Subarachnoid Hemorrhage

In depth

The one-page summary above is what the PDF prints. These sections go further.

Pathophysiology and natural history

Most nontraumatic SAH comes from a ruptured saccular aneurysm at the circle of Willis. Incidence is about 6.1 per 100,000 person-years worldwide and 11.4 in the US; women carry 1.3 times the risk, rising after 55 (AHA/ASA 2023). Hypertension and smoking are the modifiable risks; 12% of people with two or more affected first-degree relatives harbor an aneurysm.

At rupture, ICP can surge toward arterial pressure and briefly halt cerebral perfusion, causing sudden collapse, cardiac arrest (about 4%) and the global early brain injury of the first 72 h. Cisternal blood then blocks CSF absorption, irritates cortex and, as it breaks down over days, drives arterial narrowing.

DCI is no longer blamed on large-artery spasm alone: AHA/ASA 2023 lists microthrombosis, cortical spreading depolarization, blood–brain barrier breakdown, failed autoregulation and capillary transit-time heterogeneity alongside it. That is why clazosentan reduced angiographic spasm without improving outcome, while nimodipine improves outcome without reversing spasm.

Prehospital mortality is 22–26% and US in-hospital mortality about 13%, though population case fatality falls about 1.5% a year (AHA/ASA 2023); ESO 2026 puts overall case fatality near 40%. Many survivors keep cognitive, mood and quality-of-life deficits.

Presentation, diagnosis and grading

Beyond the classic thunderclap, atypical patients present with neck pain, syncope, a seizure, confusion or coma. A sentinel headache precedes 10–43%, and a missed warning bleed can be fatal (AHA/ASA 2023).

The Ottawa SAH Rule (alert patients over 15 with a new severe nontraumatic headache peaking within 1 h) calls for workup with age 40 or older, neck pain or stiffness, witnessed loss of consciousness, onset during exertion, thunderclap onset or limited neck flexion: 100% sensitive, only 13.6–15.3% specific (AHA/ASA 2023, 2b).

AHA/ASA 2023 recommends grading every patient with Hunt and Hess or WFNS (Class 1); WFNS, built on GCS and motor deficit, is less subjective. The modified Fisher scale (Frontera 2006; 1,355 patients) adds intraventricular blood to clot thickness: symptomatic vasospasm occurred in 33%, and grade 4 carried an odds ratio of 2.2 against grades 0–1.

Poor grade is not futile: 40% of 98 treated Hunt and Hess 4–5 patients had a favorable 12-month outcome (39% in a 4,506-patient meta-analysis), so treatment is reasonable after a frank family discussion (2a). Correct hydrocephalus, seizures, hyponatremia and hypothermia before judging (AHA/ASA 2023).

Preventing rebleeding: BP, reversal, securing

About 10–15% rebleed, half within 6 h (Thilak 2024; 14% of ULTRA controls), and ESO 2026 puts the risk near 40% if the aneurysm is never treated. Watch for sudden headache, a fall in GCS, new posturing, apnea, cardiac arrest or fresh blood in the EVD.

AHA/ASA 2023 (Class 1) asks for frequent BP checks and short-acting drugs to avoid severe hypertension, hypotension and swings, without a target: lower BP above 180–200 gradually, keep MAP at 65 or more and watch the exam. Observational data tie rebleeding to SBP above 160 but not below 140; older guidelines used 160 (AHA 2012) or 180 (ESO 2013), and NCS 2023 found no basis for a goal. ESO 2026 expert consensus: if SBP stays above 160–180 despite analgesia, nimodipine and a urinary catheter, lower it to about 160 with a short-acting infusion such as labetalol, no diuretics, and a MAP fall of no more than 10% in 2 h.

Nicardipine, clevidipine and labetalol titrate fast (doses on the drip cards); nitroprusside and nitroglycerin dilate cerebral vessels and can raise ICP. Reverse anticoagulants at once (Class 1): 4F-PCC plus IV vitamin K for warfarin, idarucizumab for dabigatran, 4F-PCC for factor Xa inhibitors (NCS/SCCM 2026; andexanet left the US market in December 2025), protamine for heparin. Platelet transfusion for antiplatelet users lacks aSAH evidence (AHA/ASA 2023).

Antifibrinolytics are out: in ULTRA (Lancet 2021), tranexamic acid from a median 185 min after onset until securing gave rebleeding of 10% vs 14% (not significant) and good 6-month outcome in 60% vs 64%; AHA/ASA 2023 (3: No Benefit), NCS 2023 and ESO 2026 advise against routine use.

Securing is the only proven protection: treat as early as feasible, preferably within 24 h of onset (AHA/ASA 2023, Class 1; ESO 2026 consensus), and never defer a late presenter past the DCI window. Aim for complete occlusion (Class 1); a partially secured rupture site is retreated within 1–3 months.

ISAT (2,143 patients, 97% anterior circulation, mostly WFNS 1–3) drove the 1-year coiling advantage. Later, sources differ: a post hoc 5-year analysis found no significant gap in death or dependency (AHA/ASA 2023), while ESO 2026 cites a 10-year coiling advantage (RR 0.81). Coiling rebled slightly more; clipping caused more seizures.

Hydrocephalus and CSF drainage

Blood obstructs CSF flow and absorption. Acute hydrocephalus affects 20–30%, usually within 72 h (Thilak 2024; ESO 2026 about 25%; AHA/ASA 2023 cites 15–87%), with falling arousal, headache, vomiting and impaired upgaze; it can make a patient look poor grade. For symptomatic hydrocephalus, urgent EVD, or lumbar drain when CSF pathways are open, is Class 1, placed before transfer if the first hospital cannot treat the aneurysm.

Before securing, drain gently: rapid CSF removal may provoke rebleeding, though reports conflict (AHA/ASA 2023). Follow the ordered height and clamping plan, and report CSF that turns bright red.

EVD bundles (insertion, dressing, sampling, handling, education) are Class 1: across 10 bundles, infection fell from 6.1–37% to 0–9% (AHA/ASA 2023). Only the neurosurgical team flushes a drain.

In the only weaning RCT (81 patients), direct clamping and 3-day gradual weaning gave the same shunt rate (63.4% vs 62.5%), with shorter drain and ICU days after clamping; NCS 2023 found too little evidence to choose. Shunt dependence (8.9–48%) rises with poor grade, IVH, thick clot, age, rebleeding and long drainage (AHA/ASA 2023).

Prophylactic lumbar drainage is unsettled: EARLYDRAIN (JAMA Neurol 2023; 287 patients) cut unfavorable 6-month outcome from 44.8% to 32.6% and secondary infarction from 39.9% to 28.5%. ESO 2026 (3 trials, 557 patients) found less DCI (OR 0.41) but no significant effect on poor outcome (OR 0.78) and made no recommendation.

Vasospasm and DCI: nimodipine, monitoring, rescue

DCI affects about 30%, mostly on days 4–14 (vasospasm often from day 3) with peak risk on days 6–10, far fewer than show angiographic narrowing (AHA/ASA 2023). The Vergouwen 2010 definition: a new focal deficit or a fall of 2 or more GCS points (total or one component) lasting at least 1 h, not present right after aneurysm treatment and not explained otherwise; DCI-related infarction is a new infarct on CT or MRI within 6 weeks.

Nimodipine is the only drug proven to improve outcome: in the British trial (Pickard 1989; 554 patients), 60 mg every 4 h for 21 days cut infarction from 33% to 22% and poor outcome from 33% to 20%, confirmed by a 16-trial meta-analysis and by ESO 2026 (oral nimodipine, poor outcome OR 0.56). It helps without clearly reducing angiographic spasm (NCS 2023); AHA/ASA 2023 rates it Class 1, level A.

Labels: start within 96 h; puncture capsules with an 18-gauge needle, draw the liquid into an oral syringe labeled not for IV use and flush the tube with 30 mL saline, or give the 6 mg/mL oral solution, 10 mL per dose. Avoid strong CYP3A4 inhibitors (clarithromycin, ketoconazole, ritonavir; contraindicated with the capsules) and inducers such as phenytoin and carbamazepine; cirrhosis takes 30 mg every 4 h.

Interrupted dosing tracks with more DCI (AHA/ASA 2023), so keep the daily dose: split to 30 mg every 2 h and treat BP with fluid or a vasopressor, stopping only for instability the team cannot manage; whether splitting or holding changes outcome is unknown (NCS 2023). Prophylactic IV nicardipine (906 patients) reduced symptomatic spasm but not poor outcome and caused more hypotension (34.5% vs 17.5%); NCS 2023 recommends against it.

Hypovolemia predicts DCI (58% of volume-depleted patients developed it), yet prophylactic hypervolemia and hemodynamic augmentation add pulmonary edema and heart failure without benefit (3: Harm). Aim for euvolemia with isotonic crystalloid; goal-directed monitoring cut DCI from 32% to 13% in one trial, and CVP alone misleads (AHA/ASA 2023).

For symptomatic DCI after securing, raise BP stepwise and recheck the exam at each step (AHA/ASA 2023, 2b): about 80% improve in observational series, and norepinephrine beat phenylephrine in a retrospective study (94% vs 71% improved). HIMALAIA (2018) stopped after 41 of a planned 240 patients for slow recruitment and no perfusion gain: poor outcome RR 1.0, serious adverse events RR 2.1 (not significant); NCS 2023 calls the evidence insufficient, and ESO 2026 consensus reserves it for a secured aneurysm after standard therapy fails.

Persistent deficits go to angiography for intra-arterial verapamil, nicardipine or milrinone (papaverine is avoided for neurotoxicity) and balloon angioplasty of proximal segments (AHA/ASA 2b each; ESO 2026 keeps balloons as a last line). Whether the trigger should be exam alone or exam plus imaging is unstudied (NCS 2023).

Proven not to help: starting a statin (STASH, 803 patients), IV magnesium (MASH-2, 1,203), clazosentan and tirilazad (ESO 2026 against all four; AHA/ASA and NCS 2023 against statins and magnesium, NCS also against clazosentan). NCS addresses starting statins, not stopping home ones; milrinone and intrathecal drugs remain investigational.

Sodium and volume

Hyponatremia (sodium under 135) is reported in 27–44% after aSAH (Thilak 2024), and in 49% and 57% of two endocrine cohorts (Hannon 2014; Sherlock 2006). Natriuresis and polyuria shrink plasma volume, which feeds DCI, so the classic SIADH treatment, water restriction, is dangerous here.

Which is more common is disputed: Hannon 2014 (100 patients, mostly good grade) traced 71% of cases to SIAD and 8% to acute glucocorticoid deficiency, with no salt wasting; Sherlock 2006 found SIADH in 39 and salt wasting in 4 of 62 patients with sodium under 130. Exams still favor salt wasting; after SAH the treatment is the same.

Use isotonic saline to hold euvolemia and hypertonic saline (often 3%) for a falling or symptomatic sodium, checking sodium every 4–6 h. Fludrocortisone limits natriuresis (Mori 1999: 0.3 mg/day, with less sodium and fluid needed and steady sodium); AHA/ASA 2023 rates mineralocorticoids 2a, while NCS 2023 found the evidence insufficient. Watch potassium, volume and BP; high-dose hydrocortisone adds hyperglycemia, GI bleeding and heart failure (AHA/ASA 2023).

Avoid hypotonic fluids and free water. Dilute polyuria with a rising sodium is diabetes insipidus instead, which needs desmopressin and water.

Heart, lungs, seizures and other systems

The catecholamine surge at rupture stuns myocardium. Nearly all patients show ECG changes (deep T-wave inversions, long QT, ST shifts), and troponin rises in about 28% within 24 h (Thilak 2024). Rhythm disturbances affected 35%, 5% life-threatening (Solenski 1995). Wall motion loss (28% in Banki 2006: low EF 15%, regional only 13%) crosses coronary territories, mostly basal and midventricular with the apex spared (apical ballooning also occurs), and usually recovers.

Wall motion abnormalities and troponin elevation are tied to death (RR 1.9 and 2.0) and DCI (RR 2.1 and 3.2; van der Bilt 2009). Get an ECG, troponin and echo at baseline and with instability. Support follows heart-failure principles, but no inotrope is proven better (some avoid dobutamine as an adrenergic agonist), and diuresis must not leave the patient dry (Ramnarain 2025).

Neurogenic pulmonary edema (2–31%, Ramnarain 2025), aspiration and ARDS (lung injury in up to 27%) worsen outcome. A ventilator bundle (lung-protective volumes, moderate PEEP, early feeding, protocolized extubation) is Class 1 when ventilation exceeds 24 h; prone positioning and recruitment for severe ARDS may be reasonable with ICP monitoring (2b). Keep PaCO₂ normal: hypocapnia constricts cerebral vessels.

Seizure-like events are reported in up to 26%, seizures in 7.8–15.2% in recent studies (AHA/ASA 2023). Prophylaxis may be reasonable with MCA aneurysm, poor grade, ICH, hydrocephalus or cortical infarction (2b) and is not beneficial without them. After a seizure at onset, treat for 7 days or less (2a), since longer courses do not prevent later seizures; seizures that occur after onset need longer treatment. Phenytoin prophylaxis is linked to excess morbidity, mortality and worse cognition (3: Harm), so levetiracetam is usual. cEEG for 24–48 h (longer in coma) suits a depressed or fluctuating exam (2a).

Fever is common, often central, and tied to worse outcome; seek infection and treat it, though temperature management has not improved outcome (2b). Avoid hyperglycemia and hypoglycemia (2a); insulin to 80–120 mg/dL cut infections without better outcome. SAHaRA (742 patients) found a hemoglobin trigger of 10 g/dL no better than 8 (unfavorable outcome 33.5% vs 37.7%); NCS 2023, written before it, found too little evidence for a trigger above 7.

VTE complicates 4–24%. Prophylaxis after securing is Class 1; most units start compression devices at admission. Starting heparin within 24 h of securing caused no intracranial bleeding in a case-control study (over 40% had an EVD), and enoxaparin 40 mg daily did not increase bleeding in a small RCT (AHA/ASA 2023).

Nursing care by system and transfer

• Neuro: frequent GCS or NIHSS checks (Class 1), often hourly, with pupils, speech and drift; report a 2-point GCS fall or new deficit at once.
• Cardiovascular: arterial line; BP in the ordered band for the phase; telemetry for QT and arrhythmias.
• Respiratory: swallow screen before oral intake (Class 1), aspiration precautions, HOB 30°.
• Fluids: strict I&O, daily weight, urine output and sodium trend; isotonic fluids; never fluid-restrict.
• Drugs: nimodipine on time (split only by order); bowel regimen; acetaminophen and short-acting opioids for headache.
• EVD: level, height, output and color hourly; bundle care; clamp for moves per protocol.
• Mobility: compression devices, heparin after securing, a mobility algorithm once secured (2a).
• Family: early goals-of-care talks; screen mood and cognition; finish 21 days of nimodipine.

Transfer to a high-volume neurocritical care center is Class 1 (AHA/ASA 2023; ESO 2026 suggests centers treating at least 70 aSAH patients a year, 35 in remote areas). Before transport, protect the airway, control BP, reverse anticoagulants and drain hydrocephalus; en route, keep drips titrated, manage the drain per protocol and keep nimodipine on time.

Evidence and controversies

Unsettled: a BP target before securing, nimodipine during hypotension and shorter courses, induced hypertension, triggers for angiography, lumbar drainage, mineralocorticoids, transfusion thresholds, EVD weaning and antiplatelet reversal. Older exam items still teach triple-H and salt wasting as the usual cause of low sodium.

Exam pearls

• Thunderclap headache: CT first; a negative high-quality CT within 6 h nearly excludes SAH; later, LP for xanthochromia.
• Hunt and Hess and WFNS grade severity (IV–V poor); a higher modified Fisher grade (thick clot, IVH) predicts vasospasm.
• Rebleeding is the early killer (half within 6 h): short-acting BP drips, reverse anticoagulants, secure within 24 h; no tranexamic acid.
• Nimodipine 60 mg every 4 h for 21 days, enteral only; it improves outcome, not angiographic spasm; split to 30 mg every 2 h for hypotension.
• DCI: days 3–14; a new deficit or GCS drop of 2 for 1 h; euvolemia, induced hypertension once secured, then endovascular therapy.
• Prophylactic triple-H, statins, magnesium and clazosentan do not improve outcome.
• Hyponatremia after SAH: salt and volume (saline, hypertonic saline, fludrocortisone), never fluid restriction.
• Acute hydrocephalus (drowsiness, enlarging ventricles): EVD, leveled at the tragus.
• Stunned myocardium: deep T inversions, long QT, troponin and reversible wall motion loss.
• Seizure at onset: an antiseizure drug for 7 days or less; no phenytoin prophylaxis.