Published: May 2014
Last reviewed: February 2016
Exogenous hormone therapy (HT) is used by millions of women yearly as hormone contraception (HC) or postmenopausal hormone therapy (PHT). Because these drugs are usually prescribed to healthy women, the issue of the risks of therapy has received considerable attention. This chapter reviews the risks of venous thromboembolism (VTE) associated with HT, including hormonal contraceptive preparations, postmenopausal HT, and selected estrogen receptor modulator (SERM) therapy or more aptly termed estrogen agonist, estrogen antagonists. This chapter on VTE covers pathophysiology, diagnosis, treatment, and outcomes.
Retrospective and prospective studies have shown a twofold to fourfold increase in the relative risk of VTE with the use of either HC, PHT, or SERM treatments. Prospective randomized controlled trials have shown an increase in VTE with the use of PHT and the SERM raloxifene.1 Yet clinical thrombotic events are rare in the general population, and overall they are more of a rare side effect than the main effect of exogenous hormone use and should not in general be a contraindication to initiating therapy in healthy women who have indications for either HC or PHT.
Whether to prescribe any HT depends on the net benefit for the woman after any additive risk factors are taken into account. Risk factors include differences in the thrombotic potential of various preparations, specifically oral versus transdermal, as well as a family and especially personal history of VTE. The presence of additive factors such as immobility, obesity, surgery, anesthesia, and varicose veins increase the risk of VTE.
The introduction of the hormonal contraceptive pill in 1959 was a major medical achievement that changed the way women could control reproduction and thereby achieve a greater life expectancy and better health status. HC remains the most popular form of reversible contraception worldwide.
The term hormonal contraception for many physicians is synonymous with the oral contraceptive pill (OCP). Millions of women rely on OCPs as their contraceptive method of choice. The U.S. Food and Drug Administration has reported that no other medication in the history of medical science has been studied more extensively than the OCP. Newer forms of HC include one transdermal patch (Ortho-Evra)2 one vaginal ring (NuvaRing), and one long-acting progestin-only 3-year implant (Nexplanon). The risks of VTE may be greater with transdermal HC (greater area under the curve for EE) and less with transdermal PHT (avoiding enterohepatic metabolism and hepatic coagulation factors). The risk is affected not only by estrogenic dosage but also by the type of progestin combined with the estrogen. HC have pharmacologic effects which include suppression of the hypothalamic-pituitary-ovarian axis, while PMT is approximately 4 to 5 lower in dose compared with HC and does not suppress the hypothalamic-pituitary-ovarian axis.
Most HCs are a combination of estrogen in the form of ethinyl estradiol and an androgen-derived progestin. In an effort to improve options and adherence to contraception, HC has expanded to include a weekly dermal patch (Ortho Evra), a hormonal vaginal contraceptive ring (NuvaRing) every 3 weeks, and less-androgenic oral progestins such as desogestrel and drospirenone. Also available are progestin-only HC agents: the progestin-only mini-pill including norethindrone acetate (Micronor, Nor-QD), injectable medroxyprogesterone (Depo-Provera either IM or SQ), a subdermal implantable norelgestromin device (Nexplanon), and a levonorgestrel-releasing intrauterine system (Mirena IUS) for 5 to 7 years of use. Relatively few thrombotic occurrences have been reported for progestin-only HC, and progestin-only therapy is generally presumed to have less risk than estrogen-progestin HC, just as estrogen-alone PHT has been shown to have less VTE risk than estrogen-progestin PHT.
The first case of thrombosis associated with HC occurred in 1961 when a nurse taking a high-dose estrogen OCP developed a pulmonary embolism. Myocardial infarction and stroke were reported in OCP users during the following years and were associated with older women who smoke and use HC. These early reports seemed to suggest that the thrombotic potential of the OCP was related to its relatively high estrogen content of 50 µg or more.
By the 1970s, the lower-dose OCP, containing 35 µg ethinyl estradiol, was introduced. In the 1990s, less-androgenic progestins (desogestrel, gestodene, and norgestimate) were developed, containing even lower formulations of ethinyl estradiol (20-25 µg ethinyl estradiol and even 10 µg in the Lo Loestrin Fe as well as the newer Natazia with actual estradiol valerate 2-3 mg and the progestin dienogest). The newer progestins have fewer androgenic effects than the traditional progestins (norethindrone acetate and levonorgestrel). Yet some of these newer progestins, such as desogestrel, have been found, in a number of European epidemiologic studies, to be associated with an increase in VTE compared with the older con-traceptive preparations despite the lower overall estrogen content.3 In the early 2000s, the newest progestin drospirenone (DRSP) containing HC were introduced. The anti-androgenic spironolactone analogue drospirenone (DRSP) is contained in the HC formulations of Yaz (20 mcg EE/3mg DRSP),Yasmin (30mcg EE/3mg DRSP), Beyaz (20 mcg EE/3mg DRSP with levomefolate) and Safyral (30 mcg EE/3mg DRSP with levomefolate). Questions have been raised as to risk of VTE with this popular class of HCl however, it appears in several trials the risk of VTE with DRSP containing HC is comparable to other HC and certainly less than the risk of VTE in pregnancy and the postpartum time. The greatest risk of VTE is present in women after immediately starting an HC or restarting (following a 1 month or greater hormone-free interval) with the same or a different HC. The baseline likelihood of developing a VTE in a non-pregnant, non-HC use is approximately 1-5/10,000 women years, while HC uses range from 3-9/10,000 women years while pregnancy is associated with VTE 5-20/10,000 women years and the 12-week postpartum range is 40-65/10,000 women years and is of course, higher in surgical C-section deliveries compared with vaginal deliveries.
Clinical trials designed to compare thrombogenic risks of HC have been observational and comparable. HC and the risk of VTE have mainly been investigated in case-control studies followed by a series of cohort analytic studies. The first case-control study, in 1967 by the Royal College of General Practitioners, reported that users of oral contraceptives had a threefold increased risk of venous thrombosis compared with nonusers. Most of these studies indicated that the risk was immediate, in the early phases of use, and did not further increase with a longer duration of use.
More recent studies have found a twofold to sixfold increased risk for venous thrombosis for OCP users. In the landmark Leiden Thrombophilia Study, the absolute risk for VTE was estimated at 0.8 per 10,000 per year among nonusers and 3.0 per 10,000 per year among HC users. Overall, these numbers seem to indicate a low absolute risk, translating into thousands of women needing to abstain from HC use to prevent one case of thrombosis a year.
Although HC may be associated with a significant portion of VTE events in young women, it does not account for all VTE events. Many other risk factors need to be considered in deciding whether some women have a higher risk of thrombosis when using HC, including obesity, age, a personal or family history of a VTE event, or the presence of a familial hypercoagulable state (including factor V Leiden, prothrombin gene mutation 20210A [PGM], or deficiencies of protein C, protein S, or antithrombin). With the familial thrombophilic states, including protein C, protein S, and antithrombin deficiencies, the risk of first thrombosis is significantly increased. In addition, reports of mild hyperhomocysteinemia (homocysteine >18.5 µmol/L) is a risk factor for VTE and up to 10% of the population has MTHFR mutation which can lead to hyperhomocysteinemia. Most importantly, a past history of a VTE event is the strongest indicator for recurrence in the future. Furthermore, importantly, pregnancy and the postpartum state itself increases the risk of VTE.
The factors and mechanism by which female hormones lead to a prothrombotic state are complex and not fully understood. Procoagulant factors include modest increases in the levels of factor VII, factor VIII, factor X, prothrombin, and fibrinogen, with associated decreases in the anticoagulant proteins including antithrombin and protein S. Using a thrombin-generation assay, it has been noted that women taking HC actually develop activated protein C resistance, which might provide an explanation for the increased thrombotic risk associated with HC users who are carriers of factor V Leiden mutation. It appears that the newer progestins, such as Desogestrel, might further activate protein C resistance.
Of all the inherited thrombophilias, the factor V Leiden mutation is believed to be the most common cause of congenital hypercoagulability states and may be responsible for about one half of all cases of familial venous thrombosis. Approximately 5% of all white women carry the factor V Leiden mutation. Women who are heterozygous and especially homozygous for the factor V Leiden mutation have a greater thrombotic risk when using any HC preparation. In women who carry the factor V Leiden mutation, the frequency of puerperal death from pulmonary embolism ranges from 1 in 2,000 women to 1 in 15,000 women. One of the great benefits of HC is the prevention of pregnancy and pregnancy-associated and post-partum morbidity and mortality.
Carriers of the factor V Leiden mutation had an eightfold increased risk of VTE in the landmark Leiden Thrombophilia Study. The study also suggested that the incidence of VTE was multiplied in a heterozygotic carrier of factor V Leiden who was also an HC user compared with a woman not carrying the mutation and not using HC. Higher risks have been reported for patients homozygous for the factor V Leiden mutation, with a marked in-crease in homozygous carriers using HC compared with non-carriers who were not users. Other investigations have confirmed the increased relative risk of VTE for users of HC and oral PHT who are also factor V Leiden carriers. The VTE risk for women with the factor V Leiden mutation is estimated to be increased twofold to fourfold. The adjusted risk estimate for a heterozygotic factor V Leiden carrier using HC is increased to 10-fold to 15-fold. VTE and other risk factors such as older age, obesity, surgery, immobility, prolonged travel preg-nancy and malignancy are also factors that influence this risk of VTE in mutation carriers However, simply having the factor mutation is not a reason for withholding HC from carriers of the mutation who have not themselves experienced a VTE.4
HCs have benefits beyond contraception. Limiting HC therapy can actually lead to an increase in the rate of unintended pregnancies, which themselves increase the rate of VTE compared with the nonpregnant state. HCs have an excellent risk-to-benefit balance compared with other contraceptive methods currently available. They are one of the most effective reversible means of preventing pregnancy. Discouraging the use of such an effective means of contraception could result in unwanted and unplanned pregnancies, leading to a greater risk of VTE or possibly contributing to the development of preeclampsia, abruptio placentae, fetal growth retardation, stillbirth, and miscarriage. Pregnancy itself is associated with a fivefold increase in VTE.
HC also has noncontraceptive benefits, including decreased dysfunctional uterine bleeding, reduction of dysmenorrhea, and treatment of endometriosis, acne, premenstrual dysphoric disorder, and functional ovarian cysts. HC also provides significant protection from ovarian and endometrial cancers. Carriers of the BRCA1 and BRCA2 mutations who use HC have a reduced incidence of ovarian cancer. The incidence and mortality associated with ovarian cancer (lifetime risk, 1 in 57) are greater than the attributable risk of developing VTE disease in a woman with factor V Leiden mutation and using HC (risk, 1 in 9,259 women). Denying these carriers HC might actually be harmful.
The primary role for PHT is relief of postmenopausal symptoms including vasomotor instability, genitourinary atrophy, quality-of-life issues, and bone protection. The chapter on menopause covers this in detail.5 Although the thrombogenicity of HC has been recognized since the 1960s, convincing evidence for the thrombogenic potential of PHT has emerged only since the 1990s and after the initial recruitment for the Women’s Health Initiative (WHI).
Most PHT preparations contain an estrogen and a progestin to avoid endometrial hyperplasia or stimulation. In women who have a uterus, progestin therapy is included with estrogen; estrogen-only therapy is used in women without a uterus. Common combined, continuous hormone preparations in North America include conjugated estrogens plus medroxyprogesterone (which were used in the WHI). Micronized estradiol is also available in a transdermal patch, a gel, vaginal tablets, transdermal lotion, spray and vaginal rings.
Some of the most recent evidence regarding the risk of VTE and PHT came from the terminated estrogen-progestin arm and estrogen-only arm of the WHI. Although the development of VTE was not one of the primary outcomes of this trial, multivariate analysis showed it to be the most significant risk. The use of PHT increased the risk of VTE twofold, which means PHT users have 34 VTE events annually per 10,000 women ver-sus 16 VTE events per 10,000 women who are not PHT users. The risk of VTE with PHT persisted throughout the 5-year study and was notably less in the estrogen-only arm (women with hysterectomy not needing a progestin).6
The ESTHER (EStrogen and THromboEmbolism Risk) study group, a French case-controlled study, reported that oral but not transdermal PHT estrogen was associated with increased risk of VTE in postmenopausal women.7 Their data suggest but do not confirm that transdermal estrogen may be safer than oral estrogen with respect to VTE risk. The original publication from the WHI indicated an increased risk for cardiovascular disease in PHT users. However, after central adjudication, these data did not show any significant increased risk for cardiovascular disease in PHT users. Obesity and oral HT in the context of inherited hypercoagulabilty in-creases risk of VTE while mutations carriers for factor V Leiden and PGM in the context of normal body mass index and transdermal HT do not appear to have any higher risk for VTE compared with mutation carrier controls not on any HT.
The estrogen formulations used in the postmenopausal state are used at a much lower dose than those used in HC and appear to have significantly lower biologic potency than HC. Although transdermal PHT is reported to be safer with respect to VTE, transdermal HC is reported to increase the risk of VTE at 4 to 8 per 10,000 women compared with 2 to 5 per 10,000 women using oral HC. This increase may be related to the higher total estrogen exposure. The SERM raloxifene (Evista), which is approved to prevent and treat postmenopausal osteoporosis, has been reported to increase the risk VTE twofold.
Early studies of PHT showed a slight increase in the risk of venous thrombosis, but subsequent studies did not repeat those findings. Only since the 1990s has a series of studies demonstrated that PHT users have a twofold to fourfold increased risk of venous thrombosis, and that PHT, regardless of duration of use, increases thrombosis risk. As with HC, studies have shown that the risk of VTE is highest during the first year of use.
The hemostatic effects of estrogens in oral PHT are similar to those in HC. All estrogens (oral or transdermal) seem to increase the levels of procoagulant factors VII, X, XII, and XIII and to decrease anticoagulant factors such as protein S and antithrombin, leading to a more procoagulant state not balanced by fibrinolytic activity.
The factor V Leiden mutation and prothrombin 20210A mutation are the two most common prothrombotic mutations. Women who are taking PHT and who have the factor V Leiden mutation have a 15-fold increase in the risk of VTE compared with women who do not have the mutation and do not use PHT, suggesting that with PHT, as with HC and factor V Leiden, the risk is more than additive. The Heart and Estrogen/Progestin Replacement Study and the Estrogen Replacement and Atherosclerosis trial, two randomized controlled trials that looked at PHT and secondary cardiovascular outcomes, found a 1.7% and 2.6% respective increase in VTE in women using PHT. Further investigation, through genotyping of blood samples, found that 16.7% of women with VTE carried the factor V Leiden mutation, versus 6.3% of controls.
Tamoxifen, raloxifene, and other SERMs have apparent antiestrogenic effects on the breast and varying effects on the endometrial tissue. Tamoxifen is used for breast cancer treatment and prevention, and raloxifene is used to prevent and treat osteoporosis. But there is an estrogenic effect on blood clotting by these agents, and in general they confer the same risks of VTE as does PHT. In a controlled trial of tamoxifen versus placebo, an increase was observed in the relative risk of deep venous thrombosis (relative risk, 1.6) in women with breast cancer. The MORE (Multiple Outcomes for Raloxifene Evaluation) trial has not reported any increased risk of arterial thrombotic events or early cardiovascular harm, but it has a reported a twofold or greater risk of VTE in the raloxifene-treated group.
Any woman with a personal or family history of VTE who is contemplating starting HC, PHT, or a SERM should be screened for a hereditary thrombophilia; however, screening of the general population without a personal or family history of VTE before starting HT is not recommended. The most challenging subset of women are those with known thrombophilia who have pressing indications for HC, PHT, or SERM therapy. Aromatase inhibitors (anastrozole, exemestane, letrozole) are favored in women with breast cancer who have a history of VTE because there has been no reported increased risk of VTE with these agents. There are other options for osteoporosis treatment, including bisphosphonates (risedronate, alendronate, and ibandronate and IV zoledronic acid) as well as RANKL inhibitor denosumab. These women are also candidates for nonhormonal contraception, such as barrier contraception or the copper T380A IUD, as well as the progestin-only HC.