Type of hormone replacement therapy and risk of venous thromboembolism

The results of randomised studies have been fairly consistent in finding an increased risk of venous thromboembolism with oral hormone replacement therapy. The accompanying systematic review by Canonico and colleagues finds similar results.1 The review combined data from both randomised and observational studies and found that the risk of venous thromboembolism doubled with oral hormone replacement therapy compared with placebo (pooled odds ratio 2.4, 95% confidence interval 1.9 to 3.0).

Results from the observational studies alone were then pooled to assess the risk of venous thromboembolism from different types of hormone replacement therapy and duration of use. No significant difference was seen between combined therapy and oestrogen only (2.6 v 2.2; P=0.45). Risk of venous thromboembolism was significantly higher during the first year of treatment with oral oestrogen (4.0 for a duration of less than one year; 2.1 for more than one year), and past treatment was not associated with an increased risk (1.2, 0.9 to 1.7). The results echo those from the women’s health initiative study.2 3

Canonico and colleagues’ review is timely and follows the recent publication of the final analysis of the Esther (estrogen and thromboembolism risk) study.4 This case-control study, by some of the same authors of this week’s review, added a further 67 cases of venous thromboembolism with transdermal use to the 15 that have been reported in three previous studies.

Because no randomised data exist on the risk of venous thromboembolism with transdermal oestrogen, the relation between venous thromboembolism and the route of hormone replacement therapy was derived from these four observational studies. Transdermal oestrogen did not increase venous thromboembolism (pooled odds ratio 1.2, 0.9 to 1.7)—the upper 95% confidence interval for transdermal use was lower than the lowest 95% confidence interval for oral use. Risk of venous thromboembolism was not higher with transdermal oestrogen for women at higher risk of thrombosis who had either raised body mass index or prothrombotic mutations.

Oral oestrogens are associated with prothrombotic changes in factors involved in coagulation and fibrinolysis.5 Canonico and colleagues point out that whereas transdermal oestrogen has no effect, oral oestrogen increases plasma prothrombin fragment 1+2, lowers antithrombin concentrations, and causes an acquired resistance to activated protein C. They conclude that transdermal oestrogen seems to have little or no effect on haemostasis.1

However, the effect of the route of administration may be influenced by the type of oestrogen. The contraceptive patch, which was designed to deliver a relatively low dose of ethinyl estradiol (20 μg daily), was unexpectedly found to produce 60% higher concentrations in the serum than an oral 30 μg pill.6 A 450% increase in sex hormone binding globulin, a marker of high oestrogen exposure, has also been shown with use of the contraceptive patch.7 In addition, two of the three postmarketing studies, comparing the patch with a 30 μg or 35 μg oral contraceptive, showed a doubling of risk of venous thromboembolism with transdermal delivery.8 9

Do different types of hormone influence the risk of venous thromboembolism? Conjugated oestrogens and estradiol were used in both the observational and the randomised trials that showed increased risk. Estradiol was also the oestrogen used in the randomised study that found increased recurrence in women with previous venous thromboembolism.10 One observational study found no increase with esterified oestrogens.5 Whether the risk of venous thromboembolism for combined hormone replacement therapy is influenced by the type of progestogen, as in the oral contraceptive pill, also needs further investigation. The Esther study suggested that non-pregnane derivatives were associated with an increased risk whereas micronised progesterone and pregnane derivatives were not.4 Interestingly, the pregnane derivatives included medroxyprogesterone acetate, the progestogen in the combined arm of the women’s health initiative study.

Do we have any data on the effect of transdermal oestrogen on other outcomes? The Papworth study, which randomised women with angiographically confirmed ischaemic heart disease to transdermal therapy or placebo, found no significant difference in rates of cardiac events in the transdermal group.11 The million women observational study showed similar increased risks for transdermal and oral oestrogens with respect to breast cancer.12

What about the side effect profile of transdermal oestrogens compared with oral ones? The answer to this has been hampered by a lack of trials and by incomplete and non-standardised reporting.13 Transdermal hormone replacement therapy also costs more.

As Canonico and colleagues conclude, we need further investigation into the association between venous thromboembolism and transdermal oestrogen. In the meantime, we can advise healthy menopausal women aged 50-59 that the risk of venous thromboembolism with oral preparations is about 11 additional cases per 10 000 women per year for combined therapy and two additional cases per 10 000 women per year for oestrogen only.14 Because a dose response seems to exist, these absolute risks may be lower with lower doses of hormones.5 Women with previous venous thromboembolism or a mutation affecting prothrombin should be offered alternatives to oestrogen.