It is widely recommended that consideration should be given to the therapeutic use of growth hormone (GH) in adults with GH deficiency, whether the condition is of childhood or adult onset. One reason for this recommendation is the possibility that such treatment may reduce the excess cardiovascular risk which is associated with hypopituitarism. This excess risk has been well documented, with mortality ratios of 1.7 to 2.2 being quoted in different studies, and may be a result of the insulin resistance which occurs in hypopituitarism. However, it has also been suggested that this insulin resistance may itself be the result of GH deficiency, especially as GH deficiency is accompanied by suggestive morphological features such as central adiposity. There is, however, no direct evidence that the increase in cardiovascular risk in hypopituitarism is the result of GH deficiency, and the only prospective study designed to examine the relationship failed to find a statistically significant correlation between the two. Since GH administration may also have an independent adverse effect on insulin sensitivity and could thus cause a theoretical worsening of cardiovascular risk, it is important to review the observed effects of GH administration on carbohydrate metabolism in practice. Interpretation of the literature is made difficult by many confounding factors, including differences in study duration, biochemical tools adopted, the use of selected populations and the dose-dependent effect of GH on synthesis of insulin- like growth factor-1. One of the most sensitive markers of a deterioration in insulin sensitivity is the serum insulin level. A rise in serum insulin (fasting, or post-glucose load) was reported in all studies in which it was measured. The majority of studies have also reported a rise in fasting blood glucose. A smaller proportion of reports noted an associated increase in postprandial glucose and in glycosylated haemoglobin (HbA(1c)) while a few reported new cases of either impaired glucose tolerance or frank diabetes mellitus. In general, however, the observed deterioration in insulin sensitivity was small and increases which occurred in blood glucose were small. Nevertheless, these data indicate that rather than lead to an improvement in insulin resistance in hypopituitarism, GH treatment may actually make it worse. As it is also known that even minor reductions in insulin sensitivity may be associated with a clinically significant increase in cardiovascular risk, further large-scale controlled trials are required before the efficacy and safety of GH treatment of adults can be established.