Skip to content

Subarachnoid hemorrhage

TL;DR — Aneurysmal subarachnoid hemorrhage (aSAH) is a neurovascular emergency in which rebleeding, hydrocephalus, early brain injury, and delayed cerebral ischemia (DCI) create distinct hazards over days. Sudden maximal-at-onset headache, especially with collapse, neck stiffness, vomiting, or focal deficit, requires a validated diagnostic pathway; a normal early CT is reassuring only under specific timing and image-quality conditions (Perry 2011, PMID 21768192; Perry 2013, PMID 24065011). The ruptured aneurysm should be secured as early as feasible by a multidisciplinary cerebrovascular team, using coiling or clipping according to anatomy, patient factors, durability, and treatment risk; in selected patients suitable for either, ISAT found lower 1-year death or dependency with coiling but more late rebleeding (Molyneux 2002, PMID 12414200; Molyneux 2009, PMID 19329361). Enteral nimodipine improves functional outcome even though it is not a reliable angiographic vasospasm treatment (Pickard 1989, PMID 2496789; Dorhout Mees 2007, PMID 17636626). DCI is a clinical syndrome, not a synonym for vessel narrowing: serial examination, perfusion assessment, euvolemia, and targeted rescue are more coherent than prophylactic hypervolemia or treating a Doppler number in isolation (Vergouwen 2010, PMID 20798370; Hoh 2023, PMID 37212182).

Recognition and diagnosis

The classic presentation is a headache reaching maximal intensity within seconds to a minute. Some patients present instead with collapse, seizure, meningism, vomiting, photophobia, confusion, focal deficit, or coma. A “sentinel headache” history is retrospective and cannot distinguish minor aneurysmal leakage from other thunderclap causes on its own.

The Ottawa SAH Rule was derived for alert adults with new severe non-traumatic headache reaching maximum intensity within one hour. Any of age ≥40, neck pain/stiffness, witnessed loss of consciousness, exertional onset, thunderclap onset, or limited neck flexion makes the rule positive; sensitivity was 100% (95% CI 97.2–100) but specificity only 15.3% (Perry 2013, PMID 24065011). It is a rule-out aid for its defined population, not a diagnostic test and not applicable to patients with a neurological deficit, prior aneurysm, established recurrent headache syndrome, or other exclusions.

In a prospective cohort using modern scanners, CT performed within 6 hours of headache onset was 100% sensitive (95% CI 97.0–100) when interpreted by qualified radiologists; overall CT sensitivity across all times was 92.9% (Perry 2011, PMID 21768192). Performance depends on exact timing, scanner quality, reader expertise, hemoglobin, and a neurologically intact population. When suspicion persists after a negative or delayed CT, CTA and/or lumbar puncture remain pathway choices; CTA detects aneurysms but can reveal incidental lesions, whereas lumbar puncture can detect blood breakdown products but is vulnerable to traumatic taps (Hoh 2023, PMID 37212182).

Finding Immediate implication Important alternative
Diffuse basal cistern SAH Ruptured saccular aneurysm until proven otherwise Perimesencephalic non-aneurysmal SAH
Focal convexity SAH Not the usual aneurysmal pattern Reversible cerebral vasoconstriction, venous thrombosis, amyloid angiopathy, vasculitis
Intraventricular blood / hydrocephalus CSF-flow obstruction; urgent neurosurgical review Intraventricular extension from ICH
Negative CTA with diffuse SAH Small/blister/dissecting aneurysm may be occult Catheter angiography and sometimes repeat study
Negative CT after the validated early window SAH not excluded by the “6-hour rule” LP and/or vascular imaging according to context

Severity and baseline description

The World Federation of Neurosurgical Societies scale combines GCS and motor deficit into five clinical grades (WFNS Committee 1988, PMID 3131498). The original Fisher scale linked the amount and distribution of CT blood to vasospasm (Fisher 1980, PMID 7354892); the modified Fisher scale separates thick cisternal blood and intraventricular hemorrhage and showed a more graded association with symptomatic vasospasm (Frontera 2006, PMID 16823296).

Domain Useful measure What it misses
Clinical severity WFNS grade; GCS components Sedation, hydrocephalus, seizure, intoxication and shock can depress the exam
Blood burden Modified Fisher grade Aneurysm anatomy, early brain injury, physiologic reserve
Aneurysm anatomy Size, neck, branch incorporation, location, multiplicity Wall fragility and technical difficulty are incompletely captured
Baseline brain injury Infarct, edema, global hypoxic injury, ICH Evolution after resuscitation and CSF diversion

Poor grade is not synonymous with futility. Hydrocephalus, seizures, medication, metabolic disturbance, and treatable cardiopulmonary failure can obscure recoverable neurological function. In a cohort examining withdrawal of technological life support after SAH, treatment withdrawal was independently associated with death and varied with clinical factors, illustrating the self-fulfilling-prognosis problem (Kowalski 2013, PMID 24166245).

Preventing rebleeding and securing the aneurysm

Rebleeding risk is front-loaded before definitive occlusion and is often catastrophic. Priorities are controlled analgesia, avoidance of severe blood-pressure surges and hypotension, reversal of anticoagulation, correction of major coagulopathy, and transfer to a center capable of both endovascular and microsurgical treatment. Current AHA/ASA guidance recommends securing the aneurysm as early as feasible, preferably within 24 hours, with complete obliteration where achievable (Hoh 2023, PMID 37212182).

Coiling versus clipping

ISAT randomized 2,143 patients with ruptured aneurysms considered suitable for either treatment. At one year, death or dependency occurred in 23.7% assigned endovascular treatment and 30.6% assigned neurosurgery, a relative risk reduction of 22.6% (Molyneux 2002, PMID 12414200). Long-term follow-up found a continuing survival advantage for coiling, but recurrent hemorrhage from the treated aneurysm was more frequent after coiling; absolute late rebleeding risk was low (Molyneux 2009, PMID 19329361). At 18 years in the UK cohort, more patients allocated coiling were alive and independent, while rebleeding remained uncommon in both groups (Molyneux 2015, PMID 25465111).

The pragmatic Barrow Ruptured Aneurysm Trial enrolled a broader surgical population and allowed crossovers; at 6 years, the initial functional advantage of coiling was no longer significant for anterior-circulation aneurysms, while posterior-circulation outcomes continued to favor coiling (Spetzler 2015, PMID 26115467). The contrast with ISAT reinforces that anatomy, crossover, retreatment, and time horizon matter when applying an average treatment effect.

Factor favoring endovascular treatment Factor favoring microsurgical treatment
Posterior-circulation aneurysm; older/frailer patient; favorable catheter anatomy Large intraparenchymal clot needing evacuation; branch incorporated into neck; morphology unsuitable for durable coiling
Lower immediate procedural burden in many anatomies Direct visualization and durable neck closure in selected lesions
Access to balloon/stent/flow-diversion rescue Need to avoid dual antiplatelet therapy required by some endovascular constructs

The randomized evidence applies most directly to aneurysms judged suitable for either approach, predominantly anterior-circulation lesions in patients with relatively favorable grades. Blister, dissecting, giant, complex middle-cerebral, and flow-diverter-treated ruptured aneurysms require anatomy-specific inference. Treatment choice should be documented as a joint neurointerventional and cerebrovascular surgical decision (Hoh 2023, PMID 37212182).

Antifibrinolysis is not a substitute for securing the aneurysm

ULTRA tested tranexamic acid started immediately after diagnosis and continued until aneurysm treatment or 24 hours. Good 6-month outcome occurred in 60% with tranexamic acid and 64% with usual care (adjusted OR 0.86, 95% CI 0.66–1.12); rebleeding before aneurysm treatment was 10% versus 14% (Post 2021, PMID 33357465). The reduction in pre-treatment bleeding did not translate into functional benefit in a system with early aneurysm treatment, so routine ultra-early tranexamic acid is not supported.

Nimodipine and general neurocritical care

The British Aneurysm Nimodipine Trial randomized 554 patients: cerebral infarction occurred in 22% on nimodipine versus 33% on placebo, and poor outcome in 20% versus 33% (Pickard 1989, PMID 2496789). Earlier randomized work also found fewer severe neurological deficits from vasospasm with nimodipine (Allen 1983, PMID 6338383). A Cochrane synthesis concluded calcium antagonists reduce poor neurological outcome, with the evidence driven principally by oral nimodipine (Dorhout Mees 2007, PMID 17636626).

Nimodipine is therefore continued enterally for 21 days when tolerated. Hypotension may require dose fractionation or temporary interruption, but using angiographic vasospasm as the reason to start or stop it misunderstands the evidence: outcome benefit is stronger than proof of large-vessel caliber change.

Physiological problem Management principle Avoid
Hypovolemia Isotonic replacement; track intake, output, weight, examination and perfusion Prophylactic hypervolemia and indiscriminate fluid loading
Hyponatremia / natriuresis Evaluate volume state; replace sodium and volume when depleted Fluid restriction that worsens hypovolemia in cerebral salt wasting
Fever Identify infection and control temperature Assuming every fever is central
Hyperglycemia Avoid severe hyperglycemia and hypoglycemia Intensive normalization with hypoglycemic risk
Anemia Integrate oxygen delivery, ischemia and cardiopulmonary status A universal transfusion threshold unsupported by definitive SAH evidence
Cardiopulmonary dysfunction ECG/troponin/echo when indicated; support perfusion and oxygenation Treating neurogenic myocardial injury as primary coronary occlusion without context

Fludrocortisone reduced urinary sodium and improved the efficiency of hypervolemic therapy in a small randomized study, but pulmonary edema and contemporary avoidance of prophylactic hypervolemia limit direct translation (Mori 1999, PMID 10584839). The practical target is euvolemia, not a positive fluid balance (Hoh 2023, PMID 37212182).

Delayed cerebral ischemia: syndrome, not scan label

Consensus defines clinical DCI as a new focal deficit or a fall of at least 2 GCS points lasting ≥1 hour, not immediately after aneurysm occlusion and not attributable to another cause; cerebral infarction is a separate imaging outcome (Vergouwen 2010, PMID 20798370). “Vasospasm” should describe arterial narrowing, while DCI describes clinical or infarction consequences. The terms are not interchangeable because DCI can reflect distal arteriolar dysfunction, impaired autoregulation, spreading depolarizations, microthrombosis, and systemic physiology as well as proximal vasoconstriction (Vergouwen 2011, PMID 21748502).

Surveillance signal Strength Failure mode
Serial neurological examination Directly detects clinical DCI Limited by coma, sedation, aphasia and fluctuating attention
Transcranial Doppler Repeatable bedside trend; strongest for MCA Operator and window dependence; velocity affected by flow and vessel diameter
CTA Shows proximal narrowing Anatomical vasospasm does not prove tissue ischemia
CT perfusion Maps delayed perfusion Thresholds and artifacts vary; transport burden
Continuous EEG May detect declining alpha variability or seizures Indirect and resource intensive
Catheter angiography Highest spatial resolution; permits treatment Invasive; a procedural test rather than routine screening

Meta-analysis found TCD-defined vasospasm predicted DCI with pooled sensitivity 90% and specificity 71%, useful for surveillance but insufficient as a stand-alone treatment trigger (Kumar 2016, PMID 26495942). A changed exam demands a parallel search for rebleeding, hydrocephalus, seizure, infection, hypoxia, hypotension, electrolyte disturbance, and medication effect.

Treatment of symptomatic DCI

After the aneurysm is secured, raising blood pressure may improve perfusion in selected symptomatic patients, but randomized evidence is weak. HIMALAIA stopped after 41 participants because of slow recruitment and lack of expected benefit: induced hypertension did not significantly improve overall cerebral blood flow and serious adverse events occurred in 2 of 21 treated patients versus 0 of 20 controls (Gathier 2018, PMID 29158449). Treatment should be a monitored response to a clinical perfusion deficit, not prophylactic hypertension.

For deficits refractory to optimized systemic physiology, catheter angiography can identify treatable proximal spasm; balloon angioplasty provides mechanical dilation, while intra-arterial vasodilators reach more distal vessels. A meta-analysis of intra-arterial vasodilator series reported frequent angiographic and neurological improvement but was dominated by uncontrolled observational studies, so comparative functional efficacy is uncertain (Venkatraman 2018, PMID 28663521). Intravenous milrinone protocols remain based on non-randomized experience and require controlled evaluation (Bernier 2021, PMID 33480639).

Strategies that changed a surrogate but failed the patient-centered endpoint

Strategy Trial result Lesson
Intravenous magnesium MASH-2: poor outcome RR 1.03 (95% CI 0.85–1.25) No routine magnesium for neuroprotection (Dorhout Mees 2012, PMID 22633825)
Simvastatin 40 mg STASH: no difference in 6-month ordinal mRS; common OR 0.97 Do not initiate statin specifically to prevent DCI (Kirkpatrick 2014, PMID 24837690)
Clazosentan CONSCIOUS-2: reduced angiographic vasospasm in its program but no significant improvement in vasospasm-related morbidity/all-cause mortality; more pulmonary complications, anemia and hypotension Proximal vessel caliber is not an adequate surrogate for recovery (Macdonald 2011, PMID 21640651)
Prophylactic hypervolemia No high-quality functional benefit; increases cardiopulmonary burden Maintain euvolemia instead (Hoh 2023, PMID 37212182)
Routine tranexamic acid ULTRA: less pre-treatment rebleeding signal, no functional benefit Faster definitive aneurysm treatment is the priority (Post 2021, PMID 33357465)

Hydrocephalus, seizures, and deterioration

Acute symptomatic hydrocephalus requires urgent CSF diversion, usually an external ventricular drain. Drainage strategy must balance control of intracranial pressure, rebleeding concern before aneurysm occlusion, catheter obstruction, ventriculitis, and avoidance of prolonged dependence. In a 2026 meta-analysis of six studies (1,802 patients), rapid versus gradual weaning did not significantly change shunt placement (RR 0.94, 95% CI 0.57–1.54) or infection (RR 0.99, 95% CI 0.55–1.76), but shortened hospital stay by 4.3 days (95% CI 3.0–5.7); heterogeneity and predominantly nonrandomized evidence leave the ideal protocol unresolved as of 2026-08-30 (Brown 2026, PMID 41949294; Hoh 2023, PMID 37212182).

Treat clinical or electrographic seizures. Continuous EEG is reasonable for unexplained impaired or fluctuating consciousness; routine long-term prophylaxis is not supported for every patient, while short-course prophylaxis may be considered in higher-risk phenotypes such as ruptured MCA aneurysm with intraparenchymal hematoma (Hoh 2023, PMID 37212182).

Deterioration timing Must-not-miss causes
Minutes to hours Rebleeding, hydrocephalus, seizure, expanding ICH, sedation or airway failure
Days 1–3 Early brain injury, hydrocephalus, seizures, cardiopulmonary failure, sodium disturbance
Days 3–14 DCI, infection, hydrocephalus, electrolyte/volume disturbance, venous thromboembolism
After discharge Recurrent hemorrhage, hydrocephalus, seizure, medication complications, cognitive/affective deterioration

Open questions

  • Which multimodal signature detects DCI early enough to improve outcome in patients who cannot be examined, without over-treating angiographic vasospasm (Vergouwen 2010, PMID 20798370; Kumar 2016, PMID 26495942)?
  • Does induced hypertension improve patient-centered outcomes in a selected DCI phenotype, and what cardiac or pulmonary reserve is required (Gathier 2018, PMID 29158449)?
  • Which ruptured aneurysms gain enough durability from clipping to offset the lower early morbidity seen with coiling in ISAT (Molyneux 2002, PMID 12414200; Molyneux 2009, PMID 19329361)?
  • Can a treatment targeting microcirculation, spreading depolarization, or inflammation succeed where anti-vasospasm drugs failed (Macdonald 2011, PMID 21640651; Vergouwen 2011, PMID 21748502)?
  • What drainage and weaning strategy minimizes shunt dependence, infection, and ICU duration after acute hydrocephalus (Hoh 2023, PMID 37212182)?

References

  1. Hoh BL, et al. 2023 Guideline for the management of patients with aneurysmal subarachnoid hemorrhage. Stroke. 2023;54:e314-e370. PMID 37212182.
  2. Perry JJ, et al. Sensitivity of computed tomography performed within six hours of onset of headache for diagnosis of subarachnoid haemorrhage. BMJ. 2011;343:d4277. PMID 21768192.
  3. Perry JJ, et al. Clinical decision rules to rule out subarachnoid hemorrhage. JAMA. 2013;310:1248-1255. PMID 24065011.
  4. World Federation of Neurosurgical Societies Committee. Report on a universal subarachnoid hemorrhage grading scale. J Neurosurg. 1988;68:985-986. PMID 3131498.
  5. Fisher CM, et al. Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomographic scanning. Neurosurgery. 1980;6:1-9. PMID 7354892.
  6. Frontera JA, et al. Prediction of symptomatic vasospasm after subarachnoid hemorrhage: the modified Fisher scale. Neurosurgery. 2006;59:21-27. PMID 16823296.
  7. Kowalski RG, et al. Withdrawal of technological life support following subarachnoid hemorrhage. Neurocrit Care. 2013;19:269-275. PMID 24166245.
  8. Molyneux A, et al. International Subarachnoid Aneurysm Trial of neurosurgical clipping versus endovascular coiling in 2143 patients. Lancet. 2002;360:1267-1274. PMID 12414200.
  9. Molyneux AJ, et al. Risk of recurrent subarachnoid haemorrhage, death, or dependence after clipping or coiling: ISAT long-term follow-up. Lancet Neurol. 2009;8:427-433. PMID 19329361.
  10. Molyneux AJ, et al. Durability of endovascular coiling versus neurosurgical clipping: 18-year follow-up of the UK ISAT cohort. Lancet. 2015;385:691-697. PMID 25465111.
  11. Post R, et al. Ultra-early tranexamic acid after subarachnoid haemorrhage (ULTRA). Lancet. 2021;397:112-118. PMID 33357465.
  12. Pickard JD, et al. Effect of oral nimodipine on cerebral infarction and outcome after subarachnoid haemorrhage. BMJ. 1989;298:636-642. PMID 2496789.
  13. Allen GS, et al. Cerebral arterial spasm—a controlled trial of nimodipine in patients with subarachnoid hemorrhage. N Engl J Med. 1983;308:619-624. PMID 6338383.
  14. Dorhout Mees SM, et al. Calcium antagonists for aneurysmal subarachnoid haemorrhage. Cochrane Database Syst Rev. 2007;2007:CD000277. PMID 17636626.
  15. Mori T, et al. Improved efficiency of hypervolemic therapy with inhibition of natriuresis by fludrocortisone. J Neurosurg. 1999;91:947-952. PMID 10584839.
  16. Vergouwen MD, et al. Definition of delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage. Stroke. 2010;41:2391-2395. PMID 20798370.
  17. Vergouwen MD. Vasospasm versus delayed cerebral ischemia as an outcome event. Neurocrit Care. 2011;15:308-311. PMID 21748502.
  18. Kumar G, et al. Vasospasm on transcranial Doppler is predictive of delayed cerebral ischemia: systematic review and meta-analysis. J Neurosurg. 2016;124:1257-1264. PMID 26495942.
  19. Gathier CS, et al. Induced hypertension for delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage. Stroke. 2018;49:76-83. PMID 29158449.
  20. Venkatraman A, et al. Intra-arterial vasodilators for vasospasm following aneurysmal subarachnoid hemorrhage: a meta-analysis. J Neurointerv Surg. 2018;10:380-387. PMID 28663521.
  21. Bernier TD, et al. Treatment of subarachnoid hemorrhage-associated delayed cerebral ischemia with milrinone. J Neurosurg Anesthesiol. 2021;33:195-202. PMID 33480639.
  22. Dorhout Mees SM, et al. Magnesium for aneurysmal subarachnoid haemorrhage (MASH-2). Lancet. 2012;380:44-49. PMID 22633825.
  23. Kirkpatrick PJ, et al. Simvastatin in aneurysmal subarachnoid haemorrhage (STASH). Lancet Neurol. 2014;13:666-675. PMID 24837690.
  24. Macdonald RL, et al. Clazosentan in aneurysmal subarachnoid haemorrhage undergoing surgical clipping (CONSCIOUS-2). Lancet Neurol. 2011;10:618-625. PMID 21640651.
  25. Spetzler RF, et al. The Barrow Ruptured Aneurysm Trial: 6-year results. J Neurosurg. 2015;123:609-617. PMID 26115467.
  26. Brown GW, et al. Rapid versus gradual weaning of external ventricular drains in aneurysmal subarachnoid hemorrhage: a systematic review and meta-analysis. J Neurosurg Sci. 2026;70:140-150. PMID 41949294.