Abstract
Venous thromboembolism, VTE, is a major cardiovascular disease with significant mortality and morbidity. The current management of VTE is changing as the direct oral anticoagulants have become available with more evidence emerging on their safety and efficacy profiles. This study prospectively investigated patients taking rivaroxaban, a direct oral anticoagulant, either for venous thromboembolism treatment or for prevention post-orthopaedic surgery. The pharmacodynamic behaviour of rivaroxaban in our study population was evaluated using several standard, specialised and global coagulation assays, such as calibrated automated thrombography, rotational thromboelastography and thrombin anti-thrombin complexes. The pharmacokinetics of rivaroxaban were evaluated and a population pharmacokinetic model was developed using non-linear mixed effects modelling. Finally, patient’s attitudes and beliefs, regarding rivaroxaban, were explored in the cohort of patents recruited.In total, 101 patients were recruited. The indications for taking rivaroxaban were: 58% as treatment for an initial DVT, 26% as treatment for recurrent DVT, 12% were previously receiving long-term conventional anticoagulation (vitamin K inhibitor or low molecular weight heparin, LMWH) and switched to rivaroxaban, and 4% as primary prophylaxis following an elective total hip or knee replacement surgery. Almost three quarters of the participants were Caucasian, there was an equal male:female ratio, body weight was normally distributed. 40% of participants were obese and only 1% was under weight.
Rivaroxaban concentrations were measured by anti-Xa levels, which is an accepted surrogate measurement for rivaroxaban concentration. The concentration-time graph produced from this study is representative of those graphs published in the literature, with a peak Cmax at 2-3h post rivaroxaban ingestion and an exponential decrease with time, to a trough at approximately 24h. There was a wide distribution of rivaroxaban concentrations between time zero and 3h post ingestion, reflecting the variability in absorption. There was no association between rivaroxaban concentration and body weight, lean body weight or BMI, at any time point after rivaroxaban ingestion.
Prothrombin time, PT, was measured at each blood sampling time point. The PT did not differ significantly between groups of patients, based on various demographic characteristics or body weight categories. Spearman rank coefficient of variability between PT and anti-Xa levels, was 0.9, p=0.00, and a multiple regression analysis showed that this correlation was not affected by body weight.
Activated partial thromboplastin time (APTT) was normally distributed, mean 33.4 sec and correlated with rivaroxaban concentration, r=0.306, p=0.00. D-dimer, median 350 ng/ml, displayed no significant differences between any sub-group of weight, ethnicity, indication for anticoagulation or dose of anticoagulation. Clauss fibrinogen (mean 3.47 g/L) weakly correlated with rivaroxaban concentration, r=0.155, p=0.032.
Thrombin anti-thrombin complex (TAT) measurements had no association with rivaroxaban concentration. There were differences in TAT only between the orthopaedic patients and other groups, but the significance was unreliable due to small numbers of samples. The global coagulation tests: calibrated automated thrombogram (CAT), with and without the addition of contact factor inhibitor (CTI); and rotational thromboelastography (ROTEM) were performed and analysed for each sample. CAT parameters correlated significantly with rivaroxaban concentration, in particular, lag time, r= 0.60, p= 0.00, time to peak, r= 0.67, p= 0.00, peak thrombin generation, r= -0.63, p= 0.00, velocity index r= -0.66, p=0.00. The ROTEM parameters which correlated significantly with rivaroxaban concentration, were Extem CT, A10 and MCFt and Intem CT.
Only time to peak of thrombin generation correlated with lean body weight, r= 0.14, p=0.00, but no other body weight indices showed a relationship with any CAT or ROTEM parameters.
Using NOMNEM software, a population pharmacokinetic analysis was performed on the rivaroxaban concentration data collected. A one compartment model with between subject variability on rivaroxaban clearance and volume of distribution, with a combined (additive and proportional) error model, best fitted the data. Following a full co-variate analysis, creatinine clearance on rivaroxaban clearance was found to be the only significant co-variate impacting on the pharmacokinetic profile of rivaroxaban in the dataset. Body weight (BMI, actual body weight and lean body weight) was not found to have a significant effect on rivaroxaban concentration.
| Date of Award | 1 Mar 2020 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Jignesh Patel (Supervisor) & Roopen Arya (Supervisor) |