Skip to content

Intracerebral hemorrhage

TL;DR — Spontaneous intracerebral hemorrhage (ICH) is primary bleeding into brain parenchyma; outcome is driven by initial tissue destruction, hematoma expansion, intraventricular extension, hydrocephalus, edema, and complications of critical illness. The first hours require rapid CT/CTA, blood-pressure control, immediate reversal of anticoagulation, neurosurgical and neurocritical-care assessment, and repeated examination rather than therapeutic nihilism (Greenberg 2022, PMID 35579034). Lowering systolic pressure toward 140 mm Hg is generally safe in mild-to-moderate ICH, but forcing 110–139 mm Hg did not improve outcome and increased renal adverse events (Anderson 2013, PMID 23713578; Qureshi 2016, PMID 27276234). Platelet transfusion harms non-surgical patients with antiplatelet-associated ICH, whereas four-factor prothrombin complex concentrate rapidly reverses vitamin-K antagonists and andexanet improves hematoma control in factor-Xa-inhibitor ICH at the cost of more thrombosis (Baharoglu 2016, PMID 27178479; Steiner 2016, PMID 27302126; Connolly 2024, PMID 38749032). Conventional craniotomy is not routinely beneficial, but targeted minimally invasive evacuation now has positive evidence for selected lobar hemorrhage; selection, timing, residual volume, and avoidance of self-fulfilling withdrawal-of-care decisions remain central uncertainties (Pradilla 2024, PMID 38598795).

Definition, location, and cause

“Spontaneous” ICH excludes trauma but not an underlying lesion. Deep hemorrhage in basal ganglia, thalamus, pons, or cerebellum often reflects arteriolosclerosis associated with chronic hypertension; lobar hemorrhage raises cerebral amyloid angiopathy, tumor, vascular malformation, venous thrombosis, and hemorrhagic transformation as competing explanations. Location is probabilistic, not etiologic proof.

Pattern Leading considerations Imaging or history that changes the pathway
Deep hemispheric Arteriolosclerosis; anticoagulation; stimulant-associated pressure surge CTA if young, unusual morphology, no hypertension, or disproportionate subarachnoid blood
Lobar Cerebral amyloid angiopathy; AVM/dural fistula; tumor; venous thrombosis CTA/CTV acutely; contrast MRI and susceptibility imaging after stabilization
Cerebellar Hypertensive arteriopathy; vascular lesion Brainstem compression, hydrocephalus, size and neurological trajectory determine surgery urgency
Pontine Arteriolosclerosis; cavernous malformation MRI after acute phase when appropriate
Predominantly intraventricular Small parenchymal source, AVM/aneurysm, moyamoya CTA/DSA; hydrocephalus and external ventricular drain assessment
Hemorrhage with infarct pattern Hemorrhagic transformation; venous thrombosis Arterial and venous vascular imaging

The 2022 AHA/ASA guideline recommends etiologic investigation tailored to age, location, hypertension history, and vascular-imaging findings rather than labeling every deep bleed “hypertensive” and every lobar bleed “amyloid” (Greenberg 2022, PMID 35579034).

First-hour priorities

Priority What to establish Actionable consequence
Airway and ventilation GCS trajectory, bulbar function, oxygenation Intubation when airway protection or gas exchange fails; avoid prophylactic intubation based on score alone
Non-contrast CT Location, dimensions, IVH, hydrocephalus, mass effect Establish baseline; trigger neurosurgery and repeat-imaging plan
CTA ± venous phase Macrovascular cause and active contrast extravasation DSA pathway; expansion-risk enrichment
Hemostasis Platelets, PT/INR, aPTT, fibrinogen, last anticoagulant dose, renal function Drug-specific reversal without waiting for deterioration
Blood pressure Repeated, reliable measurements; pain/agitation Smooth titratable reduction; avoid variability and overshoot
Neurological baseline GCS, NIHSS, pupils, brainstem signs Detect expansion, hydrocephalus, seizure, or herniation
Goals and prior function Values, comorbidity, advance directives Separate informed limitation from score-driven nihilism

Hematoma expansion

Expansion is a modifiable early target and a marker of ongoing bleeding. The CTA “spot sign”—contrast pooling within the hematoma—was initially associated with expansion in 8 of 9 spot-positive versus 0 of 30 spot-negative patients in a small cohort (Wada 2007, PMID 17322083). In the prospective PREDICT study, the spot sign predicted substantial growth with 51% sensitivity, 85% specificity, 61% positive predictive value, and 78% negative predictive value; it also predicted mortality and poor outcome (Demchuk 2012, PMID 22405630). A negative sign therefore does not exclude expansion.

Expansion-risk feature Interpretation Limitation
Short interval from onset to CT More time remains for continued bleeding Unknown onset weakens time-based inference
Larger baseline volume More bleeding surface and worse prognosis Volume is simultaneously predictor and outcome determinant
CTA spot sign Active contrast extravasation phenotype Moderate sensitivity; acquisition and definition vary
Anticoagulant exposure Impaired hemostasis Drug level and last dose may be uncertain
Irregular shape / heterogeneous density Non-contrast CT markers of active bleeding Less standardized than CTA
Intraventricular extension Severe phenotype and hydrocephalus risk May occur without further parenchymal growth

Serial CT is most useful when the result changes hemostatic, surgical, hydrocephalus, or prognostic management. A fixed routine schedule cannot replace urgent imaging after neurological deterioration.

Blood-pressure lowering

INTERACT2 randomized 2,839 patients within 6 hours to a systolic target below 140 mm Hg within 1 hour or guideline treatment below 180 mm Hg. Death or major disability was 52.0% versus 55.6% (OR 0.87, 95% CI 0.75–1.01; p=0.06), while prespecified ordinal mRS analysis favored intensive treatment (OR 0.87, 95% CI 0.77–1.00; p=0.04); serious adverse events were similar (Anderson 2013, PMID 23713578).

ATACH-2 tested a more aggressive 110–139 versus 140–179 mm Hg target. Death or disability was 38.7% versus 37.7% (adjusted RR 1.04, 95% CI 0.85–1.27), and renal adverse events within 7 days were 9.0% versus 4.0% (Qureshi 2016, PMID 27276234). Together, the trials support prompt, smooth lowering around 140 mm Hg for many patients presenting between 150 and 220 mm Hg, not indiscriminate pursuit of values near 110 mm Hg. Evidence is less certain for massive ICH, impending herniation, and patients requiring urgent surgery (Greenberg 2022, PMID 35579034).

Hemostatic treatment

Anticoagulant reversal

Reversal is an emergency treatment, not a response to documented expansion.

Exposure Evidence-based reversal principle Quantitative evidence
Warfarin / vitamin-K antagonist Four-factor PCC plus intravenous vitamin K; PCC preferred to plasma INCH stopped early: INR normalization within 3 h in 67% with PCC vs 9% with FFP; hematoma expansion was more frequent with FFP (Steiner 2016, PMID 27302126)
Dabigatran Idarucizumab 5 g; consider renal clearance and rebound RE-VERSE AD: median maximum reversal 100%; serious thrombotic events 6.3% by 90 days in the bleeding cohort (Pollack 2017, PMID 28693366)
Factor-Xa inhibitor Andexanet alfa or PCC according to availability, timing, drug and local protocol ANNEXA-4 achieved excellent/good hemostasis at 12 h in 82% of evaluable major bleeds; 10% had thrombosis by 30 days (Connolly 2019, PMID 30730782)
Heparin Protamine, dose tied to recent heparin exposure Drug-specific pharmacology; evidence is not from randomized ICH trials (Greenberg 2022, PMID 35579034)

ANNEXA-I directly randomized factor-Xa-inhibitor-associated acute ICH. Hemostatic efficacy was 67.0% with andexanet and 53.1% with usual care (adjusted difference 13.4 percentage points, p=0.003), mainly through less expansion. Thrombotic events were 10.3% versus 5.6%, including ischemic stroke in 6.5% versus 1.5%; 30-day mortality and modified Rankin outcomes did not differ substantially (Connolly 2024, PMID 38749032). Hemostatic superiority is therefore not equivalent to proven net functional benefit.

Antiplatelet exposure

PATCH randomized patients with spontaneous supratentorial ICH taking antiplatelet therapy, without planned emergency surgery. Platelet transfusion worsened death or dependence at 3 months (adjusted common OR 2.05, 95% CI 1.18–3.56) and increased serious adverse events (42% vs 29%) (Baharoglu 2016, PMID 27178479). This finding applies to routine transfusion in the studied non-surgical population; thrombocytopenia and emergency neurosurgery are distinct questions.

Non-specific prohemostatic drugs

Recombinant factor VIIa reduced hematoma expansion but did not improve survival or functional outcome in FAST; arterial thromboembolic serious events were more frequent with 80 µg/kg (9% vs 4% with placebo) (Mayer 2008, PMID 18480205). TICH-2 found tranexamic acid did not significantly improve 90-day functional status (adjusted OR 0.88, 95% CI 0.76–1.03), although hematoma expansion and day-7 mortality were lower (Sprigg 2018, PMID 29778325). In CTA-spot-positive STOP-AUST, growth occurred in 44% with tranexamic acid and 52% with placebo (OR 0.72, 95% CI 0.32–1.59), an underpowered neutral result (Meretoja 2020, PMID 33128912). Biological target engagement has repeatedly failed to guarantee functional benefit.

Neurosurgery and minimally invasive evacuation

Emergency surgery remains anatomically compelling for a deteriorating cerebellar hemorrhage with brainstem compression, obstructive hydrocephalus, or substantial volume; this is supported mainly by observational evidence and guideline consensus rather than a placebo-controlled trial (Greenberg 2022, PMID 35579034). For supratentorial ICH, the randomized record is heterogeneous.

Trial Population / technique Functional result What it does and does not show
STICH 1,033 supratentorial ICH; early craniotomy vs initial conservative care favorable outcome 26% vs 24%; OR 0.89, p=0.414 No overall benefit of routine early open surgery (Mendelow 2005, PMID 15680453)
STICH II 601 superficial lobar ICH, no IVH unfavorable outcome 59% vs 62%; OR 0.86, p=0.367 Neutral primary result; possible survival advantage, not proof for all lobar ICH (Mendelow 2013, PMID 23726393)
MISTIE III Catheter aspiration plus alteplase; ICH ≥30 mL mRS 0–3 at 1 y: 45% vs 41%; adjusted risk difference 4%, p=0.33 Procedure was safe and reduced mortality; achieving residual ≤15 mL was associated with better outcome, but the primary functional endpoint was neutral (Hanley 2019, PMID 30739747)
ENRICH Trans-sulcal parafascicular evacuation within 24 h; 30–80 mL lobar or anterior basal ganglia 180-d utility-weighted mRS 0.458 vs 0.374; difference 0.084 Benefit driven by lobar subgroup; basal-ganglia enrollment stopped for futility (Pradilla 2024, PMID 38598795)

ENRICH changes the question from whether “surgery” works to which route, anatomy, timing, and degree of evacuation produce benefit. Its findings should not be generalized to posterior basal ganglia, thalamic, brainstem, cerebellar, very small, or very large hemorrhage outside trial criteria.

Intraventricular hemorrhage and hydrocephalus

Intraventricular extension obstructs CSF pathways and independently worsens prognosis. External ventricular drainage treats hydrocephalus and intracranial pressure but may not clear clot rapidly. CLEAR III enrolled patients with a routinely placed drain, stable ICH under 30 mL, and obstructive IVH; intraventricular alteplase did not substantially improve mRS ≤3 (48% vs 45%; RR 1.06), but reduced mortality (18% vs 29%; hazard ratio 0.60) while increasing survival with mRS 5 (17% vs 9%) (Hanley 2017, PMID 28081952). Mortality reduction must therefore be interpreted alongside the distribution of survivor disability.

Complication prevention

  • Venous thromboembolism: In immobile stroke patients, thigh-length intermittent pneumatic compression reduced proximal DVT within 30 days from 12.1% to 8.5% (adjusted OR 0.65) and improved 6-month survival (Dennis 2013, PMID 23727163). Pharmacologic prophylaxis requires a judgment about hematoma stability.
  • Seizures: Treat clinical or electrographic seizures; depressed or fluctuating consciousness can justify continuous EEG. Routine prophylactic antiseizure medication without evidence of seizures has no established outcome benefit (Greenberg 2022, PMID 35579034).
  • Corticosteroids: Dexamethasone did not improve mortality and substantially increased complications, principally infection and diabetes, in a randomized trial (Poungvarin 1987, PMID 3574383). It has no role for perihematomal edema from primary ICH.
  • Swallowing, fever, and immobility: aspiration prevention, temperature evaluation, skin care, nutrition, and graded rehabilitation occur alongside hematoma-directed care, not after it.

Prognostication without self-fulfilling prophecy

The ICH Score assigns points for GCS, age ≥80, infratentorial origin, ICH volume ≥30 cm³, and intraventricular hemorrhage; 30-day mortality rose monotonically from 0% at score 0 to 100% at score 5 in the derivation cohort (Hemphill 2001, PMID 11283388). The FUNC score uses age, GCS, location, volume, and pre-ICH cognitive impairment to estimate functional independence (Rost 2008, PMID 18556582). Both stratify populations; neither can determine an individual's acceptable outcome or justify automatic early treatment limitation.

Early do-not-resuscitate orders and withdrawal decisions can become part of the causal pathway to death. The AHA/ASA guideline recommends avoiding premature limitation and separating DNAR status from other medical and surgical treatment decisions (Greenberg 2022, PMID 35579034). Reassessment after stabilization, sedation clearance, hydrocephalus treatment, and control of expansion provides a more defensible basis for shared decisions.

Restarting antithrombotic therapy

RESTART randomized 537 survivors of antithrombotic-associated ICH a median 76 days after the event. Recurrent ICH occurred in 4% allocated antiplatelet therapy and 9% allocated avoidance (adjusted HR 0.51, 95% CI 0.25–1.03); the trial excluded more than a modest increase in recurrence but did not prove protection (RESTART Collaboration 2019, PMID 31128924).

For atrial fibrillation, SoSTART was inconclusive for non-inferiority: recurrent intracranial hemorrhage occurred in 8% starting oral anticoagulation and 4% avoiding it (adjusted HR 2.42, 95% CI 0.72–8.09) (SoSTART Collaboration 2021, PMID 34487722). APACHE-AF found high rates of non-fatal stroke or vascular death whether apixaban was used (26%) or avoided (24%), with wide uncertainty (Schreuder 2021, PMID 34687635). Location, MRI markers, ischemic risk, blood-pressure control, and alternatives such as left-atrial-appendage occlusion remain essential to individualized decisions (cardioembolic and cryptogenic stroke).

Open questions

  • Can an expansion-enriched treatment combine spot-sign selection, ultra-early delivery, and a clinically meaningful endpoint? Spot-sign prediction is imperfect, while tranexamic acid and factor VIIa altered bleeding without improving function (Demchuk 2012, PMID 22405630; Mayer 2008, PMID 18480205; Sprigg 2018, PMID 29778325).
  • Which patients derive net benefit from andexanet, given improved hemostasis but a 5-percentage-point excess of thrombosis and no demonstrated 30-day functional benefit (Connolly 2024, PMID 38749032)?
  • Can ENRICH's lobar benefit be replicated and separated into effects of timing, access corridor, and residual volume (Pradilla 2024, PMID 38598795; Hanley 2019, PMID 30739747)?
  • Does IVH clearance improve outcomes if near-complete clot removal is reliably achieved, or does it chiefly shift death toward survival with severe disability (Hanley 2017, PMID 28081952)?
  • After ICH in atrial fibrillation, which imaging and clinical phenotype favors anticoagulation, appendage occlusion, or neither? Both pilot randomized trials remain imprecise (SoSTART Collaboration 2021, PMID 34487722; Schreuder 2021, PMID 34687635).

References

  1. Greenberg SM, et al. 2022 Guideline for the management of patients with spontaneous intracerebral hemorrhage. Stroke. 2022;53:e282-e361. PMID 35579034.
  2. Wada R, et al. CT angiography “spot sign” predicts hematoma expansion in acute intracerebral hemorrhage. Stroke. 2007;38:1257-1262. PMID 17322083.
  3. Demchuk AM, et al. Prediction of haematoma growth and outcome using the CT-angiography spot sign (PREDICT). Lancet Neurol. 2012;11:307-314. PMID 22405630.
  4. Anderson CS, et al. Rapid blood-pressure lowering in patients with acute intracerebral hemorrhage. N Engl J Med. 2013;368:2355-2365. PMID 23713578.
  5. Qureshi AI, et al. Intensive blood-pressure lowering in patients with acute cerebral hemorrhage. N Engl J Med. 2016;375:1033-1043. PMID 27276234.
  6. Steiner T, et al. Fresh frozen plasma versus prothrombin complex concentrate in vitamin-K-antagonist-related intracranial haemorrhage (INCH). Lancet Neurol. 2016;15:566-573. PMID 27302126.
  7. Pollack CV Jr, et al. Idarucizumab for dabigatran reversal—full cohort analysis. N Engl J Med. 2017;377:431-441. PMID 28693366.
  8. Connolly SJ, et al. Full study report of andexanet alfa for bleeding associated with factor Xa inhibitors. N Engl J Med. 2019;380:1326-1335. PMID 30730782.
  9. Connolly SJ, et al. Andexanet for factor Xa inhibitor-associated acute intracerebral hemorrhage. N Engl J Med. 2024;390:1745-1755. PMID 38749032.
  10. Baharoglu MI, et al. Platelet transfusion versus standard care after antiplatelet-associated spontaneous cerebral haemorrhage (PATCH). Lancet. 2016;387:2605-2613. PMID 27178479.
  11. Mayer SA, et al. Efficacy and safety of recombinant activated factor VII for acute intracerebral hemorrhage. N Engl J Med. 2008;358:2127-2137. PMID 18480205.
  12. Sprigg N, et al. Tranexamic acid for hyperacute primary intracerebral haemorrhage (TICH-2). Lancet. 2018;391:2107-2115. PMID 29778325.
  13. Meretoja A, et al. Tranexamic acid in patients with intracerebral haemorrhage (STOP-AUST). Lancet Neurol. 2020;19:980-987. PMID 33128912.
  14. Mendelow AD, et al. Early surgery versus initial conservative treatment in spontaneous supratentorial intracerebral haematomas (STICH). Lancet. 2005;365:387-397. PMID 15680453.
  15. Mendelow AD, et al. Early surgery versus initial conservative treatment in superficial lobar intracerebral haematomas (STICH II). Lancet. 2013;382:397-408. PMID 23726393.
  16. Hanley DF, et al. Minimally invasive surgery with thrombolysis in intracerebral haemorrhage evacuation (MISTIE III). Lancet. 2019;393:1021-1032. PMID 30739747.
  17. Pradilla G, et al. Trial of early minimally invasive removal of intracerebral hemorrhage. N Engl J Med. 2024;390:1277-1289. PMID 38598795.
  18. Hanley DF, et al. Thrombolytic removal of intraventricular haemorrhage in severe stroke (CLEAR III). Lancet. 2017;389:603-611. PMID 28081952.
  19. Dennis M, et al. Effectiveness of intermittent pneumatic compression after stroke (CLOTS 3). Lancet. 2013;382:516-524. PMID 23727163.
  20. Poungvarin N, et al. Effects of dexamethasone in primary supratentorial intracerebral hemorrhage. N Engl J Med. 1987;316:1229-1233. PMID 3574383.
  21. Hemphill JC 3rd, et al. The ICH score: a simple, reliable grading scale for intracerebral hemorrhage. Stroke. 2001;32:891-897. PMID 11283388.
  22. Rost NS, et al. Prediction of functional outcome in patients with primary intracerebral hemorrhage: the FUNC score. Stroke. 2008;39:2304-2309. PMID 18556582.
  23. RESTART Collaboration. Effects of antiplatelet therapy after stroke due to intracerebral haemorrhage (RESTART). Lancet. 2019;393:2613-2623. PMID 31128924.
  24. SoSTART Collaboration. Effects of oral anticoagulation for atrial fibrillation after spontaneous intracranial haemorrhage in the UK. Lancet Neurol. 2021;20:842-853. PMID 34487722.
  25. Schreuder FHBM, et al. Apixaban versus no anticoagulation after anticoagulation-associated intracerebral haemorrhage in atrial fibrillation (APACHE-AF). Lancet Neurol. 2021;20:907-916. PMID 34687635.