Author + information
- aDivision of Cardiology, Department of Internal Medicine, University of California, San Francisco, Fresno, California
- bAdvanced Cardiovascular Imaging Laboratory, Cardiovascular and Pulmonary Branch, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland
- ↵∗Address for correspondence:
Dr. Prakash Deedwania, Division of Cardiology, Department of Internal Medicine, University of California-San Francisco, Fresno, Division of Cardiology Academic Offices, 2335 East Kashian Lane, Suite 460, Fresno, California 93701.
Atrial fibrillation (AF) is the most common arrhythmia encountered in clinical practice. Patients with AF, especially those with multiple comorbid conditions and high CHA2DS2-VASc (Congestive heart failure, Hypertension, Age ≥75 years, Diabetes mellitus, previous Stroke, Vascular disease, Age 65 to 74 years, Sex category) score, are at an increased risk of developing cardioembolic stroke. Vitamin K antagonists (VKA), primarily warfarin, have traditionally been used in these patients to reduce the risk of stroke and subsequent morbidity and mortality (1). However, prescribing physicians and patients using VKA are well aware of its many limitations, such as delayed onset of action, narrow therapeutic window, numerous drug-drug and drug-food interactions, and the need for frequent and regular monitoring. Even in the monitored setting of randomized trials, anticoagulation with VKA has been unreliable. The ROCKET AF (Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation) trial reported the VKA arm to be within the therapeutic range of the international normalized ratio only 55% of the time (2). These multiple impediments have led to significant underutilization of VKA therapy (3), thereby exposing patients to risk of thromboembolic complications.
The availability of novel oral anticoagulants (NOAC) has changed this paradigm. NOAC are especially appealing because they provide reliable and in some instances superior anticoagulation compared with VKA without the need for constant monitoring. It is hoped that due to their significant ease of administration, NOAC will expand the scope of anticoagulation in AF by enrolling those patients that are ineligible, intolerant, or unwilling to use VKA as well as those that are already on VKA therapy but who are undertreated due to fluctuating international normalized ratio levels. Thus, increasing prescription of NOAC should lead to decrease in the number of untreated patients as well as those on VKA.
The GLORIA AF (Global Registry on Long-Term Oral Antithrombotic Treatment in Patients With Atrial Fibrillation) phase 2 registry published in this issue of the Journal gives us some insight regarding the use of NOAC in clinical practice (4). This registry describes oral anticoagulant (OAC) treatment patterns in newly diagnosed AF patients from around the world. The investigators report 79.9% overall compliance with OAC therapy (47.6% on NOAC and 32.3% on VKA) in 15,641 AF patients enrolled between 2011 and 2014. Those with Class I indication for anticoagulation (CHA2DS2-VASc score ≥2) are reported to receive some form of OAC 82.2% of the time. The phase 2 registry compares these trends to the pre-NOAC phase 1 registry and concludes that NOAC have now taken over VKA as the more frequently prescribed OAC in most geographical locations. Given the many advantages of NOAC, these findings are not surprising. With growing confidence among prescribers and patients, the proportion of patients on NOAC should almost certainly increase further.
Perhaps the more important and disturbing finding is that nearly 20% of AF patients are not on appropriate stroke preventive therapy. Even accounting for high bleeding risk in some patients that could potentially prevent OAC prescription (9.1% had a HAS-BLED [Hypertension, Abnormal renal/liver function; Stroke, Bleeding, Labile international normalized ratio, Elderly (age >65 years), previous Drug, alcohol, or medication usage] score ≥ 3), still at least 10% of the eligible patients are either on no therapy or inferior therapies such as aspirin. Some readers that take the “glass half-full” point-of-view may be encouraged by the findings of the GLORIA-AF phase 2 registry and argue that 80% coverage is a marked improvement from what has been reported so far. And herein comes the word of caution in interpreting these results: the reported high percentage of OAC use is not consistent across the board. In fact, unlike GLORIA-AF and EORP-AF (EURObservational Research Programme for Atrial Fibrillation) registries (5), which show 80% OAC usage, other contemporary registries have shown OAC prescription to be more in the 50% range, even after the availability of NOAC. The GARFIELD-AF (Global Anticoagulant Registry in the Field) registry, another ongoing observational study has reported 60.8% OAC use in 17,162 new AF patients from 30 countries (6). The PINNACLE (Practice Innovation and Clinical Excellence) registry reported 44.9% OAC used among 429,417 patients seen in U.S.-based clinical practices (7). Veterans Administration records show a 43.1% OAC (warfarin or dabigatran) use for new onset AF (8). Have the prescription patterns really changed as dramatically as suggested by GLORIA-AF? This appears unlikely. Hospital discharge data from 1.6 million admissions in 812 U.S. hospitals was recently presented at the American Heart Association’s 2016 scientific session (9). It showed a dismal 46% OAC prescription in new onset AF patients, compared with the 78.3% reported from North America in GLORIA-AF phase 2.
These variations suggest important methodological differences between these studies. They also highlight the inherent problems with interpreting data from registries that may not be entirely representative of the general population. GLORIA-AF has important limitations that limit generalizability. Although some of these have been mentioned by Huismann et al. (4), there is need for further elaboration.
First, this is an incident AF registry that enrolled patients with AF within 3 months of initial presentation. This cohort is not representative of most AF patients in the general population or routine clinical practice and as such these findings may not be generalizable for those with long-standing AF. Patients with paroxysmal AF (53%) appear to be over-represented in the GLORIA-AF registry with persistent (35.5%) and permanent AF (11.1%) accounting for a minority. The EORP-AF registry reported somewhat similar proportions of persistent (26%) and permanent AF (17.3%), but a lower percentage of paroxysmal AF (26.5%) (5). However, EORP-AF used the category of “first diagnosed AF” comprising another 30.3% of their total AF patients. The European Society of Cardiology AF guidelines have used this category to describe AF that has not been diagnosed before, irrespective of duration (10). Without this category, which is not part of the American College of Cardiology/American Heart Association AF guidelines (11), most of the newly diagnosed AF patients are lumped together as “paroxysmal” as seen in the GLORIA-AF phase 2 registry. It is important to note that EORP-AF registry enrolled patients with AF within the past year and may also under-represent long-standing AF patients.
Second, GLORIA-AF phase 2 registry was launched at the time of dabigatran approval and enrolled 15× more patients at many more sites than phase 1 did. It is possible that clinical sites participating in the clinical trials were more amenable to prescribing the newly marketed NOAC and participated in disproportionately greater numbers in this registry. Such over-representation, if present, would artificially elevate the number of patients receiving NOAC for stroke prevention.
Third, to describe treatment patterns in a registry, it is paramount to enroll consecutive patients. Consecutive data collection avoids potential selection bias of enrolling patients that receive the desired therapy. Though the participating sites were trained to enroll consecutive patients, they did not use screening logs to ascertain this. Thus, over-representation of patients on desired treatment within a participant site cannot be reliably excluded.
Furthermore, there are some indications within the registry that are concerning. For example, sites from Africa are reported to have an 87.4% compliance rate with OAC prescription (compared with 78.3% in North America) with only 1.5% of eligible patients not on any therapy (7.5% in North America). This incongruently superior compliance with an expensive therapy in a low-resource setting suggests selection bias and requires further explanation.
Also, the direct comparison of OAC therapy trends between phases 1 and 2 should be avoided. There was minimal, if any, overlap between patient populations represented in phase 1 and phase 2 studies. Whereas the current phase 2 registry (4) reports data predominantly from Europe (47.1%) and North America (22.5%), 67.1% of the patients enrolled in the phase 1 cohort were from China (12). In fact, no centers from North America were represented in the phase 1 cohort. Thus, any comparison of OAC use between the 2 phases may represent geographical rather than temporal variation.
Despite the limitations cited, the results from this registry are important as they advance our knowledge regarding the emerging prescription trends of OAC. The findings of this study, however, cannot be applied to populations other than those enrolled in the registry. Let us not be lulled into a false sense of security. More representative data are needed to get realistic estimates of where we stand in terms of OAC use for stroke prevention in AF. Meanwhile, from the information we have, it is apparent that much ground needs to be covered before all eligible patients with AF, especially those at high risk of stroke receive appropriate guideline-directed anticoagulant therapy. Perhaps, in this situation, it is appropriate to remind ourselves that “an ounce of prevention is worth a pound of cure.”
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Both authors have reported that they have no relationships relevant to the contents of this paper to disclose. P.K. Shah, MD, served as Guest Editor for this paper.
- American College of Cardiology Foundation
- Huisman M.V.,
- Rothman K.J.,
- Paquette M.,
- et al.,
- for the GLORIA-AF Investigators
- Lip G.Y.,
- Laroche C.,
- Dan G.A.,
- et al.
- Bassand J.P.,
- Accetta G.,
- Camm A.J.,
- et al.,
- for the GARFIELD-AF Investigators
- Hsu J.C.,
- Maddox T.M.,
- Kennedy K.F.,
- et al.
- Buck J.,
- Kaboli P.,
- Gage B.F.,
- Cram P.,
- Vaughan Sarrazin M.S.
- ↵Priedt R. Many atrial fibrillation patients missing out on blood thinners. HealthDay News, November 16, 2016. Available at: https://consumer.healthday.com/cardiovascular-health-information-20/atrial-fibrillation-959/many-atrial-fibrillation-patients-missing-out-on-blood-thinners-716879.html. Accessed November 17, 2016.
- Kirchhof P.,
- Benussi S.,
- Kotecha D.,
- et al.
- January C.T.,
- Wann L.,
- Alpert J.S.,
- et al.
- Huisman M.V.,
- Ma S.M.,
- Diener H.C.,
- et al.