Plot of the approximate odds ratios and 95% confidence intervals for death, recurrence of pulmonary embolism, or major bleeding event in the patients of the 9 studies analyzed and for the pooled data. The dotted line indicates the odds ratios across all studies.
Agnelli G, Becattini C, Kirschstein T. Thrombolysis vs Heparin in the Treatment of Pulmonary EmbolismA Clinical Outcome–Based Meta-analysis. Arch Intern Med. 2002;162(22):2537-2541. doi:10.1001/archinte.162.22.2537
Copyright 2002 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2002
In patients with acute pulmonary embolism, thrombolysis results in a more rapid resolution of pulmonary emboli than heparin treatment. Whether this advantage results in an improved clinical outcome is unclear. We sought to perform a clinical outcome–based meta-analysis of studies comparing thrombolytic and heparin treatment in patients with pulmonary embolism.
Data concerning adverse outcome events (death, recurrent pulmonary embolism, and major bleeding events) were extracted from the identified randomized studies.
A total of 56 (23.2%) of 241 patients treated with thrombolytic agents in 9 randomized trials experienced an adverse outcome event compared with 57 (25.9%) of 220 patients treated with heparin (relative risk [RR], 0.9; 95% confidence interval [CI], 0.57-1.32). In the thrombolysis group, 11 patients (4.6%) died compared with 17 (7.7%) in the heparin group (RR, 0.59; 95% CI, 0.27-1.25). Thirty-one patients (12.9%) undergoing thrombolysis had a major bleeding episode compared with 19 patients (8.6%) treated with heparin (RR, 1.49; 95% CI, 0.85-2.81). Five fatal bleeding episodes (2.1%) occurred in the thrombolysis group and none in the heparin group. Six studies provided data on recurrent pulmonary embolism. A recurrence occurred in 14 (6.6%) of 214 patients treated with thrombolytic agents and in 22 (10.9%) of 201 patients treated with heparin (RR, 0.60; 95% CI, 0.29-1.15). Recurrence and/or death occurred in 25 (10.4%) of 241 and in 38 (17.3%) of 220 patients treated with thrombolytic agents and heparin, respectively (RR, 0.55; 95% CI, 0.33-0.96; P = .03).
In patients with pulmonary embolism, thrombolysis had a lower composite end point of death/recurrence than heparin treatment. Excessive bleeding is the trade-off for improved efficacy. A comparative clinical outcome trial of thrombolysis and heparin treatment is warranted in patients with pulmonary embolism and selected for high risk of death and/or recurrence and low risk of bleeding.
PULMONARY EMBOLISM is a common and life-threatening disease. The incidence of a first episode of pulmonary embolism in acute care hospitals in the United States has been found to be 0.23 per 1000.1 In the 1960s, the introduction of heparin for treatment of pulmonary embolism cut down in-hospital mortality,2 but since then no further reduction has been achieved.3
Thrombolytic agents provide a more rapid lysis of pulmonary emboli and reduction of pulmonary hypertension than heparin.4 Whether these advantages result in an improved clinical outcome and outweigh the increased risk of bleeding complications is unknown. No conclusive clinical outcome studies aimed at comparing thrombolysis and heparin treatment in patients with pulmonary embolism are available. We performed a meta-analysis of studies comparing thrombolysis and heparin treatment in patients with pulmonary embolism with respect to the occurrence of death, recurrence of pulmonary embolism, and bleeding events.
Randomized clinical trials comparing thrombolysis and heparin for treatment of pulmonary embolism were identified by computer-aided search (MEDLINE), scan of meeting abstracts, and scrutiny of the reference lists of original research and review articles. Trials were included in this meta-analysis if data concerning death and major bleeding events were available. Data extraction from the individual trials was independently performed by the authors. Discrepancy between the results obtained by the authors concerning selection and extraction of data was resolved by consensus.
The primary outcome of this meta-analysis was the incidence of adverse clinical outcome events (death, recurrence of pulmonary embolism, and major bleeding events). Recurrence of pulmonary embolism was considered among the adverse clinical outcome events, when reported, because it has been recently shown to be a predictor of in-hospital mortality.5
An analysis of the overall bleeding events (major and minor) was performed. For the purpose of this analysis, bleeding was considered major when it was fatal, intracranial, associated with a decrease in hemoglobin level of at least 2 g/dL, or it required the transfusion of 2 or more units of red blood cells. Data on patients with intracranial and fatal bleeding were extrapolated from the more general population data of patients with major bleeding. Bleeding events were also considered separately as associated or not with invasive procedures.
No assumption regarding the comparability of patients between the different studies was used because all the statistical calculations are based on the comparison within each single trial. Nevertheless, the homogeneity of the studies was tested with an approximate χ2 statistic.
The analysis was performed cumulatively for the overall adverse outcome events, separately for each adverse outcome event (death, major bleeding event, and recurrence of pulmonary embolism), and for death and recurrence together. To avoid the overestimation of adverse events, patients with multiple events were counted only once.
The relative risk (RR), the approximate odds ratio with its corresponding 95% confidence interval (CI), and a 2-tailed P value were calculated for each trial and for the overall results. The number needed to treat and the number needed to harm were also indicated when appropriate. The results of each trial were analyzed on an intention-to-treat basis.
Eleven clinical trials comparing thrombolysis and heparin use for the treatment of pulmonary embolism were identified. Two trials were excluded from this analysis because they were nonrandomized.6,7 Therefore, 9 randomized trials,4,8- 15 2 of which were double-blind,14,15 were included in this analysis (Table 1). Patients included in double-blind studies accounted for 15.4% of the overall study population.
Overall, 461 patients were included in the analysis: 241 (52.3%) received thrombolytic therapy with streptokinase (31 patients), urokinase (102 patients), or recombinant tissue-type plasminogen activator (108 patients) in different doses and regimens, while 220 patients received heparin. The study end point was pulmonary reperfusion as assessed by angiography and/or by lung scan in most of the studies included in the analysis. With a single exception,9 the observation period was limited to the hospital stay, and no data were provided on the long-term clinical outcome (Table 1).
Adverse clinical events, death, recurrence of pulmonary embolism, and major bleeding events were first analyzed cumulatively. Across all studies, adverse clinical events occurred in 56 patients (23.2%) treated with thrombolytic agents and in 57 patients (25.9%) treated with heparin, resulting in an RR of 0.9 (95% CI, 0.57-1.32; P = .51) (Table 2 and Figure 1).
In the thrombolysis group, 11 patients (4.6%) died compared with 17 patients (7.7%) treated with heparin, corresponding to an RR of 0.59 (95% CI, 0.27-1.25; P = .51; number needed to treat, 31.6) (Table 2).
Thirty-one patients undergoing thrombolysis (12.9%) had major bleeding events compared with 19 (8.6%) treated with heparin, corresponding to an RR of 1.49 (95% CI, 0.85-2.81; number needed to harm, 23). Overall, 5 patients, all of them in the thrombolysis group (2.1%), had a fatal bleeding event. Fatal bleeding events were as follows: 3 gastrointestinal, 1 intracranial, and 1 cardiac tamponade. One additional patient undergoing thrombolysis died and had intracranial bleeding, but death was attributed to the recurrence of pulmonary embolism.
In 5 of the 9 studies, invasive procedures were performed, essentially catheter insertion for pulmonary angiography.4,8,9,11,14 The incidence of major bleeding was 15.2% (40/264) in studies requiring invasive procedures and 5.1% (10/197) in the remaining studies (RR, 2.98; 95% CI, 1.53-5.82; P<.001). The incidence of major bleeding in patients receiving thrombolytic treatment in studies with and without invasive procedures was 18.1% (25/138) and 5.8% (6/103), respectively (RR, 3.1; 95% CI, 1.32-7.30; P<.005). The incidence of major bleeding in patients receiving heparin in studies with and without invasive procedures was 11.9% (15/126) and 4.3% (4/94), respectively (RR, 2.80; 95% CI, 0.96-8.16; P<.05).
Four trials reported adequate data on overall bleeding events (major and minor),4,10,11,15 the rate ranging from 5% to 70% in the thrombolysis group and from 0% to 50% in the heparin group (Table 3). The rate of overall bleeding events was 42.6% and 26.0% in patients receiving thrombolysis and heparin, respectively, corresponding to an RR of 1.99 (95% CI, 1.61-4.58; P<.001). A statistically significant increase of overall bleeding in patients undergoing thrombolysis was observed in 1 double-blind trial.15
Six studies,4,8,10- 12,15 including 415 patients, reported the incidence of recurrence of pulmonary embolism (Table 2). This event occurred in 14 (6.6%) of 214 patients treated with thrombolysis and in 22 of the 201 patients treated with heparin, corresponding to an RR of 0.60 (95% CI, 0.29-1.15). When death and recurrence were considered together, an RR of 0.57 (95% CI, 0.31-0.89; P = .03) in favor of thrombolysis was observed. For death and recurrence taken together, the number needed to treat was 14.5.
Streptokinase, urokinase, and recombinant tissue-type plasminogen activator were used in 3, 2, and 4 studies, respectively (Table 1). The overall mortality was 8.3% (21/253) in patients included in the streptokinase and urokinase studies, while it was 3.4% (7/208) in patients included in the recombinant tissue-type plasminogen activator studies (RR, 0.41; 95% CI, 0.18-0.94). The overall incidence of major bleeding events in patients included in the streptokinase and urokinase studies was 14.6% (37/253), while it was 6.3% (13/208 patients) in patients included in the recombinant tissue-type plasminogen activator studies (RR, 0.43; 95% CI, 0.23-0.78). Among patients receiving thrombolysis, the incidence of major bleeding was 18.1% (24/133) in patients receiving streptokinase or urokinase and 6.5% (7/108) in patients receiving recombinant tissue-type plasminogen activator (RR, 0.36; 95% CI, 0.16-0.80). The incidence of major bleeding was similar in patients receiving heparin in the streptokinase or urokinase studies (10.8%) and in the recombinant tissue-type plasminogen activator studies (6.0%) (RR, 0.55; 95% CI, 0.22-1.40).
In the non–double-blind studies, 25.6% of the patients receiving thrombolytic therapy experienced an adverse event compared with 28.3% in the heparin group (RR, 0.91; 95% CI, 0.65-1.26). In the 2 double-blind studies, 11.9% of the patients receiving thrombolytic therapy experienced an adverse event compared with 10.3% in the heparin group,14,15 corresponding to an RR of 1.15 (95% CI, 0.27-4.64).
Most of the studies included in the present meta-analysis were designed to compare the effect of thrombolysis and heparin treatment on the rate and the speed of lung reperfusion in patients with pulmonary embolism. This meta-analysis focused on the occurrence of adverse clinical outcome events such as death, recurrence of pulmonary embolism, and major bleeding.
A risk reduction of 41% and 40% in favor of thrombolysis was observed for death and recurrence, respectively. When death and recurrence were analyzed together, a statistically significant risk reduction in favor of thrombolysis was observed.
The overall death rate across the studies was 6.1%. This death rate is lower than that observed in 2 recent registries in patients with pulmonary embolism.5,16 Indeed, the mortality rate was 9.6% during the in-hospital stay and 11.4% at 2-week follow-up in the Management and Prognosis of Pulmonary Embolism Trial (MAPPET) registry16 and International Cooperative Pulmonary Embolism Registry (ICOPER),5 respectively. The low death rate observed in the present meta-analysis suggests that patients at relatively low risk for adverse outcome were included in the analyzed studies. The inclusion of a low-risk population could not have favored the thrombolysis that conceivably should be particularly advantageous over heparin treatment in cases of severe pulmonary embolism.
The rate of major bleeding was 12.9% in patients receiving thrombolysis and 8.6% in patients receiving heparin treatment. These rates were influenced by the frequent use of invasive procedures, essentially catheter insertion for pulmonary angiography. This was particularly the case for patients receiving thrombolytic treatment. The rate of intracranial bleeding in patients receiving thrombolysis for pulmonary embolism was 0.8%, similar to that reported in patients with myocardial infarction treated with thrombolysis.17
The results of this analysis make critical the selection of patients with pulmonary embolism who could benefit from thrombolytic therapy. Indeed, this treatment should be reserved for patients at high risk for death and recurrence and at low risk for bleeding. The identification of a high-risk population could be facilitated by the finding of right ventricular dysfunction at echocardiography, which has been shown to be a predictor of in-hospital mortality18,19 and recurrence.20 Careful patient selection and the use of noninvasive diagnostic methods for pulmonary embolism could contribute to the reduction of adverse bleeding events.
The results of the present meta-analysis could have been influenced by the use of suboptimal thrombolytic agents and/or regimens. It should be noted that the thrombolysis treatments used in the analyzed trials differ largely in choice of thrombolytic agent and regimen. Indeed, some of the regimens used in the first studies are not conceivable nowadays. It is of interest that the incidence of death and major bleeding events was significantly higher in patients receiving streptokinase or urokinase than in patients receiving recombinant tissue-type plasminogen activator. This reading of the results leaves open the possibility that more effective and safer thrombolytic agents, administered in a more pharmacologically sound regimen, including shorter infusion time, could produce a higher clinical benefit in terms of safety and efficacy.
The studies included in our meta-analysis provide information limited to the short-term clinical outcome of patients with pulmonary embolism. Thus, it does not test the hypothesis that thrombolytic therapy could produce long-term clinical advantages such as a lower incidence of late recurrence of pulmonary embolism or chronic thromboembolic pulmonary hypertension.
In conclusion, the present meta-analysis of comparative studies in pulmonary embolism showed that death and recurrences are less frequent with thrombolysis than with heparin treatment. The excessive bleeding mainly associated with invasive procedures is the trade-off for this improved efficacy. A comparative clinical outcome trial of thrombolysis and heparin treatment in patients with pulmonary embolism and with a high risk of death and/or recurrence and a low risk of bleeding is warranted and should not be further delayed.
Accepted for publication May 2, 2002.
Corresponding author: Giancarlo Agnelli, MD, Sezione di Medicina Interna e Cardiovascolare, Dipartimento di Medicina Interna, Università degli Studi di Perugia, Via Enrico dal Pozzo, I-06123 Perugia, Italy (e-mail: email@example.com).