Ultrasonic angioplasty was recently shown to ablate thrombi and atherosclerotic plaques in vitro and to recanalize occluded arteries in experimental animal models. The goal of the present study was to examine the clinical feasibility of ultrasonic angioplasty. Intraoperative ultrasonic angioplasty was performed in vivo on totally occluded peripheral arteries (n = 7). The ultrasonic angioplasty device consists of a 1.6-mm diameter flexible wire attached to a piezoelectric crystal generating ultrasound at 20 kHz. The controls, totally occluded human atherosclerotic femoral arterial segments (n = 6), were crossed mechanically with the ultrasound wire ex vivo but without application of ultrasonic energy. Ultrasonic angioplasty achieved successful recanalization without perforation in all vessels. Angiograms of the treated arteries showed an average lumen patency of 82.5%. Histological examination of the recanalized arteries revealed that the recanalization had taken place through intima diffusely involved with complicated plaque. The treated arteries, compared with the controls, had greater area of recanalized lumen (5.9 +/- 1.8 versus 1.7 +/- 0.4 mm2, p less than 0.05) and more flow (49.3 +/- 16.0 versus 11.8 +/- 4.9 ml/min, p less than 0.03). The damage in treated and control arteries was similar. Size-distribution analysis of the plaque debris from the treated arteries showed that 41 +/- 5% of the debris was 0.2-8 microns, 48 +/- 8% was 8-30 microns, and the remainder was 30-100 microns. In the mechanically crossed arteries, there was a shift in the distribution to larger size debris with 47 +/- 1% greater than 100 microns (p less than 0.001). Ultrasonic angioplasty may be a useful clinical method for recanalization of total occlusions in patients with peripheral vascular disease. Ultrasonic energy appears to cause controlled injury to the atherosclerotic intima by selectively disrupting the ultrasound-sensitive occlusion.