Contraception for Women with Epilepsy


Address correspondence and reprint requests to Dr. Michael D. O'Brien, Guy's Hospital, St. Thomas Street, London, SE1 9RT, U.K. E-mail:


Summary:  Antiepileptic drugs that induce hepatic enzyme activity may alter the metabolism of most hormonal methods of contraception, and this affects the contraceptive regime. This effect should be considered in the choice of both the treatment of the epilepsy and the choice of contraceptive method. This review considers these interactions and offers advice about their management.

Epilepsy affects 0.5–1.0% of the population. About half of these patients are women, many of childbearing age. Many of these women will seek advice about contraception at some time. The choice of contraceptive method is influenced by many factors, and women will expect, and should receive, appropriate information about the various options for safe and effective contraception. This counseling requires a consideration of the age of the patient, their socioeconomic and educational status, the type of epilepsy, and the most suitable antiepileptic drug (AED) treatment. In this short review, we have tried to summarize those factors specific to epilepsy which influence this choice and to make recommendations about contraceptive practice in epilepsy.

In general, women with epilepsy on AEDs may take a hormonal contraceptive if they wish to do so, after a full discussion of the alternatives with their physician. The optimal choice of hormonal contraceptive is affected by whether or not the prescribed antiepileptic medication induces liver enzyme activity (enzyme-inducing anti-epileptic drug—EIAED). The combined oral contraceptive preparations (COC) achieve contraception by giving a sufficient dose of estrogen with a progestogen to inhibit ovulation. The lowest-dose estrogen products contain 20 mcg of estrogen, so in the absence of enzyme induction the usual 30 mcg preparations give some margin of safety. Some progestogen-only preparations also inhibit ovulation—medroxyprogesterone acetate (by injection) (Depo-Provera), desogestrel (Cerazette), and etonogestrel (Implanon).


The problem is that some antiepileptic drugs induce cytochrome P450 hepatic enzyme activity which increases the rate of metabolism of both estrogen and progestogens, and thus lower the blood levels of these hormones, perhaps by 50% or more (Tables 1 and 2). It is therefore very important to know whether a patient is taking an AED which affects hormonal contraception (i.e., an EIAED) before prescribing a hormonal contraceptive, and to give appropriate advice when prescribing an EIAED to women already taking hormonal contraception (Table 3).

Table 1. AEDs that affect hormonal contraception by enzyme induction
  1. Topiramate reduces the level of ethinylestradiol by about 30%, but by a different mechanism, and has a negligible effect below 200 mg/day (1).

Table 2. AEDs that do not affect hormonal contraception
Sodium valproate
Benzodiazepines (clobazam, clonazepam)
Table 3. Methods of contraception that are affected by EIAEDs
Combined oral contraceptive pill
Combined contraceptive patch
Progestogen only pill
Progestogen implant

The coprescription of lamotrigine with COCs is particularly complex. COCs reduce the blood level of lamotrigine by 40–60% if the patient is not already on an EIAED (2), so starting a COC in a patient already on lamotrigine may result in poorer control of the epilepsy. Therefore, the lamotrigine dose should be retitrated; a small increment in the dose of lamotrigine is usually all that is required. In some cases, the COC may cause the recurrence of epilepsy in a patient whose epilepsy was under very good control. Ideally, the increase in the dose of lamotrigine should take place before starting the COC. However, it is not possible to determine by how much the dose should be increased, or even if it is necessary, in any individual patient. Clearly, the alternatives should be discussed with the patient in some detail (3). There is no problem in giving lamotrigine to patients already on a COC, because the dose of lamotrigine may then be titrated to the patient's needs in the usual way.

The following procedure is therefore advised when using lamotrigine; but these procedures are not necessary if the patient is already taking an EIAED.

  • 1When starting a COC in women already on lamotrigine, increase the dose of lamotrigine by up to twice the previous dose. Note that the blood level of lamotrigine may double in the pill-free interval and this may occasionally cause side effects.
  • 2For stopping a COC and continuing treatment with lamotrigine, reduce the dose of lamotrigine by up to one half.
  • 3For starting lamotrigine in women already on a COC, use the usual AED dose escalation protocol to obtain seizure control.

Lamotrigine has no significant effect on ethinylestradiol concentrations, but does reduce levonorgestrel levels by about 20%, which is a potentially significant fall. However, there is no published evidence that lamotrigine alters the efficacy of the COC; in one study breakthrough bleeding was observed in 7 of 16 subjects when lamotrigine was added to patients taking the COC, but there was no biochemical evidence of ovulation.


In those women requiring an EIAED, one of the first three long-acting methods from Table 4 should be considered because they are not affected by EIAEDs (4). If none of these is acceptable, oral contraception can be initiated with preparations containing at least 50 mcg of estrogen. Since the withdrawal from the market of Ovran, there is no suitable 50 mcg preparation available in the U.K. Norinyl-1, which contains 50 mcg of mestranol is not suitable; because mestranol is a prodrug for ethinylestradiol, and is only 75–80% converted, it provides less than 40 mcg of estrogen. It is possible to use a 20 plus a 30 mcg pill to obtain the 50 mcg total, but the estrogens and progestogens in the two pills should be the same, such as Loestrin 20 plus Loestrin 30. However it is easier, less confusing and almost certainly more reliable to use two of the same 30 mcg preparation, such as two tablets daily of Microgynon 30, which is the commonest brand used in the U.K.

Table 4. Methods of contraception that are not affected by EIAEDs
Medroxyprogesterone acetate depot injection (Depo-Provera)
Hormone releasing intrauterine system (IUS) (Mirena)
Other intrauterine contraceptive devices (IUCD), all copper-containing
Barrier methods

Since the usual 7-day pill-free interval weakens the contraceptive effect, it is our recommendation that, in addition to increasing the daily dose of estrogen, the pill should be given in a tricycling regimen (5). These women should take three cycles of the hormonal contraceptive preparation (50–60 mcg of estrogen per day) consecutively, without a break, followed by a shorter pill-free interval of 4 days. This tricycling regimen is the most effective method of assuring contraceptive efficacy. If breakthrough bleeding occurs, it usually settles during the first two to three cycles; if not, contraception cannot be assured and the dose of estrogen may need to be increased, but should not exceed 90 mcg per day . The dose should normally not be increased because of breakthrough bleeding in the first tricycle because there is a good chance that the bleeding will subside during the first 2–3 months. If breakthrough bleeding only occurs during the third cycle, it may help to change to a two-cycle regime though still followed by the shortened 4-day pill-free interval. Additional contraceptive precautions, usually condoms, should be used until a regular regimen has been established.

Because the metabolism of both estrogen and progestogens is affected by the EIAEDs, the higher doses of estrogen should be accompanied by higher doses of the progestogen. Patients are often quite concerned about taking larger doses of hormones, fearing a higher incidence of side effects. They can be reassured on this point, because side effects with the larger doses given in combination with EIAEDs, are comparable to those associated with normal doses administered alone. The interaction between EIAEDs and hormone dose should be explained at the time of the consultation, even if it is not raised by the patient, because the concern may only arise after the patient gets home and has time to think about it.

The failure rate of COC with EIAEDs is about twice that in the general population. If, as currently believed to be the case, failure is often due to noncompliance with the correct regimen (adequate hormone dosage with tricycling), most of these failures should be preventable.

If a women taking an EIAED and the larger dose of estrogen is switched to a nonenzyme inducing AED, the higher dose of estrogen should be maintained for a further complete cycle, because of the carryover effect of enzyme induction.

It should also be noted that the tricycling of COC preparations is not a licensed indication and such prescriptions should be on a “named patient basis” (6).


The Evra patch is also affected by enzyme induction. Using two patches at a time is not advised (4), and hence patches are not suitable for long-term use by women taking EIAEDs.


Progestogens, like estrogen preparations, are also affected by EIAEDs, aggravating other difficulties associated with this form of contraception, particularly breakthrough bleeding. POP-users taking EIAEDs should be advised to consider other forms of contraception. Desogestrel (Cerazette), has an advantage over other oral POPs, especially in younger more fertile women, because it gives nearly 100% inhibition of ovulation. In addition, it provides a 12-h “missed pill window,” compared to the usual 3 h. For some young women on long-term EIAEDs it might be appropriate to take two tablets daily of desogestrel (Cerazette).


EIAEDs are believed to affect the progestogen implant, which contains etonogestrel (Implanon). Women on short-term treatment with EIAEDs are advised to use a barrier method in addition to the implant and for 28 days post-EIAED treatment. During long-term EIAED treatment, continuing the barrier or transfer to an alternative long-acting method (Table 4), with removal of the implant, is usually advised (4). Some satisfied etonogestrel (Implanon) users may resist this recommendation; an alternative unlicensed option (6) would be an additional daily progestogen pill (desogestrel—Cerazette), or an additional barrier method indefinitely, but these suggestions rather defeat the object of the implant and are therefore not recommended (4).


(“The morning after pill”)

The efficacy of the morning after pill is also affected by EIAEDs. Pharmacists should ask patients about what other medications they are taking, and if an EIAED (or self-medication—e.g., with St John's Wort) is mentioned, the patient should be referred to her physician, who should then prescribe levonorgestrel 1.5 mg (Levonelle One Step) initially followed by a second dose (1.5 mg) 12 h later (7).


The problem of enzyme induction does not apply to the use of medroxyprogesterone acetate (Depo Provera), whose metabolism is proportional to hepatic blood flow, suggesting a virtually 100% clearance on first pass through the liver (8,9). Therefore enzyme induction by EIAEDs can have no additional effect and certainly blood levels are not affected. Thus, it is not necessary to alter either the dose or the interval between injections. Irregular, frequent or prolonged bleeding with depot injections is quite common (30% of patients) in women not on EIAEDs, especially in early months of use. This effect, in patients on antiepileptic drugs, may be inappropriately ascribed to enzyme induction. This problem reduces to about 10% after 1 year and only rarely (5%) requires intervention.

A potential problem in using this depot preparation in adolescent girls is suggested by data showing that this treatment impairs the achievement of peak bone mass, which normally occurs in the early twenties. However, the Medicines and Healthcare Products Regulatory Agency advises that depot injection may be used as a first line treatment after discussion of alternatives with the patient, and with reassessment after 2 years (10). There is also a potential problem of weight gain, which is common with these preparations and very unwelcome in teenage girls. However the depot injection remains a useful option, and for some women with epilepsy there is an additional advantage of much less hormonal variation during the menstrual cycle.


(Mirena IUS)

This treatment is not materially affected by EIAEDs (11). It is highly effective and not dependent on good compliance. It is therefore a very suitable option for women with epilepsy, especially those taking EIAEDs, particularly if parous; with good counseling and expertise at the insertion, it may also be suitable for nulliparous women, including adolescents.


The usual advice about the timing of the start of the different contraceptive methods and their suitability after delivery, which should include consideration of breast feeding, is not affected by prescription of EIAEDs. However, women taking EIAEDs should consider the suitability of different methods and follow the dose protocols outlined above.


The effect of liver enzyme induction by some AEDs on most hormonal methods of contraception can be safely and effectively managed using the protocols outlined in this review. The optimal AED for the patient and the type of epilepsy can nearly always be combined with the most acceptable method of contraception, but this decision does require detailed discussion of the advantages and problems associated with the various combinations.