Migraine: a working account

About 753,000 New Zealanders live with migraine.^[1] That's not a trivial number - it's roughly one in six adults, weighted heavily toward women of reproductive age. In any given week in general practice, it's in the room. The challenge has always been that the disease is poorly named (it's not just a bad headache), inconsistently diagnosed, and until recently, inconsistently treated with medications designed for something else entirely.

This note is my working synthesis: where the science has landed, what treatments actually exist, and the frustrating gap between what's available internationally and what's accessible here.

What's actually happening

The vascular theory - migraines are caused by vasodilation - held the field for most of the twentieth century and produced triptans, which work by vasoconstriction. The neurovascular model is now the consensus, and it's more interesting.

Cortical spreading depression

Cortical spreading depression (CSD) is the likely substrate of aura and possibly of the headache phase itself. It's a slowly propagating wave of neuronal and glial depolarisation (moving at roughly 2-6 mm/min across the cortex) followed by prolonged inhibition of neuronal activity.^[2] The visual aura of classical migraine - the fortification spectra, the scotoma expanding across the visual field - maps spatially and temporally onto CSD moving across the occipital cortex, confirmed by modern functional imaging.

Susceptibility to CSD appears to be determined by the excitatory/inhibitory balance of the cortex - itself shaped by genetics, hormonal milieu, sleep, and accumulated migraine history. This is why migraine is properly a neurological disorder of hyperexcitability, not a headache disorder.

The trigeminovascular pathway and CGRP

Once CSD occurs, it activates meningeal nociceptors via the trigeminal nerve. These nociceptors release inflammatory mediators into the dural vasculature - most critically, calcitonin gene-related peptide (CGRP). CGRP is a potent vasodilator and pain sensitiser. It amplifies the signal, dilates dural blood vessels, and drives central sensitisation in the trigeminal nucleus caudalis in the brainstem.^[3]

Central sensitisation explains cutaneous allodynia - the phenomenon where patients in a migraine attack find touch, combing their hair, or even wearing glasses painful. It also explains why triptans work best early: once central sensitisation has established, vasoconstriction alone is insufficient to abort the attack.

The hypothalamus appears to be involved in the premonitory phase (yawning, food craving, fatigue hours before headache onset), and the brainstem periaqueductal grey modulates descending pain inhibition - which is deficient in migraineurs.

Genetics

Migraine is highly heritable. Genome-wide association studies have identified more than 40 loci, many relating to glutamate signalling, ion channel function, and neuronal excitability thresholds.^[4] Familial hemiplegic migraine - a rare monogenic form - involves mutations in CACNA1A, ATP1A2, or SCN1A, all encoding proteins critical to ionic homeostasis. These models have been instructive but FHM is a poor proxy for common migraine, which is polygenic and environmentally modulated.

The treatment landscape

Acute treatment

NSAIDs and paracetamol remain first-line for mild-to-moderate attacks and are often underutilised - people use them too late or at insufficient doses. The principle matters: treat early, treat adequately.

Triptans (5-HT1B/1D agonists) remain the mainstay for moderate-to-severe attacks in patients without contraindications. They cause vasoconstriction, which is why they're avoided in ischaemic heart disease, stroke, and uncontrolled hypertension. Sumatriptan 50-100mg oral is a reasonable starting point; if response is partial, try a different triptan - there's meaningful individual variability. Zolmitriptan nasal spray has the advantage of a non-oral route when nausea predominates.

The gepants represent the first genuinely new acute mechanism in a generation. Rimegepant and ubrogepant (oral CGRP receptor antagonists) and zavegepant (intranasal) work without vasoconstriction, making them suitable where triptans are contraindicated. They also appear not to cause medication-overuse headache - an important property given how common MOH is in frequent migraineurs.

Lasmiditan (a 5-HT1F agonist, or "ditan") avoids vascular receptors entirely. FDA-approved in the US, but not yet widely available in the region. Useful concept: same efficacy target (trigeminal nucleus), different receptor.

Preventive treatment - the old guard

Prevention is indicated when attacks are frequent (generally ≥4 migraine days/month), disabling, or when acute medications are being overused. Historically this meant repurposed drugs:

Agent	Mechanism	Notes for practice
Propranolol / metoprolol	Beta-blockade	First-line, good evidence. Avoid in asthma, Raynaud's. Can worsen depression.
Amitriptyline	TCA / NA reuptake inhibition	Particularly useful with co-existent tension-type headache or insomnia. Sedation is the dose-limiting factor. Start 10mg nocte.
Nortriptyline	TCA	Better tolerated than amitriptyline for many patients, similar efficacy.
Topiramate	Multiple (Na channel, GABA-A)	Good NNT (~3.5 for 50% reduction).[5] Cognitive side-effects frequent ("dopamax"). Teratogenic - contraception counselling essential. Renal stones.
Valproate	GABA enhancement, Na channel	Effective but highly teratogenic. Avoid in women of childbearing age unless extraordinary circumstances and robust contraception.
Candesartan	AT1 antagonism	Good RCT data, well tolerated, reasonable choice especially if hypertension co-exists.[6]
Flunarizine	Ca channel blocker	Not available in NZ or Australia. Widely used in Europe and Asia. Effective but weight gain and depression are issues with long-term use.

Preventive treatment - the CGRP era

The monoclonal antibodies targeting CGRP are a genuine paradigm shift. They are the first medications designed specifically for migraine. They have high specificity, monthly or quarterly dosing, and a side-effect profile that compares very favourably to older agents - most importantly, they don't cause cognitive impairment.

Four agents are now available in New Zealand, all unfunded. Long-term efficacy data are now available out to five years for erenumab, showing maintained response and favourable safety.^[7] Hepatotoxicity, constipation, and hypertension - which were initially flagged as theoretical concerns - have not materialised meaningfully in real-world data reviewed by PTAC in 2025.^[8]

Agent	Target	Dosing	NZ status	AU status
Erenumab Aimovig	CGRP receptor	70 or 140mg SC monthly	Unfunded	Not PBS-listed
Galcanezumab Emgality	CGRP ligand	240mg loading, then 120mg SC monthly	Unfunded	PBS-listed
Fremanezumab Ajovy	CGRP ligand	225mg SC monthly or 675mg SC quarterly	Unfunded	PBS-listed
Eptinezumab Vyepti	CGRP ligand	100 or 300mg IV quarterly	Not available	PBS-listed
Atogepant Aquipta	CGRP receptor (oral gepant)	60mg oral daily	Unfunded	Not PBS-listed

The NZ funding situation

This is the part that requires frank acknowledgement of where we are, because the gap between international standard of care and what's accessible here is substantial.

The contrast with Australia is instructive. Ajovy, Emgality, and Vyepti are PBS-listed across the Tasman, and since November 2024, Australian GPs can initiate CGRP biologics in consultation with specialists.^[9] The step-therapy requirements still apply in Australia (failure of three prior preventive classes), which the AHS position statement argues is not evidence-based - but at least the drugs are accessible at subsidised cost.

In New Zealand, a patient who cannot afford $350-500 per month has access to: propranolol, amitriptyline, topiramate, and valproate. For many patients, these work. For the substantial minority they don't - those with treatment-resistant migraine, contraindications, or intolerable side-effects - there is currently no funded pathway to the standard of care.

Hypotheses worth holding

Working hypotheses I find useful in the consult

Open questions I'm watching

The role of the gut-brain axis in migraine susceptibility is attracting serious research attention, with data suggesting altered gut microbiome composition in migraineurs. Whether this is causal or epiphenomenal remains unclear.

Hormonal migraine - the catamenial pattern, the perimenopause surge, the post-partum window - is mechanistically underexplained. Oestrogen withdrawal appears to lower CSD threshold. Continuous hormonal contraception as a migraine strategy is evidence-adjacent but under-studied, particularly at the low-dose end.

The question of whether frequent migraine attacks cause progressive neurological change remains open. White matter lesions on MRI are more common in migraineurs (particularly with aura), and their significance for long-term cognitive and vascular outcomes is not settled.

International context

The United States has the broadest formulary for migraine: all gepants and ditans are FDA-approved, though insurance coverage remains variable. The UK's NICE has approved CGRP antibodies for patients with ≥4 migraine days/month who have failed three prior preventives. Canada has funded erenumab and galcanezumab through provincial drug plans with varying access criteria. Japan has approved and funds multiple CGRP antibodies.

The global burden of migraine reached 1.16 billion prevalent cases in 2021 - up 58% from 1990, largely driven by population growth and better case ascertainment rather than true incidence change.^[11] It remains the leading cause of disability in people under 50 globally. The healthcare burden is disproportionately borne by women aged 30-44.^[12]

What I actually do

Some practical anchors I've settled on after enough consultations:

Diagnosis first, accurately. ICHD-3 criteria exist. Photophobia, phonophobia, nausea, unilateral location, throbbing quality, and functional impairment are the clinical markers. Migraine without aura is the most common and most missed.

Don't skip over lifestyle. Sleep regularity genuinely matters - inconsistent sleep schedules are a more reliable trigger than red wine. Caffeine is bidirectional: regular moderate use raises the threshold; withdrawal lowers it abruptly. I'm sceptical of comprehensive trigger-avoidance programmes; I'm not sceptical of sleep hygiene.

Treat early and adequately. The most common reason triptans "don't work" is that they're taken too late, at too low a dose, with nothing for nausea. Sumatriptan 100mg + domperidone 10mg, taken within 20 minutes of headache onset, works significantly better than the same drugs taken at peak headache.

Prevent if the threshold is met. Four or more migraine days per month, or any pattern causing significant disability (use the MIDAS or HIT-6 to objectify), justifies a conversation about prevention. Don't underestimate how much disability people normalise and don't volunteer.

Have the funding conversation directly. If a patient would be a good candidate for a CGRP antibody, I'd rather say "this exists, it's excellent, and it costs $X privately" than leave them without the information. Some patients will find a way. All patients deserve to know the gap between what's evidence-based and what's funded.