Endurance exercise training results in structural expansion of the vascular beds of both myocardium and skeletal muscle. In the heart, this involves increases in diameter of the large coronary arteries, and growth of new vessels at the level of resistances arteries and arterioles. Capillary growth by proliferation and sprouting has been demonstrated in trained pig hearts (White et al. 1998) but overall, evidence for exercise-induced growth of myocardial capillaries is inconsistent, occurring when it does in young rather than adult animals. This may be due to an already high capillary density, to rapid transformation of capillaries into arterioles during the growth process or to the fact that concurrent cardiac hypertrophy may mask capillary neoformation. Studies in rats, dogs and pigs have used such varying training modes (running, swimming) and periods (4-18 weeks) that it is difficult to deduce a minimum exercise intensity and duration requirement for coronary vascular adaptation. Whilst development of collateral circulation in the healthy heart as a result of exercise training is not proven, exercise-induced structural adaptations to coronary vasculature as a whole allow for improved flow and more homogeneous capillary perfusion.

In contrast to the heart, studies have repeatedly shown capillary growth in trained skeletal muscles of animals and humans, although even now it is not clear whether this occurs by proliferation and sprouting or other means such as vessel elongation or splitting. Growth of resistance arteries and arterioles within trained muscle, on the other hand, is less well established, as are the reasons for differential growth of small versus larger vessels. Both cross-sectional and longitudinal studies in humans have confirmed that endurance exercise training induces remodelling of large conduit arteries, such as the aorta and those supplying trained limbs, to have larger diameters and greater distensibility (Huonker et al. 1996). This appears to allow for augmented blood flow to exercising muscles whilst maintaining arterial wall shear stress. These structural modifications are likely to involve vascular smooth muscle and elastic tissues, but this remains to be established.

Signals generated during each bout of exercise that could initiate vascular growth and remodelling may be metabolic, relating to production of e.g. adenosine, or tissue ischaemia/hypoxia, haemodynamic effects of increased blood flow and/or pressure, mechanical stresses in cardiac and skeletal muscle arising from repetitive contraction and stretch, and production of growth factors either directly or indirectly.

In the healthy heart, increased blood flow and contraction force have been implicated, with resting bradycardia, as a consequence of training, a prime factor facilitating both of these and leading to capillary growth via vascular endothelial growth factor, VEGF (Brown & Hudlicka, 1999; Zheng et al. 1999). Increased blood flow and myocyte stretch are also important in capillary angiogenesis in skeletal muscle in conjunction with VEGF. The enlargement of coronary and peripheral conduit arteries during training is most likely related to flow-mediated increases in shear stress and production of endothelial products such as nitric oxide and prostglandins. Detailed investigations of time-dependent changes in vascular reactivity during training (e.g. Bowles et al. 2000) have established that endothelial-dependent dilatation is enhanced early on in advance of structural modifications but that in the fully adapted trained state, metabolic, myogenic and endothelial controls of vascular tone are comparatively little changed. The challenge remains in the application of knowledge of these mechanisms of vascular remodelling by exercise as a preventative strategy in the promotion of health, or to target with site-specific precision pathologies that affect the circulation.

The support of the British Heart Foundation and The Wellcome Trust is gratefully acknowledged.