ST segment resolution in ASSENT 3: insights into the role of three different treatment strategies for acute myocardial infarction

Transcription

1 European Heart Journal (2003) 24, ST segment resolution in ASSENT 3: insights into the role of three different treatment strategies for acute myocardial infarction Paul W. Armstrong a *, Galen Wagner b, Shaun G. Goodman c, Frans Van de Werf d, Christopher Granger b, Lars Wallentin e, Yuling Fu a, for the ASSENT 3 Investigators a University of Alberta, Edmonton, Canada b Duke Clinical Research Institute, Durham, NC, USA c University of Toronto, Ontario, Canada d University Hospital Gasthuisberg, Leuven, Belgium e University Hospital, Uppsala, Sweden Received 6 February 2003; received in revised form 9 May 2003; accepted 4 June 2003 This paper was guest edited by Prof. Peter L. Thompson, University of Western Australia, Perth, Australia KEYWORDS Acute myocardial infarction; Fibrinolysis; ST segment resolution Aims ASSENT 3 (Assessment of the Safety and Efficacy of a New Thrombolytic) demonstrated that the bolus fibrinolytic tenecteplase (TNK), combined with enoxaparin (ENOX) or abciximab (ABCX), substantially reduced ischemic complications of acute myocardial infarction as compared with unfractionated heparin (UH). We compared ST resolution in each of the three treatment regimens in order to evaluate the speed, extent and stability of ST segment resolution and its relationship to the primary composite endpoint(s) of the trial. Methods and results We evaluated ST segment shift and its subsequent resolution i.e. complete ( 70%), partial (<70 30%) and no resolution (<30%) in 4,304 patients 60 and 180 min after treatment. Sixty minutes after therapy there was a trend for both half-dose TNK/ABCX and TNK/UH to have more frequent complete ST resolution (P=0.072) than TNK/ENOX. Pair-wise comparison at 60 min revealed that patients receiving TNK/ABCX had significantly more complete ST resolution than TNK/ENOX (P=0.026). Further ST resolution had occurred 180 min after treatment: complete ST resolution was greater with TNK/ABCX (59.2%) than the TNK/heparin (50.8%) and TNK/enoxaparin (50.8%) (P<0.001). In the TNK/ENOX group achieving complete ST resolution by 180 min, in-hospital reinfarction was 1.9% vs 4.2% for TNK/UH (P=0.015) representing a 2.3% absolute and 55% relative reduction in reinfarction. Thirty day and one year mortality was greatest amongst those patients with <30% ST segment resolution in the TNK/ABCX group. Conclusions More rapid and complete ST resolution occurs with half-dose TNK/ABCX whereas less reinfarction occurs amongst those patients with 70% ST resolution receiving either TNK/ABCX or TNK/ENOX. These data highlight two potentially complementary mechanisms of clinical benefit associated with different pharmacologic regimens in acute myocardial infarction, i.e. more rapid versus more stable coronary patency Published by Elsevier Ltd on behalf of The European Society of Cardiology. * Corresponding author. Paul W. Armstrong, MD, 2 51 Medical Sciences Building, University of Alberta, Edmonton, Alberta, Canada T6G 2H7. Tel.: ; fax: address: paul.armstrong@ualberta.ca (P.W. Armstrong) X/03/$ - see front matter 2003 Published by Elsevier Ltd on behalf of The European Society of Cardiology. doi: /s x(03)

2 1516 P.W. Armstrong et al. Introduction Recently the ASSENT 3 (Assessment of the Safety and Efficacy of a New Thrombolytic) trial demonstrated that when the bolus fibrinolytic tenecteplase (TNK) was combined with enoxaparin (ENOX) or abciximab (ABCX), it substantially reduced the frequency of ischaemic complications of acute myocardial infarction as compared with conventional unfractionated heparin (UH). 1 This intermediate-sized trial was not primarily designed as a mortality trial but rather aimed at evaluation of treatment effects on a clinical composite end-point of 30-day death, reinfarction and refractory ischaemia. Given the prior work establishing the relationship between ST segment resolution and clinical outcomes, a systematic, prospective electrocardiographic analysis of the extent of ST segment resolution at 60 and 180 min after treatment was incorporated into the study in order to provide mechanistic insight into these end-points. 2,3 Recent availability of one-year mortality data provides additional opportunity for the evaluation of longer-term results. Our objectives were to compare the speed, extent and stability of ST segment resolution in each of the three treatment regimens and to evaluate their relationship to the primary composite endpoint(s) of the trial. Methods The ASSENT 3 trial has previously been described in detail. 1 Briefly, 6095 patients with acute myocardial infarction were randomly assigned to one of three regimens: full dose TNK and ENOX for a maximum of 7 days, half dose TNK with weightadjusted low dose unfractionated heparin and a 12-h infusion of ABCX or full dose TNK with weight-adjusted unfractionated heparin for 48 h. ECG criteria for admission were ST segment elevation 0.1 mv in at least two limb leads or 0.2 mv in two contiguous precordial leads or the presence of presumed new left bundle branch block. The ASSENT 3 ECG substudy consisted of the entire study population and pre-specified that a 12-lead ECG be recorded at baseline and again at 60 and 180 min after treatment. The ST segment measurements were evaluated centrally without reader knowledge of the treatment assignment or clinical outcome at the ECG core laboratories (Canadian VIGOUR Centre, University of Alberta, Edmonton, Canadian Heart Research Center, Toronto, and Duke Clinical Research Institute). To minimize confounding factors influencing the accuracy of the assessment of ST segment resolution, patients were excluded (Fig. 1) from this analysis if they were: (1) missing baseline, 60 or 180 min tracings (n=463); (2) without the protocol-defined amount of ST elevation on baseline ECG (n=481); or (3) had left bundle branch block (LBBB), paced rhythm, ventricular rhythm, or poor quality ECG s. Of the 463 patients with missing ECG s, 65 died within 30 days and of these, 62 died prior to the recording of the 180 min ECG. Of the 481 patients without the required ST elevation, 175 had no CK elevation and their 30-day mortality was 3.4% whereas a mortality of 7.2% for the remaining 306 patients with elevated CK was observed. Of the 847 patients with confounding ECG factors, 740 had poor quality ECG s and had a 30 day mortality of 8.4% whereas the 107 patients with either LBBB, paced rhythm or ventricular rhythm had a 30 day mortality of 13.1%. One-year mortality tracked the 30-day results and is also demonstrated in Fig. 1 Hence, the final ECG substudy group consisted of 4598 patients of whom 4304 had all three ECG s. Electrocardiographic (ECG) analysis The amount of ST segment elevation was measured manually at the J point using a hand-held caliper from leads I, avl and V1 to V6 for anterior myocardial infarction and leads II, III, avf, V5 and V6 for inferior MI. For the total ST segment deviation the sum of ST segment depression in leads II, III and avf for anterior and Fig. 1 Disposition of patients as well as their 30-day/1 year mortality.

3 ST resolution in ASSENT Table 1 that in leads V1 to V4 for inferior myocardial infarction were added. The resolution of ST segment elevation (or total deviation) at 60 min and 180 min was stratified into three categories based on Schroeder s method (2): complete resolution was defined as resolution of the initial sum of ST segment elevation 70%, partial resolution was defined as ST segment resolution <70% to 30%, and no resolution was defined as ST segment resolution <30%. In addition to this classification based primarily on the extent of ST resolution, patients were also categorized according to treatment assignment and on the relationship to the principal efficacy endpoints of ASSENT 3. Statistical analysis Descriptive statistics were summarized as medians with 25th and 75th percentiles for continuous variables and the Kruskal Wallis test was used for comparisons among treatment regimens. For categorical variables, the data were summarized in percentages and Fisher s exact test or the Chi-square test was used to assess group differences. Multivariate logistic regression analysis was used to assess the effect of baseline characteristics and extent of ST segment resolution on 30 day and one year mortality. 4 Variables examined included age, gender, presence of hypertension or diabetes, current smoking, previous MI, previous coronary artery bypass grafting (CABG) or previous PCI, systolic blood pressure, heart rate, index MI location, Killip class at presentation, time from symptom onset to treatment, and treatment assigned. All tests were 2-sided, with a 5% level of significance. All analyses were performed using SPSS (Chicago, Illinois, Version 11). Results Baseline characteristics according to treatment assignment In Table 1 the baseline characteristics of the patients in the ECG substudy are shown according to treatment assignment. As was the case in the overall study, patient characteristics were similar across the three treatment groups. The overall median time intervals (25th and 75th percentiles) from the start of bolus TNK to the acquisition of 60 min and 180 min ECG were 61 min (55, 68), and 183 min (174, 195), respectively. There were no differences in time to acquisition of ECG s at either of these time points among the three treatment groups: hence the TNK/ENOX n=1444 ½ TNK/ABCX n=1445 TNK/UH n=1415 Age (median) 60 (52, 69) 59 (51, 70) 60 (51, 70) Age >75 years % Female % Weight (kg) 77 (70, 88) 78 (68, 88) 78 (69, 88) Time from onset of symptoms to randomization (h) 2.3 (1.6, 3.5) 2.5 (1.6, 3.5) 2.5 (1.7, 3.6) Killip class >I Systolic blood pressure (mmhg) 133 (120, 150) 132 (120, 150) 132 (119, 150) Anterior infarct location % Heart rate (bpm) 73 (62, 85) 73 (63, 84) 72 (62, 84) Hypertension % Diabetes % Previous myocardial infarction % Prior CABG % Prior PCI % Current smoker % median times were 61 (55,68) and 183 (174,195) min for TNK/ENOX; 60 (52,69) and 182 (173,195) min for TNK/ ABCX, and 61 (55,69) and 183(175,194) for TNK/UH respectively. In Fig. 2a and b the data according to the extent of ST resolution at 60 and 180 min are shown for each of the three treatment groups. By 60 min there was a trend for both half-dose TNK/ABCX and TNK/UH to have more frequent complete, i.e. 70% ST resolution (P=0.072) than TNK/ENOX. Pair wise comparison at 60 min revealed that patients receiving TNK/ABCX had significantly more complete ST resolution than the TNK/ENOX group (P=0.026). Further significant complete ST resolution had occurred after 180 min had elapsed (Fig. 2b) in each of the three treatment groups: the extent of ST resolution was greater however in the TNK/ABCX arm than that evident in either of the two TNK heparin arms (P<0.001). These changes are further evident in Fig. 3 where the ST segment resolution data at both 60 and 180 min are restricted to those patients who had ECG s at both these time points as a basis for comparison with their baseline measures. The incremental improvement between 60 and 180 min was highest in patients receiving TNK/ABCX and significantly greater than that seen with TNK/UH (25.7% vs 19.1%; P=0.002): a trend for a similar incremental improvement in ST resolution between 60 and 180 min was also evident for TNK/ENOX vs TNK/UH (23.5% vs 19.1%; P=0.069). In order to track the stability of ST segment resolution the evolution of ST segment change between 60 and 180 min is shown in Fig. 4 according to both treatment group and initial extent of ST resolution at 60 min. As is evident in Fig. 4a (upper right panel) the transition from 34.5% to 50.8% of TNK/UH patients with complete ST resolution reflects that mixture of 756 (53%) patients who had already achieved ST resolution at 60 min, coupled with the 415 (29.3%) patients who had had partial resolution at 60 min and the 243 (17.2%) patients without initial ST resolution. In the lower right panel of Fig. 4a, the composition of those patients who had <30% ST resolution at 180 min is highlighted. Interestingly,

4 1518 P.W. Armstrong et al. Fig. 2 (a) Distribution of 60 minutes ST segment resolution according to treatment assignment, P=0.072 for trend comparison. (b) Distribution of 180 min ST segment resolution according to treatment assignment, P<0.001 for trend comparison. 29 (11.2%) of those with complete and 48(19.0%) of those with partial ST resolution at 60 min showed deterioration by 180 min with resultant <30% ST resolution. The ENOX group proved to be intermediate and those receiving TNK/ABCX showed the least deterioration or greatest preservation of their initial ST resolution evident at 60 min (Fig. 4b, 4c, respectively). A similar pattern was observed between 60 and 180 min for those patients with partial ST segment resolution initially as compared with those with complete or <30% ST resolution. In Table 2a and Table 2b the composite efficacy endpoints are presented by treatment assignment and categorized according to the extent of ST resolution evident at 60 and 180 min, respectively. By 60 min a clear advantage for a reduction in the composite efficacy endpoint was evident in patients with >70% ST resolution treated with TNK/ABCX vs those receiving TNK/UH. This pattern was also evident for reinfarction. The extent of ST resolution appeared to have little relationship to the benefit afforded by TNK/ENOX as compared with TNK/UH even after adjustment for key baseline characteristics. For patients with <30% ST resolution, mortality was lowest in those receiving TNK/ENOX, especially as compared with TNK/ABCX. At 180 min the findings seen at 60 min

5 ST resolution in ASSENT Fig. 3 Distribution of 60 and 180 min ST segment resolution of 4304 patients who had ECG s at both 60 and 180 min according to treatment assignment, P=0.081 and P<0.001 for trend comparison, respectively. with TNK/ABCX and TNK/ENOX became more pronounced in patients with more than 70% resolution of ST segments i.e. a significant reduction in both the composite endpoint and in-hospital reinfarction as compared with TNK/ UH. In the TNK/ENOX group achieving complete ST resolution by 180 min, in-hospital reinfarction was 1.9% vs 4.2% for TNK/UH (P=0.015) representing a 2.3% absolute and 55% relative reduction in reinfarction. Amongst patients with <30% ST resolution, thirty-day and one year mortality was lowest amongst those in the TNK/ENOX arm and highest for those receiving TNK/ABCX. Moreover within the TNK/ABCX group those patients with less than 30% ST segment resolution had a significantly longer median time from symptom onset to treatment (3.04 h) as compared to those with partial (2.9 h) or complete ST resolution (2.75 h) (P=0.006). It is noteworthy that examination across the three treatment groups amongst all patients without ST segment resolution at 180 min revealed that the median time (h) from symptom onset to treatment was significantly longer for abciximab treated patients (3.04) as compared to those treated with enoxaparin (2.5 h) or those with unfractionated heparin (2.67) (P<0.001). These patterns were also evident from analysis performed at the 60-min data point. Discussion These data derived from a systematic analysis of ST segment resolution in over 4000 patients provide new mechanistic insights into the treatment effect of the two experimental arms within the ASSENT 3 study, i.e. TNK/ ENOX and TNK/ABCX as compared with the control arm of TNK/UH. A key finding from our study is the demonstration of more rapid and complete ST segment resolution when half-dose TNK is combined with abciximab: although this was already evident 60 min after commencement of therapy it became more fully expressed and amplified at 180 min. Further insight and support for these findings is derived from the lesser tendency of patients receiving TNK/ABCX to demonstrate deterioration in complete ST resolution by 180 min once it had been achieved at 60 min following treatment. Importantly, this greater treatment effect following TNK/ABCX on an established electrocardiographic surrogate for enhanced myocardial perfusion is in concert with the lesser need for urgent revascularization demonstrated for this treatment group in the overall ASSENT 3 study. 1 It is noteworthy however that the highest mortality at 30 days and one year occurred in those patients showing failure of ST segment resolution following treatment with TNK/ABCX. The basis for this finding is unclear but could signal identification of patients with more complex thrombotic obstruction resistant to a more potent pharmacologic reperfusion regimen thereby resulting in a worse prognosis. Another potentially contributing factor was the longer time from symptom onset to treatment in this group. Interestingly, the treatment arm in ASSENT 3 that achieved the best balance between efficacy and safety, i.e. TNK/ENOX, had the least number of patients achieving complete ST resolution at 60 min: by 180 min however the ST segment resolution results in this treatment group were comparable to those in the reference arm receiving TNK/UH. These findings are consistent with those derived from the HART-2 angiographic study where no early coronary patency advantage was demonstrable with the combination of enoxaparin and accelerated

6 1520 P.W. Armstrong et al. Fig. 4 (a c) Evolution of ST segment resolution from 60 to 180 min for TNK/UH (a), TNK/ENOX (b) and TNK/ABCX (c). Patients with complete ST resolution are shown as clear, those with partial ST resolution are hatched and those with <30% ST resolution are gray. The left panel in each figure depicts the extent of ST resolution evident at 60 min. The transition from the left to the middle panel represents the evolution of ST segment resolution between 60 and 180 min. All percentages are derived from the sample size shown in the title. The composition of actual patient numbers with complete ST resolution by 180 min and for those with <30% ST resolution is outlined in the small circles in the extreme right upper and lower panels of each figure. Hence for the 50.8% of TNK/UH patients in Fig. 4a who achieved complete ST resolution at 180 min (upper right panel), 756 of these had achieved complete ST resolution by 60 min, 415 had initially 30 70% ST resolution and 243 had <30% ST resolution at 60 min.

7 ST resolution in ASSENT Table 2a All patients Efficacy endpoints by 60 min ST segment resolution according to treatment assignment TNK/ENOX 1444 ½ TNK/ABCX 1445 TNK/UH % resolution 30-day mortality, in-hospital remi, or refractory ischaemia 9.2 (41/444) 6.2 (32/513) a 11.9 (58/488) In-hospital reinfarction 2.5 (11/444) 2.1 (11/513) a 4.9 (24/488) day mortality 4.1 (18/444) 2.5 (13/513) 2.0 (10/488) year mortality 6.2 (27/439) 5.4 (27/503) 3.6 (17/471) P for trend 30 70% resolution 30-day mortality, in-hospital remi, or refractory ischaemia 11.1 (54/487) a 10.1 (46/457) a 16.0 (71/445) In-hospital reinfarction 3.1 (15/487) 2.6 (12/457) 5.2 (23/445) day mortality 3.5 (17/487) 4.6 (21/457) 5.4 (24/445) year mortality 7.3 (35/477) 6.9 (31/450) 8.2 (36/437) <30% resolution 30-day mortality, in-hospital remi, or refractory ischaemia 10.7 (55/513) a 13.3 (63/475) 16.4 (79/482) In-hospital reinfarction 2.7 (14/513) 2.5 (12/475) 2.7 (13/482) day mortality 4.3 (22/513) b 8.8 (42/475) 5.8 (28/482) year mortality 7.2 (36/501) 10.7 (50/467) 7.8 (36/462) a Denotes P for pair-wise comparisons between TNK/UH versus TNK/ENOX or ½ TNK/ABCX. b Denotes P=0.004 for comparison between TNK/ENOX and ½ TNK/ABCX. Table 2b All patients Efficacy endpoints by 180 min ST segment resolution according to treatment assignment TNK/ENOX 1444 ½ TNK/ABCX 1445 TNK/UH 1415 P for trend 70% resolution 30-day mortality, in-hospital remi, or refractory ischaemia 8.3 (61/734) a 7.0 (60/856) a 12.4 (89/719) In-hospital reinfarction 1.9 (14/734) a 2.2 (19/856) a 4.2 (30/719) day mortality 3.7 (27/734) 2.9 (25/856) 3.6 (26/719) year mortality 6.5 (47/722) 5.6 (47/840) 5.3 (37/700) % resolution 30-day mortality, in-hospital remi, or refractory ischaemia 10.6 (50/471) 11.2 (48/427) 14.8 (65/438) In-hospital reinfarction 2.8 (13/471) 2.6 (11/427) 3.4 (15/438) day mortality 4.2 (20/471) 5.9 (25/427) 3.2 (14/438) year mortality 7.0 (32/459) 7.8 (33/422) 5.2 (22/420) <30% resolution 30-day mortality, in-hospital remi, or refractory ischaemia 16.3 (39/239) 20.4 (33/162) 20.9 (54/258) In-hospital reinfarction 5.4 (13/239) 3.1 (5/162) 5.8 (15/258) day mortality 4.2 (10/239) b 16.0 (26/162) a 8.5 (22/258) < year mortality 8.1 (19/236) b 17.7 (28/158) 12.0 (30/250) a Denotes P for pair-wise comparisons between TNK/UH versus TNK/ENOX or ½ TNK/ABCX. b Denotes P for comparison between TNK/ENOX and ½ TNK/ABCX. rt-pa versus that achieved with conventional rt-pa and unfractionated heparin. 5 The relationship between clinical end-points and ST resolution at 180 min (Table 2b) is particularly instructive as it relates to potential insights into the differential mechanisms of benefit of the two experimental ASSENT 3 treatment strategies. Whereas combination therapy with TNK/ABCX achieved earlier and more complete ST resolution that appeared to closely mirror enhanced clinical efficacy in ASSENT 3, this was not the case at 60 min for TNK/ENOX. Although at 180 min a treatment effect with TNK/ENOX on ST resolution became more evident, it was virtually identical to that achieved with TNK/ unfractionated heparin. Indeed, one of the most striking effects of TNK/ENOX was on the pronounced reduction in reinfarction amongst those patients with complete ST resolution by 180 min as compared with TNK/UH. This finding is also supported by prior angiographic findings in HART-2 indicating a strong trend towards reduced reocclusion with combination enoxaparin and rt-pa versus unfractionated heparin and rt-pa in those patients achieving early coronary patency. 5 Also noteworthy was

8 1522 P.W. Armstrong et al. the striking effect of TNK/ENOX on reducing reinfarction amongst those with complete ST resolution by 180 min vs those in the same treatment arm with no ST resolution at this juncture: thereby emphasizing that once reperfusion occurs with this regimen it has a greater likelihood of being sustained than when achieved with other treatment regimens. Hence, our observations derived from a large cohort of systematically studied patients highlight two different and potentially complementary pharmacologic reperfusion regimens applied to the management of patients with acute ST segment myocardial infarction. Although our study excluded 1791 patients with incomplete electrocardiographic information whose mortality was higher (Fig. 1), the comparability of baseline characteristics within the three treatment groups in the ECG study, and the close alignment of these outcomes with those demonstrated in the overall study, provides substantial reassurance about the relevance of our findings. Additionally, a small number of patients (n=296, 17%) underwent urgent PCI in the ECG substudy within the first hospital day: although we cannot establish with certainty the timing of these procedures, it is unlikely they occurred between the two ECG points of 60 and 180 min and therefore would not be expected to have significant impact on our results. Nonetheless, given the moderate sample sizes once categorization of outcomes was divided into the three different treatment regimens and ST segment resolution categories, there is need for caution in interpreting these comparisons and we cannot be assured of a direct causal relationship to outcomes. The impressive additional improvement in complete ST resolution that occurs between 60 and 180 min raises questions about the optimal time to assess the success or failure of pharmacologic reperfusion of acute myocardial infarction and the potential added value of cointervention with mechanical revascularization within this interval for such patients. The consistently excellent outcome of patients with complete ST resolution at 180 min, irrespective of which treatment regimen was employed, provides impetus for further investigation of this issue. In particular, the evaluation of enhanced early reperfusion with the addition of a glycoprotein IIb/IIIa inhibitor coupled with the potential for reduced reinfarction and presumable reocclusion with low molecular weight heparin and fibrinolytic therapy seems provided safety concerns can be addressed a worthwhile pursuit. 6,7 Acknowledgements It is a pleasure to acknowledge the significant contribution and participation of the ECG readers from the three Core Laboratories who participated in this study: Pushpa Jagasia, Norma Hill, Quamrul Hassan, Janna Luchansky, and the coordinating assistance of Soetkin Vlassak, André Beernaert, Shirley Bestilny, Kathy Shuping and the editorial assistance of Lynne Calder. The study was supported by Boehringer Ingelheim, Germany; Genentech, South San Francisco, CA, USA; and Aventis, Bridgewater, NJ, USA. References 1. ASSENT 3 Investigators. Efficacy and safety of tenecteplase in combination with enoxaparin, abciximab or unfractionated heparin: the ASSENT 3 randomized trial of acute myocardial infarction. Lancet 2001;358: Schroder R, Wegscheider K, Schroder K et al., for the INJECT trial group. Extent of early ST segment elevation resolution: a strong predictor of outcome in patients with acute myocardial infarction and a sensitive measure to compare thrombolytic regimens. JAmColl Cardiol 1995;26: Shah A, Wagner GS, Granger CB et al. Prognostic implications of TIMI flow grade in the infarct related artery compared with continuous 12-lead ST segment resolution analysis. J Am Coll Cardiol 2000; 35: Fu Y, Goodman S, Chang W-C et al. for the ASSENT 2 Investigators. Time to treatment influences the impact of ST segment resolution on one-year prognosis: Insights from ASSENT 2. Circulation 2001; 104: Ross AM, Molhoek P, Lundergan C et al. A randomized comparison of low molecular weight heparin, enoxaparin and unfractionated heparin adjunctive to tpa thrombolysis in aspirin (HART-II). Circulation 2001; 104: GUSTO V Investigators. Reperfusion therapy for acute myocardial infarction with fibrinolytic therapy or combination reduced fibrinolytic therapy and platelet glycoprotein IIb/IIIa inhibition: the GUSTO V randomised trial. Lancet 2001;57: Antman EM, Louwerenburg HW, Baars HF et al. Enoxaparin as adjunctive antithrombin therapy for ST-elevation myocardial infarction: Results of the ENTIRE-Thrombolysis in Myocardial Infarction (TIMI) 23 trial. Circulationr 2002;105:r27 34.