Between 1988 and 1990, clinical testing was performed at 12 US institutions using the Simpson Coronary AtheroCath under an Investigational Device Exemption. Data on 1,069 lesions (873 patients) were analyzed and presented to the Food and Drug Administration (FDA) advisory panel in the summer of 1990, forming the basis for approval of this device in September 1990. Analysis of these preapproval data shows a primary success rate of 85% (defined as tissue removal, > or = 20% reduction in stenosis, < 50% residual stenosis after directional atherectomy, and no major complication), with somewhat higher primary success in prior restenosis and noncalcified lesions. Including the use of conventional angioplasty performed after atherectomy, the overall success rate was 92%. One or more major complications occurred in 4.9% of procedures, and included death (0.5%), nonfatal Q-wave myocardial infarction (0.9%), and emergency bypass surgery (4.0%). These complications were more frequent in right coronary, de novo, and diffuse (> 20-mm length) lesions. Six-month angiography results were available in 384 (77%) of 498 lesions eligible for follow-up when the registry closed and showed a restenosis rate (late stenosis > 50%) of 42%. The restenosis rate in both native vessels (30 vs 46%) and bypass grafts (31 vs 68%) was lower in primary (de novo) lesions compared with lesions that had developed restenosis after a prior intervention. Despite the use of prototype atherectomy catheters and still evolving procedural technique, this preapproval experience provided an important initial indication of the situations in which directional coronary atherectomy was most useful and helped set clear standards for performance of this procedure following FDA approval.