Other Considerations
So what do imaging studies in migraine and TTH patients teach us? Among younger patients, in the absence of any new or alarming symptoms or signs, imaging with MRI will uncover around 1% of patients with a significant abnormality, of which only a small proportion will require intervention. Most abnormal findings, even those of some potential concern such as an aneurysm, call for serial imaging rather than surgery, at least initially. This does not differ from the general, nonheadache population at this age who have had an incidental abnormality picked up by MRI for another indication. In older patients scanned for migraine, MRI is more likely to uncover deep white matter changes, and in those with MA discrete cerebellar changes. These findings do not seem to change within individuals over several years, and are unlikely to have any clinical import; at this stage they remain a research interest only.
There are many potential problems of imaging with MRI, highlighted particularly when the indication to perform the test is weak. Some patients are claustrophobic and require sedation, even a general anesthetic, if it were felt a scan was absolutely required. Many patients would be told of essentially innocuous findings, which may 'pathologize' a proportion. Some patients would be left with the conclusion that there is something wrong with them. Much of this relies on good doctor–patient communication; the utterance cavum septum pellucidum may strike fear in many yet. Furthermore, uncovering the 'incidentaloma' frames an ethical argument as to the probity of scanning in a patient with a very low pretest probability.
Annual healthcare costs of migraine in the Northern hemisphere have been estimated in the internet-based, cross-sectional IBMS study, providing an insight into the use and cost of imaging in different countries. In 2010, an MRI brain scan cost GBP£270 in the UK, €69 in France and US$242 in the USA. Although these figures may not represent the 'true' total cost of the service, they represent a proportion of the total annual cost of migraine, which in Europe is €27 billion, incorporating direct and indirect costs. Every year in the UK, 14% of chronic migraine patients and 4.6% of episodic migraine patients undergo some sort of diagnostic testing (blood tests aside) – predominately imaging, contrasted with 28 and 10% in France, respectively. In a 3-month period in the USA, 9.7% of the chronic migraine patients and 3.5% of the episodic migraine patients underwent diagnostic testing. This compares with the respective figures of 7.4 and 3.5% in Canada. The variable use of diagnostic tests among the counties surveyed in the IBMS study is also reflected in other aspects of healthcare resource use, such as presentations to the emergency department and referrals to a neurologist. However, no contingency can be drawn from any two factors, such as low rates of imaging and higher rates of emergency presentations. Regardless, given the high frequency of diagnostic testing, it may well be that imaging common headache patients has already become routine in developed countries.
Feeding into this complexity are the expectations and anxieties of patients with headache, who are often concerned they have a brain tumor or aneurysm. In one study, 150 chronic daily headache patients from a headache clinic in London (UK) were randomized to receive MRI, and followed for 1 year. Scanned patients were less worried about a serious cause for their headache (p = 0.004) at 3 months but this effect was not evident at 1 year. Approximately a third of the patients with higher psychiatric comorbidity not offered an MRI went elsewhere to have this done. Ultimately, at 1 year only the health costs of patients with higher levels of psychiatric morbidity were significantly reduced in the scanned versus unscanned groups. This was largely as a result of changing the referral pattern of the primary care doctors. Hence one could argue this finding sits comfortably in the tradition of diagnostic testing to treat the doctor, rather than the patient.