The incidence of venous thrombosis associated with estrogen treatment in male-to-female (M-->F) transsexuals is considerably higher with administration of oral ethinyl estradiol (EE) than with transdermal (td) 17-beta-estradiol (E(2)). To find an explanation for the different thrombotic risks of oral EE and td E(2) use, we compared the effects of treatment of M-->F transsexuals with cyproterone acetate (CPA) only, and with CPA in combination with td E(2), oral EE, or oral E(2) on a number of hemostatic variables [activated protein C (APC) resistance and plasma levels of protein S, protein C, and prothombin], all of which are documented risk factors for venous thrombosis. APC resistance was determined by quantification of the effect of APC on the amount of thrombin generated during tissue factor-initiated coagulation; plasma levels of total and free protein S were determined by standard ELISA; and levels of prothrombin and protein C were determined with functional assays after complete activation of the zymogens with specific snake venom proteases. CPA-only, td-E(2)+CPA, or oral-E(2)+CPA treatment produced rather small effects on hemostatic variables, whereas oral EE treatment resulted in a large increase in APC resistance from 1.2 +/- 0.8 to 4.1 +/- 1 (P < 0.001), a moderate increase in plasma protein C (9%; P = 0.012), and a large decrease in both total and free plasma protein S (30%; P < 0.005). The large differential effect of oral EE and oral E(2) indicates that the prothrombotic effect of EE is due to its molecular structure rather than to a first-pass liver effect (which they share). Moreover, these differences may explain why M-->F transsexuals treated with oral EE are exposed to a higher thrombotic risk than transsexuals treated with td E(2). Testosterone administration to female-to-male transsexuals had an antithrombotic effect.