Objective: To reevaluate the accuracy of impedance plethysmography (IPG) for the detection of proximal deep-vein thrombosis (DVT).
Patients: A total of 384 of 390 consecutive outpatients referred with their first episode of clinically suspected DVT.
Setting: University-based tertiary care medical center associated with a cancer clinic.
Design: A retrospective analysis of a cohort of patients whose data were recorded and stored prospectively on a computerized data base over a 22-month period.
Measurements: Patients were evaluated by a physician and underwent IPG testing. Patients with abnormal IPG tests and those with normal IPG results in whom there was a high clinical suspicion of DVT or in whom follow-up IPG testing was not feasible were referred for venography. Venography and IPG results were interpreted by a panel of independent observers. Two models of the IPG instrument were used (Codman 200 and Electrodiagnostic Instruments 800).
Results: Venography (or compression ultrasound) was done in 57 patients with an abnormal IPG test and in 85 patients with normal IPG results. Impedance plethysmography was abnormal in only 37 of 56 patients with confirmed proximal-vein thrombosis (sensitivity, 66%; 95% CI, 52% to 78%). Of the 57 patients with an abnormal IPG result, 37 had DVT (positive predictive value, 65%). The sensitivity for the detection of proximal DVT did not differ between the IPG 200 and 800 instruments (sensitivity, 63% and 71%, respectively; P > 0.2). Of the 19 proximal-vein thrombi not detected by IPG, 12 (63%) were occlusive and 11 (58%) involved at least the popliteal and superficial femoral veins.
Conclusions: At our center, IPG has a far lower sensitivity for proximal-vein thrombosis than has been previously reported for symptomatic outpatients. The reason for this low sensitivity is unclear. Our findings indicate that centers using IPG as the initial diagnostic test for suspected DVT should be aware of this potential problem and should consider reevaluating the sensitivity of their IPG machines by performing venography in a cohort of their patients with normal test results.