This site is intended for health professionals only

Tricky 10 minutes – ‘I want a brain scan for these headaches’

Headache is among the commonest symptoms in primary care. Headache comprises all forms of pain above the shoulders, including muzziness, tightness and pressure in the head, and facial pain. One in 1,000 of these patients in primary care have a tumour.1

It is tempting to scan everybody – a brain scan is more likely to reveal an incidental finding than a cause of the headache. A non-specific imaging policy fuels belief that only a scan can exclude a serious cause of head pain, so those with chronic symptoms re-present for further imaging. If every patient with headache is added to the routine scan waiting list then a minority will die of treatable pathology while waiting.


Diagnosis of headache is largely from the history, and most headache is migraine. This is feature-full pain (a headache plus other features), with nausea, or light and noise sensitivity, and a preference for rest. Migraine is commonly complicated by medication overuse – only a few doses a week over several weeks is enough to turn migraine from an episodic to a chronic disorder. Tension-type headache is seen by many authorities as a relatively featureless form of migraine, with which it commonly co-exists.  

A family history of brain tumour raises concern but does not increase risk. A diagnosis of rare familial tumour syndromes prompts family screening, so when this has not been arranged the family should not worry. 


A brief neurological examination is sufficient – a scheme for this is available on YouTube and has been published in Pulse.2 The key physical sign is papilloedema. Note also conscious level, eye movements, visual fields (binocular to confrontation with fingers), tendon reflexes and plantar responses.  When the history suggests a sinister cause for headache, a normal examination does not rule this out.

Red flags

Thunderclap headache – maximal within moments of onset, and lasting for at least an hour – requires same day exclusion of brain haemorrhage. On day one a normal CT brain scan effectively excludes haemorrhage, but predictive value falls rapidly over time and most headache experts struggle with management of the ambulant patient with significant thunderclap headache weeks earlier.3 These patients belong in emergency care because of the high risk of a further haemorrhage in the early weeks after the initial episode.

Other red flags from the history include a past history of cancer known to spread to brain and headache triggered by physical activity, including cough and Valsalva manoeuvre. The management of headache has been reviewed by NICE whose advice can be misinterpreted – ‘change in personality’ does not mean ‘he’s not the same person since the headaches have been bad’, but an organic neuropsychological change.4

Headache on waking is migraine if intermittent. Raised intracranial pressure causes headache on waking every day.

A headache is not a red flag for a brain tumour. Brain tumour flags include new epileptic seizures and progressive neurological impairment.

Incidental findings

These are known as incidentalomas. Their prevalence depends upon the definition – does this include an opaque maxillary sinus, or a handful of cerebrovascular scars in an arteriopath? One study found a quarter had a ‘not strictly normal’ MRI brain scan.5 A BMJ meta-analysis found ‘neoplastic, structural vascular, inflammatory lesions, cysts and other structural lesions’ but excluded white-matter hyperintensities, silent infarcts, and microbleeds.6 So the incidentaloma risk is between 1:4 (minor technical abnormality) and 1:40 (significant finding).

Incidental findings typically trigger follow-up imaging, which can drag on for years if not decades. Patients with incidentalomas often experience difficulty in obtaining insurance – those who insist upon management of incidental findings can be harmed by treatment.

Scanning – CT or MRI?

CT is preferred for the detection of acute haemorrhage. Otherwise, MRI is always preferable because it does not give ionising radiation, which following a CT scan increases cancer risk by 20% in children (only during childhood).7 Many lesions visible on MRI are invisible on CT so the worried person with a normal CT brain scan will want an MRI.

Imaging policy in headache

Some experts image everyone with headache, but I believe this sends a difficult message to the non-expert. A selective imaging policy is usually appropriate. NICE do not recommend imaging patients with migraine for reassurance, though an unblinded study suggested transient reassurance with diminished healthcare use in the early months after MRI. If you don’t know the diagnosis before the scan, you’re unlikely to know it after the scan.

For a patient with migraine or tension-type headache, a scan is not indicated. If the patient insists on a scan, explain that you think the scan will be normal and discuss the incidentaloma risk, but give way gracefully on the first request. Thereafter if symptoms do not change, do not repeat the scan.

Finally, remember that older people with new headache need a blood test for ESR and CRP to consider the possibility of cranial arteritis.

Dr Giles Elrington is a consultant neurologist at Barts Health NHS Trust and medical director at the National Migraine Centre.



  1. Kernick DP, Ahmed F, Bahra A et al. Imaging patients with suspected brain tumour: guidance for primary care. British Journal of General Practice, 2008; 58 (557): 880–885
  2. Elrington G. How to do a neurological examination in five minutes or less. Pulse, 2007; 2 October
  3. Ducros A, Bousser MG. Thunderclap headache. BMJ, 2013; 346: e8557 
  4. NICE. Headaches: diagnosis and management of headaches in young people and adults. September 2012.
  5. Weber F, Knopf H. Incidental findings in magnetic resonance imaging of the brains of healthy young men. Journal of the Neurological Sciences, 2006; 240 (1-2): 81-84
  6. Morris Z, Whiteley WN, Longstreth WT jr et al. Incidental findings on brain magnetic resonance imaging: systematic review and meta-analysis. BMJ, 2009; 339: b3016
  7. Sodickson A. CT radiation risks coming into clearer focus. BMJ, 2013; 346: f3102