Objective: To elucidate those factors that contribute to the risk of major postoperative thromboembolism and perioperative bleeding tendency.
Design: Retrospective multiple logistic regression analysis.
Setting: 7 Scandinavian hospitals (6 Swedish and 1 Norwegian).
Subjects: 2070 patients undergoing elective major abdominal surgery.
Interventions: Patients were randomised to receive 2500 or 5000 XaI units of low molecular weight heparin daily.
Main outcome measures: Major thromboembolism (proximal deep vein thrombosis confirmed by phlebography or necropsy, or pulmonary embolism confirmed by scintigraphy or necropsy, or both). Bleeding tendency (bleeding complications which were not explained by local haemorrhagic lesions or by coexisting disease).
Results: Previous thromboembolism, leg fracture or arthroplasty, present leg ulcer or malignant disease, operating time longer than 150 minutes, preoperative transfusion of 2 or more units, and preoperative hospital stay of 6 days or more (but not age, body weight, or varicose veins) were independent predictors for major postoperative thromboembolism. The risk was significantly increased with an increasing number of such risk factors. The risk of developing a diffuse bleeding complication was dependent on the dose of low molecular weight heparin, particularly in patients without risk factors.
Conclusions: The use of a narrow definition of thromboembolism lead to a pattern of risk factors which was partly different from that found in previous studies, which were usually based on diagnosis with the 125I-fibrinogen uptake test.