Growth hormone (GH) secretion in the elderly is generally diminished although there are marked individual differences ranging from normal GH secretion and normal levels of insulin-like growth factor (IGF)-I through low GH and subnormal IGF-I. It is assumed that the reduced central cholinergic activity leading to unrestrained somatostatin release leads to impaired GH secretion. The somatopause, if it occurs at all, is, in contrast to the menopause, a subtly developing physiological event. The menopause often causes severe symptoms that justify hormone replacement therapy, but the somatopause is a physiological event at the end of the lifespan with no acute symptoms that can be attributed to GH deficiency with certainty. Whether the non-specific symptoms of old age, i.e. truncal obesity, muscle atrophy, decreasing energy, and mental disorders, can be--even partially--blamed on decreased GH secretion is unclear. Thus, GH therapy in elderly patients, in the absence of pituitary disease cannot be recommended. In addition, the following has to be considered: 1) GH has to be given by subcutaneous injection, which may be technically difficult in elderly patients. 2) It is difficult to find the right individual dosage of GH since elderly patients may show increased sensitivity to GH therapy (compared with children) or may be GH-resistant. 3) Manifestation of diabetes mellitus may be enhanced in elderly patients. 4) The elevation of IGF-I levels may enhance the progression of malignant disease; it has been shown that the concentration of IGF-I in the circulation correlates to the frequency of prostatic cancer. Furthermore, acromegalic patients have a higher frequency of colonic polyps and gastrointestinal malignancies. 5) Even if problems such as dosage, mode of application and the questions of safety are resolved, the present costs of GH therapy will not allow to advocate GH treatment of all elderly patients with low levels of IGF-I. However, since some patients seem to benefit from GH therapy in senescence, further studies are needed. There may be a subset of elderly patients in whom GH treatment is useful. However, unless these patients are included in a study protocol, GH treatment should not be given to elderly patients in the absence of pituitary disease.