From 1982 to 1989, ECD was performed on 18 patients suffering from poorly-tolerated massive pulmonary embolism, for whom classic treatments (fibrinolytics and surgery) were impossible. Eleven of these 18 patients immediately improved (S group). This procedure was unsuccessful in other seven patients (F group). Thirteen patients survived (72 percent). The time lag between the first episode of pulmonary embolism and ECD was significantly shorter in the S group than in the F group (4.7 +/- 5.4 days vs 18.3 +/- 6.9 days, p = 0.0004). So was the elapsed time between the onset of hemodynamic impairment and ECD (13 +/- 12 hours vs 59 +/- 38 hours, p = 0.003). We conclude that ECD should be considered when other treatments are impossible especially when the first symptoms date back less than 15 days and the hemodynamic impairment less than 48 h.