Background Acute headache is a common reason for presentation to emergency departments. Some have significant structural pathology requiring further intervention. Emergency clinicians often rely on presenting headache features (such as thunderclap onset) to guide the need for neuroimaging and further investigation. It is unclear whether these features discriminate accurately or how the investigations of patients presenting with thunderclap headache differs internationally.

Objectives To determine the proportion of patients presenting with thunderclap onset of headache from a general headache cohort and compare demographics, investigation strategy and final diagnosis, across an international sample of patients.

Methods An international, multicentre, observational prospective cohort study. This planned sub-study focussed on patients presenting with thunderclap onset headache, with characteristics compared to the general headache cohort. The prospective observational design was chosen to capture real-world data on current international practice.

Results The study recruited 4536 patients across 67 hospitals and 10 countries during 2019. Of this, 644 patients presented with thunderclap headache onset (14.2%). Median age was 44. The majority of patients self-referred to hospital. CT brain imaging was performed in 62.7% cases and lumbar puncture in 10.6%, with wide international variation. New Zealand reported the highest rate of neuroimaging, 78.4% of patients presenting with thunderclap headache, compared to 25.0% in Romania. All cases of subarachnoid haemorrhage (SAH) were diagnosed on CT imaging results.

When compared with the parent cohort of all headache patients presenting to the ED, those with thunderclap headache had a significantly higher rate of serious cranial pathology (13.7% vs 8.5%, p<0.001) and final diagnosis of SAH (3.6% vs 0.8% p<0.001).

Conclusions Thunderclap headache presenting to the ED appears to correlate with a higher risk for serious intracranial pathology and/or SAH. Investigation strategies varied within this international cohort. Neuroimaging rates did not align with international guidelines, suggesting potential for further work on standardisation.