We recruited 169 postmenopausal, white women at the tertiary gynecology center. Menopause was defined as amenorrhea for more than 12 months with an intact uterus. The study population was described previously in detail [22]. Briefly, the exclusion criteria were: previous thrombotic events, known malignancy, diabetes mellitus; hypo- or hyperthyroidism, adrenal insufficiency, and current smoking.

We collected basic demographic and clinical data, including concomitant diseases and medication used. Both physical examination and gynecologic assessment were performed at enrollment and after 6 months of therapy. Body mass index (BMI) was calculated by dividing weight in kilograms by square of height in meters. Arterial hypertension, obesity and hypercholesterolemia were defined as previously described [22]. Diabetes mellitus was defined in accordance to the American Diabetes Association Criteria [23]. Smoking was defined as the daily use of 1 or more cigarettes. Heavy menstrual bleeding was defined as previously described [24].

The Ethics Committee of Jagiellonian University Medical College approved the study (Approval no. KBET/347/B/2012) and participants provided informed consent in accordance with the Declaration of Helsinki.

In the interventional study we enrolled 150 women who were randomly assigned in an open-label manner (1:1:1) to one of three groups: 1) standard oral HT, containing 17β-estradiol (1 mg/day) and dydrogesterone (5 mg/day; Femoston conti, Abbott, Weesp, Netherlands), n = 50; 2) ultra-low-dose HT, receiving 17β-estradiol (0.5 mg/day) and dydrogesterone (2.5 mg/day; Femoston mini, Abbott, Olst, Netherlands), n = 50; and 3) control group, who did not use HT, n = 50. Compliance with therapy was assessed by the pill count. The intervention lasted 6 months.

Venous blood samples were drawn with minimal stasis using atraumatic venipuncture at 08.00–10.00 AM, after an overnight fast. All laboratory investigations were performed at study entry and after 6 months of therapy. Routine laboratory tests including blood cell count, lipid profile, fasting plasma glucose, creatinine, antithrombin activity, thyroid-stimulating hormone, follicle-stimulating hormone (FSH), prolactin (PRL), estradiol, and alanine aminotransferase (ALT) were measured by routine laboratory techniques. Plasma fibrinogen levels were measured using the von Clauss method. High-sensitivity C-reactive protein (CRP) was measured by nephelometry (Siemens, Marburg, Germany). Factor (F)VIII was measured by one-stage clotting assay using factor-deficient plasma (Siemens).

Oxidative modification on plasma proteins was assessed according to the approach by Becatti et al. as previously described [25]. Briefly, plasma (100 μl) was incubated with DNPH (400 μl), precipitated with trichloroacetic acid and washed with a 1:1 ratio of ethanol/ethyl acetate solution. The pellet was suspended in guanidine hydrochloride. The product of the reaction was analyzed at 370 nm. PC was expressed in nmol/mL of PC per 1 mg of protein. In healthy subjects, the reference range, established in our laboratory, was 0.54–2.03 nmol/mg protein.

We measured plasma PAI-1 antigen levels (Berichrom PAI, Siemens, Marburg, Germany) and activity (Abcam, Cambridge, UK). Plasma α2-antiplasmin and plasminogen activity were measured by chromogenic assays (STA Stachrom antiplasmin and STA Stachrom plasminogen, Diagnostica Stago, Asniéres, France). Plasma TAFI activity was measured by a chromogenic assay using the ACTICHROME® Plasma TAFI Activity Kit (American Diagnostica, Stamford, CT, USA). We also measured activated and inactivated TAFI antigen in plasma (TAFIa/TAFIai; Diagnostica Stago, Asnieres, France) and results were expressed as percentage of pooled plasma from healthy volunteers.

To determine fibrinolysis and thrombin generation, blood samples (vol/vol, 9:1 of 3.2% trisodium citrate) were centrifuged (2000 × g), within 30 min since the blood draw, for 10 min. The supernatant was aliquoted and stored (at –80 °C) until analysis [26]. Analysis was performed by experienced technicians blinded to the sample origin.

Calibrated automated thrombogram (CAT, Thrombinoscope, BV, Maastricht, Netherlands) was performed using commercial reagents as described previously [22]. Briefly, 80 μL of platelet-poor plasma were mixed with 20 μL of the PPP-reagent and 20 μL of FluCa. We measured the lag phase, the maximum concentration of thrombin formed (the thrombin peak), and the area under the curve represented by the endogenous thrombin potential (ETP).

Prothrombin fragments 1 + 2 (F1 + 2) levels were assessed in citrated plasma using an ELISA (Enzygnost F1 + 2 Micro; Siemens).

A modified approach introduced by Lisman et al. [27] was used to determine clot lysis time (CLT). Briefly, citrated plasma was mixed with CaCl2 (final concentration, 15 mmol/l), recombinant human tissue factor (final concentration of 0.6 pmol/l; Innovin, Siemens), phospholipid vesicles (final concentration, 12 μmol/l), and recombined tissue plasminogen activator (rtPA; final concentration, 60 ng/ml Boehringer Ingelheim, Ingelheim, Germany). The turbidity was measured at 405 nm. CLT was defined as the time from the midpoint of the clear-to-maximum-turbid transition, which represents clot formation, to the midpoint of the maximum-turbid-to-clear transition.

All measurements were determined in duplicate by technicians blinded to the origin of the samples. Intra-assay coefficients variation was 5% and inter-assay 7%.

Statistical analysis was performed with the STATISTICA 13.0 software (StatSoft, Poland).

Categorical variables were presented as numbers and percentages. Continuous variables were expressed as mean ± standard deviation or median and interquartile range (IQR), as appropriate. The Shapiro–Wilk test was used to assess conformity with a normal distribution whereas non-normally distributed data were analyzed by Kruskal–Wallis followed by Mann–Whitney’s comparison test. Categorical variables were analyzed using either the Chi2 test or Fisher’s exact test. Pearson’s correlation coefficient (Pearson’s r) or Spearman’s rank correlation coefficient were calculated to assess the linear correlations between variables with a normal or non-normal distribution, respectively. To assess the differences between pre- and post-treatment values of continuous variables the Wilcoxon signed-ranks test was applied. For nominal variables the McNemar’s test was used. Associations between the variables were expressed as odds ratio with 95% confidence intervals. Two-sided P-values < 0.05 were considered statistically significant.