A meta-analysis on the benefit-to-risk ratio of Vitamin K antagonists (VKA) in the prevention of venous thromboembolism in major orthopedic surgery

Abstract number: P1958

Mismetti^* P., Laporte† S., Zufferey† P., Epinat† M., Decousus† H.

†Clinical Pharmacology Department, France ^*France;

Although VKA are routinely used for preventing venous thromboembolism in patients undergoing major orthopedic surgery, particularly in North America, their benefit-to-risk ratio relative to active comparators, especially low-molecular-weight heparins (LMWH), is debated. We therefore performed a meta-analysis of all available open-label, single- or double-blind prospective, randomized trials comparing adjusted-doses of VKA with placebo or no treatment, LMWH, dextran, intermittent pneumatic compression (IPC), anti-platelet agents, unfractionated heparin, or danaparoid in patients undergoing major orthopedic surgery of the lower limbs. The statistical analyses were performed on an intention-to-treat basis. Relative risks (RR) and corresponding 95% confidence intervals were calculated. In the control patients, the adjusted incidence (±SD) of deep vein thrombosis (DVT) was 52.5 ± 3.0%. VKA were more effective than placebo or no treatment for reducing DVT (n = 567, RR = 0.56 [0.37–0.84], P < 0.01) and clinical pulmonary embolism (n = 651, RR = 0.23 [0.09–0.59], P < 0.01). Although there was no significant difference in the risk of major hemorrhage, a higher rate of wound hematoma was observed in the VKA group (n = 162, RR = 2.91 [1.09–7.75], P = 0.03). In the LMWH patients, the adjusted incidence of DVT was 13.0 ± 0.5%. VKA were less effective than LMWH in preventing total DVT and proximal DVT (n = 9822, RR = 1.51 [1.27–1.79], P < 0.001; and n = 6131, RR = 1.51 [1.04–2.17], P = 0.028, respectively). The chronological evolution of the RR shows that the superior efficacy of LMWH over VKA was significant as early as 1994 and from this year onwards. Neither the type of surgery (hip vs. knee arthroplasty), nor the timing of administration of LMWH relative to surgery (pre vs. postoperatively) significantly modified the results. The differences between VKA and LMWH in major hemorrhage and wound hematoma were not significant. VKA were more effective than IPC in preventing proximal DVT (n = 534, RR 0.46 [0.25–0.82], P = 0.009), but less effective than danaparoid in preventing total DVT (n = 659, RR 2.11 [1.51–2.95], P < 0.001). There were no other significant differences between VKA and other comparators for any endpoints. Thus, in major orthopedic surgery patients, compared to no treatment or placebo, VKA are an effective thromboprophylaxis strategy at an acceptable cost in terms of bleeding complications. However, LMWH are more effective than VKA without significantly increasing the bleeding risk.