OBJECTIVE: To identify the effects of a 1-year exercise intervention on myocardial dysfunction in patients with type 2 diabetes mellitus (T2DM). DESIGN: Randomised controlled trial, the Diabetes Lifestyle Intervention Study. SETTING: University hospital. PATIENTS: 223 T2DM patients without occult coronary artery disease, aged 18-75 were randomised to an exercise training group (n = 111) or a usual care group (n = 112). Complete follow-up data were available in 176 (88 exercise, 88 usual care). INTERVENTIONS: Exercise training consisted of gym, followed by telephone-monitored home-based exercise training. MAIN OUTCOME MEASURES: Tissue Doppler-derived myocardial velocities, strain-rate and strain, body composition, glycated haemoglobin (HbA(1c)), maximum oxygen consumption (VO(2max)) and physical activity. RESULTS: Overall changes in myocardial function were not different between groups despite improvements in waist circumference, fat mass, blood glucose, HbA(1c), insulin sensitivity, VO(2max) and 6-minute walk distance in the intervention group (p<0.05). The latter also spent significantly more time in vigorous activity (p<0.05). A post-hoc analysis revealed that intervention patients who spent more time in both moderate and vigorous activity showed a significant improvement in myocardial tissue velocity (p<0.01), HbA(1c) (p = 0.03) and VO(2max) (p = 0.03) compared to controls. Myocardial strain rate (p = 0.03) and HbA(1c) improved in intervention patients with the greatest increase in moderate activity (p = 0.03). CONCLUSIONS: In patients with T2DM, current exercise recommendations led to an improvement in metabolic function, but failed to improve myocardial function in the overall group. Patients with greater increases in both moderate and vigorous activity showed improvements in myocardial function, glycaemic control and cardiorespiratory fitness. TRIAL REGISTRATION NUMBER: ACTRN12607000060448.