OBJECTIVE Mean platelet volume (MPV) is a universally available parameter with routine blood counts. It has been linked to many cardiovascular risk factors. MPV is a marker of platelet size and activity and has been linked to poor prognosis following STEMI. There has been an increasing number of reports linking diabetes mellitus (DM) to platelet dysfunction. The aim of the study was to examine the association between admission MPV and clinical outcomes in patients with DM and STEMI undergoing primary percutaneous coronary intervention (PCI). The secondary objective of the study was to evaluate whether this index can be used to determine the long-term prognosis. METHODS A total of 1,557 patients with STEMI undergoing primary PCI were enrolled and divided into two groups depending on their diabetes mellitus status: Group 1 - patients with diabetes mellitus (N=539) and Group 2 - patients without diabetes mellitus (N=1018). RESULTS MPV and peak CK-MB concentration were higher in diabetic patients as compared to non-diabetic patients. In diabetic patients, MPV was positively correlated with admission Killip class and negatively correlated with time to death during follow-up, initial TIMI flow, final TIMI flow, and erythrocyte count. In non-diabetic patients, MPV was positively correlated with the number of diseased coronary arteries, admission Killip class, and negatively correlated with time to death during follow-up and initial TIMI flow. ROC analysis revealed high diagnostic value of MPV in predicting in-hospital and one-year mortality. MPV cut-off level was lower for diabetic patients compared to non-diabetic patients. CONCLUSIONS Diabetic patients had higher MPV than non-diabetic patients. Both in diabetic and non-diabetic patients MPV proved to have good prognostic value for in-hospital and late mortality. MPV cut-off value for predicting mortality was lower in diabetic patients. Mortality rate was the highest in the fourth quartiles of MPV in both study groups.