Author + information
- Received August 27, 1996
- Revision received November 12, 1997
- Accepted November 26, 1997
- Published online March 1, 1998.
- Christopher B. Granger, MD, FACCA,
- Richard Becker, MD, FACCB,
- Russell P. Tracy, PhDC,
- Robert M. Califf, FACC, MDA,
- Eric J. Topol, MD, FACCD,
- Karen S. Pieper, MSA,
- Allan M. Ross, MD, FACCE,
- Sherryn Roth, MD, FACCF,
- Costas Lambrew, MD, FACCG,
- Edwin G. Bovill, MDC,*,1,
- for the GUSTO-I Hemostasis Substudy Group
- ↵*Dr. Edwin G. Bovill, University of Vermont Medical Center, 111 Colchester Avenue, Burlington, Vermont 05401.
Objectives. We sought to assess the effects of antithrombotic therapy after thrombolysis for acute myocardial infarction on markers of thrombin generation and activity and to determine the relation of these markers with clinical outcomes.
Background. Thrombin activation and generation often occur with thrombolysis for acute myocardial infarction. Antithrombotic regimens have been developed to reduce the resulting thrombotic complications.
Methods. We sampled plasma markers of thrombin generation and activity after thrombolysis in 292 patients. We assessed the relations of these markers with clinical outcomes at 30 days.
Results. Fibrinopeptide A (FPA), a marker of thrombin activity toward fibrinogen, was elevated at baseline (12.3 ng/ml) and increased to 18.4 ng/ml by 90 min after streptokinase and subcutaneous heparin treatment. With intravenous heparin, this increase was attenuated, but intravenous heparin did not prevent thrombin generation, as measured by prothrombin fragment 1.2 (F1.2). Heparin level, measured by anti-Xa activity, correlated with activated partial thromboplastin time (aPTT, r = 0.62 to 0.67). Thrombin activity, measured by FPA, was as closely related to aPTT as to the heparin level. Baseline levels of F1.2 were significantly related to the risk of death or reinfarction at 30 days (p = 0.008); values 12 h after enrollment also were related to 30-day mortality (p = 0.05).
Conclusions. Although intravenous heparin partly suppresses the increased thrombin activity associated with thrombolysis, it does not inhibit thrombin generation. The aPTT was as good a measure of suppression of thrombin activity as the heparin level itself. Hematologic markers of thrombin generation were found to be related to the subsequent risk of thrombotic events.
Studies suggest that thrombin activity increases after thrombolytic therapy [1–3]and that heparin can lessen this increase . After alteplase and heparin therapy, infarct-related artery patency more often occurs in patients with longer activated partial thromboplastin times (aPTT) [5, 6]. Although aPTT measurement has been the standard clinical approach to estimating the effect of heparin on thrombin inhibition, factors other than heparin affect aPTT. Fibrinogenolysis, degradation of factors V and VIII and fibrin degradation products (which inhibit thrombin activity and fibrin formation) all prolong aPTT, especially early after thrombolysis . Lower levels of fibrinopeptide A (FPA), the peptide cleaved during the conversion of fibrinogen to fibrin, and lower levels of thrombin–antithrombin complex (TAT) also have been associated with greater infarct-related artery patency after thrombolysis [4–8]. Data correlating plasma hematologic markers during acute myocardial infarction with later clinical events are lacking.
Between 1990 and 1993, the Global Utilization of Streptokinase and TPA [tissue plasminogen activator—alteplase] for Occluded Coronary Arteries (GUSTO-I) trial enrolled 41,021 patients with acute myocardial infarction in 15 countries for a randomized comparison of four thrombolytic treatment strategies . An important aspect of the study was the comparison of treatments with different degrees of systemic fibrinolysis and different intensities of adjunctive heparin. The purposes of this substudy were 1) to evaluate the effects of various thrombolytic and heparin regimens on thrombin generation and activity; 2) to assess the relation of aPTT with thrombin activity in the setting of thrombolysis; and 3) to determine the relation of various measures of thrombin activity with clinical outcomes.
1.1 Participating Hospitals
Twelve North American hospitals (10 in the United States and 2 in Canada) participating in the GUSTO-I trial enrolled patients into the hemostasis substudy; eight of these hospitals also participated in the GUSTO-I angiographic substudy .
1.2 Patient Group
Eligibility criteria for GUSTO-I have been described . Briefly, patients who were within 6 h of symptom onset of acute myocardial infarction with electrocardiographic ST segment elevation were eligible if they had no history of stroke, active bleeding, recent major operation, recent noncompressible vascular puncture, previous enrollment in the trial or previous treatment with streptokinase or anistreplase.
Patients were randomly assigned to one of four treatment strategies: streptokinase, 1.5 million U over 1 h, with subcutaneous heparin, 12,500 U twice daily begun 4 h after the start of thrombolysis; streptokinase, 1.5 million U over 1 h, with intravenous heparin; accelerated alteplase, with a bolus dose of 15 mg followed by an infusion of 0.75 mg/kg body weight (up to 50 mg) over 30 min and 0.5 mg/kg (up to 35 mg) over the next hour, with intravenous heparin; or the combination of alteplase, 1.0 mg/kg over 1 h, not to exceed 90 mg, with 10% given as a bolus, and streptokinase, 1.0 million U over 1 h, with intravenous heparin. The intravenous heparin regimen was a bolus dose of 5,000 U and then 1,000 U/h for ≥48 h; the investigator could delay the heparin infusion (not the bolus) until the completion of thrombolytic therapy if intravenous access was lacking. After ∼10,000 patients were enrolled, aPTT was found to be below the target range in about half of GUSTO-I patients at 24 h, and aPTT was much more often below the target range in heavy patients. Thus, the recommendation was made to increase the initial dose of heparin to 1200 U/h for patients who weighed >80 kg.
1.4 Monitoring aPTT and Adjusting Heparin
The aPTT was assessed at 6, 12 and 24 h after the start of thrombolytic therapy in all patients assigned to intravenous heparin. The heparin dose was adjusted according to a nomogram to achieve a target aPTT range of 60 to 85 s. Based on the 6- and 12-h aPTTs, the heparin dose was to be titrated upward for aPTTs <60 s. However, because the systemic fibrinolytic state often induced by plasminogen activators is known to prolong aPTT early after thrombolysis, the heparin dose was not to be titrated downward based on 6- or 12-h aPTTs >85 s.
1.5 Sample Acquisition and Processing
Blood samples were to be drawn at baseline (before thrombolysis) and 12 and 24 h after the start of thrombolysis. In addition, angiographic substudy centers drew additional samples 90 min after the start of thrombolysis. At centers not involved in the angiographic substudy, additional samples were drawn at 6 and 36 h, and at one center (Scarborough General Hospital, Scarborough, Ontario, Canada) at 6 and 48 h and then daily until heparin was discontinued.
Blood samples were collected for assay of heparin levels (anti-Xa, anti-IIa), thrombin generation (prothrombin fragment 1.2 [F1.2]), thrombin activity (FPA), thrombin inactivation (TAT) and fibrinogen and fibrin breakdown (fibrinogen, d-dimer).
Samples were collected through either separate venipunctures or a large-bore arterial sheath (for patients enrolled in the angiographic substudy) using a 21-gauge butterfly Vacutainer collection set (Becton Dickinson) and 12-in. (30.5-cm) tubing with multiple Luer adapters. For all phlebotomies, the first 5 ml of blood was discarded.
Two tubes were collected at all time points: a 4.5-ml, 3.8% sodium citrate tube (Becton Dickinson) and a 5-ml tube with lyophilized d-phenyl-prolyl-arginyl-chloromethylketone (2.5 × 10−5mol/liter, final concentration) (Haematologic Technologies, Inc.), EDTA (4.5 mmol/liter, final concentration) and aprotinin (150 kallikrein inhibitor U/ml, final concentration) . The samples were immediately placed on ice and spun within 1 h of sample collection at 2,000 to 3,000gfor 10 to 15 min at room temperature. The supernatant was immediately frozen at −20° to −70° C.
1.6 Laboratory Assays
All assays were performed at the core laboratory (University of Vermont). Levels of FPA were quantified using the “RIA-mat” FPA radioimmunoassay (Byk-Sangtec, Marburg, Germany) using bentonite to absorb plasma fibrinogen. The FPA antibody used has a cross reactivity to fibrinogen of <4% and a lower detection limit of 0.1 ng/ml. The reference range for our laboratory was 1.0 to 7.1 ng/ml (n = 24). Prothrombin F1.2 was determined using a two-site, enzyme-linked immunosorbent assay (Baxter Dade Diagnostics, Inc.) using monoclonal anti-F1.2 antibodies directed against antigenic sites not available in the native prothrombin molecule, and the reference range was 0.03 to 0.38 nmol/liter (n = 16). The interassay coefficient of variation for the FPA assay was 7.2% and for the F1.2 assay, 10.0%.
Levels of TAT were measured by an enzyme-linked immunosorbent assay (Diagnostica Stago, Inc.) that uses an antibody to modified antithrombin III (AT-III) found in the AT-III protease complexes. The major component of the AT-III protease complexes is the combination of AT-III with thrombin, and throughout this report the results of this assay are referred to as thrombin–antithrombin complexes, or TAT. The reference range for this assay is <20 ng/ml, and the interassay coefficient of variation is 8%.
Heparin levels were measured by anti-Xa concentration by the amidolytic method (American Bioproducts). The interassay coefficient of variation for the assay was 8.4%.
1.7 Statistical Analysis
To identify and remove values most likely representing artifact, samples for which the FPA value was markedly elevated—defined arbitrarily but prospectively as ≥50 ng/ml, before any analysis of correlation with outcomes—were removed from the analysis . This comprised 60 (9.5%) of the 12- and 24-h samples. All hematologic data for a patient at the time point when the artifact occurred were removed from the analysis. The pattern of the relations between marker levels and outcomes did not change substantially when the analyses were repeated using all samples.
Descriptive statistics included median values with interquartile ranges as measures of central tendency and variation for continuous variables and frequencies for categoric variables. Analyses of the effect of thrombolytic regimen for each of the hematologic factors were performed using univariate repeated-measures techniques. This is a special type of general linear model in which each patient at each time point is included as a new observation into the model. In this way, we can model the amount of variation that occurs within as well as between patients. The variables included in the model as predictors were the thrombolytic regimen, the time of the measure drawn (baseline to >24 h) and the interaction between time and thrombolytic regimen. From each model, we were able to obtain an overall, or global, test of the effect of treatment on outcome. If the global test was not statistically significant at p < 0.05, then no further testing was performed on that outcome. If the global test was significant, then the individual tests were performed to see where the differences occurred. We recognize that this can produce many analyses if the global tests are significant. Because problems can occur with multiple comparisons within and between the several outcomes evaluated, all interpretations of results should be considered exploratory and designed primarily to generate questions for further studies.
Spearman correlations were calculated to identify relations between the factors themselves. Plots of the average aPTT versus FPA level for deciles of FPA at 12 and 24 h illustrated the relation between these factors. Similar plots were created for aPTT versus F1.2 and TAT.
The relation of FPA and F1.2 to 30-day mortality, reinfarction and death or nonfatal reinfarction were evaluated using logistic regression techniques. Because the relations were not linear, spline transformations were used to properly model the relation seen.
Baseline characteristics, treatment assignments and outcomes for patients enrolled in this substudy, compared with those for patients not in the substudy, are shown in Table 1. The overall baseline characteristics were similar, with the exception of a higher percentage of the substudy patients being enrolled in the United States. In the hemostasis substudy, there was a higher percentage of patients of African descent and fewer patients with diabetes than in the main trial cohort.
2.1 Baseline Anti-Xa Samples and Cumulative Distributions of FPA
Heparin levels at baseline, measured by anti-Xa activity, were <0.10 U in 85.6% of patients, indicating that although samples were drawn in the vast majority of patients before heparin was given, some patients had received heparin before enrollment. Among patients assigned to receive streptokinase and subcutaneous heparin, the median anti-Xa activity was 0.01 U at 90 min and 0.06 U at 6 h.
To determine whether having arterial sheaths in place appeared to artifactually increase FPA levels, we compared the samples from the patients in the angiographic substudy, who had arterial sheaths in place for the first 18 to 24 h, with those from patients not in this substudy, who generally did not have sheaths in place. We found no evidence of higher median FPA levels among patients with sheaths versus without sheaths in place either at 12 h (6.35 vs. 8.70 ng/ml) or at 24 h (6.90 vs. 8.40 ng/ml).
2.2 Hemostasis Markers According to Treatment Group and Time
Fig. 1, A–C illustrates FPA, F1.2 and TAT levels at different times according to treatment assignment. Baseline levels of FPA, F1.2 and TAT were all elevated. For patients treated with streptokinase and subcutaneous heparin (beginning 4 h after thrombolytic therapy), there was an abrupt increase in the FPA level early after thrombolysis, in contrast to patients assigned intravenous heparin, who had a steady decrease in the FPA level over the next 12 h. When examining the 49 patients treated with streptokinase who had sequential samples at baseline and at 90 min, there was an 8.4-ng/ml increase in FPA in patients treated with streptokinase and delayed subcutaneous heparin, compared with a 0.4-ng/ml increase in patients treated with streptokinase and intravenous heparin (p = 0.03). Levels of F1.2 were elevated in the first 24 h, regardless of intravenous heparin administration. Levels of TAT also rose after thrombolysis, independent of intravenous heparin therapy. Over the first 24 h, TAT levels were lower after treatment with streptokinase and subcutaneous versus intravenous heparin (p = 0.016) and after treatment with alteplase and intravenous heparin versus streptokinase and intravenous heparin (p = 0.007).
As expected, fibrinogen levels were lower during the first 24 h among patients treated with streptokinase rather than with alteplase. Median (25th, 75th percentiles) nadir fibrinogen levels, which occurred at 90 min, were 20 mg/dl (20, 25) for streptokinase with subcutaneous heparin, 27 mg/dl (20, 48) for streptokinase with intravenous heparin, 114 mg/dl (67, 230) for accelerated alteplase and 20 mg/dl (20, 32) for the combination thrombolytic regimen. d-Dimer levels (25th, 75th percentiles) were higher 90 min after treatment with accelerated alteplase, 1,305 μg/ml (555, 2,350) versus streptokinase, 835 μg/ml (330, 1,790) (p = 0.058), but were lower by 12 h: 675 μg/ml (285, 1,670) for alteplase and 1,075 μg/ml (515, 2,690) for streptokinase (p = 0.001).
2.3 Correlation of Heparin Levels and aPTT With Hemostasis Markers
Relations of FPA, F1.2 and TAT levels with aPTT are shown in Fig. 2, with correlation coefficients of these and other relations shown in Table 2. As expected, there was a significant relation, albeit not a high degree of correlation (r = 0.62 to 0.67) between heparin level (anti-Xa) and aPTT. At 6 h, aPTT and heparin levels were negatively correlated with FPA but not with F1.2, and by 24 h there was a significant relation between higher aPTT and lower F1.2. By 24 h, the relation between aPTT and FPA was at least as significant as that between the heparin and FPA levels.
2.4 Relation of FPA and F1.2 With Clinical Outcomes
Baseline marker levels according to 30-day survival are shown in Table 3. Higher baseline levels of F1.2 were associated with a greater likelihood of death or reinfarction (p = 0.008). Relations of FPA and F1.2 at 12 h with later reinfarction and death are shown in Table 4and Fig. 3. The lower likelihood of later reinfarction with lower levels of FPA was not significant (p = 0.13), and survival tended to be higher with lower levels of F1.2 (p = 0.05).
This study reveals a series of complex relations between pharmacologic therapy, hemostatic markers of thrombin activity and clinical outcomes. One of the major goals in the GUSTO-I trial was to test the effect of aggressive intravenous heparin with thrombolysis in a large trial, using the hypothesis that more effective suppression of thrombin activity would be important for patient outcome, especially after alteplase therapy. In fact, measures of thrombin activity, albeit partially suppressed by intravenous versus subcutaneous heparin, remained elevated regardless of the heparin regimen. Moreover, similar to the findings of Galvani et al. , generation of new thrombin was increased to a similar degree with both heparin regimens. The correlation of levels of markers of thrombin activity with clinical outcomes lends support to the hypothesis that these physiologic markers may play a role in reflecting clinically important effects of antithrombotic therapy.
3.1 Hemostasis Markers According to Treatment Group and Time
Baseline FPA levels, measured by radioimmunoassay, were elevated to 10 times normal, similar to other reports of patients with acute myocardial infarction . We found an increase in FPA 90 min after thrombolytic therapy; the increase was most pronounced in the absence of intravenous heparin, where the 42% increase in FPA levels over baseline was similar to the increase observed in previous reports [1, 2]. In contrast to the experience of Rapold et al. , in which intravenous heparin caused a decrease to the normal range in FPA at 90 min, accelerated alteplase with concomitant full-dose intravenous heparin was associated with a persistent elevation in FPA, albeit less elevated than with streptokinase and delayed subcutaneous heparin. This finding suggests only partial inhibition of thrombin with intravenous heparin.
Direct comparison of the change in FPA from baseline to 6 h with or without intravenous heparin confirms the findings of Galvani et al. , that intravenous heparin attenuates the early increase after streptokinase, but that neither heparin regimen fully suppresses thrombin activity.
Although heparin was expected to decrease thrombin’s activity on fibrin formation, whether heparin would decrease the generation of additional thrombin was less certain. Thrombin generation could be reduced by suppression of the feedback activation of factors V and VIII by thrombin, or increased if the predominant effect is a decrease in the generation (and inhibitory effect) of activated protein C. Similar to the findings of Merlini et al. , thrombin generation (measured by F1.2) persisted to a similar degree, independent of intravenous heparin administration.
Similar to F1.2, TAT levels increased after thrombolysis, most prominently among patients treated with streptokinase-containing regimens and intravenous heparin regimens. The TAT complex depends on available thrombin and AT-III; thus, TAT levels would be expected to increase in concert with the increases in F1.2 levels after thrombolysis. Because heparin increases the affinity of AT-III to thrombin, TAT levels would be expected to be higher among patients treated with intravenous heparin. The lower levels of TAT after alteplase therapy suggest that this plasminogen activator may result in the exposure and availability of less thrombin than after streptokinase therapy.
As expected, the reduction in fibrinogen level was more pronounced with streptokinase-containing treatment strategies. The fibrinogen level after alteplase therapy in our study was similar to that described by Neuhaus et al. and lower than that in the Thrombolysis In Myocardial Infarction studies using 100 mg of intravenous alteplase [16, 17], suggesting more extensive fibrinogenolysis with the accelerated dosing schedule. d-Dimer, a measure of fibrin breakdown, was highest at 90 min in the alteplase group, then lowest in the same group after 6 h, suggesting an intense early fibrinolysis with alteplase, but a lesser effect after the first several hours. This is consistent with clinical evidence of more potent clot lysis with accelerated alteplase at 90-min coronary angiography , higher rates of early intracranial hemorrhage and more potent longer term effects of streptokinase—higher rates of bleeding complications other than intracranial hemorrhage and lower rates of reocclusion.
3.2 Correlation of Hemostasis Markers With Outcome
Although the goal of suppression of thrombin activity has a sound pathophysiologic rationale, studies to date have been unable to examine relations between markers of thrombin activity and outcome because of small sample sizes. We were able to show a relation between higher baseline levels of thrombin activity and generation and higher 30-day mortality. We found that higher levels of F1.2 at 12 h after thrombolysis were associated with a trend toward a higher likelihood of later reinfarction. Whether the worse outcomes were because the assays identified a group of patients at higher baseline risk [19, 20], or because of inadequate suppression of thrombin generation and activity, is unknown and cannot be definitively determined because of the small number of events. Nevertheless, the findings that heparin does not suppress F1.2 and that F1.2 may be related to higher risk of clinical events support the concept that failure to inhibit thrombin generation may be a limitation of heparin.
3.3 Correlations Between Heparin Levels and Measures of Hemostasis
The original target aPTT was derived from a rabbit model of jugular vein thrombosis, in which a heparin level of 0.2 to 0.4 U/ml by protamine titration resulted in effective anticoagulation . For most reagents, aPTT associated with this heparin level is in the range of 1.5 to 3.0 times the control value. Some investigators have argued that the optimal measure of anticoagulant effect would be a more direct measure of whether thrombin activity or thrombosis has been suppressed—a measure such as FPA . We found a highly statistically significant but relatively weak correlation between aPTT and heparin levels, perhaps in part related to factors other than heparin levels affecting aPTT after thrombolysis. Based on the results of this study, however, aPTT appears to be an equally good measure of target heparin range as the heparin level itself, because aPTT correlates at least as well as the heparin level with FPA levels at 24 h by anti-Xa assay.
Measuring heparin effect early after thrombolysis is challenging, because the systemic effect of fibrinogenolysis may increase the aPTT, independent of its effect on thrombin. The results of this study were consistent with this concept—there was a weaker relation between heparin and FPA levels at 6 h (p = 0.01) than at 24 h (p < 0.001). The finding of a significant relation of both higher aPTTs and higher heparin levels with lower FPA levels, but not with lower F1.2 levels, early after thrombolysis is consistent with the previously discussed finding that heparin inhibits activity but not generation of thrombin.
3.4 Clinical Implications
These results indicate that heparin in conjunction with thrombolytic therapy has complex, sometimes inconsistent effects on thrombin. These findings also suggest that if complete inhibition of thrombin and prevention of thrombin generation are goals, heparin is not effective early after thrombolysis. More potent direct thrombin inhibitors, with or without agents that inhibit coagulation factors earlier in the cascade (such as factors Xa or V inhibitors or tissue factor pathway inhibitors), might prevent thrombin generation more effectively.
3.5 Study Limitations
Determinations of marker levels of thrombin activity are very susceptible to artifact. The fact that multiple centers drew, processed and handled the samples, in spite of careful training, may have increased the likelihood of inconsistent quality and artifact. Although femoral artery sheaths have been found to increase FPA levels, in this study FPA levels were similar among patients with and without sheaths in place, suggesting that this was not a major factor. The relatively small overall sample size is a substantial limitation, particularly when attempting to correlate marker values with infrequent outcomes. Because of the multiple comparisons, the absent or only borderline statistical significance of some of the correlations and the observational nature of the analyses, this study should not be interpreted as definitive evidence of the effect of heparin and the significance of markers of thrombosis, but should rather be considered supportive.
Despite its shortcomings, this study represents the most detailed characterization to date of markers of thrombin generation and activity after thrombolytic therapy. The findings suggest that additional studies are needed to define the relation of both thrombin generation and activity to clinical outcomes and to address the importance of inhibition not only of thrombin activity but also of thrombin generation.
A.1 GUSTO-I Hemostasis Substudy Investigators and Study Coordinators
Scarborough General Hospital, Scarborough, Ontario, Canada: S. Roth, J. Smith; George Washington University Medical Center, Washington, DC: A. M. Ross, K. Coyne; Maine Medical Center, Portland, Maine: C. Lambrew, J. Kane; Tulsa Regional Medical Center, Tulsa, Oklahoma: E. Pickering, P. Cotham; University of Michigan Medical Center, Ann Arbor, Michigan: E. R. Bates, E. Kline-Rogers; Mother Francis Hospital, Tyler, Texas: N. Israel, R. LeBoeuf; Victoria General Hospital, Halifax, Nova Scotia, Canada: C. Kells, T. Fawcett; University of Alberta Hospital, Edmonton, Ontario, Canada: J. Burton, C. Kee; Proctor Community Hospital, Peoria, Illinois: P. Schmidt, D. Miller; St. Mary’s Hospital, Rochester, Minnesota: S. Kopecky, A. McLaughlin; Duke University Medical Center, Durham, North Carolina: C. B. Granger, E. Berrios.
↵fn1 This study was supported by Genentech, Inc., South San Francisco, California; Bayer Corporation, New York, New York; CIBA-Corning, Medfield, Massachusetts; ICI Pharmaceuticals, Wilmington, Delaware; and Sanofi Pharmaceuticals, Paris, France.
- activated partial thromboplastin time
- antithrombin III
- prothrombin fragment 1.2
- fibrinopeptide A
- Global Utilization of Streptokinase and TPA for Occluded Coronary Arteries
- thrombin–antithrombin complex
- Received August 27, 1996.
- Revision received November 12, 1997.
- Accepted November 26, 1997.
- The American College of Cardiology
- Eisenberg PR,
- Sherman LA,
- Jaffe AS
- Eisenberg PR,
- Sherman LA,
- Rich M,
- et al.
- Rapold HJ,
- Kuemmerli H,
- Weiss M,
- Baur H,
- Haeberli A
- ↵Rapold HJ, de Bono D, Arnold AER, et al., for the European Cooperative Study Group. Plasma fibrinopeptide A levels in patients with acute myocardial infarction treated with alteplase: correlation with concomitant heparin, coronary artery patency, and recurrent ischemia. Circulation 1992;85:928–34.
- Hsia J,
- Kleiman NS,
- Aguirre FV,
- Chaitman BR,
- Roberts R,
- Ross AM
- Arnout J,
- Simoons M,
- de Bono D,
- Rapold HJ,
- Collen D,
- Verstraete M
- Gulba DC,
- Barthels M,
- Westhoff-Bleck M,
- et al.
- Lawler CM,
- Bovill EG,
- Stump DC,
- Collen DJ,
- Mann KG,
- Tracy RP
- Lidon RM,
- Theroux P,
- Juneau M,
- Adelman B,
- Maraganore J
- Galvani M,
- Abendschein DR,
- Ferrini D,
- Ottani F,
- Rusticali F,
- Eisenberg PR
- Merlini PA,
- Bauer KA,
- Oltrona L,
- et al.
- Neuhaus KL,
- von Essen R,
- Tebbe U,
- et al.
- Bovill EG,
- Terrin ML,
- Stump DC,
- et al.
- Rao AK, Pratt C, Berke A, et al., for the TIMI investigators. Thrombolysis In Myocardial Infarction (TIMI) trial—Phase I: hemorrhagic manifestations and changes in plasma fibrinogen and fibrinolytic system in patients treated with recombinant tissue plasminogen activator and streptokinase. J Am Coll Cardiol 1988;11:1–11.
- ↵Gore JM, Granger CB, Sloan MA, et al., for the GUSTO Investigators. Stroke after thrombolysis: mortality and functional outcomes in the GUSTO-I trial. Circulation 1995;92:2811–8.
- ↵Berkowitz SD, Granger CB, Pieper KS, et al., for the GUSTO-I Investigators. Incidence and predictors of bleeding after contemporary thrombolytic therapy for myocardial infarction. Circulation 1997;95:2508–16.
- Lee KL, Woodlief LH, Topol EJ, et al., for the GUSTO-I Investigators. Predictors of 30-day mortality in the era of reperfusion for acute myocardial infarction: results from an international trial of 41,021 patients. Circulation 1995;91:1659–68.
- Sobel BE