In a randomized, double-blind trial, Leclerc and colleagues [1] concluded that postoperative, fixed-dose exoxaparin is more effective than adjusted-dose warfarin in the prevention of deep venous thrombosis after knee arthroplasty. However, I do not believe that their conclusions and recommendations are justified.

As a reference value, Leclerc and colleagues cited incidences of deep venous thrombosis without prophylaxis ranging from 55% to 70%. However, it is my understanding that these values relate to earlier years, in which the standard of care did not include early ambulation of the patient. Is it not possible that the incidence today, without prophylaxis, would be much lower? Unfortunately, because the authors did not include a control group that received placebo only, we have no way to know what the true reduction from the use of any prophylactic regimen would be.

The study failed to consider the effects of compliance. It is much more likely that patients will be noncompliant with a regimen that consists of a twice-daily injection than with a regimen that consists of a once-daily oral dose. This substantially increases the risk for deep venous thrombosis.

The study results were not stratified by age or other subgroup. Is it not possible that the patients who were unable to ambulate were the ones with deep venous thrombosis and that the incidence of deep venous thrombosis was unrelated to the difference between the drug regimens?

The increase in the incidence of deep venous thrombosis with warfarin assumes that the warfarin dosage was appropriate; an international normalized ratio of 2.0 to 3.0 was used. However, given the much longer half-life of warfarin and the longer time required for warfarin to reach steady state (and fully inhibit the last of the anticoagulation factors), it is likely that the patients receiving warfarin were not fully anticoagulated until 3 to 5 days after therapy began, despite "adequate" international normalized ratios. One wonders what the results would have been if the patients had been simultaneously receiving intravenous or subcutaneous heparin for 3 to 5 days while waiting for warfarin to reach its full effect. Patients who are treated for deep venous thrombosis or pulmonary embolism in hospitals traditionally receive both heparin and warfarin for 3 to 5 days before heparin therapy is discontinued.

Finally, the authors did not cite any information on cost. Although costs differ between hospitals and the cost of monitoring warfarin therapy is another variable, does the additional effectiveness of enoxaparin reported by the authors justify the added cost of this medication?

The study results appear impressive. However, after considering the added cost of enoxaparin, the effect of the drug regimen on compliance, the current incidence of deep venous thrombosis in this type of surgery, the prevalence of deep venous thrombosis within different subgroups in the population, and the incidence of deep venous thrombosis had the dosage of warfarin been appropriate, I am not convinced that the general prophylactic use of enoxaparin for knee arthroplasty is warranted.