Transgender patients and the role of the coagulation clinician

2.1 VTE risks with MTF hormonal therapy

The mainstay treatment used to transition from male to female is hormonal therapy. Estrogen drives the appearance of female characteristics such as breast growth, reduction in facial and body hair and body mass, and redistribution of body fat. This treatment is often continued indefinitely in many patients including after gender-affirming surgeries. Forms of estrogen available include combination oral contraceptives (COC) and hormone replacement formulations for menopause (HRT). Ethinylestradiol is the most commonly used form of estrogen in COC, while for HRT estradiol is now most frequently used, available in oral, parenteral, and transdermal formulations (Table 2). Estrogen affects the synthesis of coagulation factors. Levels of fibrinogen, factor VIII, von Willebrand factor, factor VII, factor X, and prothrombin increase while the level of protein S decreases.6 Acquired resistance to activated protein C can also develop, partly due to the protein S decrease;7, 8 changes in levels of sex hormone-binding globulin may be responsible.9, 10 The sum of these changes results in a prothrombotic state and an increase in VTE risk.

The risks associated with COC in ciswomen, that is, women who match the gender identity with the sex that they were assigned at birth, are well known and include a four-fold to eightfold increased risk of VTE. The VTE risk is dependent on the estrogen dose, with higher doses associated with increased risk.9 The progestin agents desogestrel, drospirenone, and cyproterone acetate further increase the risk compared to levonorgestrel for a given estrogen dose.11 Many patient-specific factors such as obesity, smoking, and inherited thrombophilias can result in a synergistic increase in VTE risk for ciswomen using COC.12-14

Although the relative risk increase of VTE associated with HRT is lower than for COC, it is not negligible. The large randomized placebo-controlled trials of the Women's Health Initiative Studies found a significantly increased risk of VTE in ciswomen taking conjugated equine estrogen with medroxyprogesterone acetate [hazard ratio (HR) 1.87, confidence interval (CI) 1.37-2.54, p < .001) and for ciswomen post hysterectomy taking only conjugated equine estrogen (HR 1.48, CI 1.06-2.07, p = .02).15 The delivery method of estrogen for HRT makes a difference in ciswomen. Oral formulations are associated with an increased risk of VTE compared to transdermal administration, theoretically due to first-pass metabolism in the liver resulting in increased synthesis of coagulation proteins.16, 17 The transdermal preparations for contraception are different from those used for HRT, containing norelgestromin/ethinylestradiol, associated with an increased risk of VTE.18 The transdermal formulation for HRT contains estradiol and has little effect on coagulation factors.19

A nested case control study of roughly 500 000 women in the United Kingdom confirms that both the estrogen type and delivery method affect VTE risk.20 Of the 5795 women who had VTE, 85% had used an oral form of estrogen resulting in an overall adjusted odds ratio of 1.58 (95% CI 1.52-1.64) compared to no use. As in the Women's Health Initiative Studies, conjugated equine estrogen with medroxyprogesterone acetate had the highest associated risk with an odds ratio of 2.10 (1.92-2.31), with increasing risk as the dose of conjugated equine estrogen increased. For estrogen use alone, oral estradiol had a 15% lower risk than that associated with conjugated equine estrogen alone, adjusted odds ratio of 1.27 (1.16-1.39) versus 1.49 (1.39-1.60). Transdermal preparations of estrogen had no increased risk of VTE compared to matched controls (adjusted OR 9.6 (8.8-1.04)); results were similar for topical estrogen gels and creams.21

How the risks associated with COC or HRT use in ciswomen apply to transwomen are not clear. Small studies in transwomen also demonstrate a lower risk of VTE with estradiol compared to other forms. In one 1997 study, the risk of VTE with the use of transdermal estradiol was 50% lower than with oral ethinylestradiol.22 Another retrospective study of 165 patients with matched controls demonstrated an eightfold increase in risk of VTE in transwomen treated with conjugated equine estrogen compared to oral or transdermal estradiol.23 One retrospective chart review of transwomen using oral estradiol found the incidence of VTE to be 7.8 per 10 000 patient-years; patients were followed for a mean of 1.9 years with only one patient diagnosed with VTE.24 Although this incidence was low, with 1286 total patient-years of follow-up, the short duration of per-patient follow-up and the fact that data were based on chart review might not indicate the true thrombotic burden, as data described later suggest that the VTE risk associated with hormone therapy in transwomen continues to increase over time.

A large database analysis of U.S. patients in an integrated managed care consortium suggests that the risks associated with hormone therapy may be different when used by transwomen than by postmenopausal ciswomen. This large cohort study compared the incidence of VTE, stroke, and myocardial infarction in 2842 transwomen and 2118 transmen over a 4-year period and matched non-transgender cases at a 10:1 ratio with reference populations of 48 686 cismen and 48 775 ciswomen.5 Cases were matched for BMI, smoking history, hypertension, and cholesterol values. Researchers found that all transwomen had a higher incidence of VTE than both reference populations. This risk increased over time, with a further increase in risk in MTF participants who initiated estrogen. The adjusted HR for VTE with estrogen use for transwomen was 3.2 (1.5-6.5) when compared to matched cismen, and 2.5 (1.2-5.0) compared to ciswomen. This risk increased with longer duration of use, with a lower HR at <2 years [1.5 (0.5-5.1) men; 1.7 (0.5-5.5) women] compared to after 2 years of use [5.1 (2.1-12.6) men; 3.2 (1.3-7.6) women]. Transwomen had a higher incidence of VTE, with 2-year and 8-year risk differences of 4.1 (95% CI, 1.6-6.7) and 16.7 (CI, 6.4-27.5) per 1000 person-years relative to the reference population of cismen and 3.4 (CI, 1.1-5.6) and 13.7 (CI, 4.1-22.7) relative to ciswomen. The increasing risk associated with estrogen use over time in MTF subjects differs from that found in postmenopausal women on HRT, in whom the risk is highest in the first year of use, then decreases over time.17, 25, 26

Transwomen using estrogens also had an increase in risk of ischemic stroke in this study that was most pronounced after 6 years of use and continued to increase over time, similar to VTE risk. Before 6 years of estrogen use the HR for ischemic stroke was 1.3 (0.6-2.9) compared to the cismen reference population and 2.3 (1.0-5.4) compared to ciswomen; after 6 years of use, the HR increased to 9.9 (3.0-33.1) compared to cismen and 4.1 (1.5-11.4) to ciswomen. The incidence of myocardial infarction was similar in transgender patients compared with matched cismen and ciswomen. Interestingly the myocardial infarction risk for transwomen was higher than the matched ciswomen cohort but was similar to that of the reference cismen cohort, suggesting that estrogen did not influence the risk of myocardial infarction. Although estradiol was the most frequently used hormone type, the authors were unable to determine types and doses of estrogen used. Importantly, the use of estrogen obtained from sources outside the database could not be determined; self-medication with estrogen could have affected the increased overall VTE rate.

Estrogen treatment alone is not sufficient in transwomen to suppress testosterone production into the normal range found in ciswomen. Agents used to suppress testosterone include spironolactone; cyproterone acetate (CPA), which has progesteronelike effects and antiandrogen activity; and gonadotropin-releasing hormone agonists. Overall, these three different types of agents have similar effects on lowering testosterone levels.27, 28 Cyproterone acetate has been shown to have equivalent associated risk for VTE to other higher-VTE-risk progestins when used with ethinylestradiol.29 The risk of VTE associated with CPA monotherapy or in combination with low-dose estradiol for MTF treatment is not as well described. In men with advanced prostate cancer, CPA monotherapy increases the risk of VTE compared with orchiectomy or luteinizing hormone-releasing hormone antagonists.30 Recent data suggest that low-dose CPA (25 mg daily) may be just as effective as higher doses (50-100 mg) at maintaining suppression of testosterone levels in transwomen; impact on thrombotic events and other adverse effects were not studied.31 Cyproterone acetate is not available in the United States or Japan. Subtle differences in spironolactone and CPA effects on lipids and prolactin levels may result in the endocrinologist's choosing one over the other for the individual patient.32, 33 Progestin agents when used alone at higher doses than for contraception, such as for dysfunctional uterine bleeding, are also associated with increased VTE risk34, 35 but are not frequently used in the transgender population.

Although the relative risk of VTE associated with hormone use is increased in transwomen compared to cismen and ciswomen, the absolute risk is low. The estimated annual incidence rates for VTE in people of European descent range from 104 to 183 per 100 000 patient-years.36 A meta-analysis estimated the crude incidence rate of VTE associated with estrogen use in transwomen to be 2.3 per 1000 patient-years; however, the heterogeneity was statistically significant.37

Most transwomen today are treated with estradiol, either oral or transdermal formulation, in association with spironolactone in the United States and Japan, as CPA is not available, or with CPA in countries where it is available. The lowest doses of estrogen that maintain serum estradiol levels in the usual physiologic female range and of CPA to suppress and maintain testosterone levels in the female range should be used.3, 4

2.2 Thrombophilia testing in MTF

The referring endocrinologist asked whether the patient should be tested for inherited thrombophilia given the family history.Discussion with the patient indicated that the brother developed VTE after major trauma. No other family history of thrombosis was known. As in the general population, the utility for screening for thrombophilia before starting hormonal therapy is low and not cost-effective. The standard approach to investigating for thrombophilia should be used, recognizing that if testing is performed results may have limited to no impact on management decisions.38, 39 Just as thrombophilia testing should not be routinely performed in women before initiating estrogen therapy, testing should not be routinely performed in transgender patients considering hormone therapy.40 This patient had used COC for >16 years without thrombotic events, making inherited thrombophilia less likely. The brother had a VTE event provoked by a major risk factor. Neither warranted thrombophilia testing. Only one retrospective study of thrombophilia testing is published in this population. Partial thrombophilia screening in 162 transwomen and 89 transmen found a low incidence in both groups, with no patient developing VTE during the 3 to 5 years of analysis despite hormone use.41

2.3 Summary Case I

This 44-year-old transwoman had used COC doses of estrogen and progestin agents for 16 years without thrombotic events, suggesting the absence of a clinically relevant inherited thrombophilia. The younger brother had a provoked VTE due to a major risk factor. No thrombophilia testing was performed. The lower risk of VTE with oral or transdermal estradiol compared to what she had previously used was discussed. The patient felt that the risk of VTE associated with using estradiol was acceptable. She was informed that the risk of VTE may increase after 2 years of use, on the basis of limited data.5 She was advised to stop smoking. Age-associated VTE risk was also reviewed,42 and signs and symptoms of VTE were discussed. She is now taking transdermal estradiol 75 micrograms applied twice a week and spironolactone 50 mg daily. See Table 3.

3.1 MTF hormone use with a history of VTE

The approach to hormonal treatment for transwomen who have had a VTE is not different from that for ciswomen with a history of VTE. Considerations for hormone therapy and VTE risks should be made in the context of the type of estrogen used and other known VTE risk factors such as obesity and smoking. Oral and transdermal estradiol preparations have lower associated risks for VTE than ethinylestradiol used in COC. Although the standard approach is not to use COC or HRT in women who have had a VTE event, many clinicians routinely continue hormones, both COC and HRT, in women who have had VTE as long as full-intensity anticoagulation is continued. In a post hoc analysis of the Einstein studies (full-dose rivaroxaban compared to vitamin K antagonist (VKA) for treatment of VTE) the risk of recurrent VTE was not different for women with (3.7%/year) and without (4.7%/year) use of hormonal contraceptives during anticoagulant treatment (adjusted HR 0.6; 95% CI 0.2-1.4).43 No specific data for transgender patients are available. Although for our present patient with a segmental pulmonary embolism, the risk of recurrent VTE is concerning, the patient's strong desire to continue with hormone treatment should not be dismissed.

The risk of major bleeding with full intensity VKA anticoagulation has been historically quoted at 1% to 3% per year. With direct oral anticoagulants, the associated risks may be lower in cases selected for low risk of bleeding. Data support a 39% relative reduction in bleeding if a direct oral anticoagulant is used to treat pulmonary embolism rather than previously used low-molecular-weight heparin/vitamin K antagonists44 with a case fatality rate of 0%.45 Data also support the efficacy of extended duration anticoagulation using reduced-dose direct oral anticoagulant, with even lower bleeding risk,46, 47 although this strategy has not been tested with concomitant hormone therapy.

3.2 Summary Case II

Given the indication to continue hormonal therapy, anticoagulant treatment was also continued. Given the patient's strong desire to continue high-risk contact sports, we chose to switch to edoxaban 60 mg od, to be taken at bedtime. Six months after the pulmonary embolism, she had no residual symptoms of shortness of breath with exertion and started to train again. We discussed skipping one dose of edoxaban on the night before a game48 or downhill skiing and accepting a slightly increased risk of recurrent VTE but maintaining her lifestyle, which was crucially important for her mental well-being.48 See Table 4.

4.1 VTE and cardiovascular risks with hormone therapy for FTM

The development of secondary male sex characteristics to transition from female to male relies on the use of testosterone. Many formulations are available including gels, patches, and intramuscular or subcutaneously administered parenteral forms. The route of administration is usually determined by patient preference as all have been demonstrated to have clinical efficacy.3 A long-acting parenteral form is associated with adverse events such as anaphylaxis and pulmonary oil microemboli and is therefore infrequently used.49 The use of testosterone results in increased muscle mass and decreased fat mass, acne, deepening of the voice, cessation of menses, and an increase in body hair, particularly on the face, chest, and abdomen.50 As in men with hypogonadism requiring testosterone supplementation, adverse events resulting from excess testosterone therapy can be seen in transmen, including erythrocytosis, sleep apnea, hypertension, excessive weight gain, salt retention, and lipid changes including hypertriglyceridemia and decreased high-density lipoprotein.51

The risk of VTE for transmen with the use of testosterone appears to be low. Many of the concerns associated with testosterone use are based on past reports of increased events in cismen taking testosterone. In cismen, a recent retrospective study of 694 hypogonadal men treated with testosterone found that only 9 developed VTE with a mean follow-up of 25 months (+/- 21 months). Six of the 9 men had other clear provoking VTE risks including lower extremity surgery or trauma, abdominal surgery, or prolonged air travel.52 In the large U.S. database study of transgender patients previously discussed, the hazard ratio estimates for VTE in all patients in the transmen cohort using testosterone compared with cismen and ciswomen were low at 1.6 (CI, 0.9-2.9) and 1.1 (CI, 0.6-2.1), respectively.5 Similar results were found in another health claims database case-control study of men with VTE and testosterone use within 15 days of VTE diagnosis during the period from 2007 and 2012,53 with adjusted OR of 0.9 (95% CI, 0.73-1.12).53 The route of administration of testosterone or preceding use of up to 60 days did not change results. Although the data are not robust, cardiovascular events with testosterone use in transmen also appear to be low. Results from many retrospective studies are confounded by inability to determine smoking status or other risk factors in studies before 2000.54 Data for testosterone replacement in hypogonadal men indicate no discernable increase in cardiovascular events when patients are supplemented to physiologic levels. Although a randomized placebo-controlled trial of testosterone replacement in hypogonadal cismen above the age of 65 was halted early because of increased cardiovascular events of all types in the treated men, there was no statistically significant difference in serious or life-threatening adverse events.55 A subsequent larger randomized trial had such low numbers of cardiovascular events in each arm that rates could not be determined.56 In the U.S. database study of transmen, results for ischemic stroke with versus without testosterone use were 1.1 (CI, 0.6-2.0) and 1.3 (CI, 0.7-2.5), and 0.7 (CI, 0.3-1.8) and 1.3 (CI, 0.5-3.9) for myocardial infarction.5 In transmen initiating testosterone the number of events was too low to determine incidence and HR for these outcomes.

4.2 Erythrocytosis

Erythrocytosis is expected with the use of testosterone in transmen but a possible increase in thromboembolic risk is concerning. In a prospective study of transmen initiating hormone therapy, the hematocrit increased over the first 12 months of testosterone use, with the largest increase in the first 3 months.57 Only 11.5% of transmen were found to have an increased hematocrit >50%, with a maximal hematocrit of 54% seen in 1 of 192 testosterone-treated patients. Increased BMI was associated with a hematocrit >50%; no differences in smoking or other risk factors were seen. No thromboembolic events were observed in the transmen cohort in this study, with follow-up of 12 to 36 months. In a retrospective study of hypogonadal cismen treated with testosterone who developed VTE, all had hematocrits <52%.52 The use of testosterone in transmen will increase the hematocrit to levels seen in cismen but does not appear to routinely exceed the normal range.

4.3 Summary Case III

The patient was reassured that his risk for VTE associated with testosterone use appeared to be minimal, was not impacted by his family history, and did not appear to be greater than that of others his age. He was counseled to avoid smoking and to follow up routinely with his endocrinologist to make sure testosterone remained in the low physiologic range, as well as for monitoring of blood pressure, complete blood count including hematocrit, lipid profile, and other assessments. Signs and symptoms of VTE and the lack of data for the long-term effects of prolonged use of testosterone in this setting were also discussed. See Table 5.