D-dimer determination as a screening tool to exclude atrial thrombi in atrial fibrillation

Abstract

Conversion of atrial fi brillation (AF) to sinus rhythm carries a defi nite risk of systemic embolism. This risk can be reduced to an acceptable level by prior anticoagulation for 3 weeks or, alternatively, by exclusion of atrial thrombi with transesophageal echocardiography (TEE) before cardioversion. 1 Although earlier cardioversion according to TEE-guided strategy has potential advantages, 2,3 TEE requires special equipment and advanced skills, which may limit its availability. D-dimer, an indirect marker of fi brin formation, has been demonstrated to be a sensitive marker of activation of coagulation in various clinical conditions. 4 –7 In AF, elevated D-dimer levels have been reported to be associated with the duration of AF, 8 left atrial appendage dysfunction, 9 and the presence of atrial thrombus. 10,11 The potential role of D-dimer in refi ning embolic risk prediction in AF remains to be determined. The aims of our study were to analyze the relation of serum D-dimer to embolic risk determined on clinical grounds in patients with AF and to assess the test’s diagnostic utility to predict the presence or absence of atrial thrombi.  In a prospective design, we measured the plasma D-dimer concentration of consecutive patients admitted to our echocardiography laboratory for TEE before cardioversion. The local ethics committee approved the protocol. The examinations were performed after obtaining written informed consent. Patients with AF of 48 hours duration or atrial fl utter with documented history of AF were eligible. The exclusion criteria included any thromboembolic events within the previous 2 months and anticoagulant therapy for 7 days. Subjects underwent transthoracic and TEE. Echocardiographic studies were performed with an ATL instrument (HDI 3500 CV, Philips Ultrasound, Bothell, Washington) using a 1.67- to 3.2-MHz, phased-array head (P4 to P2) for transthoracic examination, and a 4- to 7-MHz omniplane probe (MPT7 to MPT4) for TEE. An intravenous line was placed and blood samples were obtained without tourniquet immediately before TEE. D-dimer measurements were obtained with the use of a quantitative sandwich immunochromatographic technique (Cardiac D-dimer; Roche Diagnostics, Mannheim, Germany). The investigators and attending physicians were blinded to the D-dimer test results. The fi nal interpretations of the TEEs were the result of the consensus of the same 2 experienced investigators based on real-time observation or off-line analysis of the studies. For D-dimer results below the measurement range, 0.1 g/ml was entered. Values are presented as mean SEM. To obtain normal distribution, D-dimer values were logarithmically transformed, and unpaired 2-tailed t tests were used to compare means. The independent predictive role of embolic predictors for log D-dimer was evaluated by multiple regression analysis. Sensitivity,