Warfarin Use and Outcomes in Patients with Atrial Fibrillation Complicating Acute Coronary Syndromes

Methods 

We conducted these analyses using a pooled database of 23,208 patients with non–ST-segment elevation acute coronary syndrome who were enrolled in clinical trials and in whom information on atrial fibrillation as an in-hospital complication and use of oral anticoagulation at discharge was available. This population comprised patients from the following trials: Platelet IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy (PURSUIT),11 Platelet IIb/IIIa Antagonist for the Reduction of Acute Coronary Syndrome Events in a Global Organization Network A (PARAGON-A),12 and Superior Yield of the New Strategy of Enoxaparin, Revascularization and Glycoprotein IIb/IIIa Inhibitors (SYNERGY).13 Details of these trials were published previously. Briefly, these trials were randomized and conducted in multiple countries throughout the world. Patients were enrolled from 1995 to 1996 (PARAGON-A), 1995 to 1997 (PURSUIT), and 2001 to 2003 (SYNERGY). Inclusion criteria included chest pain, changes on electrocardiogram (ST-segment deviations), and positive biomarkers of myocardial necrosis.

Atrial fibrillation was defined as an irregularly irregular rhythm with lack of discernible P waves on an electrocardiogram any time during the hospital course after randomization. It was identified as a complication on the patient's case report form. Information about type, duration, or history of atrial fibrillation was not available.

The primary clinical outcome of this study was adjudicated death or myocardial infarction at 6 months. Myocardial infarction was defined by elevated biomarkers (2×upper limit of normal) or electrocardiographic evidence (new and significant Q waves in 2 contiguous leads).

Results 

Of the 23,208 patients enrolled in the PURSUIT, PARAGON-A, and SYNERGY trials, 5.8% (1346 patients) had atrial fibrillation as an in-hospital complication and 4.0% (917 patients) had atrial fibrillation and were discharged alive. Data on long-term outcomes were complete in 97% of the patients.

Baseline Characteristics According to Warfarin Therapy 

Patients who received warfarin at discharge more often had a history of heart failure, diabetes, myocardial infarction, and stroke and less often had a history of hypertension when compared with patients who did not receive warfarin at discharge (Table 1). Otherwise, the baseline characteristics were similar between the 2 groups.

Table 1. Baseline Characteristics According to Warfarin Use at Discharge

	Parameter	All Patients (n=917)	Warfarin (n=124)	No Warfarin (n=793)	P Value
	Age, median (25th,75th), y	69.0(62.0,74.4)	68.0(62.0,73.7)	69.0(62.0,75.0)	.33a
	Female	31.8	33.9	31.5	.60
	Race				.57
	White	90.5	92.7	90.2
	Black	4.4	4.0	4.4
	Other	5.1	3.2	5.4
	Smoking status				.80
	Never	40.7	41.0	40.7
	Previous	37.4	39.3	37.1
	Current	21.9	19.7	22.2
	Killip class				.50b
	I	82.0	80.5	82.2
	II	15.4	15.4	15.3
	III	2.3	4.1	2.0
	IV	0.3	0	0.4
	Region				.43b
	North America	56.4	61.3	55.6
	Western Europe	26.8	28.2	26.6
	Eastern Europe	10.8	8.1	11.2
	Latin America	2.4	1.6	2.5
	Australia	3.4	0.8	3.8
	Other	0.2	0	0.3
	Weight, median (25th,75th), kg	79.0(69.0,90.0)	78.5(66.4,87.5)	79.0(69.0,90.0)	.32
	Height, median (25th,75th), cm	170.2(164.5,177.8)	170.0(162.6,177.8)	170.2(165.0,177.8)	.27
	BP, median (25th,75th), mm Hg
	Systolic	130.0(116.0,145.0)	130.0(115.0,140.0)	130.0(116.0,146.0)	.49a
	Diastolic	72.0(64.0,80.0)	70.5(63.0,80.0)	72.0(64.0,80.0)	.51a
	Heart rate, median (25th,75th), beats/min	75.0(64.0,85.0)	76.0(62.0,88.0)	75.0(65.0,85.0)	.84a
	Prior CABG	12.2	11.3	12.4	.74
	Prior hypertension	61.8	58.1	62.4	.35
	Prior CHF	16.8	21.0	16.2	.18
	Prior diabetes mellitus	25.4	31.5	24.5	.10
	Prior PCI	13.5	15.3	13.2	.53
	Prior MI	34.7	38.7	34.1	.31
	Proportion of US patients	44.6	49.2	43.9	.27
	Prior CVA, stroke, or TIA	7.6	9.7	7.3	.36

BP=blood pressure; CABG=coronary artery bypass graft surgery; CHF=congestive heart failure; CVA=cardiovascular accident; MI=myocardial infarction; PCI=percutaneous coronary intervention; TIA=transient ischemic attack. Data presented as percentages unless otherwise indicated.

a Nonparametric test. b Exact test.

Medications at Discharge and In-hospital Procedures According to Warfarin Therapy 

The discharge medications and in-hospital procedures according to warfarin use are shown in Table 2. Compared with patients who did not receive warfarin at discharge, patients who did receive warfarin at discharge were more likely to be discharged with an antiarrhythmic drug, a calcium channel blocker, an angiotensin-converting enzyme inhibitor, and a nitrate, and less likely to receive aspirin and a beta-blocker.

Table 2. Discharge Medications and In-hospital Procedures According to Warfarin Use at Discharge

	Parameter	Warfarin(n=124)	No Warfarin(n=793)
	Discharge medications
	Aspirin	62.9	89.0
	Digitalis	50.5	36.9
	Ticlopidine/clopidogrel	10.5	26.4
	Beta-blockers	54.8	60.9
	Calcium channel blockers	24.2	22.6
	Nitrates	43.5	39.3
	ACE inhibitors	48.4	44.6
	Antiarrhythmics	30.5	27.1
	In-hospital catheterization	70.2	71.8
	In-hospital PCI	16.9	23.7
	In-hospital CABG	36.3	34.2

ACE=angiotensin-converting enzyme; CABG=coronary artery bypass graft surgery; PCI=percutaneous coronary intervention. Data are presented as percentages.

Warfarin Use According to Congestive Heart Failure, Hypertension, Age>75 Years, Diabetes, Prior Stroke/Transient Ischemic Attack Score and Bleeding Risk 

Warfarin use among patients with atrial fibrillation was similar across CHADS2 scores: 0, 13%; 1, 14%;≥2, 13% (Figure 1). Among patients with atrial fibrillation, aspirin and ticlopidine/clopidogrel use rates also were similar across CHADS2 scores. Of the components of the CHADS2 score, only prior heart failure had a statistically significant association with warfarin use at discharge (odds ratio 1.5; 95% confidence interval (CI), 1.02-2.28; P=.04).

View Large Image Download to PowerPoint

Figure 1. Antithrombotic use according to CHADS2 score. CHADS2=Congestive heart failure, Hypertension, Age>75 years, Diabetes, Prior Stroke/transient ischemic attack.

The use of warfarin according to bleeding risk is shown in Figure 2. Warfarin was used similarly and approximately 12% of the time, regardless of the risk of bleeding. Notably, as the risk of bleeding increased, the CHADS2 score followed the same pattern that illustrates the overlap of clinical factors associated with the risk of both stroke and bleeding. Patients at high risk of bleeding had a median CHADS2 score of 4, and 11.1% received warfarin; the rate of warfarin use was similar (11.9%) for patients at low risk of bleeding, and the median CHADS2 score was 1. Despite the increase in the median CHADS2 score, the use of aspirin and ticlopidine/clopidogrel decreased as the risk of bleeding increased (Figure 2).

View Large Image Download to PowerPoint

Figure 2. Antithrombotic use according to bleeding risk. CHADS2=Congestive heart failure, Hypertension, Age>75 years, Diabetes, Prior Stroke/transient ischemic attack.

Although patients who had prior stroke were more likely to receive warfarin at discharge (20%) than patients without a previous stroke (13%), the use of warfarin was low in both groups (Figure 3). Patients with a prior stroke were less likely to receive aspirin or ticlopidine/clopidogrel at discharge.

View Large Image Download to PowerPoint

Figure 3. Medication given at discharge based on prior stroke.

Six-Month Death or Myocardial Infarction 

Factors associated with 6-month death or myocardial infarction are shown in Table 3. Warfarin use at discharge was strongly associated with a lower risk of death or myocardial infarction within 6 months (adjusted hazard ratio 0.39; 95% CI, 0.15-0.98; P=.04) (Figure 4). In the sensitivity analysis, after adjusting for the propensity of receiving warfarin at discharge, similar results were found; warfarin use at discharge was significantly associated with lower rates of the 6-month outcome of death or myocardial infarction (hazard ratio 0.38; 95% CI, 0.15-0.97; P=.04).

Table 3. Predictors of 6-month Death or Myocardial Infarction in Patients with Atrial Fibrillation and Acute Coronary Syndrome

	Variable	HR (95% CI)	Chi-square	P Value
	In-hospital CABG	0.30(0.15–0.61)	12.11	<.01
	CHADS2 score≥2	3.51(2.03–6.06)	20.36	<.01
	Warfarin at discharge	0.39(0.15–0.98)	3.93	.04
	Aspirin at discharge	0.47(0.26–0.85)	6.62	.01

CABG=coronary artery bypass graft surgery; CHADS2=Congestive heart failure, Hypertension, Age>75 years, Diabetes, Prior Stroke/transient ischemic attack; CI=confidence interval; HR=hazard ratio.

View Large Image Download to PowerPoint

Figure 4. Adjusted 6-month death or myocardial infarction curves according to warfarin use at discharge. MI=myocardial infarction.

Discussion 

Our study has 3 main findings. First, we demonstrated that warfarin use is associated with a lower risk of 6-month death or myocardial infarction in patients with atrial fibrillation complicating an acute coronary syndrome. Second, although atrial fibrillation guidelines recommend warfarin use based on CHADS2 score,15 we showed that the use of warfarin was independent of both the estimated stroke and bleeding risks; this also was true of other antithrombotic medications. Finally, we identified several factors that are strongly associated with a risk of 6-month death or myocardial infarction, including CHADS2 score≥2.

Antithrombotic therapy is indicated for the treatment of both acute coronary syndromes and atrial fibrillation.15, 16 However, for patients with an acute coronary syndrome and atrial fibrillation, the use of combinations of different antithrombotic agents may significantly increase the risk of bleeding. Rothberg et al17 showed that the combination of aspirin and warfarin for patients with myocardial infarction is associated with lower mortality compared with the use of aspirin alone; however, bleeding rates are higher. Unfortunately, these questions could not be addressed in our study because of the lack of data on long-term bleeding. We showed that patients at high risk for bleeding are at high risk for stroke as well, and warfarin was prescribed similarly to patients at low risk for stroke. This illustrates the overlap between the clinical factors associated with both stroke and bleeding and that physicians are making the decision to use warfarin on the basis of other (unmeasured) factors. The use of antiplatelet agents, however, was lower in patients at higher risk of bleeding despite their increased risk of stroke. Thus, bleeding risk is related to the decision not to prescribe antiplatelet agents for patients with atrial fibrillation after acute myocardial infarction, even when the stroke risk is high.

In an observational study, Stenestrand et al18 showed that patients with atrial fibrillation and myocardial infarction who receive warfarin and a platelet inhibitor had lower 1-year mortality compared with those who received aspirin alone. Approximately 7.5% of their patients had atrial fibrillation and were discharged alive. The population of this study included patients with ST-segment elevation and patients with ST-segment depression acute coronary syndrome, and 30% of the patients with atrial fibrillation and acute myocardial infarction received oral anticoagulant therapy at discharge. Although our study was limited to patients with non–ST-segment elevation acute coronary syndrome and the use of warfarin was lower than in previous studies, our analysis included patients enrolled in large randomized clinical trials in which variables were prospectively defined and data were prospectively collected.

We demonstrated that although international guidelines16, 19 recommend warfarin for patients with atrial fibrillation and clear indication for anticoagulation, this recommendation tended not to be followed in those with post-acute coronary syndrome. In patients with in-hospital atrial fibrillation and prior stroke (CHADS2 score of at least 2), in whom warfarin use is generally indicated, warfarin was only prescribed approximately 20% of the time at hospital discharge (Figure 2).

Concordant with the findings of the present study, we previously showed an inverse relationship between the CHADS2 score and warfarin use in patients with post–ST-segment elevation myocardial infarction atrial fibrillation.20 This might relate to the observation that patients with a higher CHADS2 score are at an increased risk of both stroke and bleeding. Thus, the set of patients who might benefit most from warfarin tended not to receive it, perhaps because of concerns over subsequent risk of bleeding. This scenario results in an apparent paradox—patients with a higher CHADS2 score have lower warfarin use. Moreover, patients at high risk for ischemic stroke who also are at high risk for bleeding received warfarin in similar proportion to those at low risk for stroke. However, this explanation is largely speculative and needs to be verified in future studies.

We showed previously that patients with ST-segment elevation myocardial infarction and atrial fibrillation who received warfarin at discharge had lower rates of both 90-day mortality and stroke when compared with those patients who did not receive warfarin at hospital discharge.20 In the present study, we demonstrated that both warfarin and aspirin use at discharge were independently associated with lower rates of 6-month death or myocardial infarction. Therefore, although only 14% of the patients received warfarin at discharge, our study supports the guidelines in their recommendation to consider the use of warfarin and aspirin in patients with non–ST-segment elevation acute coronary syndrome and atrial fibrillation.

The timing of atrial fibrillation in the setting of an acute coronary syndrome also is an important issue that should be taken into consideration. In the present study, we described patients with non–ST-segment elevation acute coronary syndrome who developed in-hospital atrial fibrillation as a complication. We do not have information about the patients' rhythm at discharge. Siu et al21 demonstrated that transient atrial fibrillation in the setting of acute myocardial infarction is an independent predictor of future development of atrial fibrillation and stroke. We showed previously that the majority of patients who developed atrial fibrillation in the setting of an acute myocardial infarction were not in atrial fibrillation at the time of hospital discharge.20 Therefore, even if most of our patients with atrial fibrillation were not in atrial fibrillation at the time of discharge, our results raise the possibility that even if the atrial fibrillation is transient, patients have better outcomes if they receive warfarin at discharge.

Conclusions 

Warfarin is associated with better 6-month outcomes among patients with atrial fibrillation complicating an acute coronary syndrome, but its use is infrequent and not related to stroke or bleeding risks. Among the components of the CHADS2 score, the only significant predictor of warfarin use at discharge was heart failure. CHADS2 score≥2 was independently associated with worse outcomes. Antiplatelet agents, particularly ticlopidine/clopidogrel, were less often prescribed in patients at higher risk of bleeding. This study highlights the need to better understand optimal antithrombotic therapy for atrial fibrillation occurring in the setting of an acute coronary syndrome.