Author + information
- Received December 24, 1998
- Revision received June 29, 1999
- Accepted August 27, 1999
- Published online December 1, 1999.
- Allan M Ross, MD, FACC∗,* (, )
- Karin S Coyne, PhD, RN, MPH∗,
- Jonathan S Reiner, MD, FACC∗,
- Samuel W Greenhouse, PhD†,
- Cynthia Fink, MPH∗,
- Anthony Frey, MD∗,
- Eduardo Moreyra, MD, FACC∗,
- Mouhieddin Traboulsi, MD, FACC‡,
- Normand Racine, MD§,
- Arthur L Riba, MD, FACC∥,
- Mark A Thompson, MD, FACC¶,
- Steven Rohrbeck, MD, FACC#,
- Conor F Lundergan, MD∗,
- for the PACT Investigators
- ↵*Reprint requests and correspondence: Dr. Allan M. Ross, The Cardiovascular Research Institute, George Washington University, 2150 Pennsylvania Avenue, NW, Suite 4-239, Washington, DC 20037
The study evaluated the efficacy and safety of a short-acting reduced-dose fibrinolytic regimen to promote early infarct-related artery (IRA) patency during the inherent delay experienced by infarct patients referred for angioplasty as the principal recanalization modality.
Previous approaches using long-acting, full-dose thrombolytic infusions rarely showed benefit, but they did increase adverse event rates.
Following aspirin and heparin, 606 patients were randomized to a 50-mg bolus of recombinant tissue-type plasminogen activator (rt-PA) (alpha half-life 4.5 min) or to placebo followed by immediate angiography with angioplasty if needed. The end points included patency rates on catheterization laboratory (cath lab) arrival, technical results when PTCA (percutaneous transluminal coronary angioplasty) was performed, complication rates, and left ventricular (LV) function by treatment assignment and time to restored patency following angioplasty.
Patency on cath lab arrival was 61% with rt-PA (28% Thrombolysis in Myocardial Infarction trial [TIMI]-2, 33% TIMI-3), and 34% with placebo (19% TIMI-2, 15% TIMI-3) (p = 0.001). Rescue and primary PTCA restored TIMI-3 in closed arteries equally (77%, 79%). No differences were observed in stroke or major bleeding. Left ventricular function was similar in both treatment groups, but convalescent ejection fraction (EF) was highest with a patent IRA (TIMI-3) on cath lab arrival (62.4%) or when produced by angioplasty within an hour of bolus (62.5%). However, in 88% of angioplasties, the delay exceeded 1 h: convalescent EF 57.3%.
Tailored thrombolytic regimens compatible with subsequent interventions lead to more frequent early recanalization (before cath arrival), which facilitates greater LV function preservation with no augmentation of adverse events.
The recent literature regarding management of acute myocardial infarction (AMI) has emphasized the choice between intravenous (IV) fibrinolytic therapy and primary angioplasty (1–5). An alternative might be a strategy combining both approaches. The first goal of such a combination would be to overcome the only modest reperfusion success rates with thrombolytics by the im-
mediate application of early rescue angioplasty for drug failures. The second goal would be to blunt the negative impact of frequent long delays in performing primary angioplasty by producing, in some patients, very early recanalization accomplished by initial thrombolytic use. Such a combination strategy, however, is generally believed to lessen the technical success rate of primary angioplasty and result in excess bleeding; thus, the two are thought not to be compatible.
This noncompatibility notion had its origin in a small comparative trial of primary angioplasty versus thrombolysis plus immediate angioplasty, which enrolled only 121 patients (6)but documented an unacceptably high complication rate. The notion was indirectly enhanced by larger studies (7–12)showing no advantage for an “aggressive” regimen (angiography and often angioplasty after thrombolysis) over a less complicated course in conservatively managed postthrombolysis patients. In these latter trials, however, percutaneous transluminal coronary angioplasty (PTCA) if performed early (and most commonly on already patent vessels) was undertaken during the thrombolytic infusion, which potentially contributed to the higher reported rates of bleeding compared with patients assigned to a noninterventional strategy. No large trial has directly compared primary angioplasty to a strategy utilizing a very short-acting and reduced-dose thrombolytic followed by immediate angiography plus planned rescue angioplasty (as necessary for occluded infarct-related arteries [IRAs]).
The current trial was therefore conducted to interrogate several hypotheses: 1) that patients receiving a reduced bolus dose of recombinant tissue-type plasminogen activator (rt-PA) would show an increased rate of very early IRA patency compared with those who received a placebo bolus; 2) that such rapidly reperfused patients would demonstrate greater preservation of convalescent left ventricular (LV) function than those who required mechanical recanalization; 3) that the short half-life thrombolytic would not negatively influence the technical outcome of angioplasty when required; and 4) that there would not be an increase in reocclusion or clinical adverse events for rt-PA patients versus placebo patients.
The Plasminogen-activator Angioplasty Compatibility Trial (PACT) was a multicenter, randomized, double-blind, placebo-controlled study in AMI patients. Patient eligibility criteria included ischemic symptoms ≥30 min and ST-segment elevation ≥0.1 mV in ≥2 limb leads or ≥0.2 mV in ≥2 contiguous precordial leads. Exclusions were age >75, prior cerebrovascular accident or transient ischemic attack, head trauma within six months, active bleeding or bleeding diathesis, recent trauma or major surgery, prior bypass surgery, PTCA within six months, AMI symptoms >6 h, systolic blood pressure >170 or diastolic >110 mm Hg at randomization. Pregnant or lactating women were also excluded. The protocol was approved by each institution’s Review Board or Ethical Committee. All patients gave written informed consent. Oversight was provided by an independent Data Safety and Monitoring Committee (Appendix A).
The double-blinded study design is depicted in Figure 1. Eligible patients received unfractionated heparin, a 5,000-IU bolus followed by an infusion of 1,000 IU/h (1200 IU/h for patients >80 kg) and 325 mg of aspirin (chewed and swallowed). Randomization was to an IV bolus of rt-PA 50 mg (Activase, Genentech; Actilyse, Boehringer-Ingelheim, GmbH) given over 3 min or an IV placebo bolus given in an identical manner. Randomization (arranged in blocks by institution) was accomplished by selecting sealed study drug kits in prespecified sequential order.
Patients were to undergo angiography as soon as possible following study drug. On cath lab arrival, the initial injection of the IRA was used to assess patency as described in the Thrombolysis in Myocardial Infarction (TIMI) trial (13). If the IRA was occluded (TIMI flow grade 0 or 1) or open but with sluggish flow (TIMI flow grade 2), immediate angioplasty was performed (termed “planned rescue” angioplasty in rt-PA patients and “primary” angioplasty in placebo patients). If the IRA exhibited normal (TIMI grade 3) flow, a second bolus of the assigned study drug could be administered and PTCA usually deferred. Left ventriculograms were then obtained (right anterior oblique) with a radiopaque ball filmed in the left axilla for calibration. Daily aspirin was continued throughout the hospitalization, and IV heparin was maintained for a minimum of 48 h (unless severe bleeding occurred). Activated partial thromboplastin times (aPTT) were drawn at randomization and every 6 h for 48 h, with a target aPTT between 60 and 85 s (14). Other drugs and technical approaches to mechanical reperfusion were at the discretion of the investigators.
To assess for reocclusion and convalescent LV function, follow-up angiography was repeated five to seven days later. After four to six weeks, stress testing (exercise or pharmacological) was performed.
Enrolling facilities were experienced tertiary-care institutions (Appendix B)providing rapid primary angioplasty (except for one facility that transferred 10 patients to a nearby partner tertiary-care hospital). Operators met AHA–ACC (American Heart Association–American College of Cardiology) experience guidelines (15). Sites were requested to maintain screening logs of eligible AMI patients. Choices of technical angioplasty approaches and supportive pharmacological therapies (other than study drug, aspirin, and heparin) were at the discretion of the operators.
Adverse clinical events were defined as major bleeding (a need for a transfusion ≥2 U of blood or intracerebral bleeding), reocclusion, reinfarction, and death.
Angiographic core laboratory and coordinating center procedures
The IRA was selected by core laboratory angiographers, blinded to treatment, utilizing the electrocardiogram, presence of arterial thrombus, and the ventriculographic location of the contractile abnormality. The IRA flow was judged on the initial contrast injection.
Ventriculographic silhouettes were acquired digitally at end systole and diastole (borders defined by the core laboratory angiographer). Cine projector output was linked via an ITC RC Vision Plus frame grabber to an ImageComm StatVIEW workstation (Quinton Instrument, Sunnyvale, California). Volumes and ejection fractions (EFs) were calculated by the area-length method (16). The infarct zone was further characterized by the most abnormal 50% of chords in an infarct region (the number of standard deviations [SD] per chord), the number of consecutive infarct zone chords >2 SD below the norm, and the percentage of patients with no abnormal infarct zone cords (17).
Sample size requirements
To detect an absolute difference of 10% between treatments in initial patency rates of TIMI flow grade 3 with a power of 80% and ∝ = 0.05, a total of 231 patients per group were required. This sample size also allowed detection of a 20% difference in initial patency rates of TIMI flow grades 2 and 3 with a power of 95% and ∝ = 0.05. To detect an absolute difference in coronary reocclusion of 8% (6% vs. 14%) with a power of 90% and ∝ = 0.05, a total of 239 patients per group were required. A sample of 142 patients per group is required to detect an absolute difference of 5% in EF with a power of 90% and ∝ = 0.05. To account for patients without follow-up angiography (anticipated to be 25%), the total enrollment was set at 300 patients per group.
Continuous end points are expressed as mean ± SD unless otherwise noted. Chi-square tests were used to compare frequencies of categorical variables. Further pairwise comparisons were adjusted by the Bonferonni method. Continuous end points of two groups were compared by unpaired Student ttests; all p values are two-tailed. The cumulative frequency distribution was accomplished by the Kaplan-Meier method. A p value ≤0.05 was indicative of significance. All analyses were performed utilizing SAS software (SAS Institute, Cary, North Carolina). No imputation methods were employed to account for anticipated missing follow-up angiographic data.
No differences in baseline characteristics were found between the treatment groups (Table 1). Screening logs were available from institutions enrolling 75% of study patients. Sixty-three percent of patients meeting entry criteria at these hospitals were enrolled. Reasons for not enrolling were: physician preference (34.6%), patient refusal (18.4%), competing myocardial infarction (MI) trial (11.4%), and other (35.6%). No baseline characteristic differences were noted between eligible patients enrolled versus those not enrolled. Study drug kits were inadvertently opened prior to recognition of study exclusion criteria for 10 patients who were not enrolled: (2 patients age >75 years, 1 patient >6 h from pain onset, 3 patients bleeding or stroke risk, 4 patients other reasons).
For both treatment groups, patients arrived at the hospital at a median of 1.4 h following symptom onset. From hospital arrival to study drug, a median of 49 min elapsed. The median time from study drug to the initial IRA contrast injection was also 49 min. No significant differences were seen in any time interval by treatment group.
Among patients arriving at the cath lab, a patent artery (TIMI flow grade 2 or 3) was seen approximately twice as often in patients who received a thrombolytic than a placebo (61% vs. 34%; p < 0.001). TIMI flow grade 3 was present in 33% of rt-PA patients versus 15% of placebo patients (p < 0.001) (Fig. 2).
The frequency distribution of time from study drug to documenting TIMI flow grade 3 in patients who exhibited this grade at the time of the first injection into the IRA (without the need for angioplasty) and for patients with lesser initial flow grades subsequently converted to TIMI flow grade 3 by angioplasty is shown in Figure 3. The difference of 42 min between the median time points was significant (p < 0.0001).
No differences in angioplasty success between treatment groups were evident (Table 2). Following rt-PA, conversion of closed vessels (TIMI flow grades 0, 1) to patent vessels (TIMI flow grade 2 or 3) was 92.8% and to TIMI flow grade 3 alone, 76.6%; among placebo patients the values were 94.6% and 79.0% (p = 0.52 and 0.62, respectively). When the initial flow was TIMI grade 2, angioplasty restored the flow to grade 3 in 82.8% of rt-PA and 86.7% of placebo patients (p = 0.59). Among 116 patients with TIMI flow grade 3 on catheter laboratory arrival and who underwent follow-up angiography, 8 experienced reocclusion (6/57 rt-PA, 2/26 placebo). In the 262 patients with TIMI flow grade 3 restored by angioplasty and with subsequent angiograms, 10 experienced reocclusion (5/101 rt-PA, 5/151 placebo).
Balloon angioplasty was the dominant intervention (74%). Stenting was employed in 26% of patients in each group and atherectomy devices in three patients. In addition to protocol-required aspirin and heparin, abciximab was administered during the intervention to 5% of patients in each group.
Four hundred forty-four patients had an analyzable initial left ventriculogram (73.3%; 220 rt-PA, 224 placebo). Three hundred seventy-three patients (61.6%; 182 rt-PA, 191 placebo) had a follow-up left ventriculogram adequate for analysis. Missing or unanalyzable follow-up studies are detailed in Table 3. Baseline characteristics between patients with and without follow-up angiograms were nonsignificant, except women less often received a follow-up angiogram than did men (66.7% vs. 77.6%, p = 0.01) and patients without a follow-up were slightly older (60.1 years vs. 57.2 years, p = 0.003).
Left ventricular function was analyzed by treatment assignment and by method and timing of restored patency (Table 4). The EF comparison by treatment assignment, without regard for time, method of patency, or final flow grade, was similar (58.2% vs. 58.4%, respectively, p = 0.85). Measures of regional infarct zone function parallel the EF observations (Table 4).
Whereas treatment assignment did not influence ventricular function, the timing of TIMI flow grade 3 restoration did have a significant impact. Patients arriving at cath lab with TIMI flow grade 3 IRA flow already established had initial and convalescent EFs of 60.5% and 62.4% compared with a later mechanical restoration (primary or planned rescue PTCA) initial EF of 58.7% (p = 0.26) and convalescent EF of 57.9% (p = 0.004). Never achieving TIMI flow grade 3 (n = 72) was associated with the lowest initial EF (55.8%) and convalescent EF (54.7%).
The small number of patients (n = 25) in whom PTCA restored TIMI flow grade 3 patency within 1 h of study drug had a mean convalescent EF of 62.5%, not unlike those who arrived in the catheter laboratory with already established normal flow. In 88% (n = 187) of mechanically opened vessels, however, the procedure was accomplished >1 h following study drug; these patients’ mean convalescent EF was 57.3%, significantly lower than in pharmacologically reperfused and very early mechanically reperfused patients (p < 0.005).
No significant differences were observed in any adverse event on the basis of treatment assignment (Table 5). In rt-PA and placebo patients, the incidence of major hemorrhage was 12.9% and 13.5%, respectively (p = 0.84). After excluding bleeding associated with bypass surgery, the major bleeding rates were 8.5% and 8.2% for rt-PA and placebo, respectively (p = 0.92). No differences existed in bleeding rates between patients who received two boluses of study drug versus those who received one bolus. The need for later emergent revascularization was 7.3% in the rt-PA group and 7.2% in the placebo group (p = 0.98). Although not intended nor powered to be a mortality trial, it is of interest that patients arriving in the cath lab with an open IRA had a 30-day mortality rate of 2.1%; those requiring a PTCA to achieve TIMI flow grade 3, 3.1%, and those never achieving TIMI grade 3 flow, 4.7% (p = 0.48). At one year, the mortality rates were 2.8%, 5.3%, and 7.9% (p = 0.18), respectively.
The advantages of thrombolytic therapy are universal availability and short time to administration. Its chief disadvantage is the apparent limitation in restoring normal TIMI grade 3 flow, generally reported to occur in not more than 60% to 65% of recipients within 90 min of therapy (18,19). In contrast, primary PTCA has the advantage of higher TIMI flow grade 3 rates, but its availability is more limited and the delays prior to reperfusion can be substantial. Because outcome after infarction is significantly determined by the speed and degree of reperfusion restored and because the advantages and disadvantages of the two approaches appear complementary, it seemed logical to seek synergism by designing a strategy that combined both treatments. The results of PACT appear to confirm that such synergism can be achieved despite a previously widely held opinion that thrombolytics and angioplasty are not compatible.
Previous relevant studies
Interestingly, the evidence underlying this opinion of noncompatibility was never particularly strong. One previous trial directly compared primary PTCA with thrombolysis and immediate PTCA. O’Neill et al. (6)performed angioplasty during an infusion of the long-acting fibrinogen-depleting thrombolytic, streptokinase, in 58 patients comparing their course with 63 primary angioplasty patients. The success rate (TIMI flow grades 2 and 3) of PTCA in the combination treatment group was 98% versus 92% for primary PTCA. Follow-up LV function showed a trend favoring the thrombolytic group. However, an extraordinary incidence of bleeding and need for transfusion (39%) occurred in the thrombolytic group, leading the authors to voice caution regarding combination strategies. The thrombolytic regimen selected in the present trial, an initial bolus of 50 mg rt-PA, was chosen to take advantage of its fibrinogen-sparing properties and extremely short (4 to 5 min) half-life when administered as a bolus with the goal of minimizing bleeding risk.
Other studies suggesting thrombolytic-angioplasty incompatibility include TIMI-IIA and IIB, TAMI-1 and 5 (Thrombolysis and Angioplasty in Myocardial Infarction), and the European Cooperative Study Group (ECSG) investigation (7–11). None of these actually compared a combined treatment approach versus traditional primary angioplasty as done in the present trial. Rather, they studied thrombolysis (all patients) followed by a conservative course or by protocol-driven angiography and sometimes angioplasty. The interventions were more often delayed than immediate, and most angioplasties were performed in an already patent artery. Hence, they shed little light on the value of thrombolysis with immediate rescue. Only three of these studies performed interventions within the first 3 h following onset of thrombolytic therapy and directed that closed infarcts undergo PTCA attempts; such rescue procedures numbered 71 patients in the ECSG (11), 52 in TAMI-5 (9), and 37 in TIMI-IIA (7). The PTCA success rates were 80%, 89%, and 92%, respectively. The conclusion by the ECSG was that early interventions provided no benefit while augmenting risk (this study was prematurely terminated after entering 367 patients). The TIMI-IIA study, showed higher complication rates during acute intervention compared with delayed intervention (18 to 48 h) with no offsetting clinical benefits. In contrast, TAMI-5 suggested that early post-lytic intervention produced improved outcome LV function and a better overall clinical course than did conservative care. It should be noted that in most of the above studies, thrombolytic infusions were ongoing during the acute angioplasty procedures.
More recent specific-rescue PTCA reports have failed to confirm increased complications, perhaps in part the consequence of improved puncture site and sheath management and revised adjunctive anticoagulation regimens (4,20,21). The benefits of rescue angioplasty have been supported by the one randomized trial of the issue and several reported consecutive series (20–24).
The present trial has demonstrated that even a very reduced dose of fibrin-selective plasminogen activator produces an early patency advantage occurring before a mechanical reperfusion attempt can be initiated. Such very early reperfusion is associated with improved outcome LV function. When rescue angioplasty is nonetheless required, there is no decrease in the procedure’s technical success rate and no increase in adverse events. Thus, thrombolysis, as employed in this trial, and angioplasty are compatible.
The absolute temporal advantage for patients arriving in the catheter laboratory with restored complete patency can only be approximated as only the time of such patency is initially recorded. Vessels open 51 min after bolus treatment were likely open in advance of the first documented picture. In contrast, the time of mechanical flow restoration was precisely documented in the cath lab. Hence, the 42-min difference noted in Figure 3is likely to be an underestimation. In a recent report by Liem et al. (25), a 43-min “additional” delay was observed for patients transferred to a tertiary-care hospital for primary angioplasty compared with those presenting directly to the PTCA-capable facility. Though not randomized, the apparent penalty of this delay was larger infarct size by enzyme analysis (LDH 1536 IU vs. 1235 IU, p < 0.005) and a substantially lower follow-up EF six months after the MI (43% vs. 47%, p < 0.003), findings quite confirmatory of the observations in PACT and other reports (26–28).
The present study has design limitations ascribable to prior concerns regarding safety (i.e., bleeding), regarding the effectiveness of PTCA when performed with thrombolytics, and regarding an intent to minimize the delays often encountered when primary angioplasty is the selected treatment. Accordingly, the dosing of rt-PA was purposefully conservative and the trial performed in hospitals with remarkably short delays from emergency department treatment to angioplasty. One could speculate that a more aggressive thrombolytic regimen would have increased the number of patients with patency restored prior to catheter laboratory arrival. It is also reasonable to speculate that thrombolysis followed by angiography and planned-rescue PTCA would be more advantageous in clinical environments where delays from emergency department triage to PTCA are longer than those observed in PACT. The U.S. National Registry for Myocardial Infarction-2 reports a 110-min delay to primary angioplasty for in-hospital patients and 221 min for transferred patients (29). It is likely that the longer the delay, the more advantageous would have been the combination approach. Confirmatory studies with regimens tested in more “real-world” environments would be of considerable interest. The use of both stents and particularly glycoprotein IIb/IIIa antagonists in this study was less frequent than current practice, owing mainly to the recruitment of most patients in advance of published results with these approaches in ST elevation MI patients.
New strategies and therapies are scrutinized not only for their benefits but also their costs. Proof of the benefits of establishing earlier and more frequent reperfusion has been amply provided in previous investigations (19,30–32). Regarding the cost of applying two reperfusion modalities to the same patient, some increase might be offset by the immediate anatomic risk stratification provided, thereby obviating the need for other risk-stratifying tests. Savings may also be accomplished in patients angiographically identified as low risk by planning very early hospital discharge (33). In the U.S., however, the biggest cost-neutralizing factor is that more than 70% of thrombolytic-treated patients undergo coronary angiography during their initial hospitalization, with another 15% to 20% referred for angiography soon after the index hospitalization (34,35). Following the combined strategy studied in PACT, the commonly performed angiogram is moved forward in time so as to become a potentially therapeutic (rescue) as well as diagnostic intervention.
Other pharmacological strategies intended to promote IRA patency during the diagnosis-to-angioplasty delay have been subject to previous and ongoing studies. Extraordinarily high-dose unfractionated heparin initially appeared promising, but the largest and most recent evaluation (36)showed no difference in patency rates compared with standard doses of heparin such as those used in the present investigation. Patency rates following precatheter laboratory use of platelet IIb/IIIa blockade are in an early stage of evaluation, and small reports have produced variable observations (37).
The long-standing controversy between thrombolysis and primary angioplasty may no longer be the most relevant issue regarding the choice of reperfusion modalities for AMI. A judicious use of both therapies in a comprehensive combination treatment strategy is an attractive and possibly preferable approach for a significant segment of the acute infarct population, specifically those patients selected for primary PTCA or stenting but in whom very rapid accomplishment of the planned procedure does not appear likely.
We are also indebted to the following individuals for their invaluable contributions to this trial: The George Washington University Cardiovascular Research Institute staff: Anita Bhatt; Deneane Boyle, MPH; Edmund Chiong; Yasmine Draoui, MS; Melissa Hammond, RN, MS; Cheryl Lundeberg, RN; Nakia Marks; Patti Verrier, RN; and Pamela Walker, RN.
Committee Chairman: Lawrence Cohen, MD Committee Members: Spencer King III, MD David Sheps, MD Marc Verstraete, MD, PhD O. Dale Williams, PhD, MPH
☆ Funded by a combined grant from Boehringer-Ingelheim, GmbH (Ingelheim, Germany), Boston Scientific (Natick, Massachusetts), and Genentech (South San Francisco, California).
- acute myocardial infarction
- activated partial thromboplasin times
- Cath lab
- catheter laboratory
- ejection fraction
- infarct-related artery
- left ventricular
- myocardial infarction
- percutaneous transluminal coronary angioplasty
- recombinant tissue-type plasminogen activator
- Thrombolysis in Myocardial Infarction trial
- Received December 24, 1998.
- Revision received June 29, 1999.
- Accepted August 27, 1999.
- American College of Cardiology
- Global Use of Strategies to Open Occluded Arteries in Acute Coronary Syndromes (GUSTO IIb) Angioplasty Substudy Investigators
- O’Neill W.W,
- Weintraub R,
- Grines C.L,
- et al.
- Topol E.J,
- Califf R.M,
- George B.S,
- et al.
- Califf R.M,
- Topol E.J,
- Stack R.S,
- et al.
- European Cooperative Study Group for recombinant Tissue-type Plasminogen Activator (rTPA),
- Simoons M.L,
- Betriu A,
- Col J,
- et al.
- SWIFT (Should We Intervene Following Thrombolysis?) Trial Study Group
- Chesebro J.H,
- Knatterud G,
- Roberts R,
- et al.
- American College of Cardiology Cardiac Catheterization Committee,
- Douglas J.S,
- Pepine C.J,
- Block P.C,
- et al.
- Bolson E.L,
- Kliman S,
- Sheehan F,
- Dodge H.T
- Ross A.M,
- Lundergan C.F,
- Rohrbeck S.C,
- et al.
- Ellis S.G,
- da Silva E.R,
- Heyndrickx G,
- et al.
- Abbottsmith C.W,
- Topol E.J,
- George B.S,
- et al.
- Liem A.L,
- van’t Hof A.W,
- Hoorntje J.C,
- et al.
- Newby L.K,
- Rutsch W.R,
- Califf R.M,
- et al.
- Milavetz J.J,
- Giebel D.W,
- Christian T.F,
- Schwartz R.S,
- et al.
- ↵The National Registry for Myocardial Infarction-2 (Database) (1998). Genentech, Inc., South San Francisco, CA.
- Morgan C.D,
- Roberts R.S,
- Hag A.M,
- et al.
- Vogt A,
- von Essen R,
- Tebbe U,
- et al.
- Ross A.M,
- Coyne K.S,
- Moreya E,
- et al.
- Rogers W.J,
- Bowlby L.J,
- Chandra N.C,
- et al.
- Every N.R,
- Parsons L.S,
- Fihn S.D,
- et al.
- Liem A.L,
- Zijlstra F,
- Hoorntje J,
- et al.
- Antman E.M,
- Giugliano R.P,
- Gibson C.M,
- et al.