The ability of different dobutamine-induced wall motion patterns to define the anatomic status of the infarct-related artery (IRA) was evaluated in 159 patients who underwent dobutamine stress echocardiography (DSE) and coronary angiography 10 +/- 2 and 18 +/- 3 days, respectively, after hospital admission. The DSE result was classified as: (1) biphasic: improvement with a low dose followed by deterioration with a high dose; (2) worsening: direct deterioration at low or high doses; (3) sustained improvement: improvement with a low dose that was maintained at high dose; and (4) no change: no change during the entire protocol. A diameter narrowing >70% (50% for the left main stem) of major coronary arteries indicated a severe lesion. Angiograms were classified according to the jeopardy score and collateral circulation graded according to Rentrop's classification. DSE was positive in 92 patients (22 had biphasic results and 70 had worsening results) and negative in 67 patients (14 had sustained improvement and 53 had no changes). Biphasic response was associated with more frequent anterior infarction (p <0.05) and higher resting (p <0.001) and peak (p <0.01) wall motion score indexes. The IRA was totally occluded in 4 of the 92 patients (4%) with positive (worsening pattern) and 12 of the 67 patients (18%) with negative (no change pattern) tests. The biphasic pattern was associated with the highest jeopardy score and was significantly (p <0.05) more specific (100%) compared with worsening (78%) in identifying a severe stenosis of the IRA. The combination of ischemic patterns provided a significantly superior sensitivity (p <0.0001). Logistic regression analysis identified the biphasic pattern as the only significant predictor. Conversely, the prediction of total occlusion of the IRA was poor. Sustained improvement was the most specific (100%) predictor of absence of severe stenosis of the IRA, whereas the combination with no change pattern provided a significantly superior sensitivity (p <0.0001). Thus, DSE effectively predicts the residual stenosis of the IRA. In particular, the biphasic response has an excellent specificity and positive predictive value and is the only significant predictor among clinical and echocardiographic variables.