Deep venous thrombosis (DVT) is common in total knee arthroplasty (TKA). Because of the rarity of the most serious outcomes, most randomized controlled trials lack the power to analyze these outcomes. A meta-analysis was performed for agents used in DVT prophylaxis in TKA employing a Medline literature search. Study inclusion criteria were randomized controlled trials comparing prophylactic agents in elective TKA with mandatory screening for DVT by venography. Fourteen studies (3,482 patients) met inclusion criteria. For total DVT, all agents except dextran and aspirin protected significantly better than placebo (P < .0001). For proximal DVT rates, low-molecular-weight heparin was significantly better than warfarin (P = .0002). There was a trend that aspirin was better than warfarin (P = .0106). No significant difference was found for symptomatic pulmonary embolism, fatal pulmonary embolism, major hemorrhage, or total mortality.