Prior studies examined association between short-term mortality and certain changes in the admission ECG in acute myocardial infarction (AMI). Nevertheless, little is known about possible differences between patients with diabetes and without diabetes in this regard. So the aim of the study was to investigate the association between 28-day case fatality according to certain ECG changes comparing AMI cases with and without diabetes from the general population. From 2000 until 2017 a total of 9756 AMI cases was prospectively recorded in the study Area of Augsburg, Germany. Each case was assigned to one of the following groups according to admission ECG: 'ST-elevation', 'ST-depression', 'only T-negativity', 'predominantly bundle branch block', 'unspecific changes' and 'normal ECG' (the last two were put together for regression analyses). Multivariable adjusted logistic regression models were calculated to compare 28-day case-fatality between the ECG groups for the total sample and separately for diabetes and non-diabetes cases. For the non-diabetes group, the parsimonious logistic regression model revealed significantly better 28-day-outcome for the 'normal ECG / unspecific changes' group (OR: 0.47 [0.29-0.76]) compared to the reference group (STEMI). Contrary, in AMI cases with diabetes the category 'normal ECG / unspecific changes' was not significantly associated with lower short-term mortality (OR: 0.87 [0.49-1.54]). Neither of the other ECG groups was significantly associated with 28-day-mortality in the parsimonious logistic regression models. Consequently, the absence of AMI-typical changes in the admission ECG predicts favorable short-term mortality only in non-diabetic cases, but not so in patients with diabetes.