Blood in the CSF spaces is subtle and dependent — search the basal cisterns, the sulci, the sylvian fissures and the occipital horns deliberately.
Orient first
- Acute subarachnoid blood is hyperdense relative to CSF but only by a modest margin, and it sits in spaces that are normally black. You are looking for grey where black should be.
- CT sensitivity is highest in the first 6 hours and FALLS steadily thereafter as blood is cleared and becomes isodense to CSF. Time from onset changes what a negative scan means.
- The pattern suggests the cause: diffuse basal cistern blood suggests an aneurysm; perimesencephalic blood suggests a benign non-aneurysmal bleed; convexity blood suggests amyloid angiopathy, a cortical vein thrombosis or trauma.
Acquire the study
- Non-contrast CT with thin slices. Thick slices average blood away.
- Review on a SUBDURAL/intermediate window as well as the standard brain window — a slightly wider window makes thin layers against bone conspicuous.
- Look at the dependent parts: with the patient supine, blood settles in the occipital horns and the interpeduncular fossa.
The manoeuvre
- Work through the CSF spaces in a fixed order: suprasellar cistern, interpeduncular fossa, ambient and quadrigeminal cisterns, sylvian fissures bilaterally, interhemispheric fissure, cortical sulci, then the ventricles.
- Look specifically at the OCCIPITAL HORNS for a fluid–fluid level of layered blood — often the only sign of a small bleed.
- Compare the two sylvian fissures with each other; asymmetric density is the finding.
- Assess for hydrocephalus by measuring the temporal horns and the third ventricle — acute hydrocephalus changes management immediately.
- Look for an associated parenchymal haematoma, which localises the aneurysm.
- Note the PATTERN and say what it suggests.
- Look for other causes of the headache: venous sinus thrombosis (check the dense sinus), pituitary apoplexy, dissection.
What confirms it
- Hyperdense material in the subarachnoid spaces conforming to the shape of cisterns or sulci rather than to a parenchymal territory.
What licenses you to exclude it
- ⚠️ A NEGATIVE CT DOES NOT EXCLUDE SUBARACHNOID HAEMORRHAGE, and how much it lowers the probability depends heavily on the time from onset. Beyond about 6 hours the report must say so and lumbar puncture or further imaging remains indicated.
- State the time from ictus in the report where it is known — it is part of the interpretation, not background.
The classic misread
- Interpreting a negative scan at 24 hours as an exclusion.
- Missing thin convexity blood by reading only on the standard window.
- Mistaking generalised cerebral oedema with dense vessels (pseudo-subarachnoid haemorrhage) for real blood.
- Not reporting hydrocephalus, which is the treatable finding.