PROPHYLACTIC TREATMENT OF POSTOPERATIVE DEEP VENOUS THROMBOSIS IN ORTHOPEDIC-SURGERY OF THE HIP WITH ORAL ANTICOAGULANT

There are still many problems related to the prophylaxis of post-opera tive deep venous thrombosis (DVT), irrespective of the methods utilize d. The oral anticoagulant seem to be effective for this purpose, but i n general, few patients have been included in the studies and the resu lts are sometimes contradictory. We have therefore made a meta-analysi s of all the randomized trials in which Oral Anticoagulant (OA) were c ompared with a Control Group without any antithrombotic therapy, asses sing the incidence of post-operative DVT by a labelled fibrinogen test (LFT) or by phlebography, in orthopaedic surgery of the hip. The use of OA brought about a risk reduction of post-operative DVT amounting t o 64% +/- 11:p < 0.001, a risk reduction of symptomatic pulmonary embo lism of 78% +/- 21:p < 0.001, and a risk reduction of fatal pulmonary embolism of 74% +/- 18:p < 0.001, with a slight reduction of the risk of total mortality of 35% +/- 13:p < 0.01, probably explained by the o ld populations studied (over 70 years) and by the type of surgery stud ied (fracture). The risk of haemorrhagic complications on OA increased by 176% +/- 33:p < 0.001, but without any significant increase of sev ere haemorrhagic complications (p = 0.08). Thus, the prophylactic effi cacy of the OA on post-operative DVT appears to be clearly demonstrate d, but with an increased incidence of haemorrhagic complications, prob ably correlated to the high level of International Normalized Ratio (I NR) generally obtained (assumed to be more than 3). But low doses of O A, resulting in an INR between 2 and 3, seem to be just as efficacious and less haemorrhagic in orthopaedic surgery. Therefore, it could be of interest to use OA to take over early from prophylactic heparin the rapy, since the long-term incidence of post-operative DVT is not negli gible in patients at risk and because of a risk of thrombocytopenia br ought about by heparin. An optimum anticoagulation level should also b e defined through further randomized trials (INR probably between 1.5 and 3) so as to reduce the haemorrhagic risk.