Migraine: what actually helps, and what is worth investigating

Who this guide is for

If you get headaches often enough that you plan your life around them, and the advice so far has been to take painkillers and avoid cheese, this is written for you.

I am a hospital doctor who also runs a private functional medicine practice in Twickenham and Richmond, and online. I am not a neurologist, and migraine is one of the conditions where I am clearest about that boundary: a good deal of what helps is neurology's, and my job is often to make sure you get there.

What a long assessment can add is the part nobody has time for - whether the treatment is being taken correctly, whether it is now causing the problem, and what else is feeding the frequency.

First: headaches that need assessing now

Most headache is not sinister. These features are the exceptions, and they need urgent medical assessment rather than an appointment in a few weeks.

• Thunderclap headache - severe and maximal within seconds to a minute. Emergency department, same day.
• Headache with fever, neck stiffness, rash or confusion. Same day.
• New neurological symptoms: weakness, numbness, speech difficulty, visual loss, unsteadiness, or seizure.
• Headache that is worse lying down, on coughing, straining or bending, or that wakes you from sleep.
• A new headache over the age of 50, particularly with scalp tenderness, jaw ache on chewing, or visual symptoms - this can be giant cell arteritis and it can take sight.
• A significant change in a long-standing headache pattern, or a steadily worsening headache over weeks.
• New headache with a history of cancer, HIV or immunosuppression.
• Headache after head injury.

The commonest reversible cause: medication overuse headache

This is the section I would most like people to read, because it is common, it is missed constantly, and the treatment is free.

If you take acute headache medication frequently enough, the medication itself starts causing headache. The thresholds in NICE CG150 are triptans, opioids or combination analgesics on 10 days a month or more, or simple painkillers such as paracetamol, aspirin or ibuprofen on 15 days a month or more, sustained over three months.

The pattern is recognisable once you know it. Headaches that were episodic become near-daily. The medication works less well and for less time. You take it earlier and more often, which makes it worse, which means taking more. People arrive convinced their migraine has become intractable when what has actually happened is that the treatment has turned into the problem.

The treatment is withdrawal of the overused drug. It is genuinely unpleasant for a few weeks and a good proportion of people improve substantially afterwards. It should be planned, not attempted alone, and there is usually a case for starting a preventive at the same time.

I count days. Everyone who comes to me with frequent headache gets asked, precisely, how many days a month they take something. It is the most productive question in the consultation.

Getting the acute treatment right before anything else

A surprising number of people have never had a fair trial of the right drug at the right dose at the right time.

1. Take it early. Triptans work best taken as soon as the headache begins. Waiting to see whether it will be a bad one is the commonest reason they appear not to work.
2. Take enough. Under-dosing is routine, particularly with over-the-counter preparations.
3. Combine. NICE recommends a triptan with either an NSAID or paracetamol, which works better than either alone.
4. Try more than one triptan. Response varies considerably between them. Failing one does not mean failing the class, and most people are never offered a second.
5. Consider an anti-sickness drug, even without obvious nausea, because migraine slows gastric emptying and that affects absorption.
6. Count the days. Back to the section above.

Prevention: what works and what a fair trial looks like

Preventive treatment is for people having frequent or disabling attacks. The most common reason it fails is that it is stopped too early or never reaches an effective dose.

NICE lists propranolol, topiramate and amitriptyline as first-line options. Topiramate carries a serious risk in pregnancy and requires proper contraceptive counselling - that conversation is not optional. Candesartan is used off-label with reasonable evidence.

A fair trial is six to eight weeks at an adequate dose, and success is usually a reduction in frequency rather than disappearance. If you stopped a preventive after a fortnight because it had not worked, it was not given a chance.

Beyond those, the CGRP monoclonal antibodies - erenumab, fremanezumab, galcanezumab - are approved by NICE for people who have failed at least three preventives, and they have changed things considerably for that group. They are neurology-prescribed, and if you meet the criteria my job is to get you referred, not to offer you something else.

Supplements: the honest position

Three have enough evidence to be worth discussing, and none is dramatic.

• Riboflavin (vitamin B2), typically 400mg daily. A small randomised trial showed a reduction in attack frequency, it is cheap, and side effects are limited to bright yellow urine. Reasonable to try.
• Magnesium, usually as magnesium citrate or glycinate. Modest evidence, particularly in menstrual migraine. Loose stools are the limiting factor.
• Coenzyme Q10. Small trials suggest a reduction in frequency. More expensive, less convincing.

Butterbur used to be recommended and I no longer suggest it, because unlicensed preparations have been associated with liver injury and the safety of what is actually in the bottle cannot be assured.

None of these is an alternative to a proper preventive if your attacks are frequent and disabling. They are reasonable additions, and I would rather you spent your money on them than on a food sensitivity panel.

Triggers, and the chocolate problem

Trigger identification is less reliable than almost anyone expects, and the reason is interesting.

Migraine attacks begin with a premonitory phase, which can start a day or more before the headache. It produces yawning, fatigue, mood change, neck stiffness and food cravings - classically for sweet things. So the person who eats chocolate and gets a migraine hours later has usually not been triggered by the chocolate. The craving was an early symptom of an attack that had already started.

The same applies to several supposed triggers. Bright light becomes uncomfortable in the premonitory phase, so light gets blamed. Sleep is disturbed, so poor sleep gets blamed.

The ones that hold up better are more mundane: skipped meals, dehydration, caffeine withdrawal, irregular sleep, alcohol, and for many women the fall in oestrogen before a period. Those are worth acting on. A list of twenty forbidden foods usually is not, and it narrows life for no benefit.

What is worth investigating

Migraine is a clinical diagnosis and does not usually need tests. Investigation is for excluding other causes where the history warrants it, and for finding contributors that make attacks more frequent.

What I would not order: IgG food panels, hair analysis, or broad inflammatory profiles. Migraine is not an inflammatory condition in the sense those tests imply, and acting on them means eliminating foods on no evidence.

A realistic aim

Migraine is not cured. What can usually be achieved is fewer attacks, shorter and less severe ones, and confidence that a bad day will not turn into a bad fortnight.

For a good number of people, most of that comes from three unglamorous things: taking the right acute drug early enough, stopping the overuse that has built up, and giving one preventive a fair trial. If that is where you are, you may not need me at all, and your GP can do all three.

About the writer

I am Dr Andrew Greenland, GMC 4419909. I qualified in London in 1997, I am a Fellow of the Royal College of Emergency Medicine, and I still work at consultant level in an NHS hospital in west London alongside my private practice. I am a certified functional medicine practitioner.

I see patients in Twickenham and Richmond in Greater London, and online across the UK. I write to patients' GPs with their consent, refer to neurology where it is needed, and see every patient myself.