Rivaroxaban Effects Illustrate the Underestimated Importance of Activated Platelets in Thrombin Generation Assessed by Calibrated Automated Thrombography

Abstract

Background: The direct oral anticoagulant rivaroxaban inhibiting specifically activated factor X (FXa) causes delayed thrombin generation (TG) as measured by calibrated automated thrombography (CAT). The implications of these changes for assessing bleeding or residual prothrombotic risks of patients are unclear in the absence of a better understanding of the underlying mechanism.

Methods: We compared platelet rich plasma (PRP) without or with prior collagen-induced platelet aggregation (agPRP) in the CAT assay to better characterize TG in the presence of rivaroxaban.

Results: In the presence of rivaroxaban, TG curves in agPRP showed a distinct profile with a rapidly ascending phase followed with a protracted phase. Inhibition of tissue factor pathway inhibitor amplified the first phase of the curve which was also modulated by procoagulant phospholipids. Inhibition of FXIIa-dependent FXI activation revealed that aggregated platelets influenced the first phase by a combination of extrinsic and intrinsic coagulation pathway initiations. Thrombin-dependent amplification of TG (even prior collagen activation) was responsible for the second phase of the TG curve.

Conclusions: AgPRP fully includes platelet ability to support TG and reveal distinct TG phases in the presence of direct FXa inhibitors highlighting its potential use in an anticoagulated setting.

Keywords: factor Xa inhibitor; phospholipids; platelets; thrombin.

Figure 1

Effect of platelet aggregation with addition of rivaroxaban and anti-tissue factor pathway inhibitor (TFPI) on human platelet rich plasma (PRP) in thrombin generation (TG). Representative TG curves in PRP or PRP after aggregation (agPRP) triggered with 2 µg/mL of collagen with and without addition or 50 ng/mL of rivaroxaban (A) and inhibition of TFPI (anti-TFPI 5 µg/mL) (B). Lag time of thrombin generation for the different conditions (C,D). Velocity of thrombin generation under the different conditions (E,F). Results are presented as median (min-max). ** p < 0.01, *** p < 0.001, vs. PRP; ## p < 0.01, ### p < 0.001, vs. agPRP, n = 16 per group for left panels and n = 5 for right panels.

Figure 2

Modulation of camel-back shaped thrombin generation (TG) curve by the addition of phospholipids artificial vesicles. Representative curves of TG in platelet free plasma (PFP) supplemented with increasing concentrations of rivaroxaban: baseline (A), 15 ng/mL (B), 50 ng/mL (C), and 200 ng/mL (D). For each concentration, three concentrations of phospholipid vesicles (PV) were tested. PRP: Platelet rich plasma; agPRP with collagen 2 µg/mL, ETP, endogenous thrombin potential, tt peak, time to peak. Results presented as median (min-max). n = 6.

Figure 2

Modulation of camel-back shaped thrombin generation (TG) curve by the addition of phospholipids artificial vesicles. Representative curves of TG in platelet free plasma (PFP) supplemented with increasing concentrations of rivaroxaban: baseline (A), 15 ng/mL (B), 50 ng/mL (C), and 200 ng/mL (D). For each concentration, three concentrations of phospholipid vesicles (PV) were tested. PRP: Platelet rich plasma; agPRP with collagen 2 µg/mL, ETP, endogenous thrombin potential, tt peak, time to peak. Results presented as median (min-max). n = 6.

Figure 3

Effect of platelet aggregation with addition of rivaroxaban, PAR-1 and PAR-4 inhibitor to human platelet rich plasma (PRP) in TG. Representative TG curves in PRP or PRP after aggregation (agPRP) triggered with 2 µg/mL of collagen with and without addition or 50 ng/mL of rivaroxaban and inhibition of PAR-1 with vorapaxar (100 ng/mL) (A), inhibition of PAR-4 with BMS 896120 (1 µg/mL) (B) or a combination of both inhibitor (C). Lag time after addition of vorapaxar (D) BMS 986120 (E) or a combination of both inhibitor (F). Velocity after addition of vorapaxar (G) BMS 986120 (H) or a combination of both inhibitor (I). Results are presented as median (min-max). * p < 0.05, vs. PRP; # p < 0.05, ## p < 0.01, vs. agPRP, n = 4–5 per group.

Figure 4

Implication of tissue factor (TF) and thrombin positive feedback pathways in the camel-back shaped TG curve. Representative TG curves in PRP or PRP after aggregation (agPRP) triggered with 2 µg/mL of collagen with an antibody inhibiting FXIIa-dependent activation of FXI without affecting thrombin-dependent activation of FXI (14E11; 5 µg/mL) (A). Corresponding lag time and velocity (B,C). Representative curves of PRP and agPRP triggered with thrombin instead of TF (D) and corresponding lag time and velocity (E,F). Results are presented as median (min-max). * p < 0.05, vs. PRP; ## p < 0.01, vs. agPRP. n = 3–5 per group.