Why This Matters More Than Most People Realize

Migraine is not evenly spread across a lifetime. It peaks in women during their reproductive years, exactly the window when many are trying to conceive, are pregnant, or are breastfeeding. That timing collision is the whole problem. The disease is most active when the medication options feel most restricted, and roughly half of pregnancies are unplanned, so exposure to migraine drugs before anyone knows a pregnancy has started is common rather than rare.

A 2026 review in Paediatric Drugs by Smirnoff and colleagues walks through the current safety and effectiveness evidence for treating migraine across pregnancy and lactation. According to PubMed, the honest headline is that the evidence base is thinner than it should be, but it is good enough to make sound decisions, and the picture has shifted in a more reassuring direction for several of the treatments people actually reach for.

What Actually Happens to Migraine in Pregnancy

For most people with migraine without aura, attacks ease as pregnancy goes on. This is thought to track with rising and, crucially, stable estrogen levels. Migraine is highly sensitive to estrogen withdrawal, which is why attacks cluster around menstruation. In pregnancy the sharp monthly drop disappears, and the brain gets a long stretch without that trigger.

The pattern is not universal. People with migraine with aura are less likely to improve and can even worsen, partly because aura reflects cortical spreading depression, a wave of electrical and metabolic change across the cortex that responds differently to the hormonal environment than the vascular and trigeminal pathways driving typical attack pain. The postpartum period is its own hazard. The fast estrogen crash after delivery, plus sleep deprivation and dehydration from breastfeeding, brings attacks roaring back for many people within the first weeks.

What the 2026 Review Found

The Smirnoff review is a narrative synthesis rather than a meta-analysis, so it does not produce a single pooled effect size. Its value is in sorting a messy evidence base into clear, usable categories: what carries known harm, what has reassuring data, and what is simply unstudied. Three findings stand out.

"The old instinct was to strip a pregnant patient down to acetaminophen and hope the attacks fade. The better instinct, backed by the newer data, is to treat migraine as the real and treatable disease it is, matching a specific tool to a specific person rather than defaulting to undertreatment." - Cerebral Torque

Stopping an Attack: Acute Treatment

Acute treatment is about rescuing a single attack, and the layered approach used outside pregnancy still applies, just with a reshuffled toolbox. The principle is to treat early and treat adequately, because a half-dose that fails often leads to a bigger, riskier intervention later.

Acetaminophen (paracetamol) is the first-line analgesic, typically 1000 mg, and it is the most-studied option with the most reassuring record across all trimesters. It is not a strong migraine drug on its own, so it works best combined with an antiemetic.

Antiemetics pull double duty. Metoclopramide 10 mg treats the nausea that defines many attacks and has a mild pain benefit of its own, plus it speeds gastric emptying so oral drugs actually absorb during an attack. It has a long track record in pregnancy. Ondansetron is generally reserved for later trimesters given mixed first-trimester data.

Triptans are the migraine-specific rescue, and sumatriptan is the best-studied. According to PubMed, the 16-year sumatriptan, naratriptan, and Treximet pregnancy registry found a major birth defect rate of 4.2% after first-trimester sumatriptan exposure (95% CI 2.6-6.5%), which sits within the general-population baseline and showed no signal of teratogenicity. Sumatriptan does not cross the placenta well, which fits the reassuring data. It is a reasonable option when acetaminophen plus an antiemetic is not enough.

NSAIDs such as ibuprofen or naproxen have a narrow window. They are best avoided in the first trimester (a possible small miscarriage and malformation signal) and are contraindicated from about 20 weeks onward, where they can cause premature closure of the fetal ductus arteriosus and low amniotic fluid. The second trimester is the only relatively comfortable window, and even then briefly and at the lowest effective dose.

Migraine Medications in Pregnancy: A Practical Evidence Guide

A working summary of common acute and preventive options, when each fits, and how strong the evidence is. Scroll sideways on smaller screens to see every column. This is an education tool, not a prescription. Every choice belongs in a conversation with your own clinician.

Evidence key: Strong best-studied, reassuring   Supported reasonable data or expert support   Limited sparse or low-quality data   Avoid known harm or contraindicated. Doses are illustrative and must be individualized by a clinician.

Prevention: When Attacks Are Too Frequent to Rescue One at a Time

Prevention comes into play when attacks are frequent, long, or disabling enough that rescuing them individually is not working. The order of preference flips toward the lowest-exposure options first.

Start with non-drug preventives. Magnesium and riboflavin are low-risk. Consistent sleep, regular meals, hydration, steady caffeine, and stress management do real work. Biobehavioral therapies such as cognitive behavioral therapy, biofeedback, and relaxation training have solid preventive evidence and zero fetal exposure, which makes them close to ideal in pregnancy.

OnabotulinumtoxinA is a strong option for chronic migraine (15 or more headache days a month). It is injected into head and neck muscles and stays largely local, so systemic and fetal exposure is minimal, and the review counts it among the treatments now supported in pregnancy.

Oral preventives come next if needed, with propranolol the usual first choice despite its modest evidence base. If it is continued into the third trimester, the fetus gets growth monitoring and the newborn is watched for transient slow heart rate and low blood sugar.

What to Avoid, and Why

A short, firm list is more useful than a long hedged one. These are the drugs where the risk is established enough that the answer is a clear no, or a "only with a specific plan to stop."

Opioids deserve a specific mention because they still get handed out in emergency rooms for pregnancy migraine. They are poor migraine drugs, they promote medication-overuse headache, and regular use late in pregnancy risks neonatal withdrawal. Better acute options almost always exist.

Breastfeeding: A Different, Usually More Forgiving Calculus

Lactation is not pregnancy. The question changes from "what crosses the placenta" to "how much reaches breast milk, and what does that dose do to a nursing infant." For most migraine drugs the amount transferred is small, so the breastfeeding list is broader and more relaxed than the pregnancy one.

Acetaminophen and ibuprofen are both considered compatible with breastfeeding and are the everyday first choices. Ibuprofen transfers into milk in tiny amounts and has a short half-life, so the NSAID limits that applied in late pregnancy fall away after delivery.

Sumatriptan is the preferred triptan while nursing. It has poor oral bioavailability, a short half-life, and low milk transfer, so the relative infant dose is very low. Pumping and discarding is generally unnecessary, though a cautious parent can time a dose right after a feed.

Preventives that are reasonable during lactation include propranolol (long history, low milk levels) and, for chronic migraine, onabotulinumtoxinA, which does not meaningfully enter milk. Magnesium and riboflavin remain fine. Valproate is actually low in milk but is avoided more for the mother's own risk profile. Topiramate can be used with monitoring of the infant for sedation and poor feeding, though it is not a first choice.

Putting It Together: Practical Guidance

Plans work best when they are built before they are needed. If you have migraine and are planning a pregnancy, the ideal time to sort out medications is before conception, not during a first-trimester attack.

The thread running through all of this is that migraine in pregnancy and lactation is treatable, and undertreatment is a choice with its own costs. The newer evidence gives clinicians permission to be more proactive with targeted tools like triptans, nerve blocks, and onabotulinumtoxinA rather than defaulting to "tough it out."

The Bottom Line

Migraine is most active in exactly the years when pregnancy and breastfeeding are on the table, and that is why getting this right matters so much. The 2026 Paediatric Drugs review makes three things clear: a handful of familiar drugs (valproate, topiramate, ergots) carry real risk and should go before conception; the old default preventives (propranolol, verapamil) rest on weaker evidence than their reputation suggests; and several effective, targeted options (triptans, lidocaine nerve blocks, and onabotulinumtoxinA) now have enough reassuring data to be used with confidence.

None of this replaces a plan built with your own clinician who knows your history, your aura status, your blood pressure, and your goals. But it should replace the outdated idea that pregnancy means suffering through migraine with nothing but a cold cloth. You have real options, and using them well is part of a healthy pregnancy.