Hemorrhagic Complications of Thrombolytic Therapy in the Treatment of Myocardial Infarction and Venous Thromboembolism Mark N. Levine, M.D., Chainnan Samuel Z. Goldhaber, M.D., F.C.C.P. Robert M. Califf, M.D. Joel M. Gore, M.D., F.C.C.P. Jack Hirsh, M.D., F.C.C.P.
Bleeding is the major side effect of thrombolytic therapy. It usually occurs at a site of vascular invasion but can be spontaneous. Intracranial bleeding is the most feared bleeding complication occurring in between 0.3 and I percent of patients treated with streptokinase (SK), anisoylated plasminogen streptokinase activator complex (anistreplase or APSAC) or recombinant tissue plasminogen activator (rtPA) for myocardial infarction. Bleeding is an inevitable consequence of effects of plasmin-mediated proteolysis. The frequency and severity of bleeding is influenced by the use of invasive procedures, by other patientrelated factors, and by the concomitant use of anticoagulant therapy. Plasmin produces dissolution of fibrin in the hemostatic plug which is the main mechanism for bleeding. It also cleaves fibrinogen, factor V and factor VIII, to produce a systemic hypocoagulable state. The proteolysis of fibrin and fibrinogen results in elevated plasma levels of fibrin/fibrinogen degradation products which can act as thrombin inhibitors, interfere with fibrin polymerization, and become incorporated into the fibrin component of hemostatic plugs to render them more susceptible to lysis. Platelet dysfunction may also play a role in the bleeding associated with thrombolytic therapy. 1-3 Plasmin can cleave the platelet surface receptors Gplb and Gp Ilb/llla which are important for platelet adhesion and aggregation. The plasminogen activators, SK and urokinase, have minimal fibrin specificity and always induce plasma proteolysis when administered systemically in doses that induce thrombolysis. APSAC has only moderate fibrin specificity, while rtPA is fibrin-specific and can induce fibrinolysis without inducing a plasma proteolytic state; although when administered in clinically recommended doses, it usually induces a plasma proteolytic state. Despite these differences in fibrin specificity, clinical studies in patients with acute myocardial infarction have shown that bleeding complications occur with equal frequency with the non364S
specific plasminogen activators and the fibrin specific plasminogen activators; findings which suggest that lysis of fibrin in the hemostatic plug is the major mechanism for bleeding. This review determines the hemorrhagic risk of thrombolytic agents used in the treatment of patients with myocardial infarction, deep vein thrombosis, and pulmonary embolism. MYOCARDIAL INFARCTION
While there is no doubt that thrombolytic therapy increases the risk of bleeding in patients with myocardial infarction, reliable estimates of the true incidence of bleeding in different clinical situations and for different subgroups of patients are difficult to obtain. This difficulty exists because the studies in which the incidence of bleeding was carefully documented using quantitative and explicit criteria are relatively small and because the large multicenter studies focused on mortality and reinfarction as the outcome and tended to underreport bleeding complications. The very large trials do, however, provide unbiased estimates of the relative incidence of bleeding when different thrombolytic agents are compared with each other or an untreated control subject. Review of Trials
Despite the problems with definitions, we attempted to use the same methodology as previously used in this forum. 4 Only clinical trials published in manuscript form with either a control group or a comparative thrombolytic group were used, and only trials which reported transfusion and intracranial hemorrhage rates were used. In addition, only intravenous thrombolytic trials were included. In this section, the trials were divided into the following categories: trials comparing thrombolysis with conservative care or placebo; small trials comparing different thrombolytic agents, some of which used explicit criteria to assess bleeding and so provide reliable estimates on the rates of bleeding; large trials com parHemonhagic Complications (Levine et al)
Table I -Streptolcinaae in the Treatment of Myocardial Infarction• Study
Treatment
No.of Patients
Major Bleeds(%)
Intracranial Bleeds(%)
SK Placebo SK Control SK Control SK Control SK Placebo
156 159 5,860 5,852 107 112 859 882 8,592 8,595
2 (1.3) 1 (0.6) 19 (0.3)
2 (1.3) 0
European Co-op• GISSI• White et al
7
ISAM• ISIS 2"
1
(1)
51 13 46 18
(5.9) (1.5) (0.6) (0.2)
0 0 4 (0.5) 0 27 (0.31) 13 (0.15)
ing different thrombolytic agents which provide more reliable estimates of relative risks of bleeding; and trials comparing invasive and less invasive strategies in the setting of thrombolysis. Invasive trials were defined as those in which a cardiac catheterization was performed to assess patency rates. Tables 1 through 3 demonstrate the results of trials comparing thrombolytic agents with conservative management. Table 1 presents results of trials of SK 9 vs control subjects;~ Table 2 presents rtPA vs control 1 16 subjects; (}. and Table 3, anistreplase vs control subject. 11-22 The reported rates of major bleeding are highly variable, but in four of five of the large trials, the incidence of major bleeding (including intracranial bleeding) was higher in the patients receiving thrombolytic therapy. 6 •8 •9 •13•17 The exception is the GISSI-1 study in which the reported incidence of major bleeding with SK was very low (0.3 percent) with no reported intracerebral bleeds in 5,860 treated patients. 6 This very low incidence of major bleeding is an extreme example of the underreporting of bleeding Table 2-rtPA in '.&eatment of Myocardial Infarction
Invasive Collen et al 10 Verstraete et a1 11 Guerrci et ala European Cooperative13 Topoletal" TPAT"
Noninvasive ASSET'•
Treatment rtPA Placebo rtPA Placebo rtPA Placebo rt PA Placebo rtPA Placebo rt PA Placebo rtPA Placebo
Study AIMS 11 Ikram et al 18
*All studies noninvasive.
Study
Table 3-Aniatreplaae in the Treatment of Myocardial Infarction•
No. of Patients
31 14 62 62 72 66 366
355 75 25 59 56
2,512 2,493
Major Intracranial Bleeds Bleeds (%) (%)
0 0 0 1 7 5 5 1 5 1 1 3
(1.6) (10) (7) (1.4) (0.3) (5) (4) (1) (6)
0 0 0 0 0 0 5 (1.4) (0.3) (1.3) 0 0 0
35 (1.3) 12 (0.4)
7 (0.28) 2 (0.08)
Julian et al 10 Bossaert"' Croydon" Meinertz and Kaspar"'
Treatment
No. of Patients
Anistreplase Placebo Anistreplase Control Anistreplase Control Anistreplase Control Anistreplase Control Anistreplase Control
502 502 76 73 45 45 48 39 35 30 69 61
Major Intracranial Bleeds Bleeds (%) (%) 25 (5) 9 (1.8) 0 0 0 0 5 (10) 0 0 0 5 (7) 4 (6)
5 (1) 2 (0.3) 0 0 0 0 0 0 0 0 0 0
*All studies noninvasive
rates which has been a feature of some of the very large mortality trials. The incidence of major bleeding in the ISIS 2 study of 8,592 patients randomized to the SK group was also low (0.6 percent) but three times higher than the placebo group. 9 The incidence ofintracranial hemorrhage was also low (0.31 percent), but approximately twice as high as the placebo group. In contrast, in the ISAM study of over 1,700 patients, the reported incidence of major bleeding was almost tenfold higher than the ISIS 2 study for both the SK (5.9 percent) and control (1.5 percent) groups, in which the incidence of intracranial bleeding was reported to be 0.5 vs 0.0 percent, respectively. 8 Two relatively large studies have compared rtPA with placebo. The ASSET study included 5,005 patients and reported over a threefold increase in the incidence of major bleeding (1.3 vs 0.4 percent) and intracranial bleeding (0.28 vs 0.08 percent) in the rtPA group. 16 A similar relative increase in major bleeding (1.4 vs 0.3) and intracranial hemorrhage (1.4 vs 0.3) was reported in the rtPA group in the European cooperative study which included 721 patients. 13 The only study using APSAC, the AIMS study which included a relatively large sample (1,004 patients) reported an approximately threefold higher incidence of major bleeding (5 vs 1.8 percent) and of intracranial bleeding (1 vs 0.3 percent) in the APSAC group. 17 Thus, there appears to be a greater than threefold increase in major bleeding and a twofold to fourfold increase in intracranial bleeding with thrombolytic therapy. The incidence of intracranial hemorrhage in the larger studies varied between 0.3 and 1.4 percent for an absolute increase of 0.2 to 1.0 percent; this appeared to be lower in patients treated with SK than with rtPA or APSAC. Bleeding complications in small trials comparing thrombolytic agents are summarized in Table 4.23-30 There was no apparent difference in the incidence of CHEST I 102 I 4 I OCTOBER, 1992 I Supplement
365S
Table 4-Small Comparative Studiea ofThrombolytic Agents in Myocardial Infarction
Study
Treatment
No. of Patients
Invasive Bassaud et al"'
rt PA Anistreplase TIMI-I" rt PA SK Verstraete et al"' rt PA SK TAMl-5"' rt PA UK Combination TAPS" rt PA Anistreplase Noninvasive Anderson et al"' Anistreplase SK White et al"' rt PA SK PAIMS"'' rt PA SK
84
Major lntracranial Bleeds Bleeds (%) (%) (2) (I) (29) (27) (3) (6) (8.7) (10) (12) (2.8) (8.1)
0 0 0 0 0 0 4 (2.1) 3 (1.5) 0 2 (1) 2 (I)
8 (4) 7 (4) 3 (2.1) 1 (1) 8 (9) 6 (7) (1)
l (1.5) 0 2 (1.4) 0 0
2
85 143 147 64 85 191 190 194 217 216
41 39 2 4 17 19 23 6 17
188 182 135 135 86
85
major bleeding between the different agents, and there was a tendency for major bleeding to be more common in the trials which used invasive procedures. The two large comparative trials of thrombolytic agents are presented in Table 5. 31 •32 The GISSI-2/ International trial randomized 20,891 patients in a factorial design to receive either rtPA or SK in combination with subcutaneous heparin or no heparin. 31 In patients assigned to heparin, the subcutaneous dose was 12,500 U starting 12 h after the initiation of the thrombolytic agent infusion and continuing on a twice daily schedule until hospital discharge. All patients were also given aspirin, 325 mg/day. The rate of major bleeding was higher with SK (0.9 vs 0.6 percent, p
Bleeding is the major side effect of thrombolytic therapy. It usually occurs at a site of vascular invasion but can be spontaneous. Intracranial bleeding is the most feared bleeding complication occurring in between 0.3 and I percent of patients treated with streptokinase (SK), anisoylated plasminogen streptokinase activator complex (anistreplase or APSAC) or recombinant tissue plasminogen activator (rtPA) for myocardial infarction. Bleeding is an inevitable consequence of effects of plasmin-mediated proteolysis. The frequency and severity of bleeding is influenced by the use of invasive procedures, by other patientrelated factors, and by the concomitant use of anticoagulant therapy. Plasmin produces dissolution of fibrin in the hemostatic plug which is the main mechanism for bleeding. It also cleaves fibrinogen, factor V and factor VIII, to produce a systemic hypocoagulable state. The proteolysis of fibrin and fibrinogen results in elevated plasma levels of fibrin/fibrinogen degradation products which can act as thrombin inhibitors, interfere with fibrin polymerization, and become incorporated into the fibrin component of hemostatic plugs to render them more susceptible to lysis. Platelet dysfunction may also play a role in the bleeding associated with thrombolytic therapy. 1-3 Plasmin can cleave the platelet surface receptors Gplb and Gp Ilb/llla which are important for platelet adhesion and aggregation. The plasminogen activators, SK and urokinase, have minimal fibrin specificity and always induce plasma proteolysis when administered systemically in doses that induce thrombolysis. APSAC has only moderate fibrin specificity, while rtPA is fibrin-specific and can induce fibrinolysis without inducing a plasma proteolytic state; although when administered in clinically recommended doses, it usually induces a plasma proteolytic state. Despite these differences in fibrin specificity, clinical studies in patients with acute myocardial infarction have shown that bleeding complications occur with equal frequency with the non364S
specific plasminogen activators and the fibrin specific plasminogen activators; findings which suggest that lysis of fibrin in the hemostatic plug is the major mechanism for bleeding. This review determines the hemorrhagic risk of thrombolytic agents used in the treatment of patients with myocardial infarction, deep vein thrombosis, and pulmonary embolism. MYOCARDIAL INFARCTION
While there is no doubt that thrombolytic therapy increases the risk of bleeding in patients with myocardial infarction, reliable estimates of the true incidence of bleeding in different clinical situations and for different subgroups of patients are difficult to obtain. This difficulty exists because the studies in which the incidence of bleeding was carefully documented using quantitative and explicit criteria are relatively small and because the large multicenter studies focused on mortality and reinfarction as the outcome and tended to underreport bleeding complications. The very large trials do, however, provide unbiased estimates of the relative incidence of bleeding when different thrombolytic agents are compared with each other or an untreated control subject. Review of Trials
Despite the problems with definitions, we attempted to use the same methodology as previously used in this forum. 4 Only clinical trials published in manuscript form with either a control group or a comparative thrombolytic group were used, and only trials which reported transfusion and intracranial hemorrhage rates were used. In addition, only intravenous thrombolytic trials were included. In this section, the trials were divided into the following categories: trials comparing thrombolysis with conservative care or placebo; small trials comparing different thrombolytic agents, some of which used explicit criteria to assess bleeding and so provide reliable estimates on the rates of bleeding; large trials com parHemonhagic Complications (Levine et al)
Table I -Streptolcinaae in the Treatment of Myocardial Infarction• Study
Treatment
No.of Patients
Major Bleeds(%)
Intracranial Bleeds(%)
SK Placebo SK Control SK Control SK Control SK Placebo
156 159 5,860 5,852 107 112 859 882 8,592 8,595
2 (1.3) 1 (0.6) 19 (0.3)
2 (1.3) 0
European Co-op• GISSI• White et al
7
ISAM• ISIS 2"
1
(1)
51 13 46 18
(5.9) (1.5) (0.6) (0.2)
0 0 4 (0.5) 0 27 (0.31) 13 (0.15)
ing different thrombolytic agents which provide more reliable estimates of relative risks of bleeding; and trials comparing invasive and less invasive strategies in the setting of thrombolysis. Invasive trials were defined as those in which a cardiac catheterization was performed to assess patency rates. Tables 1 through 3 demonstrate the results of trials comparing thrombolytic agents with conservative management. Table 1 presents results of trials of SK 9 vs control subjects;~ Table 2 presents rtPA vs control 1 16 subjects; (}. and Table 3, anistreplase vs control subject. 11-22 The reported rates of major bleeding are highly variable, but in four of five of the large trials, the incidence of major bleeding (including intracranial bleeding) was higher in the patients receiving thrombolytic therapy. 6 •8 •9 •13•17 The exception is the GISSI-1 study in which the reported incidence of major bleeding with SK was very low (0.3 percent) with no reported intracerebral bleeds in 5,860 treated patients. 6 This very low incidence of major bleeding is an extreme example of the underreporting of bleeding Table 2-rtPA in '.&eatment of Myocardial Infarction
Invasive Collen et al 10 Verstraete et a1 11 Guerrci et ala European Cooperative13 Topoletal" TPAT"
Noninvasive ASSET'•
Treatment rtPA Placebo rtPA Placebo rtPA Placebo rt PA Placebo rtPA Placebo rt PA Placebo rtPA Placebo
Study AIMS 11 Ikram et al 18
*All studies noninvasive.
Study
Table 3-Aniatreplaae in the Treatment of Myocardial Infarction•
No. of Patients
31 14 62 62 72 66 366
355 75 25 59 56
2,512 2,493
Major Intracranial Bleeds Bleeds (%) (%)
0 0 0 1 7 5 5 1 5 1 1 3
(1.6) (10) (7) (1.4) (0.3) (5) (4) (1) (6)
0 0 0 0 0 0 5 (1.4) (0.3) (1.3) 0 0 0
35 (1.3) 12 (0.4)
7 (0.28) 2 (0.08)
Julian et al 10 Bossaert"' Croydon" Meinertz and Kaspar"'
Treatment
No. of Patients
Anistreplase Placebo Anistreplase Control Anistreplase Control Anistreplase Control Anistreplase Control Anistreplase Control
502 502 76 73 45 45 48 39 35 30 69 61
Major Intracranial Bleeds Bleeds (%) (%) 25 (5) 9 (1.8) 0 0 0 0 5 (10) 0 0 0 5 (7) 4 (6)
5 (1) 2 (0.3) 0 0 0 0 0 0 0 0 0 0
*All studies noninvasive
rates which has been a feature of some of the very large mortality trials. The incidence of major bleeding in the ISIS 2 study of 8,592 patients randomized to the SK group was also low (0.6 percent) but three times higher than the placebo group. 9 The incidence ofintracranial hemorrhage was also low (0.31 percent), but approximately twice as high as the placebo group. In contrast, in the ISAM study of over 1,700 patients, the reported incidence of major bleeding was almost tenfold higher than the ISIS 2 study for both the SK (5.9 percent) and control (1.5 percent) groups, in which the incidence of intracranial bleeding was reported to be 0.5 vs 0.0 percent, respectively. 8 Two relatively large studies have compared rtPA with placebo. The ASSET study included 5,005 patients and reported over a threefold increase in the incidence of major bleeding (1.3 vs 0.4 percent) and intracranial bleeding (0.28 vs 0.08 percent) in the rtPA group. 16 A similar relative increase in major bleeding (1.4 vs 0.3) and intracranial hemorrhage (1.4 vs 0.3) was reported in the rtPA group in the European cooperative study which included 721 patients. 13 The only study using APSAC, the AIMS study which included a relatively large sample (1,004 patients) reported an approximately threefold higher incidence of major bleeding (5 vs 1.8 percent) and of intracranial bleeding (1 vs 0.3 percent) in the APSAC group. 17 Thus, there appears to be a greater than threefold increase in major bleeding and a twofold to fourfold increase in intracranial bleeding with thrombolytic therapy. The incidence of intracranial hemorrhage in the larger studies varied between 0.3 and 1.4 percent for an absolute increase of 0.2 to 1.0 percent; this appeared to be lower in patients treated with SK than with rtPA or APSAC. Bleeding complications in small trials comparing thrombolytic agents are summarized in Table 4.23-30 There was no apparent difference in the incidence of CHEST I 102 I 4 I OCTOBER, 1992 I Supplement
365S
Table 4-Small Comparative Studiea ofThrombolytic Agents in Myocardial Infarction
Study
Treatment
No. of Patients
Invasive Bassaud et al"'
rt PA Anistreplase TIMI-I" rt PA SK Verstraete et al"' rt PA SK TAMl-5"' rt PA UK Combination TAPS" rt PA Anistreplase Noninvasive Anderson et al"' Anistreplase SK White et al"' rt PA SK PAIMS"'' rt PA SK
84
Major lntracranial Bleeds Bleeds (%) (%) (2) (I) (29) (27) (3) (6) (8.7) (10) (12) (2.8) (8.1)
0 0 0 0 0 0 4 (2.1) 3 (1.5) 0 2 (1) 2 (I)
8 (4) 7 (4) 3 (2.1) 1 (1) 8 (9) 6 (7) (1)
l (1.5) 0 2 (1.4) 0 0
2
85 143 147 64 85 191 190 194 217 216
41 39 2 4 17 19 23 6 17
188 182 135 135 86
85
major bleeding between the different agents, and there was a tendency for major bleeding to be more common in the trials which used invasive procedures. The two large comparative trials of thrombolytic agents are presented in Table 5. 31 •32 The GISSI-2/ International trial randomized 20,891 patients in a factorial design to receive either rtPA or SK in combination with subcutaneous heparin or no heparin. 31 In patients assigned to heparin, the subcutaneous dose was 12,500 U starting 12 h after the initiation of the thrombolytic agent infusion and continuing on a twice daily schedule until hospital discharge. All patients were also given aspirin, 325 mg/day. The rate of major bleeding was higher with SK (0.9 vs 0.6 percent, p
