Hemorheological effects of exercise are a triphasic phenomenon including: (a) short-term effects (hyperviscosity mostly due to fluid shifts and alterations of erythrocyte rigidity and aggregability); (b) middle-term effects (i.e., the reversal of acute effects due to plasma volume expansion (autohemodilution) that lowers both plasma viscosity and hematocrit; (c) long-term effects that further improve blood fluidity, parallel with the classical training-induced hormonal and metabolic alterations. Red cell rheology during these 3 stages is affected by white cells and oxidant stress. On the other hand, most metabolic and hormonal alterations play a role in exercise-induced hemorheological changes: among them, blood lactate appears to have opposite effects according to the training status, since it generally impairs erythrocyte fluidity while it improves it in some subgroups of highly trained athletes, a difference that could be related to membrane monocarboxylate transporter status. Body composition (mostly hydration status and the amount of fat mass) as well as its major hormonal regulating system (the growth-hormone-IGF-I axis) are both markedly modified by training and these modifications are correlated with hemorheology. Nutritional disturbances affecting caloric and proteic intake, lipids, iron, zinc, etc. also modulate the hemorheologic effects of exercise. The overtraining syndrome represents a situation of unbalance between body's possibilities, nutrition, and work load, and is associated with metabolic, hormonal, immunologic and hemorheologic disturbances. The clinical relevance of these data is underlined by studies showing that exercise training in patients suffering from metabolic and/or cardiovascular disorders (such as the insulin resistance syndrome) results in a parallel improvement of metabolism, risk factors, blood rheology and fitness. Hemorheological measurements require to be studied, at least as sensitive markers of training, and possibly as "true" risk factors highly sensitive to exercise intensification.