Subarachnoid haemorrhage

Editor – We read with interest the comprehensive article ‘Assessment of acute headache in adults – what the general physician needs to know’.¹

The authors highlight that a minority of thunderclap headaches are secondary to a non-traumatic subarachnoid haemorrhage (SAH) and refer to a study showing that in routine practice up to 20% of SAH cases may be missed by computed tomography (CT) scan, without stipulating scanning parameters and then recommending cerebrospinal fluid (CSF) xanthochromia.^1,2 In this retrospective study the diagnosis of SAH was made on clinical grounds plus the presence of CSF red blood cells.² Only four of the 11 cases missed by CT scan were imaged by modern 64 slice technology, in none of these was xanthochromia detected and only two had a vascular cause for SAH identified by angiography. The authors acknowledge that the high prevalence of aneurysms means that aneurysm identification may be incidental if the CT has not confirmed a recent bleed. Additionally, the presence of xanthochromia in the CSF is not specific to SAH.

A systematic review and meta-analysis concluded that the pretest probability for SAH has to be high to make lumbar puncture worthwhile,³ estimating the number of patients needing a lumbar puncture to identify one aneurysmal SAH amenable to treatment at anything from 250 to infinity. Though the risks are small, lumbar puncture is associated with minor morbidity (eg headache) and major morbidity (eg CSF infection, herniation). Samples for xanthochromia analysis are commonly rejected by laboratories (eg not protected from light) or uninterpretable (eg fresh blood from a traumatic tap interfering with bilirubin detection) and hospital admission is prolonged whilst results are awaited. We would therefore like to highlight the limited role of lumbar puncture over CT alone, dependant on scanning parameters, in the exclusion of SAH. However as the authors state, although aneurysmal SAH is a rare cause of headache, lumbar puncture may enable other diagnoses. Consider carefully the requirement for lumbar puncture to rule out SAH in those having been imaged by a third generation or above CT scanner with appropriately sensitive settings with a low pretest probability of a subarachnoid haemorrhage.