Abstract Background Diabetes mellitus (DM) is an established adverse prognostic factor in patients sustaining myocardial infarction (MI). However, its impact on long-term survival remains less clear. The aim of this observational study was to quantify lifetime mortality and years of life lost after MI in patients with and without DM. Methods In 1995, 2153 individuals with MI were recruited from 20 adjacent hospitals within Yorkshire, UK. Median survival, all-cause mortality at 20 years and lost years of life when compared to actuarial predictions were compared in patients with and without DM. Landmark analyses were conducted to define the ongoing impact of DM beyond specified time points. Results 13% (279/2153) had known DM. They experienced higher mortality at 30 days (33.1% vs 24.6%; p < 0.0001) and at 20 years (84.9% vs 75.7%; p < 0.0001). Overall, there was a 48% increased risk of death (p < 0.0001), which persisted after adjustment for potential confounders. There was no interaction between DM and prior MI in predicting mortality (p = 0.67). Median survival decreased by 3.3 years (p < 0.0001). The adverse impact of DM persisted in sequential landmark analyses at 1, 5 and 10 years. Presence of DM conferred 2 extra years of life lost when compared with actuarial predictions (8 vs 6 years; p < 0.0001). Conclusions DM remains an independent adverse prognostic factor in the long-term after MI. Persistently diverging survival curves support enduring efforts to reduce mortality late after MI.