Atrial fibrillation (AF) frequently occurs with acute coronary syndromes (ACS) and adds complexity to the selection of an appropriate antithrombotic strategy. We determined whether associations of antithrombotic treatment with bleeding, stroke, and death differ between patients with ACS with and without AF. Residents of Olmsted County, Minnesota, hospitalized with incident ACS during 2005 to 2010 were classified according to the presence or absence of AF either before or during the index ACS hospitalization. Antithrombotic strategy at discharge was categorized as double/triple agents versus no/single agent. Patients were followed through 2012, and propensity scores were used to estimate associations of treatment with bleeding, ischemic stroke, and mortality. Of 1,159 patients with incident ACS, 252 (21.7%) had concomitant AF (ACS + AF). Over a median follow-up of 4.3 years, 312 bleeds, 67 ischemic strokes, and 268 deaths occurred. The overall risks of bleeding, stroke, and death were similar between treatment strategies. Although limited by the small number of events, a suggestion of a lower risk of ischemic stroke for patients with ACS + AF on double/triple therapy was observed; the hazard ratios for stroke with double/triple versus no/single therapy were 0.30 (0.07 to 1.26) and 1.10 (0.52 to 2.33) in those with and without AF, respectively (p value for interaction = 0.10). In conclusion, the choice of antithrombotic strategy is not associated with the risk of ischemic stroke, bleeding, or death in patients with ACS overall. Patients with ACS + AF on double/triple therapy may experience reduced risks of stroke, although future studies are needed to confirm this finding.
Atrial fibrillation (AF) often complicates acute coronary syndromes (ACS), with at least 1 in 8 patients with ACS having AF. Selection of an optimal antithrombotic regimen for these patients is complex. Yet to date, no clinical trials have been undertaken to study the impact of antithrombotic treatment on adverse outcomes in ACS focusing specifically on differences between those with and without AF. Accordingly, we undertook this study to describe, in a well-defined community setting, the spectrum of antithrombotic treatment in patients with ACS with and without AF, and the association of antithrombotic strategies with subsequent risk of bleeding, stroke, and all-cause mortality.
Methods
This study was conducted in Olmsted County, Minnesota, using the resources of the Rochester Epidemiology Project. Virtually complete capture of health care utilization and outcomes in Olmsted County is possible because only a few providers deliver most health care to local residents and each provider uses comprehensive medical records, which are indexed through the Rochester Epidemiology Project. This study was approved by the Mayo Clinic and Olmsted Medical Center Institutional Review Boards.
Residents of Olmsted County, Minnesota, with hospital discharge diagnoses of International Classification of Diseases, Ninth Revision, Clinical Modification ( ICD-9-CM ) codes 410 to 411 from January 1, 2005, to December 31, 2010, were identified. The presence of cardiac chest pain was used to validate unstable angina (UA) using the Braunwald classification. Epidemiologic criteria incorporating cardiac pain, biomarker levels, and Minnesota coding of the electrocardiogram (ECG) were used to validate myocardial infarction (MI). According to guidelines using Troponin T in the algorithm, the presence or absence of a change between any 2 troponin measurements is defined by a difference of ≥0.05 ng/ml. As troponin can remain elevated for 2 weeks after events causing its rise, the biomarkers were downgraded from abnormal to equivocal when these conditions occurred ≤2 weeks before the MI.
AF events occurring before ACS or during the index ACS hospitalization were identified using ECGs and ICD-9-CM codes 427.31 or 427.32 assigned during inpatient or outpatient visits. The ECGs were electronically interpreted, and as part of routine clinical practice, all ECGs were subsequently verified by a cardiologist. When no ECG was present or when inconsistencies between the dates of the ECG and diagnostic code precluded the ability to determine whether AF was present at or before index, manual review of the medical record was used to validate the AF event.
Patient demographics, cigarette smoking status, procedures, and discharge medications were obtained from review of patient medical records. Antithrombotic prophylaxis therapies included warfarin, aspirin, and other antiplatelets (clopidogrel, ticlopidine, and dipyridamole). Body mass index was calculated as weight (in kilograms) divided by height (in meters) squared. Clinicians’ diagnoses were used to identify history of hyperlipidemia, hypertension, heart failure (HF), chronic obstructive pulmonary disease (COPD), cancer, or stroke or transient ischemic attack before ACS. The American Diabetes Association criteria were used to define diabetes. Glomerular filtration rate (eGFR) was estimated using the closest serum creatinine within 1 year of index using the Modification of Diet in Renal Disease Study equation. The CHADS 2 risk score for future stroke risk prediction and the ATRIA bleeding risk score were calculated.
Participants were followed through December 31, 2012, for bleeding, strokes, and deaths from any cause. Bleeding events after discharge were ascertained using ICD-9-CM codes identified by Fosbol et al as a guideline. For strokes, we excluded from our analyses individuals who had a history of ischemic stroke (n = 41), and in the remaining individuals, ICD-9-CM codes 433.x1, 434.x1, and 436 were used to identify incident ischemic strokes. Deaths were obtained from inpatient and outpatient medical records, death certificates from the state of Minnesota, and obituaries and notices of death in the local newspapers.
Statistical analyses were performed using SAS statistical software, version 9.2 (SAS Institute Inc., Cary, North Carolina). Baseline participant characteristics by the presence of AF were compared using chi-square tests for categorical variables and t tests for continuous variables. Patients were categorized based on the number of antithrombotic agents (none, single, double, and triple). Logistic regression was used to determine predictors of double/triple versus no/single antithrombotic therapy after adjustment for age and sex.
A propensity score for double/triple versus no/single antithrombotic therapy was estimated using the following variables at index ACS: age, sex, marital status, highest level of education attained, body mass index, smoking status, eGFR, family history of coronary artery disease, hypertension, hyperlipidemia, diabetes, COPD, HF, stroke or transient ischemic attack, malignancy, presence of AF, type of ACS (UA or MI), reperfusion/revascularization, peak troponin, statins, β blockers, and angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers. Inverse probability of treatment weights were assigned using a weight of 1/P for those on double/triple therapy and 1/(1 − P) for no/single therapy, where P is the propensity score for double/triple therapy. To assess the associations of double/triple versus no/single antithrombotic therapy with bleeding, ischemic stroke, and all-cause mortality, Cox proportional hazard regression, using inverse probability of treatment weighting of the propensity score and adjusting for age and sex, was modeled. In ancillary analyses, a propensity score for warfarin versus no warfarin was created, and the Cox regression analyses were repeated for warfarin versus no warfarin.
Results
From 2005 to 2010, 1,244 incident ACS were identified, of which 23 did not grant complete research authorization and 62 died in hospital, leaving 1,159 patients with incident ACS (mean age 66.5 ± 14.5 years, 60% men, 93% white) for our study. Of these, 729 (63%) had MI and 430 (37%) had UA. In addition, 252 (22%) had concomitant AF. Those with AF presented with a similar proportion of UA and MI as those without AF. However, those with ACS + AF were older, more likely to be women, less likely to be a current smoker, and had more co-morbidities ( Table 1 ).
| Variable | Total (N=1159) | Atrial Fibrillation | P-value | |
|---|---|---|---|---|
| Yes (N=252) | No (N=907) | |||
| Age (years) | 66.5 (14.5) | 76.2 (11.8) | 63.8 (14.1) | <0.001 |
| Male | 694 (59.9%) | 130 (51.6%) | 564 (62.2%) | 0.002 |
| Married | 718 (61.9%) | 132 (52.4%) | 586 (64.6%) | <0.001 |
| High school or greater education | 1020 (88.5%) | 216 (86.1%) | 804 (89.2%) | 0.162 |
| Body mass index (kg/m 2 ) | 29.4 (6.5) | 29.0 (6.4) | 29.5 (6.5) | 0.137 |
| Smoking status | ||||
| Never | 456 (39.3%) | 108 (42.9%) | 348 (38.4%) | <0.001 |
| Former | 450 (38.8%) | 112 (44.4%) | 338 (37.3%) | |
| Current | 253 (21.8%) | 32 (12.7%) | 221 (24.4%) | |
| Hypertension | 821 (70.8%) | 217 (86.1%) | 604 (66.6%) | <0.001 |
| Hyperlipidemia | 801 (69.1%) | 189 (75.0%) | 612 (67.5%) | 0.022 |
| Diabetes mellitus | 252 (21.7%) | 56 (22.2%) | 196 (21.6%) | 0.835 |
| Family history of CAD | 258 (22.3%) | 39 (15.5%) | 219 (24.1%) | 0.003 |
| Heart failure | 142 (12.3%) | 72 (28.6%) | 70 (7.7%) | <0.001 |
| Stroke/transient ischemic attack | 120 (10.4%) | 44 (17.5%) | 76 (8.4%) | <0.001 |
| Cancer | 208 (17.9%) | 73 (29.0%) | 135 (14.9%) | <0.001 |
| Chronic obstructive pulmonary disease | 164 (14.2%) | 57 (22.6%) | 107 (11.8%) | <0.001 |
| eGFR (mL/min per 1.73 m 2 ) | 69.4 (25.0) | 61.0 (24.7) | 71.8 (24.6) | <0.001 |
| Type of ACS | ||||
| Unstable angina pectoris | 430 (37.1) | 101 (40.1) | 329 (36.3) | 0.253 |
| STEMI | 168 (14.5) | 29 (11.5) | 139 (15.3) | |
| non-STEMI | 561(48.4) | 122 (48.4) | 439 (48.4) | |
| Peak troponin (ng/mL) | 1.3 (2.8) | 1.0 (2.5) | 1.3 (2.9) | 0.083 |
Nearly 80% of all patients with ACS underwent coronary angiography, but differences were observed between those with and without AF. Those with AF were less likely to have angiography (153 [63%] for those with AF and 742 [82%] for those without AF). However, of those who underwent angiography, a higher proportion of the patients with ACS + AF had ≥2 vessels with >75% stenosis compared with patients with ACS without AF (40% vs 27%, respectively).
More patients with ACS + AF were managed noninvasively than those without AF (58% vs 36%, respectively). Only 106 (42%) patients with ACS + AF received reperfusion/revascularization during the hospitalization compared with 583 (64%) patients with ACS without AF. Of those who received reperfusion/revascularization, 67 (63%) and 512 (88%) had percutaneous coronary intervention (PCI) with stent placement in the ACS + AF and no AF categories, respectively ( Figure 1 ). Most patients were prescribed β blockers (86%) and angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers (58%), which did not differ by AF. However, patients with ACS + AF were less likely to be prescribed statins (72% vs 87% in those without AF).
AF status substantially affected the choice of antithrombotic strategy at discharge ( Table 2 ). Fewer than half (48%) of the patients with ACS + AF were discharged on 2 or 3 antithrombotic agents and 33% received warfarin, usually in combination with a single antiplatelet agent, and only 10% received warfarin with 2 antiplatelet agents. In contrast, among patients with ACS without AF, 63% were discharged on either 2 or 3 agents, the majority on dual antiplatelets, with only 4% on warfarin and 1% on warfarin with 2 antiplatelets.
| Variable | Overall (N=1159) | Atrial Fibrillation | |
|---|---|---|---|
| Yes (N=252) | No (N=907) | ||
| Individual Therapy Combinations | |||
| None | 67 (5.8%) | 14 (5.6%) | 53 (5.8%) |
| Aspirin alone | 360 (31.1%) | 99 (39.3%) | 261 (28.8%) |
| Other antiplatelet alone | 20 (1.7%) | 5 (2.0%) | 15 (1.7%) |
| Warfarin alone | 18 (1.6%) | 12 (4.8%) | 6 (0.7%) |
| Aspirin + other antiplatelet | 593 (51.2%) | 50 (19.8%) | 543 (59.9%) |
| Aspirin + warfarin | 64 (5.5%) | 46 (18.3%) | 18 (2.0%) |
| Warfarin + other antiplatelet | 2 (0.2%) | 2 (0.8%) | 0 (0.0%) |
| All 3 treatments | 35 (3.0%) | 24 (9.5%) | 11 (1.2%) |
| Number of Treatments | |||
| None | 67 (5.8%) | 14 (5.6%) | 53 (5.8%) |
| Single | 398 (34.3%) | 116 (46.0%) | 282 (31.1%) |
| Double | 659 (56.9%) | 98 (38.9%) | 561 (61.9%) |
| Triple | 35 (3.0%) | 24 (9.5%) | 11 (1.2%) |
After adjustment for age and sex, the predictors of double/triple antithrombotic treatment strategies in patients with ACS + AF included younger age, nonsmoking status, overweight and obesity, hypertension, no history of COPD, higher CHADS 2 score, higher peak troponin, and reperfusion/revascularization; a trend for increased use of double/triple therapy was observed for those with lower eGFR ( Figure 2 ). In those without AF, the predictors of double/triple therapy included younger age, male gender, higher education, higher eGFR, no history of HF or cancer, MI as the index ACS event, higher peak troponin, and reperfusion/revascularization ( Figure 2 ).
Over a median follow-up of 4.3 years for all patients with ACS, 312 patients experienced a bleed, 67 incident strokes occurred, and 268 patients died. Overall, the risk of subsequent bleeding, strokes, and all-cause mortality were similar for patients discharged on double/triple and no/single antithrombotic therapy ( Table 3 ). However, a trend suggesting a lower risk for ischemic stroke was observed for patients with ACS + AF on double/triple therapy.
| Bleeding | Ischemic Stroke | All-Cause Mortality | |||||
|---|---|---|---|---|---|---|---|
| No/Single | Double/Triple | No/Single | Double/Triple | No/Single | Double/Triple | ||
| Overall | N events | 132 | 180 | 31 | 36 | 159 | 109 |
| Rate (95% CI) ∗ | 87.78 (73.44-104.1) | 72.05 (61.90-83.37) | 17.30 (11.75-24.55) | 12.12 (8.49-16.78) | 85.82 (73.00-100.2) | 35.70 (29.32-43.07) | |
| HR (95% CI) † | 1.00 (ref) | 1.00 (0.74-1.36) | 1.00 (ref) | 0.82 (0.43-1.54) | 1.00 (ref) | 0.97 (0.71-1.35) | |
| By AF Status | |||||||
| AF | N events | 44 | 47 | 10 | 4 | 73 | 40 |
| Rate (95% CI) ∗ | 139.4 (101.3-187.2) | 125.3 (92.05-166.6) | 24.71 (11.85-45.45) | 8.33 (2.27-21.33) | 171.2 (134.2-215.2) | 82.13 (58.67-111.8) | |
| HR (95% CI) † | 1.00 (ref) | 1.07 (0.60-1.91) § | 1.00 (ref) | 0.30 (0.07-1.26) ¶ | 1.00 (ref) | 0.80 (0.48-1.35) ‡ | |
| No AF | N events | 88 | 133 | 21 | 32 | 86 | 69 |
| Rate (95% CI) ∗ | 74.06 (59.40-91.25) | 62.64 (52.45-74.24) | 15.13 (9.37-23.14) | 12.85 (8.79-18.14) | 60.30 (48.23-74.47) | 26.89 (20.92-34.03) | |
| HR (95% CI) † | 1.00 (ref) | 0.99 (0.69-1.42) § | 1.00 (ref) | 1.10 (0.52-2.33) ¶ | 1.00 (ref) | 1.14 (0.76-1.71) ‡ | |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
