predisposing to collateral cerebral arteries.⁷¹ In both forms, there is irregular elastic lamina with attenuated tunica media, and a degradation of the arterial wall occurs due to caspase-dependent apoptosis.⁷² The genetic from of the disease has been associated with an inherited autosomal dominant mechanism with incomplete penetrance on chromosomes 3, 6, 8, and 17.^73–76

Arteriovenous malformations (AVMs)

AVMs cause bleeding from high flow arteriovenous shunting, gradually weakening arterial feeders, nidal vessels, and/or arterialized veins.⁷⁷ Dural arteriovenous fistulae bleed primarily from arterialized retrograde leptomeningeal venous drainage.⁷⁸ The genetic form of AVMs is associated with Hereditary Hemorrhagic Telangiectasias (HHTs), which have been linked to a loss of function in one allele of the Endoglin (ENG), Activin-like kinase receptor 1 (ALK-1), and SMAD4 genes.^79–81 ENG and Alk-1 proteins are involved in vascular stabilization, while Alk-1 and Smad4 are required for smooth muscle recruitment.^82–85 Sporadic AVMs have been associated with polymorphisms of the ALK-1 or ENG gene.⁸⁶ Murine models developed to study HHT and AVMs have shown highly unpredictable phenotypes.⁸⁷

Aneurysms

Aneurysms are dilatations of blood vessels caused by weakening of blood brain barrier angioarchitecture.⁸⁸ Cerebral aneurysms are usually sporadic but can be associated with autosomal dominant genetic disorders such as polycystic kidney disease, Marfan syndrome, fibromuscular dysplasia, or Ehlers-Danlos syndrome, as well as AVMs.⁸⁹ Risk factors for aneurysms include hypertension, cerebral blood flow patterns, and smoking, although the mechanisms leading to the development of aneurysm formation, growth, and rupture remain unclear.^{90, 91} Murine models to study the physiopathogenesis of aneurysms include ligation of the common carotid artery with induced hypertension and Angiotensin 2 with elastase treatment.⁹² Aneurysm rupture primarily causes subarachnoid hemorrhage (SAH), which is beyond the scope of this review, but can also cause ICH if they bleed into the brain parenchyma.

Neuroimaging

The computed tomographic (CT) scan is a simple and highly sensitive method for diagnosing ICH and intraventricular hemorrhage (IVH), in view of the blood’s higher density than adjacent brain and spinal fluid. The CT scan also allows accurate estimate of ICH volume using ABC method⁹³ and grading the extent of IVH,^{94, 95} both impacting clinical outcome. Earlier access to CT scan has allowed rapid diagnosis of ICH, sooner after symptom onset, and during a phase of ongoing expansion.⁹⁶ Early features on CT scan (blend, island, swirl, and black hole signs) portend a higher likelihood of hemorrhage expansion. Specific patient (age, sex, absence of hypertension and impaired coagulation) and non-contrast CT findings are incorporated into the Secondary ICH Score, which can be used clinically to estimate the added value of obtaining contrast-enhanced imaging.⁹⁷ Contrast enhanced and angiographic CT scan allows the imaging of other features such as spot sign (contrast leakage also portending subsequent hemorrhage growth) and may reveal underlying aneurysms and arteriovenous malformations (secondary ICH).^97–100

Magnetic resonance imaging allows better imaging of hematoma evolution over time and is more sensitive than CT for detecting underlying tumors, arterial or venous infarcts, and angiographically occult vascular malformations.^{101, 102} Imaging of the brain parenchyma on MRI also allows the grading of extent of associated ischemic cerebrovascular disease, and new gradient echo and susceptibility MRI sequences reveal CMBs.^{103, 104} White matter hyperintensities (WMH) and CMBs, markers of cerebral small vessel disease seen on MRI, are risk factors for spontaneous ICH^{105, 106} and are known to be associated with worse outcomes.^{103, 107} The mechanism for this relationship is unclear, as studies have found conflicting results regarding the relationship between these markers of small vessel disease and hematoma characteristics associated with poor outcomes including ICH volume and hematoma expansion (HE).^108–112

Rates by etiology

The risk of ICH increases with age, with an estimated incidence of 2/100,000/year under age 40, then increasing each decade, to 350/100,000/year over age 80.¹ This is mostly due to an increasing prevalence of untreated hypertension in ages 40–70, and a steep increase in lobar ICH from presumed CAA in older patients. Most clinical series attribute cerebrovascular malformations as the most likely cause of ICH below age 40, followed by coagulopathies, tumors, and other etiologies.¹¹³

Mechanisms of injury following ICH

Classically, injury from ICH is separated into primary injury, the initial damage from the hemorrhage itself, and secondary injury, the damage related to downstream pathways activated in the presence of intraparenchymal blood. Primary injury is related to mass effect from the initial hematoma, HE, and hydrocephalus, and occurs immediately after the hemorrhage through the first few days.^{22, 114, 115} Secondary injury, which occurs over days to weeks, is the result of activation of injurious pathways including inflammation, iron and blood-related toxicity, and oxidative stress.^{22, 116, 117} While these primary and secondary pathways of injury are pathophysiologically distinct, there is mechanistic overlap in how they cause injury to the brain parenchyma (e.g. edema development leads to additional mass effect, intraventricular blood may cause hydrocephalus both from obstruction and inflammatory effects), and as such this conceptual divide may limit the way we identify therapeutic targets. Here we will reframe this discussion to focus on damage following ICH as related to early hematoma growth, mass effect, and increased intracranial pressure (ICP); IVH and hydrocephalus; perihematomal edema and inflammation; and additional medical complications (Figure 2).

Figure 2. Mechanisms of Injury After Intracerebral Hemorrhage.

Mass effect, increased ICP, and hematoma expansion

In addition to causing disruption to the physical structure of the brain, mass effect from ICH can result in midline shift, which when severe causes herniation and compression of critical structures in the brainstem. As such, ICH volume is the strongest predictor of mortality with volume >60cm³ associated with 91% 30-day mortality.¹¹⁴ Elevated ICP, related to mass effect from the initial hematoma, HE, or obstructive hydrocephalus from IVH, can also result in neurologic worsening and herniation.

Patients who survive the initial bleeding event remain at risk for herniation and mass effect from HE and developing perihematomal edema. HE is common following ICH, with more than 70% of patients presenting within 3 hours of symptom onset demonstrating some hematoma growth within 24 hours.¹¹⁸ Studies of HE have used various percentage (≥25%, ≥33%, ≥40%) and absolute (≥3cm³, ≥6cm³, ≥12.5cm³) cutoffs to characterize hematoma growth, and all of these definitions are associated with mortality and poor functional outcome.^118–120 Early HE is common, with 26% of patients having HE on 1-hour repeat CT scan, and an additional 12% with expansion between 1-hour and 20-hour CT scans.¹¹⁹ HE is greatest among patients who undergo CT within 3 hours of symptom onset. Among these patients, 20% demonstrate expansion of ≥12.5cm³ or ≥40% within 48 hours, with 83% of this expansion occurring within 6 hours and 100% within 24 hours from symptom onset.¹²¹ In patients with oral anticoagulant-associated ICH (OAT-ICH), HE is more common, associated with greater mortality, and occurs later when compared to patients not taking oral anticoagulants at the time of their hemorrhage.^{49, 50}

IVH and hydrocephalus

Roughly 40% of ICH patients have IVH on initial imaging.¹²² Presence of IVH has consistently been shown to be associated with poor prognosis and increased mortality,^{114, 123} and greater IVH volume is associated with worse outcomes,^{124, 125} as is IVH causing obstructive hydrocephalus.¹²⁶ Initially, IVH can cause obstructive hydrocephalus, necessitating placement of an external ventricular drain (EVD) to allow removal of cerebrospinal fluid (CSF) to prevent development of and treat elevated ICP. While EVD placement can help to alleviate elevated ICP, improving survival, a benefit on functional outcome remains uncertain.^127–129 IVH has additional deleterious effects on the brain beyond those related to acute obstructive hydrocephalus. In addition to the damaging effect of intraparenchymal iron, as discussed below, intraventricular blood breakdown products have been shown to be associated with higher extracellular iron concentrations and increased secondary brain injury in subarachnoid hemorrhage.¹³⁰ In animal models of ICH, IVH has been shown to cause inflammation of the cells of the choroid plexus and ependymal lining of the ventricles, resulting in CSF hypersecretion and persistent post-hemorrhagic hydrocephalus.^{131, 132} The relationship between IVH and hydrocephalus is complex, as is the mechanism of poor outcome due to IVH.

Perihematomal edema, inflammation, and iron-induced oxidative injury

The mechanisms contributing to the development of perihematomal edema (PHE) are complex and incompletely understood. Within the first 24 to 72 hours the primary etiology is diffusion of serum proteins into the surrounding parenchyma, causing vasogenic edema.¹³³ Over the subsequent 7 to 14 days, PHE continues to expand in the setting of several classical secondary injury pathways, including release of thrombin, inflammation, and iron-induced injury.^{117, 134–136}

Thrombin, a serine protease that is a part of the coagulation cascade, is important for clot formation and prevention of HE via its role in the cleavage of fibrinogen to fibrin.¹³⁷ In animal models, pre-treatment with low dose thrombin decreases subsequent development of PHE, likely due to a combined effect of induction of heat shock proteins and endogenous thrombin inhibitors following activation of thrombin receptors.^{138, 139} However, the release of high concentrations of thrombin into the brain parenchyma after ICH contributes to the development of PHE in the first 5 days following ICH,¹⁴⁰ and treatment with the thrombin inhibitor argatroban reduces the development of PHE.^{141, 142} Additionally, injection of heparinized blood in animal models of ICH results in less perihematomal edema compared with unheparinized blood, suggesting that early blood clot formation by thrombin plays a role in the development of PHE.¹⁴³ This theory is reinforced by the finding that warfarin-related ICH is associated with less early PHE.¹⁴⁴ Several studies have suggested a role for thrombin in neuroinflammation following ICH, but the exact interaction between the coagulation cascade and the immune system in this context remains unclear. In addition to thrombin, fibrin has been shown in autoimmune and neurodegenerative diseases to play a role in neuroinflammation, and targeted inhibition of fibrin reduces neuroinflammation in these conditions.^{145, 146} In a mouse model of ICH, inhibition of fibrin formation with hirudin reduced neuroinflammation and improved neurologic outcome.¹⁴⁷ The role of the coagulation cascade in PHE and outcomes following ICH is complex, and further investigation into the mechanisms of injury and interaction with the immune system is necessary. Similar to early worsening related to mass effect from the hematoma itself and HE, several studies found an association between neurologic deterioration and mass effect secondary to PHE in the first 2 to 3 weeks following ICH.^{135, 136} Others have found no independent association between PHE and neurologic worsening when initial ICH volume is taken into account,^{148, 149} so the role of PHE in outcome remains unclear.

Following ICH, both local central nervous system (CNS) and systemic immune responses are activated, which result in several pro-inflammatory and anti-inflammatory phases that cause initial damage followed by subsequent tissue repair.¹⁵⁰ Necroptosis, inflammatory cell death, causes the release of local inflammatory factors within the first 6 hours following ICH.¹⁵¹ This process propagates an inflammatory cascade that results in the release of damaging proinflammatory factors within the perihematomal area.^{152, 153} Subsequently, matrix metalloproteinases (MMPs) are produced by CNS-resident microglia, astrocytes, and endothelial cells.^154–156 MMPs facilitate blood brain barrier breakdown and upregulation of cell adhesion molecules in the perihematomal region, which allows recruitment of systemic leukocytes including innate immune cells (NK cells, monocytes, and neutrophils), followed by cells of the adaptive immune system including T cells, which peak around day 5.^157–160 These systemic immune cells are then activated within the CNS, and they produce additional pro-inflammatory factors via the NF-κB/toll-like receptor (TLR)4 pathway.^157–159 Although the exact dynamics of this pathway are not entirely clear, it appears to play an important role in development of PHE and outcomes after ICH, as it has been shown that TLR4 inhibition improves post-hemorrhagic hydrocephalus¹⁵⁹, TLR4-deficient mice have improved functional outcome,¹⁵⁸ TNF-α-deficient mice have less edema following ICH,¹⁶¹ and administrative of minocycline reduces upregulation of TNF-α and MMP-12.¹⁵⁷After the initial pro-inflammatory phase, anti-inflammatory pathways are important for neurological recovery.^{157, 162} During this resolution phase, anti-inflammatory cytokines downregulate production of pro-inflammatory mediators to promote healing and result in improved functional outcomes.¹⁶² Additionally, erythrocyte clearance by macrophages is necessary for hematoma resolution and recovery from ICH,¹⁶³ which is partially regulated through activation of macrophages and microglia.^164–166 Neutrophils have also been shown to aid in recovery in response to IL-27.¹⁶⁷ This complex balance between damaging early pro-inflammatory cytokines and recruitment of systemic immune cells and subsequent anti-inflammatory pathways that are essential for hematoma resolution highlights the need for improved understanding of the role of the immune system following ICH for development of appropriately targeted therapeutics.

Another major contributor to both brain injury and perihematomal edema following ICH is the release of hemoglobin and its breakdown products from erythrocytes.^{168, 169} In rodent models of ICH, infusion of whole blood, blood-breakdown products, and collagenase all induce perihematomal edema, iron deposition, and neuronal injury.^168–170 Iron chelation with deferoxamine decreases edema and improves outcomes in rodent models.^169–171 In humans, imaging studies show that perihematomal iron concentration progressively increases during the first 30 days after ICH, and iron overload correlates with brain edema.^{172, 173} Oxidative DNA damage, seen in the perihematomal area in experimental models of ICH, is thought to be related to increased production of reactive oxygen species (ROS) and lipid peroxidation in this region in response to iron overload.^{174, 175} In preclinical models of ICH, upregulation of HO-1 causes an increase in oxidative cell damage and a decrease in the free radical scavengers copper/zinc superoxide dismutase and manganese superoxide dismutase.^{174, 175} In humans, elevations in oxidative DNA injury following ICH is demonstrated by the upregulation of the DNA damage product 8-hydroxy-2’-deoxyguanosine (8-OHdG) in peripheral leukocytes.¹⁷⁶ Despite promising preclinical data and preliminary human data showing safety and feasibility of treatment with the iron chelator deferoxamine in patients with ICH,¹⁷⁷ the Intracerebral Haemorrhage Deferoxamine (i-DEF) trial found that treatment with deferoxamine was safe, but found no improvement at outcomes at 90 days and a trend towards efficacy at 180 days.¹⁷⁸ It was initially thought that this perihematomal production of reactive oxygen species resulted in apoptosis of cells in the perihematomal region, but ferroptosis, a form of non-apoptotic, oxidative stress-induced cell death, has been identified.¹⁷⁹ Ferroptosis is induced by accumulation of iron-induced ROS, which results in intracellular oxidative stress, activation of activating transcription factor 4 (ATF4), and transcriptionally regulated cell death, and inhibition of neuronal ferroptosis in preclinical models of ICH reduces perihematomal cell death.^180–184

Additional medical complications

Acute seizures within the first 24–72 hours after ICH are common, with incidence in various studies between 4 and 42% and an even higher incidence when including electrographic seizures without clinical correlate.^185–189 Seizures in ICH patients are associated with lobar location of hemorrhage, presence of SAH or subdural hemorrhage (SDH), and complications including rebleeding.^{185, 186, 190} The association of seizures with outcome is less clear, with studies finding an association with acute neurologic worsening and a trend toward worse outcomes,¹⁸⁷ but other studies showing that early seizures are not associated with in-hospital mortality¹⁸⁵ or long-term neurologic outcome.¹⁸⁹

Fever is common in the first 72 hours following ICH, occurring in 30–45% of patients, and is most commonly seen in patients with intraventricular hemorrhage.^{191, 192} While some ICH patients have fevers related to infectious etiologies, neurogenic or central fevers are common following ICH and are associated with the presence of intraventricular blood.¹⁹³ Hyperpyrexia in this time period is independently associated with worse outcomes and mortality,^{192, 194, 195} as well as other complications of ICH including HE, increased ICP, PHE, early neurologic deterioration, longer ICU and hospital length of stay.^196–199

Hyperglycemia is also frequently observed in the first 72 hours following ICH, with nearly 60% of patients having elevated glucose on admission regardless of history of diabetes.²⁰⁰ Similar to fevers, hyperglycemia during this time period is independently associated with worse outcomes and early mortality in addition to HE, PHE, and elevated ICP.^200–204

Treatment

Despite the completion of many preclinical and clinical trials over the past several decades, there is no single treatment that has been shown to significantly improve mortality and neurologic outcome after ICH. ICH management guidelines focus on prevention of ICH through risk factor management, medical management to prevent worsening following initial hemorrhage, and consideration of surgical management for certain patients.

Prevention

Given the high morbidity and mortality of ICH, primary and secondary prevention of ICH through appropriate management of modifiable risk factors portends the best prognosis.

Hypertension

As noted above, poorly controlled hypertension remains a major risk factor for ICH, and ICH risk increases as degree of hypertension increases. Relative risk of ICH increased from 2.20 with baseline systolic blood pressure (SBP) 140–159mmHg to 3.78 with baseline SBP≥160mmHg in one study,⁶¹ with an adjusted risk ratio for ICH of 2.48 per 20mmHg increase in SBP in a study of Korean patients with stroke.²⁰⁵ Given the major contribution of poorly controlled hypertension to development of ICH, strict management of blood pressure remains a major factor in ICH prevention. The PROGRESS trial showed a 49% relative risk reduction in recurrent stroke when patients were treated with perindopril after ICH, compared with 26% in patients with prior ischemic stroke, suggesting that effective hypertension management is especially effective in preventing recurrent stroke in this population.²⁰⁶

Antithrombotic management

Use of antithrombotic agents remains a mainstay for prevention of cardiovascular events and ischemic stroke, but many patients with risk factors for ischemic stroke are also at risk for ICH. As such, expert consensus is that the individual risk of ICH and ischemic stroke must be considered when initiating antiplatelet agents or anticoagulants in these patients given the increased risk of ICH with their use, as discussed above.^{207, 208} Older patients taking anticoagulants are at higher risk of ICH than younger patients,^{9, 209} and this is likely due, at least in part, to increased risk of CAA in the elderly population. Several studies have found an increased risk of ICH in anticoagulated patients with imaging markers of CAA including increased risk with higher burden of CMBs,²¹⁰ presence of superficial siderosis,²¹¹ and presence of cortical SAH.²¹² Use of anticoagulation in patients with atrial fibrillation and imaging markers of CAA should be considered carefully on an individual patient basis, as traditional risk stratification scores underestimate the bleeding risk in this population.²¹³

Management of other risk factors

Several systematic reviews and meta-analyses have investigated a wide array of additional risk factors for development of spontaneous ICH, and in addition to hypertension and use of antithrombotics, heavy alcohol use, diabetes, and history of stroke or transient ischemic attack (TIA) are found to be associated with risk of ICH.^57–59 Management of these risk factors, as well as other ischemic stroke risk factors given the association of past stroke or TIA with ICH, should play an important role in prevention of ICH.

Medical management

When a patient presents with a spontaneous ICH, the mainstay of initial treatment is medical management. Management by neurointensivists in a specialized neurocritical care unit has been shown to improve outcomes in neurocritically ill patients including ICH.²¹⁴ Acute medical management is aimed at preventing worsening through prevention of HE via blood pressure management and anticoagulant reversal if indicated, and through prevention and management of secondary brain injury from seizures, elevated ICP, hyperglycemia, and fever.

Blood pressure management

The majority of patients with ICH are hypertensive on presentation,^{215, 216} and persistent hypertension during the initial hours of hospitalization is associated with HE.^{217, 218} Additionally, poor outcomes including death, dependence, and deterioration during hospitalization are associated with persistently elevated BP.²¹⁹ It was previously thought that there was an area of ischemia in the perihematomal region, so acute BP lowering after ICH was controversial. It has since been shown through animal models and imaging studies, that although there is an area of reduced cerebral blood flow surrounding the clot, this decrease does not represent ischemia and is more likely related to decreased metabolism.^220–225

As HE occurs early after ICH, early BP control is critical to preventing poor outcomes, but the ideal target BP remains controversial. Two large randomized controlled trials (RCTs) of acute BP lowering after ICH, Intensive blood pressure reduction in acute cerebral haemorrhage trial 2 (INTERACT2) and Antihypertensive treatment of acute cerebral hemorrhage II (ATACH-2), in addition to the smaller Stroke Acute Management with Urgent Risk-factor Assessment and Improvement (SAMURAI) trial investigated this question (Table 2).^226–230 Both the ATACH-2 and INTERACT2 trials compared acute intensive BP lowering to goal SBP<140mmHg to the standard of care goal SBP<180mmHg, and neither trial showed benefit according to their primary endpoint of death or disability at 3 months. INTERACT2 found improved functional outcomes with intensive BP control without adverse effects,²²⁶ leading the 2015 American Heart Association/American Stroke Association (AHA/ASA) ICH management guidelines to conclude that acute SBP lowering to <140mmHg is safe and can improve outcomes.⁵² ATACH-2, published in 2016, found no effect on outcomes or mortality with an increase in adverse renal events in the intensive group.²²⁷ A subsequent analysis of ATACH-2 patients with deep ICH found that intensive BP control was associated with decreased development of PHE and improved outcomes in patients with basal ganglia hemorrhages.²³¹ Recent studies of the relationship between BP after ICH and outcome have focused on the magnitude of BP reduction and BP variability. In a pooled analysis of patients enrolled in the INTERACT2 and ATACH-2 trials, a reduction in mean SBP to 147mmHg was associated with improved functional status, as was less SBP variability.²³² This association between increased BP variability and poor outcomes has been seen in multiple analyses.^233–236 Similarly, a retrospective cohort study of ICH patients found that greater magnitude of SBP reduction (>40 mm Hg) was associated with worse outcomes.²³⁷ In a pooled meta-analysis of individual patient data from 16 trials, BP reduction was not associated with improved functional outcomes but was associated with reduced HE.²³⁸ Taken together, these results demonstrate the need for further investigation into the optimal SBP target and management of SBP following ICH and highlight a need for individualization of SBP goals based on initial BP and hematoma characteristics, as well as attention to maintaining stable BP in the acute setting.

Table 2.

Trials of blood pressure management in ICH

Trial	Inclusion criteria	Treatment groups	Conclusions
INTERACT, 2008	- Within 6 hours of symptom onset - SBP 150–220mmHg	- SBP<140mmHg (intensive) - SBP<180mmHg (guideline-recommended)	Early intensive BP lowering is feasible, well tolerated, and may reduce hematoma expansion
ATACH, 2010	- Within 6 hours of symptom onset - SBP≥170mmHg	3 tiers: - SBP 170–200mmHg, - SBP 140–170mmHg - SBP 110–140mmHg	3-month mortality lower than expected in all tiers Early BP treatment is safe
SAMURAI, 2013	- Within 3 hours of symptom onset - SBP>180mmHg	All patients with SBP lowered to 120–160mmHg	Higher achieved SBP is associated with worse neurologic outcome
ICH-ADAPT, 2013	- Within 24 hours of symptom onset - SBP>150mmHg	- SBP<150mmHg - SBP<180mmHg	Perihematomal CBF is not reduced by acute BP lowering
INTERACT2, 2013	- Within 6 hours of symptom onset - SBP 150–220mmHg	- SBP<140mmHg (intensive) - SBP<180mmHg (guideline-recommended)	Intensive BP lowering does not reduce death or severe disability, but is associated with improved functional outcomes
ATACH-2, 2016	- Within 4.5 hours of symptom onset - SBP≥180mmHg	- SBP 110–139mmHg (intensive) - SBP 140–179mmHg (standard)	Intensive BP lowering did not reduce death or disability, Slight increase in renal adverse events in intensive group

Anticoagulant and antiplatelet reversal

Incidence of OAT-ICH has increased over the past several decades given increased use of these anticoagulant medications,⁹ and OAT-ICH is now responsible for nearly one quarter of all ICH.^{49, 239} As OAT-ICH is associated with larger hematomas and an increased risk of HE,^{49, 50} acute management of patients should focus on reversal of the anticoagulant and BP control. Reversal of vitamin K antagonists should occur as quickly as possible, as the most important determinant of reversal of international normalized ratio (INR) was time to treatment.²⁴⁰ The optimal agents for reversal of warfarin are prothrombin complex concentrates (PCCs) plus Vitamin K, which more effectively reverses INR, prevents HE, and improves outcomes compared to fresh frozen plasma (FFP) with Vitamin K or given alone.^{241, 242}

For management of ICH related to DOACs, reversal depends on the agent. PCCs are ineffective at reversing the anticoagulant-effect of dabigatran,²⁴³ but the monoclonal antibody fragment idarucizumab completely reverses the drug’s effect.^{242, 244} Reversal of Factor Xa-inhibitors is less clear. Several studies have shown reversal of the anticoagulant effect of these agents with PCCs in both in vitro and clinical studies.^{243, 245, 246} Andexanet alfa, a recombinant modified Factor Xa protein, was approved for reversal of these agents after an RCT showed that it was effective at reducing anti-Factor Xa activity and producing hemostasis in patients on Factor Xa inhibitors with major bleeding including ICH.^{247, 248} To date, a single center retrospective cohort study comparing use of andexanet alfa to PCCs for reversal of Factor Xa inhibitors showed an advantage of andexanet over PCCs in terms of mortality and hemostatic efficacy,²⁴⁹ and a recent propensity score-weighted comparative study between patients enrolled in the ANNEXA-4 trial and a retrospective cohort of OAT-related ICH patients treated with PCCs showed a benefit of andexanet in preventing hematoma expansion without a benefit on clinical outcome.²⁵⁰ Since no RCTs have compared the agents, and given the high cost of andexanet, its use has triggered significant controversy and many centers continue to use PCCs for reversal of factor Xa-related ICH.²⁵¹

Reversal of antiplatelet agents after ICH was common prior to the publication of the platelet transfusion versus standard care after acute stroke due to spontaneous cerebral hemorrhage (PATCH) trial in 2016, which showed patients receiving platelet transfusions had increased risk of death or dependence and increased serious adverse events versus those receiving standard medical care.²⁵² One possible contributing factor to adverse outcomes following platelet transfusion in ICH patients is that platelets are not cross-matched prior to transfusion, and transfusion of unmatched platelets may be responsible for these poor outcomes as a recent retrospective study showed a difference in outcomes following transfusion of matched versus unmatched platelets.²⁵³ Currently, platelet transfusion following antiplatelet-associated ICH is not recommended in patients who are not undergoing neurosurgery.²⁴² Desmopressin acetate (ddAVP) stabilizes platelet function, and consideration of use in patients with antiplatelet-related ICH is currently recommended.²⁴²

Administration of alternative hemostatic agents such as tranexamic acid (TXA) and Factor VIIa following ICH is not widely used at this time. A large RCT investigating the use of recombinant Factor VIIa in patients with ICH showed decreased HE, but no improvement in survival or functional outcome.²⁵⁴ TXA is used for hemostasis after traumatic hemorrhage, but its use in spontaneous ICH is not yet widespread as 2 large RCTs failed to show an improvement in neurologic outcome.^{255, 256}

Seizure management

As noted above, seizures are common following ICH, are associated with acute neurologic worsening, and may be associated with worse outcomes.^185–189 Given this association between seizures and acute worsening of neurologic status, as well as findings in traumatic brain injury patients showing a benefit of short-term seizure prophylaxis,²⁵⁷ many clinicians began treating ICH patients with prophylactic anti-seizure medications (ASMs) in the 1990s. Patients given prophylactic ASMs were more likely to have lobar ICH, to undergo EEG monitoring, to have higher National Institutes of Health Stroke Scale (NIHSS) scores, and to be older.^258–260 Several studies sought to investigate the efficacy and safety of this practice, with no studies finding a benefit for prophylactic treatment with ASMs after ICH and several studies showing worse neurologic or cognitive outcomes in ASM-treated patients.^258–261 Given these findings, the AHA/ASA guidelines for management of ICH no longer recommend use of prophylactic ASMs,⁵² but one study found no difference in prescribing patterns despite this recommendation.²⁵⁹

ICP management

As discussed below, definitive monitoring of ICP involves placement of an invasive ICP monitor. In patients with presumed elevated ICP (poor mental status, imaging consistent with herniation from mass effect), empiric treatment can be initiated to help stabilize the patient in the acute setting prior to placement of an ICP monitor. Standard medical management of increased ICP of any etiology includes elevation of the head of the bed to 30°, initiation of sedation, brief hyperventilation, and use of hypertonics such as mannitol or hypertonic saline.^{262, 263} Although mannitol and hypertonic saline are often used interchangeably in practice, a meta-analysis of RCTs of these agents found that hypertonic saline was more effective for management of elevated ICP,²⁶⁴ and this is reflected in the current Neurocritical Care Society Guidelines for the Acute Treatment of Cerebral Edema which recommend use of hypertonic saline over mannitol.²⁶⁵ Additionally, given low quality evidence in favor of one style of management, these guidelines recommend use of either symptom-based boluses or treatment with hypertonic saline to a goal sodium level.²⁶⁵

Management of other medical complications

As mentioned above, fevers are very common following ICH, are associated with IVH, and portend worse short and long-term outcomes,^{191, 192, 194} and the current AHA/ASA guidelines for management of ICH recommend treatment of fever in patients with ICH.⁵² Initial management generally begins with the use of antipyretics, and if these are ineffective patients can be cooled using standard cooling blankets.²⁶⁶ Use of more advanced devices such as water-circulating surface cooling or catheter-based cooling devices has been shown to be more effective at treating fevers,^{267, 268} but no specific modality has been shown to improve outcomes after ICH.

Hyperglycemia is also associated with worse outcomes and increased complications following ICH.^{200, 201, 203} Although the optimal serum glucose has not been established in randomized trials, prevention of hyperglycemia after ICH is associated with improved outcomes and decreased complications including HE, PHE, elevated ICP, and seizures.^{201, 204} This finding is reflected in the AHA/ASA guideline recommendation for monitoring of glucose following ICH and prevention of both hypoglycemia and hyperglycemia in the acute setting.⁵²

Surgical management

With volume of ICH strongly associated with mortality and poor functional outcome, surgical evacuation of ICH has long been advocated. To date no study has shown a compelling functional outcome benefit from surgical intervention. However, in properly selected patients who reach predefined thresholds of evacuation, there is a strong indication of lower mortality and improved functional outcome after evacuation of ICH.

External ventricular drainage, intracranial pressure management, and thrombolytic clearance

ICH causes extravasation of blood into the brain parenchyma, causing mass effect. As discussed above, if the hemorrhage spills into the ventricular system or compresses the cerebrospinal fluid drainage pathways this can lead to hydrocephalus, adding to the elevated ICP and compromise cerebral perfusion pressures. EVD can relieve elevated ICP, provide diversion for CSF drainage, and monitor ICPs in patients with a poor clinical exam (GCS≤8) and evidence of herniation or other signs of elevated ICP on imaging.⁵² Obstructive or communicating hydrocephalus from IVH can also be temporized with CSF diversion. The use of EVD has been associated with lower mortality after ICH/IVH in uncontrolled studies, although a benefit on functional outcome remains controversial.^127–129

The administration of thrombolytics via an EVD to clear obstructive hydrocephalus and enhance the clearance of IVH has been extensively studied and reported.²⁶⁹ The best high-quality data on IVH thrombolysis comes from the Clot Lysis: Evaluating Accelerating Resolution of Intraventricular Hemorrhage (CLEAR)-III trial, which tested the safety and efficacy of an EVD plus r-tPA versus placebo in the management and treatment of IVH causing obstructive hydrocephalus.²⁷⁰ The primary outcome of this trial was negative with no benefit on functional recovery (defined as modified Rankin scale [mRS] 0–3) at 180 days, although there was a significant benefit on mortality. There did appear to be a functional benefit in patients with IVH>20mL, as these patients had a significantly higher likelihood of achieving functional improvement if >80% of the clot was removed.²⁷¹ The use of thrombolytics through an EVD after IVH has been associated with better control of ICP in the CLEAR trial,²⁷² and with lower mortality in nationwide inpatient sample.²⁷³

Cerebellar ICH as a special compartment

The cerebellum is a frequent location for hypertensive ICH.²⁷⁴ Cerebellar hematomas are the most suited for surgical treatment as they can cause direct compression of the brainstem and/or obstructive hydrocephalus, and their evacuation routes do not transverse eloquent brain. Surgical evacuation can reverse the brainstem compression and hydrocephalus, so surgery is often recommended for hematomas >3 cm in diameter (volume ~15 mL) when assessed in the context of a poor neurological exam. Awake patients can be closely observed.^{274, 275}

The recommendation for surgical intervention for symptomatic cerebellar hemorrhages is almost ubiquitous amongst neurosurgeons as a lifesaving procedure. However, functional outcomes have recently been assessed in a study investigating individual participant data from multiple trials by Kuramatsu et al. who found that the proportion of patients with a favorable outcome (mRS 0–3) at 3 months was no different for patients that underwent surgical hematoma evacuation versus non-operative treatment.²⁷⁶ The true benefit for functional outcome will require additional studies in the future as no robust trial has demonstrated clear benefit of functional independence.

Surgical Approaches

The STICH trial (ISRCTN19976990) compared early open craniotomy (Figure 3) with conservative medical management for supratentorial ICH and found no overall statistically significant difference in mortality or functional outcome between treatment groups, but did suggest a potential benefit in lobar ICH.²⁷⁷ STICH II (ISRCTN22153967) then addressed the potential benefit of open surgery for superficial lobar hemorrhages extending within 1cm of the cortical surface, but this trial showed a nonsignificant survival advantage.²⁷⁸

Figure 3. Open surgery for hematoma evacuation of intracerebral hemorrhage that reached the cortical surface.

After demonstrated stability of the ICH via serial CT scans (ICH volume 81 mL) and etiology screening with an angiogram the patient underwent a craniotomy for clot evacuation. (A) CT showing pre-op ICH, (B) photo of open surgery with a craniotomy (C) photo showing clot evacuation cavity with minimal residual blood products. (D) CT showing 5 days post-op with residual volume of clot measured at <15 mL which predicts a favorable probability of having a good functional outcome.

With the results of the STICH trials, a paradigm shift toward less invasive approaches for ICH evacuation was undertaken. Since then, multiple novel approaches and technical adjuncts have been developed and trialed using surgical advances in image-guided and minimally invasive surgery with the goal of mitigating damage to surrounding, unaffected, or salvageable brain. The MISTIE trials explored utilizing minimally invasive catheter-based hematoma evacuation with clot thrombolysis using rt-PA (Figure 4).^{279, 280} The MISTIE III trial (NCT01827046) compared the procedure against best modern medical management and showed a survival benefit but failed to show functional improvement based on the proportion of patients with an mRS 0–3 at 1 year. However, a prespecified analysis of MISTIE III found that reducing the hematoma volume to less than 15 mL (or >70% reduction) conferred a significant probability of achieving an mRS of 0–3 at 1 year,^{279, 280} and a retrospective analysis of the STITCH II trial found similar results.²⁸¹ This identified threshold across two trials of different modalities is a key finding that should be applied to future trials in ICH.

Figure 4. Minimally invasive surgery utilizing the MISTIE protocol of a deep intracerebral hemorrhage from uncontrolled hypertension.

After demonstrated stability of the ICH via serial CT scans (ICH volume 44 mL) and etiology screening with an angiogram the patient underwent real-time clot aspiration in a procedural CT suite. (A) CT showing real-time clot aspiration, (B) photo showing aspiration of clot with gentle suction applied as per the MISTIE protocol, (C) placement of a flexible drain catheter for drainage of residual clot and (D) CT showing residual volume of clot measured at <15 mL which predicts a favorable probability of having a good functional outcome.

Other approaches have been developed for minimally invasive ICH evacuation, including promising feasibility and safety of techniques involving endoscopic evacuation with irrigation.^{282, 283} A meta-analysis of all trials to date have confirmed a likely mortality benefit, but a less robust impact on functional outcome.²⁸⁴ Different techniques did not appear to confer comparative benefit.

Timing of surgery

Timing of surgery has varied greatly in ICH case series, factors such as speed of diagnosis, transfer to the treating institution, delays related to consent and enrollment and operating room availability are variable in different healthcare system settings. Several authors have argued for faster evacuation of ICH than in prior trials,²⁸⁵ while others noted that early surgery can result in higher rebleeding or less efficient ICH evacuation.^{281, 284, 286, 287}

Analysis from the STICH I & II trials (with an intent of “early surgery”) as well as MISTIE III trial (which aimed at “initiating surgery as soon as possible after etiology screening and demonstrating ICH volume stability”) showed that the timing of surgery is nonlinear when assessing outcomes.²⁸¹ For the likelihood of achieving an mRS 0–3, a timing threshold of 62 hours was identified, beyond which a worse functional outcome was more likely. Earlier surgery within this window up to 62 hours had no impact on benefit. Time threshold for a mortality benefit was identified at 48 hours, beyond which there was a greater likelihood of mortality and no benefit from initiating surgery sooner than 48 hours. These thresholds support a broad therapeutic window for intervention for favorable functional outcome and survival. The question of timing will best be addressed in controlled comparisons in future trials.

Long-term outcomes

Survival and functional outcomes

Mortality within the first month after ICH remains as high as 27–50%,^3–5 and the majority of patients who survive are functionally dependent at the time of hospital discharge.²⁸⁸ The most common cause of death during the initial hospitalization for ICH is limitation or withdrawal of life sustaining treatment.²⁸⁹ Factors associated with early mortality make up the ICH Score and include ICH volume, older age, presence of IVH, initial Glasgow Coma Scale score, and infratentorial location of ICH.¹²³ Subsequent studies have sought to expand on this simple grading scale, but none have improved on its ability to predict outcome after ICH. ICH patients also have a high rate of mortality within the first several years following ICH,⁴ with 12-month mortality in population studies around 50%,^{5, 290–292} 5-year mortality around 70%,²⁹² and 10-year mortality greater than 80%.²⁹⁰ Risk factors for long-term mortality in patients who survive the first month after ICH include older age, male sex, heart failure, poor NIHSS at presentation, and poor modified Rankin Scale (mRS) score prior to stroke.^{4, 291} Despite this high mortality and poor outcomes at discharge, studies of long-term outcomes following ICH demonstrate improvement over time and suggest that recovery following ICH occurs over months rather than weeks. When using Barthel Index as a measure of functional outcomes, gradual improvement can be seen out to 12 months,²⁸⁸ but despite this improvement only 14–36% of patients are functionally independent at 1 year.^291–293

Cognition

In addition to more crude measures of neurologic outcome as described above, it is important to consider specific factors that may play a big role in a patient’s quality of life and ability to function independently after ICH. One of the major neurologic domains that is affected by ICH is cognition. An acute decline in cognitive function occurs at the time of the stroke, and many patients exhibit some amount of recovery, but over time it is very common for patients to experience an ongoing gradual decline in their cognitive function over several years.²⁹⁴ In patients without pre-existing dementia, 37% demonstrated cognitive decline during follow up in one study²⁹⁵ and in another study 14% of patients developed dementia at 1 year and an additional 14% by 4 years.²⁹⁶ Patients with cognitive decline after ICH were more likely to have severe cortical atrophy, CMBs, superficial siderosis, and leukoaraiosis on imaging, and were more likely to be older.^295–297

Epilepsy

ICH patients are also at risk for development of late seizures, generally characterized as seizures occurring more than 7 days after their initial bleed. The highest risk for the development of epilepsy following ICH is within the first year, with incidence in different population studies ranging from 7% to 11%.^298–300 Several studies followed patients for up to 20 years, and incidence of seizures increased slightly after the first year, with a 12% risk of seizures at 5 years²⁹⁸ and a 13.5% risk at 20 years post-ICH.³⁰⁰ Risk factors for the development of late seizures are cortical involvement of ICH, early seizures, larger ICH volume, and age <65 years old.^298–301 Post-stroke epilepsy is not independently associated with poor long-term outcomes when adjusting for other markers of cerebral small vessel disease, although further work is needed in this area.^{299, 301}

Recurrent ICH

Following initial ICH, patients are at increased risk of recurrence compared to those without prior hemorrhage. One single center study found recurrence rates of 7.8%/year for patients following lobar ICH and 3.4%/year for non-lobar ICH patients, with a total recurrence rate of about 20% in lobar ICH patients and 7% in non-lobar ICH patients during their study.³⁰² A meta-analysis of 21 ICH studies found an annualized risk of ICH recurrence of 1.8–7.4% in the first year following ICH and 2–2.4% over all of the years studied.²⁹²

BP remains a major risk factor for recurrent hemorrhage. Although classically hypertension is associated with deep hemorrhages, and lobar hemorrhages are associated with CAA, inadequate BP control is associated with recurrence of both lobar and non-lobar ICH.³⁰² This association highlights the fact that BP control is not only important in prevention of hypertensive hemorrhages, but inadequately controlled BP in a patient with abnormal vasculature due to CAA may increase their risk of ICH. Following CAA-related ICH, patients are at increased risk of recurrence of ICH compared to those with ICH related to other etiologies, and this risk increases with a higher CMB burden.³⁰³ In patients with non-CAA-related ICH, recurrent ICH risk only increases in patients with more than 10 CMBs, indicating that these patients have underlying CAA in addition to their other ICH etiology.

Ongoing and future research horizons

Prevention research

The question of when and whether to restart antithrombotics following ICH, especially in patients at high risk for ischemic stroke, remains the subject of much debate and several ongoing trials. The REstart or STop Antithrombotic Randomised Trial (RESTART) investigated the safety of restarting antiplatelet agents following ICH, and they found that the risk of recurrent ICH was likely too small to outweigh the benefits of antiplatelet therapy in patients who require them for secondary stroke prevention.³⁰⁴

Patients with atrial fibrillation requiring anticoagulation and ICH raise an important conundrum and the decision regarding restarting anticoagulants is an ongoing area of active investigation. One study used data from a prospective single center ICH cohort to compare quality-adjusted life years (QALYs) with or without restarting anticoagulation.³⁰⁵ They found that withholding anticoagulation was associated with increased QALYs for lobar ICH patients. and was therefore preferred but was less strongly correlated with increased QALYs in patients with non-lobar ICH, so they concluded that restarting these agents could be considered in patients with non-lobar ICH. In contrast, several population based and cohort studies have shown a decreased risk of thromboembolic events and lower mortality in patients who had oral anticoagulants restarted following ICH,^306–308 with only one study showing high risk of recurrent ICH.³⁰⁹ Two recent meta-analyses found a lower risk of thromboembolic events including ischemic stroke and no increased risk of ICH in patients who had anticoagulants restarted.^{310, 311} These findings highlight the need for RCTs of restarting anticoagulation following ICH, and several are currently ongoing including Anticoagulation in ICH Survivors for Stroke Prevention and Recovery (ASPIRE), Start or STop Anticoagulants Randomized Trial (SoSTART), Study of Antithrombotic Treatment After IntraCerebral Haemorrhage (STATICH), Avoiding Anticoagulation after IntraCerebral Haemorrhage (A3ICH), and Apixaban After Anticoagulation-associated Intracerebral Haemorrhage in Patients with Atrial Fibrillation (APACHE-AF).

Currently, ICH prevention in patients with CAA is focused on BP management, as there are no targeted amyloid therapies available for clinical use. A recent study using a murine model of CAA in mice expressing a human form of APOE4 showed that APOE-targeted immunotherapy with the use of an APOE antibody has the ability reduce amyloid-β (Aβ) deposition and prevent associated pathology,³¹² in contrast to studies of anti-Aβ antibodies under investigation for Alzheimer’s disease which found increased microhemorrhages.³¹³ Further investigation in preclinical and clinical models is required to investigate the potential efficacy of this antibody in stabilizing amyloid deposition and preventing CAA-related ICH.

It is likely that there are additional yet unknown genetic risk factors for ICH. Certain populations including the Japanese and Blacks are at higher risk of developing spontaneous ICH compared to whites,^{8, 10, 13} and it this risk is likely at least partially genetic. Several studies have investigated the role of genetic polymorphisms in ICH risk, with genome-wide association studies showing several polymorphisms associated with increased ICH risk in Japanese patients,^{314, 315} and increased risk of deep ICH in patients with polymorphisms associated with hypertension.³¹⁶ Recently, an antisense oligonucleotide therapy has shown promise in a mouse model of the monogenic Dutch-type CAA.³¹⁷ These studies of the genetic risk of ICH and gene-based treatment options highlight the potential for future novel preventative interventions for ICH.

Biomarker development and multi-omics research

The U.S. FDA/NIH have defined a group to define standards of rigor and categorize Contexts of Use for biomarker development.³¹⁸ Molecular or imaging signatures have been proposed for the diagnosis, etiologic and severity categorization in several neurological and neuro-vascular disorders.^{319, 320} In CA disease, several translational studies have suggested that the biological response observed in the resected lesional transcriptome following hemorrhagic activity was reflected in the plasma of patients with cavernous angioma with symptomatic hemorrhage (CASH).^{320, 321} Integrative analyses of the plasma miRNome of CASH compared to non-CASH patients showed that miR-185–5p was differently expressed. Of interest, this miRNA is a target of IL-10RA, which is also dysregulated in the lesional transcriptome of CASH.³²⁰ Decreased plasma levels of IL-10 have been reported in CCM subjects that experienced recent hemorrhagic activity, but also in age-related ICH.^{320, 322} The transcriptome of resected lesions therefore becomes an archive to identify circulating mechanistically related compounds that can be assessed in the plasma of patients, and paves the way for the development of diagnostic and prognostic biomarkers.

Next, a novel paradigm integrating multi-omic mechanistic discoveries and applying a likelihood-based computational model was developed to generate and improve biomarker candidates of disease activity.³²³ A weighted diagnostic biomarker based on the plasma levels of inflammatory and angiogenic proteins, able to differentiate CASH from non-CASH patients with up to 80% sensitivity and specificity was developed following this integrative approach.³²⁰ Further machine learning simulations showed that incorporating plasma levels of miRNAs to this protein-based diagnostic CASH biomarker may improve the diagnostic association with CASH to up to 87% sensitivity and specificity.³²⁰ There is currently a multi-site initiative to develop highly accurate as well as generalizable diagnostic and prognostic biomarkers of CASH (R01 NS114552), powered to examine the independent effects of confounders including sex, genotype, age, and lesion location. Moreover, a multimodal combination analyses between the diagnostic plasma-based CASH biomarker and permeability and perfusion features assessed in-vivo using MRI have recently showed improved sensitivity to 100% and specificity to 87% to accurately diagnose CASH.³²⁴ Integrating these findings into management guidelines may ultimately improve clinical decision-making in relation to bleeding propensity in a genetic disease of ICH. This strategy of mechanistic development of biomarkers linked to clinical contexts of use can be applied to age related angiopathy, sharing many common mechanisms with CA.⁶⁹ Other approaches of biomarker discovery have pursued a link to aging and neurocognitive changes,^{325, 326} a link between microbleeds and macrobleeds and between ischemic insults to the brain with aging and CMBs.^327–329

Treatment research

As noted above, several recent studies suggest that decreasing BP variability is important for neurologic outcomes after ICH.^{233, 234} The use of personalized BP targets is currently being investigated in patients following endovascular therapy for ischemic stroke and SAH, with studies showing that maintenance of BP within a personalized goal BP based on several measures of autoregulation is associated with better outcomes.^{330, 331} The future of BP management following acute ICH will likely focus on personalized optimization of BP targets as well.

The Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) trial investigated the role of high-dose atorvastatin in prevention of ischemic events, and a post-hoc analysis found a slight increase in non-fatal hemorrhagic stroke causing hesitation in the use of statins in patients with ICH.³³² Several retrospective studies and meta-analyses investigated the relationship between statin use and outcomes following ICH, and there is a growing body of evidence that statin use may be associated with improvements in mortality and functional outcome after ICH.^333–337 Early initiation or continuation of statins after ICH may be a simple intervention that could improve outcomes after ICH and is being investigated in SATURN (NCT03936361).

Although administration of recombinant Factor VIIa within 4 hours of ICH reduced hematoma growth in the FAST trial, its use was not associated with improved survival or functional outcome.²⁵⁴ The rFVIIa for Acute Hemorrhagic Stroke Administered at Earliest Time (FASTEST) Trial is currently enrolling patients to investigate whether administration of this agent in an enriched population (smaller ICH volumes, aged 18–80 years old, within 2 hours of symptom onset, GCS≥8) is beneficial (NCT03496883).

Another potential therapeutic target for ICH is immunodulation. Fingolimod, a sphingosine-1-phosphate receptor modulator approved for the treatment of multiple sclerosis, was tested in a small study and showed improved outcomes in patients with small to moderate sized ICH (NCT02002390).³³⁸ The use of siponimod, a selective modulator of sphingosine-1-phosphate receptors 1 and 5 was recently completed (NCT03338998). A small single-arm proof of concept study of an apolipoprotein E mimetic recently reported safety and, when compared to a control population-based cohort in a different study, improved 30 day outcomes (NCT0316858).³³⁹ Two studies are evaluating the effects of antagonism of the IL-1 receptor using anakinra (NCT03737344 and NCT04834388). A novel small molecule to suppress proinflammatory cytokine overproduction is also being tested in acute spontaneous ICH (NCT05020535). These efforts represent translation from preclinical models and hopefully will further inform the potential for targeting inflammation after ICH.

Research is ongoing to optimize the potential role of minimally invasive surgery in enhancing ICH outcome. Newer techniques utilizing endoscopy and evacuation with irrigation and aspiration are being studied, in addition to novel approaches to enhance thrombolytic clearance. Surgical performance and learning curve need further investigation in relation to different techniques, in view of the importance of the extent of ICH removal on outcome. Optimizing the timing of intervention, and potential case selection are motivating novel hypotheses in upcoming trials. Finally, pharmacologic approaches to enhance hemostasis, minimize acute injury and promote recovery may be used synergistically with surgical evacuation. Surgical approaches can also facilitate local delivery of therapeutic agents at the site of injury.

Knowledge gaps

In addition to the potential treatments for ICH mentioned above, improved understanding of the mechanisms behind ICH-related brain injury, the link between this damage and cognitive decline after ICH, and a more complete understanding of mechanisms of neurorepair after stroke will be critical for developing additional treatments.

Studies have consistently found a link between ICH and subsequent cognitive decline, including in patients without preexisting cognitive impairment or dementia.^{5, 294, 296} A recent systematic review of the literature of cognitive impairment after ICH found that post-ICH cognitive impairment gradually increases over time, and is associated with advanced age, female sex, and prior stroke, but given significant heterogeneity between studies it is difficult to draw concrete conclusions about all of the modifiable and non-modifiable risk factors for this decline.³⁴⁰ Large trials with collection of detailed demographic, clinical, biological, and imaging data will help to better clarify the risk factors for post-ICH cognitive decline to identify those at risk and intervene early to prevent worsening after ICH in the large numbers of ICH survivors.

The pathophysiologic mechanisms of post-ICH neuroprotection, neurogenesis, and repair are complex and remain an active area of investigation. Effectively preventing neuronal death mediated by ferroptosis and pyroptosis are promising areas requiring additional study. Modulation of the activation states of leukocytes and glial cells towards neuroprotective phenotypes or targeting specific mediators of pro-inflammatory injury may hold more promise than blocking all activation or recruitment of these cells, given the importance of immune responses in tissue repair. One potential avenue of interest is the in vivo enhancement of the endogenous erythrophagocytosis capabilities of macrophages and microglia. It is hypothesized neurogenesis and repair occurs following ICH, as several preclinical studies have shown improvement in rodent models of ICH when animals are treated with neural, adipose-derived, or mesenchymal stem cells.^341–344 Whether neural stem cells differentiate into functional neurons or enhance recovery through effects on the tissue milieu remains uncertain. A gap remains in the ability to longitudinally quantify reparative processes using noninvasive advanced neuroimaging, particularly in patients. The combination of minimally-invasive hemorrhage evacuation with direct instillation of neuronal protective or immunomodulatory therapies may offer both mass effect reduction, removal of a large source of oxidative injury, and augment recovery, and requires further investigation. A better understanding of the complex interaction between the neuronal resilience programs, the immune system, growth factor pathways, and stem cells involved in neurorepair will allow for development of additional novel therapeutics that can enhance endogenous repair pathways following ICH.

Conclusion

ICH remains a disease with high morbidity and mortality. However, recent advances in the identification of high-risk populations, mechanisms of vascular malformation formation, pathways contributing to progressive tissue injury after ICH, and an increase in the number of ICH clinical trials provide continued hope for reducing the burden of this disease.