Noninvasive diagnosis of deep vein thrombosis in postoperative patients.

The accuracy of noninvasive testing for the diagnosis of deep vein thrombosis (DVT) generally is less in asymptomatic patients than it is in those with symptoms suggestive of thrombosis. This is because asymptomatic DVT often is confined to the distal veins and, when it involves the proximal veins, the thrombi usually are smaller than in symptomatic patients with proximal thrombosis. Because the positive predictive value of noninvasive tests for asymptomatic DVT generally is 80% or less, abnormal results should be confirmed by venography. There are two main reasons why asymptomatic DVT is sought in the postoperative period: (1) to identify the need for full-dose anticoagulant therapy to prevent symptomatic episodes of venous thromboembolism (VTE), including fatal pulmonary embolism (this represents a form of secondary prophylaxis), and (2) to use this outcome as a surrogate for episodes of clinically important VTE in studies that are designed to evaluate methods of venous thrombosis prophylaxis. In relation to the first of these indications, evidence suggests that routine surveillance of high-risk patients to detect asymptomatic postoperative DVT does not result in improved clinical outcomes in patients who received appropriate VTE prophylaxis. In relation to the second indication, there is concern that asymptomatic VTE may not be a reliable surrogate for clinically important VTE, particularly if the effectiveness of different antithrombotic agents is being compared. Coupled with the comparatively low accuracy of noninvasive testing for asymptomatic DVT, this suggests that the results of such testing are unsuitable for the evaluation of new methods of prophylaxis in clinical trials.