Rivaroxaban Plasma Concentrations Cannot be Predicted from Coagulation Assays: Real-Life Experience From a Tertiary Hospital in New Zealand

Abstract

Background: – The use of direct oral anticoagulants, such as rivaroxaban, is preferred over vitamin K antagonists. Several clinical scenarios highlighting potential benefits of measuring plasma concentrations have been reported. This study aimed to describe the clinical experience with rivaroxaban in adult patients and develop a mathematical model to predict rivaroxaban concentrations using routine coagulation screening assays.Methods: – Data were prospectively collected over a 4-year-period, including clinical patient characteristics, total plasma rivaroxaban concentrations, and screening coagulation assays, if available. Prothrombin time, activated partial thromboplastin time (aPTT), and thrombin clotting time were included in the model to predict rivaroxaban concentrations.Results: – There were 403 rivaroxaban concentrations with a median of 66 mcg/L. The most common indications for measuring rivaroxaban concentration were routine postinitiation (78/403, 19%), breakthrough thromboembolic or bleeding events (98/403, 24%), and post-dose adjustments (42/403, 10%). Dose adjustments were made after measuring rivaroxaban concentrations in 85 patients (85/403, 21%). A statistically significant relationship existed between PT and aPTT against rivaroxaban concentrations; however, attempts to construct an accurate mathematical model using linear regression to determine rivaroxaban concentrations based on these coagulation assays were unsuccessful.Conclusions: – There was a wide range of measured rivaroxaban concentrations, reflecting the diversity of indications for measurement, with a significant minority associated with subsequent changes in rivaroxaban dosing. Given the frequency of altering dosing regimens based on the results, future efforts should focus on the correlation of concentration data with clinical endpoints to determine reference ranges. An accurate mathematical model to determine rivaroxaban concentrations based on coagulation screening assays could not be developed.