EBQ:Perry Subarachnoid Haemorrhage Study

From WikEM
Jump to: navigation, search
Under Review Journal Club Article
Perry JJ et al. "High risk clinical characteristics for subarachnoid haemorrhage in patients with acute headache: prospective cohort study". BMJ. 2010. 28(341):c5204.
PubMed Full text PDF

Clinical Question

What is the sensitivity of non-contrast head CT for detecting spontaneous subarachnoid hemorrhage (SAH), when performed on a third-generation CT scanner within 6 hours of headache onset?


Third-generation non-contrast head CT is highly sensitive for detecting spontaneous SAH when performed within 6 hours of headache onset and interpreted by a radiologist experienced in reading head CT.

Major Points

Within 6 hours of headache onset, CT had:

Sens = 100% (95% confidence interval: 97%-100%)

Spec = 100% (99.5%-100%)

For all patients studied, CT had:

Sens = 92.9% (89%-95.5%)

Spec = 100% (99.9%-100%)


ED patients in 11 tertiary care centers

Patient Demographics

  • Mean age = 45
  • 60% women

Inclusion Criteria

  • >15 years old
  • Acute headache reaching peak intensity within one hour
  • Normal neurologic exam
  • CT ordered by the treating physician to rule out SAH

Exclusion Criteria

  • Focal neurologic deficits
  • Papilledema
  • History of SAH
  • History aneurysm
  • Previous VP shunt
  • Brain neoplasm
  • Onset of headache >14 days ago
  • Recurrent headache (≥3 similar)
  • Transfer from outside hospital with confirmed diagnosis of SAH


Treating physicians worked up study subjects for SAH and initiated treatment per their usual clinical practice.


Primary Outcomes

  • To diagnose SAH based on CT, xanthochromia in CSF, or RBCs >5x106/L in final tube of CSF collected and aneurysm identified on cerebral angiography (digital subtraction, CT, or MR angiography)

Secondary Outcomes


Subgroup analysis

Subjects who underwent head CT within 6 hours of headache onset

  • Increased sensitivity for head CT (see above) when compared to study group as a whole

Subjects who did not undergo lumbar puncture vs those who did

  • No significant difference in sex, loss of consciousness, arrival by ambulance, exertional onset, vomiting, BP, HR
  • Higher mean age: 47.1 vs 43
  • Shorter time to headache maximum intensity
  • Higher incidence of neck pain
  • Higher incidence of "worst headache ever"

Criticisms and Further Discussion

  • Not all study subjects underwent a gold standard lumbar puncture, possibly resulting in an underestimation of the rate of false-negative CT. However, investigators endeavored to follow up all subjects at 6 months using medical records and phone calls; no patients contacted received a subsequent diagnosis of SAH. Only 50 patients could not be followed up by these methods. For these 50, review of referrals to neurosurgical centers and coroner reports failed to suggest that any had had a missed SAH. The sensitivity and specificity should then be viewed as for "clinically significant SAH" as non-clinically significant SAH could have been missed.
  • Meta-analysis, heavily weighted from this paper provides a negative head CT in a neurologically normal patient, performed within 6 hours of onset provides a miss rate of SAH of 2/1000 patients.[1]


  • Canadian Institutes for Health Research (grants 67107, 153742)
  • The Ontario Ministry of Health and Long Term Care
  • The Physicians of Ontario through the Physician’s Services Incorporated Foundation (01-39)


  1. Dubosh NM et al. Sensitivity of Early Brain Computed Tomography to Exclude Aneurysmal Subarachnoid Hemorrhage: A Systematic Review and Meta-Analysis. Stroke 2016. PMID: 26797666