Migraine is a “committee disease”, with multiple genetic variants. It seems extremely likely that there is no single drug that will “work” in this collection of several disorders that share some common features. In other words, prevention of migraine is a “trial and error” process.
For those who have more than 2 severe headaches/month and in patients with complicated migraine (migraine with stroke-like features), a daily medication may be worth while. These are generally highly effective (about 75% effective), but do require daily regular use. Only about 3-5% of migraine patients use prevention medications, according to Gray (1999).
These drugs fall into four major classes: anticonvulsants, antidepressants, antihypertensives, and Botox (? local anesthetic). Examples are: Amitriptyline (Elavil), Botox, Depakote and Topiramate, Inderal, Nardil and Venlafaxine, Verapamil (Calan, Isoptin). These drugs seem to work via several pathways: some are beta-blockers (e.g. Inderal, Corguard), some are calcium channel blockers (e.g. verapamil). Some work on electrical activity in the brain(e.g. topiramate). Some in mysterious ways, possibly through manipulations of serotonin(e.g. Nardil, amitriptyline, venlafaxine). More information about these is in the next section.
Migraine is a “committee disease”, with multiple genetic variants. It seems extremely likely that there is no single drug that will “work” in this collection of several disorders that share some common features. In other words, prevention of migraine is a “trial and error” process.
To use these drugs it is best to have a measure of effect. One way is to use a headache diary. Here we recommend a month/page style format, and a stoplight; color code – – red bad, yellow — fair, green good. Another is to do it online
Our approach to migraine prevention and use of these drugs can be found by clicking on the flowchart above.
We also can point the reader to an extensive online review published in 1999 by a US government agency, and found on the NCBI bookshelf. Although it is extremely dated (i.e. somebody should put out a newer edition), it still summarizes an immense amount of data.
General comments:
The author of this review usually starts patients with effexor (venlafaxine), and proceeds on to try topamax, verapamil, propranolol and then ami or nortriptyline. Botox injections and acupuncture are new avenues that we are presently trying. It is very unusual that headache control is not attained. When one “group” doesn’t work, he may combine several groups simultaneously (anticonvulsant, blood-pressure agent, antidepressant, Botox or Acupuncture). Except in unusual cases, other than Botox, drugs are stopped in women intending to become pregnant.
We would urge great caution in believing the migraine literature. In general, those that commonly publish articles about migraine treatment are funded by the drug industry (e.g. see Botox). When one looks at the lists of drug companies who have funded the writers of these articles, they may take up several paragraphs ! Drugs that might be particularly likely to have “slanted” literature are those that are expensive (which includes many) and in which there is no generic drug yet available. There are many of these medications used for migraine treatment.
We have been asked several times whether or not migraine prophylactic medications will prevent evolution of the white matter lesions that are common in migraine. As of 2010, the literature is silent on this question. There are two general ideas about the origin of the white matter lesions — 1. microemboli through a patent PFO or other type of shunt 2. Vasospasm due to some mysterious underlying neurochemical abnormality associated with migraine. If #1 is accurate, then one would think that persons with spots would be treatment resistant (as most medications we use have little to do with coagulation – -verapamil is the main exception). If #2 is accurate, one would think that persons with spots would be the same as anyone else regarding treatment, and in fact, should be treated more aggressively to prevent damage. The PFO theory has recently been downplayed.
Drugs (alphabetical).
See this link for a list of drugs organized by mechanism.
There are so many studies of these drugs, a database would seem the most logical way to organize them. We do not know of any online (yet).
Ace inhibitors (candesartin, lisinopril) Slightly effective. (2012)
ACE inhibitors are presently mainly used for hypertension, but they also have some slight utility for migraine prevention. (Ashenazi et al, 2003; Tronvik et al, 2003). Studies of these drugs suggest that they are better than placebo. Nevertheless, it is the opinion of the writer of this review that they are much less effective than beta-blockers, venlafaxine, and amitriptyline.
Aripiprazole (Abilify). Need more data.
Recently there have been anecdotal reports concerning effectiveness of this atypical antipsychotic drug for migraine prevention. Aripiprazole is a partial agonist at both the dopamine D2 receptor and the serotonin 5-HT1A receptor. Dopamine agonism would be expected to worsen migraine and nausea, while 5-HT1A agonsm, treat migraine. Like the other atypical antipsychotics, aripiprazole displays an antagonist profile at the 5-HT2A receptor. Dose is 10 mg. Common side effects include akathisia (restlessness), headache, unusual tiredness or weakness, nausea, vomiting, an uncomfortable feeling in the stomach, constipation, light-headedness, insomnia, sleepiness, shaking, and blurred vision. This is an expensive drug.
amitriptyline (Elavil). Effective and inexpensive drug, with considerable side effects, used for prevention
Tricyclics, primarily amitriptyline, have been well studied (e.g. Couch et al, 2011). Usual dose is 50 mg at night but the starting dose is usually 10 mg. Some people do well with just 10 mg. Works very well, but takes 2-6 weeks to work. Amitriptyline doesn’t lower the blood pressure. Dry mouth and sleepiness main side effects. Weight gain of as much of 25 lbs is common.Elavil is very inexpensive ! Similar tricyclic type drugs include nortriptyline, doxepin and protriptyline. While no studies are available comparing them, it is likely that they all work. Oddly enough though, although the literature is mixed, clomipramine probably does not work for Migraine (Langohr et al, 1985; Noone, 1977). Amitriptyline is more likely to have serious side effects when used by people with heart block or urination problems or persons over the age of 60. Pregnancy is category D. We mainly use this drug when our favorites fail.
Amlodipine (Norvasc). Unknown efficacy.
While there have been some sporadic reports suggesting that the calcium channel blocker, amlodipine, works for migraine. There is simply not enough data. The data for other calcium channel blockers, such as nimodipine, nicardipine, and verapamil, is not that strong.
Botox (Botulinum toxin injections). Slightly effective.
This agent has recently been FDA approved for prevention of migraine, based on a large trial funded by the drug manufactorer. The mechanism is thought to be action on sensory nerves in the head, not through paralysis of muscles in the head. It is weak and it is expensive. On the other hand, it has no systemic side effects.
carbamazepine (Tegretol) is not effective as a preventive (2012).
A related anticonvulsant, oxcarbazepine (Trileptal) has had some limited success in treating refractory migraine with about a 50% response rate (two abstracts suggest this — Johnson et al, 2002; Nett and Krusz, 2002). Oxcarbazepine is not FDA approved for this indication. We do not prescribe it in our practice for this purpose either.
Clonazepam — not effective for migraine
cyproheptadine (periactin) is a preventive medication mainly used in children.
Weight gain is common. We do not favor it in general.
Clonidine. not effective. (2012). This is in spite of reports that a close relative, guanfacine may be helpful.
Cymbalta (duloxetine). This drug may be effective in high doses. It is not effective in starting doses.
Depakote (sodium valproate). Effective but with many side effects. Pregnancy category D. (2012)
There are so many side effects that we don’t see it as a first-line. Used for prevention. Usual dose is 250, three times/day. Side effects include a prominant tremor, weight gain, and sometimes hair loss. Depakote also should not be used in women of childbearing age who are not using birth control as it is pregnancy category D. Among other things, it increases the risk of autism (Christensen et al, 2013). Some authors suggest that this drug is effective for persistent migraine aura (Rothrock, 1997). There are many possible reasons for it working and at this writing, it is not clear which one is correct (Cutrer et al, 1997).
Effexor (venlafaxine HCI), see here. We find this drug very effective.
Withdrawal is a big problem from larger doses, and we prefer to use very small doses. We always start patients on 1/3 of a capsule of the 37.5 XL (brand name is easiest to get that can be split) drug. Every week we increase the dose by 1/3, so that at the beginning of the third week, the person is on the full 37.5 capsule. Effexor is generally thought to be safe in breast-feeding (seedrugs.com site). Effexor is not approved for use in children under the age of 16. We have no data concerning whether or not it is effective in children for migraine.
It would seem logical that venlafaxine might work better in menopausal migraines than other agents, as it is also an effective medication for hot flashes.
Escitalopram (Lexapro). An SSRI. Probably little effect.
We have had little experience with use of this drug for migraine, but it has been reported effective in a one study, e.g. Tarlaci, S. (2009). Lexapro is currently expensive compared to older, generic similar nearly identical drugs such as citalopram. We don’t see a rationale why this drug would work. See the discusison of Prozac below.
Fluoxetine (Prozac). An SSRI. Midly effective.
There have been several studies of modest size, generally reporting a mild-moderate effect. This medication, a member of the SSRI family (which also includes Paxil, Zoloft, Celexa, (escitalopram) Lexapro, and Luvox), is suggested to be effective (Silberstein, 2000; Silberstein et al, 2012).Nevertheless, of roughly 7 controlled trials of SSRI’s for migraine, most have showed negligable effect. There is some evidence that another SSRI, paroxetine (Paxil) improves chronic daily headache (Langemark and Olesen, 1994; Foster and Bafaloukos, 1994).
Studies that directly compare fluoxetine to other antimigraine agents such as propranolol and amitriptyline, generally conclude that it is less effective. On the other hand, Fluoxetine has far less side effects than either.
Not everyone agrees that there is good evidence that the SSRI’s help migraine(e.g. Goadsby, 2002). We rarely prescribe them for headache in our practice.
Flunarizine (Sibellium) is available in Europe. It is a mixed mechanism drug. We rarely prescribe it because of its dopamine blocking side effects (i.e. think major tranquilizer).
Flunarizine is at least as effective as propranolol (see later). Flunarizine is likely more effective than verapamil because it combines calcium channel and dopamine blocking activity in a single preparation (Afran et al, 1998). Flunarizine has an 18 day half-life, meaning that good effects as well as adverse effects might take a long time to start and end. Flunarizine has been reported as effective in cyclic vomiting (Kothare et al, 2005), and in childhood migraine (Visudtibhan et al, 2004; Peer Mohamed et al, 2012). The usual dose is 5 to 10 mg at night.
Gabapentin (Neurontin). Not very effective, but cheap and worth a try.
Neurontin is not very potent for migraine, but it has so few side effects, that it may be worth a try anyway. This anticonvulsant is a prophylactic drug for treatment of migraine (Silberstein, 2000). Gabapentin (strangely enough) does not affect Gaba-b receptors or other commonly studied receptors. It may nevertheless increase glutamate-dependent GABA synthesis and it also binds to the calcium channel. Adverse effects include sleepiness, dizziness, fatigue and weight gain associated with increased appetite. A newer version of Neurontin is “Lyrica”. This is basically a far more expensive version of gabapentin with a few advantages. Imbalance is common as a side effect of Lyrica. Pregnancy category C.
Inderal LA (propranolol) and other beta-blockers such as timonol. Very effective. (2012)
Others include atenolol, metaprolol and nadolol, and nebivolol. There have been numerous trials of beta-blockers. No difference has been found between propranolol, metoprolol, timolol and nadolol. Beta-blockers are similarly effective to anticonvusants (e.g. topiramate, valproate) and more effective than SSRI type antidepressants (e.g. flunarizine, citalopram).
Propranolol is a very effective and cheap drug, with moderate side effects, when used for prevention. The usual dose is 60 mg LA in the evening. Works as well as Verapamil, but generally has more side effects. Pulse may be slowed. Has a mildly calming effect. Nadolol (Corgard) has a similar effect. Both Inderal and Corgard are non-selective beta blockers. More selective beta blockers include metoprolol (Lopressor, Toprol, dose 25-75 at bedtime) and Atenolol (Tenorman), and Bystolic (nebivolol). Bystolic is particularly low in side effets.
All beta-blockers have a tendency to have a withdrawall syndrome entailing hypertension, and for this reason, are best tapered off when there is a decision to stop. Betablockers also generally have a tendency to increase depression, cause weight gain, and to cause sexual dysfunction.
Selective beta blockers usually have less side effects than the unselective beta blockers. Studies of the more selective beta-blockers have shown similar efficacy to the unselective ones, and thus it would seem better to choose the ones that are more selective and have less side effects.
In general, beta-blockers shouldn’t be used by persons with asthma, depression, heart failure, diabetes, or taking allergy shots. All beta blockers have some risk of hair loss. This is fortunately rare. This is not an absolute prohibition and in some cases beta-blockers are helpful depending on the overall situation. Combined use of verapamil and beta-blockers should also, generally speaking, be avoided.
Atenolol is pregnancy category D, while metoprolol, propranolol and nadolol are all pregnancy category C. The ‘C’ agents are preferable in women of childbearing age. Beta blockers, and particularly atenolol, have been reported to increase the chance of diabetes in older people.
A recent trial in which topiramate and inderal were combined for chronic migraine, was stopped because of lack of evidence that there was any benefit (Silberstein et al, 2012). Of course, this does not mean that the combination is also ineffective in non-chronic migraine, which comprises about 97% of all migraine. Also, note that this combination competes with an alternative extremely expensive treatment for chronic migraine, Botox. In our opinion, it still is prudent to exhaust inexpensive combinations for chronic migraine, before turning to Botox.
Lamotrigine (Lamictal). Mildly effective but substantial side effects.
This anticonulsant can be used in a similar way as Depakote (see above) to prevent migraine and migraine associated vertigo (Bisdorf, 2004). Oddly enough, a recent review article said that there was evidence that Lamotrigine was not effective for migraine. (Silberstein et al. 2012). We are dubious.
Magnesium. Mildly effective, with very little side effect.
Dietary supplements of magnesium as well as intravenous injections of magnesium have been reported to be effective in migraine (Peikert et al, 1996; Mauskop, 1998). Brain magnesium has a complicated relationship to migraine (Boska et al, 2002). Magnesium is usually taken as a dietary supplement, in combination with calcium. About 500 mg/day is suggested. No prescription is necessary. Safety is unknown in pregnancy.
Memantine (Namenda). Probably not effective.
There are some very preliminary, uncontrolled studies suggesting that this drug in the usual doses for alzheimer’s reduces headache frequency by about 50% (Charles et al, 2007; Peters et al, 2007). At this writing, we are trying this out in refractory patients. It is difficult to see why it would work.
Methysergide (Sansert):
This drug is unavailable in the US as of 1/2003. Very effective but potentially dangerous. Taken in a dose of 2mg TID. Every 6 months, you MUST stop this medication for one month. There is a danger of poor circulation. This drug is a last resort. Some authors recommend getting a CT scan of the kidney area one year after initiating treatment.
The relative of sansert, methergine, is likely as dangerous as sansert.
Nimodipine, Nicardipine, Nifedipine — weakly effective. These are calcium channel blockers. See the section on verapamil for more.
There have ben many studies of nimodipine, all reporting weakly positive effects.
NSAIDS (non-steroidal anti-inflammatory drugs). Effective but substantial side effects.
Examples are aspirin, fenoprofen, flurbiprofen, ketoprofen, mefanamic acid, and naproxen. Indomethacin is not effective for prevention although perhaps effective as an abortive treatment. Naproxen is pregnancy category B, making it one of the safest and least expensive drugs in pregnancy. This entire group is under some suspicion of contributing to cardiovascular risk.
Oxcarbamazine (Trileptal) is not indicated for migraine, see comments above related to carbamazepine.
Pizotifen is a medication similar to cyproheptadine, with both antihistamine and anti-serotonin properties.
The usual dose is 0.5 mg daily. It is not FDA approved in the USA.
Propranolol – -see Inderal. Effective.
Quetiapine (Seroquel) is an atypical antipsychotic, mainly used for bipolar affective disorder.
There is a growing body of evidence that it is helpful to prevent migraine. (Krymchantowski, et al. (2010). Weight gain is a common side effect. The mechanism of action is attributed to D2 antagonism. This makes it similar to flunarizine (see above).
Tecagepant (CGRP antagonist). Don’t get your hopes up too high.
This is an investigational drug that antagonizes CGRP. A recent trial was stopped because of liver toxicity (Ho et al, 2014). Calcitonin gene-related peptide agents (CGRP) are similar in efficacy to the triptan abortive drugs. Unfortunately, they have yet to emerge into clinical medicine. Another CGRP receptor antagonist (MK-3207) was also terminated due to liver test abnormalities. These drugs are monoclonal antibodies, and if they follow the pattern of other similar pharmaceuticals, they will likely be exhorbitantly priced if/when they become available.
Topiramate (Topamax). Moderately effective but expensive and category D in pregnancy. (2012)
Topamax is effective in roughly 50% of patients with migraine. This is on the low side for effectiveness. Topiramate also is on the high side for side effects.
Unlike most headache prevention medications, Topiramate often promotes weight loss, even with low doses, due to loss of appetite. Typical doses are 25mg/day to 200 mg/d. In the author’s clinical practice in Chicago, 50-100 mg is the usual target dose, as this amount seems to have the best combination of cost/benefit. Topiramate is expensive and in large doses has peculiar cognitive effects, such as trouble finding words (Mula et al, 2003), and drowsiness. Parethesias (tingling) in the hands and taste perversion (citrus tastes metallic) is common.
Topiramate is not at all a good drug for people whose job involves manipulating words, i.e. writers, speakers, teachers, attorneys (Thompson et al, 2000). Additionally, about 50% of patients develop tingling in hands/fingers on startup. This effect usually fades out in about 2 weeks. Peak effect doesn’t occur till 3 months, so trials must be made over long periods. Like other carbonic anhydrase drugs, topiramate can increase the frequency that kidney stones are developed.
The author has encountered a few patients who became severely depressed on topiramate — they were also on Effexor (venlafaxine), so this may be a drug-drug interaction (reducing venlafaxine). On the positive side, small doses are usually side effect free. Also, topiramate does not affect blood pressure. Topiramate increases the blood levels of amitriptyline.
A recent trial in which topiramate and inderal were combined for chronic migraine, was stopped because of lack of evidence that there was any benefit (Silberstein et al, 2012). Of course, this does not mean that the combination is also ineffective in non-chronic migraine, which comprises about 97% of all migraine.
There have been many reports that topiramate can induce glaucoma, which is of the acute closed-angle type. According to Ho et al (2013), patients prescribed topiramate were found to have a 7.41 fold greater risk of being diagnosed with glaucoma during the first month after it was prescribed. However, this risk became nonsignificant after the first month. From this information, we suggest that new visual symptoms after starting topiramate, especially ocular pain, should be taken very seriously.
Pregancy is category D at present (2011) — D means dangerous. Recent data suggests a high risk of major congenital malformations such as oral cleft disorders (Hunt et al, 2008; Margulis, 2012), as is the case with most migraine medications. The odds ratio is roughly 10:1. Thus this drug is generally not suitable for migraine treatment in pregnancy. This can be an issue as the population most interested in this drug are generally women of child-bearing age.
The mechanism for this anticonvulsant drug’s effect on migraine may include pharmacological effects including enhancement of GABA, inhibition of glutamate receptors, sodium channels, and calcium channels. It also has a weak inhibition of carbonic anhydrase. Due to the combination of migraine and carbonic anhydrase activity, it can be particularly helpful in the many people who have both migraine and Meniere’s disease.
Venlafaxine (Effexor). Very effective for migraine prevention.
Please follow the link above for a detailed discussion of venlafaxine for migraine.
Verapamil (Calan ). Used for prevention. Effective, cheap, mild side effects, interacts with other drugs.
Relatively few trials have been made of calcium channel blockers, the majority of which were of nimodipine. The drug that we favor — verapamil, has largely escaped study. According to Markley (1991), verapamil “may be as effective as traditional therapies”. Greenberg (1986) also suggests that verapamil as well as nifedipine, nimodipine, and diltiazem are helpful. This suggestion is questionable as this is a very old paper, and also we have not had much success with calcium channel drugs than verapamil. Olesen (1986) suggested that studies he reviewed were unconvincing. Although there is good evidence for effectiveness of another drug with calcium channel blocking channel blocking activity, flunarizine, its main effects are likely exerted through dopamine blocking.
In our experience, verapamil is a very effective and inexpensive drug for migraine that takes about 2 weeks to work.We particularly favor this drug for persons who have high blood pressure, or who have nausea accompanying migraine. It is an excellent drug for “cyclic vomiting”. Usual dose is 120 to 240mg per day, SR or ER. SR means sustained release. We start with dose in mg roughly = weight of patient (in pounds). In other words, someone who weighs 120 lbs, would start on the 120 mg dose. We usually increase the dose if not effective at one month intervals. We do not increase beyond 240 mg/day, and we also do not increase if there is constipation or hypotension.
Verapamil is a member of the L-channel calcium channel blocker family. Other calcium channel blockers are generally ineffective (i.e. nicardipine, nifedipine), and some seem to increase migraines (the vasodilators). About 50% of users develop mild constipation. Sometimes lowers blood pressure. Verapamil is a “phenylalkylamine”, while nifedipine and nimodipine are dihydropyiridines (Greenberg, 1986).
About 1% of users develop palpitations (fluttering feeling in chest). It is usually best to stop taking this drug if you develop palpitations. Verapamil is generally safe in patients with asthma (as opposed to the beta-blocker family), and especially good in patients who also have high blood pressure.
Combined use of verapamil or other calcium channel blockers and beta-blockers should, in general, be avoided. Verapamil interacts with simvistatin (and probably other statins) and the dose of the latter should usually be reduced to 10 mg/day when both are prescribed. Verapamil is generally not for use in pregnancy (but it is category C).
Solomon (1983) in a tiny but blinded study of Verapamil in 12 patients reported a 49% effect. Verapamil is reportedly effective in hemiplegic migraine (Yu and Horowitz, 2003). Extremely large amounts of verapamil are sometimes used for Cluster headache. For this situation, heart EKG testing is recommended.
Strangely, a recent “evidence based” review article by Silberstein et al (2012), suggested that verapamil did not work for migraine. We are frankly incredulous about this assertion considering the many patients of ours who use it successfully, the lack of any studies concerning verapamil, and the inhomogeneity of migraine in general. Of course, lack of evidence for effect is not the same of evidence of lack of effect. In other words, although nobody has published a study about a drug working, this does not mean that the drug doesn’t work. It just means that nobody has published a study. This of course is a general critique of “evidence based medicine”, which translates into “no medicine”, if nobody has done a study that provides evidence.
Wellbutrin (buproprion) has been reported useful in small studies for migraine, cluster and chronic daily headaches.
Wellbutrin has no clinically significant effect on serotonin neurotransmission. We have not had much success with this drug.
Zonegran, another anticonvulsant, may have some anti-migraine effects too.
Like Topiramate, it is a carbonic anhydrase inhibitor (among other things). Zonegran may also be associated with weight loss. It is too early to say with this new drug whether it will have a role in migraine prevention. As Zonegran is more expensive, we have not had much experience with it. Zonegran is a close relative of topiramate, and one would expect it would have the same (very long) side effect profile. Topiramate is especially a bad drug for cognitive workers — we would expect zonegran to be similar.
Medications that are reportedly not effective in preventing migraine include acebutolol, clomipramine, clonazepam, clonidine, indomethacin, nicardipine, nifedipine, and pindolol (Silberstein, 2000).