Subarachnoid Hemorrhage

Subarachnoid hemorrhage (SAH) is a neurologic emergency due to bleeding into the subarachnoid space. Mortality can reach 50%. The clinical presentation is most often in the form of headache, classically defined as maximal at onset and worst of life. The most common cause is traumatic; approximately 80% of nontraumatic SAH are due to aneurysmal rupture, with the remainder from idiopathic peri-mesencephalic hemorrhage or other less common causes. Noncontrast brain computed tomography (CT) performed within 6 hours of symptom onset has sensitivity approaching 100%. Lumbar puncture may be considered after this period for definitive diagnosis if initial CT is normal.

Key points

  • Subarachnoid hemorrhage (SAH) is deadly, with 25% dying within 24 hours and an overall mortality rate of 50%.

  • Outside of trauma, SAH most commonly arises from aneurysmal rupture (80%) but may be due to peri-mesencephalic bleed or other less common causes.

  • Most patients will present with sudden, maximal headache, associated with nausea/vomiting, neck pain, and exertion. The headache is usually different than patients’ baseline headaches.

  • Diagnosis centers on head noncontrast computed tomography (CT). If conducted within 6 hours of headache onset, this test is reliable. If it is negative but patients present after 6 hours, lumbar puncture and/or CT angiography should be used.

  • Management requires rapid neurologic assessment, monitoring for intracranial pressure elevation, nimodipine, blood pressure management, analgesia, seizure treatment, and coagulopathy correction.


Subarachnoid hemorrhage (SAH) is a neurologic emergency and is defined by bleeding in the subarachnoid space, which lies between the arachnoid and pia mater. This area is normally filled with cerebrospinal fluid (CSF). Trauma is the most common cause of SAH. Most nontraumatic SAH, approximately 80%, is due to ruptured aneurysm. The causes of nonaneurysmal SAH are diverse, and the mechanism may not be identified.


Subarachnoid hemorrhage (SAH) is a neurologic emergency and is defined by bleeding in the subarachnoid space, which lies between the arachnoid and pia mater. This area is normally filled with cerebrospinal fluid (CSF). Trauma is the most common cause of SAH. Most nontraumatic SAH, approximately 80%, is due to ruptured aneurysm. The causes of nonaneurysmal SAH are diverse, and the mechanism may not be identified.


Headache accounts for approximately 2% of emergency department (ED) visits, with SAH occurring in 1% to 3% of these patients. The incidence is approximately 7 to 10 per 100,000, with mortality approaching 50%. SAH is the most common form of intracranial hemorrhage in trauma. Close to 15% of patients will die before they reach the hospital, with 25% dying within 24 hours and 45% of patients dying within 30 days. Morbidity is also severe, with only one-third of patients demonstrating full recovery after treatment.

Prognosis is predicted by level of consciousness and neurologic examination on initial evaluation, patient age (younger patients experience better outcome), and amount of hemorrhage on initial imaging (increased hemorrhage associated with worse outcome). For patients who reach the hospital, early complications of SAH account for most mortality, including rebleeding, vasospasm, seizures, increased intracranial pressure (ICP), and cardiac complications.



Most nontraumatic SAHs are due to aneurysmal rupture, and these aneurysms are usually not congenital. Most never rupture and arise at sites of arterial branching, specifically the circle of Willis in the anterior circulation. Saccular aneurysms account for 90%, and the overall prevalence of cerebral aneurysm ranges from 0.5% to 6.0% depending on the population. A systematic review including more than 56,000 patients from 23 studies found an incidence of 2.3%. Risk factors include a family history of SAH or aneurysm, smoking, hypertension, and heavy alcohol use.


Peri-mesencephalic SAH is characterized by localized blood on computed tomography (CT) without aneurysm. These bleeds are defined by hemorrhage restricted to the cisterns around the brainstem with absence of aneurysm on vascular imaging, such as CT angiography (CTA) and magnetic resonance angiography (MRA). This type has a much better prognosis than aneurysmal SAH. Other causes include vascular malformation, intracranial dissection, sickle cell disease with intracerebral hemorrhage, pituitary apoplexy, cerebral amyloid angiopathy, central nervous system tumor, cocaine use, and cerebral venous thrombosis.


SAH is a common form of intracranial bleeding in trauma. It results from disruption of the parenchyma and subarachnoid vasculature and often presents with headache, meningeal signs, and photophobia. This finding is one of the most common CT findings in patients with moderate to severe traumatic brain injury, and traumatic SAH is associated with a 3-fold increase in mortality.

Features and presentation

Most patients with SAH experience abrupt headache, often thunderclap in nature, defined by a headache that reaches maximal intensity within 1 minute. However, 10% to 25% of patients with thunderclap headache have SAH. Most of these headaches are atypical in nature and different from patients’ prior headaches. The headache may begin or worsen with exertion, and it may lateralize to the side of the bleed in 30% of patients. Key historical features are shown in Table 1 .

Table 1

Key historical features of subarachnoid hemorrhage

Feature Question

  • Did the headache reach maximal intensity suddenly rather than gradually (over 1 h or more)?

    • Sudden onset is suggestive.


  • Have you had headaches before? How does this compare?

    • Different or new headache is concerning.

Other symptoms

  • Are there other symptoms, including seizure, syncope, neck stiffness, focal neurologic deficit, vomiting, or change in vision?

    • New or different symptoms are concerning.

Sentinel bleeding may occur weeks before the maximal bleeding. Approximately 30% to 50% of patients experience this sentinel headache, which most commonly precedes a major bleed by 1 to 3 weeks. Close to 70% of patients present with headache and no neurologic deficit. Nausea and vomiting may occur in 77% of patients, though vomiting is not predictive. Carpenter and colleagues conducted a recent meta-analysis evaluating the diagnostic accuracy of history, physical examination, imaging, and lumbar puncture (LP). The key findings of this study are demonstrated in Table 2 . Of note, the absence of the worst headache of life and onset of headache that is more than 1 hour possess likelihood ratios (LRs) that are less than 1, though confidence intervals (CIs) cross 1. Other findings, such as family history of cerebral aneurysm, lethargy, history of headache, scotomata, and diplopia, demonstrate LRs with CIs crossing 1.

Table 2

Key findings of Carpenter study on subarachnoid hemorrhage

Evaluated Characteristic Likelihood Ratio (95% CI)
Neck pain history 4.12 (2.24–7.59)
Neck stiffness on examination 6.59 (3.95–11.00)
Absence of worst headache of life 0.36 (0.01–14.22)
Onset of headache more than 1 h 0.06 (0–0.95)
Noncontrast head CT within 6 h positive for SAH 230 (6–8700)
Noncontrast head CT within 6 h negative for SAH 0.01 (0–0.04)
Noncontrast head CT beyond 6 h negative for SAH 0.07 (0.01–0.61)
CSF analysis: RBC count ≥1000 × 10 6 /L 5.7 (1.4–23.0)
CSF analysis: RBC count <1000 × 10 6 /L 0.21 (0.03–1.7)
Visible xanthochromia Present: 24.67 (12.13–50.14)
Absent: 0.22 (0.09–0.54)

Abbreviations: CI, confidence interval; RBC, red blood cell.

Data from Carpenter CR, Hussain AM, Ward MJ, et al. Spontaneous subarachnoid hemorrhage: a systematic review and meta-analysis describing the diagnostic accuracy of history, physical examination, imaging, and lumbar puncture with an exploration of test thresholds. Acad Emerg Med 2016;23(9):963–1003.

Seizure at the time of onset is predictive of bleeding. Seizures occur in less than 20% of patients during or shortly after SAH. Loss of consciousness affects approximately 25% to 53% of patients; neck stiffness, or meningismus, may occur in 35% of patients after several hours as a reaction to blood in the subarachnoid space. Cranial nerve III palsy is due to direct local pressure from an aneurysm arising from the posterior communicating artery. Up to 50% of patients will have neurologic abnormalities on examination, ranging from mental status change to focal deficit.

Patients may present with a combination of symptoms that suggest another diagnosis. Isolated neck pain, fever, headache, nausea and vomiting, elevated blood pressure, or electrocardiogram (ECG) changes (deep T-wave inversions or ST changes) may occur. ECG changes may be due to catecholamine surge or autonomic vascular tone increase. Some patients may experience cardiac arrest. All of these make the diagnosis more difficult, especially in comatose patients.

Differential diagnosis

Although the classic presentation of SAH is sudden severe headache, SAH may present with vague headache and normal neurologic status. However, many other conditions may present with headache, including some with significant morbidity and mortality. This differential is demonstrated in Table 3 .

Table 3

Differential of sudden-onset, severe headache

History and Physical Examination Findings
Deadly Causes
Hypertensive encephalopathy Severe hypertension with altered consciousness
Cervical or cranial artery dissection Neck and/or face pain, usually abrupt onset, several variations of neurologic deficit
Cerebral venous and dural sinus thrombosis Headache with focal deficit or seizure with risk factors, including hypercoagulable state, pregnancy, tobacco use
Carbon monoxide poisoning Headache, nausea, and vomiting; often with multiple patients affected with similar symptoms
Idiopathic intracranial hypertension Obese females with papilledema, may have cranial nerve VI deficit
Meningitis or encephalitis Fever, headache, stiff neck in meningitis; encephalitis may present with focal deficits or seizure
Giant cell arteritis Commonly in patients aged >50 y, decreased pulse in temporal artery, temporal artery tenderness, ESR elevation, may have vision loss
Acute angle closure glaucoma Painful eye with decreased vision, corneal edema, pupil midposition
Spontaneous intracranial hypotension Headache worse when upright and improves when supine
Mass lesion (tumor, abscess, cyst) Neurologic deficit common, commonly focal; may have altered mental status
Pituitary apoplexy Headache with visual deficit, sudden onset; patient commonly with pituitary tumor
Posterior reversible encephalopathy syndrome Recurrent sudden-onset headache with nausea, vomiting, altered mental status, visual field changes; may have seizure; patients often with history of hypertension and renal disease
Stroke: hemorrhagic Sudden-onset headache with neurologic deficit, altered mental status; patients often hypertensive
Benign Causes
Migraine, tension, cluster, exertional, cough, viral sinusitis

Abbreviation: ESR, erythrocyte sedimentation rate.


Diagnosis centers on several investigations including imaging and laboratory analysis such as CSF. Misdiagnosis most commonly arises from 3 errors: failure to appreciate the full clinical spectrum of SAH, failure to obtain initial cerebral imaging, and failure to perform LP in the correct settings. Particularly in alert, neurologically normal patients, this diagnosis can be difficult, with up to 53% of patients with SAH missed on initial presentation. Missed diagnosis can result in mortality that approaches 50%, increasing to 70% in patients with rebleed.

The American College of Emergency Physicians’ (ACEP) clinical policy on the evaluation and management of adult patients with headache provides a level B recommendation for lumbar puncture (LP) in patients who present with sudden-onset, severe headache and negative noncontrast head CT. Likewise, the American Heart Association (AHA) gives a level B recommendation for LP following negative head CT noncontrast. The AHA guidelines also give a level C recommendation for CTA as a follow-up test when a noncontrast head CT is nondiagnostic in patients with suspected SAH.

Noncontrast head computed tomography

Noncontrast head CT is the primary means of diagnosis; however, early generation scanners had the potential to miss 5% of cases. Thus, LP has traditionally been advocated for those patients with suspected SAH and a negative noncontrast CT. CT technology is rapidly improving. First-generation scanners demonstrated sensitivity of 92% within 24 hours of headache onset, but more advanced-generation scanners approach a sensitivity of 100% if completed within 6 hours of symptom onset. Fig. 1 displays SAH on CT.

Fig. 1

( A, B ) Noncontrast-enhanced CT of the head; a classic example of SAH with intraventricular extension. Note the starfish appearance of the hyperdensity caused by the blood in the subarachnoid space.

( Courtesy of Michael Abraham, MD, MS, Baltimore, MD.)

A study of nearly 3000 patients by Perry and colleagues found a sensitivity of 100% for noncontrast head CT performed within 6 hours in patients with the worst ever headache, when the CT was performed on a third-generation, or newer, scanner and interpreted by a neuroradiologist. A subsequent study of 137 patients in 2012 found a sensitivity of 98.5% for noncontrast head CT performed within 6 hours, though the one miss was due to a bleeding cervical arteriovenous malformation that presented without headache. With exclusion of this patient, sensitivity for SAH was 100%. Therefore, it is reasonable to assume that CT sensitivity approaches 100% with at least third-generation scanners when an experienced radiologist interprets a scan obtained within 6 hours of headache onset. A negative head CT noncontrast within 6 hours of onset possesses a negative LR of 0.01. Beyond 6 hours the sensitivity is approximately 95%, which decreases with time to less than 90% as the time from headache onset approaches 24 hours. One important aspect is the CT should be interpreted by an experienced radiologist. With experienced radiologist interpretation, a systematic review and meta-analysis finds overall sensitivity more than 99.0%, with specificity of 99.9%, for CT obtained within 6 hours of headache onset. The pooled LR of SAH with negative CT within 6 hours is 0.010. If the scanner is an older generation and an experienced radiologist is not available, further evaluation may be considered.

Limitations of CT include anemia (sensitivity decreases when the hematocrit is <30%), smaller hemorrhage volume, CT quality, radiologist experience, and image artifacts. High-quality CT within 6 hours of headache onset interpreted by an experienced radiologist provides a risk of missing SAH with negative CT of less than 1.0% and a negative LR of 0.01.

Computed tomography angiography

The AHA’s guidelines give a level C recommendation for CTA to follow nondiagnostic noncontrast head CT. CTA can rapidly identify an aneurysm, classically with a sensitivity of 77% to 100% and specificity of 87% to 100%. A recent meta-analysis of CTA found a pooled sensitivity of 98% (95% CI: 97%–99%) and pooled specificity of 100% (95% CI: 97%–100%) for the detection of aneurysm. However, sensitivity is significantly lower for aneurysms less than 3 to 4 mm, though most aneurysms that rupture are more than 5 mm. Notably, approximately 0.6% to 5.0% of the general population will have an aneurysm on CTA, with most being asymptomatic.

Potential advantages of CTA over noncontrast head CT followed by LP include patient comfort and improved diagnostic ability to detect aneurysm. A negative noncontrast head CT and negative CTA indicate a relatively benign clinical course. Patients with negative noncontrast head CT and negative CTA have a post-test probability for aneurysmal SAH of less than 0.3%, and this combination of tests has a negative predictive value more than 99%. The potential problem with the CT/CTA approach without LP is that visualizing an aneurysm on CTA does not confirm the diagnosis of aneurysmal SAH. There is a possibility that the aneurysm is an incidental finding that will prompt unnecessary intervention, placing patients at increased risk of complications. However, patients with aneurysm and headache may be at an 8-fold higher rate of aneurysm rupture when compared with asymptomatic patients with an aneurysm. Complication rates of CTA range from 0.25% to 1.8% and include nephrotoxicity, increased radiation exposure, and allergic reaction.

CTA can provide a benefit in patients whereby LP would be difficult or not feasible (obesity, inability to cooperate, decline LP). The test will consistently show aneurysms greater than 3 to 4 mm in size, if present.

MRI/magnetic resonance angiography

A recent meta-analysis of MRI with MRA demonstrated a pooled sensitivity of 95% (95% CI: 89%–98%) and pooled specificity of 89% (95% CI: 80%–95%) for identification of cerebral aneurysms. This modality is also effective in subacute (3 days after headache onset) or chronic SAH. MRA in conjunction with MRI can diagnose aneurysms greater than 3 mm in size with more than 95% sensitivity. MRI may diagnose other conditions, such as neoplasm, multiple sclerosis, posterior reversible encephalopathy syndrome, and encephalitis. This modality does not require radiation, though several limitations exist, including limited availability in the ED, the time required for scanning, the potential for inducing claustrophobia, and the need for specialist interpretation. False-negative and false-positive aneurysms detected on MRA are often located at the skull base or the middle cerebral artery. MRI/MRA is optimal for patients who present in a subacute or chronic timeframe.

Lumbar puncture

Traditionally, LP followed nondiagnostic head CT in cases of suspected SAH. LP after negative CT is a level B recommendation by the ACEP. Of note, the ACEP’s guidelines predate the work of Perry and colleagues on the sensitivity of CT obtained within 6 hours of headache onset. However, many patients fear LP because of its painful and invasive nature as well as the risk of post-LP headache, which approaches 30%. Other issues include the time and potential difficulty of performing the test as well as the yield, which may be complicated by a traumatic tap. Contraindications include bleeding disorder or coagulopathy and increased ICP. Interestingly, less than half of patients with negative CT and acute headache undergo LP; in these studies, less than 1% of LPs are true positive when a third-generation scanner is used.

LP can add important clinical information for other diagnoses, including meningitis, spontaneous intracranial hypotension, and idiopathic intracranial hypertension. Unfortunately, LP cannot diagnose pituitary apoplexy, cerebral venous sinus thrombosis, arterial dissection, or unruptured aneurysm. Brunell and colleagues found that LP provides an alternative diagnosis in 3% of cases, though findings altered management in less than 0.5%. More than 250 LPs are required to diagnose one additional SAH missed by CT per Carpenter and colleagues’ meta-analysis.

There are several controversies in the interpretation of CSF results. Traumatic LP may occur in 15% of cases. Classically in traumatic LP, red blood cell (RBC) clearing is seen when tubes 1 and 4 are compared, with no RBCs in tube 4. However, the complete absence of RBCs in tube 4 is rare, resulting in difficulty with interpretation. Using the CSF RBC count to distinguish SAH and traumatic LP can be complex. Perry and colleagues used a threshold of 2000 × 10 6 /L in the final tube and found a sensitivity of 93% (95% CI: 66%–98%) and specificity of 93% (95% CI: 91%–95%) for aneurysmal SAH. Czuczman and colleagues found an LR for the diagnosis of SAH of 0 (95% CI: 0–0.3) with an RBC count less than 100 in the final tube and 1.6 (95% CI: 1.1–2.3) with an RBC count less than 10,000 × 10 6 /L. Combining data from these studies and a threshold of 1000 × 10 6 /L demonstrates a pooled sensitivity of 76% and specificity of 88%.


Xanthochromia is due to the in vivo breakdown of hemoglobin by normal enzymatic action, creating a yellow color in CSF. Appearance of xanthochromia takes several hours after bleeding begins (20% of patients at 6 hours) and lasts for up to 2 weeks. The sensitivity for the diagnosis of SAH approaches 90% at 12 hours after symptom onset, though classically it was thought to be 100%.

Two methods of assessment are available. The first is visual inspection, whereby the CSF is compared with water against a white background. Visible inspection possesses a pooled sensitivity of 85%, with specificity 97% based on meta-analysis. Multiwavelength spectrophotometry has poor specificity (29%–75%) but sensitivity approximately more than 95%; however, it is available in approximately 1% of US EDs. Xanthochromia in the setting of SAH greatly reduces the likelihood of traumatic LP, with a negative LR of 0.22. Physicians should consider that only 20% of those receiving an LP within 6 hours of headache onset will have positive xanthochromia.

Combination of cerebrospinal fluid, red blood cell, and xanthochromia

The use of CSF RBC count and xanthochromia together has been advocated. One prospective cohort demonstrated that the combination of absence of visual xanthochromia with less than 2000 RBCs may rule out SAH. A cutoff of 2000 RBCs has a sensitivity of 93% alone. When xanthochromia is added to this, the sensitivity was 100% (95% CI: 75%–100%), though this finding requires further validation in other populations.

Opening pressure

CSF pressure can provide valuable information in the setting of suspected SAH. Pressures greater than 20 cm H 2 O are defined as elevated, which can be found in approximately 60% of patients with SAH. Elevated pressures may also be found in cerebral venous thrombosis, idiopathic intracranial hypertension, and meningitis. Low pressure may be seen with spontaneous intracranial hypotension. Of note, pressure measurements must be taken with patients in the lateral recumbent position.

Test threshold

The test threshold is the pretest probability of disease that balances the risks of missing a diagnosis with the harms imposed by (1) the diagnostic strategy itself and (2) the treatment of those with a false-positive diagnosis. In the setting of SAH, potential evaluation and treatment entail angiography, CT, LP, and neurosurgical intervention. A meta-analysis by Carpenter and colleagues used pooled estimates of diagnostic accuracy, risks, and benefits to estimate the test threshold for the performance of LP after a negative head CT. Investigators suggest LP benefits patients with negative CT if the pre-LP probability of SAH is 5% or if the pre-CT probability of SAH is 20%. Per this meta-analysis, a 1 in 10 pre-CT probability of SAH, while assuming CT sensitivity of 95%, results in a 1 in 180 chance of missing a ruptured aneurysm.


Management involves several key steps, shown later. Neurosurgical consultation is required to arrange for definitive therapy. ED management must focus on management of the airway, hemodynamic monitoring, supportive care, and management and prevention of complications, demonstrated in Box 1 . A multidisciplinary team specialized in the care of these patients can improve outcomes, and patients may benefit from transfer to a specialized center for further care.

Box 1

  • Closely evaluate for need for airway protection/endotracheal intubation

  • Monitor for signs of increased ICP: decline in neurologic status, posturing, altered mental status

  • Treat pain and anxiety: short-acting IV analgesics, such as fentanyl

  • Treat nausea/vomiting

  • Evaluate and monitor closely for complications, including decline in mental status, herniation, seizure, ECG changes, pulmonary edema

  • BP management should target SBP less than 160 mm Hg or MAP less than 110 mm Hg

  • Maintain normothermia; avoid fever

  • Correct any coagulopathy

  • Treat any seizure

  • Provide nimodipine to decrease vasospasm risk

Abbreviations: BP, blood pressure; IV, intravenous; MAP, mean arterial pressure; systolic blood pressure.

Subarachnoid hemorrhage management considerations

Initial resuscitation

Patients with SAH, traumatic and nontraumatic, are at risk for severe complications resulting in hemodynamic compromise and neurologic decompensation. Pulmonary edema and dysrhythmia may occur in 23% and 35% of SAH, respectively, within the first 24 hours of admission. Neurologic decompensation occurs in up to 35% of patients within 24 hours. Most interventions target initial stabilization and avoidance of these complications. Airway intervention may be needed for airway protection or anticipated clinical decompensation. Indications for intubation include Glasgow Coma Scale (GCS) less than 8, signs of elevated ICP (posturing), impaired oxygenation or ventilation, and need for sedation and paralysis. The most important prognostic factors for SAH include level of consciousness at the time of hospital admission, age, and amount of blood on initial head CT. Several physiologic derangements that are common and may worsen brain injury and increase mortality include serum glucose greater than180 mg/dL, troponin elevation, fever (>100.4°F), acidosis with serum bicarbonate less than 20 mmol/L, hypoxemia with arterio-alveolar gradient greater than 125 mm Hg, blood pressure instability (mean arterial pressure [MAP] <70 or >130 mm Hg), and hypothalamic pituitary dysfunction. There are several grading scales that assess the severity of SAH; as each score increases, mortality increases. These scales are shown in Box 2 .

Dec 1, 2017 | Posted by in Uncategorized | Comments Off on Subarachnoid Hemorrhage
Premium Wordpress Themes by UFO Themes