OBJECTIVE To determine criteria for a clinically significant vein stenosis with duplex ultrasound (DU) in patients with signs and symptoms of central venous outflow obstruction. METHODS Patients referred with swelling with or without pain to the vascular laboratory to detect vein obstruction were evaluated. These were mostly patients who had liver transplant, dialysis access, and tumors. All patients had DU prior to any other imaging. Only patients who subsequently underwent phlebography with intention to treat the vein stenosis were included in the study. A phlebogram with two views, pressure measurements across the stenosis, and intravascular ultrasound in selected cases were performed in all patients with suspected stenosis on DU. Adjacent ipsilateral normal vein segments were utilized as controls. The invasive tests were performed within 2 weeks of the DU. Follow-up was performed with DU at discharge and within 6 months of the procedure. A pressure gradient of =3 mm Hg across the stenosis was used to define a >50% diameter reduction, which was also determined by phlebographic measurement. RESULTS Thirty-seven patients, 20 males and 17 females, mean age 54 years, range 27 to 79, were evaluated. Forty-one stenotic venous sites were detected with DU; inferior vena cava 14, superior vena cava 2, portal 2, iliac 11, common femoral 3, brachiocephalic 3, subclavian 5, and axillary vein 1. Phlebography identified 37 of these stenoses and demonstrated two more not seen by DU. Pressure measurements confirmed 39 of those detected by DU. The best criterion by DU to detect a >50% stenosis was a poststenotic to pre-stenotic peak vein velocity ratio of 2.5. The presence of poststenotic turbulence and planimetric calculations of the diameter reduction increased the diagnostic confidence but not the accuracy. Using the pressure gradient of >/=3 mm Hg as a reference test, there were two false positive and two false negative exams with DU, while phlebography had two false negative exams. The overall agreement of DU alone was 90% of phlebography >95% and when combined 100%. Intravascular ultrasound identified correctly all 11 lesions in 11 patients. After angioplasty and stenting, there was a dramatic reduction in the edema in most patients particularly in those that had a caval stenosis. Restenosis was identified by DU in 5/29 (17%) patients at 6 months that were confirmed by phlebography and pressure measurements. Reintervention was performed in four and it was successful in three. CONCLUSIONS DU is a sensitive method to identify a clinically significant vein stenosis. A peak vein velocity ratio of >2.5 across the stenosis is the best criterion to use for the presence of a pressure gradient of =3 mm Hg. DU can be used to select patients for intervention and also to monitor the success of the treatment during follow-up.