Abstract
It is not easy to evaluate the influence of oral contraceptives (OCs) on migraine. Many studies do not provide a clear description of the headaches that may occur in women who are taking OCs, which makes it impossible to differentiate migraine from other types of headache. While few studies differentiate between the two types of OCs available—the progestogen-only pill and the combined oral contraceptive pill—most studies refer to combined oral contraceptives (COCs), as these account for the majority of oral contraceptives.
Headache and OCs
The International Headache Society (IHS) classification of ‘Headache associated with substances but with uncertain mechanism’ (8.5.1), states: ‘Birth control pills or oestrogens: the literature on this subject is conflicting. Further study is needed’ (1).
Headache is a common side-effect reported with OC use. The relationship of the OC to the headache is not always certain. Methodological biases are numerous: the time between the onset of OC use and headache occurrence and the duration of observation are often not recorded.
Few placebo control trials are available. In one uncontrolled study, headache affected about 10% of women who reported having no headache before starting high-dose OCs (≥ 50 μg ethinyloestradiol) (2). In contrast, four double-blind placebo-controlled studies found no difference between OC and placebo in headache incidence (3).
Differences in the dose and type of oestrogen and progestogen also affect headache. Modern OCs contain ≤ 30 μg ethinyloestradiol. Studies of OC using levonorgestrel, a second-generation progestogen, report headaches in approximately 10% of all cycles (4). In contrast, the newer third-generation progestogens have less reported headache. A review of the COCs containing 30 μg of ethinylestradiol and 150 μg of desogestrel (5) found headache incidence to be only approximately 5% at the sixth cycle. Reducing the dose of oestrogen also reduces headache. With the use of 20 μg ethinylestradiol and 150 μg desogestrel, headache incidence was < 2% by the sixth cycle (6).
Influence of oral contraception on the course of migraine
The influence of oral contraception on the course of migraine is highly variable (7–14). Headaches that worsen in frequency or severity have been reported in 18–50% of cases, with most attacks occurring during the drug-free interval of the cycle. Migraine improvement has been reported in 3–35% of women and there has been no change in 39–65%. The variability has multiple explanations: (i) in older studies, the IHS criteria for migraine were not used; (ii) women were using COCs containing 50 μg of ethinylestradiol (EE). In more recent studies, the EE content is usually ≤ 30 μg; (iii) there is no definition, in most studies, of the worsening or improvement of migraine.
Migraine can begin during OC use, usually during the first cycles, and sometimes after prolonged use. After prolonged use, the responsibility of the pill is more questionable. New-onset migraines occur more often in women with a family history of migraine. However, Kudrow (10) found a family history less frequently in women whose migraine began with OC use (40%) than in those who had had migraine before OC use (72%). This suggests that not only can COCs induce attacks in susceptible persons, but can also, in some cases, play a direct role in initiating migraine. Stopping OCs when migraine starts with their use may have a variable effect: the relief may occur immediately, or after a delay of 6–12 months, and in some patients migraine can continue on a long-term basis.
In the absence of large comparative trials to determine whether some COCs influence the course of migraine in a more favourable way, empirical strategies are necessary to determine the best solution for each patient.
There is a difference in hormonal responsiveness between migraine without aura and migraine with aura (15, 16). Migraine without aura is more likely to appear for the first time at puberty and to be triggered by declining oestrogen levels such as occur with menstruation. Migraine with aura is more likely to begin with sustained high sex hormone levels during pregnancy or with oestrogen replacement (17). In the study done by Cupini (7), the influence of OCs was not significantly different in those patients who had only migraine without aura and those who had migraine with aura (who may also have attacks without aura); however, the analysis was done on patients and not on attacks. Further surveys are necessary to study the responsiveness of both types of migraine to OCs.
Possible mechanisms for increased migraine in women taking OCs
Little research has been undertaken on the possible mechanisms for the adverse changes in migraine that are associated with OC use. They might be similar to the migraine mechanisms in non-OC users.
Mazal reported on a case of a woman who developed typical migraine with aura after starting OCs. This was associated with increased platelet aggregability, which gradually returned to normal when OCs were stopped (18). Hanington studied women who developed migraine with aura after starting OCs and women who reported exacerbation of pre-existing aura. In both groups, she noted increased platelet aggregation to 5-HT that declined over 4–6 months, when OCs were stopped, paralleling a gradual improvement in migraine (19). Platelet aggregation has been implicated in aura due to the production of platelet microemboli in the pial vessels of the visual cortex (20, 21). The mechanism of attacks of migraine without aura that occur during the pill-free interval is easier to account for, as it is probably similar to a recognized mechanism of menstrual migraine, i.e. oestrogen withdrawal. However, this supposition is based on the empirical evidence that such attacks can be prevented by continuous OC use or by using natural oestrogens during the pill-free interval (see below).
Treatment of migraine in women on OCs
Treatment of OC users does not differ from treatment of women who are not on OCs. It relies first on acute treatment: i.e. analgesics, non-steroidal inflammatory drugs (NSAIDs), ergot derivatives and triptans (22–25). When a new anti-migraine drug is developed, phase 2 and 3 trials only include women using effective contraception, which is usually an OC. New anti-migraine drugs are not contraindicated in these patients. Oral contraceptives lower naratriptan clearance by 30%, but this does not make a difference in the tolerance or efficacy of this drug, and it is not necessary to adjust the dosage in women taking COCs.
Short-term prophylactic perimenstrual treatments have been proposed for from 3 to 7 days before menstruation until 1–6 days after. Various prophylactic drugs have been used, such as NSAIDs (26), beta-blockers, calcium antagonists, DHE, and magnesium, and acute treatments, such as ergotamine tartrate and sumatriptan (27), but few controlled studies have been undertaken to assess their efficacy for ‘menstrual’ migraine and none for the specific indication of migraine in pill-free interval of OCs. Commonly used prophylactic treatments have no pharmacokinetic interactions with OCs. Sodium valproate, which is the only anti-epileptic drug that has documented efficacy in migraine prophylaxis, is not associated with significant enzyme induction. Effective contraception is mandatory in fertile migranous women taking sodium valproate, because of the risk of fetal abnormalities. There have been a few studies of carbamazepine and phenytoin in migraine prophylaxis, but their efficacy has not been formally evaluated, and their enzyme-inducing action decreases oestrogen levels in women on the pill, which increases the risk of pregnancy.
Migraine is triggered by oestrogen withdrawal during the late luteal phase of the normal menstrual cycle (28, 29), or discontinuation of the synthetic EE in the pill-free interval. Oestrogen replacement during the pill-free interval is a logical approach to treating migraine occurring at this time. Continuous use of OCs, while well tolerated by some women, can induce irregular uterine bleeding, typically during the 4th cycle. Some women find it acceptable to take three consecutive packets of active pills followed by a pill-free interval. This means that they will have only five pill-free intervals each year rather than the usual 13. Although this regimen is popular in some countries, there are no clinical trial data to support its efficacy.
The advantage of using natural oestrogens alone (rather than continuous OCs) is that they allow the woman to have a break from OCs and withdrawal bleeding (due to progesterone withdrawal), providing reassuring evidence that she is not pregnant. The best treatment of menstrual migraine has been percutaneous estradiol gel, which was found superior to placebo in two double-blind control studies (30, 31). Treatment is started 48 h before the anticipated migraine attack and used daily for the next 7 days, at a dose of 1.5 mg of estradiol, which brings the mean estradiol plasma level to 80 pg/ml. Percutaneous estradiol is the treatment of choice in women who are taking oestro-progestogens, since the expected date of the attack is usually precisely known. It is usually given during the 7 days off the pill. This treatment has been shown to be most successful for pure menstrual migraine (32); the efficacy is lower in women who have menstrual migraine in addition to other attacks at any time during the cycle. Transcutaneous estradiol patches have also been studied in menstrual migraine prevention. Three dosages are available: TTS 25, which delivers 25 μg of estradiol per 24 h; TTS 50, which delivers 50 μg of estradiol per 24 h; and TTS 100, which delivers 100 μg of estradiol per 24 h. The estradiol serum levels reached with these patches are 23 pg/ml, 39 pg/ml and 74 pg/ml, respectively. TTS 50 was not found superior to placebo for the prevention of menstrual migraine in two double-blind trials (33, 34), which may be due to inadequate serum levels. TTS 100 was found superior to TTS 25 in one open trial (35). Clinical trial data of a study of oestradiol 50 μg per 24 h during the pill-free interval are not yet available.
Should the pill be stopped if migraine worsens with OC use? This important question has to be viewed in terms of the risk of ischaemic stroke.
When are OCs contraindicated?
Migraine and OCs are associated with an increased risk of ischaemic stroke (36–48). Since the absolute risk of ischaemic stroke is very low in young migrainous women (19 per 100 000/year, vs. six per 100 000/year in non-migraineurs), and the COCs currently used contain low doses of EE, their use in migraine represents a minimal risk of stroke; thus, migraine is not a contraindication to COC use in most women.
When the risk of ischaemic stroke is high, one may decide not to prescribe, or to stop, the COC. The main factors that increase the risk of ischaemic stroke are as follows:
Type of migraine: Migraine with aura increases the relative ischaemic stroke risk twice as much as migraine without aura. The differential risk due to the type of migraine in an OC user is not known, but it is likely to be greater for migraine with aura. It is not known whether the risk is different for women with pre-existing migraine with aura who start OCs compared with women who develop aura for the first time associated with OC use. Transient ischaemic attacks and migraine with aura cannot always be distinguished (20). Patients with complex or prolonged auras are believed to be at higher risk than those with simple auras, but this assumption is based more on opinion than on objective data.
Patient's age: Ischaemic stroke incidence increases significantly with age. The absolute risk is 10 times as high at the age of 40 as it was at 20.
Tobacco use: Heavy smoking (> 20 cigarettes a day) increases the relative risk of stroke in migraineurs (45).
Other vascular risk factors: Hypertension, diabetes, hyperlipidaemia, and obesity.
An increase in attack frequency or severity with OCs is itself an indication for stopping OC, whether or not it is associated with an increased risk of stroke (49). A change in the character of attacks after starting OC use is possibly of greater concern (50), but the evidence is conflicting (51). The thresholds of migraine severity and frequency that represent the contraindication or discontinuation of OCs are still to be determined, and are different for migraine with and without aura.
Recommendations for OC use in migrainous women have not been established. Some authors state that OCs should be discontinued if a migraine aura develops for the first time on the pill (52); others prefer to know the frequency of attacks with aura. Both the WHO and UK recommendations consider pre-existing migraine with focal neurological symptoms, even typical aura, to be a contraindication to OC use (53, 54).
Until well controlled prospective trials are available and guidelines developed, a decision has to be taken in each separate case. The risks and benefits of oral contraception should be discussed and a joint decision made with the patient. Oral contraceptives have, for many years, been one of the most effective contraceptive methods available. However, other methods, some of which are more effective, are now available and should be considered, as they are not associated with an increased risk of ischaemic stroke and may be better suited to the needs of some patients (55). Migraineurs who smoke heavily or have multiple thrombotic risk factors should be advised not to take COCs. When patients with migraine are placed on COCs, they should be carefully monitored. If migraines worsen or if there is new-onset migraine related to OC use, one should take into account the patient's age, the type of migraine, the frequency and severity of attacks, and the presence of other vascular risk factors. Before discontinuing OC, one may decide to modify the anti-migraine treatment by, for example, introducing a prophylactic drug. One must be more restrictive in women with migraine with aura, in smokers, and in older women, as the risk of stroke is higher in these individuals.
Any unusual headache that has a sudden onset, a long duration, or is associated with focal neurologic symptoms that differ from typical aura should prompt the immediate discontinuation of OCs, and appropriate neurologic investigations to rule out a cerebrovascular complication should be considered.