In this case-control study we investigated the preoperative risk factors and long-term results in patients undergoing emergency coronary artery bypass grafting (CABG) after failed percutaneous transluminal coronary angioplasty (PTCA). From January 1990 to March 1993, 56 emergency CABGs (Group A) were performed in patients from 7 different cardiology centres--six of these located outside our hospital within a radius of 19 km-1 h to 114 h (phi 7.3 h) after failed PTCA. No special stand-by was given and patients were operated in the next available theatre. Criteria of anatomically unfavourable stenoses for PTCA were defined as follows: Long stenoses, occlusions of other major vessels, tandem stenoses, stenoses at or near a bend or branch and, finally, eccentric stenoses. Thirty-two percent met one criterion, 35.8% two, 6.9% three and 1.8% four criteria. All patients were examined echocardiographically 14.6 +/- 8.2 months postoperatively. Patients who had been operated on electively within the same period served as a control (Group B; n = 56; case-control study). The patients were matched for age, sex, EF, LVEDP, body surface area, reoperation status, and concomitant diseases. The mortality rate for Group A patients was 10.7% (n = 6); for Group B, 1.8% (n = 1). Five of the Group A patients who died came from our own hospital and only one from a hospital 1 km away (mortality: 14.3% vs. 4.8%). There were no deaths among patients with an anatomically favourable stenosis for a PTCA; in patients with more than one criterion for an unfavourable stenosis mortality increased to 19.0% (p < 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)