Comparison of venography and ultrasound for the diagnosis of asymptomatic deep vein thrombosis in the upper body in children: Results of the PARKAA study

Deep vein thrombosis (DVT) in children occurs primarily in the upper body venous system. This prospective diagnostic study compared bilateral venography and ultrasound for detection of DVT in the upper venous system in 66 children with acute lymphoblastic leukemia. Results were interpreted by central blinded adjudication. Deep venous thrombosis occurred in 29% (19/66) patients. While 15/19 DVT were detected by venography (sensitivity 79%), only 7/19 were detected by ultrasound (sensitivity 37%). The 12 DVT detected by venography but not by ultrasound were located in the subclavian vein or more central veins. Three of 4 DVT detected by ultrasound but not by venography were in the jugular vein. We conclude that ultrasound is insensitive for DVT in the central upper venous system but may be more sensitive than venography in the jugular veins. A combination of both venography and ultrasound is required for screening for DVT in the upper venous system.