Berek Novak's Gyn 2019. Chapter 14 Family Planning

 CHAPTER 14

Family Planning

KEY POINTS

1 The most common methods of contraception used in the United States are

sterilization, oral contraceptives (OCs), and condoms, in that order.

2 Latex condoms and other barriers reduce the risk of sexually transmitted diseases

(STDs) and cervical cancer.

3 The intrauterine devices (IUDs) available in the United States are as effective as tubal

sterilization and confer no increased long-term risk of pelvic infection.

4 The combination estrogen–progestin OC, patch, and vaginal ring all provide excellent

contraception when used correctly but all can increase the risk of venous thrombosis

and thromboembolism.

5 Present low-dose estrogen–progestin combinations do not increase the risk of heart

attack among nonsmokers younger than age 35 years who have no other risks for

vascular disease.

6 The association between breast cancer and OC use remains controversial, but OCs do

not appear to substantially increase the risk of breast cancer. Furthermore, benefits

of OC use, particularly protection against pregnancy and other noncontraceptive

benefits, outweigh the risks for the majority of women.

7 Use of the progestin-only injectable or implant results in very low pregnancy rates

without the estrogen-associated risk of thrombosis.

8 Hormonal contraceptives provide extensive contraceptive and noncontraceptive

health benefits, including reduced risk for endometrial and ovarian cancer.

9 Levonorgestrel 1.5 mg and ulipristal acetate 30 mg are the most effective hormonal

means of emergency contraception in the United States. Efficacy is greatest within

24 hours of intercourse but remains high at 5 days. The copper T380A IUD within 5

days after unprotected intercourse is even more effective than hormonal methods.

10 Long-acting reversible contraceptive (LARC) methods include subdermal progestin

implants, and copper- or levonorgestrel-releasing IUDs. These offer pregnancy rates

comparable to sterilization and are among the safest methods.

11 Safe and permanent contraception is provided with laparoscopy or mini-laparotomy,

utilizing bipolar electrocautery, salpingectomy or partial salpingectomy, the Silastic

band, or the Filshie clip.

77012 Hysteroscopic sterilization techniques provide highly effective permanent

contraception for women without the use of general anesthesia or abdominal

incision. However, reports of adverse patient reactions have prompted the

manufacturer to remove the product from the market.

13 Vasectomy provides highly effective, low-cost sterilization for men and is not

associated with prostate cancer.

14 Abortion mortality rates fell rapidly with legalization; for safe abortions the overall

mortality risk is less than 1 per 100,000.

15 The risk of abortion mortality increases with gestational age. The mortality rate for

aspiration procedures was 0.3 per 100,000 at or before 8 weeks and increased to 6.7

per 100,000 at 18 weeks or greater.

The history of contraception is a long one, dating to ancient times. The voluntary

control of fertility is especially important in modern society (1). A woman who

expects to have no more than one or two children spends most of her reproductive

years trying to avoid pregnancy. Effective control of reproduction is essential to a

woman’s ability to accomplish her individual goals. From a larger perspective, the

rapid growth of the human population in modern times threatens the survival of

all. The world population is at 7.3 billion and is expected to reach 10 billion by

2056 (2). For both the individual and the planet, reproductive health requires

careful use of effective means to prevent pregnancy and sexually transmitted

infections (3).

From puberty until menopause, women are faced with challenges concerning

childbearing or its avoidance: the only options are sexual abstinence,

contraception, or pregnancy. The contraceptive choices made by couples in the

United States in 2013 are shown in Figure 14-1 (4). [1] Oral contraceptives

(OCs) were the first choice among women, used by 25.9%. Female sterilization

was the second choice, used by 25.1%. With the addition of the 8.2% of couples

relying on male sterilization, 33.3% of couples were relying on sterilization,

making this the first choice of couples. Condoms were third choice, used by

15.3%. OC use declines with age, and the rate of sterilization increases. Fortyseven percent of women under 25 who use contraceptives use OCs and only 1.6%

is sterilized. Of women aged 35 to 44 years using contraception 12.4% uses OCs

and 44.2% is sterilized, as is 17.9% of their consorts. About 10% of women use

more than one method of contraception. Although contraceptive prevalence is

high, a significant proportion of sexually active couples (6.9%) does not use

contraception, and each year, 1.46 of every 100 women aged 15 to 44 years have

an induced abortion (4,5). Abortion is an obvious indicator of unplanned

pregnancy. Abortion ratios by age group indicate that the use of abortion is

greatest for the youngest women and least for women in their late 20s and early

77130s who are most likely to continue pregnancies (Fig. 14-2) (6). The use of

abortion increases from the late 30s on. Young women are much more likely to

experience unplanned pregnancy because they are more fertile than older women

and because they are more likely to have unprotected intercourse.

FIGURE 14-1 Percent distribution of women aged 15 to 44 years who are currently using

contraception, by the type of contraceptive method used. United States, 2011 to 2013.

(From Daniels K, Daugherty J, Jones J, et al. Current contraceptive use and variation by

selected characteristics among women aged 15–44: United States, 2011–2013. Natl Health

Stat Report 2015;[86]:1–14; Figure 1.)

EFFICACY

Factors affecting the likelihood of pregnancy include the fecundity of both

partners, the timing of intercourse in relation to ovulation, the method of

contraception used, the intrinsic effectiveness of the contraceptive method, and

the correct use of the method. It is impossible to assess the effectiveness of a

contraceptive method in isolation from the other factors. The best way to assess

effectiveness is long-term evaluation of a group of sexually active women using a

particular method for a specified period to observe how frequently pregnancy

772occurs. A pregnancy rate per 100 women per year can be calculated using the

Pearl formula (dividing the number of pregnancies by the total number of months

contributed by all couples, and then multiplying the quotient by 1,200). With

most methods, pregnancy rates decrease with time as the more fertile or less

careful couples become pregnant and drop out of the calculations. More accurate

information is provided by the life-table method. This method calculates the

probability of pregnancy in successive months, which are then added over a given

interval. Problems relate to which pregnancies are counted: those occurring

among all couples or those in women the investigators deem to have used the

method correctly. Because of this complexity, rates of pregnancy with different

methods are best calculated by reporting two different rates derived from multiple

studies (i.e., the lowest rate and the usual rate), as shown in Table 14-1.

Medical Eligibility for Contraception

Since 1996, the World Health Organization (WHO) has regularly published

Medical Eligibility Criteria for Contraceptive Use (MEC). These

recommendations are based on the best evidence available supplemented by

expert opinion. The U.S. Centers for Disease Control undertook a formal process

to review and revise the WHO MEC and adapt it to US practice (7). These

eligibility criteria have become de facto US national (CDC) and international

(WHO) standards for contraceptive practice. All present methods of contraception

are assigned to one of four categories of suitability of use by women with more

than 60 characteristics or conditions. The categories are:

1. A condition for which there is no restriction for the use of the contraceptive

method;

2. A condition for which the advantages of using the method generally outweigh

the theoretical or proven risks;

3. A condition for which the theoretical or proven risks usually outweigh the

advantages of using the method;

4. A condition that represents an unacceptable health risk if the contraceptive

method is used.

773FIGURE 14-2 Percentage of total abortions, abortion rate,* and abortion ratio† by age

group of women who obtained a legal abortion—selected reporting areas.§ United States,

2013. *Number of abortions per 1,000 women aged 15–44 years. †Number of abortions per

1,000 live births. §Data are for 47 areas; excludes five areas (California, Florida, Maryland,

New Hampshire, and Wyoming) that did not report, did not report by age, or did not meet

reporting standards. (From Jatlaoui TC, Ewing A, Mandel MG, et al. Abortion

surveillance—United States, 2013. MMWR Surveill Summ 2016;65[12]:1–44; Figure 2.)

Cost

Some methods, such as intrauterine devices (IUDs) and subdermal implants,

require an initial, “up-front” investment, which in some countries can be

substantial, but then provide prolonged protection for a low annual cost. A

complex cost analysis based on the cost of the method plus the cost of pregnancy

if the method fails concludes that sterilization and IUDs are the least expensive

over 5 years (Table 14-2) (8).

Long-Acting Reversible Contraceptives

[10] Several contraceptive methods are as effective as sterilization, but are

completely reversible. All have the important advantage of being “forgettable,”

774that is, little is required of the user after the method is begun, very much in

contrast to methods like the condom that must be used with each act of

intercourse, or the OC, which must be taken daily. [10] These “forgettable”

methods have typical-use pregnancy rates of less than 1%, are effective for at

least several years without attention from the user, and are among the safest

methods. They include the etonogestrel and levonorgestrel subdermal implants,

copper-bearing IUDs such as the copper T380A and levonorgestrel-releasing

intrauterine systems (9,10).

NONHORMONAL METHODS

Coitus Interruptus

Coitus interruptus is withdrawal of the penis from the vagina before ejaculation.

This method, along with induced abortion and late marriage, is believed to

account for most of the decline in fertility of preindustrial Europe (11). Coitus

interruptus remains a very important means of fertility control in many countries.

Eighty-five million couples are estimated to use the method worldwide, yet it has

received little formal study. This method has obvious advantages: immediate

availability and no cost. The penis must be completely withdrawn from the vagina

and the external genitalia. Pregnancy has occurred from ejaculation on the female

external genitalia without penetration. Efficacy is estimated to range from 4

pregnancies per 100 women in the first year with perfect use to 22 per 100 with

typical use (Table 14-1). Jones and colleagues offer a modern review of this

practice and conclude that it likely is as effective as the condom (12).

Breastfeeding

Breastfeeding can be used as a form of contraception and can be effective

depending on individual variables. The use of contraception during lactation

should take into consideration the women’s needs and the need to maintain

lactation. Ovulation is variably suppressed during lactation. The suckling of the

infant elevates prolactin levels and reduces gonadotropin-releasing hormone

(GnRH) from the hypothalamus, reducing luteinizing hormone (LH) release and

thus inhibiting follicular maturation (13). Even with continued nursing, ovulation

eventually returns but is unlikely before 6 months, especially if the woman is

amenorrheic and is fully breastfeeding with no supplemental foods given to the

infant (14). For maximum contraceptive reliability, feeding intervals should not

exceed 4 hours during the day and 6 hours at night, and supplemental feeding

should not exceed 5% to 10% of the total amount of feeding (15). Six-month

pregnancy rates of 0.45% to 2.45% are reported for couples relying solely on this

method (16). To prevent pregnancy, another method of contraception should be

775used from 6 months after birth or sooner if menstruation resumes. Breastfeeding

reduces the mother’s lifetime risk of breast cancer (17).

Table 14-1 Percentage of Women Experiencing an Unintended Pregnancy During the

First Year of Use of Contraception and the Percentage Continuing Use at

the End of the First Year

Contraception During Lactation (Lactation Amenorrhea)

Previously, the use of combination estrogen–progestin hormonal methods (OCs,

the patch, and the ring) was not advised during lactation because of concern about

reduction in the amount and quality of breast milk. However, clinical studies

demonstrate conflicting results regarding effects on breastfeeding continuation or

exclusivity in women exposed to COCs during lactation and the use of combined

estrogen–progestin hormonal methods are category 2 in breastfeeding women

776beyond 30 days of delivery (7). Progestin-only OCs, implants, and injectable

contraception do not affect milk quality or quantity (15). Barrier methods,

spermicides, and the copper T380A IUD (ParaGard) are excellent options for

nursing mothers. Few studies have specifically looked at the use of

levonorgestrel-IUDs (LNG-IUDs) on breastfeeding. One randomized controlled

trial (RCT) raised concerns that immediate insertion of the LNG-IUD postpartum

may be associated with poorer breastfeeding performance when compared with

delayed insertion (18). However, while more studies are needed before definitive

recommendations can be made, the safety and impact on breastfeeding with the

use of other contraceptives such as implants (which produce higher blood levels)

have been previously shown (19). Given the comparatively low level of

circulating hormone in LNG-IUDs, no adverse effect on breastfeeding is

anticipated.

Table 14-2 Cost Effectiveness (C/E) of Contraceptive Methods at 5 Years

Fertility Awareness

Periodic abstinence, described as “natural contraception” or “fertility

awareness,” requires avoiding intercourse during the fertile period around

777the time of ovulation. A variety of methods are used: the calendar method, the

mucous method (Billings or ovulation method), and the symptothermal method,

which is a combination of the first two methods. With the mucous method, the

woman attempts to predict the fertile period by feeling the cervical mucous with

her fingers. Under estrogen influence, the mucous increases in quantity and

becomes progressively more slippery and elastic until a peak day is reached. The

mucous becomes scant and dry under the influence of progesterone until the onset

of the next menses. Intercourse may be allowed during the “dry days”

immediately after menses until mucous is detected. Thereafter, the couple must

abstain until the fourth day after the “peak” day.

In the symptothermal method, the first day of abstinence is predicted either

from the calendar, by subtracting 21 from the length of the shortest menstrual

cycle in the preceding 6 months, or the first day mucous is detected, whichever

comes first. The end of the fertile period is predicted by the use of basal body

temperature. The woman takes her temperature every morning and resumes

intercourse 3 days after the thermal shift, the rise in body temperature that signals

that the corpus luteum is producing progesterone and that ovulation occurred. The

postovulatory method is a variation in which the couple has intercourse only after

ovulation is detected.

A system of hormone monitoring designed to better define the fertile period

involves placement of disposable test sticks in a small battery-powered device to

detect urinary estrone-3 glucuronide and LH. Changes in these hormones reliably

predict the fertile period. These devices (Persona and Clearblue Easy Fertility

Monitor [CEFM]) can serve as aids either to becoming pregnant or avoiding it

(20). Persona is marketed in Europe for contraception. It is reported to have a

correct-use effectiveness of 94% in avoiding pregnancy. CEFM is approved in the

United States as an aid to becoming pregnant but is used off label to avoid

pregnancy. Exact effectiveness rates are not known. The CEFM is used in the

Marquette method, which combines observation of cervical mucous changes with

the CEFM results. A correct-use pregnancy rate of 2% and a typical-use

pregnancy rate of 12% were reported (21).

Efficacy

The ovulation method was evaluated by the WHO in a five-country study.

Women who successfully completed three monthly cycles of teaching were

enrolled in a 13-cycle efficacy study. There was a 3.1% probability of pregnancy

in 1 year for the small proportion of couples who used the method perfectly and

86.4% probability of pregnancy for the rest (22). A review of 15 national surveys

from developing countries estimated a 12-month gross failure rate of 24

pregnancies per 100 (23). One systematic review of RCTs that examined fertility

778awareness-based methods found that most participants in trials discontinued

prematurely and thus comparative efficacy of these methods remains unknown

(24).

Risks

Conceptions resulting from intercourse remote from the time of ovulation

more often lead to a higher likelihood of spontaneous abortion than

conceptions from midcycle intercourse (25). Malformations are not more

common.

Condoms

In the 1700s, condoms made of animal intestine were used by the aristocracy of

Europe, but condoms were not widely available until the vulcanization of rubber

in the 1840s (1). Modern condoms usually are made of latex rubber, although

condoms made from animal intestine are still sold and are preferred by some who

feel they afford better sensation. New condoms made from nonlatex materials—

such as polyurethane or synthetic elastomers that are thin, odorless, transparent,

and transmit body heat—are available. Although the nonlatex condoms may

break more easily than the latex varieties, substantial numbers of study

participants preferred them and would recommend them to others (26).

The risk of condom breakage is about 3% and is related to friction (27). The

use of water-based lubricants may reduce the risk of breakage. Petroleum-based

products such as mineral oil must be avoided because even brief exposure to them

markedly reduces the strength of condoms (28).

Sexually Transmitted Diseases

Gonorrhea, ureaplasma, and pelvic inflammatory disease (PID) and its

sequel (tubal infertility) are reduced with consistent use of barrier methods

(29–31). [2] Tested in vitro, Chlamydia trachomatis, herpes virus type 2, HIV,

and hepatitis B did not penetrate latex condoms but did cross through condoms

made from animal intestine (32). Follow-up of sexual partners of HIV-infected

individuals showed that condom use provides considerable protection (33).

Consistent condom use provides more protection than inconsistent use (34). In

one study, couples who use condoms 0% to 50% of the time had an HIV

seroconversion rate of 20.8 per 100 couple years, whereas those who used

condoms 100% of the time had a conversion rate of only 2.3 per 100 couple years

(35). Nonoxynol-9 (N-9) should not be used with condoms for HIV protection

because it is associated with genital lesions. N-9 does not add to the protection

afforded by condoms alone (36).

779[2] Condoms offer protection from cervical neoplasia (37). The relative risk

(RR) for invasive cervical cancer was 0.4 when those who used condoms or

diaphragms were compared with those who never used them (38). The

presumed mechanism of protection is reduced transmission of human

papillomavirus (HPV).

Risks

Latex allergy could lead to life-threatening anaphylaxis in either partner from

latex condoms. Nonlatex condoms of polyurethane and Tactylon should be

offered to couples who have a history suggestive of latex allergy.

Female Condom

The original female condom introduced in 1992 was a polyurethane vaginal

pouch attached to a rim that partly covered the vulva. The FC2 female condom

was U.S. Food and Drug Administration (FDA) approved in 2009 and is an

improved version made from softer synthetic latex that does not require hand

assembly during manufacture and is therefore less expensive (39). It consists of a

loose-fitting nitrile sheath with two flexible rings. It is recommended for

prevention of pregnancy and STDs, including HIV. Although breakage may occur

less often with the female condom than the male condom, slippage appears to be

more common, especially for those new to its use (40). Male condoms should not

be used with female condoms as concurrent use increases the risk of breakage.

Exposure to seminal fluid is slightly higher than with the male condom (41).

Initial US trials showed a pregnancy rate of 15% in 6 months. Subsequent

analysis found that with perfect use, the pregnancy rate may be only 2.6%. This

rate is comparable to perfect use of the diaphragm and cervical cap, the other

female-controlled barrier methods (42). As with the male condom, failure rates

fall with increasing experience. Colposcopic studies of women using the female

condom demonstrate no signs of trauma or change in the bacterial flora (43).

Vaginal Spermicides

Vaginal spermicides combine a spermicidal chemical, either N-9 or octoxynol,

with a base of cream, jelly, aerosol foam, foaming tablet, film, suppository, or a

polyurethane sponge. N-9 is a nonionic surface-active detergent that immobilizes

sperm. N-9 spermicides alone appear considerably less effective in preventing

pregnancy than condoms or diaphragms. Women using N-9 spermicides

frequently have higher rates of genital lesions than women not using spermicides.

These lesions may increase their risk for STDs and HIV (44). In the same studies

of serodiscordant couples in which condoms were proven effective in preventing

transmission of HIV, N-9 spermicides alone were not effective (35).

780Concerns were raised about possible teratogenicity of spermicides. N-9 is not

absorbed from the human vagina (45). Several large studies found no greater risk

of miscarriage, birth defects, or low birth weight in spermicide users than in other

women (46,47).

N-9 is toxic to the lactobacilli that normally colonize the vagina. Women who

use spermicides regularly have increased vaginal colonization with the bacterium

Escherichia coli and may be predisposed to E. coli bacteriuria after intercourse

(48).

Vaginal Barriers

At the beginning of the 20th century, four types of vaginal barriers were used in

Europe: vaginal diaphragm, cervical cap, vault cap, and Vimule. Vaginal

diaphragms, new varieties of cervical caps, and the synthetic sponge are used in

the United States. When used consistently, vaginal barriers can be reasonably

effective. A recent search for alternatives to condoms for HIV prevention in highprevalence areas has rekindled interest in the other vaginal barriers (49).

Diaphragm

Traditional diaphragms consist of a circular dome covered with fine latex rubber

(Fig. 14-3). There are several types of diaphragms, as determined by the spring

rim: coil, flat, or arcing. Coil-spring and flat-spring diaphragms become a flat

oval when compressed for insertion. Arcing diaphragms form an arc or half-moon

when compressed; they are easiest to insert correctly. Traditionally, diaphragms

require fitting by practitioners. Spermicide is always prescribed for use with the

diaphragm; whether this practice is necessary is not well studied.

Fitting Diaphragms

Fitting a diaphragm should be performed as follows:

1. A vaginal examination should be performed. With the first and second

fingers in the posterior fornix, the thumb of the examining hand is placed

against the first finger to mark where the first finger touches the pubic bone.

The distance from the tip of the middle finger to the tip of the thumb is the

diameter of the first diaphragm that should be tried.

2. A set of test diaphragms of various sizes is used, and the test diaphragm is

inserted and checked by palpation. The diaphragm should open easily in the

vagina and fill the fornices without pressure. The largest diaphragm that fits

comfortably should be selected. A size 65, 70, or 75 diaphragm will fit most

women.

3. The patient should practice insertion and should be reexamined to

781confirm proper position of the device. About 1 teaspoon of water-soluble

spermicidal jelly or cream is placed in the cavity of the dome. The diaphragm

is inserted with the dome downward so that the cervix will sit in a pool of the

spermicide.

4. The diaphragm can be inserted several hours before intercourse. If

intercourse is repeated, additional spermicidal jelly should be inserted

into the vagina without removing the diaphragm. The diaphragm should

be left in place at least 6 hours after intercourse to allow for

immobilization of sperm. When removed, it is washed with soap and water,

allowed to dry, and stored away from heat. It should not be dusted with talc

because genital exposure to talc may increase the risk of ovarian cancer.

In 2014, the FDA approved the Caya contoured diaphragm for use in the

United States. It is a single-size reusable silicone diaphragm. Compared to

traditional diaphragms, this contoured diaphragm does not require fitting by a

practitioner to determine the size, which in low-resource settings may be

especially important for improving access. Caya should be used with a

spermicide. It should be kept in place for 6 hours after intercourse to ensure

sperm cells are no longer active, but removed within 24 hours to prevent toxic

shock syndrome.

Risks

Diaphragm use, especially prolonged use during multiple acts of intercourse,

appears to increase the risk of bladder infections. A smaller-sized, wide-seal

diaphragm or a cervical cap can be used if recurrent cystitis is a problem,

although the problem may relate not only to mechanical obstruction but also to

alterations in vaginal flora produced by the spermicide. An epidemiologic study

comparing cases of toxic shock with controls found no increased risk from

diaphragm use (50).

Other Barriers

FemCap

This version of the cervical cap made of silicone rubber was approved by the

FDA in 2003. It looks like a sailor’s hat with the dome covering the cervix and

the brim fitting into the vaginal fornices. It is made in three sizes—22-, 26-, and

30-mm diameter—and is expected to be reusable for 2 years. It is used with

spermicide and should be left in place for at least 6 hours after intercourse, but it

may be left in place as long as 48 hours at a time. Additional acts of intercourse

require insertion of more spermicide. FemCap requires a clinician’s fitting and

prescription for use. The only available efficacy study compared the FemCap to

782the vaginal diaphragm. The 6-month pregnancy rate for the FemCap was 13.5%,

substantially higher than the 7.9% rate for the diaphragm. Both groups used N-9

spermicide with the devices (52).

FIGURE 14-3 Wide-seal diaphragm. A: Outer caudal side. B: Inner cephalad side.

Future Development in Barrier Methods

The Ovaprene ring is a nonhormonal “one-size-fits-all” intravaginal organic

silicone ring that continuously releases spermiostatic and spermicidal agents over

a 4-week period. The phase I trial demonstrated patient acceptability and safety

(51). It is expected to undergo a postcoital test clinical trial in 2018.

Sponge

The Today sponge is a polyurethane dome-shaped device containing N-9. It is

moistened with water and then inserted high in the vagina to cover the cervix. It

combines the advantages of a disposable barrier with spermicide and provides

protection for 24 hours. The contraceptive efficacy appears to differ with parity.

Nulliparous women are reported to have a perfect use pregnancy rate of 9% per

year, whereas parous women have a pregnancy rate of 20% (Table 14-1). Rates

with typical use are estimated as 12% per year in nullipara and 24% in multipara

(52). A trial comparing the sponge with a vaginal spermicide preparation used

alone without barrier showed the sponge had a slightly lower pregnancy rate (53).

Intrauterine Contraception

Worldwide over 14% of married women use intrauterine contraception (54).

In the United States, usage is increasing, but the estimates are that only 6.4% of

women use intrauterine contraception (54). Candidacy includes nulliparous

women, adolescents, and immunocompromised women. Immediate use

postpartum or after a first- or second-trimester abortion broadened usage. Five

783IUDs are available in the United States: the copper T380A (ParaGard), two 52-

mg levonorgestrel-releasing IUDs (Mirena and Liletta), a 19.5-mg levonorgestrelreleasing IUD (Kyleena), and a 13.5-mg levonorgestrel-releasing IUD (Skyla).

The copper T380A has bands of copper on the cross arms of the T in addition to

the copper wire around the stem, providing a total surface area of 380 mm of

copper, almost double the surface area of copper in earlier copper devices (Fig.

14-4). It is approved for up to 10 years of continuous use. The Mirena (Fig. 14-5)

is approved in the United States for 5 years of use, although studies through 7

years of use show no loss of efficacy (55). The Kyleena is approved for 5 years of

use, while the Skyla is approved for 3 years. Liletta is manufactured by

Medicines360 and is approved for use up to 4 years, with eventual anticipated

approval for 7 years, and possibly 10 years (56). [3] All IUDs provide safe, longterm contraception with effectiveness equivalent to tubal sterilization.

FIGURE 14-4 Copper T380A (ParaGard) intrauterine device.

Mechanism of Action

IUDs cause the formation of “biologic foam” within the uterine cavity that

contains strands of fibrin, phagocytic cells, and proteolytic enzymes. All IUDs

stimulate the formation of prostaglandins within the uterus, consistent with both

smooth muscle contraction and inflammation. Copper IUDs continuously release

a small amount of the metal, producing an even greater inflammatory response.

Scanning electron microscopy studies of the endometrium of women wearing

nonmedicated IUDs show alterations in the surface morphology of cells,

especially of the microvilli of ciliated cells (57). There are major alterations in the

784composition of proteins within the uterine cavity and new proteins and proteinase

inhibitors are found in washings from the uterus (58). These altered intrauterine

environments interfere with sperm passage through the uterus, preventing

fertilization.

The levonorgestrel in IUDs is much more potent than natural progesterone and

has a profound effect on the endometrium. The hormone is released at an initial

rate of 20 μg daily, which declines to half this rate by 5 years. For the 52-mg

LNG-IUDs, blood hormone levels are significantly lower than with other

progesterone-only contraception and remain stable at approximately 130 to 200

pg/mL (59). About 85% of cycles are ovulatory. The contraceptive effect of the

LNG-IUD is a result of thickened and scant cervical mucous, endometrial

atrophy, and an intrauterine inflammatory response (56,59).

FIGURE 14-5 Levonorgestrel T (Mirena) intrauterine device.

The IUD is not an abortifacient. The contraceptive effectiveness does not

depend on interference with implantation, although this phenomenon can occur

and is, in part, the theoretical basis for using copper IUDs for emergency

contraception. Sperm can be obtained by laparoscopy in washings from the

fallopian tubes of control women at midcycle; fewer sperm are present in the

tubal washings from women wearing IUDs (60). Ova flushed from the tubes at

tubal sterilization showed no evidence of fertilization in women wearing IUDs

(61). Studies of serum β-human chorionic gonadotropin (β-hCG) levels in women

wearing IUDs do not indicate pregnancy (62).

Effectiveness

785The copper T380A and the 52-mg LNG-IUDs have remarkably low

pregnancy rates, less than 0.2 per 100 woman-years. Total pregnancies over a

7-year period were only 1.1 per 100 for the LNG-IUD and 1.4 for the copper

T380A (55). Twelve-year data on the copper T380A showed a cumulative

pregnancy rate of only 1.9 per 100 women and no pregnancies at all after year 8

(63).

Benefits

Modern IUDs provide excellent contraception without continued effort by the

user. The copper T380A and the LNG-IUDs protect against ectopic pregnancy.

The LNG-IUD, by releasing levonorgestrel, reduces menstrual bleeding and

cramping. It is used extensively to treat heavy menstrual bleeding and is used

in Europe and the United Kingdom as an alternative to hysterectomy for

menorrhagia (64). The LNG-IUD has a beneficial effect on menorrhagia from

uterine fibroids; the benefit may be diminished with distorting submucosal

fibroids (65,66). The LNG-IUD is an effective way to deliver the necessary

progestin therapy in postmenopausal women on estrogen therapy (67).

Additional noncontraceptive benefits include a reduced risk of endometrial cancer

and improvement in symptoms of endometriosis and adenomyosis (68–71).

Risks

Infection

The Women’s Health Study found the Dalkon Shield device (withdrawn from the

market since 1974) to increase the risk of PID eightfold when women hospitalized

for PID were compared with control women hospitalized for other illnesses (72).

In contrast, risk from the other IUDs was markedly lower. Increased risk was

detectable only within 4 months of insertion of the IUD. A prospective WHO

study revealed that PID increased only during the first 20 days after insertion.

Thereafter, the rate of diagnosis of PID was about 1.6 cases per 1,000 women per

year, the same as in the general population (73).

[3] Exposure to sexually transmitted pathogens is the more important determinant

of PID risk than IUD use. In the Women’s Health Study, women who were

married or cohabiting and who said they had only one sexual partner in the past 6

months had no increase in PID (72). In contrast, previously married or single

women had marginal increase in risk, even though they had only one partner in

the previous 6 months (74). The only pelvic infection that was unequivocally

related to IUD use is actinomycosis (75). It appears that PID with actinomycosis

was reported only in women wearing an IUD. Rates of colonization with

actinomycosis increase with duration of use for plastic devices but appear to be

much less for copper-releasing IUDs. Actinomyces may be found in cervical

786cytology of up to 7% of women with an IUD. Because of the low positive

predictive value and a lack of sensitivity and specificity of cervical cytology to

diagnose this organism, antibiotic treatment and IUD removal should be reserved

for symptomatic women (76).

When PID is suspected in a woman using an IUD, appropriate cultures should

be obtained, and antibiotic therapy should be administered. Removal of the IUD

is not necessary unless symptoms do not improve after 72 hours of treatment (77).

Pelvic abscess, if suspected, should be ruled out by ultrasound examination.

Ectopic Pregnancy

All contraceptive methods protect against ectopic pregnancy by preventing

pregnancy. But when the method fails and pregnancy occurs, the risk of ectopic

pregnancy is affected by the method of contraception. IUDs and tubal

sterilization increase the probability of the pregnancy being ectopic when

pregnancy occurs, but the rate of any pregnancy is so low that women using

either of these methods have much lower rates of ectopic pregnancy than

women not using contraception (78). Risk of any pregnancy with a LNG-IUD is

between 0.1 and 0.2 per 100 woman-years. The rate of an ectopic in users of this

device is reported as 0.02 per 100 woman-years (79). In a large study of the

copper T380A, the first-year pregnancy rate was 0.5 per 100 woman-years, and

the rate of ectopic was 0.1 per 100 woman-years (80). Ectopic is a very rare event

with either IUD, but should a woman wearing one present with pelvic pain and a

positive β-hCG, an ectopic must be ruled out.

Fertility

Tubal factor infertility is not increased among nulligravid women who used

copper IUDs, but exposure to sexually transmitted pathogens such as C.

trachomatis can increase the risk (81). The contraceptive actions of the LNG-IUD

reverse soon after removal of the device. One-year life-table pregnancy rates after

removal are 89 per 100 for women less than 30 years of age—rates are similar to

women who had not been using any form of birth control (82).

Expulsion and Perforation

The 36-month cumulative expulsion rate is 10.2 per 100 IUD users and did not

vary by IUD type (LNG-IUD 10.1 vs. copper IUD 10.7, p = 0.99) (83). The

cumulative rate of expulsion is lower in nulliparous women compared to parous

women and higher in females aged 14 to 19 years compared to older women. The

risk of uterine perforation associated with insertion is dependent on the inserter.

The risk in experienced hands is on the order of 1 per 1,000 insertions or less

(84). The risk of perforation is no different between the copper and the LNGIUDs (85). Multiparity decreases the risk of perforation (84). Breastfeeding and

787time since delivery are independently associated with uterine perforation risk. The

RR of perforation in lactating compared to nonlactating women at ≤36 weeks

since delivery was 4.7 (86).

Clinical Management

Conditions listed as a category 4 (unacceptable health risk) for IUDs by the WHO

include pregnancy, puerperal sepsis, PID, active purulent cervicitis, endometrial

or cervical cancer, undiagnosed genital bleeding, gestational trophoblastic disease

with persistent elevation in β-hCG, uterine anomalies, and fibroid tumors that

distort the endometrial cavity (87). Infection with HIV is not considered a

contraindication for IUD use. There are no differences in genital viral shedding,

HIV disease progression or disease transmission among IUD users compared to

users of other contraceptives (88). Copper allergy and Wilson disease are

contraindications to the use of copper IUDs.

Candidate Selection

IUDs are appropriate for long-term contraception in most women given their

ease of use, safety, high efficacy, and favorable side effect profile.

Nulliparous women, adolescents, women undergoing a first- or secondtrimester surgical abortion, women with a recent medical abortion, and

women immediately postpartum should all be considered candidates for

IUDs (89). There is renewed interest in postpartum and postabortal insertion of

IUDs. In both circumstances, the woman is clearly no longer pregnant, she may

be highly motivated to accept contraception, and the setting is convenient for the

woman and the provider (89). Postpartum and postabortal insertions are safe.

Complications are not increased by comparison to interval insertion. The only

disadvantage is that the expulsion rate is higher. Two recent RCTs, involving

LNG-IUDs, reported expulsion rates after vaginal postplacental insertion of 24%

and 27% (90,91). Postplacental expulsion rates after cesarean sections appear to

be lower (92). While comparing postpartum to interval insertion, all the women

requesting postpartum IUD received the device, but many women scheduled for

interval insertion did not return. When surveyed at 6 months postpartum, more

women who had immediate postpartum insertion were wearing IUDs than were

those who had scheduled for interval insertion (93,94). Expulsion by 6 months

was more likely following immediate insertion (odds ratio [OR], 4.89; 95%

confidence interval [CI], 1.47–16.32); however, IUD use at 6 months was more

likely with immediate insertion than with delayed insertion (OR, 2.04; 95% CI,

1.01–4.09) (95,96). Thus, in terms of actual continuation, overall, the benefits of

IUD insertion immediately postplacental far outweigh the higher expulsion rates

associated with such insertions (97). Goodman and colleagues found many fewer

788repeat abortions among women followed after postabortal IUD insertion

compared to a cohort of women choosing non-IUD methods of contraception

after an induced abortion (98).

Insertion

At the contraceptive visit, the patient’s history is obtained and a physical

examination, screening for Neisseria gonorrhoeae and chlamydia in high-risk

women, and detailed counseling regarding risks and alternatives are provided.

Clinicians have traditionally inserted IUDs at the time of menses to ensure the

patient is not pregnant, but it can be inserted at any time in the cycle if pregnancy

can be excluded (99). The copper T380A IUD can be inserted within 5 days of

unprotected intercourse for almost 100% effective emergency contraception.

There are limited data on effective treatment of pain during IUD insertion. One

randomized double-blind placebo-controlled study demonstrated that selfadministered vaginal 2% lidocaine gel did not reduce pain with IUD insertion but

did decrease pain with tenaculum placement (100). However, studies utilizing

lidocaine–prilocaine creams have demonstrated a reduction in pain during

tenaculum placement, sound insertion, and IUD insertion (101,102). Despite

some practitioners’ preference to premedicate with misoprostol or NSAIDs, these

interventions do not reduce pain with insertion (103).

Antibiotic prophylaxis is not beneficial, probably because the risk of pelvic

infection with IUD insertion is so low. A large randomized trial of 1,985

patients receiving either oral azithromycin or placebo found no difference in rates

of IUD removal during the first 90 days after insertion and no difference in rates

of salpingitis (104). A meta-analysis found that antibiotic prophylaxis at the time

of IUD insertion did not decrease the risk of PID nor did it reduce the likelihood

of removal within the first 3 months (105). For women at high risk of STIs (e.g.,

aged 25 years or younger or having multiple sex partners), it is reasonable to

screen for STIs and place the IUD on the same day. The International Planned

Parenthood Federation, in collaboration with WHO and other international

organizations, developed guidelines that include the restriction of IUD insertion

within 3 months of puerperal sepsis (106).

The technique of insertion is as follows:

1. The cervix is exposed with a speculum. The vaginal vault and cervix are

cleansed with a bactericidal solution, such as an iodine-containing

solution.

2. Traditionally, the next step is measurement of the uterine cavity with a

uterine sound. The depth of the cavity should measure at least 6 cm from

the external os. However, a pilot study assessing a simplified insertion

789technique without the use of uterine sound or bimanual examination was found

to be safe and effective (107).

3. The use of a tenaculum for insertion is mandatory to prevent perforation.

The cervix is grasped with a tenaculum and gently pulled downward to

straighten the angle between the cervical canal and the uterine cavity. The

IUD, previously loaded into its inserter, is gently introduced through the

cervical canal.

4. With the copper T380A, the outer sheath of the inserter is withdrawn a

short distance to release the arms of the T and is gently pushed inward

again to elevate the now-opened T against the fundus. The outer sheath

and the inner stylet of the inserter are withdrawn, and the strings are cut

to project about 2 cm from the external cervical os.

5. The Mirena 52-mg LNG-IUD is inserted somewhat differently from the

copper T380A. The inserter tube is introduced into the uterus until the

preset sliding flange on the inserter is 1.5 to 2 cm from the external os of

the cervix. The arms of the T device are released upward into the uterine

cavity, and the inserter is pushed up under them to elevate the IUD up

against the uterine fundus.

6. The Liletta 52-mg LNG-IUD can be inserted using a one-handed (United

States) or a two-handed inserter (international). In both cases, the

deployment mechanism is more similar to that for Mirena than for the copper

T380A.

In nulliparous women, insertion may be more challenging because of a

narrower cervical canal than in parous women. Mechanical dilation may be

necessary. While misoprostol was initially proposed to benefit some women

during IUD insertion, a systematic review found that routine pretreatment with

misoprostol did not lead to improved ease or success of IUD insertion and

resulted in more insertional discomfort (108).

Intrauterine Devices in Pregnancy

If an intrauterine pregnancy is diagnosed and the IUD strings are visible, the

IUD should be removed as soon as possible to prevent later septic abortion,

premature rupture of the membranes, and premature birth (109). When the

strings of the IUD are not visible, an ultrasound examination should be performed

to localize the IUD and determine whether expulsion has occurred. If the IUD is

present, there are three options for management:

1. Elective abortion

2. Ultrasound-guided intrauterine removal of the IUD

7903. Continuation of the pregnancy with the device left in place

If the patient wishes to continue the pregnancy, ultrasound evaluation of the

location of the IUD should be considered (110). If the IUD is not in a fundal

location, ultrasound-guided removal using small alligator forceps may be

successful. If the location is fundal, the IUD should be left in place. When

pregnancy continues with an IUD in place, the patient must be warned of the

symptoms of intrauterine infection and should be cautioned to seek care promptly

for fever or flulike symptoms, abdominal cramping, or bleeding. At the earliest

sign of infection, high-dose intravenous antibiotic therapy should be given and the

pregnancy evacuated promptly.

Duration of Use

Annual rates of pregnancy, expulsions, and medical removals decrease with each

year of use (111). Therefore, a woman who has no problem by year 5, for

example, is very unlikely to experience problems in the subsequent years. The

copper T380A is FDA approved for 10 years and the LNG-IUD for 5 years,

although good data support use of the copper T380A through 12 years and of the

LNG-IUD through 7 years (63,85). Newer data from a prospective cohort study

support the use of 52-mg LNG-IUD into year 6 and 7 (112,113).

Choice of Devices

All IUDs available in the United States, provide protection for many years,

have remarkably low pregnancy rates, and substantially reduce the risk of

ectopic pregnancy. The LNG-IUDs reduce the amount of menstrual bleeding and

dysmenorrhea. Amenorrhea rates are 19% to 20%, 12%, and 6% for the 52-, 19.5-

, and 13.5-mg LNG-IUDs at 1 year of insertion, respectively (56,59,114,115). The

copper T380A can be expected to initially increase menstrual bleeding. It is the

most effective means for emergency contraception.

Management of Bleeding and Cramping With Intrauterine Devices

The most important medical reason that women give for requesting removal

of an IUD is bleeding and pelvic pain. These symptoms are common in the first

few months, but diminish over time, in most cases in months. Nonsteroidal antiinflammatory drugs are usually helpful. When pain and bleeding occur later, the

patient should be examined for signs of PID, partial expulsion of the device, or an

intracavitary fibroid. Two ultrasonographic studies comparing women with these

symptoms after 6 months to women with no bleeding complaints show downward

displacement of the IUDs into the cervical canal in many symptomatic women

and in some cases, intracavitary fibroids (116,117). When the patient wishes to

791continue with an IUD, removal of the displaced device and insertion of a new one

is advisable. In situations where IUDs are inexpensive and ultrasound is

expensive or not available, the best course is to offer immediate removal and

replacement of the IUD without ultrasound proof of displacement.

Future of Intrauterine Devices

VeraCept

A novel nitinol-frame low-dose copper IUD, VeraCept (ContraMed, Campbell,

California), is under investigation in a phase II multicenter study in the United

States. Because of its thinner nitinol frame, the insertion tube has a small

diameter. In a randomized comparison against copper T380S, VeraCept was

associated with less pain with insertion, higher continuation rate at 12 months,

lower expulsion rate at 12 months, and fewer removals for pain/bleeding (118).

Despite containing less than half as much copper surface area (175 mm2) as the

T380S (380 mm2), the contraceptive efficacy is similar. However, long-term

efficacy has yet to be studied.

Intrauterine Ball With Copper Pearls

IUB (OCON Medical, Modiin, Israel) copper pearls are strung on a memory alloy

nitinol wire, and when inserted into the uterine cavity, take the shape of the

uterus. This device comes in three sizes, ranging from 12 to 18 mm in diameter

with copper ranging from 300 to 380 mm2. Preliminary results from a pilot study,

however, are concerning for high expulsion rates and low user satisfaction (119).

FibroPlant

The FibroPlant 14 LNG-IUS (Contrel Research, Ghent, Belgium) is a frameless

levonorgestrel-releasing IUD which contains a nonresorbable thread with a knot

that is implanted into the myometrium of the uterine fundus. The attached implant

delivers either 14 or 20 mcg of levonorgestrel each day and is approved for 5

years of use. The reported pregnancy rate at 5 years is 0.4%, with similarly low

expulsion and perforation rates (120).

HORMONAL CONTRACEPTION

Hormonal contraceptives are female sex steroids, synthetic estrogen and

synthetic progesterone (progestin), or progestin-only without estrogen. They

can be administered in the form of OCs, patches, implants, and injectables.

The most widely used hormonal contraceptive is the combination OC

containing estrogen and progestin. Combination OCs can be monophasic, with

792the same dose of estrogen and progestin administered each day, or multiphasic, in

which varying doses of steroids are given through a 21-day or 24-day cycle.

Combination OCs are packaged with 21 active tablets and 7 placebos, or 24 active

tablets and 4 placebo tablets. The inclusion of placebos allows the user to take

one pill every day without having to count. The medication-free interval while the

user takes the placebo tablets allows withdrawal bleeding that mimics a 28-day

menstrual cycle. To begin OC use, the user takes the first pill any time from the

first day of menses through the Sunday after menstruation begins and thereafter

starts a new pack as soon as the first pack is completed. The 7-day medicationfree interval was standard for years, but studies showed that a shorter medicationfree interval is adequate to trigger cyclic withdrawal bleeding and maintains better

suppression of ovulation. Ovarian follicles mature more during the 7-day

medication-free interval than during the 4-day interval. Hence the new 24/4

combination theoretically could be more effective in preventing pregnancy than

the 21/7 combination, but this has not been demonstrated. Other variations of OC

administration are the extended-cycle and the continuous-cycle methods. Users

take active pills containing an estrogen–progestin combination for 3 months at a

time (extended cycle) or indefinitely for a year or more (continuous cycle). Users

on these regimens have more unscheduled days of spotting or bleeding than those

on 28-day cycles in the beginning, but become amenorrheic. As a result, they

experience fewer cycle-triggered symptoms such as headache and menstrual pain.

Continuous combined regimens are preferred for women with chronic pelvic pain

or when dysmenorrhea is not relieved by OCs taken in 28-day cycles (121).

Progestin-only OCs are taken every day without interruption. Other forms of

hormonal contraception include transdermal administration with the patch,

injectable progestins, injectable estrogen–progestin combinations, subdermal

implants that release progestin, and vaginal rings that release either estrogen–

progestin or progestin alone (122).

Where modern hormonal contraception is concerned, it is important to note

that: (i) ALL hormonal contraceptives, even those containing estrogen are

progestin-dominant, meaning that progestin effects are the dominant ones and (ii)

the progestin is doing the real contraceptive “work” by either suppressing

ovulation or thickening the cervical mucous. The presence of the estrogen

component assists with cycle control and adds to the ovarian suppressive effect,

but overall the progestin component dominates.

Steroid Hormone Action

Sex steroids are characterized by their affinity for specific estrogen, progesterone,

or androgen receptors, and by their biologic effects in different systems (123).

Steroids are rapidly absorbed in the gut but go directly into the liver through the

793portal circulation, where they are rapidly metabolized and inactivated. Therefore,

large doses of steroids are required when they are administered orally. The

addition of the ethynyl group to carbon-17 of the steroid molecule hinders

degradation by the liver enzyme 17-hydroxysteroid dehydrogenase and allows

potent biologic activity after oral doses of only micrograms.

Progestins

Progestins are synthetic compounds that mimic the effect of natural progesterone

but differ from it structurally. The progestins differ from one another in their

affinities for estrogen, androgen, and progesterone receptors; their ability to

inhibit ovulation; and their ability to substitute for progesterone and to antagonize

estrogen. Some are directly bound to the receptor (levonorgestrel, norethindrone),

whereas others require bioactivation as, for example, desogestrel, which is

converted in the body to its active metabolite, etonogestrel. The 17-acetoxy

progestins (e.g., medroxyprogesterone acetate) are bound by the progesterone

receptor. Norgestrel exists as two stereoisomers, identified as dextronorgestrel

and levonorgestrel. Only levonorgestrel is biologically active. Three newer

progestins (norgestimate, desogestrel, and gestodene) are viewed as more

“selective” than the other 19-nor progestins, in that they have little or no

androgenic effect at doses that inhibit ovulation (124). The FDA-approved

norgestimate- and desogestrel-containing OCs, and gestodene are available in

Europe. Gestodene is a derivative of levonorgestrel that is more potent than the

other preparations (i.e., very little of it is required for antifertility effects).

Similarly, norelgestromin is an active metabolite of norgestimate and more potent

than the parent compound. It is used in the transdermal patch. Drospirenone, a

progestin introduced in the United States, is a derivative of the diuretic

spironolactone. It has a high affinity for progesterone receptors, mineralocorticoid

receptors, and androgen receptors. It acts as a progesterone agonist but is a

mineralocorticoid antagonist and androgen antagonist (125). Comparative studies

suggest a small decrease in body weight and in blood pressure, with equivalent

cycle control and contraceptive efficacy, in women taking an OC containing 3 mg

of drospirenone/30-μg ethinylestradiol (EE) versus women taking a 150-μg

levonorgestrel/30-μg EE preparation (126). Pilot studies among women with

polycystic ovary syndrome showed good cycle control and reduction in androgen

levels with no change in weight, blood pressure, or glucose metabolism (127).

The FDA approved the 20-μg EE/3-mg drospirenone OC for premenstrual

dysphoric disorder (PMDD) in women who chose OCs for contraception. When

compared to an OC with 30-μg EE/150-μg levonorgestrel, women taking the

drospirenone OC had better relief of menstrual symptoms, better improvement in

acne, reduction in negative affect during the menstrual phase, and a greater

794feeling of well-being (128). Dienogest, another progestin introduced in the United

States, is combined with estradiol valerate, not EE. Whether it offers any

advantage over already marketed OC combinations is not yet evident. The

dienogest/estradiol valerate combination is as effective as levonorgestrel/EE as a

contraceptive, and in treating abnormal uterine bleeding (129).

Estrogens

In the United States, most OCs contain EE. An older formulation OC contains

mestranol, which is EE with an extra methyl group. It requires bioactivation in

the liver, where the methyl group is cleaved, releasing the active agent, EE. OCs

with 35 μg of EE provide the same blood levels of hormone as do OCs containing

50 μg of mestranol (130). Other contraceptives contain 17 β-estradiol (Zoely) or

its ester, estradiol valerate (Qlaira) or estradiol cypionate (Cyclofem), an

injectable.

Antifertility Effects

Combination Estrogen–Progestin Contraceptives

Ovulation can be inhibited by estrogen or by progestin alone. Pharmacologic

synergism is exhibited when the two hormones are combined and ovulation is

suppressed at a much lower dose of each agent. Combination OCs, patches, and

the NuvaRing suppress basal follicle-stimulating hormone (FSH) and LH. They

diminish the ability of the pituitary gland to synthesize gonadotropins when it is

stimulated by hypothalamic GnRH (131). Ovarian follicles do not mature, little

estradiol is produced, and there is no midcycle LH surge. Ovulation does not

occur, the corpus luteum does not form, and progesterone is not produced. This

blockade of ovulation is dose related. Newer low-dose OCs do not provide as

dense a block and allow somewhat higher baseline FSH and LH levels than

higher-dose formulations (132). This makes ovulation somewhat more likely to

occur if pills are missed or if the patient takes another medication that reduces

blood levels of the contraceptive steroids.

Progestin-Only Preparations

Highly effective contraception can be provided by progestin alone, thus

avoiding the risks of estrogen. The mode of action of progestin-only

contraceptives is highly dependent on the dose of the compound (133). With low

levels of progestin in the blood, ovulation will occur part of the time. At moderate

levels of progestin in the blood, normal basal levels of FSH and LH are present,

and some follicle maturation may occur. At higher blood levels as seen with

DepoProvera, the basal FSH is reduced, and there is less follicular activity, less

795estradiol production, and no LH surge.

Transdermal Hormonal Contraception

The patch (Ortho Evra), which adheres to the user’s skin, and the vaginal

NuvaRing contain combinations of EE and a potent progestin. Both provide

sustained release of the steroids and result in relatively constant serum levels that

are less than the peak levels seen with OCs but sufficient to prevent ovulation.

Hormonal Implants

With the levonorgestrel subdermal implants there is some follicular maturation

and estrogen production, but LH peak levels are low and ovulation is often

inhibited. In the first year of use, ovulation is believed to occur in about 20% of

cycles. The proportion of ovulatory cycles increases with time, probably as a

result of the decline in hormone release. By the fourth year of use, 41% of cycles

are ovulatory. The more potent progestin released by the etonogestrel implant is

even more effective at preventing ovulation (134). The mechanisms of action of

low-dose progestins include effects on the cervical mucous, endometrium, and

tubal motility. The scant, dry cervical mucous that occurs in women using

implants inhibits sperm migration into the upper tract. Progestins decrease nuclear

estrogen receptor levels, decrease progesterone receptors, and induce activity of

the enzyme 17-hydroxysteroid dehydrogenase, which metabolizes natural 17 β-

estradiol (133).

The sustained release offered by contraceptive implants allows for highly

effective contraception at relatively low steroid blood levels. Figure 14-6 depicts

expected steroid blood levels with implants, injectables, and OCs. An additional

mechanism for contraception was discovered with the antiprogesterone

mifepristone (RU486). In the normal cycle, there is a small amount of

progesterone production from the follicle just before ovulation. This progesterone

appears essential to ovulation, because if mifepristone is given before ovulation

this can be delayed for several days (135,136).

796FIGURE 14-6 Schematic representation of the expected pharmacokinetic profiles of

progestogens administered in different formulations. (From Landgren BM. Mechanism of

action of gestagens. Int J Gynaecol Obstet 1990;32:95–110, with permission.)

Efficacy of Hormonal Contraception

[4] When used consistently, combination OCs have pregnancy rates as low as

2 to 3 per 100 women per year. Progestin-only OCs are less effective than

combination estrogen–progestin preparations, with best results of 3 to 4

pregnancies per 100 woman-years. All methods have the potential for user

error; therefore, there may be a 10-fold difference between the best results and

797results in typical users of OCs. Injectable progestins and implants are much less

subject to user error than OCs. The difference between the best results and results

in typical users is small and is comparable to pregnancy rates after tubal

sterilization (Table 14-1). Pregnancy rates with the Ortho Evra patch and the

NuvaRing were equivalent to those of OCs; however, because it is easier to use

these methods consistently, larger studies may well demonstrate better typical

user results than with OCs (137,138). Typical use pregnancy rates with

depomedroxyprogesterone acetate (DMPA) are lower. The subdermal implants

have the lowest failure rates of any hormonal contraceptive method.

Hormonal Contraception for Obese Women

The rate of obesity worldwide is increasing. Across the United States, more than

one in three adults are obese (139). Most studies of contraceptive efficacy

intentionally excluded obese women, so the available information is limited.

Obese women are no less likely than other women to become pregnant, but they

have an increased risk of pregnancy complications (140). A systematic review of

hormonal contraception for overweight and obese women focused on 12 relevant

studies (141). The data on the risk of pregnancy in overweight and obese women

using OCs was mixed, with one study showing increased risk while another

showing decreased risk. Analysis of data from other contraceptive methods

indicated no association of pregnancy with overweight or obesity. These included

depot medroxyprogesterone acetate (subcutaneous), LNG-IUD, the two-rod

levonorgestrel implant, and the etonogestrel implant. One study evaluated a

contraceptive patch containing EE and levonorgestrel, and found that obese

women in the “treatment-compliant” subgroup had a higher reported Pearl Index

than nonobese women.

Metabolic Effects and Safety

Venous Thrombosis

[4] Women who use estrogen-containing hormonal contraceptives are at

increased risk for venous thrombosis and thromboembolism. Normally the

coagulation system maintains a dynamic balance of procoagulant and

anticoagulant systems. Estrogens affect both systems in a dose-related fashion.

For most women, fibrinolysis (anticoagulation) is increased as much as

coagulation, maintaining the dynamic balance at increased levels of production

and destruction of fibrinogen (Fig. 14-7) (142). Older studies included women

with what are now considered contraindications to the use of estrogen-containing

hormonal contraceptives: previous thrombosis, preexisting vascular disease,

coronary artery disease, cancers, and serious trauma (142,143). Low-dose OCs

have a less measurable effect on the coagulation system, and fibrinolytic factors

798increase at the same rate as procoagulant factors. Lower estrogen–dose (30- to

35-lg EE) OCs reduce the risk of a thromboembolic event when compared

with higher-dose (50-lg estrogen) OCs (144). A very large Danish study showed

for the first time that combination OCs with 20 μg of EE have an 18% further

reduction in thrombosis risk compared with 30- to 40-μg OCs after adjustment for

duration of use (145). The progesterone-only OCs and the levonorgestrelreleasing IUD were not associated with venous thrombosis.

FIGURE 14-7 Dynamic balance of hemostasis. (From Winkler UH, Buhler K, Schlinder

AE. The dynamic balance of hemostasis: implications for the risk of oral contraceptive

use. In: Runnebaum B, Rabe T, Kissel L, eds. Female Contraception and Male Fertility

Regulation. Advances in Gynecological and Obstetric Research Series. Oxfordshire, UK:

CRC Press, 1991:85–92. Reproduced by permission of Taylor and Francis Group, LLC, a

division of Informa plc.)

The absolute risk of deep vein thrombosis was strongly influenced by age,

increasing from 1.84 per 10,000 women aged 15 to 19 years to 6.59 per 10,000

for women aged 45 to 49 years with current users, former users, and never users

combined. With all types of OCs combined, the overall absolute rate of deep vein

thrombosis was 6.29 per 10,000 woman-years for current OC users compared to

3.01 among nonusers, giving an adjusted rate ratio of 2.83 (95% CI, 2.65–3.01).

799This is a higher absolute risk than the 3 per 10,000 woman-years previously

estimated and may reflect, among other things, the use of better means for

diagnosis of deep vein thrombosis (145). This population-based study includes all

Danish women aged 15 to 49 years, excluding only women with a diagnosis of

cancer or of cardiovascular disease diagnosed before the study interval.

Thrombosis risk was highest during the first year of use and decreased thereafter.

Thrombophilia

Changes in the coagulation system are detectable in all women, including those

taking lower-dose OCs; some women are genetically predisposed to thrombosis

when challenged by pregnancy or administration of exogenous estrogen. Women

with inherited deficiencies of antithrombin III (ATIII), protein C, or protein

S are at very high risk for thrombosis with pregnancy or estrogen therapy,

but they make up a very small proportion of potential OC users. A much

more common variation, factor V Leiden exists in 3% to 5% of the Caucasian

population. It codes for a one amino acid mutation in the factor V protein,

inhibiting cleavage of the protein by activated protein C (APC), which is an

essential step in maintaining the balance between coagulation and fibrinolysis

(146). Risk for a first thromboembolic episode among women using OCs was 2.2

per 10,000 woman-years for women without the factor V mutation and 27.7 per

10,000 woman-years for women with the mutation (147). There are pronounced

ethnic differences in the presence of this mutation. The Leiden allele is found in

3% to 5% of Caucasians but is rare in Africans, Asians, Amerindians, Eskimos,

and Polynesians (148). A similar mutation is found in the prothrombin gene at

position 20,210 and is described as prothrombin G20210A. This mutation occurs

in 3% of the European population and is strongly associated with venous

thrombosis in women taking OCs (149). Many more genetic conditions

predisposing to thrombosis were described. Pregnancy is an even greater

challenge for women with inherited defects of anticoagulation (150). A woman

who sustains a venous event while using OCs should be evaluated thoroughly

after she has recovered. Assessment should include at minimum the measurement

of ATIII, protein C, and protein S levels, resistance to APC, serum homocysteine,

factor V Leiden mutation, the prothrombin G20210A mutation and testing for

antiphospholipid syndrome. It should not be assumed that hormonal contraception

was the unique cause of the thrombotic episode.

Routine screening for all women before prescribing hormonal

contraception is not justified because effective contraception would be denied

to 5% of Caucasian women, and only a small number of fatal pulmonary

emboli would be prevented (151,152). Screening women with a personal or

family history of deep vein thrombosis before starting estrogen-containing

800hormonal contraception or pregnancy is strongly recommended. Women already

diagnosed as having factor V Leiden should not receive estrogen-containing

contraceptives, that is, the pill, patch, or ring.

Thrombosis and the New Progestins

Several studies found increased risk of venous thrombosis when users of OCs

containing the newer progestins desogestrel or gestodene combined with 20 to 30

μg of EE were compared with users of levonorgestrel combined with the same

doses of estrogen (153,154). Subsequent studies showed mixed results. One study

found no difference in thrombosis risk in users of OCs containing desogestrel or

gestodene compared to users of OCs with other progestins in a large study of

cases occurring between 2002 and 2006 in Austria (155). In 2013, a systematic

review combining data from 26 studies (including the aforementioned ones)

demonstrated that the RR of venous thrombosis for OCs containing 30 to 35 μg of

EE and gestodene, desogestrel, cyproterone acetate or drospirenone were 50% to

80% higher than formulations containing levonorgestrel (Fig. 14-8). However,

this risk increase corresponds to approximately a difference of 3 cases per 10,000

women-years comparing second and third generation progestin users (in contrast

to a VTE incidence of 20 cases per 10,000 in pregnancy) (156).

FIGURE 14-8 Network meta-analysis, per contraceptive plotted on a logarithmic scale.

80120LNG, 20-μg ethinylestradiol with levonorgestrel; 30LNG, 30-μg ethinylestradiol with

levonorgestrel; 50LNG, 50-μg ethinylestradiol with levonorgestrel; 20GSD, 20-μg

ethinylestradiol with gestodene; 30GSD, 30-μg ethinylestradiol with gestodene; 20DSG,

20-μg ethinylestradiol with desogestrel; 30DSG, 30-μg ethinylestradiol with desogestrel;

35NRG, 35-μg ethinylestradiol with norgestimate; 35CPA, 35-μg ethinylestradiol with

cyproterone acetate; 30DRSP, 30-μg ethinylestradiol with drospirenone; overall relative

risk (95% confidence interval) of venous thrombosis (dots [lines]); non-use, reference

group.

Ischemic Heart Disease

Ischemic heart disease and stroke were the major causes of death attributed to OC

use in the past. It is known that the principal determinants of risk are advancing

age and cigarette smoking (157). With higher-dose OCs used in the 1980s,

smoking had a profound effect on risk. Women smoking 25 or more

cigarettes per day had a 30-fold increased risk for myocardial infarction if

they used OCs, compared with nonsmokers not using OCs (158). The use of

OCs became safer because most women are taking low-dose pills and

because physicians prescribe selectively, excluding women with major

cardiovascular risk factors. A very large study in the United States confirmed

the safety of OCs as currently prescribed. A total of 187 women aged 15 to 44

years with confirmed myocardial infarction were identified during 3.6 million

woman-years of observation in the Kaiser Permanente Medical Care Program in

California from 1991 to 1994. This is a rate of 3.2 per 100,000 woman-years

(159). Nearly all of the users took OCs with less than 50 μg of EE. After adjusting

for age, illness, smoking, ethnicity, and body mass index, risk for myocardial

infarction was not increased by OC use (OR, 1.14; 95% CI, 0.27–4.72). Of heart

attack victims, 61% smoked; only 7.7% were current users of OCs. In a later

study, the same investigators pooled results from the California study with a

similar study from the Washington State. The results were the same. [5] Current

users of low-dose OCs had no increased risk for myocardial infarction after

adjustment for major risk factors and sociodemographic factors (160). Past

use of OCs does not increase risk for subsequent myocardial infarction (161).

These observations are supported by another population-based prospective study.

OC use and myocardial infarction were studied prospectively among 48,321

randomly selected women aged 30 to 49 years in the Uppsala Health Care Region

of Sweden (162). There was no association between current or past OC use and

myocardial infarction. Most current users were taking low-dose estrogen pills

(defined as less than 50 μg of EE or less than 75 μg of mestranol) with second or

third generation progestins, more than half were aged 35 or older, and 26% were

current cigarette smokers.

802Oral Contraceptives and Stroke

In the 1970s, OC use appeared to be linked to the risk of hemorrhagic and

thrombotic stroke, but these studies failed to take into consideration preexisting

risk factors (163). A rare form of cerebrovascular insufficiency, the moyamoya

disease, is linked to OC use, especially among cigarette smokers (164). The

evidence shows no risk of stroke among women who are otherwise healthy

and who use low-dose pills. One study identified all Kaiser Permanente Medical

Care Program patients aged 15 to 44 years who sustained fatal or nonfatal stroke

in California from 1991 to 1994 (165). Hypertension, diabetes, obesity, current

cigarette smoking, and black ethnicity were strongly associated with stroke risk,

but neither current nor past OC use was associated with stroke. A WHO study of

cases from 1989 to 1993 from 17 countries in Europe and the developing world

included women taking higher-dose OCs and low-dose OCs. European women

using low-dose OCs had no increased risk for either type of stroke, thrombotic or

hemorrhagic. Those taking higher-dose OCs did have measurable risk (166,167).

Women in developing countries had an apparent modest increase in risk, but this

finding was attributed to undetected existing risk factors. Another study from

Europe found less stroke risk from low-dose pills than from older, higher-dose

pills, and that the risk was less if the patient’s blood pressure was checked before

starting OCs.

Women who smoke and those who have hypertension and diabetes are at

increased risk for cardiovascular disease regardless of whether they use OCs.

The important question is whether the risk is further increased if they use lowdose OCs, and if so, by how much. The WHO study provides some insight:

smokers taking OCs had seven times the risk of ischemic (thrombotic) stroke

when compared with smokers who did not use OCs, and hypertensive women had

10-fold increased risk if they took OCs, but a fivefold risk if they did not (166).

Similarly, a study from Denmark found that women with diabetes had a fivefold

increase risk for stroke, which increased to 10-fold if they took OCs (168). These

data were not limited to low-estrogen OCs. The data suggest that although the risk

is primarily determined by the predisposing condition—hypertension, diabetes, or

cigarette smoking—the risk can be magnified by OC use, even when the OCs are

low dose. These observations were confirmed in a systematic review (169).

Hypertensive women using combination estrogen– progestin oral contraceptives

(COCs) had higher risk for ischemic stroke and acute myocardial infarction than

hypertensive women not using COCs. The US practice of limiting hormonal

contraceptives containing estrogen by women older than 35 years of age to

nonsmokers without other vascular disease risk factors is prudent (7).

Blood Pressure

803OCs have a dose-related effect on blood pressure. With the older high-dose

pills, as many as 5% of patients could be expected to have blood pressure levels

greater than 140/90 mm Hg. The mechanism is believed to be an estrogeninduced increase in renin substrate in susceptible individuals. Low-dose pills

have minimal blood pressure effects, but surveillance of blood pressure is

advised to detect the occasional idiosyncratic response.

Glucose Metabolism

Oral estrogen alone has no adverse effect on glucose metabolism, but progestins

exhibit insulin antagonism (170). Older OC formulations with higher doses of

progestins produced abnormal glucose tolerance tests with elevated insulin levels

in the average patient. The effect on glucose metabolism, similar to the effect on

lipids, is related to androgenic potency of the progestin and to its dose.

Lipid Metabolism

Androgens and estrogens have competing effects on hepatic lipase, a liver

enzyme critical to lipid metabolism. Estrogens depress low-density lipoproteins

(LDL) and elevate high-density lipoproteins (HDL), changes that can be

expected to reduce the risk of atherosclerosis (171). Androgens and androgenic

progestins can antagonize these beneficial changes, reducing HDL and elevating

LDL levels. Estrogens elevate triglyceride levels. Low-dose formulations have

minimal adverse effect on lipids, and the newer formulations (with desogestrel

and norgestimate as the progestin) produce potentially beneficial changes by

elevating HDL and lowering LDL (172,173). Although average values of a large

group show only small lipid changes with the use of current OCs, an occasional

patient may have exaggerated effects. Women whose lipid values are higher than

the mean before treatment are more likely to experience abnormalities during

treatment (172).

Other Metabolic Effects

OCs can produce changes in a broad variety of proteins synthesized by the liver.

The estrogen in OCs increases circulating thyroid-binding globulin, thereby

affecting tests of thyroid function that are based on binding, increasing total

thyroxine (T4) levels, and decreasing triiodothyronine (T3) resin uptake. The

results of actual thyroid function tests, as measured by free T4 and radioiodine

tests, are normal (174).

Oral Contraceptives and Neoplasia

Endometrial and Ovarian Cancers

Combination OCs reduce the risk of subsequent endometrial cancer and

804ovarian cancer (175,176). [8] Two-year use of OCs reduces the risk of

endometrial cancer by 40%, and 4 or more years of use reduces the risk by

60%. The evidence for this benefit continues to accumulate (177). [8] A 50%

reduction in ovarian cancer risk for women who took OCs for 3 to 4 years

and an 80% reduction with 10 or more years of use were reported (178).

There was some benefit from as little as 3 to 11 months of use. A review of all

available studies in the world published in English through 2008 concluded

that ovarian cancer risk decreased by 20% for each 4 years of use and was

seen for carriers of the BRCA1 and BRCA2 mutations as well. The benefit

persisted for at least 30 years after last use (179). A similar reduction of risk of

ovarian epithelial cancer was found in a prospective study from Norway and

Sweden, with borderline tumor risk equally reduced. Combination OCs with less

than 50 μg of EE, or less than 100 μg of mestranol and reduced doses of

progestin, provided as much protection as higher-dose pills (180). Today’s lowerdose 20-μg EE pills were not separately studied. Whether or not they provide the

same benefit remains unproven; however, progestin-only contraceptives are

reported to provide risk reduction equivalent to that of combined OCs (181).

Colorectal Cancer

There is evidence that OC use is protective against colon cancer. A casecontrol study in Italy comparing women with colon cancer with controls found a

37% reduction in colon cancer and a 34% reduction in rectal cancer (colon cancer

OR, 0.63; 95% CI, 0.45–0.87 and rectal cancer OR, 0.66; 95% CI, 0.43–1.01).

Longer use produced more protection against colon cancer (182). Results of the

US Nurses’ Health Study disclosed some degree of protection. Women who used

OCs for 96 months or more had a 40% lower risk of colorectal cancer (RR, 0.60;

95% CI, 1.15–2.14) (183). A large case-control study from Wisconsin found most

of the benefit limited to women who were less than 14 years since discontinuing

OCs (184). The mechanism of protection has not been identified. A long-term

study that prospectively evaluated women for up to 44 years detected ever use of

OCs was associated with reduced colorectal cancer (IRR, 0.81; 99% CI, 0.66–

0.99) (185).

Cervical Cancer

There may be a weak association between OC use and cancer of the cervix. A

systematic review of 28 epidemiologic studies of cervical cancer in OC users

compared with those who never used OCs reported summary RRs of 1.1 (95% CI,

1.1–1.2) at less than 5 years of pill use, 1.6 (1.4–1.7) at 5 to 9 years, and 2.2 (1.9–

2.4) at 10 or more years (186). A 2007 update by these same authors reported a

pooled RR for all studies of 1.9 (95% CI, 1.69–2.13) for invasive cervical cancer

805or CIN3/carcinoma in situ (CIS) with 5 or more years of OC use. The risk

declined after use ceased and by 10 years returned to that of never users (187). In

a study that observed women for up to 44 years, an increased risk of cervical

cancer that was seen in current and recent users appeared to be lost within 5 years

of stopping oral contraception with no evidence of increased risk in ever users

with time (185). Critics of these studies argued that causation is not proven

because few adequately control for the key behavioral factors of partners, use of

barrier contraception, and adequacy of cervical cancer screening (188). Important

risk factors for cervical cancer are early sexual intercourse and exposure to HPV.

Women who used OCs typically started sexual relations at younger ages than

women who have not used OCs and, in some studies, report having had more

partners. These factors increase one’s chance of acquiring HPV, the most

important risk factor for cervical cancer. Because barrier contraceptives reduce

the risk of cervical cancer, the use of alternative choices for contraception

compounds the difficulty in establishing an association with OC use alone (189).

The presence of HPV types 16 or 18 is associated with a 50-fold increase in risk

for precancerous lesions of the cervix (190). Adenocarcinomas of the cervix are

rare, they are not as easily detected as other lesions by screening cervical

cytology, and the incidence appears to be increasing. One study found a doubling

of risk for adenocarcinoma with OC use that increased with the duration of use,

reaching an RR of 4.4 if the total use of OCs exceeded 12 years (191). The results

of this study were adjusted for history of genital warts, number of sexual partners,

and age at first intercourse. Another summary of case-control studies includes

testing for HPV. HPV types 16 or 18 were present in 82% of the patients, yielding

an RR of 81.3 (95% CI, 42.0–157.1) for the disease. Cofactors identified with

adenocarcinoma included longer-term use of hormonal contraception. The use of

hormonal contraception by women aged 20 to 30 years is estimated to increase

the incidence of cervical cancer (any cervical cancer, or CIN3/CIS) diagnosed by

age 50 from 7.3 to 8.3 per 1,000 women in lesser-developed countries and from

3.8 to 4.5 in developed countries (192). The use of OCs is, at most, a minor factor

in causation of cervical cancer; these findings emphasize the need to immunize

against HPV and to provide cervical cancer screening worldwide. To reduce the

risk, women who are not in mutually monogamous relationships should be

advised to use barrier methods in addition to hormonal contraception. A metaanalysis demonstrated that invasive cervical cancer may be approximately onethird less frequent in women who have used an IUD (193).

Breast Cancer

There is a large volume of conflicting literature on the relationship between

OC use and breast cancer (194). No increase in overall risk is found from OC

806use, but some studies found that risk may increase in women who used OCs

before their first term pregnancy, used OCs for many years, are nulligravid, are

young at the time of diagnosis, or continue using OCs in their 40s. [6] A metaanalysis of 54 studies of breast cancer and hormonal contraceptive use reanalyzed

data on 53,297 women with breast cancer and 100,239 controls from 25 countries,

representing about 90% of the epidemiologic data available worldwide (195).

Current use of OCs was associated with a very small, but statistically stable 24%

increased RR (1.24; 95% CI, 1.15–1.33). The risk fell rapidly after

discontinuation, to 16% 1 to 4 years after stopping and to 7% 5 to 9 years after

stopping. Risk disappeared 10 years after cessation (RR, 1.01; 95% CI, 0.96–

1.05). Results did not differ in any important way by ethnic group, reproductive

history, or family history. Since the meta-analysis was published, subsequent

studies found no increased risk. A case-control study of 4,575 women with breast

cancer and 4,682 controls aged 35 to 64 years living in five cities in the United

States concluded that breast cancer risk was not increased for current or past users

of OCs and did not increase with prolonged or with higher-estrogen OC use

(196). Neither family history of breast cancer nor beginning use at a young age

was associated with increased risk. A similar study in Sweden compared 3,016

women aged 50 to 74 years who had invasive breast cancer with 3,263 controls of

the same age. [6] No relation was found between the past use of OCs and breast

cancer (197). Effect of hormone dose was explored in a 2008 US study of 1,469

women with breast cancer who were matched by race, age, and neighborhood to

community controls. The investigators administered questionnaires and

performed BRCA1 and BRCA2 testing. Subjects who began OC use during or

after 1975 were considered to have used low-dose pills. “Low dose” was not

further defined. Neither OC use overall or “low-dose” OC use was associated

with breast cancer risk, among the total group, or within any subset. Women with

BRCA1 or BRCA2 did not have higher rates of cancer whether they were OC

users or nonusers (198). [6] A prospective cohort study evaluating very–longterm risks or benefits associated with the use of combined OCs found an

increased breast cancer risk in current and recent users that appeared to be lost

within 5 years of stopping (185). The controversy over the association between

breast cancer and OC use is likely to continue. While concerns about a

connection continue to circulate, the benefits of OC use, particularly

protection against pregnancy and the other noncontraceptive benefits,

outweigh the risks for the vast majority of women.

Liver Tumors

OCs were implicated as a cause of benign adenomas of the liver. These

hormonally responsive tumors can cause fatal hemorrhage. They usually regress

807when OC use is discontinued; the risk is related to prolonged use (199). The

tumors are rare; about 30 cases per 1 million users per year were predicted with

older formulations. Presumably, newer low-dose products pose less risk. A link to

hepatic carcinoma was proposed. This cancer is closely associated with chronic

hepatitis B and C infections and is usually seen in cirrhotic livers. There are case

reports of hepatocellular carcinoma in young women with no risk factors other

than long-term OC use (200). A large study from six countries in Europe found

no association between the use of OCs and subsequent liver cancer (201). A

systematic review looked for evidence of harm associated with hormonal

contraceptive use among women already at risk because of liver disease (202).

The authors concluded from the limited data available that OCs do not affect the

course of acute hepatitis or chronic hepatitis, and do not affect the rate of

progression or the severity of cirrhotic fibrosis, the risk of hepatocellular

carcinoma in women with chronic hepatitis, or the risk of liver dysfunction in

hepatitis B virus carriers.

Oral Contraceptives and Sexually Transmitted Infections

Chlamydial colonization of the cervix appears more likely in OC users than in

nonusers but, despite this finding, several case-control studies found a reduced

risk of acute PID among OC users (203,204). In contrast, a subsequent study

found no protection with OC use (205). Whether hormonal contraceptives

influence acquisition of HIV remains under debate. A recent systematic review

increased concern about a potential causal association between DMPA use and

HIV acquisition with hazard ratio (HR) estimated at 1.5 or less (206). In response

to these findings, the WHO convened and updated recommendations for the use

of progestogen-only injectables from category 1 to category 2. Women should not

be denied the use of progestogen-only injectables but should be advised about the

uncertainty over a causal relationship. There remains no suspected increased risk

of HIV acquisition with use of OCs or contraceptive implants.

Health Benefits of Oral Contraceptives

OCs have important health benefits (Table 14-3). These include contraceptive

and noncontraceptive benefits.

Contraceptive Benefits

OCs provide highly effective contraception and prevent unwanted

pregnancy, an important public health problem. Combination and progestinonly contraceptives decrease the risk of pregnancy and therefore, decrease

the risk of ectopic pregnancy. Studies have reported an increased proportion of

pregnancies that are ectopic, in women who became pregnant while using

808progestin-only implants, IUDs, and OCs (207). A systematic review of progestinonly injectables and implants reported higher proportions of ectopic pregnancies

in the event that pregnancy occurs; however, the review confirmed that the

absolute risk of ectopic pregnancy is lower for women using these methods

compared with noncontraceptive users (208). Furthermore, the clinical and

epidemiologic evidence indicated that LNG-containing implants are more likely

to result in an ectopic pregnancy (or any pregnancy) than the etonogestrel implant

or progestin-only injectables.

Noncontraceptive Benefits

OC use produces strong and lasting reduced risk for endometrial and

ovarian cancer. In addition, protection was found for women with known

hereditary ovarian cancer. The use of OCs in BRCA mutation carriers was

associated with a significantly reduced risk of ovarian cancer (SRR, 0.50; 95%

CI, 0.33–0.75) (209). Protection increased with longer duration of use. The exact

mechanisms of action of OCs in the prevention of ovarian cancer is unknown, but

one leading mechanism stems from the “incessant ovulation” theory, according to

which chronic ovulation contributes to ovarian cancer by causing repeated trauma

of the ovarian surface epithelium (210). Therefore, OCs may decrease the risk of

ovarian cancer by suppressing ovulation.

Table 14-3 Established and Emerging Noncontraceptive Benefits of Oral

Contraceptives

Established Benefits

Menses-related

Increased menstrual cycle regularity

Reduced blood loss

Reduced iron-deficiency anemia

Reduced dysmenorrhea

Reduced symptoms of premenstrual dysphoric disordera

Inhibition of Ovulation

Fewer ovarian cysts

Fewer ectopic pregnancies

809Other

Reduced fibroadenomas/fibrocystic breast changes

Reduced acute pelvic inflammatory disease

Reduced endometrial cancer

Reduced ovarian cancer

Emerging Benefits

Increased bone mass

Reduced acne

Reduced colorectal cancer

Reduced uterine leiomyomata

Reduced rheumatoid arthritis

Treatment of bleeding disorders

Treatment of hyperandrogenic anovulation

Treatment of endometriosis

Treatment of perimenopausal changes

aOnly the low-dose EE/droperidol oral contraceptive has U.S. Food and Drug

Administration approval for premenstrual dysphoric disorder treatment.

From Burkman R, Schlesselman JJ, Zieman M. Safety concerns and health benefits

associated with oral contraception. Am J Obstet Gynecol 2004;190(Suppl):S12, with

permission.

Other documented benefits of OC use include reduction of benign breast

disease (211). The use of OCs helps relieve dysmenorrhea (212). OCs offer

effective therapy for women with menorrhagia and dysfunctional uterine

bleeding.

All combination OCs offer some protection from functional ovarian cysts,

but this is dose related (213). Although OCs may prevent cyst formation, they are

not helpful in treating large functional ovarian cysts and should not be used for

this purpose (214).

All combination OCs reduce circulating androgen levels and usually

810improve acne. Three OCs were specifically FDA approved for acne treatment:

the norgestimate/EE triphasic (Ortho Tri-Cyclen), the norethindrone/EE

multiphasic (Estrostep), and the 20-μg EE/3-mg drospirenone OC.

Fertility After Oral Contraceptive Use

After discontinuing OCs, the return of ovulatory cycles may be delayed for a few

months. Women who have amenorrhea more than 6 months after discontinuation

of OCs should undergo a full evaluation because of the risk for prolactinproducing pituitary tumors. This risk is not related to OC use but rather to the

probability that the slow-growing tumor was already present and produced

menstrual irregularity, prompting the patient to take OCs (215).

Sexuality

There is limited literature focusing on the topic of sexual function among OC

users. A 2003 Spanish study demonstrated that in a comparative cohort of women

using OCs and a cohort using IUDs, sexual desire decreased over time, but was

not affected by the contraceptive method (216). OCs containing the new progestin

drospirenone are reported to improve sexual functioning and feelings of wellbeing (217,218).

Teratogenicity

A meta-analysis of 12 prospective studies, including 6,102 women who used OCs

and 85,167 women who did not, revealed no increase in the overall risk for

malformation, congenital heart defects, or limb reduction defects with the use of

OCs (219). Progestins were used to prevent miscarriage. A large study compared

women showing signs of threatened abortion who were treated with progestins

(primarily oral medroxyprogesterone acetate) with women who were not treated.

The rate of malformation was the same among the 1,146 exposed infants as

among the 1,608 unexposed infants (220). Conversely, exposure in utero to

diethylstilbestrol, a nonsteroidal estrogen, can induce clear cell vaginal cancer and

congenital anomalies of the reproductive tract in the female offspring (221). A

literature search revealed no recent reports linking teratogenicity to hormonal

contraception.

Interaction of Oral Contraceptives With Other Drugs

Some drugs (e.g., rifampin) reduce the effectiveness of OCs; conversely, OCs

can augment or reduce the effectiveness of other drugs (e.g., benzodiazepines)

(222,223). Perhaps of greatest clinical significance are six antiepileptic drugs:

phenytoin, phenobarbital, carbamazepine, oxcarbazepine, felbamate, and

topiramate (224). These drugs and the antibiotic rifampin all induce synthesis of

811liver cytochrome P450 enzymes and reduce plasma levels of EE in women taking

OCs, increasing the likelihood of contraceptive failure. Some antiseizure agents

have no effect on the levels of contraceptive steroids in the blood. These include

valproic acid, vigabatrin, lamotrigine, gabapentin, tiagabine, levetiracetam,

zonisamide, ethosuximide, and benzodiazepines (224). St. John’s wort induces

cytochrome P450 and is reported to increase clearance of EE and norethindrone

(225). The antifungal agents griseofulvin, ketoconazole, and itraconazole induce

these hepatic enzymes and may reduce OC efficacy (225). Ampicillin and

tetracycline were implicated in numerous case reports of OC failure. They kill gut

bacteria (primarily clostridia) that are responsible for hydrolysis of steroid

glucuronides in the intestine, which allows reabsorption of the steroid through

enterohepatic circulation. It was not possible to demonstrate reduced plasma

levels of EE overall or differences in pregnancy rates (226). Certain drugs, such

as protease inhibitor atazanavir, appear to increase plasma levels of contraceptive

steroids but the clinical significance of such an interaction is unclear (227).

An example of OCs affecting the metabolism of other drugs is seen with

diazepam and related compounds. OC use reduces the metabolic clearance and

increases the half-life of those benzodiazepines that are metabolized primarily by

oxidation: chlordiazepoxide, alprazolam, diazepam, and nitrazepam. Caffeine and

theophylline are metabolized in the liver by two of the P450 isozymes, and their

clearance is reduced in OC users. Cyclosporine is hydroxylated by another of the

P450 isozymes, and its plasma concentrations are increased by OCs. Plasma

levels of some analgesic drugs are decreased in OC users. Salicylic acid and

morphine clearances are enhanced by OC use; therefore, higher doses could be

needed for adequate therapeutic effect. Clearance of ethanol may be reduced in

OC users.

The interactions of antiretroviral drugs with contraceptive steroids are complex.

Some of the drugs increase plasma steroid levels and some reduce them. A

complete list of interactions is available in the Centers for Disease Control and

Prevention’s “U.S. Medical Eligibility for Contraception” (7).

Oral Contraceptives and Clinical Chemistry Alterations

OCs have the potential to alter a number of clinical laboratory tests as a result of

estrogen-induced changes in hepatic synthesis; however, a large study comparing

OC users with pregnant and nonpregnant controls found minimal changes (228).

Hormone users took a variety of OCs containing 50 to 100 μg of estrogen, higher

doses than are used today. Compared with nonpregnant women who were not

using OCs, the OC users had an increase in T4 that is explained by increased

circulating thyroid-binding protein, no change in creatinine and globulin levels,

slight reduction in mean fasting glucose values and serum glutamic oxaloacetic

812transaminase, and a decrease in total bilirubin and alkaline phosphatase.

Choice of Oral Contraceptives

More recently introduced OCs include those containing drospirenone, more

preparations with only 20 μg of EE, new multiphasic preparations, cyclic OCs

with 24 days of active medication and 4 days of either placebo or 10 μg of EE,

and extended-cycle and continuous-cycle preparations plus branded generic

versions of most OCs. A combination OC containing estradiol valerate with a

new progestin, dienogest, was approved in 2010. There is new evidence that 20-

μg EE pills offer reduced risk for venous thrombosis (145). An approach to OC

selection for new patients is to begin with a 20-μg EE combination and then

adjust depending on the patient’s symptoms after the first 2 to 3 months. As

reported in a large systematic review of 20-μg EE OCs compared to 30- to 35-μg

EE pills, women taking the lower-dose OCs more often report changes in vaginal

bleeding, episodes of irregular bleeding and heavy bleeding, and more

amenorrhea (229). If a 20-μg pill is offered, the patient should understand that

bleeding may be a problem and that she should return if this persists and try a

different OC rather than stop the pill. The progestin component may become more

important in determining cycle control when 20-μg EE is used. A comparison of

20-μg EE OCs, one containing 100 μg of levonorgestrel, the other with 1 mg of

norethindrone acetate (NEA) found the NEA to have about twice as many days of

unscheduled vaginal bleeding during the first 3 months, a critical time period for

new users, which might be expected to lead to discontinuation (230). In a threeway trial, the 35-μg EE/norgestimate triphasic OC (Tri-Cyclen) was compared

with two 20-μg EE pills, one containing 100 μg of levonorgestrel (Alesse), the

other containing 150-μg desogestrel, followed by 2 hormone-free days and 5 days

of 10-μg EE per day (Mircette) (231). Contraceptive efficacy was not

significantly different. In the first two-pill cycles, more women taking Alesse had

breakthrough bleeding and bleeding in the second half of the cycle than those

taking the other two OCs in the first two cycles, but thereafter there was little

difference. Women taking the higher-estrogen Tri-Cyclen consistently

experienced more frequent estrogenic side effects—bloating, breast tenderness,

and nausea—than did women taking either 20-μg EE OC. These authors

concluded that for the specific OCs evaluated, changing to either one of the 20-μg

preparations would be beneficial.

For the average patient, the first choice of preparation for contraceptive

purposes is a very low estrogen OC (20-lg EE) unless there are other

considerations, for example, previous pregnancy while taking the pill.

Patients with persistent breakthrough spotting or bleeding could be offered a

pill with the same low estrogen dose, but a more potent progestin, for

813example levonorgestrel. Patients with apparent weight gain from fluid retention

while taking OCs, or with hirsutism or acne that did not respond to other OCs,

may benefit from a change to the drospirenone/EE pill. The lower-estrogen

drospirenone combination is FDA approved for treatment of PMDD and should

be considered for women with these symptoms who want hormonal

contraception. Women with acne often benefit from the reduction in

circulating testosterone that occurs with all combination OCs. Women who

experience continuing pelvic pain, dysmenorrhea, or other menstrualtriggered symptoms or who simply prefer fewer menstruations may be

offered an extended-cycle or continuous-cycle OC regimen.

Alternative Routes for Hormonal Contraception

The Ortho Evra patch and the NuvaRing provide combinations of ultrapotent

progestins with EE. Both patch and ring provide almost constant low levels of the

contraceptive steroids that are less than the peak levels seen with OCs. Both offer

greater convenience to the user, which improves compliance. The patch has a

surface area of 20 cm2. It delivers a daily dose of 150-μg norelgestromin, the

active metabolite of norgestimate, and 20 μg of EE. The patch is worn for 1 week

then replaced with a new patch for 7 days, continuing for 3 consecutive weeks

followed by a week with no patch. The patch was compared with a multiphasic

OC containing levonorgestrel 50 to 125 μg, and 30- to 40-μg EE (Triphasil) in a

randomized trial of 1,417 women (137). The overall and method failure Pearl

indices were 1.24 and 0.99 pregnancies per 100 woman-years in the patch group

and 2.18 and 1.23 in the OC group, respectively, numerically less in the patch

group but not statistically significant. Patch users had more breakthrough

bleeding or spotting in the first two cycles, but thereafter this did not differ from

OC users. Patch users reported more breast discomfort, dysmenorrhea, and

abdominal pain than the OC users, but other adverse events were uncommon and

did not differ. Perfect compliance was reported for 88.2% of patch users’ cycles

versus 77.7% of the pill users’ cycles (p <0.001). Pregnancy risk with the patch

appears to be higher for women weighing more than 90 kg.

The NuvaRing is 54 mm in the outer diameter and has a cross section of 4 mm.

It delivers daily doses of 120 μg of etonogestrel, the active metabolite of

desogestrel, with 15 μg of EE, and thus is the lowest estrogen combination

hormonal method available in the United States. The soft, flexible ring is worn in

the vagina for 3 weeks, and then removed for 1 week, after which time a new ring

is inserted. Alternatively, similar to OCs, women trying to avoid menstrual

bleeding can simply use the ring for an entire month and then change immediately

to a new ring. In a pharmacokinetic study comparing the ring with a combination

814OC containing 150 μg of desogestrel and 30 μg of EE, maximum blood levels of

EE with the ring were about one-third of those seen with the OC, and the

etonogestrel level was about 40% of that produced by the OC. Despite these

findings, ovulation was inhibited in all women studied (232). Women wearing the

ring are reported to have fewer days of irregular bleeding or spotting than women

taking an OC with 150 μg of levonorgestrel and 30 μg of EE (233). A large study

found a Pearl Index of 1.18 (95% CI, 0.73–1.80) for the vaginal ring (233). Some

women prefer to remove the ring for intercourse, although this is not necessary. It

should be reinserted within 3 hours to avoid loss of efficacy. If the ring is out of

the vagina for more than 3 consecutive hours, contraceptive efficacy may be

reduced and a barrier method should be used for 7 days.

Patch and Ring and Thrombosis

Since both methods provide constant low blood levels of EE, it was hoped that

this might reduce the risk for thrombosis. The FDA took the unusual step of

issuing a “black box warning” for the EE/norelgestromin patch after several

thrombosis cases were reported, and a small study of pharmacodynamics was

interpreted as showing higher mean EE blood levels for the patch than with the

oral route of administration. A small crossover study found increased APC

resistance when women were changed from a 30-μg EE/150-μg desogestrel pill to

either the patch or the EE/etonogestrel ring. This was interpreted as

prothrombotic (234). In another study, women on a variety of OCs had baseline

measurement of APC resistance, and protein S. Then the patients were switched

to either the patch or the ring. Those moving to the patch had changes in

laboratory parameters that could be interpreted as prothrombotic, while those

moved to the ring showed an improvement in the same studies, theoretically

reducing the risk of clotting (235). Three data-based epidemiologic studies looked

at thrombosis in patch users compared to oral contraception. The first found no

difference, but two subsequent studies did find increased risk. One study found an

overall doubling of the risk of deep vein thrombosis when patch users were

compared to users of an OC containing 30-μg EE/150-μg levonorgestrel, but the

risk estimate was of borderline statistical significance and disappeared when the

analysis was restricted to women aged 39 years or less (236). Following the

publication of these studies, the FDA convened in December 2011 to discuss the

possibly increased risk of blood clots in users of Ortho Evra compared to women

who use certain birth control pills. The Joint Advisory Committee recommended

that Ortho Evra remain in the market as an option for contraception but that the

label be revised to more clearly describe the risks and benefits associated with

Ortho Evra (237).

No similar FDA warnings were issued about the ring. There are two case

815reports in the world literature of venous thrombosis with the ring. Both were

women in their 30s who sustained cerebral venous thrombosis (238,239). It is best

to assume that the ring has the same risk for thrombosis as the other combination

hormonal contraceptives and to counsel women about risk in the same way.

Future of the Vaginal Ring

The Population Council has developed a one-year contraceptive vaginal ring that

releases 150-mcg Nestorone, a nonandrogenic progestin that is inactive orally,

and 15-mcg EE daily. It can be worn for 3 weeks, removed for 1 week, and then

reinserted for up to 13 cycles (240). Refrigeration is unnecessary, and this method

may be especially advantageous for women in low resource settings where access

to a pharmacy and electricity are unreliable. The Population Council will submit a

New Drug Application (NDA) to the U.S. FDA and pursue regulatory approval in

low- and middle-income countries.

Injectable Hormonal Contraceptives

Depot Medroxyprogesterone Acetate

DMPA, a suspension of microcrystals of a synthetic progestin, was approved for

contraception in 1992. A single 150-mg intramuscular dose will suppress

ovulation in most women for 14 weeks or longer (241). [7] The regimen of 150

mg every 3 months is highly effective, producing pregnancy rates of about 0.3 per

100 women per year. Probably because of the high blood levels of the progestin,

efficacy appears not to be reduced by administration of other drugs and is not

dependent on the patient’s weight. Women treated with DMPA experience

disruption of the menstrual cycle and have initial spotting and bleeding at

irregular intervals. Eventually, total amenorrhea develops in most women who

take DMPA; with continued administration, amenorrhea develops in 50% of

women by 1 year and in 80% by 3 years (Fig. 14-9).

The most important medical reason women discontinue the use of DMPA

and other progestin-only methods is persistent irregular vaginal bleeding. A

variety of medications are used to stop this bleeding. Many are effective in

terminating individual bleeding episodes, but a systematic review concluded that

none improved continuation rates long term (242). New approaches include

mifepristone and low-dose doxycycline. Mifepristone is of interest because

irregular bleeding with DMPA was related to the downregulation of endometrial

estrogen receptors. Treatment with 50 mg of mifepristone every 2 weeks increases

endometrial estrogen receptors and reduces breakthrough bleeding in new users of

both DMPA and the levonorgestrel progestin implant (243). Another line of

investigation concerns endometrial matrix metalloproteinase, which appears to

816play a regulatory role in the breakdown of the endometrium to produce normal

menstruation. Treatment with doxycycline inhibits matrix metalloproteinase

production in the endometrium in women after insertion of levonorgestrel

subdermal implants (244). Five-day courses of either mifepristone 25 mg twice a

day for 1 day, followed by EE 20 μg per day for 4 days, or doxycycline 20 mg

twice a day for 5 days, reduced bleeding days by about 50% when compared to

placebos during a 6-month randomized trial in women treated for prolonged or

frequent bleeding occurring with etonogestrel subdermal implants (245). Capsules

of 20-mg doxycycline are sold in the United States for treatment of periodontal

disease.

817FIGURE 14-9 Bleeding pattern and duration of use of depomedroxyprogesterone acetate

(DMPA): percentage of women who have bleeding, spotting, or amenorrhea while taking

DMPA 150 mg every 3 months. (From Schwallie PC, Assenzo JR. Contraceptive useefficacy study utilizing medroxyprogesterone acetate administered as an intramuscular

injection once every 90 days. Fertil Steril 1973;24:331–339, with permission.)

DMPA use is commonly associated with weight gain, and that is another

principal reason women discontinue use. A large study is representative of the

literature. Three cohorts of women who chose their method of contraception were

818followed for 36 months with measurements of weight and body fat. DMPA users

gained an average of 5.1 kg. The cohort using OCs over the same interval gained

only 1.47 kg, slightly less than the 2.05 kg gained by the cohort not using any

hormonal contraception. Total body fat increased 4.14 kg in the DMPA cohort,

while the increase in the OC users was 1.9 kg, only slightly more than the 1.17 kg

in the nonhormonal contraceptive cohort. Many women were followed for 2 years

after discontinuing DMPA. Those who chose nonhormonal methods after

discontinuing DMPA lost a mean of 0.42 kg each 6 months. Those who chose to

use OCs gained a mean of 0.43 kg each 6 months during the follow-up interval

(246).

Weight gain during the first 6 months and self-reporting of increased appetite

are strongly predictive of continued weight gain. Women who gained less than or

equal to 5% of body weight in the first 6 months gained a mean of 2.49 kg by 36

months, while those who gained more than 5% by 6 months gained a total mean

of 11.08 kg by 36 months (247). Studies are needed of interventions to prevent

the weight gain. At minimum, women considering DMPA need to know the

possibility of significant weight gain and be advised to avoid calorie-dense foods

and weigh themselves regularly. Ideally they should be weighed when they return

for subsequent injections so they can be counseled about the need for avoiding

further gain. Women who gain 5% of body weight by 6 months should consider

other contraceptive options.

DMPA persists in the body for several months in women who used it for longterm contraception, and return to fertility may be delayed. Studies of return to

ovulation indicated a wide variation in time to ovulation post injection with the

majority ranging from 15 to 49 weeks from the last injection for DMPA (248). In

a cohort of Thai women, 70% of former users desiring pregnancy conceive within

12 months, and 90% conceive within 24 months after terminating DMPA use

(249).

Safety

DMPA suppresses ovarian estrogen production. Prospective studies demonstrated

bone loss during DMPA therapy, with recovery of bone mass after DMPA use is

discontinued (250). Although bone loss was observed, it did not approach the

fracture threshold. Similar bone loss and then recovery occurs with lactation.

Adolescents are of special concern because they normally gain bone mass; most

of adult bone mass is attained by age 20. Estrogen injections prevent the bone loss

and allow adolescent women to gain bone density despite the use of DMPA (251).

A long-term study in adolescents documented bone density loss and confirmed

recovery of lumbosacral bone mineral density to baseline by 60 weeks after

discontinuation of DMPA and significant gain above baseline by 180 weeks.

819Recovery of density at the hip was slower, 240 weeks to significant gain (252). A

systematic review of DMPA clinical trials could find no studies where fracture

was an outcome, so whether long-term use leads to fractures is still not known

(253). The FDA black box warning added to DMPA labeling proposes that DMPA

treatment be limited to 2 years at a time, unless the patient has no other good

options for contraception. For many women, especially in developing countries,

DMPA is often the only option for highly effective contraception because it is

inexpensive and easy to administer. The issue should be discussed with women

who are considering DMPA, but DMPA should not be routinely discontinued after

2 years unless the patient wants to conceive or wants to change to another

contraceptive method for other reasons.

The effect of DMPA on plasma lipids is inconsistent; DMPA users appear to

have reduced total cholesterol and triglyceride levels, slight reduction in HDL

cholesterol, and no change or slight increase in LDL cholesterol, all of which are

consistent with a reduction in circulating estrogen levels. In some studies, the

decrease in HDL and increase in LDL are statistically significant, although the

values remain within normal ranges (254). DMPA is not associated with

myocardial infarction. Glucose tolerance tests disclose a small elevation of

glucose in DMPA users.

There is no change in hemostatic parameters, with the exception that ATIII

levels are sometimes reduced with chronic therapy (254). As noted earlier, large

epidemiologic studies have not found DMPA to be associated with thrombosis

(145). DMPA is not associated with teratogenicity (255). Nor is it associated with

affective disorders or mood changes (256).

DMPA and Lactation

There is widespread support for the use of DMPA during lactation when the

DMPA is initiated at or after 6 weeks postpartum. There is good evidence that

neither infant growth nor lactation is impaired by DMPA or progestin-only OCs

(257). There is continued controversy as to how early DMPA should be given

after delivery. Because lactation occurs in response to falling maternal estrogen

and progesterone levels after birth, administration of DMPA in the first few days

theoretically might interfere with the initiation of lactation. There is concern

about the possible neonatal effects of the progestin, but investigators were unable

to demonstrate the presence of DMPA or its metabolites in the urine of infants

whose mothers received DMPA or any other suppression of reproductive

hormones (257). In the United States, DMPA is commonly started at the time of

hospital discharge, 48 to 72 hours after delivery. In a systematic review, including

11 nonrandomized clinical trials and observational studies, progestogen-only

injectables were initiated in the first 6 weeks postpartum and most found either no

820effect on breastfeeding outcomes or improved outcomes among DMPA users

(19). The “U.S. Medical Eligibility Criteria for Contraceptive Use 2016”

considers DMPA use prior to 1 month postpartum as category 2: the benefit is

thought to exceed the theoretical risk (7).

DMPA and Neoplasia

The use of DMPA is not associated with cervical cancer (258). Neither is it

associated with ovarian cancer (259). The risk of endometrial cancer is

substantially reduced by past use of DMPA (260). A large study found no increase

in breast cancer risk among DMPA users (261).

Benefits

DMPA has many of the noncontraceptive benefits of combination OCs (262).

Decreases in anemia, PID, ectopic pregnancy, and endometrial cancer are

reported. DMPA is reported to benefit women with sickle cell disease (263).

Subcutaneous DMPA

Depo-subQ Provera 104, a lower-dose DMPA preparation for subcutaneous

administration, received FDA approval in 2005. The total dose is 30% less than

that of the DMPA intramuscular preparation. Because the dose is administered

subcutaneously, blood levels are adequate to completely suppress ovulation for

more than 13 weeks in all subjects tested, with a mean time of 30 weeks for return

to ovulatory function (264). Contraception efficacy is superb, with no pregnancies

in a total of 16,023 woman-cycles in the phase III studies done in the United

States (265). Blood levels were lower in very obese women but still sufficient to

completely suppress ovulation. The weight gain reported with the 150-mg DMPA

remains a problem with the lower-dose DMPA. Mean weight gain was 1.59 kg in

the first year of use. Loss of bone density was observed with this dosage of

DMPA, as with the larger intramuscular dose. In select countries, DMPA-SQ is

available in a single-use Uniject device and is sold under the name Sayana Press

(266). The Uniject is a small, prefilled, single-use injection device that any trained

person, including community health workers, pharmacists, and even women

themselves can administer. In an open-label observational study, current users of

DMPA IM found Sayana Press to be acceptable, with most users preferring

Sayana Press over DMPA IM, indicating promising uptake (267).

Once-a-Month Injectable

A once-a-month injectable contraception containing only 25 mg of DMPA in

combination with 5 mg of the long-acting estrogen estradiol cypionate was

briefly available in the United States, but was withdrawn by the manufacturer

821because of a packaging problem (268). Originally developed by the WHO, it is

described as Cyclofem or CycloProvera in the literature and was marketed in the

United States as Lunelle (269). Given once a month, this combination produces

excellent contraceptive effects. Monthly withdrawal bleeding is similar to a

normal menses, leading to high continuation rates despite the need for a monthly

injection. Monthly injectable combinations continue to be widely used outside the

United States.

Subdermal Implants

Three progestin-releasing subdermal implants systems are in use worldwide:

Jadelle (Bayer, Leverkusen, Germany), Sino-Implant II (Dahua, Shanghai,

China), and Nexplanon (Merck, Kenilworth, USA). All three offer long-acting

contraception that requires no continuing action by the user and are, hence,

forgettable. All are very highly effective and have no serious risk. Each can

produce irregular bleeding, which is the principal reason for discontinuation. The

mechanism of action is suppression of ovulation in the initial years of use, plus

thickening of the cervical mucous that prevents sperm penetration.

The original levonorgestrel implant (Norplant) six-rod system was replaced

with a two-rod version (Jadelle), which is identical in its release rate and clinical

activity to Norplant and is easier to insert and remove (270). Jadelle is widely

used around the world. It is approved by the FDA but not marketed in the United

States. It is approved for 5 years. The Sino-Implant II (Levoplant) is a less

expensive two-rod levonorgestrel system manufactured in China and available in

several countries. It is labeled for 3 years of use, with trials ongoing for a fourth

year (271). In June 2017, the product received WHO prequalification, which

recognizes that the product meets international quality standards for

manufacturing and clinical performance. This allows additional donors to

purchase the product for country programs. Thus, the availability of the SinoImplant II (Levoplant) is likely to increase. Nexplanon (or Implanon-NXT) is a

single-rod system containing etonogestrel, the active metabolite of desogestrel. It

was first approved in 2006 as Implanon. In 2011, Merck replaced Implanon with

Nexplanon, which contains 15 mg of barium sulfate to create a radiopaque quality

and utilizes a simpler applicator for easier, and potentially safer insertion. All

studies regarding safety and effectiveness of Implanon apply to Nexplanon.

A systematic review of 29,972 women and 28,108 woman-months of follow-up

with Norplant, Jadelle, or Implanon found no differences in pregnancies or

continuation rate over 4 years. No pregnancies occurred in any of the trials. There

were no differences in side effects or adverse events. The most common side

effect was unpredictable vaginal bleeding (272).

In 923 women followed for 20,648 treatment cycles in 11 studies, there were

822no pregnancies with the etonogestrel implant in place. Six pregnancies occurred

within 14 days of removal of the devices. [7] When these are included, as

required by the FDA, the cumulative Pearl Index was 0.38 pregnancies per 100

woman-years at 3 years (273). Irregular bleeding was a problem, but occurred

most frequently in the first 90 days of use and decreased over time. In a

randomized comparison with the six-rod Norplant, etonogestrel implant users had

less frequent vaginal bleeding, but more became amenorrheic (274). Other

commonly reported side effects are headache, weight gain, acne, breast

tenderness, and emotional lability (273). Only 2.3% of subjects discontinued

because of weight gain. Most women can use Nexplanon. The US MEC lists only

a small number of conditions as category 3, and these are based on theoretical

concerns without actual evidence of harm (7). Insertion in the postpartum period

appears to have no adverse effects on the mother or infant. Among women who

initiated the ETG implant or LNG-IUD 6 weeks postpartum, there was no

difference in the weight or height of their infants through 6 months (275). An

RCT found no difference in growth at 12 months among breastfed infants whose

mothers had etonogestrel implant inserted within 48 hours postpartum versus at 6

weeks postpartum (276). A study of etonogestrel implants compared with the

copper IUD placed 4 to 8 weeks postpartum found no differences in infant

growth, adverse events, respiratory or skin disorders, or developmental scores

(277). One population in which special considerations should be taken when

using progestin implants is HIV-positive women concurrently taking efavirenzbased antiretroviral therapy (ART), because efavirenz may reduce the

effectiveness of contraceptive implants. Pharmacokinetic studies have reported

reduced etonogestrel and levonorgestrel concentrations in implant users

concurrently using efavirenz-based ART, which raises concerns about potential

decreased contraceptive efficacy (278,279). In a retrospective cohort study of

HIV-positive women with contraceptive implants, the adjusted pregnancy rates in

efavirenz users were triple that of nevirapine users (280). Nevertheless, these

findings should not preclude women taking efavirenz from using implants

because even in this population, the implant remains more effective than nonLARC contraceptive methods.

Bone density is not affected by the etonogestrel implant, probably because

ovarian follicular activity is not totally suppressed and estradiol synthesis

continues (281). Enlarged follicular ovarian cysts are common during the first

year of use of Jadelle or etonogestrel implant and usually resolve spontaneously

(282).

[7] A comparative study of coagulation and fibrinolytic factors in users of

etonogestrel and levonorgestrel implants showed no significant changes from the

baseline, with the exception of a modest increase in ATIII and a small decrease in

823factor VII activity, changes that might reduce coagulability. Lipid levels and liver

function studies were not changed, with the exception of small elevations of

bilirubin, with somewhat more observed in levonorgestrel users than etonogestrel

users (283). Another study of hemostatic factors found modest decreases in many

measurements, within the range of normal, and a modest reduction in the

generation of thrombin in users of the etonogestrel implant (284). Taken together

these studies provide considerable reassurance that the progestin implants do not

increase thrombosis risk. With several million women now using the implants,

there are no published studies linking either implant to venous thrombosis or

myocardial infarction.

Emergency Contraception

Postcoital use of sex steroids to prevent pregnancy began in the 1960s with highdose estrogen taken daily for 5 days (285). This was replaced with the

combination OC containing EE and levonorgestrel for greater convenience (286).

[9] In the late 1990s, levonorgestrel alone became the method of choice after the

WHO showed its superiority in a large randomized trial with 1,998 women. The

pregnancy rate was 3.2% with the EE/levonorgestrel method and only 1.1% with

levonorgestrel alone (RR for pregnancy, 0.32; 95% CI, 0.18–0.70) for women

treated within 72 hours of intercourse. Nausea and vomiting occurred much less

frequently with levonorgestrel alone (23.1% vs. 50.5%, and 5.6% vs. 18.8%,

respectively) (287). The efficacy of both methods declined as the time after

intercourse increased. But even after 72 hours, the pregnancy rate with the

levonorgestrel treatment was only 2.7% (288). A single dose of 1.5-mg

levonorgestrel is just as effective as two doses of 0.75 mg, has no more side

effects, and is more convenient for the patient. Both dosing regimens are FDA

approved. While recognizing that levonorgestrel is more effective the sooner it is

taken, the WHO allows for the use of 1.5 mg of levonorgestrel as a single dose,

given up to 120 hours after intercourse (289). Research suggests that the main

mechanism of action is delay of ovulation. Noe et al. established that

levonorgestrel works only if it is administered prior to the day of ovulation (290).

No pregnancies occurred to 87 women who received levonorgestrel from 1 to 5

days prior to the day of ovulation. Seven pregnancies occurred to 35 women

treated on the day of ovulation or later. Postcoital levonorgestrel is not an

abortifacient because it is effective only when taken before ovulation.

Levonorgestrel alone is safer than the estrogen-containing preparations. There

were several case reports of thrombotic events after use of the

estrogen/levonorgestrel combination emergency contraception (291). No such

complications with levonorgestrel alone were published.

824Antiprogestins

The antiprogesterone mifepristone (RU486) is highly effective for postcoital

contraception. The usual abortifacient dose is 200 mg, but a dose of only 10 mg

is effective for emergency contraception. In one study, 2,065 women were

randomized to mifepristone, 10 mg, or levonorgestrel, two doses of 0.75 mg,

including women up to 120 hours after intercourse (292). The crude pregnancy

rate was 1.3% for mifepristone and 2.0% for levonorgestrel (p = 0.46). Side

effects were the same, and both methods were judged highly acceptable by the

patients. Mifepristone is not being developed for this use and is not available at

the appropriate dose.

[9] Ulipristal acetate, sold by the trade name of Ella, is a progesterone-receptor

modulator that was approved by the FDA and the European Union for emergency

contraception up to 120 hours after intercourse. It is a single 30-mg tablet,

available by prescription only, whereas levonorgestrel 1.5 mg has been available

without a prescription since 2013. When ovulation is imminent, ulipristal is more

effective than levonorgestrel in delaying ovulation. The odds of pregnancy for

ulipristal were 42% lower by 72 hours after intercourse and 65% lower in the first

24 hours (293). When ulipristal is administered prior to the LH peak, it delays the

rupture of the preovulatory follicle for 5 days or more, which may be its primary

mechanism of action (294). In women with BMI >30, the risk of pregnancy is

greater for those taking levonorgestrel (OR, 4.41) than for those taking ulipristal

(OR, 2.62) (295). Thus, for obese women, ulipristal is preferred over

levonorgestrel. Subsequent studies on EC and obesity have shown that the total

concentration of levonorgestrel for obese women with one dose of levonorgestrel

was significantly lower than the level observed in normal BMI women (296).

However, the concentration of ulipristal was similar between obese and normal

BMI women (297). Ulipristal is primarily metabolized by the CYP3A4 enzyme

system; hence, patients on drugs such as barbiturates, rifampin, and several of the

anticonvulsants may have reduced protection from pregnancy with it. With

regards to starting hormonal contraception following the use of emergency

contraceptive pills, hormonal contraception can be started immediately after

administration of levonorgestrel. However, in one study where women were given

either a placebo or a progestin-only pill, desogestrel, the day after ulipristal,

ovulation occurred in 45% of cycles in women receiving desogestrel compared to

3% in the placebo group (298). Therefore, in 2015, the Ella label was updated to

recommend that patients who use ulipristal wait 5 days before starting a hormonal

contraceptive method (299).

The Copper Intrauterine Device for Emergency Contraception

Postcoital insertion of a copper IUD was first reported by Lippes et al. in 1976

825(300). In initial trials, the copper IUD was inserted within 7 days of intercourse

and was more effective than steroids for emergency contraception. Subsequent

studies included women only within 5 days from intercourse. The multicenter trial

by Wu and colleagues is an example of the superb efficacy offered by the copper

T380A (301). Of 1,893 women who returned for a follow-up visit, there were no

pregnancies within 1 month of IUD insertion. [9] Efficacy for emergency

contraception is 100% when the device is inserted up to 5 days after

intercourse, and almost 100% up to 7 days after intercourse (301). An added

benefit was that 94% of the patients were continuing with the IUD at the 12-

month follow-up. There were no uterine perforations. Zhou and colleagues

reported similar excellent results in a large study with a different copper device,

the Multiload Cu-375 IUD (302). In much of the world, copper IUDs are very

inexpensive. Even in the United States where IUDs are costly, the benefit to the

patient of extremely effective emergency contraception, and long-term

contraception with one intervention, makes emergency IUDs cost effective.

Whether the LNG-IUD would work for emergency contraception is being studied.

Hormonal Contraception for Men

The same negative feedback of sex steroids that blocks ovulation in women

suppresses spermatogenesis in men, but it will produce loss of libido and

potentially extinguish sexual performance. The principle was first demonstrated

in 1974 using oral estrogen and methyl testosterone (303). Testosterone given

alone can suppress sperm production to very low levels while maintaining normal

libido and sexual performance. Over many years investigators have studied longacting testosterone salts for male contraception (304). Ethnicity is an important

predictor of efficacy of sperm suppression with testosterone therapy. Asian men

virtually always achieve azoospermia or oligospermia when treated with

testosterone undecanoate (TU) monthly injections, whereas only 86% of

Caucasian men achieved azoospermia or oligospermia or with similar testosterone

regimens (305). In a Chinese trial, 1,045 men were treated with monthly TU 500

mg. Only 4.8% failed to suppress to a sperm count less than 1 × 106 per mL. The

cumulative pregnancy rate was only 1.1 per 100 men at 30 months (306). In

Caucasian populations, testosterone was combined with progestins to further

suppress gonadotropin and improve efficacy. In an important trial with Caucasian

men, etonogestrel subdermal implants and TU injections were compared to

placebo implants and injections. Only 3% failed to suppress to a sperm count of

less than 1 × 106 (307). Side effects that were more common in the medicated

group than placebo included acne, night sweats, libido changes (usually

increased), and weight gain. In a study using Nestorone (progestin) and

826testosterone transdermal gels, 89% of men achieved suppression of their sperm

concentrations to ≤1 × 106 per mL, with minimal side effects (308).

Nonhormonal male methods are being investigated. One product called

reversible inhibition of sperm under guidance (RISUG) has been studied in both

preclinical and human research in India (309). It consists of styrene maleic

anhydride which when injected into the vasa differentia creates an incompatible

pH level for sperm and blocks transport. Vasalgel involves a similar approach but

uses a different formulation and is being developed in the United States. It is a

high–molecular-weight polymer consisting of styrene-alt-maleic acid dissolved in

dimethyl sulfoxide, which is injected into the vasa deferentia, where it remains as

a soft gel-like state that allows water-soluble substances to pass but not sperm. It

is meant to be reversible and in rabbit studies shows rapid restoration of sperm

flow after an injection of sodium bicarbonate (310). Thus far, Vasalgel has been

studied only in animal models (311). Vasalgel is expected to begin enrollment of

men in clinical trials in the United States.

STERILIZATION

Surgical sterilization is the most common method of fertility control used by

couples worldwide, with more than 250 million women and men relying on

female sterilization or vasectomy for contraception (312). Laparoscopic and

hysteroscopic techniques for women and vasectomy for men are safe and readily

available throughout the United States. The mean age at sterilization is 30 years.

Age younger than 30 years when sterilized, conflict within the marriage, and

divorce and remarriage are predictors of sterilization regret, which may lead to a

request for reversal of sterilization (313).

Female Sterilization

Hysterectomy is no longer considered for sterilization because morbidity and

mortality are too high in comparison with tubal sterilization. Vaginal tubal

sterilization, which was associated with occasional pelvic abscess, is rarely

performed in the United States. [11] Five procedures are used in the United

States.

1. Tubal sterilization at the time of laparotomy for a cesarean delivery or

other abdominal operation

2. Postpartum minilaparotomy soon after vaginal delivery

3. Interval minilaparotomy

4. Laparoscopy

5. Hysteroscopy

827Postpartum tubal sterilization at the time of cesarean delivery adds no risk

other than a slight prolongation of operating time; cesarean birth poses more

risk than vaginal birth, and planned sterilization should not influence the decision

to perform a cesarean delivery. Minilaparotomy can be performed in the

immediate postpartum state. The uterus is enlarged, and the fallopian tubes

lie in the midabdomen, easily accessible through a small, 3- to 4-cm

subumbilical incision.

Interval minilaparotomy, first described by Uchida, was rediscovered and

popularized in the early 1970s in response to an increased demand for sterilization

procedures and a simpler alternative to laparoscopy (314). Still widely practiced

in lower resource settings, it is uncommon in the United States because of

widespread availability of the endoscopic techniques.

Surgical Technique

The procedure usually elected for tubal sterilization by laparotomy is the Pomeroy

or modified Pomeroy technique (Fig. 14-10). In the classic Pomeroy procedure, a

loop of tube is excised after ligating the base of the loop with a single absorbable

suture. A modification of the procedure is excision of the midportion of the tube

after ligation of the segment with two separate absorbable sutures. This modified

procedure has several names: partial salpingectomy, Parkland Hospital

technique, separate sutures technique, and modified Pomeroy. In the Madlener

technique, now abandoned because of too many failures, a loop of tube is crushed

by cross-clamping its base, ligated with permanent suture, and excised. Pomeroy

and partial salpingectomy procedures have failure rates of 1 to 4 per 1,000 cases

(313). In contrast, pregnancy is exceedingly rare after tubal sterilization by the

Irving or Uchida methods. In the Irving method, the midportion of the tube is

excised, and the proximal stump of each tube is turned back and led into a small

stab wound in the wall of the uterus and sutured in place, creating a blind loop.

With the Uchida method, a saline–epinephrine solution (1:1,000) is injected

beneath the mucosa of the midportion of the tube, separating the mucosa from the

underlying tube. The mucosa is incised along the antimesenteric border of the

tube, and a tubal segment is excised under traction so that the ligated proximal

stump will retract beneath the mucosa when released. The mucosa is closed with

sutures, burying the proximal stump and separating it from the distal stump. In

Uchida’s personal series of more than 20,000 cases, there were no pregnancies

(314).

828FIGURE 14-10 Pomeroy technique for tubal sterilization. A: Ligation of loop of fallopian

tube. B: Subsequent excision of loop.

Laparoscopy

Laparoscopy is the most common method of interval sterilization in the

United States. In the standard laparoscopy technique, after the abdomen is

inflated with a gas, the laparoscope is inserted into the abdominal cavity through a

trocar, usually at the lower margin of the umbilicus. Although single puncture

laparoscopic sterilization procedures can be performed with an “operating

laparoscope,” it is increasingly common for a second, and sometimes third, trocar

to be inserted in the suprapubic region or lateral lower quadrants to allow the

insertion of special grasping forceps. Laparoscopic sterilization is usually

performed in the hospital under general anesthesia but can be performed under

local anesthesia with conscious sedation. Overnight hospitalization for

laparoscopy is rarely needed.

Techniques for Tubal Occlusion at Laparoscopy

Laparoscopic tubal occlusion is accomplished by any of four techniques: bipolar

electrical coagulation, application of a small Silastic rubber band (Falope ring),

829the spring clip (Hulka clip), or the titanium clip (Filshie clip). The Filshie clip,

first introduced in the United States in 1996, is used extensively in the United

Kingdom and Canada (315). It is a hinged device made of titanium with a liner of

silicone rubber tubing. Because of its lower pregnancy rate, the Filshie clip has

largely supplanted the Hulka clip (316).

In the bipolar electrocoagulation technique, the midisthmic portion of the tube

and adjacent mesosalpinx are grasped with special bipolar forceps, and

radiofrequency electric current is applied to three adjacent areas, coagulating 3

cm of tube (Fig. 14-11). The tube alone is recoagulated in the same places. The

radiofrequency generator must deliver at least 25 W into a 100-Ω resistance at the

probe tips to ensure coagulation of the complete thickness of the fallopian tube

and not just the outer layer; otherwise, the sterilization will fail (317).

To apply the Falope ring, the midisthmic portion of the tube is grasped, with

tongs advanced through a cylindrical sleeve that has the ring stretched around it

(Fig. 14-12A). A loop of tube is pulled back into the sleeve, and the outer

cylinder is advanced (Fig. 14-12B), releasing the Silastic ring around the base of

the loop of the tube, producing ischemic necrosis over time (Fig. 14-12C). If the

tube cannot be pulled easily into the applicator, the operator should stop and

change to electrical coagulation rather than persist and risk lacerating the tube

with the Falope ring applicator. The banded tube must be inspected at close range

through the laparoscope to demonstrate that the full thickness of the tube was

pulled through the Falope ring.

The Hulka clip is placed across the midisthmus, ensuring that the applicator is

at right angles to the tube and that the tube is completely contained within the clip

before the clip is closed. The Filshie clip is placed at right angles across the

midisthmus, taking care that the anvil of the posterior jaw can be visualized

through the mesosalpinx beyond the tube to ensure that the complete thickness of

the tube is completely within the jaws of the clip before it is closed (Fig. 14-13).

The cautery band or clip techniques each have advantages and disadvantages.

Bipolar coagulation can be used with any fallopian tube. The Falope ring and

Hulka and Filshie clips cannot be applied if the tube is thickened from previous

salpingitis. There is more pain during the first several hours after Falope ring

application. This can be prevented by bathing the tubes with a few milliliters of

0.5% bupivacaine just before ring placement (318). Failure of the Falope ring or

the clips generally results from misapplication, and pregnancy, if it occurs, is

usually intrauterine. After bipolar sterilization, pregnancy may result from

tuboperitoneal fistula and is ectopic in more than 50% of cases. If inadequate

electrical energy is used, a thin band of fallopian tube remains that contains the

intact lumen and allows intrauterine pregnancy to occur. Thermocoagulation, the

use of heat probes rather than electrical current, is employed extensively in

830Germany for laparoscopic tubal sterilization but is little used in the United States.

FIGURE 14-11 Technique for bipolar electrocoagulation tubal sterilization.

831832FIGURE 14-12 Placement of the Falope ring for tubal sterilization.

Laparoscopic Salpingectomy

Salpingectomy, either partial or complete, has not been traditionally used for

sterilization because of perceived comparative ease and safety of occlusion

methods. However, a large observational study compared salpingectomy with

other methods and found no increase in complication rates but an increase in

operative time of approximately 10 minutes (319). A recent study including

salpingectomies and tubal occlusion cases showed comparable immediate and

short-term complications rates across sterilization methods and an average

increase of 6 minutes of operative time for salpingectomy (320). Prophylactic

salpingectomy may reduce the risk of ovarian/ fallopian tube cancer because

most epithelial ovarian cancers arise in the distal fallopian tube, and is now

recommended when women are undergoing gynecologic surgeries for other

benign indications.

Noncontraceptive Benefits of Tubal Sterilization

In addition to providing excellent contraception, tubal ligation is associated

with reduced risk for ovarian cancer. Among the largest studies to illustrate the

relationship between tubal ligation and ovarian cancer were the Nurses’ Health

Study and Nurses’ Health Study II. These two prospective cohort studies included

29,340 women who had tubal ligation and found that tubal ligation overall was

associated with a decreased risk of ovarian cancer (HR, 0.76; 95% CI, 0.64–0.90)

(321). In a large prospective cohort study in the United Kingdom, the association

of tubal ligation with decreased risk varied by histology, with a significant

decrease in the risk of high-grade serous carcinomas (RR, 0.77; 95% CI, 0.67–

0.89), endometrioid (RR, 0.54; 95% CI, 0.43–0.69), and clear cell tumors (RR,

0.55; 95% CI, 0.39–0.77) (322).

833FIGURE 14-13 Filshie clip for tubal sterilization. (Courtesy of Femcare Ltd, Romsey,

UK.)

Furthermore, studies have explored the risks and benefits of opportunistic

salpingectomy compared to tubal ligation with regards to ovarian cancer risk

reduction. One study concluded that opportunistic salpingectomy at the time of

sterilization is cost effective with an incremental cost effectiveness ratio of

$31,432 per quality-adjusted life year (323). Another study comparing partial

salpingectomy (modified Pomeroy method) to complete salpingectomy (use of

LigaSure) at the time of cesarean delivery found no significant differences in

short-term complications (postpartum fever, wound infection, relaparotomy,

834estimated blood loss) (324). In addition, one trial which aimed to evaluate shortterm ovarian reserve by measuring antimüllerian hormone levels, detected no

significant difference between the groups, with an average increase of 0.58 ± 0.98

versus 0.39 ± 0.41 ng/mL in the salpingectomy and tubal ligation groups,

respectively (p = 0.45). Surgeries including salpingectomy were longer by an

average 13 minutes (66.0 ± 20.5 vs. 52.3 ± 15.8 minutes, p = 0.01) (325). A trend

may develop that favors salpingectomy over tubal occlusion or ligation for

sterilization.

Risks of Tubal Sterilization

Tubal sterilization is remarkably safe. The Collaborative Review of

Sterilization (CREST) study, a 1983 review of 9,475 interval sterilizations from

multiple centers in the United States, reported a total complication rate of 1.7 per

100 procedures (313). Complications were increased by the use of general

anesthesia, previous pelvic or abdominal surgery, history of PID, obesity, and

diabetes mellitus. The most common significant complication was unintended

laparotomy for sterilization after intra-abdominal adhesions were found. In

another series, 2,827 laparoscopic sterilizations were performed with the Silastic

band using local anesthesia and intravenous sedation (326). Only four cases could

not be completed (a technical failure rate of 0.14%), and laparotomy was never

needed. Rarely, salpingitis can occur as a complication of the surgery. This occurs

more often with electric coagulation than nonelectric techniques. The risk of

death with female sterilization was 1 to 2 per 100,000 sterilizations in a national

study based on data from 1979 to 1980 (313). Almost half of the deaths were

from complications of general anesthesia, usually related to the use of mask

ventilation. When general anesthesia is used for laparoscopy, endotracheal

intubation is mandatory because the pneumoperitoneum increases the risk of

aspiration. International data from the Association for Voluntary Surgical

Contraception show a similar record of safety from third-world programs: 4.7

deaths per 100,000 female sterilizations and 0.5 deaths per 100,000 vasectomies

(327).

Sterilization Failure

Many “failures” occur during the first month after surgery and are the

result of a pregnancy already begun when the sterilization was performed.

Contraception should be continued until the day of surgery, and a sensitive

pregnancy test should be routinely performed on the day of surgery. Because

implantation does not occur until 6 days after conception, a woman could

conceive just before the procedure and there would be no way to detect it.

Scheduling sterilization early in the menstrual cycle obviates the problem but

835adds to the logistic difficulty. Another cause of failure is the presence of anatomic

abnormalities, usually adhesions surrounding and obscuring one or both tubes. An

experienced laparoscopic surgeon with appropriate instruments usually can lyse

the adhesions, restore normal anatomic relations, and positively identify the tube.

In some circumstances, successful sterilization will not be possible by

laparoscopy, and the surgeon must know before surgery whether the patient is

prepared to undergo laparotomy, if necessary, to accomplish sterilization. The

CREST study reported on a cohort of 10,685 women sterilized from 1978 to 1986

at any of 16 participating centers in the United States who were followed from 8

to 14 years (313). The true failure rates for 10 years obtained by the life-table

method are given in Table 14-4. Pregnancies resulting from sterilization during

the luteal phase of the cycle in which the surgery was performed were excluded.

Of all remaining pregnancies, 33% were ectopic. The most effective methods at

10 years were unipolar coagulation at laparoscopy and postpartum partial

salpingectomy, generally a modified Pomeroy procedure. Bipolar tubal

coagulation and the Hulka-Clemens clip were least effective. The Filshie clip was

not evaluated because it was not in use in the United States. Younger women had

higher risk for failure, as would be expected because of their greater fecundity.

Table 14-4 Ten-Year Life-Table Cumulative Probability of Pregnancy per 1,000

Procedures With Different Methods of Tubal Sterilization, United States,

1978–1986

Unipolar coagulation 7.5

Postpartum partial salpingectomy 7.5

Silastic band (Falope or Yoon) 17.7

Interval partial salpingectomy 20.1

Bipolar coagulation 24.8

Hulka-Clemens clip 36.5

Total: all methods 18.5

From Peterson HB, Xia Z, Hughes JM, et al. The risk of pregnancy after tubal

sterilization: findings from the U.S. Collaborative Review of Sterilization. Am J Obstet

Gynecol 1996;174:1164; Table 2, with permission.

Since the CREST study began, sterilization by unipolar electrosurgery was

abandoned because of the risk of bowel burns and was replaced with bipolar

836electrosurgery or the nonelectric methods (tubal ring, Hulka-Clemens clip, and

the Filshie clip). An analysis of the CREST data found that bipolar

sterilization can have a very low long-term failure rate if an adequate portion

of the tube is coagulated. CREST study participants who were sterilized with

bipolar electrosurgery from 1985 to 1987 had lower failure rates than those

sterilized earlier (1978–1985). The important difference was in the application

technique of the electric energy to the tubes. Women whose bipolar procedure

involved coagulation at three sites or more had low 5-year failure rates (3.2 per

1,000 procedures), whereas women who had fewer than three sites of tubal

coagulation had a 5-year failure rate of 12.9 per 1,000 (p = 0.01) (328).

Family Health International reported large randomized multicenter trials of the

different means of tubal sterilization. The Filshie and Hulka clips were compared

in two trials. A total of 2,126 women were studied, of which 878 had either clip

placed by minilaparotomy and 1,248 had either clip placed by laparoscopy. The

women were evaluated at up to 24 months (329). Pregnancy rates were 1.1 per

1,000 women with the Filshie clip and 6.9 per 1,000 with the Hulka clip at 12

months, a difference in rates that approached statistical significance (p = 0.06).

This same group compared the Filshie clip with the Silastic tubal ring in a similar

study with a total of 2,746 women, of which 915 had the devices placed at

minilaparotomy and 1,831 at laparoscopy (330). Pregnancy rates at 12 months

were the same for the Filshie clip and the tubal ring: 1.7 per 1,000 women. The

ring was judged more difficult to apply. The Filshie clip was expelled

spontaneously from the vagina by three women during the 12 months of followup.

Hysteroscopy

In 2002, the FDA approved Essure, a hysteroscopic method of permanent birth

control. It can be provided in an office setting, with only local anesthesia or

conscious sedation and both offer the prospect of greater safety, lower cost, and

greater long-term effectiveness than the best laparoscopy [12] methods. Essure is

a microinsert consisting of a soft stainless steel inner coil and a dynamic

nickel titanium alloy outer coil (Fig. 14-14). Soft fibers of polyethylene

terephthalate run along and through the inner coil. To insert the device, a

hysteroscope is introduced into the uterine cavity, which is distended with saline.

The tubal ostia are visualized. The Essure device is inserted through the operating

channel of the hysteroscope on the end of a slender delivery wire, guided into the

tubal opening and advanced into the tube under direct vision (Fig. 14-15). Once

in place, an outer sheath is retracted, releasing the outer coils, which expand to

anchor the device in the interstitial portion of the tube. The delivery wire is

detached and removed and the procedure repeated for the other tube. When

837properly placed, three to eight of the end coils of the microinsert are visible inside

the uterine cavity. The rest are inside the fallopian tube (331,332).

[12] Essure can be installed under local anesthesia in an outpatient setting.

No incision is needed. Over time fibrous tissue grows, occluding the tubes

permanently. The Essure Confirmation Test using either a transvaginal ultrasound

(TVUS) or modified hysterosalpingogram (HSG) is required. The FDA approved

the TVUS as an alternative confirmation test in July 2015. Patients must meet all

criteria at the time of placement to be eligible for TVUS. Otherwise, they must

undergo confirmation with modified HSG. A few notable criteria include

certainty about bilateral placement, procedure time ≤15 minutes, and one to eight

trailing coils bilaterally (333). Patients who are on active immunosuppressive

therapy (systemic corticosteroids or chemotherapy) are not eligible for TVUS as a

result of concern for failure of tissue ingrowth. The modified HSG for Essure

should document bilateral placement of the devices at the uterotubal junction and

a lack of peritoneal spillage of dye. The patient with transcervical sterilization

should continue to use a reliable method of contraception until successful

occlusion is documented.

838FIGURE 14-14 Essure device for hysteroscopic sterilization. (Courtesy of Conceptus,

Inc.)

839FIGURE 14-15 Essure device with guide wire and handle. (Courtesy of Conceptus, Inc.)

Risks of Hysteroscopy

Adverse events or side effects were reported on the day of procedure in 3% of the

initial Essure patients. These consisted of vasovagal responses, cramping, nausea,

and vaginal spotting (334). Possible but uncommon risks of the hysteroscopic

tubal sterilization methods include perforation by the device at insertion and

expulsion of the device. Tubal and uterine perforation was reported to range from

0% to 1.8% (335). Other potential combinations with transcervical sterilization

are related to the hysteroscopy procedure, not the tubal occlusion process. These

include hypervolemia, injury to surrounding organs, bleeding, and infection and

occur in less than 1% of cases.

840[12] Between November 2002 and May 2015, the FDA received a total of 5,093

reports related to Essure, prompting it to convene a panel to review the safety and

effectiveness of Essure sterilization. In 2016, the FDA ordered the manufacturer

to conduct a postmarket surveillance study to determine heightened risks for

particular women, and issue a black box warning about risk of perforation,

persistent pain and suspected allergic or hypersensitivity reactions (336).

Subsequently, in July 2018, due to a decline in sales, the manufacturer removed

the device from the market.

Sterilization Failure With Essure

A meta-analysis of 31 studies including more than 12,000 procedures reported

bilateral placement rates of 81% to 98% with higher success rates reported from

studies published since 2007 (337). Reasons for failure of insertion were tubal

obstruction, stenosis, or difficulty in accessing the tubal ostia. In a review of over

66,000 Essure placements, there were 102 pregnancies (1.5 per 1,000). Seventysix percent occurred in women who did not have postprocedure imaging to

confirm tubal occlusion, did not have reliable contraception prior to confirmation

imaging or had misread imaging (338). Other studies report similar rates of

pregnancy, ranging from 1.2 to 1.5 per 1,000 (339).

Reversal of Sterilization

Reversal of sterilization is more successful after mechanical occlusion than

after electrocoagulation, because the latter method destroys much more of

the tube. With modern microsurgical techniques and an isthmus-to-isthmus

anastomosis, the first-year pregnancy rate is about 44% (340). A retrospective

review comparing microsurgical reanastomosis to in vitro fertilization (IVF)

found similar total pregnancy rates, but a higher live birthrate with IVF because

10% of the pregnancies after microsurgery were ectopic (340). Hysteroscopic

sterilization by Essure should be considered irreversible. However, reversal

following hysteroscopic sterilization has been reported using tubouterine

implantation. Thirty-six percent of those undergoing surgery subsequently

conceived naturally (341).

Late Sequelae of Tubal Sterilization

Increased menstrual irregularity and pain are attributed to previous tubal

sterilization. Study of the problem is complicated by the fact that many women

develop these symptoms as they age, even though they did not have tubal surgery,

and are treated with OCs that reduce pain and create an artificially normal

menstrual cycle. Women who discontinue OC use after tubal sterilization will

experience more dysmenorrhea and menstrual irregularity, which is unrelated to

841the sterilization. The best answer available comes from the CREST study (342). A

total of 9,514 women who underwent tubal sterilization were compared with 573

women whose partners had undergone vasectomy. Both groups were followed up

to 5 years with annual standardized telephone interviews. Women who underwent

tubal sterilization were no more likely to report persistent changes in

intermenstrual bleeding or length of the menstrual cycle than women whose

partners had vasectomy. The sterilized women reported decreases in the days of

bleeding, amount of bleeding, and menstrual pain but were slightly more likely to

report cycle irregularity (OR, 1.6; 95% CI, 1.1–2.3). In summary, the CREST

study provided good evidence that there is no “posttubal ligation syndrome.”

Vasectomy

About 500,000 vasectomies are performed each year in the United States (313).

Vasectomy is a highly effective method. The literature on efficacy is often

difficult to interpret because most studies report failure as failure to achieve

azoospermia, rather than long-term pregnancy rates among the relevant women.

Vasectomy is not effective until all sperm are cleared by the reproductive

tract, which can take up to 20 ejaculations or 3 months. It is estimated that

up to half of pregnancies after vasectomy occur during the interval between

the surgery and the clearance of all sperm (343). Not all pregnancies after

vasectomy can be attributed to the men who had the operation. [13] The best

long-term information comes from the CREST study. The cumulative probability

of failure was 7.4 per 1,000 procedures at 1 year and 11.3 at year 5, and

comparable to the failure rate of tubal sterilization (313).

[13] Vasectomy is usually performed under local anesthesia. The basic technique

is to palpate the vas through the scrotum, grasp it with fingers or atraumatic

forceps, make a small incision over the vas, and pull a loop of the vas into the

incision. A small segment is removed, and a needle electrode is used to coagulate

the lumen of both ends. Improved techniques include the no-scalpel vasectomy, in

which the pointed end of the forceps is used to puncture the skin over the vas.

This technique reduces the risk of hematoma (344). Another variation is the openended vasectomy, in which only the abdominal end of the severed vas is

coagulated while the testicular end is left open. This is believed to prevent

congestive epididymitis and sperm granuloma (345). Fascial interposition creates

a tissue barrier between the vasal ends and reduces vasectomy failure (346).

Reversibility

Vasectomy must be regarded as a permanent means of sterilization. However,

among men undergoing surgical reanastomosis (vasovasostomy), subsequent

842patency has been reported in over 90% of cases when current microsurgical

techniques are employed (347). Nevertheless, the longer the interval since

vasectomy, the poorer is the chance of reversal (348).

Safety

Operative complications include scrotal hematomas, wound infection, and

epididymitis, but serious sequelae are rare. There were no reports of deaths from

vasectomy in the United States in many years, and the death rate in a large thirdworld series was only 0.5 per 100,000. Concerns about long-term safety recurred

with the report of a possible association between prostate cancer and vasectomy

(349). [13] Large studies present the very clear conclusion that vasectomy is

not associated with prostate cancer (350–352).

843ABORTION

It is extremely likely that couples will experience at least one unwanted

pregnancy at some time during their reproductive years. In developing

countries, desired family size is larger (although decreasing), but access to

effective contraception is limited. As a result, abortion, often unsafe, is common.

Because of population growth, the absolute number of abortions increased by 5.9

million, from 50.4 million per year from 1990 to 1994 to 56.3 million per year

from 2010 to 2014 (353). Globally, 25% of pregnancies ended in abortion from

2010 to 2014. In the developed world, the percent of pregnancies ending in

abortion declined by 11 points from 39% to 28%, whereas in the developing

world, it increased significantly by 3 points from 21% to 24% (353). It is

estimated that 25 million, or nearly half of all abortions are “unsafe” where either

the persons terminating the pregnancy lack the necessary skills or the

environment in which it is performed does not conform to minimal medical

standards, or both (354).

As shown in Figure 14-16, there are large discrepancies in the rate of safe

abortions compared with unsafe abortions worldwide. In the developed countries

where there is access to legal abortion, unsafe abortions are rare. In developing

countries, where abortion is often highly restricted or completely illegal, they are

very common. The highest proportion of unsafe abortions are found in Middle

and Western Africa, where mortality rates from abortion are highest. The

subregions with the highest proportions of safe abortions (northern Europe and

northern America) include countries with less restrictive laws on abortion, high

contraceptive use, high economic development, high levels of gender equality,

and well-developed health infrastructures (354). [14] Where abortion is legal, it

is generally safe; where it is illegal, complications are common. Forty-seven

thousand women die every year from complications of unsafe abortion (355).

Societies cannot prevent abortion, but they can determine whether it will be

illegal and dangerous or legal and safe. [14] Death from illegal abortion was once

common in the United States. In the 1940s, more than 1,000 women died each

year of complications from abortion (356). In 1972, 24 women died of

complications of legal abortion and 39 died from known illegal abortions. In

2012, the most recent year for which complete data are available, there were four

deaths from legally induced abortion and no deaths from illegal abortion (abortion

induced by a nonprofessional) in the entire United States (6). The American

Medical Association’s Council on Scientific Affairs reviewed the impact of legal

abortion and attributes the decline in deaths during this century to the introduction

of antibiotics to treat sepsis; the widespread use of effective contraception

844beginning in the 1960s, which reduced the number of unwanted pregnancies; and

the shift from illegal to legal abortion (357).

The number of abortions reported each year in the United States—926,000 in

2014—has been declining since the peak level of 1.61 million in 1990 (5). In

2014, the national abortion ratio was 18.8 abortions for every 100 pregnancies,

and the national abortion rate was 14.6 per 1,000 women aged 15 to 44 years (5).

While abortion rates for all racial and ethnic groups have declined, they are

significantly higher for non-Hispanic black women. Non-Hispanic white women

have 37.5% of all abortions, non-Hispanic black women have 35.6%, and

Hispanic women 19%. Most women who obtain abortions are unmarried, 85.2%

in 2013 (6). The use of abortion varies markedly with age. In 2013, 11.3% of

women obtaining abortions were between 15 and 19 years of age, and 44.7%

were 24 years of age or younger. In 2013, the abortion ratio for women younger

than 15 years of age was 789 per 1,000 live births (Fig. 14-2) (6). The lowest

abortion ratio, 121 per 1,000 live births, is for women aged 30 to 34 years.

845FIGURE 14-16 Distribution of abortion safety categories worldwide and by region. Bars

and dots show the point estimates of the proportion of abortions in each category and

horizontal lines are 90% uncertainty intervals. (From Ganatra B, Gerdts C, Rossier C, et

al. Global, regional, and subregional classification of abortions by safety, 2010–14:

estimates from a Bayesian hierarchical model. Lancet 2017;390[10110]:2372–2381.)

Regardless of personal feelings about the ethics of interrupting pregnancy,

health professionals have a duty to know the medical facts about abortion and to

share them with their patients (358). Providers are not required to perform

abortions against their ethical principles, but they have a duty to help patients

assess pregnancy risks and to make appropriate referrals.

846Safety

[15] The overall annual risk of death with legal abortion decreased markedly,

from 4.1 per 100,000 in 1972 to 1.8 in 1976, and remained less than 1 per

100,000 since 1987. Today, legal abortion remains an extremely safe procedure,

with a mortality rate of 0.7 per 100,000 procedures (359). [15] Risk increases

exponentially with gestational age. Pregnancy-associated mortality rate with

aspiration procedures was 0.3 per 100,000 at or before 8 weeks and increased to

6.7 per 100,000 at 18 weeks or greater (359). Nevertheless, in comparison to a

maternal mortality rate of 8.8 deaths per 100,000 live births, abortion by dilation

and evacuation (D&E) beyond 18 weeks’ gestation is safer than continuing

pregnancy (360). It was estimated that 87% of the legal abortion deaths occurring

after 8 weeks would have been prevented had the woman been able get abortion

services by 8 weeks (361).

For individual women with high-risk conditions (e.g., cyanotic heart disease),

even late abortion is a safer alternative to birth. Because of the availability of lowcost, out-of-hospital, first-trimester abortion, 88% of legal abortions are

performed during the first trimester (before 13 weeks of amenorrhea), when

abortion is the safest.

Techniques for First-Trimester Abortion

Vacuum Curettage

Most first-trimester abortions are performed by vacuum curettage. Most are

performed with local anesthesia with or without moderate sedation, and usually

on an outpatient basis in a freestanding specialty clinic or doctor’s office (362).

Cervical dilation is accomplished with metal or plastic dilators, or by osmotic

cervical dilators or misoprostol 400 μg, given vaginally or by the buccal route 3 to

4 hours before the procedure (363). A plastic vacuum cannula of 5- to 12-mm

diameter is used with a manual vacuum source or an electric vacuum pump.

Manual vacuum provided by a modified 50-mL syringe is as effective as the

electric pump through 10 menstrual weeks (364). The Society of Family Planning

recommends a short course of antibiotics preoperatively to lower the risk of

infection following surgical abortion (365). A single dose of doxycycline is a safe

and effective prophylactic antibiotic and can be taken with dinner the night

preceding a procedure. A large retrospective study of women’s health care visits

within 6 weeks after an abortion using California Medicaid data estimated

abortion-related complication rates (366). Complications following first-trimester

aspiration procedures are presented in Table 14-5. Major complications, defined

as requiring hospital admission, surgery or blood transfusion, occurred in 0.16%

847of first-trimester aspiration abortions. The most common complications were

other or undetermined diagnoses, most of which led to repeat abortions. More

extensive descriptions of the management of complications are published

elsewhere (367).

Table 14-5 Distribution of Abortion-Related Complication Diagnoses Among 34,755

First-Trimester Aspiration Procedures

Complication Diagnosis

Incomplete abortion 116 (0.33)

Failed abortion 14 (0.04)

Hemorrhage 44 (0.13)

Infection 94 (0.27)

Uterine perforation 2 (0.01)

Anesthesia-related 2 (0.01)

Other or undetermineda 166 (0.48)

Total 438 (1.26)

aFor major complications, this diagnosis includes undetermined diagnoses that required

blood transfusions and surgery. For minor complications, the majority of this diagnosis

consisted of cases treated with repeat abortion, but the exact diagnosis could not be

determined. This category also includes diagnoses such as nonanesthesia-related allergic

reactions and seizures.

Data are n (%).

Adapted from Upadhyay UD, Desai S, Zlidar V, et al. Incidence of emergency

department visits and complications after abortion. Obstet Gynecol 2015;125(1):175–183.

Medical Abortion in the First Trimester

Mifepristone (RU486), an analog of the progestin norethindrone, has a strong

affinity for the progesterone receptor but acts as an antagonist, blocking the effect

of natural progesterone. Given alone, the drug was moderately effective in

causing abortion of early pregnancy; the combination of mifepristone with a low

dose of prostaglandin proved very effective, producing complete abortion in 96%

to 99% of cases (368). In 2016, the FDA approved updated labeling for

mifepristone. Changes included expansion of its use to 70 days from the start of

848the last menstrual period and a decrease in dose from 600 to 200 mg.

Furthermore, misoprostol dose was increased from 400 mcg orally to 800 mcg

buccally and home administration was allowed. Lastly, a posttreatment

examination visit to the clinic was no longer required. These label changes were

adopted because 200 mg of mifepristone was as effective as 600 mg, 800-mcg

misoprostol by vaginal or buccal routes provide higher efficacy than 400-mcg

oral dose, misoprostol can be taken at 24, 48, or 72 hours after mifepristone with

equal efficacy, and women may safely self-administer misoprostol at home (369–

373). Misoprostol can be taken by buccal or vaginal routes with similar blood

levels and similar areas under the plasma concentration curve (AUC) (374,375).

After case reports of death from Clostridium sordellii, a large cohort study was

published by Planned Parenthood describing the use of prophylactic doxycycline

100 mg twice a day for 1 week, and buccal rather than vaginal administration of

misoprostol (376). No additional cases of C. sordellii have been recently reported.

While individuals practitioners may decide to use antibiotics, the Society of

Family Planning does not believe universal antibiotics are required for all women

having a medical abortion.

Contraindications to medical abortion with mifepristone/ misoprostol include

ectopic pregnancy; an IUD in place (remove IUD first); chronic adrenal failure;

concurrent long-term corticosteroid therapy; history of allergy to mifepristone,

misoprostol, or other prostaglandins; hemorrhagic disorder; and inherited

porphyrias (377).

Methotrexate/Misoprostol and Misoprostol Alone

Alternatives to medical abortion when mifepristone is not available include

regimens with methotrexate/misoprostol and misoprostol alone. The antifolate

methotrexate provides another medical approach to pregnancy termination, but

takes longer than the technique using mifepristone/misoprostol (378). Medical

abortion can be induced with misoprostol alone, although it is less effective than

the mifepristone/misoprostol combination. Vaginal misoprostol, 800 μg, repeated

in 24 hours if fetal expulsion has not occurred, produces a complete abortion in

91% of pregnancies up to 56 days of amenorrhea (379).

Complications of Medical Abortion

Heavy or prolonged bleeding is the principal complication, with up to 8% of

women experiencing some bleeding for as long as 30 days. Need for surgical

curettage is predicted by the gestational age when the mifepristone is

administered. Two percent of women treated at 49 days or less, 3% of those

treated at 50 to 56 days, and 5% of those treated at 57 to 63 days needed curettage

for bleeding or failed abortion in a large study with 200 mg of mifepristone and

849800 μg of vaginal misoprostol (380). Late bleeding, at 3 to 5 weeks after

expulsion of the pregnancy, accounted for more than half of the curettages. Under

the updated protocol, serious complications requiring hospitalization for infection

treatment or transfusion occur in fewer than 0.4% of patients (381). The updated

regimen has been shown to cause fewer gastrointestinal side effects, such as

nausea, vomiting, and diarrhea (382).

Medication Abortion and Telemedicine

With the 2016 mifepristone FDA label update excluding the requirement of inclinic follow-up, telemedicine can facilitate the provision of medication abortions.

Removing barriers to clinic access for rural women or mitigating the need for

travel could reduce delays in receiving care and enable women to obtain an

abortion at the earliest stages of pregnancy when it is safest (383). In one study,

there were no differences in adverse events between in-person and telemedicine

patients seeking medication abortion, and telemedicine patients were more likely

to recommend the service to a friend (384).

Second-Trimester Abortion

Abortions performed after 13 weeks include those done for fetal anomalies,

medical illness, or delays resulting from pregnancy denial or difficulties making

arrangements. Young maternal age is the single greatest factor determining the

need for late abortion (6).

Dilation and Evacuation

D&E is the most commonly used method of midtrimester abortion in the United

States. Prior to surgery, it is essential to prepare the cervix (to make it soft and

pliable). There are a number of approaches to cervical ripening in this setting:

1. The cervix can be prepared by insertion of hygroscopic dilators, stems of the

seaweed Laminaria japonica (laminaria), or Dilapan-S hydrophilic polymer

rods. Placed in the cervical canal as small rods, these devices take up water

from the cervix and swell, producing mechanical dilation, and induce

endogenous prostaglandin synthesis, which aids in cervical softening. When

the dilators are removed the following day, sufficient cervical dilation is

accomplished to allow insertion of specialized forceps and a large-bore

vacuum cannula to extract the fetus and placenta (385,386). Ultrasound

guidance during the procedure is helpful (387). At the end of the midtrimester,

procedures that combine serial insertion of two or more sets of laminaria or

Dilapan-S for increased cervical dilation with a feticidal injection, or a

combination of dilators with misoprostol to induce labor, followed by assisted

850expulsion of the fetus are possible (388).

2. Misoprostol can also be used for cervical preparation prior to midtrimester

abortion. Osmotic dilator placement more effectively dilates and prepares the

cervix before D&E than misoprostol alone (389,390). However, misoprostol in

lieu of osmotic dilators has not been shown to increase the rate of serious

complications. A large case series of D&Es between 17 and 23 weeks

performed by highly experienced providers in women who received multidose

misoprostol regimen as an alternative to overnight dilation reported a major

complication rate less than 2% (391). However, the average dose of

misoprostol used was 1,200 mcg.

3. Mifepristone is another option. A recent study comparing mifepristone as an

adjunct or replacement for osmotic dilators in women undergoing D&E

between 19 and 23 weeks found that, while median total procedure time was

slightly longer for the mifepristone group compared to women who received

mifepristone and osmotic dilators or only osmotic dilators, there were no

differences in complications between the three groups (392). While some

providers insert osmotic dilators sequentially over the course of 2 days, an

RCT of D&Es from 19 to 23 weeks showed that there were no differences in

procedure times, initial cervical dilation, ease of procedure, or complications

in women who received overnight synthetic osmotic dilators with mifepristone

and misoprostol compared to women who received 2 days of dilators and

misoprostol (393).

Intact D&E is another modification useful for procedures at the end of the

midtrimester. After wide cervical dilation is achieved with serial placement of

cervical dilators, the membranes are ruptured and an assisted breech delivery is

performed, with decompression of the after-coming fetal head to allow delivery of

the fetus intact (394). In response to the federal abortion ban of 2003, an

increasing number of providers are using feticidal agents prior to late secondtrimester terminations (395). Intra-amniotic or intrafetal digoxin and intracardiac

potassium chloride are the two most common agents used for this purpose. Both

are efficacious with little maternal adverse effects (396,397). The clinical utility

of these agents is unproven. The only randomized controlled study looking at

clinical outcomes found no change in procedure time or blood loss with the use of

digoxin (398). Varying doses of digoxin were reported. In one study, 1.5-mg

intra-amniotic digoxin was always successful in inducing fetal demise within 24

hours (399).

Labor-Induction Methods

In Europe and the United Kingdom, labor induction is much more common than

851D&E for midtrimester abortion (400). Induction abortion with hypertonic saline

or urea was widely employed for labor induction abortion in the 1970s. These

were supplanted by the use of synthetic prostaglandins, and by regimens that

combine mifepristone and misoprostol.

Prostaglandins

Prostaglandins of the E and F series can cause uterine contraction at any stage of

gestation. The 15 methyl analogs of prostaglandin F2α (carboprost) and

prostaglandin E2 (dinoprostone) are highly effective for midtrimester abortion but

frequently produce side effects of vomiting, diarrhea, and with dinoprostone,

fever. Misoprostol, a 15 methyl analog of PGE1, is much less expensive than

other prostaglandins, stable at room temperature, and at doses effective for

abortion, produces fewer side effects (401). Transient fetal survival is not

infrequent after prostaglandin inductions. In the United States, it is common to

induce fetal demise before induction with regimens similar to those used in late

second-trimester D&E: intra-amniotic or intrafetal digoxin, 1 to 1.5 mg, or fetal

intracardiac potassium chloride (3 mL of a 2-mmol solution).

Midtrimester Mifepristone/Misoprostol

Mifepristone pretreatment markedly increases the abortifacient efficacy of

gemeprost and misoprostol. Mifepristone, 200 mg, is just as effective for this

purpose as 600 mg (402). A common protocol was developed using mifepristone

and misoprostol that is now recommended by the Royal College of Obstetricians

and Gynecologists (RCOG) and the WHO (403,404). Most women are cared for

as hospital day-patients without need for overnight admission, a marked

improvement over the labor induction methods of the past that often required 2 to

3 days of hospitalization. In the United States, ACOG recommends administering

mifepristone, 200 mg orally, followed in 24 to 48 hours by either 800-mcg

misoprostol vaginally followed by 400 mcg vaginally every 3 hours, or 400-mcg

misoprostol buccally every 3 hours (405). Earlier studies evaluated intervals of 24

to 48 hours between dosing mifepristone and misoprostol. A systematic review

including studies of 24- to 48-hour intervals calculated weighted median

induction times of 7.6 hours for 1-day interval and 6.8 hours for 2-day interval

(406). When comparing mifepristone–misoprostol intervals of 36 to 48 hours to

intervals of 24 hours or less, shorter intervals on average had only minimally

longer induction times (1 to 2 hours) but consistently much shorter overall total

procedure times (mifepristone to pregnancy expulsion).

Combination of Induction and Assisted Delivery

Hern developed a procedure that combines a feticidal injection of digoxin with

852serial insertion of multiple laminaria tents over 2 to 3 days, followed by

amniotomy, placement of misoprostol in the lower uterine segment, and

intravenous oxytocin to induce labor, and then an assisted delivery (407). The

procedure was successful in a large case series at 18 to 34 weeks with very few

complications.

Complications of Second-Trimester Abortion

Surgical Abortion Complications

Complications of second-trimester surgical abortion are uncommon, but risk

increases with gestational age. Complications and their frequency encountered in

almost 3,000 midtrimester abortions performed by laminaria followed by D&E on

a referral service are listed in Table 14-6. The gestational ages were 14 to 27

weeks and mean gestational age was 20.2 weeks. The most common complication

was a cervical laceration that required suturing. A major complication, defined as

one necessitating transfusion, disseminated intravascular coagulation,

reoperations with uterine artery embolization, laparoscopy, or laparotomy, was

encountered in 1.3% of patients. History of two or more cesareans, gestational

period of 20 weeks or more, and insufficient initial cervical dilation by laminaria

were independent risk factors for a major complication in a multivariate analysis

(408). The rates of complication in Table 14-6 cannot be directly compared to the

rates of complications with labor-induction abortion described below because half

of the D&E group were 20 weeks or more gestation, while only a few of the

induction patients were more than 20 weeks.

Table 14-6 Complications of 2,935 Midtrimester Dilation and Evacuation Abortion

and Intervention Rates on a Referral Service

Number

(%)

95% Confidence Interval

(%)

Complication

Cervical laceration 99 (3.3) 2.7–4.0

Atony 78 (2.6) 2.1–3.3

Hemorrhage 30 (1.0) 0.6–1.4

Other 15 (0.5) 0.3–0.8

853Disseminated intravascular

coagulation

7 (0.2) 0.1–0.4

Retained products 6 (0.2) 0.04–0.4

Perforation 6 (0.2) 0.04–0.4

Treatment of Complications

Reaspiration 46 (1.5) 1.1–2.0

Hospitalization 42 (1.4) 1.0–1.8

Transfusion 30 (1.0) 0.7–1.4

Uterine artery embolization 21 (0.7) 0.4–1.0

Laparoscopy or laparotomy 13 (0.4) 0.2–0.7

From Frick AC, Drey EA, Diedrich JT, et al. Effect of prior cesarean delivery on risk of

second-trimester surgical abortion complications. Obstet Gynecol 2010;115:762; Table 2,

with permission.

Induction Abortion Complications

The labor-induction methods share common hazards: failure of the primary

procedure to produce abortion within a reasonable time, incomplete abortion,

retained placenta, hemorrhage, infection, and embolic phenomena. With modern

protocols, these are rare. In a series of 1,002 women treated with mifepristone and

misoprostol at 13 to 21 weeks, 0.7% required a blood transfusion, 0.3% required

ergot treatment for hemorrhage, and one required laparotomy for otherwise

uncontrollable hemorrhage; 2.6% of patients received antibiotics for presumed

pelvic infection after hospital discharge and 7.9% complained of prolonged

bleeding (409). Uterine rupture has been reported in women with previous

cesarean delivery treated with misoprostol in the midtrimester. In a case series of

101 women with one or more previous cesarean births and three smaller case

series totaling 87 patients, no ruptures occurred (410). In another retrospective

study of women 14 to 26 weeks undergoing induction, there were 3 ruptures, all

in women who had 2 or more cesarean sections. These women had received 200

mcg vaginal misoprostol every 4 hours. The uterine rupture rate in these women

alone was 11.5% (411). The Society of Family Planning recommends

consideration of a 200 mcg or lower dose for women with a prior uterine scar, but

concludes that data are insufficient to advise a change in dosing interval (412).


Nhận xét