Background— Microvascular obstruction (MO) is associated with large acute myocardial infarction and lower left ventricular (LV) ejection fraction and predicts greater remodeling, but whether this effect is abolished by contemporary antiremodeling therapies is subject to debate. We examined the influence of several infarct characteristics, including MO, on LV remodeling in an optimally treated post–acute myocardial infarction cohort, using contrast-enhanced cardiac magnetic resonance. Methods and Results— One hundred patients (mean age, 58.9±12 years, 77%men) underwent contrast-enhanced cardiac magnetic resonance at baseline (≈4 days) and at 12 and 24 weeks. The effects on LV remodeling (ie, change in LV end-systolic volume index [ΔLVESVi]) of infarct site, transmurality, endocardial extent, and the presence of early and late MO were analyzed. Mean baseline infarct volume index decreased from 34.0 (21.2) mL/m ² to 20.9 (12.9) mL/m ² at 24 weeks ( P <0.001). Infarct site had no influence on remodeling, but greater baseline infarct transmurality ( r =0.47, P <0.001) and endocardial extent ( r =0.26, P <0.01) were associated with higher ΔLVESVi. Early MO was seen in 69 patients (69%) and persisted as late MO in 56 patients (56%). Patients with late MO underwent significantly greater remodeling than those without MO (ΔLVESVi, +4.1 [13.4] versus −7.0 [12.7] mL/m ² , respectively, P =0.001); those with early MO only displayed an intermediate ΔLVESVi (−4.9 [13.0] mL/m ² ). Importantly, late MO was seen frequently despite optimal coronary blood flow having been restored at angiography. Conclusions— Late MO on predischarge contrast-enhanced cardiac magnetic resonance remains an ominous predictor of adverse LV remodeling despite powerful antiremodeling therapy and may be useful in the risk stratification of survivors of acute myocardial infarction. Clinical Trial Registration— URL: http://www.clinicaltrials.gov . Unique identifier: NCT00132093.