Perry JJ, Stiell IG, Sivilotti ML, et al. High-risk clinical characteristics for subarachnoid haemorrhage in patients with acute headache: prospective cohort study. BMJ. 2010;341:c5204

Resident Reviewer

Carl Mickman

Why this study is important

Subarachnoid hemorrhage (SAH) is one of the most dangerous diagnoses made in the emergency department, and a large amount of resources are spent imaging patients to rule out this condition. Until this study, no clinical rules had been developed to assist providers in determining which patients should be imaged, and which patients could be sent home with return precautions.

What we already know about the topic

Headaches account for 2% of all emergency department visits, though SAH only accounts for between 1-3% of these headaches. The presence of neurologic defects or changes in mental status makes some of these diagnoses relatively easy, but a substantial amount of patients with SAH do not present with such obvious symptoms. It is up to providers to determine which of these patients should receive imaging.

Brief overview of the study

1999 patients were enrolled in a multicenter study in Ontario that investigated neurologically intact patients with a non-traumatic headache that peaked within one hour. Of these patients, 130 were identified to have SAH by either CT or lumbar puncture. Patients were excluded from the study if they had had three or more headaches of the same character in the last six months. Patients enrolled in the study were asked about characteristics of their headaches and statistical analysis was done to determine which characteristics were predictive of SAH. Three proposed rules, each containing four different variables were tested (several of the variables overlapped between the rules; seven total variables were found to be predictive), and found to be 100% sensitive. No subarachnoid hemorrhages were later found in patients who received no advanced testing, and the rules were found to be between 28.4% and 38.8% specific.


As many as 1050 patients may have met study criteria after chart review in these emergency departments but were not enrolled in the study (however these patients were contacted at 1 month and 6 months with 86% and 80% respectively of these patients reporting no adverse outcomes). Additionally, the initial exclusion criteria may have prevented the identification of subarachnoid hemorrhages in patients who suffered from chronic headaches. Only 105 patients out of the 1999 surveyed had two physicians independent of each other fill out the symptom survey, and results did not always correlate. Notably, there was poor correlation between the patient reported time from headache onset to peak of headache intensity, an initial exclusion parameter. The study also pointed out that one of their inclusion criteria, arrival by ambulance, was subject to geographical bias, as many areas will not have the same access to EMS services.

Take home points

This study demonstrated that in neurologically intact patients with a non-traumatic headache that peaked within one hour, if they do not possess at least one of the following symptoms, they do not warrant imaging. This clinical decision rule has become known as the Ottawa Subarachnoid Hemorrhage Rule.

  1. Age ≥40 y
  2. Neck pain or stiffness
  3. Witnessed loss of consciousness
  4. Onset during exertion
  5. Thunderclap headache (instantly peaking pain)
  6. Limited neck flexion on examination

This data was later independently verified in a 2013 study, and the seventh variable mentioned in this study, vomiting, was excluded.