Author + information
- Received July 7, 2000
- Revision received August 30, 2000
- Accepted October 4, 2000
- Published online February 1, 2001.
- ↵*Reprint requests and correspondence: Dr. Salim Yusuf, Canadian Cardiovascular Collaboration Project Office, Hamilton General Hospital, McMaster Clinic, Room 252, 237 Barton St. East, Hamilton, Ontario L8L 2X2 Canada
We sought to evaluate whether oral anticoagulant (AC) therapy given for five months was superior to standard (control) therapy in patients with unstable angina receiving aspirin.
The long-term risk of myocardial infarction (MI) or death remains high in patients with unstable angina, despite the use of aspirin. Therefore, additional treatments are necessary.
Of the 10,141 patients entering the main trial, 3,712 were randomized 12 to 48 h later to receive oral AC therapy (n = 1,848) or standard therapy (n = 1,864).
One-hundred forty patients (7.6%) suffered from cardiovascular death, MI or stroke while receiving oral AC, compared with 155 patients (8.3%) on standard therapy (relative risk [RR] 0.90, 95% confidence interval [CI] 0.72 to 1.14; p = 0.40). The rates of the primary outcomes plus refractory angina were 16.7% (n = 308) versus 17.5% (n = 327) (RR 0.95, 95% CI 0.81 to 1.11; p = 0.53). Countries were divided into good or poor compliers (based on the use of oral AC above or below 70% at 35 days), without knowledge of results by country. In good-complier countries, oral AC was discontinued in only 10.4% of patients at seven days and in 23.6% by five months, compared with 27.6% and 44.9%, respectively, in poor complier countries. There were significant reductions in the risks of both the primary (6.1% vs. 8.9%; RR 0.68, 95% CI 0.48 to 0.95; p = 0.02) and secondary outcomes (11.9% vs. 16.5%; RR 0.70, 95% CI 0.55 to 0.90; p = 0.005) with oral AC in the good-complier countries. There was little difference in the poor-complier countries (9.0% vs. 7.8% for the primary and 21.3% vs. 18.5% for the secondary outcomes, tests for interactions comparing the RRs for the primary and secondary outcomes were p < 0.02 and p = 0.002, respectively, between the two sets of countries). In the overall study, there was an excess of major bleeding (2.7% vs. 1.3%; p = 0.004), which was larger in the good-complier countries (RR 2.71) compared with the poor-complier countries (RR 1.58). There were also reductions in cardiac catheterization (RR 0.80; p = 0.004) and coronary revascularization procedures (RR 0.82; p = 0.06) in the good-complier countries, but not in the poor-complier countries (RR 0.98 and 1.06, respectively, p for interaction of 0.06 and 0.04, respectively).
Overall, oral AC led to a small, nonsignificant reduction in the risk of the primary and secondary outcomes. Stratifying the countries or centers by their rates of compliance to oral AC suggested that good compliance to oral AC could potentially lead to clinically important reductions in major ischemic cardiovascular events.
Nonocclusive coronary thrombosis has been implicated as a common cause of unstable angina (UA) (1). Antithrombotic agents, such as heparin (2)or hirudin (3), and antiplatelet agents, such as aspirin (4)or glycoprotein IIb/IIIa receptor antagonists (5), are effective in decreasing the risk of myocardial infarction (MI) or refractory ischemia. Only long-term aspirin has been shown to reduce the risk of cardiac deaths and MI in patients with UA (6). Despite the use of aspirin, ∼10% to 15% of patients still experience death or MI by one year, and ∼20% of patients are readmitted to the hospital with UA. Recent angioscopic studies indicate that coronary thrombi persist even at one month after the acute UA episode in patients taking aspirin (7), and markers of thrombin generation remain elevated in apparently asymptomatic patients. These considerations suggest that additional long-term antithrombotic strategies may be potentially beneficial.
Oral anticoagulants (ACs) have been demonstrated to reduce the risk of death, MI or stroke after acute MI (8). However, these results are derived from trials in which aspirin was not used, the intensity of anticoagulation was high (International Normalized Ratio [INR] >2.8) and a substantial degree of major bleeding occurred. We recently conducted two pilot trials using varying intensities of warfarin compared with standard (control) therapy in patients with UA receiving aspirin (9). These studies indicated that low intensity (INR <1.5) oral AC therapy was ineffective, whereas a moderate intensity (INR 2.0 to 2.5) was promising. These results are supported by the results of a previous small trial in which aspirin was concomitantly used (10)and in several post-MI trials where aspirin was not used (8). Therefore, in the Organization to Assess Strategies for Ischemic Syndromes (OASIS-2) substudy, we evaluated the effects of oral AC therapy (target INR 2.5) in addition to aspirin, compared with aspirin alone, when given up to five months, on the composite outcome of cardiovascular death, MI or stroke (primary outcome) and cardiovascular death, MI, stroke and readmission to the hospital for UA (secondary outcome).
Patients eligible for the main OASIS-2 trial, which compared a three-day regimen of hirudin vs. heparin (3), were those who could be randomized within 12 h of an episode of chest pain suspected to be due to UA or MI without ST segment elevation on their admission electrocardiogram. The diagnosis of UA was based on symptoms of angina that were new in onset, worsening or occurring with minimal activity. Detailed inclusion and exclusion criteria for the main trial have been previously published. Additional exclusions for the warfarin part of the study were a clear indication for warfarin, bleeding during heparin or hirudin, coronary artery bypass graft surgery planned within a week, normal coronary anatomy, contraindications to oral AC therapy and physician or patient reluctance.
Randomization and treatment regimens
The OASIS-2 trial utilized a partial 2 × 2 factorial design. Details of the patients’ initial randomization to heparin or hirudin and their treatment regimen have been published previously (3). The AC therapy component (warfarin in all countries, except for Hungary, where dicumarol was used) was a randomized, open trial with blinded adjudication of outcomes. Patients were randomized between 12 and 30 h after the initial randomization to heparin or hirudin, but this could be extended up to seven days in some circumstances. Patients were randomized by a toll-free telephone call to a 24-h automated randomization service. After key data were recorded, the patients were allocated to receive warfarin (loading dose of 10 mg followed by 3 mg/day for two days) or to a control group for five months. Further dosing adjustments were based on the INR values and were aimed to achieve an INR between 2.0 and 2.5. Patients were followed at one, three and five months. All patients were followed for another month to assess whether there was any “rebound” in clinical events. All patients gave written, informed consent, and the protocol was approved by the Institutional Review Board of each hospital.
The primary follow-up period was five months. The primary outcome was a comparison of the rates of cardiovascular death, MI or strokes, and the secondary analysis included these outcomes plus readmission to the hospital for UA. Definitions for all outcomes have been previously published (3).
Follow-up and blinded central adjudication of events
The one-month data are available for 3,708 patients (99.9%); three-month data for 3,703 patients (99.8%); and five-month data for 3,697 patients (99.6%). Key efficacy outcomes (death, MI, strokes and readmission to the hospital for UA) and safety outcomes (major bleeding) were adjudicated by a central committee that had no knowledge of the treatment allocation. There was a high degree of agreement in the classification of events between the clinical sites and central committee (agreement rates of 99.3% for primary outcomes and 97.7% for secondary outcomes). Therefore, the steering committee decided to accept the clinical classification for the few events that could not be adjudicated (n = 4).
We anticipated an event rate of 9% for the composite outcome of cardiovascular death/new MI/stroke and 20% for cardiovascular death/MI/stroke/readmission to the hospital for UA at five months. With 4,000 patients, we would have 80% power (2α = 0.05) to detect a 26% relative risk (RR) reduction in the primary outcome and a 21% RR reduction in the secondary outcome.
The rates of the primary and secondary outcomes were compared by using an intention-to-treat analysis with the Mantel-Haenszel (log-rank) test, stratified by center (11). Although an efficacy analysis was prespecified in the protocol, its exact nature was only determined after observing the unexpectedly large variation in compliance to oral ACs between various countries (49% to 90%) at one month. Details of the analyses were specified with no knowledge of event rates or treatment effects; only compliance rates were known.The countries were ranked solely according to their rates of compliance to allocated oral ACs at 35 days (a time when approximately one-half of the events had occurred). Those countries with ≥70% of patients receiving study medication at 35 days (six countries; n = 1,821) were considered to be “good” compliers, and those countries with <70% were considered to be “poor” compliers (eight countries; n = 1,891). The 70% cutoff also approximately corresponded to the average compliance rates in previous post-MI trials of warfarin with a positive result (11). The same analyses were conducted by dividing centers into good compliers (≥70% compliance at 35 days; 149 centers, n = 2,008) and poor compliers (<70% compliance at 35 days; 140 centers, n=1704).
An independent Data and Safety Monitoring Board conducted three interim analyses. For efficacy, the combined outcomes of cardiovascular death and new MIs were monitored using a boundary corresponding to 4 SDs for the first half of the study and 3 SDs for the second half. In addition, the boundaries had to be crossed on two successive analyses at least three months apart for consideration of early termination because of efficacy. These extreme boundaries require practically no adjustment of the p value, so p < 0.05 was considered to be statistically significant. An analysis stratified by randomization to hirudin or heparin yielded no evidence of heterogeneity. Therefore, only the overall results are presented.
During the period from August 1996 to April 1998, 10,141 patients were recruited into the main study comparing hirudin versus heparin (Fig. 1). The AC substudy was not implemented until May 1997, by which time 1,743 patients were already randomized into the main trial. Of the remaining patients, 4,686 were not included because of the physician’s reluctance to enroll patients (n = 1,814 [17.9%]), patient refusal (n = 1,246 [12.3%]), planned bypass surgery (n = 385 [3.8%]), indication for warfarin (n = 130 [1.3%]), contraindication to warfarin (n = 119 [1.2%]) or no significant coronary artery disease (n = 93 [0.9%]).
A total of 3,712 patients were randomized. Their baseline characteristics and those of the nonrandomized patients are summarized in Table 1. Apart from a higher rate of use of prerandomization heparin in excluded patients (which is explained by the varying rates of inclusion of patients by countries and the patterns of their heparin use), there were no clinically important differences in baseline characteristics.
Compliance and INR
Of the 1,848 patients assigned to oral AC therapy, the cumulative proportion of patients who had stopped warfarin was 18.3% at seven days, 28.7% at 35 days, 31.9% at three months and 36.3% at five months. The proportion of patients with INR ≥2.0 among those taking oral ACs was 55.3% at seven days, 70.2% at two weeks, 65.3% at 35 days, 71.8% at three months and 77.3% at five months. The main reasons for discontinuation of oral ACs included a reported need for cardiac catheterization, percutaneous transluminal coronary angioplasty or bypass surgery (n = 211 [11.4%]), patient refusal (n = 103 [5.5%]), bleeding (n = 84 [4.5%]) or physician decision (n = 77 [4.2%]). In the control group, 1.9% were prescribed an oral ACs at seven days, 3.4% at 35 days, 3.5% at three months and 3.3% at five months.
There were no differences in the rates of use of aspirin before randomization into the main study (64% and 65% in the warfarin and control groups, respectively). However, by discharge and during follow-up, more patients in the standard therapy arm were taking aspirin (97% vs. 92% at discharge and 93% vs. 83% at five months; p < 0.001 for both times). The use of all other drugs was similar at discharge between the two groups.
Efficacy and safety outcomes at five months (Table 2)
A total of 140 (7.6%) patients experienced the primary outcome in the oral AC group compared with 155 subjects (8.3%) in the control group (RR 0.90, 95% confidence interval [CI] 0.72 to 1.14; p = 0.40). Three hundred and eight (16.7%) patients experienced the secondary outcome in the oral AC group compared with 327 subjects (17.5%) in the control group (RR 0.95, 95% CI 0.81 to 1.1; p = 0.53). There was an excess of major bleeding (49 vs. 25; RR 2.0; p = 0.004) and minor bleeding (85 vs. 50; RR 1.73; p = 0.002) with oral ACs. The proportion of patients with life-threatening bleeding accounted for about one-half of all major bleeding cases (warfarin [n = 25] 1.4% vs. control [n = 14] 0.8%; p = 0.07). The proportion of patients undergoing cardiac catheterization was lower in the oral AC group (679 [36.8%] vs. 752 [40.4%]; p = 0.03), but there were no differences in the rates of coronary angioplasty (232 vs. 215) or bypass surgery (225 vs. 257).
Subgroup results by compliance (Table 3)
By country or center
Of the patients entering the main part (i.e., hirudin vs. heparin) of the trial, good-complier countries (35-day rates of oral AC use ≥70%) had lower rates of patient exclusions before the second randomization (AC vs. control) compared with poor-complier countries for physician (545 [13.0%] vs. 1,269 [21.3%]) or patient refusal (442 [10.5%] vs. 804 [13.5%]) or exclusions for planned bypass surgery (82 [2.0%] vs. 303 [5.1%]). These data indicate a greater willingness to participate in the oral AC part of the study in the former category of countries. This led to a higher proportion of patients being randomized in the good-complier countries (43.4%) compared with the poor-complier countries (31.8%; p < 0.001). In the good-complier countries, only 10.4% of patients had discontinued oral ACs at seven days and 23.6% at five months, compared with 27.6% and 44.9%, respectively, in the poor-complier countries (p < 0.001 at both times). The proportion of patients who achieved an INR >2.0 by hospital discharge was 64.1% in the good-complier countries compared with 35.2% in the poor-complier countries (p < 0.001).
In the good-complier countries,there was a significant reduction in the primary outcome of cardiovascular death, MI or stroke (55 [6.1%] in the oral AC group vs. 81 [8.9%] in the control group; RR 0.68, 95% CI 0.48 to 0.95; p = 0.02) compared with no effect in the poor-complier countries (85 [9.0%] vs. 74 [7.8%], respectively; RR 1.17, 95% CI 0.86 to 1.6; p = 0.33) (p = 0.02 [for heterogeneity]) (Fig. 2). A similar pattern was observed with the secondary outcome, with an apparent benefit observed in the good-complier countries (108 [11.9%] vs. 151 [16.5%], respectively; RR 0.70, 95% CI 0.55 to 0.90; p = 0.005), with no benefit in the poor-complier countries (200 [21.3%] vs. 176 [18.5%], respectively; RR 1.16, 95% CI 0.95 to 1.41; p = 0.16) (p = 0.002 [for heterogeneity]). An analysis dividing centersinto good and poor compliers provided consistent results for both the primary and secondary outcomes (Fig. 3).
The overall rates of procedures in the good compliance countries versus poor compliance countries were significantly lower (28.4% vs. 54.1%; p = 0.001). In addition, there were also reductions in the use of cardiac catheterization (22.8% vs. 28.7%; p = 0.004) and coronary revascularization (15.0% vs. 18.3%; p = 0.059) with oral ACs compared with standard (control) therapy in the good-complier countries. In contrast, there was no difference in these events between the active and control groups among the poor-complier countries (50.4% vs. 51.7% for catheterization and 32.9% vs. 30.8% for coronary revascularization) (p = 0.06 and 0.04 for interactions, respectively, comparing the results of the two categories of countries). The reasons for patients stopping medications in the poor-complier countries were more likely to be due to an invasive procedure (27.5% vs. 10.6%; p < 0.001) or physician decision (8.9% vs. 6.0%; p = 0.015). However, bleeding as the reason for stopping the medication was higher, but not significantly so, in the good-complier group (8.5% vs. 6.6%; p = 0.125), which is consistent with more intensive AC in these countries.
Compliance by patients
Figure 1provides the data on both the primary and secondary outcomes among patients who were adherent compared with those who were not in the active and control groups. The rates of primary and secondary outcomes were similar (high) in the patients who were not adherent (primary outcome: 14.9% [77 of 518] vs. 14.0% [12 of 83]; secondary outcome: 27.6% [143 of 518] vs. 26.5% [22 of 83]). However, among those adherent to the treatment allocation, both the primary and secondary outcomes appeared to be reduced substantially (primary outcome: 4.7% [63 of 1,330] active treatment vs. 8.0% [143 of 1,781] control; secondary outcome: 12.4% [165 of 1,330] active treatment vs. 17.1% [305 of 1,781] control). Although promising, there are several intrinsic limitations of this analysis, as discussed later.
Events in the month after cessation of therapy
The number of additional patients who had the primary outcome (14 vs. 15) or the secondary outcome (26 vs. 31) in the month after cessation of oral AC therapy was similar, indicating no obvious “rebound” in clinical events after cessation of warfarin.
Pooled analysis of trials of warfarin in unstable angina (Table 4)
When combining the data from the four studies that evaluated moderate-intensity oral ACs in addition to aspirin (OASIS-1 , OASIS-2, Antithrombotic Therapy in Acute Coronary Syndromes [ATACS] pilot and ATACS ), 7.4% (155 of 2,088) of patients suffered from death, MI or stroke in the warfarin group, compared with 8.8% (186 of 2,104) in the control group (RR 0.83, 95% CI 0.66 to 1.03; p = 0.10). If the pooled analysis included only the data from good-complier centers from OASIS-2, the combined results are more promising (RR 0.62, 95% CI 0.45 to 0.84), but these data should be interpreted with considerable caution.
Results vary by compliance
Our data suggest that despite the use of aspirin, the rates of cardiovascular death, MI and stroke are substantial in patients with UA during long-term follow-up. Overall, we observed a modest, nonsignificantly lower rate of ischemic events with oral ACs compared with standard (control) therapy. However, the overall results are likely to be an underestimate due to the high rates of noncompliance to oral ACs in some countries or centers. Therefore, a subgroup analysis of countries and centers based on compliance was performed. In countries or centers with high compliance to oral ACs, improved outcomes, lower rates of cardiac interventions and higher bleeding rates were observed among patients randomized to the active group compared with those in the control group. The differences in both the primary and secondary outcomes between oral AC and control therapy were also statistically heterogeneous when comparing the results among centers or countries with good versus poor compliance. The subgroup analysis by compliance was not originally specified in the protocol, as we did not anticipate the poor compliance observed in several countries. When we observed highly variable compliance rates, we ranked countries by compliance without knowledge of the results in individual countries. We used a “cutoff” of compliance at 35 days above or below 70% “blindly.” This divided the number of patients into two approximately equal halves with markedly different proportions achieving the target INR (64.1% vs. 35.2%). The 70% cutoff value also roughly corresponds to the average compliance in the post-MI trials of oral ACs (9), which indicated clinical benefit. The same cutoff value was also used in the analysis of results by compliance using each center as a unit. Although our approach does not have the force of a prospectively stated subgroup analysis, it is both biologically and pharmacologically plausible. Moreover, such an approach is methodologically valid, as it preserves an intention-to-treat analysis. Because each hospital was a strata for randomization, the subgroup analysis is a “proper” one, as it is based on a prerandomizationvariable, and the groups compared (active and control groups within each strata) would be expected to be similar (12). Second, the analysis was not “data-dredged,” but postulated and conducted without knowledge of the results. Third, the consistency of the results across three sets of efficacy outcomes (lower rates of primary outcomes of the composite of cardiovascular death/MI/stroke; lower rates of secondary outcomes of the composite of cardiovascular death/MI/stroke/readmission to the hospital for UA; and lower rates of invasive procedures), coupled with higher rates of major bleeding, is both biologically and pharmacologically plausible and consistent with external data. However, subgroup analyses may lead to an exaggeration of the magnitude of benefit observed among the good-complier countries/centers (approximately one-third risk reduction), and the true benefit in the good-complier countries may be lower. This possibility is counterbalanced by the fact that a significant proportion of patients (24%) had stopped taking oral ACs, even in the good-complier countries; only approximately one-third of patients had achieved an INR >2.0; and the proportion continuing to use aspirin after randomization was lower in the warfarin group. Therefore, the benefit among those who actually took oral ACs may be higher than that of the overall treatment and closer to the actual difference observed in the good-complier countries. This is consistent with an apparently large benefit (RR 0.59 for primary outcome and RR 0.72 for secondary outcomes) when patients who adhered to the original allocation were compared. However, such an analysis is based on a postrandomizationvariable, and the two groups (active vs. control) are not comparable. Nevertheless, these analyses are consistent with the more proper analyses (stratification by countries or by centers) and provide evidence that if a high level of adherence can be achieved, oral ACs are likely to be effective.
These exploratory analyses need confirmation in prospective trials of oral ACs that ensure high rates of compliance. When combining the overall data from OASIS-2 with data from three previous small trials of warfarin in UA/non–Q wave MI, including the OASIS pilot (9)and ATACS study (10), a promising risk reduction in death, MI and stroke of ∼17% can be seen, further supporting the possibility that oral ACs are effective when given in addition to aspirin. In this analysis, the overall OASIS-2 results were used, and the real benefit in centers that maintain a high degree of compliance is likely to be greater. In addition to the aforementioned data, the results of a U.K. trial of high risk individuals indicate a benefit of warfarin (INR ∼1.5) when added to aspirin (13).
Compliance is related to practice patterns
The data from the subgroup analyses of the OASIS-2 trial indicate clearly different patterns of practice (e.g., higher invasive procedure rates) among various countries, which is also associated with an initial reluctance to enter patients into the warfarin substudy, followed by higher rates of noncompliance once the patients were randomized. Not only do these variations in clinical practice have implications for the conduct of clinical trials of antithrombotic agents in different kinds of centers or countries, but they also may have a potentially adverse impact on patient outcomes (14). Liberal or routine use of invasive coronary procedures has not been consistently shown to be superior to a more selective strategy where cardiac catheterization is reserved for patients with provocable or recurrent ischemia (15). Further, some trials indicate a worse outcome with an aggressive approach (16). In our trial, we observed no difference in the control group’s event rates between good-complier countries and poor-complier countries, despite marked differences in intervention rates. A further potential downside associated with a practice pattern where invasive procedures are common was observed in this trial. If clinicians do not use antithrombotic agents (which are potentially beneficial) because of concerns of bleeding during cardiac catheterization, they may be denying patients the benefits of potentially helpful therapies. Therefore, there is an urgent need to conduct randomized trials that evaluate the effects of long-term antithrombotic therapies as well as determine whether a liberal invasive strategy is helpful in patients with UA. Our trial also demonstrates a general point: predefined beliefs and differences in practice patterns may also influence the efficacy of treatments. This suggests that long-term antithrombotic therapies may be best evaluated in countries or centers where there is a more conservative approach to the use of invasive procedures, unless special attention is paid to maintaining high compliance to such therapies.
Need for more effective and simple long-term therapy
The data from our trial confirm the relatively high rates of recurrent ischemic events after hospital discharge in these patients, despite the widespread use of aspirin. Conventional antithrombotic strategies have focused largely on the first week after the onset of symptoms. However, only approximately one-third of the ischemic events that we observed at six months were observed in the first week. Therefore, there is an urgent need to develop effective and acceptable long-term strategies (e.g., prolonged antithrombotic therapies) in patients with UA or non–Q wave MI. The disappointing results in the long term with the new oral glycoprotein IIb/IIIa agents (17)and with long-term use of subcutaneous low molecular weight heparin (18–20), combined with the reluctance of some physicians in our study to use oral ACs optimally, suggest that other simple and safe treatments are needed. Several trials of patients undergoing stent placement after coronary angioplasty indicate that the combination of aspirin plus thienopyridine (e.g., ticlopidine or clopidogrel) has a synergistic effect and may be superior to warfarin in preventing subacute coronary occlusion (21–24). This combination of antiplatelet agents was also safer than oral ACs. If the benefits and safety of aspirin plus a thienopyridine antiplatelet agent can be confirmed during long-term therapy after UA, this relatively simple regimen is more likely to be accepted into clinical practice. This hypothesis is being tested in the Clopidogrel in Unstable angina to prevent Recurrent Events (CURE) study (25).
Only a small, nonsignificant overall benefit was observed in patients allocated to receive warfarin. This is likely due to the poor compliance in several countries where high rates of invasive procedures are commonly used. Our exploratory analyses suggest that a regimen of moderate-intensity oral ACs in countries or centers that are likely to be compliant adds to the benefits of aspirin in patients with UA. The differences in the results between the two sets of countries or centers are biologically plausible (good compliance to ACs is associated with greater efficacy, but higher bleeding risk), statistically valid (prerandomization stratification variable was used) and internally consistent across three sets of efficacy outcomes, with significant tests of heterogeneity. Furthermore, combining the overall results of the current study with previous trials examining the effect of moderate-intensity oral ACs in addition to aspirin compared with aspirin alone is also supportive of an added benefit. This provides support for the conduct of further trials with warfarin or new antithrombotic or antiplatelet agents, or both, in patients with UA or MI or in other groups of patients at high risk of major vascular events. (Appendix)
S. Yusuf, J. Pogue and S. Anand (Hamilton, Ontario, Canada)
S. Yusuf (Chairman and Principal Investigator); G. Tognoni (Vice-Chair, Italy); K. Fox (Vice-Chair, United Kingdom); Argentina:R. Dı́az, E. Paolasso; Australia:D. Hunt, J. Varigos; Brazil:Á. Avezum, L. Piegas; Canada: S. Anand, C. Joyner, P. Theroux; Germany:H. Rupprecht, T. Wittlinger; Greece:N. Karatzas; Hungary:M. Keltai, E. Sitkei; Israel:D. Halon, B. S. Lewis; Italy:M. G. Franzosi, M. Galli, A. Maggioni, F. Mauri, G. Tognoni; Mexico:M. Ramos-Corrales; Poland:A. Budaj, L. Ceremuzynski; South Africa:P. Commerford; United Kingdom:M. Flather, K. Fox; United States:J. Anderson
Data and Safety Monitoring Board
J. Hirsh (Chairman), J. Cairns, M. Gent, R. Gorlin, J. Willerson, J. Wittes, G. Wyse
C. Joyner, M. Flather (Co-Chairs), S. Anand, P. Auger, Á. Avezum, A. Budaj, Y. K. Chan, P. Commerford, R. Dı́az, M. Flather; M. Galli, D. Halon, H. Hernandez, D. Hunt, C. Joyner, N. Karatzas, M. Keltai, M. Kyriakidis, B. S. Lewis, A. Maggioni, J. F. Marquis, F. Mauri, B. Mayosi, S. Mehta, M. Natarajan, A. Panju, E. Paolasso, L. Piegas, M. A. Ramos-Corrales, C. Rihal, H. J. Rupprecht, E. Sitkei, B. Sussex; W. Wasek, T. Wittlinger
Canadian Cardiovascular Collaboration Project Office, Hamilton Health Sciences Corporation and McMaster University, Canada:Administration:B. Cracknell, J. Lindeman; Adjudication:J. MacKay, H. Marsh; Computing:M. Anderson, M. Sloan, I. Stoica, J. Tucker; Data Management:F. Cherian, C. Christmas, C. Cuvay, I. Holadyk-Gris, S. Kotlan, J. MacKay, F. Mazur, M. Micks, K. Nair, L. Robinson, S. Seitz, S. Smith, L. Tomic; Health Economics:A. Gafni, A. Lamy; Physicians:S. Anand, S. Yusuf; Statistical Analysis:J. Pogue, C. Sigouin; Study Coordinators:J. Brown, L. Cronin; Thrombosis:M. Johnston, J. Weitz
- anticoagulation or anticoagulants
- confidence interval
- International Normalized Ratio
- myocardial infarction
- Organization to Assess Strategies for Ischemic Syndromes
- relative risk
- unstable angina
- Received July 7, 2000.
- Revision received August 30, 2000.
- Accepted October 4, 2000.
- American College of Cardiology
- Fuster V.,
- Badimon L.,
- Cohen M.,
- Ambrose J.A.,
- Badimon J.J.,
- Cheseboro J.H.
- Organization to Assess Strategies for Ischemic Syndromes (OASIS-2) Trial Investigators
- Antiplatelet Trialists’ Collaboration
- Van Belle E.,
- Lablanch J.M.,
- Bauters C.,
- Renaud N.,
- McFadden E.P.,
- Bertrand M.E.
- OASIS Pilot Study Investigators,
- Anand S.S.,
- Yusuf S.,
- Pogue J.,
- Weitz J.I.,
- Flather M.
- Antithrombotic Therapy in Acute Coronary Syndromes Research Group,
- Cohen M.,
- Adams P.C.,
- Parry G.,
- et al.
- Mantel N.,
- Haenszel W.
- ↵The Medical Research Council’s General Practice Research Framework. Thrombosis prevention trial: randomised trial of low-intensity oral anticoagulation with warfarin and low-dose aspirin in the primary prevention of ischaemic heart disease in men at increased risk. Lancet 1998;351:233–41.
- OASIS (Organization to Assess Strategies for Ischemic Syndromes) Registry Investigators,
- Yusuf S.,
- Flather M.,
- Pogue J.,
- et al.
- TIMI-IIIb (Thrombolysis in Myocardial Ischemia) Investigators
- Veterans Affairs Non–Q-Wave Infarction Strategies in Hospital (VANQWISH) Trial Investigators,
- Boden W.E.,
- O’Rourke R.A.,
- Crawford M.H.,
- et al.
- SYMPHONY Investigators
- FRagmin In unstable Coronary artery disease (FRIC) Study Investigators,
- Klein W.,
- Buchwald A.,
- Hillis S.E.,
- et al.
- Antman E.M.,
- McCabe C.H.,
- Premmereur J.,
- et al.
- Hall P.,
- Nakamura S.,
- Maiello L.,
- et al.
- Leon M.B.,
- Baim D.S.,
- Gordon P.
- Full ANTicoagulation versus ASpirin and TIClopidine (FANTASTIC) Study Investigators,
- Bertrand M.E.,
- Legrand V.,
- Boland J.,
- et al.
- ↵The Clopidogrel in Unstable angina to prevent Recurrent Events (CURE) trial programme. Rationale, design and baseline charaacteristics including a meta-analysis of the effects of the thienopyridinesin vascular disease. CURE Study Investigators. Eur Heart J 2000;21:2033–41.