Ovarian stimulation during IVF treatment results in a dramatic increase in endogenous estradiol levels of as much as 20–50 times baseline estradiol levels [6]. This supra-physiological increase in estradiol may lead to prothrombotic changes in hemostatic parameters.

The supra-physiological increases in baseline estradiol levels observed in these situations have been reported to be associated with increased levels of the procoagulant factors VIII, fibrinogen, and factor V, and von Willebrand factor (vWF). These prothrombotic changes are not balanced by corresponding increases in the naturally occurring anticoagulants antithrombin, protein C and S, which are reported to decrease following ovarian stimulation, with the net result a potentially prothrombotic state [5, 7–11]. Increasing resistance to activated protein C (APCr) appears to correlate with increasing estradiol levels, providing further support for development of a procoagulant phenotype in association with ovarian stimulation. The increased APCr does not appear to be explained by corresponding changes in the protein C/S system, suggesting another mechanism involved during IVF processes that induces acquired APC resistance during hyperstimulation [7, 11].

Fibrinolysis leads to dissolution of fibrin thrombus and enables vessel repair, and defects may predispose to venous and arterial thrombosis [12, 13]. Down-regulation of the fibrinolytic system has been observed as estradiol levels increase following ovarian stimulation. Levels of tissue plasminogen activator (t-PA), the main physiological activator of fibrinolysis, decreased with increasing estradiol levels, however, the major plasminogen activator inhibitor – 1 (PAI-1), the main physiological inhibitor of t-PA also decreased, with levels of both within normal ranges, so that the net clinical effect in terms of increased potential for thrombosis would be minimal. Rice et al. also showed down-regulation of fibrinolysis as estradiol levels increased, and also reported that thrombin-antithrombin complexes, a marker of coagulation activation leading to increased thrombin formation, remained unchanged. These observations suggest that elevated circulating estradiol does not predispose to thromboembolism as a result of hypofibrinolysis [14, 15]. Platelet function appears to remain unchanged [16].

Identification of specific thrombophilic defects does not necessarily aid in establishing overall potential thrombogenecity, and does not appear to have predictive value for recurrent thromboembolism [17–20]. More recently, attention has focused on markers of coagulation activation and global tests of haemostasis. Ovarian stimulation has been linked to increased levels of markers of in vivo coagulation activation: prothrombin fragment F1.2, which comes from in vivo cleavage of prothrombin by factor Xa, thrombin-antithrombin complexes, and D-dimer, a marker of breakdown of cross-linked fibrin/fibrinogen which rises following secondary activation of fibrinolysis after coagulation activation and fibrin formation.

In contrast to conventional tests of hemostasis which measure specific factor levels, assessment of ex vivo thrombin generation test using a calibrated automated thrombogram (CAT) system, can capture the end result of the interaction between proteases and their inhibitors and it is therefore potentially more useful and sensitive as a reflection of a haemorrhagic (low thrombin generation) or thrombotic (high thrombin generation) phenotype [21, 22]. Thrombin generation provides a global measure of thrombotic potential [23]. The thrombin generation test (TGT) provides information about the dynamics of ex vivo thrombin generation, with the thrombin generation curve described in terms of: the lag-time; the time to peak; peak thrombin concentration; and thrombin generation, with the area under the thrombin generation curve known as the endogenous thrombin potential (ETP). The ETP, a key parameter of the TGT, has been shown to have predictive value for the development of recurrent venous thromboembolism [18–20]. The ETP has been reported to be increased in women undergoing IVF [24]. This provides some support for a thrombotic phenotype associated with ovarian stimulation. Analysis of whole blood using thromboelastography has suggested changes towards hypercoagulability in women undergoing IVF treatment, although parameters remained within normal limits [25].

Despite the considerable changes in baseline concentrations of estradiol and progesterone, the clinical impact of the induced changes in haemostatic parameters during IVF treatment overall appears to be potentially limited as there is a generally a modest effect on the majority of the parameters detailed above [7, 11, 14, 26]. However, prothrombotic changes induced by a hyperestrogenic state may be potentially contributory to thromboembolism, particularly in the presence of a pre-existing prothrombotic state, and implantation failure. In women with OHSS, prothrombotic changes may be more pronounced. Excessive coagulation activation reflected by raised D-dimer levels and TAT thrombin-antithrombin complexes, or raised tissue factor and low tissue factor pathway inhibitor levels, seen in women with OHSS who do not fall pregnant even with higher oocyte yield suggests that prothrombotic mechanisms play a role in implantation failure [8, 27, 28].

OHSS is a recognized complication of assisted conception. Mild forms of OHSS are common, affecting up to 33 % of in IVF cycles and 3–8 % of IVF cycles are complicated by moderate or severe OHSS [29]. It usually develops after the administration of hCG and oocyte retrieval, and lasts for 10–14 days.

Women at high risk of developing OHSS include:

Women are classified as having mild, moderate, severe or critical OHSS based on the clinical severity at presentation. Those who develop thromboembolism lie in the critical category.

The drugs used for hormonal manipulation include follicle-stimulating hormone (FSH), human menopausal gonadotrophin (hMG), GnRH, GnRH agonist, clomiphene citrate, and hCG. Clomiphene citrate and GnRH are only rarely associated with OHSS [30].

The possibility that elevated estradiol levels may be linked to venous and arterial thrombosis stems from a number of studies on oral contraceptive pill use [31] and hormone replacement therapy [32].

Adverse pregnancy outcome in women with thrombophilia has led to speculation that these conditions may also play a role in subfertility, especially recurrent implantation failure. Proposed mechanisms include local microthrombosis at the site of implantation which impairs invasion of maternal vessels by syncytiotrophoblast and leads to implantation failure [63]. However, due to the low and varying prevalence of inherited thrombophilia, the small studies that have been conducted so far have been unable to confirm whether thrombophilia is contributory to subfertility [1].

An increased incidence of heterozygosity for the Factor V Leiden and G20210 prothrombin gene mutations was reported in women failing to conceive after three or more IVF–embryo transfer cycles [64]. In a larger study, 90 women who failed to fall pregnant after three embryo transfers were compared with two separate control groups containing women who conceived after their first attempt (n = 90) and another group containing women who conceived spontaneously (n = 100). The study group was found to have an increased incidence of homozygosity for the C677TT methylene tetrahydrofolate (MTHFR) polymorphism (and combined thrombophilias but not isolated Factor V Leiden) [65]. The C677T MTHFR polymorphism is a normal variant present in approximately 10 % of the normal population. It may, in the presence of folate deficiency, lead to hyperhomocysteinemia, which in turn may lead to thrombosis, although there is no direct association between this polymorphism and thrombosis. The phenotypic expression is silenced by oral folic acid administration. Folic acid 5 mg daily is usually advised throughout pregnancy and a dose of 400 μg considered long-term. Other studies have replicated the findings of an increased prevalence of heritable thrombophilia in women with recurrent implantation failure [66, 67] compared with those who conceived spontaneously or after their first cycle of IVF treatment. Proteins involved in fibrinolysis are necessary for trophoblast invasion into the endometrium. The precise clinical associations of polymorphisms of the gene for PAI-1, a major physiological inhibitor of fibrinolysis, with VTE and recurrent pregnancy loss remain unclear. The 4G allele is associated with higher levels of PAI-1, and might increase the risk for intravascular thrombosis. However, the contribution of this genetic variant to the risk for thrombosis, both arterial and venous, has not been established [67]. Limited data suggest that the 4G/5G polymorphism may be linked to a higher risk of implantation failure [68]. Heritable hypofibrinolysis, mediated by 4G/4G homozygosity for the PAI-1 gene, may be associated with late placenta-mediated pregnancy complications, postulated to be through thrombotic induction of placental insufficiency [69]. A meta-analysis of the PAI-1 4G/5G polymorphism showed no associations with two or three pregnancy losses [70].