Study design

This prospective study adhered to the principles of the Declaration of Helsinki and was conducted at the Institute of Cardiology, Jagiellonian University Medical College, in the Department of Coronary Artery Disease and Heart Failure and the Department of Interventional Cardiology at St. John Paul II Hospital in Kraków, Poland. The Jagiellonian University Bioethical Committee approved the final protocol (1072.6120.10.2022). Written informed consent was obtained from all patients before enrollment.

The study aimed to assess the impact of the mitral valve TEER on blood coagulation parameters, specifically thrombin generation, K_s, and clot lysis time (CLT). Blood samples were analyzed at 3 time points: visit 1 (V1) 1–2 days before the TEER procedure, V2 1–2 days after the procedure, and V3 6–8 weeks after the procedure. Flowchart of the study is presented in Figure 1.

Patients

A total of 31 patients were initially included in the study. The inclusion criteria comprised age of at least 18 years, symptomatic HF and established diagnosis of HF, severe MR, feasibility for TEER, and Heart Team qualification for the procedure. Exclusion criteria included the use of vitamin K antagonists, as well as heparin or bivalirudin administration less than 24 hours prior to blood sampling, active bleeding, known coagulation system pathologies, platelet count (PLT) below or equal to 100 × 10⁹/l, diagnosed significant liver (1.5 times above upper reference limit of alanine or aspartate transaminase) or renal failure (estimated glomerular filtration rate [eGFR] <⁠30 ml/min/1.72 m²), chronic treatment with nonsteroidal anti‑inflammatory drugs (excluding low‑dose aspirin), and systemic steroid therapy.

After obtaining informed consent, 2 patients were found to have contraindications (elevated inflammatory markers or signs of infection) for TEER. Subsequently, TEER was postponed for these individuals, leading to their exclusion from the study. For these patients, only blood samples collected before the scheduled procedure were preserved. Moreover, 2 patients died shortly after admission to the Intensive Cardiac Unit, and other 2 did not present at the follow‑up visit. Finally, out of the initially enrolled 31 patients, 27 completed V2 after the procedure, and 25 patients were evaluated over entire follow‑up (V3).

Type 2 diabetes was diagnosed according to the European Society of Cardiology guidelines on management of cardiovascular disease in patients with diabetes.¹⁹ Obesity was characterized in accordance with the criteria set forth by the World Health Organization (body mass index ≥30 kg/m²). Hypertension was delineated as systolic blood pressure equal to or above 140 mm Hg or diastolic blood pressure equal to or above 90 mm Hg, or the use of any antihypertensive medications.

Sample collection and routine laboratory tests

A fasting blood sample of 25 ml was collected from the antecubital vein and promptly preserved in tubes containing 3.2% trisodium citrate. Sample processing occurred within a 60‑minute window following blood collection. Serum levels of total cholesterol, low‑density lipoprotein cholesterol, high‑density lipoprotein cholesterol, triglycerides, glucose, and glycated hemoglobin were quantified using the Cobas 6000 biochemical analyzer (Roche, Mannheim, Germany). High‑sensitivity C‑reactive protein (hs‑CRP) was measured using latex nephelometry (Dade Behring, Marburg, Germany). N‑terminal pro–B‑type natriuretic peptide (NT‑proBNP) levels were assessed with an electrochemiluminescence immunoassay (Roche Diagnostics, Mannheim, Germany). Creatinine levels were determined using a routine laboratory technique, and eGFR was calculated based on the chronic kidney disease epidemiology collaboration formula. Complete blood count, encompassing red blood cell count (RBC), white blood cell count, hemoglobin (Hb), hematocrit, red blood cell and platelet distribution width, and PLT, was performed with a hematology analyzer (Sysmex XT2000i, Sysmex, Japan).

Laboratory investigation of coagulation parameters

Venous blood was collected from the antecubital vein between 7:00 AM and 9:00 AM using citrated tubes (9:1 ratio of 0.106 M sodium citrate; Monovette, Sarstedt, Nümbrecht, Germany). The collected blood was centrifuged at 2500 g and 20 °C for 20 minutes, yielding platelet‑poor plasma. The plasma was promptly snap‑frozen within 30 minutes of collection and stored in aliquots at –80 °C until further analysis. For individuals receiving direct oral anticoagulants (DOACs) from whom blood was collected within the preceding 24 hours, citrated plasma was treated with DOAC‑Stop (Haematex Research, Sydney, Australia) prior to assessment.²⁰

Fibrin permeation analysis

Fibrin clot permeation was determined using a pressure‑driven system as described previously.²⁰ Briefly, 20 mM calcium chloride and 1 U/ml human thrombin (Merck KGaA, Darmstadt, Germany) were added to 120 µl of citrated plasma. After 2 hours of incubation in a wet chamber, the tubes containing the clots were connected via plastic tubing to a buffer reservoir (0.01 M Tris, 0.1 M NaCl, pH 7.5), and its volume flowing through the gels was measured within 60 minutes. K_s, which indicates the pore size, was calculated from the following equation: K_s = Q × L × η/t × A × Δp, where Q is the flow rate in time (t); L is the length of a fibrin gel; η represents the liquid viscosity; A is the cross‑sectional area, and Δp stands for differential pressure.

Plasma clot lysis assay

The fibrinolysis capacity was assessed utilizing a CLT assay recommended by the International Society on Thrombosis and Haemostasis subcommittee, as detailed previously.²¹ In brief, citrated plasma was combined with 15 mM calcium chloride, 0.5 U/ml human thrombin (Merck), 15 µM phospholipid vesicles (Rossix, Mölndal, Sweden), and 18 ng/ml recombinant tissue plasminogen activator (Boehringer Ingelheim, Ingelheim am Rhein, Germany). Then, the mixture was transferred to a microtiter plate, and its turbidity was measured at 405 nm at 37 °C. The CLT was defined as the duration from the midpoint of the clear‑to‑maximum‑turbid transition, denoting the clot formation, to the midpoint of the maximum‑turbid‑to‑clear transition, signifying the clot lysis.

Endogenous thrombin potential

Thrombin generation kinetics was assessed using the calibrated automated thrombogram (Thrombinoscope BV, Maastricht, the Netherlands) following the manufacturer’s instructions. The measurements were conducted in a 96‑well plate fluorometer (Ascent Reader, Thermolabsystems OY, Helsinki, Finland) equipped with a 390/460 nm filter set at 37 °C. In short, 80 µl of platelet‑poor plasma were diluted with 20 µl of a reagent containing 5 pM recombinant tissue factor, 4 µM phosphatidylserine / phosphatidylcholine / phosphatidylethanolamine vesicles, and 20 µl of FluCa solution (HEPES, pH 7.35, 100 nM CaCl₂, 60 mg/ml bovine albumin, and 2.5 mM Z‑Gly‑Gly‑Arg‑7‑amino‑4‑methylcoumarin).

The peak thrombin represents the maximum concentration of thrombin formed during the recording time, while the area under the curve corresponds to the endogenous thrombin potential (ETP). Lag time delineates the initiation phase of coagulation, and time to peak signifies the propagation phase of thrombin generation. Each plasma sample underwent a duplicate analysis, and the intra‑assay variability was determined to be 8%.

Statistical analysis

Continuous variables were presented as mean and SD or median (interquartile range [IQR]) for normally and not normally distributed variables, respectively, while categorical variables were expressed as numbers and percentage. The normality of distributions was assessed using the Shapiro–Wilk test. All variables available at each timepoint (before TEER, 1–2 days after, and 50 days after TEER) were compared using the repeated measures analysis of variance (RM ANOVA) or the Friedman test with appropriate paired post hoc test, namely the Tukey honestly significant difference test or the Dunn test. For variables available at only 2 time points, the paired t test or the Wilcoxon signed‑rank test was used. The Greenhouse–Geisser correction was employed in the case of unequal variances in the RM ANOVA (for full records in the analyzed variables; n = 25).

Changes between the time points were calculated for each patient and presented in a plot. The Pearson or Spearman rank correlation coefficient was calculated for associations between continuous variables with normal or non‑normal distribution, respectively. Univariable linear regression analyses demonstrated relationships between clinical variables (eg, MR volume or hs‑CRP) and blood coagulation variables. Sample size was calculated for a 90% power of detecting 15% differences (increase in mean K_s from 3.5 × 10^–9 cm² to 4.025 × 10^–9 cm², with SD of 0.75 × 10^–9 cm²) in fibrin clot properties between time points (V1 and V3), at a 2‑sided level of significance of 0.05, consistent with prior research.^22-24 To achieve that level of statistical power, each group should have at least 23 patients. Statistical analysis was performed using GraphPad Prism software version 8.0.1 (San Diego, California, United States) and IBM SPSS package version 28.0 (Armonk, New York, United States).

Group characteristics

Baseline demographic and clinical characteristics of the analyzed group are shown in Table 1. The study recruited 31 patients, including 19 men and 12 women (mean [SD] age, 71.8 [8.32] years) with established HF, optimally treated, on maximally tolerated dose of guideline‑directed medical therapy, with severe MR, who were eligible for mitral TEER procedure. As suspected, most of the patients (n = 20; 64.5%) presented HFrEF, and the other (n = 8; 25.8%) had HF with preserved EF (HFpEF). Finally, the least numerous group included patients (n = 3; 9.7%) with HF with mildly reduced EF (HFmrEF). Follow‑up (V3) took place at a median (IQR) time of 50 (32.5–75.5) days after TEER. Periprocedural details, including the number and type of clips implanted, severity of MR, New York Heart Association class, and complications after TEER are available in Supplementary material, Table S1.

Laboratory evaluation

As shown in Table 1, patients shortly after TEER (1–2 days) had significantly decreased creatinine level (by 12.3%). As suspected for such procedures, most red blood cell morphologic parameters were reduced (RBC by 5.73%, Hb level by 6.1%, and hematocrit by 5.38%). Interestingly, we observed a significant drop in both analyzed platelet parameters, that is, PLT (11%) and plateletcrit (PCT) (14.29%). Moreover, hs‑CRP levels were substantially higher 1–2 days after TEER (4.8‑fold increase; Table 1). Group characteristics during follow‑up (n = 25) are available in Supplementary material, Table S2.

Echocardiographic assessment

A detailed description of baseline echocardiographic measurements can be found in Table 2. Briefly, we found that fibrinogen, ETP, CLT and K_s were significantly associated with baseline echocardiographic variables, such as left ventricular end‑diastolic diameter (LVEDd), left ventricular end‑systolic diameter (LVESd), left ventricular end‑diastolic volume (LVEDV) or EF (Supplementary material, Table S3).

Blood coagulation parameters

The only coagulation parameters that differed between the 3 time points were fibrinogen concentration and K_s (Table 3). In post hoc comparison, we found that fibrinogen level declined significantly between V2 and V3 (1–2 days after TEER vs the median of 50 days after TEER) (Figure 2). Similarly, we observed that K_s rose significantly between these 2 visits. The analysis of the other coagulation variables between V1 and V3 has not demonstrated any significant differences (Table 3 and Figure 2).

Moreover, associations emerged between baseline coagulation parameters and laboratory results, for example for hs‑CRP and fibrinogen (P = 0.01; r = 0.49), glucose and fibrinogen (P = 0.03; r = 0.42), NT‑proBNP and peak thrombin concentration (P = 0.04; r = –0.38), and ΔNT‑proBNP and CLT (P = 0.04; r = 0.33). Additionally, substantial associations were recognized between baseline platelet parameters, specifically PLT, PCT, and K_s, lag time, time to peak thrombin concentration, and fibrinogen (Figure 3). The intervariable correlation between PLT and PCT was very high (r = 0.95), leading us to present solely the results of a simple linear regression between PLT and the coagulation parameters. Other correlations between routine laboratory results and analyzed coagulation parameters are listed in Supplementary material, Table S4.

Changes in N‑terminal pro–B‑type natriuretic peptide

Given the impact of NT‑proBNP level on patient prognosis, we assessed its postprocedural changes with reference to the coagulation parameters. This allowed us to assess the impact of initial (V1) thrombin generation potential, fibrin clot permeability, and lysis on alterations in NT‑proBNP levels after TEER (V2 vs V1) and during follow‑up (V3 vs V1). We showed that only CLT correlated with the NT‑proBNP change between V2 and V1 (r = 0.4; P = 0.049; Figure 4). Finally, in the multivariable analysis, we demonstrated that baseline CLT increased the early post‑TEER NT‑proBNP levels (V2) adjusted for the initial NT‑proBNP level (R² = 0.55; estimate, 15.65; 95% CI, 2.86–28.44; P = 0.02).

Limitations

Our study has some limitations that need to be addressed. Firstly, despite the supported sample size calculation, the number of participants was restricted. Secondly, the HF group included all HF phenotypes (HFrEF, HFmrEF, and HFpEF), but the analyses primarily concentrated on MR within the entire group. Additionally, platelet activation parameters and hemolytic markers were not examined in the final analysis, limiting a more comprehensive exploration of the impact of platelet activation and shear stress on the observed effects.

Conclusions

We assessed for the first time fibrin clot characteristics and thrombin generation potential in patients undergoing the mitral TEER procedure. We demonstrated that fibrinogen levels decreased, and the permeation coefficient increased during a median 50‑day post‑TEER follow‑up, as compared with assessments shortly after the procedure. However, the analyzed blood coagulation parameters did not significantly differ from baseline vs the follow‑up assessment. Finally, CLT emerged as an independent predictor of the early decline in NT‑proBNP, underscoring its importance as a key indicator of the hemodynamic response to TEER.