Impact of subarachnoid haemorrhage Canadian clinical decision rules for investigation of acute headache: a retrospective case note review

Abstract

BackgroundPhysicians often face a dilemma to investigate alert and neurologically intact patients presenting with acute headache to emergency departments. Two Canadian prospective cohort studies evaluated high-risk clinical characteristics for subarachnoid haemorrhage in such patients and proposed clinical decision rules (Canadian rules 1, 2, 3, and Ottawa) for investigation of acute headache. We aimed to determine investigation rates for subarachnoid haemorrhage in a cohort of neurologically intact patients presenting with acute headache. MethodsWe performed a retrospective case note review of alert and neurologically intact patients presenting with acute headache to Aintree University Hospital, Liverpool, UK, between Jan 1 and March 1, 2013. The case notes of these patients were independently reviewed by two investigators to determine clinical characteristics. Criteria for inclusion were: age over 18 years, fully alert (Glasgow coma scale 15), new acute headache with no recent history of head injury (14 days), absence of focal neurological deficit or papilloedema, and absence of known cerebral aneurysm, brain neoplasm, or hydrocephalus. Canadian rules were applied retrospectively to determine the specificity, sensitivity, and negative predictive values for subarachnoid haemorrhage. Two-tailed Fisher's exact test and McNemar's test were used to determine differences. Findings403 patients presented with acute headache, of whom 162 patients satisfied the criteria for inclusion. In three (1·9%) of these 162 patients subarachnoid haemorrhage was diagnosed by CT, 11 (6·8%) had a final diagnosis of other cerebral disease, and 148 (91·3%) were diagnosed with benign causes of headaches. 69 patients (42·6%) had unenhanced CT, 28 (17·3%) had a lumbar puncture, and 25 (15·4%) had both investigations. There were no re-admissions with a subarachnoid haemorrhage for patients not fully investigated in our practice. Retrospective application of Canadian rules 1, 2, 3, and Ottawa to our cohort would have increased CT investigation rates to 54·3%, 64·8%, 50·0%, and 61·7%, respectively, compared with 42·6% in our practice (p<0·0001). If rule 3 was applied, one patient who had suffered a subarachnoid haemorrhage would have been classifed as low risk and not investigated, leading to a missed diagnosis. InterpretationThe rates of subarachnoid haemorrhage in our study were lower than those in the Canadian studies. In our smaller group of neurologically intact patients, retrospective application of the Canadian rules would have led to a significantly higher investigation rate and longer hospital stay. This study also highlights the need for further large scale prospective studies in the UK before adopting rigid decision rules for investigation of acute headache in neurologically intact patients in the National Health Service. FundingNone.