Changes in flow in the subclavian artery and vein resulting from the use of a hyperabduction maneuver during Doppler sonography in 20 volunteers were compared with retrospective findings in 16 patients clinically suspected of having thoracic outlet syndrome. Significant compression of the subclavian artery showed in the Doppler waveform as at least a doubling of peak systolic velocity or complete cessation of flow with hyperabduction; significant compression of the subclavian vein was diagnosed by complete cessation of blood flow or loss of atrial and respiratory dynamics in the waveform of the subclavian vein with hyperabduction. In volunteers, asymptomatic compression of the subclavian vein with arm abduction was seen in two (10%) and asymptomatic compression of the subclavian artery was seen in four (20%). Of the 16 patients, thrombosis of the subclavian vein was found in seven, compression of the subclavian vein with hyperabduction was found in six, and diagnoses other than thoracic outlet syndrome were established as the cause of pain in three. When duplex sonography was compared with venography, which was performed in 10 patients, one false-negative case was found because a subclavian vein thrombus had not been detected. The subclavian artery was examined in five of the six patients with positional compression of the subclavian vein; compression of the subclavian artery was found in three. The clinical significance of compression of the subclavian artery cannot be determined from our data because of the small number of patients involved. When the sonographic criteria of subclavian vein clot or compression resulting in a complete loss of normal venous phasicity with arm abduction and the clinical criterion of subsequent improvement in symptoms after curative surgery are used, color Doppler sonography is 92% sensitive and 95% specific for the diagnosis of thoracic outlet syndrome. This preliminary study shows that Doppler sonography has potential in the evaluation of thoracic outlet syndrome.