The most important finding of the present study is that in this group of type 2 diabetic patients, a twelve-week training program, which increased aerobic capacity and insulin sensitivity resulted in an improvement of systolic function, while cardiac lipid content remained unchanged. These data suggest that a decrease in cardiac lipid content in type 2 diabetic patients is not a prerequisite for improvements in cardiac function. A limitation of the current study is the relatively small number of patients included.
Exercise training is well established as a powerful tool to reduce the risk for cardiovascular disease in type 2 diabetic patients [19–23]; however it is less clear whether training can actually improve cardiac function in this population. Data on the effect of training on systolic function in type 2 diabetic subjects is limited, with only one study reporting a positive effect  while others do not [25, 26]. In diabetic animal models, however, the improvement in ejection fraction and fractional shortening with endurance training is well documented [27, 28]. We previously reported improved ejection fraction in obese but otherwise healthy human subjects in response to exercise training . Here, we show that a modest but significant improvement in ejection fraction can also occur in type 2 diabetic patients, in response to a relatively short period of training.
Rodent studies strongly support the concept that accumulation of triglycerides in cardiomyocytes causes cardiac lipotoxicity, impairing cardiac function. In addition, several human studies report negative correlations of triglyceride content with cardiac function as well [9, 11]. Some , but not all  human studies report lowered cardiac lipid content upon thiazolidinediones. Recently, we reported that in healthy overweight-to-obese subjects, the training-induced improvement in systolic function is accompanied by a decrease in cardiac lipid, in line with the cardiac lipid hypothesis . Intriguingly, though, exercise-induced improvement in cardiac contractile function was not accompanied by a decrease in lipid content in the patients with type 2 diabetes, rather cardiac lipid content tends to increase in this population, although not significantly. Why type 2 diabetic patients are unresponsive towards training-induced decrease in cardiac lipid content cannot be deduced from the present study. It is important to note that the cardiac lipid content in the present group of diabetic subjects is similar as reported in healthy overweight subjects of similar age and BMI . However, when compared to healthy lean young men, cardiac lipid content is clearly elevated by ~300% in the present group of patients . In other words, cardiac lipid content is elevated as expected based on BMI and age, however, it is not further elevated due to type 2 diabetes. This may be due to the relatively healthy patients investigated in the current study. Although diabetic, patients did not have any severe diabetes-related complications. Additionally, plasma concentrations of FFA and TG were also normal. However, although not elevated more severely than expected based on BMI and age at baseline, the cardiac lipid content was unresponsive to an exercise intervention in the diabetic subjects and interestingly, this differs from the response in healthy overweight subjects. It is yet unclear what is at the basis of this differential response, but it is well known that cardiac metabolism is altered in diabetes. For example, cardiac insulin resistance, decreased metabolic flexibility, mitochondrial dysfunction and excessive cardiac fat uptake and oxidation all have been associated with the diabetic state. Furthermore, a study in a diabetic animal model demonstrated decreased responsiveness to beta adrenergic stimulation in diabetic animals upon exercise training while beta adrenergic stimulation increased in control animals . Future studies are needed to investigate whether these factors are contributing to the differential response in diabetic patients to exercise training resulting in the unchanged cardiac lipid content.
Furthermore, the present findings demonstrate that the training-induced improvement of systolic function can occur independently of changes in cardiac lipid content. Generally, it is believed that the beneficial effect of training on systolic- and diastolic function is based on alterations of the Ca2+ regulatory systems involved in the excitation-contraction coupling and relaxation processes . According to the cardiac lipotoxicity theory, there are potential mechanisms by which lipid accumulation may interact directly with excitation-contraction coupling. We here demonstrate that this is not pivotal: the improved function with exercise training can occur independently of changes in lipid content. Apparently, the improvement of cardiac function does not necessarily require decrease in cardiac lipids in type 2 diabetic subjects. Similarly, it has been shown that pioglitazone improves diastolic function in type 2 diabetic subjects without changes in cardiac lipid content .