IntroductionClinically relevant bleeding occurs three times as frequently as recurrent venous thromboembolism in the modern early treatment of pulmonary embolism (PE) with fixed-dose, unmonitored anticoagulants. Unfractionated heparin (UFH) is monitored and adjusted to assure efficacy and minimize bleeding risk, but low molecular weight heparin (LMWH) is not. PE requires more anticoagulant than isolated deep venous thrombosis. Speculating that PE with low clot burden could lead to excess bleeding with unadjusted LMWH treatment but not with UFH, we compared PE patients receiving either UFH or LMWH with high and low clot burden for clinically significant bleeding in an observational study. Materials and MethodsPatients with acute PE at multiple Chinese teaching hospitals had been randomized to UFH or LMWH for initial treatment. These treatment cohorts had baseline measurement of pulmonary artery obstruction (PAO) score, which was prospectively separated into quartiles, lowest to highest PAO. All patients were followed for bleeding episodes, which were subsequently analyzed by quartile of PAO. ResultsTwo hundred seventy-four patients divided between the two groups had similar efficacy and safety outcomes (12 clinically significant bleeds in the UFH group vs 15 in the LMWH group). LMWH recipients with the smallest clot burdens (lowest PAO quartiles) had highest bleeding rates (Cochran-Armitage trend test, P trend=0.048), but there was no such trend for UFH recipients. ConclusionsFor UFH, excess anticoagulant pro-hemorrhagic potential is down-adjusted via activated partial thromboplastin time monitoring, but for LMWH it is not. For PE patients at high bleeding risk, UFH may be safer if the clot burden is small.