In the setting of ST-segment elevation myocardial infarction (STEMI), patients at high risk of systemic emboli who undergo primary percutaneous coronary intervention (PCI) using stents might require triple antithrombotic therapy (a combination of aspirin, thienopyridine, and vitamin K antagonist [VKA]). The risks and benefits of such therapy in the setting of STEMI have been incompletely characterized. We, therefore, assessed the outcomes of patients who received triple therapy after primary PCI in the large-scale, contemporary Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction [HORIZONS-AMI] trial. Among the 3,320 patients triaged to primary PCI, 126 (3.8%) were prescribed triple therapy and 3,194 (96.2%) were prescribed dual antiplatelet therapy. The most frequent indications for VKA treatment were a severely reduced left ventricular ejection fraction with a large akinetic area, atrial fibrillation (23.8% each), and mural thrombus (23.0%). The assignment to triple therapy was associated with older age, female gender, rhythm disturbances, Killip class >1 on admission, lower left ventricular ejection fraction, left anterior descending artery territory infarcts, and Final Thrombolysis In Myocardial Infarction flow grade <3. Patients treated with triple versus dual therapy had comparable short- and long-term ischemic outcomes but had significantly increased rates of major bleeding during the index hospitalization (17.1% vs 6.5%, p <0.0001), resulting in premature VKA discontinuation in 14.3% of those patients. In conclusion, in the setting of STEMI treated with primary PCI, the combination of aspirin, thienopyridine, and VKA results in an excess of bleeding complications and premature discontinuation of VKA. The risk of adding oral anticoagulation to patients admitted for STEMI should be carefully considered before choosing drug-eluting or bare metal stents.
Oral anticoagulation with vitamin K antagonists (VKAs) improves the outcomes and decreases the incidence of cerebral and peripheral embolism in patients who are at high risk of thromboembolic events. In the setting of ST-segment elevation myocardial infarction (STEMI), oral anticoagulants are recommended for the treatment of patients with a large akinetic region in the left ventricle, mural thrombus, or persistent or paroxysmal atrial fibrillation. When treated with primary percutaneous coronary intervention (PCI) using stents, patients at high risk of systemic emboli require triple antithrombotic therapy, including aspirin, thienopyridine, and a VKA. Such a combination carries an increased risk of bleeding events beyond that associated with each drug alone. However, the risks and benefits of triple therapy in patients with STEMI undergoing primary PCI have been incompletely characterized. We, therefore, assessed the outcomes of patients who received a combination of aspirin, thienopyridine, and VKA in the large-scale, prospective, multicenter, contemporary Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction (HORIZONS-AMI) trial.
Methods
The design and principal results of the HORIZONS-AMI trial have been previously reported. In brief, a total of 3,602 patients ≥18 years old with acute myocardial infarction within 12 hours of symptom onset and ST-segment elevation of ≥1 mm in ≥2 contiguous leads or new left bundle branch block or true posterior myocardial infarction were randomized 1:1 to bivalirudin (Angiomax, Medicines Company, Parsippany, New Jersey) versus unfractionated heparin plus a glycoprotein IIb/IIIa inhibitor. The principal exclusion criteria included contraindications to any of the study medications, previous administration of fibrinolytic therapy, bivalirudin, glycoprotein IIb/IIIa inhibitor, low-molecular-weight heparin, or fondaparinux for the present admission, current use of VKAs, a history of bleeding diathesis, conditions predisposing to hemorrhagic risk or a refusal to receive blood transfusions, stroke or transient ischemic attack within 6 months or any permanent neurologic deficit, recent or known platelet count <100,000 cells/mm 3 or hemoglobin <10 g/dl, a planned elective surgical procedure that would necessitate thienopyridine interruption within 6 months of enrollment, coronary stent implantation within 30 days, and noncardiac co-morbid conditions with a life expectancy of <1 year or that might result in protocol noncompliance.
Emergent coronary angiography with left ventriculography was performed after the randomization assignment, followed by triage to PCI, coronary artery bypass grafting, or medical management at physician discretion. After angiography, 3,006 PCI patients were randomized again in a 3:1 ratio to either Taxus Express paclitaxel-eluting stents or otherwise identical uncoated Express bare metal stents (both Boston Scientific, Natick, Massachusetts).
Bivalirudin was administered as an intravenous bolus of 0.75 mg/kg, followed by an infusion of 1.75 mg/kg/hour. Unfractionated heparin was administered as an intravenous bolus of 60 IU/kg, with subsequent boluses titrated by a nomogram to a target activated clotting time of 200 to 250 seconds. Glycoprotein IIb/IIIa inhibitors were to be administered before PCI for all patients in the unfractionated heparin arm but were allowed by protocol in the bivalirudin arm only for refractory no reflow or giant thrombus after PCI.
Aspirin 324 mg chewed or 500 mg intravenously was given in the emergency room, followed by 300 to 325 mg/day orally during the hospitalization and 75 to 81 mg/day indefinitely thereafter. A clopidogrel loading dose (either 300 or 600 mg orally per the investigator’s discretion) was administered before catheterization, followed by 75 mg/day for ≥6 months. A ticlopidine 500-mg loading dose plus 250 mg/day was administered if clopidogrel was unavailable or the patient was allergic. Dual antiplatelet therapy was recommended for ≥1 year. Treatment with VKAs within the index hospitalization or at discharge was at the physician’s discretion. The use of, and indications for, VKAs were reported in the standardized electronic case report form by the site investigator.
The clinical outcomes within the index hospitalization and at 30 days and 1 year of follow-up were compared between the patients who received triple therapy (i.e., aspirin, thienopyridine, and VKA) and those who received dual therapy (i.e., aspirin and thienopyridine) during the index hospitalization. There were 2 primary 30-day end points in the HORIZONS-AMI trial: major bleeding (not related to coronary artery bypass graft surgery) and net adverse clinical events (i.e., major bleeding or composite major adverse cardiovascular events, including death, reinfarction, target vessel revascularization [TVR] for ischemia, or stroke). The definitions of the components of major adverse cardiovascular events have been previously described.
Major bleeding included intracranial or intraocular hemorrhage, access site bleeding requiring intervention, or hematoma ≥5 cm in diameter; hemoglobin decrease ≥4 g/dl without or ≥3 g/dl with an overt bleeding source; reoperation for bleeding; and blood transfusion. An independent Clinical Events Committee who were unaware of the treatment assignment adjudicated all primary end point and stent thrombosis events using the original source documents.
The categorical outcomes were compared using the chi-square test or Fisher’s exact test. Continuous variables were compared using the Wilcoxon rank sum test. The primary event analyses were performed using time-to-event data (for which the patients were censored at withdrawal from the study or the last follow-up visit), are displayed using the Kaplan-Meier method, and were compared using the log-rank test. The 30-day and 1-year event rates are expressed as Kaplan-Meier estimates.
The multivariate analysis of predictors of major bleeding at 30 days and 1 year was performed using logistic regression analysis with stepwise selection using an entry and exit criterion of p <0.1. The candidate variables entered into the model included age, gender, co-morbidities (hypertension, hyperlipidemia, diabetes mellitus, peripheral vascular disease, and congestive heart failure), smoking status, previous PCI, previous coronary artery bypass surgery, previous myocardial infarction, Killip class on admission, baseline laboratory values (presence or absence of anemia [hematocrit value at initial presentation <39% for men and <36% for women], chronic renal insufficiency [creatinine clearance of <60 ml/min using the Cockcroft–Gault equation], platelet count, and white blood cell count), and index hospitalization details (duration from symptom onset to first balloon inflation, antithrombotic randomization, clopidogrel loading dose, use of heparin prerandomization, and prescription of VKA).
Results
Of the 3,602 patients with STEMI enrolled in the HORIZONS-AMI trial, 3,345 were triaged to primary PCI as an initial management strategy and 3,320 (99.3%) had data available on the use of aspirin, thienopyridine, and VKA. Triple therapy was initiated during the index hospitalization or at discharge in 126 (3.8%) of 3,320 patients. A total of 3,194 patients (96.2%) were prescribed dual antiplatelet therapy. The most frequent indications for VKA treatment included a severely reduced left ventricular ejection fraction with a large akinetic region (30 patients [23.8%]), atrial fibrillation (30 patients [23.8%]), and the presence of mural thrombus in the left ventricle (29 patients [23.0%]) followed by left ventricular aneurysm (12 patients [9.5%]) and deep vein thrombosis (3 patients [2.4%]). No reason for warfarin use was given for 22 patients (19.9%).
The baseline clinical characteristics, angiographic features, and procedural data are presented in Tables 1 and 2 . The assignment to triple antithrombotic therapy was associated with a lower left ventricular ejection fraction and greater rates of left anterior descending artery infarcts. Other significant associations with the use of triple therapy were older age, female gender, history of rhythm disturbances, a greater incidence of Killip class >1 on admission, and lower rates of final Thrombolysis In Myocardial Infarction (TIMI) flow grade 3.
| Variable | Triple Therapy (n = 126) | Dual Therapy (n = 3,194) | p Value |
|---|---|---|---|
| Age (years) | 62.3 (54.0–71.4) | 159.9 (52.4–69.6) | 0.05 |
| Men | 69.0% | 77.5% | 0.03 |
| Diabetes mellitus | 18.3% | 16.2% | 0.54 |
| Hypertension | 55.6% | 52.5% | 0.51 |
| Current smoking | 36.8% | 47.5% | 0.02 |
| Previous heart failure | 7.1% | 2.5% | 0.006 |
| Previous myocardial infarction | 7.9% | 8.6% | 0.34 |
| Previous percutaneous coronary intervention | 9.5% | 10.5% | 0.72 |
| History of rhythm disturbances | 7.9% | 2.7% | 0.003 |
| Killip class >1 on admission | 13.5% | 8.3% | 0.04 |
| Hemoglobin (g/dl) | 14.6 (13.8–15.6) | 14.6 (13.6–15.5) | 0.64 |
| Hematocrit (%) | 43.0 (40.5–46.5) | 43.0 (40.0–45.0) | 0.48 |
| Creatinine clearance <60 ml/min | 3.2% | 2.8% | 0.78 |
| Creatinine clearance (ml/min) | 85.9 (65.8–100.8) | 89.2 (68.4–113.9) | 0.095 |
| Variable | Triple Therapy (n = 126) | Dual Therapy (n = 3,194) | p Value |
|---|---|---|---|
| Infarct-related coronary artery | |||
| Left anterior descending | 76.1% | 39.3% | <0.0001 |
| Right coronary | 6.7% | 16.2% | 0.003 |
| Left circumflex | 14.9% | 43.0% | <0.0001 |
| Thrombolysis In Myocardial Infarction flow | |||
| Baseline | |||
| Grade 0 or 1 | 73.1% | 65.0% | 0.052 |
| Grade 2 | 14.2% | 16.1% | 0.56 |
| Grade 3 | 12.7% | 18.9% | 0.07 |
| Final | |||
| Grade 0 or 1 | 3.0% | 2.0% | 0.36 |
| Grade 2 | 9.7% | 5.8% | 0.06 |
| Grade 3 | 87.3% | 92.1% | 0.04 |
| Stent randomization | |||
| Paclitaxel-eluting stent | 71.3% | 75.2% | 0.36 |
| Bare metal stent | 28.7% | 24.8% | 0.36 |
| Randomization to drug | |||
| Bivalirudin | 51.6% | 50.2% | 0.76 |
| Unfractionated heparin plus glycoprotein IIa/IIIb inhibitor | 48.4% | 49.8% | 0.76 |
| Total stent length (mm) | 24.0 (20.0–40.0) | 24.0 (20.0–36.0) | 0.48 |
| Left ventricular ejection fraction (%) | 40 (35–49) | 50 (45–59) | <0.0001 |
| Left ventricular ejection fraction <40%, site reported | 42.9% | 13.5% | <0.0001 |
| Symptom onset to balloon time (min) | 3.77 (2.70–5.78) | 3.70 (2.67–5.57) | 0.54 |
The rates of in-hospital mortality and reinfarction did not differ significantly between the patients treated with triple or dual therapy ( Table 3 ). A total of 2 cases of ischemic stroke and 5 of definite stent thrombosis with subsequent urgent ischemic TVR occurred in patients in the triple therapy group. However, all these events occurred before VKA therapy was initiated. Protocol-defined major bleeding was nearly 3 times more common in the triple than in the dual therapy group owing to the more frequent occurrence of retroperitoneal bleeding, decrease in hemoglobin ≥3 g/dl with an overt source of bleeding, decrease in hemoglobin ≥4 g/dl without an overt source of bleeding, and the need for blood transfusions. Protocol-defined minor bleeding was also more frequent in patients treated with triple therapy, including gastrointestinal bleeding (4.8% vs 0.9%, p = 0.0004), ecchymosis (5.7% vs 2.2%, p = 0.03), and epistaxis (1.9% vs 0.3%, p = 0.05). Triple therapy was also associated with significantly greater rates of minor TIMI (10.3% vs 3.4%, p = 0.0005), major or minor TIMI (16.7% vs 6.6%, p <0.0001), Global Use of Strategies to Open Occluded Coronary Arteries (GUSTO) moderate (7.1% vs 3.1%, p = 0.02), and GUSTO moderate/severe/life-threatening bleeding (8.7% vs 3.5%, p = 0.006), and a trend toward a greater incidence of TIMI major bleeding (6.3% vs 3.2%, p = 0.07).
| Event | Triple Therapy (n = 105) | Dual Therapy (n = 3,212) | p Value |
|---|---|---|---|
| Net adverse clinical events | 23 (21.9%) | 296 (9.2%) | <0.0001 |
| Major adverse cardiac events | 8 (7.6%) | 129 (4.0%) | 0.08 |
| Death | 0 (0.0%) | 63 (2.0%) | 0.27 |
| Cardiac death | 0 (0.0%) | 57 (1.8%) | 0.26 |
| Reinfarction | 3 (2.9%) | 39 (1.2%) | 0.15 |
| Stroke | 2 (1.9%) | 12 (0.4%) | 0.07 |
| Ischemic stroke | 2 (1.9%) | 12 (0.4%) | 0.07 |
| Hemorrhagic stroke | 0 (0.0%) | 0 (0.0%) | — |
| Ischemic target vessel revascularization | 6 (5.7%) | 57 (1.8%) | 0.01 |
| Stent thrombosis | 5 (4.3%) | 49 (1.6%) | 0.046 |
| Major bleeding | 18 (17.1%) | 209 (6.5%) | <0.001 |
| Intracranial bleeding | 0 (0.0%) | 0 (0.0%) | — |
| Intraocular bleeding | 0 (0.0%) | 0 (0.0%) | — |
| Retroperitoneal bleeding | 2 (1.9%) | 12 (0.4%) | 0.07 |
| Hematoma ≥5 cm at puncture site | 3 (2.9%) | 57 (1.8%) | 0.44 |
| Decrease in hemoglobin ≥3 g/dl with overt source | 5 (4.8%) | 59 (1.8%) | 0.05 |
| Decrease in hemoglobin ≥4 g/dl without overt source | 10 (9.5%) | 108 (3.4%) | 0.004 |
| Decrease in hemoglobin >5 g/dl without overt source | 4 (3.8%) | 53 (1.7%) | 0.10 |
| Reoperation for bleeding | 1 (1.0%) | 3 (0.1%) | 0.12 |
| Blood product transfusion | 8 (7.6%) | 83 (2.6%) | 0.007 |
| Minor bleeding | 27 (25.7%) | 379 (11.8%) | <0.0001 |
The incidence of VKA discontinuation among the patients who were prescribed VKA during index hospitalization was 14.3% (15 of 105 patients).
The 30-day and 1-year outcomes of patients who received triple or dual therapy are shown in Figures 1 and 2 . The rates of all-cause and cardiac mortality, reinfarction, and ischemic TVR did not differ significantly between the 2 groups at any study point. One additional case of ischemic stroke and another of hemorrhagic stroke in the triple therapy group resulted in greater rates of stroke at 1 year of follow-up in the triple therapy group.
