Progress in the development of new oral anticoagulants (NOACs), as well as agents for their reversal, has lowered the threshold to use these therapeutics as first line agents for the management of nonvalvular atrial fibrillation and venous thromboembolism.1,2 Despite this increase in adoption, however, debate persists as to whether patients chronically maintained on vitamin K antagonists (VKAs), such as warfarin, should be switched to NOACs. The recently published research letter by Pokorney et al. assessed the stability of international normalized ratios (INRs) in patients on long-term warfarin therapy in order to address this question.3
Specifically, prospective registry data from 3,749 patients with at least three INR values in the first 6 months of therapy as well as six or more in the following year were included. Patients were deemed stable if 80% or more of their INRs were in a therapeutic range defined as an INR between 2 and 3.3 During the initiation period, only one in four patients taking warfarin had a stable INR.3 Furthermore, stability in the first 6 months was found to have limited ability to predict stability in the subsequent year (concordance index of 0.61). With regard to time in therapeutic range (TTR), only 32% of patients had a TTR of greater than 80% during the first 6 months with less than half (42%) of these patients able to maintain this in the following year.
Findings from Pokorney et al. add to the growing body of literature demonstrating the difficulty of achieving and maintaining a therapeutic INR while on warfarin therapy.4-7 Clinically, these findings are important, as deviations from TTR have been shown to be associated with increased risk of bleeding and thrombosis as well as increased health care costs.8-10 Mechanistically, patient factors such as differences in vitamin K consumption, comorbid conditions, drug-drug interactions, and medication compliance, as well as genetic differences that impact drug metabolism undoubtedly contribute to the variation of INR noted in patients on warfarin therapy.
Attempts to improve stability have included the administration of low-dose oral vitamin K. However, recent data from a multicenter randomized control trial suggests that while such therapy may help to decrease extreme variations in INR, it does not lead to an increased TTR.11 Furthermore, while significant work has been conducted in identifying specific gene variants, such as CYP2C9 and VKORC, which encode cytochrome P450 and vitamin K epoxide reductase enzymes, respectively, economic analyses suggest that testing for these gene variants would not be cost-effective.12 Additionally, clinical prediction tools, which incorporate important patient factors to help guide anticoagulation explain less than 10% of TTR variability.4
Nonetheless, some caution is warranted in the interpretation of the results reported by Pokorney and his colleagues. The proportion of registry patients treated with warfarin who had a low TTR was much lower than that previously reported by the pivotal U.S. trials of NOACs (55%-68%) and significantly lower than the results of a recent nationwide Swedish registry involving 40,449 patients.13
In the Swedish registry, the mean individual TTR was 70% with more than half the patients having a TTR of 70% or more, emphasizing the importance of health care system effects. Moreover, regardless of whether a patient is on warfarin or a NOAC, patients with a lower TTR have higher rates of diabetes, chronic obstructive pulmonary disease, heart failure, and renal failure, which may contribute to the need for additional therapies that may influence TTR.
For example, INR may be increased by ciprofloxacin or omeprazole when taken with warfarin, and CYP3A4 and P-glycoprotein (P-gp) inducers and inhibitors can result in an increased or decreased anticoagulation effect when used with NOACs. Recent reports have also highlighted variability in the safety of NOACs, particularly among patients with renal or liver insufficiency, African Americans, or patients with a prior history of GI bleeding.14-16 For these subgroups, determining NOAC activity to improve clinical safety of these agents is difficult.
PT or INR testing is largely insensitive or otherwise highly variable and the blood draw time relative to the most recent dose significantly influences the measured level of anti-Xa activity. Importantly, socioeconomic factors and family support systems also influence TTR, as important determinants of access to needed drugs or the ability to sustain related costs over time.
Taken together, prior INR stability on warfarin therapy does not ensure continued stability and, as a consequence, long-term warfarin therapy requires close monitoring in order to remain effective. To this end, further development of point-of-care coagulometers for self-testing and self-management, which have been found to be acceptable and preferred by patients, should be pursued.17 Similarly, attempts to decrease INR variability through research on optimizing computer assisted dosing programs remains warranted.18 NOACs offer an advantage over warfarin therapy in that they have a more predictable pharmacokinetic profile, which precludes the need for routine monitoring of anticoagulation parameters. However, many of the same factors, which influence TTR for warfarin do so for NOACs; NOACs have increased bleeding risk in comparison to warfarin for a number of demographic groups; and the high cost of NOACs may influence patient compliance.
Accordingly, until further data is available, consideration of the conversion of a patient on warfarin with a low TTR to a NOAC should be individualized.
Madhukar S. Patel, MD, is a general surgeon at the Department of Surgery, Massachusetts General Hospital, Boston, and Elliot L. Chaikof, MD, is Surgeon-in-Chief, Beth Israel Deaconess Medical Center, and Chairman, Roberta and Stephen R. Weiner Department of Surgery, Johnson and Johnson Professor of Surgery, Harvard Medical School. Dr. Chaikof is also an associate editor for Vascular Specialist. They have no relevant conflicts.
2. Nat Rev Cardiol. 2014;11:693-703.
5. J Thromb Haemost. 2010;8:2182-91.
6. Thromb Haemost. 2009;101:552-6.
7. Am J Cardiovasc Drugs. 2015;15:205-11.
8. Circ Cardiovasc Qual Outcomes. 2008;1:84-91.
10. J Med Econ. 2015;18:333-40.
11. Thromb Haemost. 2016;116:480-5.
12. Ann Intern Med. 2009;150:73-83.
13. JAMA Cardiol. 2016;1:172-80.
14. N Engl J Med. 2013;369:2093-104.
15. JAMA Intern Med. 2015;175:18-24.
16. J Am Coll Cardiol. 2014;63:891-900.
17. Can Fam Physician. 2011;57:e292-8.
18. J Thromb Haemost. 2008;6:935-43.