Estrogen-associated thrombosis and the road not taken

In this issue of Blood Advances, Verstraete et al¹ provide a detailed analysis of the long-term outcomes after estrogen-associated venous thromboembolism (VTE), demonstrating overall low recurrence rates. In addition to providing epidemiologic clarity, the findings raise important concerns regarding the trade-off between thrombosis and bleeding risk, as well as the utility of thrombophilia testing when deciding between the mutually exclusive paths of time-limited vs indefinite anticoagulation.

The authors capitalize on 14 years of data for patients without cancer aged <50 years from the multinational RIETE registry to characterize the initial thrombotic event and its context, as well as to compare recurrence rates among women with estrogen-associated VTE with those for men and women with non–estrogen-associated VTE. Estrogen-associated VTE was defined by exogenous estrogen use and did not include patients in the high estrogen state of pregnancy. Although the study did not limit estrogen use to combined oral contraception, this medication class reflected 97% of patients. For patients off anticoagulation, estrogen-associated VTE had a statistically significantly lower rate of recurrence at 2.07 per 100 patient-years compared with 8.70 per 100 patient-years for men and 5.05 per 100 patient-years for other women. The low rates of recurrence when off both estrogen and anticoagulation is consistent with prior research.^2,3 Most (74%) of the estrogen-associated VTEs occurred without concurrent transient risk factors for VTE.

Major and nonmajor bleeding while on anticoagulation was higher in women than in men, regardless of the etiology of VTE, which was predominantly driven by uterine bleeding. In fact, for estrogen-associated VTE, the rates of bleeding while on anticoagulation was >4 times higher than the rates of VTE recurrence off anticoagulation: 8.70 vs 2.07 per 100 patient-years. Although the duration and mortality risk from these 2 outcomes are not equal, the impact on bleeding must be weighed when considering indefinite anticoagulation, particularly in light of the low recurrence risk.

The decision to discontinue estrogen at the time of VTE diagnosis in 92% of patients likely contributed to the rates of uterine bleeding. Although eventual discontinuation of estrogen is needed to reduce the risk of recurrence, the necessity of immediate discontinuation is not evidence based: recurrence while on anticoagulation is not higher for patients remaining on estrogen.⁴ Given the ability of anticoagulation to overcome thrombosis risk in other more prothrombotic scenarios, such as pregnancy and malignancy, this fact is intuitive, which begs the question: why are clinicians in a rush? Is a lack of appreciation of the rates of and impact on uterine bleeding contributing? One hopes that the bleeding rates highlighted by Verstraete et al will help shift this practice. Women should be offered the option to remain on estrogen for as long as they intend to remain on anticoagulation.

Knowing that VTE recurrence rates are low but the consequences of recurrence can be morbid or lethal, the ability to predict recurrence remains a vital aspiration, although one that has proved remarkably difficult to achieve in practice. Thrombophilia testing has long held the allure of fulling this goal. In Verstraete et al, the authors highlight not only that thrombophilia testing was conducted more frequently in women with estrogen-associated VTE than in men or other women, 56% vs 36% and 39%, respectively, but also that testing was ultimately negative more frequently. The result of this combination is that that all 3 groups had comparable overall rates of thrombophilia identified. Notably, thrombophilia status was not a significant predictor of VTE recurrence in their multivariate model for estrogen-associated VTE. This is finding is consistent with other work that has failed to find a difference in recurrence rates based on thrombophilia status.5, 6, 7 Although these studies were not limited to estrogen-associated VTE, one also using data from the RIETE registry did specifically identify the lack of correlation with recurrence in this subpopulation.⁷

The American Society of Hematology thrombophilia testing guidelines released in 2023 have raised some controversy over the suggestion to test for inherited thrombophilias and then extend anticoagulation indefinitely for patients who test positive after estrogen-associated VTE.⁸ One must acknowledge that the data underlying the cited studies were collected before these guidelines and testing rates may be appreciably higher now. However, it is unclear that more universal testing would alter the associations found because this would require the decision to test to be correlated with the impact of thrombophilia on recurrence. With low overall recurrence rates, high bleeding rates on anticoagulation, and a lack of strong data to support the utility of inherited thrombophilia in predicting recurrence, this study by Verstraete et al may indicate it is time to revisit whether indefinite anticoagulation is appropriate to consider after most estrogen-associated VTEs. Perhaps, the tolerance of anticoagulation without bleeding, the patients’ need or desire to stay on estrogen, or factors that better predict recurrence such as family history should be centered in that decision rather than inherited thrombophilia status.

After VTE, conversations at the bedside often revolve around the negative quality of life impact of anticoagulation and the fear of recurrence. When considering indefinite anticoagulation, one is reminded of Robert Frost’s writing on how challenging it can be to decide between 2 valid paths, particularly when the consequences may only become clear in hindsight.⁹ Verstraete et al take us a step closer to having those conversations with care. We can hope that taking the time to weigh the risks and benefits of both ongoing anticoagulation and estrogen discontinuation may in turn “make all the difference.”

Conflict-of-interest disclosure: The author declares no competing financial interests.