Patients and Their Physician’s Perspectives About Oral Anticoagulation in Patients With Atrial Fibrillation Not Receiving an Anticoagulant

“Oral anticoagulation reduces thromboembolic events in patients with nonvalvular atrial fibrillation (AF); however, underuse of anticoagulation is a major issue in treating patients with nonvalvular AF at high stroke risk. Prior data from the American College of Cardiology’s (ACC) Practice Innovation and Clinical Excellence Registry (PINNACLE Registry) has found approximately 40% of patients are not receiving anticoagulants, with little change over time, despite the availability of the nonvitamin K antagonists.

Studies examining the reasons for nonuse are sparse. Because underuse may relate to both physician prescribing and patient factors, some studies have looked at physician assessment of the risk of bleeding vs risk of stroke, finding physicians tend to be risk averse. Only a handful of studies have directly assessed patient wishes, and few if any assessed patients and physicians concurrently. [..]


We established the BOAT-AF (NCT02919982) study as a prospective, multicenter study within the ACC’s National Cardiovascular Data Registry (NCDR) PINNACLE Registry. We identified sites and invited 270 investigators in PINNACLE to join this study. Among those, 25 agreed and 19 received approval from the central institutional review board and enrolled patients.

Patients were identified in the PINNACLE Registry. Patients were included if they were age 18 years or older, had nonvalvular AF and CHA2DS2-VASc score of 2 or more (calculated as congestive heart failure, hypertension, age 75 years and older, diabetes, stroke or transient ischemic attack, vascular disease, age 65 to 74 years, and sex category), were not currently treated with oral anticoagulation according to the PINNACLE Registry data, had a physician office visit within the prior 18 months, and were able and willing to complete the survey. Patients were excluded if they were no longer being followed up with at the local practice, were unable to speak or read English, or declined participation. We also evaluated a cohort of patients from the PINNACLE Registry with nonvalvular AF and a CHA2DS2-VASc score of 2 or more who were aged 18 years or older and were not using oral anticoagulation at index encounter during this same period. Race was collected from the PINNACLE registry fields, which are extracted from the electronic health record as White, Black, other, or missing. [..]

For the appropriateness of the anticoagulation adjudication, a 4-physician panel of cardiologists was established through which the PINNACLE Registry clinical data, physician, and patient surveys were reviewed. Three categories were established similar to other appropriateness7 criteria documents but because no current criteria exist, each patient’s status was adjudicated by the group based on clinical guidelines and practice: appropriate, may be appropriate, and rarely appropriate. The adjudication form and broad criteria for the 3 categories are included in eAppendix 1 in Supplement 1.

[Appropriateness categories and examples of risk profile (all patients in this cohort had CHA2DS2-VASc > 2):

  • Appropriate – patients with AF and no clear contraindication (such as prior GI bleed or intracranial hemorrhage)
  • May be appropriate – patients with a prior episode of AF but longstanding normal sinus rhythm
  • Not appropriate – patients with clear bleeding risk and especially with normal sinus rhythm and lower CHA2DS2-VASc score]


Between January 18, 2017, and May 7, 2018, 817 patients with nonvalvular AF and a CHA2DS2-VASc score of 2 or more who were not currently treated with oral anticoagulation were enrolled from 19 sites within the ACC’s PINNACLE Registry. Data from the PINNACLE Registry on the use of anticoagulation through September 30, 2019, were compared with patients at other centers in the registry. [..]

Just over half of patients reported that they had previously been prescribed oral anticoagulation but had stopped due to bleeding (145 [17.8%]), personal preference (74 [9.1%]), no further AF (74 [9.1%]), or for some other reason (112 [13.7%]). Seventy percent of patients reported that their doctor had discussed oral anticoagulation for their treatment in the past. From the patient surveys, 99 (12.3%) said they were very or extremely worried about the risk of stroke, and 273 (33.8%) were somewhat worried. When asked about bleeding, 238 (29.6%) were very or extremely worried and 214 (26.6%) were somewhat worried. On the physician survey, the top 5 reasons (not mutually exclusive) physicians cited for the patients not receiving anticoagulation were: low AF burden or successful rhythm control therapy (278 [34.0%]), patient refusal (272 [33.3%]), perceived low risk of stroke (206 [25.2%]), fall risk (175 [21.4%]), and high risk of bleeding (167 [20.4%]).

[..]  After rereview, physicians said they would consider prescribing oral anticoagulation in only 221 patients (27.1%) (44 [5.4%] with warfarin and 177 [21.7%] with novel oral anticoagulant [NOAC]). In contrast, 311 patients (38.1%) agreed or strongly agreed to the statement that they would consider taking an oral anticoagulant. In addition, 216 (26.4%) were neutral to taking oral anticoagulants, thus a total of 527 patients (64.5%) were open to consideration of anticoagulation. The majority of these were among patients in whom the physician said they would not prescribe an oral anticoagulant. Interestingly, among 272 patients in whom the physician cited patient refusal as the reason for not prescribing oral anticoagulant, 67 patients (24.6%) indicated on the survey that they would consider taking an oral anticoagulant. Of the 817 patients with AF who were not receiving anticoagulants, in 138 cases (16.9%), patients and their corresponding physicians were agreeable to starting oral anticoagulation. [..]

When looking at their treatment 1 year later, only 119 patients (14.6%) were prescribed oral anticoagulants at follow-up. Of 393 patients who had never been prescribed an oral anticoagulant prior to enrollment, 46 patients (11.7%) received oral anticoagulant at follow-up. Among 424 patients who had received an oral anticoagulant in the past but were not receiving oral anticoagulant at enrollment, 72 (17.0%) were treated with oral anticoagulant at follow-up. Among 583 patients who were receiving aspirin alone at enrollment, 93 (16%) were treated with oral anticoagulant at follow-up. [..]


[..] We found that about 50% of patients expressed fear of stroke and approximately 60% feared the risk of bleeding, but about 65% were open to reconsidering the use of anticoagulation. However, of their physicians, only 27% said they would reconsider oral anticoagulation. Even among patients who were centrally reviewed by a physician panel and considered to be appropriate for anticoagulation, less than half of physicians said they would reconsider anticoagulation. Additionally, there was a disconnect between patient and physician assessments, where many patients who the physicians thought had refused oral anticoagulation were actually open to anticoagulation when asked directly in the survey. We also found that both the patients and their corresponding physicians responded that they would be open to starting oral anticoagulation 16.9% of the time. Our data emphasize the need to revisit any prior decision against oral anticoagulation and to use shared decision-making between patient and physician to arrive at an optimal treatment plan.

[..] All of the patients in this study had a median CHA2DS2-VASc score of 4 (and all had a score of at least 2), and thus guidelines would recommend anticoagulation for these patients, as often adjudicated by the central committee. These data highlight a need for additional education on guideline recommendations. Clinicians often use the duration of episodes of AF in the decision-making process, which has led to more long-term monitoring of AF with 30-day patch monitors or implanted monitoring devices. In general, our data show that the decision to prescribe anticoagulants to a patient with low burden of AF is an ongoing question in many physicians’ minds and where additional data will help shape future recommendations. In particular, clinicians have a significantly heterogeneous practice pattern of prescribing anticoagulants to patients who have undergone successful AF ablation due to a lack of high-quality evidence to guide practice. Ongoing research will hopefully help guide best practices for this patient population. [..]


The findings of this cohort study suggest that among patients with AF who are not anticoagulated, many appear appropriate for anticoagulation, while only a small percentage were actually treated over the next year. We identified a disconnect between the perceptions of physicians and patients and the treatment decision for anticoagulation, which suggests that a fresh discussion of the risks and benefits with shared decision-making may lead to more optimal treatment and better outcomes.”

Full article, CP Cannon, JM Kim, JJ Lee et al. JAMA Network Open, 2023.4.24