The aim of the study was to assess clinical/prognostic significance of exercise-induced ischemia in patients with healed myocardial infarction. From May 1988 to January 1991, 777 consecutive patients underwent a symptom-limited (Bruce protocol) treadmill test at least 1 year after myocardial infarction. Clinical and ergometric data were entered in a prospective way in our data base. The exercise-test was positive in 231 out of 777 patients and 2 different subgroups were retrospectively identified depending on criteria of interruption: 156 patients with painless exercise-ST depression; 75 patients with painful exercise-ST depression. The main results (mean +/- SD) were analyzed with Student t test and chi 2 test. Patients with silent ischemia had longer exercise duration (547 +/- 153 s versus 395 +/- 173 s; p < 0.001) and higher double product (22.98 +/- 0.5 versus 19.71 +/- 0.4; p < 0.001) than symptomatic patients. Ischemic threshold was lower (double product: 17.98 +/- 0.4 versus 21.22 +/- 0.4; p < 0.001 with onset of ST depression at 297 +/- 148 s versus 448 +/- 147 s; p < 0.001) and time to ST normalization was longer (368 +/- 155 s versus 234 +/- 212 s; p < 0.001) in patients with painful ischemia. Patients with angina and ST depression had significantly higher prevalence of downsloping ST depression in the recovery phase (68% versus 37%; p < 0.001) and a higher prevalence of treadmill exercise score indicating high risk (49% versus 3.2%; p < 0.001). The 2 groups when compared with 99 patients with negative test post-AMI were significantly different.(ABSTRACT TRUNCATED AT 250 WORDS)