Incidence of venous thromboembolism in transgender individuals on estrogen-based gender-affirming hormone therapy

Abstract Background: Historically, the use of estrogen-based gender-affirming hormone therapy (E-GAHT) has been associated with an increased risk of venous thromboembolism (VTE). However, the origin of higher VTE rates stems from data on cisgender women and limited studies on cohorts of transgender and gender diverse (TGD) individuals using more primitive estrogen or estrogen/ progesterone preparations, and at higher doses. Currently, data regarding VTE risk associated with newer E-GAHT regimens in TGD is scarce. Methods: Duke's DEDUCE platform was used to retrospectively identify patients aged 18 years who had received care within the Duke Health Care System between January 1996 to June 2025, had been prescribed estrogen-based hormone therapy, were not assigned female at birth, and were diagnosed with VTE using ICD-based diagnostic codes. A text search for transgender related terms was conducted using DEDUCE in the above cohort to confirm the indication for estrogen was GAHT. Chart review was utilized to determine whether the patients identified were taking E-GAHT when their VTE events occurred. We then compared demographic characteristics and comorbidities of the larger transfeminine cohort with the VTE cohort using Slicer Dicer and chart review. Results: We identified 1,143 adult patients, prescribed E-GAHT and not assigned female sex at birth to create our larger transfeminine cohort. Of these, 16 patients had developed a VTE while on E-GAHT (1.4%), eleven of which included pulmonary emboli with or without deep vein thrombosis (DVT) (68.75%); five had DVT alone (31.25%). With respect to risk factors, six of the 16 individuals (37.5%) had a major transient surgical risk factor prior to the VTE event, three individuals had significant non-surgical risk factors (18%), one had cancer (6%), and six had unprovoked VTE (37.5%). Of the 16 transfeminine individuals on E-GAHT with VTE, nine (56%) were taking oral estradiol, three (18%) were on transdermal formulations, three (18%) were on intramuscular preparations and one (6%) was taking an unknown formulation. However, among the six individuals with unprovoked VTE, two were taking oral 17-beta-estradiol, one was taking oral conjugated equine estrogen, two were taking intramuscular estrogen, one patient's route was unknown, and none had been using transdermal routes. Estradiol levels at the time of VTE diagnosis were unavailable. Mean duration from E-GAHT start to first VTE event was 7.8 years for all VTE (range 0.2-19 years) and 7.9 years for unprovoked VTE (range 1.5-12.5 years). Transfeminine individuals with VTE on E-GAHT appeared to be older (51 vs 30 years), have a higher body mass index (31 vs 28), and have a higher proportion of patients with hypertension (50% vs 9%), hyperlipidemia (44% vs 11%), and diabetes mellitus (25% vs 5%) when compared with the entire transfeminine cohort. Finally, while none of the 16 patients with VTE had an underlying thrombophilia diagnosis, nine individuals in the larger transfeminine cohort on E-GAHT had an underlying thrombophilia (four with Factor V Leiden, two with polycythemia vera, and three with essential thrombocytosis). Conclusion: The incidence of VTE in our institutional retrospective cohort was lower than previously reported historical rates of VTE in TGD individuals on E-GAHT. Among those with VTE, the majority had significant risk factors suggesting that the risk of VTE with E-GAHT may be lower than previously reported. Larger prospective studies are needed to further study E-GAHT formulation-specific and patient-specific risk factors for VTE in this population.