Background Elevated copeptin, a marker for vasopressin release, has been associated with impaired prognosis in acute myocardial infarction (MI). The aim was to investigate whether this association extends beyond the acute phase and whether it is related to markers of stress (cortisol) and heart failure (NTproBNP). Methods Copeptin, cortisol and NTproBNP were measured in 926 participants (age: 76.0; male: 48.5%) in the ICELAND MI study whereof 246 had a previous MI (91 recognizable (RMI) and 155 previously unrecognizable (UMI) detected by cardiac magnetic resonance imaging). The primary endpoint was cardiovascular events (CVEs), and secondary endpoints were total mortality, heart failure and MI (median follow‐up was 9.1 years). The relation between copeptin and prognosis was assessed with the Cox proportional hazard regression (unadjusted, adjusted for cortisol and NTproBNP, respectively, and a multiple model: copeptin, cortisol, NTproBNP, age, sex, serum creatinine, heart failure). Results Copeptin was higher in participants with MI (8.9 vs. 6.4 pmol/L; P < .01), with no difference between RMI vs. UMI. Increased copeptin correlated with evening cortisol (r = .11; P < .01) and NTproBNP (r = .07; P = .04). Copeptin was associated with CVE and total mortality after adjusting for cortisol and NTproBNP separately, and remained significantly associated with total mortality in the multiple model. Conclusions Copeptin was higher in subjects with previous MI regardless whether previously recognized or not. Copeptin correlated weakly with cortisol and NTproBNP, and was independently associated with total mortality. This indicates that the prognostic implications of copeptin are not only mediated by heart failure or stress, supporting the assumption that copeptin is a marker of general vulnerability.