Subarachnoid Haemorrhage – The ED perspective.

It is a busy Thursday evening shift and you are called to the front desk to see a 34 year old female complaining of the ‘worst’ headache of her life. The headache hit her like a ‘bolt out of the blue’ and has been unremitting ever since…

Epidemiology

• Incidence about 7 per 100000.
  • Perth = 1.5 million.
  • About 2 per week in Metro area.
• 50% overall case fatality.
• 15% die prehospital.
• 33% of survivors need long-term care.
• 50% of patients <55 years old.
  • Despite being much rarer, loss of productive life similar to CVA in terms of quality-adjusted life years (QALY)

Aetiology

• 85% Ruptured aneurysms
• 10% Perimesencephalic Haemorrhages
• 5% Other – AV malformations, inflammatory, Cocaine.

Pathophysiology

• Aneurysms
  • NOT Congenital
  • Occur in about 1 in 40 people (1-5%).
  • Most never rupture.
  • Arise at sites of arterial branching.
  • Multiple aneurysms in 30%.
  • 90% = Saccular = ‘Berry’ anurysms
  • 10% = Fusiform – usually in vertebrobasilar system and present with Cranial nerve or brainstem symptoms due to pressure effects rather than haemorrhage.
  • Aneurysm risk factors:
    • Familial – first degree relatives have 5 times the risk. 2nd degree have same risk as general population.
    • Smoking
    • Hypertension
    • Heavy drinking
    • 1.6 F : 1 M
    • ?OCP

• Perimesencehalic Haemorrhage
  • Definition = haemorrhage restricted to the cisterns about the brainstem and suprasellar cistern AND a negative cerebral angiogram.
  • Has a much better prognosis than standard SAH with a much lower rate of rebleeding or vasospasm
  • 1 out of 29 patients rebled and died in one retrospective study.
  • Has a presumed venous aetiology but some neurosurgeons are sceptical of this as an entity and advocate a repeat of the angiogram.

Aneurysm Locations

Clinical Features

Headache:

• Classically presents with what’s known as a ‘thunderclap’ headache.
  • In reality only about 1 in 10 people who present to an Emergency department with a thunderclap headache will have had a SAH.
• The onset may not always be instantaneous. One study interviewed people with proven SAH who could still be interviewed and they found that the onset was:
  • Instantaneous – 50%
  • 2-60 seconds – 24%
  • 1-5 minutes – 19%
  • No idea/unable to remember – 9%
• As to how short in duration a headache can be and still be a SAH, no-one knows, however an arbitrary time of 1 hour has been suggested. The typical duration is of the order of 1-2 weeks.
• A ‘sentinel bleed’ is essentially a subarachnoid haemorrhage that the patient did not seek medical attention for, or one missed by a doctor. As a concept it has little or no use in the decision-making process in the emergency department.

Other Features:

• Vomiting is not predictive.
• Seizure at onset is.
• 2/3rds have a reduced level of consciousness.
• Neck stiffness may develop – but usually only after several hours and is due to an inflammatory reaction to the blood in the subarachnoid space, and it may not develop at all if there’s only a small amount of blood.
• 3rd nerve palsy due to an aneurysm in the posterior communicating artery.
• 1 in 7 will have intraocular haemorrhages.
• Ischaemic changes (of any type) on ECGare common
  • Possibly due to a catecholamine surge or a change in autonomic vascular tone.
• 3% will have a cardiac arrest
  • Agressive resuscitation is essential as they appear to have a high rate of ROSC (return of spontaneous circulation) and half of the survivors will regain independent living.

Investigations

The most widely accepted approach to the investigation of thunderclap headaches is a combination of CT, followed by a lumbar puncture (LP) 12 hours after onset of headache if the CT is negative.

• CT Scan
  • Modality of choice.
  • Sensitivities widely quoted in literature based on studies at least 13 years old.
    • 98% sensitivity if scanned within 12 hours of onset.
    • 93% within 24 hours.
    • 50% at 5 days – based on study from 1984.
  • Sensitivities may never improve even with improved CT quality for 2 reasons:
    • Small bleeds will be flushed away by normal CSF flow.
    • With small bleeds there may not be enough blood present to appear as hyperdense to CSF on scanning (a haematocrit of 27% is quoted as being required before a bleed will show on CT)

• MRI Scan
  • Appears comparable to CT in acute phase
  • Small studies hint it may even be better.
  • May help localise ‘CT negative, LP positive’ patients
  • May pick up pathologies not detected by CT
    • Eg: Cerebral venous sinus thrombosis (CVT), Parenchymal lesions.

• Lumbar Puncture
  • Still used in combination with CT as ‘gold standard’ in rule-out of subarachnoid haemorrhage in patients presenting to emergency departments with thunderclap headaches.
  • We (think we) know only 1 in 10 of patients presenting with thunderclap headaches will have had a subarachnoid haemorrhage.
  • A non-contrast CT within 12 hours of onset of pain is 98% sensitive for detecting SAH.
    • Therefore the risk of missing a SAH may be as low as 2 in 1000 if the patient has had a negative CT in this timeframe.
    • We accept this rate in other potentially lethal conditions (Acute coronary syndrome, pulmonary embolism)
    • Many patients will choose to accept this risk if informed consent for a LP is sought.

• Lumbar Puncture Analysis
  • Opening Pressures(normal = 8-18cm)
    • Often omitted but may provide the only clue of an alternate diagnosis eg: CVT, idiopathic intracranial hypertension. This is free information and should be performed if at all possible.
    • To get a true opening pressure reading the patient will have to be lying on their side, with their legs extended – once CSF is draining from your pencil-point LP needle ask the patient to straighten their legs out. Waiting for the CSF to rise up the manometer can be painfully slow, especially if you’re being kind and using a very fine needle .
    • Handy Hint: Hold the manometer horizontally until a reasonable amount of CSF has collected and then move the manometer so that it’s vertical you’ll get the pressure reading much more quickly.
  • Microscopy– cell counts, gram staining
    • Some centres advocate performing the LP within 12 hours of onset in a bid to diagnose SAH earlier with 3 tubes being collected for microscopy.
    • If there is no significant drop in the red cell count between the 1st + 3rd tubes, SAH is diagnosed.
    • This leads to a significant false positive rate due to bloody taps.
    • No formula to differentiate a true SAH from a bloody tap has been validated
    • Protein + Glucose (matched with concurrent serum glucose sample).
    • Culture – Pneumococcal meningitis can present with a thunderclap headache.
  • Xanthochromia
    • A fancy way of saying ‘Yellow colour’! - Bilirubin is what you’re looking for.
    • 96% sensitive if done 12 hours post headache onset.
    • Positive for at least 2 weeks (possibly up to 4 weeks)
    • 2 things cause CSF to become yellow:
      • Oxyhaemoglobin
        • Forms within 2-4 hours – both in vivo + in vitro.
        • Converted to Bilirubin over 9-15 hours.
        • Common in traumatic taps.
        • A ‘bloody’ tap analysed by the lab within 6 hours of LP may have lots of oxyhaemoglobin but should have NO Bilirubin.
      • Bilirubin
        • Only formed in vivo.
        • Degraded by UV light – ideally sample wrapped in foil until analysis is performed.
        • A lot of centres just inspect CSF visually against a white background to ‘detect’ xanthochromia
    • Spectrophotometry differentiates between a bloody tap and ‘true’ xanthochromia.
      • Pure oxyhaemoglobin will have a single ‘up and down’ peak, whereas a sample with bilrubin will have what’s referred to as the ‘bilrubin shoulder’

The Bilirubin shoulder

Management

• Supportive
  • ABCDEFG…
  • Analgesia.
  • BP control – SBP 90-140 prior to intervention.
  • Normothermia.
  • Prophylactic anti-emetics.
    • Vomiting potentially catastrophic.
  • Quiet environment with minimal visitors.
  • DVT prophylaxis
    • Stockings and pneumatic leg compression prior to intervention
    • Heparin after.
  • Prophylactic laxatives and good oral fluid intake if swallow OK (‘Straining at stool’ is probably not a good idea!)
• Specific
  • May require haematoma evacuation.
  • Prevent Rebleeding:
    • Coiling superior to clipping.
    • Not always technically possible.
  • Prevent delayed cerebral ischaemia
    • Nimodipine 60mg 4hourly x 3 weeks
    • Orally or via NG
    • Numbers needed to treat to prevent one bad outcome = 20-ish.
    • IV route unproven and may even be dangerous.
  • Seizure prophylaxis (advocated by some)
  • Antifibrinolytics, heparins and antiplatelet agents all have been proposed but remain controversial rather than definitively proven.

Complications

• Rebleeding.
  • 1 in 6 in 1st few hours resulting in 80% death/disability rate.
  • Cumulatively 40% will rebleed if untreated within 4 weeks.
• Vasospasm.
  • More accurately referred to as delayed cerebral ischaemia.
  • Typically occurs 5-14 days post bleed.
  • Incidence is related to amount of blood in CSF.
  • May present with hemiplegia or simply a reduction in level of consciousmess.
• Other
  • Seizures
  • Hydrocephalus
  • Pulmonary oedema – may be cardiogenic or neurogenic.
  • Myocardial injury
  • Hyponatraemia

Prognosis

There are 2 scales in widespread use , despite a lack of validation:

Hunt – Hess Scale + Survival

World Federation of Neurological Surgeons scale:

For more tips on the management of the sick subarachnoid haemorrhage patient, listen to Scott Weingart’s take on things: EMCrit Podcast 8 – Subarachnoid Hemorrhage.