Sensitivity of proposed clinical decision rules for subarachnoid haemorrhage: An external validation study

Subarachnoid haemorrhage (SAH) is an uncommon but important cause of sudden-onset headache. Three clinical decision rules (CDRs) for investigation in sudden headache have been proposed, but concerns were raised about the generalisability of some variables. Our aim was to determine what proportion of patients with confirmed SAH has the identified high-risk factors and the sensitivity of the proposed CDR in an Australasian cohort.

This is a retrospective cohort study of alert and neurologically intact adult patients with confirmed SAH attending two community teaching hospitals between 2000 and 2011. The outcomes of interest were the proportion of patients with each high-risk criterion (descriptive statistics) and sensitivity of each proposed CDR (%, interquartile range [IQR]).

There were 59 confirmed SAH that met the inclusion criteria. Sensitivity of proposed CDR 1 was 96.6% (95% confidence interval [CI] 88.5-99.1%), sensitivity of proposed CDR 2 was 100% (95% CI 93.9-100%) and sensitivity of proposed CDR 3 was 89.8% (95% CI 79.5-95.3%). The addition of vomiting to the criteria in CDRs 1 and 3 increased the sensitivity of both these CDRs to 100%.

CDR 2, or the refinement of CDRs 1 and 3 with the inclusion of at least one episode of vomiting as a criterion, has very high sensitivity. Although unlikely to reduce CT scan rates for patients in whom there is a clinical suspicion of SAH, they might be useful in guiding which patients require further testing (e.g. lumbar puncture) after a negative CT scan.