Randomized controlled trials assessing new drugs and devices tend to exclude subjects who are at greatest risk. The Harmonizing Outcomes With Revascularization and Stents in Acute Myocardial Infarction (HORIZONS-AMI) trial incorporated broader inclusion criteria in an attempt to include a more representative spectrum of patients presenting with ST-segment elevation myocardial infarction (STEMI). To identify the differences between this modern trial and the real world, we analyzed the characteristics and outcomes of patients with STEMI who were screened but not enrolled at a high-volume recruiting center. Of 318 consecutive patients with STEMI who were screened, 200 (62.9%) were randomized, and 118 (37.1%) were excluded. The baseline characteristics and 30-day and 1-year clinical outcomes were compared in the 2 groups. The excluded patients had numerous high-risk features compared to those randomized, including being older (67.0 ± 12.8 vs 63.0 ± 11.4 years, p = 0.004), more often had had a previous MI (34.7% vs 8.0%, p <0.001), Killip class III-IV (27.4% vs 4.0%, p <0.001), and lower hemoglobin (13.4 ± 2.3 vs 14.8 ± 1.5 g/dl, p <0.001). The excluded patients had markedly greater 30-day and 1-year rates of all-cause mortality (17.4% vs 2.0%, p <0.001, and 27.6% vs 2.5%, p <0.001, respectively), major adverse cardiovascular events (death, MI, ischemia-driven target vessel revascularization, and stroke), major bleeding, and net adverse clinical events (major adverse cardiovascular events or major bleeding). On multivariate analysis, Killip class III-IV at presentation, age, left ventricular ejection fraction, and final Thrombolysis In Myocardial Infarction flow grade 3 were independent predictors of outcome. In conclusion, despite the broadened entry criteria of the HORIZONS-AMI trial, 37.1% of all patients presenting with STEMI at a center with a high rate of enrollment were judged to be ineligible and were excluded. The excluded patients had a significantly greater risk profile and markedly increased mortality and adverse events compared to the trial-eligible group.
The treatment recommendations for treating patients with acute ST-segment elevation myocardial infarction (STEMI) rely heavily on the results of randomized controlled trials (RCTs). However, RCTs of primary percutaneous coronary intervention (PCI) in STEMI often recruit relatively low-risk patients, making it difficult to generalize the results to a “real-world” population. The Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction (HORIZONS-AMI) trial was designed to allow enrollment of a broader cross-section of patients than included in earlier RCTs. In brief, in the HORIZONS-AMI trial, 3,602 patients with STEMI symptom onset within 12 hours were randomized to either unfractionated heparin plus the routine use of a glycoprotein IIb/IIIa inhibitor or bivalirudin monotherapy. After angiography, eligible patients underwent a second randomization (3:1) to either paclitaxel-eluting stents or bare metal stents. Patients of any age, including those with cardiogenic shock at admission, previous cardiac arrest, renal insufficiency, multivessel disease, and protected left main or saphenous vein graft infarct lesions were eligible. The principal results of the HORIZONS-AMI trial have been previously reported. A screening log of excluded patients was not routinely maintained at all sites. To analyze the differences between the trial and the real world, we examined the baseline clinical and angiographic characteristics and long-term outcomes of patients screened but excluded from the HORIZONS-AMI trial at a single, high-volume, recruiting center.
Methods
The present study was conducted at Ospedali Riuniti di Bergamo (Bergamo, Italy), the second greatest recruiting center in the HORIZONS-AMI trial. Patient eligibility was sequentially determined for all consecutive patients with STEMI and an indication for primary PCI, who presented at our center during the trial recruitment period (August 10, 2005 to May 7, 2007). The exclusion criteria have been previously reported. In brief, because the trial tested an anticoagulant drug (bivalirudin), the clinical conditions determining an increased bleeding risk were considered as the main exclusion criteria. Noncardiac coexisting conditions that could limit life expectancy to <1 year or that might interfere with protocol compliance and an inability to obtain written informed consent were also considered as exclusion criteria.
Patients who did not meet any of the entry criteria for randomization but otherwise had indications for primary PCI were included in the present study as the “trial-ineligible” group. Informed consent for primary PCI was obtained from each patient or patient representative, when needed. Drug and device usage for randomized patients enrolled in the HORIZONS-AMI trial have been previously reported in detail. All excluded patients were treated with 250 mg intravenous aspirin and an intravenous bolus (70 U/kg) of unfractionated heparin before cardiac catheterization. During PCI, additional heparin boluses were given to maintain the activating clotting time at 250 to 300 seconds. The activating clotting time was routinely tested during the procedure in all patients. Glycoprotein IIb/IIIa inhibitor use was at the discretion of the operator but discouraged in the case of a high risk of bleeding. A loading dose of clopidogrel 300 mg was administered immediately before PCI in all patients. PCI was performed by the femoral or radial approach, according to the physician’s decision. The radial route was preferentially selected in presence of bleeding diathesis or known thrombocytopenia. Manual thrombus aspiration was used in patients with a large thrombus burden. Balloon angioplasty and stenting were performed according to standard practice. The choice to implant a bare metal stent or drug-eluting stent was left to operator discretion. After the procedure, the femoral arterial access sheath was removed when the activating clotting time was <180 seconds or immediately if the radial approach had been used. After successful PCI, all patients (enrolled and excluded) were treated with aspirin 100 mg once/day indefinitely and clopidogrel 75 mg/day for ≥6 months if a drug-eluting stent had been implanted (typically 12 months) or for ≥1 month if a bare metal stent had been used.
Detailed clinical, procedural, and outcomes data for the enrolled patients were collected and entered into the electronic HORIZONS-AMI case report form. A data set with common elements was collected from the excluded patients to allow a comparison of the 2 groups. In-hospital events were tracked, and any major adverse cardiovascular events (MACE; composite of death, MI, ischemia-driven target vessel revascularization, and stroke) and major bleeding events, according to the HORIZONS-AMI definitions, were evaluated and recorded. Patients alive at hospital discharge were followed by telephone interviews or outpatient clinical visits at 30 days, 6 months, and 1 year. We sought to determine the rates of all-cause and cardiac mortality at 30 days and 1 year. MACE, non–coronary artery bypass grafting (CABG)-related major bleeding, and net adverse clinical events (MACE and non–CABG major bleeding), according to HORIZONS-AMI protocol definitions, were assessed in the enrolled and excluded patients.
Continuous variables are summarized as the mean ± SD and were compared using the t test. Categorical variables are displayed as frequencies and percentages, and the groups were compared using the chi-square test. In addition, the outcomes data for the randomized and trial-ineligible groups were plotted as time-to-event curves using Kaplan-Meier estimates and compared using the log-rank test. Hazard ratios and 95% confidence intervals from the Cox proportional hazards models are displayed. Independent predictors of 1-year mortality, MACE, non–CABG major bleeding, and net adverse clinical events were evaluated using Cox proportional hazards regression analysis. Additional covariates known to affect the outcomes in STEMI were entered into the model in a stepwise fashion, including age, left ventricular ejection fraction, diabetes, Killip class, renal insufficiency, use of bivalirudin, and final Thrombolysis In Myocardial Infarction 3 flow (vs 0–2 flow). Variables that were significant at the 0.05 level were kept in the model. All analyses were 2-sided, and p <0.05 defined statistical significance.
Results
A total of 318 patients were screened for enrollment. Of these, 200 (62.9%) were enrolled and randomized in the HORIZONS-AMI trial and 118 (37.1%) were considered trial ineligible for specific reasons ( Table 1 ). The most common reasons for trial ineligibility were increased bleeding risk, intubation after cardiac arrest, and extreme hemodynamic instability. Other common reasons included ST-segment resolution on the qualifying electrocardiogram (despite earlier definite ST-segment elevation), and suspected stent thrombosis. The reason for trial exclusion was unclear in 7 patients (5.9%), and in 14 patients (11.9%), multiple exclusion criteria were present.
| Exclusion Criteria | n (%) |
|---|---|
| Increased bleeding risk | 28 (23.7) |
| Intubation after cardiac arrest or severe hemodynamic instability | 26 (22.0) |
| ST-segment resolution on baseline electrocardiogram | 16 (13.6) |
| Suspected stent thrombosis | 12 (10.2) |
| Allergy to study medications | 8 (6.8) |
| Life expectancy <1 year for noncardiac reasons | 6 (5.1) |
| Late presentation (>12 h) | 4 (3.4) |
| Cognitive impairment | 4 (3.4) |
| Consent refused | 7 (5.9) |
| Enrolled in another trial | 3 (2.5) |
| Psychiatric illness (unlikely to comply with study procedures) | 2 (1.7) |
| Human immunodeficiency virus infection | 1 (0.8) |
| Undetermined reasons for exclusion | 7 (5.9) |
| Multiple reasons for exclusion | 14 (11.9) |
The baseline characteristics of the 2 groups are listed in Table 2 . Compared to the enrolled patients, those excluded were older, more frequently had diabetes, renal insufficiency (creatinine clearance <60 ml/min as calculated using the Cockcroft–Gault equation), previous PCI and CABG, Killip class III-IV, and lower left ventricular ejection fraction, and had a greater heart rate and lower baseline hemoglobin level. Procedural success, defined as final Thrombolysis In Myocardial Infarction 3 flow, was similar in the 2 groups ( Table 3 ). Radial artery access was more common among the excluded patients. Significant differences were also present between the 2 groups in the type of stent implanted, use of procedural anticoagulation and glycoprotein IIb/IIIa inhibitors, and intra-aortic balloon counterpulsation. The mean duration of hospital stay was >3 days longer in the trial-ineligible patients (10.5 ± 14.3 vs 7.0 ± 6.0 days, p = 0.004), mainly because of the critical conditions at presentation of many of the excluded patients, such as those with cardiac arrest and Killip class III or IV. Follow-up data of the excluded patients were available for 100% at 30 days and 97.5% at 1 year. The excluded patients had markedly greater mortality at 30 days and 1 year compared to those randomized ( Table 4 and Figure 1 ). The 30-day and 1-year mortality were significantly greater for the excluded patients presenting in Killip class I or II (7.8% vs 1.0%, p <0.001, and 15.6% vs 1.0%, p <0.001, respectively). Similar trends (which did not reach statistical significance) were observed in patients presenting in Killip class III or IV (46.4% vs 25.0%, p = 0.42, and 64.3% vs 37.5%, p = 0.24, respectively). The 30-day and 1-year rates of cardiac death, MACE, non–CABG major bleeding, and net adverse clinical events were also substantially greater in the patients ineligible for trial randomization ( Table 4 and Figure 1 ). Noncardiac death did not occur during follow-up among the 6 patients excluded because of a life expectancy of <1 year. On multivariable analysis, Killip class III-IV at presentation, age, left ventricular ejection fraction, and final Thrombolysis In Myocardial Infarction flow grade 3 were the independent predictors of outcome ( Table 5 ).
| Variable | Enrolled and Randomized (n = 200) | Excluded (n = 118) | p Value |
|---|---|---|---|
| Age (yrs) | 63.0 ± 11.4 | 67.0 ± 12.8 | 0.004 |
| Age >75 yrs | 31 (15.5%) | 39 (33.1%) | <0.001 |
| Women | 43 (22.5%) | 28 (23.7%) | 0.65 |
| Diabetes mellitus | 27 (13.5%) | 27 (22.9%) | 0.03 |
| Hyperlipidemia ∗ | 95 (47.5%) | 51 (43.2%) | 0.46 |
| Hypertension | 87 (43.5%) | 62 (52.5%) | 0.12 |
| Smoker | 119 (59.5%) | 51 (43.2%) | 0.005 |
| Renal failure † | 2 (1%) | 21 (17.8%) | <0.001 |
| Previous myocardial infarction | 16 (8%) | 41 (34.7%) | <0.001 |
| Previous percutaneous coronary intervention | 17 (8.5%) | 29 (24.6%) | 0.002 |
| Previous coronary artery bypass grafting | 3 (1.5%) | 9 (7.6%) | 0.001 |
| Killip class III or IV | 8 (4.0%) | 28 (23.7%) | <0.001 |
| III | 3 (1.5%) | 7 (5.9%) | 0.043 |
| IV | 5 (2.5%) | 21 (17.8%) | <0.001 |
| Left ventricular ejection fraction (%) | 52.2 ± 7.1 | 44.1 ± 12.4 | <0.001 |
| Heart rate (beats/min) | 78 ± 17 | 85 ± 24 | 0.006 |
| Hemoglobin (g/dl) | 14.8 ± 1.5 | 13.4 ± 2.3 | <0.001 |
| Serum creatinine (mg/dl) | 1.0 ± 0.3 | 1.3 ± 1.0 | 0.053 |
∗ Defined as total cholesterol >240 mg/dl and/or low-density lipoprotein >130 mg/dl, and/or triglycerides >150 mg/dl (National Cholesterol Education Program Adult Treatment Panel III classification).
† Defined as creatinine clearance of <60 ml/min as calculated using Cockcroft–Gault equation.
| Variable | Enrolled and Randomized (n = 200) | Excluded (n = 118) | p Value |
|---|---|---|---|
| Radial arterial access | 35 (17.5%) | 31 (26.3%) | <0.001 |
| Baseline Thrombolysis In Myocardial Infarction flow grade 0–1 | 139 (72.8%) | 84 (71.2%) | 0.76 |
| Final Thrombolysis In Myocardial Infarction flow grade 3 | 184 (92%) | 105 (89.0%) | 0.37 |
| Infarct-related artery | |||
| Left main artery | 4 (2.1%) | 5 (4.2%) | 0.27 |
| Left anterior descending artery | 96 (48%) | 55 (46.6%) | 0.81 |
| Left circumflex artery | 32 (16%) | 21 (17,8%) | 0.68 |
| Right artery | 72 (36%) | 40 (33,9%) | 0.71 |
| Saphenous vein graft | 2 (1.0%) | 3 (2.5%) | 0.31 |
| Procedural anticoagulation | |||
| Unfractionated heparin | 100 (50.0%) | 118 (100%) | <0.001 |
| Bivalirudin | 100 (50.0%) | 0 (0%) | <0.001 |
| Glycoprotein IIb/IIIa inhibitor used | 107 (53.5%) | 29 (24.6%) | <0.001 |
| Interval from symptom onset to balloon inflation (h) | 4.5 ± 2.5 | 3.9 ± 2.6 | 0.03 |
| ≥1 Stents implanted | 184 (92.0%) | 105 (89.0%) | 0.37 |
| Any bare metal stent implanted | 52 (28.6%) | 76 (64.4%) | <0.001 |
| Any drug-eluting stent implanted | 134 (73.6%) | 29 (24.6%) | <0.001 |
| Thrombus aspiration | 27 (14.4%) | 19 (16.1%) | 0.68 |
| Intra-aortic balloon counterpulsation | 7 (3.5%) | 23 (19.5%) | <0.001 |
| 30-Day Outcomes | 1-Year Outcomes | |||||
|---|---|---|---|---|---|---|
| Enrolled and Randomized (n = 200) | Excluded (n = 118) | p Value | Enrolled and Randomized (n = 200) | Excluded (n = 115) | p Value | |
| All-cause death | 4 (2.0%) | 20 (17.1%) | <0.001 | 5 (2.5%) | 32 (27.6%) | <0.001 |
| Cardiac death | 4 (2.0%) | 20 (17.1%) | <0.001 | 5 (2.5%) | 26 (22.0%) | <0.001 |
| Reinfarction | 3 (1.5%) | 2 (1.9%) | 0.81 | 9 (4.6%) | 5 (5.7%) | 0.79 |
| Ischemia-driven target vessel revascularization | 2 (1.0%) | 1 (1.0%) | 0.96 | 10 (5.2%) | 5 (5.7%) | 0.88 |
| Major adverse cardiovascular events | 7 (3.5%) | 22 (18.8%) | <0.001 | 20 (10.0%) | 42 (36.6%) | <0.001 |
| Major bleeding (noncoronary artery bypass grafting) | 25 (12.5%) | 30 (25.4%) | 0.003 | 25 (12.6%) | 30 (26.1%) | 0.003 |
| Net adverse clinical events | 30 (15.0%) | 42 (35.6%) | <0.001 | 39 (19.6%) | 49 (41.7%) | <0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
