Risk of ICH with combination therapy for acute MI

Cardiology Review® Online, July 2004, Volume 21, Issue 7

Fibrinolysis is an effective and widely available treatment for acute myocardial infarction (MI). Over the past few years, manipulations of the lytic “cocktail” have been attempted to improve its clot-lysing efficacy and reduce the risk of bleeding. Modifying the plasminogen activator has simplified therapy by allowing bolus dosing, but it has not improved efficacy or reduced the bleeding risk.1,2 Efforts to improve antithrombin therapy, using a thrombin inhibitor or low-molecular-weight heparin instead of unfractionated heparin, have so far been blunted by an increase in the risk of intracranial hemorrhage (ICH) without improving coronary patency.3-5

Potentiating the antiplatelet component of the lytic cocktail, however, by adding a glycoprotein (GP) IIb/IIIa inhibitor has been shown to provide higher reperfusion rates compared with standard lytic regimens, even with reduced doses of heparin and plasminogen activator.6,7 In phase 2 trials, this higher lytic efficacy was not associated with a higher bleeding risk.8 The large Global Use of Strategies to Open Occluded Coronary Arteries (GUSTO) V mortality trial was planned to fully evaluate the benefit versus risk ratio of this combination therapy compared with a standard fibrinolytic regimen of reteplase and unfractionated heparin.

Patients and methods

The GUSTO V trial included 16,588 patients who went to the emergency department within 6 hours of onset of an ST-segment elevation acute MI.9 Patients at high risk for ICH, such as those with stroke in the past 2 years or blood pressure above 180 mm Hg systolic or above 110 mm Hg diastolic at any measurement after first medical contact, were excluded. Randomized treatments were either combination therapy with full-dose abciximab, reduced-dose heparin (60 U/kg bolus [maximum 5,000 U], followed by an infusion of 7 U/kg/hour) and half-dose recombinant plasminogen activator (r-PA; 5 U twice, 30 minutes apart), or standard fibrinolytic therapy with full doses of r-PA (10 plus 10 U) and heparin (5,000 U bolus followed by an infusion of 800 or 1,000 U/hour, according to a body weight cutoff of 80 kg). Both groups received unfractionated heparin for at least 24 hours, maintaining an activated partial thromboplastin time (aPTT) between 50 and 70 seconds using a specific heparin-adjustment nomogram. All of the patients also received aspirin at randomization and throughout the course of the study. Mortality at 30 days was the primary end point of the study.

All suspected stroke events, lasting longer than 24 hours or causing death, were evaluated by clinical and instrumental (computed tomography and magnetic resonance imaging) evidence. Stroke survivors underwent assessment of disability status between days 30 and 45 after discharge from the hospital. Any grade above 2 of the modified Rankin scale was defined as a disabling stroke. Univariate and multivariate analyses were used to identify significant predictors of stroke. Odds ratios (ORs) for the significant factors were estimated using a multiple logistic regression model of ICH.


In the whole study population, stroke was shown in 154 patients, 73 receiving reteplase (0.9%) and 81 receiving reteplase and abciximab (1.0%; OR, 1.10; 95% confidence interval [CI], 0.80—1.51). Of these strokes, 101 were hemorrhagic: 49 occurred with r-PA alone and 52 occurred with combination therapy (0.6% in both groups; OR, 1.05; 95% CI, 0.71–1.56). Median (25th–75th percentile) time from drug administration to ICH was 9.2 hours (5.9–22 hours) with r-PA alone and 5.5 hours (3.4–11 hours) with combination therapy (P = .048). Mortality at 30 days, disabling stroke, the composite of mortality and stroke, and the disability measured at follow-up were not significantly different between the two groups.

Older age, female sex, increased diastolic blood pressure, and concomitant (in the past 48 hours) treatment with digitalis were the only baseline variables significantly associated with higher ICH rates at logistic regression. Administration of heparin before randomization was also associated with a higher risk of ICH at multivariate analysis (1.21% versus 0.55% in non—heparin-pretreated patients; P = .041), but a second heparin bolus, the doses of the different boluses, postrandomization infusion rates, and aPTT values were not.

Multivariate analysis for ICH risk showed only age as having a significant association with randomized treatments. Younger age (< 70 years) was associated with a lower risk, and older age was associated with a higher risk with combination therapy (figure).


Evidence has accumulated since fibrinolytic agents began to be used for the treatment of acute MI, indicating that old age, hypertension, and previous stroke are the most powerful independent predictors of ICH following pharmacologic reperfusion therapy.10 Full comprehen-sion of the mechanisms underlying this association, however, is far from being reached. Contrary to a proven dose-related risk of intracranial bleeding with plasminogen activators and thrombin antagonists, an antiplatelet agent, such as aspirin, did not show such an effect when used in association with heparin or streptokinase.11 In addition, the experience gathered with GP IIb/IIIa blockers in tens of thousands of patients with acute coronary syndromes or undergoing percutaneous coronary interventions showed no excess risk of ICH with this powerful antiplatelet therapy.12

As far as efficacy is concerned, because the early coronary thrombus is predominantly “white,” a more antiplatelet-oriented lytic cocktail has the potential for, and actually has shown, a more powerful lytic effect. Thus, through painstaking dose-finding trials,6,7 a new thrombolytic cocktail was designed with full-dose abciximab (needed to block platelet activation almost completely) and lower doses of heparin and reteplase or alteplase. This regimen clearly showed faster and more complete reperfusion of the infarct-related artery at both the epicardial and microcirculation levels. In these trials, which included a total of slightly more than 1,700 patients, no excess risk of ICH was noted, but with very limited power.

Data from the GUSTO V trial show a low overall risk of ICH (0.6%), without significant differences between abciximab plus reteplase and reteplase only. The exclusion of patients with any previous stroke or uncontrolled hypertension may partially account for this encouraging result. Nonetheless, a significant association between age and randomized treatment was observed because younger patients receiving combination therapy were at lower risk and older patients were at higher risk compared with those receiving standard r-PA.

An explanation for this finding is not straightforward. Because the risk of ICH in lytic-treated patients has been attributed to cerebral amyloid angiopathy, a small vessel degenerative disease mostly associated with longstanding hypertension and aging, it can be hypothesized that elderly patients will have an aspecific response to a more aggressive lytic regimen. If combination therapy has been shown to provide faster and more complete recanalization of the infarct-related coronary artery, bleeding complications may also occur earlier and more frequently in the elderly, as shown by the present data, in which the group receiving combination therapy had a shorter time to ICH and an increased rate of ICH in elderly patients.

In search of a safer drug regimen, it is difficult to know whether the higher ICH risk in the elderly is attributable to abciximab or to amplified heparin activity because the aPTT controls in the GUSTO V protocol were set at fixed times after randomization, rather than at the time of any suspected stroke, and also because of the complexity of the coagulation cascade, particularly after interference with so many drugs. The overall experience obtained so far seems to implicate the heparin dose as being particularly relevant: (1) abciximab and full-dose heparin result in a higher bleeding risk compared with abciximab plus reduced-dose heparin13; (2) the bleeding risk is also increased with low-molecular-weight heparins when these agents are associated with full-dose plasminogen activators in a way that is dose-dependent and becomes concrete in patients older than 75 years5; and (3) patients in GUSTO V who were receiving heparin before randomization had a higher rate of ICH with both randomized regimens. Current consensus and guidelines point to the reduction of the heparin dose, particularly the bolus dose, to reduce the risk of major bleeding and ICH, especially in elderly and low-weight patients.14,15


Combination therapy with abciximab and a reduced-dose lytic agent is currently being tested in trials of so-called “facilitated primary angioplasty” for acute MI, a strategy of pharmacologic reperfusion during transfer for primary angioplasty. Based on experience obtained in GUSTO V, patients older than 75 years have either been excluded from these studies or they are being treated with even lower doses of r-PA and heparin compared with the GUSTO V regimen. Younger patients may benefit from the earlier and more complete myocardial reperfusion offered by this lytic regimen without paying the price of a higher bleeding risk.