Incidence and predictors of major hemorrhagic complications from thrombolytic therapy in patients with massive pulmonary embolism

PURPOSE: The risk factors for bleeding in patients receiving recombinant tissue-type plasminogen activator for massive pulmonary embolism are not known. PATIENTS AND METHODS: The hospital records of 132 consecutive patients who received recombinant tissue-type plasminogen activator for massive pulmonary embolism were retrospectively reviewed. Bleeding was estimated by using the bleeding severity index, a method previously validated in patients receiving anticoagulants. Multivariate stepwise logistic regression was used to identify independent risk factors for bleeding. Four other definitions of bleeding in large pulmonary embolism thrombolytic trials were also used, and the agreement among these criteria was assessed. RESULTS: According to the bleeding severity index, 33 patients (25%) had one or more major bleeding complications. Hemorrhage at the venous puncture site for angiography was the most frequent complication (15 patients, 11%). Major bleeding at the catheterization site was more common at the femoral site (14 of 63 patients = 22%) than at the brachial site (1 of 63 patients = 2%; P = 0.0004). The use of the five different bleeding definitions resulted in a variation in the major bleeding rate from 3% to 43%. The kappa coefficient varied from 0.07 to 0.84, indicating poor agreement between most of the classifications. CONCLUSION: The use of the femoral vein for pulmonary angiography was the only variable significantly associated with major bleeding. Most of the differences observed in the pulmonary embolism thrombolytic trials are likely related to the differences in the definition of bleeding rather than to the thrombolytic regimen.