Taylor et al provide an interesting review on the timeless conundrum of headaches and brain tumours.1 Among the discussion they seem to advocate easier access to brain imaging in primary care. This is a shame, especially since the question was answered by a well-written study (by two of the same authors) in the BJGP 7 years ago.2 I use that paper to teach medical students and GP trainees that the positive predictive value of a headache, for a brain tumour, is 0.09%. (In other words, if you have a headache there’s only a 0.09% chance of it being a brain tumour). A new onset seizure, on the other hand, has a PPV of 1.2% for a brain tumour. The problem with imaging people’s brains ‘just to make sure’ is, as the authors rightly point out, an incidental finding. A well carried out study showed 0.47% of healthy young men have an intracranial tumour,3 which is slightly more than the positive predictive value of a headache anyway. Taylor et al also, erroneously, remark ‘the commonest symptom’ of a brain tumour is headache, but in fact only 10% of people with a brain tumour ever report a headache before the diagnosis. Easier access to brain imaging for GPs will only mean one thing: more brain scans. And more brain scans means only one thing: more incidental findings.
Notes
Competing interests
The author missed a brain tumour 3 years ago in a young woman presenting with headaches (the patient is now well). He has seen a similar patient have a meningioma, erroneously attributed to the patient’s headaches, resected; the headaches persisted despite the surgery and abated once the patient’s emotional state had improved.
- © British Journal of General Practice 2014
