Skip main navigation

Patterns of Care for Atrial Fibrillation Before, During, and at Discharge From Hospitalization

From the Get With The Guidelines–Atrial Fibrillation Registry
Originally published Arrhythmia and Electrophysiology. 2021;14



Atrial fibrillation (AF) is the most common arrhythmia encountered in the hospital. However, contemporary treatment of patients hospitalized with AF, including stroke prevention, switching between these therapies, and rhythm control interventions are not well studied. We aimed to examine trends in inpatient interventions for AF, including switching oral anticoagulation (warfarin to direct oral anticoagulants [DOACs]), cardioversion, catheter ablation, and amiodarone use in hospitalized patients with AF.


Using data from the Get With The Guidelines—AFIB registry from the American Heart Association, we analyzed patterns of medication and procedure use among hospitalized patients with AF from January 3, 2013, to March 28, 2017. To identify significant predictors of switching, multivariable hierarchical regression models were developed with patient baseline characteristics and comorbidities.


Among 31 280 patients with AF from 97 participating hospitals, 47.1% were on anticoagulation at presentation (6695 warfarin and 7393 DOAC) and the majority were continued at discharge (91.1%). Of those who were not receiving anticoagulation before hospitalization, 60.6% started anticoagulation at discharge (25.0% warfarin and 75.0% DOAC). The prevalence of switching from warfarin to DOAC was 4.0% and was more likely with younger age and lower CHA2DS2-VASc. Among 28 143 patients (excluding those discharged from the emergency department or observation status), 32.0% underwent cardioversion (56.1% chemically assisted and 49.4% electrical), 6.4% AF ablation, and 1.0% left atrial appendage occlusion device implantation. Patients of White race, younger age, and lower CHA2DS2-VASc were significantly more likely to undergo cardioversion or AF ablation, while older patients with higher CHA2DS2-VASc were significantly more likely to be initiated on amiodarone.


Despite guideline recommendations prioritizing DOAC therapy, there are relatively low rates of switching from warfarin to DOAC in patients hospitalized with AF. Moreover, there is substantial variation in switching and utilization of rhythm control strategies, highlighting opportunities for performance improvement.

Graphic Abstract:

A graphic abstract is available for this article.

What Is Known?

  • Prior analyses have shown relatively low rates of direct oral anticoagulant initiation for stroke prevention in atrial fibrillation, but little is known about switching patterns among patients previously managed with warfarin and patterns of care for other atrial fibrillation-related interventions.

What this Study Adds?

  • Using a nationally representative database, we analyzed care patterns of hospitalized patients with atrial fibrillation.

  • Despite guideline recommendations prioritizing direct oral anticoagulant therapy, there are relatively low rates of switching from warfarin to direct oral anticoagulant in patients hospitalized with atrial fibrillation.

  • Substantial variation and disparities in switching and utilization of rhythm control strategies were observed, highlighting opportunities for ongoing performance improvement.

Hospitalization is a significant cost-driver for atrial fibrillation (AF) health care expenditures. Prior analyses examining patterns of AF care have largely focused on utilization of oral anticoagulation in the outpatient setting.1 Patient registries have provided important insights into contemporary patterns of outpatient AF care and identified opportunities for quality improvement,2,3 but despite accounting for the majority of AF expenditures, relatively little is known about contemporary patterns of inpatient AF care surrounding a hospitalization.

During an acute episode requiring hospitalization, patients with AF would be expected to engage with their health care teams and have consideration of a myriad treatment options. Given its ability to focus attention on the management of AF, the hospital setting provides an important lens into contemporary patterns of AF care. We sought to identify patterns of stroke prevention (anticoagulation and interventional therapies), predictors of switching between therapies, and disparities in care. We additionally examined the prevalence and predictors of undergoing in-hospital procedures including cardioversion, ablation, and newly initiating antiarrhythmic therapy, particularly amiodarone.


Data Source

We used the Get With The Guidelines—AF (GWTG-AFIB) Registry, which began in January 2013 as part of a national quality improvement initiative through the American Heart Association, now in partnership with the Heart Rhythm society. The goal of the GWTG-AFIB initiative is to improve cardiovascular health and outcomes through adherence to guideline-recommended therapies for the treatment of AF and stroke prevention. The GWTG-AFIB program, its component data elements, and the AF performance measures have been described previously.4,5 In this program, participating hospitals submit clinical data regarding medical history, hospital care, and outcomes of patients hospitalized for AF using an online case report form and patient management tool (IQVIA, Cary, NC). This registry includes consecutive patients hospitalized with a primary diagnosis of AF or atrial flutter, as well as some patients with a secondary diagnosis of AF or atrial flutter.

Participating institutions were required to comply with local regulatory and privacy guidelines. Institutions either secured institutional review board approval or were granted a waiver of informed consent under the common rule. IQVIA served as the data-coordinating center and the analytic core was the Duke Clinical Research Institute, which analyzed the aggregate de-identified data. Trained personnel abstracted the data using standardized definitions for data elements, and data edit checks were performed to ensure validity of collected data.

Population for This Study

Data was collected from January 3, 2013, to March 28, 2017. Patients had a primary or secondary diagnosis of AF and were excluded if discharge status was missing or if patients were transferred to another facility, were discharged against medical advice, or discharged to hospice care. The remaining patients were further stratified into the overall cohort and the admitted cohort with the latter group excluding patients discharged from the emergency department or following an observation stay. The overall cohort consisted of 31 280 patients from 97 sites and was used for the analyses examining patterns of anticoagulant therapy. The admitted cohort consisted of 28 143 patients from 95 sites and was used for analyses examining frequency and predictors of procedures, as well as initiation of amiodarone at discharge.

Statistical Analysis

The frequency and predictors of switching from warfarin to a direct oral anticoagulants (DOACs) were collected in the overall cohort. To define switching from warfarin to a DOAC (dabigatran, rivaroxaban, apixaban, or edoxaban), the study population was subset into patients who had been managed with warfarin before admission. Frequencies of patients on a DOAC at discharge were recorded. Patients listed as being on both warfarin and DOAC (at admission or at discharge) were excluded. Within the admitted cohort, frequencies of in-hospital procedures such as AF ablation (including type), cardioversion (chemically assisted, electrical, or TEE guided), and combinations of these procedures were reported.

For the analyses, univariate and multivariate approaches were used to identify factors associated with measures of interest. Baseline demographics, clinical data, and hospital characteristics were described as proportions for categorical variables and medians with 25th and 75th percentiles for continuous variables. Differences in these characteristics were compared using Pearson χ2 or Fisher exact tests (where applicable) for categorical row variables and Wilcoxon Rank-Sum tests for continuous row variables. Associations between patient characteristics and outcomes of interest were analyzed via logistic regression and were reported as odds ratios (ORs) with 95% CIs. Associations are considered statistically significant with a 2-sided α <0.05. Multivariate logistic regression using generalized estimating equations to account for within-hospital clustering and provide robust variance estimations were used to identify factors independently associated with switching. Multiple imputation was used for prior AF procedures of cardioversion, ablation, AF surgery, absence of past procedures, atrial arrhythmia type, labile international normalized ratio (INR; defined as unstable/high INRs or time in therapeutic range <60%), prior major bleeding, or predisposition to reduce missingness in models. All analyses are exploratory/descriptive, and there was not consideration of multiplicity.


Study Cohort

Demographic, clinical, and hospital level characteristics for the overall and admitted cohorts are shown in Tables 1 and 2. For the overall cohort, the mean age was 70.7 (SD, 13.6), 50.6% were male, 83.1% White, and 6.7% Black. Medical comorbidities included hypertension (75.0%), prior stroke (13.0%), diabetes (27.3%), heart failure (10.8%), and renal failure (6.2%). The median CHA2DS2-VASc score was 4.0 (interquartile range, 2.0–5.0) and HAS-BLED score was 2.0 (interquartile range, 2.0–3.0). Hospitals were geographically distributed across the country with Northeast accounting for 40.0% and West 11.1%. Rural locations accounted for 9.0% and academic/teaching hospitals 76.8%.

Table 1. Patient Characteristics

Overall cohort (N=31 280)Admitted cohort (N=28 143)Total on warfarin (N=6636)Switched to DOAC (N=268)Cardioversion (N=8992)First ablation (N=1338)Initiated amiodarone (N=784)
  25th percentile62626662606063
  75th percentile81818280777580
  Female15 448 (49.4%)14 055 (49.9%)3291 (49.6%)122 (45.5%)4141 (46.1%)528 (39.5%)360 (45.9%)
  Male15 832 (50.6%)14 088 (50.1%)3345 (50.4%)146 (54.5%)4851 (54.0%)810 (60.5%)424 (54.1%)
  White25 999 (83.1%)23 179 (82.4%)5679 (85.6%)210 (78.4%)7387 (82.2%)1189 (88.9%)627 (80.0%)
  Black2097 (6.7%)1941 (6.9%)438 (6.6%)31 (11.6%)563 (6.3%)36 (2.7%)65 (8.3%)
  Hispanic1908 (6.1%)1850 (6.6%)311 (4.7%)13 (4.9%)582 (6.5%)53 (4.0%)59 (7.5%)
  Asian337 (1.1%)316 (1.1%)61 (0.9%)5 (1.9%)95 (1.1%)12 (0.9%)11 (1.4%)
  American Indian of Alaska Native75 (0.2%)71 (0.3%)26 (0.4%)2 (0.8%)16 (0.2%)6 (0.5%)3 (0.4%)
  Native Hawaiian or Pacific Islander41 (0.1%)39 (0.1%)11 (0.2%)010 (0.1%)1 (0.1%)4 (0.5%)
  UTD462 (1.5%)446 (1.6%)91 (1.4%)7 (2.6%)150 (1.7%)23 (1.7%)14 (1.8%)
  Missing361 (1.2%)301 (1.1%)19 (0.3%)0189 (2.1%)18 (1.4%)1 (0.1%)
  Medicare5931 (19.0%)5505 (19.6%)1457 (22.0%)47 (17.5%)1572 (17.5%)186 (13.9%)177 (22.6%)
  Private/HMO/other medicare5904 (18.9%)5272 (18.7%)1546 (23.3%)61 (22.8%)1688 (18.8%)212 (15.8%)186 (23.7%)
  Medicaid3019 (9.7%)2738 (9.7%)745 (11.2%)38 (14.2%)777 (8.6%)86 (6.4%)81 (10.3%)
  Private/HMO/other12 903 (41.3%)11 302 (40.2%)2533 (38.2%)98 (36.6%)4225 (47.0%)716 (53.5%)320 (40.8%)
  No insurance/ND/UTD640 (2.1%)561 (2.0%)65 (1.0%)7 (2.6%)215 (2.4%)17 (1.3%)17 (2.2%)
 Medical history
  Heart failure3382 (10.8%)3123 (11.1%)1081 (16.3%)56 (20.9%)1328 (14.8%)216 (16.1%)122 (15.6%)
  Coronary artery disease8731 (27.9%)8058 (28.6%)2413 (36.4%)88 (32.8%)2486 (27.7%)370 (27.7%)241 (30.7%)
  Stroke/TIA4074 (13.0%)3749 (13.3%)1280 (19.3%)47 (17.5%)948 (10.5%)126 (9.4%)92 (11.7%)
  Diabetes8542 (27.3%)7877 (28.0%)2150 (32.4%)89 (33.2%)2363 (26.3%)326 (24.4%)230 (29.3%)
  Dialysis458 (1.5%)436 (1.6%)132 (2.0%)5 (1.9%)111 (1.2%)9 (0.7%)11 (1.4%)
  Hypertension23 471 (75.0%)21 387 (76.0%)5296 (79.8%)223 (83.2%)6765 (75.2%)972 (72.7%)587 (74.9%)
  Liver disease332 (1.1%)314 (1.1%)50 (0.8%)4 (1.5%)84 (0.9%)17 (1.3%)9 (1.2%)
  Obstructive sleep apnea4579 (14.6%)4044 (14.4%)1209 (18.2%)50 (18.7%)1680 (18.7%)337 (25.2%)107 (13.7%)
  Prior hemorrhage804 (2.6%)770 (2.7%)172 (2.6%)4 (1.5%)218 (2.4%)26 (1.9%)19 (2.4%)
  Prior MI2983 (9.5%)2751 (9.8%)800 (12.1%)24 (9.0%)815 (9.1%)128 (9.6%)75 (9.6%)
  Prior PCI3405 (10.9%)3186 (11.3%)840 (12.7%)34 (12.7%)1070 (11.9%)165 (12.3%)111 (14.2%)
  Renal disease1939 (6.2%)1845 (6.6%)551 (8.3%)15 (5.6%)423 (4.7%)62 (4.6%)54 (6.9%)
  Smoker3104 (9.9%)2866 (10.2%)448 (6.8%)24 (9.0%)956 (10.6%)107 (8.0%)87 (11.1%)
  Thyroid disease5652 (18.1%)5123 (18.2%)1352 (20.4%)48 (17.9%)1603 (17.8%)230 (17.2%)132 (16.8%)
 Other risk factors
  Labile INR862 (2.8%)772 (2.7%)711 (10.7%)39 (14.6%)138 (1.5%)22 (1.6%)21 (2.7%)
  Prior major bleeding or predisposition2459 (7.9%)2322 (8.3%)566 (8.5%)17 (6.3%)630 (7.0%)79 (5.9%)59 (7.5%)
 Prior atrial fibrillation procedure
  Prior cardioversion5426 (17.4%)4617 (16.4%)1619 (24.4%)75 (28.0%)2687 (29.9%)501 (37.4%)187 (23.9%)
  Prior ablation2894 (9.3%)2373 (8.4%)883 (13.3%)37 (13.8%)1021 (11.4%)073 (9.3%)
  Prior AF surgery (surgical Maze)234 (0.8%)214 (0.8%)118 (1.8%)1 (0.4%)107 (1.2%)23 (1.7%)6 (0.8%)
  No prior AF procedure20 822 (66.6%)18 994 (67.5%)4152 (62.6%)145 (54.1%)5159 (57.4%)677 (50.6%)548 (69.9%)
 Atrial arrhythmia type
  Permanent/long-standing persistent AF2703 (8.6%)2442 (8.7%)1259 (19.0%)27 (10.1%)313 (3.5%)101 (7.6%)30 (3.8%)
  Persistent AF4664 (14.9%)4192 (14.9%)1425 (21.5%)70 (26.1%)2071 (23.0%)398 (29.8%)128 (16.3%)
  Paroxysmal AF12 954 (41.4%)11 314 (40.2%)2965 (44.7%)125 (46.6%)3838 (42.7%)614 (45.9%)336 (42.9%)
  First detected AF6843 (21.9%)6473 (23.0%)212 (3.2%)6 (2.2%)1963 (21.8%)37 (2.8%)194 (24.7%)
  Unable to determine3947 (12.6%)3581 (12.7%)758 (11.4%)38 (14.2%)784 (8.7%)176 (13.2%)95 (12.1%)
  Missing169 (0.5%)141 (0.5%)17 (0.3%)2 (0.8%)23 (0.3%)12 (0.9%)1 (0.1%)
Risk scores
 CHADS2-VASc score*
  Median (IQR)4.0 (2.0-5.0)4.0 (2.0-5.0)4.0 (3.0-5.0)4.0 (3.0-5.0)3.0 (2.0-5.0)3.0 (2.0-4.0)4.0 (3.0-5.0)
  Mean (SD)3.7 (1.9)3.8 (1.9)4.3 (1.8)4.1 (1.8)3.5 (1.8)3.2 (1.9)3.9 (1.8)
 HAS-BLED score*
  Median (IQR)2.0 (2.0–3.0)2.0 (2.0–3.0)2.0 (2.0–3.0)2.0 (2.0–3.0)2.0 (1.0–3.0)2.0 (1.0–3.0)2.0 (2.0–3.0)
  Mean (SD)2.3 (1.2)2.4 (1.2)2.6 (1.1)2.5 (1.1)2.2 (1.1)2.0 (1.1)2.3 (1.2)

All tests treat the column variable as normal. AF indicates atrial fibrillation; DOAC, direct oral anticoagulant; HMO, Health Maintenance Organization; INR, international normalized ratio; IQR, interquartile range MI, myocardial infarction; ND, no data; PCI, percutaneous coronary intervention; STD, standard deviation; TIA, transient ischemic attack; and UTD, unable to determine.

* P values are based on χ2 rank-based group means score statistics for all continuous/ordinal row variables.

Table 2. Hospital Characteristics

Overall cohort (N=31 280)Admitted cohort (N=28 143)Total on warfarin (N=6636)Switched to DOAC (N=268)Cardioversion (N=8992)First ablation (N=1338)Initiated amiodarone (N=784)
Geographic region
 West3457 (11.1%)3154 (11.2%)813 (12.3%)39 (14.6%)968 (10.8%)179 (13.4%)180 (23.0%)
 South9253 (29.6%)8238 (29.3%)1380 (20.8%)93 (34.7%)3164 (35.2%)335 (25.0%)229 (29.2%)
 Midwest6069 (19.4%)5702 (20.3%)1583 (23.9%)71 (26.5%)2214 (24.6%)285 (21.3%)139 (17.7%)
 Northeast12 501 (10.0%)11 049 (39.3%)2860 (43.1%)65 (24.3%)2646 (29.4%)539 (40.3%)236 (30.1%)
 Academic/teaching hospital24 025 (76.8%)21 355 (75.9%)5173 (78.0%)206 (76.9%)7552 (84.0%)1038 (77.6%)540 (68.9%)
 Rural location2808 (9.0%)2232 (7.9%)1155 (17.4%)6 (2.2%)172 (1.9%)027 (3.4%)
Hospital size (No. of beds)
 500+11 274 (36.0%)10 119 (36.0%)2357 (35.5%)101 (37.7%)4118 (45.8%)741 (55.4%)308 (39.3%)
 400–4993317 (10.6%)2747 (9.8%)646 (9.7%)42 (15.7%)992 (11.0%)117 (8.7%)62 (7.9%)
 300–3993577 (11.4%)3235 (11.5%)562 (8.5%)34 (12.7%)884 (9.8%)115 (8.6%)78 (10.0%)
 200–2994184 (13.4%)3985 (14.2%)703 (10.6%)27 (10.1%)934 (10.4%)145 (10.8%)136 (17.4%)
 100–1995086 (16.3%)4420 (15.7%)1565 (23.6%)28 (10.5%)1156 (12.9%)35 (2.6%)43 (5.5%)
 50–99593 (1.9%)565 (2.0%)84 (1.3%)3 (1.1%)104 (1.2%)022 (2.8%)

Stroke Prevention Patterns of Care and Switching

Among patients in the overall cohort, 47.1%, n=14 088 were on anticoagulation on admission with 52.5% of these patients, n=7393 on DOACs (Table 1). The majority of patients admitted on anticoagulation were continued on anticoagulation at discharge (91.1%, n=12 832). For those who were not receiving anticoagulation before their hospital visit (n=15 851), anticoagulation was started at discharge in 60.6%, n=9602, with warfarin started in 25.0%, n=6695, and DOACs in 75.0%, n=6769.

The prevalence of switching from warfarin to DOACs was 4.0%, n=268 at discharge. Anticoagulant switching patterns are shown in Figure 1. Those who were switched tended to be younger, White, have private insurance, labile INRs, and paroxysmal AF. There were similar patterns of switching among men and women. Smaller hospitals, rural hospitals, and those in the Northeast were less like to switch anticoagulant therapy. In a multivariable analysis shown in Figure 2, switching was significantly less likely with increasing age (per 10-year increase: OR, 0.84 [95% CI, 0.72–0.97]), increasing stroke risk (CHA2DS2-VASc score, per 1-point increase: OR, 0.89 [95% CI, 0.83–0.96]), and rural hospitals (OR, 0.22 [95% CI, 0.09–0.54]). Hospitals in the South (versus Northeast: OR, 2.45 [95% CI, 1.68–3.57]) and larger hospitals 300 to 500 versus under 300 beds (OR, 1.61 [95% CI, 1.08–2.39]) were more likely to switch from warfarin to DOAC. Finally, left atrial appendage occlusion was performed in 1.0%, n=289.

Figure 1.

Figure 1. Patients switched from warfarin to direct oral anticoagulant (DOAC; overall cohort, n=31 280).

Figure 2.

Figure 2. Multivariate hierarchical regression model for predictors of switching from warfarin to direct oral anticoagulant (DOAC). Variables with nonsignificant associations included female gender (odds ratio [OR], 1.01 [95% CI, 0.74–1.36]), White race vs other (OR, 0.83 [95% CI, 0.58–1.18]), labile international normalized ratio (OR, 1.01 [95% CI, 0.75–1.36]), medicaid vs private/other insurance (OR, 1.35 [95% CI, 0.89–2.05]), medicare vs private/other insurance (OR, 1.31 [95% CI, 0.94–1.83]), HAS-BLED, per 1-point increase (OR, 0.92 [95% CI, 0.83–1.03]), region: Midwest vs Northeast (OR, 1.25 [95% CI, 0.83–1.88]), region West vs Northeast (OR, 1.21 [95% CI, 0.75–1.95]), academic/teaching hospital (OR, 1.06 [95% CI, 0.66–1.69]), and hospital size: 500+ vs under 300 beds (OR, 0.93 [95% CI, 0.63–1.39]).

Rhythm Control Interventions

In the admitted cohort, cardioversion occurred in 32.0%, n=8992 of patients and AF ablation in 6.4%, n=1797. Of those who had a cardioversion, 56.1%, n=5049 had a chemically assisted cardioversion and 49.4% had an electrical cardioversion. Other combinations of procedures are shown in Table I in the Data Supplement.


Baseline characteristics of patients who underwent cardioversion are outlined in Table II in the Data Supplement. Patients who underwent cardioversion tended to be younger, male, White, have private insurance, fewer comorbidities, and lower bleeding risk (Table 1). In a multivariable analysis shown in Figure 3A, significant predictors of cardioversion included White race (OR, 1.21 [95% CI, 1.08–1.35]), heart failure (OR, 1.18 [95% CI, 1.06–1.30]), and academic/teaching hospitals (OR, 1.84 [95% CI, 1.11–3.04]). Cardioversion was less likely with older age (per 10-year increase; OR, 0.85 [95% CI, 0.82–0.88]), prior stroke/transient ischemic attack (OR, 0.80 [95% CI, 0.73–0.89]), permanent AF versus first detected AF (OR, 0.47 [95% CI, 0.36–0.60]), increasing stroke risk (for each 1-point increase in CHADS2-VASc score; OR, 0.88 [95% CI, 0.85–0.91]), and increasing bleeding risk (for each 1-point increase in HAS-BLED score; OR, 0.80 [95% CI, 0.77–0.84]).

Figure 3.

Figure 3. The odds ratio (OR) and 95% CI for the association of each variable with cardioversion.A, Multivariate hierarchical regression model for predictors of cardioversion. Variables with nonsignificant associations included female gender (OR, 0.98 [95% CI, 0.92–1.04]), private/other vs no insurance (OR, 1.10 [95% CI, 0.99–1.23]), medicare vs no insurance (OR, 1.07 [95% CI, 0.92–1.25]), CAD (OR, 1.01 [95% CI, 0.94–1.09]), dialysis (OR, 0.87 [95% CI, 0.68–1.12]), hypertension (OR, 1.05 [95% CI, 0.99–1.11]), obstructive sleep apnea (OR, 1.10 [95% CI, 0.99–1.22]), prior hemorrhage (OR, 0.87 [95% CI, 0.74–1.02]), prior myocardial infarction (MI) (OR, 0.88 [95% CI, 0.79–0.98]), prior percutaneous coronary intervention (PCI) (OR, 1.02 [95% CI, 0.91–1.13]), thyroid disease (OR, 1.01 [95% CI, 0.94–1.09]), antiplatelet agent (OR, 0.98 [95% CI, 0.86–1.12]), prior atrial fibrillation (AF) surgery (OR, 0.92 [95% CI, 0.72–1.19]), paroxysmal vs first detected AF (OR, 0.96 [95% CI, 0.84–1.10]), persistent vs first detected AF (OR, 1.22 [95% CI, 0.97–1.53]), labile international normalized ratio (INR) (OR, 1.02 [95% CI, 0.86–1.20]), region: South vs Northeast (OR, 1.51 [95% CI, 0.85–2.66]), region: West vs Northeast (OR, 1.72 [95% CI, 0.78–3.79]), rural location (OR, 0.62 [95% CI, 0.35–1.11]), hospital size: 500+ vs under 300 beds (OR 1.04 [95% CI, 0.58–1.85]), hospital size: 300–500 vs under 300 beds (OR, 1.02 [95% CI, 0.52–2.00]). B, Multivariate hierarchical regression model for predictors of first AF ablation. Variables with nonsignificant associations included female gender (OR, 0.87 [95% CI, 0.74–1.03]), medicare vs no insurance (OR, 0.95 [95% CI, 0.77–1.17]), heart failure (OR, 1.03 [95% CI, 0.83–1.29]), CAD (OR, 0.94 [95% CI, 0.74–1.19]), liver disease (OR, 1.08 [95% CI, 0.78–1.49]), obstructive sleep apnea (OR, 1.49 [95% CI, 1.25–1.78]), prior hemorrhage (OR, 1.03 [95% CI, 0.69–1.54]), prior MI (OR, 0.97 [95% CI, 0.76–1.24]), prior PCI (OR, 1.25 [95% CI, 0.94–1.66]), renal disease (OR, 0.93 [95% CI, 0.69–1.24]), thyroid disease (OR, 0.93 [95% CI, 0.75–1.16]), antiplatelet agent (OR, 0.81 [95% CI, 0.60–1.10]), prior cardioversion (OR, 1.52 [95% CI, 0.63–3.70]), prior AF surgery (OR, 1.44 [95% CI, 0.78–2.66]), no prior AF procedure (OR, 0.80 [95% CI, 0.33–1.95]), labile INR (OR, 1.01 [95% CI, 0.75–1.36]), region: Midwest vs Northeast (OR, 0.86 [95% CI, 0.42–1.76]), region: South vs Northeast (OR, 0.83 [95% CI, 0.38–1.83]), region: West vs Northeast (OR, 2.15 [95% CI, 0.92–5.02]), academic/teaching hospital (OR, 0.55 [95% CI, 0.22–1.35]), hospital size: 300–500 vs under 300 beds (OR, 1.67 [95% CI, 0.77–3.61]). C, Multivariate hierarchical regression model for predictors of amiodarone initiation. Variables with nonsignificant associations included female gender (OR, 0.87 [95% CI, 0.74–1.02]), medicaid vs no insurance (OR, 1.18 [95% CI, 0.91–1.54]), medicare vs no insurance (OR, 0.83 [95% CI, 0.67–1.02]), CAD (OR, 1.01 [95% CI, 0.84–1.22]), stroke/transient ischemic attack (OR, 1.05 [95% 0.82–1.34]), hypertension (OR, 0.90 [95% CI, 0.78–1.04]), prior MI (OR, 0.88 [95% CI, 0.72–1.08]), prior PCI (OR, 1.29 [95% CI, 0.98–1.69]), thyroid disease (OR, 0.96 [95% CI, 0.80–1.14]), antiplatelet agent (OR, 0.81 [95% CI, 0.63–1.14]), prior cardioversion (OR, 0.96 [95% CI, 0.66–1.38]), no prior AF procedure (OR, 1.12 [95% CI, 0.78–1.61]), labile INR (OR, 0.73 [95% CI, 0.45–1.17]), region: Midwest vs Northeast (OR, 0.59 [95% CI, 0.23–1.52]), region: South vs Northeast (OR, 0.51 [95% CI, 0.23–1.13]), region: West vs Northeast (OR, 0.77 [95% CI, 0.38–1.55]), academic/teaching hospital (OR, 1.03 [95% CI, 0.60–1.76]), hospital size: 500+ vs under 300 beds (OR, 0.94 [95% CI, 0.40–2.19]), hospital size: 300–500 vs under 300 beds (OR, 0.75 [95% CI, 0.37–1.51]).

First-Time Ablation of AF

Characteristics of patients who underwent first AF ablation are listed in Table III in the Data Supplement. First AF ablation occurred in 5.3%, n=1338 of patients in the admitted cohort who did not have a prior ablation (Table 1). These patients tended to be younger, male, White, have private insurance, paroxysmal AF, and prior AF procedures (cardioversion and surgical Maze). Academic centers and larger hospitals were more likely to perform first-time ablation, while rural locations were less likely. In a multivariable analysis shown in Figure 3B, factors significantly associated with first AF ablation included White race (versus other; OR, 1.76 [95% CI, 1.44–2.15]), paroxysmal AF versus first detected AF (OR, 5.02 [95% CI, 3.33–7.56]), persistent AF versus first detected AF (OR, 7.81 [95% CI, 5.22–11.68]), permanent AF versus first detected AF (OR, 6.82 [95% CI, 4.59–10.14]), and larger hospital size (500+ beds versus under 300: OR, 4.31 [95% CI, 1.91–9.70]). Factors that were less likely to be associated with first AF ablation were older age (per 10-year increase: OR, 0.88 [95% CI, 0.82–0.96]), prior stroke/transient ischemic attack (OR, 0.76 [95% CI, 0.63–0.92]), increasing stroke risk (per 1-point increase in CHADS2-VASc score: OR, 0.88 [95% CI, 0.82–0.95]), and increasing bleeding risk (per 1-point increase in HAS-BLED: OR, 0.79 [95% CI, 0.71–0.87]).


Characteristics of patients discharged on amiodarone are listed in Table IV in the Data Supplement. Amiodarone was initiated at discharge in 3.8%, n=784 of those not previously on amiodarone in the admitted cohort (Table 1). These patients were more likely to be male and have paroxysmal AF. There was no statistically significant difference in age, race, or insurance type. Patients who had amiodarone newly initiated were more likely to have comorbidities (heart failure, prior percutaneous coronary intervention) and prior cardioversion. Hospital characteristics also varied for those who were started on amiodarone. A multivariable analysis in Figure 3C showed that amiodarone was more likely to be started with increasing age (odds per 10-year increase; OR, 1.30 [95% CI, 1.18–1.43]), heart failure (OR, 1.29 [95% CI, 1.10–1.51]), increasing stroke risk (for each 1-point increase in CHADS2-VASc score; OR, 1.14 [95% CI, 1.10–1.17]), and increasing bleeding risk (for each 1-point increase in HAS-BLED; OR, 1.21 [95% CI, 1.14–1.29]). Amiodarone initiation was less likely to be initiated in White individuals (White versus other race; OR 0.70 [95% CI, 0.57–0.87]), those with private insurance versus those without insurance (OR, 0.77 [95% CI, 0.63–0.95]), paroxysmal AF (OR, 0.46 [95% CI, 0.35–0.60]), persistent AF (OR, 0.48 [95% CI, 0.36–0.64]), and permanent AF (OR, 0.53 [95% CI, 0.34–0.81]) as compared with patients with first detected AF.


In this nationwide assessment of contemporary inpatient care for AF, we identified several important findings with implications for quality improvement. First, only 60% of AF patients not taking anticoagulation at the time of admission where discharged on anticoagulation. Additionally, despite guideline recommendations favoring DOACs, there were low rates of switching from warfarin to DOAC. There were several potential disparities based on race, age, and insurance status, but not sex. Finally, there were wide variations and disparities in the utilization of rhythm control strategies including ablation, cardioversion, and amiodarone initiation among hospitalized patients. This study identifies wide variation in the management of hospitalized patients with AF. A better understanding of factors that drive these decisions is needed to target improvements for better care.

A key finding from our analysis is even among hospitalized patients with AF, where there is likely to be an opportunity to fully evaluate the AF treatment plan and engage in shared decision-making, rates of switching from warfarin to DOACs are quite low, <5%. The importance of switching patients to DOACs is further emphasized in the context of the 2019 update to the American College of Cardiology/American Heart Association AF guidelines, which favors DOACs over warfarin with a class 1A recommendation.6 There are a number of potential explanations for the low rate of switching we observed. First, inpatient providers may be hesitant to change medications that they will not be following longitudinally in the outpatient setting. Along these lines, there may be resistance from patients to change medications without input from their outpatient provider. Increased cost of the DOACs compared with warfarin may be another barrier to transitioning patients at discharge.

More broadly there is therapeutic inertia, the desire to maintain current regimens and care plans in at the absence of a clear therapeutic failure or desire for change. While there were differences in study populations and event rates in the clinical trials comparing DOACs and warfarin with regards to ischemic stroke prevention and gastrointestinal bleeding, the benefits of DOACs are quite clear. These include significant reductions in intracranial hemorrhage, improved mortality, less therapeutic monitoring, and noninferior stroke prevention compared with warfarin.7 Switching appropriate patients from warfarin to DOACs may represent an important area for quality improvement efforts. Moreover, the fact that higher bleeding risk and labile INRs (defined as unstable/high INRs or time in therapeutic range <60%) were not associated with switching and there appeared to be a risk-treatment paradox with patients of lower stroke risk being transitioned to DOACs. Unfortunately, the registry is not ideally suited to capture labile INR as there is no ability to longitudinally capture INR values. The data collection form asks a yes or no question for whether the patient has a labile INR and this may have not have been fully captured by the registry.

This analysis also extends prior literature by examining predictors of switching which have not been well studied. Previous studies in outpatients found that men, older patients, and those with higher stroke risk were more likely to be switched to DOACs.5 Past studies that examined the initiation of DOACs have observed numerous disparities based on sex, lower household income, higher stroke risk, and higher bleeding risk.8–13 Interestingly, our study showed no difference in switching based on sex, but did show that patients were less likely to be transitioned if they had higher stroke risk, but bleeding risk did not impact switching.

Another key finding from our analysis is the substantial variation in the use of cardioversion, ablation, and amiodarone as well as significant disparities based on race. The pursuit of rhythm control strategies in AF has been studied in the longitudinal outpatient setting in both ORBIT-AF (Outcomes Registry for Better Informed Treatment of Atrial Fibrillation) and the PINNACLE registries, showing that approximately one-third of patients are managed with a rhythm control strategy.14,15 Our findings from a hospital-based registry build on and extend these data. The observation that White race was significantly associated with AF ablation is consistent with a body of literature demonstrating racial disparities in cardiovascular procedures and warrants ongoing surveillance and dedicated efforts to close the gap. Interestingly, subgroup data from CABANA (Catheter Ablation Versus Antiarrhythmic Drug Therapy for Atrial Fibrillation Trial) suggested that minorities appeared to have better outcomes with ablation, suggesting greater opportunity for improvement in health care.16 A study among Medicare beneficiaries, showed similar sociodemographic disparities in ablation based on race and gender, however in that study, Black patients were as likely to receive catheter ablation as White patients, but women and Hispanic patients were significantly less likely.17 Notably, we found that patients with heart failure were more likely to undergo cardioversion, but there was not difference in performance of AF ablation. It is possible that the inability to tolerate the arrhythmia in heart failure or suspected tachycardia-induced cardiomyopathy may have driven the propensity for pursuit of rhythm control strategies. Although AF ablation was not significantly different among individuals with heart failure in this analysis, the CASTLE-AF trial (Catheter Ablation Versus Standard Conventional Therapy in Patients With Left Ventricular Dysfunction and Atrial Fibrillation) showed mortality benefit and reduction in heart failure hospitalization in patients with AF with symptomatic heart failure and an ejection fraction ≤35%.18 Additionally, the AATAC trial (Ablation Versus Amiodarone for Treatment of Persistent Atrial Fibrillation in Patients With Congestive Heart Failure and an Implanted Device) showed increased AF free survival, reduction in hospitalizations, and reduction in overall mortality among patients with symptomatic heart failure with reduced ejection fraction and persistent AF who underwent AF ablation.19


There are several limitations of this analysis. First, since the GWTG-AFIB registry contains data that is abstracted from hospitals participating in a quality improvement registry, it is possible that the data may not be representative of other hospitals in the United States, as this program provides additional support to facilitate guideline-recommended treatments. Additionally, as with any clinical management decision, there are often relevant factors that fall outside of information that can be collected for the registry, such as patient preference and unmeasured practice variation. Previous studies have shown practice level variation based on provider type and additional information about providers in the registry could provide insight into treatment seen in the GWTG-AFIB registry. Other factors such as such as patient preference, nonadherence, or time to therapeutic range with warfarin may have affected anticoagulation patterns and switching to DOAC, but would be difficult to collect.

Additionally, the information about specific types of bleeding complications in the hospital were not available in the registry, and this information could provide a more complete understanding of the low number of patients discharged on anticoagulation. Also, because switching patients to a DOAC often requires that patients have an INR <2.0, it is possible that some patients were advised to switch in the outpatient setting after normalization of the INR. Another important question that was not addressed in this analysis was discharge anticoagulation after cardioversion or ablation, which is a specific question addressed by subsequent analyses. Also, continuation of anticoagulation specifically for patients with a CHADS2VASC of 2 or greater would provide important information about patients who per guidelines should be anticoagulated.

For rhythm control strategies such as cardioversion, AF ablation, and amiodarone initiation, patient symptoms are often the driving factor for these modalities and can be subjective. More information about symptom severity would be useful to determine the appropriateness of pursuing these strategies. Specifically, for treatment with amiodarone, additional information about trial previous antiarrhythmic drugs and assessment for left ventricular hypertrophy through echocardiography would be useful for determining appropriateness of amiodarone initiation.

Additionally, our analyses were exploratory/descriptive, and there was not the consideration of multiplicity. For this reason, we used an alpha level of 0.05, which could have overestimated the significance of associations given the large number of patients in the analyses.


This analysis shows that hospitalization greatly impacts treatment of patients with AF. Very few patients were switched from warfarin to DOACs with potential disparities based on race, age, and insurance status. Significant variation in rhythm control strategies including cardioversion, ablation, and amiodarone use exists among hospitalized patients with AF. These findings provide insight into contemporary patterns of care among hospitalized patients with AF.

Nonstandard Abbreviations and Acronyms


atrial fibrillation


direct oral anticoagulant


Get With The Guidelines—AF


international normalized ratio


odds ratio

Supplemental Materials

Data Supplement Tables I–IV

Disclosures Dr Desai works under contract with the Centers for Medicare and Medicaid Services to develop and maintain performance measures used for public reporting and pay for performance programs. He reports research grants and consulting for Amgen, Astra Zeneca, Boehringer Ingelheim, Cytokinetics, MyoKardia, Relypsa, Novartis, and SCPharmaceuticals. Dr Piccini receives grants for clinical research from Abbott, American Heart Association, Association for the Advancement of Medical Instrumentation, Bayer, Boston Scientific, National Heart, Lung, and Blood Institute, and Philips and serves as a consultant to Abbott, Allergan, ARCA Biopharma, Biotronik, Boston Scientific, LivaNova, Medtronic, Milestone, Myokardia, Sanofi, Philips, and Up-to-Date. The other authors report no conflicts.


For Sources of Funding and Disclosures, see page 402.

This article was sent to Andrew E. Epstein, MD, Guest Editor, for review by expert referees, editorial decision, and final disposition.

The Data Supplement is available at

Correspondence to: Nihar R. Desai, MD, MPH, Center for Outcomes Research and Evaluation, Section of Cardiovascular Medicine, Yale-New Haven Hospital, 1 Church St, Suite 200, New Haven, CT 06510. Email


  • 1. Coyne KS, Paramore C, Grandy S, Mercader M, Reynolds M, Zimetbaum P. Assessing the direct costs of treating nonvalvular atrial fibrillation in the United States.Value Health. 2006; 9:348–356. doi: 10.1111/j.1524-4733.2006.00124.xCrossrefMedlineGoogle Scholar
  • 2. Chan PS, Oetgen WJ, Buchanan D, Mitchell K, Fiocchi FF, Tang F, Jones PG, Breeding T, Thrutchley D, Rumsfeld JS, et al. Cardiac performance measure compliance in outpatients: the American College of Cardiology and national cardiovascular data registry’s PINNACLE (practice innovation and clinical excellence) program.J Am Coll Cardiol. 2010; 56:8–14. doi: 10.1016/j.jacc.2010.03.043CrossrefMedlineGoogle Scholar
  • 3. Piccini JP, Fraulo ES, Ansell JE, Fonarow GC, Gersh BJ, Go AS, Hylek EM, Kowey PR, Mahaffey KW, Thomas LE, et al. Outcomes registry for better informed treatment of atrial fibrillation: rationale and design of ORBIT-AF.Am Heart J. 2011; 162:606–612.e1. doi: 10.1016/j.ahj.2011.07.001CrossrefMedlineGoogle Scholar
  • 4. Schwamm LH, Fonarow GC, Reeves MJ, Pan W, Frankel MR, Smith EE, Ellrodt G, Cannon CP, Liang L, Peterson E, et al. Get With the Guidelines-Stroke is associated with sustained improvement in care for patients hospitalized with acute stroke or transient ischemic attack.Circulation. 2009; 119:107–115. doi: 10.1161/CIRCULATIONAHA.108.783688LinkGoogle Scholar
  • 5. Bonow RO, Bennett S, Casey DE, Ganiats TG, Hlatky MA, Konstam MA, Lambrew CT, Normand SL, Piña IL, Radford MJ, et al; American College of Cardiology; American Heart Association Task Force on Performance Measures (Writing Committee to Develop Heart Failure Clinical Performance Measures); Heart Failure Society of America. ACC/AHA clinical performance measures for adults with chronic heart failure: a report of the American College of Cardiology/American Heart Association Task Force on Performance Measures (Writing Committee to Develop Heart Failure Clinical Performance Measures) endorsed by the Heart Failure Society of America.J Am Coll Cardiol. 2005; 46:1144–1178. doi: 10.1016/j.jacc.2005.07.012CrossrefMedlineGoogle Scholar
  • 6. January CT, Wann LS, Calkins H, Chen LY, Cigarroa JE, Cleveland JC, Ellinor PT, Ezekowitz MD, Field ME, Furie KL, et al. 2019 AHA/ACC/HRS Focused Update of the 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a Report of the American College of Cardiology/American Heart Association Task Force on clinical practice guidelines and the heart rhythm society in collaboration with the society of thoracic surgeons.Circulation. 2019; 140:e125–e151. doi: 10.1161/CIR.0000000000000665LinkGoogle Scholar
  • 7. Ruff CT, Giugliano RP, Braunwald E, Hoffman EB, Deenadayalu N, Ezekowitz MD, Camm AJ, Weitz JI, Lewis BS, Parkhomenko A, et al. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials.Lancet. 2014; 383:955–962. doi: 10.1016/S0140-6736(13)62343-0CrossrefMedlineGoogle Scholar
  • 8. Marzec LN, Wang J, Shah ND, Chan PS, Ting HH, Gosch KL, Hsu JC, Maddox TM. Influence of direct oral anticoagulants on rates of oral anticoagulation for atrial fibrillation.J Am Coll Cardiol. 2017; 69:2475–2484. doi: 10.1016/j.jacc.2017.03.540CrossrefMedlineGoogle Scholar
  • 9. Desai NR, Krumme AA, Schneeweiss S, Shrank WH, Brill G, Pezalla EJ, Spettell CM, Brennan TA, Matlin OS, Avorn J, et al. Patterns of initiation of oral anticoagulants in patients with atrial fibrillation- quality and cost implications.Am J Med. 2014; 127:1075–1082.e1. doi: 10.1016/j.amjmed.2014.05.013CrossrefMedlineGoogle Scholar
  • 10. Hsu JC, Chan PS, Tang F, Maddox TM, Marcus GM. Oral anticoagulant prescription in patients with atrial fibrillation and a low risk of thromboembolism: insights from the NCDR PINNACLE registry.JAMA Intern Med. 2015; 175:1062–1065. doi: 10.1001/jamainternmed.2015.0920CrossrefMedlineGoogle Scholar
  • 11. Hsu JC, Chan PS, Tang F, Maddox TM, Marcus GM. Differences in anticoagulant therapy prescription in patients with paroxysmal versus persistent atrial fibrillation.Am J Med. 2015; 128:654.e1–654.e10. doi: 10.1016/j.amjmed.2014.11.035CrossrefMedlineGoogle Scholar
  • 12. Hsu JC, Maddox TM, Kennedy KF, Katz DF, Marzec LN, Lubitz SA, Gehi AK, Turakhia MP, Marcus GM. Oral anticoagulant therapy prescription in patients with atrial fibrillation across the spectrum of stroke risk: insights from the NCDR PINNACLE registry.JAMA Cardiol. 2016; 1:55–62. doi: 10.1001/jamacardio.2015.0374CrossrefMedlineGoogle Scholar
  • 13. Thompson LE, Maddox TM, Lei L, Grunwald GK, Bradley SM, Peterson PN, Masoudi FA, Turchin A, Song Y, Doros G, et al. Sex differences in the use of oral anticoagulants for atrial fibrillation: a report from the national cardiovascular data registry (NCDR®) PINNACLE Registry.JAHA. 2017; 6:e005801. doi: 10.1161/JAHA.117.005801LinkGoogle Scholar
  • 14. Gehi AK, Doros G, Glorioso TJ, Grunwald GK, Hsu J, Song Y, Turakhia MP, Turchin A, Virani SS, Maddox TM. Factors associated with rhythm control treatment decisions in patients with atrial fibrillation-Insights from the NCDR PINNACLE registry.Am Heart J. 2017; 187:88–97. doi: 10.1016/j.ahj.2017.02.006CrossrefMedlineGoogle Scholar
  • 15. Steinberg BA, Holmes DN, Ezekowitz MD, Fonarow GC, Kowey PR, Mahaffey KW, Naccarelli G, Reiffel J, Chang P, Peterson ED, et al. Rate versus rhythm control for management of atrial fibrillation in clinical practice: results from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) registry.Am Heart J. 2013; 165:622–629. doi: 10.1016/j.ahj.2012.12.019CrossrefMedlineGoogle Scholar
  • 16. Packer DL, Mark DB, Robb RA, Monahan KH, Bahnson TD, Poole JE, Noseworthy PA, Rosenberg YD, Jeffries N, Mitchell LB, et al; CABANA Investigators. Effect of catheter ablation vs antiarrhythmic drug therapy on mortality, stroke, bleeding, and cardiac arrest among patients with atrial fibrillation: the CABANA randomized clinical trial.JAMA. 2019; 321:1261–1274. doi: 10.1001/jama.2019.0693CrossrefMedlineGoogle Scholar
  • 17. Bhave PD, Lu X, Girotra S, Kamel H, Vaughan Sarrazin MS. Race- and sex-related differences in care for patients newly diagnosed with atrial fibrillation.Heart Rhythm. 2015; 12:1406–1412. doi: 10.1016/j.hrthm.2015.03.031CrossrefMedlineGoogle Scholar
  • 18. Marrouche NF, Brachmann J, Andresen D, Siebels J, Boersma L, Jordaens L, Merkely B, Pokushalov E, Sanders P, Proff J, et al; CASTLE-AF Investigators. Catheter ablation for atrial fibrillation with heart failure.N Engl J Med. 2018; 378:417–427. doi: 10.1056/NEJMoa1707855CrossrefMedlineGoogle Scholar
  • 19. Di Biase L, Mohanty P, Mohanty S, Santangeli P, Trivedi C, Lakkireddy D, Reddy M, Jais P, Themistoclakis S, Dello Russo A, et al. Ablation versus amiodarone for treatment of persistent atrial fibrillation in patients with congestive heart failure and an implanted device: results from the AATAC multicenter randomized trial.Circulation. 2016; 133:1637–1644. doi: 10.1161/CIRCULATIONAHA.115.019406LinkGoogle Scholar


eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.