Rationale: Starting within the first 7 days lowers the possibility of beginning the pill while she is already pregnant (although there is the possibility that the client is pregnant and implantation bleeding has been mistaken for menses) (66,104, 238).

Recommendation: For a woman having menstrual cycles, no back-up method is needed if she is in the first 7 days of her menstrual cycle and is still menstruating. If she is in the first 7 days of her cycle but is not menstruating, some programs may recommend use of a back-up method for 1 week.

COCs may be started anytime you can be reasonably sure the woman is not pregnant (see "How to Be Reasonably Sure the Woman Is Not Pregnant"). However, if COCs are started after day 7 of a regular cycle, the woman should also be counseled that:

• Her regular bleeding pattern may be altered, and
• A back-up method (or abstinence) should be used for 7 days.

Rationale: A back-up method is NOT needed if the first package of pills is started while the woman is menstruating because the risk of conception is virtually nil. After day 5 of the cycle, the risk of pregnancy begins to rise (257).

Some programs might recommend a back-up method for women who are not menstruating at the time of COC initiation because there is a very slight risk of conception from unprotected intercourse on day 7 of the cycle.

When back-up (or abstinence) is needed, it must be used for 7 days because 7 days of exposure to COCs are required to suppress follicular development (195).

Recommendation: If the client is using the 28-day pill packet, she should start a new packet the day after she finishes the previous packet (without a break). If the client is using the 21-day pill packet, she should skip 7 days before starting a new packet. If the pills are taken correctly, the client will always begin a new packet on the same day of the week.

Rationale: The longer the pill-free interval, the higher the risk of ovulation (e.g., a 10-day pill free interval confers a 10% risk of ovulation) (150, 168).

Recommendation: COCs should not be used in the first 6 weeks postpartum. COCs are considered by many experts to be the method of LAST choice during any state of lactation, especially in the first 6 weeks to 6 months. After 6 to 8 weeks postpartum, breastfeeding women desiring hormonal contraception should be encouraged to use progestin-only pills (POPs) or injectables or Norplant® implants. (Before 6 to 8 weeks postpartum, there is no risk of conception for a fully or nearly fully breastfeeding woman; see Lactational Amenorrhea Method.)

Rationale: Even low dose (30 to 35 micrograms of estrogen) COCs decrease breastmilk production (311).

Recommendation: If COCs remain the method of choice, but the woman chooses to rely on the lactational amenorrhea method (LAM) initially, start COCs when her menses return,* or when the woman is no longer fully or nearly fully breastfeeding or at 6 months postpartum, whichever comes first. COC packets may be given to the woman before this time to ensure that she is able to start the method when she needs to.

Rationale: For fully breastfeeding women, there is no known advantage to initiating COCs during LAM or while the LAM criteria apply (141, 156).

In fact, initiating COCs before they are necessary may be a disadvantage because COCs have a detrimental effect on breast milk volume and composition, which may affect the infant's health and growth (309, 311).

* In breastfeeding women, bleeding in the first 56 days (8 weeks) postpartum is NOT considered "menstrual" bleeding because it is not preceded by ovulation.

Recommendation: If she does not want to rely on LAM but is breastfeeding, she should be advised to choose a nonestrogenic method. If she still makes an informed choice to use COCs, they can be started any time after the first 8 to 12 weeks postpartum if she is still amenorrheic, or whenever the service provider can be reasonably sure that the woman is not pregnant (see "How to Be Reasonably Sure the Woman Is Not Pregnant").

Rationale: Even low dose (30 to 35 micrograms of estrogen) COCs decrease breastmilk production. Waiting at least 8 to 12 weeks postpartum permits breastfeeding to be better established. Whether exposure of the neonate (in the first 8 weeks) to exogenous estrogens and progestins may, in theory, affect neonatal growth and development is a question under study.

Rationale: The increased risk of venous thromboembolism associated with COCs may be important for women in the immediate postpartum period. However, blood coagulation and fibrinolysis are essentially normalized by 3 weeks postpartum (and are close to normal at 2 weeks postpartum) (51).

Rationale: Ovulation returns almost immediately postabortion (spontaneous or induced): within 2 weeks for first-trimester abortion and within 4 weeks for second-trimester abortion. Within 6 weeks of abortion, 75% of women have ovulated (164).

Immediate use of COCs postabortion (spontaneous or induced) does not affect return to fertility following discontinuation of COCs (161).

Recommendation: If a client has a history or current indication of excessive clotting (coagulopathy), COCs should not be recommended.

Rationale: COCs may be safely started within the first week postabortion (spontaneous or induced). Hypercoagulability of pregnancy probably does not become clinically significant until the third trimester. However, some experts recommend starting COCs exactly one week postabortion, since there is a suggestion of a slight increase in coagulation factors measurable in the first few days after first-trimester abortion, in women initiating COCs immediately postabortion. If started later than one week, COCs may not be immediately effective because the ovary resumes follicular development as soon as one week after first-trimester (spontaneous or induced) abortion (162, 163).

Incomplete abortion may also result in a condition of excessive blood clotting (disseminated intravascular coagulation), in which estrogens should be avoided.

Rationale: A rest period would disrupt the woman's preferred and successful method of contraception.

Recommendation: Stopping COCs 2 weeks before major elective surgery or after serious accidents that necessitate immobilization of the legs and resuming COCs once the woman is mobile is optimal, if she has a reliable alternative method.

Rationale: Due to the fact that estrogen may slightly increase the risk of postoperative thrombosis, it may be reasonable to stop COCs for 2 weeks before major elective surgery and resume COCs once the woman is mobile, before she resumes sexual activity. However, this small risk must be weighed against the risk of pregnancy and whether the client has a reliable alternative method (224).

Recommendation: Women over age 40 can take COCs, provided other risk factors have been considered (e.g., smoking, high blood pressure, diabetes).

Rationale: Cardiovascular risks from COC use are minimal in healthy, nonsmoking, older women (106, 262).

Recommendation: Use of COCs does not compromise future fertility.

Rationale: On average, the return to fertility after discontinuing COCs is about 2 months longer than for nonhormonal methods. The risk of amenorrhea after discontinuing COCs is small and more common in women who had irregular menses prior to COC use. Rather than causing "post-pill amenorrhea," COCs mask the irregular pattern by inducing cyclic withdrawal bleeding. Women who have irregular menses are more likely to develop secondary amenorrhea whether they take COCs or not (4, 29, 130, 262).

Rationale: Rifampin/rifampicin and griseofulvin require use of a back-up method (or increased COC dose if back-up is not possible) to compensate for hepatic micro-enzyme induction. Hepatic micro-enzyme induction by rifampin lasts for 4 weeks for short-term use and for 8 weeks for long-term use. Although anecdotal reports of failure to prevent pregnancy exist for other antibiotics, epidemiologic evidence suggests that antibiotics (except rifampin and griseofulvin) do not require a back-up method (204).

• Switch to Depo-Provera® or an effective nonhormonal method;
• A back-up method (for short-term anticonvulsant use);
• Higher-dose COCs (i.e., 50 micrograms ethinyl estradiol (EE), or two 30 to 35 micrograms EE COCs per day) for more efficient contraception and/or to produce regular menses without breakthrough bleeding.

Taking two 30 to 35 micrograms COCs per day will provide adequate estrogen to compensate for increased metabolism. Levonorgestrel levels are also reduced by phenytoin (and presumably other anti-epileptics). Therefore, doubling up on COCs that contain levonorgestrel is particularly important (204).

Rationale: The COC effect on cervical mucus is not as strong as the effect of progestin-only methods. COCs require 7 days to suppress follicular development (260).

Rationale: If 2 or more pills are missed, a back-up must be used until the client has taken 7 active pills. Missed pills may occur at the beginning of the cycle (extending the pill-free interval from 7 to 9 days and perhaps allowing escape ovulation to occur) (75, 150).

Seven days of exposure to COCs are required to suppress follicular development (107, 195).

Rationale: Acute vomiting and severe diarrhea may interfere with the effectiveness of the pill. In these cases, a back-up method is reasonable (204, 205).

Rationale: Antimalarials studied to date have not been found to decrease the efficacy of COCs. Chloroquine and primaquine have not demonstrated an effect on plasma COC hormonal levels or on ovulation inhibition. Tetracycline (which is used at low dosage in combination with quinine) has not been found to compromise the effect of COCs (15, 50, 108, 197).

A back-up method is not required unless the woman is switching to injectables or Norplant implants. The provider may want to recommend that she continue to take her COC the day she gets her first injection or the implants.

Some clinicians recommend that the woman finish her pack of pills to delay the onset of her next bleed.

Rationale: Injectables and Norplant implants are most likely effective within 24 hours, unless the woman already has fertile cervical mucus. The woman should take her pill as a back-up if she is not menstruating because there is a slight risk of conception from unprotected intercourse during those 24 hours until the injectable or implants become effective (219, 270).

If symptoms or other reasons to suspect pregnancy exist, such as missed pills, evaluate accordingly. If pregnancy evaluation cannot be performed immediately, the client can be advised to continue taking the pills until this evaluation is completed, or she can be referred to a health unit where she can be evaluated.

Rationale: Amenorrhea may be a side effect of COCs. Amenorrhea is not uncommon in women using the low dose pills, 35 micrograms or less of estrogen, due to a lack of buildup of the uterine lining.

While pregnancy is a possibility, COCs are over 99% effective when used correctly.

It is always recommended that a pregnant woman avoid unnecessary medication. However, if the woman is pregnant and is using COCs, there does not seem to be an increased risk of birth defects for the embryo (28, 114, 251).

Community-based distributors (CBD) and other nonclinical family planning providers should use screening checklists to identify conditions for which the woman can receive a limited supply of COCs and also be referred to a clinic. These screening checklists should, ideally, contain only 5 to 10 items.

Rationale: Studies show that COCs may be safely and effectively administered through nonclinical distribution (231, 246, 320).

Recommendation: If complaints or symptoms arise that are of concern to the provider or to the woman (and that may or may not be due to COCs), the woman should be referred to an appropriate facility. If the woman wants to continue COCs, they should be continued unless a serious problem with estrogen (such as excess blood clotting) is suspected.

Rationale: Much harm can be done by stopping COCs unnecessarily (e.g., risks of pregnancy and risks of abortion).

The two doses should total at least 200 micrograms of ethinyl estradiol and 1.0 mg of levonorgestrel (or 2.0 mg norgestrel). This is the Yuzpe method, which is the recommended ECP regimen.

50 micrograms EE pill (e.g., Ovral, Feminal) each with 250 micrograms (0.25 mg) levonorgestrel or 500 micrograms (0.5 mg) norgestrel:

No. of pills in first dose: No. of pills in second dose (12 hours later):

2 2

Rationale: The Yuzpe method is recommended because it has been shown to be approximately 75% effective in preventing pregnancy and because COCs are accessible and safe. The safety and efficacy of alternative methods is now under investigation (274, 291).

The effectiveness calculation of 75% is based on the expected number of pregnancies compared with the observed number of pregnancies. Expected pregnancies are calculated by matching the cycle day of intercourse with expected cycle-day-specific conception rates. Thus, if 100 women have unprotected intercourse once during the second or third week of their menstrual cycles, about 8 would become pregnant. If those same 100 women used ECPs, only 2 would become pregnant (75% reduction) (274).

Recommendation: If pills containing 30 micrograms ethinyl estradiol and 150 micrograms levonorgestrel (or 300 micrograms norgestrel) are used, four tablets should be taken followed by another four 12 hours later.

30 micrograms or 35 micrograms EE pill each with 150 micrograms (0.15 mg) levonorgestrel or 300 micrograms (0.3 mg) norgestrel:

No. of pills in first dose: No. of pills in second dose (12 hours later):

4 4

Rationale: Two doses each consisting of four 30/150 mg pills are recommended because each dose at least meets the minimum of the Yuzpe regimen of 100 mg of ethinyl estradiol and 0.5 mg of levonorgestrel (or 1.0 mg of norgestrel) per dose.

Norgestrel contains two isomers, only one of which is bioactive (levonorgestrel). Thus 0.5 mg levonorgestrel is bioequivalent to 1.0 mg of norgestrel (44, 128, 326).

Rationale: It is theorized that the efficacy of ECPs taken after 72 hours is lower than the efficacy of ECPs taken within the recommended window of 72 hours. Studies thus far have measured only the effectiveness of ECPs up to 72 hours after intercourse. If the regimen is initiated more than 72 hours after intercourse, the failure rate may be increased.

However, if the primary mechanism of action of ECPs is the prevention or delay of ovulation, variations in timing of ECP use in relation to ovulation could result in ECPs being effective for longer than 72 hours. The 72-hour limit is currently being investigated (44,105, 273, 351).

However, when available, anti-emetics may be prescribed with instructions to take them about an hour before the first dose of ECPs, particularly for a woman with a history of nausea and vomiting after taking estrogens.

For anti-emetics to be effective with ECPs, they need to be taken before the onset of symptoms.

Rationale: Approximately 50% of women who take ECPs experience nausea and 20% to 30% vomit. Use of a prophylactic anti-emetic can prevent nausea, but oral anti-emetics are not significantly helpful after nausea has developed (290, 328, 330).

In the case of severe vomiting, some providers recommend that the pills be administered vaginally.

Rationale: An effective dose of the hormones may not have been absorbed into the bloodstream within 2 hours. If vaginal administration is used, blood levels of estrogen and progestin are probably equivalent to oral administration, based on the frequency of estrogen-induced side effects and preliminary studies of effectiveness (44, 128).

Rationale: Severe diarrhea for more than 24 hours may interfere with absorption of ECPs and reduce the effectiveness of the regimen (204).

Rationale: Anticonvulsants, especially hydantoins (e.g., phenytoin), barbiturates (e.g., primidone, phenobarbital), and carbamazepine (nonbarbiturates) lead to increased metabolism, thus eliminating estrogen and progestin in the bile and decreasing the effectiveness of COCs (newly marketed anti-epileptics, such as vigabatrin, lamotrigine, and valproic acid, are not included) (204, 291, 326).

Rifampin/rifampicin (antituberculosis) and griseofulvin (antifungal) cause hepatic micro-enzyme induction, thus reducing blood levels of COCs; it is presumed the effectiveness of the ECP regimen is also reduced (204, 327).

Recommendation: Women taking liver enzyme-inducing drugs, mainly anticonvulsant treatments (phenytoin, phenobarbitol, and carbamazepine) and the antibiotic rifampicin may have to take a higher dose than the recommended ECP regimen, or use the IUD (if appropriate) for effective emergency contraception. However, an increased dose of ECPs may increase the severity or duration of side effects.

Rationale: For women taking anticonvulsants and rifampicin who require emergency contraception, some experts have recommended doubling the ECP dose, or when appropriate, using an IUD for emergency contraception (339, 350).

Recommendation: Since most anticonvulsants are associated with a risk of birth defects, prevention of unplanned pregnancy is particularly important.

Rationale: Almost all anticonvulsants are teratogenic (9, 185).

Recommendation: It is unlikely that broad spectrum antibiotics significantly affect the action of COCs, including ECPs.

Rationale: There is evidence that broad spectrum antibiotics do not decrease COC effectiveness, and so, in the absence of data for ECPs, the experts presume no clinically important effects on ECP use, either (16, 334, 337, 343).

For example, when a woman visits a provider for gynecological care, contraception, or sexually transmitted disease treatment, she can be provided with ECPs and counseled on their use.

Providing ECP information and supplies (or a prescription) in advance may be especially relevant for women relying on barrier methods or periodic abstinence.

Rationale: The Yuzpe regimen is quite safe. If the prescription guidelines are followed by the provider, it is highly unlikely that women would suffer adverse affects from the regimen. In addition, ECPs help protect a woman from pregnancy and abortion, which are more dangerous than ECP use.

Difficulty in getting access to ECPs within 72 hours of unprotected intercourse is a barrier to use. Providing ECP information and supplies (or a prescription) in advance can be convenient for both providers and women, educates women about how ECPs may be of use, eliminates the need for another clinic visit, and ensures that ECPs are available promptly after unprotected intercourse (274, 291, 336, 350, 353).

Oral contraceptives may be initiated immediately after ECP use (with routine screening). With routine screening, some providers also provide DMPA immediately, because of the low risk of pregnancy (2%) following ECP use and the low risk of teratogenic effects; other providers await the start of menses before providing injectable contraceptives.

Once one can be reasonably sure the woman is not pregnant after ECP use (e.g., arrival of menses), long-term methods such as an IUD or Norplant implants can be initiated.

If an IUD is an appropriate choice, it could be used as an emergency contraceptive. According to WHO, copper-bearing IUDs can be inserted up to five days after unprotected intercourse.

Rationale: There are no clinical data indicating that one method is more appropriate than another for use after ECPs. The choice should be made by the client and the provider. If the client was a pill user when she came in for ECPs, the reason for her missed pills should be discussed.

It is always recommended that a pregnant woman avoid unnecessary medication. However, if the woman is already pregnant or becomes pregnant due to failure of ECPs, and has chosen a hormonal method, the best evidence indicates no increased risk of birth defects for the fetus (28, 251, 291, 302).

Use of IUDs for emergency contraception is not recommended for women with an established pregnancy nor for women who are not medically eligible for continuing IUD use (302).

Rationale: It is difficult to know when ovulation occurs in a given cycle, particularly for women with irregular cycles. The risk of conception is highest between 6 days before and 1 day after ovulation (290, 355).