Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Dec 1.
Published in final edited form as: JACC Clin Electrophysiol. 2019 Oct 2;5(12):1384–1392. doi: 10.1016/j.jacep.2019.07.011

Ischemic and Bleeding Outcomes in Patients With Atrial Fibrillation and Contraindications to Oral Anticoagulation

Benjamin A Steinberg a, Nicholas G Ballew b, Melissa A Greiner b, Steven J Lippmann b, Lesley H Curtis b,c, Emily C O’Brien b,c, Manesh R Patel c,d, Jonathan P Piccini b,c,d
PMCID: PMC6927541  NIHMSID: NIHMS1543608  PMID: 31857036

Abstract

OBJECTIVES

This study sought to describe clinical outcomes among patients with atrial fibrillation (AF) and contraindications to oral anticoagulation (OAC).

BACKGROUND

Treatment with OAC prevents stroke and death in patients with AF, but may be contraindicated among patients at high bleeding risk.

METHODS

This was an observational, longitudinal analysis of a nationally representative 5% Medicare sample of patients with chronic AF and CHA2DS2-VASc (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, prior stroke or transient ischemic attack or thromboembolism, vascular disease, age 65–74 years, sex category) score ≥2. They were stratified by both the presence of high bleeding risk contraindications to OAC and by OAC use. We assessed 3-year ischemic and bleeding outcomes using multivariable Cox proportional hazards models adjusted for relevant patient characteristics.

RESULTS

Among 26,684 AF patients not treated with OAC, 8,283 (31%) had a high bleeding risk contraindication, primarily a blood dyscrasia (75%) or history of gastrointestinal bleeding (40%). Without OAC, patients with contraindications had worse ischemic and bleeding outcomes at 3 years compared with those without contraindications. We also identified 12,454 patients with OAC contraindications who received OAC. Compared with patients not receiving OAC, use of OAC was associated with reduced mortality (adjusted hazard ratio [HR]: 0.79; 95% confidence interval [CI]: 0.76 to0.83), stroke (adjusted HR: 0.90; 95% CI: 0.83 to 0.99), and all-cause hospitalization (adjusted HR: 0.93; 95% CI: 0.90 to 0.96) but increased risk of intracranial hemorrhage (adjusted HR: 1.42; 95% CI: 1.17 to 1.72).

CONCLUSIONS

High bleeding risk contraindications to OAC are common among older patients with AF, and these patients have higher mortality compared with untreated patients without OAC contraindications. The use of OAC in these patients is associated with lower rates of all-cause stroke, hospitalization, and death but higher risk of intracranial hemorrhage.

Keywords: anticoagulation, atrial fibrillation, contraindication, outcomes, Medicare


Oral anticoagulation (OAC) significantly reduces the occurrence of stroke or systemic embolism in patients with atrial fibrillation (AF) and additional risk factors (13). Yet, failure to prescribe OAC in patients with AF who are at risk for stroke is common for a variety of reasons (48). Contraindication to OAC is frequently cited as the reason for withholding therapy, often due to the perceived increased risk of bleeding in these patients (9).

We have previously described the prevalence of OAC contraindications in older U.S. patients with AF (10); however, the subsequent outcomes of these patients, when they are and are not treated with OAC have not been described. Treatment of patients with significant contraindications to OAC is challenging. The risks of stroke and bleeding among patients with significant contraindications to OAC in clinical practice are not well known, particularly for those who receive OAC despite contraindications.

The objectives of the current study were to: 1) describe the prevalence of OAC contraindications and use of OAC among patients 65 years of age and older with AF; 2) compare rates of thromboembolism and bleeding in AF patients with and without contraindications to OAC; and 3) compare outcomes in AF patients with OAC contraindications who do and do not receive OAC therapy.

METHODS

We used standard analytic files from a nationally representative 5% sample of Medicare beneficiaries and corresponding denominator files from the Centers for Medicare and Medicaid Services for 2007 to 2010. These included inpatient institutional claims covered under Medicare Part A and outpatient claims from institutional outpatient providers. We also used noninstitutional provider claims for services covered under Medicare Part B. Denominator files contained beneficiary vital status and demographic data, as well as eligibility and enrollment information.

STUDY POPULATIONS.

We defined a 2008 cohort of Medicare beneficiaries with prevalent AF not due to a reversible cause, based on claims-based diagnoses in 2007. To establish a chronic, prevalent diagnosis of AF, we required at least 2 diagnoses of AF (International Classification of Diseases-9th Revision-Clinical Modification [ICD-9-CM] code 427.31) found in any position on separate inpatient or outpatient claims ≥6 months apart. At least 1 of the claims must have been as an outpatient. Additionally, we required that beneficiaries were 65 years of age or older, living in the United States on January 1, 2008, and had continuous enrollment in Medicare fee for service in the prior colander year. This is consistent with prior, similar analyses of AF patients in Medicare datasets (1113).

The present analysis was further limited to patients with CHA2DS2-VASc (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, prior stroke or transient ischemic attack or thromboembolism, vascular disease, age 65–74 years, sex category) score ≥2 (calculated from claims between January 1 and December 31, 2007).

OAC USE.

We defined OAC use as the use of warfarin based on ≥2 of the following: prothrombin time laboratory test (Current Procedural Terminology code 85610) or home international normalized ratio monitoring instruction, equipment, or interpretation of results (Healthcare Common Procedure Coding System codes G0248, G0249, G0250) claims in any carrier or outpatient facility claims found within 90 days before the index diagnosis (January 1, 2008). The study period predated the use of direct oral anticoagulants (DOACs).

The presence of OAC contraindication was defined by the presence of any 1 or more of the following 5 conditions, as described previously (10): 1) intracranial mass; 2) severe, chronic blood dyscrasia (including thrombocytopenias, anemias, hemoglobinopathies, and hematological and lymphatic malignancies); 3) severe or major gastrointestinal (GI) bleeding; 4) intracranial hemorrhage (ICH) (traumatic and nontraumatic); and 5) end-stage liver disease. The full listing of ICD-9-CM codes used to define the study population, contraindication status, and outcomes can be found in Online Table 1 (14).

PATIENT CHARACTERISTICS.

Patient demographics (age, sex, race, and state of residence) were derived from the denominator files; comorbid conditions were defined using validated coding algorithms (15,16) for dementia, diabetes, ischemic heart disease, peripheral vascular disease, congestive heart failure, cerebrovascular disease, hypertension, chronic obstructive pulmonary disease, renal disease, stroke or transient ischemic attack, cancer, and valvular heart disease. The presence of a pacemaker or implantable cardioverter-defibrillator device was identified on the basis of Current Procedural Terminology and ICD-9-CM codes for device in situ, implantation, revision, or monitoring.

STATISTICAL ANALYSIS.

Primary analysis: contraindications and outcomes without OAC.

We initially described the prevalence of contraindications using frequencies and proportions among patients not receiving OAC. Baseline characteristics were presented for categorical variables as frequencies and continuous variables as mean ± SD. We tested for differences between contraindication groups (presence or absence) using chi-square tests for categorical variables and Wilcoxon rank sum tests for continuous variables.

Among patients not receiving OAC, we described 3-year event rates between those with and without any OAC contraindication, including transient ischemic attack, stroke (ischemic and hemorrhagic), ICH, severe GI bleeding that required a blood transfusion, all-cause hospitalization, and all-cause mortality. For mortality, we calculated incidence based on Kaplan-Meier estimates. We tested for mortality differences between groups using log-rank tests. For the other endpoints, we calculated cumulative incidence estimates. The cumulative incidence function accounts for the competing risk of mortality, which is high in this population (17,18). We tested for differences between groups on these other outcomes using Gray’s tests. For all survival analyses, we censored data for patients if they enrolled in Medicare managed care(i.e., fee-for-service end date), at the end of claims follow-up data (December 31, 2010) and at the time of death for outcomes other than mortality.

To measure the association between OAC contraindication and clinical outcomes, we used multivariable Cox proportional hazards regression models. Adjustment variables included all of the baseline demographic information and comorbid conditions described previously (the groups were significantly different for each background characteristic we investigated). The primary analysis was performed on the AF and no OAC use population in which OAC status is statically assessed as of the index date January 1, 2008 (see previous algorithm).

Secondary analyses: OAC use and contraindications.

In secondary analyses, we identified patients with OAC contraindications who nonetheless were receiving OAC. Baseline characteristics were stratified by OAC use and compared using chi-square tests for categorical variables and Wilcoxon rank sum tests for continuous variables.

To assess the association between OAC use and clinical outcomes among these patients with OAC contraindications, we used multivariable Cox proportional hazards regression models. In addition to demographics and baseline comorbidities, contraindication characteristics that significantly differed between groups (number of contraindications, blood dyscrasia, GI bleed, and ICH frequency) were also adjusted for in the models. The analysis was performed on the contraindicated AF population in which OAC status is statically assessed as of the index date January 1, 2008 (see previous algorithm).

Sensitivity analyses.

We performed multiple sensitivity analyses for both the primary and secondary analyses. The first sensitivity analyses allowed for changes in OAC status during follow-up: OAC status was recorded for every 90-day period following the index date (January 1, 2008). If a patient experienced a change in OAC status (e.g. crossover to the opposite exposure group), they were censored from the analysis. This analysis was a sensitivity analysis (as opposed to the primary) because it had the potential to produce biased results. Patients that start or stop using an OAC are likely to have a different patient profile (including OAC contraindication status) than do those that do not start to use an OAC and the differences are likely to relate to study outcomes. We also repeated the original analysis, without censoring, using an alternative definition of OAC use–based on Part D claims for warfarin use instead of international normalized ratio measurements.

Additional sensitivity analyses were performed in patients not on OAC whose only contraindication was ICH. This analysis was performed to address what many clinicians consider to be the strongest contraindication to OAC (ICH). Last, to assess for residual confounding due to healthy user bias, we performed a sensitivity analysis between OAC treatment and a negative control outcome. In this comparison, hospitalization for nutritional and miscellaneous metabolic disorders (Diagnosis Related Group 640 or 641) served as the negative control.

We used SAS software version 9.4 (SAS Institute, Cary, North Carolina) and R software version 3.53 (R Foundation for Statistical Computing, Vienna, Austria) for all analyses. We chose a 2-tailed alpha threshold of 0.05 for statistical significance. The Institutional Review Board of the Duke University Health System approved the study. This work was supported by a grant from Boston Scientific. The authors are solely responsible for the design and conduct of this study, all study analyses, and the drafting and editing of the paper and its final contents.

RESULTS

The primary study population consisted of 26,684 AF patients with CHA2DS2-VASc score ≥2 and who were not receiving OAC. Of these patients, 8,283 (31%) had at least 1 major contraindication to OAC (Table 1). The most common OAC contraindication was the presence of a blood dyscrasia (n = 6,191, 75%); however, 2,060 patients (25% of all patients with contraindications) had 2 or more OAC contraindications (Online Table 2). Patients with OAC contraindications were marginally older (mean age 81.9 years vs. 81.1 years; p < 0.0001), were more likely to be women (62% vs. 60%; p = .01), and had significantly higher mean CHA2DS2-VASc scores (5.7 vs. 4.9; p < 0.0001).

TABLE 1.

Characteristics of Patients With AF Who Were Not on OAC Medication, Stratified by Eligibility for OAC

No Contraindication (n = 18,401) OAC Contraindication (n = 8,283) p Value
Age, yrs 81.1 ± 7.9 81.9 ± 7.5 <0.001
Age groups <0.001
 65–69 years of age 1,644 (8.9) 540 (6.5)
 70–74 years of age 2,503 (13.6) 915 (11.0)
 75–79 years of age 3,547 (19.3) 1,570 (19.0)
 80+ years of age 10,707 (58.2) 5,258 (63.5)
Female 11,103 (60.3) 5,134 (62.0) 0.01
Race/ethnicity <0.001
 White 17,197 (93.5) 7,631 (92.1)
 Black 643 (3.5) 402 (4.9)
 Other 561 (3.0) 250 (3.0)
Region <0.001
 Northeast 3,795 (20.6) 1,945 (23.5)
 South 7,208 (39.2) 3,265 (39.4)
 Midwest 4,322 (23.5) 1,878 (22.7)
 West 3,076 (16.7) 1,195 (14.4)
Contraindications
 1 0(0) 6,223 (75.1)
 2 0(0) 1,958 (23.6)
 3+ 0(0) 102 (1.2)
Contraindication type
 End-stage liver disease 0(0) 65 (0.8)
 Blood dyscrasia 0(0) 6,191 (74.7)
 Intracranial mass 0(0) 237 (2.9)
 GI bleed 0(0) 3,314 (40.0)
 ICH 0(0) 647 (7.8)
Comorbid conditions
 Atrial flutter 1,310 (7.1) 833 (10.1) <0.001
 Dementia 2,122 (11.5) 1,440 (17.4) <0.001
 Diabetes mellitus 5,759 (31.3) 3,422 (41.3) <0.001
 Ischemic heart disease 9,674 (52.6) 5,463 (66.0) <0.001
 Peripheral vascular disease 5,235 (28.4) 3,395 (41.0) <0.001
 Congestive heart failure 7,378 (40.1) 4,930 (59.5) <0.001
 Cerebrovascular disease 4,594 (25.0) 3,343 (40.4) <0.001
 Hypertension 15,941 (86.6) 7,628 (92.1) <0.001
 COPD 5,866 (31.9) 3,871 (46.7) <0.001
 Renal disease 2,369 (12.9) 2,533 (30.6) <0.001
 Stroke/TIA 3,310 (18.0) 2,501 (30.2) <0.001
 Cancer 2,424 (13.2) 2,082 (25.1) <0.001
 Valvular heart disease 6,164 (33.5) 3,756 (45.3) <0.001
 CHA2DS2-VASc score 4.9 ± 1.6 5.7 ± 1.6 <0.001
  2 892 (4.8) 162 (2.0) <0.001
  3 2,526 (13.7) 502 (6.1) <0.001
  >4 14,983 (81.4) 7,619 (92.0) <0.001
 CHADS2 score 2.7 ± 1.3 3.4 ± 1.4 <0.001
  0 290 (1.6) 52 (0.6) <0.001
  1 2,707 (14.7) 550 (6.6) <0.001
  >2 15,404 (83.7) 7,681 (92.7) <0.001
Devices in place
 ICD 570 (3.1) 336 (4.1) <0.001
 Pacemaker 3,231 (17.6) 1,752 (21.2) <0.001

Values are mean ± SD or n (%).

AF = atrial fibrillation; CHA2DS2 = congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, prior stroke or transient ischemic attack or thromboembolism; CHA2DS2-VASc = congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, prior stroke or transient ischemic attack or thromboembolism, vascular disease, age 65–74 years, sex category; COPD = chronic obstructive pulmonary disease; GI = gastrointestinal; ICD = implantable cardioverter-defibrillator; ICH = intracranial hemorrhage; OAC = oral anticoagulation; TIA = transient ischemic attack.

Cumulative event rates among patients with and without OAC contraindications are shown in the Central Illustration, Parts A and B. Raw and adjusted event rates are shown in Table 2. The presence of an OAC contraindication was associated with a significantly higher 3-year adjusted hazard ratio (HR) for death (1.17; 95% confidence interval [CI]: 1.12 to 1.22), GI bleed (adjusted HR: 2.31; 95% CI: 1.64 to 3.24), and all-cause hospitalization (adjusted HR: 1.20, 95% 1.16 to 1.24).

CENTRAL ILLUSTRATION. Outomes Among for Atrial Fibrillation Patients by OAC Contraindications and OAC Use.

CENTRAL ILLUSTRATION

(A and B) Unadjusted cumulative 3-year event rates for patients with atrial fibrillation who were not on oral anticoagulation (OAC) medication. Event rates are presented by contraindication status to OAC. For mortality, we calculated the event rate as the Kaplan-Meier estimator. For all other outcomes, we calculated the event rate using the cumulative incidence function, which accounts for competing risk of death. The vertical dashed gray lines note 1 and 2 years from the index date (January 1, 2008). (C) Adjusted hazard ratios for the associations between OAC use and outcomes, among only AF patients with at least 1 OAC contraindication. GI = gastrointestinal; HR = hazard ratio; ICH = intracranial hemorrhage; OAC = oral anticoagulation; TIA = transient ischemic attack.

TABLE 2.

Oral Anticoagulation Contraindication Status and Outcomes Over 3 Years in Patients With Prevalent AF Who Were Not on OAC Medication (n = 26,684)

Event Rate Percentage at 3 Yrs* Unadjusted Adjusted
Outcome No Contraindication (n = 18,401) Contraindication (n = 8,283) p Value HR (95% CI) HR (95% CI) p Value
Death 34.6 49.6 <0.001 1.64 (1.57–1.70) 1.17 (1.12–1.22) <0.001
Stroke-broad 10.1 11.3 0.003 1.28 (1.18–1.39) 1.01 (0.93–1.10) 0.83
Ischemic stroke 5.64 6.04 0.195 1.22 (1.09–1.36) 0.98 (0.87–1.10) 0.70
TIA 2.28 2.74 0.025 1.38 (1.17–1.62) 1.08 (0.91–1.29) 0.37
ICH 1.82 2.03 0.227 1.29 (1.06–1.55) 1.08 (0.89–1.32) 0.44
GI bleed-transfusion 0.42 0.99 <0.001 2.70 (1.96–3.72) 2.31 (1.64–3.24) <0.001
Hospitalization 63.1 74.1 <0.001 1.53 (1.48–1.58) 1.20 (1.16–1.24) <0.001
*

The death event rate p value is from the log-rank test. All other outcome event rate p values are from Gray’s tests.

Adjusted for age, race, gender, region, implantable device (ICD and cardiac resynchronization therapy with defibrillator), and comorbidities (atrial flutter, dementia, diabetes, coronary heart disease, peripheral vascular disease, congestive heart failure, cardiovascular disease, hypertension, COPD, renal disease, prior stroke, cancer, and valvular heart disease).

Required a blood transfusion.

CI = confidence interval; HR = hazard ratio; other abbreviations as in Table 1.

ANTICOAGULATION USE.

In the secondary analysis among only patients with OAC contraindications, baseline characteristics according to the presence or absence of OAC therapy are shown in Table 3. Those receiving OAC were younger (mean age 80.0 years vs.81.9 years; p < 0.001), were less likely to be women (57% vs. 62%; p < 0.001), had lower mean CHA2DS2- VASc scores (5.5 vs. 5.7; p < 0.001), and less commonly had prior GI bleeding (37.5% vs. 40.0%; p < 0.001) or prior ICH (3.9% vs. 7.8%; p < 0.001).

TABLE 3.

Patient Baseline Characteristics of the AF and Contraindicated to OAC Cohorts by OAC Status

No OAC Use (n = 8,283) OAC Use (n = 12,454) p Value
Age, yrs 81.9 ± 7.5 80.0 ± 6.7 <0.001
Age group <0.001
 65–69 years of age 540 (6.5) 900 (7.2)
 70–74 years of age 915 (11.0) 1,852 (14.9)
 75–79 years of age 1,570 (19.0) 2,936 (23.6)
 80+ years of age 5,258 (63.5) 6,766 (54.3)
Female 5,134 (62.0) 7,140 (57.3) <0.001
Race/ethnicity <0.001
 White 7,631 (92.1) 11,760 (94.4)
 Black 402 (4.9) 428 (3.4)
 Other 250 (3.0) 266 (2.1)
Region <0.001
 Northeast 1,945 (23.5) 3,414 (27.4)
 South 3,265 (39.4) 4,363 (35.0)
 Midwest 1,878 (22.7) 2,966 (23.8)
 West 1,195 (14.4) 1,711 (13.7)
Contraindications <0.001
 1 6,223 (75.1) 10,278 (82.5)
 2 1,958 (23.6) 2,094 (16.8)
 3+ 102 (1.2) 82 (0.7)
Contraindication type
 End-stage liver disease 65 (0.8) 86 (0.7) .43
 Blood dyscrasia 6,191 (74.7) 9,148 (73.5) .04
 Intracranial mass 237 (2.9) 320 (2.6) .20
 GI bleed 3,314 (40.0) 4,673 (37.5) <0.001
 ICH 647 (7.8) 489 (3.9) <0.001
Comorbid conditions
 Atrial flutter 833 (10.1) 1,455 (11.7) <0.001
 Dementia 1,440 (17.4) 1,102 (8.8) <0.001
 Diabetes mellitus 3,422 (41.3) 5,203 (41.8) .51
 Ischemic heart disease 5,463 (66.0) 8,143 (65.4) .40
 Peripheral vascular disease 3,395 (41.0) 4,550 (36.5) <0.001
 Congestive heart failure 4,930 (59.5) 6,953 (55.8) <0.001
 Cerebrovascular disease 3,343 (40.4) 4,491 (36.1) <0.001
 Hypertension 7,628 (92.1) 11,370 (91.3) .04
 COPD 3,871 (46.7) 5,291 (42.5) <0.001
 Renal disease 2,533 (30.6) 3,186 (25.6) <0.001
 Stroke/TIA 2,501 (30.2) 3,223 (25.9) <0.001
 Cancer 2,082 (25.1) 3,384 (27.2) .001
 Valvular heart disease 3,756 (45.3) 6,153 (49.4) <0.001
 CHA2DS2-VASc score 5.7 ± 1.6 5.5 ± 1.6 <0.001
  2 162 (2.0) 288 (2.3) .08
  3 502 (6.1) 934 (7.5) <0.001
  ≥4 7,619 (92.0) 11,232 (90.2) <0.001
 CHADS2 score 3.4 ± 1.4 3.2 ± 1.3 <0.001
  0 52 (0.6) 77 (0.6) .93
  1 550 (6.6) 1,002 (8.0) <0.001
  ≥2 7,681 (92.7) 11,375 (91.3) <0.001
Devices in place
 ICD 336 (4.1) 753 (6.0) <0.001
 Pacemaker 1,752 (21.2) 2,947 (23.7) <0.001

Values are mean ± SD or n (%).

Abbreviations as in Table 1.

Unadjusted and adjusted outcomes comparing OAC use versus no OAC use, among patients with OAC contraindications, are shown in Table 4 (see also Central Illustration, Part C). Use of OAC among these patients was significantly associated with a lower risk of all-cause death (adjusted HR: 0.79; 95% CI: 0.76 to0.83), any stroke (adjusted HR: 0.90; 95% CI: 0.83 to0.99), and all-cause hospitalization (adjusted HR:0.93; 95% CI: 0.90 to 0.96), but a higher risk of ICH (adjusted HR: 1.42; 95% CI: 1.17 to 1.72).

TABLE 4.

OAC Use and Outcomes Over 3 Years in Patients With Prevalent AF and a Contraindication to OAC Use (n = 20,737)

Event Rate Percentage at 3 Years* Unadjusted Adjusted
No OAC (n = 8,283) OAC (n = 12,454) p Value HR (95% CI) HR (95% CI) p Value
Death 49.6 36.6 <0.001 0.65 (0.63–0.68) 0.79 (0.76–0.83) <0.001
Stroke-broad 11.3 10.4 0.03 0.81 (0.74–0.88) 0.90 (0.83–0.99) 0.03
Ischemic stroke 6.04 5.18 0.01 0.76 (0.67–0.85) 0.87 (0.77–0.98) 0.03
TIA 2.74 2.53 0.36 0.82 (0.69–0.98) 0.90 (0.75–1.07) 0.23
ICH 2.03 2.84 <0.001 1.26 (1.04–1.52) 1.42 (1.17–1.72) <0.001
GI bleed-transfusion 0.99 1.33 0.03 1.20 (0.92–1.58) 1.29 (0.98–1.70) 0.08
Hospitalization 74.1 73.5 <0.001 0.87 (0.84–0.89) 0.93 (0.90–0.96) <0.001
*

The death event rate p value is from the log-rank test. All other outcome event rate p values are from Gray’s tests.

Adjusted for age, race, gender, region, implantable device (ICD and cardiac resynchronization therapy with defibrillator), contraindication characteristics (number of contraindications, blood dyscrasia, ICH, and GI bleed), and comorbidities (atrial flutter, dementia, diabetes, coronary heart disease, peripheral vascular disease, congestive heart failure, cardiovascular disease, hypertension, COPD, renal disease, prior stroke, cancer, and valvular heart disease).

Required a blood transfusion.

Abbreviations as in Tables 1 and 2.

SENSITIVITY ANALYSES.

Results of all sensitivity analyses are shown in Online Tables 3 to 6. They were consistently similar to the overall results reported previously. The use of alternative definition for OAC use (Part D claims for warfarin vs. international normalized ratio testing) did not significantly alter the results. In the negative control sensitivity analysis, we found no significant association between OAC use and readmission for nutritional or metabolic disorders, which suggests adequate control for confounding variables in the OAC comparison models. When classification of contraindication was limited to only ICH (Online Table 4), results were qualitatively similar (though reflected the lower power with wider CIs).

DISCUSSION

There are several findings from our nationwide analysis, which may have important implications for the management of high-risk, older patients with AF and significant contraindications to OAC. First, there is a high prevalence of significant contraindications to OAC, with nearly one-third of untreated patients having ≥1 comorbid illness that could represent a contraindication to OAC (and nearly 8% overall with ≥2 such conditions). Second, patients with OAC contraindications are at increased risk of bleeding, hospitalization, and death compared with patients who do not have a contraindication (but are not treated with OAC). Finally, and perhaps most importantly, among patients with an OAC contraindication, the use of OAC was associated with lower rates of all-cause stroke, hospitalization, and death at the expense of a 42% relative increased risk of ICH. Understanding the trade-offs associated with OAC use in this population is important to optimize pharmacologic and nonpharmacologic treatment decisions and outcomes in these high-risk patients.

Risk factors for bleeding are a well-known and primary barrier to effective utilization of stroke prevention therapy in patients with AF, particularly among older individuals (19). Although several studies have attempted to define and measure the prevalence of contraindications to OAC among patients with AF, they remain highly subjective across clinicians and patients (20). Although there is agreement that several factors can increase an individual’s risk of bleeding while taking OAC (21,22), the challenge has been identifying those patients in which the associated risk of significant bleeding with OAC outweighs the benefits of stroke prevention. Some analyses have focused on the underlying bleeding risk of a population, based on bleeding scores, and demonstrated a net clinical benefit of OAC even among patients with very high bleeding risk (23). This is likely due at least in part to the fact that many of the same factors that increase bleeding risk are also associated with increased stroke risk. Furthermore, though those patients had increased risk of bleeding as defined by a bleeding score, few of those factors are commonly regarded as absolute contraindications to anticoagulation. Given such challenges in assessing the net clinical benefit of OAC in these patients, guidelines still do not recommend routine use of a bleeding risk score to determine when to withhold OAC.

In this study, we sought to analyze specific comorbidities that are often perceived by providers to indicate not only a higher risk of bleeding, but also a sufficiently high risk of bleeding that in many cases the patient may be labeled as indefinitely ineligible for OAC therapy. The presence of hematologic dyscrasias and intracranial bleeding are 2 such conditions. Yet our results demonstrate that even these conditions should be considered less stringently–they are risk factors for bleeding, which may or may not tip the scale of net clinical benefit in favor of withholding OAC. Importantly, many of these patients may instead be treated with aspirin for stroke prevention, despite evidence that it is inferior to OAC for preventing stroke and may not be of significantly lower bleeding risk compared with OAC (24).

Although risk of ICH was higher among patients with contraindications who were treated with OAC, rates of death, hospitalization, and stroke were lower than their untreated counterparts. At the population level, it appears that the higher risk of bleeding events may be outweighed by favorable stroke and mortality risk reduction. Nevertheless, there still may be individual patients for whom the risk of a fatal hemorrhagic complication truly is prohibitively high–identifying those patients is an important goal and may include those with a persistent or irreversible, high risk of ICH (e.g., intracranial mass, amyloid angiopathy). Such patients may be candidates for alternative approaches to stroke prevention in AF.

The majority of data on DOACs suggest that these agents have lower risks of the most serious bleeding, ICH, compared with warfarin (25). However, for some DOACs, there may be an increased risk of GI bleeding, compared with warfarin (26). Thus, for some patients at very high risk of bleeding, DOACs may be appropriate, and for others, the incremental risk of bleeding, compared with warfarin, may be negligible. Alternatively, patients who are at the highest risk of bleeding with long-term OAC may have the most to gain from nonpharmacological stroke prevention strategies in AF. Left atrial appendage occlusion (LAAO) has become an attractive alternative for such patients. However, the current data supporting the only Food and Drug Administration–approved LAAO device have been largely limited to patients tolerant of long-term OAC (27). This is due in part to the fact that intra-procedural parenteral anticoagulation and short-term postoperative OAC are recommended with the use of that device, and emerging data have raised the awareness for thrombotic complications of LAAO devices (28). Additional data are needed regarding the net clinical benefit and long-term outcomes of LAAO among patients with AF and high risks of both stroke and bleeding, who cannot tolerate any OAC. Studies such as the ASAP-TOO (Assessment of the Watchman Device in Patients Unsuitable for Oral Anticoagulation) trial will provide valuable additional data (29).

STUDY LIMITATIONS.

Several limitations should be acknowledged when considering these data. First, they are observational, and include only patients 65 years of age and older. Second, defining major contraindications can be challenging and is limited by lack of data on severity and a finite lookback period of 1 year. Furthermore, the definition of OAC contraindication in of itself may be subjective and often requires nuanced risk-benefit calculation in individual patients. Nonetheless, we attempted to include a set of factors that are generally considered higher risk. Although our data predate the availability of DOACs, several sensitivity analyses of only patients with Part D data were consistent with the topline results. Last, these are observational data, and despite multivariable adjustment, residual or unmeasured confounding may remain–patients with contraindications who received OAC were significantly different from those who did not receive OAC. Although incomplete adjustment, as previous, may influence the quantitative results, all the event rates are pathophysio-logically consistent and of the expected effect size, suggesting a likely true association.

CONCLUSIONS

High bleeding risk contraindications to OAC are common among older patients with AF and increased stroke risk, and these patients have higher mortality compared with untreated patients without OAC contraindications. However, the use of OAC is associated with lower rates of all-cause stroke, hospitalization, and death but higher risk of ICH. Future research should identify subgroups in which the net clinical benefit favors OAC in these high bleeding risk patients versus other potential treatment modalities such as LAAO.

Supplementary Material

1

PERSPECTIVES.

COMPETENCY IN MEDICAL KNOWLEDGE:

Patients with contraindications to OAC are at high risk of adverse events. However, the risk of OAC-associated ICH in these patients must be weighed against the lower rates of stroke, hospitalization, and mortality observed in these patients.

TRANSLATIONAL OUTLOOK:

AF-related stroke is primarily driven by cardioembolism, and OAC reduces this risk at the expense of increased bleeding risk. Among patients at very high bleeding risk, the net clinical benefit of nonpharmacological stroke prevention strategies may be preferable to OAC. However, further studies are needed to assess the optimal balance.

Acknowledgments

This work was supported by a grant from Boston Scientific. Dr. Steinberg was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health under Award Number K23HL143156. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Dr. Steinberg has received research support from Boston Scientific and Janssen; and served as a consultant for Biosense Webster, Janssen, and Merit Medical. Dr. Curtis has received grants from GlaxoSmithKline, Novartis, Boston Scientific, and St Jude. O’Brien has received grants from Novartis, Bristol-Myers Squibb, GlaxoSmithKline, Sanofi, and Janssen. Dr. Patel has received grant support from Novartis; and has served on the advisory board, as a consultant, and as a speaker for Bayer Pharmaceuticals, AstraZeneca, and Janssen. Dr. Piccini has received funding for clinical research from Abbott Medical, ARCA biopharma, Boston Scientific, Gilead, Janssen Pharmaceuticals, and Verily; and served as a consultant for Allergan, Bayer, Johnson and Johnson, Medtronic, Sanofi, and Phillips. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.

ABBREVIATIONS AND ACRONYMS

AF

atrial fibrillation

CI

confidence interval

DOAC

direct oral anticoagulant

GI

gastrointestinal

HR

hazard ratio

ICD-9-CM

International Classification of Diseases-9th Revision-Clinical Modification

ICH

intracranial hemorrhage

LAAO

left atrial appendage occlusion

OAC

oral anticoagulation

Footnotes

The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.

APPENDIX For supplemental tables, please see the online version of this paper.

REFERENCES

  • 1.Hart RG, Benavente O, McBride R, Pearce LA. Antithrombotic therapy to prevent stroke in patients with atrial fibrillation: a meta-analysis. Ann Intern Med 1999;131:492–501. [DOI] [PubMed] [Google Scholar]
  • 2.Camm AJ, Lip GY, De Caterina R, et al. 2012 focused update of the ESC Guidelines for the management of atrial fibrillation: an update of the 2010 ESC Guidelines for the management of atrial fibrillation. Developed with the special contribution of the European Heart Rhythm Association. Eur Heart J 2012; 33:2719–47. [DOI] [PubMed] [Google Scholar]
  • 3.Wann LS, Curtis AB, January CT, et al. 2011 ACCF/AHA/HRS focused update on the management of patients with atrial fibrillation (updating the 2006 guideline): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2011;57: 1330–7. [DOI] [PubMed] [Google Scholar]
  • 4.Fang MC, Stafford RS, Ruskin JN, Singer DE. National trends in antiarrhythmic and antithrombotic medication use in atrial fibrillation. Arch Intern Med 2004;164:55–60. [DOI] [PubMed] [Google Scholar]
  • 5.Gage BF, Boechler M, Doggette AL, et al. Adverse outcomes and predictors of underuse of antithrombotic therapy in medicare beneficiaries with chronic atrial fibrillation. Stroke 2000;31: 822–7. [DOI] [PubMed] [Google Scholar]
  • 6.Bungard TJ, Ghali WA, Teo KK, McAlister FA, Tsuyuki RT. Why do patients with atrial fibrillation not receive warfarin? Arch Intern Med 2000;160: 41–6. [DOI] [PubMed] [Google Scholar]
  • 7.Go AS, Hylek EM, Borowsky LH, Phillips KA, Selby JV, Singer DE. Warfarin use among ambulatory patients with nonvalvular atrial fibrillation: the anticoagulation and risk factors in atrial fibrillation (ATRIA) study. Ann Intern Med 1999; 131:927–34. [DOI] [PubMed] [Google Scholar]
  • 8.Ogilvie IM, Newton N, Welner SA, Cowell W, Lip GY. Underuse of oral anticoagulants in atrial fibrillation: a systematic review. Am J Med 2010; 123:638–45.e4. [DOI] [PubMed] [Google Scholar]
  • 9.O’Brien EC, Simon DN, Allen LA, et al. Reasons for warfarin discontinuation in the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF). Am Heart J 2014;168:487–94. [DOI] [PubMed] [Google Scholar]
  • 10.Steinberg BA, Greiner MA, Hammill BG, et al. Contraindications to anticoagulation therapy and eligibility for novel anticoagulants in older patients with atrial fibrillation. Cardiovasc Ther 2015; 33:177–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Piccini JP, Sinner MF, Greiner MA, et al. Outcomes of Medicare beneficiaries undergoing catheter ablation for atrial fibrillation. Circulation 2012;126:2200–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Piccini JP, Hammill BG, Sinner MF, et al. Incidence and prevalence of atrial fibrillation and associated mortality among Medicare beneficiaries, 1993–2007. Circ Cardiovasc Qual Outcomes 2012;5:85–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Piccini JP, Mi X, Dewald TA, Go AS, Hernandez AF, Curtis LH. Pharmacotherapy in Medicare beneficiaries with atrial fibrillation. Heart Rhythm 2012;9:1403–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Goldberg D, Lewis J, Halpern S, Weiner M, Lo Re V 3rd. Validation of three coding algorithms to identify patients with end-stage liver disease in an administrative database. Pharmacoepidemiol Drug Saf 2012;21:765–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Birman-Deych E, Waterman AD, Yan Y, Nilasena DS, Radford MJ, Gage BF. Accuracy of ICD-9-CM codes for identifying cardiovascular and stroke risk factors. Med Care 2005;43:480–5. [DOI] [PubMed] [Google Scholar]
  • 16.Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130–9. [DOI] [PubMed] [Google Scholar]
  • 17.Prentice RL, Kalbfleisch JD, Peterson AV Jr., Flournoy N, Farewell VT, Breslow NE. The analysis of failure times in the presence of competing risks. Biometrics 1978;34:541–54. [PubMed] [Google Scholar]
  • 18.Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat 1988;16:1141–54. [Google Scholar]
  • 19.McCrory DC, Matchar DB, Samsa G, Sanders LL, Pritchett EL. Physician attitudes about anticoagulation for nonvalvular atrial fibrillation in the elderly. Arch Intern Med 1995;155:277–81. [PubMed] [Google Scholar]
  • 20.O’Brien EC, Holmes DN, Ansell JE, et al. Physician practices regarding contraindications to oral anticoagulation in atrial fibrillation: findings from the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) registry. Am Heart J 2014;167:601–9.e1. [DOI] [PubMed] [Google Scholar]
  • 21.O’Brien EC, Simon DN, Thomas LE, et al. The ORBIT bleeding score: a simple bedside score to assess bleeding risk in atrial fibrillation. Eur Heart J 2015;36:3258–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJ, Lip GY. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest 2010;138:1093–100. [DOI] [PubMed] [Google Scholar]
  • 23.Friberg L, Rosenqvist M, Lip GY. Net clinical benefit of warfarin in patients with atrial fibrillation: a report from the Swedish atrial fibrillation cohort study. Circulation 2012;125: 2298–307. [DOI] [PubMed] [Google Scholar]
  • 24.Connolly SJ, Eikelboom J, Joyner C, et al. Apixaban in patients with atrial fibrillation. N Engl J Med 2011;364:806–17. [DOI] [PubMed] [Google Scholar]
  • 25.Ruff CT, Giugliano RP, Braunwald E, 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;383:955–62. [DOI] [PubMed] [Google Scholar]
  • 26.Sherwood MW, Nessel CC, Hellkamp AS, et al. Gastrointestinal bleeding in patients with atrial fibrillation treated with rivaroxaban or warfarin: ROCKET AF trial. J Am Coll Cardiol 2015;66: 2271–81. [DOI] [PubMed] [Google Scholar]
  • 27.Reddy VY, Doshi SK, Kar S, et al. 5-year outcomes after left atrial appendage closure: from the PREVAIL and PROTECT AF trials. J Am Coll Cardiol 2017;70:2964–75. [DOI] [PubMed] [Google Scholar]
  • 28.Dukkipati SR, Kar S, Holmes DR Jr., et al. Device-related thrombus after left atrial appendage closure: incidence, predictors, and outcomes. Circulation 2018;138:874–85. [DOI] [PubMed] [Google Scholar]
  • 29.Holmes DR, Reddy VY, Buchbinder M, et al. The Assessment of the Watchman Device in Patients Unsuitable for Oral Anticoagulation (ASAPTOO) trial. Am Heart J 2017;189:68–74. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1

RESOURCES