Purpose: To review noninvasive methods for diagnosis of first and recurrent deep venous thrombosis and provide evidence-based recommendations for the diagnosis of deep venous thrombosis in symptomatic, asymptomatic, and pregnant patients.
Data sources: Accuracy (comparison with contrast venography) and management (safety of withholding anticoagulants when results were normal) studies that evaluated tests for diagnosis of deep venous thrombosis were identified from a MEDLINE search, personal files, and bibliographies of reviews and original studies.
Study selection: Prospective cohort studies (accuracy and management studies) and randomized comparisons (management studies) that satisfied predefined methodologic criteria were included.
Data extraction: Sensitivity, specificity, and positive and negative predictive values were determined for accuracy studies. Rates of venous thromboembolism during long-term follow-up of patients with normal results were determined for management studies.
Data synthesis: Data from individual studies were combined under a random-effects model. The accuracy of noninvasive tests was compared, with emphasis on within-study comparisons. Recommendations for diagnosis of deep venous thrombosis were developed by a multidisciplinary group and graded according to the strength of the supporting evidence. Venous ultrasonography is the most accurate noninvasive test for the diagnosis of a first symptomatic proximal deep venous thrombosis. However, neither ultrasonography nor impedance plethysmography is accurate in asymptomatic postoperative patients. Venous ultrasonography is less accurate for symptomatic isolated distal (calf) deep venous thrombosis than for proximal deep venous thrombosis, and the clinical utility of venous ultrasonography of the distal veins is uncertain. Withholding anticoagulant therapy in symptomatic patients with suspected deep venous thrombosis who have normal results on serial venous ultrasonography or impedance plethysmography is safe. Diagnosis of recurrent deep venous thrombosis requires evidence of new thrombus formation, such as a new noncompressible venous segment detected by venous ultrasonography, conversion of a normal result on impedance plethysmography to abnormal, or presence of an intraluminal filling defect on venography. Suspected deep venous thrombosis in pregnant patients can usually be managed with serial venous ultrasonography or impedance plethysmography. In symptomatic patients with a suspected first episode of deep venous thrombosis, clinical assessment and D-dimer testing are complementary to testing with venous ultrasonography and impedance plethysmography.
Conclusions: Patients with suspected deep venous thrombosis can usually be managed with noninvasive testing. However, if the results of this testing are nondiagnostic or are discordant with the clinical assessment, venography should be considered.