Author + information
- Received September 26, 1997
- Revision received February 17, 1998
- Accepted February 25, 1998
- Published online June 1, 1998.
- Brian K Metz, MDa,
- Harvey D White, DSc∗,
- Christopher B Granger, MD†,
- R.John Simes, MD‡,
- Paul W Armstrong, MD§,
- Jack Hirsh, MD∥,
- Valentin Fuster, MD, PhD, FACC¶,
- Cynthia M MacAulay, MS†,
- Robert M Califf, MD, FACC†,
- Eric J Topol, MD, FACCa,* (, )
- for the Global Use of Strategies to Open Occluded Coronary Arteries in Acute Coronary Syndromes (GUSTO-IIb) Investigators
- ↵*Address for correspondence: Dr. Eric J. Topol, The Cleveland Clinic Foundation, Department of Cardiology, Desk F-25, 9500 Euclid Avenue, Cleveland, Ohio 44195
Objectives. We sought to show that hirudin might interact differently with streptokinase (SK) and tissue-type plasminogen activator (t-PA), which could reduce the incidence of death or reinfarction at 30 days.
Background. In a large-scale trial of patients with acute coronary syndromes, hirudin provided modest benefit compared with heparin. However, the interaction with thrombolytic agents was not specifically assessed.
Methods. Patients with symptoms of acute myocardial infarction and electrocardiographic ST segment elevation were treated with thrombolytic therapy and randomly assigned to receive hirudin or heparin.
Results. A total of 2,274 patients received t-PA, and 1,015 received SK. Baseline characteristics were balanced by antithrombin assignment. Among SK-treated patients, death or reinfarction at 30 days occurred more often in those treated with adjunctive heparin (14.4%) rather than hirudin (8.6%, odds ratio [OR] 1.78, 95% confidence interval [CI] 1.20 to 2.66, p = 0.004). Among t-PA–treated patients, the rates were 10.9% with heparin and 10.3% with hirudin (OR 1.06, 95% CI 0.81 to 1.38, p = 0.68; for treatment heterogeneity: chi-square 4.20, degrees of freedom [df] 1, p = 0.04). After adjustment for baseline differences between thrombolytic groups, the rates were 9.1% for SK with hirudin, 10.3% for t-PA with hirudin, 10.5% for t-PA with heparin and 14.9% for SK with heparin (for treatment heterogeneity: chi-square 4.5, df 1, p = 0.03), suggesting that the beneficial treatment effect of hirudin was limited to the SK-treated patients.
Conclusions. Hirudin interacts favorably with SK but not t-PA, highlighting the importance of thrombin activity after SK therapy and the potential for simulating the effects of a more potent fibrinolytic agent through direct antithrombin therapy.
Since the results of the Global Utilization of Streptokinase and Tissue Plasminogen Activator [tissue-type plasminogen activator] (t-PA) for Occluded Coronary Arteries (GUSTO-I) trial were published in 1993 (1), accelerated t-PA combined with intravenous heparin has become the thrombolytic regimen used most often to treat myocardial infarction (MI) in the United States. This regimen provided a 15% relative reduction in 30-day mortality compared with streptokinase (SK) and either intravenous or subcutaneous heparin (2). No other combination of a thrombolytic agent and an antithrombin agent has since yielded an incremental survival benefit as great.
The major limitations of heparin, the standard antithrombin therapy, are that it works indirectly, requires antithrombin III as a cofactor and is ineffective against clot-bound thrombin (3). Newer antithrombotic agents, such as hirudin (the prototypic direct thrombin inhibitor), act against both free and clot-bound thrombin and lack most of the qualities that can interfere with the action of heparin. The advent of these drugs raised the hope that their concomitant use would improve the efficacy of SK (4–7). However, neither heparin nor any other antithrombin agent has been shown, when combined with SK, to improve 30-day survival (8,9).
SK remains the least expensive thrombolytic agent, and enhancing its efficacy with adjuvant therapy would have significant public health and economic effects. The primary hypothesis of the current study was that hirudin might interact differently with SK and t-PA, which could reduce the incidence of death or reinfarction at 30 days for one of these agents. This substudy was a prospective secondary analysis of the Global Use of Strategies to Open Occluded Coronary Arteries in Acute Coronary Syndromes (GUSTO-IIb) trial (10).
The GUSTO-IIb trial enrolled 12,142 patients with acute coronary syndromes, of whom 3,457 (28.5%) received thrombolytic therapy, either at the discretion of the attending physician or within a randomized substudy (10,11). When patients were randomized to hirudin or heparin treatment, they were stratified by the presence of ST segment elevation ≥1 mm in at least two contiguous leads associated with chest discomfort within the previous 12 h. Of the 3,457 patients who received thrombolytic agents, 3,052 were considered to meet the criteria for ST segment elevation and 405 were not. Of these 405 patients who did not have initial ST elevation ≥1 mm, we included in our analysis those who had ST segment elevation ≥0.5 mm who received thrombolytic therapy within 6 h after enrollment. We excluded only 14 patients who received both SK and t-PA from analysis. The resulting cohort used for all analyses therefore included 3,289 patients, of whom 3,041 were stratified in the ST segment elevation group and 248 were not.
Patients were excluded from the GUSTO-IIb trial for active bleeding, a history of stroke, contraindication to heparin, renal insufficiency (serum creatinine concentration >2.0 mg/dl [177 μmol/liter]), systolic blood pressure >200 mm Hg or diastolic blood pressure >110 mm Hg, warfarin use at enrollment or childbearing potential. If a patient was enrolled and it was later determined that the serum creatinine at enrollment had exceeded 2.0 mg/dl, the study drug was discontinued.
In patients with ST segment elevation, after they were randomly assigned to hirudin or heparin therapy, the decision to use thrombolytic therapy was made by the attending physician, as was the choice between an accelerated t-PA regimen or SK. A subset of patients eligible for thrombolysis was enrolled in the GUSTO-IIb direct angioplasty substudy (11), in which patients were randomized to undergo coronary angioplasty or receive accelerated t-PA as in the GUSTO-I trial (2).
The study medication (hirudin or heparin) was to be infused for 3 to 5 days. The heparin doses were those used in GUSTO-I (2): a 5,000-U intravenous bolus followed by an infusion of 1,000 U/h, with no increase in infusion rate for heavier patients. The hirudin dose was a bolus of 0.1 mg/kg body weight followed by an infusion of 0.1 mg/kg per hour. The activated partial thromboplastin time (aPTT) was measured before administration of study drug; at 6, 12 and 24 h and at least daily for the duration of the infusion. The study drug infusion rate was adjusted by use of a standard nomogram to maintain the aPTT between 60 and 85 s. Study drug infusion rates could be decreased (but not increased) in response to the measures at 6 and 12 h.
All patients were given chewable aspirin (160 mg) on presentation, followed by up to 325 mg of oral aspirin daily. The use of all other medications and interventions was at the discretion of the attending physician.
The primary end point was the composite incidence of death or nonfatal reinfarction within 30 days of enrollment. The 30-day rates of mortality, nonfatal reinfarction and a composite of death or reinfarction in the subgroup of patients with ST segment elevation were prospectively defined secondary end points of the GUSTO-IIb trial (10). These end points also were assessed at 24 h because they would reflect differences in outcomes between groups during infusion of the study drug. In-hospital bleeding and selected 30-day clinical outcomes were also assessed.
Severe or life-threatening bleedingwas defined as intracranial hemorrhage or bleeding that caused hemodynamic compromise requiring intervention. Moderate bleedingwas defined as spontaneous gross hematuria, hematemesis or a decline in hemoglobin >3 g/d if associated with observed blood loss or >4 g/dl if no bleeding site was identified.
Statistical analyses included all randomized patients who met the criteria for inclusion in our substudy, according to the intention to treat principle. Comparisons of the treatment groups with regard to the primary and discrete secondary end points were performed using the conventional chi-square test. All p values are two-tailed; p < 0.05 was considered statistically significant.
A multivariable logistic regression model was used to assess the possibility of a specific thrombolytic–antithrombin interaction after adjusting for baseline differences and, if found, to investigate its effect on the composite of death or nonfatal reinfarction at 30 days. Variables included in this model were age, gender, race, smoking status, hyperlipidemia, hypertension, diabetes, peripheral vascular disease, family history of coronary artery disease, site of enrollment (inside or outside the United States), cardiac history, recent history of malignancy, height, weight, systolic and diastolic blood pressures, heart rate and Killip class (Table 1). Because this model was developed to adjust for baseline imbalances between the t-PA and SK groups, it also allowed us to compare t-PA and SK (with either hirudin or heparin) with regard to the composite end point.
Patient treatment groups
Of the 2,274 patients treated with t-PA, 1,919 received t-PA rather than SK at the discretion of their physicians, and another 355 patients were randomized to receive t-PA as part of the direct angioplasty substudy. Of these 2,274 patients, 1,151 had been randomized to receive hirudin and 1,123 to receive heparin. Of the 1,015 patients treated with SK, 500 had been randomized to receive hirudin and 515 to receive heparin.
Within either thrombolytic group, patients assigned to receive heparin and those assigned to receive hirudin did not differ significantly at baseline (Table 1). However, patients treated with t-PA were younger and more often male, smokers or hypercholesterolemic than those treated with SK. Patients treated with t-PA also had more previous angina, and their Killip classification was less severe.
Within the SK group, hirudin treatment was associated with a 34% relative reduction in 30-day mortality compared with heparin treatment (p = 0.092) (Fig. 1). There was also a relative reduction of 50% (p = 0.007) in the rate of reinfarction at 30 days with hirudin and a relative reduction of 40% in the incidence of the composite 30-day end point of death or nonfatal reinfarction (p = 0.004). Within the t-PA group, treatment groups did not differ significantly at 30 days with regard to rates of death, nonfatal reinfarction or the two combined. The test for an interaction between the thrombolytic and antithrombin agents yielded an unadjusted chi-square value of 4.20 (degrees of freedom [df] 1, p = 0.04); that is, the treatment effect of hirudin was statistically different for the SK and t-PA groups.
At 24 h, patients fared better if they had received SK and hirudin than heparin (Fig. 2). There was a 49% relative reduction in mortality with hirudin (p = 0.093). Hirudin was also associated with an 89% relative reduction in the incidence of reinfarction (p = 0.013), but there were very few events. Similarly, hirudin was associated with a relative reduction of 61% (p = 0.009) in the incidence of the composite end point of death or nonfatal reinfarction at 24 h. Within the t-PA group, patients treated with heparin or hirudin did not differ significantly with regard to the 24-h end points.
Moderate bleeding occurred more often with t-PA than with SK treatment and least often in patients who received SK and heparin (Table 2). However, patients treated with SK with heparin were the most likely, and those treated with SK with hirudin were the least likely, to have severe bleeding. With either antithrombin agent, SK yielded a lower rate for the combination of moderate or severe bleeding than did t-PA. In the SK group, fewer hemorrhagic strokes were seen in patients treated with heparin than in those given hirudin (although event rates were very low). Differences among the four groups in bleeding complications were not statistically significant, nor were differences in bleeding outcomes by antithrombin agent within each thrombolytic group.
At 6 h after thrombolytic therapy began, within each thrombolytic group the median aPTT was significantly longer for those treated with heparin than for those treated with hirudin (Fig. 3). However, within each thrombolytic group the median values at 12 and 24 h were shorter for those treated with heparin than with hirudin. Over the first 24 h, aPTT values varied less overall in patients treated with hirudin. For hirudin, there was considerably less variability from the 25th and 75th percentiles at 6, 12 and 24 h, as shown in Figure 3. Ninety-four percent of patients in this substudy did have an MI by cardiac enzyme levels. All patients classified as having reinfarction did have an MI by measurement of enzyme levels on presentation.
Among patients treated with SK, cardiogenic shock occurred less often with hirudin than with heparin treatment (4.8% vs. 7.2%, p = 0.11) (Table 3). Similarly, recurrent ischemia occurred less often in patients treated with hirudin (21.2% vs. 25.4%, p = 0.11), as did refractory ischemia (7.2% vs. 10.9%, p = 0.04). However, stroke occurred more often in patients treated with hirudin (1.6% vs. 0.6%, p = 0.12). Among patients receiving t-PA, there were no significant differences between treatment groups with regard to any of these outcomes. The incidence of congestive heart failure did not differ by antithrombin agent within the thrombolytic groups.
A multivariable regression model was applied to adjust for known imbalances in baseline characteristics between the t-PA and SK groups. After adjustment, the risk of the composite 30-day end point of death or nonfatal reinfarction was similar in three groups: patients treated with SK and hirudin (9.1%), those treated with t-PA and hirudin (10.3%) and those treated with t-PA and heparin (10.5%). However, the risk was significantly lower in patients treated with SK and hirudin than with SK and heparin (9.1% vs. 14.9%, odds ratio [OR] 0.57, 95% confidence interval [CI] 0.38 to 0.87, p = 0.009, a relative reduction of 39%). The adjusted 30-day rate of death or nonfatal reinfarction for t-PA and hirudin treatment (10.3%) was similar to that for t-PA and heparin treatment (10.5%, p = 0.91, OR 0.98, 95% CI 0.74 to 1.30). The interaction between antithrombin treatment and thrombolytic agent was significant (chi-square 4.5, df 1, p = 0.03).
The present subgroup analysis of patients treated with thrombolytic therapy for acute myocardial infarction (AMI) in the GUSTO-IIb trial reveals a beneficial effect of hirudin with SK but not t-PA. As an adjunct to SK, hirudin resulted in a lower combined incidence of death and nonfatal reinfarction than did heparin, both at 24 h and at 30 days. However, as an adjunct to t-PA hirudin did not differ significantly from heparin in its effects on these composite end points. The interaction between the antithrombin and thrombolytic agents with regard to the prevention of death and reinfarction at 30 days remained significant after adjustment for independent variables. The combination of SK and hirudin had the lowest adjusted 30-day rate of the composite end point of the four strategies (9.1%, an absolute reduction of 5.8% from the 14.9% rate with SK and heparin treatment).
Although potential mechanisms for improved outcomes with hirudin as an adjunct to SK remain uncertain, the results of an angiographic infarct-related vessel patency trial with another direct thrombin inhibitor, hirulog (the Hirulog Early Reperfusion Occlusion [HERO] trial ) showed a significant increase in early (90 min) infarct-related vessel patency with hirulog compared with heparin among patients treated with SK. These data were corroborated by another patency trial of hirulog in patients with an AMI (7). In fact, the patency rates in both trials for SK with hirulog approximated that seen with accelerated t-PA with heparin in the GUSTO-I angiographic substudy (12). Two other randomized trials of SK and hirudin also have shown favorable results for this combination with respect to early infarct-related vessel patency (5)and clinical end points (4).
An angiographic substudy was not included in the GUSTO-IIb trial, but the aforementioned trials suggest that improved early infarct-related vessel patency is the likely mechanism for improved outcomes in the patients treated with SK and hirudin. Additionally, the reduced rates of refractory ischemia, recurrent ischemia and cardiogenic shock in patients so treated, consistent with the reduced rate of the composite end point of death or nonfatal reinfarction, suggest improved blood flow in the infarct-related artery. Although the incidence of bleeding complications did not differ significantly between hirudin and heparin, the slightly greater incidence of intracranial hemorrhage with hirudin (0.6% SK/hirudin vs. 0.2% SK/heparin) is noteworthy and is consistent with more extensive clot lysis. This hemorrhage rate is similar to that with t-PA (0.7% for both hirudin and heparin) and improved when compared with the results of previous trials that used higher doses of antithrombin agents (13–15).
Thrombolytic therapy induces both thrombin generation (measured by prothrombin fragment 1.2) and thrombin activity (measured by fibrinopeptide A). A laboratory study by Brommer and Meijer (16)has shown that SK accelerates thrombin generation more than t-PA, and thrombin generation was completely halted by the addition of hirudin to SK in the test medium. A substudy of 293 patients in the GUSTO-I trial (17)showed a greater increase in thrombin–antithrombin complexes after SK than t-PA, consistent with the hypothesis that SK exposes more thrombin than t-PA. These data, which show greater amounts and activity levels of thrombin after SK than t-PA, provide a pathophysiologic rationale for our findings.
Of five previous studies of SK and direct thrombin inhibition (4–7,18), only one did not show improvement in rates of death and reinfarction with direct thrombin inhibitors. In that trial, the Thrombolysis and Thrombin Inhibition in Myocardial Infarction (TIMI-9B) study (18), 1,496 patients were randomly assigned to receive heparin and 1,506 hirudin, at dosages similar to those used in GUSTO-IIb. These patients also received either t-PA or SK at the discretion of their physicians. The incidence of the composite end point of death or nonfatal reinfarction did not differ significantly by thrombolytic or antithrombin agent. Analysis of the timing of study drug administration after thrombolysis suggests a reason for the significant benefits observed in GUSTO-IIb but not in TIMI-9B. The study drug was given a mean of 35 min after thrombolysis in GUSTO-IIb, and systemic thrombin activity has been shown (19)to peak 30 min after administration of SK in patients with an AMI. The median time from randomization to institution of study drug was 44 min in TIMI-9B, possibly explaining why the additive benefit of hirudin with SK was not observed in that study. The patients in TIMI-9B who were treated with hirudin rather than heparin within 30 min of randomization did show a strong trend toward a lower rate of death or reinfarction than patients treated with either drug after 30 min (20). This finding emphasizes that early treatment with hirudin or hirulog may be important to obtain maximal benefit from these potent thrombin inhibitors (21).
The direct thrombin inhibitors are the first antithrombin agents to improve the outcomes of patients treated with SK. Delayed subcutaneous heparin did not improve survival when used with either conventional t-PA or SK in the combined data from the Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto Miocardico (GISSI-2) and the Third International Study of Infarct Survival (ISIS-3) trials, which together included 62,000 patients treated for AMI. With the addition of subcutaneous heparin, there was no significant survival benefit at 30 to 35 days, and there was an excess incidence of major bleeding, including an excess of about one cerebral hemorrhage per 1,000 patients treated (22–25). Similarly in the GUSTO-I trial, no incremental benefit was shown for intravenous than for subcutaneous heparin with SK; the mortality rate in both groups was nearly identical at 30 days (2). Thus, the role for intravenous heparin as an adjunct to SK in AMI remains controversial (26). Although a recent meta-analysis (27,28)of 26 randomized trials of anticoagulant therapy versus control in suspected AMI has questioned the need to add heparin to a regimen of aspirin and any thrombolytic agent, including t-PA, published guidelines (29)continue to recommend that heparin be used an adjunct to t-PA.
The limitations of the present study relate to the use of a subgroup of patients for analysis (27.1% of the total enrollment of GUSTO-IIb), even though the subgroup was specified in advance. The GUSTO-IIb trial was powered to detect a significant difference between the hirudin and heparin groups for the entire patient population, including those with and without ST segment elevation MI. Because patients were not randomized to receive t-PA or SK, comparisons between these thrombolytic agents must adjust for differences in baseline characteristics. Although we used a multivariate logistic regression model to perform such adjustments, the comparison of the thrombolytic agents should be regarded as exploratory.
The results of the present analysis support the combination of the direct thrombin inhibitor hirudin and SK as a promising alternative to the current standard therapy of t-PA with heparin. The more potent, direct thrombin inhibitors may improve clinical outcomes with SK, a less effective agent than t-PA when used with intravenous heparin (2). There may also be significant cost savings with SK and hirudin because t-PA currently costs ∼$1,800 more per treatment than SK. Further large, randomized trials are needed to confirm the benefits of using direct thrombin inhibitors with SK in acute ST segment elevation MI. Nevertheless, our findings confirm the importance of thrombin activity in patients receiving SK therapy.
☆ This study was supported by Ciba-Geigy, Summit, New Jersey and Boehringer Mannheim, Gaithersburg, Maryland.
- acute myocardial infarction
- activated partial thromboplastin time
- confidence interval
- degrees of freedom
- Global Use of Strategies to Open Occluded Coronary Arteries in Acute Coronary Syndromes
- myocardial infarction
- odds ratio
- Thrombolysis in Myocardial Infarction
- tissue-type plasminogen activator
- Received September 26, 1997.
- Revision received February 17, 1998.
- Accepted February 25, 1998.
- by the American College of Cardiology
- Becker R.C,
- Gore J.M,
- Lambrew C,
- et al.
- Molhoek G.P,
- Laarman G,
- Lok D.J,
- et al.
- 6.↵White HD, Aylward PE, Frey M, et al., on behalf of the Hirulog Early Reperfusion/Occlusion (HERO) Trial Investigators. A randomized, double-blind comparison of hirulog versus heparin in patients receiving streptokinase and aspirin for acute myocardial infarction. Circulation 1997;96:2155–61.
- Theroux P,
- Perez-Villa F,
- Waters D,
- Lesperance J,
- Shabari F,
- Bonan R
- Freudenberger R,
- Fuster V
- GUSTO IIa Investigators
- Antman E.M
- Neuhaus K.-L,
- van Essen R,
- Tebbe U,
- et al.
- Granger C.B,
- Becker R,
- Tracy R.P,
- et al.
- Antman E.M
- Owen J,
- Friedman K.D,
- Grossman B.A,
- Wilkins C,
- Berke A.D,
- Powers E.R
- Simes R.J,
- Granger C.B,
- Antman E.M,
- Califf R.M,
- Braunwald E,
- Topol E.J
- Rao A.K,
- Sun L,
- Chesebro J.H,
- et al.
- ISIS-3 (Third International Study of Infarct Survival) Collaborative Group
- Collins R,
- MacMahon S,
- Flather M,
- et al.
- Ryan T.J,
- Anderson J.L,
- Antman E.M,
- et al.