PAI-1 and t-PA/PAI-1 complex potential markers of fibrinolytic bleeding after cardiac surgery employing cardiopulmonary bypass

Fibrinolysis as a Causative Mechanism for Bleeding Complications on Extracorporeal Membrane Oxygenation: A Pilot Observational Prospective Study

BACKGROUND

Extracorporeal membrane oxygenation (ECMO) is associated with a high risk of bleeding complications. The specific impact of ECMO on fibrinolysis remains unexplored. The objective of the current pilot observational prospective study was to investigate the longitudinal dynamics of fibrinolytic markers-i.e., changes over time-in the context of bleeding events in patients on ECMO.

METHODS

Longitudinal dynamics of contact phase components (kininogen and bradykinin) and fibrinolysis markers (tissue plasminogen activator [tPA], plasminogen activator inhibitor-1 [PAI-1], their complexes [tPA•PAI-1], plasmin-antiplasmin complexes, plasminogen, and D-dimer) were measured in patients undergoing venovenous and venoarterial ECMO, before implantation, at 0, 6, and 12 h after implantation, and daily thereafter.

RESULTS

The cohort consisted of 30 patients (214 ECMO days). The concentrations of tPA, D-dimer, plasmin-antiplasmin complexes, PAI-1, and tPA•PAI-1 complexes were increased, whereas plasminogen decreased compared to normal values. A noteworthy divergence was observed between hemorrhagic and nonhemorrhagic patients: in bleeding patients, D-dimer, plasmin-antiplasmin, tPA, PAI-1, and tPA•PAI-1 followed an increasing kinetics before hemorrhage and then decreased to their baseline level; conversely, nonbleeding patients showed a decreasing kinetics in these markers. Also, D-dimer and tPA followed an increasing kinetics in bleeding patients compared to nonbleeding patients (median values for D-dimer dynamics: 1,080 vs. -440 ng/ml, P = 0.05; tPA dynamics: 0.130 vs. 0.100 nM, P = 0.038), and both markers significantly increased the day before hemorrhage. A tPA concentration above 0.304 nM was associated with bleeding events (odds ratio, 4.92; 95% CI, 1.01 to 24.08; P = 0.049).

CONCLUSIONS

Contact activation induces fibrinolysis in ECMO patients, especially in patients experiencing bleeding. This finding supports the role of this mechanism as a possible causal factor for hemorrhages during ECMO and open new avenues for novel therapeutic perspectives.

EDITOR’S PERSPECTIVE

Proteins Adsorbed during Intraoperative Hemoadsorption and Their In Vitro Effects on Endothelium

(1) Background: Hemoadsorption is a method of blood purification with a wide spectrum of indications. Pre-emptive use of hemoadsorption in patients undergoing heart surgery with cardiopulmonary bypass is considered to reduce the risk of postoperative systemic inflammatory response syndrome. The current study aimed to identify the spectrum of blood proteins adsorbed on the polymer matrix of the CytoSorb hemoadsorption system and to investigate their influence on cultured endothelial cells in vitro. (2) Methods: Adsorbers used for intraoperative hemoadsorption were obtained from patients undergoing on-pump valve surgery in acute endocarditis. Proteins were extracted from the adsorbers, purified, identified with mass-spectrometry and applied to cultured human aortic endothelial cells. (3) Results: A broad range of blood proteins were identified in the material eluted from the CytoSorb adsorber. When added to cultured ECs, these protein extracts caused severe reduction in cell viability and migration. After 24 h exposure, transcriptional changes with up-regulation of multiple metabolic regulators were observed and verified on the protein level. Genes responsible for control of mitosis were significantly down-regulated. (4) Conclusions: In summary, our data reveal that intraoperative hemoadsorption allows broad spectrum removal of a wide range of molecules eliciting endothelial damage.

Biochemical markers of thrombotic complications in the acute period of ischemic stroke

Aim. To study the profile of biochemical markers of the hemostasis system, to clarify their role and relationships in the pathogenesis of the development of thrombotic complications (TC) of ischemic stroke (IS) and the associated assessment of the possibilities of their diagnostic application. Materials and methods. The study group included 302 patients (164 men, 138 women) who were admitted to the hospital with a diagnosis of IS within 24 hours of the onset of the disease. The diagnosis was confirmed by computed tomography. The average age of patients was 69 (5088) years. Blood was taken from all patients on the 1st day of the disease to determine the profile of analytes presumably associated with the pathogenesis of TC. Levels of homocysteine, protein C inhibitor, thrombomodulin, plasminogen, tissue plasminogen activator, urokinase, plasminogen activator type 1 inhibitor, t-PA/PAI-1 complex, vitronectin, plasmin-2-antiplasmin complex, D-dimer, fibronectin were determined in blood serum by ELISA. Results. TC in the acute period of IS (up to 21 days) were recorded in 32 (10.6%, 95% CI 7.3714.3) patients, of which pulmonary embolism was observed in 27 (8.94%, 95% CI 5.9812.4) patients, deep vein thrombosis in 5 (1.66%, 95% CI 0.473.47) patients. The results of the study of a panel of specific proteins involved in pathogenetic processes accompanying necrosis of brain tissue in IS demonstrated that of the entire list of markers of the hemostasis system activation selected for the study, the most significant are: the concentration of fibronectin in the prognosis of the absence of TC with a threshold value of more than 61 mkg/ml and OR 4.4 (95% CI 1.512.9, p=0.011), and the concentration of the t-PA/PAI-1 complex in the prognosis of the development of TC with a threshold value of more than 14 ng/ml and OR 11.3 (95% CI 1.18109.3, p=0.03). Conclusion. The significance of the t-PA/PAI-1 complex and fibronectin as markers of TC in IS may be due to a violation of the activation processes of the fibrinolytic link of hemostasis and the accumulation of non-deposited compounds that damage the vascular wall.

Fibrinolysis and bleeding of unknown cause

Patients with bleeding of unknown cause (BUC) present with a variety of mild to moderate bleeding symptoms, but no hemostatic abnormalities can be found. Hyperfibrinolysis is rarely evaluated as the underlying cause for bleeding in clinical practice, and well-established global assays for abnormal fibrinolysis are lacking. Few patients with definitive fibrinolytic disorders, including α2-antiplasmin deficiency, plasminogen activator inhibitor 1 deficiency, or Quebec platelet disorder, have been reported. This review aims to summarize data on established fibrinolytic disorders and to discuss assessments of fibrinolysis in prior bleeding cohorts. Furthermore, we review available global tests with the potential to measure fibrinolysis, such as turbidity fibrin clot assays and rotational thromboelastometry, and their relevance in the workup of patients with BUC. We conclude that, due to the lack of adequate global tests, hyperfibrinolysis might be an underdiagnosed cause for a bleeding disorder. The diagnosis of hyperfibrinolytic bleeding disorders would improve patient care as effective treatment with antifibrinolytic agents is available.

Von Willebrand factor in aortic or mitral valve stenosis and bleeding after heart valve surgery

OBJECTIVES

Low von Willebrand factor (VWF) increases the risk of bleeding. The objective was to assess the influence of VWF on bleeding after valvular surgery.

METHODS

We studied 82 consecutive patients in median age of 65.5 years with severe isolated aortic stenosis (AS, n = 62) or mitral stenosis (MS, n = 20), undergoing heart valve surgery in extracorporeal circulation. Preoperatively, we assessed VWF antigen (VWF:Ag) and activity (VWF:RCo), a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13 (ADAMTS13), and fibrinolysis inhibitors.

RESULTS

Compared with AS, MS patients were more frequently female (80 vs. 55%, p = 0.045) with atrial fibrillation (AF) (80 vs. 8%, p < 0.0001), with no difference in age or comorbidities. Median postoperative drainage was 420 ml for AS, and 425 ml for MS (p = 0.37). Patients with AS had lower VWF:RCo (125.8 [88.5-160.8] vs. 188.0 [140.3-207.3] IU/dl, p = 0.003) and VWF:Ag (135.8 [112.0-171.2] vs. 191.7 [147.3-236.4] IU/dl, p = 0.01) than MS patients. Mean VWF:RCo/Ag ratio was 0.88 ± 0.17, with no intergroup differences. ADAMTS13 levels and activity were similar in both groups. In AS, both VWF:RCo and VWF:Ag correlated inversely with maximal (r = -0.39, p = 0.0003 and r = -0.39, p = 0.0004, respectively) and mean (r = -0.40, p = 0.0004 and r = -0.39, p = 0.0006, respectively) transvalvular pressure gradients. There was no difference in perioperative bleeding between patients following mitral and aortic valve surgery, and bleeding was not associated with VWF:Ag or VWF:RCo.

CONCLUSIONS

In severe AS, VWF levels and activity correlate inversely with transvalvular pressure gradients, and are lower than in severe degenerative MS, but do not affect blood loss after valvular surgery in extracorporeal circulation.

The values of coagulation function in COVID-19 patients

OBJECTIVE

To investigate the blood coagulation function in COVID-19 patients, and the correlation between coagulopathy and disease severity.

METHODS

We retrospectively collected 147 clinically diagnosed COVID-19 patients at Wuhan Leishenshan Hospital of Hubei, China. We analyzed the coagulation function in COVID-19 patients through the data including thrombin-antithrombin complex (TAT), α2-plasmininhibitor-plasmin Complex (PIC), thrombomodulin (TM), t-PA/PAI-1 Complex (t-PAIC), prothrombin time (PT), international normalized ratio (INR), activated partial thromboplastin time (APTT), fibrinogen (FIB), thrombin time (TT), D-Dimer (DD), and platelet (PLT).

RESULT

The levels of TAT, PIC, TM, t-PAIC, PT, INR, FIB, and DD in COVID-19 patients were higher than health controls (p<0.05), and also higher in the patients with thrombotic disease than without thrombotic disease (p<0.05). What's more, the patients with thrombotic disease had a higher case-fatality (p<0.05). TAT, PIC, TM, t-PAIC, PT, INR, APTT, FIB, DD, and PLT were also found correlated with disease severity. Meanwhile, we found that there were significant difference in TAT, TM, t-PAIC, PT, INR, APTT, DD, and PLT in the death and survival group. Further using univariate and multivariate logistic regression analysis also found that t-PAIC and DD were independent risk factors for death in patients and are excellent predicting the mortality risk of COVID-19.

CONCLUSION

Most COVID-19 patients with inordinate coagulation systems, dynamic monitoring of coagulation parameters might be a key in the control of COVID-19 death.

Thromboelastometry for Assessing Risks of Free Flap Thrombosis in Patients Undergoing Microvascular Surgery

Introduction: Coagulation assessment is often missing in microvascular surgery. We aimed at evaluating the predictive value of thromboelastometry for free flap thrombosis in microvascular surgery patients. Materials and Methods: We enrolled 103 adult patients with traumatic injuries scheduled for microvascular free flap surgery into a prospective observational study. Thirty-six patients with recent trauma underwent surgery within 30 days (ES group), and were compared with 67 trauma patients who underwent surgery later than 30 days (late surgery, LS group) after the injury. Rotational thromboelastometry (RTE) was performed before surgery. Functional fibrinogen to platelet ratio (FPR) ≥ 42 was selected as the main hypercoagulability index. Free flap thrombosis was set as primary outcome. Thrombotic risk factors and duration of surgery related to free flap thrombosis were secondary outcomes. Statistical significance p < 0.05; not significant NS. Results: Six patients (16.7%) in the ES group and 10 (14.9%) in the LS group had free flap thrombosis (NS). In the entire cohort, free flap thrombosis rate increased in the presence of thrombogenic comorbidities (OR 4.059, CI 1.33-12.37; p = 0.014) and prolonged surgery times (OR 1.007, CI 1 - 1.012; p = 0.05). Although hypercoagulability occurred more frequently in the ES group (44.4%) than in the LS group (11.9%; p < 0.001), it was not associated with higher free flap thrombosis rate. In ES group patients with surgery times > 240 min, the risk of free flap thrombosis increased (OR 3.5, CI 1.16-10.6; p = 0.026) with 93.3% sensitivity and 86.7% specificity (AUC 0.85; p = 0.007). In contrast, in LS patients hypercoagulability increased the odds of free flap thrombosis (OR 8.83, CI 1.74-44.76; p = 0.009). Moreover, a positive correlation was found between FPR ≥ 42 and free flap thrombosis rate (r = 0.362; p = 0.003). In the LS group, the presence of thrombogenic comorbidities correlated with free flap thrombosis rate (OR 7, CI 1.591-30.8; p = 0.01). Conclusions: In LS patients with thrombogenic comorbidities, thromboelastometry supports the detection of hypercoagulability and predicts free flap thrombosis risk. In ES patients, postoperative hypercoagulability did not predict free flap thrombosis. Prolonged surgery time should be considered as a risk factor.

Risk factors of acute coagulation dysfunction after aneurysmal subarachnoid hemorrhage

Background

Although coagulopathy have been proved to be a contributor to a poor outcome of aneurysmal subarachnoid hemorrhage (aSAH), the risk factors for triggering coagulation abnormalities have not been studied after aneurysm clipping.

Methods

We investigated risk factors of coagulopathy and analyzed the relationship between acute coagulopathy and outcome after aneurysm clipping. The clinical data of 137 patients with ruptured CA admitted to our institution was collected and retrospectively reviewed. Patient demographic data (age, sex), smoking, alcohol use, hypertension, diabetes, Hunt-Hess grade, Fisher grade, operation time, intraoperative total infusion volume, intraoperative blood loss, intraoperative transfusion, intraoperative hemostatic drug treatment, calcium reduction (preoperative free calcium concentration–postoperative free calcium concentration) were recorded. Coagulation was assessed within 24 h. Postoperative hemorrhage and infarction, deep venous thrombosis (DVT), and mortality were analyzed.

Results

Coagulopathy was detected in a total of 51 cases (group I), while not in 86 cases (group II). Univariable analysis demonstrated that age, smoking, alcohol use, intraoperative total infusion volume, intraoperative blood loss, intraoperative transfusion, and calcium reduction (≥ 1.2 mg/dl) were related to coagulopathy. Non-conditional logistic regression analysis showed that age [OR, 1.037 (95% CI, 1.001–1.074); p = 0.045] and calcium reduction (≥ 1.2 mg/dl) [OR, 5.509 (95% CI, 1.900–15.971); p = 0.002] were considered as the risk factors for coagulopathy. Hunt-Hess grade [OR, 2.641 (95% CI, 1.079–6.331); p = 0.033] and operation time [OR, 0.107 (95% CI, 1.012–0.928); p = 0.043] were considered as the risk factors for hypocoagulopathy. There were 6 cases (11.7%) with cerebral infarction in group I, while 6 cases (6.98%) in group II (χ² = 0.918, p = 0.338). There were 4 cases (7.84%) with rebleeding in group I, while 5 cases (5.81%) in group II (χ² = 0.215, p = 0.643). The mortality was 9.80% (5/51) in group I, while 1.16% (1/86) in group II (χ² = 5.708, p = 0.017). DVT was not detected in all cases.

Conclusions

In conclusion, age (≥ 65 years) and calcium reduction (≥ 1.2 mg/dl) were considered as the risk factors for coagulopathy and have been proved to be associated with higher mortality after aneurysm clipping.

Hemorrhagic disorders of fibrinolysis: a clinical review

Hyperfibrinolytic bleeding can be caused by a deficiency of one of the inhibitors of fibrinolysis (plasminogen activator inhibitor type 1 [PAI-1] or α2-antiplasmin [α2-AP]), or an excess of one of the activators of fibrinolysis: tissue-type plasminogen activator or urokinase-type plasminogen activator. This review focuses on the clinical implications of these disorders. The bleeding phenotype of fibrinolytic disorders is characterized by delayed bleeding after trauma, surgery and dental procedures. Bleeding in areas of high fibrinolytic activity is also common, such as menorrhagia and epistaxis. Patients with α2-AP deficiency present with the most severe bleeding episodes. Recently, it was discovered that hyperfibrinolytic disorders are associated with a high rate of obstetric complications such as miscarriage and preterm birth, especially in PAI-1 deficient patients. Hyperfibrinolytic disorders are probably underdiagnosed because of lack of knowledge and lack of accurate diagnostic tests. A substantial part of the large group of patients diagnosed as 'bleeding of unknown origin' could actually have a hyperfibrinolytic disorder. In the case of a high index of suspicion (i.e. because of a positive family history, recurrent bleeding or uncommon type of bleeding such as an intramedullary hematoma), further testing should not be withheld because of normal results of standard hemostatic screening assays. Timely diagnosis is important because these disorders can generally be treated well with antifibrinolytic agents.

Fibrinolysis in patients with a mild-to-moderate bleeding tendency of unknown cause

In more than 50% of patients with a mild-to-moderate bleeding tendency, no underlying cause can be identified (bleeding of unknown cause, BUC). Data on parameters of fibrinolysis in BUC are scarce in the literature and reveal discrepant results. It was the aim of this study to investigate increased fibrinolysis as a possible mechanism of BUC. We included 270 patients (227 females, median age 44 years, 25–75^th percentile 32–58) with BUC and 98 healthy controls (65 females, median age 47 years, 25–75^thpercentile 39–55). Tissue plasminogen activator (tPA-) antigen and activity, plasminogen activator inhibitor type-1 (PAI-1), tPA-PAI-1 complexes, thrombin activatable fibrinolysis inhibitor (TAFI), α2-antiplasmin, and D-dimer were determined. While PAI-1 deficiency was equally frequent in patients with BUC and controls (91/270, 34%, and 33/98, 34%, p = 0.996), tPA activity levels were more often above the detection limit in patients than in controls (103/213, 48%, and 23/98, 23%, p < 0.0001). We found lower levels of tPA-PAI-1 complexes (6.86 (3.99–10.00) and 9.11 (7.17–13.12), p < 0.001) and higher activity of TAFI (18.61 (15.80–22.58) and 17.03 (14.02–20.02), p < 0.001) and α2-antiplasmin (102 (94–109) and 98 (90–106], p = 0.003) in patients compared to controls. Detectable tPA activity (OR 3.02, 95%CI 1.75–5.23, p < 0.0001), higher levels of TAFI (OR 2.57, 95%CI 1.48–4.46, p = 0.0008) and α2-antiplasmin (OR 1.03, 95%CI 1.01–1.05, p = 0.011), and lower levels of tPA-PAI-1 complexes (OR 0.90, 95%CI 0.86–0.95, p < 0.0001) were independently associated with BUC in sex-adjusted logistic regression analyses. We conclude that the fibrinolytic system can play an etiological role for bleeding in patients with BUC.

Seawater Immersion Aggravates Burn Injury Causing Severe Blood Coagulation Dysfunction

This study aimed to investigate the endothelial function in a canine model of burn injury combined with seawater immersion. The model of burn injury was established. The dogs were randomly divided into four groups including dogs with burn injury (B group), or burn injury combined with seawater immersion (BI group), or only immersion in seawater (I group), or control animals with no injury or immersion (C group). The circulating endothelial cell (CEC) count and coagulation-fibrinolysis parameters were measured. The CEC count in B group increased at 4 h, 7 h, and 10 h after injury and then reduced, whereas it continuously increased to a greater extent in BI group (P < 0.05). The von Willebrand factor (vWF) activity, plasminogen activator inhibitor (PAI-1), and the ratio of thromboxane B2 (TXB₂) to 6-keto-prostaglandin F1α (6-K-PGF_1α) in BI group had a marked increase after injury, and the tissue-type plasminogen activator (tPA) in the BI group decreased. Microscope observations revealed thrombus formation in lungs of the animals in BI group, but not in C, I, or B groups. Burn injury causes endothelial dysfunction, and seawater immersion lastingly aggravates this injury, leading to a higher risk of developing thrombosis.

Polymorphisms on PAI-1 and ACE genes in association with fibrinolytic bleeding after on-pump cardiac surgery

Background

Carriers of plasminogen activator inhibitor -1 (PAI-1) -675 genotype 5G/5G may be associated with lower preoperative PAI-1 plasma levels and higher blood loss after heart surgery using cardiopulmonary bypass (CPB). We speculate if polymorphisms of PAI-1 -844 A/G and angiotensin converting enzyme (ACE) intron 16 I/D also might promote fibrinolysis and increase postoperative bleeding.

Methods

We assessed PAI-1 -844 A/G, and ACE intron 16 I/D polymorphisms by polymerase chain reaction technique and direct sequencing of genomic DNA from 83 open heart surgery patients that we have presented earlier. As primary outcome, accumulated chest tube drainage (CTD) at 4 and 24 h were analyzed for association with genetic polymorphisms. As secondary outcome, differences in plasma levels of PAI-1, t-PA/PAI-1 complex and D-dimer were determined for each polymorphism. SPSS® was used for statistical evaluation.

Results

The lowest preoperative PAI-1 plasma levels were associated with PAI-1 -844 genotype G/G, and higher CTD, as compared with genotype A/A at 4 and 24 h after surgery. Correspondingly, 4 h after the surgery CTD was higher in carriers of ACE intron 16 genotype I/I, as compared with genotype D/D. PAI-1 plasma levels and t-PA/PAI-1 complex reached nadir in carriers of ACE intron 16 genotype I/I, in whom we also noticed the highest D-dimer levels immediately after surgery. Notably, carriers of PAI-1 -844 genotype G/G displayed higher D-dimer levels at 24 h after surgery as compared with those of genotype A/G.

Conclusions

Increased postoperative blood loss secondary to enhanced fibrinolysis was associated with carriers of PAI-1 -844 G/G and ACE Intron 16 I/I, suggesting that these genotypes might predict increased postoperative blood loss after cardiac surgery using CPB.

Electronic supplementary material

The online version of this article (doi:10.1186/s12871-015-0101-1) contains supplementary material, which is available to authorized users.

Effect of two doses of tranexamic acid on fibrinolysis evaluated by thromboelastography during cardiac surgery: a randomised, controlled study

BACKGROUND

Tranexamic acid is used to decrease bleeding and transfusions during cardiac surgery. However, dosing based on pharmacokinetic data to optimally inhibit fibrinolysis is unknown. With increasing concerns regarding seizures associated with higher doses, lower dosing schemes may be important.

OBJECTIVE

To determine the effect of two dosing schemes compared with placebo on fibrinolysis and clinical outcomes.

DESIGN

A double-blind, randomised, controlled, pilot trial.

SETTING

Single tertiary centre.

PATIENTS

Cardiac surgery patients requiring cardiopulmonary bypass.

INTERVENTION

Patients were randomised to receive a 30 mg  kg(-1) bolus and continuous infusion of 16  mg  kg (-1) h(-1) (Group HIGH), a 5 mg  kg(-1) bolus followed by 5 mg  kg(-1)  h(-1) (Group LOW) or Sodium chloride (Placebo).

MAIN OUTCOME MEASURE

Fibrinolysis was evaluated by thromboelastography and D-dimers. Secondary endpoints were blood loss, transfusion requirement and side effects.

RESULTS

Thirty-three patients were included. Significant fibrinolysis was defined by LY30 more than 7.5% based on thromboelastography and was not observed after cardiopulmonary bypass in any groups. After protamine administration, LY30 differences between groups were 0.7 [95% confidence interval (95% CI) -0.04 to 1.4] between Groups HIGH and Placebo, -0.08 (95% CI -0.82 to 0.66) between Groups HIGH and LOW, and 0.78 (95% CI 0.02 to 1.5) between Groups LOW and Placebo. A significant increase in D-dimers was observed in the Group Placebo compared with the two treatment groups. There were no differences in bleeding or transfusion requirement.

CONCLUSION

In this dose-finding study, there were no differences in fibrinolysis or clinical outcomes among the two tranexamic acid schemes and placebo. Any difference in fibrinolytic inhibition requires a larger adequately powered study.

TRIAL REGISTRATION

EudraCT number: 2010-024104-99.

D-dimer and risk of thromboembolic and bleeding events in patients with atrial fibrillation--observations from the ARISTOTLE trial

BACKGROUND

D-dimer is related to adverse outcomes in arterial and venous thromboembolic diseases.

OBJECTIVES

To evaluate the predictive value of D-dimer level for stroke, other cardiovascular events, and bleeds, in patients with atrial fibrillation (AF) treated with oral anticoagulation with apixaban or warfarin; and to evaluate the relationship between the D-dimer levels at baseline and the treatment effect of apixaban vs. warfarin.

METHODS

In the ARISTOTLE trial, 18 201 patients with AF were randomized to apixaban or warfarin. D-dimer was analyzed in 14 878 patients at randomization. The cohort was separated into two groups; not receiving vitamin K antagonist (VKA) treatment and receiving VKA treatment at randomization.

RESULTS

Higher D-dimer levels were associated with increased frequencies of stroke or systemic embolism (hazard ratio [HR] [Q4 vs. Q1] 1.72, 95% confidence interval [CI] 1.14-2.59, P = 0.003), death (HR [Q4 vs. Q1] 4.04, 95% CI 3.06-5.33) and major bleeding (HR [Q4 vs. Q1] 2.47, 95% CI 1.77-3.45, P < 0.0001) in the no-VKA group. Similar results were obtained in the on-VKA group. Adding D-dimer level to the CHADS2 score improved the C-index from 0.646 to 0.655 for stroke or systemic embolism, and from 0.598 to 0.662 for death, in the no-VKA group. D-dimer level improved the HAS-BLED score for prediction of major bleeds, with an increase in the C-index from 0.610 to 0.641. There were no significant interactions between efficacy and safety of study treatment and D-dimer level.

CONCLUSION

In anticoagulated patients with AF, the level of D-dimer is related to the risk of stroke, death, and bleeding, and adds to the predictive value of clinical risk scores. The benefits of apixaban were consistent, regardless of the baseline D-dimer level.