Warfarin, an oral anti-vitamin K anticoagulant, effectively prevents venous thrombosis in patients at risk, but requires constant laboratory monitoring to maintain a therapeutic range and reduce bleeding complications. For many indications, oral inhibitors of coagulation factor (F) Xa and thrombin (FIIa) have proven at least as effective as warfarin with comparable risk of bleeding complications, but with the advantage of administration in a fixed dose without laboratory monitoring. Although rare, thrombosis may still occur in treated patients and bleeding remains a potentially serious complication.


The assumption that an equal dose of an anticoagulant drug can achieve a comparable antithrombotic effect in different patients implies that an appropriate test of efficacy should yield results within a relatively narrow range of values. Thus, we measured the volume of platelet aggregates and fibrin deposited onto thrombogenic surfaces exposed to flowing blood to compare the antithrombotic effect of rivaroxaban (Riv), a FXa inhibitor, and dabigatran (Dab), a thrombin inhibitor, with that of warfarin (Warf) in patients undergoing total knee or hip replacement and treated to prevent deep vein thrombosis.


Blood containing 0.011 M trisodium citrate was recalcified with 5 mM CaCl2 and perfused at the wall shear rate of 300 and 1500 s-1 over a surface coated with fibrillar collagen type I, or at 300 s-1 over recombinant tissue factor (TF). Platelet aggregates and fibrin were detected in situ through distinct fluorochromes and the respective volumes were measured from stacks of confocal optical sections. We tested 12 normal controls, 12 patients treated with Warf (INR between 1.94 and 2.90), 10 patients treated with Riv and 10 with Dab between 8 and 16 days from the initiation of therapy. Statistical analysis was performed using one-way analysis of variance.


On the collagen surface at the lower shear rate of 300 s-1, the blood of Warf patients yielded an average volume of deposited fibrin significantly lower not only of control (P<0.001) but also Riv and Dab patients (P<0.01); the latter two were not different from control. All 12 Warf samples, but only 5/10 Riv and 4/10 Dab, were below the lower limit of normal values. In contrast, at the higher shear rate of 1500 s-1 the average fibrin volume was significantly lower than control in Warf (P<0.001) as well as Riv and Dab (P<0.01) blood. Notably, at both lower and higher shear rate the average volume of platelet aggregates was not decreased in Warf, Riv or Dab blood; rather, the tendency was to increased. On the TF surface, stable thrombus formation in flowing blood could only be assessed at the lower shear rate of 300 s-1, since even the reactivity of normal samples was negligible at 1500 s-1; thus we focused on blood from Riv and Dab treated patients who showed only an insignificant reduction of fibrin formation at 300 s-1 on the collagen surface. In contrast, thrombus volume in the blood of Riv and Dab treated patients was decreased on the TF surface under the same flow conditions. In the case of Dab, the average volume of both platelet aggregates and fibrin was significantly lower (P<0.01) than in control blood; with Riv, fibrin volume was also significantly reduced (p<0.01), but that of platelet aggregates was not. Of note, the volume of fibrin formed in Dab samples was only one third of that in Riv samples.


The significance of findings obtained with a test of thrombus formation in ex vivo flowing blood and the potential value for interpreting the evidence provided by clinical trials remain to be established prospectively. This notwithstanding, it is apparent that different agents have different anticoagulant effects, and those administered in a “one fits all” dose and without laboratory monitoring lead to a greater inter-individual variability of results. The influence of the thrombogenic surface on the ex vivo anticoagulant potency of drugs with distinct coagulation targets suggests that treating or preventing different thrombotic disorders may require a selective choice among therapeutic agents targeting specific coagulation pathways and physiologic inhibitors.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.