Authors
Zahid H Bajwa, MD
Jonathan H Smith, MD
Section Editor
Jerry W Swanson, MD, MHPE
Deputy Editor
John F Dashe, MD, PhD
All topics are updated as new evidence becomes available and our peer review
process is complete.
Literature review current through: Jul 2017. | This topic last updated: May 31,
2017.
Other aspects of migraine are reviewed elsewhere. (See "Acute treatment of migraine
in adults" and "Pathophysiology, clinical manifestations, and diagnosis of migraine in
adults".)
While there are no strict definitions for the precise frequency or duration of migraine
headaches that would prompt preventive therapy, more than four headaches per
month or headaches that last longer than 12 hours are generally considered
reasonable thresholds.
Choosing pharmacologic therapy — A number of drug classes are used for the
prevention of migraine. Medications that are effective in controlled trials include:
We suggest initial therapy with one of these agents for most patients with episodic
migraine (≤14 headache days per month) who have an indication for preventive
therapy (see 'Indications' above). Approximately 50 to 75 percent of patients given
any of these drugs will have a 50 percent reduction in the frequency of headache,
but the doses required may lead to intolerable side effects [7]. The choice among
preventive agents of similar efficacy should be individualized according to patient-
specific characteristics, comorbid conditions, medication side effect profile, and
patient values and preferences [8-10].
●For patients with hypertension who are nonsmokers and ≤60 years of age,
reasonable options include metoprolol, propranolol, or timolol in the absence of
contraindications to beta blockers.
●For patients with hypertension who are smokers or are >60 years of age, options
include verapamil or flunarizine. We suggest not using beta blockers as initial
therapy for migraine prevention in these patients because beta blockers may be
associated with a higher rate of stroke and other cardiovascular events compared
with other antihypertensive drugs in the primary treatment of hypertension. (See
"Choice of drug therapy in primary (essential) hypertension", section on 'Beta
blockers'.)
●For patients with depression or mood disorder, reasonable options include
amitriptyline or venlafaxine.
●For patients with epilepsy, reasonable options include valproate or topiramate.
●For patients with insomnia, amitriptyline is a reasonable option.
●For patients with obesity, topiramate is a reasonable option.
●For patients with Raynaud phenomenon, reasonable options include verapamil
or flunarizine.
The treatment of chronic migraine (≥15 headache days per month) is reviewed
separately. (See "Chronic migraine", section on 'Management'.)
●Start the drug at a low dose. Increase the dose gradually until therapeutic benefit
develops, the maximum dose of the drug is reached, or side effects become
intolerable.
●Give the chosen medication an adequate trial in terms of duration and dosage.
Clinical trials suggest efficacy is often first noted at four weeks and can continue
to increase for three months.
●Avoid overuse of acute headache therapies including analgesics, triptans, and
ergots.
●Opioids and barbiturates should not be used for the acute or preventive
treatment of migraine. Opioid use can contribute to development of chronic daily
headache and can interfere with other preventive therapies. (See "Acute treatment
of migraine in adults", section on 'Opioids and barbiturates' and "Medication
overuse headache: Etiology, clinical features, and diagnosis", section on 'Causal
medications'.)
●Women of childbearing potential must be warned of any potential risks. Avoid
valproate for this group if possible, and choose the medication that will have the
least adverse effect on the fetus. (See "Headache in pregnant and postpartum
women", section on 'Preventive therapies'.)
●Address patient expectations and consider patient preferences when deciding
between drugs of relatively equivalent efficacy. Discuss the rationale, dosing, and
likely side effects for a particular treatment. Discuss expected benefits of therapy
and how long it will take to achieve them. Preventive migraine therapy requires a
sustained commitment on the part of the patient and physician to achieve benefit.
●Migraine headaches may improve independent of treatment. If the headaches
are well controlled, slowly taper the drug if possible. Many patients experience
continued relief with either a lower dose or cessation of the medication.
While few high-quality, placebo-controlled trials have evaluated the efficacy of these
factors for migraine prevention, we recommend therapeutic lifestyle measures
because they may be effective and are unlikely to cause any harm. A headache diary
is useful for estimating the number and severity of headaches each month and for
identifying precipitating factors so that so they can be managed (form 1). The optimal
management of headache trigger factors is uncertain; traditional teaching
emphasizes avoidance, while newer approaches emphasize coping strategies. (See
'Coping with headache triggers' below.)
Antihypertensives — Data from randomized clinical trials show that beta blockers,
particularly metoprolol, propranolol, and timolol, are effective for migraine
prevention [2,11]. Lower quality evidence suggests but does not establish that
calcium channel blockers, angiotensin converting enzyme (ACE) inhibitors, and
angiotensin II receptor blockers (ARBs) are effective for migraine prevention. There
are no clinical trial data for thiazide diuretics in migraine prevention.
Blood pressure treatment appears to reduce the overall prevalence of headache in
general. This point is illustrated by the results of a meta-analysis of 94 randomized
controlled trials examining four different classes of antihypertensive medications
(beta blockers, ACE inhibitors, ARBs, and thiazides) that also included data on
headache prevalence [12]. Patients assigned to blood pressure treatment had a
significant reduction in headache prevalence compared with placebo treatment
(odds ratio 0.67, 95% CI 0.61-0.74), and headache prevalence was significantly
reduced in each of the four classes of drugs compared individually with placebo. One
limitation is that the individual trials lacked information on headache classification,
so the proportion of migraine versus other types of headaches among the subjects
is unknown. However, it is conceivable that most of the patients had migraine, as the
prevalence of headache among placebo patients (12.4 percent) is similar to the
reported prevalence of migraine in the general population. (See "Pathophysiology,
clinical manifestations, and diagnosis of migraine in adults", section on
'Epidemiology'.)
Our suggested doses for migraine prevention with beta blockers are as follows
[10,13,14]:
Although we generally do not use timolol for migraine prevention, it can be started
at 5 mg once daily; the dose range is 10 to 30 mg daily given in two divided doses.
It can take several weeks for headache improvement to accrue with these drugs [15];
the dose should be titrated and maintained for at least three months before deeming
the medication a failure.
We suggest not using beta blockers as initial therapy for migraine prevention in
patients over age 60 and in smokers. Compared with other antihypertensive drugs in
the primary treatment of hypertension, beta blockers may be associated with a higher
rate of stroke and other cardiovascular events. (See "Choice of drug therapy in
primary (essential) hypertension", section on 'Beta blockers'.)
The use of beta blockers may be also limited in patients with erectile dysfunction,
peripheral vascular disease, Raynaud phenomenon, and patients with baseline
bradycardia or low blood pressure. They must be used cautiously as well in patients
with asthma, diabetes mellitus, or depression, and in those with cardiac conduction
disturbances or sinus node dysfunction. (See "Major side effects of beta blockers".)
Calcium channel blockers — Calcium channel blockers are widely used for migraine
prevention. However, the data supporting the efficacy of calcium channel blockers
are weak and conflicting [2,10]. Two methodologically flawed studies showed a small
but significant effect of verapamil (in three divided doses starting at 120 mg daily,
dose range 120 to 240 mg daily) for migraine prevention [16,17]. Nevertheless, one
author of this topic (ZB) frequently employs verapamil as the first choice for
preventive therapy based upon ease of use and a favorable side effect profile [18].
Flunarizine, a nonspecific calcium channel blocker, also has some evidence of efficacy
[3,10]. Based on several controlled studies, the recommended dose of flunarizine is 5
to 10 mg daily. Flunarizine is not available in the United States.
Tolerance may develop with calcium channel blockers; it occurred after eight weeks
of successful therapy in 42 percent of those treated with nifedipine, 49 percent
treated with verapamil, and 4 percent treated with nimodipine in one report [19].
Tolerance may be overcome by increasing the dose of medication, or by switching to
a different calcium channel blocker.
The tricyclic antidepressants most commonly used for migraine prevention include
amitriptyline, nortriptyline, doxepin, and protriptyline. Amitriptyline is the only
tricyclic that has proven efficacy for migraine; there are insufficient data regarding
the other tricyclics [8].
Side effects are common with tricyclic antidepressants. Most are sedating, particularly
amitriptyline and doxepin. Therefore, these drugs are usually used at bedtime and
started at a low dose. Additional side effects of tricyclics include dry mouth,
constipation, tachycardia, palpitations, orthostatic hypotension, weight gain, blurred
vision, and urinary retention. Confusion can occur, particularly in older adults. (See
"Tricyclic and tetracyclic drugs: Pharmacology, administration, and side effects".)
The two largest randomized trials of topiramate for migraine each evaluated over
460 patients [24,25]. The following observations emerged from these trials [24,25]:
Data from one trial suggest that topiramate has residual benefit for migraine that
persists for up to six months after the drug is discontinued [27].
Valproate can cause nausea, somnolence, tremor, dizziness, weight gain, and hair
loss. It is contraindicated in pregnancy because of teratogenicity. In addition,
valproate has been linked to lower intelligence quotient scores in children of women
who used valproate during pregnancy (see "Risks associated with epilepsy and
pregnancy", section on 'Effect of antiseizure drugs on the fetus'). Therefore, valproate
should not be used by women of childbearing age for headache prevention [30].
Other agents — A number of other agents have been studied for migraine
prevention. Guidelines issued in 2012 from the AAN concluded that Petasites
(butterbur) is effective for migraine prevention [33]. In addition, the AAN noted that
feverfew, magnesium, certain nonsteroidal anti-inflammatory drugs, riboflavin, and
subcutaneous histamine are probably effective.
Coenzyme Q10 — Interest in coenzyme Q10 (CoQ10) and riboflavin (see 'Riboflavin'
below) for migraine treatment has been sparked by the potential role of
mitochondrial dysfunction in migraine pathogenesis [42].
Feverfew — Feverfew has been the herbal remedy most studied for the prevention
of migraine, but evidence regarding benefit is conflicting. A 2015 systematic review
of feverfew for migraine prevention found six randomized controlled trials with 561
patients that met the inclusion criteria [44]. Four of the included trials [45-48] found
that feverfew was beneficial, while two trials [49,50] found no significant difference
between feverfew and placebo [44]. There were no major adverse effects associated
with the use of feverfew. A planned meta-analysis was not performed because of
heterogeneity among the trials and the absence of common outcome measures. All
trials were considered to be at unclear or high risk of bias due to small sample size.
The review concluded that larger rigorous trials are needed before firm conclusions
can be drawn about the effectiveness of feverfew for migraine headache.
Guidelines issued in 2012 from the AAN concluded that the NSAIDs fenoprofen,
ibuprofen, ketoprofen, and naproxen are probably effective for migraine prevention
[33]. This AAN assessment was based upon an earlier systematic review and meta-
analysis from the US Agency for Health Care Policy and Research (AHCPR), which
included 23 controlled trials of 10 different NSAIDs indexed from 1966 through 1996
[63]. The AAN identified no newer randomized trials of NSAIDs for migraine
prevention [33]. However, the AHCPR report noted that the evidence was strongest
for naproxen, with multiple placebo-controlled trials showing a moderate reduction
in headache symptoms [63]. Only one trial supported the efficacy of ketoprofen [64],
while the evidence was less certain for fenoprofen. The AHCPR review identified no
controlled trials of ibuprofen [63].
The use of NSAIDs for the treatment of acute migraine headache is discussed
separately. (See "Acute treatment of migraine in adults", section on 'Nonsteroidal
anti-inflammatory drugs'.)
Acupuncture — Clinical trial evidence suggests that classic acupuncture for migraine
headache is more effective than oral pharmacologic placebos, but probably not more
effective than sham needling or traditional medical therapy [77-81]. The following
reports support this conclusion:
Although the explanation for these results is uncertain, one hypothesis is that,
compared with oral pharmacologic placebos, the placebo effect of sham acupuncture
is enhanced by contextual factors such as more elaborate treatment rituals and
higher levels of attention and physical contact [80,84,85]. Headache improvement
related to nonspecific physiologic effects of needling is another possible mechanism.
The trial screened 317 patients with frequent moderate to severe migraine headache
(with or without aura) and found 76 who were eligible for inclusion, which required
both identification of the predominant "trigger site" where "migraine pain started
and settled consistently" and a positive response to botulinum toxin injection at the
trigger site [89]. Eligible patients, who had approximately 10 migraine headaches per
month at baseline, were randomly assigned to treatment with either actual surgery
(n = 49) or sham surgery (n = 26) at their trigger site.
Actual surgery involved removal of glabellar muscles from patients with frontal
trigger sites, removal of a segment of the zygomaticotemporal branch of the
trigeminal nerve from those with temporal trigger sites, and resection of a segment
of the semispinalis capitis muscle (under general anesthesia) from those with occipital
trigger sites [89]. The sham surgery group had exposure but not resection of the
muscles and nerves through a similar incision. At 12 months, actual surgery was
statistically superior to sham surgery on a variety of outcome measures, including
complete elimination of migraine headaches (57 versus 4 percent), and a reduction
of ≥50 percent in the migraine headache index, calculated by multiplying the
headache frequency, intensity, and duration (84 versus 58 percent) [89]. Confidence
intervals were not reported. Improvement with surgery was independent of trigger
site. The surgery was well tolerated.
Given the methodologic flaws in the study design [90,91], independent confirmation
of benefit in more rigorous trials is needed. In addition, only a minority of patients
with frequent migraine (those with identifiable trigger sites and a positive response
to botulinum toxin injection) would appear to be candidates.
Closure of right-to-left cardiac shunt — There is conflicting evidence regarding a
possible association of migraine with aura and right-to-left cardiac shunts in the
setting of a patent foramen ovale (PFO) or an atrial septal defect, as reviewed
separately. (See "Pathophysiology, clinical manifestations, and diagnosis of migraine
in adults", section on 'Right-to-left cardiac shunt'.)
However, there is evidence from randomized controlled trials that PFO/ASD closure
is not an effective treatment for migraine. The multicenter double-blind MIST trial
enrolled 147 patients with a PFO who had frequent migraine with aura that was
refractory to two types of preventive medication [92]. Patients with prior stroke or
transient ischemic attack were excluded. During a three-month healing phase,
patients in both treatment groups received aspirin plus clopidogrel. At six months,
there was no difference between PFO device closure and sham treatment for the
primary end point of headache cure, which was achieved in each treatment group by
three patients (4 percent). Furthermore, the PFO closure group experienced more
serious adverse events than the sham group. The smaller open-label PRIMA trial,
which was stopped prematurely due to slow recruitment, compared PFO device
closure with medical management among subjects with migraine with aura [93]. At
one year, there was no statistically significant difference between groups for
reduction in monthly migraine days.
In earlier reports, which were relatively small and largely retrospective, improvement
in migraine after device closure of a PFO or an atrial septal defect was noted in some
(but not all) studies [94-104]. Most of the patients treated with shunt closure were
selected because of prior cryptogenic stroke or paradoxical embolism. A majority of
study patients were treated with aspirin and clopidogrel, which may have reduced
migraine frequency.
●Basics topics (see "Patient education: Migraine headaches in adults (The Basics)")
●Beyond the Basics topics (see "Patient education: Migraine headaches in adults
(Beyond the Basics)" and "Patient education: Headache treatment in adults
(Beyond the Basics)")
●The main indications for migraine prevention therapy are frequent or long lasting
migraine headaches and migraine headaches that cause significant disability or
diminished quality of life. In addition, preventive migraine therapy is indicated to
reduce the risk of neurologic damage in the presence of uncommon migraine
conditions including hemiplegic migraine, migraine with brainstem aura, and
migraine with prolonged aura. (See 'Indications' above.)
●Therapeutic lifestyle measures including good sleep hygiene, routine meal
schedules, regular exercise, and management of migraine triggers are appropriate
for all patients with migraine. (See 'Our approach' above.)
●For patients with migraine who have an indication for preventive therapy, we
suggest treatment with one of the agents effective in controlled trials (Grade 2A),
a list which includes the following (see 'Our approach' above):
•Metoprolol, propranolol, and timolol (see 'Beta blockers' above)
•Amitriptyline and venlafaxine (see 'Antidepressants' above)
•Valproate and topiramate (see 'Anticonvulsants' above)
Alternatives include calcium channel blockers including verapamil and flunarizine,
which may be less effective but are generally safe and well tolerated. (See 'Calcium
channel blockers' above.)
We suggest selecting one of these agents as initial therapy according to individual
patient characteristics and preferences, medication side effect profile, and the
presence of comorbid conditions including but not limited to hypertension,
depression or mood disorder, epilepsy, insomnia, obesity, and Raynaud
phenomenon. (See 'Choosing pharmacologic therapy' above.)
●For patients who fail to improve despite an adequate trial of initial pharmacologic
therapy, we suggest switching to a migraine prevention medication from a
different class (Grade 2C). (See 'Treatment failure' above.)
●Preventive migraine therapy requires a sustained commitment on the part of the
patient and physician to achieve benefit. In general, preventive drugs are started
at a low dose and the dose is gradually increased until therapeutic benefit
develops, the maximum dose is reached, or side effects become intolerable. Benefit
is often noted first at four weeks and can continue to increase for three months.
(See 'Principles of preventive therapy' above.)
●Nonpharmacologic measures that may be beneficial for migraine headache
prevention include aerobic exercise, biofeedback, other forms of relaxation
training, cognitive-behavioral therapies, acupuncture, and transcutaneous
electrical nerve stimulation. (See 'Nonpharmacologic interventions' above.)