Determinations of serum concentrations of total insulin-like growth factor I (tIGF-I) are important in the diagnosis, monitoring of treatment and safety evaluation of patients with growth disorders and/or metabolic disease. It is well established that tIGF-I status varies over time. Changes in tIGF-I levels in relation to an acute bout of exercise or repeated bouts, known as training, are likely to contribute to this variation. Serum tIGF-I has also been found to be of predictive value in growth prediction models employed before the start of growth hormone (GH) treatment. Furthermore, IGF-I generation tests have been suggested to be of value in the assessment of the growth response to GH administration in patients suspected of GH deficiency with or without some degree of GH insensitivity. This is discussed elsewhere in this issue. Recent progress in our understanding of growth hormone-dependent and -independent expression of the IGF1 gene in skeletal muscle and the role of sufficient energy intake during training for muscle and liver generation of IGF-I raises important questions regarding their relative contribution to the circulating pool of IGF-I. The present review is focused on circulating levels of tIGF-I in relation to a single bout of exercise or to a period of training. In addition, the expression of IGF-I locally in muscle in response to these stimuli will be discussed.