In a retrospective study of 365 chronic total occlusions that were submitted for angioplasty, we studied the influence of 27 clinical, morphologic, and procedural variables as possible predictors of successful outcomes. Success rate was shown to be significantly influenced by the following variables: operator experience (41% in early patients, first 6 months; 73% in late patients, last 6 months of entire series; p less than 0.001), duration of occlusion (less than or equal to 1 month, 89%; 1 to 3 months 87%; greater than or equal to 3 months 45%; unknown 60%; p less than 0.001), morphology of occlusion (tapered, 83%; abrupt, 51%), length of occlusion (less than or equal to 15 mm, 71%; greater than 15 mm, 60%; p less than 0.001), and bridging collaterals (present, 29%; absent, 67%; p less than 0.001). None of the other clinical, angiographic, or procedural variables correlated with the success rate of coronary angioplasty. The calculated probability for an experienced operator (greater than 100 occlusions attempted) to successfully open an occlusion with favorable morphology (less than or equal to 1 month old, short, tapered, without bridging collaterals) is 99%. An attempt by the same operator to open an occlusion with unfavorable structure (greater than or equal to 3 months old, long, untapered) has only 47% probability of success. The probability increases to 84% when the occlusion is tapered. We concluded that in addition to the duration and the length of occlusion, tapered morphology, bridging collaterals, and operator experience can predict successful angioplasty in chronic total coronary occlusion.