To determine predictors of acute coronary closure after PTCA performed with steerable catheter systems, we compared 140 procedures complicated by acute closure and 311 representative successful attempts from 4,772 procedures performed between April 1982 and March 1986. Sixteen clinical, 35 angiographic, and seven procedural variables were analyzed. Multivariate analysis found seven independent preprocedural factors related to closure: stenosis length of 2 or more luminal diameters, female gender, stenosis at a bend point of 45 degrees or more, stenosis at a branch point, stenosis-associated thrombus (filling defect or staining), other stenoses in the same vessel, and multivessel disease. In addition, four procedural factors were found to be associated with closure by univariate analysis: post-PTCA percent stenosis (p less than .001), intimal tear or dissection (p less than .001), use of prolonged heparin infusion (p less than .001), and post-PTCA gradient of 20 mm Hg or more (p = .004). Multivariate analysis of both preprocedural and procedural variables found six factors independently related to closure: post-PTCA percent stenosis, dissection, prolonged post-PTCA use of heparin, branch point location, fixed bend point location, and other stenoses in the vessel dilated. The risk of coronary closure after PTCA has many determinants. While an estimation of risk can be made before performing PTCA, the most powerful predictors of closure can only be assessed during the procedure itself.