DEMO II: A Randomized, Parallel-group, Observer-blinded Clinical Trial of Aerobic Exercise Versus Stretching Exercise for Patients With Light to Moderate Depression

The yearly incidence of depression is estimated to be 3-5%, with a lifetime prevalence of 17% in western societies.Furthermore, Major depression is associated with increased mortality from somatic disorders such as cardiovascular disease, stroke, and endocrinological diseases. WHO and others stated that unipolar depressive disorders was the leading cause of disease burden in the western world in 2002 accounting for 9,0% of all disability adjusted life years. Projecting the current development in disease patterns unipolar depression is expected to be the second highest cause of global disease burden in 2030. With remission rates of approximately 50% using medical antidepressant therapy, the search for alternative or augmentation therapy is a key issue in depressive research.

In 2001 a meta-analysis concluded that the effect of exercise training could not be determined due to lack of good quality research. The authors found that the majority of trials were without blinded outcome assessment, lacked intention-to-treat analysis, and most had short follow-up. In this context, one of the authors in 2004 launched a pragmatic randomized controlled trial investigating the effect of exercise training on depressive symptoms - the DEMO trial(10). The 165 included participants were randomized to either aerobic, non-aerobic, or relaxation training in a four-month training programme. The primary outcome was the Hamilton depression rating scale (HAM-D17) and data collection ended June 1st 2007. After four month intervention there were no overall effect of exercise on depressive symptoms; the estimate at four month between non-aerobic and relaxation training was -0.9 (95%CI: -3.2 to 1.4) and between aerobic and relaxation training was 0.99 (95%CI: -1.3 to 3.3). These results do not support any biological effect of exercise on major depression. However, the majority (115/165) of patients had received antidepressant medication for more than six weeks at baseline thus the lack of a positive result could partly be explained by a treatment resistant group. A subgroup analysis of patients not medicated at baseline estimated the effect between non-aerobic and relaxation training to -1.7 (95% CI: -6.0 to 2.5) and between aerobic and relaxation training to 0.3 (95% CI: -3.9 to 4.6). The number of patients in this important sub-group, means that any conclusion on non-medicated patients is underpowered. In addition the literature suggest that the possible effect of exercise on depression is frequency and intensity related. In the DEMO trial the participants only trained twice pr. week.

This result from the DEMO trial is somewhat in contrast with the abundant publications of animal research showing that exercise stimulates hippocampal neurogenesis and in theory improves depressive symptoms and cognitive skills. The hippocampus is involved in cognitive functions such as memory and learning processes as well as regulation of the hypothalamus. In depression and neurodegenerative diseases such as Alzheimer and Parkinson Disease, which share features like impaired memory and learning abilities, the hippocampus is characterized by atrophia. In addition to a decreased hippocampal volume, brain imaging studies have observed reduction in volume of prefrontal cortex of depressed patients, though postmortem studies have less evident results.

Prospective studies on healthy adults suggest that exercise has a beneficial effect on cognitive functions or reduces an age related decline in these. Research on rodents suggest that an increase in memory and learning in relation to exercise is mediated through an up-regulation of brain derived neurotrophic factor, which has the ability to stimulate synaptic-plasticity.

Interestingly the exercise induced cell proliferation is inhibited by peripheral blockade of either insulin growth factor 1 (IGF-1) or vascular endothelial growth factor (VEGF), which could indicate that the exercise induced neurogenesis is dependent of IGF-I/VEGF. The role of IGF-I in neurogenesis is supported by studies showing an increase in new hippocampal cells in sedentary animals given IGF-I infusion. Fibroblast growth factor 2 (FGF-2) also increases in the rat hippocampus in response to exercise, however FGF-2 injections into the adult rat hippocampus do not increase neurogenesis. The importance of IGF-I in neurogenesis is emphasized by the fact that homozygous IGF1 -/- mice at the age of to months had reduced brain weight affecting all major brain areas, especially the volume of the dentate gyrus(33) in contrary to BDNF null mutant mice that did not show major hippocampal deficits.