OBJECTIVES Risk assessment of patients with aortic stenosis (AS) undergoing aortic valve replacement (AVR) is challenging. We set out to determine the impact of myocardial late gadolinium enhancement (LGE), as detected by cardiovascular magnetic resonance (CMR), on postoperative outcomes following AVR. METHODS A prospective observational study was conducted on patients undergoing CMR using the LGE technique within 1 year of subsequent AVR. Patients were categorized into absent, mid-wall or infarct patterns of LGE by independent observers blinded to all clinical data, and data were collected with regard to 30-day mortality, major adverse cardiac and cerebrovascular events (MACCE) and postoperative complications. RESULTS A total of 63 patients were studied. Twenty-five patients had no LGE; 20 had mid-wall LGE and 18 had an infarct pattern. The incidence of MACCE, cerebrovascular accident (CVA) and heart block were significantly higher in the mid-wall group compared with the other two groups (MACCE: 25 vs. 0 vs. 5%, P = 0.014; CVA: 20 vs. 0 vs. 0%, P = 0.013; heart block: 30 vs. 4 vs. 12%, P = 0.050). Patients with no LGE had no 30-day MACCE events and no deaths up to 2 years of follow-up. CONCLUSIONS The myocardial LGE holds promise as a means of predicting risk prior to AVR for AS.