BACKGROUND: Diastolic dysfunction (DD) is associated with adverse cardiovascular outcomes. We studied the impact of an exercise-based lifestyle intervention on the evolution of DD in patients with type 2 diabetes mellitus (T2DM) and prospectively investigated the clinical correlates of DD progression. METHODS AND RESULTS: A total of 223 outpatients with T2DM were randomized to supervised exercise-based lifestyle intervention (initial gym-based program and lifestyle and diet advice followed by telephone-guided supervision) or usual care. Patients underwent echocardiographic assessment of diastolic function and metabolic and clinical evaluation at baseline and 3 years. Changes in prevalence and evolution of DD were assessed and correlations sought with clinical and metabolic variables. DD was present in 50% of patients at baseline and 54% at 3 years, with no difference between the usual care and intervention groups (60% versus 48%, P=0.10). Abnormal DD at the final visit was independently associated with older age and a decrease in peak oxygen consumption over time (P<0.05). There was no impact on glycemic control or exercise capacity. In a subanalysis restricted to patients who finished the full 3-year follow-up, control subjects were independently associated with DD at 3 years (β=0.90; odds ratio, 2.46; P=0.034), with the only other independent correlate being older age (β=0.05; odds ratio, 1.06; P=0.019). CONCLUSIONS: Despite being efficacious in the subgroup who completed 3 years of exercise-based lifestyle intervention, randomization to this program was not effective in reducing progression of subclinical DD in patients with T2DM, which may reflect the recognized difficulty of adherence to prolonged exercise intervention. CLINICAL TRIAL REGISTRATION: URL: http://www.anzctr.org.au. Unique identifier: ACTRN12607000060448.