Clinical Question: Do different formulations of estrogen similarly increase the risk of venous thrombosis?
Setting: Outpatient (any)
Study Design: Case-control
Synopsis: Investigators identified 586 perimenopausal and postmenopausal women who experienced a first venous thrombosis: 426 women with venous thrombosis alone (73 percent), 68 with venous thrombosis and pulmonary embolism (12 percent), and 92 with pulmonary embolism alone (16 percent). The women attended a large health maintenance organization in western Washington. Control patients (n = 2,268) were selected randomly from the same health plan and were matched for age, sex, treated hypertension status, and occurrence of myocardial infarction.
Information on the use of estrogen and other medications was obtained from out-patient pharmacy records. The difference in choice of estrogen was based on a pharmacy benefits program, not a decision by physician or patient. Trained medical record abstractors reviewed charts to verify all relevant clinical diagnoses.
Oral estrogen use was classified into three subgroups: (1) conjugated equine estrogen; (2) esterified estrogen; and (3) other estrogens, including micronized estradiol, which accounted for less than 1 percent of all estrogen prescriptions. Medroxyprogesterone acetate was prescribed almost exclusively as the progestin. Daily estrogen dosage and duration were calculated using computerized pharmacy data. Pertinent demographic and health status information was obtained by review of ambulatory medical records and cancer registry data.
Patients with venous thrombosis were more likely than control patients to have established risk factors for venous thrombosis, including a history of cancer, heart failure, recent hospitalization, or major fracture. A similar percentage of both groups were current users of oral estrogen with or without concomitant progestin use. However, compared with women not currently using hormones, and after adjustment for confounding factors, current users of esterified estrogen had no significant increase in venous thrombosis risk (odds ratio [OR] = 0.92; 95 percent confidence interval [CI], 0.69 to 1.22), whereas current users of conjugated equine estrogen had an elevated risk (OR = 1.65; 95 percent CI, 1.24 to 2.19). Furthermore, current use of conjugated equine estrogen was associated with an increased risk of venous thrombosis compared with current use of esterified estrogen (OR = 1.78; 95 percent CI, 1.11 to 2.84).
When analyzed for either formulation of estrogen therapy opposed and unopposed with progestin, only current users of conjugated equine estrogen with progestin had an increased risk of venous thrombosis (OR = 2.17; 95 percent CI, 1.49 to 3.14). With regard to dosage and time interval of hormone initiation, a positive dose-response relationship existed between the conjugated equine estrogen dosage and venous thrombosis risk. No association was detected between the time interval of starting either form of estrogen and venous thrombosis risk.
Bottom Line: Oral therapy with conjugated equine estrogen is associated with an increased risk of venous thrombosis in a dose-dependent fashion during the time the woman takes it. Concomitant use of medroxyprogesterone further increases the venous thrombosis risk. Esterified estrogen does not appear to increase venous thrombosis risk. Until more reliable data are available from prospective comparison trials, it makes sense to strongly consider prescribing esterified estrogen to control symptoms in perimenopausal and postmenopausal women. (Level of Evidence: 3b)