Restetenosis is still the greatest limitation of coronary angioplasty (PTCA). The systematic use of ergometry (PE) with the objective of identifying restenosis is controversial and, namely, the ACC/AHA does not recommend its routine use. Our objective was to conduct a retrospective study of the use of PE when performed late (3 to 6 months) for the detection of restenosis. As a protocol, our group performed "late" PE on all the patients without contraindications, the patients with positive ergometry or CCS class II-IV angor submitted to angiographic control. Between January 1996 and July 1997, 121 patients (pts) were submitted to PTCA. Our study population was composed of patients submitted to complete revascularisation with follow-up in our centre: 59 pts (49%) with an average age of 58 +/- 12 years, 82% male. Eighty-three percent of the pts had revascularisation in a context of unstable angina, 10% in the acute phase of myocardial infarction and 7% due to chronic angina. Stents were implanted in 42% of the pts. In the follow-up after six months, 7 pts complained of CCS class II or III angor. The ergometry showed positive electrocardiographic criteria in 11 pts (18.6%). All pts with angor had positive PE. All these pts were submitted to angiographic control; restenosis (residual stenosis equal to or above 50%) was observed in all the patients who complained of angor (100% positive predictive value); restenosis occurred in 9 pts with positive ergometry (82% positive predictive value). In asymptomatic pts, PE indicated 2 pts with restenosis (2/59-3.4%) and two false positive (2/11-18%). At six months, PE detected 22% of the pts with restenosis. In conclusion, complete post-revascularisation angor due to coronary angioplasty has a higher positive predictive value than ergometry. However, the stress test, performed systematically, can identify an additional percentage of pts with restenosis with an acceptable percentage of false positives.