Blood Clotting Markers and Heart Ultrasound in People With Device-Detected Atrial Fibrillation (DDAF-HEV)

Scientific Rationale Device-detected atrial fibrillation (DDAF) is associated with an increased risk of stroke and systemic embolism, although the risk is lower than in patients with clinically diagnosed atrial fibrillation. Consequently, it remains challenging to identify which patients are most likely to benefit from oral anticoagulant therapy while minimizing the risk of bleeding. Current stroke risk assessment in patients with device-detected atrial fibrillation relies primarily on clinical risk scores and does not incorporate haemostatic biomarkers, the burden of device-detected atrial fibrillation, or echocardiographic markers of atrial remodeling that may contribute to thromboembolic risk. This study aims to improve the understanding of thromboembolic risk in patients with DDAF and to evaluate whether integrating clinical stroke risk scores, DDAF burden, advanced echocardiographic variables, and haemostatic biomarkers may improve future stroke risk assessment.

Study Design This is a prospective, single-center observational cohort study conducted at the Department of Cardiology, Esbjerg and Grindsted Hospital, University Hospital of Southern Denmark, in collaboration with the Unit for Thrombosis Research, Department of Clinical Diagnostics.

The study consists of a single prospective observational cohort with three prespecified baseline analyses followed by a prospective 10-year longitudinal follow-up. A total of 222 participants will be enrolled, including 111 participants with device-detected atrial fibrillation and 111 control participants matched for age, sex, and indication for cardiac device implantation.

Baseline Assessments At baseline, all participants will undergo standardized blood sampling, comprehensive transthoracic echocardiography, and collection of demographic and clinical information. Device interrogation will be performed to quantify the burden of device-detected atrial fibrillation. Clinical thromboembolic risk will be assessed using the CHA₂DS₂-VASc score. In addition, the ABC-stroke score will be calculated to evaluate its associations with haemostatic biomarkers and subsequent clinical outcomes in participants with device-detected atrial fibrillation.

Independent variables in the prespecified baseline analyses include the CHA₂DS₂-VASc score, the ABC-stroke score, burden of device-detected atrial fibrillation, and advanced echocardiographic variables.

The baseline analyses are designed to address three prespecified objectives. Together, these analyses are intended to determine whether established clinical stroke risk scores, burden of device-detected atrial fibrillation, and cardiac structural and functional abnormalities are associated with a more prothrombotic haemostatic profile.

1. To investigate whether the primary haemostatic biomarkers, endogenous thrombin potential (ETP) and von Willebrand factor antigen, differ between participants with device-detected atrial fibrillation and matched controls and whether they are associated with the CHA₂DS₂-VASc and ABC-stroke risk scores. Secondary analyses will evaluate associations between additional haemostatic biomarkers and the clinical stroke risk scores.
2. To investigate whether the primary haemostatic biomarkers are associated with the burden of device-detected atrial fibrillation. Secondary analyses will evaluate associations between DDAF burden and additional haemostatic biomarkers.
3. To investigate whether the primary haemostatic biomarkers are associated with advanced echocardiographic variables, including left atrial size and function. Secondary analyses will evaluate associations between echocardiographic variables and additional haemostatic biomarkers.

The primary cross-sectional outcome measures are endogenous thrombin potential (ETP) and von Willebrand factor antigen. Secondary cross-sectional outcome measures comprise additional biomarkers of coagulation and fibrinolysis.

Longitudinal Follow-up Participants will subsequently be followed for 10 years through electronic health records, routine device interrogation reports, and Danish national health registries. Follow-up data will be collected every second year to evaluate whether baseline primary and secondary haemostatic biomarkers, DDAF burden, clinical stroke risk scores, and echocardiographic findings are associated with subsequent stroke, systemic embolism, hospitalization, death, progression to clinically diagnosed atrial fibrillation, and initiation of oral anticoagulant therapy.

Statistical Analysis The sample size was calculated to provide 80% statistical power at a two-sided significance level of 5% to detect the expected differences in endogenous thrombin potential and von Willebrand factor antigen between predefined CHA₂DS₂-VASc stroke risk groups, based on previously published effect sizes.

Continuous variables will be assessed for normality using histograms and Q-Q plots. Variables with skewed distributions will be logarithmically transformed where appropriate. Homogeneity of variances will be assessed before parametric analyses. Continuous variables will be summarized as mean ± standard deviation or median with interquartile range according to data distribution, whereas categorical variables will be summarized as frequencies and percentages.

Baseline comparisons between participants with device-detected atrial fibrillation and matched controls will be performed using appropriate parametric or non-parametric statistical methods according to data distribution. Multivariable regression models will be adjusted for predefined confounding variables as appropriate.

The prespecified baseline analyses will primarily be performed using multivariable linear regression models with endogenous thrombin potential and von Willebrand factor antigen as the primary dependent variables. Secondary analyses will evaluate associations with additional haemostatic biomarkers. Regression models will be adjusted for prespecified confounding variables. Continuous predictors will be assessed for approximate linearity, and the assumptions underlying the linear regression models will be evaluated before interpretation of the results. Sensitivity analyses will be performed to evaluate the robustness of the primary findings. Adjustment for multiple testing will be performed using the Holm-Bonferroni procedure.

Longitudinal analyses will evaluate whether baseline haemostatic biomarkers, burden of device-detected atrial fibrillation, clinical stroke risk scores, and echocardiographic findings are associated with subsequent clinical outcomes. Time-to-event analyses will be performed using Cox proportional hazards regression models when appropriate. Hazard ratios with 95% confidence intervals will be reported. Statistical significance will be defined as a two-sided p-value <0.05.