Aims/Objectives/Background Acute headache accounts for around 2% of Emergency Department attendances. Headache guidelines recommend non-contrast head computed tomography (CT) followed by lumbar puncture to exclude subarachnoid haemorrhage (SAH). Advances in imaging technology have led emergency physicians to question the necessity of routine lumbar puncture after negative CT. This systematic review assessed diagnostic strategies for neurologically intact headache patients.

Methods/Design In February 2020, 18 electronic databases (including MEDLINE and Embase) were searched for studies of any clinical decision rule or diagnostic test for assessing neurologically intact severe headache patients, reaching maximum intensity within an hour. Studies were quality assessed using the QUADAS-2 tool. Diagnostic accuracy data were extracted into 2x2 tables to calculate sensitivity, specificity, false-positive and false-negative rates. Where appropriate, hierarchical bivariate meta-analysis was used to synthesise results.

Results/Conclusions Thirty-seven studies were included. Eight studies assessing the accuracy of the Ottawa SAH clinical decision rule were pooled; sensitivity was 99.5%, specificity was 23.7%. Four studies (with neuroradiology expertise) assessing CT within six hours of headache onset were pooled; sensitivity was 98.7%, specificity was 100%. CT sensitivity beyond six hours was considerably lower (≤90%; 2 studies). Three studies assessing lumbar puncture (spectrophotometric analysis) following negative CT were pooled; sensitivity was 100%, specificity was 95.2%. LP-related adverse events were reported in 5.3–9.5% patients (2 studies).

The evidence suggests that the Ottawa Rule has limited value for ruling out SAH; the high false positive rate means that its use would potentially result in 76% SAH-negative patients undergoing further investigation with no additional benefit. Modern CT within six hours of headache onset (with images assessed by a neuroradiologist) is highly accurate and likely to be sufficient to rule out SAH. However, sensitivity reduces considerably over time. The CT-LP pathway remains a highly sensitive pathway for detecting SAH, although LP resulted in some false-positives and adverse events.