nature reviews neurology https://doi.org/10.

1038/s41582-024-01035-w

Review article Check for updates

Intracerebral haemorrhage —
mechanisms, diagnosis and prospects
for treatment and prevention
David J. Seiffge 1,7, Simon Fandler-Höfler 2,3,7
, Yang Du2,4, Martina B. Goeldlin , Wilmar M. T. Jolink
1 5
,
Catharina J. M. Klijn 6,7 & David J. Werring 2,7

Abstract Sections

Intracerebral haemorrhage (ICH) is a devastating condition associated Introduction

with high mortality and substantial residual disability among survivors. Key advances in diagnosis
Effective treatments for the acute stages of ICH are limited. However, and aetiology

promising findings from randomized trials of therapeutic strategies, Advances and uncertainties
in acute therapy
including acute care bundles that target anticoagulation therapies, blood
pressure control and other physiological parameters, and trials of Advances in prevention
measures
minimally invasive neurosurgical procedures have led to renewed
Conclusions and future
optimism that patient outcomes can be improved. Currently ongoing directions
areas of research for acute treatment include anti-inflammatory and
haemostatic treatments. The implementation of effective secondary
prevention strategies requires an understanding of the aetiology of
ICH, which involves vascular and brain parenchymal imaging; the use
of neuroimaging markers of cerebral small vessel disease improves
classification with prognostic relevance. Other data underline the
importance of preventing not only recurrent ICH but also ischaemic
stroke and cardiovascular events in survivors of ICH. Ongoing and
planned randomized controlled trials will assess the efficacy of
prevention strategies, including antiplatelet agents, oral anticoagulants
or left atrial appendage occlusion (in patients with concomitant atrial
fibrillation), and optimal management of long-term blood pressure
and statin use. Together, these advances herald a new era of improved
understanding and effective interventions to reduce the burden of ICH.

A full list of affiliations appears at the end of the paper. e-mail: [email protected]

Nature Reviews Neurology

Review article

Key points probably in part owing to recruitment challenges leading to small

sample sizes and a high risk of bias14. However, despite the continued
widespread perception that ICH is a largely untreatable disease, con-
• Intracerebral haemorrhage (ICH) is a common and severe type of siderable advances in our understanding and management of ICH have
stroke associated with severe morbidity and mortality. been made in the past few years, reigniting optimism that new treat-
ment approaches might finally lead to improved outcomes after ICH.
• Advances in acute treatment of ICH, including ultra-early blood In this Review, we provide an update on key advances and remaining
pressure control, anticoagulation reversal and minimally invasive uncertainties in ICH management related to diagnosis, aetiology, acute
neurosurgical procedures, have great potential to improve patient therapy and secondary prevention.
outcomes; bundled approaches might have particular value in acute
settings. Key advances in diagnosis and aetiology
Diagnosing causes of intracerebral haemorrhage
• Diagnosis of the underlying aetiology of ICH is crucial for under­ ICH can be caused by a diverse range of conditions. Rapid and accu-
standing the cause; diagnosis usually requires both vascular and rate diagnosis of ICH and its underlying aetiology is essential to guide
structural (magnetic resonance) imaging. rational acute treatment, prognosis and prevention strategies.
Approximately 80% of ICHs among adults aged 50 years or older
• The most important aetiologies of ICH include arteriolosclerosis, are caused by cerebral small vessel disease (SVD)4 — mainly cerebral
cerebral amyloid angiopathy and macrovascular causes. amyloid angiopathy (CAA) and arteriolosclerosis (also termed deep
perforator arteriopathy or ‘hypertensive arteriopathy’)15 (Fig. 1). In
• ICH aetiology must be determined to assess the risk of recurrent terms of prognosis, ICH due to CAA is consistently associated with a
ICH and other vascular events, which informs the use of individualized much higher risk of ICH recurrence than is arteriolosclerosis (7.4–8.5%
secondary prevention strategies and decisions on the use of antiplatelet versus ~1.0–1.7% per year)16–18. The underlying arteriopathy might also
or anticoagulant treatments for other indications. influence the risk of ischaemic vascular events, including ischaemic
stroke, which is higher than the risk of recurrent ICH for people with
ICH attributable to arteriolosclerosis (4.3% versus 1.1% at 3 months)19.
Introduction The early risk of ischaemic stroke in people with ICH attributable to
In 2019, 27% of all incident strokes globally were caused by intracerebral arteriosclerosis is as high as in people with lacunar ischaemic stroke
haemorrhage (ICH), affecting a total of 3.41 million people1. These (4.3% for ICH attributable to arteriolosclerosis versus 2.9% for people
ICH events accounted for 68 million disability-adjusted life years with lacunar stroke 3 months after the index event)19, an observation of
lost, resulting in a greater disease burden than ischaemic stroke1. potential relevance to guide secondary preventive strategies. A subset
ICH continues to be associated with marked mortality and morbid- (~10–20%) of ICH is related to macrovascular causes4,16,20–22, including
ity, with population-based studies reporting a case fatality rate of arteriovenous malformation and fistula, aneurysm, cavernous mal-
40% at 1 month2, and 1-year and 5-year survival rates of 46% and 29%, formation and cerebral venous sinus thrombosis. Rapid diagnosis of a
respectively3. Contemporary hospital-based cohorts report that only macrovascular cause of ICH might enable immediate intervention and
a third of people with ICH are functionally independent 3 months after guide decisions for subsequent management. Other, rarer, underlying
the stroke event4. causes that might warrant specific treatments include cerebral venous
The trajectory of functional recovery after ICH is perceived by sinus thrombosis, vasculitis, reversible cerebral vasoconstriction syn-
some experts to differ from that of ischaemic stroke, with a slower rate drome and moyamoya disease. As patient characteristics (for example,
of recovery, although systematic studies that support this observation age and ICH location) cannot reliably predict the underlying cause23,
are lacking. Nevertheless, post hoc analyses of randomized controlled acute diagnostic work-up for ICH should usually include imaging of
trials from the past couple of years have shown that people with ICH the brain parenchyma and intracranial vasculature.
can experience considerable functional improvement up to 12 months CT (or magnetic resonance) angiography is usually the first
after injury — well beyond the acute phase — particularly among those step in the investigation of possible macrovascular causes of ICH,
with poor early outcomes at 30 days, implying that rehabilitation and has reasonably high diagnostic accuracy compared with acute
measures should be offered well beyond the first 6 months following intra-arterial digital subtraction angiography24. However, CT angiogra-
ICH5,6. Outcome research in ICH has increasingly focused on previously phy might miss small arteriovenous malformations and fistulae24. The
neglected but important non-motor outcomes7. Long-term functional reference standard for the detection of macrovascular causes of ICH is
decline, cognitive decline and post-ICH dementia are frequent in cohort intra-arterial digital subtraction angiography; however, this procedure
studies with survivors of ICH8–11. Moreover, compared with people with is invasive, requires interventional neuroradiologist expertise and is
ischaemic stroke, anxiety, reduced participation in social roles and associated with complications, such as access site haematoma and
activities, fatigue, pain and impaired bowel function are all substan- cerebral ischaemia; the risk of such complications in people with ICH
tially more common in people with ICH, suggesting an urgent need who undergo this procedure is reported to be 3–5%25,26. Intra-arterial
for further longer-term studies in survivors of ICH7. digital subtraction angiography should therefore be performed only
Despite considerable breakthroughs in the management of ischae- when the level of suspicion for a macrovascular cause is likely to out-
mic stroke, progress in the management of ICH has not kept pace. Over- weigh the procedural risks. The DIagnostic AngioGRAphy to find vas-
all, functional outcomes after ICH are perceived to be poorer than those cular Malformations (DIAGRAM) and DIAGRAM+ scores (Box 1) can
after ischaemic stroke12 although systematic, large-scale, prospective be used to determine the probability of an underlying macrovascular
population-based data are lacking2,3,13. Moreover, therapeutic trials in cause in people with ICH27. Of note, however, the DIAGRAM score was
ICH have failed to provide high-level evidence of treatment efficacy, developed on the basis of data from a prospectively collected cohort

Nature Reviews Neurology

Review article

of people aged 18–70 years, and excluded those aged >45 years with two additional non-haemorrhagic MRI markers: significant burden of
hypertension and ICH in the basal ganglia, thalamus or posterior MRI-visible perivascular spaces in the centrum semiovale; and white
fossa24. In addition, diagnoses of SVD in the DIAGRAM study were based matter hyperintensities in a ‘multispot pattern’. These additions
on CT and not MRI, which is superior to CT in assessing the presence increase sensitivity without compromising specificity, particularly in
and severity of underlying SVD, and might have improved diagnostic non-ICH presentations35.
value in the context of identifying or excluding a macrovascular cause. One unresolved aspect of SVD classification is that of a ‘mixed’
Indeed, a study published in 2024 (ref. 28) suggests improved diag- pattern of neuroimaging markers associated with CAA and arteriolo-
nostic accuracy for macrovascular causes of ICH using the MRI-based sclerosis in the same individual. Whether this pattern truly indicates
MACRO score. The utility of more advanced non-invasive imaging tech- ‘mixed’ CAA and arteriolosclerosis or widespread arteriolosclerosis in
niques (such as dynamic 4D-CT angiography, or dynamic 4D magnetic deep and lobar small vessels remains uncertain; the latter suggestion
resonance angiography) for the diagnosis of macrovascular causes of is supported by a 2023 MRI-pathological study37, which showed that
ICH requires further investigation. strictly lobar cerebral microbleeds can, in some cases, be caused by
Arteriolosclerosis is the most frequent SVD that causes ICH, being arteriolosclerosis, and cohort studies suggesting that people with a
responsible for >90% of non-lobar ICH cases as well as about half of ‘mixed’ pattern of ICH have a similar risk factor profile to that of ICH
lobar ICH cases on the basis of data from in vivo neuroimaging17,29 and associated with arteriolosclerosis16,17,38. Furthermore, the lobar cerebral
post-mortem histopathology studies30. Several groups have proposed microbleeds in CAA might be cortical rather than juxtacortical39, which,
imaging-based criteria for arteriolosclerosis17,29,31. Histopathological if confirmed, could more accurately determine the relative contribu-
validation is available for most of the individual markers of SVD32, tions of each arteriopathy to lobar cerebral microbleeds. Cortical
but not yet for any set of specific diagnostic criteria for arteriolo- superficial siderosis seems to be highly specific for CAA36, suggest-
sclerosis, perhaps owing to the heterogeneity of histopathological ing the presence of true ‘mixed’ CAA and arteriolosclerosis in patients
findings; this heterogeneity contrasts, for example, with sporadic with lobar and deep cerebral microbleeds, ICH, or both.
CAA, in which the hallmark finding is small vessel wall amyloid-β ICH aetiology cannot always be determined when neither a mac-
deposition33. The use of a standardized neuroimaging definition for rovascular cause nor cerebral SVD is detected despite vascular and
arteriolosclerosis-related ICH, including combinations of lacunes, parenchymal neuroimaging. So-called ‘cryptogenic’ ICH affects up
moderate or severe periventricular white matter hyperintensities to 20% of patients4,16,40. A 2023 systematic review and meta-analysis
and non-lobar (deep) haemorrhages (ICH or cerebral microbleeds), of people with ICH of undetermined aetiology, defined as neither
could improve pathophysiological understanding and comparability structural nor macrovascular causes nor signs of SVD, had a tendency
across studies. towards higher mortality compared with that of people with identified
CAA is responsible for most of the other ~50% of lobar ICH cases30,34 causes of ICH16; however, this finding is likely to be confounded by
and can be diagnosed with high diagnostic accuracy in people with ICH incomplete diagnostic evaluation owing to early death. Conversely,
using acute CT (Edinburgh criteria30) or MRI (Boston 2.0 criteria35). survivors of cryptogenic ICH with standardized and complete evalua-
The Edinburgh criteria use two specific CT-defined neuroimaging tion (including brain MRI) seemed to have a very low rate of recurrent
haematoma characteristics (finger-like projections and subarachnoid ICH16. This finding, together with evidence that an aetiology-based
extension) and the APOE genotype. The Boston criteria for CAA36 use classification system for ICH is associated with a low prevalence of
MRI-defined haemorrhagic lesions, including strictly lobar cerebral cryptogenic ICH16, provides another strong argument for complete
microbleeds, ICH, cortical superficial siderosis and convexity subarach- and structured work-up (including MRI to seek underlying SVD) in
noid haemorrhage. The latest update of the Boston criteria includes most people with ICH to inform prognosis.

a
CAA
Lobar ICH b c d e

Neuroimaging markers of CAA

Arteriolosclerosis f g h i

Deep ICH

Neuroimaging markers of arteriolosclerosis

Fig. 1 | Types of cerebral small vessel diseases that cause intracerebral and white matter hyperintensities in a multispot pattern (panel e). Neuroimaging
haemorrhage. a, Lobar intracerebral haemorrhage (ICH) is most frequently markers of arteriolosclerosis include deep ICH with periventricular white matter
caused by cerebral amyloid angiopathy (CAA) or arteriolosclerosis. Deep ICH is hyperintensities (panel f ), deep cerebral microbleeds in the pons (panel g),
usually due to arteriolosclerosis. b–i, Neuroimaging markers of CAA (top row) deep cerebral microbleeds in the basal ganglia (panel h) and severe enlarged
include cortical superficial siderosis (panel b), lobar cerebral microbleeds perivascular spaces in the basal ganglia (panel i).
(panel c), enlarged perivascular spaces in the centrum semiovale (panel d)

Nature Reviews Neurology

Review article

neuropsychiatric symptoms and extensive deep SVD with white mat-

Box 1 | The DIAGRAM score ter lesions extending into the inferior temporal lobe, but ICH can also
occur, although infrequently50,51. Other monogenic SVDs that can poten-
The DIagnostic AngioGRAphy to find vascular Malformations tially be associated with ICH include cerebral autosomal recessive arte-
(DIAGRAM) study was a prospective, multicentre study that aimed riopathy with subcortical infarcts and leukoencephalopathy (CARASIL;
to assess the yield and accuracy of CT angiography, magnetic caused by mutations in HTRA1 with recessive inheritance), pontine
resonance angiography and digital subtraction catheter angiography autosomal dominant microangiopathy with leukoencephalopathy
in people with non-traumatic intracerebral haemorrhage (ICH) (PADMAL; caused by mutations in COL4A1 or COL4A2 with dominant
aged 18–70 years (n = 298)27. In this study, use of digital subtraction inheritance) and retinal vasculopathy with cerebral leukodystrophy
catheter angiography was high (52%). On the basis of the study (caused by TREX1 mutations)52. Hereditary cerebral haemorrhage
findings, the DIAGRAM score was proposed for assessment of the with amyloidosis is caused by mutations in APP on chromosome 21,
risk of a macrovascular cause in people with ICH. The DIAGRAM which encodes the amyloid precursor protein, or mutations in CST3 on
score consists of three variables: age (stratified into categories chromosome 20, which encodes cystatin C. Whereas Icelandic-type
of 18–50 years and 51–70 years), ICH location (deep, lobar or CAA is found only in some families of Icelandic ancestry, APP mutations
posterior fossa) and small vessel disease (SVD) changes on CT. have been found in Dutch (Dutch-type CAA), Italian and Flemish fami-
The DIAGRAM+ score additionally incorporated CT angiography lies, but also in families in Iowa, USA52. Interestingly, although people
findings. The highest risk of macrovascular causes was found with Down syndrome (characterized by trisomy 21) have a markedly
in people with posterior fossa ICH, younger age, absence of increased risk of Alzheimer disease, which is mainly attributed to the
SVD changes on CT and CT angiography leading to suspicion of presence of three APP genes, their risk of ICH, although elevated com-
macrovascular causes. People with deep ICH, higher age and SVD pared with that of the general population, is lower than that associated
changes on CT had a lower risk of a macrovascular cause. with Dutch-type CAA53.

Less common SVD subtypes

CAA-related inflammation (CAA-ri) is an entity that arises as an inflam-
Genetics and intracerebral haemorrhage matory response to vascular amyloid deposits54. It spans a pathological
Polymorphisms in several genes are associated with an increased risk spectrum from perivascular inflammatory infiltrate to granuloma-
of ICH, although such information is not yet widely used in clinical tous angiitis with vessel wall destruction (so-called amyloid-β-related
practice. One of the best researched is APOE, which encodes apoli- angiitis)55–57. In people with neuroimaging evidence of CAA and typi-
poprotein E — a protein involved in lipid metabolism41. There are cal acute symptoms, such as headache, encephalopathy, seizures or
three APOE polymorphic alleles41: APOE ε2 and APOE ε4 are less com- focal neurological deficits and asymmetric white matter lesions involv-
mon than APOE ε3, but are associated with an increased risk of ICH41. ing the subcortical white matter, probable CAA-ri can be diagnosed
Although APOE ε4 might be associated with CAA42, it remains unclear according to proposed neuroimaging CAA-ri criteria58. In a compari-
whether this also holds true for APOE ε2 (ref. 41). In a post-mortem son of 17 people with autopsy-confirmed CAA-ri and 37 people with
autopsy study, homozygous or heterozygous carriers of APOE ε4 alleles non-inflammatory CAA, these criteria demonstrated high specificity
had a higher vascular and parenchymal amyloid load than those who (97%) for the diagnosis of CAA-ri. However, the modest sensitivity of
did not carry APOE ε4 alleles42. In people with Alzheimer disease treated 82% and the choice of a control group without typical acute symptoms
with amyloid-β antibodies, the APOE ε4 genotype seemed to be the most associated with CAA-ri suggests that CAA-ri might be missed by these
important risk factor for amyloid-related imaging abnormalities43. criteria in practice58. Further research is therefore needed to improve
Other genetic polymorphisms potentially associated with lobar ICH the sensitivity of CAA-ri diagnostic criteria. The risk of CAA-ri recur-
occur in genes that encode glutathione peroxidase 1, complement rence in subsequent years is considerable (around 40% over 2–3 years)
C3b/C4b receptor 1, protein kinase Cη (PRKCH) and the integrin even after clinical and radiological stabilization59,60. Observational data
αV-subunit, and include variants on locus chr12q21.1 (ref. 44). indicate that early immunosuppression is associated with improved
Similarly, certain genetic variants on locus 1q22 are associated with outcomes60. Emerging evidence also suggests a role for systemic or
white matter hyperintensities in healthy populations and in people cerebrovascular inflammation in the pathogenesis of CAA and Alz-
with deep ICH45,46. Altered regulation of neurotransmitter receptors heimer disease; however, the extent to which inflammation has a role
and blood–brain barrier disruption have been suggested as poten- in the development, pattern or prognosis of ICH in people with CAA in
tial mechanisms44,45. Certain polymorphisms in genes that encode general is unknown61.
COL4A1 and COL4A2, matrix metalloproteinases 2 and 9, TNF and tissue In people with early-onset CAA (before 50 years of age) and a
inhibitors of metalloproteinases 1 and 2 might also be associated with history of human cadaveric growth hormone injection, cadaveric
deep ICH44. dural graft implantation or other potential exposures to amyloid-β,
The role of monogenetic conditions in young-onset CAA47 and iatrogenic CAA presumably due to prion-like transmission of amyloid-β
in SVD48 have been reviewed elsewhere. Suspicion of a monogenic should be considered after exclusion of potential concurrent diseases
disease should arise in people under 50 years of age with a positive (for example, genetic forms of CAA)47,62,63. Iatrogenic CAA seems to
family history of SVD-related ischaemic stroke or ICH, diffuse white manifest clinically 30–40 years after the initial exposure, which is
matter hyperintensities and lacunar or subcortical infarcts49. The most consistent with the pathophysiology of other prion diseases, such as
frequent monogenic SVD is cerebral autosomal dominant arteriopathy Creutzfeld–Jacob disease or Kuru disease62. This long latency has rel-
with subcortical infarcts and leukoencephalopathy (CADASIL), which is evance for sporadic CAA, suggesting that amyloid deposition predates
caused by mutations in NOTCH3 (ref. 50). Typically, CADASIL manifests clinical symptoms of ICH by many years, and that amyloid-β fibrils
with migraine, subcortical ischaemic stroke, cognitive impairment, propagate within the brain in a similar way to other prions.

Nature Reviews Neurology

Review article

Fluid biomarkers in the diagnosis and prognosis of SVD used clinical and neuroimaging features to define the ICH subtypes
Several blood biomarkers that are indicative of neuronal damage have arteriolosclerosis, CAA, mixed cerebral SVD, other SVD and secondary
been investigated in people with SVD64. Neurofilament light chain (NfL) causes29. Subtypes were classified according to the diagnostic level of
is part of the axonal cytoskeleton65. Plasma levels of NfL correlate with certainty, allowing for subtypes to coexist29. CLAS-ICH encourages
the SVD burden, with the presence of SVD markers on MRI, with cogni- the use of MRI in people with ICH, but also permits the use of CT29. The
tive decline in people with sporadic SVD or CADASIL65–67 and with ICH CADMUS classification (Box 2) is an MRI-based classification for ICH
recurrence in people with CAA68. associated with sporadic SVD17. CADMUS was applied in two independ-
Glial fibrillary acidic protein (GFAP) is part of the astrocytic ent cohorts of 1,180 and 313 people with ICH with good inter-rater
cytoskeleton and can be detected in blood plasma in people with astro- and intra-rater reliability. The major limitation for all classifications
cytic damage69. Plasma GFAP values were higher in people with ICH to date is the lack of histopathological validation of the underlying
than in people with ischaemic stroke in the first few hours following SVD. We believe that the classification of ICH aetiology is important,
symptom onset70,71. Although these studies suggest differences in the currently under-used and might improve management in the future,
pathophysiology of ischaemic stroke and ICH, such blood biomarkers comparable with the use of classifications in ischaemic stroke. Future
are currently of limited clinical use, as neuroimaging is a faster and research is needed to show whether CLAS-ICH or CADMUS provide
more accurate method for disease diagnosis, and because there are cur- clinical benefit.
rently no approved bedside tests or reliable thresholds for GFAP-based
diagnosis in people with a suspected stroke. Both NfL and GFAP values Advances and uncertainties in acute therapy
are also associated with incident ICH and ischaemic stroke in people Therapeutic targets
with sporadic and genetic SVD68,72. ICH results from the rupture of a blood vessel and bleeding into the
The role of cerebrospinal fluid (CSF) examination in diagnosing brain parenchyma. Although ICH is frequently dichotomized as trau-
the cause of ICH remains uncertain, although some evidence suggests matic versus non-traumatic (spontaneous) ICH, and as ‘primary’ versus
that specific patterns of amyloid-β and tau proteins can help detect ‘secondary to macrovascular causes’, an increasing body of evidence
CAA. CSF levels of amyloid-β42 and amyloid-β40 are lower in people indicates that non-traumatic ICH should neither be termed ‘spontane-
with CAA than in healthy controls or people with Alzheimer disease73, ous’, owing to the increasing recognition of trigger factors, nor ‘pri-
whereas CSF levels of total tau are mildly increased74. We recommend mary’, owing to the fact that the term is not well defined, and might
obtaining CSF when the imaging diagnosis is uncertain and when discourage further investigation, classification or the implementa-
confirming the diagnosis of CAA could affect clinical decisions, such tion of rational secondary prevention measures. Beyond underlying
as whether to give or withhold antithrombotic treatment. vasculopathies, as discussed above, potential trigger factors for vessel
rupture in people with ICH can be summarized into two main groups:
Current classifications for intracerebral short-term surges in blood pressure82,83 and infections82,84. Specifically,
haemorrhage aetiology rapid rises in blood pressure are likely to cause ICH by challenging
Several classification systems exist for non-traumatic ICH. Although vessel wall resistance in fragile, small vessels affected by SVD, whereas
stratification according to lobar or non-lobar location is common75,76, infection is thought to cause ICH through the induction of endothelial
this information alone is insufficient to provide clear insights into dysfunction and coagulopathy.
disease aetiology, as arteriolosclerosis can cause both deep and lobar At the time of vessel rupture, the ICH causes direct damage to the
ICH, and as CAA and arteriolosclerosis often co-occur30,38. Two mecha- impacted structures, resulting in the so-called primary or mechani-
nistic and risk factor-based classifications, SMASH-U77 and H-ATOMIC78, cal injury (Fig. 2). In some people, continuous bleeding during the
were proposed some years ago. SMASH-U attributes ICH to a single first few hours after ICH onset results in haematoma growth, which
aetiology based on a hierarchical algorithm77, with demonstrated dif-
ferences in 3-month outcomes according to the assigned cause4,77. The
H-ATOMIC classification incorporates multiple contributing factors78, Box 2 | CLAS-ICH and CADMUS
consistent with the high prevalence of multiple vascular risk factors
in ICH populations4. CLAS-ICH29 is a comprehensive classification that includes five
A comparative analysis found that inter-rater reliability was good defined intracerebral haemorrhage (ICH) subtypes: arteriosclerosis,
for both classifications (SMASH-U, κ = 0.82; H- ATOMIC, κ = 0.76)79. cerebral amyloid angiopathy (CAA), mixed small vessel disease
Nevertheless, these aetiological classifications have consider- (SVD), other rare forms of SVD (that is, genetic SVD) and secondary
able limitations: first, they mix causes (for instance, CAA or mac- causes (that is, macrovascular, tumour or rare cause). Patients
rovascular lesions) and risk factors (such as hypertension and in each category are scored according to the level of diagnostic
anticoagulation)80,81. Second, the classifications make assumptions evidence (that is, well defined, possible or no evidence).
about disease causality; for example, deep ICH location is part of the CADMUS17 is an MRI-based classification for SVD-related ICH
diagnostic criteria for hypertension-associated ICH, but it can also be that involves three steps. First, a secondary cause should be
caused by macrovascular lesions77,78. excluded. Second, the primary ICH location is scored (deep, lobar or
In the past 2 years, attempts have been made to classify ICH undetermined/cerebellar). Last, haemorrhagic and non-haemorrhagic
according to direct visualization and categorization of the underly- SVD markers on MRI are scored. Information on the ICH location and
ing presence and patterns of SVD biomarkers on neuroimaging; such the presence or absence of MRI-defined SVD markers is combined
a classification should preferably be based on findings from MRI32, but to classify patients into one of four mutually exclusive categories:
such an approach would be limited in that it cannot be used for people CAA, deep perforator arteriolopathy, mixed CAA (also termed deep
who cannot undergo MRI, for example, due to contraindications, perforator arteriopathy) and undetermined SVD.
early death or limited resources. The CLAS-ICH classification (Box 2)

Nature Reviews Neurology

Review article

Minimize haematoma
expansion with blood
pressure lowering and
haemostatic therapies

Neurosurgical
evacuation Neurosurgical
decompression

Primary
treatment

Anti-inflammatory Chelating or
agents antioxidant agents
Secondary
treatment

Mass effect,
tissue damage Inflammatory
response, oedema Toxicity of blood breakdown
products, oxidative stress

Minutes Hours Days Weeks Months

Fig. 2 | Treatment targets and brain damage after intracerebral the primary brain damage. Neurosurgical evacuation can also reduce the
haemorrhage. Primary treatment in the hyperacute phase of intracerebral haematoma volume. Secondary brain damage is mediated by the inflammatory
haemorrhage (ICH) involves the prevention of secondary haematoma expansion response, oedema, toxicity of blood breakdown products and oxidative stress.
by means of blood pressure lowering and use of haemostatic therapies for Several therapeutic options including anti-inflammatory treatments and
patients on therapeutic anticoagulation. Maintaining the haematoma volume decompressive hemicraniectomy might potentially mitigate these effects, but
as small as possible by preventing haematoma expansion aims to reduce evidence of their efficacy requires validation in randomized controlled trials.

probably occurs in the pre-hospital phase as well as in 20–30% of all underlying SVD influences the dynamics of haematoma expansion,
patients admitted to hospital85. Haematoma expansion is usually with more prolonged and continuous bleeding in people with CAA93.
captured on repeated imaging85,86 (Fig. 3), and might involve growth Preventing haematoma expansion is a key treatment goal of acute ICH
of the main intraparenchymal blood clot itself or new intraventricu- management85,94 (Fig. 2).
lar extensions87. The main predictors of haematoma expansion are Depending on the size and location of the ICH, intracranial pres-
time since onset, with the highest risk occurring in the first 3–6 h sure may increase. This pressure might further increase as a conse-
after symptom onset; baseline haematoma volume, with the high- quence of hydrocephalus if the circulation of the CSF is obstructed,
est risk associated with medium to large-sized haematomas; and often, but not necessarily, owing to intraventricular haemorrhage
prior anticoagulation, which is associated with a threefold increased (IVH). After the ‘primary’ injury of haematoma expansion and direct
risk of haematoma expansion86,88. Time from ICH onset and base- tissue injury, the coagulation cascade is initiated95 together with a
line haematoma volume can be combined to calculate the so-called pro-inflammatory response in the perihaematomal region96. The neu-
‘pre-scan haematoma growth rate’, which is associated with haema- rotoxic and pro-inflammatory effects of the extravascular blood are
toma expansion and poor outcomes89,90. Indeed, haematoma expan- crucially different to the pathophysiologic processes that underlie
sion is consistently associated with poor functional outcomes in ischaemic stroke. Components of erythrocytes that have been impli-
people with ICH91 regardless of the definition used (that is, absolute cated in the pro-inflammatory response include haemoglobin (and its
or relative growth). Of note, the studies that support these findings breakdown products), iron, peroxiredoxin 2 and carbonic anhydrase 1
included very few people on oral anticoagulants, in whom the risk (ref. 95). Data from animal models, supported by observations in peo-
of haematoma expansion is increased much further. The process of ple with ICH, indicate that the pro-inflammatory cascade is activated
haematoma growth has been likened to an ‘avalanche’, where bleed- within hours of haemorrhage and involves the activation of resident
ing starts from one primary vessel, which, as the haematoma grows, microglia and astrocytes, and the infiltration of leucocytes, which
causes shearing of surrounding ‘secondary’ vessels, thereby ampli- propagates neural cell death through processes that involve oxida-
fying local bleeding92. Findings from a 2022 study suggest that the tive stress, and the activation of matrix metalloproteinases and other

Nature Reviews Neurology

Review article

proteases, complement proteins and cytokines96. This early response Pre-hospital triage and ultra-acute management of people
is thought to be injurious and abundant, and to evolve over several with suspected ICH are a further challenge. Neither clinical presen-
days following the ICH96. Over time, possibly after day 3 (ref. 97), the tation nor blood-based biomarker tests reliably differentiate ICH
neuroinflammatory response is thought to change into a ‘reparative from ischaemic stroke69,106. Mobile stroke units with CT scanners
phase’ and become potentially beneficial96. Perihaematomal oedema — enable pre-hospital diagnosis of ICH107,108, which might in turn improve
which is regarded as a radiological marker of secondary injury, pre-hospital management109. The 2024 INTERACT4 trial, which investi-
including inflammation — typically peaks at day 7–10 after ICH, and gated the effect of pre-hospital blood pressure reduction in people with
is independently associated with poor functional outcomes98. Many a suspected stroke, found a significant beneficial effect on functional
different factors have a role in the inflammatory response to ICH and outcome in people with the final diagnosis of ICH110.
knowledge of the exact sequence and consequences of the neuroin- The management of people with ICH is interdisciplinary and might
flammatory process remains limited. These limitations hinder our reasonably include specialists from emergency medicine, neurology,
ability to understand how detrimental neuroinflammation can be neuroradiology, neurosurgery and intensive care. However, not all
reduced without compromising the beneficial reparative aspects of patients require all levels of care. Stroke unit care improves outcomes
this process. A 2022 systematic review of molecular markers of inflam- after ICH to a similar extent as it does for ischaemic stroke111. A 2024
mation in human brain tissue identified IL-1β as the only marker that study found that hospital type (that is, primary stroke unit without neu-
was repeatedly upregulated after ICH99. A small, randomized trial of a rosurgery versus comprehensive stroke centre with neurosurgery) did
recombinant human IL-1 receptor antagonist showed no harm but also not influence outcomes after ICH, and neither did secondary transfer
no evidence of an effect on the primary outcome of perihaematomal from a primary to a comprehensive stroke centre112. Indeed, available
oedema, although it was not powered for the latter100. A phase II trial data suggest that general treatment measures can be delivered even in
of the same therapeutic is underway101. small, primary stroke units whereas a robust referral pathway should
Accumulating evidence also suggests that, similar to ischaemic be implemented for patients in need of neurosurgery113.
stroke, ICH might benefit from early intervention, both in terms of
counteracting the primary injury and when directed at reducing the Medical treatment for intracerebral haemorrhage
secondary injury102,103. However, the optimal timing of the majority of An important goal of acute medical treatment is the prevention of
potential treatments remains to be determined. haematoma expansion85,94. Limiting secondary brain injury by targeting
oedema formation and inflammation is an additional valid goal100,101
The care bundle approach and levels of acute care (Fig. 2; see Supplementary Table 1).
The management of people with ICH includes different potential treat- Up to 70% of people with ICH have elevated blood pressure at
ments targeting a range of mechanisms94. Consequently, combining presentation, presumably due to a combination of pre-existing hyper-
different interventions, including medical and surgical management, tension and the acute response to the ICH114. As hypertension is associ-
under one ‘care bundle’ umbrella102 has been increasingly studied104,105. ated with haematoma expansion, perihaematomal oedema and poor
In the INTERACT3 trial published in 2023, a care bundle including early outcomes114, high blood pressure is a plausible treatment target.
intensive blood pressure lowering, glycaemic control, treatment of Two large, phase III randomized controlled trials investigated the
pyrexia and reversal of anticoagulation improved outcome after ICH105. influence of early intensive (INTERACT2) and very intensive (ATACH-II)
Based on this positive randomized controlled trial and on prior obser- treatment of elevated blood pressure in people with ICH115,116 (see Sup-
vational data104, a care bundle approach or ‘code ICH’ for hyperacute plementary Table 1). Although neither trial demonstrated a statistically
treatment of people with ICH is recommended and promoted by expert significant beneficial effect of antihypertensive treatment on func-
panels100,101. tional outcome, they nevertheless provided important insights into the

a b Fig. 3 | Case example of secondary haematoma

expansion. A patient on oral anticoagulation
therapy with a factor Xa inhibitor (last intake
4 h before) was admitted 45 min after onset of
intracerebral haemorrhage (ICH) symptoms.
a, Baseline CT imaging with a small parenchymal
haemorrhage in the left thalamus (1 ml). b, Follow-up
imaging showed significant haematoma expansion
(36.8 ml) and new intraventricular haemorrhage
(IVH; 12.5 ml).

Baseline CT Follow-up CT
ICH volume: 1 ml ICH volume: 36.8 ml
No IVH IVH volume: 12.5 ml

Nature Reviews Neurology

Review article

effects of acute blood pressure control in people with ICH. Specifically, thrombotic complications in the trial overall was high (18%). Moreo-
the trial findings suggested that treatment to a target of <140 mmHg ver, many people with FFP received PCC as rescue therapy during
within 1 h of ICH is safe and reduces haematoma expansion115. More the trial, making it difficult to judge the individual risk of thrombo-
intensive blood pressure control to a target of 110–139 mmHg caused embolic complications associated with each treatment. Evidence
significantly more renal adverse events, without any benefit on func- from observational studies suggests that early and aggressive reversal
tional outcome116. The pooled analysis of individual participant data of anticoagulation is associated with better outcomes than are slower
from these studies showed that achieving early and stable systolic approaches132,133. ICH associated with the use of direct oral antico-
blood pressure control is associated with a 10% increase in the odds agulants can be reversed with PCC or target-specific agents such as
of a favourable functional outcome for every 10 mmHg decrease in andexanet alfa for the oral factor Xa inhibitors134 and idarucizumab
blood pressure117. In a systematic review of 6,221 people from 16 stud- for oral direct thrombin inhibitors135. Although no randomized con-
ies of blood pressure reduction in ICH, subgroup analyses suggested trolled trials have assessed the safety and efficacy of idarucizumab in
that blood pressure interventions are most effective when applied people with ICH, the randomized, controlled ANNEXA-I trial compared
within the first few hours after the onset of symptoms and if titrated to andexanet alfa against usual care (mainly PCC)136,137, demonstrating a
a target blood pressure using α-adrenoreceptor or β-adrenoreceptor significant benefit of andexanet alfa in haemostatic efficacy, driven by
blockers118. However, neither intravenous magnesium sulfate nor a lower rate of secondary haematoma expansion, but higher rates of
transdermal glyceryl trinitrate administered in an ambulance prior thromboembolic events137. The trial found no difference in mortality
to hospital admission demonstrated beneficial effects in people with or functional outcome between the two treatments, but was not suf-
ICH119–121. By contrast, and as mentioned above, the INTERACT4 trial ficiently powered to assess this secondary outcome and the functional
reported that pre-hospital hyperacute blood pressure lowering to outcome was assessed early (30 days). A small randomized controlled
<140 mmHg within 30 min of randomization by administration of trial that assessed the antifibrinolytic drug tranexamic acid in patients
urapidil reduced the odds of a poor outcome in people with a final with direct oral anticoagulant-associated ICH on top of standard of care
diagnosis of ICH (OR 0.75; 95% CI 0.60–0.92)110. The 2022 guidelines (PCC) also failed to show any benefit138. Indirect comparison between
of the American Heart Association/American Stroke Association and people enrolled in this small trial using non-specific treatments (that
the 2020 Canadian Stroke Best Practices Recommendations advise is, PCC and/or tranexamic acid) and people enrolled in ANNEXA-4
lowering of systolic blood pressure in people with ICH within 6 h of (ref. 139), a single-arm study in which all participants received andex-
symptom onset to a target below 140 mmHg within 1 h of treatment anet alfa, found a significant reduction in the rate of haematoma
initiation, with the goal of maintaining systolic blood pressure in the expansion in patients receiving andexanet alfa and no difference in
range of 130–150 mmHg122,123, although it must be noted that these the rate of thromboembolic events140.
guidelines were published before the publication of the most recent In ICH unrelated to prior anticoagulation, several haemostatic
trials, including INTERACT4. However, a knowledge gap remains for agents have been investigated in different doses and time windows, in
specific patient groups not included in the trials, including those with subgroups with a positive ‘spot sign’141 and within the setting of a mobile
large (>30 ml) haematoma volumes, those who present >6 h after ICH stroke unit. Administration of tranexamic acid within 8 h of symptom
onset, those with systolic blood pressure >220 mmHg and those who onset had no significant effect on functional outcome at 90 days but
will undergo surgery. Whether management should differ in ICH sub- was associated with a reduction in haematoma expansion and early
groups according to the presumed underlying type of SVD93, whether death142–144. A small, phase II randomized controlled trial found that
there is a preferred route of drug administration and the rate at which tranexamic acid administered in the setting of a mobile stroke unit
blood pressure can or should be lowered are also unclear. had no effect on the rate of haematoma expansion or any other clini-
People with ICH associated with the use of antithrombotic medi- cal outcome145. Recombinant activated factor VIIa (rFVIIa) accelerates
cations (antiplatelet agents or anticoagulants) have larger baseline coagulation by directly activating factor X on the surface of activated
haematoma volumes, more haematoma expansion and higher case platelets. Although most trials with rFVIIa have not shown an effect
fatality88,124,125. Current guidelines advise that antithrombotic medi- on functional outcome at 90 days, one trial showed an effect on hae-
cations are stopped upon hospital admission126. In people with ICH matoma expansion146–148 (see Supplementary Table 1). Ongoing trials
who are using antiplatelet therapies, platelet transfusion is associ- are investigating rFVII and tranexamic acid in specific subgroups and
ated with a worse functional outcome and should be avoided127. It is earlier (<2–4.5 h) treatment windows (TICH-3 (ref. 149), Indian Trial of
unclear whether platelet transfusion should be administered to people Tranexamic Acid150, TRANSACT151, Kathmandu tranexamic acid trial152
on antiplatelet therapy for whom surgical evacuation of the haema- and FASTEST153). The results of available trials have been summarized
toma is planned. These uncertainties have led to variations in clinical in a Cochrane Review154.
practice128. Of note, a 2023 phase II trial indicated that desmopressin Numerous studies published in the past few years have tar-
might reduce bleeding in some people with ICH on antiplatelet drugs129, geted secondary brain injury, including oedema formation
justifying testing of this agent in a larger randomized trial129. and inflammation, as an approach to improve functional out-
Approximately 20% of all ICH cases are associated with prior comes after ICH. Although used in many places worldwide, man-
oral anticoagulant use, with a changing landscape from the use of nitol and corticosteroids have no proven benefit on functional
vitamin K antagonists to direct oral anticoagulants (mainly direct outcome and might cause harm155–157; randomized trials are there-
factor Xa inhibitors) during the past few years130. Reversal of vita- fore required to assess the benefits of these treatments. Two
min K antagonist-associated ICH was studied in the small, randomized small trials found the matrix metalloproteinase 9 inhibitor, mino-
controlled INCH trial131, which found that prothrombin complex con- cycline, to be safe; however, any effect on functional outcome
centrate (PCC) was superior to fresh frozen plasma (FFP) in achiev- remains to be proven158,159. Fingolimod, a sphingosine 1-phosphate
ing an international normalized ratio <1.3. PCC use was also found receptor modulator, was safe and reduced perihaematomal oedema in
to have a lower rate of haematoma expansion; however, the rate of a study of 23 people with ICH160. By contrast, a phase II trial of the iron

Nature Reviews Neurology

Review article

chelator deferoxamine mesylate did not improve clinical outcomes fewer patients with a modified Rankin scale score of 5 or 6, providing
within 90 days of ICH compared with placebo161. weak evidence (adjusted risk ratio 0.77; 95% CI 0.59–1.01; P = 0.057)
Ongoing trials are investigating the effects of fingolimod (FITCH162; that decompressive craniectomy might be superior to best medical
single dose, <24 h), anakinra (ACTION101; <8 h for 3 days), atorvastatin treatment alone. However, the intervention seemed to result in more
(STATIC163; <24 h for 7 days), edaravone dexborneol (ED-ICH162; <48 h survivors with a modified Rankin scale score of 4 (moderate severe
for 14 days) and colchicine (CoVasc-ICH164; <48 h), and others are in the disability), and no change in the number of people with a modified
planning stages (dimethyl fumarate165; <72 h for 3 days), sodium aesci- Rankin scale score of 0–3 (symptom-free to moderately disabled)
nate (REACH166; <24 h for 10 days), Chinese herbal medicine FYTF-919 who were therefore able to return home. Outcomes for decompres-
(CHAIN167; <48 h), mirabegron168 (<24 h for 14 days) and celecoxib169 sive craniectomy are worse for people with ICH than for people with
(<6 h for 21 days). ischaemic stroke183. The decision of whether to perform decompressive
Early seizures (<7 days) are frequent (3–32%) in people with ICH170. craniectomy in clinical practice should be strongly informed by the val-
Treatment with anti-epileptic medication should be considered after ues and ethical beliefs, such as prioritization of survival or autonomy,
one or more seizures122,123. Prophylactic antiseizure medication might of individuals being offered the treatment184.
be effective in preventing acute seizures, but has not been shown to Surgery for cerebellar haematomas is generally considered in
improve functional outcome so is not recommended171,172. patients with a decreased level of consciousness, brainstem symp-
toms or when the diameter of the haematoma is ≥3 cm. However, this
Surgical treatment of ICH approach has not been assessed in randomized controlled trials and
As ICH volume is an important predictor of functional outcome, whether surgery with minimally invasive techniques could improve
surgical evacuation could theoretically improve outcome by reduc- outcomes is unclear185.
ing mass effect, lowering intracranial pressure and reducing sec-
ondary injuries. However, high-quality randomized clinical trials of Advances in prevention measures
craniotomy with evacuation of the haematoma, and of minimally Recurrent intracerebral haemorrhage
invasive surgery with thrombolysis, have not demonstrated positive Assessment of ICH recurrence risk is of high priority for the communica-
outcomes in terms of functional recovery, although these interventions tion of prognosis to patients and relatives, as well as for physicians who
have been associated with reduced mortality, which has been assessed need to make decisions about optimal secondary prevention strate-
as a secondary outcome173 (see Supplementary Table 2). As a result, gies. People with ICH frequently have conditions associated with a risk
surgical evacuation of ICH is typically performed in a minority of often of future ischaemic events for which antiplatelet or anticoagulation
relatively young patients as a life-saving measure. However, an accumu- therapy is usually recommended. Whether to start, resume or discon-
lating body of evidence suggests that surgery with minimally invasive tinue these medications after ICH requires an assessment and balance
techniques, with or without endoscopy guidance, may in fact improve of the risks of ICH recurrence versus thromboembolic events. Lobar
functional outcome173. Outcomes associated with minimally invasive location of ICH is clearly and consistently associated with higher rates
surgery might be better with lower post-treatment volumes174,175 and of recurrence compared with non-lobar ICH186–188. However, this higher
earlier treatment might have an advantage over later treatment173. risk is probably related to the greater association of lobar ICH with CAA
However, ultra-early haematoma evacuation by means of craniotomy compared with non-lobar forms of ICH, suggesting that consideration
(within 4 h of symptom onset) was associated with re-bleeding in one of anatomical location may not predict recurrence risk as effectively
study176, whereas in others early endoscopy-guided surgery seemed to as consideration of the underlying causal arteriopathy18. CAA is the
be safe177,178. A 2021 study reported a 5% decrease in the odds of achieving aetiology associated with the highest risk of ICH recurrence, with an
good functional outcome at 6 months for each additional hour before estimated recurrence risk of about 7–10% per year18. The Boston version
evacuation179. Of note, assessment of outcomes at 90 days might be 1.5 criteria for probable CAA seem to be more strongly associated with
too early to capture the full potential for recovery after ICH180. The higher ICH recurrence risk than the more recently proposed version
adaptive, randomized ENRICH trial demonstrated that ICH evacuation 2.0 criteria189.
with minimally invasive trans-sulcal parafascicular surgery improved In addition to classifying the presumed underlying causal arterio-
functional outcomes at 180 days compared with guideline-based medi- pathy, individual MRI markers of SVD190 might also have an important
cal management — an effect that was mainly driven by outcomes in role in the assessment of ICH recurrence risk. Perhaps most promi-
people with lobar ICH181. Although these results are promising, many nently, cortical superficial siderosis is consistently associated with the
questions remain, including the generalizability of the results (given most markedly increased risk of recurrent ICH in people with CAA, even
that only 2.6% of the 11,603 people who were screened met the inclu- after adjusting for the presence and burden of cerebral microbleeds.
sion criteria in the ENRICH trial), the preferred surgical technique and Nevertheless, larger numbers of cerebral microbleeds also seem to
the optimal timing of surgical intervention. The results of ongoing increase the risk of recurrence in people diagnosed with CAA, arteriolo-
randomized controlled trials of different minimally invasive techniques sclerosis or both18. Small DWI-positive lesions on MRI in the acute stage
with variable inclusion criteria and in different time windows should of the initial ICH might also be associated with a higher ICH recurrence
shed further light on these questions (see Supplementary Table 2). risk, although the data on this association are overall unclear191,192.
The Swiss Trial of Decompressive Craniectomy versus Best Medical
Treatment of Spontaneous Supratentorial Intracerebral Hemorrhage Recurrent ischaemic stroke
(SWITCH), which aimed to assess whether decompressive craniectomy The annual risk of ischaemic stroke after ICH is estimated to be
could improve outcomes for people with deep ICH182, was stopped early 1–4%3,186,187,193, similar to the annual risk of recurrent ICH. Some studies
owing to lack of funding. During the 9 years in which the trial ran, 201 have reported that the risk of ischaemic stroke is lower after lobar ICH
of the target 300 people were randomly assigned to treatment, with than following non-lobar ICH186,187 — a difference that might be explained
available data demonstrating that decompressive craniectomy led to by the lower risk of ischaemic stroke associated with CAA compared

Nature Reviews Neurology

Review article

with other causes of ICH19. In people with arteriosclerosis-related Observational studies of anticoagulant therapy reinitiated after
ICH, the risk of ischaemic stroke seems to be as high as in those with ICH in people with atrial fibrillation have demonstrated a reduc-
arteriosclerosis-related lacunar stroke19. Atrial fibrillation constitutes tion in the risk of ischaemic stroke, without an increased risk of ICH
one of the most important risk factors for recurrent ischaemic stroke recurrence132,207. However, treatment biases and confounding might
in people with ICH187. have influenced these studies. More recent results from two small
randomized controlled trials (SoSTART and APACHE-AF) indicate that
Long-term blood pressure control anticoagulant therapy in people with atrial fibrillation and a history
As arterial hypertension is the most important risk factor for ICH194–196, of ICH increased the risk of recurrent ICH but tended to lower the rates of
effective blood pressure control is a priority in secondary prevention recurrent ischaemic stroke; however, the study samples were too small
measures for recurrent ICH. Several trials that have included people to draw clear conclusions about safety or efficacy208,209. An individual
with stroke have shown decreased risk of ICH associated with long-term patient data analysis (COCROACH210) found a significant reduction
intensive blood pressure control197,198. A meta-analysis showed that in the rate of thromboembolic events in people with atrial fibrilla-
intensive blood pressure control benefitted people with ICH, associ- tion using oral anticoagulants (mainly direct oral anticoagulants)
ated with a greater reduction in the risk of recurrent vascular events after ICH compared with those not using anticoagulants, along with a
than that achieved in people with ischaemic stroke199; however, tri- non-significant increase in major haemorrhagic events. However, the
als of long-term blood pressure lowering have not been performed study was not powered to provide a definite answer on the safety and
specifically in people with ICH. Long-term control of blood pressure efficacy of this intervention. Ongoing trials will provide crucial insights
is highly relevant in people with arteriolosclerosis and deep ICH as in the coming years (see Supplementary Table 3).
well as in those with lobar ICH195,200, although specific blood pressure Certain subgroups of people with ICH (those at low risk of recur-
targets have not yet been determined. A large cohort study published rent ICH and at high risk of cardioembolism) might be thought to ben-
in 2023 showed that a systolic blood pressure target of <120 mmHg efit from anticoagulation, whereas those at very high risk of recurrent
decreased the risk of recurrent ICH and other vascular events in people ICH, such as people with severe CAA (manifested as disseminated corti-
with prior ICH200. cal superficial siderosis, previous multiple recurrences or both) might
The ongoing multinational, randomized TRIDENT study will be harmed. The data safety and monitoring committee of ENRICH-AF
investigate the effect of a triple-combination antihypertensive blood recommended that the enrolment of people with lobar ICH and convex-
pressure pill (comprising telmisartan, amlodipine and indapamide) in ity subarachnoid haemorrhage — both considered to be mainly caused
addition to standard care to reduce recurrent stroke after ICH201. The by CAA — was stopped owing to the unexpected high rate of recurrent
PROHIBIT-ICH study202 compared a strategy of intensive blood pres- ICH in participants assigned to receive edoxaban211. However, no such
sure management (target <120/80 mmHg) guided by telemetric home observation was made in the smaller completed trials, and some other
monitoring with standard care to assess the feasibility of this approach ongoing trials continue to enrol this patient group. In the absence of
and its influence on MRI outcome measures including white matter definite evidence from randomized controlled trials, we suggest a
hyperintensity volume; the trial has been completed and publication potential approach to these decisions (Fig. 4).
of the results is awaited.
Utility of left atrial appendage occlusion
Antiplatelet and anticoagulation therapy after Approximately 80% of all thrombi in the left atrium form in the left
intracerebral haemorrhage atrial appendage212. Therefore, surgical or percutaneous occlusion
The safety of antithrombotic therapy has long been questioned and the (left atrial appendage occlusion (LAAO)) is an appealing option for
initiation of such therapies has traditionally been avoided after ICH people with ICH and concomitant atrial fibrillation as it does not require
out of concern for increasing the risk of recurrent ICH. The RESTART203 long-term anticoagulation. Evidence of the efficacy and safety of this
open-label pilot randomized controlled trial found no signal of harm approach is derived from three randomized controlled trials in people
if antiplatelet agents were restarted in people who had previously with high bleeding risk, which compared LAAO with warfarin (in two
taken antithrombotic therapy for the prevention of occlusive vas- studies) or direct oral anticoagulants (in one study)213–215. However,
cular disease, a median of 76 days from the onset of ICH symptoms. randomized data of people with ICH and atrial fibrillation are lacking.
Moreover, people who reinitiated antiplatelet therapy experienced Four small cohort studies (two in people with CAA and two in people
numerically fewer ICH recurrences than those who did not restart with unselected ICH) reported that LAAO had a good safety profile216,217.
therapy (adjusted hazard ratio 0.51; 95% CI 0.25–1.03; P = 0.06). The Relevant ongoing trials include the A3ICH study218, which aims to com-
results were consistent across subgroups including those stratified pare anticoagulation, LAAO or no such therapy in people with ICH, and
by ICH location, time since symptom onset, type of antiplatelet drug, the STROKE-CLOSE study219, which aims to compare LAAO with medi-
participant age and history of atrial fibrillation. Extended follow-up for cal therapy in people with ICH. Further studies include ASAP-TOO220,
up to 7 years found no significant effect of antiplatelet therapy on ICH CLOSURE-AF165 and COMPARE-LAAO221, which aim to investigate
recurrence or vascular events204. Moreover, a brain imaging sub-study LAAO in people with atrial fibrillation and contraindications to
of the RESTART trial (n = 254 with MRI)205 found that initiation of anti- anticoagulation or at high risk of haemorrhage.
platelet therapy did not lead to an increase of recurrent ICH in people
with high-risk features, such as cerebral microbleeds or fulfilling the Statins and other lipid-lowering drugs
diagnostic criteria for CAA. These findings reassure physicians about Dyslipidaemia is an important risk factor for vascular disease and pre-
the safety of using antiplatelet drugs in people who have had an ICH, sent in a considerable proportion of people with ICH222,223. However, the
although whether this strategy is beneficial remains unclear. Further use of statins in people with ICH has been controversial. Early studies
randomized controlled trials aim to address this question, including of statins in people with prior stroke showed a decreased risk of vas-
ASPIRING206. cular events; however, subgroup analyses indicated an increased risk

Nature Reviews Neurology

Review article

Individual with ICH and atrial Fig. 4 | Algorithm for deciding whether to initiate
fibrillation or resume oral anticoagulation and other
treatment options in people with intracerebral
haemorrhage and atrial fibrillation. A notable
number of people with intracerebral haemorrhage
(ICH) have concomitant atrial fibrillation. Key
1. Intensive risk factor factors in the decision to initiate or resume oral
management based on ICH anticoagulation in this group include the risk of ICH
aetiology (particularly with
regard to blood pressure control, recurrence on the basis of ICH aetiology, the severity
alcohol use and concomitant of underlying cerebral small vessel disease (SVD),
antithrombotic medication) the risk of cardioembolism, and factors and wishes
that are important to each individual. CAA, cerebral
2. Assess cause of ICH (MRI and
at least non-invasive amyloid angiopathy. aAnticoagulation with a direct
angiography) and risk of oral anticoagulant (DOAC) is recommended owing to
recurrent ICH the lower risk of intracranial haemorrhage for these
agents compared with vitamin K antagonists.

Low recurrence risk (cryptogenic Moderate recurrence risk (CAA High recurrence risk (CAA with
ICH, arteriolosclerosis without without cortical superficial disseminated cortical superficial
lobar microbleeds) siderosis, mixed SVD phenotype siderosis, recurrent ICH)
with high number of lobar
microbleeds)

Reinitiate anticoagulationa in the Individual risk/benefit Do not initiate anticoagulation

post-acute phase assessment with the patient, outside the context of a
especially considering: randomized controlled trial.
• Severity of cSVD (number of Evaluate left atrial appendage
cerebral microbleeds) occlusion considering:
• Individual thromboembolic risk • Procedural risk
• Risk factor management • Level of general health, frailty
• Comorbidities
Treatment options include no
anticoagulation, left atrial
appendage occlusion and
anticoagulation

of haemorrhagic stroke in survivors of ICH224,225. Statins may inhibit Disease-modifying therapies in cerebral small vessel diseases
platelet function, have pro-fibrinolytic capacities and decrease the Despite the massive impact of arteriolosclerosis and CAA, not limited
expression of tissue factor, thereby attenuating the generation of to a cause of ICH but also their pathogenic role in dementia and other
thrombin226–228. Furthermore, observational studies identified a higher non-motor symptoms, no disease-specific treatment options exist in
risk of ICH in people with low levels of LDL and total cholesterol229–233. clinical practice beyond risk factor management. As discussed above,
Conversely, a large meta-analysis of statin trials showed no increased intensive blood pressure management is currently the primary long-term
risk of recurrent ICH in people with previous ICH who were taking treatment option to slow the progression of SVD, including arterioloscle-
statins234. A large regional case–control registry study indicated that rosis. A subgroup of participants from the SPRINT trial who underwent
statin use was actually associated with a lower risk of ICH235. Although MRI reported a slightly smaller increase of white matter hyperinten-
one randomized controlled trial showed no difference in the risk of sities in those who were randomly assigned to the intensive blood
ICH associated with different LDL target levels in people with ischae- pressure-control group (target <120 mmHg) than those assigned to
mic stroke236, no prospective randomized controlled trials have been standard blood pressure control239. However, the PRESERVE trial, which
performed in people with ICH. The ongoing Statins in Intracerebral specifically targeted people with arteriolosclerosis, did not show such an
Hemorrhage (SATURN237) trial, which will randomly assign people with effect58. Two meta-analyses of randomized trials that assessed the effects
ICH and prior statin therapy to continuation or discontinuation of of antihypertensive medications on markers of cerebral SVD showed that
therapy, will likely shed more light on this relevant issue. the progression of white matter hyperintensities was reduced in people
No association has been found between PCSK9 inhibitor use and under (intensive) antihypertensive treatment compared with that of
long-term risk of intracranial haemorrhage238, indicating the potential people who received standard blood pressure control239,240.
utility of these agents in people with a history of ICH. However, this The LACI-2 trial showed that isosorbide mononitrate or combina-
observation is again limited by a lack of data from people with ICH, tion treatment with isosorbide mononitrate and cilostazol — which
and even more so by the exclusion of those people from large trials of have both demonstrated beneficial effects on small vessel endothelium
PCSK9 inhibitors, limiting our ability to draw any clear conclusions in experimental studies — decreased the risk of recurrent stroke and
about their use in this patient population. cognitive impairment among people who had experienced lacunar

Nature Reviews Neurology

Review article

stroke241. Although people with ICH were not included in that study, 9. Scopelliti, G. et al. Long-term anxiety in spontaneous intracerebral hemorrhage
survivors. Int. J. Stroke 17, 1093–1099 (2022).
the results indicate the need for further studies that particularly target 10. Scopelliti, G. et al. Fatigue after spontaneous intracerebral haemorrhage: prevalence
people with arteriolosclerosis-related ICH. The updated 2023 Stand- and associated factors. Neurological Sci. 15, 2127–2135 (2023).
ards for Reporting Vascular Changes on Neuroimaging (STRIVE)-2 11. Moulin, S. et al. Dementia risk after spontaneous intracerebral haemorrhage:
a prospective cohort study. Lancet Neurol. 15, 820–829 (2016).
criteria190 and the Framework for Clinical Trials in Cerebral Small Ves- 12. Tilling, E. J., El Tawil, S. & Muir, K. W. Do clinicians overestimate the severity of
sel Disease (FINESSE)242 will further guide research on SVD-related intracerebral hemorrhage? Stroke https://doi.org/10.1161/strokeaha.118.022606 (2019).
neuroimaging findings. 13. Samarasekera, N. et al. Influence of intracerebral hemorrhage location on incidence,
characteristics, and outcome: population-based study. Stroke 46, 361–368 (2015).
No disease-modifying treatment currently exists for CAA. Multi- 14. Tamborska, A., Poon, M. T. C. & Al-Shahi Salman, R. Characteristics of randomized trials
ple potential therapy targets are currently in development or under focusing on stroke due to intracerebral hemorrhage: systematic review. Stroke 49,
evaluation, including anti-amyloid antibodies and other compounds 594–600 (2018).
15. Hostettler, I. C., Seiffge, D. J. & Werring, D. J. Intracerebral hemorrhage: an update on
to remove amyloid-β plaques243 and soluble amyloid-β244 or to reduce diagnosis and treatment. Expert. Rev. Neurother. 19, 679–694 (2019).
the production of amyloid-β245,246. A trial of an intrathecal RNA inter- 16. Fandler-Höfler, S. et al. Association of the presence and pattern of MRI markers of
cerebral small vessel disease with recurrent intracerebral hemorrhage. Neurology 101,
ference molecule to reduce expression of the APP gene, and therefore
e794–e804 (2023).
amyloid-β production and vascular deposition, is underway247. Beyond 17. Goeldlin, M. B. et al. CADMUS. Neurology 102, e207977 (2024).
amyloid-targeted treatments, the fibrinolytic system might be another 18. Charidimou, A. et al. Brain hemorrhage recurrence, small vessel disease type, and
cerebral microbleeds: a meta-analysis. Neurology 89, 820–829 (2017).
potential treatment target248. Minocycline inhibits the degradation of
19. Goeldlin, M. B. et al. Small vessel disease burden and risk of recurrent cerebrovascular
perivascular extracellular matrix in cerebral small vessels249. A 2024 events in patients with lacunar stroke and intracerebral haemorrhage attributable to
study suggested that minocycline is safe, generally well tolerated and deep perforator arteriolopathy. Eur. Stroke J. https://doi.org/10.1177/23969873231193237
(2023).
associated with reduced ICH recurrence in a small group of people with
20. Bekelis, K. et al. Computed tomography angiography: improving diagnostic yield and
clinically aggressive CAA250. Future treatment options in CAA might cost effectiveness in the initial evaluation of spontaneous nonsubarachnoid intracerebral
include anti-inflammatory therapies if inflammatory activity is found hemorrhage. J. Neurosurg. 117, 761–766 (2012).
21. Delgado Almandoz, J. E. et al. Diagnostic accuracy and yield of multidetector CT
to be clinically relevant for intracranial bleeding in CAA. angiography in the evaluation of spontaneous intraparenchymal cerebral hemorrhage.
AJNR Am. J. Neuroradiol. 30, 1213–1221 (2009).
Conclusions and future directions 22. van Asch, C. J. et al. External validation of the secondary intracerebral hemorrhage score
in The Netherlands. Stroke 44, 2904–2906 (2013).
Although ICH remains a major challenge for patients, families, 23. Cordonnier, C., Klijn, C. J., van Beijnum, J. & Al-Shahi Salman, R. Radiological
carers and healthcare providers, advances in our understanding of investigation of spontaneous intracerebral hemorrhage: systematic review and
the drivers of morbidity and mortality and the identification of poten- trinational survey. Stroke 41, 685–690 (2010).
24. van Asch, C. J. et al. Diagnostic yield and accuracy of CT angiography, MR angiography,
tial treatment options provide optimism that improved outcomes and digital subtraction angiography for detection of macrovascular causes of
might finally be possible. Recognition of the underlying aetiolo- intracerebral haemorrhage: prospective, multicentre cohort study. BMJ 351, h5762
gies, effects of specific acute medical and surgical treatments, as (2015).
25. Dawkins, A. A. et al. Complications of cerebral angiography: a prospective analysis of
well as new data relevant to the secondary prevention of major adverse 2,924 consecutive procedures. Neuroradiology 49, 753–759 (2007).
vascular events should all contribute to better treatments for ICH. Organ- 26. Kaufmann, T. J. et al. Complications of diagnostic cerebral angiography: evaluation of
ization of care including pre-hospital triage and management, hospital 19,826 consecutive patients. Radiology 243, 812–819 (2007).
27. Hilkens, N. A. et al. Predicting the presence of macrovascular causes in non-traumatic
level of care and use of integrated care pathways (such as care bundles, intracerebral haemorrhage: the DIAGRAM prediction score. J. Neurol. Neurosurg.
and the ‘code ICH’ approach) are currently major implementation chal- Psychiatry 89, 674–679 (2018).
lenges. Alongside the recent surge in promising randomized controlled 28. Fandler-Höfler, S. et al. MRI-based prediction of macrovascular causes of intracerebral
hemorrhage: the MACRO score. Neurology 103, e209950 (2024).
trials, future interventional studies based on improved understanding 29. Raposo, N. et al. A causal classification system for intracerebral hemorrhage subtypes.
of ICH disease mechanisms and natural history should help answer Ann. Neurol. 93, 16–28 (2023).
many key clinical questions and improve outcomes for people with ICH. 30. Rodrigues, M. A. et al. The Edinburgh CT and genetic diagnostic criteria for lobar
intracerebral haemorrhage associated with cerebral amyloid angiopathy: model
development and diagnostic test accuracy study. Lancet Neurol. 17, 232–240 (2018).
Published online: xx xx xxxx 31. Fandler-Hofler, A. S. S. et al. Association of presence and pattern of MRI markers of
cerebral small vessel disease with recurrent intracerebral hemorrhage. Neurology 101,
References e794–e804 (2023).
1. GBD 2019 Stroke Collaborators. Global, regional, and national burden of stroke and its 32. Goeldlin, M. et al. Clinical neuroimaging in intracerebral haemorrhage related to
risk factors, 1990–2019: a systematic analysis for the Global Burden of Disease Study cerebral small vessel disease: contemporary practice and emerging concepts. Expert.
2019. Lancet Neurol. 20, 795–820 (2021). Rev. Neurother. 22, 579–594 (2022).
2. van Asch, C. J. et al. Incidence, case fatality, and functional outcome of intracerebral 33. Greenberg, S. M. & Vonsattel, J. P. Diagnosis of cerebral amyloid angiopathy. Sensitivity
haemorrhage over time, according to age, sex, and ethnic origin: a systematic review and specificity of cortical biopsy. Stroke 28, 1418–1422 (1997).
and meta-analysis. Lancet Neurol. 9, 167–176 (2010). 34. Jäkel, L., De Kort, A. M., Klijn, C. J. M., Schreuder, F. & Verbeek, M. M. Prevalence of
3. Poon, M. T., Fonville, A. F. & Al-Shahi Salman, R. Long-term prognosis after intracerebral cerebral amyloid angiopathy: a systematic review and meta-analysis. Alzheimer’s
haemorrhage: systematic review and meta-analysis. J. Neurol. Neurosurg. Psychiatry 85, Dement. 18, 10–28 (2022).
660–667 (2014). 35. Charidimou, A. et al. The Boston criteria version 2.0 for cerebral amyloid angiopathy:
4. Goeldlin, M. B. et al. Etiology, 3-month functional outcome and recurrent events in a multicentre, retrospective, MRI-neuropathology diagnostic accuracy study. Lancet
non-traumatic intracerebral hemorrhage. J. Stroke 24, 266–277 (2022). Neurol. 21, 714–725 (2022).
5. Foster, L. et al. Effect of deferoxamine on trajectory of recovery after intracerebral 36. Linn, J. et al. Prevalence of superficial siderosis in patients with cerebral amyloid
hemorrhage: a post hoc analysis of the i-DEF trial. Stroke https://doi.org/10.1161/ angiopathy. Neurology 74, 1346–1350 (2010).
strokeaha.121.037298 (2022). 37. Perosa, V. et al. Histopathological correlates of lobar microbleeds in false-positive
6. Shah, V. A. et al. One-year outcome trajectories and factors associated with functional cerebral amyloid angiopathy cases. Ann. Neurol. 94, 856–870 (2023).
recovery among survivors of intracerebral and intraventricular hemorrhage with initial 38. Pasi, M. et al. Mixed-location cerebral hemorrhage/microbleeds: underlying
severe disability. JAMA Neurol. 79, 856–868 (2022). microangiopathy and recurrence risk. Neurology 90, e119–e126 (2018).
7. Ozkan, H. et al. Prevalence, patterns, and predictors of patient-reported non-motor 39. Jolink, W. M. T. et al. Histopathology of cerebral microinfarcts and microbleeds in
outcomes at 30 days after acute stroke: prospective observational hospital cohort study. spontaneous intracerebral hemorrhage. Transl. stroke Res. 14, 174–184 (2023).
Int. J. Stroke https://doi.org/10.1177/17474930231215660 (2023). 40. Malhotra, K. et al. Prevalence, characteristics, and outcomes of undetermined
8. Pasi, M. et al. Long-term functional decline of spontaneous intracerebral haemorrhage intracerebral hemorrhage: a systematic review and meta-analysis. Stroke 52, 3602–3612
survivors. J. Neurol. Neurosurg. Psychiatry https://doi.org/10.1136/jnnp-2020-324741 (2020). (2021).

Nature Reviews Neurology

Review article

41. Falcone, G. J. & Woo, D. Genetics of spontaneous intracerebral hemorrhage. Stroke 75. Smith, E. E. et al. MRI markers of small vessel disease in lobar and deep hemispheric
https://doi.org/10.1161/strokeaha.117.017072 (2017). intracerebral hemorrhage. Stroke 41, 1933–1938 (2010).
42. Caselli, R. J., Walker, D., Sue, L., Sabbagh, M. & Beach, T. Amyloid load in nondemented 76. Charidimou, A. et al. The Cerebral Haemorrhage Anatomical RaTing inStrument
brains correlates with APOE ε4. Neurosci. Lett. 473, 168–171 (2010). (CHARTS): development and assessment of reliability. J. Neurol. Sci. 372, 178–183
43. Filippi, M. et al. Amyloid-related imaging abnormalities and β-amyloid-targeting (2017).
antibodies: a systematic review. JAMA Neurol. 79, 291–304 (2022). 77. Meretoja, A. et al. SMASH-U: a proposal for etiologic classification of intracerebral
44. Guo, H. et al. Genetics of spontaneous intracerebral hemorrhage: risk and outcome. hemorrhage. Stroke 43, 2592–2597 (2012).
Front. Neurosci. 16, 874962 (2022). 78. Marti-Fabregas, J. et al. The H-ATOMIC criteria for the etiologic classification of patients
45. Woo, D. et al. Meta-analysis of genome-wide association studies identifies 1q22 as a with intracerebral hemorrhage. PLoS ONE 11, e0156992 (2016).
susceptibility locus for intracerebral hemorrhage. Am. J. Hum. Genet. 94, 511–521 (2014). 79. Marti-Fabregas, J. et al. SMASH-U versus H-ATOMIC: a head-to-head comparison for
46. Fornage, M. et al. Genome-wide association studies of cerebral white matter lesion the etiologic classification of intracerebral hemorrhage. J. Stroke Cerebrovasc. Dis. 27,
burden: the CHARGE consortium. Ann. Neurol. 69, 928–939 (2011). 2375–2380 (2018).
47. Banerjee, G. et al. Clinical considerations in early-onset cerebral amyloid angiopathy. 80. Jackson, C. A. & Sudlow, C. L. Is hypertension a more frequent risk factor for deep than
Brain. 146, 3991–4014 (2023). for lobar supratentorial intracerebral haemorrhage? J. Neurol. Neurosurg. Psychiatry 77,
48. Manini, A. & Pantoni, L. Genetic causes of cerebral small vessel diseases: a practical 1244–1252 (2006).
guide for neurologists. Neurology 100, 766–783 (2023). 81. Seiffge, D. J. et al. Small vessel disease burden and intracerebral haemorrhage in patients
49. Yamamoto, Y., Craggs, L., Baumann, M., Kalimo, H. & Kalaria, R. N. Review: molecular taking oral anticoagulants. J. Neurol. Neurosurg. Psychiatry https://doi.org/10.1136/
genetics and pathology of hereditary small vessel diseases of the brain. Neuropathol. jnnp-2020-325299 (2021).
Appl. Neurobiol. 37, 94–113 (2011). 82. van Etten, E. S. et al. Trigger factors for spontaneous intracerebral hemorrhage:
50. Bersano, A. et al. The role of clinical and neuroimaging features in the diagnosis of a case-crossover study. Stroke 53, 1692–1699 (2022).
CADASIL. J. Neurol. 265, 2934–2943 (2018). 83. Sallinen, H., Putaala, J. & Strbian, D. Triggering factors in non-traumatic intracerebral
51. Liao, Y. C. et al. Intracerebral hemorrhage in cerebral autosomal dominant hemorrhage. J. Stroke Cerebrovasc. Dis. 29, 104921 (2020).
arteriopathy with subcortical infarcts and leukoencephalopathy: prevalence, 84. Elkind, M. S. V., Boehme, A. K., Smith, C. J., Meisel, A. & Buckwalter, M. S. Infection as a
clinical and neuroimaging features and risk factors. Stroke https://doi.org/10.1161/ stroke risk factor and determinant of outcome after stroke. Stroke 51, 3156–3168 (2020).
strokeaha.120.030664 (2021). 85. Morotti, A. et al. Intracerebral haemorrhage expansion: definitions, predictors, and
52. Sondergaard, C. B., Nielsen, J. E., Hansen, C. K. & Christensen, H. Hereditary cerebral prevention. Lancet Neurol. 22, 159–171 (2023).
small vessel disease and stroke. Clin. Neurol. Neurosurg. 155, 45–57 (2017). 86. Al-Shahi Salman, R. et al. Absolute risk and predictors of the growth of acute
53. Buss, L. et al. Intracerebral haemorrhage in Down syndrome: protected or predisposed? spontaneous intracerebral haemorrhage: a systematic review and meta-analysis of
F1000Res 5, https://doi.org/10.12688/f1000research.7819.1 (2016). individual patient data. Lancet Neurol. 17, 885–894 (2018).
54. Kozberg, M. G., Perosa, V., Gurol, M. E. & van Veluw, S. J. A practical approach to the 87. Yogendrakumar, V. et al. Redefining hematoma expansion with the inclusion of
management of cerebral amyloid angiopathy. Int. J. Stroke 16, 356–369 (2021). intraventricular hemorrhage growth. Stroke 51, 1120–1127 (2020).
55. Chu, S., Xu, F., Su, Y., Chen, H. & Cheng, X. Cerebral amyloid angiopathy (CAA)-related 88. Seiffge, D. J. et al. Meta-analysis of haematoma volume, haematoma expansion and
inflammation: comparison of inflammatory CAA and amyloid-β-related angiitis. mortality in intracerebral haemorrhage associated with oral anticoagulant use. J. Neurol.
J. Alzheimers Dis. 51, 525–532 (2016). 266, 3126–3135 (2019).
56. Salvarani, C. et al. Aβ-related angiitis: comparison with CAA without inflammation and 89. Rodriguez-Luna, D. et al. Ultraearly hematoma growth in active intracerebral
primary CNS vasculitis. Neurology 81, 1596–1603 (2013). hemorrhage. Neurology 87, 357–364 (2016).
57. Scolding, N. J. et al. Abeta-related angiitis: primary angiitis of the central nervous system 90. Rodriguez-Luna, D. et al. Ultraearly hematoma growth predicts poor outcome after acute
associated with cerebral amyloid angiopathy. Brain 128, 500–515 (2005). intracerebral hemorrhage. Neurology 77, 1599–1604 (2011).
58. Auriel, E. et al. Validation of clinicoradiological criteria for the diagnosis of cerebral 91. Dowlatshahi, D. et al. Defining hematoma expansion in intracerebral hemorrhage:
amyloid angiopathy-related inflammation. JAMA Neurol. 73, 197–202 (2016). relationship with patient outcomes. Neurology 76, 1238–1244 (2011).
59. Antolini, L. et al. Spontaneous ARIA-like events in cerebral amyloid angiopathy-related 92. Schlunk, F. & Greenberg, S. M. The pathophysiology of intracerebral hemorrhage
inflammation: a multicenter prospective longitudinal cohort study. Neurology 97, formation and expansion. Transl. stroke Res. 6, 257–263 (2015).
e1809–e1822 (2021). 93. Seiffge, D. J. et al. Cerebral amyloid angiopathy and the risk of hematoma expansion.
60. Regenhardt, R. W. et al. Association between immunosuppressive treatment and Ann. Neurol. 92, 921–930 (2022).
outcomes of cerebral amyloid angiopathy-related inflammation. JAMA Neurol. 77, 94. Seiffge, D. J. & Anderson, C. S. Treatment for intracerebral hemorrhage: dawn of a new
1261–1269 (2020). era. Int. J. Stroke 19, 482–489 (2024).
61. Low, A., Mak, E., Rowe, J. B., Markus, H. S. & O’Brien, J. T. Inflammation and cerebral small 95. Xia, F. et al. The fate of erythrocytes after cerebral hemorrhage. Transl. Stroke Res. 13,
vessel disease: a systematic review. Ageing Res. Rev. 53, 100916 (2019). 655–664 (2022).
62. Banerjee, G. et al. Iatrogenic cerebral amyloid angiopathy: an emerging clinical 96. Xue, M. & Yong, V. W. Neuroinflammation in intracerebral haemorrhage:
phenomenon. J. Neurol. Neurosurg. Psychiatry https://doi.org/10.1136/jnnp-2022-328792 immunotherapies with potential for translation. Lancet Neurol. 19, 1023–1032
(2022). (2020).
63. Kaushik, K. et al. Iatrogenic cerebral amyloid angiopathy post neurosurgery: frequency, 97. Ohashi, S. N. et al. Role of inflammatory processes in hemorrhagic stroke. Stroke 54,
clinical profile, radiological features, and outcome. Stroke 54, 1214–1223 (2023). 605–619 (2023).
64. Khalil, M. et al. Neurofilaments as biomarkers in neurological disorders. Nat. Rev. Neurol. 98. Cliteur, M. P. et al. The association between perihaematomal oedema and functional
14, 577–589 (2018). outcome after spontaneous intracerebral haemorrhage: a systematic review and
65. Qu, Y. et al. Association of plasma neurofilament light with small vessel disease burden in meta-analysis. Eur. Stroke J. 8, 423–433 (2023).
nondemented elderly: a longitudinal study. Stroke 52, 896–904 (2021). 99. Loan, J. J. et al. Secondary injury and inflammation after intracerebral haemorrhage:
66. Duering, M. et al. Serum neurofilament light chain levels are related to small vessel a systematic review and meta-analysis of molecular markers in patient brain tissue.
disease burden. J. Stroke 20, 228–238 (2018). J. Neurol. Neurosurg. Psychiatry 93, 126–132 (2022).
67. Egle, M. et al. Neurofilament light chain predicts future dementia risk in cerebral small 100. Parry-Jones, A. R. et al. Phase II randomised, placebo-controlled, clinical trial of
vessel disease. J. Neurol. Neurosurg. Psychiatry 92, 582–589 (2021). interleukin-1 receptor antagonist in intracerebral haemorrhage: BLOcking the Cytokine
68. Cheng, X. et al. Neurofilament light chain predicts risk of recurrence in cerebral amyloid IL-1 in ICH (BLOC-ICH). Eur. Stroke J. 8, 819–827 (2023).
angiopathy-related intracerebral hemorrhage. Aging 12, 23727–23738 (2020). 101. Cliteur, M. P. et al. Anakinra in cerebral haemorrhage to target secondary injury resulting
69. Perry, L. A. et al. Glial fibrillary acidic protein for the early diagnosis of intracerebral from neuroinflammation (ACTION): study protocol of a phase II randomised clinical trial.
hemorrhage: systematic review and meta-analysis of diagnostic test accuracy. Int. J. Eur. Stroke J. 9, 265–273 (2024).
Stroke 14, 390–399 (2019). 102. Parry-Jones, A. R. et al. Acute care bundles should be used for patients with intrac­
70. Katsanos, A. H. et al. Plasma glial fibrillary acidic protein in the differential diagnosis of erebral haemorrhage: an expert consensus statement. Eur. Stroke J. https://doi.org/
intracerebral hemorrhage. Stroke 48, 2586–2588 (2017). 10.1177/23969873231220235 (2023).
71. Luger, S. et al. Diagnostic accuracy of glial fibrillary acidic protein and ubiquitin 103. Li, Q. et al. Code ICH: a call to action. Stroke 55, 494–505 (2023).
carboxy-terminal hydrolase-L1 serum concentrations for differentiating acute 104. Parry-Jones, A. R. et al. An intracerebral hemorrhage care bundle is associated with lower
intracerebral hemorrhage from ischemic stroke. Neurocrit. Care 33, 39–48 (2020). case fatality. Ann. Neurol. 86, 495–503 (2019).
72. Chen, C. H., Cheng, Y. W., Chen, Y. F., Tang, S. C. & Jeng, J. S. Plasma neurofilament light 105. Ma, L. et al. The third intensive care bundle with blood pressure reduction in acute
chain and glial fibrillary acidic protein predict stroke in CADASIL. J. Neuroinflammation cerebral haemorrhage trial (INTERACT3): an international, stepped wedge cluster
17, 124 (2020). randomised controlled trial. Lancet 402, 27–40 (2023).
73. Verbeek, M. M. et al. Cerebrospinal fluid amyloid β40 is decreased in cerebral amyloid 106. Luger, S. et al. Glial fibrillary acidic protein serum levels distinguish between
angiopathy. Ann. Neurol. 66, 245–249 (2009). intracerebral hemorrhage and cerebral ischemia in the early phase of stroke. Clin. Chem.
74. Renard, D. et al. Interest of CSF biomarker analysis in possible cerebral amyloid 63, 377–385 (2017).
angiopathy cases defined by the modified Boston criteria. J. Neurol. 259, 2429–2433 107. Cooley, S. R. et al. Mobile stroke units facilitate prehospital management of intracerebral
(2012). hemorrhage. Stroke 52, 3163–3166 (2021).

Nature Reviews Neurology

Review article

108. Walter, S. et al. European Stroke Organisation (ESO) guidelines on mobile stroke units for 138. Polymeris, A. A. et al. Tranexamic acid for intracerebral hemorrhage in patients on
prehospital stroke management. Eur. Stroke J. 7, xxvii–lix (2022). non-vitamin K antagonist oral anticoagulants (TICH-NOAC): a multicenter, randomized,
109. Schwabauer, E. et al. Effects of Mobile Stroke Unit dispatch on blood pressure placebo-controlled, phase 2 trial. Stroke 54, 2223–2234 (2023).
management and outcomes in patients with intracerebral haematoma: results from 139. Connolly, S. J. et al. Full study report of andexanet alfa for bleeding associated with
the Berlin_Prehospital Or Usual Care Delivery in acute Stroke (B_PROUD) controlled factor Xa inhibitors. N. Engl. J. Med. 380, 1326–1335 (2019).
intervention study. Eur. Stroke J. 9, 366–375 (2024). 140. Siepen, B. M. et al. Andexanet alfa vs non-specific treatments for intracerebral hemorrhage
110. Li, G. et al. Intensive ambulance-delivered blood-pressure reduction in hyperacute in patients taking factor Xa inhibitors — individual patient data analysis of ANNEXA-4 and
stroke. N. Engl. J. Med. 390, 1862–1872 (2024). TICH-NOAC. Int. J. Stroke https://doi.org/10.1177/17474930241230209 (2024).
111. Langhorne, P. et al. Stroke unit care benefits patients with intracerebral hemorrhage: 141. Demchuk, A. M. et al. Prediction of haematoma growth and outcome in patients with
systematic review and meta-analysis. Stroke 44, 3044–3049 (2013). intracerebral haemorrhage using the CT-angiography spot sign (PREDICT): a prospective
112. Marti-Fabregas, J. et al. Influence of hospital type on outcomes of patients with acute observational study. Lancet Neurol. 11, 307–314 (2012).
spontaneous intracerebral hemorrhage: a population-based study. Neurology 103, 142. Sprigg, N. et al. Tranexamic acid for hyperacute primary IntraCerebral Haemorrhage
e209539 (2024). (TICH-2): an international randomised, placebo-controlled, phase 3 superiority trial.
113. Seiffge, D. J. Good treatment of hyperacute intracerebral hemorrhage is as easy as ABC. Lancet 391, 2107–2115 (2018).
Neurology 103, e209614 (2024). 143. Meretoja, A. et al. Tranexamic acid in patients with intracerebral haemorrhage
114. Willmot, M., Leonardi-Bee, J. & Bath, P. M. High blood pressure in acute stroke and (STOP-AUST): a multicentre, randomised, placebo-controlled, phase 2 trial.
subsequent outcome: a systematic review. Hypertension 43, 18–24 (2004). Lancet Neurol. 19, 980–987 (2020).
115. Anderson, C. S. et al. Rapid blood-pressure lowering in patients with acute intracerebral 144. Liu, J. et al. Tranexamic acid for acute intracerebral haemorrhage growth based on
hemorrhage. N. Engl. J. Med. 368, 2355–2365 (2013). imaging assessment (TRAIGE): a multicentre, randomised, placebo-controlled trial.
116. Qureshi, A. I. et al. Intensive blood-pressure lowering in patients with acute cerebral Stroke Vasc. Neurol. 6, 160–169 (2021).
hemorrhage. N. Engl. J. Med. 375, 1033–1043 (2016). 145. Yassi, N. et al. Tranexamic acid versus placebo in individuals with intracerebral
117. Moullaali, T. J. et al. Blood pressure control and clinical outcomes in acute intracerebral haemorrhage treated within 2 h of symptom onset (STOP-MSU): an international,
haemorrhage: a preplanned pooled analysis of individual participant data. Lancet double-blind, randomised, phase 2 trial. Lancet Neurol. 23, 577–587 (2024).
Neurol. 18, 857–864 (2019). 146. Mayer, S. A. et al. Efficacy and safety of recombinant activated factor VII for acute
118. Moullaali, T. J. et al. Early lowering of blood pressure after acute intracerebral intracerebral hemorrhage. N. Engl. J. Med. 358, 2127–2137 (2008).
haemorrhage: a systematic review and meta-analysis of individual patient data. J. Neurol. 147. Gladstone, D. J. et al. Effect of recombinant activated coagulation factor vii on
Neurosurg. Psychiatry 93, 6–13 (2022). hemorrhage expansion among patients with spot sign-positive acute intracerebral
119. Saver, J. L. et al. Prehospital use of magnesium sulfate as neuroprotection in acute stroke. hemorrhage: the SPOTLIGHT and STOP-IT randomized clinical trials. JAMA Neurol. 76,
N. Engl. J. Med. 372, 528–536 (2015). 1493–1501 (2019).
120. van den Berg, S. A. et al. Prehospital transdermal glyceryl trinitrate in patients with 148. Yogendrakumar, V., Mayer, S. A., Steiner, T., Broderick, J. P. & Dowlatshahi, D. Exploring
presumed acute stroke (MR ASAP): an ambulance-based, multicentre, randomised, hematoma expansion shift with recombinant factor VIIa: a pooled analysis of 4
open-label, blinded endpoint, phase 3 trial. Lancet Neurol. 21, 971–981 (2022). randomized controlled trials. Stroke 54, 2990–2998 (2023).
121. Investigators, R.- Prehospital transdermal glyceryl trinitrate in patients with ultra-acute 149. ISRCTN Registry. Tranexamic acid for very early bleeds in the brain. https://doi.org/10.1186/
presumed stroke (RIGHT-2): an ambulance-based, randomised, sham-controlled, ISRCTN97695350 (2024).
blinded, phase 3 trial. Lancet 393, 1009–1020 (2019). 150. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
122. Greenberg, S. M. et al. 2022 guideline for the management of patients with spontaneous NCT05836831 (2024).
intracerebral hemorrhage: a guideline from the American Heart Association/American 151. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
Stroke Association. Stroke 53, e282–e361 (2022). NCT03044184 (2019).
123. Shoamanesh, A. et al. Canadian stroke best practice recommendations: management 152. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
of spontaneous intracerebral hemorrhage, 7th edition update 2020. Int. J. Stroke 16, NCT04742205 (2024).
321–341 (2021). 153. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
124. Tsivgoulis, G. et al. Neuroimaging and clinical outcomes of oral anticoagulant-associated NCT03496883 (2024).
intracerebral hemorrhage. Ann. Neurol. 84, 694–704 (2018). 154. Eilertsen, H. et al. Haemostatic therapies for stroke due to acute, spontaneous
125. Goeldlin, M. B. et al. Intracerebral haemorrhage volume, haematoma expansion and intracerebral haemorrhage. Cochrane Database Syst. Rev. 10, Cd005951 (2023).
3-month outcomes in patients on antiplatelets. A systematic review and meta-analysis. 155. Sun, S. et al. The effect of mannitol in the early stage of supratentorial hypertensive
Eur. Stroke J. 6, 333–342 (2021). intracerebral hemorrhage: a systematic review and meta-analysis. World Neurosurg.
126. Christensen, H. et al. European stroke organisation guideline on reversal of oral https://doi.org/10.1016/j.wneu.2018.11.249 (2018).
anticoagulants in acute intracerebral haemorrhage. Eur. Stroke J. 4, 294–306 (2019). 156. Shah, M. et al. Effect of hyperosmolar therapy on outcome following spontaneous
127. Baharoglu, M. I. et al. Platelet transfusion versus standard care after acute stroke due intracerebral hemorrhage: Ethnic/Racial Variations of Intracerebral Hemorrhage (ERICH)
to spontaneous cerebral haemorrhage associated with antiplatelet therapy (PATCH): study. J. Stroke Cerebrovasc. Dis. 27, 1061–1067 (2018).
a randomised, open-label, phase 3 trial. Lancet 387, 2605–2613 (2016). 157. Wintzer, S., Heckmann, J. G., Huttner, H. B. & Schwab, S. Dexamethasone in patients with
128. Sondag, L. et al. Variation in medical management and neurosurgical treatment of spontaneous intracerebral hemorrhage: an updated meta-analysis. Cerebrovasc. Dis. 49,
patients with supratentorial spontaneous intracerebral haemorrhage. Eur. Stroke J. 6, 495–502 (2020).
134–142 (2021). 158. Chang, J. J. et al. Minocycline and matrix metalloproteinase inhibition in acute
129. Desborough, M. J. R. et al. Desmopressin for patients with spontaneous intracerebral intracerebral hemorrhage: a pilot study. Eur. J. Neurol. 24, 1384–1391 (2017).
haemorrhage taking antiplatelet drugs (DASH): a UK-based, phase 2, randomised, 159. Fouda, A. Y. et al. Minocycline in acute cerebral hemorrhage: an early phase randomized
placebo-controlled, multicentre feasibility trial. Lancet Neurol. 22, 557–567 (2023). trial. Stroke 48, 2885–2887 (2017).
130. Siepen, B. M. et al. Intracerebral haemorrhage in patients taking different types 160. Fu, Y. et al. Fingolimod for the treatment of intracerebral hemorrhage: a 2-arm proof-of-
of oral anticoagulants: a pooled individual patient data analysis from two national concept study. JAMA Neurol. 71, 1092–1101 (2014).
stroke registries. Stroke Vasc. Neurol. https://doi.org/10.1136/svn-2023-002813 161. Selim, M. et al. Deferoxamine mesylate in patients with intracerebral haemorrhage
(2024). (i-DEF): a multicentre, randomised, placebo-controlled, double-blind phase 2 trial.
131. Steiner, T. et al. Fresh frozen plasma versus prothrombin complex concentrate in patients Lancet Neurol. 18, 428–438 (2019).
with intracranial haemorrhage related to vitamin K antagonists (INCH): a randomised 162. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
trial. Lancet Neurol. 15, 566–573 (2016). NCT04088630 (2024).
132. Kuramatsu, J. B. et al. Anticoagulant reversal, blood pressure levels, and anticoagulant 163. Gao, D. et al. Statins for neuroprotection in spontaneous intracerebral haemorrhage
resumption in patients with anticoagulation-related intracerebral hemorrhage. JAMA. (STATIC): protocol for a multicentre, prospective and randomised controlled trial.
313, 824–836 (2015). BMJ open. 14, e079879 (2024).
133. Sheth, K. N. et al. Time to anticoagulation reversal and outcomes after intracerebral 164. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
hemorrhage. JAMA Neurol. 81, 363–372 (2024). NCT04890379 (2023).
134. Demchuk, A. M. et al. Hemostatic efficacy and anti-FXa (factor Xa) reversal with 165. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
andexanet alfa in intracranial hemorrhage: ANNEXA-4 substudy. Stroke 52, 2096–2105 NCT03463317 (2021).
(2021). 166. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
135. Pollack, C. V. Jr., Reilly, P. A. & Weitz, J. I. Dabigatran reversal with idarucizumab. N. Engl. J. NCT05263167 (2022).
Med. 377, 1691–1692 (2017). 167. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
136. Huttner, H. B. et al. Hematoma expansion and clinical outcomes in patients with factor-Xa NCT05066620 (2021).
inhibitor-related atraumatic intracerebral hemorrhage treated within the ANNEXA-4 trial 168. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
versus real-world usual care. Stroke 53, 532–543 (2022). NCT05369351 (2024).
137. Connolly, S. J. et al. Andexanet for factor Xa inhibitor-associated acute intracerebral 169. US National Library of Medicine.ClinicalTrials.gov https://clinicaltrials.gov/study/
hemorrhage. N. Engl. J. Med. 390, 1745–1755 (2024). NCT05434065 (2024).

Nature Reviews Neurology

Review article

170. De Herdt, V. et al. Early seizures in intracerebral hemorrhage: incidence, associated 201. Anderson, C. S. et al. Triple therapy prevention of recurrent intracerebral disease events
factors, and outcome. Neurology 77, 1794–1800 (2011). trial: rationale, design and progress. Int. J. Stroke 17, 1156–1162 (2022).
171. Peter-Derex, L. et al. Safety and efficacy of prophylactic levetiracetam for prevention of 202. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
epileptic seizures in the acute phase of intracerebral haemorrhage (PEACH): a randomised, NCT03863665 (2021).
double-blind, placebo-controlled, phase 3 trial. Lancet Neurol. 21, 781–791 (2022). 203. RESTART Collaboration. Effects of antiplatelet therapy after stroke due to intracerebral
172. Tran, Q. K. et al. Preventing seizure occurrence following spontaneous intracerebral haemorrhage (RESTART): a randomised, open-label trial. Lancet https://doi.org/10.1016/
haemorrhage: a systematic review and meta-analysis of seizure prophylaxis. Seizure 87, s0140-6736(19)30840-2 (2019).
46–55 (2021). 204. Al-Shahi Salman, R. et al. Effects of antiplatelet therapy after stroke caused by
173. Sondag, L. et al. Neurosurgical intervention for supratentorial intracerebral hemorrhage. intracerebral hemorrhage: extended follow-up of the RESTART randomized clinical trial.
Ann. Neurol. 88, 239–250 (2020). JAMA Neurol. 78, 1179–1186 (2021).
174. Gregson, B. A. et al. Volume reduction with surgery for ICH: when is it effective? Analysis 205. Al-Shahi Salman, R. et al. Effects of antiplatelet therapy on stroke risk by brain imaging
of the CT scans from the STICH II trial. Br. J. Neurosurg. 1–8 https://doi.org/10.1080/ features of intracerebral haemorrhage and cerebral small vessel diseases: subgroup
02688697.2023.2207643 (2023). analyses of the RESTART randomised, open-label trial. Lancet Neurol. 18, 643–652
175. Polster, S. P. et al. Intracerebral hemorrhage volume reduction and timing of intervention (2019).
versus functional benefit and survival in the MISTIE III and STICH trials. Neurosurgery 88, 206. ISRCTN Registry. The use of aspirin or clopidogrel for people who survive stroke due to
961–970 (2021). bleeding in the brain. https://doi.org/10.1186/ISRCTN16705062 (2024).
176. Morgenstern, L. B., Demchuk, A. M., Kim, D. H., Frankowski, R. F. & Grotta, J. C. Rebleeding 207. Biffi, A. et al. Oral anticoagulation and functional outcome after intracerebral
leads to poor outcome in ultra-early craniotomy for intracerebral hemorrhage. Neurology hemorrhage. Ann. Neurol. 82, 755–765 (2017).
56, 1294–1299 (2001). 208. Schreuder, F. et al. Apixaban versus no anticoagulation after anticoagulation-associated
177. Sondag, L. et al. Safety and technical efficacy of early minimally invasive intracerebral haemorrhage in patients with atrial fibrillation in the Netherlands
endoscopy-guided surgery for intracerebral haemorrhage: the Dutch Intracerebral (APACHE-AF): a randomised, open-label, phase 2 trial. Lancet Neurol. 20, 907–916 (2021).
haemorrhage Surgery Trial pilot study. Acta Neurochir. 165, 1585–1596 (2023). 209. Effects of oral anticoagulation for atrial fibrillation after spontaneous intracranial
178. Kuo, L. T. et al. Early endoscope-assisted hematoma evacuation in patients with haemorrhage in the UK: a randomised, open-label, assessor-masked, pilot-phase,
supratentorial intracerebral hemorrhage: case selection, surgical technique, and non-inferiority trial. Lancet Neurol. 20, 842–853 (2021).
long-term results. Neurosurg. Focus. 30, E9 (2011). 210. Al-Shahi Salman, R. et al. Effects of oral anticoagulation in people with atrial fibrillation
179. Kellner, C. P. et al. Time to evacuation and functional outcome after minimally invasive after spontaneous intracranial haemorrhage (COCROACH): prospective, individual
endoscopic intracerebral hemorrhage evacuation. Stroke 52, e536–e539 (2021). participant data meta-analysis of randomised trials. Lancet Neurol. 22, 1140–1149
180. Dammers, R., Beck, J., Volovici, V., Anderson, C. S. & Klijn, C. J. M. Advancing the surgical (2023).
treatment of intracerebral hemorrhage: study design and research directions. World 211. Shoamanesh, A. Anticoagulation in patients with cerebral amyloid angiopathy. Lancet
Neurosurg. 161, 367–375 (2022). 402, 1418–1419 (2023).
181. Pradilla, G. et al. Trial of early minimally invasive removal of intracerebral hemorrhage. 212. Cresti, A. et al. Prevalence of extra-appendage thrombosis in non-valvular
N. Engl. J. Med. 390, 1277–1289 (2024). atrial fibrillation and atrial flutter in patients undergoing cardioversion: a large
182. Beck, J. et al. Decompressive craniectomy plus best medical treatment versus best transoesophageal echo study. EuroInterventi 15, e225–e230 (2019).
medical treatment alone for spontaneous severe deep supratentorial intracerebral 213. Osmancik, P. et al. Left atrial appendage closure versus direct oral anticoagulants in
haemorrhage: a randomised controlled clinical trial. Lancet 403, 2395–2404 (2024). high-risk patients with atrial fibrillation. J. Am. Coll. Cardiol. 75, 3122–3135 (2020).
183. Klijn, C. J., Dammers, R. & Sprigg, N. Decompressive craniectomy for deep 214. Reddy, V. Y. et al. Percutaneous left atrial appendage closure vs warfarin for atrial
intracerebral haemorrhage: a SWITCH towards better outcomes? Lancet 403, 2351–2353 fibrillation: a randomized clinical trial. JAMA 312, 1988–1998 (2014).
(2024). 215. Holmes, D. R. et al. Percutaneous closure of the left atrial appendage versus warfarin
184. Lin, J. & Frontera, J. A. Decompressive hemicraniectomy for large hemispheric strokes. therapy for prevention of stroke in patients with atrial fibrillation: a randomised
Stroke 52, 1500–1510 (2021). non-inferiority trial. Lancet 374, 534–542 (2009).
185. Kuramatsu, J. B. et al. Association of surgical hematoma evacuation vs conservative 216. Horstmann, S. et al. Left atrial appendage occlusion in atrial fibrillation after intracranial
treatment with functional outcome in patients with cerebellar intracerebral hemorrhage. hemorrhage. Neurology 82, 135–138 (2014).
JAMA 322, 1392–1403 (2019). 217. Schrag, M. et al. Left atrial appendage closure for patients with cerebral amyloid
186. Banerjee, G. et al. Longer term stroke risk in intracerebral haemorrhage survivors. angiopathy and atrial fibrillation: the LAA-CAA cohort. Transl. Stroke Res. https://doi.org/
J. Neurol. Neurosurg. Psychiatry 91, 840–845 (2020). 10.1007/s12975-020-00838-5 (2020).
187. Li, L. et al. Risks of recurrent stroke and all serious vascular events after spontaneous 218. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
intracerebral haemorrhage: pooled analyses of two population-based studies. Lancet NCT03243175 (2023).
Neurol. 20, 437–447 (2021). 219. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
188. Boe, N. J. et al. Major cardiovascular events after spontaneous intracerebral hemorrhage NCT02830152 (2024).
by hematoma location. JAMA Netw. open. 6, e235882 (2023). 220. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
189. Fandler-Höfler, S., Gattringer, T., Enzinger, C. & Werring, D. J. Comparison of Boston NCT02928497 (2024).
criteria v2.0/v1.5 for cerebral amyloid angiopathy to predict recurrent intracerebral 221. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/
hemorrhage. Stroke 54, 1901–1905 (2023). NCT04676880 (2024).
190. Duering, M. et al. Neuroimaging standards for research into small vessel 222. Kittner, S. J. et al. Ethnic and racial variation in intracerebral hemorrhage risk factors
disease-advances since 2013. Lancet Neurol. 22, 602–618 (2023). and risk factor burden. JAMA Netw. Open. 4, e2121921 (2021).
191. Wiegertjes, K. et al. Diffusion-weighted imaging lesions and risk of recurrent stroke after 223. Lioutas, V. A. et al. Assessment of incidence and risk factors of intracerebral hemorrhage
intracerebral haemorrhage. J. Neurol. Neurosurg. Psychiatry 92, 950–955 (2021). among participants in the Framingham Heart Study between 1948 and 2016. JAMA Neurol.
192. Murthy, S. B. et al. Diffusion-weighted imaging lesions after intracerebral hemorrhage 77, 1252–1260 (2020).
and risk of stroke: a MISTIE III and ATACH-2 analysis. Stroke 52, 595–602 (2021). 224. Amarenco, P. et al. High-dose atorvastatin after stroke or transient ischemic attack.
193. Murthy, S. B. et al. Association between intracerebral hemorrhage and subsequent N. Engl. J. Med. 355, 549–559 (2006).
arterial ischemic events in participants from 4 population-based cohort studies. 225. Chaturvedi, S. et al. Effect of atorvastatin in elderly patients with a recent stroke or
JAMA Neurol. 78, 809–816 (2021). transient ischemic attack. Neurology 72, 688–694 (2009).
194. Fischer, U. et al. Acute post-stroke blood pressure relative to premorbid levels in 226. Rosenson, R. S. & Tangney, C. C. Antiatherothrombotic properties of statins: implications
intracerebral haemorrhage versus major ischaemic stroke: a population-based study. for cardiovascular event reduction. JAMA 279, 1643–1650 (1998).
Lancet Neurol. 13, 374–384 (2014). 227. Takemoto, M. & Liao, J. K. Pleiotropic effects of 3-hydroxy-3-methylglutaryl coenzyme a
195. Biffi, A. et al. Association between blood pressure control and risk of recurrent reductase inhibitors. Arterioscler. Thromb. Vasc. Biol. 21, 1712–1719 (2001).
intracerebral hemorrhage. JAMA 314, 904–912 (2015). 228. Undas, A., Celińska-Löwenhoff, M., Kaczor, M. & Musial, J. New nonlipid effects of statins
196. Biffi, A. et al. Impact of uncontrolled hypertension at 3 months after intracerebral and their clinical relevance in cardiovascular disease. Thromb. Haemost. 91, 1065–1077
hemorrhage. J. Am. Heart Assoc. 10, e020392 (2021). (2004).
197. PROGRESS Collaborative Group. Randomised trial of a perindopril-based 229. Falcone, G. J. et al. Genetically elevated LDL associates with lower risk of intracerebral
blood-pressure-lowering regimen among 6,105 individuals with previous stroke or hemorrhage. Ann. Neurol. 88, 56–66 (2020).
transient ischaemic attack. Lancet 358, 1033–1041 (2001). 230. Georgakis, M. K. et al. Genetic determinants of blood lipids and cerebral small vessel
198. Benavente, O. R. et al. Blood-pressure targets in patients with recent lacunar stroke: disease: role of high-density lipoprotein cholesterol. Brain 143, 597–610 (2020).
the SPS3 randomised trial. Lancet 382, 507–515 (2013). 231. Ma, C. et al. Low-density lipoprotein cholesterol and risk of intracerebral hemorrhage:
199. Kitagawa, K. et al. Effect of standard vs intensive blood pressure control on the risk a prospective study. Neurology 93, e445–e457 (2019).
of recurrent stroke: a randomized clinical trial and meta-analysis. JAMA Neurol. 76, 232. Mayerhofer, E. et al. Genetically predicted on-statin LDL response is associated with
1309–1318 (2019). higher intracerebral haemorrhage risk. Brain 145, 2677–2686 (2022).
200. Teo, K. C. et al. Blood pressure control targets and risk of cardiovascular and 233. Yu, Z. et al. Lipids, apolipoproteins, statins, and intracerebral hemorrhage: a Mendelian
cerebrovascular events after intracerebral hemorrhage. Stroke 54, 78–86 (2023). randomization study. Ann. Neurol. 92, 390–399 (2022).

Nature Reviews Neurology

Review article

234. Ziff, O. J., Banerjee, G., Ambler, G. & Werring, D. J. Statins and the risk of intracerebral 249. Voigt, S. et al. Minocycline for sporadic and hereditary cerebral amyloid angiopathy
haemorrhage in patients with stroke: systematic review and meta-analysis. J. Neurol. (BATMAN): study protocol for a placebo-controlled randomized double-blind trial. Trials
Neurosurg. Psychiatry 90, 75–83 (2019). 24, 378 (2023).
235. Boe, N. J. et al. Association between statin use and intracerebral hemorrhage location: 250. Bax, F. et al. Minocycline in severe cerebral amyloid angiopathy: a single-center cohort
a nested case–control registry study. Neurology 100, e1048–e1061 (2023). study. J. Am. Heart Assoc. 13, e033464 (2024).
236. Amarenco, P. et al. A comparison of two LDL cholesterol targets after ischemic stroke.
N. Engl. J. Med. 382, 9 (2020). Author contributions
237. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/ S.F.-H., M.B.G., W.M.T.J. and D.J.W. researched data for the article and contributed substantially
NCT03936361 (2024). to discussion of the content. D.J.S., S.F.-H., M.B.G., W.M.T.J., C.J.M.K. and D.J.W. wrote the
238. Sanz-Cuesta, B. E. & Saver, J. L. Lipid-lowering therapy and hemorrhagic stroke risk: article. S.F.-H., Y.D., M.B.G., W.M.T.J. and D.J.W. reviewed and/or edited the manuscript before
comparative meta-analysis of statins and PCSK9 inhibitors. Stroke 52, 3142–3150 submission. D.J.S., S.F-H., C.J.M.K. and D.J.W. contributed equally.
(2021).
239. van Middelaar, T. et al. Effect of antihypertensive medication on cerebral small vessel Competing interests
disease: a systematic review and meta-analysis. Stroke 49, 1531–1533 (2018). DJW reports speaking honoraria from Bayer; speaking and chairing honoraria from Alexion
240. Lai, Y. et al. Effect of intensive blood pressure control on the prevention of white matter and NovoNordisk; and consultancy fees from Alnylam, Bayer and NovoNordisk. He is Co-Chief
hyperintensity: systematic review and meta-analysis of randomized trials. J. Clin. Investigator for the PROHIBIT-ICH trial. He has participated as a member of the LACI-2, TICH-2,
Hypertens. 22, 1968–1973 (2020). TICH-3, RESTART, MACE-ICH and PLINTH Trial Steering Committees. The other authors declare
241. Wardlaw, J. M. et al. Isosorbide mononitrate and cilostazol treatment in patients with no competing interests.
symptomatic cerebral small vessel disease: the Lacunar Intervention Trial-2 (LACI-2)
randomized clinical trial. JAMA Neurol. 80, 682–692 (2023). Additional information
242. Markus, H. S. et al. Framework for clinical trials in cerebral small vessel disease Supplementary information The online version contains supplementary material available at
(FINESSE): a review. JAMA Neurol. 79, 1187–1198 (2022). https://doi.org/10.1038/s41582-024-01035-w.
243. Leurent, C. et al. Immunotherapy with ponezumab for probable cerebral amyloid
angiopathy. Ann. Clin. Transl. Neurol. 6, 795–806 (2019). Peer review information Nature Reviews Neurology thanks C. Anderson and E. Sandset for
244. Greenberg, S. M. et al. A phase 2 study of tramiprosate for cerebral amyloid angiopathy. their contribution to the peer review of this work.
Alzheimer Dis. Assoc. Disord. 20, 269–274 (2006).
245. Chauhan, N. B. & Siegel, G. J. Antisense inhibition at the β-secretase-site of β-amyloid Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
precursor protein reduces cerebral amyloid and acetyl cholinesterase activity in Tg2576. published maps and institutional affiliations.
Neuroscience 146, 143–151 (2007).
246. van Veluw, S. J. et al. Vasomotion as a driving force for paravascular clearance in the Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this
awake mouse brain. Neuron 105, 549–561.e5 (2020). article under a publishing agreement with the author(s) or other rightsholder(s); author
247. US National Library of Medicine. ClinicalTrials.gov https://clinicaltrials.gov/study/ self-archiving of the accepted manuscript version of this article is solely governed by the
NCT06393712 (2024). terms of such publishing agreement and applicable law.
248. Mutimer, C. A. et al. Cerebral amyloid angiopathy and the fibrinolytic system: is plasmin a
therapeutic target? Stroke 52, 2707–2714 (2021). © Springer Nature Limited 2024

Department of Neurology, Inselspital University Hospital Bern and University of Bern, Bern, Switzerland. 2Stroke Research Centre, Department of Brain
1

Repair and Rehabilitation, UCL Queen Square Institute of Neurology, University College London, London, UK. 3Department of Neurology, Medical
University of Graz, Graz, Austria. 4Department of Neurology, Xiangya Hospital of Central South University, Changsha, Hunan, China. 5Department of
Neurology, Isala Hospital, Zwolle, Netherlands. 6Department of Neurology, Donders Institute of Brain, Cognition and Behaviour, Radboud University
Medical Centre, Nijmegen, Netherlands. 7These authors contributed equally: David J. Seiffge, Simon Fandler-Höfler, Catharina J. M. Klijn, David J. Werring.

Nature Reviews Neurology