Head-to-head comparison between the non-vitamin K antagonist oral anticoagulants for treatment of venous thrombo-embolism: do we really need it?

This editorial refers to ‘Safety and effectiveness of rivaroxaban and apixaban in patients with venous thrombo-embolism: a nationwide study’, by C. Sindet-Pedersen et al., on page 220. Treatment of venous thrombo-embolism (VTE) has been changed since the introduction of the non-vitamin K antagonist oral anticoagulants (NOACs), based on the results of several large randomized trials, including >27 000 patients with acute VTE (Table 1). All four NOACs (apixaban, dabigatran, edoxaban, and rivaroxaban) have been proved to be non-inferior to comparators, in terms of efficacy on recurrent VTE or VTE-related death, for the treatment of acute VTE, while some of the NOACs (apixaban, dabigatran, and edoxaban) have been proved to be superior to comparators in terms of reduction of clinically relevant non-major bleeding.1–5 Moreover, apixaban and rivaroxaban are preferred over dabigatran and edoxaban, since both dabigatran and edoxaban require bridge treatment with heparin.2,3 Comparison of design and main results of phase III trials for the treatment of acute venous thrombo-embolism VTE, venous thrombo-embolism; DVT, deep vein thrombosis; PE, pulmonary embolism; HR, hazard ratio; CI, confidence interval; CRNM, clinically relevant non-major. aAcute DVT study; bacute PE study. Comparison of design and main results of phase III trials for the treatment of acute venous thrombo-embolism VTE, venous thrombo-embolism; DVT, deep vein thrombosis; PE, pulmonary embolism; HR, hazard ratio; CI, confidence interval; CRNM, clinically relevant non-major. aAcute DVT study; bacute PE study. Currently, there is no head-to-head comparison between apixaban and rivaroxaban. Some evidence has been provided by two meta-analyses comparing the AMPLIFY and EINSTEIN studies.6,7 In terms of efficacy, the absolute rate of recurrent VTE was 2.3% on apixaban in the AMPLIFY study,1 and 2.1% on rivaroxaban in the EINSTEIN studies,4,5 this difference being non-significant.6,7 Similarly, in terms of safety, the absolute rate of major bleeding was 0.6% on apixaban in the AMPLIFY study,1 and 0.8–1.1% on rivaroxaban in the EINSTEIN studies (acute deep vein thrombosis study and acute pulmonary embolism study, respectively),4,5 this difference also being non-significant.6,7 However, in one of the meta-analyses, apixaban was associated with a significantly lower risk of major bleeding or clinically relevant non-major bleeding [hazard ratio (HR) 0.47; 95% confidence interval (CI) 0.36–0.61].6 Consequently, for treatment of acute VTE, the European guidelines recommend (class I, level of evidence B) either apixaban or rivaroxaban as an alternative to combination of parenteral anticoagulation with a vitamin K antagonist (VKA), and either dabigatran or edoxaban as an alternative to VKA, following acute-phase parenteral anticoagulation, without making any difference between the four NOACs.8 A study published in this issue of the European Heart Journal – Cardiovascular Pharmacotherapy aimed to compare effects of apixaban vs. rivaroxaban in patients with VTE on the risk of death, recurrent VTE, and hospitalized bleeding in a Danish nationwide cohort.9 In this registry analysis, including 8187 patients with VTE treated with either apixaban (1504 patients; 18%) or rivaroxaban (6683 patients; 82%), there were no significant differences at 180 days in the risk of death, recurrent VTE, or hospitalized bleeding. There are several differences between patients from this registry study and patients included in the randomized AMPLIFY and EINSTEIN trials:1,4,5 (i) they were older (by a median of 10 years); (2) consequently, they had more co-morbidities, including cancer; (iii) they had a higher rate of pulmonary embolism; whereas (iv) they had lower rates of previous VTE and thrombophilia. Therefore, all-cause mortality (5.1% for apixaban and 4.6% for rivaroxaban) was different from that reported in the AMPLIFY (1.5%) and EINSTEIN studies (0.3 and 2.4%).1,4,5,9 Meanwhile, definition of bleeding (as ‘in-hospital admission with a bleeding diagnosis’) was different from those used for major bleeding and clinically relevant non-major bleeding in the AMPLIFY and EINSTEIN trials, making the comparison difficult. This registries-based study has several strengths: (i) it compares the efficacy and safety of apixaban vs. rivaroxaban in a nationwide cohort; and (ii) it included all patients with a diagnosis of VTE between January 2015 and June 2017, excluding any selection bias. Therefore, it reflects everyday clinical practice better than the randomized trials. However, there are some major limitations of this study: (i) the size of the study and number of events are too small and, therefore, the study is underpowered to detect a difference between the two drugs; moreover, there is a major discrepancy between the size of the apixaban (18%) and rivaroxaban (82%) groups; (ii) lack of the adjudication of the events, a limitation similar for all observational studies; (iii) major differences between the general characteristics of the two groups, patients on apixaban being older, with more co-morbidities, and with a higher rate of pulmonary embolism as an index event, than patients on rivaroxaban; although the authors adjusted in their analysis for these differences, there might be also other residual confounders, such as body mass index, smoking status, blood pressure, lipid status, etc., not available in the registries; and (iv) definition of bleeding is different from those used in the randomized trials, making the comparison difficult and prone to underestimation of the number of all bleeding events. With all these limitations, we can admit that this registries-based study provides an insight into the absolute risks of death, recurrent VTE, and hospitalized bleeding, when apixaban and rivaroxaban are used in a nationwide cohort. Due to the small differences in terms of efficacy and safety endpoints between the NOACs (Table 1), confirmed by this ‘real-world’ study, a head-to-head randomized trial would need a very large number of patients followed for a long period of time, in order to have enough power to demonstrate differences. Since all four NOACs are comparable in terms of efficacy and safety, and equally recommended by the current guidelines, the energy and costs implied by such a trial would probably be unacceptable. However, more pragmatic trials, based on ‘real-world’ data, might provide more useful results, and will be of high relevance for current clinical practice. Conflict of interest: The author has received research grants and consultancy speaker fees from Pfizer/BMS, Boehringer Ingelheim, Daiichi Sankyo, and Bayer.